Fishery Bulletin f?*ari*« H t Sr ATES O* + 1983 r Wooos Hote, Jv»ass. Vol. 81, No. 1 January 1983 SMITH, TIM D. Changes in size of three dolphin (Stenella spp.) populations in the eastern tropical Pacific 1 LANGTON, RICHARD W. Food habits of yellowtail flounder, Limanda ferru- gi>iia (Storer), from off the northeastern United States 15 DUNN, JEAN R. Development and distribution of the young of northern smooth- tongue, Leuroglossus sckmidti (Bathylagidae), in the northeast Pacific, with com- ments on the systematics of the genus Leuroglossus Gilbert 23 KATZ, S. J., C. B. GRIMES, and K. W. ABLE. Delineation of tilefish, Lopholatilus chamaeleonticeps, stocks along the United States east coast and in the Gulf of Mexico 41 RICHARDS. R. ANNE, J. STANLEY COBB, and MICHAEL J. FOGARTY. Effects of behavioral interactions on the catchability of American lobster, Ho- marus americanus, and two species of Cancer crabs 51 PARSONS, GLENN R. The reproductive biology of the Atlantic sharpnose shark, Rhizoprionodon terraenovae (Richardson) 61 APPELDOORN, RICHARD S. Variation in the growth rate of Myaarenaria and its relationship to the environment as analyzed through principal components analysis and the a> parameter of the von Bertalanffy equation 75 SHAKLEE, JAMES B., RICHARD W. BRILL, and ROBIN ACERRA. Bio- chemical genetics of Pacific blue marlin, Makaira nigricans, from Hawaiian waters 85 BARTOO, NORMAN W., and KEITH R. PARKER. Stochastic age-frequency estimation using the von Bertalanffy growth equation 91 JOHNSON, ALLYN G., WILLIAM A. FABLE, JR., MARK L. WILLIAMS, and LYMAN E. BARGER. Age, growth, and mortality of king mackerel, Scom- beromorus ca valla, from the southeastern United States 97 HAN AN, DOYLE A. Review and analysis of the bluefin tuna, Thunnus thynnus, fishery in the eastern North Pacific Ocean 107 SWARTZMAN, GORDON L., and ROBERT T. HAAR. Interactions between fur seal populations and fisheries in the Bering Sea 121 DURBIN, ANN GAIL, EDWARD G. DURBIN, THOMAS J. SMAYDA, and PETER G. VERITY. Age, size, growth, and chemical composition of Atlantic menhaden, Brevoortia tyrannus, from Narragansett Bay, Rhode Island 133 (Continued on back cover) V. Seattle, Washington U.S. DEPARTMENT OF COMMERCE Malcolm Baldrige, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION John V. Byrne, Administrator NATIONAL MARINE FISHERIES SERVICE William G. Gordon, 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 bulle- tin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodi- cal, issued quarterly. In this form, it is available by subscription from the Superintendentof 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. EDITOR Dr. Carl J. Sindermann Scientific Editor, Fishery Bulletin Northeast Fisheries Center Sandy Hook Laboratory National Marine Fisheries Service, NOAA Highlands, NJ 07732 Editorial Committee Dr. Bruce B. Collette Dr. Donald C. Malins National Marine Fisheries Service National Marine Fisheries Service Dr. Edward D. Houde Dr. Jerome J. Pella Chesapeake Biological Laboratory National Marine Fisheries Service Dr. Merton C. Ingham Dr. Jay C. Quast National Marine Fisheries Service National Marine Fisheries Service Dr. Reuben Lasker Dr. Sally L. Richardson National Marine Fisheries Service Gulf Coast Research Laboratory Mary S. Fukuyama, Managing Editor The Fishery Bulletin (ISSN 0090-0656) is published quarterly by Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, Bin C15700, Seattle, WA 98115 Second class postage paid to Finance Department, USPS, Washington, DC 20260. 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 i| 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 Management and Budget through 1 April 1985. , Fishery Bulletin CONTENTS Vol. 81, No. 1 January 1983 SMITH. TIM D. Changes in size of three dolphin (Stenella spp.) populations in the eastern tropical Pacific 1 LANGTON, RICHARD W. Food habits of yellowtail flounder, Limanda ferru- ginia (Storer), from off the northeastern United States 15 DUNN, JEAN R. Development and distribution of the young of northern smooth- tongue, Leuroglossus schmidt i (Bathylagidae), in the northeast Pacific, with com- ments on the systematics of the genus Leuroglossus Gilbert 23 KATZ, S. J., C. B. GRIMES, and K. W. ABLE. Delineation of tilefish, Lopholatilus chamaeleonticeps, stocks along the United States east coast and in the Gulf of Mexico 41 RICHARDS, R. ANNE, J. STANLEY COBB, and MICHAEL J. FOGARTY. Effects of behavioral interactions on the catchability of American lobster, Ho- marus americanus, and two species of Cancer crabs 51 PARSONS, GLENN R. The reproductive biology of the Atlantic sharpnose shark, Rhizoprionodon terraenovae (Richardson) 61 APPELDOORN, RICHARD S. Variation in the growth rate of Mya arenaria and its relationship to the environment as analyzed through principal components analysis and the w parameter of the von Bertalanffy equation 75 SHAKLEE, JAMES B., RICHARD W. BRILL, and ROBIN ACERRA. Bio- chemical genetics of Pacific blue marlin, Makaira nigricans, from Hawaiian waters 85 BARTOO, NORMAN W., and KEITH R. PARKER. Stochastic age-frequency estimation using the von Bertalanffy growth equation 91 JOHNSON, ALLYN G., WILLIAM A. FABLE, JR., MARK L. WILLIAMS, and LYMAN E. BARGER. Age, growth, and mortality of king mackerel, Scom- beromorus cavalla, from the southeastern United States 97 HAN AN, DOYLE A. Review and analysis of the bluefin tuna, Thunnus thynnus, fishery in the eastern North Pacific Ocean 107 SWARTZMAN, GORDON L., and ROBERT T. HAAR. Interactions between fur seal populations and fisheries in the Bering Sea 121 DURBIN, ANN GAIL, EDWARD G. DURBIN, THOMAS J. SMAYDA, and PETER G. VERITY. Age, size, growth, and chemical composition of Atlantic menhaden, Brevoortia tyrannus, from Narragansett Bay, Rhode Island 133 (Continued on next page) Seattle, Washington 1983 For sale by the Superintendent of Documents, U.S. Government Printing Office. Washington. DC 20402— Subscription price per year: #21.00 domestic and $20.25 foreign. Cost per single issue: $6.50 domestic and $8.15 foreign. ( 'ontents— continued Notes VREELAND, ROBERT R., and ROY J. WAHLE. Homing and fisheries contri- bution of marked coho salmon, Oncorhynchus kisutch, released at two Columbia River locations 143 COLTON, DOUGLAS E., and WILLIAM S. ALE VIZON. Movement patterns of bonefish, Albula vulpes, in Bahamian waters 148 KLEPPEL, G. S., and E. MANZANILLA. Analyses of feeding in two marine copepods from Santa Monica Bay, California 154 SHENKER, JONATHAN M. Distribution, size relationships, and food habits of juvenile king-of-the-salmon, Trachipterus altivelis, caught off the Oregon coast 161 BOND, CARL E., TING T. KAN, and KATHERINE W. MYERS. Notes on the marine life of the river lamprey, Lampetra ayresi, in Yaquina Bay, Oregon, and the Columbia River estuary 165 MENZ, FREDRIC C, and DONALD P. WILTON. An economic evaluation of the St. Lawrence River-eastern Lake Ontario bass fishery 168 The National Marine Fisheries Service (NMFS) does not approve, recommend or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication fur- nished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary prod- uct 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 pur- chased because of this NMFS publication. ^ m- AWARD AWARD Best NMFS publications for 1981 The Publications Advisory Committee of the National Marine Fisheries Service has announced the best publications authored by the NMFS scientists and published in the Fishery Bulletin and the Marine Fisheries Review for 1981. Only effective and interpretive articles which significantly contribute to the understanding and knowledge of NMFS mission-related studies are eligible, and the following papers have met this requirement. For the Fishery Bulletin, the paper "The spawning energetics of female northern anchovy, Engraulis mordax" by J. Roe Hunter and Roderick Leong was awarded as the best publication, appearing in the Fishery Bulletin 79(2): 215-230. Hunter and Leong are both fishery biologists with the Southwest Fisheries Center La Jolla Laboratory, NMFS, NOAA, La Jolla, Calif. For the Marine Fisheries Review, the paper "Low temperature preserva- tion of seafoods: A review" by Louis J. Ronsivalli and Daniel W. Baker II was awarded the best publication, appearing in the Marine Fisheries Review 43(4):1-15. Ronsivalli, now retired, was the former Director of the Northeast Fisheries Center Gloucester Laboratory, NMFS, NOAA, Gloucester, Mass.; Baker is a mechanical engineering technician with the same laboratory. N fr CHANGES IN SIZE OF THREE DOLPHIN (STENELLA SPP.) POPULATIONS IN THE EASTERN TROPICAL PACIFIC Tim D. Smith 1 ABSTRACT Dolphins from three populations, one of Stenella attenuata and two of S. longirostris, have been killed incidentally in the yellowfin tuna purse seine fishery in the eastern tropical Pacific, two populations since about 1959 and the other since about 1969. Size changes in these populations are estimated from numbers killed each year, population size estimates in 1979, and net recruitment rates. Ranges of values for some parameters are considered, accounting for some uncertainties. Assuming central values of the ranges of maximum net recruitment rate (3%) and the population level giving maximum net productivity (65%), one S. longirostris population, the eastern spinner dolphin, is near 20% of pre-exploitation levels; the S. attenuata population, the northern offshore spotted dolphin, is between 35 and 50%; and the second S. longirostris population, the whitebelly spinner dolphin, is between 58 and 72% of pre-exploitation levels. Purse seine fishing for tuna in the eastern tropi- cal Pacific often involves dolphins found in asso- ciation with yellowfin tuna. Tuna fishermen pursue and capture the dolphin-yellowfin tuna complex, releasing the dolphins from the net while retaining the tuna (Green et al. 1971). Mor- tality of dolphins occurs incidental to this fishing process. Purse seine fishermen were using dolphin schools to catch tuna by 1959; there is anecdotal information suggesting limited use as early as the 1940's (anonymous reviewer). Starting in the mid-1960's the Bureau of Commercial Fisheries, predecessor of the National Marine Fisheries Service (NMFS), conducted limited research to document the situation and to collect data on numbers and kinds of dolphins killed. This re- search expanded in the 1970's, especially after passage of the Marine Mammal Protection Act (MMPA) of the United States in 1972, and con- tinues. Substantial research efforts were mount- ed to assess the status of the dolphin stocks and to develop procedures for reducing incidental mor- tality and injury. Two assessments of the condition of dolphin populations involved in the yellowfin tuna purse seine fishery have been completed in recent years. 2,3 1 describe the results of the latest assess- ment of the three populations most affected by the fishery; calculation of population sizes from 1959 through 1978 is emphasized, based on esti- mates of the population size in 1979, on annual numbers killed from 1959 through 1978, and on net recruitment rates. These results, based on data available through the end of 1980, do not necessarily represent NMFS policy, which in- volves additional considerations. A third assess- ment of these populations is scheduled for 1984 and will include information since 1980. POPULATION MODEL Methods developed in 1976 (footnote 2) for esti- mating pre-exploitation abundance are based on a simple recursive relationship Nm = N t - K t + R t (N t -%K t ), where t denotes the year; N, the abundance; K, the number of animals killed; and R, the net re- cruitment rate. This model assumes that the population size in the next year is simply the present population size, minus the present inci- dental kill, plus the net number of individuals re- cruited to the population during the year. This latter quantity is taken to be the net recruitment 'Southwest Fisheries Center, National Marine Fisheries Service, NOAA, P.O. Box 271, LaJolla, Calif.; present address: Northeast Fisheries Center, National Marine Fisheries Ser- vice, NOAA, Woods Hole, MA 02543. 2 Southwest Fisheries Center La Jolla Laboratory. 1976. Report of the workshop on stock assessment of porpoises in- volved in the eastern Pacific yellowfin tuna fishery. Natl. Mar. Fish. Serv., NOAA, Admin. Rep. LJ-76-29, 53 p. 3 Smith, T. D. (editor). 1979. Report of the workshop on status of porpoise stocks, La Jolla, Calif., 27-31 August 1979. Southwest Fish. Cent. La Jolla Lab., Natl. Mar. Fish. Serv., NOAA, Admin. Rep. LJ-70-41, 120 p. Manuscript accepted June 1982. FISHERY BULLETIN: VOL. 81, NO. 1. 1983. FISHERY BULLETIN: VOL. 81. NO. 1 rate (birth rate less natural death rate) multi- plied by the number of animals actually repro- ducing in a given year. The number of repro- ducing animals is approximated by assuming that one-half the animals killed in a year repro- duce before dying. Solving this relationship for N t , one obtains (i) Repeatedly applying this equation to estimate the population size for any number of years (s) prior to the year (c) for which an independent estimate of population size {N ) is available yields in general N s = N c -z 1 - . (2) n(i + r\ j=1 n/i + r) The 1979 workshop (footnote 3) extended this procedure by calculating the recruitment rate Ri, i years prior to the present, using the density- dependent relationship (Allen 1981) Ri — -R TO <1 (3) N p is the estimated population size at the begin- ning of the first year of exploitation, p years earlier; R,„, the maximum net recruitment rate; Z, the density-dependent exponent; and N,- and N p , estimated from Equation (2). Because N p in Equation (3) is not known until the series in Equation (2) has been calculated, an iterative procedure is required to solve the equations for historical population size. Equations (1) and (3) together form a special case of the generalized production model of Pella and Tomlinson (1969). In Equation (3) the net recruitment rate is maximum (R m ) when the population size ap- proaches zero, decreasing to zero as the popula- tion size approaches N p . Z determines the popu- lation size at which the rate of change of the population is maximum, the maximum net pro- ductivity level (MNPL). The values of Z corre- spond to the MNPL approximately as (Polacheck 1982) MNPL N P (1 + Zf z (4) If Z=l, then the MNPL is one-half the equilibri- um population size; if Z is >1, then MNPL is greater than one-half the equilibrium size. The fraction of the maximum reproductive rate, R m , realized at a given population size, increases as the value of Z increases. Statistical properties of the estimate of N p and the ratio N c /Np are examined in detail in Smith and Polacheck (1979), wherein methods are de- veloped for calculating the variances. Tests of sensitivity of the estimates of N p to the values JV C , K t , and R m show that the estimates are most sen- sitive to the value of present abundance and least sensitive to the net recruitment rate. Examina- tion via simulation of the shape of the sampling distribution shows that if N has a symmetrical sampling distribution, then so does the estimate Np. Several estimates of each parameter required by the model are available in working documents and technical memoranda prepared by the staff of the Southwest Fisheries Center. I rely on the most current estimates, primarily minor revi- sions of those used by the 1979 workshop, with reference to papers describing earlier estimates as needed to document methods. POPULATIONS Populations affected most by the yellowfin tuna purse seine fishery are of the genus Stenella, and are found in the area from just south of the Equator to an approximate lat. 20°N and west from the Mexican and Central American coasts to an approximate long. 150°W. Two populations of spotted dolphins, S. attenuata, and three popu- lations of spinner dolphins, S. longirostris, are found in this region. The two spotted dolphin populations are re- ferred to as "offshore" and "coastal" forms. The coastal spotted dolphin population occurs near- shore and around islands, while the offshore spotted dolphin ranges from nearshore to an ap- proximate long. 150°W. The two forms overlap in range near the coast. Perrin (1975) distinguished these two forms of S. attenuata morphologically. He noted that 1) the larger coastal form occurs seaward to 50 km while the offshore form occurs as nearshore as 20 km, and 2) the coastal form was involved in only 7 of 1,373 purse seine sets on dolphins observed be- tween 1971 and 1974. Additional data collected since then, including reexamination of speci- mens collected during sets in the years 1971-74, SMITH: SIZE CHANGES OF THREE DOLPHIN POPULATIONS indicate that through 1978, a total of 22 sets have been observed on coastal spotted dolphins, out of a total 9,672 observed overall (about 0.2%). The yellowfin tuna purse seine fishery was concentrated nearshore in the early 1960's, and many sets were made in the area probably occu- pied by both coastal and offshore spotted dol- phins. Direct observations in the 1960's distin- guishing these forms of the spotted dolphin are not available. Based on observations made in the 1970's where these forms were distinguished, however, it appears that the coastal form has never been significantly involved in this fishery. Since 1978, sighting data collected by scientific observers aboard tuna vessels have been edited, using consistent criteria of school size, body size, and coloration for distinguishing coastal and off- shore spotted dolphins. In 1979, for example, within about 50 km of the coast there were 46 sightings of coastal spotted schools, 160 sightings of offshore spotted schools, and 25 sightings of spotted dolphin schools which could not be dis- tinguished to form with the available data. These three school types were subsequently set on in 2, 73, and 6 instances, respectively. Even assuming that all schools not identified to form were coastal spotted dolphins, the proportion of sighted coast- al spotted dolphin schools subsequently set on is much smaller than that of the offshore form (0.11 vs. 0.46, P<0.001). This differential selection exists even though the catch of yellowfin tuna in sets on coastal spotted dolphins has been approxi- mately twice that on offshore spotted dolphins. If coastal spotted dolphins were a significant part of this fishery, one would expect their involve- ment in sets to be proportional to the rate at which they are encountered. In addition, 18 of the 22 sets on coastal spotted dolphins occurred in 1973, and, except for one set, these were made by two vessels in the Gulf of Nicoya, a small area off the Costa Rican coast. Based on this information, I have assumed that the coastal spotted dolphin has been involved only rarely in this fishery. Two spinner dolphin populations, referred to as the "eastern" and "whitebelly" forms, are in- volved in the yellowfin tuna purse seine fishery. A third form, termed the Costa Rican spinner, occurs near the coast from Mexico to Panama, but is not involved in the fishery. The eastern and whitebelly forms overlap broadly in range, with the whitebelly spinner dolphin generally occur- ring more seaward. The eastern form has been involved with this fishery since 1959, whereas the whitebelly spinner dolphin population ap- parently became increasingly involved as the fishery expanded seaward in the 1960's. The whitebelly spinner and the offshore spotted forms have Southern Hemisphere populations (Perrin et al. 1979). These populations have been involved only recently with the yellowfin tuna purse seine fishery, as it has expanded south- ward, and are only lightly exploited. Data on reproductive condition of these southern popula- tions are used as estimates of reproductive rates for unexploited or equilibrium populations. 1979 POPULATION SIZE ESTIMATES Holt and Powers (1982) gave estimates of abun- dance based on aerial and research-vessel sight- ing surveys and data from scientific observers aboard fishing vessels. Estimates of the size, N,, of the i th population in their survey area are based on the equation N, = P t S t D P, A, (5) where P, denotes the proportion of dolphin schools containing dolphin of the genera Stenella, Delphinus, and Lagenodelphis; S t , the mean size of these schools; D, the estimated density of all dolphin schools sighted; P,, the fraction of schools containing dolphins of the ith population; and A, the area inhabited. This equation is applied to 1) a nearshore stratum, surveyed using both an air- plane and research vessels, and 2) an offshore stratum, surveyed only by research vessels. The nearshore stratum extends seaward from the coastline about 800 km, and from lat. 22°N to 12°S. The offshore stratum extends from the outer edge of the nearshore stratum to the bound- ary of the dolphin range. Approximate areas of the maximum historical range of the three dolphin populations are used for the area inhabited, A in Equation (5). These are estimates of the area enclosed by a smooth curve which includes most locations where dol- phins of different species have been reported by both fishing vessels and research vessels, as de- scribed in Holt and Powers (1982). While occasional sightings of dolphin schools have been reported outside these areas, the areas are overestimated in that ". . . at any point in time it is likely that each of the various dolphin species FISHERY BULLETIN: VOL. 81, NO. 1 only occupies a portion of its historical range." 4 Overlap between coastal and offshore forms of spotted dolphin is not reflected in the population estimates given by Holt and Powers (1982). Due to the large differences in areas inhabited, how- ever, adjustments to account for the unknown degree of overlap would increase the offshore spotted dolphin population estimate by 3% at most, which is insignificant for the general re- sults being presented here. The density estimate for the nearshore area is obtained from line transect theory applied to aerial survey sighting data. This follows earlier applications (Smith 1981), but with several im- provements. For instance, the aircraft we used had superior downward visibility; right-angle distance from the aircraft trackline to the sighted dolphin schools was determined directly, either by electronic navigation equipment or visually for shorter distances, rather than being calcu- lated from visual estimates of range and bearing; and the originally used negative-exponential sighting model was replaced with the superior Fourier series model (Burnham et al. 1980). The density estimate for the offshore area, which could not be surveyed by air, is obtained by com- paring relative dolphin school sighting rates from research vessel surveys in nearshore and offshore areas with absolute density estimates from the nearshore area. The resulting density estimate of all dolphin schools of >15 animals in the nearshore area is about 3.6 schools/ 1,000 km 2 , while the density estimate in the offshore area is about one-half that value. The school size estimate is about 200 animals, based on visual estimates of the size of schools seen during the aerial survey. The accuracy of these visual estimates has been confirmed by counts of individual dolphins from aerial photo- graphs, and the accuracy of the counts from these photographs has been confirmed by counts of dolphins released from a purse seine (Allen et al. 1980). This estimated school size also includes an adjustment for the tendency of larger schools to be more readily visible at greater distances from the aircraft, and hence to be overrepresent- ed in the sample. Allen et al. (1980) also demonstrated that accu- rate school size estimates could be made from ships. Although not used by Holt and Powers (1982), the mean school size estimated from re- search vessel sighting data was about 180, not significantly different from the value derived from aerial data described above. In contrast, the mean school size estimated from tuna vessel sighting data collected by scientific technicians was about 580, significantly higher (P<0.001) than the other two values. This difference im- plies either nonrandomness of the sample of dol- phin schools encountered by the tuna vessels, or biases in the estimation techniques used by the technicians. P ( for each of the 22 populations involved in the yellowfin tuna purse seine fishery can be esti- mated from data collected aboard either tuna vessels or research vessels. Fishing vessels en- counter significantly more schools composed pri- marily of spotted and spinner dolphins than do research vessels. The reason for this difference is not known, but it is possible that fishing vessels encounter spotted and spinner dolpin schools more frequently than would be expected under random search because they are searching for tuna, which occur with these two schools more frequently than with other species of dolphins. Studies of the searching process of tuna fishing vessels are being conducted which should help resolve this question. Because the proportions P, are different for unknown reasons, Holt and Powers (1982) gave several sets of estimates of total abundance, depending on the estimates of Pi from different combinations of the research vessel and tuna vessel data. Two sets of estimates are considered here (Table 1), one using research vessel data alone and the other using combined tuna vessel and research vessel data. Aerial survey procedures used in the present population-size estimates are still being refined. For instance, a field study was completed in mid- Table 1.— Population size estimates (thousands) at the beginning of 1979 for three populations of dol- phins in the eastern tropical Pacific, using estimates of the species mix from research vessel data alone, and from combined tuna vessel data and research vessel data, with standard deviations in parentheses (Holt and Powers 1982). Population Research vessel data only Fishing and research vessel data 4 Hammond, P. S. (editor). 1981. Report of the Workshop on Tuna-Dolphin Interactions, Managua, Nicaragua, April 1981, p. 5. IATTC Spec. Rep. 4, Inter-Am. Trop. Tuna Comm., c/o Scripps Inst. Oceanogr., La Jolla, CA 92093. Offshore spotted 1,682.0 (471.8) 2,775.0 (761.4) Eastern spinner 292 7 (71.0) 2929 (64.4) Whitebelly spinner 216.0 (67.4) 3804 (134.9) SMITH: SIZE CHANGES OF THREE DOLPHIN POPULATIONS 1980 to determine the effect of sea state and sun position on the visibility of dolphin schools di- rectly on the trackline. Data from this experi- ment, which have not yet been completely ana- lyzed, will be of use in the design of future surveys and in the evaluation of earlier sur- veys. INCIDENTAL KILL ESTIMATES Incidental kill (K, ) of dolphins in year t is esti- mated by multiplying the mean kill of dolphins per set in year t ( KPS,) by the total number of net sets involving dolphins made by the tuna fleet in year t (NSETS,), as K, = KPS, NSETS,. (6) These estimates are obtained for each year with the data stratified by vessel fish-carrying capac- ity, amount of tuna caught in the net set, and geographic location of the set, following the gen- eral approach described by Lo et al. (1982). Kill rate information is available from a lim- ited set of tuna fishing trips in the 1960's and from a more extensive set in the 1970's collected by scientific observers placed aboard a large pro- portion of the U.S. fishing vessels. To illustrate the data, some mean kill rates, stratified by amount of tuna caught, are shown in Table 2. Higher kill rates are apparent in successful (>% ton tuna caught) than in unsuccessful (<% ton tuna caught) sets, as are marked declines in kill rates over time. Numbers of dolphin sets and fishing trips on which observations of numbers of dolphins killed were made are shown in Table 3. Observations of the numbers of dolphins killed in the 1960's were made by both the crew and the scientists. Although few observations were made, there is no consistent difference between kill Table 2.— Observed mean kill of dolphins per net set (KPS) by U.S. tuna purse seiners 1964-78, for successful and unsuccessful net sets, with sample sizes (iV), from NMFS records. Successful sets Unsuccessful sets (>% KPS t tuna) N (< 'A t tuna) Year KPS W 1964-72 55.7 343 79 25 1973. 226 576 06 130 1974 15.7 753 24 261 1975 18.9 778 2.5 169 1976 16.1 627 57 126 1977 3.4 2,706 0.9 495 1978 4.2 1,434 3.9 249 rates reported by both types of observers (59 and 52, respectively); this suggests the presence of a noncrew-member observer had no significant effect on the kill rate of dolphins in the 1960's. All data on kill rates of dolphins for the period 1971-78 were collected by noncrew-member sci- entists, precluding a direct comparison of kill rates between fishing trips with (observed) and without (unobserved) scientific observers for this period. Groom, 5 however, reported dolphin kill rates on a fishing trip in 1979 with no scientific observer on board; his kill rates were about 4 times higher than the average rate in 1979 for scientist-observed trips and were approximately 20 times higher than on previous and succeeding observed fishing trips by the same vessel and captain. This difference in mean kill rates was due to the significantly lower proportion of sets with few dolphins killed on Groom's trip than on the scientist-observed trips. For instance, the proportion of sets with zero dolphins killed was 0.23 on Croom's trip with 0.76 for observed trips. Although limited information is available, it appears that kill rates on some unobserved ves- sels were higher in the late 1970's, and that this could result in the observed kill rates being lower than the actual rates. If there has been an "ob- server effect," it most likely occurred in the late 1970's, because regulations were adopted in the United States in 1976 requiring the use of cer- tain dolphin-release procedures, and because sci- entific observers were then used to collect regu- lation compliance information. If kill estimates for the last few years were revised with this in mind, it would only slightly affect the calcula- tions presented here, since the large number of animals killed through 1975 tends to dominate in Equation (2). However, such revisions to the kill estimates could markedly change our perception of the current rate of change of these populations. In addition to the known direct kill of dolphins in the fishery, research has been conducted to estimate both the number of dolphins injured and released alive from the purse seines, and the possible number of dolphins which, while not ex- hibiting injuries, die or suffer reduced viability from stress of capture and handling in the purse seining operation. Observations of the number of injured dolphins have been made aboard tuna vessels since 1975; estimates of the number in- 5 Croom, M. M. 1980. The tuna-porpoise problem: Man- agement aspects of a fishery. M.S. Internship Rep., Marine Resour. Manage. Program, Oreg. State Univ., Corvallis, 41 p. FISHERY BULLETIN: VOL. 81. NO. 1 jured fluctuate around 4.8% of the number killed directly, ranging from 3 to 7%. The problem of stress-induced mortality or debility was explored in a workshop of experts on large mammal physi- ology and pathology, and research plans to ap- proach this problem were developed. 6 Subse- quently one aspect of this problem was examined with dolphin specimens collected aboard tuna vessels. 7 Reproductive tracts were examined for evidence of spontaneous abortion, and muscle tissue for myopathy; no evidence of either was found. No estimates of the magnitude of such effects have been made, and currently no re- search is underway to investigate stress-induced mortality. As a conservative measure, given our limited knowledge, I assume in the estimates of total dolphin mortality given here that all of the injured dolphins subsequently die of their in- juries. Thus estimates of total kill of dolphins are the sum of the estimated numbers killed directly and the numbers injured. Numbers of net sets made by the tuna purse seine fleet have been recorded by the Inter- American Tropical Tuna Commission (IATTC) from logbooks kept by the fishermen (Table 3). In the logbooks the type of each net set may be re- corded, along with tuna catch, location, and other information. The three major types of sets are 1) those known to involve dolphins, 2) those known not to involve dolphins, and 3) those for which the data indicate neither the presence nor absence of dolphins. Types of sets not involving dolphins include "floating object sets" (e.g., a rope, board, log, etc.), a "school fish set" (i.e., a net set on tuna sighted at or near the surface), and a "porpoise set." The logbook data are incomplete, however, because some members of the fleet do not report and because, in some cases, only lim- ited information was recorded by the fishermen. The logbook coverage rate, however, is high. The data in columns D, N, and U in Table 3 have only recently become available, and analy- ses are proceeding to use this information di- rectly to estimate the total number of sets made on dolphins. Preliminary results for the total numbers of sets for each year 8 are similar to 6 Stuntz, W. E., and T. B. Shay. 1979. Report on capture stress workshop, La Jolla. California, May 1979. Southwest Fish. Cent. La Jolla Lab., Natl. Mar. Fish. Serv., NOAA, Admin. Rep. LJ-79-28, 24 p. 7 Cowan, D., and W. Walker. 1979. Disease factors in Ste- nella attenuata and Stenella longirostris taken in the eastern tropical Pacific yellowfin tuna purse seine fishery. South- west Fish. Cent. La Jolla Lab., Natl. Mar. Fish. Serv., NOAA, Admin. Rep. LJ-79-32, 21 p. 8 Inter-American Tropical Tuna Commission. 1981. Tuna- Table 3.— Number of tuna purse seine sets. 1959-78, (D) known to have been made on dolphins in the eastern tropical Pacific, (N) known not to have been made on dolphins, and ( U) unknown if made on dolphins (IATTC text footnote 8), with (is) estimates of the total number of sets made on dolphins (Smith text foot- note 3). Also shown are the numbers of observed fishing trips and purse seine sets on porpoise from NMFS records. Sets made Sets observed Trips Year D N U £ observed 1959 132 759 2,985 1.037 1960 1.644 1,256 7,390 5,696 1961 3,617 3.825 8,694 8,247 1962 2.886 8.830 4,337 4,060 1963 3,290 9.266 6,322 4.687 1964 5,933 7.681 4,745 8.090 67 2 1965 6,172 7,176 5,631 7,981 1966 5,443 7,001 5,247 7,250 28 1 1967 3,510 10,018 3,594 4,478 1968 3,833 8,988 1,642 4,271 15 1 1969 7,664 6.552 2,055 8.678 1970 7,912 9.692 1,664 8,552 1971 4.816 10,728 3,404 5,039 78 5 1972 8.193 4,682 3,514 9,036 272 12 1973 8,686 9,463 3,672 9,998 752 22 1974 7.955 11.669 4,835 8,539 1,120 36 1975 8.172 13,396 4,902 8,951 1,049 31 1976 7.481 17,789 5,184 7,910 1,295 45 1977 7,485 15,005 7,643 9.757 3,335 94 1978 5,174 21,527 5,639 5,910 1,771 102 those given in column E of Table 3, but the re- sults are not yet available in the stratified form needed to estimate numbers killed, described be- low. Earlier estimates of the total number of sets made on dolphins (column Uoi Table 3) were ob- tained indirectly for the years prior to 1970, based on the catches of tuna, and include an adjustment for nonreported sets. For the Period 1959-72 Estimating the annual rate of dolphin kill dur- ing the period 1959-72 is difficult because obser- vations were few, especially in the early part of the period; consequently, extrapolation of infor- mation on kill rates is necessary. One effect on rate of kill is the development and improvement of the "backdown" dolphin-release procedure (Coe and Sousa 1972; Barham et al. 1977), by which the vessel moves in reverse during a short portion of the purse seine retrieval, thereby pulling the net out from under the dolphins. Barham et al. (1977) reported that the "backdown" dolphin-re- lease procedure was developed aboard one vessel in 1959 and 1960, and transferred to a second ves- sel in 1961. Subsequently, the use of the proce- dure expanded rapidly within the fleet, although dolphin investigations. Background paper 6, prepared for the 39th meeting of the IATTC, Paris. October 1981. Inter- Am. Trop. Tuna Comm., c/o Scripps Inst. Oceanogr., La Jolla, CA 92093, 17 p. 6 SMITH: SIZE CHANGES OF THREE DOLPHIN POPULATIONS full use was not evident even by 1964. Comparing kill rates with and without "backdown" is compli- cated however, because the effectiveness of the release procedure has increased over time. No information is available on kill rates from non-U. S. vessels during 1959-72, but the non- U.S. fleet was small. There is little reason to sus- pect that these kill rates were different, because fishermen of both fleets were still learning how to use purse seine gear for catching tuna in asso- ciation with dolphins and how to release the caught dolphins. The available kill rate data for this period were stratified, for use in Equation (6), by amount of tuna caught, size of the vessel, and fre- quency of use of the "backdown" procedure. The data were pooled across the years 1964-72 and extrapolated back to the years 1959-63 when no kill rate data were collected. These stratified kill rates were multiplied by the number of sets made on dolphins in each stratum to estimate the total number of dolphins, of all populations, killed directly in this fishery. Estimating proportions of the total kill of dol- phins from each population for this period is dif- ficult because the yellowfin tuna purse seining was expanding westward and because data on the species of dolphins observed killed are avail- able only for 1971 and 1972. Prior to 1969 this fishery operated shoreward of the range of the whitebelly spinner dolphin, primarily within the range of the eastern spinner and offshore spotted dolphins. The total kill estimates are prorated to population for the years 1959-72, based on ob- served proportions in the 1971-72 data of 70, 23, and 3% for offshore spotted, eastern spinner, and whitebelly spinner dolphins, respectively. The other 4% consisted of several species, primarily common dolphins, Delphinus delphis, which are not considered in this study. Although the tuna purse seine fishery was expanding seaward throughout the 1960's toward the range of the whitebelly spinner dolphin, a major seaward shift occurred in 1969. Lacking detailed data, I assume this year to be the first significant in- volvement of the whitebelly population. Some additional data on the species of dolphins involved in each set has recently become avail- able from the IATTC, suggesting a declining proportion of sets involving spinner dolphins and an increasing proportion involving spotted dol- phins throughout the 1960's. Preliminary exami- nation of these data indicates that the overall proportions of sets involving each species are not greatly different from the 1971-72 observer data. Direct use of these new data will involve making a number of assumptions about species-specific kill rates. Using the above proportions based on the 1971- 72 data and increasing the estimates of total number killed by 4.8% to account for those dol- phins possibly dying of injuries, I estimated the total numbers of dolphins killed, by population (Table 4). These are revisions of estimates used by the 1979 workshop (footnote 3). Table 4. — Estimates of numbers (in thousands) of dolphins killed by all fleets in the eastern tropical Pacific, 1959-78, for three populations of dolphins (Smith text footnote 3). Offshore Eastern Whitebelly Year spotted spinner spinner 1959 71 27 1960 357 133 1961 402 150 1962 167 62 1963 183 69 1964 306 115 1965 337 126 1966 306 115 1967 206 77 1968 178 67 1969 365 122 15 1970 355 118 14 1971 176 59 7 1972 288 96 12 1973 131 32 33 1974 95 26 47 1975 105 45 34 1976 47 9 20 1977 22 5 5 1978 19 2 4 For the Period 1973-80 Substantially more data exist on kill rates for the period 1973-78 than for the period 1959-72. The 1973-78 data are more reliable because they were collected by NMFS employees trained spe- cifically for obtaining kill information. Starting in 1974 fishing trips were randomly selected for observation to obtain a representative sample. Greater cooperation by the fishing fleet resulted in an increasing proportion of selected trips actually observed from 1974 to 1976. However, it was not until 1976 that fishing trips begun after July were sampled. In the early 1970's fish- ing tended to occur farther offshore later in the year; because kill rates are generally higher in the offshore areas, the failure to collect data from late-season trips probably resulted in an under- estimate of actual dolphin kill rates in those years. This problem is partially accounted for by stratifying the data by area. The species composi- FISHERY BULLETIN: VOL. 81. NO. 1 tion of the kill was also recorded, allowing direct estimates of total kill of dolphins, from Equation (6), for each population. The number of dolphins killed per set from 1973 to 1976 for successful and unsuccessful sets was about 18 and 3, respectively, a decrease from the 1964-72 levels of 56 and 8. The number killed in successful and unsuccessful sets in 1977 and 1978 was again lower, about 4 and 2, respectively (Table 2). These decreases occurred as U.S. regu- lations were developed and eventually imple- mented, and as methods for more effective use of backdown and other dolphin-release procedures were developed and used. The decreases in kill rates were apparently due, at least in part, to wider adoption of procedures for dolphin release. The non-U. S. tuna purse seine fleet increased markedly during this period. First observations of the kill rate for this fleet were in 1979, which showed that the rate was very similar to that of the U.S. fleet (Allen and Goldsmith 1981). Given this similarity in 1979, it is reasonable to assume that during the earlier part of the 1970's the non- U.S. kill rate declined, as did the U.S. kill rate (Table 2), as dolphin-release technology devel- oped by the U.S. fleet became known. If such a decline in the non-U. S. kill rate occurred, how- ever, it would probably have been somewhat slower than that for the U.S. fleet, because of lack of legal pressure to reduce the incidental kill and time lags in technology transfer. Following the procedure developed by the 1979 workshop (footnote 3), I estimated the non-U. S. kill by assuming 1) the same kill rate in 1971-72 for the non-U. S. fleet as that observed aboard U.S. ves- sels in those years; 2) the same kill rate in 1973 for the non-U. S. fleet as that of the U.S. fleet in 1975; and 3) a linear convergence of the two rates toward the 1979 U.S. rate. Estimates of numbers of dolphins killed by non-U. S. vessels obtained under these assumptions are used here. How- ever, additional study is needed, especially since the recorded kill rate for the non-U. S. fleet in 1980 was somewhat higher than that for the U.S. fleet (Allen and Goldsmith 1982). These kill rates, stratified by vessel size, amount of tuna caught, and area fished, are used in Equation (6), along with the estimated num- ber of sets on dolphins, to estimate total direct kill by population for each year. These estimates are then increased by 4.8% to account for dol- phins assumed to die of their injuries (Table 4). The results in Table 4 are slight revisions of the estimates used by the 1979 workshop (footnote 3). NET RECRUITMENT RATE ESTIMATES Maximum net recruitment rate (i?,„) is re- quired to estimate historical abundance. This is calculated as the difference between gross pro- duction of calves and the natural mortality rate, assuming that natural mortality does not change, when a population is reduced substantially be- low its equilibrium level. Gross Reproductive Rates Gross recruitment rates can be estimated as the product of the female fraction of the popula- tion, the mature female fraction, and the annual pregnancy rates. Estimates of these parameters are given in Table 5, based on samples of dol- phins collected by scientific observers aboard tuna vessels from 1973 to 1978. Two methods were used to estimate the annual pregnancy rate: The first method (I) is the observed proportion of pregnant females in the population divided by the gestation period; the second method (II) is similar, but uses additional information on fre- quency of nursing calves in the samples from each net set (Perrin et al. 1977a, b, c). There are known sampling biases in these data for spotted dolphin because of the fishing pro- cess, partly accounted for by using data for spot- ted dolphin recruitment rates from only those sets where more than 40 dolphins were killed. In addition, the observed fraction of the mature, pregnant female dolphins has varied among years, with a general decline in offshore spotted dolphin and a large degree of variability in east- ern spinner dolphin. Age-specific effects are not accounted for in the analyses so far, however, particularly the Table 5. — Proportion of sampled dolphins (female and ma- ture) of three populations and estimates of annual pregnancy rate (P) and gross reproductive rate (G), using two methods. 1 See text for details. Annual production Proportion Method I Metf P lod II Population Female Mature P G G Offshore spotted 0.56 0.56 0.38 0.119 0.32 0.100 Eastern spinner 0.51 0.43 0.34 0.075 0.45 0.099 Whitebelly spinner 0.51 0.52 036 0.096 0.33 0.088 'Henderson, J. R..W. F. Perrin, and R. B. Miller. 1980. Ratesof gross annual production in dolphin populations (Stenella spp and Delphinus delphis) in the eastern tropical Pacific, 1973-1978 Southwest Fish. Cent. La Jolla Lab., Natl. Mar Fish. Serv , NOAA. Admin Rep. LJ-80-02. 51 p. 8 SMITH: SIZE CHANGES OF THREE DOLPHIN POPULATIONS lower pregnancy rate that probably occurs in older animals. New methods are being developed for age determination, and an effort is being made to apply these methods to age the samples of dolphins. With accurate data on age of ani- mals, a more detailed examination of sampling biases will be undertaken. Natural Mortality Rates No direct estimates of natural mortality rates exist for the eastern Pacific dolphin populations, as might be obtained from tagging data or from a sampled age structure. Ohsumi (1979) presented a statistical relationship between natural mor- tality rate and body length for cetaceans, from which can be derived an annual, natural mortal- ity rate of around 0.14 for the eastern Pacific dolphin populations. However, this estimate is obtained by extrapolating the relationship out- side the range of his data, and consequently is unreliable. Another method of estimating natural mortal- ity rate is from information on gross reproduc- tive rate for a population in equilibrium with its environment, assuming natural mortality does not change with population size. This approach was used in the 1976 workshop (footnote 2). An estimate of gross reproductive rate of 0.09 (Ka- suya et al. 1974) for a population off Japan, thought to be lightly exploited, was used as the natural mortality rate estimate for the eastern tropical Pacific populations. It now appears that the population off Japan had, in fact, been ex- ploited to a greater degree than was thought, and that there is segregation of prepubertal dolphins into separate schools (footnote 3, p. 41). The as- sumption, consequently, of a natural mortality rate of 0.09 is probably not valid. In the 1979 workshop (footnote 3), estimates of the gross reproductive rate of lightly exploited Southern Hemisphere populations of spotted and spinner dolphins in the eastern tropical Pacific were used as estimates of natural mortality rates. These rates were 0.098 and 0.067 for spot- ted and spinner dolphins, respectively. Net Rates Net recruitment rates for the offshore spotted, eastern spinner, and whitebelly spinner dolphin populations can be estimated as the differences between the gross reproductive rate estimates, listed in Table 5, and the corresponding natural mortality rate estimates given above. Using method I estimates of pregnancy rates, one ob- tains estimated net reproductive rates of 0.021, 0.008, and 0.029 for these three populations, re- spectively. Using method II estimates of preg- nancy rates, one obtains estimates of 0.002, 0.032, and 0.021, respectively. These highly variable estimates are unsatisfactory, because they are based on data with known sample biases, and they differ among populations in unexpected ways. In particular, it is not expected that the net reproductive rate of the whitebelly spinner dol- phin, which has been relatively less exploited, should be higher than that of the more heavily exploited eastern spinner dolphin popula- tion. Due to these uncertainties, specific point esti- mates were not obtained by the 1979 workshop participants. Rather, a range of values from 0.0 to 0.04 were considered equally likely, given the available information. The lower value of 0.0 was selected by the 1979 workshop to reflect uncer- tainties about unexpected changes in some repro- ductive rates, and the small magnitude of the estimates of net reproductive rates. This range compares with the estimates from the 1976 work- shop of 0.02-0.06, with a midpoint estimate of 0.04. Although higher rates of increase of ceta- cean populations have been reported, contrary to the conclusions in the 1979 workshop report, there are no reliable estimates of rates of increase for dolphin populations which can be used with confidence. Pending better information, the range of estimates considered in the two work- shops will be used here, recognizing that higher rates may be possible. Rate Dependent on Population Size The evidence on which to base an estimate of the value of Z in Equation (3) for dolphin popula- tions is limited. Fowler (1981) argued that for large, long-lived mammals, Z is greater than unity. He based this conclusion on a review of empirical data, primarily from terrestrial popu- lations, and on an analysis of the demographic constraints which come with long life and ex- tended parental care. McCullough (footnote 3, p. 8) gave preliminary estimates of maximum net productivity level (MNPL), and hence Z, for four large terrestrial mammal populations. His esti- mates agree with Fowler's conclusions that Z is greater than unity, and that later reproducing animals would have higher values of Z. FISHERY BULLETIN: VOL. 81. NO. 1 The 1976 workshop (footnote 2) concluded that the available information implies MNPL is with- in the range of 50-70% of the equilibrium popula- tion size, corresponding to values of Z from 1 to 5.1. The 1979 workshop recognized that "There had been a shift of scientific opinion in recent years [since 1976] towards accepting the idea that relative net productivity in mammals, especially large, K-selected species, is a non-linear func- tion of population size," (footnote 3, p. 7) and con- cluded that MNPL for these dolphin populations is probably in the range of 65-80% of the equilib- rium population size (Zfrom 3.5 to 11.5). I con- sider the values for MNPL of 50-80% (Zfrom 1 to 11.5) of equilibrium population size in order to explore the sensitivity of the calculations to this uncertainty. HISTORICAL TRENDS IN ABUNDANCE Estimates of population sizes prior to 1979 from Equations (2) and (3) for each population are shown in Table 6. Values are given using 1) two different estimates of present (1979) abun- dance (from combined research vessel and fish- ing vessel data, and from research vessel data alone), and 2) the parameters MNPL = 65% and R,„ — 0.03. For this range of parameter values, the offshore spotted dolphin population in 1959 Table 6.— Estimates of population size (in thousands) of off- shore spotted, eastern spinner, and whitebelly spinner dolphins from 1979 back to 1959, using Equations (2) and (3) and param- eters MNPL = 65% and R,„ = 0.03. 1979 estimates are based on species proportions from (FR) combined research vessel and fishing vessel data and (R) research vessel data alone. Offshore spotted Year FR Eastern spinner FR = R Whitebelly spinner FR 1979 1978 1977 1976 1975 1974 1973 1972 1971 1970 1969 1968 1967 1966 1965 1964 1963 1962 1961 1960 1959 2,775 2,719 2,668 2,673 2,675 2,679 2,754 2,967 3,064 3,340 3,264 3,720 3,844 4.071 4,335 4,574 4,695 4,803 5,169 5,519 5.590 1,682 1,653 1,628 1,629 1,686 1,732 1,824 2,046 2,164 2,457 2,756 2,865 3,001 3,239 3,520 3,754 3,879 3,991 4,358 4,708 4,779 293 287 283 284 320 336 358 443 488 591 695 742 799 893 998 1,092 1,141 1,185 1,323 1.454 1.481 380 376 373 386 434 456 486 494 499 512 527 527 527 527 527 527 527 527 527 527 527 216 215 214 229 258 301 331 340 345 358 373 373 373 373 373 373 373 373 373 373 373 was between about 4,800,000 and 5,600,000 ani- mals. The eastern spinner dolphin population in 1959 numbered about 1 ,500,000, while the white- 1.0 0.9 LU 0.8 N (/) ( Z o 1- 06 < _l 3 Q. Oh o a. LU 04 > 1- < 3 _l LU OC 0.2 0.1 0.0 Whitebelly \Spinner _l I L_ Eastern Spinner _L 1960 1965 1970 YEAR 1975 Figure 1. — Relative population sizes of whitebelly spinner, offshore spotted, and eastern spinner dolphins, 1959-79, using population estimates based on species proportions from com- bined research and fishing vessel data, and assuming R „, = 0.03 and MNPL = 65% of equilibrium abundance. Population sizes are relative to estimated population sizes in 1959. 1.0 0.9 0.8 LU N (/) Z o 0.7 0.6 a 0.5 O a. iu 0.4 > < 0.3 _i LU K 0.2 0.1 0.0 Whitebelly \Spinner \ Eastern N — _ Spinner _L_I I i_ 1960 1965 1970 YEAR 1975 Figure 2.— Relative population sizes of whitebelly spinner, offshore spotted, and eastern spinner dolphins, 1959-79, using population estimates based on species proportions from re- search vessel data alone, and assuming R ,„ = 0.03 and MNPL - 65% of equilibrium abundance. Population sizes are relative to estimated sizes in 1959. 10 SMITH: SIZE CHANGES OF THREE DOLPHIN POPULATIONS belly spinner dolphin population in 1969 num- bered between 400,000 and 500,000. The offshore spotted and eastern spinner dolphin populations declined rapidly in the 1960's and early 1970's in the face of kills which were, for example, on the order of 7-12% of the 1965 population sizes. The whitebelly spinner dolphin population declined most rapidly in 1974 when the kill was between 11 and 16% of its population size. These estimates of absolute population sizes are shown in Figures 1 and 2 relative to the equi- librium population size (N t /N,,), so that the trend in abundance of these populations can be exam- ined. For all of the parameter values considered, these dolphin populations have declined substan- tially relative to their pre-exploitation sizes. The ratio of 1979 to pre-exploitation popula- tion sizes for different values of R,„ and MNPL (and hence Z) shows the sensitivity of the calcula- tions to changes in parameter estimates (Table 7; Figs. 3,4). The value of MNPL when R,„ is zero is not meaningful, as the estimate of pre-exploi- tation population size (Equation (2)) collapses to the sum of the present population size estimate and the total numbers killed over all years. This is reflected in Figures 3 and 4 in the convergence of the lines when R m is zero. Table 7.— Estimates of 1979 relative population sizes of off- shore spotted, eastern spinner, and whitebelly spinner dolphin populations, using two estimates which differ in species pro- portions from (FR) combined fishing and research vessel data and from (R) research vessel data alone, for ranges of maxi- mum net recruitment rate (R,„ ) and maximum net productiv- ity level (MNPL). Offshc re Eastern Wh itebelly MNPL (%) spotted spinner FR = ft S| Dinner Rm FR ft FR ft 0.00 — 0.40 0.29 0.17 066 053 0.03 50 0.45 0.32 0.18 0.69 0.55 65 0.50 0.35 020 0.72 0.58 80 0.53 0.37 0.21 0.77 0.61 0.06 50 049 0.35 0.20 0.71 0.57 65 060 0.42 023 0.78 0.63 80 0.68 0.47 0.25 086 0.69 1.0 0.9 - UJ N 0.8 7 O 0./ 1- < -J 0.6 3 0. O 0.5 LU > P 0.4 < -1 LU 0.3 0.2 0.1 0.0 0.00 0.03 R m 0.06 1.0 i- 0.9 - B 0.8 i °' 7 H < 0.6 - Q. O 0.5 Q. LU > 0.4 I- < uj 0.3 h K 0.2 - 0.1 - 0.0 m 5>- ef*a •-- =B) Easternjg-^^ESS 0.00 0.03 'm 0.06 Figure 3.— Population size of offshore spotted dolphins in 1979 relative to 1959 (Nn/Nss) as a function of maximum recruit- ment rate (i?,,, = 0, 3, 6%) using two current population estimates which differ in species proportions from (FR) combined fish- ing and research vessel data and from (R) research vessel data alone. MNPL values of 50% (dashed lines), 65% ( solid lines), and 80% (dot-dashed lines) are shown. Figure 4.— Population sizes of eastern spinner and whitebelly spinner dolphins in 1979 relative to 1959 (Ni 9 /N 59 ) as a function of maximum recruitment rate(.R„, =0, 3, 6%) using two current population estimates which differ in species proportions from (FR) combined fishing and research vessel data and from (R) research vessel data alone. MNPL values of 50% (dashed lines), 65% (solid lines), and 80% (dot-dashed lines) are shown. 11 FISHERY BULLETIN: VOL. 81. NO. 1 DISCUSSION AND CONCLUSIONS The three populations of dolphins involved with the yellowfin tuna purse seine fleet in the eastern tropical Pacific have declined since 1959 and the decline was not arrested until recently (Figs. 1, 2). Assuming the historical kill level, and the central values for R m and MNPL, the whitebelly spinner dolphin population has de- clined to between 58 and 72% of its pre-exploita- tion levels; the offshore spotted dolphin popula- tion has declined to between 35 and 50% of its pre-exploitation size; and the eastern spinner dolphin population has declined to around 20% of its pre-exploitation size. Examination of Figures 3 and 4 shows that the numerical values of the estimates of relative abundance in 1979 for offshore spotted dolphin and whitebelly spinner dolphin are relatively more sensitive to changes in the maximum net recruitment rate and the maximum net produc- tivity level parameters than are the estimates for the eastern spinner dolphin. Also, the sensitivity of these calculations to the maximum net pro- ductivity level increases markedly as the value of the maximum net recruitment level increases. The sensitivity (in percent change) in the ratio of present to pre-exploitation abundance, however, is largest for the offshore spotted dolphin and least for the whitebelly spinner dolphin. This is due in part to the shorter time span over which the whitebelly spinner dolphin has been exploit- N t ed, and in part to the lower — — ratio for the east- ern spinner dolphin ratio, which makes smaller differences result in a larger percentage. Although there are a number of uncertainties about specific parameter estimates used in these calculations, the general declines in abundance change relatively little over the ranges of param- eter estimates explored. For example, rather rapid declines in the 1960's, followed by decreas- ing rates of decline in the 1970's, are evident for all parameter values considered. Specific aspects of these declines in abundance, however, depend to a greater degree on the actual parameter val- ues. For example, the estimated changes in popu- lation sizes from 1975 to 1978 vary with the spe- cific values of maximum net recruitment rate, while the estimated changes in population sizes in the 1960's are relatively insensitive to this parameter. In order to improve our estimates of reproduc- tive and mortality rates, a complete review of vita! rates for these dolphin populations and for cetaceans in general should be carried out. Sev- eral approaches to this problem have been identi- fied, including a detailed review of the eastern tropical Pacific dolphin data and of the existing data for other cetacean populations. Given the gaps in our knowledge of cetacean reproductive processes, analyses of alternate mathematical models of such processes will be fruitful. Although improvements in estimates of abun- dance and kill levels are needed, these areas are generally much better understood than the re- cruitment process. Population-size estimation techniques are still being improved upon; cur- rent emphasis is on testing the assumptions needed in applying line transect theory to aerial sighting survey data and in estimating dolphin school size. Future work will emphasize im- proved shipboard sighting methodology for pos- sible application of line transect theory. Marked improvements in the estimates of num- bers of dolphins killed are not anticipated; key areas needing additional information are the kill rates both in the non-U. S. fleet and on unobserved fishing trips. Neither of these areas is readily amenable to study, although further analysis of the kill rates on unobserved trips may provide some basis for exploring this uncertainty. The possible levels of indirect mortality or debility due to the stress of chase and capture are also of concern. Because of the large numbers of dol- phins captured and released each year, even very low rates of indirect mortality could have a sig- nificant effect on the population. ACKNOWLEDGMENTS This assessment of the status of the dolphin populations is built on data collected by many in- dividuals. The collection of these data has been made possible in large measure by the coopera- tion of the U.S. tuna fishing fleet. In addition, many individuals have contributed to the analy- sis of the data, including National Marine Fish- eries Service staff and numerous scientists from various organizations. It is not possible to ac- knowledge the contributions of specific individ- uals to information presented here because of the large numbers of people who have been involved, but without their efforts the present analysis would not be possible. I also wish to acknowledge the very helpful reviews of an earlier draft of this paper by Douglas Chapman, John Gulland, Linda Jones, and Jeff Breiwick. 12 SMITH: SIZE CHANGES OF THREE DOLPHIN POPULATIONS LITERATURE CITED Allen, K. R. 1981. Application of population models to large whales. In C. W. Fowler and T. D. Smith (editors), Dynamics of large mammal populations, p. 263-275. John Wiley, N.Y. Allen, R. L.,D. A. Bratten.J. L. Laake, J. F. Lambert, W. L. Perryman, and M. D. Scott. 1980. Report on estimating the size of dolphin schools, based on data obtained during a charter cruise of the M/VGina Anne, October 11-November 25, 1979. In- ter-Am. Trop. Tuna Comm., Data Rep. 6, 28 p. Allen, R. L., and M. D. Goldsmith. 1981. Dolphin mortality in the eastern tropical Pacific incidental to purse seining for yellowfin tunas, 1979. Rep. Int. Whaling Comm. 31:539-540. 1982. Dolphin mortality in the eastern tropical Pacific incidental to purse seining for yellowfin tunas, 1980. Rep. Int. Whaling Comm. 32:419-422. Barham, E., W. K. Taguchi, and S. B. Reilly. 1977. Porpoise rescue methods in the yellowfin purse seine fishery and the importance of Medina panel mesh size. Mar. Fish. Rev. 39(5):1-10. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density for line transect sampling of biological populations. Wildl. Monogr. 72, 202 p. Coe, J., and G. Sousa. 1972. Removing porpoise from a tuna purse seine. Mar. Fish. Rev. 34(11-12):15-19. Fowler, C. W. 1981. Density dependence as related to life history strat- egy. Ecology 62:602-610. Green, R. E., W. F. Perrin, and B. P. Petrich. 1971. The American tuna purse seine fishery. In H. Kristjonsson (editor), Modern fishing gear of the world 3, p. 182-194. Fish. News (Books) Ltd., Lond. Holt, R. S., and J. E. Powers. 1982. Abundance estimation of dolphin stocks involved in the eastern tropical Pacific yellowfin tuna fishery de- termined from aerial and ship surveys. U.S. Dep. Com- mer., NOAA Tech. Memo. NOAA-TM-NMFS SWFC- 23. Kasuya, T., N. Miyazaki, and W. H. Dawbin. 1974. Growth and reproduction of Stenella attenuata in the Pacific coast of Japan. Sci. Rep. Whales Res. Inst. (Tokyo) 26:157-226. Lo, N. H., J. Powers, and B. Whalen. 1982. Estimating and monitoring incidental dolphin mortality in the eastern tropical Pacific tuna purse seine fishery. Fish. Bull., U.S. 80:396-401. Ohsumi, S. 1979. Interspecies relationships among some biological parameters in cetaceans and estimation of the natural mortality coefficient of the southern hemisphere Minke whale. Rep. Int. Whaling Comm. 29:397-406. Pella, J. J., and P. K. Tomlinson. 1969. A generalized stock production model. [In Engl. and Span.] Inter-Am. Trop. Tuna Comm.. Bull. 13: 421-496. Perrin, W. F. 1975. Distribution and differentiation of populations of dolphins of the genus Stenella in the eastern tropical Pa- cific. J. Fish. Res. Board Can. 32:1059-1067. Perrin, W. F., J. M. Coe, and J. R. Zweifel. 1977a. Growth and reproduction of the spotted porpoise, Stenella attenuata, in the offshore eastern tropical Pa- cific. Fish. Bull., U.S. 74:229-269. Perrin, W. F., D. B. Holts, and R. B. Miller. 1977b. Growth and reproduction of the eastern spinner dolphin, a geographical form of Stenella longirostris, in the eastern tropical Pacific. Fish. Bull., U.S. 75:725- 750. Perrin, W. F., R. B. Miller, and P. A. Sloan. 1977c. Reproductive parameters of the offshore spotted dolphin, a geographical form of Stenella attenuata, in the eastern tropical Pacific, 1973-1975. Fish. Bull., U.S. 75:629-633. Perrin, W. F., P. A. Sloan, and J. R. Henderson. 1979. Taxonomic status of the "southwestern" stocks of spinner dolphin, Stenella longirostris and spotted dol- phin, S. attenuata. Rep. Int. Whaling Comm. 29:175- 184. Polacheck, T. 1982. Local stability and maximum net productivity levels for a simple model of porpoise population size. U.S. Dep. Commer., NOAA Tech. Memo. NOAA-TM- NMFS SWFC-17. Smith, T. D. 1981. Line transect techniques for estimating density of porpoise schools. J. Wildl. Manage. 45(3):650-657. Smith, T. D., and T. Polacheck. 1979. Analysis of a simple model for estimating histori- cal population sizes. Fish. Bull., U.S. 76:771-779. 13 FOOD HABITS OF YELLOWTAIL FLOUNDER, LIMANDA FERRUGINEA (STORER), FROM OFF THE NORTHEASTERN UNITED STATES Richard W. Langton 1 ABSTRACT Stomachs of 1,021 yellowtail flounder caught in 1973-76 contained primarily polychaetes (43%) and crustaceans (18%) as a percentage weight of total contents. The most important prey were Spiophanes bombyx (9.68%) and Unciola sp. (13.65%). Predator size had little effect on diet com- position whereas geographic distribution did. Spiophanes bombyxwas three times more important as prey on Georges Bank than in southern New England, and amphipods were more important in southern New England than on Georges Bank. From the middle Atlantic to southern New England to Georges Bank the total weight of stomach contents increased from 0.12% to 0.14% to 0.21% of the fishes' body weight. Year-to-year differences were inconsistent; however, fish stom- achs from spring cruises contained more food, 0.20%, than those from autumn cruises, 0.14% body weight. During a composite 24-hour day, peak stomach content weight occurred in the afternoon to early evening. Polychaetes accounted for less of the stomach contents at night while amphipods increased in importance during the night. Sex of the fish had no effect on diet composition although the stomachs of females were fuller than males, 0.15% vs. 0.11% body weight. Neither diet composition nor the percentage of empty stomachs were related to gonadal maturity stages, but stomachs from spawning fish contained the least amount of prey, 0.06%, while resting-stage fish contained the most, 0.24% body weight. Over a 12°C temperature range there was little change in diet composition, but between 3° and 8°C a greater percentage of stomachs contained prey and a larger quantity of prey than between 9° and 15°C. Over a 220 m depth range the stomach content weight increased with depth for smaller fish (<15 cm), while the percentage of empty stomachs increased for larger fish (>21 cm). Diet composition showed the greatest effect of depth with S. bombij.r dominating the diet in the 74-110 m depth zone (26.6% of the stomach content weight) and Crangon septemspinosa, also being dominant in a single depth zone, comprising 39.6% of the diet at 147-183 m. The yellowtail flounder, Limanda ferruginea (Storer), is a right-handed, thin-bodied flounder that occurs along the eastern seaboard of North America from Labrador to Chesapeake Bay (Bigelow and Schroeder 1953; Royce et al. 1959). It has contributed significantly to the total flat- fish catch, primarily from southern New Eng- land and Georges Bank, since about 1935 (Royce et al. 1959; Sissenwine et al. 1978 2 ). Biological in- formation has been summarized by Bigelow and Schroeder (1953) and updated by Lux and Liv- ingston (in press). These summaries qualitatively describe the diet as consisting of small crusta- ceans, worms, and molluscs. Quantitative work ■Northeast Fisheries Center Woods Hole Laboratory, Na- tional Marine Fisheries Service, NOAA, Woods Hole, Mass.; present address: Maine Department of Marine Resources, Ma- rine Resources Laboratory, West Boothbay Harbor, ME 04575. 2 Sissenwine, M. P., B. E. Brown, and M. M. McBride. 1978. Yellowtail flounder (Limanda ferruginea): Status of the stocks, January 1978. Northeast Fisheries Center Woods Hole Laboratory Reference No. 78-02, 27 p. on the diet is limited. Inshore yellowtail flounder have been examined by Libey and Cole (1979) off Cape Ann in Massachusetts while Efanov and Vinogradov (1973) surveyed the offshore feeding pattern of yellowtail flounder in southern New England and on Georges Bank. Langton (1979 3 ), Grosslein et al. (1980), and Langton and Bowman (1981) described the diet of fish from the middle Atlantic to western Nova Scotia, and Pitt (1976) conducted a study on the Grand Banks. These papers generally agree that crustaceans, par- ticularly amphipods, and polychaetes are major prey items. However, the absolute quantities of prey in the stomachs differ, being influenced by both biological and abiotic factors. Only one of the studies lists the stomach contents by predator size (Pitt 1976) and none of the studies evaluate comprehensively all factors influencing the diet Manuscript accepted June 1982. FISHERY BULLETIN: VOL. 81, NO. 1. 1983. 3 Langton, R. W. 1979. Food of yellowtail flounder, Li- ma ndaferruginea( Storer). International Council for Explor- ation of the Sea. CM. 1979/G:54, 10 p. 15 FISHERY BULLETIN: VOL. 81. NO. 1 of this predator. The purpose of the present paper is to describe the stomach contents of yel- lowtail flounder and quantitatively evaluate fac- tors influencing the quantity and composition of the animal's stomach contents. METHODS Yellowtail flounder stomachs were collected on eight bottom trawl survey cruises conducted from 1973 through 1976. The dates of the cruises are as follows: 16 March-15 May 1973; 26 Sep- tember-20 November 1973; 12 March-4 May 1974; 20 September-14 November 1974; 4 March- 12 May 1975; 15 October-18 November 1975; 4 March-8 May 1976; 20 October-23 November 1976. Fish collections were made from the RV Albatross IV or RV Delaware II, using a #36 Yankee otter trawl for autumn surveys and a #41 Yankee otter trawl for spring surveys. A scheme of stratified random sampling was carried out in the continental shelf waters between Nova Scotia and Cape Hatteras, N.C. For survey purposes this region has been divided into five geographic areas, which are further subdivided into depth strata as depicted in Clark and Brown (1977) and described by Grosslein (1969 4 ). Yellowtail flounder were selected from the catch in, primarily, two of the five geographic areas, i.e., southern New England and Georges Bank which include the three major fishing grounds and major yellowtail flounder stocks in U.S. waters (Lux 1963). Stomachs were labelled according to vessel, cruise, station, length, sex, and sexual maturity and were preserved individ- ually in a gauze wrapping in 10% Formalin 5 . The sampling strategy was designed to collect fish, more or less at random, from the population with- out bias towards a specific length, except as de- scribed below. We attempted to collect 50 fish per geographic area per cruise for fish both above and below 12 cm TL (total length). Twelve centimeters in length approximates the length of 1- to 2-yr-old fish, and these smaller fish were preserved intact after the body cavity was cut open to insure fixation of the contents. In the laboratory, individual stomachs were opened, and the contents emptied onto a fine mesh screen and rinsed with seawater. The vari- 4 Grosslein, M. E. 1969. Groundfish survey methods. Northeast Fisheries Center Woods Hole Laboratory Reference No. 69-2, 34 p. 5 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ous items were sorted and identified to the lowest possible taxa. Each distinct group was blotted dry and immediately weighed. In the text and tables these weights have been expressed as a percentage of the total weight of stomach con- tents. In the text these percentages are often given in brackets after the mention of taxa to quantify their relative importance. Twelve percent of the fish collected fell into the three smallest size classes (Table 1) with a mean length of 7.6 cm. Fish >15 cm TL were equally distributed around the 31-35 cm size class with 70% (h =715) of all fish examined falling between 26 and 40 cm TL. The average length of all fish comprising this peak is 32.8 cm. For some analy- ses two size-related groupings of fish, represen- tative of this bimodal distribution, have been differentiated while in other cases the data are presented by 5 cm length classes or expressed as a percentage of the fishes' body weight according to the length/weight equation in Wilk et al. (1978). [W=aL h where a = 0.4514" 5 , 6 = 3.1257, and L is in millimeters.] RESULTS Food Of the 1,021 stomachs examined, 684 contained prey which weighed in total 422 g. The overall mean fish length and standard deviation was 29.4+10.5 cm. The prey were allocated into 148 different categories, which included all taxonom- ic levels of identification and such miscellaneous categories as sand and unidentifiable animal re- mains. The most important major taxonomic groupings were polychaetes and crustaceans (Table 1). Polychaetes accounted for 43% of the stomach contents. The families Spionidae (13.27%), Lum- brinereidae (1.90%), Sabellidae (1.42%), and Nephtyidae (1.19%) were all of some importance. Spiophanes bombyx was the major prey, making up 9.68% of the weight of the total stomach con- tents. Other polychaetes (17.24%) and polychaete tubes (7.94%) accounted for the remainder of the prey in this taxon. Crustaceans (18.0%) were second in impor- tance, the amphipods (13.65%) being the major prey group. Unciola sp. (4.41%), Leptocheirus pinguis (2.25%), and Byblis serrata (1.72%) were important amphipod prey. Other gammarids (1.92%), ampeliscids (1.56%), and corophiids (0.3%) made up most of the remaining amphipod 16 LANGTON: FOOD HABITS OF YELLOWTAIL FLOUNDER Table 1.— Principal items in stomachs of yellowtail flounder, Limandaferruginea, by 5 cm length classes. Data are expressed as a percentage of the total weight of stomach contents (+ indicates present but <0.01%). Fish length intervals in centimeters Stomach contents 1-5 6-10 11-15 16-20 21-25 26-30 31-35 36-40 41-45 46-50 51-55 Anthozoa 1.09 0.69 4.24 8 10 3.69 971 Other Cnidaria 0.01 Sptophanes bornbyx 11.74 5.82 0.93 1028 11.97 11.99 11.06 Spionidae 0.60 1.77 2.71 4.91 9.80 Sabellidae 2.02 0.97 3.34 088 0.16 Annelida tubes 1.01 4.46 12.12 916 14.35 Lumbrinereidae 0.54 3.05 349 1.05 0.11 2.22 Nephtyidae + 0.54 0.28 280 0.92 0.23 039 Other polychaetes 10.82 11.91 13.64 29.19 2335 21.83 14.51 1646 1226 0.33 Byblis serrata 1.73 2.02 0.59 1.24 1.24 2.35 2.50 2.34 0.03 Other Ampeliscidae 1471 0.79 1.81 0.71 2.45 2.25 1.40 0.54 0.92 Unciola sp + 2.22 3.64 1528 1286 1203 5.36 3.44 0.26 0.60 Other Corophndae 4.02 3.31 0.93 0.33 1.34 0.32 0.07 0.01 Gammandae + 3.77 4.96 1.47 1.61 421 2.14 1.35 2.79 0.12 Leptocheirus pmguis + 8.51 4.77 3.81 1.92 3.11 1.96 Other Amphipoda 6.24 0.93 1.43 1.27 2.55 1.03 1.53 0.18 3.87 Crangon septemspmosa 41.38 9.09 11.98 3.25 7.14 2.53 0.71 001 Dichelopandalus leptocerus 31.83 0.64 1.96 066 009 Other crustaceans 58.62 19.16 4.37 1928 6.94 2.32 1 96 0.72 1065 0.16 Animal remains + 15.39 7.28 28.42 2446 19.30 1609 15.47 3679 0.59 Sand 0.93 0.60 1.81 10.19 822 10.22 17.34 23.02 1030 8845 Other groups + 0.19 18.40 5.39 0.31 2.28 3.38 1.25 1.99 0.93 0.88 Number examined 39 77 21 23 63 187 337 191 60 18 2 Number empty 22 25 6 9 16 63 108 62 20 4 Mean weight per stomach (g) 00001 0.021 0072 0.103 0230 0242 0375 563 0.950 2.445 9652 Mean length (cm) 4.0 8.3 120 18.1 23.4 283 32.7 377 42.3 47.5 51.0 prey. Only two other crustaceans were of signifi- cance in the yellowtail flounder's diet, namely, the shrimps Crangon septemspinosa (1.89%) and Dichelopandalus leptocerus (0.94%). All other taxonomically distinct groups con- tributed only 4.96% of the weight of stomach con- tents. Unidentifiable animal remains (17.18%) and sand ( 16.92%) accounted for the remainder of the total weight of stomach contents. Size-Related Feeding Habits Amphipods were the most important prey for the smaller yellowtail flounder although stom- achs from every size class of fish contained am- phipods (Table 1). Polychaetes comprise a greater percentage of the stomach contents of the larger fish but, like amphipods, they occur in stomachs from most every size class. The occurrence of anthozoans in the larger size fish (>26 cm) might reflect a tendency for larger yellowtail flounder to be selecting "wormlike" prey. Geographic Comparison Composition of the diet of yellowtail flounder in southern New England and on Georges Bank was similar, with polychaetes and amphipods accounting for 50 to 70% of the total weight of stomach contents in both areas (Table 2). Poly- chaetes were the major prey in both regions with TABLE 2.— Principal items in stomachs of yellowtail flounder, Limandaferruginea, by geographic area in the northwest At- lantic. Data are presented as a percentage of the total weight of stomach contents (+ indicates present but <0.01%). Southern Middle New Georges Stomach contents Atlantic England Bank Anthozoa 1.93 3.52 Other Cnidaria 0.01 Spiophanes bornbyx 4.35 13.18 Spionidae 4.41 3.12 Sabellidae 3.30 0.25 Annelida tubes 5.47 7.53 824 Lumbrinereidae + 2.57 1.50 Nephtyidae 266 0.28 Other polychaetes 9.80 22.74 13.89 Byblis serrata 8.15 2.18 1.34 Other Ampeliscidae 1.55 2.44 1.01 Unciola sp. 1.50 7.01 2.81 Other Corophiidae 5.36 0.36 0.19 Gammaridae 0.77 2.25 1.72 Leptocheirus pm- guis 4.38 3.25 1.58 Other Amphipoda 0.54 1.38 1.58 Crangon septemspi- nosa 2.16 1.75 Dichelopandalus leptocerus 1.90 0.35 Other crustaceans 1.82 1 34 Animal remains 41.57 15.53 1785 Sand 20 09 7.75 22.65 Other groups 0.83 2.46 1.86 No. fish examined 16 502 502 No empty stomachs 4 163 169 Mean weight per stomach ±SD (g) 0242±0.324 0.323±0.578 0.512±1.452 Mean length ±SD (cm) 28.2±8.2 29.2±9.2 29.7±11.7 Spiophanes bornbyx being the most important species identified. On Georges Bank S. bornbyx was three times more important as prey than in 17 FISHERY BULLETIN: VOL. 81. NO. 1 southern New England. The other major differ- ence between areas was in the quantity of other polychaetes, but a large percentage of this group was unidentified remains (12.93% in southern New England and 11.61% on Georges Bank). The diversity of polychaete prey was very similar in the two areas; 27 families of polychaetes in the stomach contents of fish from southern New Eng- land and 24 different families on Georges Bank. Eleven different genera of polychaetes were iden- tified in each area. Six of these were common to both regions, but only Spiophanes contributed >1% to the total stomach contents weight. Amphipods made up almost twice the percent- age of the weight of stomach contents in southern New England than on Georges Bank (18.87% vs. 10.23%). The same species were important in both areas (Table 2). There was, however, a slightly greater reliance on Unciolasp. and Lep- tocheirus pinguis in southern New England than on Georges Bank. The diversity of amphipod prey was greater on Georges Bank, 16 genera as opposed to 11 genera, although yellowtail floun- der from the two areas preyed on 9 of the same genera. Crustaceans such as C. septemspinosa and D. leptocerus played a minor role in the diet of yel- lowtail flounder as did all other arthropod groups except the amphipods. The only other category of stomach contents that differed substantially be- tween areas was the quantity of sand in the stom- achs. This might be related to the heavy preda- tion on «S. bombyx on Georges Bank, since this polychaete is reported to prefer a fine sand sub- strate (Light 1978). The percentage of empty stomachs was virtu- ally the same in southern New England and on Georges Bank, but was less in the Middle Atlan- tic (Table 2). The mean weight per stomach in- creased from the Middle Atlantic to Georges Bank and the mean fish length also increased from south to north(Table2). This size difference did not counterbalance the increase in stomach content weight. The mean weight of stomach con- tents ranged from 0.12% in the Middle Atlantic to 0.14% in southern New England and 0.21% body weight on Georges Bank. Yearly, Seasonal, and Diurnal Variation Data were collected over a 4-yr period in both the spring and autumn and throughout the day- night cycle. It is, therefore, possible to examine the influence of the time of capture on the com- position of the diet as well as on changes in the absolute quantity of prey in the stomachs. On a year-to-year basis, polychaete worms were always the most important prey, between 36 and 44% of the diet, followed by amphipod crusta- ceans, 10 to 33% of the diet. Within these two taxa the actual percentage composition of the various groups fluctuated, but no systematic changes in diet were discernible. Within the Polychaeta, for example, S. bombyx made up between 2 and 12% of the diet from 1973 to 1974 and ranged from 9 to 11% between 1975 and 1976, respectively. At the family level, Spionidae, the range increased from 2 to 16% for the first 2 yr and 9 to 18% for the lat- ter 2 yr. When spionids were most important, there was also a very large percentage of sand in the stomachs, 20 and 27% for 1974 and 1976, re- spectively, which probably relates to predation on these particular polychaetes. Among the am- phipods, Unciola sp. showed the greatest fluctua- tion, ranging from 16% of the diet in 1973 to 1% in 1975 but increasing to just under 5% in 1976. The mean weight of prey showed an increase from 1973 to 1976, but when this was corrected for fish size, there was no pattern evident in these changes. The slightly larger mean fish lengths occurring in 1975 and 1976 counterbalanced the increase in the mean weight of stomach contents. The percentage of empty stomachs also showed no consistent yearly change, fluctuating around the overall mean value of 33%. Species composition of the diet showed no dras- tic shift between spring and autumn. Polychaetes were more important in the spring (49%) than in the autumn (35%), and the same was true for am- phipods, 19% vs. 13%. Both of the changes may, however, simply reflect the higher percentages of unidentified animal remains and sand in the fish stomachs collected in the autumn. In all years, except 1976, stomachs collected on spring cruises contained a greater mean weight of prey than stomachs from fish collected in the autumn. Although the mean length offish in the spring was only slightly larger (30.0 cm vs. 28.8 cm). The 4-yr mean weight of prey in the stom- achs was 0.505 g (0.20% body weight) for the spring and 0.298 g (0.14% body weight) in the autumn. The percentage of empty stomachs was also lower in the spring than the autumn; 22.7% of the 574 stomachs examined from spring cruises versus 46.3% of 447 stomachs examined from autumn cruises. An examination of the data for a composite 24-h day revealed a diurnal feeding pattern. 18 LANGTON: FOOD HABITS OF YELLOWTAIL FLOUNDER Although there was a certain degree of hour-to- hour variability, a peak in the weight of stomach contents occurred during the afternoon-early eve- ning period (Fig. 1). In Figure 1 the day has been divided into four periods— dawn (0300-0800 h), day (0900-1400 h), dusk (1500-2000 h), and night (2100-0200 h)— which accounts for seasonally variable day length in the dawn and dusk period. Despite a seasonal change in day length, the com- position of the diet also changed over a 24-h peri- od. Polychaetes were less important prey during the night than during any of the other three time periods. They dropped from values ranging from 41-47% to 24% as a percentage of the weight of stomach contents. Conversely, crustaceans, am- phipods in particular, were more important at night (values ranging from 15 to 23% vs. 34%). Unidentifiable animal remains also accounted for their smallest percentage of the diet (13.0%) in the dusk period when the fish stomachs were fullest. The greatest percentage of empty yellow- tail flounder stomachs was found during the night (46%) and the smallest ( 19%) during the day with intermediate levels occurring at dawn (34%) and dusk (26%). To evaluate whether the diurnal feeding pat- tern shown in Figure 1 is statistically significant, an analysis of variance, including time of day and seasonal factors, was conducted. The results of this analysis are given in Table 3 for trans- formed data using an inverse hyperbolic sine transformation (Y' = sin h" 1 i\fY)) to account for the extreme skewness of the data (i.e., a large number of empty and almost empty stomachs) (see Bartlett 1947). Both time of day and season are significant factors in determining the weight of stomach contents for yellowtail flounder. This analysis confirms that there are statistically sig- nificant differences in stomach content weight over a 24-h period. These results are, however, influenced by the level of interaction between time of day and season, such that it is not clear which of these two factors is the most important Table 3.— Analysis of variance of the weight of stomach contents for yellowtail flounder, ex- pressed as percent body weight, for time of day and season. See text for details. Source df Sum of squares F Time of day Season Interaction Error 3 1 3 1,011 0.03376 0.02464 0.00894 0.78005 1459" 31 93* 386* 'Significant, P = 0.05 "5- 0.30 cu 5 O CD 3 ^Z cd -t-» CO -C CD he S 3 3 cd c-3, 3 C D> O CD- £ CO. » CO 3 « C Ol CO o 3 « C o> CO £ en CO PS I — .O LI O ai (1) C u a.— LU CD >.T3 III 3 CO £ o 5 « c I ai •5£ o Ol c c C0.3> 3 «£ o o 2 ai <= ~ £ c CO ™ » c£-p £ c E Q. CD ~ E~2 oi> E c eg p O' o CO CO Ol TT CO LO Tf in in in cd N N O) CM CM Tf LO CO Tf Tf CM CM CM CM CO CO Tf 1- — ^ CO C m co co co O 00 CO -^ CM Tf ° 2 CO o LO CO CD Tf CO CO Tf LO LO Tf o co m o CM co co yf Tf m CD I s - r» r- CO I s - Ol I s - Oi o> O CM r- O y- CM CM CM CO CO •~ — 5 C-O CM y- 0> CO ? 5" S CM CO CO CO CO Ol CO O i- CO o 00 CD T- CM TT o CD CM co r- o Ol co in co o m CO N N MO O ■- >- CM r- t- CM CM CM CM C- CO CO CM ■** co co o co in co -^ in in lo co *-r CM CM CO CO CO f CD f- in CM Ol LO CM CO Tf CO CO CM CO CO co in CM CM CO CO Ol Ol Tf CM o CO LO CM O CM Ol LO CM TT 01 CD CO CD O O CM CM CO Tf Tf LO O CO CM CO OWSO"- CM CM CM CO CO cm in co cm cd CM CM CO CO CO d d d d d >-OCJ)i-f n coir in m ■a o I s - w O) 00 Tt Tt d d ■<* ■<»■ CD o a CO CD O o o o o CO LO CO O 0) CO CO CO O CO o p CO Ol CM p >- CO m CD CM O CD o co co in id co s CO CM o CM CO CD Cli LO CM CM CO CO CO CM CM O CO Tf CM CM CO I-- CM LO Oi CM Ol Ol CO CO CO CO CM CO Tf Tf lh co Tf r- co cm co -- T- T- CM CO Ol CM o o in o r- co CM CO CO CD Ol CO CD CO i- CO o o o o o o o o o CM CM CM CM CM CM CO CO CO CO CO Tf Tf Tf Tf Tf Tf m cd co oi o c\j in in oi r-- co cm t- | co 1- in CO 5 Tf in t- C0 CD in m in CO LO LO CD CO f-- CM ai o in r— co o i- T" CO CO CO CO Ol 1— co i in ^f in in CO CO CO CO CO Ol CD CO CO CO CM CO f- Ol CO y- 00 O o o o o o o o o o O o O O CM CM CM cm co Tf Tf in o cm" in in CM CO C\J CD cm 1 eg co co CO CD CO in" P r^ co ■a- un CO LO o CD o. LO Ol CD 1^ CM t^ CO 1— t^- y- CO Ol Ol s ^ CO O T- CM t- Tf CO in co CO LO 1 CO Tl- m in ai in O LO O o Tf m oo o d odd o o o o O o O o O o o CM CM CO co co Tf Tf m cm in in O O) CO Tf CO C\J i- CM CO CO o d d (6 d CO O) co tt Tf co co r- cm co CO CC CO o Ol Tf in i- co o in in d d o in CO o 35 CD O m o LO o CO d CO O Ol Ol Ol o o> CO CO O CO Ol o CT> T-; in ■- f i- CM CO 5" n in co cm ^ in ^ 9 9 t CO CO Tf CD O CO Cli I s Tf in O CM LO CM LO LO CM CO LO CM CM CO CD CO CM r^ CO o cq CO in CD Ol Tf ai CM CM r- o O CM O <- CO o co in -^ co lo ^r oi in co o cd in CO CD TT CO Tf t- CO Tf CM CM CM CM CM CO CO CO •» i T in T in in co co cd r- r- r— f- r- r-- o CO Ol o O y- Tf in CD co m in tt co co in co r- CM J- >- CO CO Ol CO in P CD -- CM i- o CM I s - CM Ol CM CM Tf Ol CM in CM m Ol CM CO CM CO CO Tf Tf CM CD Tf CD CM CO O O - tt in in co I s - o o in co un CO m CM 1 CO LO o LO CD O CD CM CD CD O 00 CO CM 1- CM CO LO CM CO CD Tj- CM CO ^t CO CO Tf CD r- Tf CM CO Tf lO o o CO LO CO i-ooioo TT in CD N CO Ol O 1- CM CO ^r LO CD h~ CO O) O r~ CM CM CM CM CO Tf in CD N CO CM CM CM CM CM CM Ol O CM CO CO CM CO CO CO Tf LO CO CO CO CO CO Tt CO CD N N t- in co co Tf in lo *- *- co «- h- co O CO CM CO in LO CO LO in in in in in m in co in in in in in CO o> a> Ol o> Ol Ol Ol o> O CT> Ol Ol Ol Ol Ol Ol Ol Ol Ol Ol Ol Ol Ol O) Ol Ol Ol Ol Ol Ol Ol Ol Ol Ol oi oi oi oi oi 't in cbsco Ol o >- CM CO ^3- LO CD r~ cd cr> O r- CM CM CM CM co tj in co s co CM CM CM CM CM CM Ol o CM CO CO CM CO CO CO tt m CO CO CO CO CO Tf CO cd r-- r*- y- in co co Tf in in o o o o o o o o O o O O O O o o O O O o o o o o o o o o O O o o o o o o o o o o ^r in co s co Ol O T- CM CO •t LO CD r^ cd O r- CM CM CM CM co tt in co s co CM CM CM CM CM CM Ol o CM CO CO CM CO CO CO Tf in coo CO CO CO Tf CO CD N N t- m co co Tf lo in 3) and rays develop. Rays develop in the pectoral and pelvic fins. Distinctive "muscle bands," which overlap the base of the caudal fin rays de- velop in transforming and juvenile specimens (Figs. 2C, 3). The adipose fin develops after trans- formation (Fig. 3). 29 FISHERY BULLETIN: VOL. 81. NO. 1 Osteology c - — - 3 > C sj ■D a> c E _o o '-£ .5 > CO c CO . 13 H CD CD E 2 T3 3 <° T3 C rt *j m c a) cu CO £ CD to 0) 3 > CO cfi CO > 0) be c c § CO 0) > be c/) 0) 3 e« . to > -s CD 3 0) CD en 03 CO •~ 0) > > •*- .c -a ts CO •~ c pare c o X CD LL so CD CD CO ■° 3 _o E "> "Si 3 w £ s S3 ^ «** o CO 0) CO 'S 11) CD 3 Hi CO > 3 > CO CO > '"— i T3 c o C cfl X 0) 0) "oj oa > c3 a CD £ ■° 3 o CO is C .o +j >-, o ft o i~ ft >, -a o E m CD I — oj w j 09 < H cocom"<*r--h--^.intpm CO CO CO CM CM i- cocomi-ocor-co co o cq p ^r -^ in •r- t- CM CM LO if" T^ -H -H -H -H -H -H -H O CM CO CM O ^ CM r*- m r^ io in *- O) OO CO -<* CD O o o i- 1 r 1 d ^ -H -H -H -H -H -H co n n ^r s o OO)0)O1CT>C7>O")CJ)O)O)O)O)CJ> o co ^ COO) ^tco^- mo cb ow in to ' r ~ cococmcoco co-^tco i- q co coco aiO) 7 cb^rt^r^^ cm co ^ i-h-T-i-r- t-CMCO CO ^T if) ~ oS cm a> u m co in 00 ' ' ' i t-ttO ip'-cOt-cocO'-Nmcoa) r-' C\i (N (O ^' CO i- O i- t- r-' -H-H+l-H+l+l-H-H+i-H-H MOO)TtlD(DOONOlO NCNi^cicb'-OCDCDO)Tf N CM CNJ CM CM CO i- in S in I I CO N CO t CM CO ^T coscpcoscomi T CO t ID <£> t- *? IT) T ty ^ t-IOt-CVJi-t-i-CD^ I I I I I IT) TJ CO t- CD r- r- ] co i-cocMco-sroo -H-H-H-H-H-H-H-H oiownit onoi Loaicbc3cDcbc6-^ h-. t- CM CM CM C\J CVJCMCMCMCMCMCMCMCM — O ^ CO CD ^ co cm en oi CMCOO ' CO "^

^ f~- . a> co ^ciSCMCO-^4 T CDS -- '^r 't CO 'CD O) CO I I I I CO o CO i- O co o m CMCOi-r-.-CO.-0 -H-H-H+t-H+l-H-H cosooicococtico S CD S CO CO CO _1 CO > CO > CO c O) c CD CO c £ - I c o>£ oi £ "O I co CO — c £ ^ I _ — CO o>£ £| 2 O c 0)~ - E we? a" CD t - •D « o <3> Q. CO © 0) _ CJ o <0 ™ ;• c C m O O O §3-D J =5 23££333 OcOcDCDCDOQ.Q.000 COCOlLULUicOCIOcOCOCO 3 3 T3 T3 CD CD a. o. TO TO T3 -D 3 3 CO CO O O (Tables 3, 4; Figure 4) Although a few structures ossify in relatively small L. schmidti larvae, a number of skeletal elements do not calcify until the larvae trans- form into juveniles. Other portions of the skele- ton do not completely ossify until well into the juvenile stage. The general sequence of ossifica- tion is as follows: cleithrum; dentary and vomer- ine teeth; pharyngeal teeth; parasphenoid, den- tary, maxillary, vomer, premaxillary; most other bones of the cranium; certain elements of the caudal fin and, at or near transformation, axial skeleton; dorsal, anal, and paired fins; and gill rakers and secondary caudal fin rays. Teeth Teeth on the dentary and vomer begin to form in 8 mm larvae (Table 3), increasing in number as the larvae grow and doubling in number at transformation. Dentary teeth appear to increase in number with growth in transformed juveniles, but vomerine teeth remain relatively constant in number. In specimens about 12 mm long, one pharyngeal tooth develops on each plate (Table 4); after transformation the teeth then increase to three. Borodulina (1969) also reported three teeth on the pharyngeal plate in adult L. schmidti. Teeth on the glossohyal and palatine develop in larvae 16-18 mm long (Table 3). Glossohyal teeth disappear during transformation and are absent in juveniles and adults: hence, the common name "smoothtongue" for L. schmidti. A single palatine tooth is present during the larval stage; the num- ber of palatine teeth increases after transforma- tion (Table 3). Skull The dentary, maxillary, parasphenoid, and operculum begin to ossify in 14-15 mm larvae (Table 3). In larvae 18-21 mm long, the premaxil- lary, pre-, sub-, and interopercle bones, vomer, symplectic, branchiostegal rays, and urohyal begin ossifying in some larvae (Tables 3, 4). These structures, however, are not consistently calci- fied until larvae reach about 28 mm long (32 mm for the urohyal). Certain bones of the olfactory, orbital, otic, and oromandibular regions and much of the hyoid arch begin to ossify in some larvae 24-26 mm 30 DUNN: DEVELOPMENT AND DISTRIBUTION OF LEUROGLOSSUS SCHMIDTI Table 3.— Development of meristic and other structures in Leuroglossus schmidti larvae and juveniles. Mean data are given for specimens in the specified length range. Specimens between dashed lines are undergoing notochord flexion; those between solid lines are juveniles. SL (mm) 5.9-7.9 8.0-8.9 9.0-9.9 10.0-10.9 11.0-11 9 120-129 Sam- ple size Dorsal fin rays Anal fin rays Pectoral fin rays Pelvic fin rays Principal caudal rays Secondary caudal rays Upper Lower Upper Lower Neural spines Haemal spines Centra Abdom. Caudal Total 12 6 5 1 1 2 13.0-13.9 14.0-14.9 15.0-15.9 16.0-16.9 17.0-17.9 18.0-18.9 19.0-19.9 20.0-20.9 21.0-21.9 22.0-229 23.0-23.9 24.0-24.9 250-25.9 26 0-26.9 27.0-27.9 28.0-289 29 0-299 30 0-309 31.0-31.9 32.0-32.9 33.0-33.9 34 0-349 31.0-31 9 33.0-33.9 35.0-35.9 37.0-37.9 44.0-449 51.0-51 9 SL (mm) 5.9-7.9 8.0-8.9 9.0-9.9 10.0-10.9 11.0-11 9 12.0-12.9 22 2.0 4.3 4.5 0.3 0.3 10.0 9.0 20 2.0 20 1.8 0.4 0.4 4.0 3.6 1.0 1.2 0.4 5.0 4.5 1.0 1.3 0.3 80 7.2 1.2 18 02 6.7 60 0.3 03 67 6.0 0.7 1.0 07 7.5 6.8 3.0 3.0 0.8 100 9.0 3.0 3.3 0.5 08 100 9.0 3.3 48 08 2.0 10.0 90 3.0 4.0 0.5 10.0 9.0 2.0 1.0 10.0 9.0 5.0 5.0 2.0 2.0 10.0 9.0 4.0 5.0 1.0 0.3 03 05 1 1.0 10.0 10.0 10.0 10.0 100 100 13.0 13.0 12.5 11.0 11.0 12.0 9.0 7.0 7.5 90 6.0 9.0 8.0 9.0 9.0 9.0 9.0 9.0 100 10.0 10.0 10.0 10.0 10.0 9.0 9.0 9.0 9.0 9.0 9.0 13.0 13.0 14.0 15.0 15.0 15.0 14.0 12.0 13.5 15.0 13.0 15.0 490 51.0 495 490 49.0 50.0 230 24.0 23.5 240 23.0 24.0 27.0 28.0 27.0 27.0 27.0 27.0 230 24.0 23.5 24.0 23.0 24.0 Sample size Hypurals Epurals Uroneurals Branchios- tegal rays Gill rakers Teeth 0.7 04 2.0 0.7 0.4 1.0 0.3 '12.8 0.5 1.0 1.0 50.0 52.0 50.5 51.0 50.0 51.0 Upper Lower Total Dentary Glossohyal Vomer Palatine 13.0-13.9 14.0-14.9 15.0-159 16.0-16.9 17.0-17.9 0.3 2.0 3.0 2.0 5.0 3.7 2.0 1.7 1.0 1.0 0.7 3.0 1.3 0.3 180-189 19.0-19.9 20.0-20.9 21.0-21.9 220-22.9 23.0-23.9 24.0-24.9 25.0-259 26.0-269 27.0-27.9 280-28.9 29 0-29.9 30 0-30.9 31.0-31.9 32.0-329 33.0-33.9 34.0-349 31.0-31.9 33.0-33.9 35.0-35.9 37.0-37.9 44.0-449 51.0-51 9 1.8 7.0 1.4 2.8 3.0 1.2 2.3 5.3 4.8 5.0 6.0 7.0 7.0 0.5 2.0 04 1.0 1.3 1.6 1.3 1.3 1.5 20 2.0 20 20 2.0 2.0 0.7 0.7 1.0 20 0.4 1.2 2.0 12 1.7 1.3 2.0 2.0 2.0 2.0 2.0 2.0 2.0 1.0 2.8 38 4.0 4.3 10.0 10.0 9.0 10.4 10.8 11.8 12.0 11.0 13.0 11.8 14.0 13.0 14.0 12.0 11.0 1.3 1.3 2.5 3.0 3.0 2.8 23 24 2.3 3.0 2.8 28 3.0 2.5 3.0 3.0 3.0 2.5 2.0 3.0 30 3.0 3.0 3.0 3.0 30 3.0 30 3.0 30 3.0 3.0 3.0 3.0 7.0 7.0 7.0 7.0 7.0 7.0 0.2 7 08 2.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 2.0 2.0 7.0 150 22.0 240 20 6.0 3.0 2.0 20 7.0 14.0 21.0 28.0 9.0 4.0 2.0 2.0 7.0 14.5 21.5 27.5 6.5 4.0 1.0 2.0 2.0 7.0 14.0 21.0 34.0 8.0 7.0 2.0 2.0 7.0 17.0 24.0 32.0 10.0 4.0 1.0 2.0 2.0 9.0 17.0 260 46.0 9.0 5.0 'Haemal spines not fully differentiated on one specimen with 50 centra ossifying. It was therefore not possible to determine the number of precaudal vertebrae in that specimen 31 FISHERY BULLETIN: VOL. 81, NO. 1 c — a> en c o> .22 c c 09 E £ o o c c C3 e« o _o "5 *5 c <£ • ~ en OB ° c .c — u TJ IE Ji 5 C/j *j « n! <£ >>~ .Q 0) O) w en .5 =3 .2 .S *e .* 22 «t- ._ •— ««-. 02 ■- en en e g !s >> .■tt *3 J to H c _: o ■C o -« c ^ O) 00 E = _o; &« V- to s « o -a en C a> •- ~ cv c c """* — T3 O C m a! cd QJ .. > ro f*> O cr> n pj q tn pi p] □ 09 o o PI p] c * E ■S 8 c a. O en '-£ : aS S o a! 5 - 0) 3 £ o" * CO .5 ^.2 > » gjs < c H .22 3 K X c> I- c r c- c- c* c- Olc o s! E n J m id- gCL(/)CLC0C0Q.a. O u. D y co a. 0- UJ UJ = £ £ ci a 5 fl « " s « Q- K « 3 a O lu m co .|Er' i P K i y E 0> « c m E of J £ £ a O 5q.lu05<20 mm. In a 23.1 mm speci- men (Fig. 4e) all seven hypural bones, 10+9 prin- cipal rays, and 2+2 secondary rays are ossifying. Both pairs of uroneurals are ossified. A single unossified epural and a number of unossified neural and haemal spines can be observed. The 34 HS HY, HY 2 HY3 HY 4 HY 2 HY 3 HY 2 H Y 3 Figure 4.— Development of the caudal fin of Leuroglossus schmidti: A, 6.4 mm SL; B, 11.5 mm SL; C, 15.3 mm SL; D, 18.4 mm SL; E, 23.1 mm SL; F, 30.6 mm SL; G, 35.1 mm SL; H, 51.6 mm SL. Ossified elements are stippled. NC = noto- chord; HY = hypurals; EP = epural; HS = haemal spine; UN = uroneural; NS = neural spine; U = ural centra; PU = preural centra; SNP = specialized neural process; PCR = principal caudal rays; SCR = secondary caudal rays. DUNN: DEVELOPMENT AND DISTRIBUTION OF LEUROGLOSSUS SCHMIDTI £ NS UN, EP U 7 N 2, HS HY, NS Ui U , Ni U 2 EP UN 2 UN, EP SCR- UN1 EP UN 2 hy 7 SCR 35 FISHERY BULLETIN: VOL. 81, NO. 1 adult complement of 10+9 principal caudal rays is consistently ossified in 29-30 mm larvae, but secondary caudal rays are not completely de- veloped until after transformation. In a 30.6 mm specimen (Fig. 4f), ural centra 1 and 2 are ossify- ing as are two neural spines associated with preural centra 1 and 2 and haemal spines associ- ated with preural centra 1-4. The bases of pre- ural centra 1-8 are beginning to ossify. In a 35.1 mm transformed juvenile, all centra of the caudal complex, preural and ural, are ossi- fied as is the specialized neural process (Hollister 1936) dorsad to Ui (Fig. 4g). Ural centra 1 and 2 are still separate. Neural and haemal spines and principal and secondary caudal rays are fully ossified. The hypural bones are separate and autogenous in this specimen. A 51.6 mm juvenile has ural centra 1 and 2 fused into a single uro- style (Fig. 4h). The single epural is beginning to ossify and hypural bones 1-3 are fused together at their bases and ankylosed to the urostyle. Hypural bones 4-7 are still autogenous in this specimen. In this specimen, the 10 superior prin- cipal caudal rays are distributed as follows: hypural 7, one ray; hypural 6, two rays; hypural 5, four rays; hypural 4, three rays. The inferior nine principal rays included one ray on hypural 3, five rays on hypural 2, two rays on hypural 1, and one ray on the posteriormost haemal spine. During the development of the caudal fin, the ventralmost principal caudal ray is associated with hypural 1; it apparently is displaced to ar- ticulate with the ultimate haemal spine in some, but not all, juvenile specimens. Dorsal and anal fin rays ossify rapidly during transformation as do pectoral and pelvic fin rays (Table 3). It was not possible to follow the se- quence of ossification of individual rays in these fins. The adult complement of dorsal, anal, and pelvic fin rays is present on the smallest trans- formed specimen. The adult complement of eight to nine pectoral rays, however, is not present in all transformed specimens examined. Scales are not present in a 55.0 mm juvenile, but they are reported to be deciduous (Hart 1973), and hence, may have been lost during capture. COMMENTS ON THE VALIDITY OF THE GENUS LEUROGLOSSUS GILBERT AND THE SPECIFIC STATUS AND NAME OF L. SCHMIDTI The results of this study offer support for the validity of the genus Leuroglossus Gilbert which has been questioned. Gilbert (1890) erected the genus Leuroglossus in the family Argentinidae for specimens from the Gulf of California that he described as Leuroglossus stilbius. Chapman (1943) reviewed Leuroglossus Gilbert, removed it from Argentinidae, and placed it in Bathylagi- dae. Cohen (1964) synonymized Leuroglossus Gil- bert with Bathylagus Giinther. Borodulina (1968) stated Leuroglossus lacked an orbitosphenoid bone, although Cohen (1964) said it was present in Leuroglossus and used its presence as a gener- ic character for Bathylagus. Borodulina (1969) later described the osteology of L. schmidti and compared it with B. paeificus (based on Chap- man 1943). She stated that Leuroglossus, in con- trast to Bathylagus, lacked an orbitosphenoid, possessed teeth on the palatine, had three den- ticles on the last pharyngobranchial, and pos- sessed antorbitals. Ahlstrom (1969) described differences in the movements of oil globules be- tween Bathylagus and Leuroglossus eggs which, with the lack of an orbitosphenoid in Leuroglos- sus (as reported by Borodulina 1968), he felt, lent additional support to the validity of Leuroglossus as a genus distinct from Bathylagus. My samples of Leuroglossus lacked an orbito- sphenoid, whereas those cleared and stained spec- imens I examined of Bathylagus did possess an orbitosphenoid. Specimens of B. paeificus, B. ochotensis, and B. milleri I examined lacked teeth on the pharyngobranchials, whereas Leuro- glossus possesssed three teeth on the fourth pharyngobranchial. Based in part on the lack of an orbitosphenoid in Leuroglossus, the presence of one in Bathyla- gus and the differences in the movements of oil globules in the eggs of these two genera, as re- ported by Ahlstrom (1969), I follow Borodulina (1969) and Ahlstrom (1969) in considering the two genera distinct. The number of valid genera in the Bathylagidae and analysis of their rela- tionships, however, await further study of the entire family. This study provides additional evidence to rec- ognize L. schmidti as a species distinct from L. stilbius. Rass (1955) described a northern sub- species, L. stilbius schmidti, from the Kurile- Kamchatka Trench, based on morphometric measurements which differed from measure- ments described by Gilbert (1890). Cohen (1956) synonymized L. s. schmidti with L. stilbius, as- serting that the proportions used by Rass to de- scribe L. s. schmidti were size dependent. Boro- dulina (1968) pointed out that L. stilbius had 36 DUNN: DEVELOPMENT AND DISTRIBUTION OF LEUROGLOSSUS SCHMIDTI 39-42 vertebrae, whereas L. schmidti possessed 49-51 vertebrae. Borodulina (1968) considered L. schmidti to be a subspecies of L. stilbius, although she suggested that L. schmidti might subse- quently be recognized as a separate species, a suggestion that she did not pursue because of insufficient material. She also considered L. uro- tranus (Bussing 1965), described from the Peru- Chile Trench, to be another subspecies of L. stil- bius. Ahlstrom (1968) noted the differences in vertebral counts in L. stilbius and L. schmidti. Subsequently Ahlstrom (1969) pointed out the differences in egg size, pattern in migration of oil globules during embryonic development, larval pigment, and body proportions between L. stil- bius and L. schmidti which, along with differ- ences in vertebral counts, he felt, enabled recog- nition of L. schmidti as a distinct species. Peden (1981) examined vertebral numbers in Leuroglossus from samples collected from Mexi- co to the Aleutian Islands and westward to Japan. He noted that samples of Leuroglossus from Brit- ish Columbia waters had an average of 8.5 more vertebrae than those samples collected off Ore- gon. He therefore recognized L. schmidti as distinct from L. stilbius stilbius. As presently known, the geographical ranges of the two spe- cies do not overlap, as discussed below. Based on the differences in vertebral counts in the two nominal species reported by Borodulina (1968, 1969) and Peden (1981), the evidence presented by Ahlstrom (1968, 1969), and the results of this study, I consider L. schmidti specifically distinct from L. stilbius. The valid name of the northern smoothtongue is considered here to be L. schmidti, rather than Therobromus callorhini or Leuroglossus callo- rhini. Therobromus callorhini was described by Lucas (in Jordan and Gilbert 1899) from bones extracted from fur seal stomachs collected in the Bering Sea. He noted that the specimens had 26 precaudal and 23 caudal vertebrae and placed the species in Osmeridae. Chapman (1941) showed that T. callorhini was notanosmerid and later (Chapman 1943) he suggested that T. callorhini (emended to callorhinus) was most likely identical with either Bathylagus pacificus or B. alascanus (= B. milleri). Cohen (1964) syn- onymized Therobromus Lucas with Bathylagus Gunther. Ahlstrom (1968, 1969) suggested that the correct name of L. schmidti was Leuroglossus callorhini, but did not formally propose such a synonomy. If the two names do refer to the same species, then L. schmidti is a junior synonym of T. cal- lorhini. The type material of Therobromus cal- lorhini Lucas apparently no longer exists (ac- cording to D. M. Cohen 7 ). However, as the name T. callorhini has apparently not been used as a senior synonym in more than 50 yr (Chapman 1943), T. callorhini constitutes a nomen oblitum according to the International Code of Zoological Nomenclature (Stoll et al. 1964). Hence, I con- sider the valid name of the northern smooth- tongue to be Leuroglossus schmidti. OCCURRENCE OF EGGS AND LARVAE OF LEUROGLOSSUS SCHMIDTI (Figure 5) Eggs and larvae of L. schmidti are broadly distributed in near-coastal waters from about southern Vancouver Island, British Columbia, to the central Bering Sea. In midocean they are apparently distributed as far south as lat. 46°N, since eggs and larvae of L. schmidti were col- lected in 1951 and 1955 from about lat. 46°-57°N and long. 149°-179°W (Ahlstrom 1969; Moser 8 ). They have apparently not been found in coastal waters off Oregon, however, as the relatively few specimens of Leuroglossus larvae, collected off Oregon by Oregon State University, consisted exclusively of L. stilbius stilbius (Richardson 1973; Washington 9 ). The results of an ichthyo- plankton survey conducted in October-November 1971 from off Washington (lat. 46°45'N) to Dixon Entrance, British Columbia (lat. 54°30'N), were reported by Naplin et al. (footnote 4). Eggs of L. schmidti were found only north of lat. 53°N off Queen Charlotte Islands, whereas only L. stilbius stilbius eggs were collected south of lat. 51°N off Vancouver Island and coastal Washington. The few Leuroglossus larvae collected during this cruise were all L. schmidti, and they were taken only north of lat. 54°N. Possibly the eggs identi- fied as L. stilbius stilbius off Vancouver Island 7 D. M. Cohen, National Systematics Laboratory. National Marine Fisheries Service, National Museum of Natural His- tory Washington, D.C. (present address: NWAFC, NMFS, NOAA, 2725 Montlake Blvd. East, Seattle, WA 98112), pers. commun. July 1980. 8 H. G. Moser, Southwest Fisheries Center La Jolla Labora- tory, National Marine Fisheries Service, NOAA, La Jolla, CA 92038, pers. commun. March 1980. 9 B. B. Washington, School of Oceanography, Oregon State University, Corvallis, Oreg. (present address: Gulf Coast Re- search Laboratory, Ocean Springs, MI 39564), pers. commun. November 1980. 37 FISHERY BULLETIN: VOL. 81. NO. 1 65 00N ^\£ Naplin et i Larvae Naplin el al Eggs Matson and Wing 11978) Larvae □ Kendall et al Larvae O Kendall et al Eggs Waldron 119811 Larvae 1 I Moser Eggs ^ Moser Larvae # Ahlslrom 119691 Eggs 60 00N - 55 00N 50 00N ▲ • 45 OON 175 00E 175 OOW 165 00W 155 00W 1 45 OOW 135 OOW 125 OOW FIGURE 5. — General areas where eggs and larvae of Leuroglossus schmidti have been reported. Key: Naplin et al. (text footnote 4); Matson and Wing (1978); Kendall et al. (text footnote 3); Waldron [1981]; Moser (text footnote 8); Ahlstrom (1969). couid have been transported northward from more southerly spawning areas. During this cruise, surface geostrophic currents indicated a 6 cm/s northward flow offshore from about lat. 47°N to 51 °N (Naplin et al. footnote 4). Leuroglossus schmidti larvae were the third most abundant fish larvae collected in plankton samples from coastal waters of southeastern Alaska (lat. 56°50'-59°28'N, long. 133°10'-135° 23'W) in April-November 1972 (Mattson and Wing 1978). This species accounted for 4.5% of the total catch of fish larvae; abundance was high from May to August, peaking in June and July. Plankton sampling in Kodiak Island shelf waters from November 1977 through March 1979 re- vealed that eggs of L. schmidti were found prin- cipally at the shelf break (water depth >200 m); abundance was greatest in the fall, but eggs were found in small numbers in summer and winter (Kendall et al. footnote 3). Larvae were also most abundant over the shelf break in the fall, but sea- sonal abundance was not determined. Waldron [1981] summarized available distribution data on larvae and juveniles of L. schmidti occurring in the eastern Bering Sea from 1955 to 1978. Based on plankton sampling conducted by the United States, U.S.S.R., and Japan (primarily during summer) utilizing a variety of sampling devices, larvae identified as L. schmidti were most frequently reported over the shelf break. ACKNOWLEDGMENTS A number of people at N WAFC assisted in this study. Arthur W. Kendall, Jr., engaged in help- ful discussions with the author, constructively reviewed the manuscript, and assisted in numer- ous other ways. Beverly Vinter illustrated the larvae and caudal fin development and shared her knowledge of larval bathylagids. Ann C. Matarese reviewed the manuscript and made many useful comments. Bernie Goiney and Jay Clark rendered technical assistance. Eleanor S. Uhlinger efficiently and rapidly processed sev- eral drafts of the manuscript. Bruce W. Wing, NWAFC, Auke Bay, Alaska, loaned larvae of L. schmidti. H. Geoffrey Moser, SWFC, La Jolla, Calif., kindly made available unpublished data on the occurrence of L. schmidti and provided unpublished illustrations of this species; he also made available his collection of radiographs and cleared and stained bathylagid specimens. T. W. Pietsch, University of Wash- ington (UW), Seattle, loaned bathylagids. Alex 38 DUNN: DEVELOPMENT AND DISTRIBUTION OF LEUROGLOSSUS SCHMIDTI Peden, British Columbia Provincial Museum, Victoria, kindly made available unpublished data on vertebral counts of Leuroglossus spp. Kevin M. Howe, UW, shared his knowledge of fishes, radiographed specimens, made literature available, and reviewed the manuscript. Daniel M. Cohen, National Systematics Laboratory, NMFS, Washington, D.C. (present address: NWAFC, Seattle), imparted generously his knowledge of bathylagids, provided unpublished data on L. schmidti, and made helpful comments on the manuscript. LITERATURE CITED Ahlstrom, E. H. 1965. Kinds and abundance of fishes in the California Current region based on egg and larval surveys. Calif. Coop. Oceanic Fish. Invest. Rep. 10:31-52. 1968. What might be gained from an oceanwide survey of fish eggs and larvae in various seasons. Calif. Coop. Oceanic Fish. Invest. Rep. 12:64-67. 1969. Remarkable movements of oil globules in eggs of bathylagid smelts during embryonic development. J. Mar. Biol. Assoc. India 11:206-217. 1972. Distributional atlas of fish larvae in the California Current Region: six common mesopelagic fishes — Vinci- guerria lucetia, Triphoturus mexicanus, Stenobrachius leucopsarus, Leuroglossus stilbius, Bathylagus wesethi, and Bathylagus ochotensis, 1955 through 1960. Calif. Coop. Oceanic Fish. Invest. Atlas 17, 306 p. Ahlstrom, E. H., and H. G. Moser. 1976. Eggs and larvae of fishes and their role in system- atic investigations and in fisheries. Rev. Trav. Inst. Peches Marit. 40:379-398. BORODULINA, O. D. 1968. Taxonomy and distribution of the genus Leuroglos- sus (Bathylagidae, Pisces). Probl. Ichthyol. 8:1-10. 1969. Osteology of Leuroglossus stilbius schmidti Rass (Bathylagidae). Probl. Ichthyol. 9:309-320. Bussing, W. A. 1965. Studies of the midwater fishes of the Peru-Chile Trench. In G. A. Llano (editor), Biology of the Antarctic Seas II, p. 185-227. Antarct. Res. Ser. 5. Natl. Acad. Sci. Nat. Res. Counc. Publ. 1297. Chapman, W. M. 1941. The osteology and relationships of the osmerid fishes. J. Morphol. 69:279-301. 1943. The osteology and relationships of the bathypelagic fishes of the genus Bathylagus Giinther with notes on the systematic position of Leuroglossus stilbius Gilbert and Therobromus callorhinus Lucas. J. Wash. Acad. Sci. 33:147-160. Cohen, D. M. 1956. The synonymy and distribution of Leuroglossus stilbius Gilbert, a North Pacific bathypelagic fish. Stanford Ichthyol. Bull. 7:19-23. 1964. Suborder Argentinoidea. In Fishes of the West- ern North Atlantic, Part 4, p. 1-70. Mem. Sears Found. Mar. Res., Yale Univ. Dingerkus, G., and L. D. Uhler. 1977. Enzyme clearing of alcian blue stained whole small vertebrates for demonstration of cartilage. Stain Technol. 52:229-232. Dunn, J. R. 1983. The utility of developmental osteology in taxonomic and systematic studies of teleost larvae: a review. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 450. 19 p. Gilbert, C. H. 1890. A preliminary report on the fishes collected by the steamer Albatross on the Pacific coast of North America during the year 1889, with descriptions of twelve new genera and ninety-two new species. Proc. U.S. Natl. Mus. 13:49-126. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180, 740 p. Hollister, G. 1936. Caudal skeleton of Bermuda shallow water fishes. I. Order Isospondyli: Elopidae, Megalopidae, Albulidae, Clupeidae. Dussumieriidae, Engraulidae. Zoologica (N.Y.) 21:257-291. Jordan, D. S., and C. H. Gilbert. 1899. The fishes of Bering Sea. In D. S. Jordan, The fur seals and fur-seal islands of the North Pacific Ocean, Part 3:433-492. Gov. Print. Off., Wash.. D.C. Mattson, C. R., and B. L. Wing. 1978. Ichthyoplankton composition and plankton vol- umes from inland coastal waters of southeastern Alaska, April-November, 1972. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-723, 11 p. Monod. T. 1968. Le complexe urophore des poissons teleosteens. Mem. Inst. Fond. Afr. Noire 81, 705 p. Moser, H. G. [1981]. Morphological and functional aspects of marine fish larvae. In R. Lasker (editor), Marine fish larvae: Morphology, ecology, and relation to fisheries, p. 89-131. Univ. Wash. Press, Seattle. Norden, C. R. 1961. Comparative osteology of representative salmonid fishes, with particular reference to the grayling (Thy- mallus arcticus) and its phylogeny. J. Fish. Res. Board Can. 18:679-791. Peden, A. E. 1981. Recognition of Leuroglossus schmidti and L. stil- bius (Bathylagidae, Pisces) as distinct species in the North Pacific Ocean. Can. J. Zool. 59:2396-2398. Rass, T. S. 1955. Glovokovodnye ryby Kurilo-Kamchatskoi vradin. (Deep-sea fishes of the Kurile-Kamchatka Trench). [In Russ.] Akad. Nauk SSSR, Tr. Inst. Okeanol. 12: 328-339. (Transl. by Nat. Hist. Mus., Stanford Univ., Calif.). Richardson, S. L 1973. Abundance and distribution of larval fishes in wa- ters off Oregon, May-October 1969, with special empha- sis on the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 71:697-711. Russell, F. S. 1976. The eggs and planktonic stages of British marine fishes. Acad. Press, Inc., Lond., 524 p. Sanzo, L. 1931-1933. Salmonoide. Fauna Flora Golfo Napoli, Mongr. 38(4 parts), 1064 p. Schmidt, J. 1906. On the larval and post-larval development of the Argentines {Argentina silus [Ascan.] and Argentina 39 FISHERY BULLETIN: VOL. 81, NO. 1 sphyraena Linne) with some notes on Mallotus villosus adopted by the XV international congress of zoology. [0. F. Muller]. Medd. Komm. Havunders. (Fisk.) 2(4): Int. Trust Zool. Nomencl., Lond., 176 p. 1-20. Waldron, K. D. 1918. Argentinidae, Microstomidae, Opisthoproctidae. [1981.] Ichthyoplankton. In D. W. Hoodand J. A. Calder Mediterranean Odontostomidae. Rep. Dan. Oceanogr. (editors), The eastern Bering Sea Shelf: oceanography Exped. Mediterr. 2(A5):l-40. and resources. Vol. I. p. 471-493. U.S. Dep. Com- Stoll, N. R., R. P. Dollfus, J. Forest, N. D. Riley, C. W. mer.. Natl. Oceanic Atmos. Admin., Off. Mar. Pollut. Sabrosky, C. W. Wright, and R. V. Melville (editors). Assess. 1964. International code of zoological nomenclature 40 DELINEATION OF TILEFISH, LOPHOLATILUS CHAMAELEONTICEPS, STOCKS ALONG THE UNITED STATES EAST COAST AND IN THE GULF OF MEXICO S. J. Katz, 1 C. B. Grimes, 2 and K. W. Able 3 ABSTRACT Tilefish, Lopkolatilus chamaeleonticeps, are an important commercial species in the Mid-Atlantic Bight and the focus of developing fisheries in the South Atlantic Bight and the Gulf of Mexico. Attempts were made to delineate stocks over this range by analyzing for variation in morphology (28 meristic and morphometric characters) and electrophoretic migration of eye, liver, and muscle proteins. Morphological and electrophoretic data (liver isocitrate dehydrogenase and liver esterase) consistently supported a separate Mid-Atlantic Bight stock. Electrophoretic data suggested that South Atlantic Bight and Gulf of Mexico samples belonged to a separate, single stock. This was not consistently supported by the more variable morphometric characters. It was suggested that Mid- Atlantic Bight populations be treated as a separate stock and, as a working hypothesis, that South Atlantic and Gulf of Mexico populations be considered as a second stock. Tilefish, Lopholatilus chamaeleonticeps, are dis- tributed from southern Nova Scotia (Leim 1960; Markle et al. 1980) south to off Surinam, South America, (Wolf and Rathjen 1974) and through- out the Gulf of Mexico (Bigelow and Schroeder 1947; Hoese and Moore 1977) but exclusive of the Caribbean Sea (Dooley 1978). The tilefish is the basis for a valuable bottom longline fishery in the Mid-Atlantic Bight (Grimes et al. 1980), and this fishery is developing elsewhere along the east coast of the United States and in the Gulf of Mexico. This paper investigates tilefish popula- tions to determine if separate stocks can be iden- tified over this range. There are several reasons to suspect that dis- tinct stocks of tilefish may occur. Tilefish prob- ably have a restricted habitat. They are reported from rather narrow temperature ranges (9°- 14°C) at the edge of the continental shelf along the east coast (Goode 1884; Rathburn 1895; Bige- low and Schroeder 1953) and in the Gulf of Mexi- co (Nelson and Carpenter 1968; Wolf and Rath- jen 1974). Also, preliminary tagging studies (Grimes et al. in press) suggested that individual 'Ecology Graduate Program, Rutgers University, New Brunswick, NJ 08903. 2 Forestry and Wildlife Section and New Jersey Agricultural Experiment Station, Rutgers University, New Brunswick, NJ 08903. 3 Biological Sciences and Center for Coastal and Environ- mental Studies, Rutgers University, New Brunswick, NJ 08903. tilefish moved <2 km in over 1 yr. These obser- vations are supported by submersible observa- tions which suggest that tilefish are resident in temporally stable burrows of their own con- struction (Able et al. 1982). In the Mid-Atlan- tic Bight, tilefish are caught the year-round which also suggests that these may be resident populations. In addition, the prevailing current patterns, temperature regimes, and species dis- tribution patterns along the east coast suggest that important faunal boundaries may exist at Cape Hatteras and around the Florida peninsula (see Briggs 1974 for discussion). This study re- ports on morphological and electrophoretic char- acteristics of tilefish from the U.S. east coast and the Gulf of Mexico. The distribution of the char- acters were used to test the null hypothesis that there are no differences among these popula- tions. MATERIALS AND METHODS Tilefish samples were obtained from commer- cial fishermen or collected by hook and line on exploratory fishing cruises (National Marine Fisheries Service RV Oregon ID during 1978 and 1979 (Fig. 1) (Katz 1982). Information on physi- cal conditions at collection were unavailable, but temperature is known to be relatively constant throughout the range (see above). Fish were transported fresh, on ice, or frozen, depending on distance of collection from the laboratory. Manuscript accepted July 1982. FISHERY BULLETIN: VOL. 81. NO. 1. 1983. 41 FISHERY BULLETIN: VOL. 81. NO. 1 Figure 1.— Sample locations for tile- fish along the U.S. east coast and the Gulf of Mexico. Submarine canyons are identified in the inset. Electrophoresis Eye, liver, and muscle tissues were removed from individual fish and frozen as soon as pos- sible. Vertical starch gel electrophoresis was used to detect protein variation. Initially only tis- sues of fish from the most distant collection lo- calities (Hudson Canyon and off Texas) were screened for 28 enzymes to maximize the chance of finding polymorphic enzymes. Of the 28 en- zymes screened during the initial electrophore- sis, several were scorable; however, most ap- peared monomorphic (malate dehydrogenase, lactate dehydrogenase, xanthine dehydrogenase, creatin kinase, adenylate kinase, peptidase, alco- hol dehydrogenase, malic enzyme, 6-phosphoglu- conate dehydrogenase, and glyceraldehyde 3- phosphate dehydrogenase) and only two [liver isocitrate dehydrogenase (IDH) and liver ester- ase (EST)] were polymorphic. Liver tissues from all collections were then run for both IDH and EST with an amine citrate buffer (pH 6.0) (Clay- ton and Tretiak 1972) for 17 h at 140 V and 40°C, and allelic frequencies were determined for all populations. Allelic frequencies were compared with their Hardy-Weinberg expectations by a chi-square test (Spiess 1977). We evaluated dif- ferences between sample locations, by chi-square contingency tests of electromorph distribution between sample locations. This test does not as- sume Hardy-Weinberg equilibrium and com- pares n samples with k classes to determine whether the individual A" classes are in the same relative proportion throughout the n samples. Length (age)-related differences in genotype distribution were tested (chi-square) on the largest sample with a wide range of sizes ( n = 40, west side of Hudson Canyon). Fish were divided into two size classes (<550 mm fork length and >550 mm) based on the approximate size at sex- ual maturity. Morphology Seven meristic (number of dorsal fin spines and rays, anal fin spines and rays, pectoral fin rays, upper and lower gill rakers on the first arch) and 21 morphometric (fork, standard, total, pectoral fin, pelvic fin, upper jaw, snout, adipose flap, barbel, snout to vent, snout to anal origin, snout to dorsal origin, snout to incurrent nostril, lengths; orbit diameter, interorbital width, head width, height of first, second, and third dorsal fin spines, caudal peduncle depth, and suborbital depth) characters were counted or measured following Hubbs and Lagler (1967), 42 KATZ ET AL.: DELINEATION OF TILEFISH STOCKS with two exceptions: Barbel length was mea- sured from its posterior tip to the junction with the lower lip, and the suborbital depth was mea- sured from the lower margin of the infraorbitals to the junction of the articular and interopercu- lar bones. Morphometric characters were mea- sured to the nearest millimeter with dividers and a tape measure. These characters were chosen on the basis of a preliminary study of two specimens of tilefish by Bigelow and Schroeder (1947) and a systematic study of the Branchiostegidae by Dooley (1978). Morphological data was determined from fish of dissimilar lengths (Fig. 2), so we used analysis of covariance to remove the size effects as sug- gested by Atchley et al. (1976). A linear relation- ship to standard length (SL) was determined for most morphological characters with the excep- tion of adipose flap length where an additional coefficient of standard length squared was in- cluded in the model because of allometry. For the final size-corrected comparisons between sam- ple locations we used sample location least square means for each morphological character ( Barr et Gult of Mexico- Campeche Banks MALES Jl FEMALE ,ll s N 10 -40 -20 Gulf of Mexico- off Texas J. L 15 -20 Gulf of Mexico- off West Florida .dk k 9 -40 -20 South Atlantic Bight- off South Carolina .iL ,h 10 -40 ^20 Mid-Atlantic Bight- west side of Hudson Canyon J 33 1-6O -40 -20 Mid-Atlantic Bight- i- east of Hudson Canyon ^_| i 26 hLL 12 -20 Mid-Atlantic Bight- east of Block Canyon L 4, ii 54 -40 -20 Mid-Atlantic Bight - west of Veatch Canyon IX. " li. 24 -40 -20 — n J- _L 200 400 600 800 1000 200 400 600 800 1000 Standard Length (mm) Figure 2.— Length-frequency histograms of tilefish samples used to conduct the morphological analysis. See Figure 1 for approximate locations. al. 1976). Least square means are estimates of arithmetic means that would be predicted had samples with the same size composition been ob- tainable from each sampling location. We conducted analysis of covariance on each morphological character to test for differences between sampling locations. Sex, sample loca- tion, and all interactions were initially included in the covariance model, but all nonsignificant (P<0.01) interactions were removed from the final model. The difference between sample loca- tion least square means for each morphological character for each sex was tested by comparison with the west Hudson Canyon sample using a t test. Significant differences were determined conservatively, using a high significance level (P<0.001), because the possibility of finding dif- ferences increases with the number of tests run. To further test for differences between sample locations we used discriminant function analysis (Jolicoeur 1959; Seal 1964) to determine the level of distinctness of fish from each location. The discriminant function was computed using both raw and size-corrected data for males and fe- males separately, because the analysis of covari- ance indicated sexual dimorphism. Only linearly related morphological characters were used in the raw discriminant function (Seal 1964). Size correction of morphological characters was ac- complished using the average value of standard length (SL) of all samples, and linear and quad- ratic regression coefficients (Bi, B 2 ) obtained from covariance analysis for each morphological character according to the following formula: corrected = raw - B x (SL-SL) - B 2 (SL-SL) 2 . This correction removed size effects by displac- ing each morphological observation towards the average, while allowing sample location and in- teraction effects to remain. RESULTS Electrophoretic Data The genetic basis of protein variation in tile- fish was implied from the electrophoretic band- ing patterns. IDH showed a dimeric pattern (heterozygote was three banded) with medium, slow, and fast bands. The rare fast form occurred only as a heterozygote in 10 out of 226 fish in the Mid- Atlantic Bight samples; therefore it has been left out of the statistical analysis. The EST 43 FISHERY BULLETIN: VOL. 81, NO. 1 locus exhibited a monomeric pattern (heterozy- gote was two banded) with fast and slow bands. Our interpretation of the dimeric and monomeric nature of these enzymes is consistent with past studies of their molecular structure (Manwell and Baker 1970). Distribution of EST and IDH electromorphs was not significantly different than expected from Hardy-Weinberg equilibri- um (Tables 1 , 2), which is additional support for a single-locus, two-allele genetic model. However, it should be noted that the chi-square test is not very sensitive at small sample sizes(<200)(Fair- bairn and Roff 1980). There was no significant difference in geno- type distribution as a function of length (age) for both enzymes (EST X 2 = 1.16, P>0.6, n = 20; IDH x 2 = 2.93, P>0.2, n = 20) in the sample ex- amined (west side of Hudson Canyon). There were distinct patterns of variation among the populations sampled (Fig. 3). Chi- square contingency tests revealed no significant differences in genotype distribution within the Mid- Atlantic Bight or southern sampling loca- tions (South Carolina, west Florida, Texas, and Campeche) (within Mid- Atlantic Bight EST X 2 5 , 3 = 8.77, 0.25 CL O ^ 0) ° II T3 en * D) o C - _i O — i j ■■3 P 6 e« o 7 §y CL) T, a - a V n CL a » 3 £5 e« en £ c 1.1 »- o T3 CD £ E en Q. C en _CD _C at en o. .2 g as «> £ +3 =■> ™ O <= en o. c (71 a; a «<-> £ .2 o 5 oj £ - en -r a) £ 5 C en m 3> .S c en O C >} CD CO o CL~ X c b s cfi -<-» 2 to ra .c 03 0; o c: o> ^ 3 N t* o CD CD .— CS Q_ CO • — *- is O to _ > C en C . r ai (/) TO r- c o -Com ndep 15 o o en i H- c CO en Q. W 'en E J >> CO oa <$ in < c H co 17) c I ns i o m o X r- X 0) o CD Q. ^ o E o t- 505" O CD CO 00 ID CO - CO in CM 00 CO 00 CO CO CD oS i- ,- 1- CO CD o >. c CO i O i co .*' 7: j= c H c 2 ■DCDuOlOoioCJ) ) w ,. CO i^ ^ '" — 10 -C C I c o >s c o o cm O CM co d o •xt o CO CD O CO a> LD o r~ LO o o en o ■J CO CO CD r- o CD CO b d ° I ra i. O U £, i 'OffliicfiOoiocn a>-^ 17) >< "- 1= CO cu X 3 C5 UqQJT3lli^lu urn „ q;^^o^o3o°o» „ ra £ - x - I = CD - > •5 V> ^. ^ ^ ~- c_ c^c^ cu .f5oJ5oi5C) o 45 FISHERY BULLETIN: VOL. 81. NO. 1 o O a 03 < to >> ccj o £ ■o 2 Q. C ■O <"i> mi? < _ co Q) I CD CO -D ' O 1> n Q. il t 1) i- CO *~ -So* 2 <= o) o — c CD 0) CL — CD CO CO CO t- o o ■* ^ m C\J C\J OJ CJ) CO CM 1 *r CO W b CD CO CO T CD O CO O) in O CM o og CO o ID CO *r CO CO CD r- CM r- CO CO CO CO o r^ C\J o CO CO CO CO O) o CO CO CO CD O r- o in O) o o CO t- CO t- d en t- o) CD CD o CM CO in o CO co o 1 O) in o The nature of the variation in the morphomet- ric characters examined varied between sexes and locations (Tables 3, 4). For several characters the least square mean values appeared to vary clinally. This was most evident for male adipose flap height and orbit diameter as seen in plots of raw data (Figs. 4, 5), female interorbital width and male head length. The values for other char- acters showed less consistent patterns and in some cases could be interpreted to suggest two distinct groups with the South Carolina samples most similar to Mid-Atlantic Bight groups (Tables 3, 4). This was most obvious for male pec- toral fin length and female pectoral fin length, caudal peduncle depth, and head length. Clinal variation was also suggested by the increasing number of significantly different morphological characters with increasing geographic distance between compared samples. The discriminant function analysis was con- ducted with both raw and size-corrected data. In each case the results were virtually identical with two exceptions (males, east Hudson Can- yon - 60% correct classification with size correct- ed vs. 23% raw data, and Campeche - 86% correct classification vs. 43% raw data). We believe neither of these significantly affects the overall interpretation of the results, and we report the raw data results here (Tables 6, 7). The discriminant function analysis suggests a similar clinal pattern of variation for both males and females (Tables 6, 7). There was generally low differentiation within the Mid-Atlantic Bight samples, and where misidentification oc- curred it was to other Mid-Atlantic Bight or South Carolina samples and infrequently to west Florida and the Gulf of Mexico off Texas. Gulf of Mexico samples naturally had higher percent- age correct classification (sample locations were more widely separated geographically) and in- correct classifications were usually to other Gulf of Mexico samples. Classifications for South Carolina samples had a high correct classifica- tion, and where misclassification occurred it was to both Mid-Atlantic Bight and Gulf of Mexico locations. DISCUSSION For purposes of interpreting the significance in allelic frequencies observed for IDH and EST we are assuming that the genetic variation ob- served is neutral (Allendorf and Phelps 1981; Ihssen et al. 1981). Thus, based on the patterns 46 KATZ ET AL.: DELINEATION OF TILEFISH STOCKS E E en c a. CO Q.d)CU<0 -. E o t- o * *- 03 c o>_ o O 5 u to o o y o X U- c ^ re •o CD CD Ol o CD CD o CD -g o ° o. re ^5 _ro o co--*;--;-';- From sample locations 5O5050 go O O O co < w < 5 5 Gulf of Mexico- Campeche Banks Gulf of Mexico- off Texas Gulf of Mexico— off west Florida South Atlantic Bight— off South Carolina Mid-Atlantic Bight- west of Hudson Canyon Mid-Atlantic Bight- east of Hudson Canyon Mid-Atlantic Bight- east of Block Canyon Mid-Atlantic Bight- west of Veatch Canyon N 43 14 43 _7 17 83 _6 18 12 70 V7_ 9 75 8 8 12. 9 55 18 9 9 22 4 15 23 23 19 16 26 2 5 7 5 54 27 41 12 4 4 17 50 13 24 Table 7.— Percent female tilefish classified to sample loca- tions by discriminant function analysis. To sample locations From sample locations Gulf of Mexico- Campeche Banks Gulf of Mexico- off Texas Gulf of Mexico- off west Florida South Atlantic Bight- off South Carolina Mid-Atlantic Bight- west of Hudson Canyon Mid-Atlantic Bight- east of Hudson Canyon Mid-Atlantic Bight- east of Block Canyon Mid-Atlantic Bight- west of Veatch Canyon c c c c c ~ re •c c , re . O c ~ re 1 re O c CD P t SZ c -C O r re 1 , "O rr rn rr O CD x c o> re 8b 1) — y - 1 CD O 5= CD m CO cu > >s 5 x CD 5 "> z: 3 < O CO O re ™ CO 'o b E ^_ re 30 TO TO < 0] < w J, re T3 a) < to J, re T3 0) < 0) 5 O O C3 CO > ^ :> 2> N 20 20 87 22 50 13 78 10 70 10 10 10 1°. 15 _9 10. 3 6 45 3 33 9 33 17 17 33 33 12 2 4 11 6 3 67 9 54 4 17 8 5 33 33 24 Mexico populations. In the discriminant function analysis South Carolina samples of both sexes classified correctly a high percentage of the time but misclassification occurred to both Mid-At- lantic Bight and Gulf of Mexico samples. The variability in the pattern of morphological char- acters can be accounted for by clinal variation in these characters, or, less likely, by two distinct groups that are only weakly differentiated. The interpretation of the morphological data may also be hampered by the small samples for more southern populations and the great distances be- tween them. Other life history data for tilefish in the Mid- 48 KATZ ET AL.: DELINEATION OF TILEFISH STOCKS Atlantic Bight are in accord with the concept of a separate stock. As we have previously mentioned, they are resident because they are taken year- round in the fishery (Grimes et al. 1980), appar- ently move short distances in the course of a year (Grimes et al. in press), and construct temporally stable burrows that may be occupied for the life of a fish (Able et al. 1982). In addition, they are known to reproduce in the Mid-Atlantic Bight because gonads show seasonal patterns of devel- opment and decline (Idelberger et al. 1981) and eggs and larvae have been collected (Fahay and Berrien 1981). The prevailing current patterns and hydro- graphic regimes over the study area are consis- tent with our delineation of the stocks. While there is a southwesterly drift of shelf water with- in the Mid-Atlantic Bight (Miller 1952; Bumpus 1973) that would provide mixing of eggs and lar- vae, it is unlikely that egg or larval transport occurs between the Mid-Atlantic and South At- lantic Bights. The Gulf Stream turns eastward at Cape Hatteras so that its axis is located 250 km east of the shelf break in the Mid-Atlantic Bight (Emery and Uchupi 1972). This difference in Gulf Stream effects produces distinct northern and southern continental shelf water masses (Stefansson et al. 1971; Emery and Uchupi 1972). Thus it is unlikely that egg and larval transport between these two areas would commonly occur, although Cox and Wiebe (1979) have suggested that anticyclonic eddies could provide a mech- anism for transporting oceanic larvae across the Gulf Stream to Mid-Atlantic Bight waters. Prevailing current systems in the southern United States may provide the means for larval mixing between the Gulf of Mexico and the South Atlantic Bight as suggested by the similarities in allelic frequencies for samples from these two areas. The Gulf of Mexico Loop Current (Maul 1977) provides a means for tilefish larvae to be transported out of the Gulf of Mexico and into the South Atlantic Bight as it joins the Florida Cur- rent and eventually forms the Gulf Stream. In addition to prevailing currents, periodic mass mortality may have contributed to the dif- ferences between distinct stocks. Following their discovery by a cod fisherman off southern New England in 1879, tilefish experienced a mass mortality in 1882 (a few billion fish reported floating at the surface; Bumpus 1898) probably caused by a sudden temporary intrusion of cold water (McLellan et al. 1953; Hachey 1955). This mortality may have resulted in a "founder effect" phenomenon and thus be responsible for stock differences we have noted. In summary, we believe that the available data suggest that Mid-Atlantic Bight tilefish popula- tions represent one unit stock and that South Atlantic Bight and Gulf of Mexico populations be considered another stock, at least as a working hypothesis. However, the wide geographic sepa- ration of the latter two areas may necessitate managing them as two stocks. Because the elec- trophoretic results suggest that gene flow may occur between Gulf of Mexico and South Atlantic Bight populations, this should be done with cog- nizance that Gulf of Mexico populations could serve as a source of recruits to South Atlantic Bight populations. ACKNOWLEDGMENTS This work was initiated upon the urgings and assistance of our friend, the late Lionel Walford, and could not have been accomplished without the cooperation and assistance of other persons whom we wish to gratefully acknowledge. S. Turner and M. Horvath assisted in processing samples. R. Trout provided valuable statistical council and R. Vrijenhoek availed us of his elec- trophoresis facilities and expertise. Fran and Lou Puskus and commercial tilefish fishermen in Barnegat Light, N.J., helped us obtain Mid- Atlantic Bight samples. M. Godcharles, Marine Research Laboratory, Florida Department of Natural Resources, helped us obtain samples from west Florida. Samples from off Texas, South Carolina, and the Yucatan Peninsula (Campeche Bank) were obtained from National Marine Fisheries Service RV Oregon //cruises. Financial support was provided by a New Jersey Sea Grant (R/F-2), a small grant from Rutgers University Research Council, the New Jersey Agricultural Experiment Station (Project No. 12409), and the Center for Coastal and Environ- mental 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, Lo- pholatiluschamaeleonticeps, in Hudson Submarine Can- yon. Environ. Biol. Fish. 7:199-205. Allendorf, F. W., and S. R. Phelps. 1981. Use of allelic frequencies to describe population structure. Can. J. Fish. Aquat. Sci. 38:1507-1514. Atchley, W. R., C. T. Gaskins, and D. Anderson. 1976. Statistical properties of ratios. I. Empirical re- sults. Syst. Zool. 25:137-148. 49 FISHERY BULLETIN: VOL. 81. NO. 1 Barr, A. J., J. H. Goodnight, J. P. Sall, and J. T. Helwig. 1976. A users guide to SAS 76. SAS Inst. Inc., Raleigh, N.C., 329 p. BlGELOW, H. B., AND W. C. SCHROEDER. 1947. Record of the tilefish, Lopholatilus chamaeleonti- ceps Goode and Bean, for the Gulf of Mexico. Copeia 1947:62-63. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53:1-567. Briggs, J. C. 1974. Marine zoogeography. McGraw-Hill, N.Y., 475 p. BUMPUS, D. F. 1973. A description of the circulation of the continental shelf of the east coast of the United States. In B. War- ren (editor), Progress in oceanography, Vol. 6, 4:111-157. Pergamon Press. Bumpus, H. C. 1898. On the reappearance of the tilefish (Lopholatilus chamaeleonticeps). Bull. U.S. Fish Comm. 18:321- 333. 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. Cox, J., AND P. H. Wiebe. 1979. Origins of oceanic plankton in the Middle Atlantic Bight. Estuarine Coastal Mar. Sci. 9:509-527. Dooley, J. K. 1978. Systematics and biology of the tilefishes (Perci- formes: Branchiostegidae and Malacanthidae), with de- scriptions of two new species. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 411, 78 p. Emery, K. 0., and E. Uchupi. 1972. Western North Atlantic Ocean: topography, rocks, structure, water, life and sediments. Am. Assoc. Pet. Geol., Tulsa, Okla., 523 p. Fahay, M. P., and P. Berrien. 1981. Preliminary description of larval tilefish (Lopho- latilus chamaeleonticeps). In R. Lasker and K. Sher- man (editors), The early life history of fishes: recent studies, p. 600-602. Rapp. P.-V. Reun. Cons. Int. Ex- plor. Mer 178. Fairbairn, D. J., and D. A. Roff. 1980. Testing genetic models of isozyme variability with- out breeding data: can we depend on the X 2 ? Can. J. Fish. Aquat. Sci. 37:1149-1159. Goode, G. B. 1884. The tile-fish family— Latilidae. In The fisheries and fishery industries of the United States. Section I. Natural history of useful aquatic animals, p. 360-361. U.S. Comm. Fish Fish. Grimes, C. B., K. W. Able, and S. C. Turner. 1980. A preliminary analysis of the tilefish, Lopholatilus chamaeleonticeps, fishery in the Mid- Atlantic Bight. Mar. Fish. Rev. 42(11):13-18. Grimes, C. B., S. C. Turner, and K. W. Able. In press. A technique for tagging deepwater fish. Fish. Bull.. U.S. 81(3). Hachey, H. B. 1955. Water replacements and their significance to a fishery. Pap. Mar. Biol. Oceanogr., Deep-Sea Res., Suppl. to Vol. 3, p. 68-73. Hoese, H. D., and R. H. Moore. 1977. Fishes of the Gulf of Mexico: Texas, Louisiana, and adjacent waters. Texas A&M Univ. Press, College Sta- tion, 327 p. Hubbs, C. L., and K. F. Lagler. 1967. Fishes of the Great Lakes region. Univ. Mich. Press, Ann Arbor, 186 p. Idelberger, C, C. B. Grimes, and K. W. Able. 1981. The reproductive biology of tilefish. Lopholatilus chamaeleonticeps in Middle Atlantic and southern New England waters. (Abstr.) Bull. N.J. Acad. Sci. 26(2): 67. Ihssen, P. E., H. E. Booke, J. M. Casselman, J. M. McGlade, N. R. Payne, and F. M. Utter. 1981. Stock identification: materials and methods. Can. J. Fish. Aquat. Sci. 38:1838-1855. Jolicoeur, P. 1959. Multivariate geographical variation in the wolf Canis lupus L. Evolution 13:283-299. KATZ, S. J. 1982. Identification of tilefish, Lopholatilus chamaeleon- ticeps, stocks along the east coast of the United States and the Gulf of Mexico. M.S. Thesis, Rutgers Univ., New Brunswick, N.J. Leim, A. H. 1960. Records of uncommon fishes from waters off the maritime provinces of Canada. J. Fish. Res. Board Can. 17:731-733. Manwell, C, and C. M. A. Baker. 1970. Molecular biology and the origin of species; hetero- sis, protein polymorphism and annual breeding. Univ. Wash. Press, Seattle, 394 p. Markle, D. F., W. B. Scott, and A. C. Kohler. 1980. New and rare records of Canadian fishes and the influence of hydrography on resident and nonresident Scotian Shelf ichthyofauna. Can. J. Fish. Aquat. Sci. 37:49-65. 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. McLellan, H. J., L. Lauzier, and W. B. Bailey. 1953. The slope water off the Scotian shelf. J. Fish. Res. Board Can. 10:155-176. Miller, A. R. 1952. A pattern of surface coastal circulation inferred from surface salinity-temperature data and drift bottle recoveries. Woods Hole Oceanogr. Inst. Ref. 52-28, 14 p. Nelson, W. R., and J. S. Carpenter. 1968. Bottom longline explorations in the Gulf of Mexi- co — A report on "Oregon IPs" first cruise. Commer. Fish. Rev. 30(10):57-62. Rathbun, R. 1895. Physical inquiries. Off coast of southern New Eng- land and the Middle States. In Report upon the in- quiry respecting food-fishes and the fishing-grounds, p. 32-35. U.S. Comm. Fish Fish., Part 19, Rep. Comm. 1893. Seal, H. L. 1964. Multivariate statistical analysis for biologists. Metheun and Co., Ltd., Lond., 209 p. Spiess, E. B. 1977. Genes in populations. Wiley, N.Y., 780 p. Stefansson, U., L. P. Atkinson, and D. F. Bumpus. 1971. Hydrographic properties and circulation of the North Carolina shelf and slope waters. Deep-Sea Res. 18:383-420. Wolf, R. S., and W. F. Rathjen. 1974. Exploratory fishing activities of the UNDP/FAO Caribbean Fishery Development Project, 1965-1971: A summary. Mar. Fish. Rev. 36(9):l-8. 50 EFFECTS OF BEHAVIORAL INTERACTIONS ON THE CATCHABILITY OF AMERICAN LOBSTER, HOMARUS AMERICANUS, AND TWO SPECIES OF CANCER CRAB R. Anne Richards, 1 J. Stanley Cobb, 1 and Michael J. Fogarty 2 ABSTRACT Intraspecific and interspecific behavioral interactions may affect the probability of capturing Cancer irroratus, C. borealis, and Homarus americanus in lobster traps. To test this hypothesis, the catch per unitof effort(CPUE)of eachof these species in trapsstocked with C. irroratus, C. borealis, or H. americanus was compared with that obtained from empty baited traps (controls). In traps stocked with lobsters, the catch of all three species was significantly reduced. Traps stocked with 8 lobsters caught significantly fewer crabs than traps containing 3 lobsters. The only effect of stocking traps with crabs was to increase the catch of C. borealis in traps stocked with 3 crabs of either species. Results of laboratory experiments comparing crab CPUE in control traps with crab CPUE in traps stocked with 8 lobsters concurred with the field results. When H. americanus was stocked in the holding section (parlor) of the trap, a greater proportion of the crab catch was found in the entrance section (kitchen ). This behavioral response may facilitate escape of crabs from traps containing H. americanus. The distribution of the lobster catch was un- affected by stocking H. americanus or Cancer crabs in the parlor. Behavioral mechanisms underlying reductions in crab CPUE were investigated by laboratory observation of an actively fishing trap. When H. americanus was stocked, C. borealis avoided entering traps. Cancer irroratus entered the kitchen of traps containing H. americanus, but the proportion entering the parlor was reduced. The escape rate of both crab species increased in traps stocked with H. americanus. The position underneath the entrance to the parlor was preferred by all species. When both H. americanus and Cancer crabs were present in the trap, H. americanus occupied that position. A number of environmental and biological fac- tors are known to affect the probability of cap- turing crustaceans in traps. Water temperature and salinity are positively correlated with cap- ture rates of rock lobster, Panulirus cygnus, (Morgan 1974), and a linear relationship between temperature and the catchability of American lobster, Homarus americanus, was found by McLeese and Wilder (1958). Biological rhythms and physiological changes, such as those asso- ciated with the molt cycle (e.g., Chittleborough 1975), may affect feeding and other activities (e.g., Bennett 1974; Morgan 1974) and thus cause fluctuations in catchability. In addition, behav- ioral attributes such as avoidance of dead con- specifics (Hancock 1974; Morgan 1974; Chapman and Smith 1979), intraspecific attraction (re- viewed in Hancock 1974), or competitive relations 'Department of Zoology, University of Rhode Island, King- ston, RI 02881. 2 Rhode Island Department of Environmental Management, Division of Fish and Wildlife. 150 Fowler St., Wickford, RI 02852; present address: Northeast Fisheries Center Woods Hole Laboratory, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. 0.05, Table 1). Therefore the catches from both field locations were com- bined according to treatment. The number of trap hauls for each stock species was made equal by randomly deleting observations. The hypothe- sis that the CPUE of C. irroratus, C. borealis, and H. americanus is not affected by the pres- ence of other animals inside traps was tested by comparing the total catch of each species in stocked traps with the total catch in control traps. Catches obtained after 24 h immersion time were compared using a x 2 goodness of fit test (Zar 1974). In traps containing 8 or 3 lobsters, the total catch of C. irroratus, C. borealis, and H. ameri- canus was significantly reduced (x 2 <2> = 277.8, 35.1, 18.2, respectively, P<0.001) (Table 2). In addition, the catch of both species of crabs was significantly lower in 8-lobster treatments than in 3-lobster treatments (C. irroratus, \ 2 w = 22.9, Table 1.— x 2 values for 3 X 2 contingency tables comparing strings of each treatment type for Homarus americanus (Ha), Cancer irroratus (Ci), and C. borealis (Cb) between locations. A separate contingency table was made for each species caught. * = P<0.05, @ = expected frequency of one cell was <5. Corr parison of locations for Species caught Ha treatments Ci treatments Cb treatments C. borealis C. irroratus H. americanus 0.980 @ 3880 0.348 @ 0.920 48.357* 0146 2.675 2.594 1.816 P<0.001; C. borealis, x \d = 6.1, P<0.025). The catch of lobsters was not affected by the density of stocked lobsters (x 2 u> = 2.42, P>0.05). The only effect of stocking traps with crabs was to increase the catch of C. borealis in traps stocked with either 3 C. borealis or 3 C. irroratus (for both treatments, x 2 (d = 8.6, P<0.005). Stocking traps with crabs had no effect on the catch of lobsters (P>0.05). The average size of animals captured did not differ between treatments for any of the species (Student's t test, P>0.05) (Table 3). The results of the laboratory experiments in which lobsters were stocked concurred with those from the field. The catch of both C. irrora- tus and C. borealis was significantly reduced when H. americanus was in the parlor (Table 4). Behavior Location Within Trap The spatial distribution of animals caught in a trap may be affected by behavioral interactions among the trap occupants. To test this hypothe- sis, the proportion of the catch found in the entry section, or "kitchen," in control traps was com- pared with the proportion in the kitchen in stocked traps. All comparisons of proportions were made using the normal approximation for differences between two proportions (Zar 1974). Stocked animals were placed in the parlor. In both field and laboratory experiments, a Table 2.— Total numbers of Cancer irroratus, C. borealis, and Homarus americanus caught after 24-h immersion time in field experiments. Catch per trap haul is indicated in parentheses; control = empty baited traps; treatment refers to species stocked; n = no. of trap hauls for each treatment level. H. amer/canus-stocked C borea//s-stocked C. /rrorafi/s-stocked Species caught Control 3 8 Control 3 8 Control 3 8 C. irroratus 319(7.60) 100(2.38) 42(1.00) 300(8.82) 371(10.91) 300(8.82) 342(9.50) 365(10.14) 355(986) C. borealis 70(1.67) 36(0.86) 17(0.40) 61(1.79) 99(2.91) 78(2.29) 65(1.81) 102(2.83) 70(1.94) H. americanus 54(1.29) 31(0.74) 19(0.45) 23(0.68) 21(0.62) 33(0.97) 29(0.81) 29(0.81) 29(0.81) n 42 42 42 34 34 34 36 36 36 In = 336 Table 3. — Average size (mm) and standard deviation (SD) of Homarus americanus. Cancer bore- alis, and C. irroratus caught in all traps, locations combined. Size of crabs is carapace width; size of lobsters is carapace length. H. amencanus-stocked C. borea//s-stocked C. /rrorafus-stocked Species caught Control 3 8 Control 3 8 Control 3 8 C. irroratus X 91.7 91.8 92.2 90.6 91.1 92.2 91.5 89.5 92.1 SD (10.1) (11.5) (13.2) (9.9) (11.3) (10.4) (10.6) (118) (8.6) C. borealis X 92.8 94 8 94.8 93.3 94.5 92.3 94.4 94.6 92.7 SD (9.5) (9.2) (6.5) (10.6) (8.4) (8.1) (7.9) (6.9) (9.0) H. americanus X 68.3 73.4 74.8 722 71.1 73.2 71.2 72.2 71.8 SD (7.9) (6.8) (8.1) (6.7) (9.1) (96) (7.6) (7.0) (12.5) 54 RICHARDS ET AL.: BEHAVIORAL INTERACTIONS OF AMERICAN LOBSTER AND CANCER CRABS Table 4. — Total number of Cancer irroratus or C. borealis caught in 10 laboratory trials of each treatment and catch spe- cies. ** = P<0.001, \ 2 goodness of fit test. Treatment Species caught Control 8 Homarus amencanus x 2 C. irroratus C. borealis 49 66 15 20 238" 14.9" greater proportion of the crab catch was found in the kitchen of 8-lobster treatments than of con- trols (Tables 5, 6). Stocking traps with 3 lobsters had no effect on the distribution of crabs, and lob- sters were unaffected by either stock density of lobsters (P>0.05). Interspecific interactions between C. irroratus and C. borealis apparently influenced the distri- bution of these species inside traps. In traps stocked with either 3 or 8 C. irroratus, the pro- portion of the C. borealis catch found in the kitchen was significantly greater than in con- trols (Z = 2.50, P<0.01). In traps containing 3 C. borealis, the proportion of the C. irroratus catch found in the kitchen was significantly greater than in controls (Z = 2.50, P<0.01), but no effect was seen in traps stocked with 8 C. borealis (P>0.05) (Fig. 2). LJ .14 X o I- .12 X. .10 z .08 z o .06 I- cr .04 o Q_ o .02 rr *" C. irroratus a 5^ C 4) ^S ■o ro u) CO w o *-* ** o / V/ \ ' realis s d s\ IC.bo stocke ** 1 - z 8 C. bore trol — o / , O / / O o u / / n = n = -n = - n = 'rU' "n s^ 35 -34' -37^ 38 A0\36. C. borealis SPECIES CAUGHT Figure 2.— Proportion of Cancer irroratus and C. borealis found in the kitchen of traps stocked with congeners in field experiments. All data obtained after one setover day are in- cluded, n = number of trap hauls, ** = significant difference (P<0.01) between treatment and control, using normal ap- proximation for differences between two proportions (Zar 1974). Table 5.— In field experiments, spatial distribution of Cancer irroratus, C. borealis, and Homarus americanus catch in traps stocked with H. americanus (Ha). All data obtained after one setover day are included. Proportion of catch found in the kitchen of stocked traps was compared with controls using normal approximation for differences between two propor- tions (Z) (Zar 1974). n = number of trap hauls; * = P<0.05, ** = P<0.001. Proportion Species cau ght Treatment n in kitchen z C. irroratus 8 Ha 42 0.29 8.67" 3 Ha 54 0.06 0.91 ns Control 51 0.03 C. borealis 8 Ha 42 0.35 2.00' 3 Ha 54 009 67 ns Control 51 0.13 H americanus 8 Ha 42 000 0.94 ns 3 Ha 54 00 1 25 ns Control 51 003 Table 6. — In laboratory experiments, spatial distribution of the Cancer crab catch in traps stocked with Homarus ameri- canus (Ha). Proportion of catch found in the kitchen of stocked traps was compared with controls using normal approxima- tion for differences between two proportions (Z) (Zar 1974). n = no. ot trap hauls, * = f< U.UUU1. Species cau ght Treatment n Proportion in kitchen Z C. irroratus C. borealis 8 Ha Control 8 Ha Control 10 10 10 10 0.27 0.08 0.70 009 5.19* 527- Competition Inside Traps To further investigate how the location of ani- mals in a trap is affected by behavioral interac- tions, competition for preferred areas in the trap was studied in the laboratory. Frequency of occu- pation was used as an index of preference and was measured as the number of times a given position was occupied when censused every 15 min. The observed distribution of animals was compared with an expected uniform distribution using a x 2 goodness of fit test. For lobsters and for each crab species in the absence of lobsters, the preferred position in the parlor was underneath the entry head (C. irroratus, x 2 w = 202.0, P< 0.001; C. borealis, x \v =51.8, P<0.001; H. ameri- canus, x 2 (4> = 744.2, P<0.001). When lobsters were present, the number of crabs in the parlor decreased sharply, so comparisons between lob- ster-stocked and control traps were made using proportions. In the presence of lobsters, the preference of both crab species changed (C. irro- ratus, Z = 2.26, P<0.01; C. borealis, Z = 5.97, P<0.001). Cancer irroratus occupied the middle of the parlor, and C. borealis occupied the cor- ners most frequently when H. americanus was 55 FISHERY BULLETIN: VOL. 81. NO. 1 present (C. irroratus, x \v = 82.3, P<0.001; C. borealis, x 2 (4, = 52.5, P<0.001) (Table 7). Space inside the trap was partitioned into ver- tical strata. Both crab species showed a signifi- cant increase in occupation of the top part of the trap when lobsters were present (C. irroratus, 0.47 vs. 0.79, Z = 4.87, P<0.001; C. borealis, 0.21 vs. 0.38, Z = 1.76, P<0.05). This contrasts with 99% occurrence of lobsters in the bottom portion of the trap. from the parlor did not increase in lobster- stocked traps for either species (C. irroratus, Z= 1.37, P>0.05; C. borealis, Z = 0.37, P>0.05). DISCUSSION Trap Efficiency The results of the field and laboratory experi- ments demonstrate that the presence of lobsters Table 7.— Laboratory-observed frequency and relative frequency of occupation of positions in the parlor by Cancer irroratus, C. borealis, and Homarus americanus. Counts were weighted to com- pensate for unequal availability of positions due to trap design. * = significant (P<0.01) x 2 val- ues for frequency of occupation and preferred positions; + = significant (P<0.01) differences in occupation of a particular position in lobster-stocked traps and controls; ctl = control; lob = 5 lob- sters stocked. Position i sccupied Corner Under head & arner by head Side Middle Species caught ctl lob ctl lob ctl lob ctl lob ctl lob C. irroratus Frequency 129* 27 12 9.3 32 8 9 8 60 57' Relative frequency 0.53 0.25 + 005 0.09 0.13 0.07 0.04 0.07 025 0.52 + C. borealis Frequency 93* 3 32 42.5* 42.5 17.3 23 9 42 15 Relative frequency 0.40 0.04 + 0.14 0.49 + 0.18 0.20 0.10 010 0.18 0.17 H. americanus Frequency — 555* — 204.8 — 38.6 — 109 — 471 Relative frequency — 0.40 — 0.15 — 0.03 — 0.08 — 0.34 Trap Entry and Escapement Laboratory observations revealed that C. ir- roratus and C. borealis respond differently to traps stocked with H. americanus. The presence of H. americanus did not affect the number of C. irroratus entering the kitchen (39 vs. 33, x 2 u> = 0.35, P>0.05); however, significantly fewer C. borealis entered when H. americanus were stocked (35 vs. 8, x 2 (d = 18.2, P<0.001). The proportion of C. irroratus which moved from the kitchen to the parlor was significantly reduced in lobster-stocked traps (0.81 vs. 0.23, Z= 2.73, P<0.0001). The proportion of C. bore- alis entering the parlor did not decrease signifi- cantly when H. americanus was present (0.53 vs. 0.31, Z = 0.58, P>0.05); however, the number of C. borealis that had entered the kitchen was relatively low. The proportion of both C. irroratus and C. bo- realis which escaped the kitchen increased sig- nificantly in the presence of H. americanus (C. irroratus, 0.23 vs. 0.55, Z = 2.86, P<0.005; C. bo- realis, 0.26 vs. 0.63, Z= 1.97, P<0.025). Escape reduces the CPUE of crabs, and provide a pos- sible explanation for the inverse relationship be- tween lobster and crab catches seen in other studies (e.g., Stasko 1975; Krouse 1978; Fogarty and Borden 1980). This effect appears to be den- sity-dependent since fewer crabs were captured when a large number of lobsters were present. Factors other than behavioral interactions could cause negative correlations between lob- ster and crab catch rates. Cancer irroratus is often spatially separated from C. borealis and H. americanus in Narragansett Bay (Jeffries 1966; Fogarty 1976). Such discontinuous distributions could result in inverse catches of C. irroratus and H. americanus, or of C. irroratus and C. borealis, but do not explain the differences seen in the catch of adjacent traps in this study. Other fac- tors known to affect catchability (e.g., size, sex, reproductive condition, molt stage) were held constant among stocked animals used in the dif- ferent treatments. Temperature changed little over the course of the study (average surface tem- perature, 21.9°±2.15°C). This and other environ- mental variables would have affected all treat- 56 RICHARDS ET AL.: BEHAVIORAL INTERACTION'S OF AMERICAN LOBSTER AND CANCER CRABS ments equally. The nonrandom arrangement of treatment levels within strings could have biased catch rates through gear competition. However, we feel the assumption that equal numbers of animals were attracted to all traps is valid for the following reason. If gear competition caused the reduced crab catches in lobster-stocked strings, a similar pattern of catch rates would have been seen in crab-stocked strings. This was not the case. Cancer irroratus is a prey item for lobsters (Squires 1970; Weiss 1970; Scarratt and Lowe 1972; Ennis 1973), suggesting that the decreased catch of this species in traps containing lob- sters may be the result of predator-avoidance be- havior. Cancer borealis and H. americanus are thought to compete for shelter space in rocky subtidal habitats (Stewart 1972; Fogarty 1976; Cooper and Uzmann 1977; Wang 1982). In labo- ratory studies (Fogarty 1976), H. americanus dominated C. borealis for possession of shelter. This dominance appeared to be the result of avoidance by C. borealis rather than overt ag- gressive interactions. Such behavior may cause reduced catches of C. borealis in traps containing lobsters. The reduction in lobster CPUE when lobsters were stocked is not surprising since lobsters are known to be highly aggressive and generally in- habit shelter alone under natural conditions (Cobb 1971; Cooper and Uzmann 1980). Trap sat- uration apparently becomes important for lob- sters at relatively low catch levels since traps stocked with 8 and 3 lobsters were equally effec- tive in reducing the lobster catch. In a laboratory experiment reported by Smolowitz (1978), a re- duction in trap entry was seen with only 1 or 2 lobsters in the trap. Reduced entry was probably important in the present study since escapement of stocked lobsters was low (10.1%). Stock rates used for crabs were low compared with crab catches in control traps. At higher den- sities, crabs might have had a more significant effect on the catch of lobsters. An increased lob- ster catch might be expected in traps containing C irroratus, a lobster prey item (Squires 1970; Weiss 1970; Ennis 1973; McLeese 1974). How- ever, the presence of live prey may not signifi- cantly increase the attractiveness of an already baited trap. No evidence was seen of lobster predation on crabs in traps. Similarly a decrease in lobster catch might be expected in traps con- taining a competitor (C. borealis). However, C. borealis is less aggressive than H. americanus (Fogarty 1976; Wang 1982) and occupies mutu- ally desirable shelters through passive means rather than active displacement, as shown in Stewart's (1972) study. Trap saturation apparently was not an impor- tant factor for crabs at the stock levels used, since crab catches in crab-stocked traps were not re- duced below the level of control traps. In labora- tory observations, Miller (1978, 1979a, 1980) noted that intraspecific agonistic interactions among C irroratus, Hyas araneus, and C. pro- ductus aggregating downstream from baited traps often resulted in departure from the trap area. He suggested that trap saturation in these three species was due in part to "intimidation" of crabs outside the trap by those inside. However, at relatively low catch densities, the effects of aggression may be minimal. The increased C. borealis catch in traps stocked with 3 crabs of either species is difficult to ex- plain. Release of attractants from the bait by feeding activity could enhance trap entry. As crab density inside the trap increases, such enhancement may be countered by increased aggression, reducing trap entry rates and in- creasing escapement. These speculations do not explain why the C. irroratus catch was not simi- larly increased by a low stock density of either crab species. Behavior Location Within Trap Behavioral interactions apparently affected the spatial distribution of animals in traps. A greater proportion of the crab catch was found in the kitchen when 8 lobsters were stocked in the parlor. This may have been the result of the avoid- ance responses discussed above and may enhance escapement of crabs from traps containing lob- sters. Cancer borealis shifted to the kitchen in both density levels of C. irroratus-stocked traps, but the distribution of C. irroratus changed sig- nificantly only in traps stocked with 3 C. bore- alis. Perhaps the generally greater activity of C. irroratus (Jeffries 1966; pers. obs.) serves as a deterrent to parlor entry by C. borealis. Both spe- cies may be influenced by prior residence effects in which an advantage is conferred upon the in- dividual^) initially utilizing a resource (e.g., Sinclair 1977; Davies 1978; O'Neill and Cobb 1979). Such an effect may have been caused by the stocking procedure. 57 FISHERY BULLETIN: VOL. 81, NO. 1 Competition Inside Traps During scuba diving observations of lobster traps, Pecci etal. (1978) noted an apparent domi- nance of crabs over lobsters in occupation of mutually desirable "niches" in traps. They re- ported that when both crabs and lobsters were present in traps, crabs always occupied positions that were evidently preferred by both species. The observations of this study contradict those of Pecci et al. Both crab species were displaced by lobsters. It is possible that our results re- flect a prior residence advantage conferred on lobsters by the stocking procedure. However, our findings agree with what is known of the relative aggressiveness of H. americanus, C. borealis, and C. irroratus (Fogarty 1976; Wang 1982). Escapement could be a significant factor in re- ducing the efficiency of traps. Skud 5 considered this the most likely explanation for declining catch rates for lobster over time. High escape rates for two species of Cancer have been ob- served by Miller (1979b) and High (1976). In this study, escape of both crab species from the kitchen increased when lobsters were present in the parlor, probably due to the behavioral inter- actions described above. Escape of crabs from the parlor did not increase when lobsters were stocked. This may reflect both the design of the parlor head, which makes escape more difficult, and the small sample size resulting from a low rate of entry to the parlor. In summary, the behavioral mechanisms in- volved in reducing crab catches in traps contain- ing lobsters were Trap Entry and Escapement In the laboratory, the presence of H. america- nus in a trap did not affect the number of C. irro- ratus entering the kitchen, but did decrease the number of C. boreal is entering. Just the opposite might have been expected in light of the preda- tor-prey relationship between C. irroratus and H. americanus. We observed no interactions be- tween animals inside the trap and those outside; thus the sensory basis for avoidance by C. bore- alis of traps containing lobsters is unknown. The proportion of C. irroratus moving from the kitchen to the parlor was reduced in lobster- stocked traps. The decrease in parlor entry rate for C. borealis was not statistically significant; however, the number of C borealis that had entered the kitchen was relatively low. Reduced parlor entry appeared to be the direct result of interactions between animals in the two trap compartments. These typically consisted of a lob- ster displaying (meral spread) or lunging at a crab climbing up the parlor head, resulting in retreat to the kitchen by the crab. In several in- stances, crabs hanging from the parlor head con- tacted a lobster, which responded by displaying or attacking the crab. The crab then pulled back up into the parlor head and returned to the kitchen. General lobster activity (fighting, ex- ploring, etc.) had a similar effect on crabs in the parlor head. Only 24% of C. irroratus and 10% of C. borealis entering the parlor head actually entered the parlor when lobsters were stocked. Parlor entrants increased to 60% and 67%, re- spectively, in control traps. 1 ) For C. borealis, entry to the trap is reduced, and escapement of those that enter the kitchen is increased. 2) For C. irroratus, trap entry is not reduced, but entry to the parlor decreases and rate of escape from the kitchen increases. SUMMARY This study demonstrated that behavioral inter- actions between animals attracted to traps can have significant effects on the probability of their capture. The CPUE of American lobsters and of two species of commercially harvested Cancer crabs was significantly reduced in traps containing lobsters. Such effects may be density- dependent, since significantly fewer crabs were caught in traps containing 8 lobsters than in traps containing 3 lobsters. The proportion of captured crabs occupying each trap section changed significantly when lobsters were stocked, and behavioral observations indicated that lobsters occupy the mutually preferred posi- tions in traps. The behavioral mechanisms re- sponsible for decreased crab catches included both reduced entry (C. borealis) and increased escapement (C. irroratus and C. borealis). These results reflect the behavioral and ecological rela- tions of the three species. 5 Skud, B. E. 1976. Soak-time and the catch per pot in an offshore fishery for lobsters (Homarus ameriranus). ICES Special Meeting on Population Assessments of Shellfish Stocks, No. 8, 25 p. 58 RICHARDS ET AL.: BEHAVIORAL INTERACTIONS OF AMERICAN LOBSTER AND CANCER CRABS ACKNOWLEDGMENTS We are grateful to Saul B. Saila and H. Perry Jeffries for their helpful suggestions throughout the project. The Rhode Island Department of En- vironmental Management generously provided a boat and laboratory facilities. Additional lab- oratory space was provided by the University of Rhode Island's Graduate School of Oceanog- raphy. Funding was provided by the University of Rhode Island Sea Grant Development fund. The thoughtful reviews of Jay K rouse, Robert Elner, and a journal reviewer were much appre- ciated. LITERATURE CITED Bennett. D. B. 1974. The effects of pot immersion time on catches of crabs. Cancer pagurus (L.) and lobsters, Honiara* gam- marus (L.). J. Cons. Int. Explor. Mer 35:332-336. Chapman, C. J., and G. L. Smith. 1979. Creel catches of crab, Cancer pagurus L. using dif- ferent baits. J. Cons. Int. Explor. Mer 38:226-229. Chittleborough, R. G. 1975. Environmental factors affecting growth and sur- vival of juvenile western rock lobsters Panulirus longi- pes (Milne-Edwards). Aust. J. Mar. Freshw. Res. 26: 177-196. Cobb, J. S. 1971. The shelter-related behavior of the lobster, Homar- us americanus. Ecology 52:108-115. Cooper, R. A., and J. R. Uzmann. 1977. Ecology of juvenile and adult clawed lobsters, Ho- marus americanus, Homarus gammarus, and Nephrops norvegicus. In B. F. Phillips and J. S. Cobb (editors), Workshop on lobster and rock lobster ecology and physi- ology, p. 187-208. CSIRO, Aust., Div. Fish. Oceanogr., Circ. 7. 1980. Ecology of juvenile and adult Homarus. In J. S. Cobb and B. F. Phillips (editors), The biology and man- agement of lobsters, Vol. II, p. 97-142. Acad. Press, N.Y. Davies, N. B. 1978. Territorial defense in the speckled wood butterfly (Pararge aegeria): the resident always wins. Anim. Behav. 26:138-147. Ennis, G. P. 1973. Food, feeding, and condition of lobsters, Homarus americanus, throughout the seasonal cycle in Bonavista Bay, Newfoundland. J. Fish. Res. Board Can. 30:1905- 1909. FOGARTY, M. J. 1976. Competition and resource partitioning in two spe- cies of Cancer (Crustacea, Brachyura). M.S. Thesis, Univ. Rhode Island, Kingston, 94 p. FOGARTY, M. J., AND D. V. D. BORDEN. 1980. Effects of trap venting on gear selectivity in the inshore Rhode Island American lobster, Homarus amer- icanus, fishery. Fish. Bull., U.S. 77:925-933. Hancock, D. A. 1974. Attraction and avoidance in marine inverte- brates—their possible role in developing an artificial bait. J. Cons. Int. Explor. Mer 35:328-331. High, W. L. 1976. Escape of Dungeness crabs from pots. Mar. Fish. Rev. 38(4):19-23. Jeffries, H. P. 1966. Partitioning of the estuarine environment by two species of Cancer. Ecology 47:477-481. Kennedy, D., and M. S. Bruno. 1961. The spectral sensitivity of crayfish and lobster vision. J. Gen. Physiol. 44:1089-1102. Krouse, J. S. 1978. Effectiveness of escape vent shape in traps for catching legal-sized lobster, Homarus americanus, and harvestable-sized crabs, Cancer borealis and Cancer ir- roratus. Fish. Bull, U.S. 76:425-432. McLeese, D. W. 1974. Olfactory responses of lobsters (Homarus ameri- canus) to solutions from prey species and to seawater extracts and chemical fractions offish muscle and effects of antennule ablation. Mar. Behav. Physiol. 2:237- 249. McLeese. D. W., and D. G. Wilder. 1958. The activity and catchability of the lobster (Ho- marus americanus) in relation to temperature. J. Fish. Res. Board Can. 15:1345-1354. Miller, R. J. 1978. Crab (Cancer irroratus and Hyas araneus) ease of entry to baited traps. Can. Fish. Mar. Serv. Tech. Rep. 771, 8 p. 1979a. Entry of Cancer productus to baited traps. J. Cons. Int. Explor. Mer 38:220-225. 1979b. Saturation of crab traps: reduced entry and es- capement. J. Cons. Int. Explor. Mer 38:338-345. 1980. Design criteria for crab traps. J. Cons. Int. Ex- plor. Mer 39:140-147. Morgan, G. R. 1974. Aspects of the population dynamics of the western rock lobster, Panulirus cygnus George. II. Seasonal changes in the catchability coefficient. Aust. J. Mar. Freshw. Res. 25:249-259. O'Neill, D. J., and J. S. Cobb. 1979. Some factors influencing the outcome of shelter competition in lobsters (Homarus americanus). Mar. Behav. Physiol. 6:33-45. Pecci, K. J., R. A. Cooper, C. D. Newell, R. A. Clifford, and R. J. Smolowitz. 1978. Ghost fishing of vented and unvented lobster, Ho- marus americanus, traps. Mar. Fish. Rev. 40(5-6):9-43. RlCKER, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Fish. Res. Board Can. Bull. 191, 382 p. Scarratt, D. J., and R. Lowe. 1972. Biology of the rock crab (Cancer irroratus) in Northumberland Strait. J. Fish. Res. Board Can. 29: 161-166. Sinclair, M. E. 1977. Agonistic behavior of the stone crab, Menippe mer- cenaria (Say). Anim. Behav. 25:193-207. Smolowitz, R. J. 1978. Trap design and ghost fishing: Discussion. Mar. Fish. Rev. 40(5-6):59-67. Squires, H. J. 1970. Lobster (Homarus americanus) fishery and ecol- ogy in Port au Port Bay, Newfoundland, 1960-65. Proc. Natl. Shellfish. Assoc. 60:22-39. 59 FISHERY BULLETIN: VOL. 81. NO. 1 Stasko, A. B. 1975. Modified lobster traps for catching crabs and keep- ing lobsters out. J. Fish. Res. Board Can. 32:2515-2520. Stewart, L. L. 1972. The seasonal movements, population dynamics and ecology of the lobster, Homarus americanus (Milne- Edwards), off Ram Island, Connecticut. Ph.D. Thesis, Univ. Connecticut, Storrs, 112 p. Wang, D. 1982. The behavioral ecology of competition among three decapod species, the American lobster, Homarus ameri- canus, the Jonah crab, Cancer borealis, and the rock crab, Cancer irroratus in rocky habitats. Ph.D. Thesis, Dep. Zoology, Univ. Rhode Island, Kingston, 105 p. Weiss, H. M. 1970. The diet and feeding behavior of the lobster, Ho- marus americanus, in Long Island Sound. Ph.D. The- sis, Univ. Connecticut, Storrs, 104 p. Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, N. J., 620 p. 60 THE REPRODUCTIVE BIOLOGY OF THE ATLANTIC SHARPNOSE SHARK, RHIZOPRIONODON TERRAENOVAE (RICHARDSON) Glenn R. Parsons 1 ABSTRACT Atlantic sharpnose sharks, Rhizoprionodon terraenovae (Richardson), were collected in the north central Gulf of Mexico from June 1979 to May 1980. The principal sampling devices employed were longline, trawl, and rod and reel. From a total of 215 Atlantic sharpnose sharks obtained during the study, 144 were female and 71 were male, ranging from 30 to 107 cm total lengths. The reproductive anatomy of both male and female sharpnose sharks is described. Atlantic sharpnose sharks differ from other carcharhinids in that the ovary is developed on the left side in females and overlapping siphon sacs are present in males. Clasper development suggests that males mature at about 80 cm total length, while ovarian egg diameters show that female maturation occurs at about 85 cm. Matings occur primarily between mid-May and mid-July. Embryonic growth is rapid immediately after fertilization during summer and fall but declines during winter and spring. Gestation requires 10 to 11 months and parturitions probably peak in June. Pups are released near shore at an average total length of 32 cm. Statistical analyses reveal a positive relationship between adult total length and litter size, with the largest individuals being the most fecund. An inverse relationship was observed between the numbers of embryos per uterus and embryo size. Mechanical "packing" within the uterus is proposed to explain the relationship. The seasonal distribution of sharpnose sharks was found to be determined by an inshore-offshore migration. The data indicate that during winter months in deeper offshore waters, aggregates of predominately adult female sharpnose sharks may be encountered. The sex ratio at birth was found to be 1:1 but among adults collected a 1:2.8 ratio was observed. Studies dealing with the reproductive biology of elasmobranchs have fallen far behind the volu- minous amount of data that have accumulated on reproduction in the teleostean fishes. The north- ern Gulf of Mexico has been an area of particu- lar neglect with only a few rather generalized studies (Springer 1938, 1940, 1950; Baughman and Springer 1950). Springer's (1960) classic work on the natural history of the sandbar shark, Careharhinus milberti (Eulamia milberti), con- tains a great deal of reproductive information that might be applied to carcharhinid sharks in general. Likewise, Clark and von Schmidt's (1965) survey of the sharks of the central gulf coast of Florida provided valuable reproductive data. The understanding of the life history of the blue shark, Prionace glauca, was furthered by Pratt's (1979) examination of its reproductive biology. Data concerning the life history of Rhizoprion- odon terraenovae are scarce. Rhizoprionodon spe- cies are believed to be born in the late spring and •Department of Biological Sciences, University of South Alabama, Mobile, AL 36688; present address: Department of Marine Science, University of South Florida, 140 Seventh Ave- nue South, St. Petersburg, FL 33701. Manuscript accepted June 1982. FISHERY BULLETIN: VOL. 81, NO. 1. 1983. summer. Bigelow and Schroeder (1948) reported that recently born specimens can be collected from Florida in July and that they were also present off the mouth of the Mississippi River in August. Skocik (1969) reported that pups are usually born in the spring but no data were avail- able on mating season or gestation period. Rhizoprionodon species are viviparous, the embryos obtaining nourishment via a placental connection (sometimes called a "pseudo- or yolk- sac placenta") between mother and embryo. Fe- cundity in Rhizoprionodon has been variously reported. Baughman and Springer (1950) report- ed four embryos for R. terraenovae. Bass et al. (1975) found an average of 4.7 embryos with a range of two to eight in R. acutus. Skocik (1969) reported a litter size of 12 for R. terraenovae, while Bigelow and Schroeder (1948) reported the same number for R. terraenovae taken around Cuba. Clark and von Schmidt (1965) briefly sur- veyed R. terraenovae off Englewood, Fla., and found one 83 cm female with five eggs. They also reported that all adult females examined had functional left ovaries. Compagno (1978) report- ed a range of one to four embryos for R. porosus. The pups of R. terraenovae have been reported to be 11 to 16 in (27.9 to 40.6 cm) at birth (Baugh- 61 FISHERY BULLETIN: VOL. 81. NO. 1 man and Springer 1950). Bigelow and Schroeder (1948) reported that specimens from Texas showing traces of the umbilical scar were from 280 to 407 mm long. Among R. terraenovae populations, adults are commonly 26 to 30 in (66 to 76 cm) total length (TL) (Baughman and Springer 1950), but the size at which male and female Atlantic sharp- nose sharks mature is unknown. In his revision of the genera Scoliodon, Loxodon, and Rhizoprion- odon, V. G. Springer (1964) reported that insuffi- cient information was available to establish the size at which males first mature but it appeared that maturation occurs at >640 mm TL. Bass et al. (1975) reported that male R. acutus mature between 68 and 72 cm and females at 70 to 80 cm TL. The present study is an attempt to clarify some of the known aspects of R. terraenovae reproduc- tive biology as well as to provide additional in- formation. The reproductive "strategy" of the Atlantic sharpnose shark is also examined. METHODS AND MATERIALS Atlantic sharpnose sharks, Rhizoprionodon terraenovae (Richardson), were collected in the north central Gulf of Mexico from June 1979 to May 1980. The principal sampling devices em- ployed were longline, trawl, and rod and reel. Floating longline generally gave the best re- sults (Table 1). The technique, as used by Japa- nese fishermen, is described by Lopez et al. (1979). Because of the hazard to navigation that a floating longline represents, longlining opera- tions were undertaken exclusively in deep waters offshore (Fig. 1). Longline sets were made in 10 to 28 fathom (18 to 51 m) depths, approximately due south of Dauphin Island, Ala. A trawl was used to collect specimens both inshore as well as offshore. Rod and reel, gill net, and seine were used exclusively inshore. Specimens were immediately weighed and sexed. Total, fork, and standard lengths were 30 u 30' 30" 15' 30 u 00' 29 u »5' 29 u 30' MOBILE O. O 0^2) DAUPHIN q ISLAND fort Morgan' o ••-S GULF OF MEXICO 10 fm 20 fm 30 fm 15' 00' 87° k5' Figure 1.— Coastal Alabama study area of the Atlantic sharp- nose shark. Offshore points (closed circles) represent longline and trawl sites. Inshore points (open circles) represent trawl, gill net, rod and reel, and seine sites. measured to the nearest 0.1 cm. Lengths of the claspers and siphon sacs were measured on all male specimens. All specimens were dissected immediately in the field by an incision starting at the cloaca and extending to the midpectoral region. Notes on reproductive condition in males Table 1. — Landings of Atlantic sharpnose sharks by month and by method. Longline and trawl produced more than 60% of the sharpnose shark specimens. Sharpnose sharks were collected in 10 of the 12 mo of the study period. — indicates no collections; indicates collections attempted but no sharks landed. Jan. Feb Mar Apr May June July Aug. Sept. Oct. Nov. Dec Totals Longline — 1 — 2 — — 14 — 4 19 35 75 Trawl — 1 — 21 8 — 6 8 6 9 59 Rod/reel — — 8 1 38 1 2 50 Gill net — — — 2 15 4 4 — — — — 25 Seine — — — — — 6 — — — — — 6 Totals — 2 — 25 31 7 48 27 8 4 28 35 215 62 PARSONS: REPRODUCTIVE BIOLOGY OF ATLANTIC SHARPNOSE SHARK were taken, using those indicators of maturity reported by Clark and von Schmidt (1965). Dis- sections of males allowed examinations of the re- productive systems and measurements of testicu- lar length, weight, and volume. Testes and epididymides were removed from some specimens, preserved in 10% Formalin 2 , and returned to the laboratory. Histological sec- tions of testes as well as epididymides were pre- pared. The tissues were embedded in paraffin, sectioned at 7 jum, stained with hematoxylin and eosin, and examined with phase contrast micros- copy. Sperm smears were also examined under the microscope. After obtaining weight and total, fork, and standard lengths, female specimens were dis- sected and their reproductive organs examined. Ovarian lengths as well as the number of ovarian eggs and their diameters were recorded. When embryos were present, the number, sex, total length, and wet weight were determined for each uterus. When appropriate, the data were keypunched and statistically evaluated, using the McGill University System for Interactive Computing (MUSIC) time sharing system. The STATPAK computer program, a statistical package con- taining 23 statistical analyses and data modifica- tion routines, was used to analyze the data. RESULTS AND DISCUSSION Reproductive Anatomy Ovarian Structure Forty-two Atlantic sharpnose shark ovaries were examined during the study period. E lasmo- branchs possess a great deal of variability in the structure of the ovary (Dodd 1972). The ovary of the adult Atlantic sharpnose shark is an un- paired, tear-shaped organ, 6 to 10 cm long and 3 to 5 cm wide. Unlike other carcharhinids, the ovary of the sharpnose shark is developed on the left side only. Structure and location of the sharp- nose shark ovary (aside from its position on the left side of the body cavity) are similar to that found in the blue shark (Pratt 1979). The adult sharpnose shark's ovary, during most of the year, is filled with many small (ca. 2.0 to 5.0 mm) oocytes embedded in dense connective tissue. Outside the breeding season the ovary of the adult female contains an average of about 30 oocytes greater than ca. 2 mm in diameter. These oocytes serve as a "pool" from which the next gen- eration of eggs will be drawn. In some ovaries, unusual, bright red, fluid-filled structures were found, ranging from about 2 to 8 mm in diame- ter (Fig. 2). These structures are assumed to be oocytes in a state of atresia that had failed to ovu- late during the most recent breeding period. These preovulatory structures may be "corpora atretica," which are derived from egg-containing follicles. In Cetorhinus maximus the corpora atretica are believed to arise from follicles that have attained a diameter of about 1.0 mm (Dodd 1972). The corpora atretica consist of vacuolated peripheral cells and a central cavity and are well vascularized (Dodd 1972). ANTERIOR Normal Developing Ovum Atretic Ovum 2 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. POSTERIOR FIGURE 2.— Diagram of an Atlantic sharpnose shark ovary taken in December from a 93 cm gravid female. A red, fluid filled (atretic?) ovum can be seen in the center of the ovary. Ovulation As ovulation approaches, rapid yolk deposi- tion occurs in four to eight of the many smaller oocytes. The "selected" oocytes are preferentially yolked, while the others undergo atresia. At or near ovulation the ovary appears highly vascu- larized and the large, yellow oocytes fill the en- tire ovary (Fig. 3). Measurements of both ovarian and uterine oocytes suggest that ovulation occurs at an egg diameter of about 20 mm. After ovulation, the eggs move through the body cavity into the ostium tubae which forms the anterior end of the oviduct. In most cases 63 FISHERY BULLETIN: VOL. 81, NO. 1 FULL SIZE OVARIAN EGG ANTERIOR I cm POSTERIOR OVARY Figure 3.— "Ripe" ovary of an Atlantic sharpnose shark. The ovary contains ca. 20 mm ova that are ready to ovulate. an equal number of ova enter both oviducts, although in some instances greatly dispropor- tionate numbers of embryos were found between right and left uteri. The eggs move through the oviducts to the oviducal gland where fertilization probably takes place. The oviducal gland (Fig. 4) in the Atlantic sharpnose shark is a paired struc- ture located at the forward end of the oviduct. The oviducal glands are the source of the egg case, and in some sharks the glands may be the cranial oviduct lumen opening caudal oviduct ANTERIOR POSTERIOR 1.0mm Figure 4.— Diagram of an oviducal gland taken from a mature female Atlantic sharpnose shark. site of long-term sperm storage (Pratt 1979). Viable sperm can be found within the lumen of those tubules within the gland which secretes the egg shell (Wourms 1977). As no histological sec- tions of adult sharpnose sharks' oviducal glands were prepared, the question of sperm storage in sharpnose sharks remains unresolved. Prasad (1944), however, noted the presence of spermato- zoa in the oviducal glands of Scoliodon sorra- kowah, a closely related Indian Ocean species. This observation suggests that the oviducal gland may have at least a short-term storage capacity. After moving through the oviducal gland the fertilized eggs then move to the uterus where they become implanted in depressions in the uterine wall. At this point the eggs are found en- cased in a thin, yellowish shell with pointed ends (Bigelow and Schroeder 1948). Within the uterus the eggs are elongate, averaging about 18 mm wide and about 32 mm long. Fertilization is ap- parently very efficient since in examination of 315 embryos only two unfertile eggs were noted (0.6%). Placentation and Structure of the Umbilical Cord During the first 2.5 to 3.0 mo of gestation, the Atlantic sharpnose shark embryos depend upon the yolk sac for nourishment. After about 3 mo the yolk sac has become intimately associated with the uterine wall to form a yolk-sac placenta. October embryos, i.e., 3 mo old, were ca. 16 to 20 cm and had well-developed placentas with little yolk material remaining. By November, 4 mo into gestation, embryos were 19 to 23 cm long and no yolk material remained in the placenta. In a related Indian Ocean species, Scoliodon sor- rakowah, Mahadevan (1940) described a very thick vascularized area of the uterine wall, re- ferred to as a trophonematous cup, which forms to receive the yolk sac of the foetus. This vascu- larized area was also noted in the Atlantic sharp- nose shark. Development of the umbilical cord closely par- allels placentation. The umbilical cord is con- nected on the embryo's ventral surface in the midpectoral region. Very early in development the umbilical cord is virtually naked. By the time the embryos have grown to about 6.0 cm TL the umbilical cord has developed many knoblike ap- pendages which give it a "pipe-cleaner" appear- ance. The appendages are about 1 mm long, and terminate in one or a cluster of several grapelike 64 PARSONS: REPRODUCTIVE BIOLOGY OF ATLANTIC SHARPNOSE SHARK distentions. Budker (1971) suggested that in ad- dition to placentally derived nutrients, these appendages may allow the embryo to absorb di- rectly nutritive substances that are secreted by the uterine lining. This type of nutrition is termed histotrophic. As gestation progresses the append- ages of the sharpnose shark's umbilical cord lengthen and change morphologically. Full-term embryos possessed umbilical cords about 10 to 12 cm long with appendages about 10 mm. The pro- jections at this time have a foliose appearance, i.e., flattened, extensively branched, and termi- nating in rounded, flat expansions. This differs from the fingerlike shape described for the pro- jections found on the umbilical cord of Sphyrna tiburo (Schlernitzauer and Gilbert 1966). Structure of Claspers and Siphon Sac The paired claspers of the adult male Atlantic sharpnose shark are much the same as those of other carcharhinid sharks. The claspers are rigid, calcified, intromittent organs that rotate freely around their attachment base. The tip, or rhipidion, expands whereupon the rigid carti- lages of the tip are directed at right angles to the main axis of the clasper. This expansion is be- lieved to function as an anchor, holding the clasp- er in the oviduct during copulation. Under nor- mal circumstances the claspers are directed posteriorly. Springer (1960) has suggested that just prior to mating the claspers of large carcha- rhinid sharks such as Eulamia milberti (Carcha- rhinus milberti) rotate in and forward. Expan- sion of the rhipidion occurs independently after insertion of the clasper into the oviduct of the fe- male. This apparently also occurs in the Atlantic sharpnose shark, since a live specimen captured in December had one clasper oriented in this fashion, with the rhipidion expanded, probably a result of trauma. The clasper gradually returned to normal after about 3 min. The siphon sac in the adult Atlantic sharpnose shark is a muscular, subdermal organ which be- gins at the base of the claspers, extends anteriorly along the ventral surface, and ends just short of the coracoid bar. The sac in adults ranges from about 20 to 28 cm long and 1 to 2 cm wide. Unlike other shark species which have paired separate siphon sacs, Atlantic sharpnose sharks possess overlapping sacs which communicate with the claspers via an opening located at the base of each clasper. Springer (1960) suggested that the siphon sac is filled with water just prior to mating and is used to flush sperm along the clasper groove and into the oviducts during copulation. The clasper siphon of adult spiny dogfish, Squa- lus acanthias, has been found to be a rich source of serotonin. This suggests that the siphon-sac secretion may play a role in affecting the mech- anism of copulation and ejaculation in the male, or by eliciting contractions of the female repro- ductive tract, thus influencing passage of sperm and fertilization (Mann 1960). Structure of the Testes and Epididymides The testes in the adult male Atlantic sharpnose sharks are paired, elongate, flattened organs (Fig. 5). Depending on the season and the size of the adult, the testes range from 13 to 20 cm long, 1 to 2 cm wide, and 0.5 to 1.0 cm thick. The testes are located dorsal to the lobes of the liver at the anterior end of the peritoneal cavity. The organs are supported here by a mesorchium. Microscopic examination of a mature testis of the sharpnose shark shows that the organ is filled ANTERIOR ^i POSTERIOR cm TEST I S Figure 5.— Diagram of a "ripe" Atlantic sharpnose shark tes- tis. The testis is turgid indicative of the reproductively active condition. 65 FISHERY BULLETIN: VOL. 81. NO. 1 with spherical seminiferous ampullae, much the same as are found in spiny dogfish (Simpson and Wardle 1967) and blue shark (Pratt 1979). Histo- logical sections of mature testes demonstrate that these ampullae contain spermatozoa in vari- ous stages of development (Fig. 6). Viewed in cross section, the heads of the mature spermato- zoa are arranged in discrete groups around the periphery of the spherical ampullae. The spermatozoa leave the testis by way of the efferent ductules and enter the epididymis. The epididymis is a paired organ located above the testis against the dorsal wall of the abdominal cavity. The sharpnose shark's epididymis is about 15 cm long, 1.0 cm wide, and 0.5 cm thick. Histological sections of an epididymis from a re- productively active sharpnose shark reveal great numbers of spermatozoa present in the tubules of the organ (Fig. 7). Maturation Males Maturity in animals can generally be deter- mined by comparing external secondary sex characters in adults with the same characters in smaller individuals. Using two indicators of sex- ual maturity (i.e., clasper growth and siphon-sac development), it was determined that matura- tion of the male Atlantic sharpnose shark begins at about 60 to 65 cm TL and is complete at about 80 cm. At <65 cm TL the clasper length represents about 2.5% of the adult total length. Regression analysis shows that the claspers undergo a period of rapid growth with a major inflection in the line occurring at 65 to 70 cm TL (Fig. 8). The claspers quickly elongate, growing 3 cm within a short period of time to represent 7 to 8% of the total length. The smallest mature males exam- ined were about 80 cm long and their claspers represented about 7.8% of total length. There were many individuals examined between 75 and 80 cm TL that possessed elongated claspers, but incomplete calcification of the claspers indi- cated that the specimens were not mature. The clasper grows faster than the total length at the onset of maturation and for a short period into adult life. Regression analysis indicates that from about 85 to 95 cm TL the relationship is un- changing, but after 95 cm there is a period of Figure 6.— Histological section of a testis from a mature Atlantic sharpnose shark (X440). The cross sections show that the heads of the mature spermatozoa are arranged in discrete groups around the periphery of the spherical seminiferous ampullae. 66 PARSONS: REPRODUCTIVE BIOLOGY OF ATLANTIC SHARPNOSE SHARK ¥!*?& . FIGURE 7.— Histological section of an epididymis from a mature Atlantic sharpnose shark (X 140). Large num- bers of spermatozoa are present within the tubules of the structure. negative allometric growth. The claspers, after attaining their functional length, do not continue to grow or at least grow very little. This is a ten- able hypothesis since continued growth would not necessarily enhance the claspers' utility. Development of the siphon sacs coincides close- ly with the rapid increase in clasper length (Fig. 9). This muscular, subdermal organ is nonexis- tent until the onset of maturity. The siphon sacs develop quickly and represent about 28% of the 88t 74 5 59 0. 4 5 31 •6t 30 o OBSERVED • REGRESSION ESTIMATED '8-8- o o -o OO o • O \Q 8 7 O 8\D 2b o/°° o ' ° \° _l o < 20 in 15 z o o I Q. 35 7 44 1 52-4 60 8 69 2 77 5 85 9 94 2 102 6 TOTAL LENGTH (cm) Figure 8. — The maturation of male Atlantic sharpnose sharks as evidenced by clasper development. The regression line indi- cates that maturation occurs between 80 and 85 cm total length, AT = 70. 10 5 35-7 44 I 52-4 60-8 69-2 77-5 85 9 94 2 102 6 TOTAL LENGTH (cm) Figure 9.— The maturation of male Atlantic sharpnose sharks as evidenced by siphon-sac development. The scatter diagram suggests that maturation occurs at about 80 cm total length, AT = 35. 67 FISHERY BULLETIN: VOL. 81. NO. 1 total length at maturity. The smallest mature individuals were about 80 cm and possessed si- phon sacs about 23% of total length. Females Maturation in females was determined by ex- amining the developing ovary and ovarian eggs. Females were found to mature at a greater total length than males. The ovary does not begin to develop until the individual reaches about 60 cm TL. Figure 10 shows that development reaches an asymptote between 85 and 90 cm TL. Even among individuals of the same size taken during the same month there is a high degree of varia- tion in ovarian length. For this reason ovarian length is not considered a good indicator of ma- turity in Atlantic sharpnose shark. Changes in the diameter of ovarian eggs were found to be a reliable indicator of the beginning of maturation. Figure 11 shows the first genera- tion of ovarian eggs produced by the subadult population. Increase in egg diameter begins at 60 to 65 cm TL, at about the same time the length of the ovary begins to increase. The eggs increase in diameter until the first ovulation, which oc- curs at about 85 to 90 cm TL. Most female sharp- nose sharks mature within this size range. Several female sharpnose sharks that had re- cently matured were examined. One individual of 88 cm TL, collected in late May, had full-sized ovarian eggs and had apparently recently mated due to the numerous mating scars that were ob- served in the region between the first and second I0-4T >- cr 88 I 7-3 + 5 57 + 2 6 o o o o I . • o o o o o o oo •••••• . ooo o • I o o o g oo o o o o o OBSERVED • REGRESSION ESTIMATED 50 60 70 80 90 TOTAL LENGTH fcrW 100 110 Figure 10.— Regression analysis showing development of the ovary in Atlantic sharpnose sharks, N = 42. Maturation is esti- mated to be complete at 85 to 90 cm total length. CT (li H ld < Q (a to < cr o 25 T 20 15- 10 • 1st OVULATION •-• •- • • -+- ■+- -+- 40 50 60 70 80 TOTAL LENGTH (cm) 90 100 Figure 11. — Maturation of female Atlantic sharpnose sharks as evidenced by the increase in ovarian egg diameter. Hand-fit curve approximates the increase in ovarian egg diameter from juvenile to first ovulation, N = 63. dorsal fins. An 86 cm individual, collected in early July, possessed six ova (8 to 10 cm), while another 89 cm female, collected in mid-July, pos- sessed uterine eggs. In late August, all mature females examined contained embryos. The small- est gravid specimens were 87, 88, and 89 cm TL and contained 11, 8, and 6 cm embryos, respec- tively. These observations further support the 85 to 90 cm estimated size at maturity. Mating Season Twenty-three reproductively active male At- lantic sharpnose sharks were examined to delin- eate the mating season. A gonadosomatic index (GSI), testis weight expressed as percent total body weight, was found to be the best indicator of mating season. The GSI provided a defined mating season for male sharpnose sharks (Fig. 12). Reporting on central gulf coast of Florida populations, Clark and von Schmidt (1965) suggested that small shark species (such as Mustelus norrisi and Scoli- odon terraenovae = Rhizoprionodon terraenovae) mate and bear young in the late winter and early spring. In the north central gulf, contrary to Clark and von Schmidt's findings for Florida, male sharpnose sharks appear to be reproduc- tively active during late spring and summer. From about September to March, the GSI was found to be low, 0.2 to 0.37. During these months specimens were observed to have reduced testes 68 PARSONS: REPRODUCTIVE BIOLOGY OF ATLANTIC SHARPNOSE SHARK 0.6t 0.5 0.4 o 0.3 < 5 0.2 0.1 I 2 HAND FIT CURVE Figure 12. — Mating season of adult male Atlantic sharpnose sharks as evidenced by the seasonal increase in gonadosomatic index (GSI). The data suggest that male sharpnose sharks are reproductively active during late spring and summer. The closed circles represent mean values and the numbers indicate sample sizes, N= 20. and no visible sperm or semen in the seminal vesicles. In late April the GSI had risen to 0.51, but there was little sperm present in the seminal vesicles. During mid- to late May the GSI aver- aged 0.47. All mature individuals had enlarged testes, turgid seminal vesicles, and copious amounts of sperm present in the claspers as evi- denced by microscopic examination. This condi- tion was found to persist through June and July with GSI equalling 0.59 and 0.57, respectively. Several adult males examined in August were found to have large quantities of sperm in the seminal vesicles. A single GSI determination in- dicated a slight decline from previous months. The mating season in female sharpnose sharks was evidenced by an increase in ovarian egg di- ameter (Fig. 13). From August to December the average egg diameter increased from ca. 3.0 to 4.2 mm. In almost every ovary examined during November and December, a few eggs were be- ginning to visually dominate the other oocytes. In February, the mean oocyte diameter equalled 5.0 mm, with some eggs reaching 11 mm. In Feb- ruary, all mature ovaries contained four to eight oocytes that were noticeably larger than sur- rounding eggs. From mid-February to late May or June, there was a rapid increase in egg diame- ter to about 20 mm at ovulation. The information indicates that the mating sea- son for male and female sharpnose sharks in the northern Gulf of Mexico coincides, although male sharpnose sharks are reproductively active earlier in the year. Assuming that females do not mate when gravid and that ovulations occur after copulation, then the mating season must occur between mid-May and mid-July. Most adult females still carried near-term embryos in mid-May, and by mid-July all females examined had uterine eggs. Considering the peak of partu- rition for gravid females (see Embryonic Growth and Development section), the subsequent ap- pearance of uterine eggs, and the occurrence of the first detectable embryos, the peak of mating most likely occurs from mid-June to mid-July. Embryonic Growth and Development Embryos representing various stages of devel- opment were weighed, sexed, and measured in total length. Conceptions were estimated to be at a peak in early to mid-July. At this time several sharpnose sharks that possessed recently ovu- lated uterine eggs but no visible embryos were examined. In late August, gravid females were collected, and they contained embryos ranging from about 4 to 11 cm TL. The smallest embryos examined were still dependent upon the yolk sac. They had prominent branchial gill filaments, undeveloped fins, and the anterior end was en- larged in relation to the rest of the body. Pratt (1979) suggested that growth of embryonic Prio- nace glauca is linear. Increase in length of sharp- nose shark's embryos approximates a sigmoid curve as evidenced by polynomial regression 25 1 IT LU < O O 19 LlI z < o 20- 15- 10 5- UTERINE • MEAN I RANGE •H-l Figure 13.— Mating season of adult female Atlantic sharpnose sharks as evidenced by the seasonal increase in ovarian egg diameter, N = 1,260. The data suggest that the mating season for females occurs from mid-June to mid-July. 69 FISHERY BULLETIN: VOL. 81. NO. 1 analysis (Fig. 14). After conception there is a pe- riod of rapid growth through the remainder of the summer and fall. By November the embryos have attained an average of 21.3 cm and appear almost completely developed. There is a notice- able inflection in the regression line in Novem- ber. The increase in length declines through the winter and spring months, although a slight in- crease may occur just before parturition in May or June. Pups are born at an average of about 32 cm TL. Skocik (1969) reported a total length of 25 cm for sharpnose shark at birth, and Bigelow and Schroeder (1948) stated that newborn sharp- nose sharks are generally about 275 to 400 mm long. The largest embryo recorded during the study period was 36 cm TL and the smallest free- living specimen was 32 cm. Increases in weight of the sharpnose shark's embryo differed from the increases in total length (Fig. 15). Embryo weight increased slowly during the period from estimated conception (mid-July) to October. Thereafter, however, until parturition in late May or June, an almost linear increase of about 16 g/mo occurred. Parturition occurs most likely between about 95 and 150 g. By using the above information, it was possible to estimate the gestation period. Atlantic sharp- nose shark's embryos require a 10 to 11 mo gesta- tion period, beginning in July or August and ending in May or June of the following year. Relationships Between Adult Females and Embryos A significant relationship was observed be- tween total length of the gravid female and the number of offspring produced. This is note- worthy since other works have failed to show such a relationship among carcharhinids (Springer 1960; Clark and von Schmidt 1965). Figure 16 shows that the total length of the adult is correlated with litter size ( ANOVA significant at <0.01). There is a direct relationship between fecundity and the size of the adult with the largest individuals being the most fecund. Grav- id females produce an average of 5 pups/litter per year (one to seven), but in most cases either four or six embryos will be present. It was anticipated that a relationship between litter size and embryo size could be detected. An optimal clutch size has been demonstrated in some species of birds (Lack 1954, 1966, 1968). Compared with small and large clutches, inter- mediate-sized clutches leave proportionately 36 5 28 2 199 11-6 o > a. 33 -4-9 1 * • REGRESSION ESTIMATE I RANGE Figure 14.— Growth of embryonic Atlantic sharpnose shark. Regression analysis shows the increase in embryo total length from fertilization to parturition, N = 300. 145-6 124-7- I03-8-- 829- x S2 Id S o £ 62-0 m 2 Id 412- 203- -0 6-»-« • REGRESSION ESTIMATE I RANGE -t 1 f -i 1 1 1 Figure 15. — Growth of embryonic Atlantic sharpnose shark. Regression analysis shows the increase in embryo weight from fertilization to parturition, N = 300. more offspring that survive to maturity. Birds from large clutches are smaller in size than birds from intermediate-sized clutches. After evalu- ating the data, an "optimal litter size" could not be demonstrated for the Atlantic sharpnose sharks. However, when the right and left uteri of adults collected during a single sampling trip (to cancel out seasonal differences) were treated separately, an inverse relationship was observed between the numbers of embryos per uterus and 70 PARSONS: REPRODUCTIVE BIOLOGY OF ATLANTIC SHARPNOSE SHARK oz 4 LU 3 2 • MEAN I RANGE N = 78 85-90 90-95 95-100 100-105 105-110 ADULT TOTAL LENGTH (cm) Figure 16.— Relationship between adult total length and litter size of the Atlantic sharpnose sharks. The plot indicates that fecundity increases significantly as adult total length increases (F = 9.216, P<0.00001). 24.1 + 22.9 I i- o z LU _l _l s 2 1.7- *~ 20.4 - o CE HI s LU 19.2 18. X • MEAN 1 RANGE EH 95% CONFIDENCE INTERVAL N = 300 I 3 4 No. PER UTERUS Figure 17.— Relationship between numbers of embryos per uterus and embryo total length of the Atlantic sharpnose sharks. Embryo total length decreases significantly with in- creasing number per uterus, AT = 89. embryo size (Fig. 17). The figure indicates that at the 95% confidence limits significant differ- ences exist between the total lengths of the em- bryos. Embryos were found to be largest when one or two are present per uterus. However, in only one case was there a single embryo found within a uterus. It is conceivable that mechanical "packing" within the uterus causes "intra-uterine competi- tion" for nutrients. As already discussed, in addi- tion to placentally derived nourishment, sharp- nose shark embryos may be able to absorb directly nutrients which are produced by the uterine epithelium. An increase in the number of embryos within the uterus above some optimal value might result in competition for this "uter- ine milk" and a decrease in embryo size. In sharpnose sharks, the parents that produce what might be termed an "optimal" number of embryos per uterus are producing the largest embryos. If we assume that these size differences are retained until birth, and thereafter, these larger embryos will result in progeny of highest individual fitness. Larger offspring cost more to produce, but they are also worth more (Pianka 1978). It would be interesting to examine the repro- ductive strategy of tropical sharpnose shark populations, since these sharks have been report- ed to have litters with as many as 12 embryos (Bigelow and Schroeder 1948; Skocik 1969). Based on this study, it would be a logical extra- polation to predict that these litters would result in smaller offspring. A litter of 12 must be approaching maximum fecundity for sharpnose sharks. Seasonal Distribution In this study it was determined that migratory behavior of the Atlantic sharpnose shark is pri- marily limited to an inshore-offshore movement. From late April to September of 1979, 93 sharp- nose sharks were collected from shallow inshore waters. During the period from late October 1979 to April 1980, despite numerous attempts, no sharpnose sharks were collected inshore. Sharpnose sharks may be encountered offshore year-round; however, the data indicate that the concentration of sharks is greatest during the fall and in particular, winter months. From Octo- ber 1979 to February 1980, 59 sharpnose sharks were collected during offshore longlining. Fig- ure 18 shows that the number of sharpnose sharks landed in deep water, as well as the catch per unit effort (CUE), is low in spring and sum- mer (CUE = 1.2 and 2.4, respectively) and in- creases to a high in winter (CUE = 7.3). The above data suggest that the migration from inshore to offshore begins around October or November. Atlantic sharpnose sharks appar- ently remain in deeper waters during the colder months and return inshore again in April and May. 71 FISHERY BULLETIN: VOL. 81, NO. 1 10 1 LjJ o 5 T 40 23. 14 12 IA ia. Li 2£l 30 ■20 •10 D to o z S z < Id co O z Q. < I co SPRING FALL Figure 18.— Catch per unit effort (CUE) in sharks/100 hooks per hour and number of Atlantic sharpnose sharks landed dur- ing longline operations. Ninety percent of these offshore land- ings were gravid females. Since adult female Atlantic sharpnose sharks were collected inshore only during summer months, the data suggest that females migrate inshore in late spring or summer to pup and mate, whereupon they return offshore again to overwinter. During June and July sharpnose shark pups with a fresh umbilical scar (in some cases the scar was actually an open slit) could be collected from the littoral zone. It is likely that special nursery areas exist for many shark spe- cies (Springer 1967), although the existence of specific pupping or nursery grounds for the At- lantic sharpnose sharks could not be conclusively established from this study. However, since new- born pups were never taken from deep waters in spite of intensive trawling, it is reasonable to suppose that the pups were born in shallow water. Perhaps the shallows of the northern Gulf of Mexico's extensive barrier island system serve as pupping/nursery grounds for the Atlantic sharpnose shark. Sex Ratio Sex of the Atlantic sharpnose sharks could be determined by clasper examination in embryos as small as 5.0 cm TL. The sex ratio through most of gestation could therefore be determined. The sex ratio early in development and of near-term embryos was found to be 1:1. One-hundred and fifty male and 155 female embryos were exam- ined. These data suggest that the sex ratio at par- turition is also 1:1. Among adults sampled, the sex ratio was found to be one sided in favor of females. During this study 33 adult male and 91 adult female sharpnose sharks were collected representing a 1:2.8 ratio. During offshore longlining 90% of the catch consisted of gravid adult female sharpnose sharks. This condition in sharpnose shark is not without precedent, as it has been observed in other shark species. Springer (1940), discussing Carcharhinus milberti and Carcharhinus ob- scurus, stated that in both species females out- number males. Clark and von Schmidt (1965) found a similar situation in Galeocerdo cuvieri. ACKNOWLEDGMENTS This research was funded by a fellowship awarded to the author by the Dauphin Island Sea Laboratory. Robert Shipp provided helpful as- sistance. Special thanks are due David Nelson for reviewing the manuscript, Snead Collard for supplying shark rodeo data, Benny Rohr for sup- plying NMFS data, and John Gourley for supply- ing data. I am also very grateful for the assistance pro- vided by the faculty and staff of the Dauphin Island Sea Laboratory. Likewise, I wish to thank the graduate students of the University of South Alabama, especially Rick Blaise, Steve Branstet- ter, Don Marley, and Austin Swift. 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 Mustelus and Car- charhinus) and Sphyrnidae. South Afr. Assoc. Mar. Biol. Res., Invest. Rep. 38, 100 p. Baughman, J. L., and S. Springer. 1950. Biological and economic notes on the sharks of the Gulf of Mexico, with especial reference to those of Texas, and with a key for their identification. Am. Midi. Nat. 44:96-152. Bigelow, H. B., and W. C. Schroeder. 1948. Sharks. In A. E. Parr and Y. H. Olsen (editors), Fishes of the western North Atlantic, Part 1, p. 59-546. Mem. Sears Found. Mar. Res., Yale Univ. 1. Budker, P. 1971. The life of sharks. Columbia Univ. Press, N.Y., 222 p. Clark, E., and K. von Schmidt. 1965. Sharks of the central Gulf coast of Florida. Bull. Mar. Sci. 15:13-83. Compagno, L. J. V. 1978. Sharks. In W. Fischer (editor), FAO species iden- tification sheets for fishery management purposes; west- ern central Atlantic, Vol. 5, unpaginated. Lack, D. 1954. The natural regulation of animal numbers. Ox- 72 PARSONS: REPRODUCTIVE BIOLOGY OF ATLANTIC SHARPNOSE SHARK ford Univ. Press, Lond., 343 p. 1966. Population studies of birds. Oxford Univ. Press, Lond., 341 p. 1968. Ecological adaptations for breeding in birds. Methuen, Lond., 409 p. Lopez, A. M„ D. B. McClellan, A. R. Bertolino, and M. D. Lange. 1979. The Japanese longline fishery in the Gulf of Mexi- co, 1978. Mar. Fish. Rev. 41(10):23-28. Mahadevan, G. 1940. Preliminary observations on the structure of the uterus and the placenta of a few Indian elasmobranchs. Proc. Indian Acad. Sci., B, 11:1-44. Mann, T. 1960. Serotonin (5-hydroxytryptamine) in the male re- productive tract of the spiny dogfish. Nature (Lond.) 188:941-942. PlANKA, E. R. 1978. Evolutionary ecology. Harper and Row, N.Y., 397 p. Prasad, R. R. 1944. The structure, phylogenetic significance, and func- tion of the nidamental glands of some elasmobranchs of the Madras coast. Proc. Natl. Inst. Sci. India, Part B, Biol. Sci. 11:282-302. Pratt, H. L. 1979. Reproduction in the blue shark, Prionace glauca. Fish. Bull., U.S. 77:445-470. Radcliffe. L. 1916. The sharks and rays of Beaufort, North Carolina. Bull. U.S. Bur. Fish. 34:239-284. SCHLERNITZAUER, D. A., AND P. W. GILBERT. 1966. Placentation and associated aspects of gestation in the bonnethead shark. Spin/nut tiburo. J. Morphol. 120:219-231. Simpson, T. H., and C. S. Wardle. 1967. A seasonal cycle in the testis of the spurdog, Squalus acanthias, and the sites of 3/3-hydroxysteroid dehydro- genase activity. J. Mar. Biol. Assoc. U.K. 47:699-708. Skocik, R. 1969. The sharks around us. Star Publ. Co. Inc., 206 p. Springer, S. 1938. Notes on the sharks of Florida. Proc. Fla. Acad. Sci. 3:9-41. 1940. The sex ratio and seasonal distribution of some Florida sharks. Copeia 1940:188-194. 1950. A revision of North American sharks allied to the genus Carcharhinus. Am. Mus. Nat. Hist. Novit. 1451, 13 p. 1960. Natural history of the sandbar shark, Eulamia mtiberti U.S. Fish Wildl. Serv., Fish. Bull. 61:1-38. 1967. Social organization of shark populations. /« P. W. Gilbert, R. F. Mathewson, and D. P. Rail (editors), Sharks, skates, and rays, p. 149-174. John Hopkins Press, Baltimore. Springer, V. G. 1964. A revision of the carcharhinid shark genera Scoli- odon, Loxodon, and Rhizoprionodon. Proc. U.S. Natl. Mus. 115:559-632. WOURMS, J. P. 1977. Reproduction and development in chondrichthyan fishes. Am. Zool. 17:379-410. 73 VARIATION IN THE GROWTH RATE OF MYA ARENARIA AND ITS RELATIONSHIP TO THE ENVIRONMENT AS ANALYZED THROUGH PRINCIPAL COMPONENTS ANALYSIS AND THE co PARAMETER OF THE VON BERTALANFFY EQUATION Richard S. Appeldoorn 1 ABSTRACT Age-length data and environmental parameters were obtained for 25 populations of the soft-shell clam, Mija arenaria. Growth rates were analyzed for 20 of the populations and variations in the growth rates were related to differences in the environment. The analysis of growth was based on Gallucci and Quinn's w parameter for the von Bertalanffy equation. Environmental variability was analyzed, using principal components analysis which yielded three environmental factors: North- ness, siltiness, and sedimentary hydrocarbons. Growth was found to be significantly related to each of the three components. A distinct latitudinal growth relationship was observed, with growth de- creasing towards the north. Temperature, tidal height, tidal position, and edaphic conditions sys- tematically varied with latitude, with temperature being the dominant factor affecting growth. Growth was negatively correlated to both siltiness and sedimentary hydrocarbons. The growth of the soft-shell clam, Mya a renaria, has been studied by many investigators (Wilton and Wilton 1929; Belding 1930; Newcombe 1936; Swan 1952; Brousseau 1979; and others), and much work has been done in assessing the im- portance of various environmental factors in the growth process. These factors include water cur- rent and quality, food, temperature, salinity, various edaphic parameters, and pollution. In the past, investigators were obliged to study these factors individually even though it was realized that many were interrelated (Belding 1930). Because of local variations researchers often disagreed on the relative importance of each of these factors, and overall trends have not been firmly established. The purpose of this study was to investigate various factors contributing to growth rate vari- ations in soft clam populations and to demon- strate a methodology incorporating the analysis of multiple factors applicable to the above inves- tigation. Of specific interest was the demonstra- tion of a latitudinal trend in growth and the fac- tors responsible for it, since such a relationship had yet to be quantified (Brousseau 1979). Prin- cipal components analysis was used to analyze 'Graduate School of Oceanography, University of Rhode Island, Kingston, R.I.; present address: Department of Marine Sciences, University of Puerto Rico, Mayagiiez, Puerto Rico 00708. multivariate environmental data, and the von Bertalanffy model was used for the analysis of growth, using the recently introduced growth rate parameter w of Gallucci and Quinn (1979). This study represents one of the first applica- tions of w to investigate growth rate variations. MATERIALS AND METHODS Samples of Mya arenaria and environmental data were obtained from 25 sites located along the east coast of North America, from Maryland to Nova Scotia (Fig. 1). The sites were initially chosen and sampled as part of a study to investi- gate the relationship between environmental quality and neoplasia (Brown 1980), and as a re- sult 1) the sites varied greatly in their environ- mental quality, 2) the sampling design employed was not specifically designed for the present study, and 3) it was therefore necessary to use proxy data in some cases to represent certain en- vironmental characteristics. These drawbacks were not severely limiting, since the particular statistical techniques used could control much of the induced variability in the data. Estimates of the following environmental parameters were obtained: Salinity, tidal position, tidal range, average annual temperature, sedimentary grain size, dispersion and skewness of grain sizes, per- cent silt-clay, percent organic matter, and total sedimentary hydrocarbons. Manuscript accepted June 1982. FISHERY BULLETIN: VOL. 81, NO. 1, 1983. 75 FISHERY BULLETIN: VOL. 81. NO. 1 75°W 70° W 65 8 W 45° N - 40° N - r l" A / RS \ hzcr / PY \ ^k ^2r V £ NORTH AMERICA RB-~^^ BN, nbA WC\\ AH-^ SH ^^ SP-^\. \ : T \ gc \\vS PT Dl \ji/ / pi JL \ WK 3R y / oi * 02 \ WP TS — J _jZ U' A NR ATLANTIC OCEAN ^AR 1 l - 45°N - 40° N 75°W 70°W 65°W Figure 1.— Location of sampling sites. Site codes are given in Table 1. Salinity, at low tide, was measured by a refrac- tometer; tidal position was estimated on a scale of 0-1, where = subtidal and 1 = full exposure. Estimates of the average annual temperature near each site were obtained from various litera- ture sources, and estimates of the tidal range were obtained from National Ocean Survey (1978). Sediment samples (composites of two surface cores 21 cm 2 X 8 cm depth) were collected and analyzed to determine grain size distribution and organic content. The sand fraction was ana- lyzed by dry sieving; silt-clay by the hydrometer method (American Society for Testing Materials 1963). The particle size distributions obtained from the two analyses were pooled, and the cumu- lative frequency versus grain size () was plotted for each sample. From the graphs the following summary statistics were obtained: Median grain size (Md4>), quartile deviation (QD), and skew- ness (Skq) (Buchanan 1971). The results were reported in phi notation rather than millimeters [(f) = — log2(mm)], as this scale is commonly used to describe grain size characteristics and be- cause it allows for greater discrimination in the silt-clay range which may be more meaningful biologically. The percent organic matter was determined by measuring the percent weight loss of a small aliquot upon ignition at 550°C for 4 h (Buchanan 1971). Estimates for total sedimentary hydrocar- bons through infrared analysis were obtained from C. Brown. 2 The sites and their environmen- tal parameters are given in Table 1 along with their dates of collection, latitude, and code. Clams were sampled from one or more trenches dug by a standard clam hoe or shovel, with the exception of the Chesapeake Bay sites where a commercial hydraulic escalator dredge was used. All clams excavated were retained for analysis. For each individual, shell length (maximum shell dimension) was measured by vernier calipers to the nearest millimeter. Age structure was determined via length-fre- quency analysis for 19 of the populations. Simi- lar information for six of the sites (BN, WF, SP, GC, PY, JL) was available from Appeldoorn (1981), though only the West Falmouth (WF) growth data were used in subsequent analyses since major growth interruptions resulting from pollution events occurred at the other sites. 2 C. W. Brown, Professor, Department of Chemistry, Univer- sity of Rhode Island, Kingston, R.I., pers. commun. May 1979. 76 APPELDOORN: VARIATION IN GROWTH RATE OF MYA ARENARIA ccj u — c E c o S» '> c S- '53 -1-3 T3 C eS be c "a. E I w < E Q. to «- > O i_ c £ TO — CT <° O E # = JS Ed I- CL — 3?I CD a> _ d o> co t; ^ 5 = ra o co Q. — E .1 CD ~ a; q. 5 E Q co 2% w 8 i-'-^'T»-'-T-(DS(\jcoc\j(MO)oinsc\joifiu)ocoo)a) i-i-CMC\JCOCOCOi-CMi-C\JCOC\J^COC\JC\JCOC\JC\JC\ICOC\JCMCM i- o cm 0'-oina)(Dcjc\j 1—1- m^i-coi-m^-m Oi -^ -COCMCOC\ICO cbbcoinr-^moScob r-a)omtx)r^aoa> | tom<3)Oono)'-c\Ttrioco y- cvi i- ^ co W c\i ^ *t ' b i- cm b ^CONSincOCOlDCO '-cocotoommtm i-i-i-i-c\ji-i-b^-" tDO'-OJCDOJi-^S mc7)mm(\cvjijc\j(o i-i-Oi-OOOOCNJO I I I o o o m o m in m m in co m co co o o o b o ooinooooooinmmmmm mmcocMmmmcocococococococo ooooboboobbbooboobb f^r^coinojc\jc\jc\jincoco^-co^rr^inotnc\ji-i-cocococo ssococDiocpO)itc\j( , )incNini-sc\i^'itininit;'tc\ic\j ooT-^obbbi-i-i-i-i-i-i-i-cocococococbbi- • *— intninoooooh-omooocor^-^-cvii-mmoooo (NicNicocnwcNJwnir^^in^-incoinO'-'-'-t-'-'-Trit i- i- i- i- i- CM CNJ oooococococoino^j-inmmmi-i-^r»-i-i-r^cooo inincocowwcviwo^cNjdor-dr-c^sNsscoicDcTiai ^~ t— **"" ^~ t - t- r- i - T~ T~ ^- t- v— ^- i— ^- (0010000>0>- r- "-•-C>JCN*<*i(*)c*)OV(OVlDiDVlO , W .-.-.- .- .- WT-SC^NcOCOCOCONOrOOlCOCAJC^CDCOSCO^fOCDCOa) ininNinojcorocNjcONCNJcooincocncooNCDcoNOincD 0)oco^cococo^rininco(0- h ca ro "- 5 QC ~> - CL CO O CO o ~ C/) o « CO O CO z s c Z 5 en CD CD o CO -* - O a; O > 0)1 w 3 - r o t= o S 3 __ O lli m C- CD E cn O CO E o CO CL CD a s r 5 o D 2? E CO ^ "D CD CC C f~ CO O CN CD CD O) CD CO ^ CO ^> co^ „E CO o — CD i- C CM _-i2» 0) o> c o c -^ >• o = H o E ^ CD -^- c° 5ilu0ll10$ o m , m co tr co SM£M m - <-> *z — *i : ^ in co r» 77 FISHERY BULLETIN: VOL. 81, NO. 1 Length-frequency analysis was chosen because it could be applied to all samples, thus facili- tating the comparison between samples. The use of shell annuli is unreliable south of Cape Cod (Mead and Barnes 1904; Shuster 1951), and MacDonald and Thomas (1980) found little sup- port for the technique in a Prince Edward Island population. Constraints on the sampling design precluded mark-recapture methods. For each population the modes on a length-fre- quency histogram were broken down into a se- ries of normal curves (Tesch 1971; MacDonald and Pitcher 1979) by a Dupont 3 310 Curve Re- solver, an analog computer which allows one to break down a complex distribution into its basic components in a graphical fashion (Appeldoorn 1981). From the resulting graphs the mean and standard deviation of the curve which represents each mode of the histogram can be obtained. The curve resolver also determines the percentage of the whole sample under each curve. Length-frequency analysis assumes that spawning and settlement are discrete relative to growth such that the length distributions of co- horts are separable. Ropes and Stickney (1965), Pfitzenmeyer (1962), and Brousseau (1978) found that periods of both spawning and settlement of each cohort were discrete events. In the latter study, closely spaced cohorts within the same year were separable by length-frequency analy- sis using probability paper. In the present study, discrimination of cohorts within a year class was also possible. By inspection of the histograms and subse- quent age-length curves and through considera- tion of local recruitment processes and sampling efficiency, ages were assigned to each cohort (Brothers 1980; Schnute and Fournier 1980). When possible, results were corroborated by comparing them with previously published age- length data for the same or nearby areas (e.g., Belding 1930; Pfitzenmeyer 1972; Mead and Barnes 1904; Gilfillan and Vandermuelen 1978; Brousseau 1979), by comparison of adjacent areas (e.g., the two Quonochontaug Pond sites), by comparison of multiple samplings (Allen Har- bor, Deer Isle), and by counts of shell annuli (Portland, Deer Isle). The ages assigned were relative rather than absolute; the time beyond the last yearly incre- ment represents the fraction of expected yearly growth already obtained (Appeldoorn 1981). This process results in a smoother growth curve, since it linearizes seasonal growth variations which would otherwise necessitate the use of a more complex growth model (Cloern and Nichols 1978). The analysis of growth differences can be sim- plified by comparing model parameters rather than the direct age-length observations (Rao 1958). Growth was modeled by fitting the von Bertalanffy growth function (VBGF) to the age- length data. The VBGF is described by the equa- tion: L, = U (1 _ -ftt-to) 3 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. where t = time, L, = length at time t, L^ = maxi- mum asymptotic length, K = growth constant, and to = time when L, = 0. The single growth parameter of Gallucci and Quinn ( 1979) is obtain- ed by w = KL X . Recent studies on the statistical comparison of VBGF's (Allen 1976; Bayley 1977; Gallucci and Quinn 1979; Kimura 1980; Misra 1980; Kappen- man 1981) and on the VBGF's biological basis (Pauly 1979, 1981) have removed most of its past criticism (Roff 1980). Dickie (1971) considered the VBGF applicable for modeling population growth even when individual growth did not fit the model. The VBGF has been previously ap- plied to Myo armaria by Munch-Petersen(1973), Brousseau (1979), and Brethes and Desrosiers (1981). The co parameter was chosen for analysis be- cause, as a single parameter, it was easily calcu- lated, tractable to further analysis, statistically comparable, interpretable in both a biological and statistical sense, and more robust than either Kor L^ (Gallucci and Quinn 1979). A major bene- fit of applying the VBGF is that only estimates of length at known time intervals are required to determine K, L tt , and hence co. Absolute age at length is only required to estimate to. However, to is of less importance here, since it is not a mea- sure of growth, but only a location parameter. The VBGF was fitted to the data according to the methods of Gallucci and Quinn (1979), using the NLIN procedure of SAS79 (Helwig and Council 1979) which yielded estimates of the parameters, their asymptotic standard errors, and the correlation coefficient of /f and L x . From these estimates the co parameter and its variance were calculated (Gallucci and Quinn 1979). The regression procedure incorporated the size and 78 APPELDOORN: VARIATION IN GROWTH RATE OF MYA ARENARIA variance of each age mode. Therefore, variation in the original data is reflected in the variance estimates of the model parameters, and poorly represented age modes, where estimates of mean length and variance might be subject to error, are weighted less. The resulting growth curves are based on the assumption that growth varies from year to year only to the extent expected owing to normal fluctuations in growing condi- tions. Hence, they are an estimation of "average" growth within a population, representing an in- tegration of several variable processes affecting growth. The environmental data listed in Table 1 were used to characterize Mya armaria habitats. These data were subjected to principal compo- nents analysis (PCA) to reduce the observed var- iables to a more meaningful and manageable number of factors without excessive loss of infor- mation. PCA locates hidden components which have generated dependence in the observed var- iables (Morrison 1976). Each resulting compo- nent is a composite variable — a linear combina- tion of the original variables. The components are independent and ordered, so that the first component accounts for most of the observed var- iation, the second for most of the residual varia- tion, and so on. The loadings given for each com- ponent represent the correlation coefficient (r) between a variable and a component. The analy- sis was run on the Pearson product-moment cor- relation matrix of the environmental parameters (to allow for standardization of the units of mea- sure) by using the CORR, FACTOR, and SCORE procedures of SAS79 (Helwig and Council 1979). The components produced by PCA are limited by the input data and can only reflect the factors represented by those data. In the present study the selection of factors was constrained by the sampling design, and no direct measurements were made on a number of factors which would be expected to influence growth (e.g., current flow, food concentration). However, several of the factors represent an integration of processes, incorporating factors not measured directly. For example, current flow is represented to some de- gree by tidal range, tidal position, and sediment characteristics (see Discussion). This integration effect will help offset the limitations of the input data. The growth rate parameter was transformed to log 10 (co) for the analysis of growth variations. Since logio(ZO and logio(L J are inversely propor- tional (Pauly 1979), it isfeltthatlogio(aj)isamore suitable measure of growth (Appeldoorn in press). A difference in logio(co) would then indi- cate a fundamental difference in growth — not just a reciprocal change in Kand L^. [See Pauly 1979, 1980 for a discussion of the analogous P = \ogw{K-W x ) parameter of the VBGF for weight.] Variations in growth rate were analyzed using a stepwise functional regression of logio(a;) on the components generated by PCA, where the resid- uals of the regression of the logio(ou) on Compo- nent 1 were regressed against Component 2 and so on. The geometric mean functional regression was deemed appropriate because of variability in both ai and the components, small sample size, and uncertainties about the distribution of the data (Ricker 1973; Laws and Archie 1981). In normal predictive regressions the regression co- efficient (slope) is b\ functional regression yields a coefficient of v — b/r where r is the correlation coefficient. The standard error of v (SE,) equals the standard error of 6(SE/,) and 95% confidence limits on rare approximated by v± 2SE, (Ricker 1973). Estimates of b, r 2 , and SE* were obtained using the GLM procedure of SAS79 (Helwig and Council 1979) and used to calculate v and its 95% confidence limits. The significance of the regres- sion is tested by determining if the confidence limits bracket v = 0. If not, the null hypothesis Ho: v = is rejected. RESULTS The mean lengths at age as determined through length-frequency analysis are given in Appendix Table 1 for the 19 populations analyzed here. The parameters of the VBGF and logio(to) are given in Table 2. Using the 95% confidence limits around logmM, statistically significant growth differences become readily apparent. A functional regression of logioMon latitude yield- ed: logio(to) = 4.8184 - 0.0878 latitude with r = 0.8220. Although the regression accounts for the majority of the observed variation in growth, it does not indicate what underlying processes may be responsible for this relationship. The results of the PCA are shown in Table 3. The terms used in the table follow the definitions in Morrison ( 1976). In order to simplify the table, those loadings <0.30 have been left out, although all variables contribute to all components to some degree. The first five components have been retained and account for 88% of the observed variation. Of these, the first three were exam- ined in greater detail. 79 FISHERY BULLETIN: VOL. 81. NO. 1 Table 2.— Estimates and standard errors for the von Bertalanffy constants. Site code K L^ to Logio(w) + 95% confidence interval TS AR NR RB WP Q1 Q2 SR WK CR AH WF NB EG WC PT Dl SH RS PI 0.2530 0.2740 3016 0.1829 02992 0.1175 0.1069 0.2119 0.1811 0.1997 0.0903 0.0917 0.1532 0.1377 0.1411 1468 0.1255 00565 0.1623 00986 (00597) (0.0520) (0.0162) (00986) (0.0114) (0.0194) (0.0134) (0.0229) (0 0155) (0.0114) (0.0184) (0.0162) (00198) (0 0425) (00246) (0.0077) (0.0114) (00083) (00287) (0 0248) 111.05 107.13 7969 81.50 73.27 93.23 1 1 1 .00 72.34 111 80 97.75 113.20 13673 8928 91.95 87 18 67.91 67 96 135.71 73.13 81 55 (1118) ( 7.22) ( 1.10) (22.08) ( 0.89) ( 9.20) ( 696) ( 2.48) ( 4.21) ( 1.60) (13.11) (1488) ( 4.30) (18.20) ( 7.37) ( 1.39) ( 2.46) (12.34) ( 4.52) (1078) -1.188 -1.440 -0.718 -1.450 -0 400 -1.104 -1.205 -0.445 -0436 -0 990 -1.668 -1.357 -1.571 -0.914 -1.549 0.836 -0.781 -0980 -0.745 -0.171 (0263) (0.268) (0.095) (0.558) (0.058) (0.148) (0.191) (0.225) (0.127) (0.143) (0.288) (0.184) (0.304) (0.186) (0.236) (0.122) (0.218) (0.336) (0.434) (0.432) 1.4486 1.4677 1 3808 1.1734 1.3418 1.0396 1.0743 1.1855 1.3066 1.2905 1.0095 1 0982 1.1360 1.1025 1 .0899 9986 0.9311 08847 1.0754 09053 1 3839 1.4166 1.3473 1.1202 1.3253 09848 1.0045 1.1417 1.2724 1.2512 0.9147 1 0056 1.0902 1 0439 09965 09778 08974 07663 1.0275 07839 1.5050 1.5134 1.4119 1.2207 1 3577 1.0882 1.1344 1.2253 -1.3383 1.3265 -1.0873 -1.1746 -1.1774 -1.1541 1.1668 1.0186 -0.9623 •0.9776 1.1167 09674 Table 3.— Results of the principal components analysis on environmental data. Loadings <0.30 have been omitted for clarity. Environmental Principa components Commu- parameter 1 2 3 4 5 nality Average temperature -0 938 0925 Tidal range 0.806 383 0.819 Tidal position 0817 0.855 Md0 -0.503 0.725 0.891 QD0 0.808 -0.382 0872 Skq0 0.386 -0.858 0.933 % silt-clay 609 0.675 0880 % organic matter 0.521 0.765 0.942 Total hydrocarbons 0.802 0855 Salinity 0.396 0.750 0.840 Eigenvalues 3 588 1.875 1.363 1.134 0.851 % variance 35.9 18.7 13.6 11.3 8.5 % cumulative variance 35.9 546 68 3 796 88.0 The first component is interpreted as repre- senting latitude, since the major contributing variables vary with latitude. Average annual temperature, as might be expected, shows the highest correlation. It decreases with latitude. To avoid confusion the first component will be re- ferred to as "northness." The second component is sediment siltiness. Grain size (negatively cor- related) and percent silt-clay (positively cor- related) are the main contributing variables. The high correlation of salinity may reflect the role of flocculation and estuarine circulation in the distribution of silts and clays in estuarine sediments (Krumbein and Sloss 1963; Knauss 1978). The third component, positively correlated with hydrocarbons and percent silt-clay, is sedi- mentary hydrocarbons. The higher silt-clay com- ponent (also reflected to some degree by positive skewness) provides a greater sedimentary sur- face area for the retention of hydrocarbons (Ly tie and Lytle 1977). The first three components were used for the further analysis of growth to try and deduce fac- tors which could have contributed to the latitudi- nal trend and to point out secondary growth affecting factors. Several sites were omitted from this analysis because missing values pre- cluded the calculation of the component scores. The results of the stepwise regression analysis are given in Table 4. As expected, growth was found to be negatively correlated with northness. The second regression showed a negative rela- tionship between siltiness and growth. The last Table 4.— Results of the stepwise regression of growth (w) on the first three principal components. The slope of the predictive regression (6) can be found by 6 = vr. Regression Intercept v = slope Approximate 95% confidence limits Log,o(w) vs. northness 0.693 1st residual vs. siltiness 0.184 2d residual vs sedimentary hydrocarbons 0217 1.1137 -0.1653 -0.2269 parameter proved useful, both for its de- scription of growth rate and ease of manipula- tion in further analyses. Incorporating both the sample size and variance of the age-length deter- minations into the nonlinear regression protected oj against random errors in the location of the age-length modes. In addition, since w is more robust than either Kor L x , it is further protected against inaccuracies in their estimation — an advantage over using K during the subsequent growth analyses. The trends resulting from the subsequent regressions were also protected from random inaccuracies in the estimation of to by virtue of the large number of sample sites used. However, it is important to recognize that such random errors do exist and not to continue the analysis past its potential limits. ACKNOWLEDGMENTS The author wishes to express his gratitude to those people who assisted in the clam collection, particularly R. L. Dow, M. Richards, J. M. Hick- ey, M. L. H. Thomas, and M. Worobec; and to those people who were involved in the sample processing, notably R. S. Brown, J. Keller, A. Miller, S. Polofsky, C. Brown, and E. Franklin. S. B. Saila, A. N. Sastry, R. C. Bullock, and two journal reviewers gave helpful criticism. Finan- cial support for sample collection was provided by the American Petroleum Institute. LITERATURE CITED Allen, R. L. 1976. Method for comparing fish growth curves. N.Z. J. Mar. Freshwater Res. 10:687-692. American Society for Testing Materials. 1963. Grain-size analysis of soils. D422-63. A.S.T.M. Standards 11:205-216. Appeldoorn, R. S. 1981. Response of soft-shell clam (Mya arenaria) growth to onset and abatement of pollution. J. Shellfish Res. 1: 41-49. In press. Relationships between the parameters of the von Bertalanffy growth function among bivalves as re- vealed with the auximetric grid. [Abstr.] J. Shellfish Res. Bayley, P. B. 1977. A methodfor finding the limits of application of the von Bertalanffy growth model and statistical estimates of the parameters. J. Fish. Res. Board Can. 34:1079- 1084. Beaven, G. F. 1960. Temperature and salinity of surface water at Solo- mons, Maryland. Chesapeake Sci. 1:2-11. Belding, D. L. 1930. The soft-shelled clam fishery of Massachusetts. Mass. Dep. Conserv., Div. Fish. Game, Mar. Fish. Sect., Mar. Fish. Ser. 1, 65 p. Brethes, J.-C. F., and G. Desrosiers. 1981. Estimation of potential catches of an unexploited stock of soft-shell clam (Mya arenaria) from length com- position data. Can. J. Fish. Aquat. Sci. 38:371-374. Brothers, D. T. 1980. Age and growth studies on tropical fishes. In S. B. Saila and P. M. Roedel (editors), Stock assessment for tropical small scale fisheries, p. 119-136. Int. Cent- Mar. Resour. Dev., Kingston, R.I. Brousseau, D. J. 1978. Spawning cycle, fecundity, and recruitment in a population of soft-shell clam, Mya arenaria, from Cape Ann, Massachusetts. Fish. Bull., U.S. 76:155-166. 1979. Analysis of growth rate in Mya arenaria using the von Bertalanffy equation. Mar. Biol. (Berl.) 51:221- 227. Brown, R. S. 1980. The value of the multidisciplinary approach to re- search on marine pollution effects as evidenced in a three-year study to determine the etiology and patho- genesis of neoplasia in the soft-shell clam, Mya arenaria. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 179:125-128. Buchanan, J. B. 1971. Sediments. In N. A. Holme and A. D. Mclntyre (editors). Methods for the study of marine benthos, p. 30- 52. Blackwell Sci. Publ., Oxford. Cloern, J. E., and F. H. Nichols. 1978. A von Bertalanffy growth model with a seasonally varying coefficient. J. Fish. Res. Board Can. 35:1479- 1482. Dickie, L. M. 1971. Addendum: Mathematical models of growth. In W. E. Ricker (editor). Methods for assessment of fish production in fresh waters, p. 126-130. Blackwell Sci. Publ., Oxford. Dow, R. L. 1975. Reduced growth and survival of clams transplant- ed to an oil spill site. Mar. Pollut. Bull. 6:124-125. Dow, R. L., and J. W. Hurst, Jr. 1975. The ecological, chemical and histopathological evaluation of an oil spill site. Part I. Ecological studies. Mar. Pollut. Bull. 6:164-166. Dow, R. L., and D. E. Wallace. 1961. The soft-shell clam industry of Maine. U.S. Fish Wildl. Serv., Circ. 110, 36 p. Gallucci, V. F., and T. J. Quinn II. 1979. Reparameterizing, fitting, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Gilbert, M. A. 1973. Growth rate, longevity and maximum size of Ma- coma balthica (L.). Biol. Bull. (Woods Hole) 145:119- 126. Gilfillan, E. S., D. Mayo, S. Hanson, D. Donovan, and L. C. Jiang. 1976. Reduction in carbon flux in Mya arenaria caused by a spill of No. 6 fuel oil. Mar. Biol. (Berl.) 37:115-123. Gilfillan, E. S., and J. H. Vandermeulen. 1978. Alterations in growth and physiology of soft-shell clams, Mya arenaria, chronically oiled with Bunker C 82 APPELDOORN: VARIATION IN GROWTH RATE OF MYA ARENARIA from Chedabucto Bay, Nova Scotia, 1970-76. J. Fish. Res. Board Can. 35:630-636. Helwig, J. T., and K. A. Council. 1979. SAS user's guide, 1979 ed. SAS Inst. Inc.. Ra- leigh. N.C., 494 p. HiCKS, S. D. 1963. Physical oceanographic studies of Narragansett Bay, 1957 and 1958. U.S. Fish Wildl. Serv., Spec. Sci. Rep.— Fish. 457, 30 p. Jeffries, H. P. 1962. Environmental characteristics of Raritan Bay, a polluted estuary. Limnol. Oceanogr. 7:21-31. Kappenman, R. F. 1981. A method for growth curve comparisons. Fish. Bull., U.S. 79:95-101. Kellogg, J. L. 1905. Conditions governing existence and growth of the soft clam (Mya armaria). Rep. Comm. year ending June 30, 1903, 29:195-224. KlMURA, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull., U.S. 77:765-776. Knauss, J. A. 1978. Introduction to physical oceanography. Prentice- Hall, Inc., Englewood Cliffs, N.J., 338 p. Krumbein, W. C, and L. L. Sloss. 1963. Stratigraphy and sedimentation. W. H. Freeman, San Franc, 660 p. Laws, E. A., and J. W. Archie. 1981. Appropriate use of regression analysis in marine biology. Mar. Biol. (Berl.) 65:13-16. Lytle, J. S., AND T. F. Lytle. 1977. Sediment hydrocarbons as environmental indi- cators in the northeast Gulf of Mexico. In D. A. Wolfe (editor), Fate and effects of petroleum hydrocarbons in marine ecosystems and organisms, p. 404-412. Per- gamon Press, N.Y. MacDonald, B. A., and M. L. H. Thomas. 1980. Age determination of the soft-shell clam Mya are- naria using shell internal growth lines. Mar. Biol. (Berl.) 58:105-109. MacDonald, P. D. M., and T. J. Pitcher. 1979. Age-groups from size-frequency data: A versatile and efficient method of analyzing distribution mixtures. J. Fish. Res. Board Can. 36:987-1001. Marine Research. 1975. Charleston study annual report. April 1974 - March 1975. Mar. Res. Inc., Falmouth, Mass., 241 p. Mead, A. D., and E. W. Barnes. 1904. Observations on the soft-shell clam (fifth paper). R.I. Comm. Inland Fish., 34th Annu. Rep., p. 29-68. MlSRA, R. K. 1980. Statistical comparisons of several growth curves of the von Bertalanffy type. Can. J. Fish. Aquat. Sci. 37:920-926. Morrison, D. F. 1976. Multivariate statistical methods. 2d ed. Mc- Graw-Hill, Inc., N.Y., 415 p. Munch-Petersen, S. 1973. An investigation of a population of the soft clam (Mya arenaria L.) in a Danish estuary. Medd. Dan. Fisk. Havunders., New Ser. 7:47-73. National Ocean Survey. 1978. Tide tables 1979: High and low water predictions. East coast of North and South America, including Greenland. Natl. Oceanic Atmos. Admin., Natl. Ocean Surv., Rockville, Md„ 293 p. Newcombe, C. L. 1936. A comparative study of the abundance and the rate of growth of Mya arenaria L. in the Gulf of St. Lawrence and Bay of Fundy regions. Ecology 17:418-428. Newcombe, C. L., and H. Kessler. 1936. Variations in growth indices of Mya arenaria L. on the Atlantic coast of North America. Ecology 17: 429-443. Pauly, D. 1979. Gill size and temperature as governing factors in fish growth: a generalization of von Bertalanffy's growth formula. Berichte aus dem Institut fur Meere- skunde 63, 156 p. 1980. A new methodology for rapidly acquiring basic in- formation on tropical fish stocks: growth, mortality, and stock recruitment relationships. In S. B. Saila and P. M. Roedel (editors), Stock assessment for tropical small scale fisheries, p. 154-172. Int. Cent. Mar. Resour. Dev., Kingston, R.I. 1981. The relationships between gill surface area and growth performance in fish: a generalization of von Ber- talanffy's theory of growth. Meeresforschung 28:251- 282. Pfitzenmeyer, H. T. 1962. Periods of spawning and setting of the soft-shell clam, Mya arenaria, at Solomons, Maryland. Chesa- peake Sci. 3:114-120. 1972. Tentative outline for inventory of molluscs: Mya arenaria (soft-shell clam). Chesapeake Sci. 13(suppl.): S182-S184. Rao, C. R. 1958. Some statistical methods for comparison of growth curves. Biometrics 14:1-17. RlCKER, W. E. 1973. Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. Ropes, J. W., and A. P. Stickney. 1965. Reproductive cycle of Mya arenaria in New Eng- land. Biol. Bull. (Woods Hole) 128:315-327. ROFF, D. A. 1980. A motion for the retirement of the von Bertalanffy function. Can. J. Fish. Aquat. Sci. 37:127-129. Sameoto, D. D. 1972. Yearly respiration rate and estimated energy bud- get for Sagitta elegans. J. Fish. Res. Board Can. 29: 987-996. Schnute, J., and D. Fournier. 1980. A new approach to length-frequency analysis: Growth structure. Can. J. Fish. Aquat. Sci. 37:1337- 1351. SHOREY, W. K. 1973. Macrobenthic ecology of a sawdust-bearing sub- strate in the Penobscot River estuary (Maine). J. Fish. Res. Board Can. 30:493-497. Shuster, C. N. 1951. On the formation of mid-season checks in the shell of Mya. [Abstr.] Anat. Rec. 111:543. Swan, E. F. 1952. The growth of the clam Mya arenaria as affected by the substratum. Ecology 33:530-534. Tesch, F. W. 1971. Age and growth. In W. E. Ricker (editor), Meth- ods for assessment of fish production in fresh waters, p. 98-130. Blackwell Sci. Publ., Oxford. 83 FISHERY BULLETIN: VOL. 81, NO. 1 Thomas, M. L. H. 1978. Comparison of oiled and unoiled intertidal com- munities in Chedabucto Bay, Nova Scotia. J. Fish. Res. Board Can. 35:707-716. Turner, H. J., Jr. 1948. Report on investigations of the propagation of the soft-shell clam, Mya arenaria. Rep. Invest. Shellfish. Mass. 1:3-9. Welch, W. R. 1981. Monthly and annual means of sea surface tempera- ture: Boothbay Harbor, Maine 1905 through 1980. Maine Dep. Mar. Resour., Res. Ref. Doc. 81/8. Wilton, M. H., and H. I. Wilton. 1929. Conditions affecting the growth of the soft shell clam, Mya arenaria L. Contrib. Can. Biol. Fish. 4:81- 93. Appendix Table 1. — The age (yr), length ±1 standard deviation (mm), and percent of sample in each age class. Percentages may not total 100 due to roundoff or exclusion of some outliers from the analysis. Sample size is given in parentheses. — = undefined. % of % of % of Age Length sample Age Length sample Age Length sample Navesink River (103) Coonamessett River (124) Portlanc (367) 0.67 26.012.3 3 1.15 340 — 0.5 1.5 19010.8 1 1 33 35.2+25 14 2.15 42.8+2.0 4 2.5 26.711.3 4 1.67 42.5+2.0 23 285 52.011.0 7 3.5 31.511.6 17 2.33 47.311.8 18 3.15 56.112.0 7 4.5 37.111.6 21 2.67 51.8+1.4 10 4 15 64.011.4 18 5.5 41.311.6 29 3.33 55.311.4 11 5.15 69.411.2 9 6.5 44.511.2 10 3.67 59.611.6 2 600 72.811.2 16 7.5 48.111.2 15 4.50 62.1 + 1.4 2 7 15 77.411.2 18 95 52.710.7 3 5.50 65.411.2 2 8.15 82.311.4 17 10.5 55.010.5 1 6.50 69.711.5 4 10.15 88211.1 5 11.5 56.710.4 1 8.50 74.711 2 4 11.15 94 - 0.5 13.5 60 210.5 1 9.50 78.011 2 6 Quonochontaug Pond-2 (146) Saugatucket River (140) Deer Isle (318) 2.15 33.712.7 41 20 29 011.3 2 3.0 25.711.9 4 3.15 41.6+2.0 28 26 33.810.9 5 4.0 31.211.1 7 4.15 49.0+1.3 6 3.0 37.611.6 9 5.0 35.211.3 21 5.15 55.111.8 6 3.8 43.312.2 42 6.0 39 2+1 3 18 6.15 60.211 4 4.8 48.911.5 7 7.0 42.1 + 1 1 9 7.15 65811.3 6 58 52.811.1 12 8.0 44.611.0 10 8.15 70.510.8 4 7.0 56.110.7 6 9.0 47.8±1.0 12 9.15 73.510.8 3 8.0 60 011.2 6 10.0 51.5+1.0 9 10.15 79 011.1 4 9.0 64.511.2 3 Quonochontaug Pond-1 (198) Allen Harbor (144) Potato I sland (201) 1.33 23.9+2.1 33 1.85 30 011.8 11 2.5 25.310.7 2 2.33 30.7+2.1 43 2.85 37.712.6 24 3.5 30.911.3 8 3.33 38.311.9 13 3.85 44.511.5 15 4.5 34.611.6 25 4.33 44.911.3 7 4.85 50.411.8 18 5.5 39 211.5 41 5.33 49.611.2 2 5.85 55.211.3 14 6.5 43 7±1.0 13 6.33 54.211.4 3 685 59.711.4 8 7.5 47 110.8 6 733 58.611.4 1 7.85 65.811.4 9 8.5 49.410.9 3 Watchemoket Cove (90) Big Annemessex River (177 Robinston (190) 1.15 27 612.9 24 1.33 57.011.3 14 3.67 37.911.6 8 2.15 34.812.0 31 1.80 60 911.8 30 4.67 42 511.6 16 3.15 42.4±1.7 19 2.33 688+2.0 33 5.67 47.211.6 41 4.15 48 212 5 11 280 737+0.8 4 6.67 51 711.0 20 6.15 57.111.4 8 3.33 77.211.3 7 7.67 54.610.9 7 7 15 62.511.0 6 4.33 86 811.5 2 8.67 57 011.2 4 New Bedford (180) Stockton Harbor (164) Tangier Sound (166) 1.85 32 011.0 05 3.0 31.011.8 2 1.33 546±2.3 36 2.85 43.711.2 3 40 38.211.8 2 1.80 62.011.6 26 3.56 48.311 6 8 5.0 44.8+1.4 11 2.33 67.911.2 13 3.85 51.210.8 15 6.0 49211.1 13 2.80 73.411.3 11 4.56 53 910.9 16 7.0 54.1 ±1.5 17 3.33 78.210.9 3 4.85 56 .210.9 21 8.0 58 211.2 21 3.80 81.610.8 1 5.56 58911.2 16 90 62 811 3 13 4.33 86.210.7 2 5.85 61.210 8 11 10.0 66.511.4 11 6.85 64.410.9 6 11.0 70.311.2 3 Raritan 3ay (200) 7.85 69.010.6 3 12.0 75.011.3 5 1.33 30410.8 3 Winnapau 3 Pond (229) Wickford (203) 1.67 36.411.8 23 09 24.211.6 3 0.20 7.0 — 05 2.33 40.211.7 39 1.5 31.511.0 2 2.00 37.311.1 2 2.67 43811.6 25 19 37.411.2 5 2.67 48.711.4 4 3.33 47 410.8 10 25 41.4+1.2 7 3.00 53.911.4 10 2.9 46.211.3 14 380 60411.8 18 East Greenwich Cove (192) 3.5 50.611.8 24 4.80 68 1 ±2.5 36 1.0 20.813.5 30 4.5 56.0+1.6 29 580 75.1 + 1.8 13 2.0 30811.0 9 5.5 60.511.2 8 600 80211.1 7 3.0 38.412.8 49 6.5 64.111.4 4 7.00 84.710.9 2 4.0 45.011.5 11 7.5 67.210.7 4 800 885106 2 5.0 50.911.1 3 84 BIOCHEMICAL GENETICS OF PACIFIC BLUE MARLIN, MAKAIRA NIGRICANS, FROM HAWAIIAN WATERS 1 James B. Shaklee, 2 Richard W. Brill, 3 and Robin Acerra 4 ABSTRACT An electrophoretic survey of 35 enzyme-coding gene loci in Pacific blue marlin was accomplished to determine levels of genetic variation and the feasibility of using electrophoresis to study stock structure in this species. Polymorphism (P99) in the marlin was 0.26 and the average heterozygosity (H) was 0.06. Allele frequencies at 11 variable loci were determined for a sample of 95 fish from Kona, Hawaii. The observed levels of polymorphism and heterozygosity suggest that a biochemical genetic analysis of blue marlin stock structure is possible and may reveal stock heterogeneity. The Pacific blue marlin, Makaira nigricans, is the predominant billfish species in the central tropical Pacific. As such, it is an important com- mercial species and the object of a considerable sport fishery. The average annual catch of this species in the Pacific exceeds 14,000 t (metric tons) (Shomura 1980). The Pacific blue marlin is primarily distributed in equatorial areas, al- though Japanese longliner catch records indi- cate that its range extends from lat. 48°N to 48°S. During the Southern Hemisphere summer (December through March) a center of concen- tration occurs in the western and central South Pacific (between lat. 8°S and 26°S). In the North- ern Hemisphere summer (May through October) a center of concentration occurs in the central North Pacific (between lat. 2°N and 24 °N). Dur- ing April and November the fish appear to be concentrated equatorially between lat. 10°N and 10°S (Rivas 1975). There is currently no direct evidence of migration of blue marlin within the Pacific. However, a general movement to the northwestern Pacific during the Northern Hem- isphere summer and to the southeastern Pacific during the Southern Hemisphere summer has been postulated by Howard and Ueyanagi (1965) 'Contribution No. 652 from the Hawaii Institute of Marine Biology. 2 Hawaii Institute of Marine Biology and Department of Zool- ogy, University of Hawaii, Honolulu, Hawaii; present address: Division of Fisheries Research, CSIRO Marine Laboratories, 233 Middle Street, Cleveland, Queensland 4163, Australia. :i Pacific Gamefish Foundation, Honolulu, Hawaii; present address: Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 3830. Honolulu, HI 96812. 4 Hawaii Institute of Marine Biology, University of Hawaii. Honolulu, Hawaii; present address: Southampton College, Long Island University, Southampton, NY 11968. on the basis of the shifting abundance patterns of the fish. Little is known about spawning, other than that Pacific blue marlin appear to spawn throughout the year in an area 10°-20° on either side of the Equator, and up to 30° on either side of the E qua- tor during the Northern and Southern Hemis- pheres' respective summer months. In general, the highest spawning densities occur in the west- ern Pacific, with the density decreasing eastward (Strasburg 1970; Matsumoto and Kazama 1974; Rivas 1975). Because of the apparently single equatorial Pacific spawning area, it has been as- sumed that the species consists of a single unit stock (Yuen and Miyake 1980; Yoshida 1981), yet there has been no direct test of this assumed stock structure. The most recent report available on the condition of the Pacific blue marlin stock considers it to be badly overfished. Yuen and Miyake ( 1980) calculated that the present fishing effort (commercial longliner effort only, since no data are available on recreational fishing effort) is about twice that suitable for maximum sus- tainable yield. Because the catch per unit effort of Pacific blue marlin has steadily declined over the past 10 yr, in spite of a fairly constant level of effort, Yuen and Miyake (1980:19) concluded "...that continued fishing at high levels will con- tinue to reduce the abundance of the stock and a recruitment failure will become a distinct possi- bility." The importance of being able to define sub- populations or stocks of fishes with respect to the formulation of appropriate fishery management schemes has long been recognized (Marr 1957). This problem is especially acute for species (such as Pacific blue marlin) which are highly migra- Manuscript accepted June 1982. FISHERY BULLETIN: VOL. 81. NO. 1, 1983. 85 FISHERY BULLETIN: VOL. 81. NO. 1 tory, subjected to an oceanwide multinational fishery, and which, because of relatively low catches, are not well suited to tag-recapture studies. In fact, the pressing need to understand blue marlin stock structure has been recognized for some time (Shomura 1980; Yoshida 1981). The electrophoretic analysis of protein poly- morphisms in natural populations can be a pow- erful approach for analyzing genetic aspects of population structure in sexually reproducing or- ganisms. For this reason, the technique has been applied to the study of racial or subpopulation differentiation in numerous invertebrates and vertebrates (Ayala 1976). Because of the basic importance of information on subpopulation or stock structure to fisheries management (Berst and Simon 1981), population genetic studies have been conducted for many species of fishes [re- viewed by de Ligny (1969) and Allendorf and Ut- ter (1979)]. Most of the fishes investigated to date have been freshwater species or marine forms which are either inshore shallow-water species or demersal species. Stock heterogeneity for oceanic species has not generally been reported (but see Fujino 1976; Fujinoet al. 1981). Although open water, pelagic species may be characterized by large panmictic cosmopolitan populations, this pattern has not yet been clearly established. One problem in test- ing this hypothesis has been the unusually low levels of genetic variability observed to date in several large marine vertebrates such as skip- jack tuna (Fujino 1970) and seals (McDermid et al. 1972; Bonnell and Selander 1974). Indeed, Selander and Kaufman (1973) have even sug- gested that large, mobile vertebrates may gener- ally have low levels of heterozygosity— a charac- teristic which, if true, would preclude definitive stock analysis using electrophoretic techniques (but see Ryman et al. 1980). The general lack of progress in defining stock structure in oceanic fishes, such as scombroids, using electrophoretic methods, is attributable to several factors. Many of the reports in the literature have been prelimi- nary in nature dealing with small samples of fish and few variable loci. Although such small sam- ple sizes are not unexpected given the remote, far-seas nature of many of the commercial fish- eries, they severely limit the subsequent statisti- cal treatment of the data. Similarly, the analysis of only one or two polymorphisms reduces the likelihood of demonstrating any population sub- division which may exist. Finally, the schooling and/or highly migratory nature of many of these fishes makes it difficult to plan and execute ade- quate sampling programs. The study described in the present report was designed to determine the suitability of utilizing electrophoretic techniques to study stock struc- ture in the Pacific blue marlin. Three specific questions were addressed: 1 ) How much and what kind of electrophoreti- cally detectable genetic variation is there in the Pacific blue marlin? Specifically, is there enough genetic variation to allow an electro- phoretic analysis of stock structure in this species? 2) What combinations of enzymes, tissues, and buffer systems can be utilized in a study of genetic variation in this species? 3) What allele frequency distributions charac- terize the population of Pacific blue marlin in Hawaii? MATERIALS AND METHODS Muscle, liver, heart, eye, and brain samples were dissected from Pacific blue marlin landed at the Hawaiian International Billfish Tourna- ment held at Kailua-Kona, Hawaii, in August 1980. All tissue samples were taken immediately after each fish had been weighed, and all fish had been dead for at least 1 h but <8 h. The dissected tissues were initially placed on ice and subse- quently transferred to a freezer within 12 h. The time delay between fish capture and the freezing of dissected tissues did not seem to adversely affect any of the polymorphic enzymes screened with the possible exception of L-iditol dehydro- genase which could only be scored in 84 of the 95 fish analyzed. Tissues were stored frozen at -20°C until extracted. Tissue extracts were prepared by homogeniza- tion using a loose-fitting, motorized stainless steel pestle in polycarbonate centrifuge tubes. The extraction buffer consisted of 0.1M Tris-HCl pH 7.0 containing 1 X 10" 3 M EDTA and 5 X 10~ 5 M NADP\ After homogenization, the ex- tracts were centrifuged at 25,000 X gfor at least 30 min. Supernatants were transferred to indi- vidually labeled glass vials, capped, and stored at — 75°C until the electrophoretic analysis was completed. The supernatants were subjected to horizontal starch gel electrophoresis (modified from Selan- der et al. 1971), using some 15 different buffers. The gels were made using Lot 60F-0558 starch 86 SHAKLEE ET AL.: BIOCHEMICAL GENETICS OF PACIFIC BLUE MARLIN (Sigma Chemical Co., St. Louis, Mo.) at a concen- tration of 12% w/v. After electrophoretic separa- tions, enzyme patterns were visualized using standard histochemical staining recipes modi- fied from Shaw and Prasad (1970), Selander et al. (1971), and Siciliano and Shaw (1976). All zymograms were photographically recorded. Patterns of enzyme variation which were con- sistent with the subunit structure of the enzyme (when known) and simple models of Mendelian inheritance were scored and recorded as geno- types. Names of enzymes and Enzyme Commis- sion numbers follow the recommendations of the Commission on Biochemical Nomenclature ( 1973). For multilocus enzyme systems, loci were given alphabetic designations when appropriate (e.g., Gpi-A) or were simply assigned a number beginning with 1 for the most anodally migrat- ing isozyme. The most common allele at each locus was designated 100, and all other alleles at that locus were numbered according to their elec- trophoretic mobility relative to the 100 allele. Negative numbers refer to alleles with cathodal migration. The putative genotype data were sum- marized as genotype and allele frequency distri- butions. The genotype distributions were exam- ined for internal consistency with the Mendelian inheritance model by chi-square testing of good- ness-of-fit of observed genotype ratios with those expected for a single random mating population in the absence of differential selection among the alleles. The expected ratios were computed from observed allele frequencies using Levene's (1949) unbiased method for small samples. Heterozy- gosity for each locus (h) was calculated as h = 1 -- XX i 2 where Xi is the frequency of the /th allele. Average heterozygosity (H) was calculated as the mean of h over all loci examined. RESULTS Tissue samples from 95 Pacific blue marlin were analyzed. A total of 23 enzyme systems representing 35 gene loci were satisfactorily re- solved using extracts of muscle, liver, and eye (Table 1). Heart and brain tissue did not add sig- nificantly to this total. Eleven loci exhibited de- tectable genetic variation in the sample of 95 fish analyzed. The enzymes adenosine deaminase (Ada), mannosephosphate isomerase (Mpi), and phosphoglucomutase (Pgm) all behaved as mono- mers with two-banded heterozygotes. Aspartate aminotransferase ( Aat-1), alcohol dehydrogenase (Adh), glucosephosphate isomerase (Gpi-A), mus- cle glycerol-3-phosphate dehydrogenase (G-3- Pdh-2), liver isocitrate dehydrogenase (Idh-1), phosphogluconate dehydrogenase (Pgdh), and umbelliferyl esterase (Umb) behaved as dimers exhibiting triple-banded heterozygous patterns. L-iditol dehydrogenase (Iddh), often referred to as sorbitol dehydrogenase in the literature, ap- peared to be a tetramer as heterozygotes exhib- ited a five-banded phenotype. Two of the 1 1 variable loci were represented by only a single heterozygous individual out of the 95 fish screened. The remaining nine loci were TABLE 1.— Electrophoretic analysis of Makaira nigricans from Hawaii. M = muscle, E = eye, L = liver. Enzyme Loci Name (Enzyme Commission number) Abbr. Tissue Invariant Variable aspartate aminotransferase (2.6.1.1) adenosine deaminase (3.5.4.4) alcohol dehydrogenase (1.1.1.1) creatine kinase (2.7.3.2) enolase (4.2.1.11) esterase (3.1.1.—) glyceraldehy de-phosphate dehydrogenase (1.2.1.12) glutamate dehydrogenase (1.4.1.2) glucosephosphate isomerase (5.3.1.9) glycerol-3-phosphate dehydrogenase (1.1.1.8) hexose diphosphatase (3.1.3.11) L-iditol dehydrogenase (1.1.1.14) isocitrate dehydrogenase (1.1.1.42) lactate dehydrogenase (1.1.1.27) malate dehydrogenase (1 .1.1.37) malate dehydrogenase (NADP*) (1.1.1.40) mannosephosphate isomerase (5 3.1.8) peptidase (3.4.11 . — ) phosphogluconate dehydrogenase (1.1.1.44) phosphoglucomutase (2.7.5.1) superoxide dismutase (1.15.1.1) umbelliferyl esterase xanthine dehydrogenase (1.2.1.37) Aat L 1 Ada M Adh L — Ck M + E 2 Eno M 1 Est L 2 Gapdh M 2 Gdh L 1 Gpi M 1 G-3-Pdh M+L 1 Hdp L 1 Iddh L — Idh M+L 1 Ldh M + E 3 Mdh M 3 Mdh(Nadp+) M 1 Mpi M — Pep M 2 Pgdh M — Pgm M Sod L 1 Umb M Xdh L 1 87 FISHERY BULLETIN: VOL. 81, NO. 1 polymorphic by the normal criteria (common allele at a frequency of 0.99 or less) with five loci (Ada, Mpi, Pgdh, Pgm, and Umb) having the most common allele at a frequency of between 0.95 and 0.99 and four loci (Aat-1, Adh, G-3-Pdh- 2, and Iddh) having the most common allele at a frequency of <0.95 (Table 2). All 11 variable loci exhibited two or three alleles except for Aat-1 which had five different alleles, three of which were reasonably common. Heterozygosity values for the individual loci ranged from zero for all of the apparently mono- morphic loci to 0.494 for G-3-Pdh-2. The average heterozygosity (H) across all 35 loci was 0.0605. Where possible, the observed genotype distri- butions were tested for goodness-of-fit to Hardy- Weinberg equilibrium expectations (Gpi-A, Idh- 1, Mpi, and Pgm were not tested because of the very small number of observed variants in the sample). Where necessary, rare alleles were pool- ed prior to the tests. All tests were nonsignificant except that for Adh ( x 2 = 6.97, df = 1; P<0.01) where there was a significant deficiency of het- erozygotes. Further analysis of the Adh data was under- taken to attempt to identify the major contribu- tors) to this significant chi-square value. Since sex linkage of a locus can result in a deficiency of heterozygotes, the sample of blue marlin was subdivided into males and females. A chi-square test of the Adh genotypes of the 81 male fish also revealed a significant deficiency of heterozygotes (x 2 = 9.36, df = 1; P<0.005). Another possible source of the deficiency of heterozygotes could be the pooling of different year classes which actu- ally had different frequencies of the Adh alleles. Indeed, year class fluctuations of allele frequency have been reported in other fishes (Williams et al. 1973; Mitton and Koehn 1975; Smith et al. 1978; Smith 1979). In the absence of growth data for this species, the only subdivision we could make was on the basis of size. The 95 blue marlin were subdivided into two groups: 1) 100-200 lb total weight and 2) 201-450 lb. There were 74 fish in the 100-200 lb group and the statistical analy- sis of this group once again revealed a deficiency of heterozygotes ( x 2 = 9.95, df = 1; P<0.005). The similarity of the results for small fish and for males is not unexpected since all of the female fish (N = 13) were in the large size class (>200 lb). Therefore, the deficiency of heterozygotes seems to characterize the overall sample and cannot be attributed to sexual or gross age (= size) differences. Table 2.— Allele frequencies and heterozygosities for 11 variable loci in Makaira nigricans from Hawaii. Locus' (heterozygosity) Allele Frequency Aat-1 250 0.016 (h = 0.4497) 145 0.145 100 0.720 27 0.102 -30 0.016 Ada 107 0005 (h = 0.0716) 100 0.963 92 0.032 Adh -220 0.344 (h = 0.4556) -100 0.656 Gpi-A 100 0995 (h = 0.0099) 86 0.005 G-3-Pdh-2 100 0.595 (h = 0.4944) 75 0389 60 0.016 Iddh 147 0.006 (h = 0.4219) 100 0.702 22 0.292 ldh-1 100 0.995 (h = 0.0099) 84 0005 Mpi 104 0005 (h = 0.0316) 100 0.984 90 0.011 Pgdh 155 0.016 (h = 0.0529) 100 0.093 67 0.011 Pgm 144 0.011 (h = 0.0316) 100 0.984 33 0005 Umb 100 0.953 (h = 0.0896) 87 0.047 'Sample size = 95 fish for each locus ex- cept for Mpi where N = 94, Aat-1 and Adh where N = 93, and Iddh where N = 84. DISCUSSION In spite of the low levels of genetic variation re- ported in the literature for skipjack tuna(Fujino 1970; Fujino et al. 1981; Lewis 1981; Richardson in press) and suggested by Selander (1976:34) that "...levels of variability are unusually low in large marine vertebrates such as tuna fish and porpoises," the above data clearly indicate that the Pacific blue marlin does not have abnormally low levels of genetic variation. The level of poly- morphism (P.99) observed for marlin in the pres- ent study (P = 0.26; i.e., 9 out of 35 loci), although slightly lower than the average for fish (P = 0.31) reported by Selander (1976), is higher than the averages for reptiles, birds, and mammals (0.23, 0.15, and 0.23, respectively). Furthermore, the average heterozygosity (H) of 0.0605 for the 35 loci screened in the Pacific blue marlin is greater than the average of 0.0494 calculated by Nevo (1978) for 135 species of vertebrates and the aver- age of 0.0478 calculated by Winans (1980) for 82 species of fishes. Although perhaps somewhat un- 88 SHAKLEE ET AL.: BIOCHEMICAL GENETICS OF PACIFIC BLUE MARLIN expected, the relatively high level of genetic variation reported above for blue marlin is not unique among large marine vertebrates as sev- eral other scombroid fishes (e.g., white marlin, southern bluefin tuna, and Spanish mackerel) ex- hibit similar or even higher levels of variation (Edmunds 1972; Smith and Jamieson 1980; Lew- is 1981; Shaklee unpubl. data). The observed pattern of genetic variation can be used to subdivide the 11 variable loci into two general categories. Four loci (Aat, Adh, G-3-Pdh- 2, and Iddh) form one group characterized by high heterozygosities (0.4219-0.4944) due to the presence of at least two relatively common alleles. The second group, which is composed of the re- maining seven variable loci (Ada, Gpi-A, Idh-1, Mpi, Pgdh, Pgm, and Umb), is characterized by low heterozygosity per locus (0.0099-0.0896) and the presence of a single common allele at a fre- quency of at least 0.95. In reality, both groups of loci are of utility in population analyses because the power of the statistical tests for detecting sig- nificant differences in allele frequency between pairs of samples actually increases somewhat as the frequencies approach the extremes (i.e., and 1.0) compared with samples having fre- quency distributions close to 0.5. The close agreement between observed and ex- pected genotypic frequencies for all but one of the variable loci is consistent with all 95 fish analyzed belonging to a single panmictic popula- tion. However, the significant (P<0.01) defi- ciency of heterozygotes observed at the Adh locus is not. Although this observation may be due to selection or simply be an anomaly, another poten- tial explanation is that this heterozygote defi- ciency is due to the mixing of two or more differ- ent stocks of blue marlin which have different frequencies of the two Adh alleles (Wahlund effect). Such potential stock mixing would not be unreasonable given the presumed migratory na- ture of Pacific blue marlin. The fundamental significance of the above ob- served levels of genetic variation is that they are adequate to allow a biochemical genetic analysis of stock structure in Pacific blue marlin. We are now in the process of initiating just such an anal- ysis and are employing an experimental design which should allow us to detect stock heteroge- neity which has either a stable geographical basis or a temporally shifting geographic basis. The former basis for stock heterogeneity, namely the localization of two or more stocks in different re- gions of the species' range is by far the most com- monly observed form of population subdivision among organisms. We will be testing for this type of heterogeneity by analyzing samples from different localities (Hawaii, Guam, Samoa, etc.) throughout the range of the Pacific blue marlin. However, this type of analysis is complicated by limited access to marlin caught in many areas of the range, by the shifting patterns of abundance which characterize this species, and by the vir- tual impossibility of obtaining simultaneous sam- ples of blue marlin from multiple localities throughout the range. Indeed, it is the apparent migratory nature of the species which suggests that the second type of population structuring, that based on both temporal and geographic iso- lation, may be occurring in this species. Our ap- proach to this problem is to sample continuously in the Hawaiian Islands to look for significant seasonal shifts in allele frequency such as might be expected if different stocks of blue marlin mi- grate past the Hawaiian Islands at different times of the year. Given the present potentially overfished nature of billfish stocks and the diffi- culties associated with alternative forms of stock analysis such as tag-recapture studies, this bio- chemical genetic approach may well represent the only practical means of gathering informa- tion on stock structure in a time frame compat- ible with the urgent need for the formulation of meaningful management programs for this spe- cies. ACKNOWLEDGMENTS We wish to gratefully acknowledge the support and encouragement provided by the staff and Board of Governors of the Hawaiian Interna- tional Billfish Association and the cooperation of all the participants in the 1980 Hawaiian In- ternational Billfish Tournament. The valuable help provided by Jerry Kinney and his staff at the Volcano Isle Fish Co. is also sincerely appre- ciated. LITERATURE CITED Allendorf, F. W.. and F. M. Utter. 1979. Population genetics. InW. S Hoar, D. J. Randall, and J. R. Brett (editors). Fish physiology, Vol. 8, p. 407- 454. Academic Press, Inc., N.Y. Ayala, F. J. (editor). 1976. Molecular evolution. Sinauer Assoc, Inc., Sun- derland, Mass., 277 p. Berst, A. H.. and R. C. Simon. 1981. Introduction to the Proceedings of the 1980 Stock Concept International Symposium (STOCS). Can. J. 89 FISHERY BULLETIN: VOL. 81. NO. 1 Fish. Aquat. Sci. 38:1457-1458. BONNELL, M. L., AND R. K. SELANDER. 1974. Elephant seals: Genetic variation and near extinc- tion. Science (Wash.. D.C.) 184:908-909. Commission on Biochemical Nomenclature. 1973. Enzyme nomenclature. American Elsevier Publ. Co., Inc., N.Y., 443 p. De Ligny, W. 1969. Serological and biochemical studies on fish popula- tions. Oceanogr. Mar. Biol. Annu. Rev. 7:411-513. Edmunds, P. H. 1972. Genie polymorphism of blood proteins from white marlin. U.S. Dep. Inter., Fish Wildl. Serv., Bur. Sport Fish. Wildl., Res. Rep. 77:1-15. Fujino, K. 1970. Immunological and biochemical genetics of tunas. Trans. Am. Fish. Soc. 99:152-178. 1976. Subpopulation identification of skipjack tuna speci- mens from the southwestern Pacific Ocean. Bull. Jpn. Soc. Sci. Fish. 42:1229-1235. Fujino, K., K. Sasaki, and S. Okumura. 1981. Genetic diversity of skipjack tuna in the Atlantic, Indian, and Pacific Oceans. Bull. Jpn. Soc. Sci. Fish. 47:215-222. Howard, J. K., and S. Ueyanagi. 1965. Distribution and relative abundance of billfish- es (Istiophoridae) of the Pacific Ocean. Stud. Trop. Oceanogr., Inst. Mar. Sci. Univ. Miami 2, 134 p. Levene, H. 1949. On a matching problem arising in genetics. Annu. Math. Stat. 20:91-94. Lewis, A. D. 1981. Population genetics, ecology and systematics of Indo-Australian scombrid fishes, with particular refer- ence to skipjack tuna (Katsutvonus pelamis). Ph.D. Thesis, Australian National Univ., Canberra, 314 p. Marr, J. C. 1957. The problem of defining and recognizing subpopu- lations of fishes. U.S. Fish Wildl. Serv., Spec. Sci. Rep.— Fish. 208:1-6. Matsumoto, W. M., and T. K. Kazama. 1974. Occurrence of young billf ishes in the central Pacific Ocean. In R. S. Shomura and F. Williams (editors), Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 2, p. 238- 251. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. McDermid, E. M., R. Ananthakrishnan, and N. S. Agar. 1972. Electrophoretic investigation of plasma and red cell proteins and enzymes of Macquarie Island elephant seals. Anim. Blood Groups. Biochem. Genet. 3:85-94. Mitton, J. B.. and R. K. Koehn. 1975. Genetic organization and adaptive response of allo- zymes to ecological variables in Fundvlus heteroclitus. Genetics 79:97-111. Nevo, E. 1978. Genetic variation in natural populations: Patterns and theory. Theor. Pop. Biol. 13:121-177. Richardson, B. J. In press. The distribution of protein variation in skipjack tuna (Katsuwonus pelamis) from the central and south- western Pacific. Aust. J. Mar. Freshw. Res. 34. Rivas, L. R. 1975. Synopsis of biological data on blue marlin, Maka ira nigricans Lacepede, 1802. In R. S. Shomura and F. Williams (editors). Proceedings of the International Bill- fish Symposium, Kailua-Kona, Hawaii. 9-12 August 1972. Part 3, p. 1-16. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. Ryman, N., C. Reuterwall, K. Nygren, and T. Nygren. 1980. Genetic variation and differentiation in Scandi- navian moose (Alces alces): Are large mammals mono- morphic? Evolution 34:1037-1049. Selander, R. K. 1976. Genie variation in natural populations. In F. J. Ayala (editor), Molecular evolution, p. 21-45. Sinauer Assoc, Inc., Sunderland, Mass. Selander, R. K., and D. W. Kaufman. 1973. Genie variability and strategies of adaptation in animals. Proc. Natl. Acad. Sci. USA 70:1875-1877. Selander, R. K., M. H. Smith, S. Y. Yang, W. E. Johnson, and J. B. Gentry. 1971. Biochemical polymorphism and systematics in the genus Peromyscus. I. Variation in the old field mouse (Peromyscus polionotus). Stud. Genet. VI, Univ. Texas Publ. 7103:49-90. Shaw, C. R., and R. Prasad. 1970. Starch gel electrophoresis of enzymes— a compila- tion of recipes. Biochem. Genet. 4:297-320. Shomura, R. S. (editor). 1980. Summary report of the billfish stock assessment workshop Pacific resources. U.S. Dep. Commer.. NOAA-TM-NMFS-SWFC-5. Siciliano, M. J., and C. R. Shaw. 1976. Separation and visualization of enzymes in gels. In I. Smith and J. W. T. Seakins (editors), Chromato- graphic and electrophoretic techniques Vol. 2, Zone elec- trophoresis, 4th ed., p. 185-209. Halsted Press. Smith, P. J. 1979. Esterase gene frequencies and temperature rela- tionships in the New Zealand snapper, Chrysophrys auratus. Mar. Biol. (Berl.) 53:305-310. Smith, P. J., and A. Jamieson. 1980. Protein variation in the Atlantic mackerel Scomber scombrus. Anim. Blood Groups Biochem. Genet. 11: 207-214. Smith, P. J., R. I. C. C. Francis, and L. J. Paul. 1978. Genetic variation and population structure in the New Zealand snapper. N.Z. J. Mar. Freshw. Res. 12: 343-350. Strasburg, D. W. 1970. A report on the billfishes of the central Pacific Ocean. Bull. Mar. Sci. 20:575-604. Williams, G. C, R. K. Koehn, and J. B. Mitton. 1973. Genetic differentiation without isolation in the American eel, Anguilla rostrata. Evolution 27:192-204. WlNANS, G. A. 1980. Geographic variation in the milkfish Chanos chan- os. I. Biochemical evidence. Evolution 34:558-574. Yoshida, H. O. 1981. Status report: Pacific striped, blue, and black mar- lins. In Status reports on world tuna and billfish stocks, p. 277-300. U.S. Dep. Commer., NOAA-TM-NMFS- SWFC-15. Yuen, H. S. H., and P. Miyake. 1980. Rapporteurs' reports, blue marlin, Makaira nigri- cans. In R. S. Shomura (editor), Summary report of the billfish stock assessment workshop Pacific resources, p. 13-19. U.S. Dep. Commer., NOAA-TM-NMFS- SWFC-5. 90 STOCHASTIC AGE-FREQUENCY ESTIMATION USING THE VON BERTALANFFY GROWTH EQUATION Norman VV. Bartoo 1 and Keith R. Parker 2 ABSTRACT The method of estimating age frequency from length frequency via the von Bertalanffy growth equation is deterministic and yields biased results. Most of the bias can be removed by incorporating a stochastic element in the von Bertalanffy relationship. The stochastic element is based on esti- mated probabilities of lengths by intervals at age, the probabilities being estimated from variances in lengths-at-age. Based on age-length samples from the Pacific bonito fishery the stochastic method gives improved age-frequency estimates over those obtained by the deterministic method. The sto- chastic application may be generalized to all growth models including discontinuous growth such as in crustaceans. Complex population dynamics techniques rely heavily on age-structure information. Frequent- ly, appropriate assessment techniques for a stock require an estimate of the age frequency of that stock. For example yield-per-recruit analysis (Ricker 1958) is computed on the dynamic rela- tionship between growth and mortality: Mortal- ity rates when computed via cohort analysis (Murphy 1965) are based on estimated age fre- quency. For some species accurate aging methods are not available. When feasible, determining the age of fish and consequently computing an age frequency are most accurately accomplished by visual inspection of scales, otoliths, or other struc- tures (Ricker 1958). Such visual inspection is time consuming and often expensive. To reduce the cost and time of estimating the age structure of a fisheries catch, age frequency is usually esti- mated from sampled length frequency, the age- length relationship being described by either an age-length key or a growth curve such as the von Bertalanffy growth curve (Ricker 1958). The growth curve method is used when there are in- sufficient data to construct an age-length key. Age-length keys work on the principle that age can be estimated from length using information contained in a previously or concurrently aged sample from the population. As long as the pro- portion of length-at-age remains the same for all ages, then the age-length key will yield unbiased 'Southwest Fisheries Center La Jolla Laboratory. National Marine Fisheries Service, NOAA, 8604 La Jolla Shores Drive, La Jolla, CA 92038. 2 1837 Puterbaugh Street, San Diego. CA 92103. estimates of age for any sampled lengths from that population. However, since the estimated parameters of an age-length key — proportions of age-at-length— are dependent on the sampled population used to construct the key, the applica- tion of the key to the population with altered age structures can yield inaccurate results. Kimura (1977) and later Westrheim and Ricker (1978) demonstrated that under conditions of varying year-class strength and substantial overlap of lengths between ages, age-length keys can yield nearly useless estimates of numbers-at-age. Clark (1981) effectively removes age-length key bias by first proportioning numbers in length intervals at age over time and then using the ma- trix of these proportions standardized over time to compute least-squares estimates of age fre- quency from the vector of length frequency. Effective applications of many stock assessment growth and mortality based methods require that ages are expressed in fractions of years (Ricker 1958; Lenarz et al. 1974). The large num- ber of aged fish required to construct a sufficient key for a large number of ages is difficult and ex- pensive to attain. Even with Clark's bias correc- tion procedure, the construction of a sufficient key can present difficulties due to data needs. In this paper we deal specifically with the von Bertalanffy growth equation and the application of stochastic methods to reduce or eliminate bi- ases. However, it should be noted that the method presented here may be applied to any growth equation as well as to cases where no growth equation has been fitted or where growth is dis- continuous as in crustaceans. The von Bertalanffy growth equation mathe- Manuscript accepted June 1982. FISHERY BULLETIN: VOL. 81. NO. 1. 1983. 91 FISHKRY BULLETIN: VOL. 81. NO. 1 matically models the relationship between age and length, length being the dependent variable (see Equation (1)). As suggested by Gulland (1973), age is estimated from length by algebrai- cally rearranging the growth equation so that age is the dependent variable (see Equation (2)). Re- gardless of whether length or age is the depen- dent variable, the von Bertalanffy relationship is deterministic: There is a one-to-one correspon- dence between age and length. For the von Bertalanffy growth equation, age frequency is estimated from a length sample as follows: 1) For each length compute the corresponding age. 2) For each age interval, usually the interval between midpoint ages of adjacent ages, sum the number of aged fish falling within the in- terval. 3) The age frequency is then the total number of aged fish falling within each age interval. Use of the von Bertalanffy growth equation for age-frequency estimation results in several types of biases, different from those inherent in age- length keys. This paper documents these biases and proposes a method for their resolution. BIASES When growth is modeled according to the von Bertalanffy age-length relationship (Brody 1945; Ricker 1958), Lt = L x (1 -exp[-k(t - h)]), (1) then age, t, can be converted to length: t = h + \n (\ - L t IU)l(-k) (2) where Lt L x k to length at age t the asymptotic length the rate at which length reaches L hypothetical age at which fish would have zero length. When computing numbers-at-age from Equa- tion (2), estimation bias occurs. One bias is due to L x being a fitted parameter. Thus, all numbers- at-length greater than L x must either be elimi- nated or arbitrarily distributed to older ages. Bias also results when lengths approach L^ and are mathematically allocated to ages above those attainable by fish within the stock. As lengths (L) approach L x , Equation (2) will yield unreason- ably old ages. Additional bias results from the deterministic nature of the von Bertalanffy equation: Back cal- culations of length to age, Equation (2), are on a one-to-one basis. Thus, for any length there is a determined age. In reality, there can be a num- ber of possible ages for any given length, the most probable age-at-length being that with the highest relative contribution of numbers-at- length. Since these back calculations are without probabilistic arguments, the determined age is not necessarily the most probable for the given length. Back calculations of length to age also result in a mathematical estimation bias due to the switch- ing of independent and dependent variables in going from Equation (1) to Equation (2). The de- gree of bias is likely to be a function of the amount of residual error in fitting Equation (1). The bias will probably not be consistent between cases and the degree of bias will have to be con- sidered separately for each case. Consequently this bias is not specifically dealt with in this paper. A computer model can readily demonstrate this bias. For von Bertalanffy parameters: L x = 90.0 units, to = 0.0 units, and k = 0.30, predeter- mined numbers-at-age are arbitrarily distrib- uted normally with a standard deviation equal to 3 units about the von Bertalanffy length-at-age, Equation (1), for ages I through X. A length-fre- quency vector is then generated by 1) multiply- ing the number-at-age times the probability of age occurring within each 0.5 unit length inter- val, thus for each age generating a vector of num- ber-at-length for length intervals between and 100 units, and 2) accumulating numbers-at- length for each length interval over all ages. The numbers-at-age are then deterministically esti- mated from Equation (2) by accumulating num- bers-at-length over the length intervals at age. The bias from this model is illustrated in Table 1. Input and back-calculated numbers-at-age and their differences are listed in columns 2, 3, and 4, respectively. The input numbers-at-age represent a sample age distribution where either catchability, recruitment, mortality, or some combination thereof, are age-class variant. Dif- ferences, column 4, indicate a strong bias which increases with overlap of length distributions at age. One hundred and eleven fish were aged to be 92 BARTOOancl PARKER: STOCHASTIC AOE-FREUUKNCY ESTIMATION Tablk 1.— Input and estimated numbers-at-age for both the deterministic (col. 3) and stochastic (col. 5) models, with the in- put numbers-at-age in column 2. The differences between the input numbers-at-age and the deterministic estimates are given in column 4. Numbers -at-age Age Input Deterministic Difference Stochastic (1) (2) (3) (4) (5) I 200 199 1 200 II 400 399 1 400 III 800 760 40 800 IV 200 267 -67 200 V 600 441 159 600 VI 300 378 78 300 VII 400 320 80 400 VIII 300 258 42 300 IX 100 164 -64 100 X 100 68 32 100 >x — 111 - 111 — Inf. — 35 -35 — greater than the maximum age, 10. Thirty-five had lengths greater than L x and, consequently, were not classifiable. BIAS RESOLUTION With estimated variance of length-at-age a sto- chastic model can be built from the von Berta- lanffy relationship: For any age the probability of a specific length interval is the probability of that interval taken over all length intervals con- taining that age. Thus for all ages a probability matrix ("P"-matrix) of dimension r by c can be computed, where r = the number of rows, or length intervals, and c — the number of columns, or ages, then P (1,1) = (max. length, min. age). If the number-at-age vector is "a" (ai = (min. age)) and the number-at-length vector is L (L\ = (max. length)), then P a = L. (3) And as long as r > c then the numbers-at-age vec- tor can be uniquely solved via least-squares: a = (P'PY'P'L (4) Applying this stochastic method (Equation (4)) to the previous example, the numbers-at-age generated from the number-at-length vector are given in column 5 of Table 1. Since the probabili- ties of the P-matrix are the same as those used to generate the number-at-length vector, it is not surprising that the solution yields unbiased re- sults. This computed example illustrates that the stochastic method yields unbiased estimates of age frequency. PACIFIC BONITO For the Pacific bonito, Sarda chiliensis, of the eastern tropical Pacific, Campbell and Collins (1975), using ages determined from otoliths, esti- mated the von Bertalanffy growth parameters to be L^ = 76 87 cm, t = -0.785 yr. and k = 0.6215. Numbers-at-length for 1 cm intervals for ages I through V are shown in Figure 1 with the corre- sponding length-frequency plot in Figure 2. These numbers represent the 1973 catch from California waters and are a subset of the data used to estimate the von Bertalanffy parameters. This example serves to demonstrate bias and illustrate application of the stochastic method. If desired a variance-covariance matrix can be gen- erated (Draper and Smith 1981) to estimate pre- cision in the resulting age structure. 80 r- 70 - E o h- 60 O z HI 50 - 40 1 2 3 5 2 1 9 11 8 ^S^lV, 5 >*^ 12 1 — 1 6* r 4 2 A 2 1 3/ 9* 15 3 1 /38 /41 8 2 1 J / 25 / 15 ' 3 1 — 1 / 13 / 26/ 44/ 48 71 100=76.870111 T=-0.785yr ft 13 3 K = 0.6215 2 1 1 1 i i IV V AGE (years) Figure 1.— Numbers-at-age by length in 1 cm intervals for the Pacific bonito. Sarda chiliensis. Data from 1973 California landings (Campbell and Collins 1975). The length-frequency information and von Bertalanffy parameters are used to generate both deterministic and stochastic estimates of numbers-at-age. The estimated length-at-age 93 FISHERY BULLETIN: VOL. 81, NO. 1 80 r LENGTH (cm) Figure 2.— Length frequency for the Pacific bonito, Sarda chiliensis, from 1973 California landings (Campbell and Col- lins 1975). and sample standard deviations are Age Length Standard deviation I 51.5 cm 2.7 cm II 63.3 cm 2.1 cm III 69.5 cm 2.3 cm IV 72.9 cm 2.0 cm V 74.8 cm 1.9 cm Deterministic estimates were made on lengths rounded to the nearest 0.1 cm. From Equation (2) the deterministic numbers-at-age are shown in column 3 of Table 2 with the difference between the true and estimated numbers in column 4. While the estimates are reasonably close over the first two ages, they become increasingly dispar- ate for older ages. Thirteen fish had lengths greater than those at the maximum age. Seven had lengths greater than L and consequently were unclassifiable. Quarter centimeter intervals were used to com- pute the stochastic estimates of age frequency. The results are shown in column 6 of Table 2, with the difference between true and estimated numbers in column 7. For all ages, especially the older ages, the stochastic estimates are closer and less biased than those of the deterministic method (column 3). Some insight into the improvement of the sto- chastic age estimates over the deterministic age estimates can be gained by inspection of Figure 1. Lengths of age I fish overlap those of age 1 1 fish and vice versa. Since the deterministic cutoff point for age I fish is 58.7 cm (1.5 yr), all overlap is lost in the deterministic model. In contrast, for the stochastic model, overlaps in lengths-at-age are shared between ages, the degree of sharing being relative to the probabilities of length inter- vals at the respective ages. With increasing age, the extent of relative overlap and, consequently, misaging increases for the deterministic model; allocation of lengths to ages becomes more sensitive. Only if the de- gree of overlap between adjacent ages is equal do accurate estimates of numbers-at-age result from the deterministic model. In the present ex- ample varying year-class strength and random variability in lengths-at-age offset this sensitive compensatory mechanism needed for accurate estimation with the deterministic model. Fish lengths above 75.3 cm, the length at age 5.5 yr, are misclassified either as older or of in- finite ages for the deterministic model. Since for the stochastic model probabilities of length inter- vals at age exist for all ages and lengths, even for lengths above L x , fish at lengths above the 75.3 cm cutoff point are distributed to all ages rela- tive to their respective probabilities for length intervals. DISCUSSION Calculations of age from length via the von Bertalanffy growth equation result in several Table 2.— Deterministic (col. 3) and stochastic (col. 6) estimates of numbers- at-age with their respective differences from the true numbers-at-age in col- umns 4 and 7 for the Pacific bonito. Sarda chiliensis. from 1973 California landings (Campbell and Collins 1975). Numbers-at ■age Deter- Differ- % Dif- Sto- Differ- % Dif- Age True ministic ence ference chastic ence ference (D (2) (3) (4) (5) (6) (7) (8) I 424 411 13 3 1 415 9 2.1 II 158 167 -9 -5.7 162 -4 -2.5 III 54 39 15 27.8 49 5 93 IV 80 71 9 11 3 85 -5 -6.3 V 21 29 -8 38.1 26 -5 -238 >v — 13 13 Inf — — 0.0 Inf. — 7 -7 Inf. — — 00 94 BARTOO and PARKER: STOCHASTIC AGE-FREQUENCY ESTIMATION types of bias. The degree of bias is proportional to overlap in lengths-at-age and changes with weak or strong year classes. When overlap increases with age, age-frequency estimates will generally be more biased for older than younger ages. When overlap occurs, biases will always result, since the numbers-at-length will be allocated to unreasonably old ages. Any numbers-at-length for lengths greater than L x will be undetermined in age estimation, resulting in downward biases for those ages contributing such lengths. Age estimation biases can be effectively re- moved by creating a stochastic model based on a matrix of length interval probabilities at age. The probability matrix (P-matrix) is indepen- dent of year-class strength and will effectively remove all sources of estimation bias except that due to random variation in length-frequency esti- mation. As long as the von Bertalanffy growth parameters remain the same over time, the sto- chastic method based on accurate estimates of variance in length-at-age will always yield un- biased results. A probability model of the distribution of length-at-age with estimated parameters is nec- essary for estimating probabilities of length intervals at age for the P-matrix. If age infor- mation is unavailable then variances can be esti- mated from visually separable length-frequency modes. In the case where modes are separable for the first few ages only, there will be a problem in estimating variances for older ages: A model re- lating the variance in length-at-age with age can be used in estimating variances for these older ages. Ricker (1969) proposed that while distribu- tions in lengths-at-age remain normal, variances increase during the first few years, stabilize, and then decrease over the final years. The trend in variances with age for a similar species might also be substituted in cases where variances are unavailable. The principal strengths of the stochastic meth- od are that few fish are required to be aged to estimate the P-matrix and that existing von Ber- talanffy growth relations can be used. Accurate estimates of variance in length-at-age can prob- ably be achieved with as few as 20 to 30 fish/ age, which is likely to be a much smaller number of fish than needed to estimate accurate propor- tions of age-at-length necessary to construct an age-length key. Von Bertalanffy growth parameters have been estimated for many species. Since most stocks have variant year-class strength, overlaps in lengths-at-age, and lengths exceeding the up- per bound for the last age attainable, conversion to a stochastic model may be necessary, if unbi- ased estimates of age frequency are desired. Re- examination of age-length data used to estimate the von Bertalanffy parameters may be useful in estimating variances in lengths-at-age for the P- matrix. Taking additional age-length samples may be a cost-effective way of improving age-fre- quency estimation. In fishery management, the overestimation of maximum age by the deterministic von Berta- lanffy equation may produce underestimates of mortality rates which may result in overesti- mates of population size and recruitment. Fur- ther, the deterministic method tends to "fill in" weak year classes which results in underesti- mates of year-class variability and overestimates of recruitment stability. In general, all of these affect accuracy of a stock assessment and con- tribute to improper advice for fishery manage- ment. Application of the stochastic method shown here to cover other growth equations and situa- tions, such as discontinuous growth, is handled by simply estimating appropriate elements in the P-matrix for each case. ACKNOWLEDGMENTS We thank Douglas Chapman and Alec MacCall for helping to define the problem and evaluating the solution. Mark Farber, Joseph Powers, Lewis J. Bledsoe, and Gary Sakagawa provided critical review and comment for which the authors are grateful. We are also grateful to R. A. Collins of the California Department of Fish and Game for providing data. LITERATURE CITED Brody, S. 1945. Bioenergetics and growth. Reinhold Publ. N.Y.. 1023 p. Campbell, G., and R. A. Collins. 1975. The age and growth of the Pacific bonito, Sarda chiliensis, in the eastern North Pacific. Calif. Fish Game 61:181-200. Clark. W. G. 1981. Restricted-least squares estimate of age composi- tion from length composition. Can. J. Fish. Aquat. Sci. 38:297-307. Draper, N. R., and H. Smith. 1981. Applied regression analysis. 2d ed. John Wiley and Sons, Inc.. N.Y.. 709 p. Gulland, J. A. 1973. Manual of methods for fish stock assessment. Part 95 FISHERY BULLETIN: VOL. 81, NO. 1 1. Fish population analysis. FAO Man. Fish. Sci. 4. Board Can. 22:191-202. 154 p. RlCKER, W. E. KiMi'RA, D. K. 1958. Handbook of computations for biological statistics 1977. Statistical assessment of the age-length key. J. of fish populations. Fish. Res. Board Can. Bull. 119, Fish. Res. Board Can. 34:317-324. 300 p. Lenarz, W. H., W. W. Fox, Jr., G. T. Sakagawa, and B. J. 1969. Effect of size-selective mortality and sampling Rothschild. bias on estimates of growth, mortality, production, and 1974. An examination of the yield per recruit basis for a yield. J. Fish. Res. Board Can. 26:479-541. minimum size regulation for the Atlantic yellowfin WESTRHEIM, S. J., AND W. E. RlCKER. tuna, Thunnus albacares. Fish. Bull., U.S. 72:37-61. 1978. Bias in using an age-length key to estimate age-fre- Murphy, G. I. quency distributions. J. Fish. Res. Board Can. 36:184- 1965. A solution of the catch equation. J. Fish. Res. 189. 96 AGE, GROWTH, AND MORTALITY OF KING MACKEREL, SCOMBEROMORUS CAVALLA, FROM THE SOUTHEASTERN UNITED STATES 1 Allyn G. Johnson, William A. Fable, Jr., Mark L. Williams, and Lyman E. Barger 2 ABSTRACT Age, growth, and mortality of king mackerel, Scomberomorus en ml In, from the southeastern United States were studied. Otoliths from 1,449 fish were used to estimate age composition, growth rates, and mortality rates of this species. Age composition varied between locations (Texas, Louisiana, Florida, South Carolina, and North Carolina). The majority of older fish were found in Louisiana waters. The oldest females were 14 + years old and the oldest males were 9+ years old. Compensatory growth was found in both sexes. The von Bertalanffy growth equations were as follows: Males (all areas) 1, =965(1 — e' ' 17) ); females from Louisiana 1 , = 1 ,529 (1 — e '); and females (excluding Louisiana) 1 ( = 1,067 ( 1 — e -0 291 1 -0 97) where 1 = fork length (mm) and / = years. The mean annual mortality rate determined by six methods of analysis ranged from 0.32 to 0.42. The length-weight relations of king mackerel were for males: W = 0.8064 X lO^L 29928 ; for females: W = 0.8801 X 10" 5 L 29827 . where W = weight in grams and L = fork length in millimeters. King mackerel, Scomberomorus cavalla, is a major recreational and commercial fisheries re- source in the southeastern United States (Ma- nooch 1979). Age, growth, and mortality informa- tion has been based on small specimens collected from a limited geographical area (Beaumariage 1973). A need has existed to reexamine age, growth, and mortality from broader geographi- cally based samples. King mackerel of Brazil have been studied in- tensively, but the great distance separating these Brazilian fish from those in the United States makes application of their results to king mack- erel in United States waters a questionable prac- tice (see Manooch et al. 1978 for annotated bib- liography on this species). A geographically comprehensive sampling of king mackerel in U.S. waters was initiated by us in 1977. Recreational landings were sampled be- cause the sport fishery is less localized than the commercial fishery. We utilized samples from Texas to North Carolina to meet our objectives of determining the age composition, growth rates, length-weight relationships, and mortality rates of king mackerel from U.S. waters. •Contribution No. 82-29-PC, Southeast Fisheries Center Panama City Laboratory, National Marine Fisheries Service, NOAA, Panama City, Fla. 2 Southeast Fisheries Center Panama City Laboratory, Na- tional Marine Fisheries Service, NOAA, 3500 Delwood Beach Road. Panama City, FL 32407-7499. METHODS AND MATERIALS King mackerel (7,723 fish) were collected from Texas, Louisiana, Florida, South Carolina, and North Carolina from June 1977 through August 1979 (Fig. 1). They were caught by recreational hook and line, except for some small individuals, which were caught in shrimp trawls at Cape Canaveral, Fla., in December 1978. The trawl- caught fish were used in determining the rela- tion between otolith radius and fish length. In 1979, 121 fish samples were taken only in north- west Florida and were used to supplement exist- ing samples for the marginal increment analysis. Processing the fish samples involved several steps. The fish were sexed when possible, mea- sured to the nearest millimeter of fork length (FL), and weighed to the nearest gram. Otoliths were removed from the fish, cleaned, and stored either dry or in 100% glycerin. The otoliths were examined under reflected light in a black-bottomed watch glass containing 100% glycerin with a binocular dissecting micro- scope at 28X. The otolith radius (OR) was mea- sured on the posterior surface from the focus to the distal margin along the axis approximating the extension of the sulcus acousticus. All mea- surements were made in ocular micrometer units (1 om/j. = 0.0363 mm). Marks were counted and measured along the radius to their distal edge. The marks were opaque (light) under re- Manuscript accepted June 1982. FISHERY BULLETIN: VOL. 81. NO. 1. 1983. 97 FISHERY BULLETIN: VOL. 81. NO. 1 NORTH CAROLINA SOUTH CAROLINA CAPE CANAVERAL Figure 1. — Location of king mackerel, Scomberomorus cavalla, sampling sites. fleeted light, while the interspaces were hyaline (dark). Otoliths were classified into age groups accord- ing to the number of opaque nonmarginal marks (following the method of Beaumariage 1973). Each otolith was examined by two readers. If the readers did not agree on the age of a fish, data for that fish were not used. We determined the time of mark formation by comparing frequency per month of otoliths with opaque margins. A high percentage of opaque margins indicated recent mark formation. Comparison of age estimations was made, based on surface (whole) and internal (sectional) examination of 133 otoliths. Three to 10 otoliths from each age (0+ through 14+) were used for the comparison. Three to six sections, each 0.15 mm thick, were made through the focus of each otolith, using a Norton 3 diamond blade (SD519- N50m-l/8) rotating at about 285 rpm on an Iso- met low-speed saw. The otolith was mounted in thermoplastic (quartz) cement (No. 70C Lake- side) and cooled with mineral spirits during sec- tioning. Later the cement was dissolved by soak- ing in 50% isopropanol. The free sections were then mounted on glass slides using Piccolyte ce- 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ment and examined with a binocular dissecting microscope. The relationship of the size of the aging struc- ture (OR) to the size of the fish (FL) was deter- mined by using least-square regressions with both linear and power curves. Once the relation- ship was established, fork lengths at earlier ages were back-calculated from surface otolith mea- surements, using methods adopted from Tesch (1971), Ricker (1975), and Everhart et al. (1975). Otolith measurements were analyzed for im- plications of compensatory growth. A frequency distribution of otolith lengths from the focus to the proximal edge of the first opaque mark was developed. Both slow- and fast-growing fish were separated from those that grow at inter- mediate rates, and lengths at earlier ages were back-calculated for both the slow and fast grow- ers. A computer program by Abramson (1971) was used to fit von Bertalanffy theoretical growth curves. Each age was given equal weight, and mean back-calculated lengths were used in the computations. Length-weight equations were developed for the entire king mackerel collection, and for males and females separately, by a computer program following Ricker's (1975) suggestions. Nonlogarithmic length intervals (50 mm) and 98 JOHNSON ET AL.: AGE. GROWTH. AND MORTALITY OF KING MACKEREL weight intervals (computed by the program) were used. A maximum of 20 length-weight val- ues was randomly selected for the analysis with- in each qualifying length and weight interval. If any length or weight interval contained fewer than 20 values, all were utilized. Estimates of annual mortality rate A (after Ricker 1975) were developed by catch-curve analysis of south Florida length-frequency data. These data were used because they best repre- sented the king mackerel in U.S. waters accord- ing to Trent et al. (1981). Since these data were not separated by sex, two age-length keys were developed, one combining males and females as- suming a 1:1 sex ratio and the other assuming a 1 male:2 female ratio (the approximate ratio in our collection). The length-frequency data were converted to age-frequency distributions {N, = number of fish caught in age-class i ) by applying each of the combined age-length keys. Age classes I through X of the resultant catch curves were analyzed by 1. Heincke's (1913) method; 2. Jackson's (1939) method; 3. Rounsefell and Everhart's (1953) method; 4. Beverton and Holt's (1957) method, using the mean of values computed with their equation 13.4 between successive age groups; 5. Robson and Chapman's (1961) method, un- corrected for possible age-length key bias; and 6. finding the slope (m) of a regression line fitted to In (N,) and i and substituting in the equation A — 1 — e m . RESULTS AND DISCUSSION Age The validity of using otoliths for estimating the age and past growth history depends on these structures being directly correlated with the growth of the fish and on otolith mark formation being periodic. We found the otolith radii to be closely correlated to fork lengths, especially when the data were transformed to represent a "power" function. The "power curve" equation, FL = 1.232 OR 1331 with correlation coefficient r = 0.987, had a better fit than the linear equa- tion, FL = 5.559 OR + 84.818 with r = 0.847. This close correlation of OR and FL satisfied the first criterion for validation of otoliths as an age determination structure. The second criterion, mark formation of known periodicity, needed further investigation. Beaumariage (1973) found king mackerel with opaque margins during 8 mo of the year (February-September); the highest percentage of otoliths with opaque margins oc- curred in May. He concluded, "Most otolith mar- gins become opaque (form annuli) during April. May, and June...." Fish in our collections exhib- ited opaque margins in 11 moof the year with the peak during May (54%); however, few fish were collected during the winter months (November- February). No month had a high percentage (over 75%) of fish with opaque margins, and only one month (March) lacked fish whose otoliths had opaque margins (Table 1). In recent years the use of whole otoliths for estimating the age of fish has been questioned. Beamish ( 1979) indicated that a fish's age may be underestimated using surface examination and that otolith sections are more reliable. However, we found 96.5% agreement between king mack- erel age estimates (number of opaque marks) comparing surface and sectional readings. This indicates that our age estimations for whole oto- liths are similar to those of sectioned ones. The agreement betw r een tw r o readers about the number of marks on king mackerel otoliths was 98%. The number of otoliths found to be usable was 1,449. Age and Size Composition Age composition of king mackerel varied greatly among the areas (Table 2). Younger fish were taken in northwest Florida, while older fish were caught off Louisiana, particularly in 1978. Fish of intermediate age were landed primarily in Texas, South Carolina, and North Carolina. The oldest females in our sample were 14+ yr (over 1,400 mm FL) and the oldest males were 9+ yr (970 mm FL;. Much age variation occurred within a single length group in our data (Tables 3, 4) as it did in Beaumariage's (1973) data. For example, we found females 850-899 mm FL were 1-8 yr old (Table 3). Back-Calculated Growth The weighted means of the back-calculated fork lengths for male and female king mackerel from all areas and years sampled in this study are shown in Tables 5 and 6. Differences in mean 99 FISHERY BULLETIN: VOL. 81. NO. 1 Table 1.— Percentages by month, area, and year of king mackerel otoliths having opaque margins. ( ber of fish. : total num- Area Year Jan. Feb. Mar Apr May June July Aug Sept Oct. Nov Dec. Texas 1977 — — — — — 26.7 (15) 286 0.0 (5) — - — — 1978 ~ 0.0 (5) 0.0 (17) 25 (40) — — Louisiana 1977 — — — — — 00 (4) — — 00 (15) 00 (22) 00 (18) — 1978 0.0 167 0.0 40.6 00 154 65 135 00 00 50 143 (7) (6) (43) (32) (2) (26) (62) (37) (5) (51) (20) (7) NW Florida 1977 — — — — — 18.2 (11) 94 (64) 3 1 (65) 0.0 (73) 4.3 (46) — — 1978 " " ~ 0.0 (15) 00 (160) 00 (97) 00 (107) 11 1 (135) — — 1979 " 61.2 (62) 200 (20) 19.2 (27) 00 (12) — — SE Florida 1978 — — 500 (6) 1979 83 3 (6) ~ — — South Carolina 1978 " " 2.9 (104) — — North Carolina 1978 00 (5) 63.6 (22) 38.5 (13) 26.7 (15) 3.8 (53) 8.9 (313) — — Total 38.5 16.7 0.0 40.6 543 23 7 7.2 44 08 7.2 2.6 333 (13) (6) (43) (32) (70) (118) (364) (271) (253) (671) (38) (13) Table 2.— Percentages of king mackerel by area and year within each age group, developed from age-length keys and length- frequency distributions. Area Age in years Year 10 11 12 13 14 No fish Males Texas 1977 1978 Louisiana 1977 1978 NW Florida 1977 1978 South Carolina 1978 North Carolina 1978 Total males Females Texas 1977 1978 Louisiana 1977 1978 NW Florida 1977 1978 South Carolina 1978 North Carolina 1978 Total females 0.8 2.0 0.6 — 69 24.1 24.1 2.6 1 9 13 5 165 26 9 31.3 167 20 5 93.1 2.7 1 2 04 21.1 88 21 8 136 5.2 5.2 183 35 7 488 86 105 126 _ 27 9 48.8 70 4 1 8.5 5.8 37.3 04 08 12.6 289 — 0.4 13 60 39.6 30.4 12.5 10.0 850 5.9 2 5 2 1 17 3 3.6 26.5 21.7 4.5 37 19 7 20 4 37,9 10.9 99 11 1 27 6 3.5 6 9 6.9 _____ 2 9 20.6 32.5 3.6 3.3 5.8 — — — — 533 100 — _______ 10 20.0 24 36.0 8.0 12 — — — — — 25 2 2.6 — — — 498 0.8 - — 1.107 136 19 7 1 4 — — — — — 147 20.0 8.6 3.5 3.5 — — — — 115 7 7 6.5 17 14 14 ___ 2.507 9.3 4.7 2.3 — — — — — 43 23.6 99 10.8 — — — — — — 780 30 1 10.9 6.7 2.9 6.7 — — — — 239 144 24 4 11.9 7.7 77 109 88 44 13 08 479 5.8 6 0.6 0.4 1 — — 1,393 16 8 0.1 — — 1,463 5.6 112 4 4 5 6 2.4 8 0.9 — 249 19 2 16 4 8.5 4 3.2 — 4 — — 402 86 0.3 4 2 2 1.7 1.7 1.1 07 0.2 1 5,216 length occurred from year to year and from area to area. Only data for Louisiana, however, where five or more individuals were used in computing a mean, showed the range of means within an age group to vary more than 100 mm. In 2 yr of sampling in Louisiana, over 300 fe- males were sampled, but too few males were col- lected to back-calculate size at previous ages. Generally, the Louisiana fish were also much larger than those taken elsewhere, and we con- cluded that this must be an anomalous group of fish. We separated Louisiana females from other females for growth computations, except those dealing with compensatory growth. 100 JOHNSON ET AL.: AGE. GROWTH. AND MORTALITY OF KING MACKEREL Table 3.— Length composition (%) of female king mackerel by age group (locations combined). Length group (mm FL) Age in years Total 1 2 3 4 5 6 7 8 9 10 11 12 13 14 fish 350-399 1000 1 400-449 33.3 667 6 450-499 43 5 565 23 500-549 100.0 48 550-599 100 90 600-649 96 4 36 112 650-699 77.5 197 28 71 700-749 25.3 65 1 7.2 1 2 1.2 83 750-799 30 360 430 160 20 100 800-849 24 11 36.2 31 5 134 39 1 6 127 850-899 1 6 08 18.9 33.6 320 98 25 08 122 900-949 1 11.0 22.0 250 28.0 90 4 100 950-999 2 5 23.4 31 2 260 143 1.3 1 3 77 1,000-1.049 167 23.1 34.6 11 5 64 38 2.6 1 3 78 1,050-1.099 4 1 286 265 102 102 163 4 1 49 1.100-1.149 1.9 11.5 40 4 13.5 192 7.7 5.8 52 1.150-1.199 11 9 21.4 33.3 95 95 7 1 48 2.5 42 1.200-1.249 29 15 2 21 2 21.2 9 1 15.2 6.1 9 1 33 1.250-1.299 125 8 3 42 16 7 333 8.3 16.7 24 1.300-1.349 43 4 3 130 8.7 21.7 26.3 130 87 23 1,350-1,399 50 150 30.0 350 5.0 50 50 20 1,400-1,449 26.7 133 333 200 6 7 15 1,450-1,499 143 57.1 143 143 7 1.500-1.549 1,550-1,599 50.0 500 2 Table 4.- —Length composition (%) of male king mackerel by age group (locations combined). Length Age in years Total no. (mm FL) 1 2 3 4 5 6 7 8 9 10 11 12 fish 400-449 100 450-499 15.2 84.8 500-549 100.0 550-599 983 1.7 600-649 93.0 5.3 1 7 650-699 37 5 37 5 146 104 700-749 11.9 35.7 31.0 16.6 2.4 24 750-799 111 27 8 463 130 1 8 800-849 2.0 15.4 346 21 2 192 38 38 850-899 150 50 350 30.0 100 5.0 900-949 14.2 429 429 950-999 25.0 250 250 25.0 1.000-1.049 25.0 75.0 1.050-1,199 1,200-1,249 100.0 4 33 51 60 57 48 42 54 52 20 7 4 4 1 Table 5. — Weighted means of back-calculated fork lengths (mm) for female king mackerel from all areas, 1977-78. South North Age class Texas Lou siana NW Fl 1977 orida 1978 Carolina 1978 Carolina 1977 1978 1977 1978 1978 1 487 457 504 502 463 443 415 393 II 688 673 718 714 670 687 638 627 III 777 748 824 824 755 764 750 738 IV 847 811 906 909 805 838 809 798 V '805 853 970 983 866 895 864 844 VI '849 937 990 1,045 '897 '934 916 891 VII '932 '885 '1,097 1,096 '963 941 939 VIII '1,203 1,148 996 992 IX '1.361 1,202 1.033 '1,000 X 1.252 '1,034 XI 1.311 XII 1.332 XIII '1.350 XIV '1,399 'Lengths based on less than 5 samples. 101 FISHERY BULLETIN: VOL. 81, NO. 1 Table 6.— Weighted means of back-calculated fork lengths (mm) for male king mackerel from all areas, 1977-78. South North Age class Te xas Lo jisiana NW Fl orida Carolina 1978 Carolina 1977 1978 1977 1978 1977 1978 1978 1 414 413 — 473 407 373 385 II 588 574 635 665 607 614 III 659 658 686 '734 715 702 IV 703 720 736 '746 746 747 V 747 790 '798 '769 781 VI 1 754 829 '850 '821 795 VII '803 '896 '810 VIII '789 '951 IX '943 'Lengths based on less than 5 samples Back-calculations for male king mackerel from all areas combined are shown in Table 7. Growth is rapid until the third year of life, after which time the annular growth increment de- creases and stabilizes at an average 42 mm FL. Females from the combined areas (Table 8), excluding Louisiana, also showed rapid growth in the first 3 yr, after which the annual growth increment decreased to an average 40 mm FL. Females were larger than males for all ages. Fish from Louisiana ( all females) exhibited an impressive growth rate (Table 9). They averaged 69 mm longer than other females at age 1 , and by age 10 were 218 mm longer than their counter- parts. The yearly growth increment was over 60 mm to age 6, an increment not maintained by other females, or males, past age 3 in other loca- tions. Our combined back-calculated data were com- pared with those from Beaumariage (1973) (Table 10). His data were converted to fork lengths from standard lengths (SL) using his equation: FL = 1 .096 SL - 17.143. Disregarding Louisiana females, both male and female mean Table 7.— Average back-calculated fork lengths (mm) at age for male king mackerel from all areas, 1977-78. Mean length Age at capture class (mm FL) Ag e in years N 1 2 3 4 5 6 7 8 9 1 570.3 206 425.0 II 708.6 41 4226 667 3 III 7670 41 408 5 6187 737.6 IV 772.5 44 403.3 5946 677.5 747 9 V 820.4 22 375 1 590 5 6690 733.9 796.1 VI 832.6 16 349.2 559.4 641 9 700 1 7558 808.3 VII 852 3 3 3896 579.3 6486 717 7 752.7 802.1 8383 VIII 9200 2 415.2 5788 649 2 7145 773 1 8173 8625 896.0 IX 9700 1 476.6 5603 623.3 754.1 796.2 830.2 864 7 8994 943.4 Weighted mean 376 414.1 6134 6892 734 777 4 809.3 850 8 897 1 943 4 Annual increment 199 3 75.8 448 434 31.9 41.5 46.3 46.3 Table 8.— Average back-calculated fork lengths ( mm) at age for female k ing mackerel from all areas except Louisiana, 1977-78. Age class Mean length at capture (mm FL) N Ag e in years 1 2 3 4 5 6 7 8 9 10 1 6048 315 456.4 II 741.2 112 427 9 693 8 III 8096 105 435.8 6454 7744 IV 858 7 100 426.2 6489 753 5 830.2 V 897.1 79 4993 635 3 7296 8007 8654 VI 933.7 44 405.1 630 3 727.2 791.6 848.2 9083 VII 9602 21 363 613.3 703 3 760.5 827.4 884.5 937.5 VIII 1,028 8 3924 635.0 732.5 7966 852.2 910.0 9553 1,020.9 IX 1,056.0 6 337.4 609 2 732.0 790.9 847.3 893 9 938.1 987.7 1,034.6 X 1.062 1 2 325.5 557.8 683 1 747 4 796 9 833.9 883 6 9347 9784 1,033.6 Weighted mean 792 433.9 652.0 747.1 806.5 853.5 8994 938.5 997 7 1,020.6 1,033.6 Annual increment 218.1 95.1 59.4 470 459 39.1 59.2 22.9 130 102 JOHNSON ET AL.: ACE. OROWTH. AND MORTALITY OF KINO MACKEREL oo t- 05 o S3 O) s3 3 3 bt s o T3 CD 03 -Q CD bt s3 •-_ ai > I ai w j B5 < If* O) en rf co CO o tn cm a) r-- co o r*- co o CO CO CO o r-* o> c\i t- (D O T O O CO CO 0) CO CD CO CO CO CO o 05 r~ o r~ CO CO CO C\J CO CO CO CO en en CO IT) t a> a> en CO i-~ co CM CO CD CO CO CO 0) CO o CO o CO r\j o O) CO in m r-~ o DO (0 CM CO CO r^- ■-cr o CO CO CO CO o o CO CO 0) o o 0) CD O in o CO 'CO o LO un o CO CO o m 0) o nid •- to in co o •** r- t co i c\i 1- c\i ^- co h- r- o ro n O) m ^ co cd tt cd O O G) O O CD OO O c\jc\jcocdcdcococom-c\j co m i-WNOcnmcD'-T-co t^ c\i --OlCDtDLnCMN^TrOCO CO I s - OCDCOCDOOCRCDCDOOCO CO COOltDOCOCOCOtmr-tfi CO t- NcDco^-coifii-ocDn'T oo m OJCNJCNJCMNCOinOCD^r- o CO coo^cr>cr>cocoaoa>cococo oo OTj-^h-oir-cNJcncMoaico COCOCOCOCONCONCONNN CO i- '-ocnocomoioO'-iomo) m co NlfiinCOS'-COCD^lOi-COr- tj- i-^ co^ojcMCNjT-tDcoN'-comrn i- i- r-h-r-.r*-r-r*-cococOr*-cotoco r- eg mcooO)cDNc\iocO'-oro^o co i-oniONdiococjcoiriaios c\j NOCMCO'-CNJO'tCDNOCDCOCO O mininifiifim^^ir^LnTjcoco m '-^IDOCOCO'JmOIDCMt-t-OJCNJ i- i-co^-r-c\jt-*-c\ji-*- co ocococDCMio^-ifiNcomO'-m mmdmcriQO)^'- co ct> oi -^ a> CO'-OinciMnotDNCDCDO'- CDCOCOCDOO'-CVJCMCNJCOCO'CJ-TJ 03 E — — XX - = > ~' ^ X X 5 < Table 10. — Mean back-calculated fork length (mm) at ages, from Beaumariage (1973) and this study. Beaumariage'sdata were transformed from standard length by his formula FL = 1.096 SL- 17.143. Males Females (t Beau- 'xcept La ) Beau- Johnson Johnson Age manage et al. manage et al. 1 457 414 491 434 2 643 613 703 652 3 705 689 793 747 4 752 734 857 807 5 795 777 928 854 6 822 809 986 899 7 839 851 1,033 939 fork lengths at age were smaller in our study than in his in all eases but one (7-yr old males). Several explanations for the differences seem reasonable. First, our back-calculations em- ployed a power curve, whereas his employed a linear equation. Secondly, our fish were sampled from a wide geographical range, which yielded fish with wide variation in age composition, whereas Beaumariage sampled from a more re- stricted area. Lastly, our sampling occurred almost 10 yr after his, and various changes may have occurred in the population owing to exploi- tation or other influences. Compensatory Growth Compensatory growth (Ricker 1975) appeared to occur in both male and female king mackerel. Length-frequency distributions of otolith mea- surements from the focus to the proximal edge of the first opaque mark in both sexes showed a nor- mal distribution of values. After examination of the distributions, we defined slow-growing fish (both sexes) as those with an incremented 50 om^u or less, fast-growing males as those with an in- crement of 81 om/u or more, and fast-growing fe- males as those with an increment of 86 om/i or more. Back-calculated lengths for these fish are shown in Table 11. While fast-growing males grew 525 mm in year 1, they grew only 135 mm in year 2. The slow-growing males grew 303 mm in their first year, but made up some of their size difference by growing 285 mm in their second year. Females showed a similar trend, with fast- growing fish having a first-year increment of 559 mm and a second-year increment of 184 mm. The slow-growing females grew 282 mm in year 1 and 334 mm in year 2. Beyond age 2, yearly growth increments were similar within each sex. Growth compensation in king mackerel is 103 FISHERY BULLETIN: VOL. 81, NO. 1 Table 11.— Annual fork length in- crements (mm) computed from back- calculations on fast- and slow-grow- ing male and female king mackerel (from all areas combined). Males Femai es Age Fast Slow Fast Slow 1 525 303 559 282 2 135 285 184 334 3 87 85 101 99 4 53 72 89 67 5 104 63 75 63 6 49 64 66 7 46 75 8 52 65 9 53 47 10 44 47 11 51 67 12 34 10 13 67 35 14 11 100 probably the result of an extended spawning sea- son. Long spawning seasons and multiple spawns are discussed by Beaumariage (1973) and would result in great size variation in young-of-the- year king mackerel. Some of that size variation would be decreased as the smaller fish continue to grow at a higher rate in their second year than do larger fish in their second year. Although the slow-growing fish make up some difference in size during year 2, they remain smaller than the fast growers throughout their lives. Theoretical Growth The von Bertalanffy theoretical growth param- eters computed from back-calculated fork lengths are shown in Table 12, along with those reported by other authors. The von Bertalanffy (1938, 1957) growth equation is the following: 1* = L„(l -*<< to)\ where 1; = length at age t, Loo = asymptotic length, k = growth coefficient, and Table 12.— von Bertalanffy growth parameters for king mackerel. k t. fo Author value (mm FL) (yr) Males Johnson et al ., all areas 0.28 965 -1.17 Beaumariage (1973) 0.35 903 -2.50 Nomura and Rodrigues (1967) 0.18 1,160 -0.22 Females Johnson et al.. excl La 0.29 1,067 -0.97 Johnson et al.. La. 0.14 1,529 -2.08 Beaumariage (1973) 0.21 1,243 -2.40 Nomura and Rodrigues (1967) 0.15 1,370 -0.13 to = time when length would theoreti- cally be zero. Our theoretical growth parameters are be- tween those calculated by Beaumariage (1973) and Nomura and Rodrigues (1967). Beaumar- iage 's theoretical growth parameters were calcu- lated by employing observed sizes offish at each age, while Nomura and Rodrigues apparently combined both back-calculated lengths and em- pirical lengths in their calculations. We employed mean back-calculated lengths at age in our com- putations, which may account for some of the dif- ferences between our values and those of the other investigators. Length-Weight Relationship The length-weight values for king mackerel computed for the equation W = a L b , where W is weight in grams and L is fork length in milli- meters, are presented in Table 13. Male length- weight values from our study were within the confidence intervals set by Beaumariage (1973), but for both our female and combined sexes, length-weight values were below his lower con- fidence intervals. Mortality Mortality estimates are presented in Table 14. The mean annual mortality rate (A = 0.37) is low- er than Beaumariage's (1973) estimate (A = 0.54). We feel that our results are more concor- dant with generally accepted techniques of catch- curve analysis, in that our catch-curves were de- veloped from age-frequency data, as opposed to the length-frequency catch-curve used by Beau- mariage. We also feel that our results are less in- fluenced by the effects of gear selectivity than Beaumariage's results, since Trent et al. (1981) stated that commercial hook-and-line gear ex- cludes small and large king mackerel to a great- er extent than does recreational hook-and-line gear. Nevertheless, there are many difficulties in using catch-curve analysis in our study. Spe- cific problems are related to the Beverton and Holt ( 1957 ) and Robson and Chapman ( 1961 ) tech- niques. The first technique involves using sev- eral consecutive years of data, which were un- available in our study. With the second technique, we used age-length keys as the basis for our catch-curves but were unable to make correc- tions for the bias when such keys were used ( Rob- 104 JOHNSON ET AL.: AGE, GROWTH. AND MORTALITY OF KING MACKEREL Table 13.— Summary of length-weight relations of U.S. king mackerel. W - weight ir i gram s; L = tc >rk : le n) *th in mill lmeters. No Range (mm FL) w a L" 95% confidence i interval Correlation Sex fish 3 b Lower Upper ir) Male 701 428- 1.355 08064 X 10 -5 2.9928 2.9572 30284 0.9909 Female 2.023 351- 1,554 08801 • 10 29827 29562 30092 0.9910 Sexes combined 2.821 351- 1,554 8464 X 10 29881 3.0153 30153 0.9899 Table 14.— Estimated annual mortality rate (A) by estimation technique, assuming 1:1 and 1:2 male:female ratios. Estimation technique Male: Female ratio Heincke (1913) Jackson (1939) Rounsefell Beverton & Everharl & Holt (1953) (1957) Robson & Chapman (1961) Regression analysis Mean A 1:1 1:2 0.35 034 034 035 42 0.42 042 042 032 033 0.35 036 037 0.37 son and Chapman 1961). This was a result of the age-length keys being developed for a different fish sample than the one being analyzed for mor- tality rates. The difficulties in applying Robson and Chapman's technique resulted in an implica- tion that king mackerel are not fully recruited into the south Florida recreational fishery until age 7, after which the annual mortality rate is 0.53. This mortality estimate is similar to Beau- mariage's (A = 0.54), but the age at recruitment was found by Beaumariage to be 2-3. His esti- mate was based on a smaller age range (0-7) than was ours. This difference probably influenced the resulting mortality estimates. Many difficulties are also involved in the basic concept of using catch-curve analysis to estimate mortality in king mackerel. Rounsefell and Ever- hart (1953) emphasized that catch-curve analy- sis is based on false assumptions when applied to most pelagic species, including mackerel. Rob- son and Chapman (1961) reiterated this warning, stating, "if year classes.. .vary in strength and survival rates vary from year class to year class and age to age, then the age-frequency distribu- tion in the catch of a single season provides no identifiable information whatsoever regarding [mortality rates]...." These comments force us to state our mortality findings with some wariness. ACKNOWLEDGMENTS We thank Michael Crow, Mark Farber, and Dennis Lee of the National Marine Fisheries Service, Miami Laboratory, Miami, Fla., for their constructive reviews of this manuscript. ADDENDUM Fischer (1980) reported on the length-weight relationship of king mackerel off Louisiana. His length-weight values are similar to ours. LITERATURE CITED Abramson, N. J. 1971. Computer programs for fish stock assessment. FAO Fish. Tech. Pap. 101, 149 p. Beamish. R. J. 1979. Differences in the age of Pacific hake (Merluccius productus) using whole otoliths and sections of otoliths. J. Fish. Res. Board Can. 36:141-151. Beaumariage, D. S. 1973. Age, growth, and reproduction of king mackerel, Scomberomorus cavalla, in Florida. Fla. Mar. Res. Publ. 1. 45 p. Beverton, F. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish populations. Fish. Invest. Minist. Agric. Fish. Food (G.B.), Ser. II. 19, 533 p. Everhart, W. H., A. W. Eipper, and W. D. Youngs. 1975. Principlesof fishery science. Cornell Univ. Press, Ithaca. N.Y., 288 p. Fischer, M. 1980. Size distribution, length-weight relationships, sex ratios, and seasonal occurrence of king mackerel (Scom- beromorus cavcdla) off the southeast Louisiana coast. La. Dep. Wildl. Fish. Tech. Bull. 31:1-21. Heincke, F. 1913. Investigations on the plaice. General report. I. Plaice fishery and protective measures. Preliminary brief summary of the most important points of the re- port. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 16, 67 p. Jackson, C. H. N. 1939. The analysis of an animal population. J. Anim. Ecol. 8:238-246. 105 FISHERY BULLETIN: VOL. 81. NO. 1 MANOOCH. C. S., III. 1979. Recreational and commercial fisheries for king- mackerel, Scomberomorus cavalla, in the South Atlantic Bight and Gulf of Mexico, U.S.A. In E. L. Nakamura and H. R. Bullis, Jr. (editors), Proceedings of the mack- erel colloquium, p. 33-41. Gulf States Mar. Fish. Comm., Brownsville, Tex. Manooch, C. S., Ill, E. L. Nakamura, and A. B. Hall. 1978. Annotated bibliography of four Atlantic scom- brids: Scomberomorus brasiliensis, S. cavalla, S. macu- latus, and S. regalis. U.S. Dep. Commer., NOAA Tech. Rep. NMFSCirc. 418, 166 p. Nomura, H.. and M. S. S. Rodrigues. 1967. Biological notes on king mackerel, Scomberomorus cavalla (Cuvier), from northeastern Brazil. Arquivos do Estocao de Biologia marinha do Universidade Fed- eral do Ceara. 7(l):79-85. RlCKER. W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Fish. Res. Board Can., Bull. 191, 382 p. Robson, D. S., and D. G. Chapman. 1961. Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. ROUNSEFELL, G. A., AND W. H. EVERHART. 1953. Fishery science: its methods and applications. John Wiley and Sons, Inc., N.Y., 444 p. Tesch, F. W. 1971. Age and growth. In W. E. Ricker (editor). Meth- ods for assessment of fish production in fresh waters. 2d ed., p. 98-130. Blackwell Sci. Publ., Oxford. Trent, L., R. 0. Williams, R. G. Taylor, C. H. Saloman, and C. S. Manooch III. 1981. Size and sex ratio of king mackerel. Scomberomor- us cavalla, in the southeastern United States. U.S. Dep. Commer.. NOAA Tech. Memo. NMFS-SEFC-62, 59 p. National Marine Fisheries Service, Panama City, Fla. von Bertalanffy, L. 1938. A quantitative theory of organic growth (inquiries on growth laws. I). Hum. Biol. 10:181-213. 1957. Quantitative laws in metabolism and growth. Q. Rev. Biol. 32:217-231. 106 REVIEW AND ANALYSIS OF THE BLUEFIN TUNA, THUNNUS THYNNUS, FISHERY IN THE EASTERN NORTH PACIFIC OCEAN Doyle A. Hanan 1 ABSTRACT Northern bluefin tuna migrate from waters near Japan to the eastern North Pacific where they are fished primarily by purse seine. While annual catches fluctuate greatly, two major periods are identified. The average annual catch in the second period ( 1950-present) is nearly double that for the first period (1921-50) and is attributed to increased fishing effort by the "high-seas" tuna fleet oper- ating off Baja California. The declining catch per unit effort in the second period and declining catches after 1963 are assumed to indicate declining abundance of bluefin tuna in the eastern North Pacific. Length-frequency analysis reveals 1) significantly smaller bluefin tuna in U.S. waters than in waters off Baja California and 2) significant variation in mean lengths among years. Analysis of tag-recapture data confirms seasonal northward migration and vulnerability to the fishery for as many as three fishing seasons. A catchability coefficient of 1.66 X l(T 4 /boat-day and an annual instantaneous total mortality rate of 2.07, both estimated from the tag-recapture data, are used with summaries of fishing effort to calculate an average annual exploitation rate of 30% for bluefin tuna in the eastern North Pacific. Purse seining for northern bluefin tuna, Thun- nus thynnus Linnaeus, in the eastern North Pa- cific Ocean began about 1914, with the first large commercial landings in 1918 (Whitehead 1931). Prior to the development of this purse seine fish- ery, a sport fishery existed off southern Califor- nia at Santa Catalina Island; and since bluefin tuna are difficult to catch by hook and line, elab- orate fishing methods evolved such as using a kite to make the bait (flying fish) skip across the water (Clemens and Craig 1965). The Tuna Club of Avalon at Santa Catalina Island even awarded "blue buttons" to its members for catching the large and wary prize. Because of this difficulty in hooking bluefin, the commercial "high-seas" fleet did not fish for bluefin until the late 1950's, when most of the fleet had converted from pole- and-line gear to purse seines (Bell 2 ). Currently the bluefin fishery consists of a "wet- fish" fleet, principally out of San Pedro, Calif.; a high-seas fleet mostly out of San Diego, Calif.; and since 1975, an expanding Mexican fleet most- ly out of Ensenada, Baja California. The bluefin 'California Department of Fish and Game, c/o Southwest Fisheries Center, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. 2 Bell, Robert R. 1970. Bluefin tuna Thunnus thynnus ori- entalis in the northeastern Pacific Ocean. Unpubl. manuscr. Calif. Dep. Fish Game, 350 Golden Shore, Long Beach, CA 90802. Manuscript accepted Julv 1982. FISHERY BULLETIN: VOL. 81, NO. 1. 1983. fishery extends along the coast of North America from Cabo San Lucas, Baja California, to Point Conception, Calif., and occasionally farther north (Table 1). The bluefin catch is composed mainly of 1-, 2-, and 3-yr-old fish, which appear to mi- grate to the eastern North Pacific from the west- ern Pacific near Japan (Schultze and Collins 1977); however, older and much larger bluefin are reported and occasionally caught in the east- ern North Pacific. This paper reviews and analyzes the bluefin tuna fishery in the eastern North Pacific, using data collected by the California Department of Fish and Game (CFG) in cooperation with the Inter-American Tropical Tuna Commission (IATTC), and the National Marine Fisheries Service (NMFS) of the U.S. Department of Com- merce. CATCH AND EFFORT ANALYSIS Although annual bluefin catches have fluctu- ated considerably in the eastern North Pacific (Table 2), two major periods are identified in the catch by a plot of a 10-yr running average (Fig. 1). During the first period, about 1921-50, total landings averaged 5,066 t (metric tons)/yr and were declining toward the end of the period. During this time, bluefin were landed almost ex- 107 FISHERY BULLETIN: VOL. 81, NO. 1 Table 1.— Total number of months in which bluefin tuna catch exceeded 50 t within a 1° area of latitude and longitude for the years 1957-69 and 1974. Each latitude and longitude indicates the southeast corner of the 1° area of considera- tion. Asterisks indicate the coastline. Lati- Long tude tude 125 120 119 118 117 116 115 114 113 112 111 110 40 1* 34 1 * 33 1 3 7 2* 32 7 12 14* 31 3 11 5* 30 8 11 • 29 3 6 2 • 28 7 3 7 * 27 1 8 4- 26 2 10 7 • 25 1 12 6* 24 6 8 1* 23 4 3 1* Table 2.— Total landings of bluefin tuna by commercial (in metric tons (t)) and sport fisheries (no. fish) in the eastern North Pacific Ocean, 1918-81. Asterisks indicate no data available. Land ngs Landings Commer- Sport Commer- Sport Year cial (t) (no. fish) Year cial (t) (no. fish) 1918 2,722 1950 1.242 27 1919 6.800 1951 1,752 7,142 1920 4,776 1952 2,076 145 1921 894 1953 4.433 4.276 1922 1.275 1954 9,537 966 1923 1.460 1955 6,173 8.179 1924 1.470 1956 5,727 34.187 1925 1,725 1957 9,215 6.428 1926 2.960 1958 13,934 884 1927 2.222 1959 6,914 1,330 1928 6.215 1960 5.422 97 1929 3.414 1961 9,603 2,268 1930 9.943 1962 14,651 2,453 1931 1.603 1963 14,189 737 1932 486 1964 10.642 693 1933 254 1965 7,556 92 1934 8,327 1966 16,846 1.998 1935 11.418 1967 6,601 3.166 1936 8.584 2,920 1968 6,063 1,231 1937 5.758 4,020 1969 7,172 1.470 1938 8.041 11,927 1970 4,024 1,833 1939 5,369 9,909 1971 8,415 749 1940 9,058 6,878 1972 13.390 1,470 1941 4,318 1973 10,576 5,347 1942 5,826 1974 5,748 5.765 1943 4,617 1975 9,578 3,348 1944 9,228 1976 10,561 2,040 1945 9.341 1977 5.151 1,838 1946 9,993 528 1978 5.903 479 1947 9,452 2,194 1979 6,743 1.087 1948 2,961 104 1980 3,128 729 1949 1,991 1,841 1981 1,016 clusively by the San Pedro wetfish fleet, which seasonally targets fishing effort on sardines, an- chovies, mackerel, bonito, bluefin tuna, and other fishes, depending on fish availability, mar- ket price, and market demand (cannery orders). During the second period, about 1950-present, annual landings increased to 16,846 t in 1966, then declined to 1,016 1 in 1981, averaging 9,076 1 for the period. At the beginning of this period, many of the high-seas boats that had converted to purse seining began catching large numbers of bluefin off Baja California, although they target- ed their fishing on yellowfin tuna, Thunnus alba- cares, and skipjack tuna, Katsuwonus pela- m is. Two sources of data were used in summarizing total catch by area. The first, landings reported by CFG, separates pounds landed in California for 1918-79 into those caught in California wa- ters and those caught south of California waters. These data reveal an overall decreasing trend in bluefin catch north of the international border; and, until about 1963, there was an overall in- creasing trend in total catches south of the inter- national border (Fig. 2). The second source of catch data also includes effort information and was compiled into a data base for summary and analysis. These data, rep- resenting about 87% of the catch during the peri- ods 1954-69 and 1971-74, came from summaries of skippers' logs, from interviews with skippers and engineers, from CFG landing receipts, and from IATTC summaries of the high-seas fleet. Catch and effort in the data base are recorded by 1° areas of latitude and longitude. For this study, one boat-day or part of a boat-day of effort is as- signed to a seiner for each day or partial day of purse seining or searching for tuna in the bluefin fishing range (north of lat. 22°N) during months in which bluefin were caught. Catch data, sum- marized by areas north and south of lat. 32°N (the parallel nearest the international border), show trends similar to the reported California landings for the same years (Fig. 3). For comparison with the CFG reported Cali- fornia landings and for future consideration of the effects of Mexican regulations concerning 108 HANAN: BLUEFIN TUNA FISHERY 18,000 15,000 BLUEFIN TUNA • • ANNUAL CATCH o o 10-YEAR RUNNING AVERAGE 1925 1945 1950 YEAR 1955 1970 FIGURE 1.— Annual catches of northern bluefin tuna in the eastern North Pacific Ocean for the years 1918-81. I8,000r 1920 1930 1940 1950 1960 1970 1980 YEAR Figure 2.— Annual California landings (metric tons) of north- ern bluefin tuna caught north and south of the United States- Mexico international border, 1918-79. South is represented by solid circles and lines, north by open circles and broken lines. the 200-mi exclusive economic zone, much of the data in this paper are separated into areas north and south of lat. 32°N. A better division from a biological standpoint would be north and south of lat. 29°N (Fig. 4). The increase in California landings of bluefin caught south of the border during the 1957-66 period can be attributed to increased fishing effort, but the decline in catch north of the border cannot be explained by declining effort, since ef- fort remained comparatively level throughout the period (Fig. 5). Because bluefin are valuable ($l,180/short ton in 1981) and because fishing ef- fort north of the 32d parallel remained fairly con- stant, the decline in northern catches is attrib- uted to a decrease in abundance in that area. During this period, increased catches south of the border appear to have offset the decline in catches to the north and to indicate the fish were intercepted before migrating northward. If this is true, the recent catch decline south of the bor- der indicates declining bluefin abundance in the eastern North Pacific. Catch and effort data summarized by latitude show a bimodal distribution centering just north of the 25th and 32d parallels (Table 3). The catch- es are concentrated in the period June-Septem- ber, with the largest catches shifting northward during the fishing season (Fig. 4). Early in the 15,000 12,500 - p 10,000h I 7,500 - o < <-> 5,000 2,500 \ ,\ / \ NORTH v v V __----" _L J_ 1957 1960 1963 1966 1969 1972 1975 YEAR Figure 3.— Logged annual catches (metric tons) of bluefin tuna north and south of lat. 32°N for 1957-69 and 1974. 109 FISHERY BULLETIN: VOL. 81. NO. 1 MONTH JFMAMJJASOND 6,000 <25 25-300 >300 Figure 4. — Mean monthly catches of bluefin tuna (^_) for the 13 period 1957-69 in metric tons per latitude. Totals are for areas between a given parallel and the next higher parallel. Table 3.— Mean catch of bluefin tuna (metric tons (t)) and mean effort (boat-days) per latitude for the years 1957-69 and 1974. Mean catch Mean effort Latitude (t) (boat-days) 36 3 29 1.16 35 000 0.29 34 2306 6.67 33 452.60 75.27 32 2,160.34 460.39 31 1.048.67 214.39 30 62263 155.11 29 544.82 13596 28 561.82 170 37 27 61546 253.89 26 734.33 359.86 25 981.61 62956 24 543.07 348.16 23 234.57 353 10 22 0.44 122.02 season there are catches both in northern and southern parts of the bluefin range, whereas there are relatively few catches late in the season in the southern part of the range, thus indicating northward movement. This shift is also apparent in the number of occurrences of recorded bluefin 1957 1960 1963 1966 1969 1972 1975 YEAR Figure 5. — Logged annual effort (boat-days/year) for bluefin tuna north and south of lat. 32°N for 1957-69 and 1974. catch per month and latitude during the 1957-69 period (Table 4). The northward shift in location of the largest catches does not reflect a shift in fishing effort, since effort remains high in the south throughout the season (Fig. 6). Apparently, bluefin move northward or there is a shift in bluefin vulnerability towards the north during the fishing season. Catch and effort data for 1957-69 summarized by vessel size indicate that seiners of 101-300 ton capacities accounted for more than 70% of blue- fin landings and that smaller vessels tended to be phased out of the fishery and replaced by larger ones (Table 5). CATCH-PER-UNIT-EFFORT ANALYSIS Catch per unit effort (CPUE) is calculated for each year as total catch divided by total effort (Table 6). The relationship between the CFG data and the IATTC data (Bayliff and Calkins 1979) is expressed as a ratio which includes the origin. The ratio estimator 2,y/%x, obtained from years for which both CFG and IATTC measures of CPUE are available (1966-74), yielded a value of 1.01, by which the CFG values (1954-65) were multiplied to obtain IATTC equivalents. These equivalent CPUE values were then plotted, and a regression line fit to them reveals a decline in CPUE with time (Fig. 7). This observed decline is probably conservative because fishing effort, which was not standardized, has most likely be- come more effective with time (Pella and Psaro- pulos 1975). CPUE values were highest in the northern 110 HANAN: BLUEFIN TUNA FISHERY Table 4. — Total occurrences of recorded bluefin tuna catch per latitude and month during 1957-69 and 1974. Totals are for the areas between a given parallel and the next higher par- allel. Lati- M Dnth tude J F M A M J J A S O N D Total 36 1 1 2 35 34 2 2 1 5 33 2 1 5 9 9 4 2 32 32 1 3 2 6 11 13 12 4 2 1 55 31 2 9 11 8 1 31 30 8 10 6 2 26 29 3 1 1 3 9 6 1 2 26 28 7 5 4 5 7 9 11 7 4 1 3 2 65 27 3 9 7 2 1 1 23 26 9 10 1 1 21 25 1 3 11 11 3 1 30 24 3 10 7 20 23 1 3 6 4 14 22 1 1 2 Total 8 8 5 13 19 65 101 70 41 13 6 3 MONTH FMAMJ JASOND LU Q <1 1-3 >3 Figure 6.— Mean monthly effort for bluefin tuna (Jil.) for the 13 period 1957-69 in boat-days per latitude. Totals are for areas between a given parallel and the next higher parallel. part of the bluefin range (Figs. 8-10), and is at- tributed to differences in searching and fishing methods between the wetfish and the high-seas fleets. Because effort data are not available for the Y = 10.5-0.1 14X 1955 1960 1965 1970 1975 1980 YEAR Figure 7.— Annual catch per unit effort (metric tons/boat- day) of bluefin tuna plotted by year ( 1954-78) with a regression line fitted to the curve. MONTH J FMAMJ JASOND 0*5 6-75 >75 Figure 8.— Mean monthly catch per unit effort of bluefin tuna m < i 'per latitude (metric tons/boat-day) for the period 1957-69. 13 Totals are for areas between a given parallel and the next high- er parallel. 1979-81 period, the rapid decline in total catches during those years cannot be explained by direct 111 FISHERY BULLETIN: VOL. 81, NO. 1 Table 5. — Yearly catch (metric tons) of bluefin tuna and effort (in parentheses) by vessel size class for 1957-69 and 1974, from logbook data. Hold capacity in short tons: 0-50 = Class 1; 51-100 = Class 2; 101-200 = Class 3; 201-300 = Class 4; 301-400 = Class 5; over 400 = Class 6. Vessel class Year 1 2 3 4 5 6 Total 1957 205 2,537 4,614 35 — — 7,391 (20) (328) (679) (24) (-) (-) (1,051) 1958 276 1,840 6,767 646 — — 9.529 (37) (433) (1.266) (42) (-) (-) (1.778) 1959 164 1.912 2,468 522 330 — 5,396 (29) (408) (1,352) (267) (117) (-) (2,173) 1960 5 287 2,318 1,067 1.081 69 4,827 (33) (194) (1.495) (730) (341) (39) (2,832) 1961 21 526 5.325 2,331 1.015 4 9,222 (4) (171) (2.222) (925) (352) (12) (3,686) 1962 14 959 7,061 2,840 1.498 — 12,372 (8) (185) (2,447) (1,515) (603) (32) (4,790) 1963 — 544 5,483 4,228 3,055 87 13,397 (-) (85) (1,667) (1.729) (1,051) (56) (4,588) 1964 18 523 3,641 2,937 1.565 — 8,684 (4) (77) (1.367) (1,577) (749) (19) (3,793) 1965 60 294 2,538 2,242 1.312 36 6,482 (12) (51) (1.641) (1.338) (753) (13) (3.808) 1966 21 429 5,576 5,400 3,107 561 15,094 (16) (112) (1,479) (1,299) (580) (38) (3,524) 1967 60 289 1.318 2,530 1,804 51 6.052 (10) (33) (1.103) (1.936) (1.435) (270) (4.787) 1968 — 399 2,038 1,481 1,300 293 5.511 (-) (69) (1.162) (895) (493) (71) (2,690) 1969 32 175 3,370 2,338 605 448 6,968 (7) (40) (1.200) (1,280) (479) (232) (3,238) 1974 60 257 1,712 905 719 677 4,330 (3) (81) (672) (450) (500) (251) (1.957) Total 936 10,971 54,229 29,502 17,391 2,226 115,255 (183) (2,267) (19,752) (14,007) (7,453) (1.033) (44,695) Total 1% 10% 47% 26% 15% 2% 100% (0.4%) (5%) (44%) (31%) (17%) (2%) ( 1 00%) 14 12 10 3 a. o ' NORTH 1957 1960 1963 1966 1969 1972 1975 YEAR Figure 9. —Annual catch per unit effort of bluefin tuna (metric tons/ boat-day) for 1957-69 and 1974 north and south of lat. 32 °N. CPUE evidence. However, if it is assumed that effort remained at about the same levels, CPUE 6.0 4.5 2 3.0 - 1.5 - 0.0 L z> a. O 20° _L 24° 28° 32° LATITUDE 36° 40° Figure 10.— Bluefin tuna catch per unit effort (metric tons/ boat-day) by latitude for 1957-69 and 1974. Latitude area is that lying between a given latitude and the next higher lati- tude. would have declined by an even greater rate than that predicted by the trend in Figure 7. This in- dicates that bluefin abundance in the eastern North Pacific has declined severely. 112 HANAN: BLUEFIN TUNA FISHERY Table 6.— Bluefin tuna CPUE values from this study (CFG) for the years 1954-74 and from IATTC (Bayliff and Calkins 1979) for the years 1954-78. CFG values converted to IATTC equivalent values are in parentheses. CPUE values CPUE values Year CFG IATTC Year CFG IATTC 1954 4 49 (4.55) 1967 1.26 1 63 1955 5.44 (5.52) 1968 2.05 2.35 1956 3.59 (364) 1969 2 15 1.96 1957 7 03 (7 13) 1970 — 1 71 1958 5.36 (5 44) 1971 2 31 2.11 1959 2.48 (2.52) 1972 3.61 3.23 1960 1 70 (1 72) 1973 3.15 289 1961 250 (2.54) 1974 2 21 1 75 1962 258 (2.62) 1975 2.73 1963 292 (2.96) 1976 298 1964 2.29 (2.32) 1977 1 86 1965 1 71 (1.73) 1978 1.62 1966 428 5.40 LENGTH-FREQUENCY ANALYSIS Length-frequency data summaries (Figs. 11- 14) were obtained from two CFG data sets of fork-length samples taken as frozen bluefin were unloaded at Terminal Island, Calif., canneries. Set 1 (1952-65) represents random samples of 50 fish/seiner; set 2 (1963-71 and 1974) represents random samples of 20 fish for every 200 short tons landed from each 1° area of latitude and lon- gitude. Set 2 samples were taken for an age de- termination study. Although a smaller number of bluefin were sampled, they appear to repre- sent the same population as the first data set, z O 4 a LU a 2 1952 N=1,140 60 80 100 120 140 160 180 120 140 160 180 60 80 100 120 140 160 1 180 O 4 1954 N = 3,680 ML T — i 1 60 80 100 120 140 160 180 60 80 100 120 140 160 1! FORK LENGTH (cm) FORK LENGTH (cm) Figure 11.— Bluefin tuna percent length frequencies, 1952-57. 113 FISHERY BULLETIN: VOL. 81. NO. 1 60 80 100 120 140 160 180 60 80 100 120 140 160 180 10 8 - Z 6 HI o a LU 4 1960 N = 3.498 80 100 120 140 FORK LENGTH (cm) FORK LENGTH (cm) FIGURE 12.— Bluefin tuna percent length frequencies. 1958-62. 1 r 160 180 Table 7. — Mean length frequencies of bluefin tuna, north and south of lat. 32°N, 1952-65. Mean length Year South North Combined 1952 737 70.2 732 1953 67.1 63.8 68.1 1954 79.7 66.3 76.3 1955 83.1 72.3 78.8 1956 90.4 65.8 83.1 1957 83.7 71.5 73.0 1958 81.3 77.7 78.6 1959 85.8 90.6 90.3 1960 112.1 968 105.6 1961 72.3 71 2 71 7 1962 73.5 64.0 68.8 1963 80.3 68.9 76.4 1964 70.4 62.8 679 1965 79.8 658 76.0 when overlapping years (1963-65) and composite samples for both data sets are compared (Fig. 15). Analysis of fish lengths from the first data set shows a decrease in mean length with increasing latitude. These data (1952-65) were also sum- marized by year for areas north and south of the 32d parallel (Table 7) for a two-way analysis of variance. The analysis shows significant differ- ences (P<0.01) among years and between areas. These results show that bluefin caught in the north are smaller than those to the south (Fig. 16) and that mean lengths vary considerably, as much as 39.8 cm/yr. 114 HANAN: BLUEFIN TUNA FISHERY 1963 LWA N = 1,533 Ih-Mj ' I 1— 120 140 160 180 14 - 1964 N = 4,812 12 J 10 PERCENT O) 00 4 1 2 A %■■ i i T " T 1 f 60 14 I 1964 LWA N- 1,446 12 10 8 1 6 1 4 1 2 L l if " I i i 80 100 120 140 160 180 60 80 100 120 140 160 180 O 4 100 120 140 160 180 FORK LENGTH (cm) 4 - 1965 LWA N = 1,380 120 140 160 180 FORK LENGTH (cm) FIGURE 13.— Bluefin tuna percent length frequencies, 1963-65. Graphs to the left are based on length-frequency samples only, whereas those to the right are based on length-weight-age frequency samples. TAGGING DATA ANALYSIS From 1953 to 1958, 186 bluefin were tagged and released by CFG and IATTC in the eastern North Pacific incidental to tagging other spe- cies. From 1962 to 1968 a tagging cooperative of CFG, U.S. Bureau of Commercial Fisheries (NMFS), and the Mission Bay Research Founda- tion of San Diego tagged and released 2,836 blue- fin. Of these, 565 (20%) were recaptured in the eastern North Pacific, including 7 by sport fish- ing and 9 in the western Pacific (Clemens and Flittner 1969). Bluefin for tagging were caught by purse seine and tagged with spaghetti-loop tags prior to 1960 and with spaghetti-dart tags since then. Bluefin are caught within about 200 mi of the coast, thus spatial analysis of tag returns is ex- pressed only by latitude. Of the 565 tagged blue- fin caught in the eastern North Pacific, recovery latitude information is available for 540 returns. Data from tagged fish recovered during the sea- son in which they were released (62%) show a general movement northward (Table 8); how- ever, many were caught near the release point and to the south (Table 9). Tagged fish recaptured during the second and third fishing seasons after tagging were well dispersed throughout the fish- 115 FISHERY BULLETIN: VOL. 81. NO. 1 O 4 0. 2 60 80 100 120 140 160 180 60 80 100 120 140 FORK LENGTH (cm) 160 60 80 100 120 140 160 60 80 100 120 140 160 180 60 80 100 120 140 160 FORK LENGTH (cr Figure 14.— Bluefin tuna percent length frequencies for 1966-71 and 1974 taken from length-weight-age samples. 180 Table 8. — Bluefin tuna tags returned during tagging season (1958-68) summarized by latitude of release and of return. Totals are for areas be- tween a given parallel and the next higher parallel. Return latitude latitude 33 32 31 30 29 28 27 26 25 24 Total 33 13 28 41 32 85 28 1 1 115 31 20 16 2 2 40 30 16 19 7 4 2 48 29 1 2 1 3 7 28 2 1 1 5 1 10 27 2 8 5 1 1 17 26 25 24 1 1 2 7 23 2 22_ 58 Total 13 154 72 18 7 16 9 23 2 22 336 116 HANAN: BLUEFIN TUNA FISHERY 4 - 10 60 80 100 120 140 160 180 H 3 Z UJ o DC UJ 0. 2 1963-1971 and 1974 N = 9,980 ,1,LJJ " 1 60 80 100 120 140 160 180 FORK LENGTH (cm) Figure 15.— Composite bluefin tuna percent length frequen- cies. Upper graph summarizes length-frequency samples for 1952-65. and lower graph summarizes length-weight-age sam- ples for 1963-71 and 1974. ing grounds and fishing season, indicating good mixing with the untagged population. Gulland (1963) described a method of estimat- ing fishing mortality from tagging experiments; this method was modified and applied to the z UJ o °- 4 NORTH OF 32° N. N = 30.444 60 80 100 120 140 160 180 10 8 - z UJ o oc UJ 0. 4 SOUTH OF 32° N. N=31,773 60 80 100 120 140 160 180 FORK LENGTH (cm) Figure 16.— Bluefin tuna percent length-frequency compos- ites for 1952-65, north (top) and south (bottom) of lat. 32°N. bluefin data. It was assumed that the number of tags returned per unit of effort is proportional to the CPUE, and no provision was made for immi- gration or emigration. For any period following tagging, an estimate of catchability (q) would be the number of tags returned per unit of effort di- vided by the initial number released. When these Table 9. — Total number of returned bluefin tags summarized by latitude of release and of return. Totals are for areas between a given parallel and the next higher parallel. Release Retu rn latitude latitude 33 32 31 30 29 28 27 26 25 24 23 Total 33 13 28 1 4 3 4 3 9 3 68 32 87 30 5 9 5 7 6 20 7 1 177 31 21 17 3 2 4 1 2 8 1 1 60 30 17 27 14 10 3 2 6 8 3 90 29 3 2 9 2 8 4 28 28 5 2 7 7 8 5 2 36 27 2 8 5 4 1 1 21 26 25 24 1 1 2 7 25 2 22. 60 Total 13 163 87 41 34 37 29 39 57 38 2 540 117 FISHERY BULLETIN: VOL. 81. NO. 1 estimates are plotted against time, the intercept at time zero is an estimate of (/ for bluefin in the eastern North Pacific. As tagged bluefin were not fully dispersed dur- ing the season of tagging, monthly estimates of q were calculated as the monthly mean, per 1° area of latitude and longitude, for 1° areas from which tagged fish were caught. For the second and third seasons, when tagged fish appeared to be fully dispersed, monthly estimates were calcu- lated for the entire bluefin range; then, the nat- ural logarithms of these values and those for the first season were plotted (Fig. 17). Effort and therefore q are expressed in boat-days. The re- gression line fitting these points (Y= —8.7363 — 0. 1725 X, R 2 « 68%) was weighted by the number of tagged fish released each year, since the num- ber of tagged fish varied between 35 and 960/yr. The best estimate of q from the tag-recapture data is the antilogarithm of the regression line intercept, 1.66 X 10~ 4 /boat-day with a 95% con- fidence interval of 0.99 X 1(T 4 to 2.63 X KT 4 / boat-day corrected for geometric mean bias (Beauchamp and Olson 1973). The slope of the re- gression (—0.17, S 2 = 0.02) is an estimate of the monthly instantaneous mortality coefficient (Z), and was expanded to estimate the yearly instan- taneous mortality (Z = 2.07, S 2 = 0.24) including immigration and emigration. This estimate com- pares favorably with Bayliff and Calkins' (1979) and Bayliff s (1980) estimates (Z = 2.08, S 2 = 0.8) for 1962-66. They call these estimates "rates 1 2 8 9 10 11 12 13 RETURN MONTHS 222324 Figure 17.— Natural logarithms of adjusted return rates for tagged bluefin tuna plotted against number of months between tagging and recapture, for the years 1962-64, 1966, and 1968. The predicted catc liability coefficient (q) from straight-line re- gression and the 95% confidence interval around (§) are shown at the zero-month intercept. of attrition," since immigration and emigration are included. The ratio of fishing mortality to instantaneous total mortality is an estimate of the exploitation ratio (Ricker 1975) and was calculated as a mean for the period 1962-70 because q was also calcu- lated for that period. The mean annual fishing effort in that period was 4,215 boat days which, multiplied by (/, estimates a fishing mortality of 0.7/yr. Dividing this value by estimated Z(2.07/ yr) yields an exploitation ratio of 0.34, and then multiplying by the annual mortality or "attri- tion" (0.87) yields a 30% exploitation rate. DISCUSSION The review and analysis of data concerning the bluefin tuna fishery in the eastern North Pacific show large fluctuations in the catch to be a major part of two important phases. The decline in catch near the end of the first phase (1921-50) is offset by the development of a "high seas" purse seine fleet and the resultant increased catch of bluefin off Baja California. The current decline (1963-present) is probably due to a decline in the abundance of bluefin as indicated by CPUE evi- dence. The effect on the resource of Mexico's 200- mi regulations was not assessed at this time; how- ever, the apparent decline in catch and CPUE cannot be attributed to such regulation since it has been enforced only recently. The declines in catch and CPUE in the eastern North Pacific are significant and are reflected by an even greater decline in catch and nomi- nal CPUE in the western Pacific (Figs. 18, 19). 50 O o o I o 40 30 20 o 10 _L 1948 1956 1964 1972 1980 YEAR Figure 18.— Annual Japanese landings of northern Pacific bluefin tuna for the years 1951-59 (metric tons X 1.000) and 1962-79 (thousands of fish). 118 HANAN: BLUEFIN TUNA FISHERY 2.80 2.10 UJ Q. O 1.40 0.70 1960 1964 1968 1972 YEAR 1976 1980 FIGURE 19.— Annual Japanese catch per unit effort (metric tons/boat-day) from long-line catches of northern bluefin tuna for the years 1962-79. Although those data (Anonymous 1981; Yama- naka and Staff 1963) represent only a portion oi the fishing effort in the western Pacific, they in- dicate a need for more extensive and explicit data from that area. With improved data, mathe- matical models for estimating sustainable yields can be used to describe the status of the bluefin resource throughout the North Pacific Ocean. Based on strong evidence of declining stock abundance, the bluefin tuna fisheries in the Pa- cific Ocean should receive an extensive analyti- cal review, and nations fishing bluefin, especially Japan, Mexico, and the United States, should consider needed actions. If management to con- serve this valuable resource is to be taken, it should be soon, so that the resource can return to an optimal level of abundance. ACKNOWLEDGMENTS I sincerely thank Alec MacCall (CFG) for his help and recommendations during this study. Norman Bartoo (NMFS) provided valuable ad- vice in the preparation of the manuscript. The editorial reviewers, especially Harold Clemens (CFG) and William Bayliff (IATTC), are thanked for their comments and suggestions. Funds for the study were provided by contract to the South- west Fisheries Center, NMFS. LITERATURE CITED Anonymous. 1981. Annual report of catch and effort statistics by area on Japanese longline fishery 1979. Fish. Agency Jpn., Res. Dev. Dep., 234 p. Bayliff, W. H. 1980. Synopsis of biological data on the northern bluefin tuna, Thunnus thynnus (Linnaeus, 1758), in the Pacific Ocean. Inter-Am. Trop. Tuna Comm., Spec. Rep. 2: 261-294. Bayliff, W. H., and T. P. Calkins. 1979. Information pertinent to stock assessment of north- ern bluefin tuna, Thunnus thynnus, in the Pacific Ocean. Inter-Am. Trop. Tuna Comm., Intern. Rep. 12, 78 p. Beauchamp, J. J., and J. S. Olson. 1973. Corrections for bias in regression estimates after logarithmic transformation. Ecology 54:1403-1407. Clemens, H. B., and W. L. Craig. 1965. An analysis of California's albacore fishery. Calif. Dep. Fish Game, Fish Bull. 128, 301 p. Clemens, H. B., and G. A. Flittner. 1969. Bluefin tuna migrate across the Pacific Ocean. Calif. Fish Game 55:132-135. GULLAND, J. A. 1963. The estimation of fishing mortality from tagging experiments. Int. Comm. Northwest Atl. Fish., Spec. Publ. 4:218-227. PELLA, J. J., AND C. T. PSAROPULOS. 1975. Measures of tuna abundance from purse-seine oper- ations in the eastern Pacific Ocean, adjusted for fleet- wide evolution of increased fishing power, 1960-1971. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull. 16:281-400. RlCKER, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Fish. Res. Board Can., Bull. 191, 382 p. SCHULTZE, D. L., AND R. A. COLLINS. 1977. Age composition of California landings of bluefin tuna, Thunnus thynnus, 1963 through 1969. Calif. Dep. Fish Game, Mar. Res. Tech. Rep. 38, 44 p. Whitehead, S. S. 1931. Fishing methods for the bluefin tuna (Thunnus thynnus) and an analysis of the catches. Calif. Dep. Fish Game, Fish Bull. 33, 32 p. Yamanaka, H., and Staff. 1963. Synopsis of biological data on kuromaguro Thun- nus orientalis (Temminck and Schlegel) 1842 (Pacific Ocean). FAO Fish. Rep. 6(2):180-217. 119 INTERACTIONS BETWEEN FUR SEAL POPULATIONS AND FISHERIES IN THE BERING SEA Gordon L. Swartzman and Robert T. Haar 1 ABSTRACT In this paper we consider fur seal-fisheries interaction in the Bering Sea by asking whether the slower than originally predicted recovery of the fur seal stock from female fur seal harvest during 1956-68 might be a result of a reduction in carrying capacity because of the large fishery harvest of walleye pollock and Pacific herring— fish which are important fur seal prey. The changes we found occurring in the fur seal population did not support the hypothesis that fur seal carrying capacity was reduced by the fisheries. In fact the population parameters changed little, or changed in a direction opposite to that proposed by the hypothesis. Study of the fur seal diet data indicated that walleye pollock comprised a larger part of the fur seal diet in the 1970s, after the establishment of the fishery, than earlier, although average pollock size appeared to drop significantly. This trend may have been induced by an increased harvest of older fish. Since walleye pollock are cannibalistic, the removal of the older fish by the fishery could result in lower mortality among the younger pollock stocks, the outcome being an increase in the pollock resource available to both the fishery and the fur seal. In this paper we assess and clarify possible rela- tionships between fur seals and fisheries in the Bering Sea. The event most prominent in focus- ing concern on fur seal-fisheries interactions was the failure of the Pribilof Islands' fur seal herd to recover as predicted from large female harvests during 1956-68. While the present herd appears to have stabilized, it has stabilized at a population 30% below the maximum sustained productivity estimates made in 1955 (York and Hartley 1981). A number of possible explana- tions for this have been presented, including re- duced fur seal carrying capacity. In this paper we 1) briefly summarize and highlight the available fur seal and fish data, in- cluding studies of cases of other known marine mammal-fish interactions, 2) consider the evi- dence about fur seal population dynamics and seal-fish interactions, and 3) suggest analyses of existing data and further field sampling needed to clarify the effect of the Bering Sea fishery on fur seal populations. AVAILABLE DATA The relevant data may be divided into fur seal data, Bering Sea fish stock and fishery data, and anecdotal marine mammal-fish interaction data. 'Center for Quantitative Science, University of Washington, Seattle, WA 98195. The fur seal data consist of 1) annual fur seal col- lections at sea during 1958-74 in the eastern North Pacific Ocean and the eastern Bering Sea conducted jointly by the United States and Can- ada under terms of the Fur Seal Interim Conven- tion (Kajimura et al. 1979, 2 1980 3 ); 2) harvests from 1950 to 1978 on the Pribilof Islands of sub- adult males (Lander 1981) and counts of harem and nonharem bulls from 1905 to 1978 on other island rookeries; 3) estimates of pup production on the Pribilof Islands from 1912 to 1924 and from 1951 to 1979 (Johnson 1975; Lander 1981), and counts of dead pups from 1950 to 1979 (Lan- der 1981); and 4) studies of fur seal rookery be- havior (Bartholomew and Hoel 1953; Gentry 4 ), food habits (Spalding 1964; May 1937; Wilke and 194-1**- Manuscript accepted July 1982. FISHERY BULLETIN: VOL. 81. NO. 1. 1983. 2 Kajimura, H.. R. H. Lander, M. A. Perez. A. E. York, and M. A. Bigg. 1979. Preliminary analysis of pelagic fur seal data collected by the United States and Canada during 1958- 74. Report submitted to the 22d Annual Meetingof the Stand- ing Scientific Committee, North Pacific Fur Seal Commission, 247 p. Northwest and Alaska Fisheries Center National Ma- rine Mammal Laboratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, WA 98115. 3 Kajimura, H., R H. Lander, M. A. Perez, A. E. York, and M. A. Bigg. 1980. Further analysis of pelagic fur seal data collected by the United States and Canada during 1958-74. Part 1. Submitted to the 23d Annual Meetingof the Standing Scientific Committee, North Pacific Fur Seal Commission, 94 p. Northwest and Alaska Fisheries Center National Marine Mammal Laboratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, WA 98115. 4 R. Gentry, Northwest and Alaska Fisheries Center National Marine Mammal Laboratory, National Marine Fisheries Ser- vice. NOAA, 7600 Sand Point Way NE., Seattle. WA 98115, pers. commun. May 1980. 121 FISHERY BULLETIN: VOL. 81, NO. 1 Kenyon 1957; Fiscus 1979; Kajimura et al. foot- notes 2, 3), and fertility (Abegglen and Roppel 1959). Bering Sea groundfish and pelagic fisheries data, which give estimates of relative abundance, life history parameters, and migratory patterns of important fish stocks, are contained in a num- ber of Northwest and Alaska Fisheries Center (NWAFC) reports (Pereyra et al. 1976 5 ; Favorite et al. 1979 6 ; Pruter 1973; Bakkala et al. 1979 7 ). These data cover the period of development of the large foreign groundfish fishery in the eastern Bering Sea (1954-78) and include catch, catch per unit effort (CPUE), mortality, seasonal mi- gration patterns, and diets for a number of com- mercially important fish, including walleye pol- lock and Pacific herring, important food sources for the fur seal in the eastern Bering Sea. Fur Seal Data Synopsis Seal Data Collected at Sea Fur seal migration patterns were deduced from fur seals sampled at sea from 1958 to 1974. Adult males remain year-round in the Bering Sea and Gulf of Alaska, while females migrate south in winter, with smaller (younger) females tending to migrate the farthest south. Many sub- adult males also migrate south, but not nearly so far as the females. Females begin returning to the rookeries of the Pribilof Islands in June, and the rookeries are almost completely established by the end of July (Kajimura et al. footnotes 2, 3). Pelagic data were also used to construct a fur seal life table (Lander 1981) which, along with a pup production estimate, gave an overall fur seal biomass estimate for the Pribilof Islands stock of 29,000 t or 1.25 million animals. Seasonal pat- terns of growth were also computed from the 5 Pereyra, W. T., J. E. Reeves, and R. G. Bakkala. 1976. Demersal fish and shellfish resources of the eastern Bering Sea in the baseline year 1975. Proc. Rep., 619 p. Northwest and Alaska Fisheries Center Seattle Laboratory, National Marine Fisheries Service, NOAA, 2725 Montlake Blvd. E., Seattle, WA 98112. "Favorite, F., W. J. Ingraham, Jr., K. D. Waldron, E. A. Best, V. G. Wespestad, L. H. Barton, G. B. Smith, R. G. Bakkala, R. R. Straty, and T. Laevastu. 1979. Fisheries oceanog- raphy — eastern Bering Sea Shelf. Proc. Rep. 79-20, 481 p. Northwest and Alaska Fisheries Center Seattle Laboratory, National Marine Fisheries Service, NOAA, 2725 Montlake Blvd. E., Seattle, WA 98112. 7 Bakkala, R., L. Low, and V. Wespestad. 1979. Condition of groundfish resources in the Bering Sea and Aleutian area. NMFS Northwest and Alaska Fisheries Center report sub- mitted to the International North Pacific Fisheries Commis- sion, 106 p. pelagic survey data (Lander 1981). Stomach con- tent data were pooled over years by region and by month, and were presented as the frequency of occurrence (proportion of stomachs containing a particular food item), the volume and the percent of total food volume comprised by each prey type, and the number of specimens of each prey type and their percent of the total diet. Diet compo- sition of fur seal stomachs by percent volume (which we consider to be the most reliable mea- sure of prey abundance in predator stomachs) in the eastern Bering Sea is given in Table 1 (modi- fied from Kajimura et al. footnote 3) pooled by month over all years of data collection. Table 1.— Major species in fur seal diets in the east- ern Bering Sea (percent volume), June-September. (Kajimura et al. footnote 3). Species June July August September Herring — 0.2 13.2 0.2 Capelin 699 16.4 170 15 2 Pollock 4.1 50.9 26.1 383 Deepsea smelt — 4.0 3.5 8.6 Atka mackerel 19 4 1.5 1.7 18 Squid 4.9 22.0 29.4 17.5 Fur seals are pelagic feeders and are highly opportunistic (Kajimura 1981 8 ), feeding on a wide variety of species. Of their major prey only pollock and herring are target species for a fish- ery. Data on fur seal diets outside the eastern Bering Sea corroborate the pattern of fur seals feeding primarily on schooling fish. South of British Columbia, hake replaces pollock in seal stomachs and herring and sand lance are increas- ingly important, while capelin decreases in im- portance. Anchovy is the most important fur seal food off California. Since fur seals and fisheries both tend to exploit schooling species, a possible competitive relationship may exist between fur seals and fisheries. Most fur seal feeding in the Bering Sea is done by lactating females during the summer pupping period, so the importance of food during this period cannot be overempha- sized. Since this is the period of rapid pup growth and is also the period of maximum growth for nonpregnant females and subadult males (Fig. 1), food limitation during this period could have drastic consequences to pup survival, especially after they leave the rookeries. 8 Kajimura, H. 1981. The opportunistic feeding of north- ern fur seals off California. Unpubl. manuscr., 46 p. North- west and Alaska Fisheries Center National Marine Mammal Laboratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, WA 98115. 122 SWARTZMAN and HAAR: INTERACTIONS BETWEEN FUR SEALS AND FISHERIES 130 120 110 100 - 90 80 ?0 7 years 6 years 5 years _1_ Jan Feb Mar Apr May June July Aug See Oct Nov Dec Figure 1.— Seasonal pattern of growth in mean length (cm) of nonpregnant female fur seals of age 1-7. Curves are drawn by inspection with the restriction of no downward curvature. An x designates <10 seals. From Lander (1981). Sampling on the Fur Seal Rookeries The herds on the Pribilof Islands (St. Paul and St. George Islands and Sea Lion Rock) are esti- mated to comprise 80% of the total world fur seal population. Every year from 1912 to 1924 and since 1950 some census of pup births has been made. Dead pup counts have also been made. Harvests of subadult males on the island hauling grounds have yielded information on weights, lengths, and age composition of these animals as well as limited food data from stomach samples. An estimate has also been made annually of num- bers of harem bulls. From 1956 to 1968 almost 300,000 females were harvested from St. Paul and St. George Islands, presumably to increase the sustained productivity of the herds. The herd subsequently failed to achieve a higher sustained productivity as was postulated from higher pregnancy and survival rates predicted from population projec- tions (Abegglen et al. 1956 9 ). From 1912 to 1924, pup populations were esti- mated from direct counts. Fur seal populations increased steadily over this period at an 8% an- nual rate, as they recovered from heavy losses due to pelagic sealing in the late 19th and early 20th centuries. Direct counts were discontinued from 1924 to 1948, but an 8% annual population increase was assumed. However, estimates of pups in 1948 showed that the 8% increase had not continued. In 1947, tagging studies were set up to estimate pups and were continued until 1961. In 1960 an estimation procedure involving pup shearing and direct counts was initiated to re- place the tagging method. Estimates of the num- ber of pups born were computed by adding live pup estimates to dead pup counts. The 1951-61 tagging studies are presently thought to have greatly overestimated actual pup abundance because of procedural difficul- ties and lost tags (Chapman 1973). The pup shear- ing procedure, although shown to be unbiased by comparing pup estimates with direct counts on small rookeries (Chapman and Johnson 1968), may be biased for large rookeries in such a way as to underestimate actual pup numbers (Fowler 10 ). Age-specific survival and weight at age were estimated from the weighing and aging of the preadult males harvested annually on the rook- eries. Male harvest was discontinued on St. George Island in 1972 to study the effect of the male population density on seal population dy- namics. Recent pup survival on St. George Island appeared lower than on St. Paul Island (Lander 1981), and this has been linked to the increased abundance of idle males on the rookeries (Fowler footnote 10). Bering Sea Fish Data Data on commercially important Bering Sea fish stocks by species have been compiled by the NWAFC. Catch data from Japanese, Russian, Korean, Polish, United States, and Canadian fishing operations have been included. The ma- jor species (in order of magnitude of catch) are walleye pollock, Theragra chalcogramma; yel- lowfin sole, Limanda aspera; Pacific herring, Clupea harengus pallasi; Pacific salmon, Onco- rhynchus spp; Pacific cod, Gadusmacrocephalus; sablefish, Anoplopoma fimbria; Pacific halibut, Hippoglossus stenolepis; other flatfish (rock sole, flathead sole, Alaska plaice, Greenland turbot, 9 Abegglen, C. F., A. Y. Roppel, and F. Wilke. 1956. Alas- ka fur seal investigations, Pribilof Islands, Alaska. Manuscr. rep., 143 p. Northwest and Alaska Fisheries Center National Marine Mammal Laboratory, National Marine Fisheries Ser- vice, NOAA, 7600 Sand Point Way NE., Seattle. WA 98115. 10 C. W. Fowler, Head, Fur seal investigations group, North- west and Alaska Fisheries Center National Marine Mammal Laboratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, WA 98115, pers. commun. June 1980. 123 FISHERY BULLETIN: VOL. 81. NO. 1 and arrowtooth flounder); and Pacific ocean perch, Sebastes alutus. Pacific herring and wall- eye pollock (hereafter referred to as herring and pollock) are the most important of these species in the diet of fur seals in the Bering Sea, and have been heavily fished (as have yellowfin sole, hali- but, and Pacific ocean perch). The intensity of fishing on herring and pollock suggests the pos- sibility of fur seal stock depletion due to de- creased food abundance, although stock deple- tions can also have other causes. Figure 2, adapted from Pereyraetal. (footnote 5) and Favorite et al. (footnote 6), gives the total catch for pollock and herring as well as an index of relative abundance (CPUE) based on research trawl surveys conducted by the International Pa- cific Halibut Commission, the National Marine Fisheries Service (NMFS), and the Japanese Fishery Agency. Pollock stocks have been heavily fished since 1964, with peak yields coming in the early 1970's. A steady increase in CPUE between 1964 and 1968 may have been due in part to improvements in fishing gear and tactics, but must also have been due to higher levels of recruitment of young fish (Pruter 1973), possibly because of reduced cannibalism. Pruter ( 1973) pointed out that, since only a few age groups of pollock are utilized in 1-8 to c o sz 10 -r- CO 3 o 160- herring — catch 120- '/ i '/ i '/ X '/ i '/ i '/ i 1 / * — CPUE 80- i / * / I / 1 / 1 I 1 I 40- \ \ 1 1 i i «T 20 c o o CD E 3 « 8 3 o sz "O CD &_ "O c 3 sz A s\ 15- 10- / / / / 1 1 1 § 1 J \f\ r * \ / * \ / % / \ y \ 5- 1 1 1 pollock* — catch -—CPUE 1 1 1 -4 -2 -30 20 3 O JZ CD LU Q. Z> en D. c O 2 o "55 E 10 3 o si CD LU q. 3 ; CD < 5.5 5.0 j i i l i i i i i i 54 56 58 60 62 64 66 Year class Figure 3.— Estimated average age at first reproduction of fe- male northern fur seals based on females pregnant at least once for the 1954-64 year classes. From Kajimura et al. (text foot- note 3). 126 SWARTZMAN and HAAR: INTERACTIONS BETWEEN FUR SEALS AND FISHERIES entire population would leave the nonrookery population with a higher proportion of imma- tures which would then affect the samples taken at sea. Another difficulty with these data is that only 2 yr of pre- 1956 age class data were avail- able from the pelagic cruises, and the other pre- 1956 data reported by Kajimura et al. (footnote 3) may not have used the same index of maturity as Kajimura et al. (footnote 3). Other possible sources of bias in the age at maturity estimate were the tendency of the pelagic fur seal samples to contain a higher number of older individuals than expected, and the underlying assumption that survival rates of pregnant and nonpregnant females are the same (Kajimura et al. footnote 3). Growth With Age Preliminary analysis by the National Marine Mammal Laboratory (NMML) (Fowler footnote 10) of the data from 3-yr-old males harvested on the Pribilof Islands showed a statistically signifi- cant increase in weight over time from 1964 to 1970 in contrast to growth rate reductions to be expected under a reduced fur seal carrying ca- pacity. Kajimura et al. (footnote 2) plotted the average length of pregnant females against age for the time periods 1958-62, 1963-68, and 1969-74. Their results (Fig. 4) indicate that growth rates were greater from 1963 to 1974 than from 1958 to 130 125 120 1963-68 \ ^ ,-10 seals. From Kajimura et al. (text footnote 2) 15 J. Berdine, Judson Hall, Room 621, 53 Washington Square South, New York, NY 10012, pers. commun. August 1980. 127 FISHERY BULLETIN: VOL. 81, NO. 1 on St. Paul Island. Gentry (footnote 4) made simi- lar observations on nursing fur seals in the late 1970's and found no significant change in time at sea from those of Bartholomew's study. Pup Survival to Age 2 Lander (1981) calculated early survival rates to age 2 for male fur seals from the 1950-70 year classes. York and Hartley (1981) analyzed these estimates, using Mann-Whitney and Student's t tests, and found pre- 1956 rates to be significantly lower than post-1956 rates (0.32 vs. 0.40 aver- age). This does not appear to support the hypothe- sis of reduced carrying capacity. Time Trends in Fur Seal Diets Fur seal stomach contents taken in 1960, 1962- 64, 1968, and 1973-74 cruises were used to inves- tigate trends in fur seal diets to see whether these might have changed after development of the pollock fishery. These data were summarized by month. Figure 5 indicates that the age composition in catch in the pollock fishery shifted from a mode of 4 yr in 1964 to 3 yr in 1974 with the 2-yr-old catch also being strongly represented. H. Kaji- mura, who was present on the cruises, suggested that the size of pollock in fur seal stomach sam- ples decreased from 1964 to 1974. Examination of average volume per pollock specimen in fur seal stomachs (Unpubl. data 16 ; Table 2) corrobo- rates this observation, with average specimen size decreasing significantly between 1968 and 1973-74. We also note that the percentage volume of the total stomach content comprised of pollock was consistently high in 1973-74 (>48%), while earlier, especially before 1968, pollock comprised a variable and usually low percentage of the diet (<20% in 8 of the 11 mo sampled). These data indicate that there may have been an interaction between fur seal diets and the pol- lock fishery. As fishing pressure on pollock in- creased, fishing out of older age classes reduced the average size of the fish and increased the average growth rate of the pollock. Furthermore, young pollock survival may have been increased through reduced cannibalism. These increased en a on =3 o H z to -J a > a z i— i O 8 4 O 6 4 8 4 O 8 4 8 4 8 O 8 W 4 O 8 O 8 4 O 8 O 8 4 O 1964 CW 8.53 CN 15,214 1965 CW 9.25 CN 15,462 1966 CW 6.83 CN 12,009 1967 CW 8.23 CN 14,764 968 CW I 1.84 CN 21,574 1969 CW 10.41 CN 18,965 19 70 CW 8.99 CN 19,453 1971 CW 5.79 CN 12,392 1972 CW 9.33 CN 20,493 973 CW 6.25 CN 16,347 1974 CW 5.38 CN 15,51 23456789 AGE Figure 5.— Age composition in catch per unit effort (CPUE) of walleye pollock from the Japanese trawl fishery in the eastern Bering Sea. Japanese trawl fishery includes the mothership fishery and the North Pacific trawl fishery, but not land-based- dragnet fishery. From Salveson and Alton (text footnote 12). C W = CPUE in weight in metric tons; CN = CPUE in number. 16 Data obtained from Dr. M. Tillman, Director, Northwest and Alaska Fisheries Center National Marine Mammal Lab- oratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, WA 98115. stocks of smaller fish were reflected by the in- crease in abundance of pollock in fur seal diets after 1968 and by a marked decrease in the aver- age size of fish taken by the fur seals. This in- 128 SWARTZMAN and HAAR: INTERACTIONS BETWEEN FUR SEALS AND FISHERIES Table 2. — Fur seal dietof walleye pollock from pelagic samples in theeastern Bering Sea. (Unpubl. data (text footnote 16).) Number of Volume of pollock diet Number of Percent of Pollock volume/ stomachs i n pollock in diet total specimen (cm 3 ) Date with food cm 3 Percent numbers June 1960 4 385 12.3 19 5.2 20.26 July 1960 152 39.807 61 403 9.8 98.7 Aug 1960 61 37.124 75 148 10 251 June 1962 53 295 24 2 0.16 147.5 July 1962 137 4.343 126 45 1.1 96.5 Aug. 1962 277 17,266 183 323 3.1 53.45 Sept. 1962 111 10,342 28 235 5.4 44.0 July 1963 256 11.188 14.16 62 0.56 180.45 Aug 1963 536 9,758 5 163 0.59 59.9 Sept. 1963 17 700 11.06 1 0.11 700 July 1964 97 2.354 95 7 0.27 336 Aug 1964 213 29,296 15.4 792 98 37 July 1968 78 31,901 76.9 384 14.3 83 Aug. 1968 53 11.206 37.4 30 1.21 3735 July 1973 148 72,427 907 1.418 33.0 51.07 Aug 1973 191 36,564 60.7 1,305 15.1 43.34 Sept. 1973 178 32,511 48.5 2,172 23.7 14.9 July 1974 52 13,658 87 4 244 58.6 36.0 Aug 1974 110 15,198 63.2 390 20.2 38.9 crease in total stock biomass, mostly in the youn- ger age classes, can account both for the increased fur seal diets on (mostly smaller) pollock and the continued high yield of the fishery after over 10 yr of heavy fishing pressure. Table 2 indicates that both fur seals and the fishery may have exploited the same pollock re- source, since both show a drop in size of "catch" over time. We suspect that the trend toward greatly increased abundance of pollock juveniles in the Bering Sea has also resulted in larger schools (patches) of juvenile pollock, which has made them an easier target for the fur seals and also the fishery, than previously. One possible, dangerous consequence of future increased fish- ing pressure on pollock, however, is that most of the catch will be of premature individuals. With continued heavy fishing pressure, this might re- sult in inadequate recruitment to maintain the stock. A possible alternative explanation for why pol- lock were so consistently taken by fur seals in 1973-74 is that these were relatively cold years with pollock aggregating more on the outer shelf than in warmer years (Pereyra et al. footnote 5). Another possible explanation is that the Pribilof area, where the bulk of the 1973 and 1974 stom- ach samples were taken (unlike the earlier sam- ples which did not focus as heavily on this area), is a nursery area for young-of-the-year pollock, which may account for the reduced average size and increased abundance of pollock in fur seal stomachs during 1973 and 1974. Despite these possible alternatives, the most plausible hypothe- sis is that pollock has increased in importance in fur seal diets since the initiation of the pollock fishery. Energetics Approach to Fur Seal Food Consumption The total amount of food consumed by fur seals (and other marine mammals as well) in the east- ern Bering Sea has been estimated by a num- ber of individuals (Laevastu and Larkins 1981; McAlister and Perez 1976 17 ; Anonymous 1979 18 ). McAlister and Perez (footnote 17) estimated that fur seals eat 378,000 t of fish and squid every year. They used an estimated feeding rate of 7.5% body weight daily while Miller ( 1978) 19 suggested that 14% body weight daily may be more appro- priate to support seals at 7°C, the average sum- mer temperature in the Bering Sea. Miller based his arguments on metabolic studies in which he recorded oxygen consumption at different tem- peratures in the laboratory for a number of juve- nile seals and also conducted feeding studies using food most commonly found in the diet of "McAlister, W. B., and M. A. Perez. 1976. Preliminary estimates of pinniped-finfish relationships in the Bering Sea. Background paper for the 19th meeting of the North Pacific Fur Seal Commission, 29 p. Northwest and Alaska Fisheries Center National Marine Mammal Laboratory, National Ma- rine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, WA 98115. 18 Anonymous. 1979. Draft environmental impact state- ment of the Interim Convention on Conservation of North Pa- cific Fur Seals. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Seattle, Wash., 39 p. 19 Miller, L. K. 1978. Energeticsof the northern fur seal in relation to climate and food resources of the Bering Sea. U.S. Marine Mammal Commission Report MMC-75/08, 27 p. 129 FISHERY BULLETIN: VOL. 81. NO. 1 fur seals in the Bering Sea. Using Miller's esti- mate for consumption would give an estimate of 705,000 t eaten annually by fur seals. Laevastu and Larkins ( 1981 ) gave an estimate of 513,000 t taken by fur seals annually in the eastern Bering Sea, with an additional 368,000 t taken in the Aleutian region. The latter estimates were based on runs of the PROBUB (prognostic bulk bio- mass) model. Estimates of fur seal populations of the Bering Sea and the Aleutian Islands and their mean consumption rates, given in Table 3 (Anonymous footnote 18), were used to compute a total fur seal consumption of 219,000 t. These estimates can be compared with annual fish catches in the eastern Bering Sea and Aleu- tian Islands (North Pacific Fishery Management Council 20 ). Between 1968 and 1976, annual fish catches varied between 750,000 and 2,100,000 t in the eastern Bering Sea and between 40,000 and 80,000 t near the Aleutian Islands. These fig- ures indicate that fish harvests by marine mammals and by man in the Bering Sea are com- parable and that the marine mammals' harvest exceeds man's in the Aleutian Islands' area. It is important to note, however, that fur seals prey on a larger number of species than man, and thus a part of their harvest is not in direct competition with man's. As a consequence of the fur seals' greater ability to switch prey when abundances of preferred prey species are low, total fur seal consumption is probably fairly steady from year to year, while man's is highly variable. It has been estimated (Anonymous footnote 18: table 12) that 9.8% of fish standing stock in the eastern Bering Sea and the Aleutian Islands is consumed annually by marine mammals, 5% by man, and 1.8% by birds (1.9% by fur seals). Lae- vastu and Larkins (1981) estimated a total com- mercial fish standing stock of 24,880,000 1 in the Bering Sea and Gulf of Alaska, which implies that 3.5% of all commercial fish stocks are taken by fur seals annually and 10.7% by all marine mammals. The fur seal figures are deceptive, since fur seal impact on fish stocks is relatively localized. Thus, fur seals near the Pribilof Islands are probably consuming considerably more fish than man is, though man may be harvesting some different species than fur seals. This ener- getics computation is inconclusive with respect to fur seal-fishery interaction, except to show that competition between the two is possible. DISCUSSION Suggested Analyses of Existing Data Population Indices -"Data available from North Pacific Fishery Management Council, 333 W. Fourth Ave., Suite 32, P.O. Box 3136DT, An- chorage. AK 96813. Following Eberhardt and Siniffs (1977) sug- gestion that a population's response to impact may be reflected by various indices, we suggest Table 3. — Fur seal population estimates at sea (June-November) in the eastern Bering Sea and Aleutian area (Anonymous text footnote 18). June-Nov. Estimated Estimated eastern percent of population Mean 2 Mean 3 daily Population 1 Bering Sea time at sea at sea weight consumption Age class total and Aleutian (June-Nov.) (June-Nov ) (kg) rate (%) Pups 349.000 '321,000 10 32,100 1000 14.00 M+F, age 1 174,000 67,000 5 90 78,300 9.54 13.76 M+F, age 2 122,000 61.000 5 75 45,750 1669 12 32 F. age 3 55,000 23.000 5 80 22,400 1880 12.53 F, >age 4 582,000 46,000 6 79 368.140 35.64 11.76 M. age 3-7 101,000 71,000 7 10 7.100 3260 7.60 M, >age 7 11,000 9,000 7 10 900 105.25 701 Total 1,394,000 1,043,000 "(754,100) 554.690 2992 11.71 'Average 1969 to 1974. 2 Based on National Marine Mammal Division, NMFS pelagic research data, 1958-74, N = 13,772, except average weight for pups (10 kg) based on observations in the Pribilof Islands during September; total mean weight based on an effective fishery population 754,000, on time spent on land and at sea for each class during June and September. 'Weighed by mean animal weight of estimated body weight for animals weighing <10 kg or <45 kg in waters colder than 15°C; 7% for >10 kg on land or >45 kg at sea. "Based on the ratio of males to females (0.085) in the eastern Bering Sea during June-November from National Marine Mammal Division, NMFS pelagic research data, 1958-74 (N = 4,451). 5 8% mortality, pups estimated to feed at sea only 18 d (10% of time) during September-November 6 These percentages represent proportions of the total population of the respective age class not on the rookeries during the breeding season 7 Based on percent of time out of 130 d not on rookery. 8 Effective fishery population (June-November) 130 SWARTZMAN and HAAR: INTERACTIONS BETWEEN FUR SEALS AND FISHERIES that the available data from which these indices are computed be also studied for trends. Indices that are most easily obtained for the fur seal are pup birth estimates, dead pup counts, male sur- vival to age 3 (from male harvest data), and length at age for preadult males (from harvested males). Fur Seal Diet Trend We have suggested a relationship between fur seals and the fishery via greatly increased abun- dance of juvenile pollock (Table 2). The data used, however, were already combined in such a way that we were unable to separate the data by region where the data were collected and the de- gree of digestion of the prey. We suggest that the original data be used to conduct a complete sta- tistical analysis with corrections made for the area in which the sample was taken and, if pos- sible, the time of day the samples were taken (assuming that the correlations found between the proportion of the stomachs empty and time of day the samples were taken also applies to the percentage of food digested). Variance estimates can also be computed and used to make statisti- cal tests for time trends both in the average size of pollock in fur seal stomachs and in the percent- age of the total diet comprised of pollock. Role of Patchiness in Seal Feeding Although we suspect from survey data on pol- lock (Smith 1979) that pollock are quite patchily distributed in the eastern Bering Sea, the survey data need to be reexamined for an indication of the size of patches or degree of aggregation. An . attempt should be made to represent this patchi- ness stochastically (in terms of probability). One important question to be considered with these data is whether or not there has been a trend in pollock school size from 1963 to 1974 in the east- ern Bering Sea. Another approach to consider patchiness is to use the abundance of pollock in fur seal stomachs collected at different locations as an index to the spatial separation and size of pollock schools. Suggested Future Data Collection We suggest that a fish trawl survey targeting on pollock be conducted between the Pribilof Islands and Unimak Pass from June to Septem- ber with study designed to focus on areas of high pollock density to determine the size distribution of pollock, the size of the schools, and, if possible, to observe fur seal feeding intensity around the schools. The pollock and fur seals might be tracked by using multibeam sonar techniques. Additional stomach samples of fur seal taken in conjunction with the trawl survey would give useful insight into fur seal food selectivity. CONCLUSIONS In summary, we see rookery fur seal behavior and multispecies, age-classed, patch-feeding models as directions for future study. Before pro- ceeding in this direction we recommend further detailed analyses of the fur seal stomach content data, to explore more fully the interaction be- tween the fur seal and the walleye pollock fishery (Table 2), and to elucidate other interactions with fisheries of which we may be unaware at this time. The available fur seal and fishery data, while limited, appear to be the best mammal-fishery data in the world and as such deserve to be fully archived and fully utilized. ACKNOWLEDGMENTS The authors wish to acknowledge the help of Anne York, Jerry Hornof, Mike Perez, Hiro Kajimura, Bruce McAlister, Ron Ryel, Jim Ber- dine, and Mike Tillman in conducting the work which led to this report. We are especially grate- ful to Chuck Fowler for his ideas, support, and his making data and other resources openly avail- able to us. Others whom we appreciate for help in review- ing the manuscript and/or providing suggestions on our work include Gary Smith, Robert Francis, and Taivo Laevastu at NWAFC; Nigel Bonner of Cambridge, England; Douglas Chapman at the University of Washington; Robert Hofman and Peter Major of the Marine Mammal Commis- sion; and Lee Eberhardt at Battelle Pacific Northwest Laboratory. In our office Ed Small and Stan Clark supplied computer support and Pat Sullivan editorial and clerical assistance. This work was funded in part by the Marine Mammal Commission. LITERATURE CITED Abegglen, C. E., AND A. Y. Roppel. 1959. Fertility in the northern fur seal, 1956-57. J. 131 FISHERY BULLETIN: VOL. 81. NO. 1 Wildl. Manage. 23:75-81. Bartholomew, G., and P. Hoel. 1953. Reproductive behavior of the Alaska fur seal, Cal- lorhinus ursinus. J. Mammal. 34:417-436. Chapman, D. G. 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 random- ized sampling. Trans. Am. Fish. Soc. 97:264-270. Eberhardt, L. L., and D. B. Siniff. 1977. Population dynamics and marine mammal man- agement policies. J. Fish. Res. Board Can. 34:183- 190. Fiscus, C. H. 1979. Interactions of marine mammals and Pacific hake. Mar. Fish. Rev. 41(10):l-9. Fowler, C. W. 1980. Indices of population status. In C. W. Fowler, W. T. Bunderson, M. B. Cherry. R. J. Ryel. and B. B. Steele, Comparative population dynamics of large mam- mals: A search for management criteria, p. 282-295. U.S. Mar. Mammal Comm., Wash., D.C. 1981. Density dependence as related to life history strat- egy. Ecology 62:602-610. Johnson, A. M. 1975. The status of northern fur seal populations. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 169:263-266. Lander, R. H. 1981. A life table and biomass estimate for Alaskan fur seals. Fish. Res. l(1981/1982):55-70. Laevastu, T., and H. A. Larkins. 1981. Marine fisheries ecosystem: its quantitative evalu- ation and management. Fish. News (Books) Ltd., Farn- ham, Engl., 161 p. Lett, P. F., and T. Benjaminsen. 1977. A stochastic model for the management of the northwestern Atlantic harp seal (Pagophilus groenlan- dicus) population. J. Fish. Res. Board Can. 34:1155- 1187. May, F. H. 1937. The food of the fur seal. J. Mammal. 18:99-100. Pruter, A. T. 1973. Development and present status of bottomfish re- sources in the Bering Sea. J. Fish. Res. Board Can. 30: 2373-2385. Salveson, S. J., and M. S. Alton. 1976. Pollock (Family Gadidae). In W. T. Pereyra, J. E. Reeves, and R. G. Bakkala (principal investigators), Demersal fish and shellfish resources of the eastern Be- ring Sea in the baseline year 1975, p. 369-391. U.S. Dep. Commer., NOAA, Northwest and Alaska Fisheries Center, Seattle, Wash. [Proc. Rep.] Simenstad, C. A., J. A. Estes, and K. W. Kenyon. 1978. Aleuts, sea otters, and alternate stable-state com- munities. Science (Wash., D.C.) 200:403-411. Smith, G. B. 1981. The biology of walleye pollock. In D. W. Hood and J. A. Calder (editors). The Eastern Bering Sea Shelf: oceanography and resources, Vol. 1, p. 527-551. U.S. Dep. Commer., NOAA, Seattle, Wash. Spalding, D. J. 1964. Comparative feeding habits of the fur seal, sea lion and harbor seal on the British Columbia coast. Fish. Res. Board Can. Bull. 146, 52 p. Wilke, F., and K. W. Kenyon. 1957. The food of fur seals in the eastern Bering Sea. J. Wildl. Manage. 21:237-238. York. A. E.. and J. R. Hartley. 1981. Pup production following harvest of female north- ern fur seals. Can. J. Fish. Aquat. Sci. 38:84-90. 132 AGE, SIZE, GROWTH, AND CHEMICAL COMPOSITION OF ATLANTIC MENHADEN, BREVOORTIA TYRANNUS, FROM NARRAGANSETT BAY, RHODE ISLAND Ann Gall Durbin, Edward G. Durbin, Thomas J. Smayda, and Peter G. Verity 1 ABSTRACT Age and size were determined for 2,015 Atlantic menhaden caught in Narragansett Bay, R.I., during 1976. Atlantic menhaden were predominantly age 2 and age 3, and in all age groups were significantly smaller than fish caught from Long Island Sound to the Gulf of Maineduring 1955-71. The chemical composition of the Atlantic menhaden, as determined from analysis of selected sub- samples, was ash— 10.94. carbon— 56.61, and nitrogen— 8.03% of dry weight; kilocalories— 6. 238 per gram dry weight and 7.002 per gram ash-free dry weight; and dry weight — 33.4% of wet weight. Instantaneous annual growth rates during the years 1970-75 were estimated from back- calculated fork lengths and wet weights at the time successive scale annuli were formed. Instanta- neous daily growth rates of Atlantic menhaden in Narragansett Bay during 1976 were estimated from the growth of the scale margin beyond the 1976 annulus, and from the increase in mean fork length and wet weight of the fish as the season progressed. Growth rates of age 2 and age 3 Atlantic menhaden in 1976 were considerably greater than the respective average growth rates estimated for previous years, suggesting significant differences in age-specific growth rates of Atlantic menhaden in different regions and different years. The Atlantic menhaden, Breroortia tyrannus, is a schooling, plankton-feeding clupeid which ranges inshore along the Atlantic coast from Florida to Maine. It makes extensive seasonal migrations, moving north during spring and south during fall (Nicholson 1971, 1972, 1978). Atlantic menhaden are usually present in Narragansett Bay, R.I., from April to Novem- ber, with peak abundance from June to mid- September. Here we report measurements of age, size, and chemical composition of menhaden caught in Narragansett Bay during 1976. We also report the first calculations of instantaneous growth rates in fork length and wet weight, as measured from scale annuli of individual fish. These data are part of a larger study to deter- mine the energy budget of adult menhaden in Narragansett Bay. METHODS Atlantic menhaden were sampled from the catch of two purse seiners, operating from Point Judith, R.I. During 1976, fishing activity fluc- tuated considerably, according to abundance and availability of Atlantic menhaden in Nar- 'Graduate School of Oceanography, University of Rhode Island, South Ferry Road, Narragansett, RI 02882. ragansett Bay. Most of the catch was obtained during early June and from late July to early September. All samples were collected during these two periods, with two additional samples collected on 7 October and one on 4 November. Random samples of fish from a purse seine set were stored on ice and returned to the laboratory at the end of the day. A total of 2,262 fish were sampled from 24 purse seine sets. An average of 94 fish were collected per set, with a maximum of 2 sets sampled on a given day. About 11% (247 fish) that had regenerated scales, and therefore could not be aged, were excluded from further analysis. Wet weight and fork length were recorded, and several scales were collected for age deter- mination (June and Roithmayr 1960). Every fifth fish from each sample was collected into a subset of five fish and frozen for dry weight de- termination or chemical analysis. Dry weights, for the calculation of wet weight:dry weight ratios, were determined by drying groups of these frozen fish at 105°C to constant weight. Fish used for chemical analysis were homogen- ized, while still frozen, with an equal volume of distilled water. Ash, carbon, nitrogen, and calor- ic contents were determined for subsamples of the freeze-dried homogenate. The ash content was measured by combusting samples at 475°C Manuscript accepted Julv 1982. FISHERY BULLETIN: VOL. 81. NO. 1, 1983. 133 FISHERY BULLETIN: VOL. 81. NO. 1 for 4 h (4-8 replicates). The carbon and nitrogen contents were determined with a Hewlett- Packard 2 Model 185B CHN Analyzer (3 repli- cates) and the caloric content with a Parr adiabatic bomb calorimeter (4 replicates). Five scales from each fish were mounted dry between acetate sheets and examined under a Wild M5 dissecting microscope at 18X. Annuli were counted, and distances from the focus to each annulus and to the scale margin were measured with an optical micrometer on the most symmetrical and clearly marked scale. Condition factor (CF) was calculated from the following equation: CF wet weight (g) X 10 5 fork length (cm 3 ) (1) Length-weight relationships were determined from functional regression of logio wet weight on logio fork length (Ricker 1973. 1975b; Jolicoeur 1975). Functional regressions were used because experimental error existed in both the x and y values. Growth of the fish during 1976 was determined by regressing the size of the fish (y) against the date of capture (.r). Here, ordinary regressions were used because error was associated only with the y values. Table 1. — Size and condition of Atlantic menhaden caught in Narragansett Bay, R.I.. during 1976, compared with those caught in the North Atlantic during 1955-71. Means and 95% confidence limits are shown for the Rhode Island data. Size of menhaden during the years 1955-71 are taken from June and Reintjes (1959, I960); June (1961); June and Nicholson (1964); Nicholson and Higham (1964a, b. 1965a. b); Nicholson (1975). Narragansett Bay All fish 3 June-4 Nov. North Atlantic 1955-62 1955-71 Age 1 Fork length (mm) 233 Wet weight (g) 238 Condition factor 1.87 No. of fish 2 Age 2 Fork length (mm) 233+1.5 272 268 Wet weight (g) 241 ±4.8 363 — Condition factor 1.86±0.011 1 89 — No. of fish 633 Age 3 Fork length (mm) 238±1 4 296 288 Wet weight (g) 260±3.6 452 — Condition factor 1 8910 008 1.89 — No of fish 1,224 Age 4 Fork length (mm) 249±3.7 312 306 Wet weight (g) 303+14.4 545 — Condition factor 1 9110023 1.90 — No. of fish 134 Age 5 Fork length (mm) 272+10.3 321 317 Wet weight (g) 384±39.7 604 — Condition factor 1 89±0.068 1 90 — No of fish 18 Age 6 Fork length (mm) 274±1.2 329 325 Wet weight (g) 407±1.6 657 — Condition factor 1.94+0.295 1 91 — No. of fish 4 RESULTS AND DISCUSSION Atlantic Menhaden Age Structure, Size, and Condition Factor Atlantic menhaden taken from Narragansett Bay during 1976 were predominantly age 2 and age 3 (Table 1), and the relative proportions of the different age groups in the catch remained approximately constant throughout the sam- pling period. The high proportion (31.4%) of age 2 menhaden taken in the Narragansett Bay catch during 1976 was unusual, based on records from previous years. During 1955-71, age 2 menhaden usually did not migrate in significant numbers north of Long Island, although in some years large numbers were observed in New England waters (June and Reintjes 1959, 1960; June 1961; June and Nicholson 1964; Nicholson and Higham 1964a, b, 1965a, b; Nicholson 1975). Also, the age distribution in the 1976 Narragan- 2 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. sett Bay catch (Table 1 ) was quite different from that in 1975 (Ganz 1975), where, in a sample of 1,100, age 1 = 0.2%, age 2 = 14.6%, age 3 = 70.7%, age 4 = 13.4%, and age 5 = 1.6%. Age 4 and older menhaden contributed signifi- cantly in numbers and in biomass to the North Atlantic catch prior to 1966 (Nicholson 1975). However, during the mid-1960's these older age groups dwindled until they became a negligible part of the catch (Nicholson 1975). Small numbers of age 4+ menhaden in Narragansett Bay catches of 1975 (15.0%) and 1976 (7.7%) indicate that the relative abundance of these age "groups continues to be low. Menhaden caught in Narragansett Bay in 1975 (mean weight 297.6 g (Ganz 1975)), and in 1976 (this study), were considerably smaller than fish of the same age caught during 1955-71 in the North Atlantic area (Long Island Sound to Gulf of Maine) (Table 1). However, the condition factor of the 1976 fish was similar to that of fish previously caught in the North Atlantic (Table 1), implying that the basic length-weight relationship was the same. 134 DURBIN ET AL: MENHADEN AGE. GROWTH, AND CHEMICAL COMPOSITION The relationship between wet weight and fork length in the 1976 fish was determined by regressing logio wet weight on log™ fork length. The functional regressions determined for age groups 2-5 were not significantly different in slope or elevation (P<0.05), and a common rela- tionship for all ages combined was therefore determined; where W = wet weight (g) and L = fork length (mm): log™ W=- 5.3055 +3.2441 logic L (2) r= 0.9615 n = 2,015. Back-Calculated Size-at-Age and Growth Rate The fork length, at the time a menhaden formed each of its scale annuli, was calculated by direct proportion by: L S. (3) where L, = fork length ( mm) at the time scale an- nulus i was formed . Si = width of scale ( mm ) from focus to an- nulus i L c — fork length (mm) of the fish at time of capture S c = width of scale (mm) from focus to outer margin, at time of capture. Mean back-calculated fork lengths of each age group at the time of annulus formation are presented in Table 2. The overall length-weight relationship (Equation (2)) was used to convert the back-calculated fork lengths of each fish to wet weight; mean values for each age group are presented in Table 2. These back-calculated lengths and weights were then used to calculate the annual instan- taneous growth rate of each fish during previous years (Table 3), where G t (L) = log,, L(i+i, - log, L (i) (4) where G,(L) = instantaneous yearly growth rate Table 2.— Mean back-calculated fork length and wet weight of Atlantic menhaden caught in Rhode Island waters during 1976, using the overall length-weight relationship (Equation (2)). Mean ±95% confidence limit, back calculated at annulus at age Age Year (1976) class n 1 2 3 4 5 6 Fork length: 2 1974 633 103.5±1.9 179.0±1.2 3 1973 1,224 91.4+1,0 150 8+1.0 191.3+1 4 1972 134 904+3.2 1465+3.9 185 0+4.0 217.3±4.6 5 1971 18 100.9 + 103 161.4±10.6 193 5±11 6 220.7±13.2 248.5±12.5 6 1970 4 114.2+18.7 163.9±30.6 186 6±34.6 214.6±36.4 236.7±48.2 252.0+52.5 Wet weight 2 1974 633 20.4+1.2 103.5±2.4 3 1973 1,224 13.1+0.5 60.9±1.4 129.1+2 5 4 1972 134 12.8+1.8 57.3±5.0 118.7±8.4 1997+14.0 5 1971 18 180±6.2 76.2±15.2 136.1+25.9 2084+37 4 3020+45.5 6 1970 4 24 .1+12.1 784±42.5 119.4+70.9 186.9+95.6 260.6+170.1 319.5±208.7 Table 3.— Mean annual growth in fork length (L) 1 and in wet weight (W) 2 of each age group of Atlantic menhaden during previous years. These individual growth rates were then averaged to provide an estimate of the mean growth of each age group during successive years of its life. Mean ±95% confidence limit, instantaneous yearly growth at age Age Year (1976) class n 1 2 3 4 5 Fork length: 2 1974 633 0.5745+0.0189 3 1973 1.224 0.5129±0.0092 0.2406±0.0047 4 1972 134 0.4906±0.0267 0.2385±0.0148 0.1608±0.0103 5 1971 18 0.481 4±0.0923 0.1831±0.0408 0.1313±0.0244 0.1207±0.0269 6 1970 4 0.3600±0.1317 0.1299±0 1024 0.1407+0.1479 00964±0.0525 0.0620±0.0282 Wet weight: 2 1974 633 1.8636+0 0614 3 1973 1,224 1.6639+0.0300 0.7805±0.0152 4 1972 134 1.5917±0.0865 0.7737±0.0480 0.5218±0.0335 5 1971 18 1.5617±0.2993 0.5941±0.1325 0.4260±0.0791 0.391 7±0 0872 6 1970 4 1.1678+0.4272 0.4213±03322 0.4563±0.4799 0.31 28±0. 1702 0.201 1±0.0914 'Growth was calculated individually for each fish from its back-calculated fork length at the time of annulus formation in 2 succes- sive years, where instantaneous yearly growth rate, G, (L) = log. Li,+n - log. Lw 2 Growth was calculated individually for each fish from its back-calculated wet weight at the time of annulus formation, where instantaneous yearly growth rate, G, (IV) - log, Wum - log, Wai. 135 FISHERY BULLETIN: VOL. 81. NO. 1 in fork length for a fish age i L( () = back-calculated fork length at the time annulus i was formed L ((+ D = back-calculated fork length at the time annulus i + 1 was formed. The instantaneous annual growth in wet weight, Gj (W), was similarly calculated from back- calculated wet weight of each fish at the time each annulus was formed (Table 3). Growth calculated in this way is the "true growth rate" of the individual fish, as opposed to the "population growth rate" derived from the mean size-at-age of a fish population and which generally underestimates the true rate (Ricker 1975a). However, individual growth rates cal- culated according to Equation (4) may still underestimate growth of the average individual in prior years, if the back-calculations of size-at- age are affected by Lea's Phenomenon. Although Lea's Phenomenon has been observed in men- haden (June and Roithmayr 1960; Nicholson 1972), we are unable to assess the importance of this potential bias in Tables 2 and 3, because we lack the necessary information on actual mean size and seasonal growth rates of the menhaden population during 1970-75. Among age groups 3-6 (1970-73 year classes), the mean back-calculated size-at-age and the annual instantaneous growth rates of fish of equivalent age were not significantly different (P<0.05) (Tables 2, 3). Annual growth rates declined with increasing age of the fish. The mean back-calculated size-at-age of age 2 men- haden (1974 year class) was, however, signifi- cantly larger (P<0.05) than that of fish of earlier year classes (Table 2), indicating that age 2 men- haden had grown significantly more at age and age 1 than fish from the older age groups. Further information on total menhaden population movements and on age and size structure during 1976 is needed in order to evaluate the Narragansett Bay data in terms of the population as a whole. However, some pre- liminary conclusions may be drawn, based on comparisons with data from 1955 to 1971. The summer distribution of the Atlantic men- haden is discontinuous, with a southern group ranging from Florida to Virginia and a northern group (composing the main body of the popula- tion) ranging from Chesapeake Bay to Maine (June and Reintjes 1959, 1960; June 1961; June and Nicholson 1964; Nicholson and Higham 1964a, b, 1965a, b; Nicholson 1971, 1975). During summer the northern group is age-stratified along the coast, with younger fish in the more southern part of the range and older fish pre- dominating in the north. Nicholson (1971) con- cluded that age 1 menhaden were most abundant from Chesapeake Bay to New Jersey; age 2 from New Jersey to the south shore of Long Island; age 3 from Long Island Sound to Nantucket Sound; and age 4+ from Nantucket Sound to Maine. The average size of individuals within each age group also increased with latitude, especially with age 1 and age 2 fish. This size stratification was much less pronounced for age 3 and older menhaden. Since Rhode Island is located within the summer population center of age 3 menhaden, Rhode Island landings should provide a good estimate of the mean size of age 3 menhaden in the population. However, since Rhode Island is near the northern limit of the age 2 fish, we would expect the landings to represent only the larger members of this age group. Records from 1955 to 1971 suggest that age 2 menhaden caught in Narragansett Bay during 1976 were probably the larger members of the 1974 year class and were not representative of the year class as a whole. The comparatively large size-at-age and the growth rates back- calculated for the age 2 menhaden at age and age 1 (Tables 2, 3) are consistent with this sug- gestion. Menhaden of all ages (including age 2) caught in Narragansett Bay during 1976 were among the smallest fish for their age ever recorded, and resembled the very small menhaden typically caught in Chesapeake Bay in earlier years (June and Reintjes 1959, 1960; June 1961; June and Nicholson 1964; Nicholson and Higham 1964a, b, 1965a, b; Nicholson 1971). The back-calculated fork lengths of the 1976 fish demonstrated that they had been small since age 1. Size differences between age groups were also greatly reduced (Tables 1, 2, 3). The reason for the small size of menhaden caught in Narragansett Bay during 1975 and 1976 is not known. Present results are open to two interpretations: 1) Migratory patterns during 1976, and possibly 1975, did not follow the pattern observed in earlier years, and therefore the size of the menhaden from Narragansett Bay was not representative of any age group in the overall population; or, 2) there has been a significant, overall reduction since 1971 in size-at-age within the Atlantic men- 136 DURBIN ET AL: MENHADEN AGE, GROWTH. AND CHEMICAL COMPOSITION haden population. Such a reduction in size-at- age could result from a number of factors, in- cluding poor growth during age only, followed by normal growth rate; an overall decline in the mean growth rate of all age groups; or a shift in the relative proportions of different spawning groups within the population (see June 1965; Nicholson 1972), where faster growing individ- uals have declined and been replaced by slower growing individuals. Growth During 1976 Instantaneous Daily Growth Rate Mean instantaneous daily growth rates of menhaden caught in Narragansett Bay during 1976 were estimated from the seasonal increase in mean size of the fish. Such estimates, based on successive samples from a population, assume that the fish were initially of similar size and that there was no significant influx of new fish, with different growth histories, into the region during the study period; these conditions are difficult to meet with a free-ranging fish such as the menhaden. However, we have evidence that these conditions were met, at least for a 1-mo period during the study. First, back-calculated fork lengths at the most recent annulus indicated that menhaden caught in Narragansett Bay were of similar length at the start of the 1976 growing season (Fig. 1). Second, daily observa- tions by the menhaden spotter pilots suggest that our samples collected between 3 August and 1 September were derived from a single group of menhaden. Many large schools were observed moving into Narragansett Bay during the week of 26 July. No significant additional movement of schools into or out of the bay was observed until 7 September, when large schools were again seen entering the bay. Uniformity of the back- calculated fork lengths of the menhaden sampled during this period (Fig. 1) supports the fishermen's opinion that the same group of fish was being sampled. The influx of new fish into the area, observed by the commercial fishermen on 7 September, was accompanied by an abrupt shift in the mean and variance of back-calculated fork lengths of age 3 menhaden on 7-8 Septem- ber, presumably because of the mixing of new arrivals with those already present (Fig. 1). Daily growth rates of age groups 2 and 3, the most abundant age groups in the samples, were estimated for the period 3 August- 1 September 270 250 230 210 190 170 ^ 150 * 270 O U. 250 E E I h- o < UJ 230 210 19 170 150 Age 3 • x fork length at capture o x back-calculated fork length ----Slope = 0.002l23 Slope = OOOI854 9 n% //* § ni _L Age 2 ~ m x fork length at capture o x back-calculated fork ----Slope = 0002244 Slope = 0.002l79 J //- // f*T J^ 10 20 30 10 20 30 JUNE JULY - 1 ' 4 ' // "L //"L 10 20 30 10 7 4 AUG SEPT OCT NOV Figure 1.— Mean fork length ±95% confidence limits of At- lantic menhaden collected from Narragansett Bay during 1976. Curves depict the instantaneous daily growth in length (Table 4, Equations (9)-(12)). from 1) rates of increase in mean fork length and wet weight during this period (Figs. 1, 2) and 2) growth rate of the scale margin beyond the 1976 annulus (Fig. 3). en 380 340 300 260 220 180 140 3 100 "o3 340 3 300 IX 260 220 180 140 100 Age 3 ---Slope = 00093l5 -Slope = 0005888 -I L- J L, I//-L//- Age 2 --Slope = OOI0324 Slope =0007133 itf'\»\ 10 20 30 10 20 30 JUNE JULY 10 7 4 10 20 30 AUG SEPT OCT NOV Figure 2.— Mean wet weight ±95% confidence limits of At- lantic menhaden collected from Narragansett Bay during 1976. Curves depict the instantaneous daily growth in wet weight (Table 4, Equations (13H16)). 137 FISHERY BULLETIN: VOL. 81. NO. 1 04 - 03 - 02 - .c -C 1 u -a S * CO ^ CD o> 04 C 3 03 - 02 01 Age 3 Slope =0.002608 Slope =0.002010 M-t// ? // e ^?-^ J I u I Age 2 Slope = 0002695 Slope =0002140 ,^H' : // r // 10 20 JUNE 30 10 20 30 JULY 10 20 AUG 30 10 7 4 SEPT OCT NOV Figure 3.— Seasonal growth of the scale margin beyond the 1976 annulus in Atlantic menhaden collected from Narragan- sett Bay during 1976. Means ±95% confidence limits are shown. Curves depict the instantaneous daily growth of the scale margin (Table 4, Equations (5)-(8)). Growth rates in fork length and wet weight were determined by regressing log e fork length and log? wet weight vs. the date of capture (Figs. 1, 2; equations are in Table 4). Mean instantane- ous daily growth rates were equal to the slopes of the relationships. Growth of the scale margin was determined for each fish from G = log — e S. (4) where G = instantaneous growth increment S r = total width (mm) of the scale at time of capture Si = width (mm) to the most recent (1976) annular ring. The value of G provides an independent estimate of the total amount of growth by that fish during 1976, up to the time of capture. If the exact date were known when fish resumed growth during the spring of 1976, the mean daily growth rate for the entire season could be determined for each individual fish. However, since this date is unknown, the mean daily growth rate can be estimated only for the overall population, by re- peatedly sampling that population and regress- ing the individual values of G against the date of capture (Fig. 3). This approach is analogous to that already described for estimating daily growth in fork length and wet weight. Instantaneous daily growth rates of age 2 and age 3 menhaden caught within the bay during 3 August- 1 September were 0.27 and 0.26%/d in the growth of the scale margin, 0.22 and 0.21%/d growth in fork length, and 1.03 and 0.93%/d growth in wet weight (Table 4). There were no significant differences (F<0.05) between these measures of growth for age 2 and age 3 menhaden, probably because the two age groups were very similar in size. The mean daily growth rate of the scale margin did not differ signifi- cantly (F<0.05) from that of fork length, indicating that both grew in the same propor- tion. Daily growth was also estimated, as described above, for all fish collected between 3 June and 8 September (Table 4). These growth estimates were lower than those derived from fish thought to have remained within the bay during August, but only growth estimates in wet weight were significantly different (P<0.05). Table 4.— Linear regressions from which the instantaneous daily growth rates of Atlantic menhaden in Narragansett Bay during /scale widthA 1976 may be calculated, where jc = date of capture ( 1 June = day and 8 Sept. = day 100) and (A) y = log,( ; r— 77 l;(ti) (/-log, ^ \ scale width,/ fork length ( mm); and (C) y = log, wet weight (g). y values are the means of each sample of fish; n = the number of samples. The instan- taneous daily growth rate equals the slope of each regression relationship. Reg ession statistics Age 2 Age 3 Eq. Eq no. Intercept Slope±95% C L. r n no. Intercept Slope±95% C.L. r n (A) Growth of scale margin 3 June-8 Sept. (5) 11201 0.002140±0.000203 09810 21 (6) 0.08157 0.00201 0±0 000265 09641 21 3 Aug -1 Sept. (7) 0.06990 0.002695±0.001058 08767 12 (8) 0.04600 0.002608±0 000562 0.8767 12 (B) Growth in fork length 3 June-8 Sept. 0) 5.29509 0.0021 79±0.000147 09902 21 (10) 5.33835 0.001 854±0 000225 9693 21 3 Aug.-1 Sept. (11) 5.28420 0.002244±0.000490 09541 12 (12) 5.30698 0.0021 23±0.000520 09468 12 (C) Growth in wet weight 3 June-8 Sept. (13) 4.96608 0.0071 33±0.000638 09830 21 (14) 5 12665 0.005888±0.000815 0.9606 21 3 Aug.-1 Sept. (15) 4.69906 0.01 0324±0.001 949 0.9650 12 (16) 4.82850 0.00931 5±0.001 729 09660 12 138 DURBIN KT AL: MENHADEN AGE. GROWTH. AND CHEMICAL COMPOSITION The mean dates on which growth was initiated during 1976 were calculated as 10 April and 21 April for age 2 and age 3 menhaden, respectively (Equations (5) and (6) in Table 4). These esti- mates fell within the time period (March-early May) during which growth is believed to resume and the annular ring is formed (June and Roithmayr 1960). Seasonal Growth Rate In addition to these short-term estimates of daily growth rate described above, the total seasonal growth increment was determined for individual fish from the amount of growth of the scale and from back-calculations of growth in fork length and wet weight, since the 1976 annulus was formed. By early June, age 2 and age 3 menhaden had already grown considerably since their 1976 annulus was formed (Table 5). Age 2 menhaden had grown more in length and weight, and showed a greater exponential incre- ment in size, than age 3 menhaden. These results mean that during the spring of 1976, either the Table 5. — Seasonal growth of age 2 and age 3 Atlantic men- haden caught in Narragansett Bay during 1976. Absolute growth (in mm fork length and g wet weight), and the instan- taneous growth increment since the formation of the 197(i annulus are shown. Age 2 Age 3 Date of capture x±95%C.L x±95°/oC.L. 3-10 June fork length at capture (mm) 203.7±1.5 212.2±1.2 back-calculated fork length at 1976 annulus (mm) 1803±1.9 193.3±1.4 growth (mm) 23.4 18.9 exponential increment 1220 0933 wet weight at capture (g) 153.0±3.9 1790±3.9 back-calculated wet weight at 1976 annulus (g) 1020±3.6 1 132.4±3 1 2 growth (g) 51.0 466 exponential Increment 04055 0.3016 8 September fork length at capture (mm) 247 9 3 250 6" back-calculated fork length at 1976 annulus (mm) 1790±1.2 5 191. 3 + 1. 5 growth (mm) 689 59.3 exponential increment 03256 2700 wet weight at capture (g) 2928 6 303. 5 7 back-calculated wet weight at 1976 annulus (g) 101.0±2.4 1 130 5±2.4 2 growth (g) 191 8 173.0 exponential increment 1.0644 8440 'Using the length-weight relationship for age 2 menhaden, where log, o fork length = -5 4799 + 3 3166 log,o wet weight, r = 0.963, and n =633. 2 Using the length-weight relationship for age 3 menhaden, where log,o fork length = -5 2138 + 3 2062 log,o wet weight, r= 0.956, and n =1,224 3 Average size on this date (Table 4, Equation (9)). 'Average size on this date (Table 4. Equation (10)). 5 Based on data from all fish (Table 2) 6 Average size on this date (Table 4, Equation (13)). 'Average size on this date (Table 4, Equation (14)). age 2 fish had a higher instantaneous daily growth rate than age 3 menhaden, or they resumed growth in the spring earlier than the age 3 fish, or both. By 8 September the mean growth of age groups 2 and 3 during 1976 was considerably greater than the average yearly growth rates of age 2 and age 3 menhaden in other years, as estimated from the back-calculations of size-at- age (Tables 2, 3). For example, during 1976 the scale annuli of age 3 menhaden indicated that when these fish were age 2, their total exponen- tial increments in fork length and wet weight (i.e., their instantaneous yearly growth rates) were 0.2406 and 0.7805, respectively. In com- parison, by 8 September the mean exponential increments in fork length and wet weight of age 2 menhaden during 1976 were 0.3256 and 1.0644. Similarly, age 4 menhaden caught during 1976 increased in fork length by 0.1608 and in wet weight by 0.5218 as age 3 fish during 1975. During 1976, the increments in fork length and wet weight of age 3 menhaden were 0.2700 and 0.8440 by 8 September. Some additional growth may have taken place after 8 September; June and Roithmayr (1960) found that growth of the scale margin in Atlantic menhaden continued until September or October. Results indicate that significant differences in the growth rate of menhaden occur, probably because menhaden, found over an extensive geo- graphic area during the summer, experience a wide range of temperature and food conditions" that could affect growth. Further investigation into regional and annual differences in the instantaneous growth rates may provide a basis for determining which geographic regions can potentially contribute most to menhaden pro- ductivity and could provide considerable insight into ways of maximizing the yield from this fishery. Chemical Composition The mean carbon, nitrogen, caloric, and ash contents and dry weight of menhaden from Narragansett Bay are summarized in Table 6. The ratio of dry weightwet weight remained fairly constant in all samples; otherwise, there was a consistent trend in those fish with a high caloric content toward high carbon content and low nitrogren and ash content as a percent of dry weight (Fig. 4). Ash, caloric, and moisture contents of the men- 139 FISHERY BULLETIN: VOL. 81. NO. 1 Table 6.— Chemical composition of Atlantic menhaden col- lected from Narragansett Bay, R.I., between 3 June and 8 September 197R. Determinations were made on groups of five fish. No of Constituent x±95% C.L. samples Dry wtwet wt 0.334 ±0018 19 Ash. proportion of dry wt 1094 ±0.0292 21 C, proportion of dry wt 05661 ±0.0671 18 N, proportion of dry wt 08028±0 00349 18 Kcal (g dry wt fish)" 1 6.238 ±1006 20 Kcal (g ash-free dry wt fis h)" 1 7.002 ±0 942 20 5000 6000 7000 5000 5400 5800 6200 CALORIES (g dry wt )" 6600 I 7000 Figure 4.— Carbon, nitrogen, and ash contents (percent of dry weight) as a function of caloric content in Atlantic menhaden collected from Narragansett Bay. haden in the present study are similar to those re- ported for Atlantic menhaden from Beaufort, N.C., (Thayer et al. 1972) and Chesapeake Bay (Dubrow et al. 1976). Menhaden are compara- tively higher in percentage of dry weight and in caloric content than most other fish species (Dahlberg 1969; Perkins and Dahlberg 1971; Mayer et al. 1973; Sidwell et al. 1974; Small 1975; Kitchell et al. 1977; Foltz and Norden 1977). ACKNOWLEDGMENTS We wish to thank Harold A. Loftes, Captain of Ocean State, and Charles Follett, Captain of Cindy Bet, for their assistance in collection of samples and observation of the abundance and movements of the Atlantic menhaden in Narragansett Bay. This research was supported by National Science Foundation Grants OCE 76 02572 and OCE 79 19551. LITERATURE CITED Dahlberg, M. I). 19R9. Fat cycles and condition factors of two species of menhaden, Brevoortia (Clupeidae), and natural hybrids from the Indian River of Florida. Am. Midi. Nat. 82: 117-126. Dubrow, D., M. Hale, and A. Bimbo. 1976. Seasonal variations in chemical composition and protein quality of menhaden. Mar. Fish. Rev. 38(9): 12- 16. Foltz, J. W., and C. R. Norden. 1977. Seasonal changes in food consumption and energy content of smelt (Osmerus mordax) in Lake Michigan. Trans. Am. Fish. Soc. 106:230-234. Ganz, A. R. 1975. Observations of the Narragansett Bay menhaden fishery. R.I. Dep. Nat. Resour.. Div. Fish Wildl., Mar. Fish. Sec. Leafl. 45:1-21. JOLICOEUR, P. 1975. Linear regressions in fishery research: some com- ments. J. Fish. Res. Board Can. 32:1491-1494. June, F. C. 1961. Age and size composition of the menhaden catch along the Atlantic coast of the United States, 1957, with a brief review of the commercial fishery. U.S. Fish Wildl. Serv., Spec. Sci. Rep.— Fish. 373, 39 p. 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 W. R. Nicholson. 1964. Age and size composition of the menhaden catch along the Atlantic coast of the United States, 1958, with a brief review of the commercial fishery. 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. 1960. Age and size composition of the menhaden catch along the Atlantic coast of the United States, 1956, with a brief review of the commercial fishery. U.S. F"ish Wildl. Serv., Spec. Sci. Rep.-Fish. 33R. 38 p. June, F. C, and C. M. Roithmayr. 1960. Determining age of Atlantic menhaden from their scales. U.S. Fish Wildl. Serv., Fish. Bull. 60:323- 342. Kitchell, J. F., J. J. Magnuson, and W. H. Neill. 1977. Estimation of caloric content for fish biomass. Environ. Biol. Fish. 2:185-188. Nicholson. W. R. 1971. 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- 140 DURBIN ETAL: MENHADEN AGE. GROWTH. AND CHEMICAL COMPOSITION calculated length frequencies. Chesapeake Sci. 13:161- 174. 1975. Age and size composition of the Atlantic men- haden, 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. Nicholson, W. R., and J. R. Higham, Jr. 1964a. Age and size composition of the menhaden catch along the Atlantic coast of the United States, 1959, with a brief review of the commercial fishery. U.S. Fish Wildl. Serv., Spec. Sci. Rep.-Fish. 478, 34 p. 1964b. Age and size composition of the 1960 menhaden catch along the U.S. Atlantic coast, with a brief review of the commercial fishery. U.S. Fish Wildl. Serv., Spec. Sci. Rep.-Fish. 479, 41 p. 1965. Age and size composition of the menhaden catch along the Atlantic coast of the United States, 1961, with a brief review of the commercial fishery. U.S. Fish Wildl. Serv., Spec. Sci. Rep.-Fish. 495. 28 p. 1966. Age and size composition of the menhaden catch along the Atlantic coast of the United States. 1962, with a brief review of the commercial fisherv. U.S. Fish Wildl. Serv., Spec. Sci. Rep.-Fish. 527. 24 p. Perkins, R. J., and M. D. Dahlberg. 1971. Fat cycles and condition factors of Altamaha River shads. Ecology 52:359-362. Ricker, W. E. 1973. Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. 1975a. Computation and interpretation of biological statistics of fish populations. Fish. Res. Board Can. Bull. 191:1-382. 1975b. A note concerning Professor Jolicoeur's com- ments. J. Fish. Res. Board Can. 32:1494-1498. Sidwell. V. D., P. R. Foncannon, N. S. Moore, and J. C. Bonnet. 1974. Composition of the edible portion of raw (fresh or frozen) crustaceans, finfish, and mollusks. I. Protein, fat, moisture, ash. carbohydrate, energy value, and cholesterol. Mar. Fish. Rev. 36(3):21-35. Small. J. W., Jr. 1975. Energy dynamics of benthic fishes in a small Ken- tucky stream. Ecology 56:827-840. Thayer, G. W.. W. E. Schaaf, J. W. Angelovic, and M. W. LaCroix. 1973. Caloric measurements of some estuarine organ- isms. Fish. Bull.. U.S. 71:289-296. 141 NOTES HOMING AND FISHERIES CONTRIBUTION OF MARKED COHO SALMON, ONCORHYNCHUS KISUTCH, RELEASED AT TWO COLUMBIA RIVER LOCATIONS In 1970 we conducted an experiment to deter- mine if coho salmon, Oiieorhynchus kisutch, re- leased away from the rearing site would return to the release area and contribute to the fisheries there (Vreeland et al. 1975). We found the coho salmon returned almost exclusively to the re- lease area and contributed to the fisheries near the release site. However, because the single fin marks applied were duplicated by other experi- menters on the Pacific coast, we could not evalu- ate the contribution of the two groups to the ocean fisheries. We also surmised a possible detrimen- tal effect of transportation on the survival of the group released downstream from the hatchery. In 1972 we initiated a study with 1971-brood coho salmon to 1) confirm the homing results of the previous study, 2) eliminate possible differ- ences in survival due to transportation, and 3) determine the contribution of the release groups to the Pacific coast fisheries. Methods We chose coho salmon originally from Klaska- nine Hatchery in Oregon, the same fish stock used in the previous study. Hatchery personnel collected adults and took eggs at Little White Salmon National Fish Hatchery, located near Cook, Wash., on the Little White Salmon River about 1.5 km (1 mi) upstream from its confluence with the Columbia River and 242 km (150 mi) from the Pacific Ocean (Fig. 1). Coho salmon were reared at Willard National Fish Hatchery, 4.5 km (3 mi) up the Little White Salmon River from Little White Salmon Hatchery. The two groups offish were hatched and raised under uniform conditions in hatchery ponds. Fin clipping took place in September 1972 at Wil- lard Hatchery. We applied two marks to the fish: adipose right ventral (Ad-RV) and adipose left ventral (Ad-LV). Youngs Bay (Fig. 1) was selected as the release site, situated about 19 km (12 mi) upstream from the mouth of the Columbia River and fed by four small rivers: Lewis and Clark, Walluski, Youngs, and Klaskanine Rivers. We transported the Ad- RV marked coho salmon 253 km (157 mi) in about 4 h to Youngs Bay on 14 and 15 May 1973, where they were released at a public launch ramp. We transported the fish in two tank trucks, each 3,785 1 (1,000-gal) capacity. Each truck was loaded with 462 kg (1,018 lb) of fish at 57.8 fish/ kg (26.2 fish/lb) or about 26,700 fish. During the 2 d, we transported 106,852 Ad-RV marked coho salmon weighing 1,847 kg (4,072 lb) from Willard Hatchery to Youngs Bay (Table 1). To maintain similar handling procedures and equalize any possible effects of transportation on survival, we transported the Willard Hatchery release for a time and distance similar to the Youngs Bay release. On 16 and 17 May 1973, we hauled 107,707 Ad-LV marked coho salmon weighing 1,835 kg (4,045 lb) in the same two tank trucks used for the Youngs Bay release. The fish were transported about 161 km (100 mi) for 3 h and 35 min on 16 May and 182 km (113 mi) for 3 h and 50 min on 17 May. Each truck contained about 458 kg ( 1,010 lb) of coho salmon. The hatch- ery crew released all the coho salmon from Wil- lard Hatchery into the Little White Salmon River on 17 May. We used catches of marked coho salmon in the fisheries and hatchery return data to determine the effect of release site on contribution and hom- ing. Sampling for fin-marked coho salmon took place in 1973 and 1974 in the major Pacific coast salmon fisheries of Alaska, Washington, Oregon, and California, the Columbia River fisheries, and at potential hatchery return sites on the Columbia River. State fishery personnel sampled the Alaska troll fishery, the California, Oregon, and Washington ocean sport and troll fisheries, and the Columbia River gill net fishery. Person- nel from National Marine Fisheries Service sam- pled catches from the Youngs Bay gill net fishery at two fish processing plants. Table 1.— Numbers of marked coho salmon released in the Columbia River for the homing experiment. Ad-RV = adipose right ventral; Ad-LV = adipose left ventral. Fin Releases Fish/kg Release date Release mark No kg location Ad-RV Ad-LV 106.852 107.707 1,847 1,835 57.9 58.7 14-15 May 1973 17 May 1973 Youngs Bay Willard Hatchery FISHERY BULLETIN: VOL. 81. NO. 1. 1983. 143 SCALE IN KILOMETERS Columbia River y ASTORIA RELEASE SITE, ASTORIA YACHT CLUB NEW HIGHWAY 101 BRIDGE FIGURE 1.— Columbia River study area showing lo- cation of Willard and Little White Salmon National Fish Hatcheries and detailed features of the Youngs Bay region. KLASKANINE FISH HATCHERY Returns of coho salmon to hatcheries near the two release sites were examined for marked fish to assess the effect of release site on hom- ing. Hatchery personnel examined all returns to Little White Salmon National Fish Hatchery for marked coho salmon in the fall of 1973 and 1974. A series of waterfalls blocks access to Willard Hatchery; therefore, coho salmon released from Willard return to Little White Salmon Hatchery. In addition, State personnel examined all returns at the following hatcheries for marked coho salmon: Klaskanine Salmon Hatchery on the Klaskanine River (a tributary of Youngs Bay); Big Creek Salmon Hatchery on Big Creek near Knappa, Oreg.; the Elokomin Salmon Hatchery on the Elochoman River near Cathlamet, Wash.; and the Grays River Salmon Hatchery on the Grays River near Grays River, Wash. (Fig. 1). Results Homing We compared 1) the location of catch within the Columbia River and 2) return sites of the two marked groups to determine the accuracy of homing to the release site. The fish in this study returned almost exclu- 144 sively to the area of release, similar to fish in pre- vious studies (Rounsefell and Kelez 1938; Taft and Shapovalov 1938; Donaldson and Allen 1957; Ellis 1968 1 ; Jensen and Duncan 1971; Mahnken and Joyner 1973; Vreeland et al. 1975; Scholz et al. 1976). No Willard Hatchery release fish were caught in the Youngs Bay fishery, but 199 Youngs Bay release fish were caught in the fish- ery. Only two Youngs Bay releases were seen in hatchery returns, one at Klaskanine Hatchery and the other at Little White Salmon Hatchery (Table 2). Hatchery personnel observed only 26 Willard releases at Little White Salmon Hatch- ery. Construction in 1974 of a new barrier dam and fish ladder at the hatchery may have pre- vented some coho salmon from entering the hatchery ponds. However, the hatchery biologist at Little White Salmon Hatchery believed most fish entered the adult holding ponds prior to the ladder closure. 2 The specificity of the homing we observed is apparently linked to the physiological stage of parr-smolt transformation. Work by Hasler (1966) and Carlin(1968) indicated the imprinting process occurs rapidly at the time of parr-smolt transformation. With steelhead trout, Salmo gairdneri, Wagner (1969) hypothesized the hom- ing imprint is acquired rapidly before and/or during downstream migration. Mighell (1975) 3 found fish exposed to a new water source for as little as 4 h will imprint on the new source. Coho salmon released in a Lake Michigan tributary strayed extensively (Peck 1970). Hasler et al. (1978) postulated that this was due to releasing the fish after smolting had taken place. Jensen and Duncan (1971) described accurate homing with coho salmon released afsmolt size." Cooper et al. (1976) found a 2-d exposure to morpholine at the onset of smolting imprinted fish to the chemical as well as did a 30-d exposure. W. S. Zaugg (1975), 4 who has attempted to define more Table 2.— Number of 1971-brood Youngs Bay and Willard Hatchery release coho salmon recovered at five Columbia River hatcheries. 1973 and 1974. Ad-RV = adipose right ven- tral; Ad-LV = adipose left ventral. 'Ellis, C. H. 1968. A return of adult coho salmon demon- strating a high degree of selectivity in homing, In Proceed- ings of the Northwest Fish Culture Conference, December 4-6, 1968, Boise, Idaho, p. 40-42. Unpubl. manuscr. Wash. Dep. Fish., 115 Gen. Admin. Bldg., Olympia, WA 98504. 2 S. L. Leek, U.S. Fish and Wildlife Service, Little White Salmon National Fish Hatchery, Willard, WA 98605, pers. commun., September 1978. 3 Mighell, J. 1975. Some observations on imprinting of ju- venile salmon in fresh and saltwater. In Summary notes from papers presented at homing workshop. Unpubl. manuscr., p. 11-12. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 2725 Montlake Blvd. E.. Seattle, WA 98112. 4 W. S. Zaugg, Northwest and Alaska Fisheries Center, Na- tional Marine Fisheries Service, NOAA, 2725 Montlake Blvd. E., Seattle, WA 98112, pers. commun., November 1975. Your gs Bay release Willa rd Hatchery release (Ad-RV) (Ad-LV) Hatchery 1973 1974 Total 1973 1974 Total Klaskanine 1 1 Big Creek Grays River Elokomin Little White Salmon 1 1 2 24 26 Total 1 1 2 2 24 26 closely the onset of the parr-smolt transforma- tion, feels the imprinting will not occur until a certain stage of the transformation is reached. Unfortunately, none of the authors (nor do we) in- dicate a stage of the parr-smolt transformation at time of release. Time of smolting and imprint- ing has yet to be defined closely enough to predict the homing location of fish released in different areas. Until more is learned, we expect varying results could occur with homing studies depend- ing on when the fish are released. Fishery Contribution We examined ocean and Columbia River catches of coho salmon to determine the contri- bution of both release groups to the Pacific coast fisheries. Fishery samplers saw 350 Youngs Bay releases and 78 Willard releases in 1973 and 1974 (Table 3). No coho salmon from either release were observed in the catches of Alaska commer- cial fisheries. Fisheries samplers in Canada did not examine coho salmon for multiple fin marks; however, on the average, Canadian fishermen land only 6% of all Columbia River hatchery coho salmon (Wahle et al. 1974). Catches of the two marked groups occurred primarily in the Wash- ington, Oregon, and California marine fisheries and the Columbia River gill net fishery. Total estimated catches for 1973 and 1974 of Youngs Bay and Willard release groups are 2,455 and 598, respectively. Catches in the Oregon and California troll fisheries contained over 50% of both marked groups (55% Ad-RV, Youngs Bay and 61% Ad-LV, Willard releases). Washington marine recoveries occurred primarily near the Columbia River, except for catches of Willard re- lease coho salmon at LaPush on the north Wash- ington coast. Landings of Willard release fish at LaPush comprised nearly one-half of the release caught in the Washington troll fisheries. 145 Table 3.— Observed (in parentheses) and estimated catches of marked 1971-brood coho salmon released from the two Columbia River sites and recovered in Pacific coast fisheries, by fishery type and year of capture. 1 Ad-RV = adipose right ventral; Ad-LV = adipose left ventral. Fishery type Youngs Bay release (Ad-RV) Willard Hatchery release (Ad-LV) Location 1973 1974 Total Percent 1973 1974 Total Percent Alaska Commercial ( 2 ) ( 2 ) British Columbia Commercial < 3 ) ( 3 ) ( 3 ) — ( 3 ) ( 3 ) ( 3 ) — Sport C) ( 3 ) ( 3 ) — ( 3 ) ( 3 ) ( 3 ) — Washington Ocean Troll 144(18) 144 6 47(9) 47 8 Sport 219(32) 219 9 65(11) 65 11 Puget Sound Commercial Sport Oregon Troll 668(84) 668 27 229(29) 229 38 Sport 117(24) 117 5 12(1) 41(7) 53 9 California Troll 694(65) 694 28 139(18) 139 23 Sport 49(7) 49 2 5(1) 5 1 Columbia River Gill net 365(14) 365 15 60(2) 60 10 Indian 4 Sport Youngs Bay Gill net 6(2) 193(104) 199 8 Total 6(2) 2.449(348) 2,455 100 12(1) 586(77) 598 100 'Data obtained from: "1973 fin-mark sampling and recovery report for salmon a fisheries" and "1974 Wire tag and fin-mark sampling and recovery report for salmon fisheries," Fish Commission of Oregon, Clackamas, Oreg. 2 Not sampled. 3 No sampling for multiple fin-marked coho. "Setnet and dip net fisheries. nd steelhead from various Pacific Coast and steelhead from various Pacific Coast Overall fishery contribution rates for this study are lower than rates reported in studies conducted in the 1960's with coho salmon from Columbia River hatcheries. For all fisheries com- bined, the Youngs Bay release contributed 23.0 fish/1,000 released, and the Willard release con- tributed 5.6 fish/1,000 released. In a diet test at Washougal Hatchery (Senn and Noble 1968), the contribution of 1961-brood coho salmon, fed a diet similar to that fed the 1971-brood, was 51 fish/1,000 releases to the Pacific coast fisher- ies. Wahle et al. (1974) found the average con- tribution to the fisheries of 1965 and 1966 brood coho salmon was 55 fish/1,000 releases at Colum- bia River hatcheries. Fishery contributions of marked groups of 1967-, 1968-, and 1969-brood coho salmon at Cowlitz Hatchery ranged from 21 to 52 fish/1,000 releases. 5 In the earlier 1968- brood study, the Willard Hatchery release con- tributed 7.7 fish/1,000 releases to the Columbia River and Youngs Bay fisheries. We do not know the reasons for the poorer survival of the 1971- brood fish. The release site significantly affected fishery contribution despite the low survival. We believe the Youngs Bay release survived at a higher rate than the Willard release fish because the Youngs Bay release contributed more heavily to all fish- eries sampled than did the hatchery release. The contribution ratios of the Youngs Bay release to the Willard Hatchery release by fishery are 3.2:1 for Washington marine fisheries, 2.8:1 for Ore- gon ocean fisheries, 5.2:1 for California ocean fisheries, 9.4:1 for the Columbia River fisheries, and 4.1:1 overall. Differences between contribu- tion rates when all fisheries are combined are significant (x 2 = 137.36). We postulated two possible reasons for the higher fishery contribution of the Youngs Bay release. The Youngs Bay release possibly had a higher survival to the estuary than did the Wil- lard hatchery release because the former group avoided downstream-migration mortalities from predation, gas bubble disease, and from passing over spillways or through turbines at the Bonne- ville Dam. A number of authors have reported the adverse effects of Columbia River dams on survival of juvenile salmonids (Schoeneman et al. 1961; Bell et al. 1967 6 ; Long et al. 1968 7 : Bell and DeLacy 1971 8 ; Ebel et al. 1973; Slatick et al. 5 Hopley, C. W. 1975. Informal interim report on portions of 1967-, 1968-,and 1969-brood Cowlitz River cohostock timing- evaluation. In Coho marking program on the lower Columbia River. Unpubl. manuscr., p. 9-43. Wash. Dep. Fish., 115 Gen. Admin. Bldg.. Olympia, WA 98504. 6 Bell,M. C, A. C. DeLacy. and G.J. Paulik. 1967. A com- pendium on the success of passage of small fish through tur- bines. Unpubl. manuscr., 268 p. U.S. Armv Corps Eng., Portland Dist., Fish Eng. Res. Program, P.O. Box 2946, Port- land, OR 97208. 7 Long, C. W., R. F. Krcma, and F. J. Ossiander. 1968. Re- search on fingerling mortality in Kaplan turbines— 1968. Unpubl. manuscr., 7 p. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 2725 Montlake Blvd. E., Seattle. WA 98112. K Bell, M. C, and A. C. DeLacy. 1971. A compendium on 146 1975; Collins et al. 1975 9 ; Collins 1976; Ebel and Raymond 1976). Ebel (1970) found groups of fall chinook salmon, Oncorhynchus tshawytscha, re- leased below Bonneville Dam, had over twice the survival rate to the Columbia River estuary com- pared with a group released above the dam. The low flow of the Columbia River in 1973 caused a particularly serious passage and survival prob- lem for juvenile salmon because most of the river flowed through the turbines at the dams. 10 A second possible reason for the higher contri- bution of the Youngs Bay release is that the bay may provide a better rearing area than the hatch- ery release site because food is more abundant. A large concentration of the amphipod Corophium salmonis occurs in Youngs Bay, particularly in May, and is a major food item for coho salmon in the bay. 11 Abundant food could have given the Youngs Bay release an initial survival advan- tage. Summary We conducted this study to confirm previous results on the feasibility of creating or enhancing a fishery in a specific area by releasing hatchery salmon into that area. We compared the location of return and contribution with the Pacific coast fisheries of coho salmon released at two locations on the Columbia River. Two groups each of about 100,000 1971-brood coho salmon at Willard Na- tional Fish Hatchery were fin clipped: In May 1973 one group was released at Youngs Bay near Astoria, Oreg., and the other at Willard Hatch- ery. Both groups were transported an equal time and distance prior to release to equalize any pos- sible effects of transportation on survival. Marine sport and commercial salmon fisheries of the Columbia River and Youngs Bay, as well the survival of fish passing through spillways and conduits. Unpubl. manuscr., 144 p. U.S. Army Corps Eng., Portland Dist., Fish Eng. Res. Program. P.O. Box 2946, Portland, OR 97208 9 Collins, G. B., W. J. Ebel. G. E. Monan, H. L. Raymond, and G. K. Tanonaka. 1975. The Snake River salmon and steel- head crisis, its relations to dams and the national energy crisis. Unpubl. manuscr., 30 p. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 2725 Montlake Blvd. E., Seattle. WA 98112. '"Columbia River Fisheries Council. 1978. Recommenda- tions of Columbia River Fisheries Council for in-stream flows in the Columbia and Snake Rivers. Unpubl. manuscr., 24 p. Columbia River Fish. Counc, Suite 1240, Lloyd Bldg., 700 N.E. Multnomah St., Portland. OR 97232. "Durkin, J. T., S. J. Lipovsky, G. R. Snyder, and M. E. Tut- tle. 197V. Environmental studies of three Columbia River estuarine beaches. Unpubl. manuscr., 78 p. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 2725 Mont- lake Blvd. E., Seattle, WA 98112. as Columbia River hatchery returns, were sam- pled for marked coho salmon in 1973 and 1974. Over one-half of both groups of marked fish were caught by Oregon and California marine sport and commercial fishermen. Recoveries of the re- maining marked fish occurred in Washington, Columbia River, and Youngs Bay fisheries. The Youngs Bay release contributed 23 fish/ 1,000 releases to the Pacific coast fisheries, and the Willard Hatchery release contributed 5.6 fish/ 1,000 releases. The fish homed to the release site with little straying. Only one Youngs Bay release returned to Little White Salmon National Fish Hatchery. Acknowledgments We thank Ron and Lois Saling, as well as per- sonnel of the Barbey Packing Corporation, New England Fish Company, and Jessie's Ilwaco Fish Company for their cooperation and assistance in conducting this study. Our appreciation is ex- tended to the owners of the Astoria Yacht Club for the use of their boat launch ramp. We are in- debted to the hatchery crew at Willard Hatchery for their assistance with tagging and fish trans- portation operations and the hatchery superin- tendents and personnel at Klaskanine, Big Creek, Elokomin, Grays River, and Little White Salmon Hatcheries for their assistance in examining coho salmon returns for our marked fish. We also thank fishery biologists and samplers in Wash- ington, Oregon, and California for providing re- covery efforts leading to the catch data necessary for this report. Helpful editorial comments were contributed by Reino Koski, Roger Pearson, and Rae Mitsuoka, National Marine Fisheries Ser- vice, and Technical Writer-Editor Mary Lee Sibley-Armour. A special thanks to Alma Follis for help in preparation of the report. Literature Cited Carlin, B. 1968. The migration of salmon. In Atlantic Salmon Association Centennial Award Fund, series of lectures, p. 14-22. Atl. Salmon Assoc, Montreal, Que. Collins, G. B. 1976. Effects of dams on Pacific salmon and steelhead trout. Mar. Fish. Rev. 38(ll):39-46. Cooper, J. C, A. T. Scholz, R. M. Horrall, A. D. Hasler. and D. M. Madison. 1976. Experimental confirmation of the olfactory hy- pothesis with homing, artificially imprinted coho salm- on (Oncorhynchus kisutch). J. Fish. Res. Board Can. 33:703-710. 147 Donaldson, L. R.. and G. H. Allen. 1957. Return of silver salmon, Oncorhynckus kisutch (Walbaum) to point of release. Trans. Am. Fish. Soc. 87:13-22. Ebel, W.J. 1970. Effect of release location on survival of juvenile fall chinook salmon, Oncorhynckus tshawytscha. Trans. Am. Fish. Soc. 99:672-676. Ebel, W. J., D. L. Park, and R. C. Johnsen. 1973. Effects of transportation on survival and homing of Snake River chinook salmon and steelhead trout. Fish. Bull., U.S. 71:549-563. Ebel, W. J., and H. L. Raymond. 1976. Effect of atmospheric gassupersaturation on salm- on and steelhead trout of the Snake and Columbia Rivers. Mar. Fish. Rev. 38(7):1-14. Hasler, A. D. 1966. Underwater guideposts, homing of salmon. Univ. Wis. Press, Madison, 155 p. Hasler, A. D., A. T. Scholz, and R. M. Horrall. 1978. Olfactory imprinting and homing in salmon. Am. Sci. 66:347-355. Jensen, A. L.. and R. N. Duncan. 1971. Homing of transplanted coho salmon. Prog. Fish- Cult. 33:216-218. Mahnken, C, and T. Joyner. 1973. Salmon for New England fisheries. Part III: De- veloping a coastal fishery for Pacific salmon. Mar. Fish. Rev. 35(101:9-13. Peck, J. W. 1970. Straying and reproduction of coho salmon, Onco- rhynckus kisutch, planted in a Lake Superior tributary. Trans. Am. Fish. Soc. 99:591-595. ROUNSEFELL, G. A., AND G. B. KELEZ. 1938. The salmon and salmon fisheries of Swiftsure Bank, Puget Sound, and the Fraser River. Bull. U.S. Bur. Fish. 48:693-823. SCHOENEMAN, D. E., R. T. PRESSEY, AND C. 0. JUNGE, JR. 1961. Mortalities of downstream migrant salmon at McNary Dam. Trans. Am. Fish. Soc. 90:58-72. Scholz, A. T., R. M. Horrall, J. C. Cooper, and A. D. Hasler. 1976. Imprinting to chemical cues: the basis for home stream selection in salmon. Science (Wash., D.C.) 192: 1247-1249. Senn, H. G., and R. E. Noble. 1968. Contribution of coho salmon Oncorhynckus kisutch from a Columbia River watershed hatchery. Wash. Dep. Fish., Fish. Res. Pap. 3(l):51-62. Slatick, E., D. L. Park, and W. J. Ebel. 1975. Further studies regarding effects of transporta- tion on survival and homing of Snake River chinook salmon and steelhead trout. Fish. Bull., U.S. 73:925- 931. Taft, A. C, and L. Shapovalov. 1938. Homing instinct and straying among steelhead trout (Salmo gairdnerii) and silver salmon {Oncorhyn- ckus kisutch). Calif. Fish Game 24:118-125. Vreeland, R. R., R. J. Wahle, and A. H. Arp. 1975. Homing behavior and contribution to Columbia River fisheries of marked coho salmon released at two locations. Fish. Bull., U.S. 73:717-725. Wagner, H. H. 1969. Effects of stocking location of juvenile steelhead trout, Salmi) gairdneri, on adult catch. Trans. Am. Fish. Soc. 98:27-34. Wahle, R. J., R. R. Vreeland, and R. H. Lander. 1974. Bioeconomic contribution of Columbia River hatch- ery coho salmon, 1965 and 1966 broods, to the Pacific salmon fisheries. Fish. Bull., U.S. 72:139-169. Robert R. Vreeland Roy J. Wahle Environmental anil Technical Services Division Columbia River Fisheries Development Program National Marine Fisheries Service, NOAA 847 N.E. 19th Avenue, Suite 350 Portland. OR 97232 MOVEMENT PATTERNS OF BONEFISH, ALBULA VULPES, IN BAHAMIAN WATERS The regular daily movement patterns of fishes appear closely related to predictable changes in their environment. Factors such as tidal fluctua- tions (Dodson and Leggett 1973; Stasko et al. 1973), light levels (Yuen 1970; Collette and Talbot 1972; Standora et al. 1972; McFarland et al. 1979), and temperature (Coutant 1975; Kelso 1976; Haynes et al. 1978; Langford et al. 1979) have been found to influence the cyclic move- ment of fishes. Until recently, most information on such movement patterns has been obtained primarily through direct observation. However, there are many situations in which direct visual methods are not feasible. An alternate means of obtaining such information has been provided by recent advances in the use of ultrasonic telem- etry as a research tool. Ultrasonic telemetry has become a valuable technique both in freshwater and deep marine environments. However, the use of ultrasonics in coastal waters is still in the early developmental stages. Rapid signal attenuation occurs under such conditions because of combined effects of the high conductivity of the water, vegetative growth, turbulence, and bottom reflection (Stasko and Pincock 1977). This research attempted to use ultrasonics to determine movements and daily activity pat- terns of the bonefish, Albula vulpes, in Bahamian waters. The only prior attempt at scientifically studying bonefish movements in the western Atlantic region was by Bruger, 1 who initiated a 'G. E. Bruger. Research Biologist, Florida Department of Natural Resources, Marine Research Laboratory, 100 Eighth Ave. SE., St. Petersburg, FL 33701. pers. commun. May 1980. 148 FISHERY BULLETIN: VOL. 81. NO. 1. 198:i comprehensive conventional tagging program on Florida bonefish. He met with no success, however, presumably due to the failure of the dart tags used. This current research thus repre- sents the first attempt to use ultrasonic telem- etry for this purpose. Methods and Study Sites This investigation was conducted in waters around a series of small islands or "cays" at the East End of Grand Bahama Island. The general environment here consists of mangrove, sand flats, creeks, lagoons, and offshore reefs. The north shores of these cays border the shallow waters of the Little Bahama Bank, while their south shores merge with the waters of the North- west Providence Channel. Areas in which bone- fish populations were frequently observed were selected as tagging and tracking sites; these areas are locally known as McLean's Town, Big Harbor, Little Harbor, Thrift Harbor, and Big Creek (Fig. 1). Each site represents a somewhat different habitat type: Portions of several are situated in protected lagoon areas between Abaco Islands and Grand Bahama Islands, and portions of others are located in shallow backwaters of East End, Grand Bahama, while two other locations are adjacent to open ocean and coral reefs. Figure 1.— Grand Bahama Island with inset of study areas in the vicinity of East End. Arabic numerals represent tracking sites at the following locations: 1) McLean's Town, 2) Big Harbor. 3) Little Harbor. 4) Thrift Harbor, 5) Big Creek. f, Little! <3 t Bahama ^J v J Bank Great Abaco Northwest Providence Channel 1 2 _i i Scale 1 50.000 [■'.] - tracking sites 149 The short-term (daily) movements of bonefish were monitored by an ultrasonic tracking sys- tem. Fish were captured by angling and gill nets. Bonefish were only minimally injured by the netting procedure, since the mesh size of 6.25 cm was chosen to restrain the fish without injury to the gills. Captured fish were removed by hand from the gill net and were held in a hand net for further treatment. Individuals, selected accord- ing to size (>2 kg) and physical condition, were equipped with ultrasonic transmitters. In track- ing studies prior to 1981, the transmitter was placed in the stomach with a glass plunger (Henderson et al. 1966; Yuen 1970). During 1980, this technique often resulted, 3 out of 5 times, in disgorgement of the transmitter. Therefore, surgical implantation of the transmitter in the body cavity was used during 1981. Here, the fish was restrained ventral side up. Several scales were removed posterior to the pelvic fins and lateral to the midline, and an incision of 3-4 cm was made with a surgical scalpel. The trans- mitter was then inserted, and the incision sutured. This procedure is similar to that used by Hart and Summerfelt (1975). To aid in recovery, the fish was slowly worked forward and back in the water by hand to aerate the gills. The majority of the fish appeared to survive the implant and recovered without noticeable effect, provided predators of the bonefish were not in the immediate vicinity at time of release. Several individuals held in a saltwater holding tank for periods of 24-96 h showed no noticeable ill effects. Conclusions drawn from the observed movements of fish immediately after release are of questionable value, since behavior and movements may be strongly influenced by the process of capture and handling. Thus, only tracks initiated 24 h or more after release were considered to reflect normal behavior. The transmitters were 58 mm long and 15 mm in diameter, weighed 3-4 g in water, and oper- ated at a frequency range of 74-77 kHz. They were manufactured by either Smith Root Inc. or Sonotronics, 2 and were pulsed at different inter- vals (1-2 pulses/s), so that individual fish could be distinguished when several transmitters were operating in the same general area. Power for the pulse intervals was supplied by mercury batteries with a useful life of about 7-14 mo. Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Range was as wide as 0.5 km at times, but much narrower when the water was turbulent. A Smith Root TA-50 and a Sonotronics digital (pulse/frequency display) receiver, with their respective hydrophones, were used to receive the signals. All tracking was conducted from a 4.5 m skiff equipped with two foot-controlled variable- speed electric motors. The hydrophones were mounted off the bow about 0.5 m below the sur- face, allowing the direction of a transmitting fish to be ascertained by pointing the bow in the di- rection of the strongest signal. Data recorded during tracking included location, water depth and temperature, tide state, time, and wind speed and direction. This information was gen- erally recorded at about 30-min intervals, but more frequently when a tracked fish was moving rapidly. Location was accurately recorded on Bahamian land survey maps by using chartered landmarks in conjunction with depth. To investigate long-term movements, a con- ventional tag and release program was initiated in February 1980. At the outset, Monel metal strap tags were crimped into the lower jaw. This method was replaced (January 1981) by the use of dart tags (FD 68B PVC) inserted adjacent to the dorsal fin, a procedure requiring less time and handlingof the fish. These tags were of much heavier construction than those used by Bruger (1974). Tagging was concentrated in areas frequently fished and/or areas in which schools of bonefish were consistently seen. Monthly collections of 20-30 bonefish were obtained from the study areas by nets and angling from June 1980 through December 1981, except September 1980. These data pro- vided information on size distribution of cap- tured individuals over the yearly cycle. Collec- tions were obtained each month from the same general areas (indicated in Fig. 1). Results Between August 1980 and November 1981, 13 bonefish were implanted with ultrasonic trans- mitters and released. Of these, only three fish were relocated more than 24 h from time of re- lease. Two of these fish, from McLean's Town Creek (50.5 and 53.5 cm FL (fork length)), were tracked for a period of 5 d each, with total track- ing times of 16 and 30 h, respectively. The fish from Big Creek Lake (61.0 cm FL) was followed over a 100-d period for a total tracking time of 32 h (Fig. 2). Water depths in these areas ranged 150 Little Thrift Harbor Cay 07 Appro*. ,„ water ,A/ i depth 15 at mean 2.0- high «e tide " ao Distonce (meters) 350 700 Figure 2.— Movements of a 61.0 cm FL bonefish in Big Creek Lake during 1981. The solid line denotes move- ment during falling tides, and the dashed line movement during a rising tide. The Arabic numerals denote the starting point and day of track, with '/ 2 -h intervals and direction of movement indicated by arrows. The time periods and dates of the individual tracks are: 1)1630-1700, 14 April; 2) 0855-0925, 1615-1650, 15 April; 3) 1350- 1435, 16 April; 4) 1557-1640, 18 April; 5) 1600-1700, 21 April; 6) 1200-1300, 22 April; 7) 1500-1530, 16 May; 8) 1200-1300, 20 May; 9) 0900-1000, 20 June; 10) 1200-2400, 23 June; 11) 1220-2420, 23 July. from 0.1 to 4 m. Tracking occasionally extended into areas of <0.1 m depth, at which times move- ments were visually monitored by observing the exposed dorsal and caudal fins. Visual observa- tions indicated that the fish generally remained near the substrate (<1 m). The range of water temperatures measured during a single track of any fish was no more than 8°C, with a low of 24°C and a high of 32°C. The general pattern of daily movements was a retreat to deeper water on an ebbing tide and a movement into shallow water on a rising tide. This pattern can be clearly seen in the track of the fish from Big Creek monitored for 100d(Fig. 2). A similar pattern was obtained by tracking the other two fish for 5 d each at McLean's Town. However, some variability was noted in the observed depth of fish movements as compared with the depth range available at the two locations. The fish at Big Creek was observed to move consistently into very shallow water (<1 m) with the rising tide. In contrast, the McLean's Town fish showed a variable response in depth- related movement. Also, "mudding" (a common term used to describe the turbidity resulting from fish feeding in bottom sediments) was ob- served only during low tide at Big Creek, but throughout the tidal cycle at McLean's Town. Nocturnal movements closely followed the same pattern. From January 1980 through December 1981, 214 bonefish were tagged with Monel metal or dart tags and released in the same channels, bays, and flats of Deep Water Cay as they were captured. None of these fish were recaptured more than 24 h from time of release. Only a single collection resulted in recapture of tagged fish; this was made 4 h after the fish had been tagged. Collection data provided a record of fish lengths and weights for each month (excluding September 1980) over a 19-mo period (June 1980- December 1981). The proportion of large fish (>55.5 cm FL) in these collections showed a pro- nounced regular seasonal change, with a strong 151 inverse relation to inshore water temperature (Fig. 3). That this change represents a movement of large fish rather than small fish from the flats during summer is strongly supported by numerous conversations with the guides, man- agers, and avid anglers of the Deep Water Cay Club. All of these persons made it clear that the catching of large (>55.5 cm FL) bonefish on the flats, although not common in winter months, is extremely rare in summer. The measurement of 55.5 cm FL, used in this paper to distinguish large from small fish, corresponds to the division between the fifth and sixth age-class of bonefish from the Florida Keys (Bruger 1974). Discussion Information gained from extended ultrasonic tracking of three individuals in two different areas suggests that bonefish display a regular pattern in daily movements in response to tidal changes. These movement patterns, although monitored on individuals, are probably repre- sentative of school movement because trans- mitter-implanted fish generally returned and remained with schools of bonefish (3-20 individ- uals) within 24 h of release. The observed differ- ences in daily movements of bonefish in the two different areas may indicate the effects of differ- ences among the two locations in such factors as bottom topography, food resource distribution, and predation. However, this point is in need of further research. Information derived from ultrasonic tracking, conventional tagging, and repetitive collecting effort in specific areas indicates that movements of bonefish on a long-term basis are highly variable and without apparent pattern. Ultra- sonic tracking has indicated that individual fish usually remain in a given localized area for less than a week. The two fish tracked for 5 d apparently left the McLean's Town area after that time, since extensive searching on the sixth day, up to 2 km from the area last observed, resulted in no relocation of the fish or the trans- mitter. Subsequent searches of the same area weeks and months later also were unsuccessful. Six other fish equipped with transmitters and released in apparently good condition were never relocated 24 h after release. Another strong indication of the transient nature of bonefish movements is the lack of return of conventionally tagged fish, although a concentrated tag and recapture effort was made a. E sample sue ,Tau } A s [no individuals] 34 25 O 18 N 19 'J 20 20 20 25 22 J 25 A S O N D 25 24 20 22 21 O - temperature • - /o individuals Summer 1980 Summer 1981 Months FIGURE 3. — Proportions of large individuals found in monthly collections of bonefish in the waters around Deep Water Cay, Grand Bahama. Each data point represents the percent of individuals col- lected each month exceeding 55.5 cm FL. No collection was obtained in September 1980. Open circles = temperature, solid circles = percent individuals. 152 in relatively restricted areas over a period of 18 mo. On one occasion only, fish (three individuals released 4 h prior to recapture) that had been previously tagged and released were recovered again. Failure to relocate fish conventionally tagged or fitted with transmitters could be the result of factors other than fish leaving the gen- eral area, such as mortality due to predation or shock of capture and handling, or tag fail- ures. However, none of the evidence gained in this study suggests these factors were respon- sible. The observed reduction in the proportion of large bonefish present on the flats during warm- water periods may correspond to a general offshore movement in preparation for spawning. Summer temperatures in the shallow areas of Thrift Harbor Creek have exceeded 34 °C (Fig. 3). Although thermal requirements for Albula vulpes have not yet been experimentally deter- mined, abnormally high temperatures are known to be deleterious to gamete formation among vertebrates (Guyton 1976; Langman 1981). Among fishes, it has been generally estab- lished that thermal requirements are even more restrictive for the reproductive process than for either growth or survival (Brett 1956). The hypothesis that large bonefish move offshore during summer is supported by the lore of the local Bahamian fishermen. They believe that larger individuals undergo a regular movement into deeper (15-25 m) waters at this time. During fall (October-November), these fish return inshore and aggregate in large numbers to spawn in shallow creeks. Erdman (I960 3 ) reported a similar observation by commercial fishermen in Puerto Rico. At the time of this in- shore movement, the fish are said to be lighter in color, with a highly silvery appearance. Personal examination by the senior author of fish collected by anglers from such aggregations revealed that nearly all individuals were sexually ripe. Addi- tional evidence of seasonal offshore movements of bonefish comes from scuba divers in the Free- port area (pers. commun.), who have reported observing schools of thousands of bonefish sus- pended above the reefs. Bohlke and Chaplin (1968) cited a similar observation occurring off the Tongue of the Ocean, Green Cay, Baha- mas. 3 Erdman, D. S. 1960. Notes on the biology of the bonefish and its sports fishery in Puerto Rico. Paper prepared for the 5th Int. Game Fish Conf. Miami Beach, Fla., 11 p. Summary Bonefish appear to remain in a specific loca- tion (e.g., creek, small bay, channel, etc.) for a period usually not exceeding several days, and then move on to other locations. While at a given location, there is a distinct pattern to daily movements in response to tidal fluctuations, but long-term movements appear to be highly variable, with no definable pattern seen. In summer, larger individuals are rarely found on the flats. Their reappearance in the fall concurs with a rapid drop in water temperature at that time of the year. Acknowledgments We gratefully acknowledge the generous fi- nancial support of Perkin Sams, Mr. and Mrs. Lewis Murdock, D. H. Ruttenburg, and the Slocum-Lunz Foundation. The authors would also like to thank the Murdocks, John Adams, Richard Shaul, Patrick Pitts, Marie Colton, and guides of Deep Water Cay Club for their assis- tance in the field. We thank the Bahamian Government for permission to conduct research in their territorial waters. A very special thanks to John and Ann Dickinson whose enthusiasm and generosity made the field work possible. Literature Cited Bohlke, J. E., and C. C. G. Chaplin. 1968. Fishes of the Bahamas and adjacent tropical waters. Livingston Publ. Co., Wynnewood, Pa., 771 P- Brett, J. R. 1956. Some principles in the thermal requirements of fishes. Q. Rev. Biol. 31:75-87. Bruger, G. E. 1974. Age, growth, food habits, and reproduction of bone- fish, Albula vulpes, in south Florida waters. Fla. Mar. Res. Pub!. 3, 20 p. COLLETTE, B. B., AND F. H. TALBOT. 1972. Activity patterns of coral reef fishes with emphasis on nocturnal-diurnal changeover. In B. B. Collette and S. A. Earle (editors). Results of the Tektite program: Ecology of coral reef fishes, p. 98-124. Nat. Hist. Mus. Los Ang. Cty. Sci. Bull. 14. COUTANT, C. C. 1975. Temperature selection by fish - a factor in power plant impact assessments. In Environment effects of cooling systems at nuclear power plants, p. 575-597. Int. At. Energy Agency Symp., Vienna. DODSON, J. J., AND W. C. LEGGETT. 1973. Behavior of adult American shad (Alosa sapidis- sima) homing to the Connecticut River from Long Island Sound. J. Fish. Res. Board Can. 30:1847-1860. 153 GUYTON, A. C. 1976. Textbook of medical physiology. 5th ed. W. B. Saunders Co., Phila., 1194 p. Hart, L. G., and R. C. Summerfelt. 1975. Surgical procedures for implanting ultrasonic transmitters into flathead catfish (Pylodictis olivaris). Trans. Am. Fish. Soc. 104:56-59. Haynes, J. M., R. H. Gray, and J. C. Montgomery. 1978. Seasonal movements of white sturgeon (Acipenser transmontanus) in the mid-Columbia River. Trans. Am. Fish. Soc. 107:275-280. Henderson, H. F., A. D. Hasler, and G. G. Chipman. 1966. An ultrasonic transmitter for use in studies of movements of fishes. Trans. Am. Fish. Soc. 95:350-356. Kelso, J. R. M. 1976. Movement of yellow perch (Perca flavescens) and white sucker (Catostomus commersoni) in a nearshore Great Lakes habitat subject to a thermal discharge. J. Fish. Res. Board Can. 33:42-53. Langford, T. E.. A. G. P. Milner, D. J. Foster, and J. M. Fleming. 1979. The movements and distribution of some common bream (Abramis brama) in the vicinity of power station intakes and outfalls in British rivers as observed by ultrasonic tracking. C.E.R.L. L44. Note RD/L/N, p. 785-788. Leatherhead, Surrey, U.K. Langman, J. 1981. Medical embryology. 4th ed. Waverly Press, Inc., Bait., 384 p. McFarland, W. N., J. C. Ogden, and J. N. Lythgoe. 1979. The influence of light on the twilight migrations of grunts. Environ. Biol. Fish. 4(l):9-22. Standora, E. A., T. C. Sciarrotta, D. W. Ferrel, H. C. Carter, and D. R. Nelson. 1972. Development of a multichannel, ultrasonic telem- etry system for the study of shark behavior at sea. Caiif. State Univ. Long Beach Found. Tech. Rep. 5, 69 p. Stasko, A. B., and D. G. Pincock. 1977. Review of underwater biotelemetry, with empha- sis on ultrasonic techniques. J. Fish. Res. Board Can. 34:1261-1285. Stasko, A. B., R. M. Horrall, A. D. Hasler, and D. Stasko. 1973. Coastal movements of mature Fraser River pink salmon (Oncorhynchus yorbushcha) as revealed by ultra- sonic tracking. J. Fish. Res. Board Can. 30:1309-1316. Yuen, H. S. H. 1970. Behavior of skipjack tuna, Katsuwonus pelamis, as determined by tracking with ultrasonic devices. J. Fish. Res. Board Can. 27:2071-2079. Douglas E. Colton William S. Alevizon Department of Biological Sciences Florida Institute of Technology Melbourne, FL 32901 ANALYSES OF FEEDING IN TWO MARINE COPEPODS FROM SANTA MONICA BAY, CALIFORNIA Understanding the feeding strategies of herbiv- orous, planktonic copepods is an important step in determining how primary production is parti- tioned in coastal marine food webs. The condi- tions under which selective feeding occurs among these animals vary, and are defined both by the species and the environment (Poulet 1974; Poulet and Marsot 1980; Donaghay 1980). Although it is desirable to study feeding be- havior in natural zooplankton assemblages, this is often difficult. Identification of phytoplankton in the gut by standard dissection and microscopic techniques is labor intensive, and usually quali- tative. Furthermore, it is impossible to identify many of the soft-bodied organisms which might have been consumed. For this reason, much of the work on food selection in copepods has been restricted to the laboratory, where cultivated foods (Frost 1972) or natural particles (Poulet 1978) have been offered to the animals. While such studies have provided valuable informa- tion, they have been limited by the variety of foods which can be offered and by other technical problems (Mullin 1963; Harbison and McAlister 1980). Studies employing gut contents analysis of animals collected in the field using gut fullness (Hayward 1980; Huntley 1980) or chlorophyll a fluorescence as an estimate of total phytoplank- ton biomass in the gut (Mackas and Bohrer 1976; Boyd et al. 1980) have answered questions about when and where certain zooplankton feed, but usually provide only indirect data on the kinds of phytoplankton actually ingested. Dagg and Grill (1980) showed that the rate of particle ingestion is often not solely a function of concentration and suggested that food quality may be important in explaining the variability observed in the rela- tion between feeding rate and particle concen- tration. To understand the processes involved in food selection it is necessary to determine directly the types of materials in the guts of the copepods be- ing studied. Such an analysis must be capable of detecting soft-bodied phytoplankton as well as diatoms and armored dinoflagellates, and of pro- viding some indication of the relative importance of different taxa in the diet at a given time. We have been especially interested in the importance of the green algae to zooplankton feeding in coastal waters. Information in this area is rela- 154 FISHERY BULLETIN: VOL. 81, NO. 1. 1983. tively scarce despite the periodic importance of green algae in the coastal flora (C. Lorenzen un- publ. data). In September 1980, the cyst (phycoma) stage of Halosphaera sp. (Prasinophyceae) was observed in Santa Monica Bay, Calif., providing an oppor- tunity to study its importance in the feeding of two calanoid copepods, Acartia tonsa and Calan- us pacificus. Since chlorophyll b is present only in the green algae (Chlorophyceae, Prasinophy- ceae, Euglenophyceae) and chlorophyll c is pres- ent in the diatoms, dinoflagellates, chrysomo- nads, Haptophyceae, and Cryptophyceae (Meeks 1974; Parsons et al. 1977), we sought to compare water column concentrations of chlorophyll pig- ments with those in the guts of animals collected in various parts of the bay. Methods Samples were collected at two of three stations in Santa Monica Bay (Fig. 1) on 12 and 26 Sep- tember 1980. On 12 September, stations 7B and N6 were sampled. On 26 September, stations 7B and N4 were occupied. All samples were taken between 0700 and 1200 h. Depth integrated water samples were collected by lowering a submersible pump through the water column (to the same depth as zooplankton were collected; see below) at a constant rate and by pumping into a 122 1 plastic container. The contents were mixed thoroughly, and 1 1 samples were withdrawn and fixed in 3% buffered For- malin 1 for phytoplankton counting and identifi- cation, using the method of Palmer and Maloney (1954). Five hundred ml water samples were frozen for pigment analysis. In the laboratory, these were passed through 0.45 ^m filters (Nu- cleopore) at low vacuum (<100 mm Hg), and pro- 1 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 1 118°40'W 1 i i 118 30'W MALIBU "■*■>■ ' '-■&; \L SANTA MONICA \ 34°00'N- V^^PLAYA DEL REY N4 \ - f O 7B O \ EL SEGUNDO N6 V r v \f *0 L. REDONDO IV. BEACH - Js t 33°50'N - 1 2 4 6 8 10 I s " **. j Kilometers 1 1 j li: Figure 1.— Station locations. 155 cessed for chlorophylls a, b, and c by the trichro- matic method (Strickland and Parsons 1972), using the equations of Parsons and Strickland (1963). Unfortunately, these equations do not give accurate estimates of chlorophyll b or c, and despite numerous attempts to improve their ac- curacy, no set of equations has been completely satisfactory (Jeffrey 1968, 1981; Jeffrey and Humphrey 1975). However, if the errors for gut and water samples are assumed to be the same, then comparisons can be made between pigment concentrations in guts and water. To test this, Acariia tonsa were starved for 24 h and a sample of animals was examined microscopically to en- sure that the guts were empty. Pure cultures of the diatom Thalassiosira fluviatilis were added to half of the flasks containing the copepods. The animals in the remaining flasks were not fed. After 4 h, the animals in all of the flasks were processed for chlorophyll pigments as de- scribed below. We chose to feed a diatom in this experiment to find out if chlorophyll a or c might, during digestion, be converted into a product absorbing at the wavelengths used in measuring chlorophyll b (not found in diatoms). Table 1 sug- gests that this did not occur. Low levels of chloro- phyll b were detected both in the culture and the fed animals but not in the starved animals, indi- cating that there was some contaminant in the culture or a small error in the equation at high chlorophylls a and c concentrations. Pigment ratios in the culture and the guts of animals fed from the culture were fairly stable. Copepods for gut contents analysis were col- lected from 7 to 10 vertical tows of a CalCOFI vertical tow net (335 /jm mesh). Tows were made from 70 m at station 7B and from near-bottom at stations N4 and N6. The tows took 2-3 min each to complete and were made in rapid succession. The ship was kept on station during the entire sampling period. On the cruise of 12 September, Table 1.— Ratios of chlorophyll pig- ments in a culture of Thalassiosira flu- viatilis and in Acartia guts when the animals were starved or fed the phyto- plankton culture. Chlorophyll pigments Culture Acartia guts Starved Fed a/b' a/c b/c 5691 1 21 002 ND 60.00 0.30 1 .09 0.02 Chlorophyll a levels per milliliter of culture = 0.24 /yg; per animal = 12 ng. ND = a ratio that could not be computed, since chlorophyll b was not detected. adult A. tonsa and C. pacificus were immedi- ately separated from the rest of the catch. Half of each sample was washed in filtered (0.5 ^m) sea- water and frozen, and the other half was main- tained in aerated, filtered seawater for 24 h allowing them to clear their guts of food prior to freezing. This empty gut group was used to de- tect absorbance at wavelengths used in chloro- phyll analysis that was not due to chlorophyll. On 26 September, collections were made in a similar manner, except that half of the entire catch was washed with filtered seawater and frozen and the other half was maintained alive for 24 h in fil- tered seawater prior to freezing. Specimens for analysis were separated from the rest of the catch in the laboratory. Chlorophyll a analysis of gut contents was con- ducted by macerating 50-200 animals of each species of full and empty gut groups in 90% ace- tone and by reading absorbances in triplicate at wavelengths of 750, 665, 645, and 630 nm on a Beckman model 34 spectrophotometer. Chloro- phylls a, b, and c concentrations were computed by the trichromatic equations of Parsons and Strickland (1963). Chlorophyll degradation prod- ucts (pheopigments) in the gut contents were computed as described by Strickland and Par- sons (1972) and were considered as part of the total chlorophyll because the processing of food in the gut rapidly degrades chlorophylls. Since our interest was in how much plant material was present and not in the rate of food processing, we include both chlorophyll and its degradation products as a single indication of plant biomass in the gut. Gut fullness was estimated in A. tonsa collected on 26 September. The animals were "cleared" by immersion in 85% lactic acid for 30 min and then examined under 25X magnification (Hay ward 1980). Gut fullness was estimated independently by each author and the average of the two esti- mates was recorded. Attempts were made to esti- mate gut fullness in C. pacificus, but there was disagreement between estimates because the lactic acid did not clear the animals well. Results Phytoplankton density and community struc- ture were similar at station 7B on both sampling dates. Cell density averaged 8-10 X 10 3 cells/1. The majority of phytoplankton species found were dinoflagellates (35% of the community), dominated by Gymnodinium spp., and diatoms 156 (35% of the community), dominated by Skeleto- nema costatum. Cysts of Halosphaera sp. com- posed about 20% of the community on both dates; the motile form was not detected. Coccolithophor- ids composed about 10% of the cells counted. Mean chlorophyll a concentrations at the station were 0.56 and 1.16 iig/\ on 12 and 26 September, respectively. At stations N4 and N6 there were about 10 5 phytoplankton/1 on each occasion. About 60% of the phytoplankton in the samples were diatoms of the genus Chaetoceros. Other diatom species composed about 25% of the community, and dino- flagellates made up 10%. Halosphaera cysts, small unidentifiable spherical cells (some of which probably contained chlorophyll b) and coccolithophorids made up about 5% of the com- munity. Mean chlorophyll a concentrations were 1.54 and 1.92 yug/1 on 12 and 26 September, re- spectively. Figure 2a, b summarizes gut fullness estimates for 50 A. tonsa from stations N4 and 7B. At sta- tion N4 (Fig. 2a), 70% of the animals exhibited >40% gut fullness; mean gut fullness was 55.5%. At station 7B (Fig. 2b), about 20% of the animals exhibited >40% gut fullness; the mean was 31% fullness. Gut chlorophyll a concentrations (corrected for the absorbance of empty gut animals) are shown in Table 2. Comparative water column data are also provided. The concentration of chlorophyll a in the gut contents of A. tonsa increased with seaward distance. Animals col- lected at station 7B had, on average, 40 times more chlorophyll in their guts than the same spe- cies at nearshore locations. The chlorophyll b and c content of the water column diminished slightly with distance from shore (Fig. 3a). In the guts of Acartia these pig- ments increased sharply from nearshore to off- shore stations (Fig. 3b). a) 100r 80 H e TO 14-1 o o\° b) 0) 60 40 20 100r 80 6 60 to 14-4 o 40 20 A. tonsa N-4 Mean Fullness = 55.5% — T A. tonsa 7-B Mean Fullness = 31.0% 20 40 60 80 % Fullness 100 Figure 2.— Percent gut fullness of Acartia tonsa at stations a) N4 and b) 7B on 26 September 1980. AT = 50. Table 2.— Chlorophyll a in two copepod species, Acartia tonsa and Calanus pacificus, at stations of varying distances from shore, with comparative water column 1 values. Distance from Chlorophyll a Species Gut (±SD) Background Water (±SD) Station shore (km) (sample size) (ng per animal) (ng per animal) (/ug per liter) 2 N4 0.6 A tonsa (200) 12 (0.01) 0.16 1.92 (0.36) 3 N6 0.9 A tonsa (160) 0.18 (001) 1.54 3 7B 18.0 A. tonsa (200) 6.10 (0.003) 1.16 (0.15) 2 7B 180 *C. pacificus (55) 13 24 (0 003) 0.001 0.56 'Water column chlorophyll a values are mean ± standard deviation (in parentheses) from water column composite samples. On 12 September only one sample was analyzed; on 26 September. 5 sub- samples of the water column composite were analyzed. a Data from cruise on 26 September 1980. 3 Data from cruise on 12 September 1980 "Numbers of C. pacificus at the nearshore stations were too low (<5 animals/tow) for the analysis to be conducted. 157 al 100 10 o en 1 0.1 0.1 O chl c, water • chl b, water V j_ bl CO E c < O 1 10 Km from Shore 100 0.1 100 Figure 3.— a) Mean chlorophyll b and c concentrations in water samples plotted relative to distance from shore, b) As in a) but for pigments in the gut contents of Acartia tonsa and Calanus pacificus. If the ratio of the chlorophyll b or c to its sum, T (= b + c), is the same in the gut of a copepod as it is in the water, then it might be reasoned that feeding on phytoplankton was not selective. Vari- ations from unity would be interpreted as an in- dication of food selectivity. We define relative selectivity indices for chlorophyll b (RSIb) and chlorophyll c (RSI C ) as: RSIt RSIc = (b/T). (b/T) w (c/T) e (c/T) w (1) (2) where g and w represent the ratios in the gut and water, respectively. RSI values, presented in Table 3, indicate selec- tivity for chlorophyll b-bearing organisms by Acartia at stations N4 and N6. At station 7B, Acartia evidenced a weak selection of chlorophyll c-bearing organisms and Calanus pacificus selected for chlorophyll b-bearing organisms. Discussion There were clear differences in the gut con- tents of Acartia from near- and offshore loca- tions. Gut fullness was higher in copepods from nearshore than from those offshore (Fig. 2), but the amount of chlorophyll a in the guts of animals collected nearshore was substantially lower than in the guts of animals from offshore locations (Table 2). Apparently materials other than phyto- plankton composed a relatively large portion of Table 3.— Relative selectivity indi- ces for chlorophyll b (RSI b ) and chlorophyll c (RSI C ) by Acartia tonsa and Calanus pacificus. Station Species RSI RSIc 'N4 A tonsa 2 04 0.62 2 N6 A tonsa 2.48 061 2 7B A lonsa 0-73 1.08 '7B C pacificus 1.20 0.93 1 Data from cruise on 26 September 1980 2 Data from cruise on 12 September 1980. the diet of the nearshore animals. Evidence from laboratory studies (Poulet 1973; Heinle and Flemer 1975; Richman et al. 1977; Roman 1977) suggests the possibility of a detrital or animal component in the diet of Acartia when these foods are available. The RSI indicates that C. pacificus was feeding selectively on phytoplankton containing chloro- phyll b at station 7B. The only green alga detected by microscopic analysis of water samples at this location was Halosphaera sp. Although it is pos- sible, even likely, that other green algae were present, the typical chlorophytes and euglenoids were not observed, and, unlike the nearshore sta- tions, nanoplanktonic green algae appeared to be absent. We assume, therefore, that Halosphaera was at least the dominant source of chlorophyll b in the water, and constituted the greater portion of the chlorophyll b signal in the C. pacificus gut. Since we cannot test this assumption, what fol- lows must be considered somewhat speculative. However, we suggest that under the conditions observed in Santa Monica Bay at the time, selec- tive feeding by Calanus on Halosphaera cysts 158 would be energetically advantageous to the ani- mal. Although many calanoid copepods, including C. pacificus, are recognized omnivores (Landry 1980), there have been numerous reports that C. pacificus will remove certain types of particles from the water, apparently in preference to others (Gifford et al. 1981 ). Therefore, the indica- tion of selective feeding is not surprising. It is difficult, however, to explain the mechanisms driving this selection. It has been held that food selection is often passive in nature. For instance, the intersetal distance may facilitate the capture of certain-sized particles over others (Frost 1972; Wilson 1973), and accidental encounter may re- sult in the most abundant particles being most commonly ingested (Poulet 1974). However, ex- planations based on passive feeding modes have been inadequate in several situations (Huntley 1980), and the work of Poulet and Marsot(1980) and Friedman (1980) suggests that morphologi- cal adaptations exist among the copepods which would permit a high degree of food selection based on the active detection of mechanical and chemical stimuli. Most enlightening have been the cinematic evidence and physical arguments of Koehl and Strickler (1981) that copepods used the feeding appendages as paddles to move water to the sec- ond maxillae, rather than as strainers to filter it. This being the case, the selection of large parti- cles, observed by Frost (1972), Gifford et al. (1981), and many others, would seem due to an active preference for these particles under cer- tain conditions rather than the passive collection of material in the appendages. This is not to imply that copepods never ingest nanoplankton or feed passively, as we know they do. Rather, we suggest that active food selection may be quite common, even typical, in C. pacificus. To understand the adaptive significance of selective feeding on large particles, it is neces- sary to consider the circumstances under which this sort of feeding might be most useful. Landry (1981) suggested that when the abundance of diatoms decreases in the water, adult C. pacifi- cus begins to prey on copepod nauplii. An expla- nation of this behavior would be that when small particles (diatoms) become scarce and nauplii relatively abundant, it is energetically efficient to capture the larger biomass units (nauplii). The low phytoplankton density observed dur- ing the present study is characteristic of Santa Monica Bay in the fall (Kleppel and Manzanilla 1981). We can extend Landry's (1981) argument somewhat by suggesting that the waning of dia- tom-sized particles might cause a shift in feeding to large biomass units represented by the cysts of Halosphaera. To get a feeling for the advantage of feeding on these cysts in relation to diatoms, we can compare rough estimates of the carbon in a diatom with that of the Halosphaera cyst and its rosettes (the individual units of the cyst which will mature into 200-550 motile cells), using equa- tions based on cell volume (Strathmann 1967). We stress that such estimates have wide confi- dence intervals and should be considered on the basis of scale rather than accuracy. The diameter of a mature Halosphaera cyst ranges from 200 to 800 ^m, depending on species (Parke and den Hartog-Adams 1965; Boalch and Mommaerts 1969). The cysts we observed were somewhat smaller, 100-150 ^m, indicating that they were not mature. This may explain why no motile cells were detected. Using the smaller measured diameter ( 100 ^m), we calculate a car- bon content of 0.031 /ig/cyst. Considering only the rosettes (diameter based on literature values = 15-20 /xm for the smallest units; Parke and den Hartog-Adams 1965) and assuming them to be round discs, 2 //m thick, we calculate the car- bon content of one rosette to be 56-92 pg. If there are 200 rosettes/cyst, then the carbon content of the rosettes in one cyst is 0.011-0.018 /ig. Using the volume of Skeletonema costatum (the dominant diatom at station 7B) equal to 1,390 yum 3 (Parsons et al. 1961), the cellular carbon con- tent estimated by the Strathmann equation is 91 pg. Since S. costatum typically forms chains 4-10 cells in length, the carbon content of a chain would be 3.7 X 10" 4 to 9.1 X 10 -4 M g- This is nearly two orders of magnitude lower than the carbon content of one Halosphaera cyst or its rosettes. Although we stress that these estimates are crude and we recognize that numerous factors will affect the actual carbon content of a cell, the magnitude of the difference between the esti- mated carbon in Halosphaera and Skeletonema nonetheless seems significant. It would appear that selective feeding on Halosphaera would have a distinct advantage for C. pacificusby pro- viding a large energy ration with each capture. This would seem of obvious value in ecosystems characterized by patchy food supplies. Acknowledgments We thank T. Hayward and D. Kiefer for read- 159 ing and criticizing the manuscript and F. Reid for discussions about the taxonomy of Halo- sphaera. H. Stubbs and L. Horita provided tech- nical assistance, and J. Nagano, K. Mikawa, and S. Cheng of Los Angeles provided ship time and laboratory space. Literature Cited BOALCH, G. T., AND J. P. MOMMAERTS. 1969. A new punctate series of Halosphaera. J. Mar. Biol. Assoc. U.K. 49:129-139. Boyd, C. M., S. L. Smith, and T. J. Cowles. 1980. Grazing patterns of copepods in the upwelling sys- tem off Peru. Limnol. Oceanogr. 25:583-596. Dagg, M. J., and D. W. Grill. 1980. Natural feeding rates of Centropages typicus fe- males in the New York Bight. Limnol. Oceanogr. 25: 597-609. DONAGHAY, P. L. 1980. Grazing interactions in the marine environment. In W. C. Kerfoot (editor), Evolution and ecology of zoo- plankton communities, p. 234-240. Univ. Press New Engl., Hanover. N.H. Friedman, M. M. 1980. Grazing interactions in the marine environment. /// W. C. Kerfoot (editor), Evolution and ecology of zoo- plankton communities, p. 185-197. Univ. Press New Engl., Hanover, N.H. Frost, B. W. 1972. Effects of size and concentration of food particles on the feeding behavior of the marine planktonic cope- pod Calanus pacificus. Limnol. Oceanogr. 17:805-815. GlFFORD, D. J., R. N. BOHRER, AND C. M. BOYD. 1981. Spines on diatoms: Do copepods care? Limnol. Oceanogr. 26:1057-1061. Harbison, G. R., and V. L. McAlister. 1980. Fact and artifact in copepod feeding experiments. Limnol. Oceanogr. 25:971-981. Hayward, T. L. 1980. Spatial and temporal feeding patterns of copepods from the North Pacific central gyre. Mar. Biol. (Berl.) 58:295-309. Heinle, D. R., and D. A. Flemer. 1975. Carbon requirements of a population of the es- tuarine copepod Eurytemora affinis. Mar. Biol. (Berl.) 31:235-247. Huntley, M. 1980. Yellow water off La Jolla, California in July 1980: Zooplankton grazing. (Abstr.) Annu. Conf., Calif. Coop. Oceanic Fish. Invest., Oct. 20-23, 1980, Idyllwild, Calif. Jeffrey, S. W. 1968. Quantitative thin layer chromatography of chloro- phylls and carotenoids from marine algae. Biochem. Biophys. Acta 162:271-285. 1981. An improved thin-layer chromatographic tech- nique for marine phytoplankton pigments. Limnol. Oceanogr. 26:191-197. Jeffrey, S. W., and G. F. Humphrey. 1975. New spectrophotometric equations for determining chlorophylls a. b, c u c 2 in higher plants, algae, and natural phytoplankton. Biochem. Physiol. Pflanz. 167: 191-194. Kleppel. G. S., and E. Manzanilla. 1981. Phytoplankton abundance and distribution in Santa Monica Bay. In W. Bascom (editor), Biennial Report, 1979-1980, p. 265-273. Southern California Coastal Water Research Project, Long Beach, Calif. Koehl. M. A. R., and J. R. Strickler. 1981. Copepod feeding currents: Food capture at low Reynolds Number. Limnol. Oceanogr. 26:1062-1073. Landry, M. R. 1980. Detection of prey by Calanus pacificus: Implica- tions of the first antennae. Limnol. Oceanogr. 25:545- 549. 1981. Switching between herbivory and carnivory by the planktonic marine copepod Calanus pacificus. Mar. Biol. (Berl.) 65:77-82. Mackas, D., and R. Bohrer. 1976. Fluorescence analysis of zooplankton gut contents and an investigation of diel feeding patterns. J. Exp. Mar. Biol. Ecol. 25:77-85. Meeks, J. C. 1974. Chlorophylls. In W. D. P. Stewart (editor). Algal physiology and biochemistry, p. 161-175. Univ. Calif., Berkeley. Mullin, M. M. 1963. Some factors affecting the feeding of marine cope- pods of the genus Calanus. Limnol. Oceanogr. 8:239- 250. Palmer, C. M., and L. C. Maloney. 1954. A new counting slide for nannoplankton. Limnol. Oceanogr. Spec. Publ. 21. Parke, M., and I. den Hartog-Adams. 1965. Three species of Halosphaera. J. Mar. Biol. Assoc. U.K. 45:537-557. Parsons. T. R., K. Stephens, and J. D. H. Strickland. 1961. On the chemical composition of eleven species of marine phytoplankters. J. Fish. Res. Board Can. 18: 1001-1016. Parsons, T. R., and J. D. H. Strickland. 1963. Discussion of spectrophotometric determination of marine-plant pigments, with revised equations for ascertaining chlorophylls and carotenoids. J. Mar. Res. 21:155-163. Parsons, T. R., M. Takahashi, and B. Hargrave. 1977. Biological oceanographic processes. 2d ed. Per- gamon Press, N.Y., 332 p. POULET, S. A. 1973. Grazing of Pseudocalanus minubus on naturally occurring particulate matter. Limnol. Oceanogr. 18: 564-573. 1974. Seasonal grazing of Pseudocalanus minutus on particles. Mar. Biol. (Berl.) 25:109-123. 1978. Comparison between five coexisting species of ma- rine copepods feeding on naturally occurring particulate matter. Limnol. Oceanogr. 23:1126-1143. Poulet, S. A., and P. Marsot. 1980. Chemosensory feeding and food-gathering by om- nivorous marine copepods. hi W. C. Kerfoot (editor). Evolution and ecology of zooplankton communities, p. 198-218. Univ. Press New Engl., Hanover, N.H. Richman, S., D. R. Heinle, and R. Huff. 1977. Grazing by adult estuarine calanoid copepods of the Chesapeake Bay. Mar. Biol. (Berl.) 42:69-84. Roman, M. R. 1977. Feeding of the copepod Acatiia tonsa on the dia- tom Nitzsckia closterium and brown algae (Fucus vesi- culosa) detritus. Mar. Biol. (Berl.) 42:149-155. 160 Strathmann, R. R. 1967. Estimating the organic carbon content of phyto- plankton from cell volume or plasma volume. Limnol. Oceanogr. 12:411-418. Strickland. J. D. H.. and T. R. Parsons. 1972. A practical handbook of seawater analysis. 2d ed. Fish. Res. Board Can. Bull. 167, 310 p. Wilson, D. S. 1973. Food size selection among copepods. Ecology 54: 909-914. G. S. Kleppel University of Southern California Institute for Marine and Coastal Studies University Park- Los Angeles, CA 90007 Present address: Fish Harbor Marine Laboratory University of Southern California 820 South Seaside Are. Terminal Island, CA 90731 E. Manzanilla Southern California Coastal Water Research Project 646 W. Pacific Coast Highway Long Beach, CA 90806 Present address: University of Southern California Institute for Marine and Coastal Studies University Park Los Angeles, CA 90007 DISTRIBUTION, SIZE RELATIONSHIPS, AND FOOD HABITS OF JUVENILE KING-OF-THE-SALMON, TRACHIPTERUS ALTIVELIS, CAUGHT OFF THE OREGON COAST The king-of-the-salmon is a strikingly colored rib- bonfish of the family Trachipteridae that occurs in the oceanic and coastal waters of the eastern Pacific Ocean, from Chile to Alaska. Captures have been recorded from the coastal regions and offshore halfway to the Hawaiian Islands. Specimens have also been taken in coastal waters and estuaries along the United States and Cana- dian shores on rare occasions (Hart 1943; Walker 1953). Their lower depth limit is not known, but individuals have been taken from the surface down to at least 650 m (Fitch 1964). Spawning apparently occurs in the open ocean throughout the year, but is probably concen- trated in the spring. Plankton surveys off Cali- fornia have recorded the largest catches of larvae during the months of June and July (Fitch 1964). Bongo net and neuston net collections from north- ern California, Oregon, and Washington fre- quently contained eggs in April and May 1980, but larvae were rarely taken (Kendall and Clark 1 ). August 1980 samples contained relative- ly few eggs (Kendall and Clark 2 ). Egg densities during the spring sampling reached 25 eggs/10 m 2 , and the eggs were found from 5 to 320 km offshore (Kendall 3 ). Throughout the early life stages, allometric growth reduces the proportionate size of the fins and alters the body form by increasing the relative size of the posterior portion of the fish (Sette 1923; Hubbs 1926). Fitch (1964) examined the otoliths of five individuals to deter- mine their ages. His fish ranged from a 400 mm juvenile with an estimated age of 1 yr to a 1.5 m adult with an age of 7 yr. The stomach contents of several adults show that these fish eat whole micronectonic organ- isms (e.g., small squid, epi- and mesopelagic fishes) as well as macrozooplankton such as euphausiids (Fitch 1964). Roedel (1938) pre- sented a qualitative list of the gut contents of five juveniles (about 100-200 mm long) taken from the stomach of a longnose lancetfish, Alepisaurus ferox, caught off Santa Monica, Calif. Copepods were found in three of the stomachs, while polychaetes and fish larvae were each found in one stomach. During 1980 and 1981, 44 juvenile king-of-the- salmon were collected with a purse seine during a study of the ecology and migration of juvenile salmonids off the Oregon coast. This paper presents an analysis of the spatial distribution, size relationships, and the feeding habits of these unusual fish. 'Kendall, A. W.. Jr., and J. Clark. 1982. Ichthyoplankton off Washington, Oregon, and Northern California April-May 1980. Northwest and Alaska Fish. Cent. Process. Rep. 82-1 1. 44 p. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA. 2725 Montlake Blvd. East. Seattle, W A 98112. ^Kendall, A. W.. .Jr., and J. Clark. 1982. Ichthyoplankton off Washington, Oregon, and Northern California August 1980. Northwest and Alaska Fish. Cent. Process. Rep. 82-12, 43 p. Northwest and Alaska Fisheries Center. National Marine Fisheries Service, NOAA. 2725 Montlake Blvd. East. Seattle. W A 98112. 3 Arthur W. Kendall. Northwest and Alaska Fisheries Center. National Marine Fisheries Service. NOAA, 2725 Montlake Blvd. East, Seattle, WA 98112. pers. commun. Jan- uarv 1983. FISHERY BULLETIN: VOL. 81, NO. 1. 1983. 161 Materials and Methods In June 1980 and May, June, July, and August 1981, 10-d sampling cruises were conducted. In 1980, 44 collections were made with a purse seine. Six sets were made during the night at two stations in 2,100-2,200 m of water about 100 km offshore from the mouth of the Columbia River. Thirty-five sets were made during the day along three transects located north and south of the mouth of the Columbia River and off the mouth of the Yaquina River. These transects ex- tended from the 40 m isobath to 40 km offshore. An additional three sets were made along the shore 40-50 km south of the Columbia River. During the 1981 cruises, 273 sets of the purse seine were made along 12 transects, from north of the Columbia River to south of Coos Bay (Fig. 1). Transects were sampled from the 40 m isobath to distances ranging from 10 to 50 km offshore. Be- cause of time constraints, not all 12 transects were sampled on each cruise. Both day and night samples were taken in 1981. Secchi depths were determined during day hauls made in June, July, and August 1981. Samples were collected with herring purse seines operated from chartered commercial fish- ing vessels. The cruises in 1980 and in May and June 1981 used a 457 m long purse seine bor- rowed from the National Marine Fisheries Service in Seattle. This net fished about 9 m deep and sampled about 150,000 m 3 of water. A 457 m long commercial herring seine was used in July and August 1981; it fished about 15 m deep and enclosed about 250,000 m 3 of water. Both nets were constructed of 30 mm stretched mesh. Immediately after capture, king-of-the- salmon were preserved in 10% Formalin 1 and re- turned to the laboratory. Preserved lengths and weights were then measured, and stomachs were removed for analysis. Stomach contents were sorted to the lowest practical taxonomic level. Prey were then counted, blotted dry, and weighed to the nearest 0.001 g. Results and Discussion In June 1980, 22 juvenile king-of-the-salmon were collected in five of the six night sets made 100 km offshore from the mouth of the Columbia River. No other specimens were collected in any 4 Reference to trade names does not imply endorsement by the National Marine Fisheries Service. NOAA. i — i — i — i — r 7 15 ' v' d ' OR -1 # June 1980 □ May 1981 O June 1981 A August 1981 Figure 1.— Transects off the Oregon coast sampled during 1980 and 1981, and locations of capture of juvenile king-of-the- salmon. Numbers above symbols indicate how many juveniles were taken at that station. of the sets made closer to the shore in 1980. In 1981, specimens were taken in both day and night sets. Sixteen juveniles were taken from the Coos Bay region in June and one was taken in August, which were the only months that this area was sampled. Four juveniles were taken off- shore from the Columbia River in May and June, and a single fish was caught at the westernmost station of the Yaquina River transect in June (Fig. 1). 162 Forty-three of the 44 juveniles collected throughout this study were taken in May and June, while none were caught in July, and only one was taken in August. The high abundance of these fish in the late spring and early summer samples may indicate the presence of seasonal variation in their distribution. During the sampling of each transect, a dis- tinct boundary separating green coastal water from blue oceanic water was generally observed. All but one of the juveniles were taken west of this front, on or beyond the 150 m isobath. Secchi depths at the locations of capture of juveniles in June and August 1981 ranged from 11 to 25 m. In contrast, Secchi depths in the green coastal water were generally <10 m. The abundance of juveniles in the Coos Bay region is probably a re- flection of the narrow continental shelf there and the steep depth gradient within several kilo- meters of shore. Blue oceanic water with Secchi depths of 10-25 m was found to extend to within 5-10 km of the coast in June, and right up to the beach in August. The fish taken offshore in 1980 ranged in size from 68 to 509 mm SL, and weighed from 1.0 to 78.4 g. Juveniles collected inshore in 1981 ranged in length from 70 to 245 mm SL and weighed from 1.8 to 17.5 g. All 11 of the specimens >250 mm SL were taken offshore. The preserved length-weight relationship of 40 undamaged specimens can be summarized by a power curve regression equation: W = 2.04 X 10" 4 L 206 (r = 0.99; Fig. 2). The specimens collected offshore in 1980 and inshore in 1981 relied, as would be expected, on 80- 60- 40 5 20- W = 204 x 10'" L 206 r 2 r0.99 100 200 300 400 Standard Length (mm) — I - 500 Figure 2. — Length-weight relationships of 40 undamaged, preserved juvenile king-of-the-salmon. different planktonic food sources. All specimens contained at least some items in their stomachs, but the total biomass per stomach was generally <0.20 g and never exceeded 0.85 g. These low weights are more a reflection of the size and morphology of the fish than of low feeding rates. Many of the specimens had their simple, tubelike stomachs fully packed with prey. The offshore specimens fed extensively on an hyperiid amphipod, Phronima sp. (Table 1). Prey identified as Phronima were found in 15 of the 21 stomachs examined, with a maximum of 16 Phronima per stomach. Crustacean parts were found in 20 of these stomachs. These parts, particularly leg and chela segments, generally closely resembled Phronima. Other hyperiids Table 1.— Frequency of occurrence of prey taxaand maximum abundance of prey taxa in juve- nile king-of-the-salmon stomachs collected at offshore stations (1980; N = 21) and inshore stations (1981; N = 20) off the Oregon coast. 1980 1981 Number Maximum Maximum Number Maximum Maximum Prey of number per biomass per of number per biomass per taxa stomachs stomach stomach (g) stomachs stomach stomach (g) Unidentified material 21 — 0220 17 — 0.087 Crustacean parts 20 — 0512 2 0027 Amphipods Phronima 15 16 0.495 1 1 0.001 other 6 4 0.029 2 1 0.001 Copepods 12 12 0002 18 184 0.215 Euphausiids 3 2 0003 10 37 148 Shrimp larvae 1 1 0.001 Crab megalops 1 1 0.044 3 5 0021 Squid (tentacle) 1 1 0.027 Chaetognaths 2 5 0.011 Fish larvae 13 19 0.241 scales 11 6 004 163 were occasionally eaten, but did not constitute a major component of the diet. Copepods were present in 12 stomachs, but were in low numbers and probably were not very important as a dietary item. Fish scales were taken from 11 stomachs. The scales did not appear to come from other fish collected in the same net hauls and may indicate that these small-toothed juve- niles consume scales floating free in the water. One fish stomach contained a piece of a squid tentacle, further suggesting that these fishes occasionally act as scavengers by picking up debris from predation events. This reliance on Phronima as the dominant food organism is notable because of the parasitoid relationship between the Phronimi- dae and gelatinous zooplankton. Laval (1980) summarized the data known about this relation- ship and showed that Phronima spp. generally mature and live within the bodies of pelagic salps and siphonophores. Both the hosts and the amphipods are virtually transparent, and ex- ceptional visual acuity is probably necessary to locate these prey. Traces of the hosts were not found in the fish stomachs, indicating that the fish either rapidly digest the host, pick the amphipods from the host, or eat the amphipods while the amphipods are moving between the hosts. The inshore fishes caught in 1981 consumed a more varied range of prey (Table 1). Copepods were the most important prey item and were found in 18 of the 20 stomachs analyzed, in numbers ranging up to 184 copepods per stomach. Fish larvae were another important component of the diet and were found in 13 stomachs. These larvae ranged from tiny (2-3 mm) unidentifiable fish to 20 mm flatfish larvae (Hippoglossoides sp.). Up to 19 larvae were taken from a single stomach. Juvenile and a few adult euphausiids (Euphausia pacifica and Thysan- oessa spinifera) were taken from 10 stomachs, in numbers up to 35 euphausiids per stomach. Unlike the oceanic specimens, the inshore fish rarely ate hyperiid amphipods and never consumed fish scales. The dietary differences observed between the offshore and inshore collected specimens are an expected feature that reflects the availability of different prey taxa in different environments. The offshore stations had blue, clear water of relatively low particulate content, while the in- shore stations were influenced by higher coastal productivity as well as river and estuarine input. The 1981 juveniles were collected in the transi- tion zone between the oceanic and coastal en- vironments. Utilization of this ecotone perhaps enabled these fish to take advantage of a portion of the coastal productivity and yet remain in a relatively clear oceanic habitat. Acknowledgments I would like to thank W. W. Wakefield, J. Fisher, and the captains and crews of the fishing vessels Flamingo, Kristen Gail, and Soupfin for their assistance in collecting the specimens. I also thank R. Brodeur and W. Pearcy for their comments on the manuscript and A. Kendall for his data on egg distributions. This research was made possible by funding provided by the Oregon Department of Fish and Wildlife, Oregon State Sea Grant College Program, Oregon Aqua Foods Inc., Crown-Zellerbach Inc., and Ana- dromous Inc. Literature Cited Fitch, J. E. 1964. The ribbonfishes (Family Trachipteridae) of the eastern Pacific Ocean, with a description of a new species. Calif. Fish Game 50:228-240. Hart, J. L. 1943. Katsuwonus and Trachipterus in British Colum- bia. Fish. Res. Board Can. Prog. Rep. Pac. Coast Stn. 56:16. Hubbs. C. L. 1926. The metamorphosis of the California ribbon fish Trachypterus rex-salmonorum. Pap. Mich. Acad. Sci., Arts Lett. 5:469-476. Laval, P. 1980. Hyperiid amphipods as crustacean parasitoids associated with gelatinous zooplankton. Oceanogr. Mar. Biol. Annu. Rev. 18:11-56. ROEDEL. P.M. 1938. Notes on the ribbon-fish, Trachypterus rex- salmonorum. Calif. Fish Game 24:422-423. Sette, O. E. 1923. The occurrence of Trachypterus rex-salmonorum at Monterey, and notes on its post-larval growth. Copeia 1923:93-96. Walker. E. T. 1953. Records of uncommon fishes from Puget Sound. Copeia 1953:239. Jonathan M. Shenker School of Oceanography Oregon State University Con-all is. OR 97331 164 NOTES ON THE MARINE LIFE OF THE RIVER LAMPREY, LAMPETRA AYRESI, IN YAQUINA BAY, OREGON, AND THE COLUMBIA RIVER ESTUARY 1 The river lamprey, Lampetra ayresi, although uncommon in Oregon, is collected occasionally in the surface waters of the ocean and in estuaries. The species appears to be most abundant in the Columbia River estuary and is often found in Yaquina Bay. Systematic sampling programs in those two estuaries, carried out by the National Marine Fisheries Service (NMFS) in the Columbia River estuary and by Oregon State University in Yaquina Bay, have provided suffi- cient specimens (225) so that a preliminary assessment of the saltwater life of the species in Oregon can be attempted and comparisons made with its life history in British Columbia as reported by Beamish (1980). The capture of river lampreys and the sam- pling program by which specimens were ob- tained are described or outlined by Dawley et al.. 2 Durkin et al., 3 and Myers (1980). River lampreys were usually caught incidentally in studies of other species and were taken by means of beach seine, purse seine, lampara net, and 'Technical Paper No. 6201, Oregon Agricultural Experi- ment Station, Oregon State University, Corvallis, OR 97331. 2 Dawley, E. M., C. Sims, R. D. Ledgerwood, D. R. Miller, and J.G.Williams. 1981. Study to define the migrational char- acteristics of chinook and coho salmon in the Columbia River estuary and associated marine waters. Progress report of coastal zone and estuarine studies. Pacific Northwest Regional Commission and Coastal Zone and Estuarine Studies Division. Northwest and Alaska Fisheries Center, National Marine Fisheries Service. NOAA, Seattle. WA 98195. 3 Durkin, J. T., T. C. Coley, J. T. McCabe, Jr., W. D. Muir. K. Verner. and R. L. Emmett. 1981. Non-salmonid, salmonid fishes. In Columbia River Estuary Data Development Pro- gram, 1979-80 Annual Report, Vol. 2, p. 1-24, Pacific North- west River Basins Commission. National Marine Fisheries Service, NOAA. Hammond. OR 97121. bottom trawl. Mesh sizes of the nets employed were usually 6.5 mm or 9.5 mm bar measure, thus selection for larger individuals was prob- able. Additional specimens were obtained from a variety of sources. Specimens are held in the fish collection of the Department of Fisheries and Wildlife, Oregon State University (OS). Downstream Migration In British Columbia, river lampreys entering saltwater from late April to early July averaged 110 mm total length (TL); the range of lengths was 40-190 mm (Beamish 1980). We have no downstream migrants from freshwater, but we have two lots (OS 7320-1) that include specimens 115 mm long taken in marine waters on 21 May 1980. The earliest collection of the year of marine specimens in Oregon was made 5 May. One specimen measuring 161 mm long (OS 7370) from the Pacific Ocean and another measuring 206 mm (OS 4630) from Yaquina Bay were collected on that day. Both were immature and had been feeding. Because early May corre- sponds to the spawning season, the two feeders must have migrated early and apparently would have matured after the summer feeding season. From mid-May to mid-June, specimens taken from Yaquina Bay with a 9.5 mm-mesh seine ranged in length from 141 to 245 mm (Table 1). In the same period, specimens taken by various nets (including some of 6.5 mm mesh) from the Columbia River estuary ranged from 115 to 278 mm. Specimens captured in the Pacific Ocean between mid-May and 25 June ranged from 145 to 237 mm. The distribution over the size range is sparse so that modes are difficult to recognize, except that in the Columbia estuary series (OS 6852, 6856, 6857) for 4 June (n = 110) 62% of the specimens fall between 160 and 210 mm. Table 1.— Ranges and means of total length of river lampreys captured in saltwater off Oregon (by half-month periods, all years combined). Col umbia es tuary Yaquina Bay Pacific Ocean Period n Range X n Range X n Range X 5/1-15 2 157-200 178-5 4 160-206 1790 1 161 161 5/18-31 22 115-285 157.7 2 184-185 184.5 3 124-187 163 6/1-15 110 120-278 1860 3 141-245 192.7 — — — 6/16-30 2 163-167 165.0 1 255 255 4 145-237 198 7/1-15 9 125-171 214.3 4 159-231 179.0 7/16-31 — — — 8 133-241 179 6 8/1-15 5 192-310 243.1 4 193-255 217.8 8/16-31 25 176-304 236 .1 8 184-247 2159 9/1-15 6 259-282 2677 1 260 260 9/16-30 1 240 240 10/1-15 — — — 11/1-15 1 205 205 FISHERY BULLETIN: VOL. 81. NO. 1. 1983. 165 Maturation Individuals captured May through August in saltwater show little development of the gonads, except for specimens >250 mm taken 31 August 1979 (OS 6858). These specimens have gonads visibly larger than those of smaller individuals. In addition, at least one of the allometric changes associated with sexual maturity is evident. The eyes of this 250-304 mm group constitute <25% of the preorbital length, whereas in 181-245 mm specimens from OS 6858 the eye constitutes be- tween 25 and 33% of preorbital length. One specimen (OS 17) of 205 mm TL captured 14 November 1949 in Yaquina Bay had developing gonads. The season of spawning in the Columbia and Yaquina systems is deduced to be April and May, based on four specimens as follows: OS 112, 267 mm, March 1940, mature migrant, Bonne- ville Dam, Columbia R.; OS 343, 263 mm, 30 April 1958, mature migrant. Tongue Point, Columbia R.; OS 537, 181 mm, 15 April 1959, spawner, Yaquina R.; OS 471, 203 mm, 9 May 1959, spawner, Simpson Cr. (trib. Yaquina R.). Vladykov and Follett (1958) suggested that spawning of the species took place in April and May. Beamish (1980) reported spawning in hold- ing tanks during May. October if these animals grow at the rate observed by Beamish (1980) in British Columbia. In that study, an increase of 100 mm from mid- June to mid-August was noted. In the present study, a rough estimate of growth in the Columbia can be made by comparing early June samples (n = 110), which had a mean length of 186 mm, with combined samples from 31 August and 2 September {n — 31), which averaged 242 mm. In a system such as the Columbia, assessment of size and growth is complicated by factors other than sampling problems. Some individuals may spend more years as larvae than others, some may transform and migrate to saltwater earlier in the year than others, some may feed in freshwater before entering saltwater (Beamish 1980), and those destined to migrate back to distant tributaries might have the genetic capacity for rapid growth and early departure from the feeding grounds. Kan (1975) noted that Pacific lampreys showed a rough correlation between size and distance of migration in the Columbia, but in that species large size can be reached not only by fast growth but by spending up to 3 or 4 yr in marine waters, rather than the few to several months spent by the river lamprey. Growth and Upstream Migration Although occasional adult specimens of the river lamprey have been taken from Yaquina Bay during October and November, no river lampreys have been captured in the Columbia River estuary from early September to May. The Pacific lamprey, Lampetra tridentata, has appeared December to June in catches from the Columbia estuary, intimating that the gear used during the winter is capable of capturing lampreys and that the absence of the river lamprey from the catch indicates their absence from the estuary. We suggest the absence means that river lampreys move into freshwater in early autumn. Judging from the specimens caught from mid- August on, adult river lampreys must move into freshwaters of the Columbia system at lengths of of about 200 mm to >300 mm. Those that feed in Yaquina Bay probably leave saltwater at similar sizes, although the largest specimen captured there was 260 mm. Specimens up to 255 mm have been taken in Yaquina Bay in June, thus lengths of 300 mm could be reached by September or Ecological Observations All but two of the eight ocean-caught river lampreys were taken in tows or hauls made within 34 m of the surface. The remaining two were taken close to the surface by anglers. Speci- mens from Yaquina Bay were taken by seine (3 m deep), but usually by lampara net (21 m deep) from subtidal channels. Specimens from the Columbia estuary were taken from shallow water by purse seine and beach seine. "Pelagic" coloration of blue to black on the back and silver on sides and belly appears to be typical of actively feeding L. ayresi, as reported by Kan (1975) and Beamish (1980). This contrasts sharply with the grey coloration of the deep- dwelling Pacific lamprey. Water temperature in Yaquina Bay at times of capture of river lampreys ranged from 13° to 21°C. Salinity ranged from 12 to 29%« (Myers 1980). Associated fishes in Yaquina Bay were usually American shad, Alosa sapidissima; Pacific herring, Clupea harengus pallasi; juvenile coho salmon, Oncorhynchus kisutch; juvenile chinook salmon, O. tshawytscha; surf 166 smelt, Hypomesus pretiosus; and shiner perch, Cymatogaster aggregata. Scars from attacks by lampreys were occasionally seen on juvenile salmonids, usually just below the dorsal fin. Scars were noted less commonly on other species, but some were noticed on a wide range of sizes of fish, including adult pile perch, Rhacochilus rur- al. Two of the ocean-caught lampreys were taken while attached to a herring and a smelt of un- known species that anglers were using for bait. Feeding Habits Beamish (1980) presented data on the feeding habits of the river lamprey, mentioning sal- monids and Clupea as common prey. Miller 4 ob- served what he considered significant predation by the river lamprey on chinook salmon 60-120 mm long in Elliott Bay, Wash. In the present study, 141 of the 225 specimens from marine water were examined for evidence of feeding. Only four had empty guts. Gut con- tents of 30 specimens (OS 6857) captured 4 June 1979 from the Columbia River estuary were ex- amined for identifiable material. Fragments of muscle tissue, intestine, liver, ovary, scales, and bones were present in some combination in all guts examined. Scale and bone fragments identi- fied as clupeid were found in 14 guts, one of which also contained a worn lamprey tooth lamina and a scale from a salmonid. The sal- monid scale had an ocean-type nucleus and re- sembled scales of chum salmon, Oncorhynchus keta. Clupeid scales from five guts were identi- fied as being from American shad, which migrate up the Columbia in great numbers during June. Scale fragments from six guts were thought to represent Pacific herring. One gut had no recognizable clupeid remains, but held a small salmonid scale with two freshwater annuli, thus probably being from a smolt steel- head, Sal mo gairdneri. The guts of 9 of 10 specimens (OS 6858) taken 31 August 1979 from the Columbia estuary con- tained recognizable clupeid remains. One con- tained an American shad scale and three held fragments thought to be from Pacific herring scales. Seven contained forked intermuscular bones. In addition to clupeid remains, two guts held fragments of unidentified salmonid scales. Acknowledgments We are grateful to Terry Durkin, Greg Hamman, Richard Ledgerwood, David Miller, and Walter Receconi for their help in supplying information and specimens. James D. Hall, Howard F. Horton, and Richard A. Tubb re- viewed the manuscript and made helpful sug- gestions. Literature Cited Beamish, R. J. 1980. Adult biology of the river lamprey (Lampetra ayresi) and the Pacific lamprey {Lampetra tridentata) from the Pacific coast of Canada. Can. J. Fish. Aquat. Sci. 37:1906-1923. Kan. T. T. 1975. Systematics, variation, distribution, and biology of lampreys of the genus La mpetra in Oregon. Ph.D. Dis- sertation, Oregon State Univ., Corvallis, 204 p. Myers. K. W. 1980. An investigation of the utilization of four study areas in Yaquina Bay, Oregon, by hatchery and wild juvenile salmonids. M.S. Thesis, Oregon State Univ.. Corvallis, 234 p. VLADYKOV, V. D., AND W. 1. FOLLETT. 1958. Redescription of Lampetra ayresii (Giinther) of western North America, a species of lamprey (Petromy- zontidae) distinct from Lampetra fluviatilis (Linnaeus) of Europe. J. Fish. Res. Board Can. 15:47-77. Carl E. Bond Department of Fisheries and Wildlife Oregon State University Corrallis, OR 97331 Ting T. Kan Department of Fisheries anil Wildlife Oregon State University Corrallis, OR 97331 Present address: Papua Neiv Guinea University of Technology Lae, Papua New Guinea Katherine W. Myers Department of Fisheries ami Wildlife Oregon State University Corrallis, OR 97331 Present address: Fisheries Research Institute University of Washington, Seattle, WA 9H195 4 Denny M. Miller, formerly Research Assistant, University of Washington, Seattle, WA 98195, pers. commun. April 1968. 167 AN ECONOMIC EVALUATION OF THE ST. LAWRENCE RIVER-EASTERN LAKE ONTARIO BASS FISHERY The St. Lawrence River-eastern Lake Ontario bass fishery has long been known as one of the finest sport fisheries in North America. Despite its well-documented popularity, there has been little research on this recreational fishery's eco- nomic value. Furthermore, recent interest has focused on the fishery's trout and salmon angling opportunities, which have been significantly en- hanced since the early 1970's through the man- agement efforts of New York's Department of Environmental Conservation. This study pro- vides information on the economic importance of the bass fishery, considered by many to be one of the best smallmouth bass fisheries in the world. The economic value of this recreational fishery should be taken into account in decisions affect- ing use of the St. Lawrence River and for plan- ning and evaluating management of this re- source. The current study estimated the economic value of the St. Lawrence River-eastern Lake Ontario bass fishery to licensed New York resi- dent anglers. Benefits to out-of-state anglers (including Canadians) and nonlicensed anglers were not evaluated, nor were Canadian sites in the region included in this study. In addition, general recreational benefits of the fishery to tourists and others were not considered. Though a recreational fishery may be of value from a number of perspectives, it has long been estab- lished on conceptual grounds that economic eval- uation of recreation benefits should be based on the willingness of users to pay for services pro- vided. However, willingness to pay for outdoor recreation facilities cannot be estimated through the normal procedure of observing market de- mand because the typical practice is to provide these facilities to users free of charge. This study used the so-called travel cost meth- od to estimate demand for the angling services of the St. Lawrence River-eastern Lake Ontario bass fishery. The first section of this article dis- cusses the method that was used to estimate the fishery's economic value. It includes a descrip- tion of the fishery and a discussion of the travel cost method and the data. The second section pre- sents the empirical findings. The concluding sec- tion discusses the implications of the results for management policy. Methods Determining the Value of Recreation Facilities There is a substantial body of literature on esti- mating economic value to users of outdoor recre- ation. Two approaches have been widely used to obtain information for estimating economic val- ue. The first asks individuals to reveal directly their willingness to pay for use of a recreation site. An important problem with this approach is the incentive to misstate true preferences, pos- sibly leading to inaccurate estimates of economic value (Freeman 1979). The other procedure for estimating economic value is the travel cost method, first applied to outdoor recreation by Clawson (1959) and Clawson and Knetsch(1966). The hypothesis of the travel cost method is that outdoor recreation demand can be estimated by observing how visitation to a specific site varies with differences in costs of traveling to the site. Travel costs are viewed as a charge for use of a resource's services, and the pattern of visitation by geographical area indicates the willingness to pay for its use. The travel cost method is a two-stage estima- tion procedure. The first stage predicts site visi- tation as a function of travel costs and other explanatory factors. Then a demand curve is de- rived showing how visitation would vary in re- sponse to a price (or entrance fee) charged for use of the site, assuming that users view an increase in price as equivalent to the additional costs needed to travel greater distances to the site. The site's net economic value (NEV) in its current use is equal to the area under the demand curve above the level of travel costs (Clawson and Knetsch 1966; Dwyer et al. 1977). 1 The Participation Equation Visitation patterns to the St. Lawrence River- eastern Lake Ontario area (Fig. 1) during the 1976-77 year form the basis for this analysis. The equation for predicting visitation to the fishery was based on a survey of licensed New York resi- dent anglers (New York Department of Environ- mental Conservation 1976). The sample was lim- ited to 904 anglers (from 51 of New York's 62 'The travel cost method assumes that users derive benefits from the recreation site itself rather than the trip (Brown et al. 1965). 168 FISHERY BULLETIN: VOL. 81. NO. 1. 1983. St. Lawrence County s i tes on the St . Lawrence Ri ver Jefferson County sites on the St. Lawrence Ri ver Jefferson County sites on Lake Ontario Oswego County sites on Lake Ontario Wayne and Cayuga County sites on Lake Ontario Figure 1.— Map of St. Lawrence River-eastern Lake Ontario bass fishery. counties) who spent >5% of their time fishing for either smallmouth or largemouth bass at one of the designated sites. The study area comprised five sites chosen on the basis of geography, avail- ability of data, and observed visitation. Two sites were on the St. Lawrence River and three were on eastern Lake Ontario. The participation equation is equivalent to an ordinary demand function for a marketed com- modity where quantity (visits to a site) is a func- tion of prices (travel costs), income, and qualita- tive characteristics. The participation equation for the bass fishery was: A, =f(TCa, It, PF„ S h Aj, TC U ) (1) where D,j = total days angling at site j by re- spondents from county of ori- gin i for the 1976-77 fishing season / = a symbol representing an expli- cit functional relationship be- tween Dij and the explanatory variables Tdj = travel costs from county of origin i to site j; calculated by mea- suring road distance from the midpoint of each county to the PF; Si A< = TC lk = midpoint of each site and mul- tiplying the measured distance by an estimate of the cost per mile average annual income of an- glers from county of origin ;' average preference level for bass of anglers from county of ori- gin i; preference level repre- sents the percentage of total angling time spent fishing for the species of interest number of anglers to whom the questionnaire was sent in coun- ty of origin i; a constant per- centage of the angler popula- tion across all counties relative attractiveness of site j; the amount of shoreline miles at site j relative to the total miles available at all sites an index of travel costs from county of origin i to substitute angling sites in the study area. This demand function relates participation at sites not only to their own prices and quality, but also to the attributes of comparable substitute 169 sites. Travel costs were assumed to be a function of both monetary expenditures and the cost of travel time. 2 Ignoring time costs will cause biased estimates of demand and economic value (Cesar- io and Knetsch 1970). Cost of travel time was cal- culated by multiplying estimated travel time en route to the site by an hourly wage rate (Knetsch et al. 1976). Sample size was included as an inde- pendent variable in the participation equation because others have found that visitation in- creases at a nonlinear rate with increases in popu- lation (Cesario and Knetsch 1976; Grubb and Goodwin 1968). Travel costs to substitute sites, TC, k , were represented in an index of travel costs reflecting the availability of substitute angling opportunities. 3 The attractiveness of available recreation sites can also be an important determi- nant of visitation patterns. The decision to visit a particular site depends, in part, on the attractive- ness of that site compared with other available sites. Site attractiveness measures used by others have included angling success rates (Stevens 1966), size of the recreation area (Ravenscraft and Dwyer 1978), congestion at the site(McCon- nell 1977), and accessibility (Cesario and Knetsch 1976). Data limitations reduced the possible choices for attractiveness variables in this study to fishing success rates and shoreline distance. Site Demand and Economic Value The second step of the travel cost method de- rives the demand for and economic value of the recreation site from the participation equation. The usual procedure is to derive a demand curve for a specific site by estimating demand from each origin and aggregating over all origins for "'Travel costs were converted to price per angler day by taking into account travel distance and whether lodging ex- penditures were reported by anglers. Analysis of the survey data indicated that anglers who resided at a (one-way) dis- tance between 125 and 175 mi from the site generally incurred lodging expenditures, indicating an overnight stay at the site. Accordingly, price per angler day was assumed to equal one- half the estimated travel costs for anglers residing more than 150 mi from a site. For anglers closer to the site, price per angler day was assumed to equal estimated travel costs. Mone- tary costs were assumed to be IOC/mile. Travel time costs were calculated by multiplying estimated travel time at 50 mph by a value equal to 35% of the wage rate in the angler's county of origin. Hotel costs were not included in the cost esti- mates since they could not be determined on a per angler day basis. 3 Use of an index reflects the overall availability of substi- tutes. Dividing the index by four would give the average price of a substitute site in this fishery. A generalized approach to the treatment of substitute sites is preferable to a specific sub- stitute site in a regional travel cost model (Cesario and Knetsch 1976; Dwyer et al. 1977: Ravenscraft and Dwyer 1978). each increment of a hypothetical fee until aggre- gate demand for the resource is reduced to zero (Grubb and Goodwin 1968; Cesario and Knetsch 1976; Knetsch et al. 1976). This study estimated NEV for each origin using a separate site-spe- cific demand curve. Then the site's total NEV was found by numerical aggregation across all origins. This procedure estimates NEV more accurately than the usual procedure because there is less aggregation in deriving the site de- mand curve (McConnell and Norton 1976; Menz and Wilton 1982). 4 Demand was estimated from the participation equation for each site with the following: Dij = C tJ + pi (Tdj + p) + e (2) where D,j = the observed days of participa- tion when the fee is zero 5 Tdj = travel costs from county of origin i to site j C>j = the composite of all other vari- ables p — the hypothetical fee charged for use of the site e = an error term. The site's NEV to anglers in each origin was ob- tained by integrating the demand equation be- tween the limits of current travel costs and the cost at which Dij would become zero. Results Some anglers may fish exclusively for small- mouth bass, others for largemouth bass, and some may be unconcerned about the specific type of bass caught. Therefore, three separate analy- ses were conducted: one each for the smallmouth and largemouth bass fisheries and one for the "combined" bass fishery. The value of the com- bined fishery was determined in a separate anal- ysis because addition of the smallmouth and largemouth bass results would double-count anglers who fish for both species. The same fish- ing sites were used for each analysis. Characteristics of anglers and sites are pre- sented in Tables 1 and 2. Smallmouth and large- "This method will be more accurate than if an aggregate de- mand curve were used, but it will not provide as accurate an estimate of economic value as aggregation of individual eco- nomic values (Brown et al. 1965; Smith 1975a). 5 The value of Dy.was set equal to zero whenever a negative quantity resulted from the calculation. 170 Table 1.— Characteristics of New York resident anglers in the St. Lawrence River- eastern Lake Ontario bass fishery, 1976. Combined bass Smallmouth bass Largerr outh bass Standard Standard Standard Mean deviation Mean deviation Mean deviation Preference (%) 378 23.7 306 14,5 278 13.3 Experience (No. of years; 297 14.3 31 8 239 229 153 Education (No of years] 133 29 133 29 13.2 2.9 Annual income ($) 18,100 9,500 18,600 9.900 16.900 8,500 Table 2.— Characteristics of sites and angler participation in St. Lawrence River-eastern Lake Ontario bass fishery, 1976. Ai igler pa rtici pation Dista ice traveled Success rate Shoreline (mi) (d sh per angler day) Standard Site (mi) Mean deviation Si nallmou th Lar gemouth Combined St. Lawrence River: St Lawrence County (Site 1) 64 171 93 3 0.85 020 0.84 Jefferson County (Site 2) 48 149 698 0.88 31 1 03 Lake Ontario: Jefferson County (Site 3) 63 99 63.8 1.30 031 1 39 Oswego County (Site 4) 24 62 57.5 64 026 071 Wayne & Cayuga Counties (Site 5) 32 35 29 7 073 25 080 Entire fishery 231 110 82.0 094 0.27 1.02 mouth bass anglers were similar in socioeconom- ic characteristics, but the average smallmouth bass angler had more angling experience. Aver- age one-way distance traveled by anglers to the sites varied from 35 mi for the Wayne and Cayuga County sites on Lake Ontario to 171 mi for the St. Lawrence County sites on the St. Lawrence River. Angling success rates were highest at the Jefferson County sites on Lake Ontario. The Participation Equation There does not appear to be any theoretical justification for a particular functional form of the relationships for estimation (Smith 1975b). Various functional forms of the participation equation (Equation (1)) were estimated. The final form was as follows: log (Di 3 + 0.8) = /So + 0i log TC tJ + fo log I, + 3 log PFi + 4 log Si + 05 l0gi4; + 06 log TC lk + e (3) where the terms are parameters to be estimated and e is the random component. 6 The double log- arithmic model produced more significant pa- rameter estimates and also exhibited greater explanatory power than linear and semilogarith- mic forms, so it was used to derive the estimates for this part of the analysis. 7 The results for the participation equation (Equation (3)) are presented in Table 3. Because assumptions about monetary and time costs of travel could influence the results, alternative participation equations were estimated using dif- ferent values for these cost components. The re- sults are also shown in Table 3. The effect and significance of the explanatory variables re- mained virtually unchanged, suggesting that confidence can be placed in the results from this stage of the analysis. The estimates are consistent with theoretical expectations and are similar for the three fishery specifications. Most of the estimated coefficients were statistically significant at the 1% level and 6 The quantity, 0.8, in Equation (3) is added to the fee to pre- vent the use of the logarithm of zero. All logarithms are natural logarithms. The objective in specifying the participation equation was to obtain reliable estimates of parameters rather than a high R 2 (Gum and Martin 1975). Other studies that have used the double log format are Grubb and Goodwin (1968). Smith (1975b), and Smith and Kopp(1980). 171 E o O o £ CD O E £ CO E o O •a c Q £ o E ca E CO £3 £ o O o £ co o £ "5 E CO cd — r-- — o — co — i-irirco6t-i-cb tJ- --.CD — C\J , — . LO — CMNCOSCONN'J 0> O CD O O r- n i- co s n ^r ,- _ o> — ro io h- o i- c\j cd co cm t- in *-; rr d n r- co n 6 I I i- I CM CO — LO - — . r^- — O — - ^T — Q) — CM — - , ^ TttncNjcoTrCTJunco ^r co cd co cm io a) uo toioo^cDi-w ^r lo m cnj co s o t T--«3 : T- : C\JO , ^T-cb r- IT) O CNJ LO *— CM O ll -- -- ~ " I I I l CO CO— ~ O — i- — . CM—- CO — COCOCOLOr--CDCDO LOCO TTONCSJ^CT)'-^ ^ f"- T^LOOCSJO^-T-CO i- LO i- — LO co — ^r — ^ — - CMOOICMCOCOOOCO TTNOJNCOfOOCO lo — lo — cd — cd — lococolotj-ojcdco ^0)C0O^0)t-0J T^Tj-'ocvioT^-^cb cm — r- — l- O i- *- O) cnj r-- CO i- CM CO CO -^ O CO i- O CD o I I I I w r- — oj — lo LO CO O 00 G) CO CO CNJ O CM CM -^ d ro n c\j s d I I I I CM O) — *- — CD T-lfiOtMOT-i-CO cd— ^ — r*- — co — CMCOCDNCOOINCD ^rcD'^-'-cocooco i- -^ i- ^r d i- i- r- cnj — t- — r- — cm. — . LOr^LOLO^TOlCDLO TT O)C0lOtJ-O)t-CNJ r^vocNid^^cb tT — f- CNJ CO r- oi s coo om LO CM CO CD CNJ ^r d cm \ t- o co o I I I I CO co . — — ai — i- loco O) o co ro to n Tt N N i- CO C\J N ^J T-id d co" co i- cd d ■^- LO T^ LO O CO O) CO N O CO — CO — LO en n t co o LD LO CM CM CO i- TT CM TJ- i- CO O u. CO C o a CD *■* — • b CO o i u_ C ro u « o- cd UJ 2 t- ro o t"cd c; *~ > E CD ^ *2 ro -s CD t/i in CD -C 3 .t = «$.? ro E>o in u. Z all except those noted as such in Table 3 were sig- nificant at the 5% level. Travel costs from origin i to site j(TC u ) and the measure of site attractive- ness (Aj) were found to be highly significant de- terminants of participation. The effect of substi- tutes on site visitation depends on their location and attractiveness relative to the site being studied (Burt and Brewer 1971; Cicchetti et al. 1976; Dwyer et al. 1977). The negative and statis- tically significant coefficient for TC ik suggests that the sites in this fishery serve as complements for one another and that anglers are drawn to the fishery as a whole instead of to a particular site. Economic Value of the Fishery Table 4 presents the estimated net economic benefits to New York resident bass anglers for the fishery. Values were estimated for each site and for each species of bass on a per angler day basis and as an annual total. The annual total for each site was calculated by multiplying the value per angler day by the estimated number of an- gling days as given in Table 5. B and C of Table 4 show the effect of alternative assumptions about distance and time costs. It can be seen that the results vary widely from site to site and with different assumptions con- cerning the monetary component of travel costs. Variation among sites is due to the relative at- tractiveness of the sites, size of population in nearby counties, and other factors affecting visi- tation patterns. These factors affect the willing- ness of anglers to pay for the sites' services and the number of anglers attracted. Highest values per angler day were estimated for St. Lawrence County sites on the St. Lawrence River. At two sites the NEV per angler day for largemouth bass exceeded the NEV per angler day for small- mouth bass. However, due to a greater number of estimated angling days for smallmouth bass at these two sites, the total NEV of the smallmouth bass fishery exceeded that of the largemouth bass fishery for every site. 8 The value of the combined bass fishery at each site is less than the total of the individual smallmouth and largemouth bass values because the fisheries for 8 It should be emphasized that the total value of the fishery equals the estimated number of angling days at each site times the per angler day value. This assumes that the angler day is entirely attributable to the site's bass fishery. To reduce pos- sible bias from this assumption, the sample population was limited to anglers who fished at one of the five sites and indi- cated that they had spent more than 5% of their time fishing for bass. 172 Table 4.— Net economic value of the St. Lawrence River-eastern Lake Ontario bass fishery to New York resident anglers, 1976. Location Type of fishery Site 1 Site 2 Site 3 Site 4 Site 5 Total A. Estimates using travel costs of 10d Gn k R N Assimilation efficiency (dimension- less) Volume searched during feeding (1/ fish per hour) Volume swept clear during feeding (1/fish per minute) Food particle length (/im) Filtration efficiency of the gill rakers (dimensionless) >G, OPT S K, OPT Swimming speed during feeding ("foraging speed") (cm/second) Concentration of plankton in the wa- ter (kcal/1; mg N/1) Duration of the daily feeding period ("foraging time") (hours/day) Foraging speed which maximizes growth rate at a given concentration of plankton Foraging speed which maximizes gross growth efficiency for a given foraging time. Atlantic menhaden are highly specialized plank- tivores which feed on suspended particulate material (phytoplankton, zooplankton, and detritus). During feeding, an Atlantic menhaden swims with its mouth open and gill opercula flared, causing the comblike gill rakers, which otherwise lie flat inside the mouth, to swing inward and form a fine-meshed screen across the throat (Peck 1894). Water entering the mouth is filtered through the rakers before exiting through the gill arches. Adult Atlantic menhaden do not pursue individual prey (Durbin and Durbin 1975). Instead they filter the column of water that lies direct- ly ahead. Although the menhaden are size-selective, this merely reflects the mesh size of the gill rakers and does not represent active selection for specific types of prey. Laboratory studies have shown that Atlantic menhaden change their swimming and feeding behavior according to the concentration of food in the water (Durbin et al. 1981). In the absence of food the fish swam at a characteristic speed of 0.47 body lengths/s, with a routine respiration rate of 0.1 mg 2 /g wet weight per h. The menhaden increased their swimming speed and respiration rate severalfold during feeding. Foraging speed increased asymp- totically with increasing food concentration, while respiration rate increased exponentially with in- creasing foraging speed. The fish initiated and ter- minated feeding at distinct threshold concentrations of plankton that were inversely related to particle size. Exogenous nitrogen excretion in the Atlantic menhaden was proportional to the nitrogen content of the ration (Durbin and Durbin 1981). Digestion rates were rapid, and assimilation efficiency was high. The menhaden were evidently adapted for the efficient processing of large amounts of particulate material which is ingested during prolonged periods of continuous feeding. These observations provide the basis for the development of the energy and nitrogen budgets and will be discussed in more detail below. In accordance DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN with the experimental data, the budgets are de- veloped for the case of an adult Atlantic menhaden, 26 cm FL (fork length) and weighing 302 g wet and 101 g dry, which feeds upon the diatom Ditylum brightwelli. The temperature is 20°C. DERIVATION OF ENERGY AND NITROGEN BUDGETS Energy Budget (Model I) The general equation for the energy budget is pre- sented in Equation (1). Energy Intake TOTAL DAILY RATION, R K (KCAL/G DRY WEIGHT PER DAY).— The daily ration, R K (kcal/ fish per d) which is obtained by an Atlantic menhaden will be equal to the volume searched (i\ 1/fish per h), times the efficiency (e, dimensionless) with which particles are removed from the volume searched, times the concentration (c, kcal/1) of food particles in the water, times the duration (h, h/d) of the feeding period. F e = — v (5) #„ v e c h (kcal/fish per d). (3) Volume searched (v).— During feeding, the mouth is held continuously open and the fish swim in school formation, travelling along a straight or curvilinear path without changing course to pursue individual prey. Thus each fish filters the column of water which lies directly ahead. The volume searched is equal for all prey types, and may be adequately described as a cylinder, or, more accurately, an ellipsoid, with a cross-sectional area equal to that of the fish's open mouth and a length equal to the distance covered by the fish per unit time, i.e., the foraging speed (s, cm/ s) . For an Atlantic menhaden averaging 26 cm FL, the gape was approximately elliptical, with major and minor axes of 3.91 and 2.90 cm, respectively; the to- tal cross-sectional area of the mouth was therefore 8.93 cm 2 (Durbin and Durbin 1975). Thus v= 32.148 s (1/fish per h). (4) In feeding experiments (Durbin and Durbin 1975) the mean value of F for Ditylum brightwelli was 5.8 1/ fish per min, while v was estimated to be 23.3 1/fish per min. This gives a value of e = 0.25 for D. brightwelli. Filtration efficiencies for different-sized particles may be calculated from an equation describing the relationship between filtration efficiency and food particle length (Durbin and Durbin 1975): F = 8.290 \og l0 L - 9.733 (1/fish per min). (6) In the experiments the fish were unable to filter par- ticles smaller than about 13 /on. Incorporating the appropriate values for u and e into Equation (3), the ingested ration, R K , for D. brightwelli would be given by Filtration efficiency (e).— Filtration efficiency is the efficiency with which the Atlantic menhaden filters particles of a given size from the water and is equal to the observed removal rate or volume swept clear, F (1/fish per min) , divided by the total volume searched, v (1/fish per min), i.e., p = 8.037 sc h (kcal/fish per d). (7) In the model the Atlantic menhaden weighed 302 g wet = 101 g dry (Durbin and Durbin 1981). Thus r k = 0.079574 sch (kcal/g dry weight per d). (8) ASSIMILATED RATION, P R K (KCAL/G DRY WEIGHT PER DAY).— If the fecal losses, F K , are subtracted from the ingested ration, R K , a measure of the assimilated ration is obtained. The assimilated ration can also be determined by multiplying R K by the assimilation efficiency, p, i.e.,pR K , where R K ' (9) In our experiments with the Atlantic menhaden, we observed slight changes in the overall assimilation efficiency of a meal, depending on meal size (Durbin and Durbin 1981). However, because the observed differences in overall assimilation efficiency were small and because of the uncertainty about the significance of these differences, we assumed a con- stant assimilation efficiency for the model and took the means of the experimentally determined values. For Atlantic menhaden feeding onD. brightwelli, the mean assimilation efficiency, p, equalled 0.8636 for carbon, 0.9240 for nitrogen, and 0.8954 for calories (Durbin and Durbin 1981). Substituting Equation (9) into Equation (1) we may rewrite the general equation for the energy budget: G K = pR K -T K -E 1 (10) 179 FISHERY BULLETIN: VOL. 81. NO. 2 where the assimilated daily ration, pR K , is given by pR K = 0.8954 R K (kcal/g dry weight per d) (11) = 0.071250 sc h (kcal/g dry weight per d). Energy Output (12) RESPIRATION, T K (KCAL/G DRY WEIGHT PER DAY).— In the absence of food, the Atlantic menhaden swam at a characteristic speed of 1 2.2 cm/ s (0.47 body lengths/s), with a routine respiration rate of 0.10 mg 2 /g wet weight per h (Durbin et al. 1981). During feeding the fish increased their swimming speed by a factor of 2.4- to 3.5-fold above the non- feeding rate, depending on the plankton concentra- tion in the water. Both swimming speed and respiration rate increased abruptly with the onset of feeding, and stabilized within a few minutes. One of the more interesting aspects of the Atlantic men- haden feeding behavior was that they would maintain a virtually constant swimming speed throughout the entire 7-h experimental feeding period, if the input of food remained constant. When the food input was stopped, the fish quickly consumed the remaining plankton in the tank, decreasing their swimming speed as the plankton concentration dropped. Thus the return to the routine swimming speed following feeding was quite rapid. In low-ration experiments, respiration rates declined to the routine, prefeeding rate almost immediately after feeding. In high-ration experiments, respiration rate remained slightly el- evated above baseline for 2-5 h after feeding. The amount of energy expended above routine during the postfeeding period was small and did not show any clear relationship with food ration size. It has therefore been omitted for the purpose of the en- ergy budget. Based on these considerations, the respiratory costs for the energy budget are considered separate- ly for periods of feeding and nonfeeding. Thus T K = T r> K + 7) K (kcal/g dry weight per d) (13) where T K = total daily expenditure for respiration T r K — routine respiration during the non- feeding period T f K = respiration during feeding. Oxygen consumption rates were converted to caloric equivalents by means of oxycalorific coef- ficients in Elliott and Davison (1975). The appro- priate coefficients were determined from the ratios of oxygen consumed: nitrogen excreted by Atlantic menhaden before, during, and after feeding (Durbin and Durbin 1981). During feeding, Atlantic men- haden swimming at their preferred speed of about 41.3 cm/s appeared to be catabolizing protein. An ox- ycalorific coefficient of 3.20 X 10~ 3 kcal/mg 2 was therefore used during periods when the fish were feeding. Nonfeeding menhaden catabolized about 28% protein and 72% fat (where Q ox = 3.28 X 10" 3 kcal/mg 2 ), and the combined oxycalorific coeffi- cient was 3.258 X 10" ;i kcal/mg 2 . Routine respiration rate, T r K . — The routine respi- ration rate of quietly swimming, nonfeeding Atlantic menhaden was 0.10 mg 2 /g wet weight per h = 0.2 99 mg 2 /g dry weight per h = 0.000974 kcal/g dry weight per h (Durbin et al. 1981). Thus the daily routine respiration during the nonfeeding period is given by T rK = 0.000974 (24-/2) (kcal/g dry weight per d)(14) where h is the duration of the feeding period (h/d). Respiration during feeding, T f K . — The respiration rate increased significantly during feeding. This in- crease could be attributed to three sources: The higher voluntary swimming speed, the possible effect of excitement, and the specific dynamic effect of the food (SDA). The swimming speed was clearly the dominant factor, and accounted for 84.37c of the in- creased respiratory rate during feeding and 73.3% during the postfeeding period. Excitability was dif- ficult to quantify, but our qualitative observations of the behavior of the fish indicated that they were least excitable during feeding and most excitable during the postfeeding period when they continued to hunt for food after the input to the tank had been stopped. SDA is considered to represent mainly the loss of en- ergy during the deamination of protein, and it ap- pears to constitute a fixed proportion of the energy content of a particular type of food (Muir and Niimi 1972). The energy cost of SDA is usually determined by monitoring the metabolic rate of the fish following a meal. Unfortunately in the present study we were unable to measure SDA separately because of the prolonged feeding period, during which ingestion and digestion occurred simultaneously. However, since about 80% of the ration was digested and assimilated during the 7-h feeding period (Durbin and Durbin 1981), most of the respiratory cost of SDA was included in the measurement of the total respiration rate during feeding. 180 DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN The total respiration rate during feeding increased exponentially with increasing foraging speed (s, cm/ s), where T^—IO 002948 *- 15342 (mg 2 /g wet weight per h) (15) 7}.* = 2.994 (10002948 s-1.5342) ( m g 2 /g dry weight per h). (16) Converting to calories and Watts 1974). The caloric equivalents of these compounds are: 5.51 X 10 3 kcal/mg urea-N (Elliott and Davison 1975), 13.32 X 10~ 3 kcal/mg creatine- N, and 41.3 X 10" 3 kcal/mg trimethylamine-N (Weast 1977). The mean value for these compounds was 20.04 X 10" 3 kcal/mg DON. Endogenous nitrogen excretion, E 1)K .— The endoge- nous excretion rate equals 10.72 jug N/g dry weight per h (Durbin and Durbin 1981). The daily endoge- nous nitrogen excretion was therefore T fK = 0.00958 (10o-o2948.-i.5342) (kcal/g dry weight per h). (17) The daily energy expenditure for respiration during feeding is therefore T t K = 0.00958 h 10 002948s - dry weight per d). 1.5342 ) (kcal/g (18) Total daily respiration, T K . — Combining Equations (14) and (18) we obtain an expression for the total respiratory expenditure per day as a function of the foraging speed (s, cm/s) and the foraging time {h, h/ d): T K =h [0.00958 (10 ,l0294 » -15342) - 0.000974] + 0.02338 (kcal/g dry weight per d). (19) NITROGEN EXCRETION, E K (KCAL/G DRY WEIGHT PER DAY).— Energy is lost through the excretion of nitrogenous compounds. In the absence of food the fish excreted nitrogen at a low rate (basal or endogenous excretion, E b N ). Nitrogen excretion increased as a result of feeding (exogenous excretion, E fN ). The total daily nitrogen excretion (£ N ) is thus: E s = £ b N + E f< N (mg N/g dry weight per d). (20) The energy equivalent of the excreted nitrogen was determined as follows: Of the total nitrogen excreted by menhaden, 69.6% was in the form of ammonia and 30.4% was in the form of dissolved organic nitrogen (DON) (Durbin and Durbin 1981). The caloric equi- valent of ammonia nitrogen is 5.94 X 10 3 kcal/mg NH 3 - N (Elliott and Davison 1975). The individual compounds comprising the DON excreted by Atlan- tic menhaden were not determined. For the purpose of the energy budget, the DON was assumed to con- sist of equal parts of urea, creatine, and trimeth- ylamine, the major organic nitrogen compounds which are known to be excreted by teleosts (Watts E bN = 0.257 (mg N/g dry weight per d). Converting to calories (21) E b< K = 0.0026282 (kcal/g dry weight per d). (22) Exogenous nitrogen excretion, E f K . — The exogenous nitrogen excretion of menhaden fed D. brightwelli was directly proportional to the total nitrogen con- tent of the ration, R s (mg N/g dry weight per d) (Dur- bin and Durbin 1981): E fili =0.616 # N - 0.020 (mg N/g dry weight per d). (23) Converting to calories E fK = 0.006299 R s - 0.0002045 (kcal/g dry weight per d). (24) The nitrogen content of a ration of D. brightwelli, i? N (mg), may be converted to kilocalories, R K (kcal), ac- cording to the following relationship R K = 0.06158 R s (25) (Durbin and Durbin 1981). Thus if the daily ration is expressed in kilocalories rather than nitrogen, the daily exogenous nitrogen excretion becomes E UK = 0.1023 R K - 0.0002045 (kcal/g dry weight per d). (26) Total daily nitrogen excretion, E K .— Combining Equations (21) and (23) we obtain an expression for the total daily nitrogen excretion rate E N = 0.616 R N + 0.237 (mg N/g dry weight perd). (27) Combining Equations (22) and (26), the daily ni- trogen excretion rate is expressed in calories 181 FISHERY BULLETIN: VOL. 81, NO. 2 E K = 0.1023 R K + 0.002423 (kcal/g dry weight per d). (28) where K B (log,, 10)7) (35) Since the total daily ration is given by R K — 0.079574 s c h (kcal/g dry weight per d), we can sub- stitute and obtain an expression for the total energy lost per day through nitrogen excretion, as a function of the foraging speed (s, cm/s) of the Atlantic men- haden, the concentration of food (c, kcal/1) and the foraging time (/?, h/d): E K = 0.008140 sc h + 0.002423 (kcal/g dry weight per d). (29) Growth Rate, G K and Gross Growth Efficiency, K x K Equations (12), (19), and (29) may be combined to provide an estimate of the daily growth rate, G K (kcal/g dry weight per d), as a function of menhaden foraging speed (s, cm/s), the concentration of plank- ton in the water (c, kcal/1), and the foraging time (h, h/ d), since G K = pR K — T K - E K (kcal/g dry weight per d) G K = h [0.06311 s c - 0.00958 (io 002948 s " 1 - 5342 ) + 0.000974] - 0.025803 (kcal/g dry weight per d). (30) The gross growth efficiency, K l is calculated ac- cording to K - G Thus K l in calories is equal to Equation (30) Equation (8) K\.K ~ (31) (32) From Equation (30) we can also determine the foraging speed which maximizes growth rate (s G 0PT ), for any given values of c and h. First restating Equa- tion (30) in a more general form, replacing the con- stants by A,B, C,D,E,J, andM, G K = h\Asc- B(10ids-e>) +J]-M (kcal/g dry weight per d). (33) We then differentiate Equation (30) with respect to s, i.e., set —j—= 0, and we find ds \og l0 K + E >6\OPT + l ~ log 10 c (34) In the present study where D. brightwelli is the food, s g ,opt= H9.4433 + 33.9213 log 10 c. (36) To determine the equation for the swimming speed which maximizes gross growth efficiency (s KO pt)> i-e., when = 0, we use the following general equa- ls tion: Ki - Equation (30) Equation (8) _ /? \Asc -fl(10'»>-^) +J] P s c h - M (37) where P is the constant in Equation (8), i.e., in the present example, P — 0.079574. We next define the new constants A' = *'=f B" = E (log,.10)L> '-$ v-£. i.nd thus K _B'10-«- 1 SC -J' M sc dK, r i^ont (38) (39) (40) (41) (42) (43) ds ^ -J'= (B"s -B') 10<*-fi> (44) This identity must be solved iteratively for s KOPT by using a given value of h and trial values of s. In the present example using D. brightwelli, we find 0.32426 0.01224 = (0.0081722 s - 0.12039) X 10<°-02948.s-1.5342) _ (45) Each term in the energy budget has now been defined in the same three variables: The foraging speed (s), the food concentration (c), and the foraging 182 DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN time (/?). Model I describes the potential interactions among these three variables, and their effects on menhaden energy intake, energy expenditure, growth, and growth efficiency. Energy Budget (Model II) Model II is a special case of Model I which incor- porates information on the swimming and feeding behavior of the Atlantic menhaden in response to plankton concentration. Laboratory observations have shown that Atlantic menhaden adjust their foraging speed according to the concentration of food in the water. When D. brightwelli was the food, the threshold concentration for the onset of feeding was about 1 jig chlorophyll a/1. Between about 1 and 4 ju.g chlorophyll a/1, the menhaden increased their forag- ing speed roughly in proportion to increasing plankton concentration. Above 4 /xg chlorophyll a/1, however, swimming speed remained nearly constant at about 4 1.3 cm/s(1.6 body lengths/s), independent of further increases in plankton concentration. Thus the relationship between the Atlantic menhaden foraging speed and Ditylum chlorophyll a (a, ju.g/1) was approximately asymptotic, where 29.62 (a - 1) 0.396 + (a - 1) + 12.2 (cm/s) (46) (Durbin et al. 1981). The equation includes the feed- ing threshold for Ditylum (1 jug chlorophyll a/1) and the routine (nonfeeding) swimming speed of the fish (12.2 cm/s), which represents the lower limit of the foraging speed. The chlorophyll a content of D. brightwelli may be converted to kilocalories according to the following relationship: 1 jug chlorophyll a = 6.06 X 10" 4 kcal . (47) Thus Equation (46) becomes 48,873 c - 29.62 s = 1,650 c - 0.604 + 12.2 (cm/s) (48) where c (kcal/1) is the plankton concentration. By substituting Equation (48) fors in Equations (8), (12), (19), (29), (30), and (32) for R K ,pR K , T K ,E K ,G K , and A', K , respectively, we are able to eliminate as a variable and rewrite the menhaden energy budget solely in terms of food concentration (c, kcal/1) and foraging time (h, h/d). This is Model II. Nitrogen Budget (Model I) The general equation for the nitrogen budget pre- sented in Equation (2) may be rewritten: G N = pR N — £ N (49) where p is the assimilation efficiency for nitrogen = 0.9240 (Durbin and Durbin 1981). The nitrogen budget is controlled by the same three variables as the energy budget: The foraging speed (s), the food concentrations (c or n), and the foraging time (h). The total dialy ration, R N (mg N/g dry weight per d), equals R s = 0.79574 s n h (mg N/g dry weight per d) (50) where n is the plankton concentration (mg N/1). The assimilated daily nitrogen ration, pR N , equals pR N = 0.073526 s n h (mg N/g dry weight perd). (51) The endogenous, exogenous, and total daily nitro- gen excretion rates, E b _ N , E f _ N , and E N (mg N/g dry weight per d) are presented in Equations (21), (23), and (27), respectively. Substituting Equation (27) into Equation (49), we obtain the following expression for the daily growth rate, G N : G N = 0.308 R N - 0.237 (mg N/g dry weight per d). (52) Gross growth efficiency, K lN , equals 0.308 i? N - 0.237 ^l.N R, (mg N/g dry weight per d) (53) where i? N is calculated according to Equation (50). If the ration is converted from units of nitrogen to kilocalories (Equation (25)), then Equations (52) and (53) become G N = 5.0016 R K - 0.237 (mg N/g dry weight per d) ^l.N — 5.0016 R k - 0.237 (54) (55) 16.239 R K where R K is calculated according to Equation (8). Nitrogen Budget (Model II) The empirical relationship between foraging speed, s (cm/s), and plankton concentration, a(pig/l) (Equa- 183 FISHERY BULLETIN: VOL. 81, NO. 2 tion (46)), is expressed in units of nitrogen through the following relationship: Thuss 1 jug chlorophyll a = 0.00984 mg N. (56) 3,010.2 n - 29.62 101.63 n - 0.604 + 12.2 cm/s (57) where n is the plankton concentration in mg N/1. The nitrogen budget can then be expressed solely in terms of plankton concentration, n (mg N/1), and foraging time, h (h/d), by substituting Equation (57) into Equations (50), (51), (27), (52), and (53) to com- pute R N ,pR N , E N , G N , and X liN (mg N/g dry weight per d), respectively If the food ration or the plankton concentration is expressed in kilocalories rather than units of ni- trogen, Equation (48) is substituted into Equation (8), and then Equation (8) into Equations (54) and (55) for the calculations of G N and /f liN> respectively. The Model II nitrogen budget, like energy budget II, is thus controlled by only two variables, c and h. RESULTS Energy Budget The energy budget is presented in two forms, a general model (Model I) and a special case of this model which incorporates information on the swim- ming and feeding behavior of the fish in response to plankton concentration (Model II). Model I, which defines the range of values which the energy budget could theoretically assume, is a function of the forag- ing speed (s), the concentration of plankton in the water (c) , and the foraging time (h ) . In Model II, forag- ing speed is a dependent function of plankton con- centration, and the energy budget is defined simply in terms of the variables c and h. Thus the two models describe the potential, and the actual, bioenergetic ranges within which the menhaden operate. In the following examples to illustrate the models, the variables s, c, and h assume values from to 50 cm/s, to 0.0090 kcal/1, and to 24 h/d, respectively, which should encompass the range of these variables in nature. In examples where s is assumed to be con- stant, a value of 41.3 cm/s was selected, because in the experiments this was the average foraging speed of the Atlantic menhaden at moderate to high plankton concentrations, where s was nearly independent of food level. Where c = constant, a value of 0.0030 kcal/1 was used, which is slightly above the threshold value of c at which s becomes food-concentration independent. We lack information on the foraging time of adult Atlantic menhaden in the wild. However, 184 since they feed continuously in the laboratory when food is present, when h — constant, we assigned it a value of 14 h, which is approximately equal to the number of daylight hours during the summer at the latitude of Narragansett Bay. In the experimental studies from which the budgets were derived, the variables s,c, and h took the follow- ing values: /i = 7h,c- 0.0010 to 0.0065 kcal/1, and s = 29.3 to 43.3 cm/s (1.1 to 1.7 body lengths/s). Within this relatively narrow range in foraging speed, the respiration rate increased from 2.2- to 5.4-fold over the routine rate. Slower foraging speeds (<29 cm/s) were observed during the transition period of declining phytoplankton concentration, after the in- put of food was terminated. The minimum foraging speed was greater than the routine swimming speed (12.2 cm/s), but was not closely determined in this study. The total ration ranged from 0.015 to 0.147 kcal/g dry weight, which corresponded to a feeding rate of 0.00217 to 0.02065 kcal/g dry weight per h. Using Model I we have described how foraging speed, food concentration, and the duration of feed- ing affect the menhaden energy budget (Fig. 1). InFigure 1, Al-A4,s increases, whilec and/7 remain constant. The total and the assimilated daily food in- take {R K and pR K ) increase linearly with increasing values of s (Fig. 1, Al). Among the energy expendi- ture terms, the exogenous nitrogen excretion (E f K ) increases linearly, the endogenous nitrogen excre- tion (E b K ) and the routine metabolic rate (T r A ) re- main constant, and the respiration during feeding {T f K ) increases exponentially with increasing s (Fig. 1, A2). Thus the assimilated daily ration increases linearly, whereas the total energy expended in- creases curvilinearly. If these two curves are drawn on the same axes, we find that they intersect twice, at a low and a high foraging speed (here, about 7 and 5 1 cm/s) (Fig. 1, A3). These intersections, where the en- ergy intake is balanced by the output and G = 0, define a range of foraging speeds within which the en- ergy intake exceeds expenditure, and positive growth takes place. At foraging speeds outside this range, the energy expenditures exceed the energy intake and the fish must draw upon stored energy reserves, thus undergoing negative growth. Within the defined range of foraging speeds, the growth curve (G K ) is convex upwards, increasing curvilinearly from zero to reach a maximum value at an intermediate swim- ming speed, then declining back to zero (Fig. 1, A4). The growth efficiency curve (K l K ) shows a similar pattern, but reaches its maximum value at a different foraging speed than that for maximum growth. InFigure 1, Bl-B4,c increases, whiles and/j remain constant. The energy intake (R and pR) increases DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN UJ < > cr UJ z UJ Z3 a. O S0.3 0.2 - 0.1 0.3 0.2 0.1 c = 0.0030 kcol/i h=l4 hours /doy F K □ PR K E3 ® 5 = 41.3 cm/sec h = 14 hours/day % Body kcal/doy (A4) " / 40 S K,0PT c=0.0030 kcal/* @ s=4l.3 cm/sec ^■- , '>.". l ';r'-"' K 0.6 0.4 0.2 >- o z UJ o u. u. Ul 5 o cr o c/j 10 o or 20 40 FORAGING SPEED (s, cm/sec) 0.0030 0.0060 PLANKTON CONCENTRATION (c, kcal/£) 8 16 FORAGING TIME (h, hours/day) 24 Figure 1. — Model I energy budget for the Atlantic menhaden at 20 C, where: A, foraging speed) s) increases, while plankton concentration (c) and foraging time (h) remain constant; B, plankton concentration increases, while foraging speed and foraging time remain constant: and C. foraging time increases, while foraging speed and plankton concentration remain constant. Al, Bl, and CI represent energy intake (R K and pR K ); A2, B2, and C2, the energy output (T r K , T, K , E K ); A3, B3.and C3 compare the intake and output of energy and show the surplus energy which is available for growth: A4, B4, and C4 illustrate growth and gross growth efficiency. 185 FISHERY BULLETIN: VOL. 81. NO. 2 linearly with increasing c (Fig. 1, Bl). The energy ex- penditure to exogenous nitrogen excretion (E f K ) also increases linearly, whereas E h N and respiration (T r K and T f K ) are constant (Fig. 1, B2). The curves rep- resenting energy intake and expenditure both in- crease linearly with increasing values of c (Fig. 1,B3), and thus growth (G K ) increases linearly and gross growth efficiency increases asymptotically (Fig. 1, B4). InFigure 1, Cl-C4,/i increases, whiles andc remain constant. Here, also, the energy intake (R andpR) in- creases linearly with increasing h (Fig. 1,C1). The en- ergy expenditure to endogenous nitrogen excretion (E b K ) remains constant, while exogenous nitrogen excretion (E fJi ) and the respiration during feeding (T /A ) increase linearly, and the routine respiration (T r K ) declines linearly (Fig. 1, C2). The curves describing the energy intake and expenditure increase linearly with increasing values of h, (Fig. 1, C3), and again we find that growth (G K ) increases linearly, and gross growth efficiency (K" 1-A -) increases asymptotically (Fig. 1, C4) These examples demonstrate that in order for an Atlantic menhaden, which forages ats cm/s for/? h/d, to obtain a maintenance ration, the concentration of food must equal a minimum threshold value, c min (i.e., 0.0021 kcal/1 in Fig. 1, B3-B4). Similarly, a menhaden foraging ats cm/s when the plankton con- centration = c kcal/1, must feed for some minimum period h mm (in Fig. 1, C3 and C4; 6.2 h/d) in order to obtain a maintenance ration. There will also be a minimum foraging speed, s min , required to obtain a maintenance ration for each combination of c and h (in Fig. 1, A3 and A4; 7.0 cm/s). If growth is to occur, s, c, and h must exceed s mm , c^ n , and h min . The general rule is that for any swimming speed (s), the more abundant the food, the smaller the maintenance ra- tion, and the shorter the feeding time required to ob- tain the ration (Fig. 2, A, B). If an Atlantic menhaden forages at 41.3 cm/s, for example, the lowest concen- tration of Ditylum at which it could obtain a main- tenance ration would be about 0.0018 kcal/1, assuming that it fed for 24 h/d. The maintenance ra- tion would be about 0.143 kcal/g dry weight per d. With an increase in plankton concentration, the re- quired feeding time and the maintenance ration decline very rapidly, reaching 4 h/d and 0.05 1 kcal/g dry weight per d at c = 0.0039 kcal/1, and declining more slowly thereafter to 1 .3 h/d and 0.038 kcal/g dry weight per d at c — 0.009 kcal/1. An interesting feature of the energy budget is that for any combination of c and/?, there is a single forag- ing speed which will maximize the growth rate(s (; 0PT ) (Fig. 1, A4). Similarly, growth efficiency reaches its 186 UJ j 12 < ^ 04 s s = Observed s = 4i3 cm/sec _i i I I i_ s : Observed s= 41 3 cm/sec 00006 00018 00030 00042 00054 00066 00078 00090 PLANKTON CONCENTRATION (c.kcol//) FIGURE 2. — A, relationship between the concentration of plankton and the maintenance ration of Atlantic menhaden which are assumed to swim at a constant speed of 41.3 cm/s (Model I) and at their actual speeds in response to plankton concentration (Model II). B, the foraging time required for the Atlantic menhaden to obtain a maintenance ration at different concentrations of plankton, assum- ing that they swim at 41.3 cm/s (Model I) or at the actual speed which has been observed in the laboratory (Model II). maximum value at a unique foraging speed (s A0PT ), which is always less thans (; 0PT . s G 0PT increases cur- vilinearly with increasing food concentration (Fig. 3), but is independent of the duration of feeding (Fig. 4, Equation (36)). In contrast, s AOPT declines as the duration of feeding increases (Fig. 4), but is indepen- dent of food concentration (Fig. 3, Equation (45)). It should be remembered however that the values of G. and K UK when the fish swim at s G 0PT and s, 'K 'A, OPT are determined by both c and /?. For example, if c — 0.0030 kcal/1, a fish will maximize its growth rate if it swims at 33.9 cm/s although the actual rate of growth 0030 00060 0090 PLANKTON CONCENTRATION (c. kcal/Z) FIGURE 3. — The relationship between plankton concentration and the foraging speed which maximizes the Atlantic menhaden's growth rate (s L; opt'- s G OPT ' s independent of foraging time (h). DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN 4 8 12 16 20 FORAGING TIME (h, hours/day) FIGURE 4.— The relationship between foraging time and the foraging speed which maximizes the Atlantic menhaden's gross growth efficiency (s K 0FT ). s(s K 0PT ) is independent of plankton concentra- tion (c). depends onh. Similarly, a fish feeding for 14 h/d will maximize its growth efficiency if it swims at 23.8 cm/ s; however the resulting values of K Y will depend on c. The foregoing examples demonstrate that the rela- tive size of each component in the energy budget {R K , pR K , T K , E K , and G K ) will vary according to the values of s, c, and h. Since the different elements retain no fixed proportions within the overall energy balance, there is no single "standard" energy budget which can be described for the Atlantic menhaden. It can also be seen that in Model I, a change in either food concentration or the duration of feeding has a direct, proportional effect on the growth rate, because total energy intake and expenditure are linear functions of c and h, when s = constant. However, a change ins has a nonlinear impact on the growth rate. This is because the respiration rate is an exponential function of swimming speed, and thus a change in swimming speed causes a proportional change in energy intake but a more-than-pro- portional change in total energy output. In the Model II energy budget, s is no longer an in- dependent variable, but is a dependent function of food concentration c, according to the experimental- ly derived relationship in Equation (48). The foraging speed is nearly constant at moderate to high concen- trations, but is reduced at low plankton abundance. The threshold concentration (0.0006 kcal/1) at which the fish stop feeding on Ditylum is also included in this model. The effect of reducing the foraging speed, when plankton concentration is low, is illustrated in Figure 5, which provides a comparison of Model LI with Model I, where s = constant =41.3 cm/s. (This foraging speed was chosen for the Model I example because it provides the best overall fit to Model II, facilitating the comparison between the two. The choice of another value for s would cause Model I to depart further from the actual behavior of the fish and would increase the difference between the two models.) In Model I, we found that when s and h were con- stant, the curves describing R K ,pR K , T K , E K , and G K as a function of increasing c were all linear or constant (Fig. 1, B1-B4; Fig. 5, A1-A4). In Model II, these curves are nearly linear or constant at moderate to high plankton concentrations, where s ~ constant. How- ever, they become increasingly curvilinear at lower concentrations, when s is changing rapidly (Fig. 5, B 1 -B4). Thus we find that Model II is quite similar to Model I where s = 41.3 cm/s, when the food concen- tration is above c mm in the Model I example (-0.002 1 kcal/1 for h = 14 h/d). The models diverge signifi- cantly as c declines below c min . If the Atlantic menhaden were to continue to swim at their "preferred" speed when the plankton concentration is low, a significant deficit in the energy budget would result (Fig. 5, A3). However, Model II shows that by reducing their foraging speed when food concentra- tion is low, the Atlantic menhaden act to regulate their energy expenditure to remain close to their rate of energy uptake (Fig. 5, B3). Reducing the foraging speed has this effect, because of the exponential relationship between respiration and swimming speed. A reduction in foraging speed causes the res- piration term to decline more rapidly than the inges- tion term. The resulting change in the energy balance enables the fish to obtain a maintenance ration in less time, and at a lower concentration of food, than would have been possible had they continued to forage at the higher speed. The growth rate and growth ef- ficiency are thereby enhanced at low concentrations (compare Fig. 5, A4 and B4). This effect can also be seen in Figure 2. At the threshold concentration (0.0006 kcal/1) where Atlantic menhaden cease feeding on Ditylum, it can be seen (Fig. 5, Bl and B2) that the routine metabolic costs alone are greater than the energy which could be derived from feeding. The behavior of the fish apparently reflects the fact that it is not bioenergetically profitable to feed at such a low plankton density. Nitrogen Budget In the nitrogen budget there are three loss terms: The endogenous excretion, which is a constant, and the exogenous excretion and the fecal losses, which are proportional to the nitrogen content of the daily ration. The remaining nitrogen from the ration is retained as growth. Thus we find that the nitrogen 187 Model I FISHERY BULLETIN: VOL. 81, NO. 2 Model H Figure 5.— A comparison of the Model I energy budget, where foraging speed and foraging time are constant while plankton concentration increases, with the Model II budget incorporating the actual volun- tary swimming speed of the Atlantic menhaden at each concentration of plank- ton. Panels numbered 1, 2, 3, 4 are as in Figure 1. 0.0030 0.0060 0.0030 0.0060 0.0090 PLANKTON CONCENTRATION (c.kcal/i) budget, though functionally simpler than the energy budget, is controlled by the same three variables: The foraging speed (s, cm/s), the concentration of food (c, kcal or mg N/1), and the foraging time {h, h/d). In the Model I nitrogen budget, R N , pR N , E fN , and G N all increase linearly, E bN remains constant, and K 1N increases asymptotically with increasing values of s, c, and h (Fig. 6). However, as we found in the energy budget, these curves in the Model II nitrogen budget are nearly linear at plankton concentrations suf- ficiently high that s ~ constant, but become in- creasingly curvilinear at low plankton concentrations because of the decline in the foraging speed (Fig. 7). Here, also, the reduction in foraging speed enables the Atlantic menhaden to obtain a maintenance ra- tion in less time and at a lower concentration of 188 plankton, and to increase their growth rate and growth efficiency, relative to the case in Model I where foraging speed was assumed to remain con- stant at 41.3 cm/s. The nitrogen and energy budgets differ in some im- portant ways. First, we have seen that in the energy budget, with an increase in swimming speed (s), the growth rate and growth efficiency increase from zero, reach a maximum, then decline back to zero (Fig. 1, A4). However, in the nitrogen budget, growth in ni- trogen increases linearly (i.e., indefinitely), and growth efficiency increases asymptotically (Fig. 6, A4) with increasing swimming speed. Second, for any given s, c, and h, the predicted growth efficiency in calories is usually significantly different from that in nitrogen (Fig. 8). Figure 8 shows that differences ex- DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN 8 20 40 00030 00060 FORAGING SPEED PLANKTON CONCENTRATION (c, kcol//) 8 16 FORAGING TIME (h, hours/day ) (s, cm/sec) FIGURE 6. — Model I nitrogen budget for the Atlantic menhaden. Panels are as in Figure 1. ist not only in the maximum or asymptotic values of the two growth efficiencies, but in the x-intercepts of the curves as well. The x-intercepts are of particular interest since they define the minimum requirements ( s min» c min» ^min) f° r the fish to obtain a maintenance ra- tion in nitrogen or calories. In Figure 8 A, for instance, where c = 0.0030 kcal/1 and h = 14 h/d, s min K is respectively. Thus in these examples the menhaden would show positive growth in ni- trogen at a lower food concentration, and with a short- er foraging time, than they would in calories. Atlantic menhaden can exercise direct control over the variables s and h, but they may or may not have any impact on the environmental variable c. Thus it is of interest to consider how a change in the values of s and h will affect the minimum plankton concentra- tion required for the Atlantic menhaden to obtain a maintenance ration in calories and nitrogen. The curve in Figure 9, calculated from the Model I budget, shows the combinations of s and h at which c min K = c min,N- For all combinations of s and h, which fall below 189 FISHERY BULLETIN: VOL. 81, NO. 2 Model I Model H 0.4 - UJ 0.3 < 1- z > 0.2 o cc l±J z UJ 0.1 0.3 K => n h- 3 0.2 O > ID >< o QC ■60.1 LU \ Z *— UJ i >> •5 o> \ >- °0.3 o o J£ or ^_ UJ z UJ 0.2 1- Ul z s = 4l.3 cm/sec h= 14 hour/day i • '- 1 - - 'f.K r,K _ L K 0.3 0.2 o or 0.1 % Body kcal/Day (A4) - 4.0 - 2.4 0.8 s = Observed h=l4 hours/day - 1 ' I- 1.'- tl*"". 1 " * »%2^i "'"*.Cr **?*,' 0.0030 0.0060 0.0030 0.0060 0.0090 PLANKTON CONCENTRATION (c.kcal/*) o z 0.6 w o Ul 0.4 x I- =S o 0.2 g CO CO o or o FIGURE 7.— Comparison of the Model I nitrogen budget where foraging speed and foraging time remain constant while plankton concentration increases, with the Model II budget incorporating the observed swimming speeds of the Atlantic menhaden at each plankton concentration. 190 DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN c =00030 kcol/i h =14 hours/doy S=4I 3 cm/sec h = 14 hours/doy c=00030 kcol// S = 4I3 cm/sec 20 40 FORAGING SPEED (s,cm/s«c) 00030 00060 PLANKTON CONCENTRATION (Ckcol/O 8 16 FORAGING TIME (h, hours/doy) FIGURE 8. — Comparison of the gross growth efficiency of the Atlan- tic menhaden in calories (K, K ) and nitrogen (K, »•) where A, forag- L 1.K l.N' ing speed(s) increases while plankton concentration (c) and foraging time (h) remain constant; B, plankton concentration increases while s and h are constant; and C, foraging time increases while s and c are constant. z <^ „ < 4 cc O c min, N ' c min, K c min, K L min, N 10 20 30 40 50 FORAGING SPEED (s, cm/sec) 60 FlGl'RE 9. — Boundary curve defining the combinations of foraging time and foraging speed at which the minimum plankton concentra- tion required for the Atlantic menhaden's growth in calories (c min K ) is less than, and greater than, that required for growth in nitrogen ( c min,.\)- this boundary, c mm A -will be lower than c min N . Atlantic menhaden will be able to grow in calories at a lower food concentration than they can in nitrogen. Con- versely, where s and h are greater than the boundary values, c mmN will be lower than c mmK . Atlantic menhaden can grow in nitrogen at a lower food con- centration than they can in calories. Next we consider how the actual foraging speeds of the Atlantic menhaden compare with the boundary curve in Figure 9. Figure 10A shows the foraging speed in relation to food concentration. Figure 10B shows that for all values of h up to 24 h/d, Atlantic menhaden forage at speeds such that their minimum food requirement for growth in nitrogen is lower than for calories, i.e., c min _ N is A -. Thus at low plankton concentrations, the growth efficiency in nitrogen is greater than in calories. However, it can be seen from Figure 9 that.K\ N remains > K 1K only over a narrow range of food concentrations immediately above c mm. v #1, k increases very rapidly above c mui K and soon overtakes K lN . Thus in most circumstances where the fish are growing, growth efficiency in calories will be considerably higher than in nitro- gen. DISCUSSION Functioning of the Energy and Nitrogen Budgets These models permit a detailed analysis of the en- ergetics of the Atlantic menhaden, by showing how energy intake (ingestion), as well as energy losses and expenditures (feces, excretion, respiration) vary with the concentration and size of the food particles, the foraging speed of the fish, and the duration of feed- ing. These different components of the model, and the predicted growth rate and growth efficiency, are dis- cussed in more detail below. Energy Intake VOLUME SEARCHED.— The volume searched by the Atlantic menhaden can be described in very z ^. 1 • p & li- o Q V> UJ ~~ UJ UJ £ Q- m in O 50 30 1— vJT 0.6 *5 o tr > o o z 0.4 o SF o UJ 0.2 - A ' f t i i i i i i \< P K I,K K, N ^ 24 — x^\^^ 5 / / / 2A - -vA" "/ ----"~ 5 /I /I / *»**""" / / / /y ' Iff 1 \/f i i i i 0.0030 0.0060 00090 PLANKTON CONCENTRATION FIGURE 10.— A, laboratory defined relationship between Atlantic menhaden voluntary foraging speed and plankton concentration. B, gross growth efficiency of menhaden which forage at these swim- ming speeds for different periods of time. 191 FISHERY BULLETIN: VOL. 81. NO. 2 simple terms, i.e., an ellipsoid with cross-sectional area equal to that of the fish's open mouth, and length equal to the distance travelled by the fish per unit time. The volume searched is equal for all types of prey. With other species of filter-feeding fishes, a slight modification of this basic formula may be necessary, according to the mode of feeding. For ex- ample, a number of species (northern anchovy, Leong and O'Connell 1969; alewife, Janssen 1978; gizzard shad, Drenner et al. 1978) are described as rhythmically opening and closing the mouth during feeding, apparently producing a suction which draws in particles located outside the perimeter of the mouth. Here, the cross-sectional area of the volume searched is somewhat larger than the mouth area; also, a correction factor is needed to account for the proportion of the time the fish's mouth is closed and not actually filtering. Nevertheless, the basic sim- plicity of the volume searched by a filter feeder is in marked contrast to the case of a predatory fish or par- ticulate planktivore. Since these fishes visually lo- cate and capture their prey, the volume searched is complex and depends on a variety of factors, includ- ing the visual capacity and adaptations of the fish, the inherent visibility and behavioral characteristics of the prey, and the nature of the underwater visual en- vironment (quantity and quality of the illumination, clarity of the water). Thus the volume searched by a particulate feeder is different for different types of prey, and even if a fish were to swim at constant speed and feed on a single prey type, the volume searched will continually change according to variables such as the time of day, and the depth at which the fish swims (Durbin 1979). FORAGING SPEED.— Foraging speed affects both the energy intake and expenditure terms in the energy budget, but only the energy intake in the ni- trogen budget. Foraging speed is the principal deter- minant of the volume searched for food, since the cross-sectional area of the mouth in an Atlantic menhaden of a given size is constant. Foraging speed in the Atlantic menhaden increases asymptotically with increasing food concentration. Because of this there will be two critical levels of abundance for each prey species: c t , the threshold concentration at which the menhaden are stimulated to feed, and t\, the con- centration at which foraging speed becomes approx- imately independent of food concentration. With Ditylum, the value of c, was about 4.5 jxg chlorophyll a/1 (0.0027 kcal/1), and the fish swam at an average speed of 41.3 cm/s. From Figure 2 it is seen that when c = 0.0027 kcal/1, the fish swimming at 41.3 cm/s would obtain a maintenance ration in slightly more than 7 h. At higher food concentrations the required feeding time would be much less, generally <4 h. These results suggest that Atlantic menhaden feed- ing onDitylum will swim at their "preferred" speed as long as the concentration is sufficiently high to en- able the fish to meet their daily energy requirements in <8 h of feeding. At lower food concentrations the fish conserve energy by swimming more slowly dur- ing feeding. Whether these results are fortuitous and apply only to Ditylum, or instead imply a funda- mental relationship between foraging speed and foraging time which is applicable to different food types, cannot be determined from present informa- tion. FILTRATION EFFICIENCY.— The effective vol- ume searched will be determined by the filtration ef- ficiency {e). As described earlier (Equation (6)) filtration efficiency is fairly high for zooplankton- sized particles, but in the range of phytoplankton- sized particles declines sharply to a minimum size threshold of about 1 3 jum. This means that the Atlan- tic menhaden cannot directly exploit the < 20 ju.m size fraction of phytoplankton, which forms the greater part of the total phytoplankton biomass on their sum- mer feeding grounds (Durbin et al. 1975). Menhaden exploit this food resource indirectly, however, by feeding upon the zooplankton. ASSIMILATION EFFICIENCY.— The efficiency with which food is assimilated further modifies the energy intake by the Atlantic menhaden and will af- fect the predicted growth rate and growth efficiency in the model. If assimilation changes with different meal sizes or rates of feeding, then the proportion of ingested energy which is available for metabolism and growth will also change. Most investigators have found that assimilation efficiency is independent of ration size (Gerking 1955; Menzel 1960; Pandian 1967; Birkett 1969; Iwata 1970; Beamish 1972; Kelso 1972; Staples and Nomura 1976). However, Elliott (1976) and Solomon and Brafield (1972) found a slight decrease in assimilation efficiency as meal size increased. (In the latter study the authors suggest that the change may have been an artifact arising from the incomplete recovery of a small amount of fecal material in the tank.) For the Atlantic menhaden we assumed a constant assimilation ef- ficiency with different ration sizes. The mean assimilation efficiencies observed for the Atlantic menhaden feeding on phytoplankton were quite high (86.4% for carbon, 92.4% for nitrogen, and 89.5% for calories). For Atlantic menhaden feeding on zooplankton the values were similarly high (86.7, 192 DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN 91.3, and 87.7%, respectively). The high values for zooplankton were consistent with results from other fishes (Gerking 1955; Pandian 1967; Beamish 1972; Kelso 1972). Few measurements of carbon, nitrogen, or caloric assimilation exist for marine herbivorous fishes. Menzel (1959) found that Holacanthus as- similated 85% of the nitrogen and 77.7% of the calories from two species of macroalgae. The lower assimilation inHolacanthus may have been related to the type of food. However, there do not appear to be any comparable studies with marine phytophageous fishes, which would indicate whether the high as- similation efficiency of the Atlantic menhaden is typical of this trophic group. Energy Losses RESPIRATION.— The major energy outputs by the Atlantic menhaden are respiration and excre- tion. Respiration by the menhaden was divided into feeding (T fK ) and nonfeeding components {T rK ). SDA was not included as a separate component, but for reasons discussed earlier was included as part of the feeding respiration rate. SDA is thought to be a fixed proportion of the energy content of the food ra- tion, and in carnivorous fishes has been estimated at about 12.7-16% (Muir and Niimi 1972; Beamish 1974; Pierce and Wissing 1974; Schalles and Wis sing 1976). Partitioning T fK into its components, T sK and T SDA K , would have caused some minor changes within the energy budget, but would not have sig- nificantly affected the predictions of growth rate and growth efficiency. The most important change would be in a case analagous to Figure IB, where food con- centration increases while s and h remain constant. Here, the ingested ration automatically increases in proportion to c because Atlantic menhaden filter a constant proportion of particles from the water. T fK in this illustration is constant, which reflects the fact that its major component T sK is constant. However, if SDA were included separately we would actually ex- pect to see a small linear increase in T fK because T SDA k should presumably increase in proportion to the ration R K . For Atlantic menhaden, the metabolic cost of feed- ing appears to be high (Durbin et al. 1981). This is because of the very rapid increase in respiration rate per unit increase in foraging speed. This rate of in- crease was about 2.5 times greater than has been ob- served in other (nonfilter feeding) species during forced long-term swimming (Beamish 1978). Thus even minor changes in the foraging speed can have a significant impact on metabolic expenditures and the overall energy balance. The energy budget demonstrates that for an active species such as the menhaden, it is not possible to use a constant multiplier of the standard metabolism, as recommended by Winberg (1956), to estimate meta- bolic expenditures in the field. Not only is the suggested multiplier of 2 times the standard rate too low (in our studies the routine rate was 3.4 times the estimated standard rate, and the average feeding rate 2.3-4.8 times routine, or about 8-17 times stan- dard), but also the relative size of the respiration component within the overall energy budget is also a variable, changing according to the values of s, c, and h. EXCRETION.— Excretion, the other major energy output, is similarly a variable. In contrast to respira- tion which depends on swimming speed and foraging time, excretion depends on the amount of food eaten. Excretion, therefore, will follow no constant relation- ship to respiration in the energy budget (Model I). The linear relationship between ration size and ex- ogenous nitrogen excretion is similar to results in other studies (Gerking 1971; Savitz et al. 1977), although the proportion of nitrogen excreted will de- pend on the balance of amino acids in the food rela- tive to the requirements of the fish. Growth Rate and Growth Efficiency The rates of energy intake and expenditure deter- mine the amount of energy which is available for growth. Atlantic menhaden must invest considerable time and energy in feeding. The Model I energy and nitrogen budgets show that if foraging speed remains constant, then growth will increase linearly with in- creasing ration size, regardless of whether this is brought about by an increase in food concentration or foraging time. Consequently, gross growth efficiency increases asymptotically with increasing ration size. Model II demonstrates that given the actual swim- ming behavior of the menhaden, the relationship be- tween ration size and growth is in fact very nearly linear at moderate-high plankton densities were s ~ constant, but becomes significantly curvilinear at lower plankton levels because of the decreasing for- aging speed. With the reduction in foraging speed, the energy balance changes because proportionally less of the ingested ration is used to support meta- bolism, which leaves more energy available for growth. Ivlev's (1960) bioenergetic model of the bleak, Alburnus alburnus, showed that in this particulate- feeding planktivore, growth increased asymptotically, rather than linearly, with increasing food concen- 193 FISHERY BULLETIN: VOL. 81, NO. 2 tration. These results reflect basic differences in the ingestion process between filter- and particulate- feeding planktivores. Since a filter feeder like the Atlantic menhaden removes a constant proportion of the particles in the water per unit of time, without the necessity to capture and handle each item of prey in- dividually, the ingestion rate increases linearly with increasing food concentration and swimming speed. In contrast, with the particulate planktivore, feeding is a series of discrete events and there will be a max- imum ingestion rate set by the time required to cap- ture and handle each prey. Thus, as Ivlev has shown experimentally (Ivlev 1960, 1961), ingestion rate in- creases asymptotically with increasing food concen- tration. This causes an asymptotic growth curve. There does not appear to be any information avail- able to describe the ingestion pattern of a particulate feeder as a function of swimming speed. However, based on Holling's predation model (Holling 1966), an increase in the swimming speed of a particulate planktivore will increase the encounter frequency and hence the feeding rate. Based on this model we could expect that with increasing swimming speed, the ingestion rate will increase asymptotically towards a maximum rate set by the handling time. In most laboratory studies of the relation between feeding and growth, the fish are given a fixed ration for a specified period, after which the amount of growth is determined. The food is made readily avail- able to the fish, and hence the time and energy expended for feeding is presumably small. In the majority of these studies, growth was linearly related to ration size, which implies that assimilation efficien- cy and the increment in metabolism and growth per unit of ration remained constant at all ration levels (Pandian 1967; Birkett 1969; Gerking 1971; Jones and Hislop 1972, 1978; Niimi and Beamish 1974; Staples and Nomura 1976; Stirling 1977). Where reported, growth efficiency increased asymptotically with increasing ration size; this is a consequence of the observed linear growth-ration relation. In several studies the relationship between growth and meal size appeared to be slightly curvilinear, however, with the growth rate somewhat depressed at high rations (Carline and Hall 1973; Elliott 1975; Wurtsbaugh and Davis 1977). Under these con- ditions, growth efficiency increased curvilinearly from zero at the maintenance ration to a maximum value, and thereafter declined curvilinearly. Warren and Doudoroff (1971) suggested that such a phe- nomenon could be caused either by a reduction in assimilation efficiency at high rations, or by a change in the energy balance within the fish, in which the metabolic component increased (higher SDA, or greater spontaneous activity) at the expense of the energy available for growth. Another possible cause of departure from linearity could arise from changes in the wet weight: dry weight ratios (Staples and Nomura 1976). These investigators found that fish at high ration levels increased in percent of dry weight relative to fish on low rations. Thus measurements of growth based on wet weight will overestimate the true growth of fish at low rations, and underestimate growth at high rations, which can lead to an apparent curvilinearity in the growth-ration relationship. The growth of sockeye salmon on fixed rations in- creased nearly linearly with increasing ration size, in keeping with results from other similar studies (Brett et al. 1969; Brett and Shelbourn 1975). However the latter investigators found that if they included growth data from fish fed "excess rations," where voluntary food intake continually declined as the fish grew, the overall relationship between growth and increasing ration size was asymptotic, making the growth ef- ficiency curve convex upwards. The Model I prediction of a linear relation between ration size and growth in the Atlantic menhaden, when swimming speed is constant (i.e., activity = constant), and the slight departure from linearity by Model II, is therefore supported by most experimen- tal studies of feeding and growth in other fish species. It should be noted that if assimilation efficiency in Atlantic menhaden were to decline at high feeding rates beyond the range of the experimental data, we would expect that growth rate will approach an asymptote, and growth efficiency will decline with further increases in ration size. However since the ex- periments covered the range of plankton concen- trations which the fish might be expected to encounter in nature (Durbin and Durbin 1981), the possible decline in assimilation at very high feeding rates would not appear to be meaningful for Atlantic menhaden under most circumstances in the wild. It should also be noted that since the foraging costs of obtaining a ration of a particular size will vary ac- cording to s, c, and h, there will not be a single (unique) relationship between ration size, growth rate, and growth efficiency in Atlantic menhaden. The models predict that over most of the range of plankton concentrations where growth is possible, growth efficiency will be higher for calories than for nitrogen. These findings are consistent with field ob- servations that the fat and caloric composition of the menhaden increases relative to protein during its season of growth (Dahlberg 1969; Dubrow et al. 1 976) . At low plankton concentrations the fish forage at speeds such that growth in nitrogen is possible even when there is an overall net energy deficit. This 194 DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN. suggests that protein is conserved when food levels are low. It should also be noted that since the foraging costs of obtaining a ration of a particular size will vary according to s, c, and h, there will not be a single unique) relationship between ration size, growth rate and growth efficiency in Atlantic menhaden. The models predict that over most of the range of plankton concentrations where growth is possible, growth efficiency will be higher for calories than for nitrogen. These findings are consistent with field observations that the fat and caloric composition of the menhaden increases relative to protein during its season of growth (Dahlberg 1969; Dubrow et al. 1976). At low plankton concentrations the fish forage at speeds such that growth in nitrogen is possible even when there is an overall net energy deficit. This suggests that protein is conserved when food levels are low. Optimal Foraging by Planktivores In a landmark study, Ware (1975) combined Ivlev's (1960) data on Alburnus with Holling's (1966) preda- tion model to develop a bioenergetic model of this particulate planktivore, which could be used to test different theories of optimal foraging. Ware was the first to demonstrate the existence of s G0PT and s K0PT , and showed the importance of swimming speed in determining the energy balance within the fish. His analysis demonstrated that the swimming speeds of fish in nature can be extremely useful and sensitive indicators of how different species respond to and exploit changes in their food resource. An in- teresting feature of Ware's (1975) model of a particu- late planktivore was that as c increased, s G0PT curvilinearly increased to a maximum at a single food concentration, and thereafter declined, whereas s kopt declined monotonically with increasing values of c. These changes in s G 0PT and s K 0PT were due to the effect of handling time on the rate of ingestion in the Holling (1966) model. In contrast the present study, which extends Ware's concepts of s G0PT and s K 0PT to a filter feeder, shows that since handling time is negligible in a filter feeder, s G0PT increases asymptotically with increasing values of c, whereas s K 0PT is solely a function of h and independent of c. It is interesting that for both particulate and filter- feeding planktivores, distinct foraging strategies are required in order to achieve maximal growth rate or growth efficiency. The experimental data from the Atlantic menhaden make it possible to determine whether the foraging behavior of this species is directed towards enhanc- ing some measure of ecological fitness such as growth rate or growth efficiency. This may be done by com- paring the growth rates and growth efficiencies calculated for the observed swimming speeds of the menhaden with those that would result if the fish were to swim at speeds equivalent to either s GO pt or s k,opt- The comparison is made with s G 0PT in Figure 11 for the case where h = 14 h/d and with s A'.OPT m Figure 12 for the case where c = 0.0030 kcal/1. Figure 1 1 demonstrates that the growth of Atlantic menhaden which swim according to the laboratory derived relationship in Equation (48) is very close to the maximum possible growth at each concentration of plankton. This suggests that foraging speed in the adult Atlantic menhaden is a behavioral adaptation to maximize growth rate. In contrast, at any given concentration of food the observed foraging speed was always >s K0PT , which resulted in submaximal values of K x K (Fig. 12). This is evidence that the fish were not acting to maximize growth efficiency. To maximize growth efficiency the fish would have had to regulate their foraging speed according to the duration of feeding. This was not ob- served in Atlantic menhaden in the laboratory, where foraging speed at a given concentration of food remained constant for periods of up to 7 h. Further, we have shown that foraging strategies which regu- late swimming speed in order to maximize growth rate and growth efficiency are mutually exclusive. Figures 4 and 11 provide an explanation for the hyperbolic nature of the plankton concentration- foraging speed relationships in Equation (48). s G0PT changes most rapidly at low concentrations of plank- ton, and it is in this region where Atlantic menhaden most strongly regulate their foraging speed, s G 0PT o a. h= 14 hours/day — S = S G.OPT s = observed s = 30cm/sec s= 50 cm/sec O 00006 O.OOI8 00030 0.0012 00054 00066 00078 00090 PLANKTON CONCENTRATION (c.kcal//) FIGURE 11.— A comparison of the growth of the Atlantic menhaden at different concentrations of plankton, when the fish swim accord- ing to s G 0PT ; their actual voluntary speeds; and constant speeds of 30 and 50 cm/s. Foraging time is 14 h/d. 195 FISHERY BULLETIN: VOL. 81, NO. 2 changes less rapidly at moderate-high plankton abundance, and in fact the constant preferred speed of the Atlantic menhaden (41.3 cm/s) is sufficiently close to s G 0PT that growth remains nearly maximal over a very broad range of plankton abundance. Thus there is no great "penalty" if the fish swim at constant speed rather than exactly at s G 0PT within this region of the curve. The choice of this preferred speed is fairly exacting, however. As can be seen in Figure 11, at speeds not greatly different from 41 cm/s (30 and 50 cm/s), growth will be suboptimal over much of the plankton concentration range. How much of a sacrifice in growth efficiency is im- plied if the fish swim at s COPT ? Figure 12 indicates that K x K , though suboptimal, is still reasonably high when the fish swim ats G 0PT . However, as the for- aging speed increases above s G0PT , there is an in- creasingly rapid decline in K 1 K , as can be seen in Figure 12 where s = constant = 50 cm/s. In conclusion, the present results, which demon- strate a very close agreement between the predicted relationship between s G 0PT and food concentration, and the observed relationship between foraging speed and c, indicate that the foraging speeds of the adult Atlantic menhaden have evolved over time towards maximizing growth rate. This optimization of growth rate has necessarily resulted in a submax- imal growth efficiency. In his analysis of data for the bleak, Ware (1975) showed that the observed forag- ing speed when c ~ 0.000808 kcal/1 was also quite close to the value of s G 0PT predicted from his model. However, there was insufficient information in Ivlev's (1960) original study to indicate whether the bleak adjusts its foraging speed to remain near s G 0PT at different plankton concentrations. Studies y 05 o U. 04 UJ X 0.3 I- 3 o <£ 02 e> in o i o cr c =0.003 kcal ll s ° 5 K.0PT s = 30 cm/sec s = observed —-""'' s = 50 cm/sec ^ <*- *~ s' . — ' ' ^- ■" ** ..^ s .• — ''' ___— — jy ,,- " P ^~~ f ^ — ' / ^^ y /^ / s' / / 4 8 12 16 20 24 FORAGING TIME (h, hours/day) FIGURE 12. — A comparison of the gross growth efficiency of the Atlantic menhaden as a function of foraging time, when the fish swim according to s K qpt ; their actual voluntary speeds; and constant speeds of 30 and 50 cm/s. Plankton concentration is 0.0030 kcal/1. demonstrating selective feeding in planktivores (e.g., Brooks 1968; Leong and O'Connell 1969; O'Connell 1972; Werner 1974; Werner and Hall 1974; O'Brien et al. 1976; Eggers 1977; Confer et al. 1978) indicate that foraging strategies, which result in the max- imization of energy intake, may be a more general phenomenon among these fishes. However, it should be pointed out that these feeding studies only con- sider energy intake and not energy expenditures, so that the extent to which these fishes are following op- timal strategies for growth or growth efficiency can- not really be determined. Extension of the Model to Particles of Different Size Observations using several phytoplankton species as food (Durbin and Durbin 1975) indicated that the preferred (concentration independent) foraging speeds were similar for these species. However these estimates of swimming speed, made with a stop- watch, were not sufficiently accurate to distinguish the small changes in foraging speed that have been found to be significant in the energy budget. Thus it would be desirable to verify this observation using a more precise method, such as video or cinema- tography, to determine the swimming speeds. In the same study it was, however, clear that the threshold concentration for the onset of feeding (c ( ) and the concentration at which foraging speed became approximately independent of food concen- tration (c,) were quite different for plankton particles of different size. The inverse nature of this relation- ship is consistent with the fact that when an Atlantic menhaden forages at a given speed, its energy expen- diture is the same for all food types, yet its energy in- take declines with decreasing food particle size because of the declining efficiency of the gill rakers. This means that a higher concentration of small par- ticles is needed in order for a fish to satisfy its minimum energy requirement, and thus we would ex- pect an increase in c t and c, as particle size declines. The constants in the equations presented here have been specified for Ditylum brightwelli, which is about 80 jiim long. A change in particle size will change the filtration efficiency (e), which will necessitate recal- culation of some of the constants in the equations for R,E,G, s G0VT , and s K 0PT . This is a simple matter ex- cept for the last two quantities, and for these we have presented the steps in the integration of the equa- tions in sufficient detail (Equations (33) to (45)) to permit recomputation for different particle sizes. It is of particular interest to consider how s G0PT changes with a change in food particle size. It has 196 DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN been shown (Fig. 4) thats Gi0PT increases with increas- ing food concentration. This is because with increas- ing c, the rate of energy intake increases per unit of energy expenditure. An increase in food particle size affects the ingestion rate in a manner analagous to an increase in particle abundance, and thus we find that s G 0PT increases with increasing particle size as well (Fig. 13). s G0PT is most strongly affected by food par- ticle size in the range of 20-60 ftm, moderately af- fected within the range of 60-300 jum, and relatively unaffected by further increases in particle size above about 300 /Am. In other words, s G 0PT is strongly size- dependent in the range of phytoplankton particles, less so in the range of microzooplankton, and is for practical purposes independent of particle size in the range of copepodites and late- stage nauplii. This pat- tern, of course, reflects the filtration efficiency curve of the gill rakers (Equation (6)). PARTICLE LENGTH l/im| 0009 Kcol// Chi o(mg/m ! 1200 1600 2000 *Pl d^ weight (mg/m 3 ) PLANKTON CONCENTRATION FIGURE 13.— The effect of particle size on the relationship between the foraging speed which maximizes the Atlantic menhaden's growth rate (s G 0PT ) and plankton concentration. Figure 13 illustrates the need for information on how the Atlantic menhaden responds to mixtures of different-sized particles. For example, do the men- haden respond to the total biomass of particles, or do they key in on certain size classes, ignoring the remainder even though they may filter these particles simultaneously with the larger prey? We have seen that Atlantic menhaden feeding on a single food type will alter their energy expenditures according to the abundance of food, such that they maximize their growth rate at each level of food abundance. However there is a need for further investigation of their feed- ing behavior on different sizes and mixtures of par- ticles to determine the degree to which they act as "optimal foragers" in a mixture of plankton species. Application of the Atlantic Menhaden Models to the Field The energy and nitrogen budgets have been derived in terms of three controlling variables, each of which can be determined from direct field measurements: The foraging speed (s), the concentration of plankton (c), and the foraging time (h). Foraging speed can be measured in the field using acoustic techniques, and this procedure can be used to verify our predictions of swimming speed based on laboratory inves- tigations of the relationship between s and c. If con- firmed in the field, these laboratory studies will enable us to eliminate s as an independent variable and define the budgets simply in terms of c and h. However, as mentioned previously, before we can use this approach in the field, where the fish feed on a variety of particle sizes, additional laboratory work is needed to quantify the foraging speed- food-concen- tration relationships for different types and sizes of plankton. The foraging time {h) could be determined from diel surveys of stomach contents to determine gut fullness and the state of digestion of the food (the latter is an indicator of how recently the food was in- gested). If h proves to be relatively invariant, or under simple control of an external variable such as day length, it may ultimately become possible to describe the energy and nitrogen budgets of the Atlantic menhaden solely as a function of the average concen- tration of different-sized plankton in the water. The effects of body size and temperature also need to be considered in applying the models to the field. Hettler (1976) has investigated the effects of body size, temperature, and salinity on routine metabo- lism in juvenile Atlantic menhaden. The influence of these variables on the swimming and feeding behavior of the Atlantic menhaden, and on the other com- ponents of the energy budget, must be investigated as well, before a general energy and nitrogen budget for the Atlantic menhaden can be described. Another point to consider in applying the present energy budget to the field is that Atlantic menhaden in nature may have additional energy expenditures beyond those of the laboratory fish, principally the costs of predator avoidance, spawning activity, and the energy cost of migration. The first two activities would increase respiratory expenditure, and corre- spondingly reduce the amount of surplus energy that is available for growth. It is not clear to what extent seasonal migrations of the Atlantic menhaden (Nichol- son 1971, 1978) represent an additional energy cost, however, since it is possible that the Atlantic men- haden continue feeding as they move along their migratory routes. In addition, the seasonal migration 197 FISHERY BULLETIN: VOL. 81, NO. 2 does not require elevated swimming speed since Atlantic menhaden swimming at a routine speed of 12.2 cm/s could accomplish the distance between Rhode Island and Cape Hatteras, N.C., well within the 3-4 mo duration of the spring and fall migra- tions. ACKNOWLEDGMENT We thank Peter Verity and Thomas Smayda for their assistance in the experimental studies on menhaden, Captains Harold Loftes and Charles Follett for enabling us to obtain the menhaden, and Theodore Smayda for permitting us to use his laboratory facilities. We also thank Ernesto Lorda and Christopher Langdon for their derivation of Equations (33)-(45), Elizabeth Watkins and Marian McHugh for drafting the figures, and Patricia Degi- dio for preparation of the manuscript. We express our thanks to the National Science Foundation for support of this research through Grants OCE 7602572 and OCE 7919551. LITERATURE CITED Beamish, F. W. H. 1972. Ration size and digestion in largemouth bass, Microp- terus salmoides Lacepede. Can. J. Zool. 50:153-164. 1974. Apparent specific dynamic action of largemouth bass Micropterus salmoides. J. Fish. Res. Board Can. 31:1763- 1769. 1978. Swimming capacity. In W. S. Hoar and D. J. Randall (editors), Fish physiology, Vol. VII, p. 101-189. Acad. Press, N.Y. BlRKETT, L. 1969. The nitrogen balance in plaice, sole and perch. J.Exp. Biol. 50:375-386. Brett, J. R., and J. E. Shelbourn. 1975. Growth rate of young sockeye salmon, Oncorhynchus nerka, in relation to fish size and ration level. J. Fish. Res. Board Can. 32:2103-2110. Brett, J. R., J. E. Shelbourn, and C. T. Shoop. 1969. Growth rate and body composition of fingering sock- eye salmon, Oncorhynchus nerka, in relation to tempera- ture and ration size. J. Fish. Res. Board Can. 26:2363- 2394. Brooks, J. L. 1968. The effects of prey size selection by lake plankti- vores. Syst. Zool. 17:272-291. Carline, R. F., and J. D. Hall. 1973. Evaluation of a method for estimating food consump- tion rates of fish. J. Fish. Res. Board Can. 30:623-629. Confer, J. L., G. L. Howick, M. H. Corzette, S. L. Kramer, S. FlTZGIBBON, AND R. LANDESBERG. 1978. Visual predation by planktivores. Oikos 31:27-37. Dahlberg, M. D. 1969. Fat cycles and condition factors of two species of menhaden, Brevoortia (Clupeidae), and natural hybrids from the Indian River of Florida. Am. Midi. Nat. 82:1 17- 126. Drenner, R. W., J. R. Strickler, and W. J. O'Brien. 1978. Capture probability: the role of zooplankter escape in the selective feedings of planktivorous fish. J. Fish. Res. Board Can. 35:1370-1373. Dubrow, D., M. Hale, and A. Bimbo. 1976. Seasonal variations in chemical composition and pro- tein quality of menhaden. Mar. Fish. Rev. 38(9):12-16. DURBIN, A. G. 1979. A review of the principles of food selection by plankton- feeding fishes. In R. Stroud (editor), Symposium on predator-prey systems in fish communities and their im- plications in fisheries management, p. 203-218. Sport Fishing Inst, Wash., D.C. DURBIN, A. G., AND E. G. DURBIN. 1975. Grazing rates of the Atlantic menhaden Brevoortia tyrannus as a function of particle size and concen- tration. Mar. Biol. (Berl.) 33:265-277. Durbin, A. G, E. G. Durbin, P. G. Verity, and T. J. Smayda. 1981. Voluntary swimming speeds and respiration rates of a filter-feeding planktivore, the Atlantic menhaden, Bre- voortia tyrannus (Pisces: Clupeidae). Fish. Bull., U.S. 78:877-886. Durbin, E. G., and A. G. Durbin. 1981. Assimilation efficiency and nitrogen excretion of a filter-feeding planktivore, the Atlantic menhaden, Bre- voortia tyrannus (Pisces: Clupeidae). Fish. Bull, U.S. 79:601-616. Durbin, E. G., R. W. Krawiec, and T. J. Smayda, 1975. Seasonal studies on the relative importance of dif- ferent size fractions of phytoplankton in Narragansett Bay (USA). Mar. Biol. (Berl.) 32:271-287. EGGERS, D. M. 1977. The nature of prey selection by planktivorous fish. Ecology 58:46-59. Elliott, J. M. 1975. The growth rate of brown trout (Salmo trutta L.) fed on reduced rations. J. Anim. Ecol. 44:823-842. 1976. The energetics of feeding, metabolism and growth of brown trout (Salmo trutta L.) in relation to body weight, water temperature and ration size. J. Anim. Ecol. 45:923- 948. Elliott, J. M., and W. Davison. 1975. Energy equivalents of oxygen consumption in animal energetics. Oecologia (Berl.) 19:195-201. GERKING, S. D. 1955. Influence of rate of feeding on body composition and protein metabolism of bluegill sunfish. Physiol. Zool. 28:267-282. 197 1. Influence of rate of feeding and body weight on protein metabolism of bluegill sunfish. Physiol. Zool. 44:9-19. HETTLER, W. F. 1976. Influence of temperature and salinity on routine meta- bolic rate and growth of young Atlantic menhaden. J. Fish Biol. 8:55-65. HOLLING, C. S. 1966. The functional response of invertebrate predators to prey density. Mem. Entomol. Soc. Can. 48:1-86. IVLEV, V. S. 1960. On the utilization of food by planktophage fishes. Bull. Math. Biophys. 22:371-389. 1961. Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven, 302 p. IWATA, K. 1970. Relationship between food and growth in young Cru- cian carps, Carassius auratus cuvieri, as determined by the nitrogen balance. Jpn. J. Limnol. 31:129-151. 198 DURBIN and DURBIN: ENERGY AND NITROGEN BUDGETS FOR ATLANTIC MENHADEN JANSSEN, J. 1978. Feeding- behavior repertoire of the alewife, Alosa pseudoharengus, and the ciscoes Coregonus hoyi and C. ar- tedii. J. Fish. Res. Board Can. 35:249-253. Jones, R., and J. R. G. Hislop. 1972. Investigations into the growth of haddock, Melano- grammus aeglefinus (L) and whiting, Merlangius mer- langus (L) in aquaria. J. Cons. Int. Explor. Mer 34:174-189. 1978. Further observations on the relation between food in- take and growth of gadoids in captivity. J. Cons. Int. Explor. Mer 38:244-251. Kelso, J. R. M. 1972. Conversion, maintenance, and assimilation for walleye, Stizostedion vitreum uitreum, as affected by size, diet, and temperature. J. Fish. Res. Board Can. 29:1181-1192. LEONG, R. J. H., AND C. P. O'CONNELL. 1969. A laboratory study of particulate and filter feeding of the northern anchovy (Engraulis mordax). J. Fish. Res. Board Can. 26:557-582. Menzel, D. W. 1959. Utilization of algae for growth by the angelfish, Hola- canthus bermudensis. J. Cons. Int. Explor. Mer 24:308- 313. 1960. Utilization of food by a Bermuda reef fish, Epinephelus guttatus. J. Cons. Int. Explor. Mer 25:216-222. MUIR, B. S., AND A. J. NlIMI. 1972. Oxygen consumption of the euryhaline fish aholehole (Kuhlia sandvicensis) with reference to salinity, swim- ming, and food consumption. J. Fish. Res. Board Can. 29:67-77. NICHOLSON, W. R. 197 1. Changes in catch and effort in the Atlantic menhaden purse-seine fishery 1940-68. Fish. Bull., U.S. 69:765- 781. 1978. Movements and population structure of Atlantic men- haden indicated by tag returns. Estuaries 1:141-150. Ndmi, A. J., and F. W. H. Beamish. 1974. Bioenergetics and growth of largemouth bass (Microp- terus salmoides) in relation to body weight and tem- perature. Can. J. Zool. 52:447-456. O'BRIEN, W. J., N. A. SLADE, AND G. L. VlNYARD. 1976. Apparent size as the determinant of prey selection by bluegill sunfish (Lepomis maerochirus). Ecology 57:1304- 1310. O'CONNELL, C. P. 1972. The interrelation of biting and filtering in the feeding activity of the northern anchovy (Engraulis mordax). J. Fish. Res. Board Can. 29:285-293. PANDIAN, T. J. 1967. Transformation of food in the fish Megalops cyp- rinoides II. Influence of quantity of food. Mar. Biol. (Berl.) 1:107-109. Peck, J. I. 1894. On the food of the menhaden. Bull U.S. Fish. Comm. 13:113-126. Pierce, R. J., and T. E. Wissing. 1974. Energy cost of food utilization in the bluegill (Lepomis maerochirus). Trans. Am. Fish. Soc. 103:38-45. Savitz, J., E. ALBANESE, M. J. Evtnger, and P. Kolasinski. 1977. Effect of ration level on nitrogen excretion, nitrogen retention and efficiency of nitrogen utilization for growth in largemouth bass (Micropterus salmoides). J. Fish Biol. 11:185-192. Schalles, J. R., and T. E. Wissing. 1976. Effects of dry pellet diets on the metabolic rates of bluegill (Lepomis maerochirus). J. Fish. Res. Board Can. 33:2443-2449. Solomon, D. J., and A. E. Brafield. 1972. The energetics of feeding, metabolism and growth of perch (Perca fluviatilis L.). J. Anim. Ecol. 41:699-718. Staples, D. J., and M. Nomura. 1976. Influence of body size and food ration on the energy budget of rainbow trout Salmo gairdneri Richardson. J. Fish Biol. 9:29-43. Stirling, H. P. 1977. Growth, food utilization and effect of social interaction in the European bass Dicentrarchus labrax. Mar. Biol. (Berl.) 40:173-184. Ware, D. M. 1975. Growth, metabolism, and optimal swimming speed of a pelagic fish. J. Fish. Res. Board Can. 32:33-41. Warren, C. E., and P. Doudoroff. 1971. Biology and water pollution control. W. B. Saunders Co., Phila., 434 p. Watts, R. L., and D. C. Watts. 1974. Nitrogen metabolism in fishes. In M. Florkin and B. T Scheer (editors), Chemical zoology Vol. VIII. Deutero- stomians, cyclostomes, and fishes, p. 369-446. Acad. Press, N.Y. WE AST, R. C. (editor). 1977. CRC handbook of chemistry and physics. 58th ed. CRC Press, Inc., Cleveland. Werner, E. E. 1974. The fish size, prey size, handling time relation in several sunfishes and some implications. J. Fish. Res. Board Can. 31:1531-1536. Werner, E. E., and D. J. Hall. 1974. Optimal foraging and the size selection of prey by the bluegill sunfish (Lepomis maerochirus). Ecology 55:1042-1052. WlNBERG, G. G. 1956. (Rate of metabolism and food requirements of fish- es.] Nauchnye Tr. Belorussk. Gos. Univ. Minsk, 253 p. (Transl. 1960. Fish. Res. Board Can. Transl. Ser. 194.) Wurtsbaugh, W. A., and G. E. Davis. 1977. Effects of temperature and ration level on the growth and food conversion efficiency of Salmo gairdneri, Richard- son. J. Fish Biol. 11:87-98. 199 REPRODUCTION AND EMBRYONIC DEVELOPMENT OF THE SAND TIGER SHARK, ODONTASPIS TAURUS (RAFINESQUE) 1 R. Grant Gilmore 2 , Jon W. Dodrill\ and Patricia A. Linley 2 ABSTRACT The capture of one ripe male, 191.5 cm TL, and 26 pregnant female, 236.6-274.3 cm TL, sand tiger sharks, Odontaspis taurus, from the east-central coast of Florida from 1946 to 1980 has permitted examination of early reproductive activity and embryonic development in this species. Variations in ovulation rates and oviducal gland activity produce six distinct egg capsule types at varying times during gestation. Some egg capsules produced during early gestation contain only ovalbumin and/or mucus while others contain several fertilized ova. As gestation proceeds, more capsules contain unfertilized ova and ovulation rates increase. These latter capsules serve principally as food for the surviving embryo. Sixty-two embryos, 13-1,060 mm TL, provided information on intrauterine development which allowed classification of seven developmental periods based on gestation time, embryonic anatomy, posture, activity, and source of nutrition. Initially, embryos 13-18.5 mmTL obtain nutrition from internal coelomicyolk sup- plies during a period of early tissue differentiation. In embryos between 18.5 and 51 mmTL, consumption of encapsulated yolk supplies occurs until hatching, between 49 and 63 mm TL. After hatching, the embryo ab- sorbs yolk-sac nutritive supplies and may also consume uterine fluid. At about 100 mm TL, the embryo begins to hunt and consume other intrauterine embryos. Seven to nine months into gestation, ova are no lon- ger fertilized. In each uterus, the single remaining embryo, 334-1,060 mm TL, consumes enlarged yolk cap- sules containing 7-23 unfertilized ova. Just prior to parturition the maternal ovary is greatly reduced in size, few egg capsules are found within the uteri, and in each uterus the remaining embryo exhibits reduced yolk consumption and an enlarged liver. Parturition observed in captivity typically takes place from December through March, after 9-12 months of gestation. Newborn juveniles are about 100 cm long. The sand tiger shark, Odontaspis taurus (Rafinesque, 1810), is a cosmopolitan species distributed in sub- tropical and temperate waters at depths < 60 m (Bass et al. 1975). In the western Atlantic, adult sand tiger sharks occur from the Gulf of Maine to Brazil (Bigelow and Schroeder 1948). Although sand tiger sharks have been captured on both coasts of Florida (Springer 1938, 1948, 1963; Clark and von Schmidt 1965), captures have been more common along the Florida east coast (Dodrill 4 ). Unlike the adults, free-swimming juvenile 0. taurus in the western Atlantic are restricted only to tem- perate (Bigelow and Schroeder 1953) and warm- temperate waters, extending as far south as northern Florida. Juveniles 109.3-157.7 cm in total length (TL) have been recorded in neritic waters from the 'Contribution No. 305, Harbor Branch Foundation, Inc., Fort Pierce, Fla. 2 Harbor Branch Foundation, Inc., R.R. 1, Box 196, Fort Pierce, FL 33450. 'District V Naturalist, Division of Recreation and Parks, Florida Department of Natural Resources, Rt. 1, Box 107-AA, Clermont, FL 32711. 4 Dodrill, J. W. 1977. A hook and line survey of the sharks found within five hundred meters off shore along Melbourne Beach, Brevard County, Florida. Unpubl. M.S. Thesis, 304 p. Fla. Inst. Technol., Melbourne, FL 32901. vicinity of Fernandina Beach (lat. 30°40'N, Nassau County) on the Florida Atlantic coast, from Cedar Key (lat. 29°15'N, Levy County) in the northeastern Gulf of Mexico (Don Hoyt 5 ), and from the northern Gulf of Mexico (Branstetter 1981). In the western Atlantic, females with near-term em- bryos have been captured off eastern Florida and in the northern Gulf of Mexico (Springer 1948; Hoyt footnote 5; Robert Jenkins 6 ). At parturition, two young are born (95-110 cm TL), one developing in each uterus (Springer 1948; Cadenat 1956; Sadowsky 1970; Bassetal. 1975). Published observations on the early intrauterine development of O. taurus are limited to the accounts of Coles (1915), Springer (1948), Cadenat (1956), andBassetal. (1975). Springer (1948) was the first to observe embryonic oviphagy in O. taurus. He found large quantities of yolk in the stomachs of embryos dissected from females from the northern Gulf of Mexico and east-central Florida. Bass et al. (1975) described an intact 40 mm embryo found in the Manuscript accepted September 1982. FISHERY BULLETIN: VOL. 81, NO. 2, 1983. 5 Don Hoyt, Florida Shark Club, Inc., Jacksonville, FL 322 1 1 , pers. commun. 1967-77. 'Robert Jenkins, Marineland Inc., St. Augustine, FL 32084, pers. commun. 1977. 201 FISHERY BULLETIN: VOL. 81, NO. 2 stomach of a 1 70 mm embryo dissected from a female from Natal, South Africa. These were the smallest embryos yet recorded from O. taurus and provided the first description of embryonic cannibalism in this species. The capture of 28 pregnant O. taurus from various locations on the east coast of Florida (1946-80) pro- vided 62 embryos, 13-1,060 mmTL (Table 1, Fig. 1). These specimens have allowed a more detailed de- scription of early embryonic development in this species than was possible previously. This study de- scribes the various developmental stages in 0. tau- rus based principally on embryonic anatomical de- velopment and changes in maternal gonadal mor- phology. METHODS All adult 0. taurus specimens examined were cap- tured either on rod and reel sport fishing gear or on static 10-30 hook set lines. Fourteen specimens were captured 200 m to 19 km from shore in neritic waters off Melbourne Beach, Brevard County, Fla. (lat. 28°00'N, long. 80° 33' W). All specimens came from depths of 5-12 m. A 15th specimen was caught at lat. 27°25'N, long. 80°12'W, east of Fort Pierce Inlet, St. Lucie County, Fla. A 16th specimen, a 240 cm female, gave birth to two pups at Sea World of Orlan- do, Fla., and all three were examined. This latter adult female was captured on 2 1 August 1 980 at Port Canaveral, Brevard County (lat. 28°24.5'N). Eleven other specimens were captured prior to our study; these data and, in some cases, embryos from these specimens were included (Table 1). Embryos and adult reproductive tracts were pre- served in 10% Formalin 7 and stored in 10% buffered Formalin or 70% ethanol, or were frozen. All of these specimens were entered and catalogued into the In- dian River Coastal Zone Museum (IRCZM). Egg diameters and embryos < 1 30 mm TL were measured using vernier calipers to the nearest 0.1 mm. All length measurements including total length (TL) follow Bass et al. (1975). 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table 1.— Uterine embryo and egg capsule data for Odontaspis taurus, from the Florida east coast, arranged chronologically by month of examination of embryos, 1947-81. Adult No. of egg Encaps ulated Damage d (a) or consumed (b) Hatched embryos Total size (cm, TL) capsules Left in uteri Right embryos mm. TL) embryos mm. TL) (mm TL) no. of Date Left Right Left Right Left Right embryos 15 Mav 1977 254.5 20 20 16 May 1977 254.9 18 19 28 May 1977 2603 26 27 5 June 1978 264 2 8 8 5 June 1976 258 1 35±2 34±2 41 42 57 ? 3 5 June 1976 262.5 20±2 20±2 38* 38* 2+? 5. June 1976 263 2 20-35 20-35 38' 38* 2+> 6 June 1978 274 3 8 8 9 June 1976 249.5 29±4 29±4 27 31 2 28 June 1976 254.1 47 53 27, 34 27, 38. 46 63 62 7 8 July 1978 274,2 66** 69" 13. 18 y 49(a) 45(a). 49(a) 131 131 7 18 July 1976 271.5 78 81 34 51(a) 127 100 4 27 July 1975 263.0 ? ? ? ? 317 317+10 2 29 July 1977 254.0 77" 77" 271 227 2 5 Aug. 1976 2366 68 65 9(b). 22(b), 35(a). 36(b) 30(b) 41(a) 41(b) 334 320 9 4 Sept. 1970 282.5 17.5 18 5 ? 7 ? 2+? 4 Sept 1970 2692 ? ? ? ? 330±10 330±10 2 3 Nov 1962' ? ? ? ? ? ? 650 1 8 Nov 1954' ? ? ? ? ? 830 890 2 24 Nov. 1947 2 273.0 970 960 2 24 Nov. 1947 2 2390 825 1 12 Dec 1976 3 266 7 ? ? ? ? 1.000±10 1,000±10 2 30 Dec. 1958' 261.6 > ? ? ? 1.025 1.033 2 22 Jan 1947 2 ? > ? ■> ? 1,000±10 1.000±10 2 22 Jan 1947 2 ? > ' ? ? ? 15 Feb. 1959' 261.5 > -> ? ? 1,060 > 1,060 2 9 Mar 1947 2 ' 4 272.0 > ? ? ? 1,050 1.030 2 22 Mar 1981 5 240.0 } ? ? ? 910 95 2 "Length given as 1.5 inches, therefore not accurately determined. **Blastodiscs were observed on some eggs. 7 Egg capsules and embryos could have been present but were not recorded. 1 F. G. Wood, formerly of Marine I and Inc., St. Augustine, FL 32084, pers. commun. 1 976-77 2 Spnnger 1948. 3 E. Herbert, Florida Shark Club. Jacksonville, FL 3221 1, pers. commun. 1976-77 4 A. McBrtde, Curator, Manneland Inc., St. Augustine, FL 32084, unpubl. data, 1947. 5 Specimens were still living in captivity April 1983 at Sea World of Orlando, Fla. 202 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS The entire reproductive tract was removed and ex- amined as fresh, frozen, or preserved material. Uterine fluid volume was determined by tying off both ends of the uterus in a fresh specimen, removing the uterus, making a small incision in the uterine wall, and allowing the contained fluid to drain into a graduated flask. Selected preserved ovaries were cut into sections which were weighed to the nearest 0. 1 g. Ovarian egg counts were made by counting all mac- roscopic eggs in two preserved sections from an ovary of known weight. These ova counts were then multiplied by the ratio of total ovarian weight/section weight, to predict the total number of ova in the entire ovary. A 13.0 mm TL embryo taken from an egg capsule^ from a shark caught on 8 July 1978 was embedded in paraffin, cut on a rotary microtome at 6 jam on a sagit- tal plane, and stained with a Cason modification of the Mallory-Heidenhain stain (Humason 1972). Fresh sperm samples were fixed in 2.5% glutar- aldehyde, prepared for scanning electron micros- copy, and examined on a Zeiss Novascan. Several Polaroid electron micrographs were taken for sperm descriptions. 1,100 1,000 Apr. May June ' July ' Aug. ' Sept. ' Oct. ' Nov. ' Dec. Jan Feb. Mar. Month FIGURE 1.— Recorded lengths of embryos taken from Odontaspis taurus females captured 1946-80 along the Florida east coast, with hypothetical growth curve (above) of embryos and monthly numbers of adult males and females (bar graph below) captured during the same 34-yr period. 203 FISHERY BULLETIN: VOL. 81, NO. 2 Drawings and Kodachrome transparencies were made of various embryos, egg capsules, and re- productive organs. OBSERVATIONS AND DESCRIPTIONS Mating Activity (Mating Period, Location, and Spermatozoa) The occurrence of similar-size males or females in unisexual groups has been documented on several occasions (records of the Florida Shark Club show 107 0. taurus landings, Burton 1932 8 ; Sadowsky 1970; Bass et al. 1975; Hoyt 1976-77 see footnote 5; Wood 1976-77 9 ). These observations show that female groups of 0. taurus make coordinated seasonal coastal movements possibly for breeding, gestation, and eventually parturition (Fig. 1). Females captured at the same time and location ten- ded to have embryos in the same state of develop- ment, suggesting coordinated breeding activity and postbreeding migrations. Observations of many an- nual cycles from 1947 to 1981 established winter- spring as a breeding period off the Florida east coast and provided comparisons of data on gestation (i.e., embryonic development rates and seasonality). A 191.5 cm TL ripe male 0. taurus, captured 8 Feb- ruary 1980 in shallow water (10 m depth) in the vicinity of Fort Pierce Inlet, St. Lucie County, Fla. (lat.27°25.7'N, long. 80° 12.5'W), showed evidence of recent mating activity. His claspers were turgid and hematose, with sperm and seminal fluid actively flowing from the clasper tip. The testes were also en- larged (22.5 X 3.5 cm, 0.68 kg). A larger 203 cm TL male examined from Fort Macon, 1.5 km west of Beaufort Inlet, N.C.(lat.34°40'N, 10 January 1978), contained testes which were considerably smaller (8.0 X 5.0 cm, 0.064 kg). Several scanning electron micrographs were made of the sperm from the 8 Feb- ruary 1980 male specimen. A single sperm had a typi- cal chondrichthian helical head structure 3 1 ju.m long and a tail 40.3 ju,m long (Fig. 2B). The entire length of the sperm was 69-7 1.5 jLtm. Living sperm were obser- ved to rotate about their long axes, propelled by the circular motion of the extended tails. Mating scars resulting from copulatory activity 8 E. M. Burton, The Charleston Museum, Charleston, S.C., pers. commun. 24 Oct. 1932 to J. T. Nichols, American Museum of Natural History, N.Y. (made available by Stewart Springer, Mote Marine Lab., Sarasota, FL 33577). T. G. Wood, Marineland Inc., St. Augustine, FL 32084, pers. com- mun. 1976-1977. have been commonly observed in female "galeoid" sharks; however, it appears that courtship scars on males are rare (Springer 1967; Stevens 1974; Pratt 1979). Springer (1960) had noted the presence of fresh cuts on female Eulamia milberti (= Car- charhinus plumgeus) in correlation with the presence of early embryos. Springer (1963) found that most of the 0. taurus taken in a shark fishery operating in the Atlantic off east-central Florida were females with a high incidence of courtship scars; but no dates were given for these observations. Odontaspis taurus females we captured on 9 June and 5 August 1976 (Table 1) had tooth puncture wounds between the 1st and 2d dorsal fins. The 191.5 cm male, taken on 8 February 1980 off Fort Pierce Inlet, had been recent- ly raked by another shark along the upper left side of the body behind and above the gill openings (Fig. 2 A). This wound consisted of eight incisions, created by a narrow, long tooth rather than a flat, wide blade tooth, typical of many carcharhinid sharks. As 0. taurus has a long narrow tooth cusp, it is possible that the wound was the result of either an attack by, or copulation with, another sand tiger shark. These ob- servations indicate that copulatory activity may take place off the Florida east coast and therefore account for the following observations of the earliest em- bryonic development in specimens from this geo- graphical region. Early Gonadal and Embryonic Developmental Period (January-September; 0-60 mm TL) General Female Anatomy The female reproductive tract of 0. taurus may be divided into the ovary, ostium, anterior oviduct, oviducal gland, isthmus, uterus, and vagina, typical of most galeoid sharks. Only the right ovary is functional and enlarged. Above the ovary and at- tached to it via membranous connective tissue (mesovarium) is the ostium which collects ovulated ova and distributes them to the oviducts. The two oviducts (paired, right and left) bifurcate from the os- tium. The anterior oviducts are about 9 mm in diameter and 300 mm in length from ostium to ovidu- cal glands in a 254 cm female. The heart-shaped oviducal glands (53 X 93 mm in the same female) function in egg capsule formation. Much larger than the anterior oviduct, the portion of the oviduct following the oviducal gland known as the isthmus is 20-34 mm in diameter, allowing for the passage of multiple encapsulated ova. The isthmus opens into 204 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS •XT > * , FIGURE 2.— A 191.5 cm TL ripe male Odontaspis taurus captured 8 February 1980 off Fort Pierce Inlet, Fla. (A) Tooth rake scars along upper left side of body behind and above gill openings. (B) Scanning electron micrograph of a 69-71.5 ^m sperm, with head structure 31 /im long and tail 40.3 jum long (1.950X magnification). 205 FISHERY BULLETIN: VOL. 81, NO. 2 the uterus which is heavily folded and vascularized near its opening. The 7-8 mm uterine wall in 0. taurus does not function in placentation as in carcharhinid sharks. The paired uteri unite posteriorly to form a common vagina. Ovarian Activity This period begins from January to April with in- semination of the female 0. taurus and extends into the following September as exhibited by the pro- longed fertilization of ova via stored sperm. Ova fer- tilization apparently occurs in the anterior oviduct or oviducal gland (Fig. 3Ba, b) prior to egg capsule for- mation. The oviducal gland then produces a variety of collagen egg capsules, some of which contain fer- tilized ova (Figs. 4, 5). Egg capsules are then deposited in the uterus. Although encapsulated em- bryos are present in the uterus for 5-6 mo, the development of a single embryo from fertilization to hatching in utero takes about 3-4 mo. The number of ova and the general overall size of the ovary increased during early pregnancy. During this period, ova diameters ranged from 2.0 to 10.2 mm and weights ranged from 1.6 to 410 mg. A 254.5 cm TL female 0. taurus captured 15 May 1977 con- tained a 4.6 kg ovary with 22,180 ova 1.3-10.0 mm in diameter (Table 2). Encapsulated fertilized ova (i.e., blastodiscs were evident) were present in the uterus, but no embryos. All 11 sand tiger sharks examined between June and August possessed greatly hyper- trophied right ovaries (left ovaries are atrophic and nonfunctional) weighing between 3.7 and 8.5 kg and taking up considerable space (360-455 mm in length) in the body cavity (Fig. 3, Table 2). The largest ovary (8.5 kg) came from an 8 July 1978 sand tiger shark, which also had two embryos that were past the "ear- ly" uterine developmental stages and in the "post- hatch" cannibalistic stage during which consump- tion of ova would be their primary means of nutrition. Oviducal Gland Activity The paired oviducts of 0. taurus may be divided into four basic sections (Fig. 3B). The anterior portion (a) is a narrow tube lined with ciliated columnar epithe- lial cells, extending between the ostium and the oviducal or nidamental gland. This anterior tube is 310 mm long in a 254 cm sand tiger shark and about 9 mm in diameter. The oviducal gland (b) secretes mucus, ovalbumin, and the major elasmobranch egg case component, collagen (Wourms 1977). Neither the anterior portion of the oviduct nor the oviducal glands were sectioned and examined in detail for sperm storage; therefore, the exact site of fertiliza- tion in O. taurus remains unknown. However, fer- tilization must occur prior to encapsulation of the ova in the shell membrane or collagen egg capsule. En- capsulation takes place within the oviducal gland. TABLE 2. — Comparative reproductive data for female Odontaspis tauruf capture, 1947-78. arranged chronologically by month of Location Shark (cm, TL) Ovary Ova Embryos in both Date Weight (kg) Length (cm) No. Size (mm) uteri 24 Feb. 1 960 Gulf of Mexico 1 296 — 40 > 1,000 >10 15 May 1977 West Atlantic Brevard Co.. Fla. Melbourne Beach 2545 4.6 36 22.180 1.3-10.0 9 June 1976 West Atlantic Brevard Co., Fla. Melbourne Beach 2495 3 7 — 13.200 >5-10 2 28 June 1976 West Atlantic Brevard Co., Fla. Melbourne Beach 254.1 45.5 > 1 ,000 7 8 July 1978 West Atlantic Brevard Co., Fla. Melbourne Beach 274.2 8.5 41.4 24,290 2.5-10.2 7 27 July 1947 Gulf of Mexico 2 Chandleur Is., La. 312.5 — — 24,000 >1-10 2 29 July 1977 West Atlantic Brevard Co., Fla. Flondana Beach 254.0 — 45.7 > 1.000 2-9.5 2 5 Aug. 1976 West Atlantic Brevard Co., Fla. Melbourne Beach 2366 4.5 31.0 12,810 3-10 9 24 Nov. 1947 West Atlantic' Brevard Co., Fla. Off Cape Canaveral 273 >100 10 1 1 Clark and Von Sch midt 1965. 'Springer 1948. 206 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS FIGURE 3.-A 254 cm female Odontaspis taurus captured 29 July 1976 at Melbourne Beach, Fla. (A) Enlarged ovary with ova (a and b) extending through damaged portions of ovarian membrane. (B) Oviduct consisting of (a) thin tube leading from ostium to (b) oviducal gland; (c) isthmus, and (d) uterus containing embryos and egg capsules. 207 FISHERY BULLETIN: VOL. 81. NO. 2 FIGURE 4. — Dissection from a female Odontaspis taurus of (A) oviducal gland (right) and isthmus containing a Type I egg capsule which had 16 ova; (B) two Type V "short tail" gel capsules leaving the oviducal gland. This bulbous organ varies in size with reproductive activity and produces a wide variety of egg capsules (Figs. 4, 5). Egg capsules leave the oviducal gland and proceed down the elastic narrow isthmus (Fig. 3Bc), 250-350 mm in length, 20-34 mm in diameter, con- necting the gland with the expanded uterus (Fig. 3Bd). There is an increase in vascularization and folding of the inner epithelial lining where the isthmus joins the enlarged uterus. The size of the uterus, the volume of the uterine fluid, and the length of the isthmus increased during early gestation (June to July). The volume of fluid in a single uterus increased from 260 ml, to 325 ml, to 1,600 ml in specimens from 15 May, 28 May, and 208 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS FIGURE 5. — (A) Type V "short tail" gel capsules from Odontaspis taurus, containing ovalbumin and/or mucus; (B) Type IV "long tail" gel cap- sules; (C) Type I blastodisc capsule, containing 16 ova; (D) Type II ovoid yolk capsules, containing 18 ova each. (Photo courtesy Marineland Inc., St. Augustine, Fla.) 29 July, respectively. The uterine fluid also increased in relative cloudiness and contained numerous rup- tured egg capsules and yolk fragments. During early gestation the oviducal gland produced at least six distinct types of egg capsules (Figs. 4, 5): Type I blastodisc capsules (Figs. 4A, 5C) — contain 7- 18 ova, 1-14 of which have visible blastodiscs. This capsule type was more prevalent during early gestation, as 83% of the intact capsules, examined in a 16 May 1977 specimen, contained blas- todiscs, 25% in a 28 May 1977 specimen, 22% in a 28 June 1976 specimen, and none in a 5 August 1976 specimen. The overall capsule number was generally low, 15 or fewer per uterus. Type II ovoid yolk capsules (Fig. 5D) — consist of a light amber shell membrane enclosing a rounded bulbous head containing the ova and a flattened transparent amber tail 40-58 mm long. We found these capsules to contain a large yolk volume con- sisting of 7-18 ova (mean = 11), 10 mm in diameter, with no sign of fertilization (i.e., no blas- todisc). Springer (1948) found 16-23 ova (mean = 19) per capsule of this type in a female containing 260.4-266.7 mm embryos. Ovoid yolk capsules in- creased in numbers as Type I blastodisc capsules declined. Dimensions of Type II capsules ranged from 2 1 to 29 X 78 to 1 18 mm and weight from 8.6 to 19.4 g. Ova in the egg mass during the first 2-3 mo of pregnancy comprised only 60-80% of the capsule volume, while the remainder consisted of ovalbumin and/or mucus adjacent to the tail. A gelatinous ovalbumin/mucoid substance also lined the inner walls of the egg capsule. Type III reduced yolk capsules — have the same dimensions as Type LI ovoid yolk capsules but contain only 1-3 ova. Type III capsules were ob- served only during the first 3 mo of gestation. Type IV "long tail" gel capsules — contain amber, green, or white gelatinous material, fluid, and no ova. Although not determined, the variably 209 FISHERY BULLETIN: VOL. 81. NO. 2 colored gelatinous material probably contains ovalbumin and mucus, in different proportions. The dimensions are usually similar to those of the Type II capsules but may vary, as total lengths of up to 1 70 mm were observed in capsules with very long tails (Fig. 5B). These capsules were most common during the first 3 mo of gestation. Type V "short tail" gel capsules (Figs. 4B, 5A) — are the smallest capsules, are generally flattened, and contain only gelatinous ovalbumin/mucoid ma- terial. These capsules were also most common during the first 3 mo of gestation. Type VI embryo capsules — contain an embryo and a reduced volume of yolk. Despite the presence of multiple ova and several blastodiscs in embryo capsules, dissection of all Type VI capsules failed to show more than one embryo developing within a single capsule. Prior to entering the uterus, egg capsules of the same type were found in similar positions in both oviducts of a particular adult. No matter how many eggs were ovulated, encapsulation of albumin would occur synchronously in each oviducal gland, thus pro- ducing egg capsules of the same type at the same time. Calculation of egg capsule production rates, based on changes in uterine capsule numbers, in- dicates that capsule formation takes place at 24-36 h intervals. Initial egg capsules contain ovalbumin and/ or mucus derived from the oviducal gland. As the ovulation rate and the number and volume of ova in- creased during later stages of gestation, more ova were present in the oviducal gland when encapsula- tion occurred. At this time only ovoid yolk capsules, Type II, were found in the oviduct and uterus. Embryonic Development Multiple embryos from 0. taurus develop in each uterus during the early stages of gestation. However, the maximum number of capsules containing mac- roscopic embryos is low (no more than 9% or 2-7 of all capsules in both uteri combined at any given time). Encapsulated embryos were found from June to Sep- tember. The maximum number of embryos in a single uterus was seven, ranging in size from 19 to 334 mm TL. Four of these seven were found in the mouth and stomach of the largest embryo. After June, the num- ber of undamaged encapsulated embryos and the percentage of capsules with blastodiscs declined. B 1mm Figure 6. -Three views of a 13 mm embryo (IRCZM 103179) taken from an adult Odontaspis taurus, 27 4. 2 cm long, captured 8 July 1978. Left side; (B) dorsal; (C) ventral. (A) 210 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS 13 MM EMBRYO (IRCZM 103179, Figs. 6-8).— The 13 mm embryo is described from one of four em- bryos, 13-131 mm, taken from the left uterus of a 274.2 cm sand tiger shark caught 8 July 1978 (Table 1). This and an 18 mm embryo were undamaged and encapsulated, while three other embryos partially encapsulated or free within the same sand tiger shark were damaged by attacks from two larger 131 mm embryos, one in each uterus. The 13 mm embryo was the smallest examined. It contained yolk both inter- nally and in a yolk sac. The embryo was obviously re- stricted in mobility appearing as little more than a yolk mass with a head, notochord, and minute pec- toral fin buds. The 13 mm embryo resembles an amphibian embryo after gastrulation and formation of primary organ rudiments. It does not resemble the early embryos described for other elasmobranchs [e.g., Heterodontus japonicus (Smith 1942); Chlamy- doselachus anguineus (Gudger 1940); Mustelus canis (TeWinkel 1950, 1963)]. Histological sections showed an incomplete connection between internal yolk sup- plies and an external yolk sac (Fig. 8A). A membrane at the junction of the yolk stalk and the yolk sac ap- pears to isolate the yolk-sac yolk from the yolk stalk and coelomic yolk supplies in the 1 3 mm embryo. The coelomic cavity, cardiac stomach, valvular intestine, and pericardial cavity all contained yolk. The max- imum horizontal diameter of the embryo was 9 mm, due principally to the contained yolk. This diameter was greater than that of the yolk sac (6.0 mm). The gill arches and mouth cavity were open, but the latter was lacking dentition. No retinal tissue was seen and gonadal tissue was undifferentiated. 18.5 MM EMBRYO (IRCZM 103134, Fig. 9).— The 18.5 mm embryo was from the right uterus of a 282.5 cm TL female 0. taurus captured 4 September 1970. Although encapsulated, the embryo and the capsule had been greatly damaged. This embryo was similar to the 13 mm embryo but differed in having less internal yolk and greater differentiation of exter- nal features. A spiracle was present as were first and second dorsal, caudal, anal, and pelvic fin buds in ad- dition to the pectoral fin buds which had developed earlier. The yolk sac was 6.0 mm in diameter as in the 13 mm embryo. FIGURE 7.-Angle horizontal sagittal view of a 13 mm Odontaspis taurus embryo (IRCZM 103179), head and branchial region: (b) brain; (o) orbit; (ga) gill arches; (ysy) yolk sac yolk. 211 FISHERY BULLETIN: VOL. 81, NO. 2 FIGURES.— Angle horizontal sagittal view of head section of a 13 mm Odontaspis taurus embryo (IRC ZM 103179). (A) Pericardial and anterior coelomic cavities: (cy) coelomic yolk; (yss) yolk-sac stalk; (ysy) yolk-sac yolk; (ysm) yolk-sac membrane. (B) Yolk stalk, yolk sac, and lower coelomic cavity: (ypc) yolk in pericardial cavity; (cs) cardiac stomach; (cy) coelomic yolk; (vi) valvular intestine; (ga) gill arches. 31.0 MM EMBRYO (IRCZM 103139, Fig. 10).— This encapsulated embryo was the only one present in the right uterus of a 249.5 cm female 0. taurus cap- tured 9 June 1976. The 7.5 mm diameter yolk sac was slightly larger than that of smaller embryos ex- amined. All fin buds had developed further. External gill filaments were present. 49.0 MM EMBRYO (IRCZM 103102, Fig. 11).— The 49.0 mm embryo was found free in the uterus of a 274.2 cm TL female O. taurus caught 8 July 1978. The emaciated condition, numerous small puncture wounds, and absence of large numbers of branchial filaments on this embryo indicated that it had been attacked by the larger 131 mm embryo also present in the uterus. Although the 49 mm embryo is near the size range of other recently hatched embryos (i.e., SI- OS mm) from 0. taurus females caught during June, it also could have been torn from its egg capsule by the larger embryo. Apparently also damaged by attacks from the larger embryo, the yolk sac of this embryo was only 4 mm in diameter. Erect wide triangular teeth lacking basal denticles were clearly visible. The stiff, sharp structure of these teeth indicated that they were functional and could have enabled the em- bryo to hatch from the egg capsule. Gill filaments ex- tended from the gill arches, although many were damaged and probably removed when the embryo was attacked. 57 MM EMBRYO (IRCZM 103145, Fig. 12). — The 57 mm embryo was found free along with an unhatched 4 1 mm embryo in the left uterus of a 258. 1 cm TL female O. taurus captured 5 June 1976. This embryo revealed maximum development in external branchial filaments. Numerous long filaments extend- ed from both the gill openings and spiracle (Fig. 12 A). A single 3.7 mm filament extended from the cornea at the dorsal edge of the iris (Fig. 12B). Rudimentary claspers were evident on the inner margin of the pelvic fins, indicating secondary sex characteristics were developing. 212 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS 1mm 1mm • 1 FIGURE 9.— Two views of an 18.5 mm Odontaspis taurus embryo (IRCZM 103134) taken from the right uterus of a 282.5 mm TL female cap- tured 4 September 1970, showing damage by intrauterine attacks from larger embryo. Posthatch and Intrauterine Cannibalistic Period (June-September; 60-334 mm) This period is characterized by hatching of the largest encapsulated embryos, consumption of yolk- sac yolk supplies, and active cannibalism by the largest hatched embryo upon other intrauterine en- capsulated or small hatched embryos until only one embryo remains. These events occur simultaneously in each uterus. From June to September this developmental period overlaps the latter part of the early gestation phases of other sibling embryos. Two hatched embryos, 62 and 63 mm, (Fig. 13) from each uterus of a late June sand tiger shark were noticably more robust than five 27-46 mm embryos still encapsulated in these uteri. However, there was no evidence that the larger embryos had begun to feed upon other egg capsules, encapsulated em- bryos, or other free embryos. The 62 and 63 mm specimens still possessed 5.5-6.0 mm diameter yolk sacs and branchial filaments. At about 100 mm, the embryo has consumed the contents of the yolk sac and begins obtaining nourishment through adelphophagy and oophagy. Evidence of intrauterine cannibalism was found in the uterus of a 271.5 cm female 0. taurus, caught 18 July 1976, which contained a large hatched embryo (100 mm) that had attacked and badly damaged 213 FISHERY BULLETIN: VOL. 81, NO. 2 FIGURE 10.— View of a 31.0 mm Odontaspis taurus embryo (IRCZM 103139) taken from the right uterus of a 249.5 cm female captured 9 June 1976. FIGURE 1 1.— Two views of a 49 mm Odontaspis taurus embryo (IRCZM 103 1 02) taken from a 274.2 cm TL female captured 8 July 1978, show- ing emaciation and injuries from intrauterine attacks by a larger 131 mm embryo. 214 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS FIGURE 12.— (A) A 57 mm Odontaspis taurus embryo (IRCZM 103145) taken from a 258.1 cm TL female captured 5 June 1976; (B) enlarge- ment of orbit and spiracle showing associated filaments. FIGURE 13.— Hatched 62 mm Odontaspis taurus embryo with 6 mm yolk sac taken from right uterus of a female caught 28 June 1976, Melbourne Beach, Fla. 215 FISHERY BULLETIN: VOL. 81. NO. 2 (puncture wounds and torn gut) a 51 mm embryo (drawn to scale; Fig. 14A). Having already developed teeth, the 51 mm embryo (see Figure 11 of a 49 mm embryo) had a potential for competitive interaction with the larger 100 mm embryo, although at a decided size disadvantage. It is possible that the 51 mm embryo had not hatched prior to the attack. However, empty and broken egg capsules were not found in the uterus. There is no evidence that the 100 mm embryo had tried to consume any of the other 8 1 egg capsules in the uterus, nor were there broken or damaged capsules in the opposite uterus which con- tained a 127 mm hatched embryo. We obtained further evidence that hatched em- bryos and/or encapsulated embryos are selectively preyed upon by their larger siblings within the uterus. Two embryos (45 and 49 mm) in the right uterus of an 8 July 1978 female 0. taurus were badly damaged by the attack of a 131 mm male embryo. Six empty egg capsules were found within the same uterus. None of the other 63 egg capsules were damaged (some of which contained fertilized ova). In the left uterus, a 49 mm embryo had been mutilated by a 131 mm em- bryo and two of the 66 egg capsules were empty. A 334 mm embryo from the left uterus of a 5 August 1976 adult 0. taurus had four embryos 9-36 mm TL within its pharynx. Two damaged capsules still con- tained two embryos (35 and 41 mm), both of which had been punctured numerous times through the capsule membrane. Sixty-eight undamaged capsules did not contain embryos. None of the 65 undamaged capsules in the right uterus contained embryos. However, this uterus contained an intact 41 mm em- bryo with an egg capsule fragment within the stomach of the largest embryo (320 mm). 100 MM EMBRYO (IRCZM 103137, Fig. 14B).— This male embryo was found in the right uterus of a 271.5 cm adult 0. taurus captured 18 July 1976. It had well-developed fin rudiments and a particularly well-developed caudal fin. The gill slits were large and without external filaments. Both upper and lower labial furrows were prominent. The yolk sac was absent although an attachment scar was present. Erect teeth, more slender than in previous embryos, were present in multiple rows. The teeth lacked lateral secondary basal cusps (basal denticles) typi- cal of adult 0. taurus. The teeth of this embryo were obviously functional because punctured and torn egg capsules and a damaged (tooth-marked) 51 mm em- bryo were found in the same uterus. 131 MM EMBRYO (IRCZM 103103, Fig. 14C).— A male embryo, from an 8 July 1978 sand tiger shark, 1 cm Figure 14.— (A) A 51 mm Odontaspis taurus embryo attacked and damaged by (B) a 100 mm male embryo inside the uterus of a 271.5 cm female captured 18 July 1976 (both IRCZM 103137). (C) A 13 lmm male embryo (IRCZM 103103) taken from the uterus of afemale captured 8 July 1978. This embryo had attacked and damaged the 49 mm embryo shown in Figure 1 1. 216 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS resembled the 100 mm embryo, except that all fins but the pectorals were similar to those of the adult and the gut was more distended with yolk. This em- bryo had attacked the 49 mm (Fig. 11) and 45 mm embryos present in the same uterus. 227 AND 271 MM EMBRYOS (IRCZM 103101, Fig. 15A, B).— The 227 mm female and larger 271 mm male embryo came from a 29 July 1977 sand tiger shark. The snout was narrow and had lengthened, resembling that of the adult as did other anatomical features, including the fins. In both embryos the en- tire digestive tract and abdominal wall were dis- tended from the consumption of yolk. Many broken egg capsules were also found within the uteri. 334 MM EMBRYO (IRCZM 103135, Fig. 15C) .— This was a female embryo from a 5 August 1976 sand tiger shark. The stomach was distended with yolk. Many "adultlike" features were apparent. This em- bryo contained four smaller embryos (9-36 mm) in its pharynx. Late Gestation, Postcannibalistic, Oophagous, Preparturition Period (September-March; 334-1,000 mm) After fertilization of 0. taurus ova has ceased and all other developing embryos have been consumed by the surviving embryo, unfertilized ova become the primary source of nutrition. This transitional period begins in August-September when embryo lengths reach 330-340 mm. Embryonic growth and development rates are rapid during this period (Fig. 15C, Table 3). A 330 mm em- bryo in September may attain 650-890 mm by late October or early November and 830-970 mm by late November. During this period the embryo consumes large quantities of yolk and a length of 1.0 m may be reached in December (Figs. 15D, 16). Embryos reaching 1.0 m are near parturition which may take place between December and March, after a gesta- tion period of 9-12 mo. A maximum size of 1.2 m TL may be reached before birth (Cadenat 1956). A 272 FIGURE 15.— Four specimens of embryonic Odontaspis taurus showing progressive abdominal distention from consumed yolk: (A) A 227 mm female embryo (IRCZM 103101) from the right uterus and (B) a 271 mm male embryo (IRCZM 103101) from the left uterus, of a female cap- tured 29 July 1977; (C) a 334 mm female embryo (IRCZM 103135) taken from a female captured 5 August 1976; and (D) an 80-100 cm embryo. 217 FISHERY BULLETIN: VOL. 81. NO. 2 Table 3.— Postparturition growth [total length (TL) and total weight] of two captive juvenile Odontaspis taurus from observations made by F. G. Wood at Marineland Inc., St. Augustine, Fla. NR = not recorded. Male Female TL Weight TL Weight Date (cm) (kg) (cm) (kg) Born 15 Feb. 1959 NR NR NR NR 17 Feb. 106 6.2 NR NR 9 Oct. 126 12.6 NR NR 29 Dec. 137 5 19.1 139 19.1 30 Aug. 1960 NR NR 145 NR 1 2 Dec. NR NR NR 37.5 28 Dec. NR NR 167 NR 16 June 1961 NR NR 175.5 40.7 died 17 Mar. 1962 167 5 NRdied 1 Mean growth rate (TL) 1.62 cm/mo 2.03 cm/mo 19.44 cm/yr 24.36 cm/yr '37 mo old. claspers extended 7.5 cm past pelvic fin tip. cm 0. taurus female was captured 10 April 1946 and kept in an aquarium for 11 mo; it died on 9 March 1947. Her autopsy revealed two decomposing near- term embryos 103-105 cm TL (6.1 and 6.4 kg) (McBride 1947 10 ; Springer 1948). The oophagous stage in development is preceded by an increase in ovary size, ovulation rate, number of ova per capsule, and number of Type II capsules pro- duced. The number of ova per capsule increased to a maximum of 23 ova/capsule during the fall and win- ter (Fig. 5D). During late gestation the embryos swallowed such great quantities of yolk that their stomachs became greatly distended. Cadenat (1956) found 1.5 kg of yolk (18.8% total body weight) in a near-term 0. taurus embryo weighing 8 kg. This dis- tention of the abdomen has precipitated the term "yolk stomach" used by earlier authors, particularly for the oophagous embryos of Lamna nasus ("Dot- termagen" of Lohberger 1910). The distention of the embryonic stomach declines in the final days near parturition. At birth the young O. taurus do not have excessive amounts of yolk within the digestive tract. We examined a 91.0 cm, 3.75 kg dead female pup (Fig. 17) from a 240 cm female O. taurus held captive since 21 August 1980, in a display tank ("Shark Encounter") at Sea World of Orlando. The pup died immediately after birth on 22 March 1981. The stomach and intestine of the newborn shark were not distended with yolk, although yolk was present. Another pup, born simul- taneously with the other uterus, lived and is presently on display (April 1983). Simultaneous to the decline in yolk consumption is an increase in the size of the embryo's liver. The left and right lobes of the liver of the specimen from Sea World of Orlando measured 20.3 and 23.7 cm, re- spectively, with a total liver weight of 372 g(9.9% of to- tal body weight). Cadenat (1956) found the liver of a near-term embryo to be relatively large, contributing 6.43% of the total body weight, in a 110 cm specimen. The large liver in the near-term embryo compares favorably with the largest liver recorded in adults at 7.54% total body weight (Cadenat 1956). A similar condition of large liver size and reduced yolk con- sumption has been observed in a near-term oophagous embryo (97 cm TL) of Isurus paucus (Gilmore in press). The increase in size of the embryo's liver corre- sponds to an observed decline in maternal ovarian ac- tivity and ovary size near the end of gestation (Springer 1948). The liver of the pregnant near-term female sand tiger shark also reaches a minimum size at this time (2.88% total body weight, Cadenat 1956), revealing the maximum uilization of the adult's nutri- tive materials to support the two large, ravenous embryos. Nutritional supplies stored within the embryo's liver can then be utilized during the last few days of gestation and after birth preceding the first capture of prey. The surviving newborn female 0. taurus from Sea World of Orlando did not eat until 25 d after birth. She first ate (two pieces of clam) a day after she attacked and killed another small shark (Diakis semifasciata, Frank Murru"). After the initial feeding the young sand tiger shark ate dead clams, squid, and fish (blue runner, Caranx crysos, sardines, herrings, "smelt", and mackerel) during daily feeding periods. Fortunately O. taurus has been kept in captivity for extended periods (up to 10 yr, 2 mo; R. van der Elst 12 ). Several births have taken place both in a South African aquarium (van der Elst footnote 12) and American aquaria (Wood footnote 9; Murru foot- note 11). Wood (footnote 9) made the following ob- servations of the birth of O. taurus pups in an aquarium at Marineland, St. Augustine, Fla., on 15 February 1959 from a female captured 1 1 November 1958 (Fig. 18): "The head of the first pup was first observed about 0945 extending3 to 4 inches (7. 6 to 10.2 cm] from the cloaca. The head came out a little further during the next 30 minutes. The pup was born c. 1015. '"A. F. McBride, formerly with Marineland Inc., St. Augustine, Fla., pers. commun, 8 Nov. 1947 to Stewart Springer, Mote Marine Lab.. Sarasota, FL 33577. "F. Murru, Curator of Fishes, Sea World of Orlando, FL 32809, pers. commun. 1981. |: R. van der Elst, S. Afr. Assoc. Mar. Biol. Res., Durban, South Africa, pers. commun. 1977. 218 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS FIGURE 16.-Two views of an Odontaspis taurus embryo (80- 1 00 cm) dissected from a dead female, showing extent of prepar- turition yolk consumption. Note adultlike color pattern on embryo. Measurements not taken. (Photos courtesy of Marine- land Inc., St. Augustine, Fla.) 219 FISHERY BULLETIN: VOL. 81, NO. 2 . SHE Figure 17. (Upper) Lateral view of a 91.0 cm female Odontaspis taurus (IRCZM 103182) born 22 March 1981 at Sea World of Orlando, Fla.; (lower) view of dentition of same empryo. "The female had been swimming between 5 and 8 feet [1.5 to 2.4 m] off the bottom in the center section. The pup was born c. 6 ft [1.8 m) above tjje bottom. It immediately swam off. The mother shark did not alter course or speed at the time the pup fell free. "Within less than a minute after the first pup was born, about 3 inches [7.6 cm] of tail appeared. The end of the tail disappeared 10 to 12 minutes later. Approximately 1 minutes later the tip of the sec- ond pup's snout emerged following 3 to 4 inches [7.6 to 10.2 cm] of the head. The head disappeared a few minutes later. It appeared from this and the distortions of the female shark's belly that the pup turned several times inside of her in the course of half an hour or so. "The tip of the tail appeared and disappeared again, then the snout began to emerge about an hour after the first pup had been born. This was followed by gradual emergence to [of] the head to 220 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS •■''-"■':' FIGURE 18.— Aquarium birth oWdontaspis taurus embryo, 15 February 1959, at Marineland Inc., St. Augustine, Fla. (A) Adult female with distended abdomen; (B) initial emergence of embryo snout; (C, D) inverted emergence of head to gill openings prior to completing birth. (Photos courtesy of Marineland Inc., St. Augustine, Fla.) about the second gill slit. For about 40 minutes the pup came no farther, then it gradually moved out to the origin of its pectorals. Five to 8 minutes later the mother abruptly speeded up and banked in the water with her belly outward. The pup popped out at 1233, rose to the surface, then came back to the bottom. "Both pups swam rapidly and rather erratically until caught ... ." Other births observed by Wood (footnote 9) were not so prolonged and were more difficult to analyze, e.g., a birth occurred on 30 December 1958, within 7 min following a cloacal discharge. Complete emer- gence of the embryo took 2-3 s. Regardless of the length of birthing time, embryos have been consist- ently observed to emerge headfirst. This is in con- trast to recent observations of tail-first births of carcharhinoids [e.g., Carcharhinus milberti (Wass 1973); Sphyrna mokarran (Mooney 1975); Galeocer- do cuvieri (Bravo 1980)]. Increase in length and weight after birth in captivity can be seen in Table 3. Newborn 0. taurus gain con- siderable weight during the first few months. A 106 cm, 6.2 kg pup born on 15 February 1959 was 137.5 cm and 19.1 kg by 29 December 1959. This same pup survived in captivity until 17 March 1962. Notes taken by Wood (footnote 9) point out that this specimen, a male, appeared to be nearing sexual maturation. At an age of 37 mo and length of 167.5 cm (Table 3) the shark's claspers extended 75 mm past the pelvic tips and the "general appearance" of the testes indicated the shark was becoming sexually ma- ture. Our observations indicate males are mature when at least 191.5 cm (see Observations and De- scriptions section). These data indicate that western Atlantic 0. taurus may mature earlier than South Af- rican specimens which were found to first mature at lengths of 220 cm (Bass et al. 1975). South African observations of captive 0. taurus indicate that "maturity is attained after about 8 years in the females . . . although the five year old male that we have is not far from maturity" (van der Elst footnote 12). Our pregnant females from the east coast of Florida ranged in size between 236.6 and 274.3 cm TL. These sizes are within the range of 240-272 cm for pregnant South African female 0. taurus (Bass et al. 1975). 221 FISHERY BULLETIN: VOL. 81, NO. 2 DISCUSSION AND SUMMARY Reproduction in Odontaspis taurus is typified by the occurrence of both synchronous group and syn- chronous individual physiological activities. Unisex- ual male and female groups converge on a mating ground, and intersexual behavioral activities such as biting (i.e., typically male biting female) may serve as a precopulatory release mechanism (Springer 1967; Stevens 1974). Over several years some variation is apparent, but the simultaneous presence of several females in a similar reproductive state off the Florida east coast indicates a definite seasonality for reproductive activity. After mating, the oviducal glands produce six basic types of egg capsules. Capsules without ova are pro- duced initially, suggesting that oviducal gland activi- ty precedes ovulation. Ova-laden egg capsules are produced during the latter half of gestation, prin- cipally as a food source for the remaining embryo in each uterus. The synchronous occurrence of egg capsules of the same type in the oviduct and the variation in ova numbers per capsule could be partially explained by three hypothetical physiological mechanisms, por- tions of which have been documented in various elasmobranchs: 1) Extrinsic stimuli may cause the pituitary gland to secrete hormones which eventually cause ovarian ova to maturate. (Removal of the pituitary in Scliorhinus caniculus prevents ovulation, Dodd et al. 1960.) During the period of ova maturation, luteal tissue may form (TeWinkel 1950; Chieffi 1967) and could possibly secrete hormones which initiate oviducal gland activity preceding ovulation. E gg cap- sules would then be produced initially without ova. TeWinkel (1950) similarly deduced that in Mustelus canis, ". . . it is not unlikely, therefore, that ovarian hormones present at the time of ovulation or slightly preceding it, stimulate the secretion of a single egg- case by each oviducal gland irrespective of the num- ber of ova discharged." Sperm would have to be stored if mating activity were the extrinsic stimuli af- fecting the pituitary and if ova maturation took some time. Although we have not documented if or where sperm is stored in 0. taurus, the most likely location would be the oviducal gland which has been shown to be the site for sperm storage in other elasmobranchs (Metten 1939; Prasad 1945; Pratt 1979). 2) Extrinsic stimuli may cause the pituitary to se- crete hormones which eventually cause ovarian ova to maturate and, in addition, directly affect oviducal gland activity. Steroid sex hormones (e.g., estrogen) have been shown to directly affect the secretory ac- tivity of the oviduct in Squalus caniculus (Hisaw and Abramowitz 1938; Dodd et al. 1960; Simpson et al. 1963). Mobilization of egg capsule production in the oviducal gland may take less time than ova matura- tion, therefore producing egg capsules without ova. 3) Sperm arriving at the oviducal gland may stimu- late the gland to secrete ovalbumin and collagen cap- sules preceding pituitary hormone release. However, pituitary hormones and/or luteal hormones may maintain ovarian and oviducal gland activity through gestation. The staggered development of the 0. taurus em- bryos indicates that sperm had been stored for 2-4 mo, and either fertilization of some ova took place as late as July and August or development of fertilized capsules was somehow delayed. Embryonic development may be divided into sever- al phases within the developmental periods already discussed, based on anatomical characteristics and nutritive strategies (Fig. 19) Encapsulated early embryos derive nutrition from internal coelomic yolk supplies, although a yolk sac and stalk are pres- ent. The presence of yolk sacs 6.0 mm in diameter or larger in embryos 13-57 mm demonstrates little ap- parent change in the external yolk supply during a period of extensive growth and differentiation within the egg capsule. In the 13 mm embryo, external con- sumption of other encapsulated ova is improbable, 1000 CANNIBALISTIC PHASE - YOLK SAC - UTERINE FLUID PHASE HATCHING 123456789 10 11 MONTH FIGURE 19. — Embryonic growth curve and nutritional phases in development of Odontaspis taurus. 222 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS because cellular differentiation and organ formation were still in a primitive phase of development. When they have developed sufficiently to consume external food, larger early embryos (20-63 mm) may consume other ova contained within their own capsule. Therefore, following the consumption of internal, en- docoelomic yolk, the embryo may enter another nu- tritional phase while still encapsulated. These obser- vations suggest that initial internal coelomic yolk supplies and other encapsulated ova and albumin contribute more to initial embryonic growth and dif- ferentiation in embryos 49-57 mm TL than does the yolk of their own yolk sac. Although several blas- todiscs and ova are observed in a single capsule, only one embryo develops indicating that the activity of one blastodisc somehow reduces or arrests the activi- ty of other blastodiscs. After developing functional teeth and hatching at 49-63 mm, the embryo may utilize a variety of nutri- tive sources. It is possible that intrauterine fluid, as well as the yolk remaining in the yolk sac, may be a food source. The 62 and 63 mm specimens still possessed a 5.5 mm diameter yolk sac and well- developed branchial filaments. Uterine fluid was found to increase in volume after hatched embryos were found. It is possible that this fluid may be ab- sorbed through the extensive branchial filaments found in these embryos. However, these filaments also may have a respiratory function. Of the many anatomical features observed in the developing em- bryos, the presence of a filament attached to the cor- nea of the 57 mm embryo was among the most in- teresting. Its presence on the cornea suggests a res- piratory rather than a nutritive function. The normal- ly high metabolic demand of retinal tissue suggests that there may be a need for such a filament. After the embryo hatches, the yolk sac eventually declines in size demonstrating the utilization of this nutritive source. Uterine fluids were observed to in- crease in volume when newly hatched embryos were present. This fluid could also be consumed by the embryo. Activity of the hatched embryo within the uterus may cause uterine hormones to induce in- creased ovarian activity, since ovulation rates and uterine yolk capsules increase after the first embryo hatches. Other embryos also developing in some of these capsules were not attacked when hatched em- bryos were only 17-40 mm larger than encapsulated embryos. The size advantage of a hatched 63 mm em- bryo over a 46 mm encapsulated embryo may not be great enough for an active attack, even though the potential prey is restricted in movement due to its en- capsulation. The first embryo to hatch apparently does not begin to hunt for and detect other encap- sulated embryos until it reaches about 100 mm in length. Initially only those capsules containing em- bryos are attacked, while up to 81 capsules without embryos are undamaged. Attacks are made by puncturing and cutting the capsule membrane with teeth. These attacks may also puncture and tear the embryo within the capsule, as we found punctured, dead embryos still encapsulated. The encapsulated embryo that was attacked is probably consumed later after the capsule is eventually opened by repeated attacks from the larger embryo. It is apparent from these data that the first embryo to hatch and reach a length approximating 100 mm would be most likely to survive. By the time the em- bryo reaches a length of 227-340 mm, during August and September, it will have consumed its in- trauterine competitors. If the embryo first to develop dies in utero before consuming all other embryos, the next largest embryo will probably become the domi- nant predator and continue the developmental pat- tern. The two 320 and 334 mm embryos from 5 August 1976 had consumed other embryos and also contained 7.5-9.0 g of yolk in their stomachs. After reaching 300-400 mm and having consumed all smaller embryos, the embryo begins attacking egg capsules which contain 7-23 unfertilized ova. In most cases the capsules were not consumed but were torn open near the posterior portion of the capsule and the ova or gelatinous material had been removed. Em- bryos 131 mm or greater in length were found to con- tain varying quantities of yolk in both their stomachs and valvular intestines. The embryo increases significantly in size (i.e., from 334 to 1 ,060 mm) by consuming uterine yolk supplies and uterine fluid. After the embryos reach a length of about 1.0 m and weights of 3.8-10.0 kg, parental ovarian activity is reduced, stomach yolk content of the embryo declines, and its liver increases in size. After 9-12 mo of gestation, birth occurs. Teeth in the newborn 0. taurus are well developed, extending beyond the gums (Fig. 17B). The teeth in the newborn 91 cm female pup we examined had well-developed lateral tooth denticles typical of adult specimens. However, Taniuchi(1970) reported no O. taurus <100 cm with lateral tooth denticles. Although only two young are produced at the end of a lengthy gestation period, they have several selec- tive advantages as top predators in marine food webs. The newborn sand tiger sharks are large at birth and are comparable in size to many common adult neritic predators (e.g., scombrids and caran- gids). They are also larger than the young of most other galeoid sharks (45-60 cm, Wourms 1977). Their larger size as a top predator also allows a 223 FISHERY BULLETIN: VOL. 81, NO. 2 greater range of available prey for consumption. The predation rate on young 0. taurus will be lower as few fish are larger. A similar argument has been made by Wourms (1977) for the selective advantages of viviparity in chondrichthyan fishes in general. However in 0. taurus, not only is the near-term em- bryo quite large but also it is conditioned in utero to hunt, attack, and consume prey. At birth they are "ex- perienced young" (Springer 1948). The young sand tiger sharks, one from each uterus, having already killed for survival before birth, may have a selective advantage during competitive interactions with other interspecific predators of similar age or size (except possibly other lamnoid and some galeoid sharks). The advantage in interspecific competition may have been demonstrated, although under cap- tive conditions, in the lethal attack of a 25 d-old 0. taurus pup on Triakis semifasciata. ACKNOWLEDGMENTS We would like to thank Stewart Springer, Charles Richardson, Kenneth Moore, Frank Murru, Rolando Cavazos, and John Collins for donating specimens. F. G. Wood of the Naval Ocean Systems Center, San Diego, provided valuable notes on his observations of 0. taurus held in aquaria at Marineland Inc., St. Augustine, Fla., while he was employed there. Stewart Springer graciously made available his cor- respondence on reproduction of the sand tiger sharks and also made comments on the manuscript. Robert Jenkins, Curator of the Marineland Research Laboratory, provided photographs taken at Marine- land Inc. of the sand tiger shark embryos and parturi- tion. Frank Murru, Curator of Fishes, Sea World at Orlando, kindly provided notes on the condition, feeding, and general activity of captive 0. taurus specimens in the Shark Encounter exhibit. Rudy van der Elst of the South African Association for Marine Biological Research, Ocean Research Institute, Dur- ban, provided information on captive specimens and photos of embryos. Don Hoyt of the Florida Shark Club, Inc., Jacksonville, Fla., made available the club's landing records for 0. taurus. Robert Jones of the Harbor Branch Foundation and two anonymous reviewers made creative and helpful comments on the manuscript. LITERATURE CITED Bass, A. J., J. D. D'Aubrey, and N. Kistnasamy. 1975. Sharks of the east coast of southern Africa. IV. The families Odontaspididae, Scapanorhynchidae, Isuridae, Cetorhinidae, Alopiidae, Orectolobidae and Rhiniodon- tidae. Oceanogr. Res. Inst. (Durban), Invest. Rep. 39, 102 p. BlGELOW, H. B., AND W. C. SCHROEDER. 1948. Lancelets, Cyclostomes, and Sharks. In A. E. Parr (editor), Fishes of the western north Atlantic, Part 1, 576 p. Mem. Sears Found Mar. Res., Yale Univ. 1. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Branstetter, S. 1981. Biological notes on the sharks of the north central Gulf of Mexico. Contrib. Mar. Sci. 24:13-34. Bravo, R. 1980. Upstaging the film stars: tiger shark gives birth. Sea Front. 26:170-171. CADENAT, J. 1956. Note d'ichtyologie ouest-africaine. XIV.— Remarques biologiques sur le Requin-sable Carcharias (Odontaspis) taurus Rafinesque 1810. Bull. Inst. Fr. Afr. Noire 18:1249-1256. CHIEFFI, G. 1967. The reproductive system of elasmobranchs: Develop- mental and endocrinological aspects. In P. W. Gilbert, R. F. Mathewson, and D. P. Rail (editors), Sharks, skates and rays, p. 553-580. Johns Hopkins Press, Bait. Clark, E., and K. von Schmidt. 1965. Sharks of the central gulf coast of Florida. Bull. Mar. Sci. 15:13-83. Coles, R. J. 1915. Notes on the sharks and rays of Cape Lookout, N.C. Proc. Biol. Soc. Wash. 28:89-94. Dodd, J. M., P. J. Evennett, and C. K. Goddard. 1960. Reproductive endocrinology in cyclostomes and elas- mobranchs. Symp. Zool. Soc. Lond. 1:77-103. GlLMORE, R. G. In press. Observations on the embryos of the longfin mako, Isurus paucus, and the bigeye thresher, Alopias super- ciliosus. Copeia 1983: GUDGER, E. W. 1940. The breeding habits, reproductive organs, and external embryonic development of Chlamydoselachus based on notes and drawings left by Bashford Dean. In E. W. Gudger (editor), Bashford Dean memorial volume - archaic fishes, Art. 7, p. 521-646. Am. Mus. Nat. Hist, N.Y. Hisaw, F. L., and A. A. Abramowitz. 1938. Physiology of reproduction in the dogfishes, Mustelis canis and Squalus acanthius. Woods Hole Oceanogr. Inst. Rep. 1938, p. 22. HUMASON, G. L. 1972. Animal tissue techniques. 3d ed. W. H. Freeman Co., San Franc, 641 p. Lohberger, J. 1910. Uberzwei reisuge embryonen von Lamma (Beitra zur Naturgeschichte Ostasiens). Abh. Bayer, Akad. Wiss. 4(Suppl. 2):l-45. METTEN, H. 1939. Studies on the reproduction of the dogfish. Philos. Trans. R. Soc. Lond., 230 (Ser. B):217-238. MOONEY, M. J. 1975. Hammerheads born in captivity. Sea Front. 21:359- 361. Prasad, R. R. 1945. The structure, phylogenetic significance, and function of the nidamental glands of some elasmobranchs of the Madras coast. Proc. Inst. Sci. India 11:282-302. 224 GILMORE ET AL.: REPRODUCTION AND EMBRYO DEVELOPMENT OF SAND TIGER SHARKS Pratt, H. L., Jr. 1979. Reproduction in the blue shark, Prionaceglauca. Fish. Bull.. U.S. 77:445-470. SADOWSKY, V. 1970. On the dentition of the sand shark, Odontaspis taurus, from the vicinity of Cananeia, Brazil. Bolm Inst. Ocean- ogr. S. Paulo. 18(l):37-44. Simpson, T. H., R. S. Wright, and S. V. Hunt. 1963. Sex hormones in fish. Part II. The oestrogens of Scyliorhinus caniculus. J. Endocrinol. 26:499-507. Smith, B. G. 1942. The heterodontid sharks: Their natural history and the external development of Heterodontus (Cestracion)japon- icus based on notes and drawings by Bashford Dean. In E. W. Gudger (editor), Bashford Dean memorial volume - archaic fishes, Art. 8, p. 651-770. Am. Mus. Nat. Hist., N.Y. Springer, S. 1938. Notes on the sharks of Florida. Proc. Fla. Acad. Sci. 3:9-41. 1948. Oviphagous embryos of the sand shark, Carcharias taurus. Copeia 1948:153-157. 1960. Natural history of the sandbar shark, Eulamia milber- ti. 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 survival, p. 95-113. D. C. Heath and Co., Boston. 1967. Social organization of shark populations. In P. W. Gilbert, R. F. Mathewson, and D. P. Rail (editors), Sharks, skates and rays, p. 149-174. Johns Hopkins Press, Bait. Stevens, J. D. 1974. The occurrence and significance of tooth cuts on the blue shark (Prionoce glauca L.) from British waters. J. Mar. Biol. Assoc. U. K. 54:373-378. Taniuchi, T. 1970. Variation in the teeth of the sand shark, Odontaspis taurus (Rafinesque) taken from the East China Sea. Jpn. J. Ichthyol. 17:37-44. TeWinkel, L. E. 1950. Notes on ovulation, ova, and early development in the smooth dogfish, Mustelus canis. Biol. Bull. (Woods Hole) 99:474-486. 1963. Notes on the smooth dogfish, Mustelus canis, during the first three months of gestation. II. Structural mod- ifications of yolk-sacs and yolk-stalks correlated with in- creasing absorptive function. J. Exp. Zool. 152:123- 137. Wass, R. C. 1973. Size, growth, and reproduction of the sandbar shark, Carcharhinus milberti, in Hawaii. Pac. Sci. 27:305-318. WOURMS, J. P. 1977. Reproduction and development in chondrichthyan fishes. Am. Zool. 17:379-410. 225 COPEPODS AND SCOMBRID FISHES: A STUDY IN HOST-PARASITE RELATIONSHIPS Roger F. Cressey, 1 Bruce B. Collette, 2 and Joseph L. Russo 2 ABSTRACT Host specificity of the copepods parasitic on scombrid fishes is the basis for an analysis of the host-parasite relationship. A total of 46 different species of parasitic copepods were collected from 47 species of Scom- brinae (the monotypic Gasterochismatinae is excluded). A revised host-parasite list is presented, including new data by R. F. Cressey and H. B. Cressey. Those copepod species present on more than one host species have preferred hosts, and indicate tendencies to being host specific. The copepods present an American species of Scomberomorus suggest evolutionary trends in that group. Two species (ancestral S. cavalla and ancestral S. sierra) were probably present prior to the separation of the Atlantic and Pacific Oceans. The present Atlantic S. maculatus andS. brasiliensis arose from aS. sierra ancestor. Copepod data suggest that the Indo-West Pacific S. eommerson is the most primitive extant species, while S. multiradiatus is the most advanced. The copepods parasitic on Sarda species indicate the origin of that genus in Australasia, with the Atlantic S. sarda being the most advanced species. The genus Allothunnus, previously regarded as a member of the tribe Sardini, is shown to have affinities with the Thunnini and may be the most primitive member of that tribe. A cladistic analysis of the copepod genus Unkolax correlates well with current hypotheses of the phytogeny of scombrid genera. Host-parasite relationships of the Scombrinae are compared with those found in a previous study of host-parasite relationships in needlefishes (Belonidae). Parasite- based host phytogenies follow the methods of Brooks. In this paper we test the validity and application of several parasitological theories regarding host-para- site relationships of copepods parasitic on scombrid fishes. As in our earlier joint effort (Cressey and Collette 1970), in which we treated the relationships of parasitic copepods and needlefishes, the analyses are enhanced by the collaboration of specialists re- presenting each animal group (Cressey — parasitic copepods, Collette and Russo — scombrid fishes). Parasite taxonomy on which the present paper is based has been published separately (Cressey and Cressey 1980). Additional material collected since that publication and an updated list of hosts and copepods, because over 200 additional scombrids have been examined, are included in this paper. Examples of 10 genera of copepods are illustrated (Fig. 1) to indicate the kinds of copepods that para- sitize scombrids. Because many earlier reports on parasitic copepods contain misidentifications of both host and para- site, we rely on our own collections or direct examina- tion of specimens used in published accounts. The often repeated "Fahrenholz rule" (Noble and Noble 1973:548) suggests that related parasites 'National Museum of Natural History, Smithsonian Institution, Washington, DC 20560. 2 Systematics Laboratory, National Marine Fisheries Service, NOAA, Smithsonian Institution, Washington, DC 20560. Manuscript accepted August 1982. FISHERY BULLETIN: VOL. 81, NO. 2, 1983. are found on related hosts, thus indicating host phy- logeny. This generalization we now know is an oversimplification. Hennig (1966:109-110) illustrated how it is pos- sible to have the same parasite species on hosts of polyphyletic origin through incomplete parallelism. Cautions on the use of parasites as indicators of host phylogeny, echoed by Mayr (1957), Hennig (1966), Noble and Noble (1973), and others, are well-found- ed. We feel, however, that these problems can be minimized by studying comprehensive collections of both hosts and parasites, using the maximum num- ber of parasite groups on the hosts. Presence of parasites on any host may reflect host ecology, chorology, or phylogeny. We believe that information on host-parasite phylogeny has increased validity as sample size, and the numbers of parasite species from different parasite groups (Crustacea, Trema- toda, Protozoa, etc.) available for study increases. When a parasite group is taxonomically well under- stood, it can be treated as a host character with as much validity as host morphology, serology, and ecology. Objections or reservations regarding the parasite approach to host phylogeny raised by Mayr (1957) and Hennig (1966) are based on studies or examples, using a relatively small number of parasite species, usually within one parasite taxon (genus or family). If, however, one repeats the analysis of the same hosts 227 FISHERY BULLETIN: VOL. 81, NO. 2 using numerous parasite groups, the parasite taxa that do not parallel the host phylogeny are likely to become apparent. Another parasitological theory we have tested is "Szidat's rule," which suggests that primitive (gen- eralized) parasites are found on primitive hosts and that advanced (specialized) parasites are found on advanced hosts. We provide an example supporting this concept when we consider the scombrid host preferences of the copepod genus Unicolax on scom- brid hosts (p. 254). SAMPLING ADEQUACY AND HOST SIZE Before considering host specificity, it is necessary to know whether enough hosts were examined to pro- vide samples of all species of the usual parasite fauna. Individual collections of copepods from each scombrid species were recorded on cards sequential- ly, enabling us to consider the question: "How many specimens of a host species should be examined before all parasitic copepod species are likely to have been collected?" Examples are given in Ta- ble 1. Table 1 . — Number of specimens that had to be examined in order to find all known copepod species. Specimens Total examined No. of specimens until all copepod Species examined collected spp. Scomberomorus commerson 130 53 9 Scomberomorus sierra 116 12 3 Sarda sarda 106 35 4 Euthynnus aff/nis 74 44 8 Auxis spp. 68 60 6 Scomberomorus concolor 47 2 3 Of the six species presented in Table 1, the two spe- cies of Scomberomorus endemic to the eastern Pa- cific (S. sierra and S. concolor) required a relatively small number of individuals to be examined (2-12 specimens), until all parasitic copepods were collect- ed. Wider ranging species (S. commerson, Sarda sar- da, Euthynnus affinis, and Auxis spp.) required FIGURE 1. — Examples of copepods parasitic on scombrids: a) Uni- colax anonymous, female; b) Holobomolochus asperatus, female; c) Shiinoa inauris, female and males; d)Caligus bonito, female; e) Ely- trophora brachyptera, female; f) Gloiopotes hygomianus , female; g) Tuxophorus cybii, female; h) Pseudocycnus appendiculatus , female; i) Pseudocycnoid.es armatus, female; j) Brachiella thynni, female and dwarf male attached. examination of a greater number of specimens (35- 60) before we collected all of their copepod species. The two endemic species have fewer species of parasitic copepods than the nonendemic species. Other scombrids with restricted distributions (Scomberomorus multiradiatus , S. sinensis, and S. munroi) also have fewer parasitic copepod species than related species with wider distributions. When collecting parasitic copepods from hosts with wide distributions, specimens must be examined from throughout the range. We found that the num- ber of parasite species is usually less at the periphery of the host's range, so that conclusions relative to to- tal parasite fauna for a species cannot be based on geographically limited collections. We also examined the relationship between host size and infestation density in order to determine its importance in sampling adequacy. It is generally ac- cepted that larger individuals of host species usually support a greater parasite fauna, both in number of species and individuals. Although little work has been done on the ectoparasite fauna in relation to host size (age), Dogiel et al. (1961:9) noted an in- crease in the numbers of Ergasilus sp. on the gills of Esox lucius on larger fish. Cressey and Collette (1970) found that specialized copepods (those pos- sessing holdfasts or that are very host specific) are found mainly on larger needlefish, while generalized copepods (less host specific and not highly modified) are found most often on smaller needlefish indi- viduals. In the present study, copepods of the families Pseudocycnidae, Bomolochidae, and Shiinoidae parasitic on three species of Scomberomorus were considered (Table 2). We chose these copepod spe- cies for the study because they remain attached in preserved specimens. Pseudocycnids (Fig. lh, i) are firmly attached to gill filaments; bomolochids (Fig. Table 2. -Infestation densities of Scomberomorus commerson, S. maculatus, and S. brasiliensis for three copepod groups, Pseudocycnoides, Bomolochidae, and Shiinoa. Range No. of Pseudocycnoides Bomolochidae Shiinoa of hosts No. of % No. of % No. of % (mm FL) hosts parasites density parasites density parasites density 100-200 32 64 2.0 31 1.0 201 -300 47 202 4.3 89 19 1 0.02 301 -400 17 35 2.1 11 0.7 5 0.3 401 + 16 34 2.1 8 0.5 25 1.6 228 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES 229 FISHERY BULLETIN: VOL. 81, NO. 2 la, b) are in the nasal sinuses and can only be collect- ed by cutting open the nares; shiinoids (Fig. lc) are firmly attached to lamellae of the nasal rosettes. Other copepods, such as caligids, are not as firmly at- tached, and many specimens are undoubtedly lost during handling and preservation of the hosts. The Scomberomorus species were represented by a rea- sonable number of specimens with adequate size- range coverage. The apparent optimum size for infestation by the two species of pseudocycnids and the two bomolo- chids is between 201 and 300 mm FL (fork length). Infestations of Pseudocycnoides armatus and P. buc- cata seem to remain at the same levels (about 2 per fish) in groups with smaller and larger size in- dividuals with about twice that infestation rate in the optimum size range. Infestations of the bomolochids Unicolax ciliatus (from S. commerson) and Holobo- molochus divaricatus (from S. maculatus and S. brasiliensis) apparently decrease with increased host size after 300 mm FL; no Scomberomorus over 500 mm FL examined was parasitized by bomolochids. The two species of Shiinoa (S. inauris horn Scomber- omorus brasiliensis and S. maculatus and Shiinoa occlusa from Scomberomorus commerson), on the other hand, are not found on smaller fish, and the greatest infestation rate occurs on fish over 400 mm FL. The change in infestation rate with host size in some of these parasite species may be due primarily to mechanical factors. In order for female pseudo- cycnids to remain attached to the gill filaments, the lateral lobes of the cephalon must partially encircle the filament. Until a prospective host reaches an opti- mum size, the filament may be too small for the adult copepod to secure itself. As the host fish grows, the filaments may become too large for the parasite to re- main attached. Two very large S. commerson (1,115 and 1,150 mm FL) from New South Wales, Australia, were parasitized by several P. armatus. These cope- pods were considerably longer than average for the species (8.1 vs. 4.9 mm), which may account for their ability to infest a larger size host. Shiinoa attaches to its host by piercing a nasal lamella with its recurved second antennae which are opposed by an elongate and recurved rostrum. The combination results in a ring through the lamella, with the rest of the parasite hanging free. It may be necessary for the host to at- tain a minimum size (275 mm FL in our data) before the lamella is large enough to accommodate the parasite. (Shiinoa males attach to female copepods rather than the host.) The presence of bomolochid species on 100-200 mm hosts cannot be as easily correlated with mechanical 230 factors. Bomolochids are not firmly attached to their hosts. Those species considered here are found loose within the nasal sinuses and are capable of moving about possibly as scavengers more than as true para- sites. Possibly the reduction in infestation of bomo- lochids in larger fish is associated with the increased presence of Shiinoa in the nasal sinuses of hosts larger than 300 mm. ECOLOGICAL RELATIONSHIPS To determine the influence of ecological relation- ships as opposed to phylogenetic host specificity of parasitic copepods found on scombrids, we examin- ed the literature records of parasitic copepods from fishes with habits similar to those of scombrids (large size, open ocean, fast swimming, predatory, etc.). We compiled data for the following fish groups: Billfishes (Istiophoridae and Xiphiidae), sharks, A lepisaurus, Lampris, Coryphaena, several genera of Carangidae, Rachycentron, Pomatomus, and the gempylids, Ru- vettus and Thyrsites. We have tried to use discretion in evaluating the reliability of literature records. For example, Bere (1936) reported Caligus bonito from Pomatomus sal- tatrix, Lutjanus griseus, Mugil cephalus, Oligoplites saurus, Scomberomorus maculatus, and S. cavalla. She indicated in her report that the copepod material was identified by C. B. Wilson. The first author of this paper examined the specimens, deposited in the Smithsonian (USNM 79090), in order to verify the Pomatomus record. Bere presumably sent Wilson the material separated by host. Wilson apparently put together all specimens that he identified as a single species. The collection contains about 15 Caligus specimens with no host names and represents three species — Caligus bonito, C. mutabilis, and males of a third species. It is impossible to verify the occurrence ofC. bonito onPomatomus, and the record must be ig- nored. Another record (Capart 1959) of C. pelamy- dis from Pomatomus is questionable because Capart's illustration does not appear to be of C. pela- mydis. Eliminating unreliable reports leaves C. cor- yphaenae, a relatively distinct species, as the only copepod common on scombrids which also occurs on many ecologically similar species. It has been record- ed from the following nonscombrid genera: Caranx, Elagatis, Coryphaena, Xiphias, Squalus, Seriola, Isurus,Echeneis, and Sphaeroides. There have been a few reports of Caligus productus and C. pelamydis from nonscombrid hosts, but both of these copepods have been often confused with closely related spe- cies. Rohde (1980) reported C. pelamydis from 3 of 88 specimens of Trachurus trachurus and 22 of 122 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES specimens of Scomber scombrus with C. pelamydis from Helgoland (these copepod identifications were verified by G. Boxshall of the British Museum (Natural History)). As the record shows, most species of copepods com- mon on scombrid hosts are restricted to scombrids. Caligus coryphaenae apparently is the only common scombrid parasitic copepod whose host choice is in- fluenced by ecological rather than phylogenetic factors. There is evidence that in some cases the presence of a species of parasitic copepod on two or more host species which are not closely related may be the result of an association between the hosts. The para- sitic copepod Pumiliopes jonesi (= P. capitulatus) is common on the eyes of scombrids of the tribe Scom- brini {Rastrelliger and Scomber) and on the clupeids Clupanodon punctatus and Herklotsichthys dis- plonotus. Both groups are filter-feeding schooling fishes. Another example is Caligus macaroui (= C. fulvipur- pureus) common on the Pacific saury, Cololabis saira (Hotta 1962), but reported on Auxis as well by Gussev (1951). Cololabis feeds primarily on plank- tonic crustaceans with eggs and larvae of fishes form- ing secondary diet items (Hotta and Odate 1956; Taka et al. 1980). Auxis feeds on a wide variety of small fishes, cephalopods, and planktonic crus- taceans (Uchida 1981). We are unaware of any rec- ords of Auxis preying on Cololabis, but sauries are common food items of billfishes. HOST SPECIFICITY Host specificity is concerned with the predilection of a parasite species for one or a few species of host or hosts. The comprehensive data on which this study is based demonstrate host specificity. The occurrence of a species of parasite in a variety of host species does not necessarily imply a lack of host specificity. Careful analysis of collection data with reference to percent of host individuals parasi- tized by a particular parasite species will usually show that one or a few host species are heavily infest- ed, some occasionally infested, and some rarely in- fested with the parasite species. Dogiel et al. (1961) referred to these groups as main, secondary, and ac- cidental hosts. Holmes (1979) referred to the three groups as required hosts, suitable hosts, and unsuit- able hosts. Holmes considered required and suitable hosts as those with which the parasite can develop to maturity (or to an infective stage in intermediate hosts), and unsuitable hosts as those with which the parasite cannot develop, but may be transported to a suitable or required host. Not enough is known of the life histories of most parasitic copepods to evaluate their state of "well being" on respective hosts. Col- lection data, however, indicate that species found on several host species vary in infestation rate in ways suggesting the host categories of Dogiel et al. and Holmes. In addition, unpublished data based on par- asitic copepod collections by the first author from fishes of the Gulf of Mexico indicate the same cate- gories of infestation. The recently published revised data on the parasitic copepods of scombrids (Cressey and Cressey 1980) enable us to compare data based on a synoptic re- view of literature records of copepods parasitic on scombrids (Silas and Ummerkutty 1967) with a sur- vey based solely on verified host and parasite iden- tifications (Cressey and Cressey 1980). We have used the same format as that of Holmes and Price (1980) except we have considered specificity at the generic level rather than the family level (our data are based only on the Scombridae). Comparisons of the two analyses (Tables 3, 4) point out the inadequacies of an unverified data base. Data based on the literature survey of Silas and Ummer- kutty (1967) indicate that 60% of the copepod spe- cies parasitic on scombrids are specific to 1 genus, 5 % to 2 genera, 1 1 % to 3 or 4 genera, 2 % to 5 or more genera, and 23% were also recorded from nonscom- brid hosts. The data based on Cressey and Cressey (1980) and additional records in this paper indicate 54% specific to 1 genus, 18% to 2 genera, 9% to 3 or 4 genera, 9% to 5 or more genera, and only 9% are also found on nonscombrids. Clearly, the latter is a better TABLE 3.— Host specificity of scombrid copepods based on data from Silas and Ummerkutty (1967). Number of host species infested No. of Scombrid genera and infested 1 2 3-4 5-8 nonscombnd 1 28 3 3 2 2 1 3-4 3 3 5+ 1 Nonscombr ids 13 Table 4.— Host specificity of scombrid copepods based on data from Cressey and Cressey (1980) and later. Number of host spec ies i nfested No. of genera infested 1 2 3-4 5-8 9+ Scombrid and nonscombnd 1 11 6 2 4 1 2 3 2 3 3-4 5+ 2 1 1 4 Nonscombrids 4 231 FISHERY BULLETIN: VOL. 81, NO. 2 index to host specificity at the generic level than that based solely on literature. The Silas and Ummer- kutty data indicate a higher specificity at the level of 1 genus of host; they also indicate a higher percent- age of "generalists" (36% with 3 or more genera plus nonscombrids). Furthermore, the Cressey and Cres- sey and later data indicate a gradual transition from greater to lesser host specificity, whereas the data based on Silas and Ummerkutty do not. Comparison of percent specificity (percent species with only one host, see Price 1980:123) shows a wide range of specificity per genus of scombrid copepod parasites (Table 5). Specificity to a genus of hosts seems more meaningful to us, so we have also cal- culated these figures. Six of the seven families that contain scombrid copepod parasites show relatively high percent specificity at the generic level (50-75%) while the Caligidae is distinctly lower (35%). Scombrinae The subfamily Scombrinae is composed of two groups of two tribes. The more primitive mackerels (Scombrini) and Spanish mackerels (Scomberomo- rini) have a distinct notch in the hypural plate, lack any bony support for the median fleshy caudal pe- duncle keels, and do not have the penultimate verte- bral centra greatly shortened. Scombrini The tribe Scombrini contains the two genera of mackerels, Scomber and Rastrelliger. Mackerels have small conical teeth and a large number of gill rakers. Characters differentiating the two genera have been given by Matsui (1967:table 4). Copepod fauna: 9 species in 7 genera. Bomolochid copepods can be separated into two subgroups based on the presence of one or two major setae (in addition to the remainder of the normal complement) on each caudal ramus. The genera found on Scomber and R. brachysoma (Pumilopes, Orbitacolax, and Nothobo- molochus) are members of the group with one major terminal seta. Although members of this same cope- pod subgroup are found on other fish families, none are found on other scombrids. This host specificity of some members of that subgroup to the Scombrini distinguishes the true mackerels from the other scom- brid tribes. Pumiliopes jonesi is the only copepod found in both genera of Scombrini and nowhere else, occurring in the orbits of two species of each genus. The infestation rate in Rastrelliger was 13%, in Scom- ber only 2%. TABLE 5.— Percent specificity (percent species with only one host) and percent generic specificity (percent species with hosts only in one genus) in genera of copepod parasites of scombrid fishes. Percent No. of Percent generic Copepod genus species specificity specificity Bomolochidae (12) (33) (58) Holobomolochus 3 33 100 Unicolax 5 20 75 Ceratocolax 1 Nothobomolochus 1 100 Orbitocolax 1 100 100 Pumiliopes 1 Shunoidae (2) (0) (50) Shiinoa 2 50 Caligidae (12) (17) (35) Caligus 12 17 35 Euryphoridae (4| (75) (75) Elytrophora 2 50 50 Cloiopotes 1 100 100 Caligulus 1 100 100 Tuxophoridae (3) (67) (67) Tuxophorus 3 67 67 Pseudocycmdae (4) (25) (75) Pseudocycnus 1 Pseudocycnoides 3 33 100 Lerneopodidae (4) (25) (75) Brachiella 2 Clavellisa 1 100 100 Clavellopsis 1 100 100 Scomber Linnaeus We follow most recent authors (Fraser-Brunner 1950; Collette and Gibbs 1963; Matsui 1967) in con- sidering Pneumatophorus a synonym of Scomber. Scomber differs from Rastrelliger in a number of ana- tomical characters which have been summarized by Matsui (1967:table 4). Copepod fauna: 5 species in 4 genera. Only the lerneopodid Clavellisa scombri is restricted to Scomber, occurring on gills of Scomber japonicus and S. australasicus in our material. It was originally described from a host identified as S. scom- brus from Trieste, but we failed to find it in 97 speci- mens of that species. Matsui (1967) recognized three species oiScomber. S. scombrus Linnaeus in the North Atlantic and Med- iterranean; S. australasicus Cuvier in the western Pacific from Japan to southern Australia east to the Hawaiian Islands, and across the eastern Pacific bar- rier to Socorro Island off Mexico; and S. japonicus Houttuyn, a worldwide antitropical species. All the copepod species known from the three species have been found on S. japonicus, of which we have examin- ed about 500 specimens. Rastrelliger Jordan and Starks Matsui (1967:table 4) summarized the diagnostic characters of Rastrelliger. Copepod fauna: 5 species in 5 genera. Pumiliopes jonesi and two other bomo- 232 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES lochids were found in two species of Rastrelliger, 0. aculeatus in the orbits, and N. kanagurta on the gills. Matsui (1967) recognized three species of Rastrel- liger: R. faughni Matsui from Taiwan, the Philippine Islands, Indonesia, and western India; R. brachysoma (Bleeker) in the same general area of the western Pacific asi?. faughni but extending east to Fiji; andR. kanagurta (Cuvier) which is widespread throughout the Indo- West Pacific from Taiwan, the Philippines, Samoa, and Australia east throughout the Indian Ocean to Madagascar and the Red Sea. At least one individual has gone through the Suez Canal into the eastern Mediterranean Sea (Collette 1970). All but one of our copepod records are from R. kanagurta and R. faughni. Our only lernanthropid was a female Lernanthropus kanagurta from a Bornean specimen of/?, brachysoma. This is probably not a usual scom- brid parasite (Cressey and Cressey 1980:45). Scomberomorini This is the most speciose tribe in the family, con- taining 20 of the 48 species. Most of these (18 spe- cies) belong to Scomberomorus, the Spanish mack- erels and seerfishes; the other 2 species belong to the monotypic genera Acanthocybium and Grammator- cynus. Copepod fauna: 25 species in 8 genera. The copepod genus most characteristic of the Scomber- omorini is Shiinoa, found attached to the nasal ro- settes of Acathocybium, Grammatorcynus, and 10 species of Scomberomorus. (Shiinoa was also found on one specimen of Gymnosarda, but we do not be- lieve Gymnosarda is a usual host for this copepod.) Scomberomorus Lacepede Scomberomorus differs from the other two genera in the tribe, Acanthocybium and Grammatorcynus, by usually lacking a swim bladder. The genus is com- posed of 18 species (Collette and Russo 1 980). There is one species in the Gulf of Guinea and Mediterra- nean Sea — S. tritor (Cuvier); four in the western At- lantic — cavalla (Cuvier), regalis (Bloch), maculatus (Mitchill), and brasiliensis Collette, Russo, and Za- valla-Camin; and two in the eastern Pacific — con- color Lockington and sierra Jordan and Starks. The remaining 1 1 species are in the Indo- West Pacific: guttatus (Bloch and Schneider); koreanus (Kishinou- ye); lineolatus (Cuvier); plurilineatus Fourmanoir; commerson (Lacepede); sinesis (Lacepede); semifas- ciatus (Macleay); queenslandicus Munro; niphonius (Cuvier); munroi Collette and Russo; and multiradiatus Munro. Copepod fauna: 23 species in 7 genera. In addition to two species of Shiinoa, Scom- beromorus is commonly parasitized by the pseudo- cycnid genus Pseudocycnoides (buccata, armatus, scomberomori) , the bomolochid genera Holobomolo- chus (diuaricatus , asperatus, nudiusculus) , and Uni- colax (U. ciliatus), and several species of Caligus (especially C. biseriodentatus, C. infestans, and C. cybii in the Indo-West Pacific, C. mutabilis and C. productus in the western Atlantic, and C. omissus in the eastern Pacific). The speciose nature of Scom- beromorus and its copepod parasites requires further discussion, by regions. ORIGINS AND EVOLUTION OF AMERICAN SCOMBEROMORUS.— Six species of Scomber- omorus occur in American waters. (Figs. 2, 3). Two of these, S. sierra and S. concolor, are restricted to the eastern Pacific from about lat. 10° to 40°N. Scom- beromorus concolor presently occurs only in the Gulf of California. The four Atlantic species areS. cavalla, found from about lat. 30° S to 45°N; S. brasiliensis, a southern coastal species (Belize to southern Brazil); S. maculatus, a northern coastal species (Yucatan to Massachusetts); and S. regalis, a largely insular species (most abundant in the Bahamas and West Indies). The six species of American Scomberomorus are parasitized as a group by the following species of copepods: H. asperatus (S. cavalla), H. nudiusculus (S. sierra, S. concolor), H. divaricatus (S. brasiliensis, S. maculatus, S. regalis), Shiinoa inauris (Scom- beromorus maculatus, S. brasiliensis, S. regalis), C. mutabilis (S. cavalla, S. brasiliensis, S. maculatus), C. omissus (S. sierra, S. concolor), and P. buccata (all species mentioned in this paragraph). To use parasitic copepods as indicators of host phy- logeny we determined the pleisiomorphy-apomor- phy of certain taxonomic characters. This is possible within a closely related group of parasites based on reduction and modification of characters for parasi- tism. It seems reasonable to assume that, as species of a parasite group evolve, the later (more recent) species are more specialized or reduced than the old- er species. If we assume that hosts and parasites evolve together, the information on the evolution of one group should provide evolutionary information about the other group. Four genera of copepods para- sitic on Scomberomorus lend themselves to analysis and are discussed below. Three species oiHolobomolochus parasitic on Amer- ican species of Scomberomorus and a fourth species from Caranx hippos form a subgroup of the genus (see Cressey and Cressey 1980:8). In these species, 233 FISHERY BULLETIN: VOL. 81, NO. 2 FIGURE 2.— Distribution of Atlantic and eastern Pacific species of Scomberomorus. the last exopod segments of legs 2-4 bear a number of plumose setae, heavily sclerotized spines, and short- er nonplumose setae with armature intermediate to that on spines and setae. Long plumose setae (adap- tations for free swimming) are primitive, whereas heavily sclerotized spines (adaptations for attach- ment) are advanced characters. The three Holobomo- lochus from Scomberomorus show a transition in the numbers of each of these character states. Holo- bomolochus asperatus (parasite of S. cavalla) bears 1 8 long plumose setae and 7 sclerotized spines on the last exopod segments of legs 2-4. The same append- ages of H. nudiusculus (on eastern Pacific Scom- beromorus) bear 16 plumose setae, 2 intermediate setae/spines, and 7 spines. The same appendages of H. divaricatus (on all western Atlantic Scomberomo- rus except cavalla) bear 14 setae, 4 intermediate setae/spines, and 7 spines. This transition in de- creased numbers of long plumose setae and increase in intermediate setae/spines within these three para- site species suggests H. asperatus to be the most primitive, H. nudiusculus intermediate, and H. di- varicatus to be most advanced. If the hosts reflect the phylogeny of the parasites, then this suggests that S. cavalla is the most primitive; the two eastern Pacific species — S. sierra and S. concolor — are interme- diate; and the three western Atlantic species — S. regalis, S. maculatus, and S. brasiliensis — are the most advanced of the American species of Scomber- omorus. Holobomolochus has 23 currently recognized spe- cies in the western Atlantic and eastern Pacific and 1 species from the eastern Atlantic (a species from In- dia is not a Holobomolochus, as reported by Pillai 1973). Unicolax ciliatus, a species of another bomo- lochid genus, is found on 9 species of Scomberomo- rus in the Indo-West Pacific and on S. tritor in the eastern Atlantic. Four remaining species of Unicolax, including Atlantic and eastern Pacific species, are found only on non-Scomberomorus scombrids. This parasite distribution and host affiliation suggest that Holobomolochus was already well established on American Scomberomorus before the appearance of Unicolax in this area. Based on the evidence that U ciliatus has not undergone further speciation on 10 Scomberomorus species despite the geographic isola- tion of one of those species (S. tritor from the eastern Atlantic) and the presence of Holobomolochus on the American Scomberomorus, it can be assumed that Holobomolochus is older than Unicolax. 234 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES FIGURE 3. — Distribution of Scomberomorus cavalla, S. commerson, and S. sinensis. Unicolax appears to be more advanced than Holo- bomolochus by possessing a heavily sclerotized mod- ified seta on the first antenna and having 1 seta rather than 2 setae on the mid-endopod segment of leg 3. The highly modified copepod genus Shiinoa (Shiin- oidae) is comprised of three species: Shiinoa occlusa from Indo-West Pacif ic A canthocybium, Grammator- cynus, Scomberomorus, and Gymnosarda and the eastern Atlantic S. tritor; Shiinoa inauris from west- ern Atlantic Scorn beromorus (except S. cavalla); and Shiinoa elagatis from Indo-Pacific Elagatis (Caran- gidae). The first author is describing a fourth spe- cies from the Indian Ocean jack, Caranx malabaricus. Of the three described species S. elagatis with 3- segmented rami of legs 1 and 2 is the most primitive. Shiinoa occlusa from Indo-West Pacific scombrids is intermediate with 3-segmented rami of legs 2 and 3 but with fewer spines and setae and reduced body segmentation compared with S. elagatis. Shiinoa in- auris from three of the four western Atlantic Scom- beromorus (all except S. cavalla) is most advanced with only 2 segments in the exopods of legs 2 and 3 of the females and 2 segments in both rami of legs 2 and 3 of the males. Infestations by the western Atlantic S. inauris and its speciation probably did not occur until after the last geologic separation of the eastern Pacific. On scombrids, Shiinoa has differentiated into only two species. Although this genus is recorded from 1 spe- cies of Scomberomorus, the highest rates of infesta- tion among scombrid hosts are in Grammatorcynus and Gymnosarda. Shiinoa occlusa, from Indo-West Pacific scombrids, is more primitive than the western Atlantic S. inauris, indicating the latter' s probable derivation from Indo-Pacific stock. The presence of the highly specialized siphono- stome copepod parasite, P. buccata, on all species of American Scomberomorus with relatively high infes- tation rates (30- 63%) indicates that this parasite was present before the separation of Atlantic and eastern Pacific Oceans, but, in spite of the present isolation, the two populations have not differentiated (unlike the three Holobomolochus species). From this it appears that dispersal and some specia- tion of American Scomberomorus occurred prior to their being parasitized by bomolochid and shiino- id copepods. The evidence derived from an analysis of the cope- pods parasitic on the six American Scomberomorus species suggests the following sequence of events: 235 FISHERY BULLETIN: VOL. 81, NO. 2 1. During the period when the eastern Pacific and Atlantic Oceans were continuous, two species of Scomberomorus were probably present, an ancestral S. cavalla and an ancestral S. sierra. Both of these were infested with species of Holobomolochus and P. buccata. 2. As the land mass of Central America separated the Atlantic from the Pacific, the two ancestral forms were divided into four populations. The Atlantic pop- ulation of S. cavalla persisted while the Pacific pop- ulation disappeared. The Pacific S. sierra population persisted and gave rise to S. concolor, while the At- lantic population subsequently divided into a south- ern species, S. brasiliensis, and a northern species, S. maculatus. The derivation of S. regalis was also prob- ably from a sierra ancestor. The origin of pre-caval- la and pre-sierra populations was probably derived from the Indo-Pacific S. commerson line and the S. tritor line, respectively (Fig. 4). FIGURE 4. — Tentative cladogram of the Scomberomorini. Numbers refer to morphological characters from Collette and Russo (text footnote 3). 3. The population of ancestral S. sierra in the At- lantic differentiated to produce ultimately the north- ern coastal species S. maculatus and the southern coastal species S. brasiliensis and insular S. rega- lis. 4. Some species of copepods differentiated as either new host species were formed, or populations of related hosts were isolated. 5. An additional genus {Shiinoa) of parasitic cope- pod became established on three of the Atlantic spe- cies of Scomberomorus (brasiliensis, maculatus, and regalis) after the formation of a land barrier separat- ing the eastern Pacific from the Atlantic. The ab- sence of Shiinoa on Scomberomorus cavalla may indicate that S. cavalla, derived from the S. commer- son line, may have occupied the Atlantic prior to the parasitization of scombrids by Shiinoa. The later infestations of Shiinoa in the western Atlantic may have been derived from Scomberomorus tritor and consequently occur only on the three western Atlan- tic species of Scomberomorus derived from the tri- tor line. Based on the anatomy of Scomberomorus, the American species belong to different species groups. Scomberomorus cavalla is the western Atlantic re- placement for S. commerson, which is widespread in the Indo-West Pacific. The other five American species, plus S. tritor from the eastern Atlantic, form the S. regalis species group (Fig. 4), defined by the presence of nasal denticles (Collette and Russo manuscr. in prep. 3 ). These five American species share a unique specialization of the fourth left epi- branchial artery (Collette and Russo footnote 3), which indicates that these species were derived from an 5. tritor ancestor. This pattern of relationships is fully compatible with that derived from the cope- pod data. INDO-WEST PACIFIC SCOMBEROMORUS. - There are 11 recognized species of Indo-West Paci- fic Scomberomorus (Collette and Russo 1980; Figs. 3, 5, 6). Four genera of parasitic copepods are common on Indo-West Pacific species of Scomberomorus (Ta- ble 6): Unicolax, parasitic in the nasal sinuses; Shiinoa, attached to the nasal lamellae; Pseudocycnoides , at- 3 Bruce C. Collette and Joseph L. Russo. Systematics and mor- phology of the Spanish mackerels (Scomberomorus). Manuscr. in prep., 400 p. Systematics Laboratory, National Marine Fisheries Service, NO AA, Smithsonian Institution , Washington , DC 20560. Table 6. — Infestation of Indo-West Pacific species of Scomberomorus with parasitic copepods. Host species arranged from most infested (most primitive'.') to least infested (most specialized?). The eastern Atlantic & tritor is included for comparison. Total copepod Total Common Species n species genera genera commerson 130 9 6 4 semifasciatus 26 5 4 4 queenslandicus 39 5 4 4 guttatus 58 4 4 4 plunltneatus 14 5 5 4 wphonius 19 4 4 4 munroi 19 3 3 3 koreanus 6 4 2 2 Itneoiatus 14 3 3 3 sinensis 10 3 2 1 muitiradiatus 29 2 2 2 tritor 21 4 3 3 ^Unicolax, Pseudocycnoides. Shiinoa, Callous. 236 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES FIGURE 5. — Distribution of Scomberomorus lineolatus, S. plurilineatus, S. munroi, and S. niphonius. tached to the gill filaments; and several species of Caligus, found in the gill area, mouth, and on the body surface. The generally accepted theory that the more primi- tive members of a host group usually harbor more species of parasites than those that evolved later in- dicates the following. Scomberomorus commerson is the most widespread species occurring from the east- ern Mediterranean (recent Suez migrant) eastward throughout the Indian Ocean into the western Pacific Ocean (see Figure 3). Nine species of copepods, from four genera cited above plus two additional genera (Ticxophorus and Brachiella), have been collected from S. commerson. No other species of Scom- beromorus harbors more than seven species and six genera of copepods. Thus, the parasite data indicate S. commerson to be the most primitive member of the Indo- West Pacific Scomberomorus. If the converse is true, the data suggest that S. multiradiatus with only two copepod species is the most advanced (special- ized). The data further suggest that the origin of S. com- merson was in the Indo-Australian Archipelago, be- cause all nine species of copepods are reported from specimens in that area with a decrease in the num- ber of parasite species to the north and west (Fig. 7). Scomberomorus niphonius is unusual among the Indo- West Pacific members of the genus in its cope- pod parasites. Most Indo-West Pacific Scomberomo- rus are parasitized by P. armatus. Scomberomorus niphonius is commonly parasitized by a closely re- lated species, P. scomberomori, which has more pri- mitive characters than P. armatus, and is apparently specific to S. niphonius. This suggests that S. ni- phonius may be primitive compared with the other Indo-West Pacific species. Scomberomorus nipho- nius might also be considered primitive based on one of its morphological characters (Fig. 4). It is the only species in the genus to have a straight intestine. Most other species of Scomberomorus have two bends (and three sections) to the intestine. One species, S. koreanus, has three bends (and five sections), pre- sumably a specialized condition. Two of the 19 specimens of S. niphonius were para- sitized by C. pelamydis (the only Caligus so far reported from it) which is found on several other scombrids, most commonly on species of Sarda. Caligus cybii, closely related to C. pelamydis, has been reported from sue Indo-West Pacific species of 237 FISHERY BULLETIN: VOL. 81, NO. 2 S. queenslandicus _2£_ 150 FIGURE 6. — Distribution of Scomberomorus guttatus.S. koreanus,S. multiradiatus, S. semifasciatus, and S. queenslandicus. Scomberomorus, including species whose ranges overlap those of S. niphonius, S. koreanus, and S. sinensis. The first author cannot ascertain with cer- tainty which of these two copepods, based on their morphology, may be the more primitive, but the re- duced specificity of C. pelamydis and the apparent restriction of C. cybii to Indo- West Pacific Scom- beromorus suggest C. pelamydis to be more primitive. If true, this supports the indication of the primitive nature of S. niphonius provided by the two species of Pseudocycnoides. A single specimen of C. pelamydis has also been collected by us from S. sinensis. This might be used to argue thatS. commerson andS. niphonius arose from a common ancestor, with S. niphonius now restricted to the northwest Pacific (colder water) and S. com- merson, together with other species, occupying the more temperate and tropical waters. Scomberomorus commerson and S. sinensis both have prominent dips in the lateral line, but the dip is under the second dor- sal finlets in the former species and under the first dorsal fin in the latter species; this similarity may be due to convergence rather than close relationships. These three species (S. commerson, S. niphonius, and 5. sinesis), S. cavalla, and S. queenslandicus all ap- pear to be relatively primitive (Fig. 4). Grammatorcynus Gill Although included in the Scomberomorini by re- cent works such as Collette (1979), the exact sys- tematic position of this monotypic genus is in doubt (Collette and Russo 1979), because it also shares some characters with the Scombrini. It has the same number of vertebrae as do the Scombrini (3 1), usual- ly 13 precaudal plus 18 caudal. Its possession of an extra, ventral lateral line is unique in the family. The double-lined mackeral, G bicarinatus (Quoy and Gaimard) is known from much of the tropical Indo- West Pacific, particularly near coral reefs from the Marshalls and Carolines, Philippine Islands, Aus- tralia, and the East Indies east to the Red Sea. Copepod fauna: 5 species in 2 genera, Shiinoa and Caligus. Only one species of Caligus, C. asymmet- ricus, is at all common on Grammatorcynus (14.9%). This copepod has been found on nine scombrids in the Indo- West Pacific and is perhaps more charac- teristic of the Sardini {Cybiosarda elegans, Sarda 238 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES Figure 7. — Numbers of copepod species on Scomberomorus commerson in different areas of its distribution (large numbers represent number of copepod species; small numbers represent number of hosts examin- ed). orientalis, S. australis) with infestations of 8-12%. Acanthocybium Gill This monotypic genus appears to be a specialized offshoot of Scomberomorus and does not merit place- ment in its own subfamily or tribe as has been ad- vocated by some previous authors (e.g., Starks 1910). It is closest to the Cyhium group of Scomberomorus (S. cavalla andS. commerson), according to Conrad (1938) and MagoLeccia (1958). The wahoo, A. solan- dri (Cuvier), is a large species (reaching over 1,500 mm SL) and has a well-developed swim bladder. It is a high-seas epipelagic species found round the world in tropical and subtropical waters. Copepod fauna: 6 species in 5 genera. A canthocybium is similar to the other Scomberomorini in being parasitized by Shi- inoa and Tuxophorus, but the rate of infestation is very low. The most common two copepods are the eu- ryphorid Gloiopotes hygomianus (infestation rate of 42% of our 64 specimens, 54% of the 100 fish from the Line Islands examined by Iverson and Yoshida 1957) and the lerneopodid Brachiella thynni (61% of our specimens, 98% of those examined by Iverson and Yoshida). The other four species of Gloiopotes are parasites of billfishes (Istiophoridae). Some workers in the past (e.g., Liitken 1880) and the present (G. David Johnson, pers. commun. 4 ) be- lieve that Acanthocybium is closely related to the billfishes. We feel that the parasite data are best in- terpreted as evidence of ecological similarity between the groups (fast swimming, high-seas species) rather than as evidence of phylogenetic relationships. Bra- chiella thynni was also found on three species of Thunnus (T. obesus, T. albacares, and T. thynnus) and two of Scomberomorus (S. regalis and S.plurilin- eatus). This species has been reported from a variety of hosts, usually attached in the axil of the pectoral fin. A second species of Brachiella is known only from two western Pacific species of Scomberomorus . There seems little ecological or phylogenetic information that can be drawn from parasitism by Brachiella. Parasitic copepods of the genera Tuxophorus and Gloiopotes suggest relationships between Scomber- omorus and Acanthocybium of the Scomberomorini 4 G. David Johnson, South Carolina Wildlife and Marine Re- sources Department, Charleston, SC 29412. 239 FISHERY BULLETIN: VOL. 81, NO. 2 and the Istiophoridae (Table 7). Three species of the copepod genus Tuxophorus are parasitic on the body surface of species of Scomber vmorus and Acantho- cybium in the Atlantic and Indo-West Pacific Oceans. When the paper by Cressey and Cressey (1980) went to press, these three species, T. cybii, T. cervicornis, and T. collettei, were retained in Tuxophorus because they conformed to the diagnosis of that genus. Subse- quent considerations by the first author lead to the conclusion that they are not members of Tuxophorus but represent a new genus closely related to Gloiopo- tes or are possibly members of Gloiopotes. The pre- sence of frontal lunules on these three species is the only character separating them from Gloiopotes, as it is presently defined. An earlier work on the parasitic copepods of lizardfishes (Cressey and Cressey 1979) gave an example of a caligid genus (Abasia), which showed a transition series of six species with a grad- ual reduction in the frontal lunule from well devel- oped to absent. This indicates the possibility that the presence or absence of the frontal lunule is not always a valid generic character. The genus Tux- ophorus was described by Wilson (1908) for T cali- godes, based on material collected from Atlantic Rachycentron canadus and Echeneis naucrates. The second species, T. wilsoni, was described by Kir- tisinghe (1937) from the carangid, Chorinemus, from Sri Lanka. Four of the five species of Gloiopotes are found on the body surface of various species of istiophorids; the fifth, G. hygomianus, is restricted to A, solandri. The occurrence oiGloiopotes on A canthocybium and istiophorids might be used as evidence to support re- lationships between the two groups. The question is: TABLE 7. — Host-parasite records for Tux- ophorus cybii, T, collettei, T. cervicornis, and Gloiopotes spp. Host-parasite Area Tuxophorus cybii Acanthocybium solandri Indian Ocean Tuxophorus cervicornis Scomberomorus commerson Indo-Pactfic Tuxophorus collettei Scomberomorus regalis Atlantic Gloiopotes hygomianus Acanthocybium solandri Cosmopolitan Gloiopotes amencanus Istiophorus amencanus Atlantic Gloiopotes ornatus Tetrapturus albidus Atlantic Makaira nigricans Atlantic Gloiopotes huttom Tetrapterus audax Indo-Pacific Makaira indicus Indo-Pacific Istiophorus platypterus Indo-Pacific Gloiopotes watsom Tetrapterus audax Indian Ocean Makaira nigricans Indo-Pacific Makaira indicus Indian Ocean Istiophorus platypterus Indo-Pacific "Are these relationships ecological or phylogenetic?" The morphological similarities between Acanthocy- bium and the Istiophoridae seem best explained as convergences; those between Acanthocybium and Scomberomorus indicate that A canthocybium is the specialized sister-group of Scomberomorus (Fig. 4). Thus, we argue that the presence of Gloiopotes on Acanthocybium and istiophorids is an ecological re- lationship, but that the occurrence of three species of Tuxophorus on Acanthocybium and Scomberomorus reflects shared phylogeny. Support for this argument could come from the presence oiGloiopotes on some open ocean, fast-swimming host but we have no such data. The explanation for the occurrence of species of Gloiopotes only on Acanthocybium and istiophorids must remain uncertain for the present. Sardini The bonitos consist of eight species placed in five genera (Collette and Chao 1975). Except for Allo- thunnus, the Sardini differ from the Thunnini in lack- ing prominent prootic pits on the ventral surface of the cranium. Collette and Chao (1975:table 14) sum- marized the characters distinguishing the five genera of Sardini. Copepod fauna: 11 species in 5 genera. Caligus bonito has been found on all. Unicolax collat- eral was found in Orcynopsis, Cybiosarda, and two species of Sarda. Orcynopsis Gill The monotypic Orcynopsis and Cybiosarda show several characters that distinguish them from Sarda and Gymnosarda (Collette and Chao 1975). Orcynop- sis is a short-bodied and short-headed bonito. Orcy- nopsis unicolor (Geoffrey St. Hilaire) is an eastern Atlantic endemic whose range is centered in the Med- iterranean Sea but extends south to Dakar, Senegal, and north to Oslo, Norway (Collette and Chao 1975: fig. 69). Copepod fauna: 1 specimen of U. collateralis and 1 specimen of Caligus bonito. Cybiosarda Whitley As noted above, the monotypic genera Cybiosarda and Orcynopsis share a suite of characters that differ- entiate them from Sarda and Gymnosarda (Collette and Chao 1975). Cybiosarda elegans (Whitley) is vir- tually an Australian endemic; is found along the northern three-quarters of the continent from Perth, Western Australia, to Sydney, New South Wales (Collette and Chao 1975:fig. 69); and occurs along the south coast of Papua New Guinea (Collette 240 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES 1979). Copepod fauna: 3 species in 2 genera, the same species as in Orcynopsis plus Caligus asymmet- ricus, which is found on various species in three of the four tribes. Sarda Cuvier The four species of Sarda all have several dorsal stripes, ranging from horizontal to oblique in orienta- tion. Sarda and Gymnosarda share a number of char- acters that distinguish them from Orcynopsis and Cybiosarda (Collette and Chao 1975). Collette and Chao (1975) recognized four species of Sarda (Fig. 8): Sarda australis (Macleay) is restricted to the east coast of Australia, Norfolk Island, and New Zealand; S. chiliensis inhabits the eastern Pa- cific where it is divisible into two subspecies, S. c. chiliensis (Cuvier) from Peru and Chile and S. c. lin- eolata (Girard) from Alaska to Baja California; S. orientalis (Temminck and Schlegel) is widespread in the Indo-Pacific from South Africa and the Red Sea east to Japan, China, the Philippine Islands, the Hawaiian Islands, and across into the eastern Pacific from Baja California to Peru; andS. sarda (Bloch) is found throughout tropical and temperate waters of the Atlantic Ocean including the Gulf of Mexico and the Mediterranean and Black Seas (Collette and Chao 1975; Fig. 8). A summary of the 26 most important characters used in distinguishing the species of Sarda was pre- sented by Collette and Chao (1975:table 17). Copepod fauna: 9 species in 3 genera. In addition to the two widespread bonito parasites, U. collateralis and Caligus bonito, three other copepods are com- mon on species of Sarda; Ceratocolax euthynni, Cal- igus pelamydis, andC. asymmetricus. The presence of five common copepods on species of Sarda presents an opportunity for further analysis. Over 200 specimens of the four species of Sarda were examined with an overall infestation rate of 7 5 % (156 of 206 specimens examined). It is thought that as a host species or related group of host species dis- perses from its place of origin it loses parasites in the process (see discussion of Scomberomorus commer- son above). When one examines the infestation rates of the individual Sarda species, first with all of its copepod parasites and secondly each species with its individual parasite species, the change in infestation rates from one Sarda species to another may reflect the speciation of Sarda species away from the center of origin of the genus. An analysis of these data (Table 8) indicates an origin of the genus in Australasia (S. australis, S. orientalis, or an ancestor of theirs) with the eastern Pacific S. chiliensis derived from S. australis and the Atlantic S. sarda from S. chiliensis. The infestation rates of C. bonito, C. asymmetricus, and U. collateralis suggest that the copepod parasites of S. sarda could have been derived from those of S. orientalis. The oc- currence of C. pelamydis onS. sarda, however, and its absence on S. orientalis reinforce the idea that S. sarda may have been derived, along with its para- sites, from S. australis or S. chiliensis but not from S. orientalis. Sarda sarda has the lowest overall infesta- tion rate (68%) and has lost one Caligus species (asymmetricus) and replaced U. collateralis with the Atlantic scombrid bomolochid copepod Ceratoco- lax euthynni. The overall infestation rates of the four species of Sarda are 8. australis, 90%; S. orientalis, 82%; S. chil- iensis, 76%; and S. sarda, 68%. These data support the proposal that species radiation progressed from Indo- West Pacific to eastern Pacific to Atlantic with- in the genus. The 26 morphological characters used by Collette and Chao (1975:table 14) to distinguish the species of Sarda tend to support the evolutionary hypothesis deduced from the copepod data. Sarda sarda is the most specialized of the four species in its increased numbers of vertebrae and other correlated meristic characters. Sarda australis appears most primitive in such characters as number of dorsal and anal finlets. It shares some primitive characters, such as the oc- casional presence of vomerine teeth, with S. sarda. If other similarities between these two species (loca- tion of first closed haemal arch, length of haemal pre- and postzygapophyses, shape of vertical wing of pelvic girdle, etc.) can also be considered primitive, then S. chiliensis and S. orientalis are in a relatively intermediate evolutionary position. In some cases, Table 8. — Infestation rates by four species of copepods on the four species of Sarda (arrows indicate direction of decrease). Sarda spe cies Copepod species orientalis australis chiliensis sarda Caligus bonito Caligus pelamydis Caligus asymmetricus Un/colax collateralis 36.4 12.1 36.4 - 59.1 500 9.1 9.1 _► 55.6 89 _ 31.1 7.5 241 FISHERY BULLETIN. VOL. 81, NO. 2 a -a Is <3 CO u c -c 3 ■S? e c o 3 £1 Q I W O 251 FISHERY BULLETIN: VOL. 81, NO. 2 3. S comber omorus niphonius is possibly the most primitive species of the genus based on Caligus and other copepod parasites (Pseudocycnoides scomber- omori, a copepod specific to ^ u o a a T3 o o c E o 1/1 o O < o >- D I/) O SCOMBRINI SCOMBEROMORINI SARDINI THUNNINI GASTEROCHISMATINAE SCOMBRINAE FIGURE 16. — Diagram of relationships of scombrid fishes modified from Collette and Russo (1979:fig. 1). S. Farris (following Farris 1970 and Farris et al. 1970). Infestation by a given copepod species was, somewhat arbitrarily, considered primitive; absence, specialized. A transformation series was used to indi- cate decreasing amounts of parasitism by a given copepod species across a matrix of scombrid genera. The Wagner tree was rooted at Rastrelliger, one of the most primitive members of the Scombrinae. The resulting Wagner tree (Fig. 17) shows major differ- ences from the diagram of relationships based on host morphology (Fig. 16). The only concordant sis- ter groups produced in this tree are A canthocybium and Scomberomorus. There are at least two problems with coding the in- festation data in this manner. Use of copepod species ignores information concerning the relationships of the species. Another difficulty is coding copepod in- festation as a two-state character (present or absent in a host species), when Caligus infestation data can only be interpreted as host preference (relative per- cent of infestation) rather than as host specificity (see previous section on Caligus). The program was rerun using infestation by genera of copepods and defining Caligus presence as more than 5% infestation to cor- rect for this problem. This Wagner tree (Fig. 18) is much closer to the diagram based on host morphol- ogy. Several concordant sister groups are present: Scomberomorus-Acanthocybium defined by the ac- quisition of Tuxophorus at node (5), Grammator- cynus- Scomberomorus + Acanthocybium defined by the acquisition of Caligus at node (4); Katsuwonus- Thunnus, loss of Ceratocolax at node (9); and Eu- thynnus-Katsuwonus + Thunnus, acquisition of Pseudocycnus at node (8). There are also several differences between this Wagner tree and the diagram of relationships based on host morphology. Gymnosarda is associated with Grammatorcynus-Acanthocybium group based on the presence of Shiinoa in all four genera. However, we found Shiinoa in only one specimen of Gymnosar- da, so not much reliance can be placed on this asso- ciation. We found only two other copepods on Gymnosarda, single occurrences of C. bonito and C. productus, which were omitted in this run of the pro- gram. There was only one common copepod on Allo- thunnus (Elytrophora), but there were also records of the same two species of Caligus as in Gymnosarda. Perhaps examining more specimens of Gymnosarda and Allothunnus (we examined only seven of each) would yield more copepods that would cluster these two genera with the natural group of the Sardini plus Thunnini. We turned from attempts at producing a cladistic classification of all scombrids, using the infestation data, and decided to use only a portion of the data, in- festation by the nasal bomolochids of the genus Uni- colax. The five known species of Unicolax are all parasites in the nasal sinuses of scombrid fishes. The first author compared characters within the species of Unicolax with those in the related outgroup genus 254 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES CD co "O CD CO O o to 3 C :», o o TO E E CD <3 a o c o 3 C C -C 3 C C 3 -C ^- E 3 .3 o E CO 1 o o o c o -U -a * c r- 3 CO u o o CO H 00 * i2 x 3 CO 3 cr c- co ^>- o c '-, .-. w UJ FIGURE 17. — Wagner tree of scombrid hosts basedon infestation by copepod species. Synapomorphies (gain, loss, or reduction in infes- tation rate of copepod species) occurred at the following nodes: 1) loss of Lernanthropus kanagurta, Orbitacolax aculeatus, andNo- thobomolochus kanagurta; 2) loss of Pumiliopes jonesi; 3) gain of Caligus asymmetricus and Unicolax collateralis; 4) gain of Caligus bonito; 5) gain of Caligus productus and C. coryphaenae; 6) gain of Brachiella thynni; 7) gain of Shiinoa occlusa and reduction of infesta- tion of Caligus productus; 8) reduction of infestation of Caligus asym- metricus; 9) gain of Caligus pelamydis, Unicolax mycterobius, U. collateralis, and reduction of infestation of C. productus and C. cory- phaenae; 1 0) gain of Caligus bonito, Ceratocolax euthynni, and reduc- tion of infestation of Unicolax mycterobius. Bomolochus. The eight characters used are as fol- lows: Number of setae on the exopod of leg 4 (many = plesiomorphic, few = apomorphic); presence or ab- sence of surface ornamentation on the abdomen and caudal rami (presence = plesiomorphic, absence = apomorphic); first exopod segment of leg 2 with long hairs or short spinules (hairs = plesiomorphic, spi- nules = apomorphic); number of setae on the first maxilla (4 = plesiomorphic, 3 = apomorphic); num- ber of setae on exopod last segment of leg 2 (5 = plesiomorphic, 4 = apomorphic); number of seg- ments in first antenna (7 = plesiomorphic, 6 = apo- morphic) ; endopod segments with a row of short hairs (plesiomorphic) or patch of fine spinules (apomor- phic); exopod spines of leg 2 with fine hairs (plesio- morphic) or mostly toothed (apomorphic). Phylogenetic relationships of the copepod para- sites of the genus Unicolax are represented in the branching diagram (Fig. 19), generated with the CO CD cc CO 3 C CO 3 E CO 3 CD :x u O F 5 CD CO w o o CO T3 CO 3 CD w 3 r 01 C CO CD CD o U CD C n -o 3 O b b Q -c O CO C 5 3 CO o § C E E o c CD c o O 5 CO CD -c ii -*- >, o O 3 CD 3 to •a. o (3 00 ^ u u H oo UJ *: CO 3 C C 3 -C FIGURE 18. — Wagner tree of scombrid hosts based on infestation by copepod genera. Synapomorphies (gain or loss of copepod genera) occurred at the following nodes: 1) Orbitocolax, Nothobomolochus, andLernanthropus; 2) loss of Pumilopes; 3) gain of Shiinoa; 4) gain of Caligus; 5) gain of Tuxophorus; 6) gain of Unicolax; 7) gain of Cera- tocolax and Caligus; 8) gain of Pseudocycnoides; 9) loss of Cerato- colax. Autapomorphies are A) gain of Clavellisa; B) gain of Elytrophora; C) gain olHomobomolochus, Unicolax, and Pseudocyc- noides; D) loss of Unicolax and gain of Elytrophora. 3 .CO 3 O | o c CD co .3 -Q O co CO 3 to "§ cb u FIGURE 19.— Cladogram showing relationships of Unicolax species. The nodes (1-4) represent the following: 1 = species of Unicolax; 2 = teeth on leg 2 exopod spines; 3 = endopod segments with patches of spinules; 4 = fewer than 5 setae on fourth leg exopod. WAGNER 78 program using characters of copepod morphology. The additive binary matrix of this tree is presented in Figure 20. Phylogenetic relationships of the scombrid hosts (Fig. 2 1) are adapted from Collette and Russo (1979) and represent a monophyletic sub- 255 ttwttwtthw*^^^ 1 2 3 4 5 6 7 8 9 1 1 2 1 3 1 1 4 1 1 1 5 1 1 1 6 1 1 7 1 8 9 FIGURE 20. — Additive binary matrix based on relationships of Unicolax parasites. to I 1 e .a to ■% .to 1 5 o to to o Q. A3 to ■c o o c 1 ■a f c 1 H o o CO o I 5 -C to 3 c X. s ■c to 3 to Uj v: c c •20- U L,J L_„_l L LJ J UJ ,J .19. ■21- I FIGURE 21. — Cladogram of scombrid hosts based on host morphol- ogy. Nodes 12-21 represent hypothetical ancestors. set of the Scombridae. The additive binary matrix of this tree is presented in Figure 22. In Figure 23 we have indicated the scombrid genera in the tribes Scomberomorini, Sardini, and Thunnini parasitized by Unicolax, based on the phylogeny of the Scombrinae proposed by Collette and Russo (1979). The copepod species are ranked from the most plesiomorphic (generalized) to the most apo- morphic (specialized), based on the Wagner tree of Unicolax (Fig. 19). As stated earlier, parasite phylogenies can be coded as characters and used to generate host trees; con- versely, host phylogenies can be coded as characters and used to generate parasite trees (Brooks 1981). In cases where a host has more than one parasite or a parasite has more than one host the character states for the two series are inclusively OR'd (Copi 1972) and a single series is used. By logically OR'ing two characters, a character state is said to be present in the union of two groups, if and only if it is present in one or both groups. For example, in Figure 20, Auxis harbors U. collaterals (2) and U. mycterobius (4). The character states for a host bearing U. collateralis can be determined by reading across line 2 of the additive character matrix, that is a one or logical true for states 2, 8, and 9 and not true for the others. The character states for a host bearing U. mycterobius can be deter- mined by reading across line 4, that is a one or logical true for states 4, 6, 7, 8, and 9 and not true for the others. Logically OR'ing the two rows of the matrix results in the character states 2, 4, 6, 7, 8, and 9 being true and the others being not true. Referring to the parasite tree (Fig. 19), these character states repre- sent the host, A axis, as having or having had during the course of its evolution (sensu lato) parasitic taxa (2) U. collateralis, (4) U. mycterobius, and hypotheti- cal ancestors (1), (2), (3), and (4). Proceeding in this manner for each host, a parasite (parasite ancestor) by host matrix is constructed. This matrix was subjected to cladistic character anal- ysis using the WAGNER 78 program for optimiza- tion. The resulting Wagner tree (Fig. 2 4) is rooted at a hypothetical host ancestor without Unicolax parasites. According to Brooks' (1981) methodology, this tree is an estimate of host phylogeny in lieu of host mor- phological data. It estimates host phylogeny based on phylogenetic events of their parasites. Because we have a host phylogeny based on morphological data, a direct comparison between the two trees is pos- sible. We attempt to explain the source of differences between the estimate of host phylogeny based on parasites and a cladogram based on host morphol- ogy. The most notable difference is that the base, node (5) of the host by parasite tree (Fig. 24), is formed by an unresolved multicotomy. This has resulted be- cause it is more parsimonious to assume that the four scombrid taxa, which lack Unicolax, never had them than to assume they were first acquired then lost. Node (4) is a subset of node (21) on the host phy- logeny (Fig. 21) and is based on a common Unicolax ancestor [node 1, (Fig. 19)]. Node (3) is a subset of node (19) on the host phylogeny and is based on the presence of ancestor (2) and parasite (2), U col- lateralis. An unresolved tricotomy is present at node (3) because the only parasite shared by the hosts Cybiosarda and Orcynopsis is U. collateralis, which is present below node (3) and is therefore treated as 256 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 1 1 1 2 1 1 3 1 1 1 4 1 1 1 5 1 1 1 6 1 1 1 7 1 8 1 1 9 1 1 1 10 1 1 1 1 11 1 1 1 1 12 1 13 1 14 1 15 1 16 1 17 1 18 1 19 1 20 1 21 FIGURE 22.— Additive binary matrix based on scombrid relationships. Numbers 1-11 are host taxa and 12-21 are hypothetical ancestors represented in Figure 21. synplesiomorphous. Node (2) is based on ancestors (2) and (4). Node (1) is an unresolved tricotomy and does not represent a subset of the host phylogeny because it includes Sarda. This node is based on the presence of parasite (4), U. mycterobius. Events which are not shared (autopomorphous) include the acquisition of (1), Unicolax ciliatus in Scomber- omorus; the acquisition of (5), U. reductus and the loss of (2); U. collateralis, in Katsuwonus; and the ac- quisition of (3), U. anonymous, in Euthynnus. The loss of the parasite U. collateralis in Katsuwanus is the only homoplasy in the host by parasite tree. The above procedure can also be used to generate a parasite phylogeny by using a data matrix construct- ed from information concerning host phylogeny. The parasite host tree (Fig. 25) is rooted at a non- scombrid ancestor based on the assumption that the common ancestor of Unicolax was from a nonscom- brid. This tree (Fig. 25) can be compared with the tree representing parasite phylogeny, which is based on an analysis of parasite morphological characters (Fig. 19). Node (4) on the parasite by host tree (Fig. 25) is comparable with node (1) on the parasite phy- logeny (Fig. 19). Unicolax ciliatus is the sister group of all other parasitic taxa in both trees. Node (3) of the parasite by host tree contains all elements of node (2) on the parasite phylogeny; however, U reductus is removed as the sister group of other taxa on the 257 FISHERY BULLETIN: VOL. 81, NO. 2 U ciliatus U- collaterahs U. anonymous j U mycterobius U. reductus [1] [2] [2] [5] [2,4] [2.4] [2.3,4] FIGURE 23. — Occurrence of species of Unicolax on scombrids in the tribes Scomberomorini, Sardini, and Thunnini. Copepods are rank- ed from most plesiomorphic (top) to most apomorphic (bottom). Scombrids are arranged to depict hypothesized phylogenetic re- lationships. parasite by host tree whereas U. collateralis occupies this position on the parasite phylogeny. This dis- crepancy occurs because parasites U. anonymous, U. collateralis, and U. mycterobius are all found on the host Euthynnus (9), at node (2) on the parasite by host tree (Fig. 25). Unicolax collateralis and U. myc- terobius are then grouped because they co-occur on host taxa 5 (Sarda) and 8 (Auxis) as well as nodes (16) and (18) of the host phylogeny (Fig. 21). Hypo- thesized hosts, which are not shared (autapomorph- ies), include Scomberomorus and host node (20) for U. ciliatus, Katsuwonus and host node (12) for U. reductus, and Orcynopsis, Cybiosarda, and host node (17) for U. collateralis. The parasite by host tree (Fig. 25) presents no homoplasy. If we make the assumption that the host and para- site phylogenies, which are based on morphological data, are both true, how do we explain the current dis- tribution of parasites on hosts? This question is anal- ogous to questions of biogeography. We know by generating a host tree from parasitic phylogenetic in- formation and by generating a parasite tree from host phylogenetic information that the two data sets are s 2 CO 3 co 3 C c to 3 C -Q •o F CO TO o TO e CO Q c o co CD Q. 3 -c O -Q o CO o c TO c E E o c o .O ■5 »_, O ^ o *-. :x «r. FIGURE 24. — Host tree based on parasitic phylogenetic information. Numbers in brackets at top of figure represent infestation by 1) Unicolax ciliatus, 2) U. collateralis, 3) U. anonymous, 4) U. mycterobius, and 5) U. reductus. Numbers crossing branches on tree represent acquisitions of parasites or parasite ancestors, except for number 2 leading to Katsuwonus which indicates a loss. CO 3 CO 3 o .CO 2 o 3 o -S5 03 0) CO o CJ 5 10 12 3 4 jg 2 — ' J' I 1 19 I A 1 FIGURE 25. — Parasite tree of Unicolax species based on host phylogenetic information. Numbers crossing branches on tree re- present historic infestations of hosts or host ancestors by parasites or parasite ancestors. 258 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES not concordant. We also know that several parts of these data sets are in agreement, that is to say, some evolutionary events in Unicolax are correlated with speciation (vicariant) events in the Scombridae. These events are easily explained by models of allo- patric speciation and hypotheses of dispersal are un- necessary. Before we can suggest a dispersal event, we must first factor out host-parasite relationships which are due to cospeciation events. They may be done by overlaying parasitic phylogenetic data in the form of a character state tree on the host phylogeny. This procedure is similar to the generation of the host by parasite tree (Fig. 24), with the exception that the parasite phylogenetic information is forced onto the host cladogram. In our example the scombrid host tree was coded as a character state tree. A character by scombrid taxon matrix was constructed so that each character was repeated a number of times. To this we added the characters from the parasite phylogeny by host data matrix used to generate the host by parasite tree. The repetition of the character by scombrid taxa matrix has the effect of forcing the tree into a particular shape, in our case, the original host cladogram. The number of replicates is large enough so the parasite phylogeny data does not alter the outcome of the tree. This combined data matrix was submitted to the WAGNER 78 program and a most parsimonious tree was generated. This tree (Fig. 26) is the same shape as the original host phylogeny, and characters relating to historical events of the parasites are over- layed or forced onto the tree in a parsimonious con- figuration. The overlay presented in Figure 26 indicates that parasite evolutionary events (— 2), (17), (— 8), and (—9) (indicated as characters circled in broken lines) were reversed or lost in several host taxa or lineages. This indicates the loss of a parasite or a hypothetical ancestral parasite. The only independent acquisition of parasites or hypothetical ancestral parasites oc- curred between Sarda, node (8), and Auxis on the cladogram. In both cases parasite 4, U. mycterobius, and its hypothetical ancestors (6) and (7) not only were independently acquired but also must have been independently evolved. In this case it is more reasonable to invoke an hypothesis of dispersal and to explain the infestation of Sarda by U. mycterobius by dispersal from another scombrid host. This hypo- thesis is more parsimonious than the coevolutionary hypothesis in that it requires one dispersal event rather than a series of independent identical evolu- tionary events (having serious taxonomic im- plications for parasitic taxa, i.e., if two taxa evolve in independent lineages they must be considered sepa- 3 E O 3 CD E -5 "C w CD o o CD CO "O cS o W a CD -Q £ O o CO E o c CD CD C E c o 3 u O CD :v s* to "=t C/3 CD 6 O CO a c c 3 •x. 3 CO 3 C C iv. ■C 3 Uj CO 3 C O 5 CD CO 3 c 3 •C I— 1 Kr M4* (&* U 11 J 7 = J J J I Figure 26. — Overlay of historical parasite information on host phylogeny. Negative numbers indicate losses and numbers circled in broken lines indicate independent acquisitions or losses of para- sites or parasite ancestors. rate, possibly sibling species). It must be noted that an hypothesis of independent evolutionary events leading to the establishment of U. mycterobius on Sarda may in itself require a dispersal event earlier in its evolutionary history. The coevolution of Unicolax and its scombrid hosts can be reconstructed as follows. The three higher tribes of the Scombrinae (Scomberomorini, Sardini, and Thunnini) share Unicolax, indicating that this genus arose from a more primitive bomolochid after the ancestors of these three tribes evolved from the Scombrini. Unicolax ciliatus, the most primitive species of Unicolax, is present only in the most primi- tive of the three tribes, the Scomberomorini. Uni- colax collateralis is found on members of the tribes Sardini and Thunnini. Infestation by U anonymous yields little information because it is restricted to Eu- thynnus alletteratus from both sides of the Atlantic. It is apparently a more recently derived species that has not spread far geographically or host-wise. Un- icolax mycterobius is restricted to the two most primi- tive genera of the Thunnini (Auxis and Euthynnus) except for its presence on two specimens of Sarda orientalis from Japan. This seems best explained as 259 K1SHEKY tSULLtlLM: VUL. 81, l^U. '1 dispersal from its usual host. It may be an example of a parasite species utilizing an alternate host in the ab- sence of its preferred host. Finally, U. reductus, the most specialized species of Unicolax, has been found only on a highly specialized host, Katsuwonus. This indicates that Katsuwonus evolved from the Euthyn- nus stock, and U. reductus evolved from the ancestor of U. mycterobius. It should be noted that, in each of the three tribes, Unicolax was not found in the most specialized scom- brid genus. In Thunnus this may be the result of com- petition resulting from heavy infestations of the monogenetic trematode, Nasicola klawei (Stunkard), in the nasal sinuses of the host fish. There is no evidence, however, that parasite competition is a fac- tor in Acanthocybium, Gymnosarda, and Allothun- nus. It may be that as each of the tribes evolve, the most specialized members lose parasites. This con- cept is consistent with other data presented else- where in this paper (see Scomberomorus infestation data in Table 6 and Sarda parasite discussions). COMPARISON OF COPEPOD PARASITES IN SCOMBRIDAE AND BELONIDAE After completing the analysis of the parasitic cope- pods of the Scombridae, it seemed instructive to make comparisons with those of the Belonidae, the only other family of fishes that has been studied in a similar manner (Cressey and Collette 1970). The Scombridae (48 species) is a larger family than the Belonidae (32 species). All scombrids are marine species, although several enter estuaries and only Scomberomorus sinensis is found far up the Mekong River. Four genera of Belonidae (Belonion, Potomor- rhaphis,Pseudotylosurus, andXenentodon) plus three species of Strongylura are restricted to freshwater, and populations of several other species of Stron- gylura invade freshwater long enough to acquire freshwater parasites. Thus, parasites of the family Ergasilidae (nine species) must be omitted in any comparisons because they are restricted to hosts in freshwater. Several other families of copepods can- not be used because their habitat does not occur in the host group. Species of Shiinoidae live inside the nasal cavities of their host, attached to the lamellae of the nasal rosettes. Belonidae have an open nasal pit with no place for a Shiinoa to attach. Scombrid species breathe largely by ram-jet ventilation of the gills and so have small oral valves in the upper and lower jaws, apparently too small to house the caligid copepod Caligodes which was found on seven species of Belonidae. Several species of the caligid genus Caligus were found on needlefishes but all in small numbers, partially because Cressey and Collette's study used mostly preserved specimens that were prone to lose parasites like Caligus, which are mostly external. Two ecological habitats, parasitized by three fami- lies of copepods in the two families of fishes, seem comparable — gills and oropharyngeal cavity. Bomolo- chid copepods are found in the oropharyngeal cavity of both host families (and also in the nasal cavities of the Scombridae). Species of the closely related families Lernanthropidae and Pseudocycnidae at- tach permanently to the gills of belonids and scom- brids, respectively (Table 10). Comparison of the parasitic copepod fauna of the most speciose genera of each family, Strongylura and Scomberomorus, reveals some interesting dis- tributional patterns. Bomolochus bellones, the com- mon bomolochid of Strongylura, extends from the Table 10. — Comparison of parasitic copepod fauna on gills (Lernanthropidae and Pseudocycnidae) and oropharyngeal cavities (Bomolochidae) in genera from the Belonidae (Strongylura) and Scombridae (Scomberomorus). Strongylura 5 species Bomolochus bellones (5/5) Bomolochus sinensis (1/5} Nothobomolochus digitatus ( 1 /5) Lernanthropus be/ones (3/5) Lernanthropus tylosun (5/5) S senegalensis Bomolochus bellones 3 species Bomolochus bellones (3/3) Lernanthropus be/ones (3/3) Lernanthropus tylosun (2/3) 2 species Bomolochus constrictus (2/2) Bomolochus ensiculus (2/2) Lernanthropus belones (2/2) Lernanthropus tylosun (1/2) Indo-West Pacific Eastern Atlantic Western Atlantic Eastern Pacific Scomberomorus 1 species Unicolax ciliatus (9/10) Pseudocycnoides armatus (8/10) Pseudocycnoides scomberomon (1/10) S. tritor Unicolax ciliatus 4 species Holobomolochus divancatus (3/4) Hofobomolochus asperatus (1/4) Pseudocycnoides buccata (4/4) 2 species Holobomolochus nudiusculus (2/2) Pseudocycnoides buccata (2/2) 260 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES Indo- West Pacific through the eastern Atlantic to the western Atlantic Ocean. It is replaced by two species of bomolochids in the eastern Pacific — B. constrictus and B. ensiculus. Unicolax ciliatus, the common bomolochid of Scomberomorus, extends from the Indo- West Pacific to the eastern Atlantic. It is re- placed in the western Atlantic by H. divaricatus and H. asperatus and in the eastern Pacific by H. nudi- usculus. The gill parasites, Lernanthropus and Pseudocyc- noides, show a similar pattern. The two species of Lernanthropus, being circumglobal, extend farther than Bomolochus does. Pseudocycnoides armatus is found on species of Scomberomorus in the Indo-West Pacific. It is replaced in the western Atlantic and eastern Pacific by P. buccata. No Lernanthropus or Pseudocycnoides were found on the single host species of Strongylura and Scomberomorus in the eastern Atlantic. Host specificity at the generic level depends on fac- tors such as the number of species in a given host genus, maximum body size of the host species, and distribution of the host species. The most speciose genera in each family (Scomberomorus with 18 of 47 species in the Scombrinae and Strongylura with 14 of 32 species in the Belonidae) have the most copepod species, 50 and 857c, respectively, of the total para- site fauna recorded for these two families (Table 1 1). However, if one calculates a mean number of cope- pod species per host species, a different picture em- erges. In both fish families, monotypic genera, including large pantropical species, contain the most copepod species per host species, Acanthocybium and Katsuwonus in the Scombridae with 6 of 46 species of copepods and Ablennes in the Belonidae with 9 of 21. The genera with the next highest number of cope- pod species per host species are moderate-sized species, Euthynnus (three species) with 3.7 copepod species per host species in the Scombridae and Platybelone (monotypic) with 7 of 21 in the Belo- nidae. The three genera with the lowest number of parasitic copepods per host species in the Belonidae (0-0.5) are a special case, without parallel in the Scombridae, small (4-28 cm body length) freshwater South American species. No copepods were found on the South African monotypic Petalichthys but only a few host specimens were examined. ACKNOWLEDGMENTS Most of the copepod collections, which provide the data on which this study is based, were reported in Cressey and Cressey (1980). We reiterate our thanks to the many scientists and staff of the institutions cited in the earlier paper for making fish collections available to us for study. Additional specimens of Scombridae were examined in South Africa by the second author through the assistance of Rudy van der Elst (Oceanographic Research Institute, Dur- ban), Philip Heemstra and Margaret M. Smith (J. L. B. Smith Institute of Ichthyology, Grahamstown), and P. A. Hulley (South African Museum, Cape- town). M. A. A. Baker of the J. L. B. Smith Institute recently loaned us a significant collection of South African scombrid copepods, for which we are grate- ful. Drafts of the manuscript were reviewed by Daniel M. Brooks, Daniel M. Cohen, Robert H. Gibbs, Jr., Ju-shey Ho, and Klaus Rohde. LITERATURE CITED Baudcm Laurencin, F. 197 1. Crustacees et helminthes parasites de l'albacore (Thun- Table 11. — Number of parasitic copepod species per genus and maximum size of Belonidae and Scombridae. (Maximum size of belonid species given in cm body length and of scombrid species in cm fork length. Beloni.dae Sc ombndae Max. x no. Max. X No. size No spp. copepod spp. No. size No. spp. copepod spp. Genus spp. (cm BL) copepods /host spp. Genus spp. (cm FL) copepods /host spp. Ablennes 1 73 9 9 Acanthocybium 1 183 6 6 Belone 2 32-53 3 15 Allothunnus 1 96 3 3 Belomon 2 4 Auxis 2 40-50 6 3 Petalichthys 1 21 Cybiosarda 1 42 3 3 Platybelone 1 30 7 7 Euthynnus 3 64-100 11 3.7 Potamorrhaphis 3 10-16 1 05 Grammatorcynus 1 100 5 5 Pseudotylosurus 2 20-28 1 0.5 Gymnosarda 1 108 3 3 Strongylura 14 23-59 18 1.3 Katsuwonus 1 100 6 6 Tylosurus 5 39-130 11 2.2 Orcynopsis 1 130 2 2 Xenentodon 1 17 2 2.0 Rastrelliger 3 20-35 5 0.7 Sarda 4 50-85 9 2.3 Scomber 3 40-50 6 2.0 Scomberomorus 18 30-220 23 1.3 Thunnus 7 90-300 10 1.4 Totals 32 21 0.68 47 46 0.98 261 FISHERY BULLETIN: VOL. 81, NO. 2 nus albacares) du Golfe de Guinee - Note preliminaire. Doc. Sci. Cent. Rech. Oceanogr. Abidjan 2:11-30. BERE, R. 1936. Parasitic copepods from Gulf of Mexico fish. Am. Midi. Nat. 17:577-625. Brooks, D. R. 1979. Testing the context and extent of host-parasite coevolu- tion. Syst. Zool. 28:299-307. 1981. Hennig's parasitological method: A proposed solution. Syst. Zool. 30:229-249. Capart, A. 1959. Copepodes parasites. In Expedition oceanographique Beige dans les eaux cotieres Africaines de l'Atlantique sud (1948-1949). Resultats scientifiques 3(Part 5):57- 126. Inst. R. Sci. Nat. Belg. Carey, F. G, J. M. Teal, J. W. Kanwisher, K. D. Lawson, and J. S. Beckett. 1971. Warm-bodied fish. Am. Zool. 11:135-143. Causey, D. 1960. Parasitic Copepoda from Mexican coastal fishes. Bull. Mar. Sci. Gulf Caribb. 10:323-337. COLLETTE, B. B. 1970. Rastrelliger kanagurta, another Red Sea immigrant into the Mediterranean Sea, with a key to the Mediterranean species of Scombridae. Bull. Sea Fish. Res. Stn. Haifa 54, p. 3-6. 1979. Adaptations and systematics of the mackerels and tunas. In G. D. Sharp and A. E. Dizon (editors), The physiological ecology of tunas, p. 7-38. Acad. Press, N.Y. COLLETTE, B. B., AND L. N. CHAO. 1975. Systematics and morphology of the bonitos (Sarda) and their relatives (Scombridae, Sardini). Fish. Bull., U.S. 73:516-625. COLLETTE, B. B., AND R. H. GlBBS, JR. 1963. A preliminary review of the fishes of the family Scom- bridae. FAO Fish. Rep. 6, p. 23-32. COLLETTE, B. B., AND J. L. RUSSO. 1979. An introduction to the Spanish mackerels, genus Scomberomorus. In E. L. Nakamura and H. R. Bullis, Jr. (editors), Proceedings of the mackerel colloquium, p. 3- 16. Gulf States Marine Fisheries Commission. 1980. Scomberomorus munroi, a new species of Spanish mack- erel from Australia and New Guinea. Aust. J. Mar. Freshw. Res. 31:241-250. Conrad, G M. 1938. The osteology and relationships of the wahoo [Acanthocybium solandri), a scombrid fish. Am. Mus. Nat. Hist. Novit. 1000, 32 p. COPI, I. M. 1972. Introduction to logic. 4th ed. Macmillan Publ. Co., N.Y., 540 p. CRESSEY, R. F., AND B. B. COLLETTE. 1970. Copepods and needlefishes: A study in host-parasite relationships. Fish. Bull, U.S. 68:347-432. Cressey, R. F., and H. B. Cressey. 1979. The parasitic copepods of Indo-West Pacific lizard- fishes (Synodotidae). Smithson. Contrib. Zool. 296:1-71. 1980. Parasitic copepods of mackerel and tuna-like fishes (Scombridae) of the world. Smithson. Contrib. Zool. 311:1-186. Dogiel, V. A., G. K. Petrushevski, and Yu. I. Poly- ANSKI (editors). 1961. Parasitology of fishes. Oliver and Boyd, Edinb., 384 p. |Transl. from Russian by Z. Kabata.] Farris, J. S. 1970. Methods for computing Wagner Trees. Syst. Zool. 19:83-92. Farris, J. S., A. G. Kluge, and M. J. Eckardt. 1970.' A numerical approach to phylogenetic systematics. Syst. Zool. 19:172-189. Fitch, J. E., and W. L. Craig. 1964. First records for the bigeye thresher (Alopias super- ciliosus) and slender tuna (Allothunus fallai) from Califor- nia, with notes on eastern Pacific scombrid otoliths. Calif. Fish Game 50:195-206. FITCH, J. E., AND P. M. ROEDEL. 1963. A review of the frigate mackerels (genus Auxis) of the world. FAO Fish. Rep. 6, p. 1329-1342. Fraser-Brunner, A. 1950. The fishes of the family Scombridae. Ann. Mag. Nat. Hist., Ser. 12, 3:131-163. Gibbs, R. H., Jr., and B. B. Collette. 1967. Comparative anatomy and systematics of the tunas, genus Thunnus. U.S. Fish Wildl. Serv., Fish. Bull. 66:65- 130. GODSIL, H. C. 1954. A descriptive study of certain tuna-like fishes. Calif. Dep. Fish Game, Fish Bull. 97, 185 p. Graham, J. B. 1973. Heat exchange in the black skipjack, and the blood-gas relationship of warm-bodied fishes. Proc. Natl. Acad. Sci. U.S.A. 70:1964-1967. 1975. Heat exchange in the yellowfin tuna, Thunnus albacares, and skipjack tuna, Katsuwonus pelamis, and the adaptive significance of elevated body temperatures in scombrid fishes. Fish. Bull., U.S. 73:219-229. GUSSEV, A. B. 1951. Parasitic Copepoda of some marine fishes. Collect. Pap. Parasit. Zool. Inst., Acad. Sci. SSSR 13:394-463. HELLER, C. 1865. Crustaceen. Reise der Osterreichischen Fregatte Novara um die Erde in den Jahren 1857, 1858, 1859. Zool. Theil. 2,280 p. Hennig, W. 1966. Phylogenetic systematics. Univ. 111. Press, Urbana, 263 p. Holmes, J. C. 1979. Parasite populations and host community structure. In B. B. Nichol (editor), Host-parasite interfaces, p. 27-46. Acad. Press, N.Y. Holmes, J. C, and P. W. Price. 1980. Parasite communities: The roles of phylogeny and ecology. Syst. Zool. 29:203-213. Hotta, H. 1962. The parasitism of saury (Cololabis salra) infected with parasitic copepods, Caligus macarovi Gussev, during fish- ing season in 1961. (In Jpn., Engl, summ.] Bull. Tohoku Reg. Fish. Res. Lab. 21:50-56. Hotta, H., and K. Odate. 1956. The food and feeding habits of the saury, Cololabis saira. |In Jpn., Engl. abstr.J Bull. Tohoku Reg. Fish. Res. Lab. 7:60-69. Iversen, E. S., and H. O. Yoshida. 1957. Notes on the biology of the wahoo in the Line Islands. Pac. Sci. 11:370-379. Kabata, Z. 1965. Copepoda parasitic on Australian fishes. IV. Genus Caligus (Caligidae). Ann. Mag. Nat. Hist., Ser. 13, 8:109- 126. 262 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES KlRTISINGHE, P. 1937. Parasitic copepods of fish from Ceylon. II. Parasitolo- gy 29:435-452. KlSHINOUYE, K. 1923. Contributions to the comparative study of the so-called scombroid fishes. J. Coll. Agric, Imp. Univ. Tokyo 8:293- 475. L0TKEN, C. 1880. Spolia Atlantica. Bidrag til kunkskab om form- forandringer hos fiske under deres vaext og unvikling, saerligt hos nogle af atlanterhavets h«(jsfiske. Dan, Vidensk. Selks. Skr. Kjtfbenhaven, Ser. 5, 12:409-613. Mago Leccia, F. 1958. The comparative osteology of the scombroid fishes of the genus Scomberomorus from Florida. Bull. Mar. Sci. Gulf Caribb. 8:299-341. Margolis, L., and J. R. Arthur. 1979. Synopsis of the parasites of fishes of Canada. Fish. Res. Board Can., BulL 199, 269 p. Margolis, L., Z. Kabata, and R. R. Parker. 1975. Catalogue and synopsis of Caligus, a genus of Cope- poda (Crustacea) parasitic on fishes. Fish. Res. Board Can., Bull. 192, 117 p. Matsui, T. 1967. Review of the mackerel genera Scomber and Rastrel- liger with description of a new species of Rastrelliger. Copeia 1967:71-83. Matsumoto, W. M. 1960. Notes on the Hawaiian frigate mackerel of the genus Auxis. Pac. Sci. 14:173-177. 1976. Second record of black skipjack, Euthynnus lineatus, from the Hawaiian Islands. Fish. Bull., U.S. 74:207. Matsumoto, W. M., and T. Kang. 1967. The first record of black skipjack, Euthynnus lineatus, from the Hawaiian Islands. Copeia 1967:837-838. Mayr, E. 1957. Die denkmoglichen Formen der Artentstehung. Rev. Suisse Zool. 64:219-235. Noble, E. R., and G. A. Noble. 1973. Parasitology; the biology of animal parasites. 3d ed. Lea and Febiger, Phila., 617 p. PlLLAI, N. K. 1973. Three new bomolochids parasitic on fishes of the Kerela coast. Indian J. Fish. 20:487-496. Price, P. W. 1980. Evolutionary biology of parasites. Princeton Univ. Press, Princeton, 237 p. Regan, C. T. 1909. On the anatomy and classification of the scombroid fishes. Ann. Mag. Nat. Hist, Ser. 8, 3:66-75. Richards, W. J., and J. E. Randall. 1967. First Atlantic records of the narrow-corseleted frigate mackerel, Auxis thazard. Copeia 1967:245-247. ROHDE, K. 1980. Host specificity indices of parasites and their applica- tion. Experientia 36:1369-1371. Silas, E. G., and A. N. P. Ummerkutty. 1967. Parasites of scombroid fishes. Part 2. Parasitic Cope- poda. Proc. Symp. Scombroid Fishes I (Part 3):876-993. Mar. Biol. Assoc. India, Symp. Ser. STARKS, E. C. 1910. The osteology and mutual relationships of the fishes belonging to the family Scombridae. J. Morphol. 21:77- 99. TAKA, S., M. KlTAKATA, and T. Wada. 1980. Food organisms of saury, Cololabis saira (Brevoort) and vertical distribution of zooplankton in the southeast waters off Kuril Islands in July, 1976-1978. Bull. Hok- kaido Reg. Fish. Res. Lab. 45:15-41. UCHIDA, R. N. 1981. Synopsis of biological data on frigate tune, Auxis tha- zard, and bullet tuna, A. rochei. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 436, 61p. Wade, C. B. 1949. Notes on the Philippines frigate mackerels, family Thunnidae, genus Auxis. U.S.FishWildl. Serv.,Fish. Bull. 51:229-240. Walters, V., and H. L. Fierstine. 1964. Measurements of swimming speeds of yellowfin tuna and wahoo. Nature (Lond.) 202:208-209. Webb, B. F. 1976. A record of the copepod parasite Elytrophora brachyp- tera from slender tuna (AUothunnus fallai) taken in Tasmanian waters. Tasmanian Fish. Res. 10:28. Wilson, C. B. 1908. North American parasitic copepods: New genera and species of Caliginae. Proc. U.S. Natl. Mus. 33:593-627. 1937. Parasitic copepods taken during the third Hancock Expedition to the Galapagos Islands. Allan Hancock Found. Pac. Exped. 2:23-30. 263 FISHERY BUEEETTOI VOL. «1, NU. '2 APPPENDIX Below is a list of scombrid hosts and their parasitic copepods. Numbers after fish names indicate the number offish examined. Numbers after copepod names indicate number of fish infested. Asterisks indicate new record since Cressey and Cressey (1980). SCOMBRINI Rastrelliger brachysoma (33) Lernanthropus kanagurta (1) Rastrelliger faughni ( 1 4) Pumiliopes jonesi (2) Nothobomolochus kanagurta (2) Orbitacolax aculeatus (2) Rastrelliger kanagurta (124) Pumiliopes jonesi (20) Nothobomolochus kanagurta (7) Caligus kanagurta (2)* Orbitacolax aculeatus (2)* Scomber australasicus (55) Pumiliopes jonesi (5)* Clavellisa scombri (4) Scomber japonicus (500) Clavellisa scombri (9) Pumiliopes jonesi (8) Caligus pelamydis (1) Caligus mutabilis (1) Clavelopsis saba (1) Scomber scombrus (97) Caligus pelamydis (1) SCOMBEROMORINI Acanthocybium solandri (64) Brachiella thynni (39) Gloiopotes hygomianus (27) Caligus productus (11) Shiinoa occlusa (2) Caligus coryphaenae (1) Pennella species (1) Tuxophorus cybii (1) Grammatorcynus bicarinatus (47) Shiinoa occlusa (9) Caligus asymmetricus (7) Caligus regalis (4)* Caligus bonito (1)* Caligus pelamydis (1)* Caligus productus (1)* Scomberomorus brasiliensis (62) Pseudocycnoides buccata (39) Holobomolochus divaricatus (14) Caligus mutabilis (4) Shiinoa inauris (3) Scomberomorus cavalla (36) Pseudocycnoides buccata (18) Holobomolochus asperatus (10) Caligus mutabilis (2) Caligus productus (1) Scomberomorus ccmmerson (130) Pseudocycnoides armatus (25) Unicolax ciliatus (23) Caligus cybii (16) Shiinoa occlusa (15) Caligus biseriodentatus (12) Caligus infestans (7) Tuxophorus ceruicornis (3) Brachiella magna (2) Caligus asymmetricus (2) Tuxophorus cybii (1) Scomberomorus concolor (47) Pseudocycnoides buccata (14) Holobomolochus nudiusculus (13) Caligus omissus (7) Scomberomorus guttatus (58) Caligus biseriodentatus (17) Unicolax ciliatus (14) Pseudocycnoides armatus (3) Shiinoa occlusa (1) Scomberomorus koreanus (19) Caligus cybii (11) Pseudocycnoides armatus (4) Unicolax ciliatus (1)* Scomberomorus lineolatus (14) Unicolax ciliatus (3) Caligus biseriodentatus (1) Pseudocycnoides armatus (1) Scomberomorus maculatus (77) Pseudocycnoides buccata (27) Holobomolochus divaricatus (25) Shiinoa inauris (7) Caligus mutabilis (2) Scomberomorus munroi (6) Caligus cybii (3) Unicolax ciliatus (2) Caligus biseriodentatus (1)* Caligus productus (1) Scomberomorus multiradiatus (29) Pseudocycnoides armatus (8) Caligus biseriodentatus (7)* Scomberomorus niphonius (19) Pseudocycnoides scomberomori (6) Unicolax ciliatus (3) Caligus pelamydis (2) Shiinoa occlusa (1) Scomberomorus plurilineatus (14) Pseudocycnoides armatus (12) Unicolax ciliatus (4)* Brachiella thynni (1)* Caligus asymmetricus (1)* Shiinoa occlusa (1)* Scomberomorus queenslandicus (39) Caligus biseriodentatus (12) Unicolax ciliatus (3) Pseudocycnoides armatus (2) Caligus cybii (1)* Shiinoa occlusa (1) Scomberomorus regalis (38) Pseudocycnoides buccata (12) Holobomolochus divaricatus (11) Shiinoa inauris (5) Caligus productus (3) Caligus bonito (1) Brachiella thynni (1) Tuxophorus collettei (1) Scomberomorus semifasciatus (26) Pseudocycnoides armatus (5) Unicolax ciliatus (4) Caligus cybii (3) Shiinoa occlusa (2)* Caligus biseriodentatus (1)* Scomberomorus sierra (116) Pseudocycnoides buccata (48) Caligus omissus (39) Holobomolochus nudiusculus (28) Scomberomorus sinensis (10) Caligus cybii (2) Brachiella magna (1) Caligus pelamydis (1)* Scomberomorus tritor (21) Unicolax ciliatus (4) Shiinoa occlusa (1) Caligus productus (1) Caligus diaphanus (1) 264 CRESSEY ET AL.: COPEPODS AND SCOMBRID FISHES SARDINI Sarda sarda (106) Caligus bonito (33) Ceratocolax euthynni (21) Caligus pelamydis (8) Caligus productus (1) Sarda australis (22) Caligus bonito (13) Caligus pelamydis (11) Caligus asymmetricus (2) Unicolax collateralis (2) Sarda chiliensis (45) Caligus bonito (25) Caligus pelamydis (4) Caligus productus (1) Sarda orientalis (33) Unicolax collateralis (12) Caligus bonito (12) Caligus asymmetricus (4) Caligus kanagurta (2)* Caligus productus (1) Caligus coryphaenae (1)* Unicolax mycterobius (1)* Gymnosarda unicolor (7) Caligus bonito (1) Caligus productus (1) Shiinoa occlusa (1) Cybiosarda elegans (38) Unicolax collateralis (16) Caligus asymmetricus (3) Caligus bonito (1)* Orcynopsis unicolor (7) Unicolax collateralis (1) Caligus bonito (1)* Allothunnus fallai (7) Elytrophora brachyptera (5) Caligus bonito (2)* Caligus productus (1)* THUNNINI Auxis species (68) Unicolax collateralis (19) Unicolax mycterobius (9) Caligus productus (2) Caligus asymmetricus (1) Caligus coryphaenae (1) Caligus pelamydis (1) Euthynnus affinis (74) Unicolax collateralis (32) Caligus asymmetricus (5) Caligus regalis (5) Pseudocycnus appendiculatus (4) Unicolax mycterobius (3) Caligus pelamydis (2) Caligus productus (1) Caligus bonito (1) Euthynnus alletteratus (64) Caligus coryphaenae (9) Unicolax collateralis (8) Ceratocolax euthynni (7) Caligus productus (5) Caligus bonito (4) Pseudocycnus appendiculatus (3) Unicolax mycterobius (3) Unicolax anonymous (2) Caligus pelamydis (1) Euthynnus lineatus (15) Unicolax collateralis (4) Caligus bonito (3) Katsuwonus pelamis (135) Caligus productus (54) Caligus coryphaenae (51) Pseudocycnus appendiculatus (8) Unicolax reductus (3) Caligus bonito (2)* Caligus asymmetricus (1) Thunnus alalunga (13) Elytrophora brachyptera (8) Caligus coryphaenae (1) Caligus productus (1) Pseudocycnus appendiculatus (1) Thunnus albacares (112) Caligus productus (51) Elytrophora brachyptera (39) Caligus coryphaenae (32) Pseudocycnus appendiculatus (21) Brachiella thynni (8) Caligus asymmetricus (1) Thunnus atlanticus (76) Caligus productus (70) Caligus coryphaenae (9) Elytrophora brachyptera (1) Thunnus maccoyii (7) Elytrophora brachyptera (5) Caligus productus (1)* Pseudocycnus appendiculatus (1)* Thunnus obesus (42) Elytrophora brachyptera (20) Caligus coryphaenae (18) Brachiella thynni (10) Elytrophora indica (11) Pseudocycnus appendiculatus (3) Thunnus thynnus (57) Caligus coryphaenae (16) Caligus productus (16) Elytrophora brachyptera (11) Pennella species (3) Brachiella thynni (2) Caligus bonito (1) Pseudocycnus appendiculatus (1) Thunnus tonggol (29) Pseudocycnus appendiculatus (7) Caligus kanagurta (1)* 265 POPULATION ASSESSMENT OF THE GRAY WHALE, ESCHRICHTIUS ROBUSTUS, FROM CALIFORNIA SHORE CENSUSES, 1967-80 Stephen B. Reilly, 1 ' 2 Dale W. Rice,' and Allen A. Wolman' ABSTRACT Estimates of abundance by year were developed for the California-Chukotski stock of gray whales, from a 13- year consecutive series of shore censuses, conducted near Monterey, Calif. Annual estimates of population size range from a low of 10,414 for 1971-72 to a high of 17,577 for 1979-80. Standard errors are about 10% of pop- ulation estimates. During the 13 years censused, the population increased annually by 2.57c, concurrent with a 1.2% harvest in the Soviet subsistence fishery, indicating a 3.7% net annual productivity. Seasonal migratory timing was relatively constant during the study period. Gamma probability density func- tion models of the annual migrations past Monterey had an overall mean day of 9 January, with a range from 8 to 19 January. A slight depression in mean hourly count for0070-0800 h, during 1978-79 and 1979-80, contrast- ed with a constant mean hourly count through 10 daylight hours during the previous 1 1 years. Aerial surveys of the offshore distribution of southward migrating whales during 1979-80 agreed closely with those reported for 1978-79, indicating that 40% pass within 1 mile (1.6 km) of shore and 907c within 2 miles (3.2 km). In the shore censuses, about 20% of the passing whales were missed due to their distance offshore. The estimation of population size for large whales has traditionally been based upon information de- rived from exploitation, e.g., catch per unit effort, mark-recapture, or related data (Allen 1980). Be- cause of the recent decline in exploitation of marine mammals, assessment techniques based upon sight- ing surveys are increasing in importance (Eberhardt et al. 1979). The annual migration of the California stock of gray whales, Eschrichtius robustus (Lill- jeborg 1861), makes it especially well suited to assessment by means of sighting surveys. Assess- ment studies on this stock can potentially aid in the development of sighting survey field and analysis techniques, especially those in which the observer is stationary and the population mobile. This paper presents some recent developments in the use of shore-based census data for whale population as- sessment, and the results of the 1979-80 gray whale census. Revised population estimates for the pre- vious 12 annual censuses are also reported, along with a consideration of change in population siz e dur- ing the period 1967-80. Each year during the northern winter the California stock of gray whales migrates from feeding waters in 'National Marine Mammal Laboratory, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, WA 981 15. 2 National Marine Mammal Laboratory, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE., Seattle, Wash.; present address: Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, 8604 La Jolla Shores Drive, La Jolla, CA 92038. Manuscript accepted October 1982. FISHERY BULLETIN: VOL. 81, NO. 2, 1983. the Bering and Chukchi Seas, south along the west coast of North America, to calving areas in Mexical waters (Fig. 1); the stock returns to the Arctic in the spring (Rice and Wolman 1971). In many places along the route, the whales pass very close to land (Gilmore 1960; Pike 1962; Rice and Wolman 1971; Rugh and Braham 1979). Consequently, it is feasible to census the migrating whales visually from strategic points along the shore. Early shore-based censuses were summarized by Reilly et al. (1980). Systematic censuses of south- ward migrating gray whales were initiated during the winter of 1967-68 at both Point Loma (lat. 32°40'N; 130 m above sea level) in San Diego, Calif., and at Yankee Point (lat. 36°29'N; 23 m above sea level) near Monterey, Calif. The San Diego count was con- ducted intermittently until 1977-78, for a total of 5 yr. The San Diego data were not analyzed in this study because an unverified proportion of the pop- ulation passes far offshore south of Point Conception (Rice 1965) and because the migration route may have been influenced by increased boat traffic (Rice 1965; Reeves 1977). The Monterey census was con- ducted each year for 13 yr up to and including 1979- 80. Beginning in 1975-76 the counting station was moved 3.7 km south to Granite Canyon (21m above sea level) due to real estate development of the Yan- kee Point site. The Monterey data were used as the basis for this study, because they form a continuous time series and are less complicated by coastal geography and boat traffic than the San Diego data. 267- FISHERY BULLETIN: VOL. 81, NO. 2 150° 160° 170° 180° 170° 160° 150° 140° 130° 120° 110° 50° - FIGURE 1.— The approximate migration route of the California stock of gray 20 CANADA GRAY WHALE MIGRATION ROUTES - 20° whales. 150° 160° 170° 180° 170° 160° 150° 140° 130° 120° 110° To estimate total abundance by extrapolating from recorded counts of passing whales one must deter- mine the following: 1) What proportion of the population, if any, passes beyond sight of the observers? Does this change with time or experience? How does the observer's accuracy in estimating the distance to passing whales vary with distance? 2) Are there diel variations in migration rate? How can daylight counts be used to estimate the num- ber of whales passing at night? 3) How do weather (visibility) conditions affect census results? 4) Does the observer's ability to count the number of individuals within a passing group vary with group size? 5) Are the initiation and termination of the migra- tion fully represented in the data? During the 1978-79 southward migration we con- ducted two types of verification experiments aimed at addressing the questions of points 1 and 4 above. These were reported in detail in Reilly et al. (1980). In one experiment we tested 12 observers simul- taneously for accuracy in estimating distances to and numbers within 50 events in which whales passed the Granite Canyon station. The observers estimated the distance offshore to within one of seven predefined distance intervals, as during the actual annual cen- suses (see Methods). We found significant hetero- geneity between observers for both distance and count estimates. Given this heterogeneity, there were also consistent biases recorded: In placing whales to within correct intervals out to 1 mi (1.6 km) and beyond 1.5 mi (2.4 km), and in estimating the true number of individuals present in groups of one whale, and four or more. Further analysis of this data (Reilly 1981) indicated that "experienced" observers were on average no more accurate than inexperienced ob- servers, but somewhat more precise. A second experiment was conducted during 1978- 79 to characterize the width of the migration corridor offshore from the Monterey counting stations (Reilly et al. 1980). A small aircraft flew a series of transects perpendicular to the coast in the vicinity of the stations, recording locations of sighted whales (Fig. 2). The results indicated that, contrary to previous assumptions and characterizations of 95% of the population passing within 1.6 km (Rice and Wolman 1971; Sund and O'Connor 1974), we found only about 40% within 1.6 km, with significant numbers passing offshore between 1.6 and 4.5 km. This ex- periment was repeated during 1979-80, with results reported here. Regarding night migration rate (point 2 above), af- ter a review of all available information, we accepted an assumption of a constant 24-h rate. Contrary to the earlier report of Ramsey (1968), we found no evidence of a diurnal fluctuation from the shore cen- sus data. During the 1979-80 migration a new (pro- totype) infrared image sensor, supplied by the U.S. Department of Defense, was tested at Granite Can- yon. As with previously tested night-vision devices (Reilly et al. 1980), it proved unsatisfactory. The possible effect of visibility conditions on cen- 268 REILLY ET AL.: POPULATION ASSESSMENT OF GRAY WHALE 122° 00' 122° 55' 36° 30' 36° 25' 36° 20' Carmel Lobos Yankee Point Station Granite Canyon Station CALIFORNIA 122° 50' 1 - FIGURE 2.— The California coast south of Monterey, showing census stations and aerial transect lines for gray whale study. 269 FISHERY BULLETIN: VOL. 81, NO. 2 sus results was not addressed in Reilly et al. (1980). We report here a quantitative appraisal of this effect, and account for it in our abundance estimation. METHODS Field Methods: Shore Census The exact seasonal duration of the annual census changed only slightly from year to year, but it usually began on or before 10 December and ended on or af- ter 6 February (59 d). The watch was conducted be- tween 0700 and 1 700 h, 7 d a week, by two observers who alternated 5-h shifts. The observers watched to the north for southward swimming whales to come into view. At first sighting of a whale or group of whales the time was recorded and an initial estimate was made of the number of whales in the group. The whales were kept under ob- servation until they were directly offshore from the station, usually about 0.5 h later. At that time a final estimate of the number present was recorded, along with the time and an estimate of the distance of the animals offshore. Distance estimates were classified in seven intervals: 0-0.25; 0.25-0.50; 0.50-0.75; 0.75- 1.0;1.0-1.5; 1.5-2.0; 2.0+ mi. Beaufort Sea state, wind direction, and notes on visibility conditions were recorded continuously throughout the day. Binoculars (7 X 50) were used regularly. Beginning in 1978-79, visibility conditions were assigned one of six ordinal categories (Table 1) for each pod ob- served. For data prior to 1978-79, visibility conditions were classified to within these categories during the analysis, based upon information recorded sys- tematically during the censuses. TABLE 1. — Gray whale census — Granite Canyon visibility codes. Code Condition Description 01 Excellent 02 Very good 03 Good 04 05 06 Fair Poor None Clear day. or high clouds. No glare. Horizon visi- ble. Effective sighting distance = 3+ mi Clear or some cloud cover. Some glare, surface ripple. Effective sighting distance = 2-3 mi. Some fog, haze, low clouds. Some interference from chop, surf, or glare. Effective sighting dis- tance = 1-2 mi. Fog, full overcast, light rain, haze with glare. Fre- quent whitecaps. Effective sighting distance = 0.5-1 mi. Moderate rain or fog, large surf, bad glare, etc. Effective sighting distance = 0.25-0.5 mi. Combination of conditions make it very difficult or impossible to see even the closest (within 0.5 mi.) whales. Heavy rain, dense fog, near dark- ness, etc. Analysis Methods: Shore Census Occasionally during the censusing, only one of the standard two sighting times per group (when first seen and when directly offshore) was recorded. Fre- quently when an observer came on duty at 0700 h there were whales directly offshore and no "north time" was recorded. In addition, at the end of the day at 1 700 h, whales which had not yet passed directly in front of the station were often sighted to the north, and no "south time" was recorded. To correct for missing time records, a mean difference between the two times was calculated for each observer in- dividually. Missing time records were then generated from this average, and the single time record avail- able. The time when the animals were directly offshore was then used to categorize data for time of day analyses. Only sightings with this time falling between 0700 and 1700 h were used for abun- dance estimation. The results of the 1978-79 and 1979-80 half-day observation periods were investigated by analysis of variance (ANOVA) for differences between observ- ers and between morning vs. afternoon periods on rate of recording animals, as was previously done (Reilly et al. 1980) for the 1967-68 through 1977-78 data. We also examined the two most recent censuses for possible changes in hourly rates of recorded counts, as done previously for the 1967-68 through 1977-78 data. Again, we looked for significant de- pressions in the counts both at the ends of the 5-h ob- server periods (as an indication of observer fatigue) and at the beginning and end of the day (as an indica- tion of daylight-mediated change in migration rate) . For any migratory species which can be censused feasibly from a fixed point, the distribution of daily counts, transformed to proportions for each migra- tion, can be viewed profitably as a time-density dis- tribution and modeled by various probability density functions (Mundy 1979). We previously assumed a normal distribution (Reilly et al. 1980) for all years pooled. Problems with this approach were that mean days between years were not equal and that a slight but consistent skewness occurred causing lack of fit. Consequently we have replaced the normal distribu- tion with the more flexible gamma distribution (Pear- son's Type III; Bury 1975) and modeled each year separately. The time-density model for each migra- tion was then employed in three ways: 1) To estimate the number of whales having passed the station before the first and after the last day of the census (the "tails"). 2) As a standard for comparison with observed dai- ly results, in a determination of if, and to what degree, conditions associated with the six visi- bility categories affect census results. 270 REILLY ET AL.: POPULATION ASSESSMENT OF GRAY WHALE 3) To estimate the proportion of the population passing the census station on days for which the visibility conditions were worse than a critical value, as determined by the results of the visibility analysis (2). The data on pod-size estimation from all years were examined both for differences between years and for a pattern in distance from shore. The offshore distance frequency distribution of ob- servations was investigated for significant differ- ences between the two locations, as a preliminary to post facto application of correction factors for whales missed offshore. Field Methods: Verification Experiments The aerial transects to determine the offshore dis- tribution of the migratory corridor were repeated in 1979-80 following our previous methods (Reilly et al. 1980). We flew a Cessna 172 3 aircraft at 305 m (1,000 ft) altitude, at a speed of 145 km/h (90 mi/h), along a series of predefined tracklines (Fig. 2). These lines were situated along a 25 km stretch of the coast which included both the Yankee Point and Granite Canyon census stations. Distances of whales from shore were calculated from the timed difference between their position and the shore edge, and the plane's speed. During 1979-80 we flew a total of 13 flights for 34 h, in periods of good to excellent visibility. Flights were continued until a number greater than the minimum sample size of whales was obtained (330) for 90% precision in correctly classifying the population into the seven distance intervals used in the shore census (Reilly et al. 1980). Sample-size determination was based upon Cochran's (1977:74-76) formulae for sampling for proportions. Data from the 1974-75 shore counts were used as a presample of the propor- tions expected within the distance intervals from shore. The seven-interval experimental design also presented the opportunity to analyze the data in a pooled, less demanding interval scheme, with result- ing higher precision in estimating the within-inter- val proportions. Additive bias corrections were previously deter- mined from the results of the observer bias ex- periments regarding estimation of the number of whales present in passing groups. Specifically for es- timates of group size n (see Appendix 1 for explana- tion of notation) E[n] = n + b n { n + 0.350 n = 1 n + 0.00 n = 2,3 (1) n + 0.333 n > 4 with variances as in Appendix 2. Analysis Methods: Verification Experiments Aerial sightings were analyzed for effects on off- shore distance estimates from: differences between the two individual observers; the side of the plane from which the whales were seen; and the period of day (morning or afternoon flight) by ANOVA. The distance distributions from the 2-yr surveys were test- ed by x 2 (chi-square) for the possibility of pool- ing. To address the misclassification bias suggested by the results of the 1978-79 experiments, the data from those experiments were reanalyzed by using a less demanding classification scheme of three broad in- tervals: 0-0.75 mi (1.2 km); 0.76-1.5 mi (2.4 km); 1.6 mi + (2.6 km). From this characterization, a series of reclassification parameters (probabilities) were cal- culated, p ab , being the proportion of whales estimated to be within interval a, that were determined to be ac- tually passing within interval 6. The actual census data, structured in the same three intervals, were re- structured by application of these parameters as m b = Z,(m a p ab ), (2) where m a includes the whales originally classified in- to interval a, and m b comprises the whales redis- tributed into interval b, which were originally (erroneously) estimated to be in a. For example, for a = 1 and 6=1, sightings correctly classified into inter- val 1 are summed into the newm b=l . For a =2,6= 1, sightings incorrectly classified during the censuses into interval 2 are reclassified, or summed, into rh 6=1 . Inthecaseofa= 2,6= l,p 2 i = 0.2367 of the whales originally put in interval 2 would be placed into inter- val 1. The redistributed census data were then com- pared with the "true" distribution from the aerial surveys. As a simple correction factor, the ratio of the cumulative proportions seen within 2.4 km (1.5 mi) was calculated for each year {k): h{k) = CJC p- (3) 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. A necessary assumption of this method is that at least during periods of good or better visibility, all groups of whales passing within 2.4 km (1.5 mi) were recorded. 271 FISHERY BULLETIN: VOL. 81, NO. 2 Analysis Methods: Estimation of Abundance In fitting the probability density functions to the census data, the unit used was the estimate of the proportion of the population passing during a 24-h day. The number passing on day ;' was estimated as ftj= (SENAy) 24, (4) where E[n\ is the expected value of n, i.e., the es- timate of the number per group, corrected for bias as in Equation (1). The relative proportion passing on day; was estimated as Pj = nj/JMj. (5) Model parameters were first estimated for each year using all data points regardless of recorded visibility conditions. Data were fit by the two- parameter gamma model fU\**P) = a-r(/3) Ufa)?-* exp{-;/a} (6) for each migration separately. The parameters of the gamma distribution, their variances and covariance, were estimated by the method of maximum like- lihood (Chapman 1956; Greenwood and Durand 1960). Equality of parameters between years was tested by the.F statistic (Chapman 4 ), F = £ (jc - x) 2 /n - 1 £ var (x)/n (7) for x = a, [i. The distribution ofp ; for each year was then used to determine the effect of visibility conditions on census results. An average visibility condition was calcu- lated for each day from all of the recorded codes (Ta- ble 1). The difference (residual) between the observed and predicted relative proportions for each day was also calculated. An ANOVA was performed on the residuals with visibility categories as groups, along with multiple range tests (Duncan's, Student- Newman-Kuels, Scheffe's). These results were used, along with an examination of the mean squared errors for each category, to set a critical level of visibility conditions beyond which there was significant inter- ference with accurate censusing. The data were then 4 D. G. Chapman, Director, Center for Quantitative Science, College of Ocean Fishery Sciences, University of Washington, Seattle, WA 98195, pers. commun. March 1980. refit by the gamma distribution using only days with visibility codes less than the critical value as points. The new set of daily predictors (p' ; ) from the fitted gamma model were used in the further estimation procedures. Then, as an alternate to Equation (2), the abun- dance for day; was n, = f [(LE[n])A,] 24 WW) vis < critical value (8a) vis > critical value. (8b) That is, for days with visibility conditions less than or equal to some critical level (with levels defined as in Table 1) the average hourly sighting rate, correct- ed for counting bias, multiplied by 24 h, was used as the estimate of the total number of whales passing. For days with visibility conditions worse than some critical value, the estimate of the number passing came from the expected proportion for the day (from the gamma distribution model of migratory timing for that year, p'j) multiplied by the sum of the daily es- timates from the first fitting of the gamma model. For estimating the "tails" of the migration, a slight modification of the method of Mundy (1979) was used. This method was developed to predict total run size for salmon from intermediate results of counts, given that migratory timing can be modeled. The to- tal "run" Nj was predicted by minimizing the least squares error function / En, . , (9) which was solved for Nj {N estimated by data cumu- lative to day;') by N,=?(2>i ; ) 2 /2^, (10) Here Mundy uses 6 ] as the cumulative proportion ex- pected to have passed by day;, and we define 0, as that quantity less the predicted proportion missed before the first day of each census. The final form of the abundance estimate for each year k was then, A^ = {Z(Zn//(En,)-0,)/7(*). (ID The variance for Equation (11) was estimated in two ways. The first, S5, outlined in Appendix 2, was derived from the component variances of the pa- rameters used in the model, employing the Delta Method (Seber 1973). In the second method the data were subsampled in five 2-h samples/d. The five 272 REILLY ET AL.: POPULATION ASSESSMENT OF GRAY WHALE estimates for the year were then calculated using Equation (11). A simple variance of these estimates about the mean estimate (S 2 R ) was then calculated. Variances were compared for equality (H :S| = SI) by the test statistic (" ~ PS si (12) where \ 2 i s distributed approximately as chi- squared (Freund 1962:371) with rejection regions X 2 > XW2,n-l or X 2 < Xl - a /2fl-V Analysis Methods: Trends in Population Size In order to test for a trend in population size during the 13-yr study period, two models were chosen for regression analysis. This first model was simple linear regression, the second was a weighted log,, model: N t = N e rt , (13) where N t is population size in year t,N is year zero, or 1967 for the shore census time series. Equation (13) was fit linearly as InN, = lnN + rt, (14) CO c CO CO D O LU Z> o LU DC 10 8 CS 6 > o 4- i*i i ill i iiiii ^! nin^J^^^^^ 1 2 3 4 5 6 7+ GROUP SIZE ESTIMATE FIGURE 3. — Frequencies of group size estimates from Monterey gray whales census, 1967-68 through 1979-80, n = 23,678 obser- vations. with weights calculated as an inverse function of the estimated variance of iV, in the log model: Var(lnfy) \f'(N t )\ 2 Var(N t ) = Var(N t )/Nf (15) = ,M-1 = W, RESULTS Shore Census Data Base A histogram of group sizes as recorded from the 13 annual censuses is presented in Figure 3. The overall mean was 2.086 (S 2 = 1.974, n = 23,749). The mean group sizes by year are listed in Table 2. An ANOVA indicates that there are significant differences be- tween the mean pod sizes recorded by year (F = 8.282 > F\2.~,Q.ob)- Multiple range tests (Duncan's, Student- Newman-Kuel's, Scheffe's) show that 1967-68 and 1977-78 are different from each other and the rest, while all the others are homogeneous. In the 1967-68 census the unusually high mean is attributable to one of the two observers that year. His individual mean pod was 3.123 (S 2 = 2.651), and was significantly dif- TABLE 2.— Mean pod size estimates by year for the Monterey gray whale censuses, 1967-68 through 1979-80. Group mem- bership identifies placement by multiple range tests into one of three nonsignificantly different subgroups. Mean pod Group Year estimate SD n membership 1967-68 2.4970 3.5520 1239 3 1968-69 2.1471 22550 1509 2 1969-70 2.0900 2.2790 1643 2 1970-71 20330 1 6110 1652 2 1971-72 2 1630 1 8700 1272 2 1972-73 2.1400 1.6780 2041 2 1973-74 20990 1 7980 1859 2 1974-75 2.0710 1.9170 1855 2 1975-76 2.0620 1.7210 2086 2 1976-77 20660 1.5930 2296 2 1977-78 1.8250 1.2470 1996 1 1978-79 2.0040 2.3750 1960 2 1979-80 2.1030 1 7120 2341 2 Overall 2.0855 1.9736 23.799 Analysis of variance Source df ss ms F Between groups 12 385.7561 32.1463 8. 283' Within groups 23.736 92.119.1877 8.8810 Total 23.748 92,504 9438 'Significant at a = 0.05. ferent from the other observer that year, whose mean was 1.886 (S 2 - 1.959:^= 24.528>f . . 05 ). In 1977-78 however, the two observers did not differ significantly 273 FISHERY BULLETIN: VOL. 81, NO. 2 from each other in mean pod size estimated (1.842, 1.829, t= 1.1442 < t.0.05) and, consequently, the dif- ference of this year's data from others cannot be credited to one aberrant observer. There was a significant increase in mean group size as a function of distance from shore (Fig. 4) (F = 97.28 > F 5f 23i). A significant linear increase in the pooled data (Fig. 4) was also noted in 10 of the 13 individual years. In the remaining 3-yr data ( 1968-69, 1972-73, 1978-79), the average pod size peaked at about 0.6-0.9 km (1-1.5 mi) from shore, and de- creased thereafter. This may be a real between-year difference in whale behavior, but is more likely a function of the varying abilities of the observers themselves. There are highly significant differences between years in the frequency of observations recorded within offshore distance intervals (x 2 = 2,340, df = 24). For this analysis, a pooled three-interval dis- tribution was used in light of the observer bias tests discussed above. Within both the Yankee Point loca- tion subset of years and the Granite Canyon subset there also exists significant heterogeneity in the offshore distribution (x 2 = 1,077, df= 14;x 2 = 1,025, df= 8, respectively). Given this, a difference between locations pooled overyears (x 2 = 239, df = 2) is not surprising and also not particularly meaningful.Con- sequently, given the range of interyear variation, we cannot adequately test for interlocation differences in the migratory corridor and therefore have applied distance estimation corrections equally to data from both locations. Within each year, the distribution of distance es- timates was tested for a within-season change, since our verification experiments were conducted during roughly the middle third of the migration. For this, the data were divided into early (10-29 December), mid (30 December-18 January), and late (19 Janu- ary-6 February) time periods. As with the first 1 1-yr data (Reilly et al. 1980), the 1978-79 and 1979-80 distributions have no seasonal differences indicated by contingency table analysis (x 2 = 8.54, 7.13, < X% 0.05)1 but do have significantly different mean dis- tance observations (F = 16.34,26.91 > F 2 xooh ). Con- sequently, as with the first 1 1 yr, only data from the middle third of the migration were used for com- parison with aerial results in Equation (3). No significant period differences were indicated for the 1978-79 and 1979-80 censuses, in the ANOVA testing for effects on numbers of whales recorded per 5-h shift, from variation between observers and from period (morning or afternoon). Similar results were obtained in the comparison of observers within each year (F = 1.242, 2.003, F 1I18 ). The data were 4.00 3.50 u 1- < 3.00 2 1- 2.50 LU N tn 2.00 Q. 1.50 0.5 _l_ _J_ _L JL 0.25 0.50 0.75 1.00 1.25 1.50 1.75 DISTANCE FROM SHORE (nautical miles) 2.00 FIGURE 4. — Mean pod size estimates by distance from shore, with 95% confidence limits, from 13 annual gray whale census, 1967-68 through 1979-80, n = 23,678 observations. therefore considered homogeneous for pooling over these factors. The results from 1978-79 are somewhat different than the results from the first 11 yr, in the rate of whales recorded per hour of day. The mean counts show significant differences in an ANOVA (F= 3.717 >F 9, », 0.05 ) which are due to the depressed value for 0700-0800 h (Fig. 5). Multiple range tests (Duncan's, Stude