^^^TE^ O^ ^ Fishery Bulletin ^ i National Oceanic and Atmospheric Administration • National Marine Fisheries Service % 1 9 1 A Vol. 73, No. 1 y/ooJs hole; January 1975 MAY, ROBERT C. Effects of temperature and salinity on fertilization, embryonic development, and hatching in Bairdiella icistia (Pisces: Sciaenidae), and the effect of parental salinity acclimation on embryonic and larval salinity tolerance .... 1 FOX, WILLIAM W., JR. Fitting the generalized stock production model by least- squares and equilibrium approximation 23 SMAYDA, THEODORE J. Net phytoplankton and the greater than 20-micron phytoplankton size fraction in upwelling waters off Baja California 38 ANDERSON, LEE G. Optimum economic yield of an internationally utilized com- mon property resource 51 CARR, WILLIAM E. S., and JAMES T. GIESEL. Impact of thermal effluent from a steam-electric station on a marshland nursery area during the hot season 67 LINDALL, WILLIAM N., JR., WILLIAM A. FABLE, JR., and L. ALAN COLLINS. Additional studies of the fishes, macroinvertebrates, and hydrological conditions of upland canals in Tampa Bay, Florida 81 LOUGH, R. GREGORY. A reevaluation of the combined effects of temperature and salinity on survival and growth of bivalve larvae using response surface techniques 86 HORN, MICHAEL H. Swim-bladder state and structure in relation to behavior and mode of life in stromateoid fishes 95 McEACHRAN, JOHN D., and J. A. MUSICK. Distribution and relative abundance of seven species of skates (Pisces: Rajidae) which occur between Nova Scotia and Cape Hatteras 110 KJELSON, MARTIN A., DAVID S. PETERS, GORDON W. THAYER, and GEORGE N. JOHNSON. The general feeding ecology of postlarval fishes in the Newport River estuary 137 KNIGHT, MARGARET D. The larval development of Pacific Euphausia gibboides (Euphausiacea) 145 WING, BRUCE L. New records of Ellobiopsidae (Protista (incertae sedis)) from the North Pacific with a description of Thalassomyces albatrossi n.sp., a parasite of the mysid Stilomysis major 169 BERRIEN, PETER L. A description of Atlantic mackerel. Scomber scombrus, eggs and early larvae , ittt 186 (Continued on back cover) Seattle, Washington U.S. DEPARTMENTOFCOMMERCE Frederick B. Dent, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Admiryistrator NATIONALMARINE FISHERIES SERVICE Robert W. Schoning, Director Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 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. Reuben Lasker Scientific Editor, Fishery Bulletin National Marine Fisheries Service Southwest Fisheries Center La Jolla, California 92037 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. William H. Bayliff Inter-American Tropical Tuna Commission Dr. Daniel M. Cohen National Marine Fisheries Service Dr. Howard M. Feder University of Alaska Mr. John E. Fitch California Department of Fish and Game Dr. Marvin D. Grosslein National Marine Fisheries Service Dr. J. Frank Hebard National Marine Fisheries Service Dr. John R. Hunter National Marine Fisheries Service Dr. Arthur S. Merrill National Marine Fisheries Service Dr. Virgil J. Norton University of Rhode Island Mr. Alonzo T. Pruter National Marine Fisheries Service Dr. Theodore R. Rice National Marine Fisheries Service Dr. Brian J. Rothschild National Marine Fisheries Service Mr. Maurice E. Stansby National Marine Fisheries Service Dr. Maynard A. Steinberg National Marine Fisheries Service Dr. Roland L. Wigley National Marine Fisheries Service Kiyoshi G. Fukano, Managinp; Editor The Secretary of Commerce has determined that the publication of this periodical is necessary in the transact the public business required by law of this Department. Use of funds for printing of this periodical has been app by the Director of the Office of Management and Budget through May 31, 1977. ion of approved Fishery Bulletin CONTENTS Vol. 73, No. 1 January 1975 MAY, ROBERT C. Effects of temperature and salinity on fertilization, embryonic development, and hatching in fia/rrfieZ/a icistia (Pisces: Sciaenidae), and the effect of parental salinity acclimation on embryonic and larval salinity tolerance .... 1 FOX, WILLIAM W., JR. Fitting the generalized stock production model by least- squares and equilibrium approximation 23 SMAYDA, THEODORE J. Net phytoplankton and the greater than 20-micron phytoplankton size fraction in upwelling waters off Baja California 38 ANDERSON, LEE G. Optimum economic yield of an internationally utilized com- mon property resource 51 CARR, WILLIAM E. S., and JAMES T. GIESEL. Impact of thermal effluent from a steam-electric station on a marshland nursery area during the hot season 67 LINDALL, WILLIAM N., JR., WILLIAM A. FABLE, JR., and L. ALAN COLLINS. Additional studies of the fishes, macroinvertebrates, and hydrological conditions of upland canals in Tampa Bay, Florida 81 LOUGH, R. GREGORY. A reevaluation of the combined effects of temperature and salinity on survival and growth of bivalve larvae using response surface techniques 86 HORN, MICHAEL H. Swim-bladder state and structure in relation to behavior and mode of life in stromateoid fishes 95 McEACHRAN, JOHN D., and J. A. MUSICK. Distribution and relative abundance of seven species of skates (Pisces: Rajidae) which occur between Nova Scotia and Cape Hatteras 110 KJELSON, MARTIN A., DAVID S. PETERS, GORDON W. THAYER, and GEORGE N. JOHNSON. The general feeding ecology of postlarval fishes in the Newport River estuary 137 KNIGHT, MARGARET D. The larval development of Pacific Euphausia gibhoides (Euphausiacea) 145 WING, BRUCE L. New records of Ellobiopsidae (Protista {incertae sedis)) from the North Pacific with a description of Thalassomyces albatrossi n.sp., a parasite of the mysid Stilomysis major 169 BERRIEN, PETER L. A description of Atlantic mackerel, Scomber scombrus, eggs and early larvae 186 (Continued on next page) Seattle, Washington For sale by the Superintendent of Documents, U.S. Government Printing Office, Washing- ton, D.C. 20402 — Subscription price: $11.80 per year ($2.95 additional for foreign mail- ing). Cost per single issue - $2.95. Contents — continued GILMARTIN, MALVERN, and NOELIA REVELANTE. The concentration of mer- cury, copper, nickel, silver, cadmium, and lead in the northern Adriatic anchovy, Engraulis encrasicholus, and sardine, Sardina pilchardus 193 ANDERSON, WILLIAM W., JACK W. GEHRINGER, and FREDERICK H. BERRY. The correlation between numbers of vertebrae and lateral-line scales in western Atlantic lizardfishes (Synodontidae) 202 RICE, STANLEY D., and ROBERT M. STOKES. Acute toxicity of ammonia to several developmental stages of rainbow trout, Salmo gairdneri 207 Notes TILLMAN, MICHAEL F. Additional evidence substantiating existence of northern subpopulation of northern anchovy, Engraulis mordax 212 LOOSANOFF, VICTOR L. Comment. Introduction of Codium in New England waters 215 EFFECTS OF TEMPERATURE AND SALINITY ON FERTILIZATION, EMBRYONIC DEVELOPMENT, AND HATCHING IN BAIRDIELLA ICISTIA (PISCES: SCIAENIDAE), AND THE EFFECT OF PARENTAL SALINITY ACCLIMATION ON EMBRYONIC AND LARVAL SALINITY TOLERANCE^ Robert C. May^ ABSTRACT Eggs and larvae of the sciaenid fish bairdiella, Bairdiella icistia, were obtained from fish matured in the laboratory by photoperiod manipulation and induced to spawn by hormone injections. The effects of temperature and salinity on fertilization, embryonic development, hatching, and early larval survival were studied with the material thus obtained, and the effects on gametes of parental salinity acclima- tion were also investigated. Fertilization took place over a wide range of temperatures and salinities, but was completely blocked at salinities of 10%o and below. A low level of spermatozoan activity may have accounted for the lack of fertilization at low salinities. Successful embryonic development occurred between temperatures of approximately 20° and 30°C, and salinities of 15 and 40%o. The production of viable larvae was estimated to be optimal at a temperature of 24.5°C and a salinity of 26.6''/oo. An interaction of the two factors was apparent, development at high salinities being most successful at low temperatures and development at high temperatures being most successful at low salinities. The stage of maturity of the spawning female had a great influence on the overall viability of the eggs produced, as well as on their response to temperature and salinity. Adult bairdiella matured sexually in dilute seawater with a salinity of 15%o, and the salinity tolerance of the eggs produced by these fish was unaltered. The bairdiella, Bairdiella icistia (Jordan and Gil- bert), is a sciaenid fish native to the Gulf of California. In 1950 the species was successfully introduced into the Salton Sea, a large saline lake in southern California (Whitney 1961). Salton Sea water has an ionic composition different from that of ocean water (Carpelan 1961; Young 1970), and its overall salinity, now approximately 38%o,^ is rising at a rate of about l%o every 3 yr (U.S. Department of the Interior and the Resources Agency of California 1969). This rising salinity has caused concern that the present sport fishery in the Salton Sea (based on several fish species, including bairdiella) will fail when the upper sa- linity tolerances of the fishes are exceeded 'Based on a portion of a dissertation submitted in partial satisfaction of the requirements for the Ph.D. degree at the University of California at San Diego, Scripps Institution of Oceanography. ^Hawaii Institute of Marine Biology, P.O. Box 1346, Kaneohe, HI 96744. ^This value varies somewhat with season and location in the Salton Sea. (Walker et al. 1961). Lasker et al. (1972) found that the survival of bairdiella eggs and early lar- vae was severely inhibited by Salton Sea water at a salinity of 40%o; thus, at the present rate of salinity increase, the bairdiella population may suffer a loss in recruitment within the next 10 yr. The work reported in this paper was undertaken to provide more information on the salinity toler- ance of bairdiella during early development, espe- cially as influenced by temperature and by the acclimation of spawning parents to different salinities. Because of poor embryonic and larval survival in Salton Sea water (May 1972), these experiments were all conducted in seawater of ordinary ionic composition. The effects of Salton Sea water per se and their implications for the population of bairdiella in the Salton Sea will be discussed elsewhere (May in preparation). Bairdiella normally spawn during April and May in the Salton Sea (Whitney 1961; Haydock 1971). However, thanks to the work of Haydock (1971), bairdiella can be induced to mature and spawn in the laboratory at any time of the year, Manuscript accepted March 1974. FISHERY BULLETIN: VOL. 73, NO, 1, 1975. making bairdiella eggs and larvae extremely favorable material for experimentation. In addi- tion to providing a year-round supply of eggs, laboratory spawning techniques have permitted maintaining bairdiella at different salinities dur- ing maturation and spawning in order to test the effect of parental salinity acclimation on the salin- ity tolerance of the gametes, embryos, and larvae. MATERIAL AND METHODS Capture and Maintenance of Fish Methods used for collecting and maintaining bairdiella were nearly identical to those described by Haydock (1971). Adult bairdiella were cap- tured with a 60-m beach seine on the west coast of the Salton Sea, just north of the Salton Bay Yacht Club. Rectangular fiberglass tanks of 2,000-liter capacity were used to hold fish in the laboratory and were supplied with continuously flowing warm (22°C) seawater from the Southwest Fisheries Center system (Lasker and Vlymen 1969). Water was filtered through pol5T)ropylene GAF'* snap-ring filter bags of 50- ^^m pore size (GAF Corp., Greenwich, Conn.). Mercury lamps provided illumination (Haydock 1971) and the photoperiod was controlled as desired by timers. The fish were fed ad libitum twice each day with ground squid, supplemented by ground red crab, Pleuroncodes planipes , at a ratio of approximately 1 part of crab to 6 of squid (wet weight). The red crabs were intended as a source of carotenoids because some authors have indicated that paren- tal carotenoid deficiency may affect the viability of offspring (Hubbs and Stavenhagen 1958). Several outbreaks of the parasitic ciliate, Cryp- tocaryon irritans Brown, occurred (Wilkie and Gordin 1969) and were effectively controlled by adding copper sulfate at 0.2 ppm as Cu^ ^ in the morning and late afternoon, allowing the chemi- cal to be diluted in the interim by the continuously flowing seawater. Whenever fish were handled, they were subsequently treated with Furacin an- tibiotic (Eaton Veterinary Laboratories, Norwich, N.Y.) at 130 ppm, which was gradually diluted in the open seawater system. This precaution effec- tively controlled bacterial infections and allowed repeated handling of fish without adverse conse- quences. FISHERY BULLETIN: VOL. 73, NO. 1 Induced Maturation and Spawning Fish which had ripe gonads when captured were maintained in this condition for several months by exposing them to a photoperiod of 16 h light, 8 h darkness (16L:8D) at approximately 22°C (Haydock 1971). Prolonged exposure of female fish to long days resulted in eventual resorption of the ova. After a group offish had been spawned out or had begun gonadal resorption, they were shifted to a short photoperiod (9L:15D) and colder water (15°C). After being held on short days for a few months, fish could then be brought to maturity by increasing the photoperiod at a rate of 30 min per day until 16L:8D was reached; after about 3 mo on 16L:8D at 22°C, the fish had developed mature ovaries and were ready to spawn. Successful spawning could be induced over a period of at least two or three more months before gonadal resorp- tion began. Photoperiod manipulation was effec- tive in inducing ovarian maturation regardless of the time of year, and the experiments described in this paper were conducted in the summer, fall, and winter instead of during the normal spring spawn- ing period. Bairdiella kept in the laboratory vary consider- ably in their ovarian development (Haydock 1971). In the present study the maturity of female fish was assessed from ovarian biopsies taken with a glass capillary tube (Stevens 1966). At first only the maximum oocyte diameters were recorded immediately after sampling, along with qualita- tive notes concerning the amount of ovarian stroma in the sample. When it became apparent that this was not a sufficiently sensitive measure of the state of maturity, the samples were pre- served in 3% Formalin (in 50% seawater) and all oocyte diameters of 175 /^m or greater were mea- sured with an ocular micrometer a day or so later,^ giving an oocyte size-frequency distribution based on measurements of approximately 100 oocytes. The fish which had been biopsied in this manner were marked individually on the lower jaw with injections of the dye, National Fast Blue 8GXM ( = Fast Turquoise PT) (Kelley 1967; Haydock 1971). Mature female fish weighing 100-150 g were injected in the epaxial musculature near the dor- sal fin with 100 lU of gonadotropin from pregnant mare's serum (PMS; Sigma Chemical Co., St. Louis, Mo.) in a carrier of Ringer's solution, after ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 'No measurable oocjrte shrinkage occurred even after a week of preservation. MAY: EFFECTS ON BAIRDIELLA ICISTIA being anesthetized with MS-222 (tricaine methanesulfonate) at 150 ppm. Haydock (1971) found that salmon pituitary glands and PMS were both effective in inducing ovulation in bairdiella. PMS was used here because it had a standardized activity and was more readily available and easier to prepare than salmon pituitaries. The injected fish were checked for ovulation 30 h after injection and at hourly intervals thereafter until ovulation took place (Haydock 1971). In the vast majority of cases, ovulation occurred 30 or 31 h after the hor- mone injection. Spawning bairdiella of the size used in these experiments will jdeld 100,000 or more eggs (Haydock 1971). Male fish remained in a running ripe condition in the laboratory and did not require hormone injections. Bairdiella do not spawn spontaneously in captivity, whether in- jected or not, and gametes must be obtained by stripping. Haydock (1971) demonstrated that eggs must be fertilized 1 or 2 h after ovulation if maxi- mal viability is to be retained. Fertilization Approximately 1,000 to 3,000 eggs were squeezed from an anesthetized, freshly ovulated female and added to a petri dish containing 75 ml of water of the desired temperature and salinity. When fertilizations under a number of conditions were to be made, eggs were added to all petri dishes before sperm was added. Sperm from a lightly anesthetized male fish was taken up in a pasteur pipette which was immediately filled and flushed with water from a petri dish containing eggs. Eggs and sperm were swirled in the dish for several seconds. This procedure was repeated for every dish, fresh sperm being obtained each time. After cleavage had begun, random samples (usu- ally 100 to 300 eggs) were taken from each petri dish, preserved in 3% Formalin and later ex- amined, and the number cleaving recorded. The percentage of eggs cleaving was taken as the per- centage fertilized. Spermatozoan activity was measured in various salinities by placing a drop of sperm under a cover slip, focusing on it with a compound microscope at 430 X and adding seawater of the desired salinity. At frequent intervals after hydration, the activity of spermatozoa was rated on an arbitrary scale of 0 to 5, 0 being no activity and 5 being maximal activity. All such tests were conducted at approx- imately 25°C. More than 70 runs were made utiliz- ing spermatozoa from nine fish, each run compris- ing between 4 and 15 observations, depending on the duration of activity. Incubation Developing eggs from the fertilization dishes were counted out by pipette under a dissecting microscope, rinsed with clean water of the test salinity to remove sperm, and transferred to in- cubators. The transfer of eggs was usually com- pleted by the time the blastula stage had been reached, within 3 or 4 h after fertilization. One hundred developing eggs were placed in each in- cubator, and there were two replicate incubators for each experimental treatment. Each incubator (Figure 1) consisted of a 400-ml Pyrex beaker with an insert made from a truncated pol3T)ropylene beaker with its bottom covered by Nitex nylon mesh (350- /i m mesh opening). Three hundred mil- liliters of water were added to each incubator. A slow stream of air bubbles in a centrally positioned glass tube created a flow of water such that eggs which rested on the bottom at low salinities were bathed by a continuous flow of aer- ated water (Figure 1). One or two days before each experiment, seawa- ter with a salinity of approximately 60°/oo was made by adding artificial sea salts ("Instant Ocean"; Aquarium Systems, Inc., Wickliffe, Ohio) to HA Millipore-filtered seawater. This solution was filtered through paper (Whatman No. 1) to eliminate a residual cloudiness and then diluted with deionized water to the desired test salinities. Batches of seawater were aerated for several Figure 1. — Egg incubator. A) Parafilm cover; B) polyethylene air tube; C) 400-ml Pyrex beaker; D) water line; E) 250-ml IX)lypropylene beaker, cut off at bottom; F) glass chimney; G) Nitex mesh. Arrows indicate direction of water flow. FISHERY BULLETIN: VOL. 73. NO. 1 hours before each experiment to stabilize oxygen tension and pH. Potassium penicillin G (50 lU/ml) and streptomycin sulfate (0.05 mg/ml) were added to the water just before it was placed in the in- cubators. Salinities were calculated by multiply- ing chlorinity values (Schales and Schales 1941) by 1.80655 (Johnston 1964) and remained within ±0.5%o of the original salinity during an experi- ment. Temperatures were maintained within ±0.2°C of the desired value by immersing petri dishes and incubators in water baths equipped with cooling coils and thermostatically controlled heaters. The incubators were illuminated con- tinuously from fluorescent room lamps which gave an intensity of from 320 to 480 Ix at the water surface. Dissolved oxygen concentration in the in- cubators decreased with increasing temperature and salinity, and measured concentrations were wathin 2 or 3% of the saturation values given by Kinne and Kinne ( 1962). The highest oxygen con- tent (at 18°C and 15%o) was 6.24 ml/liter, and the lowest (at 30°C and 55%o) was 4.05 ml/liter. The pH in the incubators increased with increasing salinity and decreasing temperature, varying be- tween 8.08 and 8.27. The percentage hatching and the condition of the larvae at hatching were recorded for each in- cubator. Supplementary containers (20-ml petri dishes) with 30 fertilized eggs each, were provided at each treatment to allow examination of eggs during development vdthout disturbing the eggs in the incubators. Hatched larvae were not fed; some were kept in 400-ml beakers (without the polypropylene inserts used prior to hatching) and the pattern of mortality of the starved larvae re- corded, and some were used in experiments on the temperature and salinity tolerance of yolk-sac larvae (May 1972). During an early experiment, histological prep- arations were made of newly hatched larvae from different salinities at 25°C. Larvae were fixed in Bouin's solution, dehydrated in ethanol-normal butyl alcohol, embedded in paraffin, and sectioned transversely at 8 idm. Sections were stained with Mayer's hemalum and eosin. Experimental Series Two series of experiments on fertilization suc- cess, embryonic development, and hatching success were conducted, each series involving observations at 25 different combinations of tem- perature and salinity. Each series included two separate hormone-induced spawnings offish held under identical conditions. The two spawnings in each series constituted a composite factorial array of treatments (a 3 x 5 plus a 2 x 5 factorial); this design, similar to those employed by Alderdice and his colleagues (Alderdice and Forrester 1967, 1971a, b; Alderdice and Velsen 1971), allowed coverage of a large factor space without utilizing all possible combinations of treatments. The ranges of temperature and salinity employed cov- ered the viable ranges for bairdiella eggs, as de- termined in preliminary experiments. Table 1 in- dicates the temperatures and salinities in which eggs were fertilized and incubated in the two spawmings of each series. The fish utilized for Series A were captured to- ward the end of the spawning season in the Salton Sea on 7 June 1971 and maintained on a 16L:8D photoperiod in 22°C water until the first hormone-induced spawning of the series on 23 August 1971 and the second on 1 September 1971. Ovarian biopsies indicated that the eggs were ready for spawning at this time, but only max- imum oocyte diameters were measured and no oocyte size-frequency distributions were obtained. The tests were repeated in a second series of exper- iments. Series B. A group offish captured in the Salton Sea on 20 May 1970 was shifted gradually Table 1. — Dates and temperature-salinity conditions for exper- iments in Series A and B, 1971. There were two spawnings, performed at different dates, in each series; each spawning utilized eggs and sperm from different fish. Temper- Series A Series B ature (C) Salinity (°/oo) 23 Aug. 1 Sept. 25 Nov. 3 Dec. 18 10 X X 20 X X 30 X X 40 X X 50 X X 21 15 X X 25 X X 35 X X 45 X X 55 X X 24 10 X X 20 X X 30 X X 40 X X 50 X X 27 15 X X 25 X X 35 X X 45 X X 55 X X 30 10 X X 20 X X 30 X X 40 X X 50 X X 4 MAY: EFFECTS ON BAIRDIELLA ICISTIA from a short photoperiod to a 16L:8D photo period between 25 June and 10 July 1971. Half of these fish were transferred gradually to 15%o and al- lowed to mature in that salinity as described below, while the other half were kept in sea water (approximately 33%o) and used to supply eggs for the Series B experiments. Prior to these spawn- ings, ovarian biopsies were taken and oocjrte size- frequency distributions determined to assure that the fish were fully mature. Acclimation of Spawning Fish to Low Salinity These fish came from the same collection as those used to supply eggs in the Series B experi- ments and were brought to maturity simultane- ously with them. The salinity was lowered to 15%o over a period of 8 days by mixing seawater with an increasing proportion of fresh water. The day length was then increased from 9 to 16 h in 30-min increments, and the temperature was raised from 16° to 22°C over the same period (Fig- ure 2). The tap water had been dechlorinated by passage through a commercial charcoal filter, and the mixed tap water and seawater flowed through the fish tank at 1 ,000 liters per hour (the same flow rate was maintained in the tank receiving straight seawater). Salinity was monitored daily in the seawater and low-salinity tanks. Variations were relatively slight during the period of gonadal maturation, monthly means ranging from 32.7 3 16 I uj 10 35 30 15 - , • • '^^ — TemperQtu Sahnity- ° o 0 o -I 1 I I I I L. J i I \ L 24 O - 22 lij 20 3 18 2 U 16 Q- 2 8 10 12 14 16 18 20 22 24 26 28 30 2 4 6 8 10 12 14 ^ JUNE 'I JULY 1 Figure 2. — Day length, temperature, and salinity during tran- sition period, when fish were transferred to low-salinity water, warm temperatures, and long days. to 33.3%o in the seawater tank and from 15.2 to 15.7°/oo in the low-salinity tank. Female fish living at 15%o were injected with PMS on 25 October, 8 November, and 16 November 1971. Eggs were fertilized (with sperm from males also acclimated to 15"/oo) and incu- bated as described above, at salinities of 10, 15, 20, 30, 40, 45, and 50%o. The temperature was 24.0°±0.2°C in all experiments with eggs from fish acclimated to low salinity. Hatched larvae were kept in 400-ml beakers at their original salinity to determine the percentage surviving to yolk exhaustion. The activity of spermatozoa from fish acclimated to 15%o was assessed at various salinities as described above. RESULTS Spermatozoan Activity Bairdiella spermatozoa measured 40 ^m in total length, the head being about 2.5 Mm long. In distilled water and dechlorinated tap water, spermatozoa showed at most only slight move- ment, usually in the form of very slow undulations which lasted at least 10 min. After approximately 1 min, the heads of many of these spermatozoa seemed to acquire bright rings, which an oil- immersion lens revealed to be the tail curled around the head, still undulating slowly. Bairdiella spermatozoa became activated im- mediately upon contact with seawater (Haydock 1971), and the intensity of activity varied with salinity and time after initial contact with water. Spermatozoa were most active at the higher salinities but remained active longest at the lower salinities. At 10 and 15%o, a small smount of ac- tivity remained even as long as 10 min after hydration, but at 10%o spermatozoa seldom showed activity above level 3 and at 15%o they only rarely and briefly attained level 5 (Figure 3). At 25%o all activity ceased by 4 min after hydra- tion, and at 35%o no activity was usually seen after 3 min. At 45 and 55"/oo, activity had com- pletely stopped by 1.5 min after hydration. On rare occasions, at salinities between 15 and 55%o, slow undulations of some spermatozoa were ob- served after other movements had ceased. No dif- ference was noted between spermatozoan activity in seawater and in Salton Sea water, nor between the activity of spermatozoa from fish acclimated to a salinity of 15%o and from those kept at 33%o. Spermatozoan activity in dilute suspensions of FISHERY BULLETIN: VOL. 73, NO. 1 2 3 4 TIME (minutes) Figure 3. — Spermatozoan activity in four salinities as a func- tion of time after hydration. The activity levels are described in the text. sperm was the same as when hydration was car- ried out underneath a cover slip, indicating that the high concentration of spermatozoa in the lat- ter case did not seriously affect the level or dura- tion of their activity. Maturity of Spawning Fish Examination of many fish during this project showed that 500 fj.Tn was approximately the max- imum diameter attained by oocytes in bairdiella before gonadal hydration. During hydration, which occurs in the laboratory only after an injec- tion of gonadotropic hormone, the accession of water swells the eggs to 700 iimor more, the size at spawning. Ovarian biopsies showed that the two female fish used to supply eggs in the Series A experiments had oocjd^es as large as 500 jjun before injection. The ooc3d;e size-frequency distributions for the fish used in Series B (Figure 4) also showed maximum diameters of about 500 mn, and there were modes at 420 to 455 /am for the first fish and 385 to 455 fim for the second in Series B. The much poorer fertilization and hatching suc- cess in Series A (see below) indicates that max- imum oocyte diameter is not necessarily a good index of readiness for spawning. By this method it is impossible to tell whether there is a mode at the large end of the size-frequency distribution, as is characteristic of fish which are ready to spawn. The fish used to supply eggs in Series A were probably captured after the peak of spawning in the Salton Sea and their gonads at that time were 30 >- 20 O z UJ o ^ '0 1 ii _^. L 175 2 10 245 280 315 350 385 420 455 490 OOCYTE DIAMETER dim) 30 r >■ 20 o 3 a liJ 10 cr I l« M I 175 210 245 280 315 350 385 420 455 490 OOCYTE DIAMETER (pm) Figure 4. — Oocyte size-frequency distributions, based on ovarian biopsies, from fish used in Series B experiments, a) fish spawned on 25 November 1971, b) fish spawned on 3 December 1971. probably either partly spent or beginning to be resorbed (see Haydock 1971). It was hoped that subsequent exposure to long days would induce ovarian recrudescence, but instead this treatment over a period of 2.5 mo apparently maintained the gonads at a suboptimal state of maturity or al- lowed them to regress even further (see Haydock 1971). A postspawning refractory period (Har- rington 1959; Sehgal and Sundararaj 1970) may exist in bairdiella, but it cannot be very pro- nounced, since not only were eggs obtained from these fish after hormone injections in August and September, but at least 60% of the eggs could be fertilized under optimum conditions (see below). The fish used in Series B had completely regressed gonads when they were first exposed to long days in July 1971. By November 1971 or earlier they had developed ovaries capable of producing a large proportion of viable eggs, showing as much as 90% fertilization. Fertilization Although fertilization did take place at a salin- ity of 15%o, it was completely blocked at 10%o (Table 2). In order to examine this phenomenon further, unfertilized eggs were placed in 10%o water for various periods of time and then trans- 6 MAY: EFFECTS ON BAIRDIELLA ICISTIA Table 2. — Percentage fertilization at various combinations of temperature and salinity in Series A and B. Temperature (°C) Salinity (°/oo) Percentage fertilization Table 4. — Survival and hatching of fertilized eggs transferred from 20''/oo to 10%o at various stages. Eggs were incubated in 20-ml petri dishes. The stages are described in Table 6. Series A Series B 18 21 24 27 30 10 20 30 40 50 15 25 35 45 55 10 20 30 40 50 15 25 35 45 55 10 20 30 40 50 0 14.5 48.5 24.4 14.8 13.4 43.5 35.8 38.8 1.9 0 63.1 43.2 7.4 5.7 33.7 52.6 60.1 16.8 0 0 48.7 15.3 2.7 0 0 28.9 81.6 87.5 41.6 18.4 69.9 49.4 59.9 52.5 0 62.3 87.8 81.3 50.8 30.3 77.2 76.7 82.0 67.4 0 74.6 89.8 68.7 23.1 ferred to 20%o and immediately exposed to sperm. The results (Table 3) showed that 10%o water did not render the eggs infertile: even after 20 min at 10%o, a large proportion of the eggs could be fer- tilized at 20%o and develop to hatching, although there was no fertilization in controls kept at 10%o . It was also found that eggs fertilized at 20%o could be transferred to 10%o and develop to hatching (Table 4). Thus the actual process of fertilization was somehow blocked at 10%o. In Series A, fertilization was much more sensi- tive to high salinities and there seemed to be a greater temperature-salinity interaction than in Series B, with fertilization being more successful at high salinities when the temperature was low (Table 2). In Series B, at salinities above 10%o, Table 3. — Effect of exposure to IC/oo water for various periods of time on fertilizability of bairdiella eggs at 20''/oo . At each time interval, between 200 and 400 eggs were transferred from 10 to 20''/oo, exposed to sperm, and later examined for fertilization. Thirty fertilized eggs from each group were followed until hatch- ing. Time Fertilized Hatching at lO^/oo at 20''/oo at 20<'/oo (%) (%) 45 s 92.5 60.0 2 min 90.2 66.7 5 min 75.4 74.2 10 min 48.4 43.3 20 min 44.8 75.9 Survival to Stage at Number of eggs stage VI Hatching transfer transferred (%) (%) lie 30 96.7 63.3 IV 31 93.6 41.9 V 30 100 66.7 VII 29 — 62.1 fertilization was in nearly all cases over 50% , the few exceptions being at low temperature/low sa- linity and high temperature/high salinity com- binations. A maximum of 89.8% fertihzation was observed at 30°C-30%o in Series B. The thermal limits for fertilization in Series B were evidently beyond the range tested (18°-30°C). Normal Development It will be helpful to outline the normal pattern of development of bairdiella eggs before discussing alterations in this pattern induced by various combinations of temperature and salinity. Newly spawned bairdiella eggs are approximately 725 jum in diameter and contain an oil globule with a diameter of about 18 pm. Occasionally there are two or three smaller oil globules instead of a single one. Like most pelagic eggs, bairdiella eggs float with the animal pole downward. The development o^Bairdiella icistia eggs (Table 5, Figure 5) follows the pattern typical for small pelagic fish eggs and is not greatly different from that of 5. chrysura as described by Kuntz (1915). Ahlstrom's numerical designation of developmental stages (Ahlstrom 1943) has been adopted here (Table 5), although some slight modifications of his scheme were necessary, and some of the stages have been broken down into substages. The times required to reach certain stages are listed (Table 5) for eggs at 33%o at 25°C, based on observations made during a preliminary experiment in 1970. The newly hatched larvae are approximately 1.7 mm in length (snout to tip of notochord) and in ordinary seawater float upside down near the surface of the water. Incubation Time The time between fertilization and hatching varied with temperature and with salinity, and the patterns of hatching determined from the sup- plementary containers in Series B are shown in FISHERY BULLETIN: VOL. 73. NO. 1 Table 5. — Normal development of bairdiella eggs. Designation of stages in general follows Ahlstrom ( 1943), and times required to reach various stages are given for eggs in 33%o water at 25°C. Approximate Ahlstrom Sub- time after Description stage stage fertilization 1 a — Unfertilized egg b 2 min blastodisc II a 40 min 2 blastomeres b 50 min 4 blastomeres c 60 min 8 blastomeres d 2fi Morula e 3h Blastula. periblast very apparent III a 6h Early gastrula. germ ring encircles as much as 1/3 of yolk, embry- onic shield rudimentary b 7h Mid gastrula, embryonic shield expands, germ ring encircles as much as 2/3 of yolk IV 8h Late gastrula, primitive streak forms V 9h Blastopore closes, optic vesicles and Kupf- fer's vesicle form VI a lOh Somites begin to form: scattered melanophores appear, most dorsally behind optic vesicles, a few extending posterlad along notochord b 12h Lens and otic vesicles form, tip of tail reaches oil droplet VII 15h Tail has moved beyond oil droplet and lifted off yolk: finfold apparent VIII 17h Tail well beyond oil drop- let; embryo twitches occasionally; heartbeat regular — 20 ti Hatching Figure 6. Series A showred similar patterns, but due to poorer survival the data are less complete and are not shown. In Figure 6 the cumulative percentage hatched has been plotted on a proba- bility scale against time on an arithmetic scale; a straight line in this type of plot indicates a normal distribution (Sokal and Rohlf 1969), which is to be expected if differences in hatching time are due simply to random individual variation. At 30°C hatching was normally distributed for all salinities, but this was not true at the lower tem- peratures. At 27°C there was a plateau at 25°/oo, indicating that the hatching of certain eggs was delayed. At 24°C, hatching was distributed ap- proximately in a normal fashion at 20, 40, and 50%o, but at 30%o there was an inflection, the rate of hatching being slower after 23 h than be- fore. At 21°C, hatching was distributed normally for 15, 35, and 45%o, but at 25"/oo hatching took place in two phases separated by a 3-h period dur- ing which no hatching took place. The time required for 50% of the larvae to hatch, estimated by graphical interpolation, decreased from 35.2 h at 21°C-25%o to 16.0 h at 27°C-25%o. The estimated time at 50% hatching was slightly later at 30°C than at 27°C, although hatching began 2 h earlier in the former (see Figure 6). No clear-cut effect of salinity on median hatching times is discernible, but Figure 6 shows that hatching was completed more rapidly at the higher salinities (35%o and above). The duration of hatching (the time between the appearance of the first and last hatched larvae) tended to be greater at the lower salinities and temperatures. Embryonic Mortality In certain treatments some surviving embryos failed to hatch but continued to develop wathin the chorion. Alderdice and Forrester (1971b) intro- duced the apt term, "postmature unhatched eggs" to describe such cases. Almost without exception, the postmature unhatched embryos were de- formed in some way, usually bent and abnormally small. Often in such eggs part of the chorion was eventually digested away (Figure 7f), presumably by hatching enzymes, but the weak embryo was incapable of breaking completely free. Postma- ture unhatched eggs were most common at the low salinities, and the greatest proportion occurred at 30°C-20%o (Table 6). Eggs in Series A showed much higher mortality than those in Series B, especially at the higher temperatures and salinities (Table 7). The follow- ing description of embryonic mortality refers primarily to the eggs in Series B, which are consid- ered more representative of normal, healthy eggs. The higher mortality in Series A usually showed up very early in embryonic development (prior to stage V); otherwise the two series showed similar trends. No eggs hatched at 18°C and nearly all died during stage lie (blastula). After the second cleav- age at 18°C the blastomeres assumed a clover- leaf appearance which was not seen at higher temperatures (Figure 7; cf. Figure 5). Subsequent cleavages at 18°C were rather irregular, and dur- ing the blastula stage much of the cytoplasm gathered into isolated clumps, and the periblast became unusually large (Figure 7). Nearly all eggs stopped developing at this stage. 8 MAY: EFFECTS ON BAIRDIELLA ICISTIA n Figure 5.— Normal developmental stages of Bairdiella icistia at 25°C-33»/oo. a) stage lb, 4 min after fertilization; b) stage Ila, 40 min; c) stage lib, 50 min; d) stage lie, 60 min; e) stage lid, 2 h; f) stage He 3 h- g) stage Ilia, 6 h; h) stage Illb, 7 h; i) stage IV, 8 h; j) stage V, 9 h; k) stage VI, 12 h; 1) stage VII, 15 h- m) stage VIII, 17 h; n) newly hatched larva. FISHERY BULLETIN: VOL. 73, NO. 1 -30C- -27C 24C ir -2IC- 1 ju^ 11 I r- I I I III I I „| I I I, I I, I I I, I , I, I ,1 I I I I I. I I I I I ,1 I — I I I I I I 'III I I I I I I I I I I I I I I I I l_l l_L 14 16 18 20 16 18 20 22 24 26 28 30 32 34 36 38 40 42 TIME (hours after fertilization) Figure 6. — Cumulative percentage of larvae hatching, as a function of time after fertiliza- tion, for the Series B experiments. Percentage hatching is plotted on a probabihty scale; lines were fitted by eye. Table 6. — Percentage of postmature unhatched eggs at various combinations of temperature and salinity in Series B. There are two replicates at each treatment combination. Series A showed similar trends. Salinity (°/oo) Temperature 21 24 27 30 15 20 25 30 35 40 45 50 55 17.9 23.5 2.1 2.0 4.0 7.5 0 40 19.2 13.9 7.8 4.4 0 0.9 16.4 11.1 4.0 3.2 7.2 5.2 5.5 9.9 43.4 31.8 6.7 23.0 3.1 2.1 At 30°C, all eggs died at or before gastrulation at 50%o; at 40°/oo, a small proportion of the eggs survived the high early mortality but most of these failed to hatch, only 4-6% hatching success- fully in Series B (Table 7). At 20 and 30%o at 30°C, most embryonic mortality occurred after the em- Table 7.— Percentage total and viable hatch of fertiHzed eggs in various combinations of temperature and salinity in Series A and B. In each series there were two repUcate groups of eggs (a and b) at each treatment combination. Temper- ature (°C) 18 21 24 27 30 Salin- ity (O/oo) 10 20 30 40 50 15 25 35 45 55 10 20 30 40 50 15 25 35 45 55 10 20 30 40 50 Percentage total hatch Percentage viable hatch Series A a b Series B a b Series A a b Series B a b 0 0 0 0 0 66.7 25.0 27.6 1.9 0 0 62.6 34.3 24.2 8.0 67.7 22.6 40.9 11.9 0 0 12.0 1.0 0 0 0 0 0 0 0 82.5 40.4 37.0 7.1 0 0 39.8 53.5 51.4 9.9 66.0 34.1 41.5 4.5 0 0 14.4 1.0 0.9 0 0 0 0 0 0 88.4 94.7 72.7 72.8 0 0 96.0 75.6 66.3 66.3 76.2 79.8 784 42.9 1.0 0 85.9 62.2 4.1 0 0 0 0 0 0 76.5 92.9 68.8 47.5 0 0 96.0 89.0 80.0 41.2 89.9 88.4 79.2 31.9 2.0 0 73.8 60.0 6.2 0 0 0 0 0 0 2.0 20.0 16.3 0 0 0 33.3 28.3 10.1 1.3 7.1 22.6 30.7 0 0 0 7.0 0 0 0 0 0 0 0 0 7.5 29.3 35.0 2.0 0 0 23.5 44.6 38.3 2.8 4.1 30.6 29.8 1.5 0 0 3.1 0 0 0 0 0 0 0 0 77.9 77.9 50.6 18.1 0 0 70.0 63.0 34.6 0 68.2 70.8 57.5 0 0 0 38.2 15.2 0 0 0 0 0 0 0 68.4 59.7 38.8 11.9 0 0 80.2 70.2 53.8 0 81.1 81.0 54.7 0 0 0 28.6 14.6 0 0 bryos had developed pigmentation, although at 30%o abnormal development was apparent in many eggs during cleavage and gastrulation 10 MAY: EFFECTS ON BAIRDIELLA ICISTIA v-jj Figure 7. — Developmental abnormalities. A) stage lib, at 18°C-30''/oo , showing unusual clover- leaf appearance of blastomeres; B) stage He, 18°C-30"/oo, showing enlarged periblast and clumped cytoplasm; C) stage He, 18°C-30%o, showing clumped cytoplasm; D) stage He, 30°C-30''/oo, showing irregular cleavage pattern; E) stage Ilia, 30°C-30°/oo, showing abnormal germ ring and clumping of cytoplasm; F) deformed embryo unable to free itself completely from the chorion, 24°C-20<'/oo. (Figure 7), and some eggs showed clumping of the cytoplasm similar to that observed at 18°C. Lar- vae hatching at 30°C were inactive. By following the development of individual eggs in the sup- plementary containers, it was noted that, no mat- ter what the temperature or salinity, irregularly cleaving eggs usually died before completing gas- trulation, and none ever hatched. At 21°, 24°, and 27°C, hatching was generally poorer at the higher salinities (Table 7). At 55%o virtually no hatching 11 FISHERY BULLETIN: VOL. 73. NO. 1 took place. A maximum of 96^^ hatching of fer- tihzed eggs was observed at 24°C-20%o. Deformed Larvae Immediately after hatching, larvae often had curved bodies reflecting the curvature necessi- tated by confinement within the chorion (Figure 8), but such larvae soon straightened out. Some larvae, however, had sharply bent or kinked notochords at hatching, a deformity which was irreversible and which prevented normal swim- ming. These deformities were most common at high salinities (40%o and above) and at 30°C. In Series A, salinities of 15 and 20%o produced a high proportion of larvae with a strange deforma- tion, in which the tail was recurved and fused to Figure 8. — Newly hatched larvae. A) ventral view of a normal larva, showing curvature often seen just after hatching, 24°C-30%o; B) lateral view of a larva with a recurved tail, 24°C-20»/oo, Series A. the trunk (Figure 8). Up to 789^ of the larvae hatching at 27°C-15%o showed this irreversible deformity in Series A, but the figure was only about 15% at 21°C-15%o and less than 10% at 24°C-20%o; with one or two minor exceptions, other treatments in Series A did not produce this particular distortion, and it was not observed in any treatment in Series B. A greater proportion of late-hatching larvae in a given treatment dis- played deformities than early-hatching larvae. Larvae hatching at 15 and 20"/oo showed pro- nounced edema (Figure 9). Histological sections showed that the size of the subdermal space was inversely related to sahnity (Figure 10), an os- motic phenomenon which Battle (1929) also ob- served in larvae of Enchelyopus cimbrius. The yolk sac of newly hatched bairdiella larvae was larger and contained more water at lower salinities (May 1972). Survival of Starved Larvae Besides showing deformities, at high tempera- tures and salinities many larvae died before ex- hausting their yolk supplies. At 45 and 50%o all larvae in Series A were dead within 1 day after hatching, and the same was true of the few hatched larvae at 40%o at 30°C (Figure 11). The time of major mortality and the maximum surviv- al time of starved larvae were inversely propor- tional to temperature and salinity. Because some of the larvae from Series B were used in tests of temperature and salinity tolerance (May 1972), a complete set of survival curves is not available for them. However, estimates of the percentage of larvae surviving to yolk absorption were obtained from the remaining larvae and from larvae in the least stressful conditions in the tolerance experi- ments, and these estimates indicated better larval survival in Series B than in Series A. For example, the Series B curves for 27°C (Figure 12) did not show the high mortality before yolk exhaustion at 25 and 35%o seen in Series A, and a similar differ- ence between the two series occurred at 21°C. At 24°C-40%o, an estimated 70% of the larvae were alive at yolk exhaustion in Series B, compared with only about 20% in Series A. At the highest temperatures and salinities, however. Series B showed heavy early mortality similar to Series A. Viable Hatch The percentage hatching of viable larvae (Table 12 MAY: EFFECTS ON BAIRDIELLA ICISTIA B Figure 9. — Two-day-old larva, 24°C-20%o, with enlarged subdermal space. A) side view, B) dorsal view. 7), calculated from the preceding information, may be considered the ultimate criterion of suc- cessful development in these experiments. Viable larvae are defined here as morphologically normal larvae capable of surviving to yolk absorption, since all other larvae would not survive in nature. Series A showed a much lower viable hatch than Series B at very high and very low temperatures and salinities. Even for the best eggs, it is clear that salinities above 40%o are detrimental to early survival, and that 30°C is extremely stress- ful. Survival at higher salinities was considerably better at low temperatures. The various observa- tions on embryonic and larval survival in Series B are summarized (Figure 13) in the manner of Al- derdice and Forrester (1967). Response Surfaces It has become customary to describe a biological 13 FISHERY BULLETIN: VOL. 73, NO. 1 B I D 14 MAY: EFFECTS ON BAIRDIELLA ICISTIA Figure 10. — Transverse sections of newly hatched larvae incu- bated in various salinities at 25°C. Serial sections were made of each larva, and the sections illustrated were located two sections posterior to the anus. A) 20%o, B) 330/00, C) 450/00, D) 50%o. 100 3 4 5 6 7 AGE (days) 12 3 4 5 AGE (days) Figure 11. — Survival curves for unfed larvae at various temperatures and salinities in Series A. There were two replicate groups of larvae at each treatment, and the vertical dashed lines indicate the time of complete yolk absorption at each temperature. 100 0 12 3 4 AGE (days) Figure 12. — Survival curves for unfed larvae in various salinities at 27°C, Series B. Vertical dashed line indicates the time of complete yolk absorption. response to temperature and salinity by fitting a second order polynomial to the data and present- ing response surfaces calculated from this equa- tion (e.g., Costlow et al. 1960; Alderdice and For- rester 1967; Haefner 1969). This procedure was applied by computer to the results for fertilization, total hatch, and viable hatch, and the resulting equations are given in Table 8. Analysis of vari- ance (ANOVA) showed that, although regression accounted for most of the variance in these data, deviations from regression were highly significant for all equations. This probably reflects the difficulty of fitting a second order polynomial to data of this sort, especially when abrupt thresholds are present, as between 10 and 15%o and 18° and 21°C. A higher order polynomial, or a nonlinear model (Lindsey et al. 1970), would no doubt yield a better fit. Nonetheless, the second 15 Ui (E < UJ a. Z UJ •E fcrliliiolion .[0] .[ol .0 \ high proportion v '*^^ rtigh morTolity ^^-^ .^ ^^ol beni rwlochordt \ \ during ernttryon.c V --^j^,^ "^- — X V;^ -,^ N^evelopmen) \_mortol.ly "" — -.. ^^"v \ y prior 10 _>, '^^ \ . , gostruloiion •0 + .s •^ \ \ -[0] •@ Imortolify r- — ^_ Iduring jolh- ^. ~ ^soc sioge \ ' ^ ^nmnlmtm mAftnlil^ •{Oj complete mortality prior to goslrulation .[0] •0 .[0] 10 20 30 40 SALINITY (%o) 50 Figure 13. — Summary of the effects of temperature and salinity on early development of bairdiella. Closed circles identify treat- ment combinations utilized in the experiments, and the numbers in squares beside them give the mean values for viable hatch in Series B. The cross marks the estimated position of maximum viable hatch. FISHERY BULLETIN: VOL. 73, NO. 1 significance of interaction by existing statistical techniques. Acclimation of Spawning Fish to Low Salinity On 20 October 1971 it was discovered that only 4 of the 26 fish acclimated to 15%o seawater were females, whereas 15 of the 26 fish at 33%o were females. The random assignment offish to the two tanks had somehow resulted in a great disparity in their sex ratios. Two of the four female fish from 15%o biopsied on 20 October 1971 had well- developed ovaries, showing that gonadal matura- tion can take place in a salinity of 15%o. The two well-developed females, as well as one of the poorly developed ones, were spawned with hor- mone injections; the oocyte size-frequency dis- tributions from biopsies of the three fish shortly before injection are shown in Figure 14. Table 8. — Multiple regression equations for percentage fertilization, total hatch, and viable hatch, as functions of temperature and salinity. Y = arcsin (percentage) '^,X = temperature (C),X2 = salinity C/oo). Series A; Series B: y = Y = -3.89030 + 0.25964X, -2.76156 + 0.14033X, Fertilization + 0.10770X2 - 0.00476X,2 - + 0.12125X2 - 0.00240X,2 - - 0.001 25X2^ - - 0.00149X2^ - 0.001 28X,X2 0.00055X,X2 Series A: Series B: Y = Y = -8.55800 + 0.72293X, -13.40397 + 1.06566X, Total hatch + 0.04177X2 - 0.01482X,2 - + 0.10817X2 - 0.021 15X, 2 - - 0.00073X2^ - - 0.001 57X2^ - - 0.0001 4X,X2 - 0.00070X,X2 Series A; Series B: Y = Y = -3.91134 + 0.31755X, -9.99277 + 0.81039X, Viable hatch + 0.02932X2 - 0.00645X,2 + 0.07829X2 - 001620X,2 - 0.00046X2^ - 0.001 17X2^ - 0.00017X,X2 - 0.00066X,X2 order equations are useful in that they allow com- putation of optimal conditions (Box 1956). The resulting values (Table 9) show a thermal op- timum at about 24°C for total and viable hatch in both series, and optima of 23° and 25°C for fertili- zation in Series A and Series B, respectively. The calculated salinity optimum for fertilization was considerably higher in Series B than in Series A (36 vs. 31%o), but in both series the optimal salinities for hatching were below those for fertili- zation, ranging from 26 to 29%o. The optimal re- sponses estimated at these points from the equa- tions (Table 9) are below the maximal values ac- tually recorded (cf Tables 2 and 7), another indi- cation of the lack of fit of the second order polyno- mial. The calculated positions of the optima, how- ever, are the best available estimates of the true optima. These experiments were designed primar- ily to cover wide ranges of the two factors under consideration, and the arrangement of treatments unfortunately does not allow testing of the Table 9. — Optimum temperatures and salinities for fertiliza- tion, total hatch, and viable hatch, estimated from the regression equations (Table 8). Also listed are the optimum percentage fertilization, total hatch, and viable hatch, calculated from the regression equations at the estimated temperature and salinity optima. Item Temperature CO Salinity {°/oo) Percentage Fertilization: Series A Series B 23.1 25.1 31.3 36.1 50.3 85.8 Total hatch: Series A Series B 24.3 24.7 26.3 28.9 47.6 94.3 Viable hatch: Series A Series B 24.3 24.5 27.4 26.6 11.2 67.3 The freezing point depression of blood serum from fish acclimated to 15%o, determined by the melting point method of Gross (1954), was 0.64 ± 0.066°C (mean± SD, n = 12 fish), and that offish from 33«/oo was 0.63 ± 0.076°C (n = 12 fish). The two groups did not differ significantly, nor was 16 MAY: EFFECTS ON BAIRDIELLA ICISTIA 30 ^ Fish I 20 - O 10 - Mil J^ J i t .£23 175 210 245 280 315 350 385 420 455 490 50 iS40 O 30 2 UJ O20|- UJ a: u- 10 - Fish n JZZL I 75 210 245 280 315 350 385 420 455 490 sS30r o20h z LiJ 10 o llJ q: Fish HI 1^1 i^i^iiiii^ ill I 75 210 245 280 315 350 385 420 455 490 OOCYTE DIAMETER (;jm) Figure 14. — Oocyte size-frequency distributions, based on ovarian biopsies, from fish acclimated to IS^/oo. Table 10. — Fertilization success for eggs obtained from fish ac- climated to 15%o. Salinity (°/oo) Percentage fertilization 10 15 20 30 40 45 50 FishI Fish II Fish III 0 0 0 19.7 73.8 42.2 24.2 89.8 53.4 2.5 31.1 88.5 0.9 18.7 79.2 0 21.2 85.3 0 20.4 63.3 Table 11. — Percentage total and viable hatch of fertilized eggs at various salinities for eggs from fish acclimated to 15''/oo. For each fish, there were two replicate groups of eggs at each salin- ity. Salinity Fish I Fish I Fish I . (°/oo) a b' a b a b Total hatch ■ 15 23.4 22.0 87.1 79.8 97.0 92.6 20 30.5 44.2 86.7 86.3 97.0 95.7 30 10.9 2.0 46.9 66.0 97.9 94.9 40 — — 80.0 80.9 84.8 68.8 45 — — 71.2 71.4 65.7 76.3 50 — — 35.9 31.0 22.8 45.8 Viable hatch 15 17.0 20.0 66.3 50.5 79.2 72.6 20 27.1 26.9 62.7 70.3 82.7 76.5 30 8.7 0 37.5 49.2 84.3 79.9 40 — — 43.0 56.1 55.3 44.7 45 — — 18.1 12.6 19.4 22.9 50 — — 0 0 0 0 there a significant difference between sexes within each group (Mann- Whitney U test; Siegel 1956). The fish with poorly developed ovaries (Fish I) became listless and swam in a disoriented manner after the hormone injection; 5 days after spawn- ing, it still spent most of its time resting on its side on the bottom of the tank. At this point the fish was sacrificed and dissected, revealing some large, hydrated eggs with coalesced yolk, 665-735 A^m in diameter, along with many unhydrated eggs still in their follicles, measuring 350 jum in diameter. Eggs obtained from the hormone-induced spawn- ing of this fish showed low fertility, significant numbers being fertilized only at 15 and 20%o, with a maximum of 24.2% fertilized at 20"/oo (Table 10). The hatching success of fertilized eggs was also poor, with a maximum total hatch of 44.2% at 20%o (Table 11). A few embryos and larvae produced by this fish displayed various de- grees of cyclopia, a deformity rarely seen in other batches of eggs. As expected, the two ripe fish produced much better eggs, with maximum fertilization percent- ages of almost 90% (Table 10). Eggs from Fish II had a lower optimum salinity than those from Fish III and were more sensitive to high salinities (Table 10). No fertilization took place at 10%o, as was the case with eggs from fish living at 33%o. Hatching at 15, 20, and 30%o was better in eggs from Fish III than in those from Fish II, despite the better fertilization success of the latter at 15 and 200/00 (Table 11). Hatching at 40, 45, and 50%o was comparable in the two batches of eggs. Eggs from Fish II hatched more successfully at 15 and 20%o than at 30%o, whereas those from Fish III hatched equally well at 15, 20, and 30%o. The incidence of postmature unhatched eggs was simi- lar to that in eggs from Series A and Series B, with most appearing at 15 and 20%o, few at 30%o, and very few or none above 30%o. The hatching success at various salinities (20, 30, 40, and 50%o) of the best batch of eggs from fish living at 15%o (i.e., from Fish III) was com- pared with that of the best batch of eggs from fish at 33%o (Series B) at the same temperature (24°C) and salinities, by ANOVA (an arcsin-square root transformation was applied to the percentages). 17 FISHERY BULLETIN: VOL. 73, NO. 1 Neither total nor viable hatching success differed significantly between the two groups. Therefore, acclimation of spawning fish to a low salinity did not affect the salinity tolerance of the eggs in any detectable way. Effects of acclimation salinity on egg size and buoyancy will be discussed elsewhere (May in preparation). DISCUSSION Fertilization and early development in Bair- diella icistia are stenothermal and stenohaline processes. The approximate limits for successful development, from fertilization to yolk exhaus- tion, are 20° to 28°C and 15 to 40%o, although a certain interaction of the two factors is apparent, development being more successful at the higher salinities when the temperature is relatively low, and at the higher temperatures when the salinity is relatively low. The limits within which success- ful reproduction can take place are defined by the most sensitive stages and events in development. The lower limit of salinity for bairdiella reproduc- tion is defined by fertilization, since eggs cannot be fertilized at 10%o or below, even though eggs fertilized at a higher salinity will develop at 10%o . However, the lowest salinity at which eggs remain buoyant may in some cases determine the lower salinity threshold for successful reproduction (May 1972). The upper salinity limit, and both the upper and the lower limits of temperature, are defined by the abilities of the embryos to develop. Fertilization is successful at 18°C but develop- ment is not; likewise, fertilization does take place at 30°C, and at salinities of 45%o and above, but the hatching of viable larvae is greatly curtailed. Fertilization in bairdiella is more limited by salinity than temperature over the ranges studied. The complete block to fertilization which occurs at 10%o may be related to an inability of spermatozoa to function properly at this salinity. Although the egg itself seems to be unharmed by water of IC/oo, at this salinity spermatozoa never attain the high intensity of activity that they do at higher salinities. At 15%o, where spermatozoan activity is more intense than at 10%o but less intense than at higher salinities, fertilization oc- curs but is poorer than at higher salinities. Fairly high salinities seem to aid fertilization: the calcu- lated optimum salinities for fertilization were higher than the optima for hatching in both Series A and Series B. It is possible that low calcium levels at low salinities inhibit the activity of 18 spermatozoa (Yanagimachi and Kanoh 1953). In general, the greater the intensity of spermatozoan activity, the shorter is the overall duration of ac- tivity (Figure 3). Thus a shortlived but extremely high level of spermatozoan activity may be neces- sary for fertilization in bairdiella, perhaps be- cause penetration of the micropyle requires a con- siderable expenditure of energy on the part of the spermatozoa. This implies that the actual process of fertilization takes place during the first few seconds after hydration of the sperm, when sper- matozoan activity is maximal. Haydock ( 1971) re- ports that bairdiella spermatozoa are no longer able to fertilize eggs 30 s after sperm hydration. In such a situation, experimental technique could have a marked influence on the success of artificial fertilization, since a delay of a few seconds be- tween hydration of the sperm and contact of the sperm with eggs could significantly reduce the percentage of eggs which become fertilized. A technical problem of this sort may explain the puzzling differences in fertilization success be- tween eggs from Fishes II and III, acclimated to 150/00 (Table 10). Several previous investigations of salinity ef- fects on spermatozoan activity in other fishes pro- vide interesting contrasts with the present results. Ellis and Jones (1939) found that sper- matozoa of Atlantic salmon, Salmo salar, a fish which spawns in fresh water, were active for over 180 min in seawater diluted to 15 and 20% and that the duration of activity dropped off sharply above and below these salinities. Working with the longjaw mudsucker, Gillichthys mirabilis, Weisel (1948) observed that spermatozoa showed only feeble activity in seawater diluted to 17-24%, but activity was intense in 25% seawater and above; the duration of spermatozoan activity was maximal (over 50 h!) in 25% seawater and de- creased at higher salinities, as it did in the case of bairdiella. Yamamoto (1951) found that sper- matozoa of the flounder, Limanda schrenki, were active in normal seawater and in seawater diluted to 50%, but showed no activity (and no fertilizing capability) in 25% seawater. Hines and Yashouv (1971), on the contrary, found that mullet, Mugil capita, spermatozoa exhibited a gradual increase in duration of activity with increasing salinity up to the salinity of normal seawater, rather than a threshold. Dushkina (1973) reported that sper- matozoa of Pacific herring, Clupea harengus pallasi, were most active at higher salinities (17-23"/oo), but remained active longest at the MAY: EFFECTS ON BAIRDIELLA ICISTIA lowest salinities (0.3-0.5%o), as was true for bair- diella; however, in herring the duration of sper- matozoan activity was much longer (4-8 days at 6-7°C) and some fertilization occurred even in fresh water. Spermatozoan activity and its re- sponse to salinity appear to be extremely variable among fish species, which is hardly surprising in view of the diversity of habitats and modes of reproduction of fishes. The large proportion of postmature unhatched eggs at low salinities (Table 6) reflects a high incidence of malformations under these condi- tions, the embryos being physically unable to break from the chorion. Edema seen among larvae in low salinities suggests that deformities and the inability to hatch may be related to osmotic prob- lems. Battle (1929) noted a similar difficulty in hatching among embryos of fourbeard rockling, Enchelyopus cimbrius, in low salinities and attrib- uted it to abnormally developed musculature, which prevented movements required to free the embryo from the egg case. An inability to complete hatching at low salinities has been reported for other species as well (Ford 1929; McMynn and Hoar 1953; Alderdice and Forrester 1967; Dush- kina 1973). The generalization that gastrulation and hatching are the two developmental stages most sensitive to physical disturbance (e.g., Holli- day 1969) seems valid in the case of bairdiella. The finding that unfed bairdiella survive longest in low salinities and low temperatures is not unique. Nakai ( 1962) and Hempel and Blaxter (1963) likewise found that starving larvae of Sardinops melanosticta and Clupea harengus survived longer at lower salinities, and more rapid mortality among unfed larvae at higher tempera- tures has been observed on a number of occasions (e.g., Qasim 1959; Bishai 1960; Hempel and Blax- ter 1963; Alderdice and Velsen 1971; Hamai et al. 1971). High temperatures increase metabolic rate, accelerate yolk absorption (May 1972), and no doubt hasten death from starvation. The effect of high salinities on larval physiology is less cer- tain: a salinity effect on embryonic or larval ox- ygen consumption has not been demonstrated ex- cept after abrupt transfer (HoUiday 1969), and salinity has only a small effect on the rate of yolk absorption in bairdiella (May 1972). High salinities may increase larval mortality by caus- ing osmotic or ionic changes in the interior milieu, although the larvae of some species have proved capable of osmoregulating over rather wide ranges of salinities (Holliday 1969). Lower levels of activity have been observed among larvae of some species in low salinities (Hempel and Blax- ter 1963; Holliday 1965), and may reduce their metabolic demand and thus extend their survival time (Holliday 1965). The salinity tolerance of bairdiella eggs is not significantly affected by acclimation of the parent fish to low salinity (15%o). This might suggest that the enhanced survival at low salinities which Solemdal ( 1967) observed in eggs from the Finnish population of flounder, Pleuronectes flesus, has a genetic basis. If acclimation of spawning fish to low salinities does not cause an increase in em- bryonic tolerance to low salinities, one might ex- pect that high-salinity acclimation would be simi- larly ineffectual in aiding embryonic survival at high salinities. This supposition should be verified experimentally; but, if valid, it implies that salin- ity responses determined on eggs from fish living in ordinary seawater should be accurate predic- tors of reactions to different salinities in nature, except where genetic adaptation has occurred. This could be a significant advantage in cases where it is important to estimate the effects of rising salinities in specified habitats, such as the Salton Sea or the Gulf of California, where high salinities may in the future pose a threat to exist- ing stocks of fish. Because of the unusual chemical nature of the Salton Sea, it is impossible to estimate the salinity tolerance of bairdiella eggs in Salton Sea water from the present data concerning their responses in ordinary seawater. There is evidence that the ionic composition of Salton Sea water has a del- eterious effect on the survival of eggs and larvae (Lasker et al. 1972; May 1972), so that the upper salinity limits defined in the present study are probably higher than those which hold for bair- diella in the Salton Sea. The spawning season of bairdiella occurs dur- ing a period of rapidly rising temperatures. In the Salton Sea this species spawns mainly in April and May, with a peak of spawning probably in mid-May (Whitney 1961; Haydock 1971). Max- imum surface temperatures in the Salton Sea are plotted in Figure 15, where the spawning time of bairdiella is also shown. It is clear that some bair- diella may spawn in water of 30°C or higher, al- though most spawning is probably finished before temperatures reach this level. Whitney (1961) re- ports finding bairdiella eggs in 1955 as late as 1 August, which means they could have been ex- posed to the undoubtedly lethal temperature of 19 FISHERY BULLETIN: VOL. 73, NO. 1 ASONDJ FMAMJ J ASONDJ FN. AMJ J 1954 1955 1956 MONTH a YEAR Figure 15. — Maximum surface temperatures in the Salton Sea. Open circles: measurements made at Sandy Beach, Salton Sea, from August 1954 to July 1956 (after Carpelan 1961). Closed circles: measurements made at various stations on the Salton Sea during 1967 (after Young 1970). The shaded areas indicate the major spawning period oi Bairdiella icistia, and the vertical dotted lines indicate the latest records of bairdiella eggs in the plankton in 1955 and 1956, according to Whitney (1961). 35°C. Such late spawning by bairdiella seems un- likely, however, and Whitney may have collected eggs of the orangemouth corvina, Cynoscion xanthulus, which spawns during the summer and probably produces similar eggs, rather than bair- diella. In any event, it seems possible that late spawning bairdiella in the Salton Sea could re- lease their eggs in water with a temperature high enough to reduce embryonic and larval survival severely. In view of the temperature-salinity in- teraction which occurs in the case of both em- bryonic and larval tolerance, bairdiella which spawn at relatively low temperatures early in the season will probably have a selective advantage as the salinity of the Salton Sea rises. In the absence of detailed information on the distribution of bairdiella and the physical condi- tions obtaining in its native habitat, the Gulf of California, it is difficult to apply the present findings to the ecology of this species in that area. However, the utilization of Colorado River water for irrigation has caused an increase in the river's salinity (Wolman 1971); if this, and the accom- panying reduced flow of fresh water into the upper Gulf of California, results in a significant rise in salinity in areas where bairdiella spawn, the com- bined action of salinity stress and heat in this arid region could adversely affect early survival in the local bairdiella population. The warm brine effluent from a proposed desalination plant in this area (Thomson et al. 1969) could aggravate the situation considerably if dispersal of the effluent is not adequate. ACKNOWLEDGMENTS I take great pleasure in thanking Reuben Lasker, Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, for advice and encouragement throughout this study and for providing equipment and laboratory space and making available the excel- lent aquarium facilities, without which this work would have been impossible. Irwin Haydock and David Crear introduced me to techniques for maturing and spawning bairdiella in captivity, and Robert G. Hulquist of the California Depart- ment of Fish and Game facilitated collecting ef- forts at the Salton Sea. James R. Zweifel, South- west Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, helped in the statistical treatment of some of the data, and Dale Mann drafted the figures. Finan- cial support was provided by the University of California Institute of Marine Resources. LITERATURE CITED Ahlstrom, E. H. 1943. Studies on the Pacific pilchard or sardine, (Sardinops caerulea). 4. Influence of temperature on the rate of development of pilchard eggs in nature. U.S. Fish Wildl. Serv., Spec. Sci. Rep. 23, 26 p. AXDERDICE, D. F., AND C. R. FORRESTER. 1967. Some effects of saUnity and temperature on early development and survival of the English sole (Parophrys vetulus). J. Fish. Res. Board Can. 25:495-521. 1971a. Effects of salinity and temperature on embryonic development of the petrale sole (Eopsettajordani). J. Fish. Res. Board Can. 28:727-744. 1971b. Effects of salinity, temperature, and dissolved ox- ygen on early development of the Pacific cod {Gadus macrocephalus). J. Fish. Res. Board Can. 28:883-902. Alderdice, D. F., and F. P. J. Velsen. 1971. Some effects of salinity and temperature on early development of Pacific herring (Clupea pallasi). J. Fish. Res. Board Can. 28:1545-1562. Battle, H. L 1929. Effects of extreme temperatures and salinities on the development of Enchelyopus cimbrius (L.). Contrib. Can. Biol. Fish., New Ser., 5:109-192. BisHAi, H. M. 1960. Upper lethal temperatures for larval salmonids. J. Cons. 25:129-133. Box, G. E. P. 1956. The determination of optimum conditions. In O. L. 1 20 MAY: EFFECTS ON BAIRDIELLA ICISTIA Davies (editor). The design and analysis of industrial ex- periments, p. 495-578. Oliver and Boyd, Lond. Carpelan, L. H. 1961. Physical and chemical characteristics. In B. W. Walker (editor), The ecology of the Salton Sea, California, in relation to the sportfishery, p. 17-32. Calif. Dep. Fish Game, Fish Bull. 113. CosTLOw, J. D., Jr., C. G. Bookhout, and R. Monroe. 1960. The effect of salinity and temperature on the larval development of Sesarma cinereum (Rose) reared in the laboratory. Biol. Bull. (Woods Hole) 118:183-202. DUSHKINA, L. A. 1973. Influence of salinity on eggs, sperm and larvae of low-vertebral herring reproducing in the coastal waters of the Soviet Union. Mar. Biol. (Berl.) 19:210-223. Ellis, W. G., and J. W. Jones. 1939. The activity of the spermatozoa of Salmo salar in relation to osmotic pressure. J. Exp. Biol. 16:530-534. Ford, E. 1929. Herring investigations at Plymouth. VII. On the artificial fertilisation and hatching of herring eggs under known conditions of salinity, with some observations on the specific gravity of the larvae. J. Mar. Biol. Assoc. U. K., New Ser., 16:43-48. Gross, W. J. 1954. Osmotic responses in the sipuncuhd Dendrostomum zostericolum. J. Exp. Biol. 31:402-423. Haefner, p. a. 1969. Temperature and salinity tolerance of the sand shrimp, Crangon septemspinosa Say. Physiol. Zool. 42:388-397. Hamai, I., K. KyOshin, and T. Kinoshita. 1971. Effect of temperature on the body form and mortal- ity in the developmental and early larval stages of the Alaska pollack, Theragra chalcogramma (Pallas). Bull. Fac. Fish. Hokkaido Univ. 22:11-29. Harrington, R. W., Jr. 1959. Photoperiodism in fishes in relation to the annual sexual cycle. In R. B. Withrow (editor), Photoperiodism and related phenomena in plants and animals, p. 651-667. Am. Assoc. Adv. Sci., Publ. 55, Wash., D.C. Haydock, I. 1971. Gonad maturation and hormone-induced spawning of the gulf croaker, Bairdiella icistia. Fish. Bull., U.S. 69:157-180. Hempel, G., and J. H. S. Blaxter. 1963. On the condition of herring larvae. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 154:35-40. HiNES, R., and a. Yashouv. 1971. Some environmental factors influencing the activity of spermatozoa oiMugil capita Cuvier, a grey mullet. J. Fish Biol. 3:123-127. Holliday, F. G. T. 1965. Osmoregulation in marine teleost eggs and larvae. Calif Coop. Oceanic Fish. Invest., Rep. 10:89-95. 1969. The effects of salinity on the eggs and larvae of teleosts. In W. S. Hoar and D. J. Randall (editors). Fish physiology. Vol. 1. Excretion, ionic regulation, and metabolism, p. 293-311. Academic Press, N.Y. HUBBS, C, AND L. StaVENHAGEN. 1958. Effects of maternal carotenoid deficiency on the via- bility of darter (Osteichthyes) offspring. Physiol. Zool. 31:280-283. Johnston, R. 1964. Recent advances in the estimation of salinity. Oceanogr. Mar. Biol. Annu. Rev. 2:97-120. Kelly, W. H. 1967. Marking freshwater and a marine fish by injected dyes. Trans. Am. Fish. Soc. 96:163-175. Kinne, O., and E. M. Kinne. 1962. Rates of development in embryos of a cyprinodont fish exposed to different temperature-salinity-oxygen combinations. Can. J. Zool. 40:231-253. KUNTZ, A. 1915. The embryology and larval development of Bairdiella chrysura and Anchovia mitchilli. Bull. U.S. Bur. Fish. 33:3-19. Lasker, R., AND L. L. Vlymen. 1969. Experimental sea-water aquarium. U.S. Fish Wildl. Serv., Circ. 334, 14 p. Lasker, R., R. H. Tenaza, and L. L. Chamberlain. 1972. The response of Salton Sea fish eggs and larvae to salinity stress. Calif Fish Game 58:58-66. Lindsey, J. K., D. F. Alderdice, and L. V. Pienaar. 1970. Analysis of nonlinear models — the nonlinear re- sponse surface. J. Fish. Res. Board Can. 27:765-791. May, R. C. 1972. Effects of temjjerature and salinity on eggs and early larvae of the sciaenid fish, Bairdiella icistia (Jordan and Gilbert). Ph.D. Thesis, Univ. California, San Diego, 281 p. McMynn, R. G., and W. S. Hoar. 1953. Effects of salinity on the development of the Pacific herring. Can. J. Zool. 31:417-432. Nakai, Z. 1962. Studies relevant to mechanisms underlying the fluctuations in the catch of the Japanese sardine, Sardinops melanosticta (Temminck & Schlegel). Jap. J. Ichthyol. 9:1-113. Qasim, S. Z. 1959. Laboratory experiments on some factors Etffecting the survival of marine teleost larvae. J. Mar. Biol. Assoc. India 1:13-25. SCHALES, O., AND S. S. SCHALES. 1941. A simple and accurate method for the determination of chloride in biological fluids. J. Biol. Chem. 140:879-884. Sehgal, a., and B. I. Sundararaj. 1970. Effects of various photoperiodic regimens on the ovary of the catfish, Heteropneustes fossilis (Bloch) during the spawning and the postspawning periods. Biol. Reprod. 2:425-434. Siegel, S. 1956. Nonparametric statistics for the behavioral sci- ences. McGraw-Hill, N.Y., 312 p. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Co., San Franc, 776 p. SOLEMDAL, P. 1967. The effect of salinity on buoyancy, size and develop- ment of flounder eggs. Sarsia 29:431-442. Stevens, R. E. 1966. Hormone-induced spawning of striped bass for res- ervoir stocking. Prog. Fish-Cult. 28:19-28. Thomson, D. A., A. R. Mead, and J. R. Schreiber. 1969. Environmental impact of brine effluents on Gulf of 21 California. U.S. Dep. Int., Off. Saline Water, Res. Dev. Prog. Rep. 387, 196 p. U.S. Department of the Interior and the Resources Agency OF California. 1969. Salton Sea project, California. Fed.-State Recon- naissance Report, October 1969, 160 p. Walker, B. W., R. R. Whitney, and L. H. Carpelan. 1961. General considerations and recommendations. /n B. W. Walker (editor), The ecology of the Salton Sea, California, in relation to the sportfishery, p. 185-192. Calif. Dep. Fish Game, Fish Bull. 113. Weisel, G. F., Jr. 1948. Relation of salinity to the activity of the sper- matozoa of Gillichthys, a marine teleost. Physiol. Zool. 21:40-48. Whitney, R. R. 1961. The bairdiella, Bairdiella icistius (Jordan and Gil- bert). In B. W. Walker (editor). The ecology of the Salton Sea, California, in relation to the sportfishery, p. 105-151. Calif. Dep. Fish Game, Fish Bull. 113. FISHERY BULLETIN; VOL. 73, NO. 1 Wilkie, D. W., and H. Gordin. 1969. Outbreak of cryptocaryoniasis in marine aquaria at Scripps Institution of Oceanography. Calif. Fish Game 55:227-236. WOLMAN, M. G. 1971. The nation's rivers. Science (Wash., D.C.) 174:905-918. Yamamoto, K. 1951. Studies on the fertilization of the egg of the flounder. 1. Effects of salt concentration in the fertilization. J. Fac. Sci. Hokkaido Univ., Ser. 6, 10:253-259. Yanagimachi, R., and Y. Kanoh. 1953. Manner of sperm entry in herring egg, with special reference to the role of calcium ions in fertilization. J. Fac. Sci. Hokkaido Univ. 11:487-494. Young, D. R. 1970. The distribution of cesium, rubidium, and poteis- sium in the quasi-marine ecosystem of the Salton Sea. Ph.D. Thesis, Univ. California, San Diego, 213 p. ♦ 22 FITTING THE GENERALIZED STOCK PRODUCTION MODEL BY LEAST-SQUARES AND EQUILIBRIUM APPROXIMATION ^ William W. Fox, Jr.^ ABSTRACT A least-squares method for fitting the generalized stock production to fishery catch and fishing effort data which utilizes the equilibrium approximation approach is described. A weighting procedure for providing improved estimates of equilibrium fishing effort and an estimator of the catchability coefficient are developed. A computer program PRODFIT for performing the calculations is presented. The utility and performance of PRODFIT is illustrated with data from a simulated pandalid shrimp population. The production model approach to fish stock as- sessment is simply an adaptation of the Lotka- Volterra population equations into the situation of a population exploited by man. The earliest such adaptation was by Graham (1935) in assess- ing the potential production from North Sea fish stocks. The major development of this approach in fisheries management, though, is due to Schaefer (1954, 1957) who initiated it as a management tool for the yellowfin tuna fishery of the eastern tropi- cal Pacific Ocean. While there has been an at- tempt at a detailed extention of the production model approach to multispecies fisheries (Lord 1971), the usual application has been on a single species stock. Mathematical formulation of the production model begins with the general differential equa- tion dPIdt = P,g (P,) - PMO (1) where P, is the population size at time ^ Ptg (Pf ) is the population production function encompassing the effects of reproduction and natural mortality (and growth in weight if biomass is the population unit), and h (/",) is the fishing mortality coefficient exerted by/", units of fishing effort. Fishing effort is assumed to be standardized from nominal fishing effort such that qf^ = F^ , where F, is the instantaneous coefficient of fishing mortality and <7 is a constant (the catchability coefficient), giving QftPf - dCldt, the rate of catch. At equilibrium, that is dPIdt = 0, the catch rate equals the produc- 'Adapted, in part, from a Ph.D. dissertation, College of Fisheries, University of Washington, Seattle, WA 98195. ^Southwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA, P.O. Box 271, La Jolla, CA 92037. tion rate such that an equilibrium yield, Y , is obtained Y = qfP = PgiP). (2) The most general assumptions about the form of PfgiPf) are that it should 1) approach zero as P, approaches some environmental capacity, P^ax' and 2) increase to some maximum at a population size smaller than the environmentally limited size. Practically, the function should be simple, since in any case the approach is a gross simplification of population dynamics. The most fiexible, simple function advanced for Ptg (P,) is a simple case of Bernoulli's equation (Chapman 1967; Pella and Tomlinson 1969) PtgiPt) =HPr-KP, (3) where H, K, and m are constant parameters.^ Equation (3) includes the logistic function when m = 2 (Schaefer 1954, 1957) and the Gompertz function [K'P^ - H'P.lnPJ as m^l (Fox 1970). Equation (3), hereafter referred to as the generalized stock production model after Pella and Tomlinson (1969), approaches zero at Pmax = {K/H) !'•"' 1' and has a maximum Popt = [m^'^'-'^n -P^ax- Three equilibrium relationships can be derived by the substitution of Equation (3) in Equation (2) to obtain 1) Yield and population size Y =HP"' - KP, 2) Population size and fishing effort (4) ^When formulated as in Equation (3), H and K are positive for m < 1, but are negative for m > 1. Manuscript accepted May 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. 23 FISHERY BULLETIN: VOL. 73. NO. 1 3) Yield and fishing effort (Kq q ^X"! 1 (5) (6) The critical points, useful as management impli- cations and previously derived by Pella and Tom- linson (1969), are: /■.p. = K{^ - l)/9 Popt = [KI{mH)] m - I (7) (8) and m - I m - I H[K/(mH)] - [K"'/imH)] , (9) where f^^^ is the amount of fishing effort required to produce Ymax, the maximum sustainable aver- age yield (MSAY),^ and Popt is the equilibrium population size obtained atf^^^. Figure 1 demon- strates the flexibility of the generalized stock pro- duction model with three values for m (0.5, 2.0, 4.0); each curve has the same value for Pmax and Y -* max • In utilizing the production model for analysis of the status of a particular population, the usual basic assumptions are that 1) the model is being applied to a closed single unit population, 2) the concept of equilibrium conditions^ applies to the population under analysis, and 3) the age-groups being fished have remained, and v^ll continue to remain, the same. If one is able to obtain data which represent equilibrium conditions at three or more population levels, then no additional as- sumptions are needed to fit the production model. In most fishery data sets, however, no real period of equilibrium conditions will exist. Using data from the transitional states of a population re- quires the additional assumptions that both 1) time lags in processes associated with population change and 2) deviations from the stable age *i^max is usually referred to as the maximum sustainable jaeld (MSY). The term MSY, however, does not convey that in reality the yield will fluctuate due to changes in the population even if the fishing effort and catchabillty coefficient remain constant. Hence, the "equilibrium yield" curve represents a curve of yield that is sustainable at some average level. *The definition of equilibrium adopted here, essentially that of Beverton and Holt (1957), is: 0ven a constant rate of fishing, including zero, a population will achieve a state where, on the average, it will not change in size or characteristics. structure at any population level have negligible effects on the production rate, Ptg (P)< (Schaefer and Beverton 1963). Schaefer (1954, 1957) pioneered the use of transitional state data for fitting a production model (the logistic form) to catch and fishing effort data. Schaefer's (1957) method for estimating the parameters consisted of approximating differen- tial equation (1) with two finite difference equa- tions and then iteratively solving them. Pella and Tomlinson (1969) greatly improved upon Schaefer's method by demonstrating that a catch history of a fishery could be predicted from the fishing effort history, initial estimates of the pro- duction model parameters, and the integrated form of Equation (1). Then final parameter esti- mates could be obtained by a pattern search rou- tine which finds those parameters which minimize the residual sum of squared differences between UJ POPULATION SIZE *'^*«- UJ >» •Cr-~. N z ""■"-"^^ o "H. ^'^Ss.^^^ 1- '^V*^*'**^,^ < _J 3 \ ^'^V;;:^^__m = 0.5 n. ^'^'^N..,,^^ o »m = 4 ^^^m = 2 a. 1 ^^^. FISHING EFFORT FISHING EFFORT Figure 1.— Equilibrium relationships of the generalized stock production model for three values of m. (A) Equilibrium yield and population size; (B) population size and fishing effort; (C) equilibrium yield and fishing effort. 24 FOX: FITTING THE GENERALIZED STOCK PRODUCTION MODEL the observed and predicted catches. While these two estimation methods are very different in their degree of sophistication, they are fundamentally the same in that both methods utilize the predic- tion of population transitional state changes by the production model. For convenience, this ap- proach will be subsequently referred to as the transition prediction approach. Gulland (1961) established a second approach to fitting production models with transitional state data. Gulland's approach estimates the level of fishing effort which, if equilibrium obtained, would produce, on the average, the observed level of catch per unit effort in each year of the fishery. Then the set of paired catches per unit effort and estimated equilibrium fishing effort units are fitted to one of the equilibrium relationships given by, or derived from, Equation (4), (5), or (6). This approach will be referred to subsequently as the equilibrium approximation approach. Clearly, the transition prediction and equilib- rium approximation approaches are basically dif- ferent. The transition prediction approach is obvi- ously intimately based upon the transition state population assumptions. On the other hand, the degree to which the equilibrium approximation approach is dependent on these assumptions is unclear. This paper presents a least-squares method and a computer program PRODFIT, which uses the equilibrium approximation ap- proach to estimate the parameters (and indices of their variability) of the generalized stock produc- tion model. A weighting procedure for providing improved estimates of equilibrium fishing effort and an estimator of the catchability coefficient are developed. The utility and performance of com- puter program PRODFIT is illustrated by fitting deterministic and stochastic data from a simu- lated pandalid shrimp population. Some cursory comparisons between the equilibrium approxima- tion and transition prediction approaches are made by repeating the pandalid shrimp simulated data fits with GENPROD, the computer program written by Pella and Tomlinson (1969). FITTING METHOD The equilibrium approximation approach was first outlined in Gulland (1961), but is more fully explained in Gulland (1969:120). Gulland's method involves relating the annual catch per unit effort in year i, Ui, to the fishing effort aver- aged over some number of years, T. Gulland (1961) first defined T as the mean life expectancy of an individual in the fishable population, orZ ~^, where Z is the instantaneous total mortality coefficient and the value of Z "^ is rounded off to the nearest integer. Subsequently, Gulland (1969) defined T as the average fishable duration of a year class (again to the nearest whole year) — he provided the following example: if recruitment is at 4 yr and if most of the catch in year / consists of 4 to 9 yr-old fish, then the average fishable duration is about 3 yr so U, would be related to an average off,, fi - 1 and/*! - 2. The general formulation for the averaged fishing effort in year i is l = h 2^ (7) J = I r + 1 A discussion of the rationale for, and performance of, Gulland's averaging method is given by Gul- land (1969:120). Weighted Average Fishing Effort Method In this paper a different tack is taken which results in approximating equilibrium fishing ef- fort with a weighted average. The catch per unit effort of the incoming year class J in year i, U,j, is related to the amount of effort in year i; that of the previous year class, C/, j _ i , is related to the fishing effort in years i and / - 1; that of the year class which entered 2 yr previously, [/, j - 2 . is related to the fishing effort in years i,i — 1, and i - 2; and so forth. The catch per unit effort of the total fishable population, assuming equal catchability, is + f/, k + 1 for k year classes. For the simplest case where the incoming year class is recruited at the beginning of each year's fishing season, therefore, [/, - {A: • f, + (^ - 1) • /; _ 1 + (8) Equation (8) suggests a weighted average of fishing effort over the total number of years that a year class contributes significantly to the fishery, or + f=[k ■ /•, +(^ -1)./;. [k + {k - I) + ■ ■ ■ + i] + f. k + (9) 25 FISHERY BULLETIN: VOL. 73, NO. 1 An arithmetic average rather than a geometric average is suggested because most appHcations are on catch in weight, i.e. while year classes de- cline exponentially in terms of numbers they con- comitantly increase in terms of mean weight per individual. The weighting procedure can be more precise if it is knowTi when during the year of record that recruitment occurs. For example, if recruitment occurs at midseason during the year of record for a fishable population of three year classes, f, changes from ( 3 /; + 2 /": _ ^ + /; _ 2 ) /6 to {2.5/; + 1.5 /; - 1 + 0.5/; 2! /4.5. Further pre- cision is gained if k is variea from year to year with the level of fishing effort, since at high fish- ing rates fewer year classes will contribute sig- nificantly to the catch than at low fishing rates. Further adjustments can be made for unequal catchability among the year classes. The unweighted method of averaging the fishing effort. Equation (7), and the new weighted method, Equation (9), will be compared in a sub- sequent section of this paper. Estimation Procedure critical points in terms of the parameters of Equa- tion (11) are /"opt = (a - a/n)/(m|3) C^opt = (aim) Vim - 1) (12) (13) and (a — am) ia/m) l/(m - 1) (14) Given the data set {u, , f,] , where i = 1. . .n observations, the least-squares criterion for es- timating the parameters a, /3, and m is to minimize the function •5 = 2 w^.(u, - U,)2 (15) i = 1 where the W, are statistical weights, and f7, are the predicted equilibrium catches per unit effort from Equation (11). The statistical weights. w^ = iuy\ (16) Gulland (pers. commun.)^ prefers an eye-fitted curve for estimating the equilibrium relationship between C/, and /", because of the over- simplification of the method and the errors as- sociated with usual catch and effort data. How- ever, these reasons should not defer the seeking of a more precise method of fitting a curve nor the taking advantage of error estimation schemes, if the simplifications and assumptions are kept in mind. On the contrary, it will be demonstrated that, at least for some controlled conditions, the equilibrium approximation approach provides reasonably good results. Equation (5) may be written in terms of catch per unit effort and averaged fishing effort as m - 1 U, = [iKq IH) + {q IH)f^ ] - .1/ (m - 1) or simply 1) (10) (11) Equation (11) is a nonlinear function with three parameters which does hot require simultaneous estimation of the catchability coefficient, q. The *John A. Gulland, Food and Agricultural Organization, Rome, Italy. are derived from the assumption of the multiplica- tive error structure as suggested by Fox (1971). Weighting as in Equation (16) will usually give the greatest weight to observations at the highest level of averaged fishing effort; in many cases these also will be the most recent observations. Giving greater weight to observations at high ef- fort levels will tend to give the greatest weight to observations vidth the greatest temporal and spa- tial coverage of the population. In addition, giving the greatest weight to the most recent data is especially advantageous when approaching the ^max level during a period increasing fishing effort because the observations nearest the Fmax level receive the greatest weight. Up to now no mention has been made on the estimation of the catchability coefficient, q. This is because experience with GENPROD and stochas- tic simulation studies have indicated that poor results are frequently obtained from the simul- taneous estimation of q (Pella and Tomlinson 1969; Fox 1971). Once that a, |8,andm have been estimated, q may be treated as a conditional prob- abilistic variable and estimated as a mean value. Two tacks were selected, the difference method and the integral method. 26 FOX: FITTING THE GENERALIZED STOCK PRODUCTION MODEL The difference method involves writing Equa- tion (1) as a finite difference equation for the pro- duction model in terms of catch per unit effort and the estimates for a , |3 , and m as ^At/, 9 M a J u. -f.u. (17) for each year i;M is taken as one unit, Equation (17) is divided through by (/,, summed over the n - 2 yr that A [7, can be estimated, and then solved for q^ , % 1=2 t^i P 1 = 2 - (n - 2) f - 2 /■,] P 1=2 where (18) (19) This method has provided reasonable estimates with the logistic (m = 2) and Gompertz {m—>l) forms of the production model for several fisheries (Fox 1970). Pella and Tomlinson (1969) observed that Equa- tion (19) can be a poor estimator of the change in stock size during year i under certain circum- stances. The integral method avoids this problem by writing Equation ( 17) as a differential equation dU : = q dt, (20) ui-^ -r + 4[/"'-^) where f* , the effective effort having been exerted between years i and i +1, is estimated by r = (/; +/",.i)/2. (21) The integral of Equation (20) after rearranging some terms is q, = \n[\{zUy"'+ 4 )/uu! :T + ^mzm -z) (22) where 2 = -d//? - f*. The fact that Equation (22), as an estimator of g, gives negative values when the stock changes in one direction, depending on whether m is greater or less than 1, is remedied by taking the absolute value of q. Also, since q is constrained against being less than zero, the geometric mean will probably be a better estimator than the arithmetic mean (this will be demonstrated to be so in at least one case), such that q, = e n - 1 2 In I o. I Kn - 1) ( = 1 ' (23) becomes the integral estimator. Variability Measures Some measure of the variability of the parame- ter estimates can be made using the "delta" method (Deming 1943). If S is the weighted re- sidual sum of squares for the final parameter es- timates, a variability index matrix, V, is com- puted by V = {X'WX)'^SI{n - 3) (24) where W is sltx n hy n diagonal matrix of the statistical weights, X is an n by 3-parameter ma- trix of first partial derivatives of Equation (11) with respect to each parameter (given in the Ap- pendix). The diagonal elements of V are variabil- ity indices of the parameter estimates and the off-diagonal elements of V are covariability indi- ces. Since Equation (11) is nonlinear, the indepen- dent variable is not without error, the errors in the dependent variable are correlated, and the statis- tical weights are random variables, it is virtually impossible to make probability statements about the accuracy of the parameter estimates (Draper and Smith 1966). However, V gives some index of the variability inherent in the data which is useful largely for comparative purposes between differ- ent fisheries and data sets. For convenience, an error index may be formulated as E, = [100 ^V {X)]lx (25) where X is the estimated parameter and V (X) is its corresponding variability index. Variability and error indices of Y max/opt. and t/opt also may be computed by the "delta" method (see Appendix) and the elements of V (Equation 24). Program PRODFIT A computer program PRODFIT, in FORTRAN IV language, was written to perform the calcula- tions described above. A brief description of the program's options and mode of operation is given below. DATA INPUT OPTION. Option l.—A catch 27 and fishing effort history, ^Ci , /", | , of i = 1 . . .n years length and a vector of significant year class numbers [k, | are read in. There may be embedded zeros, if they are true zeros and do not simply reflect a lack of information. The only real problem with unreal zeros, however, occurs in the estimation of g^. The catch per unit effort vector is computed internally and the averaged fishing effort vector is computed by Equation (9) with SUBROUTINE AVEFF. Option 2 . — If one wishes to compute the aver- aged fishing effort vector by another method or if data are obtained which represent equilib- rium conditions, then this option is selected and the vectors of catch per unit effort and averaged (or equilibrium) fishing effort |t/, ,//} are read in directly. No estimate of q- can be made, how- ever. STARTING VALUES OPTION. Option 1 .— Initial estimates of the parameters are com- puted in SUBROUTINE INEST and the user provides the starting estimate for m, either 0, 1, or 2. Option 2. — Occasionally the data are so variable that INEST does not provide compati- ble starting values for the parameters. In this case, or in any case, the user may opt to enter directly all the initial parameter estimates. MODEL OPTION. The user may allow PROD- FIT to estimate m to any desired precision. Fre- quently, however, the data are so variable that no significant reduction in the residual sum of squares is obtained by varying m . The user then has the option to fix m at 2, the logistic model (Schaefer 1957); at 1, the Gompertz model (Fox 1970); or at 0, the asymptotic yield model. WEIGHTING OPTION. The user may select the statistical weights as Equation (16) or may choose to not weight the observations, i.e., Wi = 1 for all i. CATCHABILITY COEFFICIENT. The catch- ability coefficient, q, is estimated by Equation (22), but both the geometric and arithmetic av- erages are computed. Program PRODFIT uses an adaptation of the same pattern search optimization routine, MIN, as contained in GENPROD (Pella and Tomlinson 1969) to locate the least-squares parameter esti- mates. A more sophisticated Taylor series ap- FISHERY BULLETIN: VOL. 73, NO. 1 proach (Draper and Smith 1966) was attempted initially, but severe distortion of the sum-of- squares space prevented reasonable convergence. In order to facilitate termination of the searching procedure, the sum-of-squares space is searched with m and a transformation of the parameters a and |3to Ur l/(m - 1) a (g - a.m){alm) m0 l/(m (26) (27) where Umax is the unexploited population size in terms of catch per unit effort. Neither Umax nor Y max change greatly with moderate changes in m . The output of PRODFIT provides a listing of the input data, the transformed data, initial parame- ter estimates, the iterative solution steps, the final estimates of a, |3, and m and their variability indi- ces, the management implications of the final model Umax, U opt, f opt, and Ymax and their variabil- ity indices, the observed and predicted values and error terms, and estimates of the catchability coefficient, q. A listing of program PRODFIT and a user's guide are available on request from the author. COMPARATIVE EXAMPLES OF THE EQUILIBRIUM APPROXIMATION METHODS Two methods of averaging fishing effort which attempt to approximate equilibrium conditions have been presented, the unweighted method (Equation 7) and the new weighted method (Equa- tion 9). In order to compare these two methods, catch histories for a simulated pandalid shrimp fishery (Fox 1972) were generated using a generalized exploited population simulation model GXPOPS (Fox 1973). It should be noted, however, that the comparisons are, for the most part, simply illustrative. It is virtually impossible to demonstrate conclusively which is the better method because there is an infinite choice of life histories, parameter values, fishing effort his- tories, and stochastic variation representations. Equilibrium values for the unexploited popula- tion biomass in terms of catch per unit effort (^max), the maximum equilibrium yield {Ymax), and optimum fishing effort (/"opt), were determined empirically by running the simulation model 28 FOX: FITTING THE GENERALIZED STOCK PRODUCTION MODEL (Table 1). The catchability coefficient, q, was as- sumed to be 1.0. Figure 2 presents the equilib- rium values of catch per unit effort and yield at fishing effort values ranging from 0.0 to 1.3 for the simulated shrimp population. Above/" =1.3 the population level did not stabilize in 25 yr of simu- lation and aif = 2.0 the population was definitely extinguished. The equilibrium data for/" =0.0 to 1.3 were fit to the generalized stock production model with PRODFIT to illustrate the obtainable degree of correspondence. The generalized stock production model very closely approximates the equilibrium values for the simulated pandalid Table 1. — Empirical management implications for the simu- lated pandalid shrimp population and those estimated for the generalized stock production model with PRODFIT. Method V ' max Umax 'opt q m Empirical 5.60 17.96 1.02 1.0 — PRODFIT 5.56 17.91 1.11 — 0.604 18 C 15 - \ A H CC „ O 11^ Ok, li. li. UJ c . ^ X O - ^^ 1- 6 - ^^^o^^ < ^^~- 3 m 13 O UJ J o J 9 ® J o o o « « 8 J • J L J_ X o o o 8 ° _L J I L 5 _l_ 0.0 0.4 0.8 12 0.0 04 0.8 12 FISHING EFFORT Figure 4. — Results (dots) of five stochastic simulated catch trials for the equilibrium approximation approach to fitting the generalized stock production model with the weighted averaging method. Circles are the true values. structure is the multiplicative error model (Fox 1971) C, = C- (28) where C, is the observed catch in year i, Ci* is the expected catch, and e, is a random variable with an expected value of 1 and standard deviation o . In practice, however, the e, are usually correlated because some (or all) of the component sources of variability do not meet the assumptions. An ideal (i.e., in the sense that the e, are inde- pendent and random) error structure was chosen to illustrate the estimation ability of the two equilibrium approximation methods, because the "true" error structure of any given population and fishery is unique and largely unknown. Five inde- pendent sets of 12 pseudorandom, normally dis- tributed variables, 6, as with an expectation of zero and a standard deviation of 0.1 were pro- duced with the Library Subroutine RAND (Uni- versity of Washington Computer Center, Seattle). The sets of 5's were used to produce five stochastic catch data sets from the deterministic catch his- tory (Figure 3) and Equation (28), with e, de- fined as 1 + 6 , . The results of fitting the five replicate sets of catch and effort data by the weighted (Equation 9) and unweighted (Equation 7) averaging methods are given in Table 5. The effects of even moderate variability on the parameter estimates for both averaging methods are apparent. On the average, two (m and i^max) of the three determining parameters (m, Ymax, and Umax) are closer to the empirical values for the weighted effort averaging method. The important observation, however, is that all the unweighted estimates of i^max fall above the empirical value and that the average over the five replicates is significantly different from the empirical value with probability greater than 0.999. Plots of the empirical equilibrium yields and those determined from the generalized stock pro- duction model parameters estimated by the weighted average method are compared in Figure 4. Equilibrium yield, for the most part, is esti- mated reasonably well in each replicate for the range of estimated "equilibrium" fishing effort, 0.0 to 1.0 (Table 3). The exception is replicate 4 where the empirical equilibrium yield is substan- tially underestimated above f = 0.8. Beyond the range of data, f = 1.0 to 1.3, the equilibrium yield is estimated reasonably well on the average, but not individually. None of the fitted models, of 31 FISHERY BULLETIN: VOL. 73, NO. 1 course, reveal that there is no equilibrium yield in the range of f = 1.6 to 2.0 for the simulated shrimp population (Figure 2). Table 6 provides a comparison of the catchabil- ity coefficient estimates by three techniques for each fishing effort averaging method. Clearly the best estimates were produced by the geometric mean for the integral method, with the mean es- timate by the weighted average fishing effort pro- cedure being slightly better than that of the un- weighted average procedure. COMPARATIVE EXAMPLES OF THE EQUILIBRIUM APPROXIMATION AND TRANSITION PREDICTION APPROACHES Computer program GENPROD (Pella and Tom- linson 1969) was employed to fit the deterministic and stochastic catch and effort histories of the simulated shrimp to compare the results of the transition prediction and equilibrium approxima- tion approaches. The reader is cautioned, as in the previous section, that these results are largely illustrative and should not be misconstrued as being valid for all cases in which a production model may be employed. Deterministic Comparison The comparison of equilibrium parameters in Table 7 reveals that the equilibrium approxima- tion approach provided estimates that were closer to all the empirical values except m , where the two approaches estimated the same value as the em- pirical equilibrium fit. GENPROD estimated parameters which predicted the simulated catch history (Figure 3) extremely well— the largest error was only 0.05, the sum of squared errors was 0.00659, and the R statistic, a measure of im- provement in the fit over simply using the mean catch as a predictor (Pella and Tomlinson 1969), was 0.99994. Utilizing the empirical equilibrium parameters in the generalized production model, however, resulted in a poorer, but still good, pre- diction of the transition state catches — the max- imum error was 0.50, the sum of squared errors was 0.48515, and the R statistic was 0.99544. Ap- parently due to failure of the assumptions regard- ing population lag and age structure shifts or problems with precision in the numerical integra- tion, the accuracy of some equilibrium parameter estimates by the transition prediction approach Table 6.— Estimates of the catchability coefficient, q, from the five replicated stochastic catch histories by three methods for the weighted and unweighted fishing effort averaging procedures. Actual value of q is 1.0. Effort method Estimation method Mean q Range of q Weighted r Integral method^ Geometric mean 1.008 0.53-1.56 Arithmetic mean 1.660 1.27-2.41 Difference method^ 1.503 1.35-1.77 Unweighted T' Integral method Geometric mean 1.028 0.65-1.74 Arithmetic mean 1.546 1.12-2.11 Difference method 4.459 2.22-10.47 'Equation (9); k ^Equation (22). ^Equation (18). "Equation (7); T = 4. = 2. were sacrificed in order to reduce the sum of squared errors by nearly 99%. Stochastic Comparison The results of fitting the five replicate stochastic catch histories by the equihbrium approximation and transition prediction approaches are given in Table 8. Of the four common parameters (m, Y max, C/max > and g), the equilibrium approximation ap- proach estimates were closer to the empirical val- ues of m , y^ax > and q , both on the average and for most of the replicates. The transition prediction approach estimates were closer, on the average, to the empirical value for L^'max- The transition pre- diction approach provided one extremely poor es- timate ofq (replicate 3) and all replicate estimates are above the empirical value — the latter phe- nomenon could be related to the accuracy of the numerical integration scheme in GENPROD (Fox 1971). The additional parameter required by GENPROD, the ratio of the initial to unexploited population size (Po/Pmax), was estimated very well. There is considerable variability in the esti- mates of the most frequently desired parameter, ^max, by either approach (Table 8) in spite of as- suming an ideal error structure (independent, Table 7.— Empirical and estimated parameters for the simu- lated pandalid shrimp catch history using the equilibrium ap- proximation and transition prediction approaches. Approach m 0, Empirical '060 5.60 Equilibrium approximation^ 0.60 5.67 Transition prediction^ 0,60 5.92 17.96 17.97 17.69 1.00 0.87 1.32 1.00 1.16 'Estimated, Table 1. ^Program PRODFIT, k = 4. unweighted estimates option. ^Program GENPROD, KK = 3, DEL = 3, unweighted estimates. 32 FOX: FITTING THE GENERALIZED STOCK PRODUCTION MODEL Table 8. — Empirical and estimated parameters for the five replicated stochastic catch histories using the equilibrium approximation and transition prediction approaches. Mean ^^,^^~ squared Method Replicate m ' max Umax Q Pq/P max error Empirical — '0.60 5.60 17.96 1.00 1.000 20.0100 Equilibrium 1 1.03 5.80 17.49 0.53 — 0.0136 approximation 2 0.00 8.65 18.99 1.56 — 0.0107 approach^ 3 0.60 5.73 17.97 0.87 — 0.0083 4 1.04 5.07 18.68 0.95 — 0.0047 5 0.24 6.68 18.40 1.13 — 0.0145 Mean 0.58 6.39 18.30 1.01 — 0.0104 4SEx 0.21 0.62 0.26 0.17 — 0.0018 Transition 1 1.7 5.81 17.72 1.34 0.738 0.0105 prediction 2 0.0 9.09 18.15 1.18 1.095 0.0131 approach^ 3 2.1 6.69 17.29 3.97 1.211 0.0086 4 1.7 5.26 17.83 1.40 1.313 0.0053 5 0.0 9.34 19.21 1.52 0.797 0.0126 Mean 1.10 7.24 18.04 1 88 1.031 0.0100 SEx 0.45 0.84 0,32 0.52 0113 0.0014 'Estimated, Table 1. ^Assumed value ^Program PRODFIT; /( = 4, weighted estimates option. "Standard error of the mean. ^Program GENPROD; KK = 3, DEL = 3. weighted estimates. The program was modified slightly from the version of Pella and Tomllnson (1969) by replacing /opt with Vmax as one of the determining parameters to allow fittmg the case where rfi = 0 (i.e. ?opt = =c at m =0). Identical solutions were obtained for the remaining three cases with either version. random and with constant expectation and vari- ance), the observed catch being within 20% of the expected catch with probability 0.95, and the fishing effort being known without error. The maximum error for the equilibrium approxima- tion approach was +54% (replicate 2) and for the transition prediction approach was +67% (repli- cate 5). The problem with these maximum errors (as well as an additional replicate of the transi- tion prediction approach) was estimating m as 0.0, where Ymax occurs at infinite fishing effort. It is not unreasonable, however, to obtain m = 0.0 since the data series is so short and the best value for m is about 0.60. Considering these results and the true relationship between yield and effort (Figure 2) it would be prudent to adopt an alter- native m estimation strategy for short data series. Alternative strategies which could be adopted for short data series are 1) to consistently assume one of the special cases of the generalized stock production model, either the logistic form (m = 2) or the Gompertz form (m -^ 1), or 2) fit both special cases and select the one with the least sum of squared errors. Table 9 presents the parameters estimated by the two approaches through fixing the value for m at 1 (actually 1.001) and 2. For comparative purposes, the results of these alter- native strategies are summarized in Table 10. Fix- ing m at 1 or 2 resulted in average estimates of ^max nearer the empirical value with less vari- ability than obtained by allowing m to be freely estimated for both the equilibrium approximation and transition prediction approaches. The empiri- cal value of m is 0.6; hence assuming m ^ 1 produced estimates nearer the empirical value of ^max than assuming m — 2. For any given data set, however, one could not determine a priori which value of m to assume. The strategy of fitting both m ^^ 1 and m = 2 and then selecting that which provided the least-squares parameter estimates worked very well in comparison with freely es- timating m under three criteria: 1) more accurate average estimate, 2) smaller average percentage error, and 3) smaller maximum overestimate. Comparing the equilibrium approximation and transition prediction approaches with the same three criteria reveals that the equilibrium approx- imation approach was superior [ 1) 0.5% vs. 5.2%, 2) 3.6% vs. 8.5%, and 3) 3.6% vs. 18.4%)]. DISCUSSION The simple, illustrative calculations on the simulated pandalid shrimp population, of course, did not determine which of the approaches was better for general use in fitting the generalized stock production model. However, some additional guidance can be gained through examining some of their relative weaknesses with regard to the number of data points and the number of parameters they require. The moving average of fishing effort in the equilibrium approximation approach results in the exclusion of points at the beginning of the data 33 FISHERY BULLETIN: VOL. 73, NO. 1 Table 9. — Estimated parameters for the five replicated stochastic catch histories using the equihbrium approximation and transition prediction approaches for fixed estimates of m. Mean ^^^ — -^ squared Method m Replicate r max Umax Q Po/Pmax error Equilibrium 1 1 5.80 17.51 0.54 0.0136 approximation 2 5.63 17.99 1.36 — 0.0155 approach' 3 5.56 17.66 0.99 — 0.0087 4 5.05 18.72 0.92 — 0.0048 5 5.78 17.73 0.95 — 0.0166 Mean 5.57 17.92 0.96 — 0.0118 2 1 6.27 16.73 0.29 — 0.0182 2 6.27 17.17 1.14 — 0.0267 3 6.06 16.92 0.88 — 0.0139 4 5.86 17.65 0.77 — 0.0115 5 6.35 16.94 1.19 — 0.0260 Mean 6.16 17.08 0.85 — 0.0193 Transition 1 1 5.44 18.18 1.10 0.778 0.0107 prediction 2 6.00 18.16 1.79 1.014 0.0170 approach^ 3 6.39 17.81 2.97 1.058 0.0096 4 4.66 18.08 1.07 1.162 0.0058 5 6.10 18.40 1.87 0.715 0.0142 Mean 5.72 18.13 1.76 0.945 0.0115 2 1 5.95 17.61 1.38 0.720 0.0107 2 6.47 17.83 2.40 0.962 0.0201 3 6.63 17.37 3.62 1.210 0.0084 4 5.30 17.73 1.35 1.361 0.0051 5 6.51 17.89 2.29 0.604 0.0165 Mean 6.17 17.69 2.21 0.971 0.0122 'Program PRODFIT; /( = 4, weighted estimates option. ^Program GENPROD; KK = 3, DEL = 5, weighted estimates. Table 10. — Summary of Ymax estimates by alternative strategies with the equilibrium approximation and transition prediction approaches for five replicated stochastic catch histories. Empirical value of y^gx is 5.60. Standard Average 'max error of percentage Method/strategy Mean mean error Range Equilibrium approximation approach' 1. Estimate m 6.39 0.62 17.8 5.07-8.65 2. Assume m — » 1 5.57 0.14 3.6 5.05-5.80 3. Assume m = 2 6.16 0.09 10.0 5.86-6.35 4. Least-squares, m = 1 or 2 5.57 0.14 3.6 5.05-5.80 Transition prediction approach^ 1. Estimate m 7.24 0.84 31.7 5.26-9.34 2. Assume m —> 1 5.72 0.31 10.0 4.66-639 3. Assume m = 2 6.17 0.25 12.4 5.30-6.63 4. Least-squares, m = 1 or 2 5.89 0.24 8.5 5.30-6.63 'Program PRODFIT. ^Program GENPROD. set unless either there was no fishing prior to the first record of the set or some information is avail- able on the approximate level of catch and effort. One should check carefully to ensure that critical points (those being the only points at high, low, or intermediate levels of fishing) are not excluded or that the fitted model does not deviate greatly from where they might reasonably be expected to lie. If fishing effort was reasonably constant or negligi- ble prior to the first record, dummy data of length ^ - 1 can be employed to allow use of the first few data points. Also, since the average fishable dura- tion, T, is less than the number of significant fishable year classes, k , the unweighted averaging method will result in fewer data being excluded In any case, the sensitivity of the parameter esti- mates to alternative averaging times should be explored. No data points are excluded with the transition prediction approach, a positive factor which should be considered even if one is satisfied with the parameter estimates obtained with the equilibrium approximation approach. On the other hand, the transition prediction approach utilizes five parameters while the equilibrium ap- proximation approach utilizes only three, so that vdth few significant year classes in the fishable population there is little difference between the required number of data points. For example, the transition prediction approach statistically 34 FOX: FITTING THE GENERALIZED STOCK PRODUCTION MODEL requires six points, while the equihbrium ap- proximation approach with four significant year classes will require, in general, seven points. With a large number of significant year classes in the fishable population or a relatively high age at first capture, however, the major concern for either approach is the likelihood of failure of the transition state population assumptions. The results summarized in Table 7 illustrate a general shortcoming in simultaneously estimat- ing a large number of parameters, i.e. large devia- tions from model can be statistically reduced in a least-squares estimation procedure at the expense of the accuracy of certain "desired" parameters. The transition prediction approach, fitting a "free-form" type of curve with five parameters, is relatively more susceptible than the equilibrium approximation approach which fits a monotoni- cally decreasing curve with only three parame- ters. On the other hand, estimates from the equilibrium approximation approach can be very sensitive to the placement of one data point in certain cases (e.g., a data point at an intermediate level of fishing with clusters of points at both low and high levels of fishing). Utilizing the production model approach for as- sessing the effects of exploitation presents significant problems in addition to choice of the parameter estimation procedure or the length of the data series. These additional problems are 1) maintaining a constant catchability coefficient throughout the data series, 2) assessing the effects of changes in the constitution of the fishery, and 3) assessing the effects of time lags in population production processes. The basic components of the overall effective catchability coefficient are 1) the relative ef- ficiency of various types and classes of fishing gear and 2) the manner in which the gear is employed relative to the availability and vulnerability of the population, and its subunits, to capture. Heterogeneity in the efficiency of various gear classes, or vessels, within a fishing season can be alleviated by adjusting for their estimated rela- tive fishing powers — currently the best method for estimating fishing power is by analysis of variance with the computer program FPOW (Berude and Abramson 1972). The major problem remaining, however, is adjusting for among-year changes in efficiency of the standard gear. Rothschild (1970) discussed and provided examples of problems as- sociated with changes in the catchability coefficient related to areal deployment of the fishing gear. The expansion of fishing across a gradient of population density will increase or de- crease the effective catchability coefficient de- pending on the direction of the density gradient and fishing expansion. Year-to-year shifts in the population location and density relative to the fishing effort deployment also could likewise create trends in the catchability coefficient. Age-specific differences in the catchability would cause shifting of the overall effective catchability coefficient with changes in fishing effort. For ex- ample, if the catchability offish declined with age, then the overall effective catchability of the fishable population would increase with increas- ing fishing effort since the relative proportion of younger age groups would most likely increase. Alterations in the constitution of the fishery probably are the most difficult problems to over- come satisfactorily. Expansion of the fishery across several stocks, either independent or with some mixing, can result in rather large shifts in the productivity estimates (Joseph 1970; Inter- American Tropical Tuna Commission 1972). Changes in the relative levels of fishing effort exerted by different gear types which exploit dif- ferent age groups of the population, either volun- tarily or through a change in minimum size limit regulations, can similarly have significant impact on the shape of the production model curve (Le- narz et al. 1974). The latter problem identifies a major shortcoming of the production model ap- proach; i.e., the impact on total yield by altering the selectivity of fishing gear can not be assessed a priori without considerable additional informa- tion. The effects of time lags in population production processes (e.g., reproduction, growth, and mortal- ity, both density-independent and density- dependent) can result in either overestimation or underestimation of the population productivity, or in population cycling which may never result in reaching an equilibrium state (Wangersky and Cunningham 1957; Walter 1973). In summary, both the equilibrium approxima- tion and the transition prediction fitting methods are useful, one or the other more so under condi- tions outlined above. Application of the produc- tion model to catch and fishing effort data is rela- tively simple, the primary virtue of the approach. The interpretation of the results and the formula- tion of advice for managing the resource, however, can be extraordinarily complicated by a variety of 35 FISHERY BULLETIN; VOL. 73, NO. 1 factors. Therefore, the proper perspective of pro- duction model analysis is that it is little more than a regression model, yet very useful for making "first estimate" projections of the relationship be- tween the level of exploitation and expected equilibrium yield. ACKNOWLEDGMENTS Douglas G. Chapman and Gerald J. Paulik of the University of Washington, Seattle, and Brian J. Rothschild of the Southwest Fisheries Center, National Marine Fisheries Service, NOAA, La Jolla, Calif reviewed an early manuscript and offered useful suggestions for improvement. LITERATURE CITED Berude, C. L., and N. J. Abramson. 1972. Relative fishing power, CDC 6600, FORTRAN IV. Trans. Am. Fish. Soc. 101:133. Beverton, R. J. H., AND S. J. Holt. 1957. On the dynamics of exploited fish populations. Fish. Invest. Minist. Agric, Fish. Food (G.B.), Ser. 2, 19, 533 p. Chapman, D. G. 1967. Statistical problems in the optimum utilization of fisheries resources. Int. Stat. Inst., Bull. 42(l):268-290. Deming, W. E. 1943. Statistical adjustment of data. John Wiley & Sons, N.Y., 261 p. Draper, N. R., and H. Smith. 1966. Applied regression analysis. John Wiley & Sons, N.Y., 407 p. Fox, W. W., Jr. 1970. An exponential surplus-yield model for optimizing exploited fish populations. Trans. Am. Fish Soc. 99:80-88. 1971. Random variability and parameter estimation for the generalized production model. Fish. Bull., U.S. 69:569-580. 1972. Dynamics of exploited pandalid shrimps and an evaluation of management models. Ph.D. Thesis, Univ. Washington, Seattle, 223 p. 1973. A general life history exploited population simulator with pandalid shrimp as an example. Fish. Bull., U.S. 71:1019-1028. Graham, M. 1935. Modem theory of exploiting a fishery, and applica- tion to North Sea trawling. J. Cons. 10:264-274. GULLAND, J. A. 1961. Fishing and the stocks offish at Iceland. Fish. In- vest. Minist. Agric, Fish. Food (G.B.), Ser. 2, 23(4), 52 p. 1969. Manual of methods for fish stock assessment. Part 1. Fish population analysis. FAO (Food Agric. Organ. U.N.) Man. Fish. Sci. 4, 154 p. Inter-American Tropical Tuna Commission. 1972. Annual report of the Inter-American Tropical Tuna Commission, 1971, 129 p. Joseph, J. 1970. Management of tropical tunas in the eastern Pacific Ocean. Trans. Am. Fish. Soc. 99:629-648. Lenarz, W. H., W. W. Fox, Jr., G. T. Sakagawa, and B. J. Rothschild. 1974. An examination of the yield jjer recruit basis for a minimum size regulation for Atlantic yellowfin tuna, Thunnus albacares. Fish. Bull., U.S. 72:37-61. Lord, G. 1971. Optimum steady state exploitation of a multispecies population with predator-prey interactions. Univ. Wash., Fish. Res. Inst., Cent. Quant. Sci. For. Fish. Wildl., Quant. Sci. Pap. 29, 8 p. Pella, J. J., AND p. K. Tomlinson. 1969. A generalized stock production model. Inter-Am. Trop. Tuna Comm., Bull. 13:419-496. Rothschild, B. J. 1970. A systems view of fishery management with some notes on the tuna fisheries. Univ. Wash., Cent. Quant. Sci. For. Fish. Wildl., Quant. Sci. Pap. 14, 78 p. Schaefer, M. B. 1954. Some aspects of the dynamics of populations impor- tant to the management of commercial marine fisheries. Inter- Am. Trop. Tuna Comm., Bull. 1:25-56. 1957. A study of the dynamics of the fishery for yellowfin tuna in the eastern tropical Pacific Ocean. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull. 2:245-285. Schaefer, M. B., and R. J. H. Beverton. 1963. Fishery dynamics — their analysis and interpreta- tion. In M. N. Hill (editor). The sea, Vol. 2, p. 464-483. John Wiley & Sons, N.Y. Walter, G. G. 1973. Delay-differential equation models for fisheries. J. Fish. Res. Board Can. 30:939-945. WaNGERSKY, p. J., AND W. J. CUNNINGHAM. 1957. Time lag in population models. Cold Spring Harbor Symp. Quant. Biol. 22:329-338. i I 36 FOX: FITTING THE GENERALIZED STOCK PRODUCTION MODEL APPENDIX Miscellaneous Equations for PRODFIT Elements of the X -matrix Lett/, = (a + 0/,) Then 9f/, ^ ,, - (2 - m)/(m - 1) g^ =l/(m - l)(a + )3/-,) a^ ' bq ^^i , . - l/(m - 1) _ -g7^ = - (a + 0/;) X ln(a + (3/",)[l/(m - DP Derivatives for the Delta Method Variance Estimates y -' max "_£_max _ TT- r , a - i'^ maxim/ (m - l)]/a a ay a^ •^ max'P 9"^ ma ^^^r^ = ^max X In (m/a)/(m - 1)2 Z'. opt ^/'opt 9~- = (1/m - 1)//? 9/ opt a /"opt 3^r- = -a/((3m2) t^opt at/opt a a f^opt/[a(m -1)] ■^ - -t/opt [m In (a/m) + m -l]/[m(m - 1)2] 37 NET PHYTOPLANKTON AND THE GREATER THAN 20-MICRON PHYTOPLANKTON SIZE FRACTION IN UPWELLING WATERS OFF BAJA CALIFORNIA Theodore J. Smayda^ ABSTRACT Between 26 March and 6 April 1973 various phytoplankton studies were carried out during the MESCAL II survey in an area measuring circa 105 km x 30 km, and centered approximately at Punta San Hipolito, Baja California. Upwelling was then in its early stages. The composition of 22 collections of net phjftoplankton (No. 20 net), and the composition and abundance of the non-setose (i.e., excluding Chaetoceros, Bacteriastrum) size fraction >20 nxn collected at various depths at 13 stations are reported here. The mean carbon content in the upper 50 m contained in the >20 pm non-setose size fraction was 533.5 mg C/m^ for all stations, and ranged from 306 to 1,022 mg C/m* at individual stations. Based on a C/Chl a ratio of 40: 1, the mean concentration in the euphotic zone represents circa 12% of the total phytoplankton carbon present. Lauderia annulata (28%) and several Coscinodiscus species (33%) accounted for most of the carbon found in the >20-Mni size fraction, even though the latter comprised only about 10% of the mean population expressed as cell number. The mass occurrence of Coscinodiscus reported previously for Magdalena Bay during summer upwelling was not observed. The Coscinodiscus population and the non-setose component of the >20- /jm size fraction contributed only 1.2% and 4%, respectively, of the daily caloric ingestion estimated for the crab, Pleuroncodes planipes, previously reported to graze heavily on Coscinodiscus. Sinking rates (61 to 144 m/h) of Pleuroncodes fecal material exceeded by onefold to fourfold those rates estimated for the various sizes of Coscinodiscus and zooplankton fecal pellets sampled during the survey. The abundant crab population is, thus, important in causing an exceptionally rapid deposition of unassimilated phytoplankton frustules and organic material onto the sea floor. Floristic changes accompanying upwelling were detectable. The occurrence of the unique diatoms Coscinodiscus (Brenneckella) eccentricus and Planktoniella muriformis in these waters is apparently reported for the first time. The present data together with earlier observations suggest that the net diatom community is similar in the coastal waters extending from San Diego, Calif, to the Gulf of Panama. The data do not support the idea that the abundance of Pleuroncodes in this upwelling system is causally linked to that of Coscinodiscus. There is little information on the composition and abundance of the phytoplankton community in the upwelling waters off Baja California. The available data are mostly qualitative (Allen 1924, 1934, 1938; Balech 1960; Cupp 1930, 1934), aside from recent, cursory observations on phyto- plankton cells >25 A( m which are grazed by the red crab, Pleuroncodes planipes (Longhurst et al. 1967). Unique mass blooms of Coscinodiscus have been observed by Longhurst et al. in the upwelled waters of Magdalena Bay. This phenomenon and the great abundance of Pleuroncodes, which grazes on Coscinodiscus cells, may be distinctive characteristics of this upwelling system. Smith et al.2 have concluded that this crab is an impor- 'Graduate School of Oceanography, University of Rhode Is- land, Kingston, RI 02881. ^Smith, K. L., Jr., G. R. Harbison, G. T. Rowe, and C. H. Manuscript accepted May 1974. FISHERY BULLETIN: VOL. 73, No. 1, 1975. tant herbivore in the California Current upwell- ing system, where its role is comparable to that of the anchovy, Engraulis ringens, in the Peru Current. Longhurst (1968) has evaluated the potential fishery for this galatheid crab, which occurs in both the benthic and pelagic zones; some crabs exhibit diurnal migrations (Longhurst et al. 1967). Pleuroncodes is generally distributed through- out this region (Blackburn 1969), while informa- tion on the regional distribution and abundance of Coscinodiscus is lacking. It is therefore unknown whether Coscinodiscus indeed generally charac- terizes the phytoplankton community in these upwelled waters. Clarification of this is relevant Clifford. Respiration and chemical composition of Pleuroncodes planipes (Decapoda: Galatheidae): Energetic significance in an upwelling system. Manuscr., 22 p. Woods Hole Oceanogr. Inst., Woods Hole, Mass. 38 SMAYDA: NET PHYTOPLANKTON IN UPWELLING WATERS to the question of whether Pleuroncodes' occur- rence is causally linked to that of Coscinodiscus. Coscinodiscus and other heavily silicified diatoms sink to the sea floor, as documented for the Gulf of California (Round 1967, 1968). This deposition contributes skeletal remains (i.e., to the thanatocoenosis) and organic matter to the sediments. The abundance and sinking charac- teristics of this diatom population are also of in- terest, since Pleuroncodes also occurs in the benthos. During this benthic residence, when population densities up to 250 individuals/m^ have been found (Smith et al. footnote 2), it may feed on detrital material (Longhurst et al. 1967). Various studies on phytoplankton were carried out during the MESCAL II expedition of the RV Thomas G. Thompson in 1973 to study upwelling off Baja California, as a continuation of 1972 ac- tivities in this area (Walsh et al 1974). These included the routine, shipboard examination of both net phytoplankton and the >20- Mm size frac- tion filtered from quantitative samples. The dis- covery that natural populations of a Ditylum brightwelli and, possibly, Biddulphia mobiliensis exhibited diel cell division has been reported (Smayda in press a). Examination of the >20-/jm size fraction was partly motivated by the need to know its composi- tion and abundance, particularly that for Coscino- discus. This was to evaluate the aforementioned relationship possibly occurring between Pleuron- codes and Coscinodiscus, and to establish the latter's importance during the initial stages of upwelling. This latter objective was prompted by the remarkable bloom found in Magdalena Bay during a later stage of the upwelling cycle. Finally, such data are needed to evaluate the sinking and turnover rates of the more heavily silicified and dissolution-resistant components of this size fraction which sink faster and repre- sent an energy source for benthic secondary production. METHODS Between 26 March and 6 April 1973, 22 collec- tions of net phytoplankton were made at 15 sta- tions (multiple sampling on different days at some). A No. 20 (mesh opening of 69 /am) net 30 cm in diameter sampled the upper 50 to 100 m (de- pending on depth) for 30 min by repeated vertical oscillations, during which the net was lowered at a rate of ca. 30 m/min and retrieved at a rate of 10 m/min. The samples were examined microscopi- cally soon after collection, after placing onto a slide an aliquot of the unpreserved, sedimented material from an unagitated sample. Of the 15 stations sampled, 13 were located in a sampling block measuring about 105 km long and 30 km wide centered approximately at Punta San Hipolito off the coast of Baja California (Figure 1). The coordinates of the northern- and southern- most stations are lat. 27°6.7'N, long. 114°21.2'W andlat. 26°28.5'N,long. 1 13°45. 5 'W, respectively. The stations extended offshore from within sight of land to within, or near, the California Current; the inner- and outermost stations were at lat. 26°55.2'N, long. 114°02.2'W and lat. 26°51.2'N, long. 114°10.7'W, respectively. Stations 1 and 2 (not shown in Figure 1) were located about 460 km north of this main sampling area at lat. 30°57.8'N, long. 116°32'W and lat. 28°8.2'N, long. 115°39.2'W, respectively. Quantitative samples were also collected at 13 stations from the surface to 50 m at 10-m inter- vals, and at 75 m with 5-liter Niskin Bottles. Seven stations (18 to 24) were sampled at 6-h intervals while following a drogue. From samples collected in the upper 30 m, 2 liters were usually filtered through a 20-/im mesh net, and 3 liters from greater depths. The apparatus used is illus- trated in Durbin et al. (in press). The material JO- 5' 30' 15' II4'>00' 45' 30j_. 114° \. iZ-'^MEXICO IS' 50 (m I00-- - *._fm , — A' SAN PABLO PT 6i, 30» ^AN MBL0>T 30° 15' '~-~.' j\ tS>iA-., NAUTICAL MILES 0 5 10 15 10 ABREOJOS i ^24 *°''" 100 (m 45' 30' 0 10 20 KILOMETERS 1 1 4 5' 30' 15' II4<>00' 45' 2 0' Figure 1. — Location of stations ■where collections of net phyto- plankton ( + ) and net and water bottle samples (•) were made from 26 March to 6 April 1973 (except that net and water bottle collections were made only at Stations 26 and 38 at the fre- quently sampled station located off Punta San Hipolito). A rep- resents stations, along with Station 27, used to illustrate the occurrence of upwelling in Table 1; the outermost station is Station 29. 39 FISHERY BULLETIN: VOL. 73, NO. 1 retained by the net was concentrated to 30 ml, preserved with hexamine + Formalin^ and 1 ml of the concentrate then enumerated on board ship using a Sedgwick Rafter Counting Chamber. The concentrate was obtained by stopping filtration to leave about 1 cm of water above the filter. As stated in the Introduction, cells in the size class >20/umare frequently heavily silicified and sink to the sea bed. Chaetoceros and Bacteriastrum are usually not represented in the latter (Round 1968), presumably because of rapid dissolution of their silicon frustules. The various objectives of the present and other, ongoing studies during MESCAL II emphasized the Coscinodiscus and other non-setose genera, and also required real-time data for proper execution of the program. Shipboard enumeration of phyto- plankton was therefore necessary. Quantitative shipboard enumeration of specimens belonging to genera characterized by setae is difficult; their setose nature makes them prone to movement within the counting chamber in response to the ship's vibrations and movement. For these various reasons, during the numerical census repre- sentatives of the genera Chaetoceros and Bacteri- astrum and a species similar in general appear- ance to (but probably not) Nitzschia frigida were not enumerated. Numerical abundance was transformed into carbon equivalents. From 10 to 40 cells of each species (depending on abundance) were measured to obtain the mean dimensions required to calcu- late cell volume using appropriate mensuration formulae. The carbon content was then calculated from Strathmann's (1967) equation log C = 0.758 (log V) - 0.352 where V is the cell volume in jum^. From this cellular estimate (pg per cell), the population car- bon was computed. The constant 0.352 differs slightly from Strathmann's given value, and is based on additional diatom analyses (Eppley, pers. commun.). The mean population per liter (C) in the upper 50 m was calculated from: C=^ [(Co+Ci)(Z, -Zo) + (Ci +C2)(Z3-Z2) + . . . + (C4 + C5) (Z5 + Z4)] where Co, Ci, etc. are the observed cell (carbon) concentrations per liter at the surface (Zq) and 10 m (Zi), etc., down to 50 m (Z5). Concentrations per square meter of sea surface down to 50 m were obtained from (C) (5 x lO'*). Samples were collected at 75 m at only 9 of the 13 quantitative stations because of depth. This, together with the sparse populations usually found there, accounts for the emphasis on the upper 50 m. RESULTS Upwelling occurred during the field program. Table 1 presents some representative physical and chemical parameters along a transect of three sta- tions sampled on 3 and 4 April near Punta San Hipolito (Figure 1). The inflow and upwelling of cold, nutrient-rich water at the nearshore station (27) is evident. Upwelling was usually more pro- nounced near and shorewards of the 50-fathom isobath. Details of this upwelling, which was in its early stages, and associated biotic responses will be presented elsewhere (Walsh, Kelley, Whit- ledge, Huntsman, and Pillsbury in prep.). Net Phytoplankton The species identified in the net material are listed in Appendix Table 1. Throughout the ship's track of ca. 700 km the No. 20 net phytoplankton was characterized by the genus Chaetoceros Table 1. — Hydrographic conditions along a transect off Punta San Hipolito showing the occurrence of upwelling during 3 and 4 April 1973 (Stations shown in Figure 1). I ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Depth (m) fiQ aflite r °C o/oo 0, PO4 NO3 8102 Station 29 (lat 26°48N, long. 114°07.5'W) 0220 0 16.04 34122 25.08 0.47 0.98 1.23 10 16.04 34.119 25.08 0.54 0.98 1.10 20 15.67 34.127 25.17 0.55 0.98 0.98 30 15.65 34.112 25.16 0.63 1.31 1.53 40 13.68 33.957 25.47 1.05 3.94 7.52 50 11.87 33.736 25.65 1.40 9.18 13.34 75 11.51 33.821 25.79 1.65 14.43 18.83 Station 28 (lat. 26°51.5'N, long. 114°04.8'W) 0100 0 15.69 34.044 25.10 0.51 0.66 1.75 10 15.25 34.002 25.17 0.63 0.66 2.48 20 14.35 33.912 25.29 0.82 1.31 4.91 30 13.64 33.930 25.45 1.06 3.94 8.52 40 11.94 33.690 25.60 1.22 8.20 11.29 50 11.84 33.922 25.80 1.67 14.43 16.95 Station 27 (lat. 26°55.2'N, long. 114°02.2'W) 1720 0 13.53 34.080 25.59 1.15 7.03 11.87 10 13.26 34.086 25.65 1.32 7.91 12.39 20 13.04 34.092 25.70 1.57 10.94 15.36 30 12.26 34.063 25.83 1.83 14.28 18.85 40 11.81 34.141 25.98 2.12 20.08 22.68 50 11.36 34.243 26.14 2.34 23.72 27.05 40 SMAYDA: NET PHYTOPLANKTON IN UPWELLING WATERS {affinis, curvisetus, debilis, didymus, socialis) in species and abundance. The genus Coscinodiscus was a conspicuous co-dominant, but varied in rela- tive abundance from station to station. The re- markable colonial diatom Planktoniella murifor- mis (Loeblich et al. 1968; Round 1972) was also prominent throughout this region. Nonetheless, some apparent regional differences are notewor- thy. At Station 1 located near Punta Kolnett a very rich, diverse net plankton community occurred on 26 March dominated by Chaetoceros and Nitzschia spp. and Thalassiothrix frauenfeldii . Asterionella japonica, Eucampia cornuta, and Lithodesmium undulatum were other abundant diatoms. This community stands out from others in the importance of Asterionella (many small pennate diatoms were attached to the colonies), which was not found in subsequent net tows. Also unlike subsequent stations, Phaeocystis cf. pouchetii was common while Coscinodiscus spp. were not. Allen (1945) has reported extensive blooms of Phaeocystis off southern California. This colonial haptophycean is well known for its apparent adverse effects on certain fisheries in the North Sea during mass blooms. The lack of nutrient data at Stations 1 and 2 prevents assessment of possible upwelling. How- ever, when sampled on 27 March the upper 50 m of the latter station was considerably warmer (15.29°C at 0 m, 14.94°C at 50 m) than at Station 1 (14.53°C at 0 m, 11.42°C at 40 m). The net phyto- plankton community was considerably poorer and dominated by Ceratium spp.; peridinians were frequent, and the diatoms Biddulphia mobil- iensis, Coscinodiscus spp., and Planktoniella sol were common. This community suggests that up- welling was weak, if occurring. The principal features of the net collections {n = 20) made in the intensive survey area (Figure 1) are: 1) the community at the deepwater stations (16, 31, 32) located seaward of the 50-fathom isobath was less abundant and differed somewhat relative to the shallower stations; 2) the composi- tion at the latter stations was generally similar; and 3) a slight change in apparent species domi- nance occurred by the end of the 12-day sampling period. At the outer, deepwater Station 16 (30 March) Chaetoceros affinis and curvisetus dominated; Ceratium and Peridinium spp. were also common. At Stations 31 and 32 (4 April), Bacteriastrum dominated together with the above Chaetoceros species and decipiens and socialis. Coscinodiscus spp. were subordinate; Asterolampra marylandica and cf Pyrocystis lunula were frequent. The lower relative abundance and the difference in domi- nant species at these outer stations are also reflected in the quantitative samples (Table 2). The lowest mean concentration occurred at Sta- tion 32 (quantitative samples were not collected at Stations 16 and 31). The physical-chemical data indicate that upwelling was not occurring at Sta- tion 16 and was insignificant, if taking place, at Stations 31 and 32. At the nearshore Station 34 (4 April), where the hydrographic conditions were similar to Station 27 (Table 1), the Bacteriastrum component important at Stations 31 and 32 was absent and Thalassiosira rotula dominated along with the Chaetoceros spp. This increased importance of Thalassiosira rotula relative to samples collected a week earlier is also noted in the series collected near Punta San Hipolito (Stations 3 to 38) (Figure 1). The nearshore communities were otherwise dominated by different proportions of Chaetoceros and Coscinodiscus spp. The Coscinodiscus compo- nent was especially prominent at Stations 10, 17, and 19, for example. (The net tows frequently contained pennate diatoms which might have been scoured from bottom sediments during upwelling.) The apparent differences in net community composition, abundance, and species succession during the 10-day sampling period in the inten- sive survey area probably reflect variations in in- tensity of upwelling, which was just beginning based on aerial reconnaissance of sea-surface temperatures prior to the ship's arrival in the sur- vey area. Between 28 March (Station 3) and 30 March (Station 13) cold water ascended 10 m at the fixed station near Punta San Hipolito (Figure 1; Pillsbury, pers. commun.). Quantitative Samples Numerical Abundance The results of the quantitative census of the non-setose species in the >20-/jm size fraction are presented in Table 2. The mean population level in the upper 50 m ranged from about 2,110 to 9,800 cells/liter. Lauderia annulata dominated numeri- cally (from 35 to 60% of total abundance) through- out the area, except at the last station (38) sam- pled (3%) where Thalassiosira rotula dominated 41 FISHERY BULLETIN: VOL. 73, NO. 1 Table 2. — The mean, non-setose population (cells/liter and Mg C/liter) in the >20- /um size class in the upper 50 m. Lower value in cell abundance (i.e. nin) represents number of dead cells. Sta- tion Time c o 1 o o w II 10 c 11 E 3 . ■c a. to Q. Coscinodiscus (Brennecketia) eccentricus w 3 0 « ■6 0 0 1 II Q CD (D UJ CD to TO 0 £ to 0 S CO to SI -J Is H -J (0 to §1 '5 c 0 c 0 (0 CO s: 2 c 0 CO 0 Q. N Q. £ tn Qj to 0 :::; ^- CO » .0 to io CO :3 CO C 1 = tr LU I 1- 0 d. Q. CO 0) CO. a 11 "2 " to CO Q> 0 ■DO 13 1145 4,067 29/5 115/1 119 33 630/54 198/7 24 3 2,015/2 330 32 308/25 5/1 173 48 5 8.5 11.05 0.07 0.72 0.25 0.08 4.83 1.49 0.01 0.01 2.68 0.48 0.20 0.17 0.06 18 1800 3,995 47/6 254/15 29 35 631/49 284/6 57 16 1,396 589 87 335 3 177 48 7 7.8 13.85 0.11 1.58 0.06 0.09 6.61 2.14 0.02 0.07 1.86 0.85 0.22 0.18 0.06 19 0000 4,576 94 171 12 30 520/63 267 14 43 2,942 508 31 220 6 448/17 24 5 11.5 16.13 0.23 1.06 0.02 0.08 7.26 2.01 0.004 0.18 3.92 0.74 0.15 0.45 0.03 20 0600 3,321 22/3 59/2 12 7/1 309/27 125 2 22 2,008 306 20 197/3 5 116/34 114 3 8.7 7.38 0.05 0.37 0.02 0.02 2.38 0.94 0 0.09 2.67 0.44 0.13 0.12 0.15 21 1200 2,456 29/2 68/2 5 29 332/35 137 — 37 1,171/2 181/18 28 194/5 36/2 168 37 4 10.5 8.31 0.09 0.42 0.01 0.07 4.37 1.03 — 0.15 1.56 0.26 0.13 0.17 0.05 22 1800 9,806 45 232 10 55 479/29 552/15 123 91 5,665 719 37 746 11 841/19 190 10 6.1 20.44 0.11 1.44 0.02 0.14 3.96 4.16 0.04 0.38 7.55 1.04 0.50 0.85 0.25 23 0000 4.476 26/1 76/2 10 40 318/50 148/2 77 94 2,512 425 36 288/12 130/2 236/50 56 4 15.7 8.96 0.06 0.47 0.02 0.10 2.27 1.16 0.02 0.39 3.35 0.62 0.19 0.24 0.07 24 0600 3,595 21/1 80/2 17 97 191/17 64 — 101 1,957/10 314 7 538/13 52 113 38/5 5 8.9 7.11 0.05 0.50 0.04 0.24 1.79 0.48 — 0.42 2.61 0.46 0.36 0.11 0.05 25 1200 4,915 48 76 36 44 257/42 179 190 42 2,616 461 30 569 28 276 46 17 16.3 9.13 0.12 0.47 0.07 0.11 1.90 1.35 0.06 0.18 3.48 0.67 0.38 0.28 0.06 26 1800 3,248 26/2 88 1 12 508/50 78/2 10 8 1,526 148 6 46 — 170 618 3 9.8 7.75 0.06 0.55 0 0.03 3.23 0.59 0.003 0.03 2.03 0.21 0.03 0.17 0.82 32 1800 2,110 8 59 43 101/1 230/43 53/1 — 109 819 120 6 552 29 7 10 7 18.7 6.12 0.02 0.37 0.09 0.25 2.88 0.40 — 0.46 1.09 0.17 0.37 0.01 0.01 34 2140 8,653 10/2 54/1 12 11 406/16 272 46 43 3,835 486 12 151/2 7 1,318 1,990 — 3.9 15.00 0.02 0.34 0.02 0.03 2.47 2.05 0.01 0.18 5.11 0.70 0.10 1.33 2.64 38 0720 4,926 53/2 6 49 4 623/65 40 18/60 26 136/7 602/19 8 82 5 618 2,639/2 17 10.4 9.03 0.13 0.04 0.10 0.01 3.10 0.30 0.005 0.11 0.18 0.87 0.05 0.62 3.51 X = 10.5 (54%). The latter was also important at Station 34 (23%); the maximum, mean abundance of Schroederella delicatula (1,318 cells/liter) was also found here. Coscinodiscus spp. were usually next to Lauderia in numerical importance, and composed a maximum of 10 to 15% of the mean population. Table 3 lists the species of Coscinodiscus found. (Coscinodiscus "large species" may represent several species difficult to identify properly in the counting chamber.) Planktoniella sol contributed from about 2 to 25% of the mean population, and Lithodesmium un- dulatum usually around 7 to 12%. The absolute abundance of the unique colonial aggregate, Planktoniella muriformis, which can have up to at least 530 cells/colony (Loeblich et al. 1968), is unknown; individual cells in the colonies were not counted. The mean number of colonies per liter in the upper 50 m ranged from 6 to 92, with the low levels (6 to 10) persistent at stations made after Station 25 (Table 2). The mean vertical distribution of this species shows a similar abun- dance (22 to 30 colonies/liter) in the upper 40 m (Table 4; Figure 2), contrary to expectations, and will be reconsidered later. Species ofRhizosolenia >20 ^^m were not abun- dant, and included: bergoni ,calcar avis ,imbricata var. shrubsolei , stolterfothii . Diatoms which are included in OTHERS in Table 2, and their max- imum abundance (cells per liter) are: Asteromphalus heptactis (12) Corethron pelagicum (19) Dactyliosolen sp. (72) Hemidiscus cuneiformis (84) Leptocylindrus danicus (228) Paralia sulcata (79) Skeletonema costatum (152) Stephanopyxis cf turris (192) Thalassionema nitzschioides (126) Thalassiothrix cf. mediterranea var. pacifica (16) 42 SMAYDA: NET PHYTOPLANKTON IN UPWELLING WATERS Table 3. — The mean population as cells/liter (a) and ng C/liter (b) of the different Cosdnodiscus species >20 fjm in the upper 50 m. Station CO Q. E o (0 "t5 C c o c o o to o c O O c u "co « o (U Q. 0) 2 13a 33 118 3 465 3 41 663 b 0.082 0.568 0.015 1.361 0.061 2.822 4.91 18 a 35 209 95 262 2 63 666 b .087 1.005 .467 .767 .041 4.337 6.70 19 a 30 202 103 135 4 77 551 b .075 .972 .506 .395 .082 5.300 7.33 20 a 7 152 36 104 0 17 316 b .017 .731 .177 .304 0 1.170 2.40 21 a 29 182 18 94 5 44 372 b .072 .876 .088 .275 .102 3.029 4.44 22 a 55 204 57 184 5 30 535 b .137 .981 .280 .539 .102 2.065 4.10 23 a 40 144 28 116 10 13 351 b .100 .692 .138 .340 .205 .895 2.37 24 a 97 64 51 64 4 14 294 b .241 .308 .251 .187 .082 .964 2.03 25 a 44 167 13 64 1 12 301 b .109 .803 .064 .187 .020 .826 2.01 26 a 12 176 18 283 11 18 518 b .030 .847 .088 .828 .225 1.239 3.26 32 a 101 125 17 42 34 20 339 b .251 .601 .084 .123 .696 1.377 3.13 34 a 11 95 24 271 0 16 417 b .027 .457 .118 .793 0 1.101 2.50 38 a 4 60 19 528 0 17 628 b .009 .289 .093 1.545 0 1.170 3.11 These species are listed only to indicate their pres- ence; their actual abundance is probably greater, since most of these would routinely pass through a 20-Mni mesh net depending on orientation of the cells during filtration. Ceratium furca usually dominated the dinoflagellates >20-)Um ; populations of Ceratium fusus were persistent. Reproductive stages simi- lar to those depicted by von Stosch (1964) for some ceratians were frequent. Pyrocystis was present, including an organism quite reminiscent oi Py- rocystis lunula (vide Figure 559 in Schiller 1937) in shape and stages found. Maximum abundance was 60 cells/liter in the upper 10 m at Station 38 (13.96° to 14.31°C, otherwise similar to Station 27 (Table 1)). Various stages of the cf. Pyrocystis lunula cycle were also found during growth exper- iments carried out with mixed, natural popula- tions. The dinoflagellate population was usually sparse, however, with no indication of red tide in the >20-/;m size fraction either visually or mi- croscopically. However, several weeks later, fol- lowing temporary subsidence of upwelling, a red- tide outbreak occurred in these waters (Walsh, pers. commun.) similar to pre-upwelling blooms encountered during MESCAL I in March 1972 (Walsh et al. 1974). A coccolithophorid similar to Syracosphaera apsteini (15 cells/liter) was found occasionally. Noctiluca scintillans was frequently encoun- tered in the samples, especially at Station 38, with evidence of active predation of the phytoplankton by Noctiluca. Carbon Abundance The mean carbon content in the upper 50 m for the dominant non-setose diatom component >20 jum, exclusive of Planktoniella muriformis, Rhizosolenia spp., and OTHERS is given in Table 2. The reason for excluding Planktoniella muriformis is because of the great difficulty to enumerate the cells within the colonies, whose size varied considerably. Insufficient specimens of the rarer Rhizosolenia and "other" species pre- vented reliable cell sizing to calculate cell volume. The mean carbon content in the upper 50 m ranged from 6.12 to 20.44 /ug C/liter at the various stations; the overall mean was 10.67 /U g C/liter (Tables 2 to 4). Comparison of the percent of the total population represented by a species on a numerical and carbon basis shows an inherent inadequacy of the numerical census as a popula- tion monitor. For example, the Coscinodiscus spp. as carbon contributed from 16.7 to 53.4% of that in the >20-iJ.m size fraction (exclusive of the non- setose species which were not monitored), while numerically they composed only from 4.8 to 16.7%. The corresponding means for all stations were about 36% and 11%, respectively. The six most abundant species as carbon {x = 10.67 ug C/liter) compared to their numerical (x = 4,732 cells/liter) importance in the upper 50 m are: Ug CI liter % cellsl liter % Lauderia annulata 2.97 27.8 2,227 47 Coscinodiscus "large species" 1.84 17.2 27 0.6 Ditylum brightwelli 1.38 12.9 184 3.9 Coscinodiscus cf. asteromphalus 0.71 6.7 148 3.1 Biddulphia mobiliensis 0.64 6.0 103 2.2 Thalassiosira rotula 0.61 5.7 454 9.6 For Coscinodiscus (Brenneckella) spp., the means are 3.53 ug C/liter (33.1%) and 458 cells/liter (9.7%). The Coscinodiscus (Brenneckella) spp. and the four other species given above compose 9.13 43 FISHERY BULLETIN: VOL. 73, NO. 1 Table 4. — Mean vertical distribution as cells/liter and as equivalent carbon ( ^ig C/liter) of the >20- jim non-setose size fraction at all stations between 30 March and 6 April 1973 in MESCAL 11 survey area (n = 12 (0 m), n = 13 (10-50 m), n = 9 (75 m)) Depth (m ) X upper Species 0 10 20 30 40 50 75 50 m Actinoptychus undulatus 36 41 48 36 18 13 8 34 0.08 0.10 0.12 0.09 0.04 0.03 0.02 0.081 Biddulphia mobiliensis 177 185 121 66 47 19 5 103 1.10 1.15 0.75 0.41 0.29 0.12 0.03 0.64 Ceratium spp. 66 68 28 5 0 2 0 27 0.14 0.14 0.06 0.01 0 0.005 0 0.057 Coscinodiscus (Brenneckella) 73 67 43 22 17 9 2 38 eccentricus 0.16 0.17 0.11 0.05 0.04 0.02 0.005 0.092 Coscinodiscus of. asteromphaius 286 228 150 96 112 18 33 148 1.38 1.10 0.72 0.46 0.54 0.09 0.16 0.709 Coscinodiscus ? concinnus 49 64 46 13 30 17 16 37 0.24 0.31 0.23 0.06 0.15 0.08 0.08 0.182 Coscinodiscus eccentricus 296 346 254 151 85 62 15 203 0.87 1.01 0.74 0.44 0.25 0.18 0.04 0.593 Coscinodiscus of. granii 11 8 3 10 2 2 0 6 0.23 0.16 0.06 0.20 0.04 0.04 0 0.119 Coscinodiscus ("large species") 26 28 60 16 15 3 5 27 1.79 1.93 4.13 1.10 1.03 0.21 0.34 1.838 2 Coscinodiscus (Brenneckeila) 741 741 556 308 261 111 71 458 4.67 4.68 5.99 2.31 2.05 0.62 0.63 3.53 Ditylum brightwelli 304 346 247 98 73 3 4 184 2.29 2.61 1.86 0.74 0.55 0.02 0.03 1.38 Eucampia cornuta 85 24 28 85 33 6 0 43 0.03 0.01 0.01 0.03 0.01 0 0 0.015 Guinardia flaccida 69 109 66 24 17 2 0 50 0.29 0.46 0.28 0.10 0.07 0.008 0 0.211 Lauderia annulala 3,393 3,761 3,597 1,479 580 46 30 2,227 4.52 5.01 4.79 1.97 0.77 0.06 0.04 2.97 Lithodesmium undulatum 584 667 589 259 211 42 8 408 0.85 0.97 0.85 0.38 0.31 0.06 0.01 0.59 Planktoniella muriformis (colonies) 24 24 27 32 30 12 11 26 Planktoniella sol 506 664 406 195 113 26 11 329 0.34 0.44 0.27 0.13 0.08 0.02 0.007 0.22 Rhizosolenia spp. 24 29 64 11 6 1 1 25 Schroederella delicatula 601 456 705 220 116 47 0 364 0.61 0.46 0.71 0.22 0.12 0.05 0 0.37 Thalassiosira rotula 442 420 969 585 57 34 0 454 0.59 0.56 1.29 0.78 0.08 0.05 0 0.61 2 cells/liter 7,028 7,100 7,424 3,403 1.532 352 138 4.732 £ Mg C/llter 15.51 16.59 16.98 7.17 4.37 1.04 0.77 10.67 /Jig c/liter, or 86% of the mean, and 3,426 cells/ liter (72%). Vertical Distribution The mean vertical distribution of the species numerically and as carbon is given in Table 4. Selected examples of the types of vertical distribu- tion characterizing certain species are given in Figure 2. The standing stock declined sharply between 20 and 30 m; a uniform abundance characterized the upper 20 m. Both numerically and as biomass, the mean population at 30 m was about 45% of that at 20 m (Table 4). Expressed as carbon, and relative to the populations at 20 m, the mean populations at greater depths were only 25% (40 m), 6% (50 m) and 4.5% (75 m). About 62% of the mean carbon content of 533.5 mg C/m^ in the upper 50 m oc- curred in the upper 20 m, where a mean of 16.42 /Jg C/liter is calculated. This pattern in vertical dis- tribution is consistent with the mean compensa- tion depth of about 23 m determined from Secchi disc measurements at 17 stations during this cruise leg. The mean vertical carbon distribution of certain species (Figure 2) illustrates that peak abundance usually occurred in the upper 20 m. The photo- taxic ceratians are most concentrated in the upper 10 m, with a rapid decrease (as percent of mean maximum abundance) with depth. The possibility that the "working distance" vertically within the water column varies between species is suggested by the representative distributions illustrated in Figure 2. The depth at which 50% of the mean maximum abundance occurred ranged from about 20 to 35 m between species, and from 25 to 55 m for the 25% level. Differences in light requirements, particularly that of growth at low intensities, might account for the observed distributions, if a physiological explanation can be applied. How- ever, such distributions can also reflect differences in sinking rates, differential grazing, etc. Thus, while the underlying reasons are obscure, it is evident that the biomass distribution within and 44 SMAYDA: NET PHYTOPLANKTON IN UPWELLING WATERS 0 20 40 60 80 100 % Coscmodiscus Loudena 00 % Figure 2. — The mean vertical distribution at all stations of Actinoptychus undulatus, Ditylum brightwelli, Lauderia annulata , Planktoniella muriformis , and P. sol, and for the com- bined Ceratium and Coscinodiscus species. Abundance is given as percent of the maximum mean abundance for each species presented in Table 4. below the euphotic zone differs between species of phytoplankton. BIOGEOGRAPHICAL COMMENTS Planktoniella muriformis Loeblich et al. (1968) described Coenobiodiscus muriformis as a new genus and species from north San Diego Bay, Calif., where blooms occur, and where it was reported to be in every sample col- lected since its first sighting in July 1966. Cul- tures were also established at 23° to 25°C. This unique, colonial diatom comprised up to 530 cells embedded in a one-cell thick gelatinous matrix which linked the cells in the girdle region. The circular to subcircular colonies have concave- convex shape and can be at least 500 yu m in diame- ter. Round (1972) recently described a similar or- ganism from the harbor at Tema in Ghana, Africa. (Environmental data were not given.) It differed from the San Diego population in the presence of considerably fewer cells per colony and slight mi- crostructural variations. Nonetheless, Round concluded that these taxa were similar, and trans- ferred this species to the genus Planktoniella. This unique organism was conspicuous in the present survey, both in the vertical net tows and quantitative samples along the approximately 700-km track at temperatures ranging from about 11.5° to 16°C. In experiments to be described else- where in greater detail (Smayda in press b), the growth rate for colony increase was 2.9 and 2.0 "doublings" per day at ca. 15° to 18°C. These com- pare with daily colony doubling rates of 1.3 to 1.6 for cultured populations at 23° to 25°C calculated from data presented in Loeblich et al. (1968). The principal value of these data is the indication that active growth occurred under the upwelling condi- tions. Loeblich et al. and Round disagree as to whether all cells in the colony divide to produce a new colony, or whether growth without new col- ony formation can also occur. The maximum recorded abundance of Planktoniella muriformis was 205 colonies/liter at the surface at Station 18. It was very common in the net tows. Thus, given its relative abundance at this time, its noteworthy appearance, and the long-term program of frequent net collections (especially during this time of year) in the coastal waters of southern and Baja California (including this survey area), carried out by Allen and Cupp, their failure to comment in any fashion on its presence is puzzling. Nor is it cited in any way in their periodic species lists for these waters (Cupp 1934; Allen 1938), or for the Gulf of California (Cupp and Allen 1938; Gilbert and Allen 1943), where floristic similarities are evident. Neither does Round (1967) mention it in his recent report on the net phytoplankton in the Gulf of California. Further, only this species and Thalassiosira rotula, of those found during this survey, were not found in the Gulf of Panama (Smayda 1966). Thus, the present observations suggest that Planktoniella muriformis is presently distributed in the Pacific Ocean from San Diego south to Punta Abreojos. But it is uncertain whether its presence and/or distribution in these coastal wa- ters are relatively recent phenomena. Its apparent general rarity in nature and intriguing global dis- tribution (off Baja California and Ghana) are also puzzling, although possibly an artifact of sam- pling. (The recent discovery of another remark- able colonial diatom, Thalassiosira partheneia, in the upwelling waters off Cape Blanc, Africa may also illustrate this latter problem (Schrader 45 1972).) It is also possible that differences in habitus account for this enigma. Loeblich et al. (1968) report that solitary cells present in cultures were unable to form colonies, and under certain conditions colonies reproduced themselves as "clusters of cells" or "a solitary pattern of growth" occurred. If the occurrence of variations in habitus correctly explains these biogeographical issues, then the factors triggering colony formation be- come of interest. Upwelling does not appear to be detrimental in this regard, at least during its initial stages in the survey area. From its size, thickness, and concave-convex shape, it might a priori be presumed that Planktoniella muriformis is particularly well adapted for flotation and has a near-surface niche. However, the equal distribution in colony abun- dance in the upper 40 m is noteworthy, and con- trasts to Planktoniella sol's concentration in the upper 20 m (Figure 2; Table 4). Coscinodiscus (Brenneckella) eccentricus In the Gulf of Panama a unique centric diatom was found identified as Brenneckella sp. (Smayda 1966). It was characterized by an "outer, gelatin- ous" ring surrounding the girdle region in or on which coccolithophorids and other particulate matter were embedded. This organism was also commonplace in the present material (Tables 1,4), and grew actively in one experiment when 2.9 divisions/day were measured (Smayda in press b). Gaarder and Hasle (1961) have reviewed its tax- onomic history, the limited information on its dis- tribution, and the potential relationships between the attached organisms and the host diatom. Based on electron microscopy, they concluded that the two species of Brenneckella described earlier are conspecific with Coscinodiscus eccentricus , a synonomy which is followed here. Nonetheless, it is listed separately as Coscinodiscus (Bren- neckella) eccentricus in Tables 2 and 4 where mean values for the Coscinodiscus spp. are given. Gaarder and Hasle suggest that the attachment of cocolithophorid cells to this diatom is a mere agglutination without any symbiotic significance. While this may be so, the relationship still re- mains intriguing. One may ask why other Coscinodiscus species, or centric diatoms, includ- ing Planktoniella sol characterized by an outer membrane, seemingly are invariably devoid of such epibionts. FISHERY BULLETIN: VOL. 73, NO. 1 DISCUSSION Allen (1924, 1934, 1938) and Cupp (1930, 1934; Cupp and Allen 1938) carried out a long-term sur- vey (approximately 1921 to 1937) of the net phyto- plankton in the coastal, surface waters of southern and Baja California. These data are valuable prin- cipally in their suggestion that the net diatom community in these waters from San Diego to the Gulf of Panama is similar, inclusive of the Gulf of California (Cupp and Allen 1938; Gilbert and Allen 1943). Subsequent quantitative observa- tions in the Gulf of Panama (Smayda 1963, 1966), net collections in the Gulf of California (Round 1967), and the present survey generally support this. Diatoms dominated the net community (Table 4) in response to upwelling, then in its early stages. A red-tide outbreak occurred during mid- April in the survey area following a temporary subsidence of upwelling (Walsh pers. commun.). During the MESCAL I survey of 1972 in this same region the dinoflagellate Gonyaulax polyedra was dominant in March (Walsh et al. 1974). Its abundance then also probably reflects the occur- rence of limited, if any, upwelling. Thus, annual variations in time of inception of upwelling in these waters, as well as variations within a given upwelling cycle, are reflected in the relative im- portance of dinoflagellates vis-a-vis that of dia- toms. An abundance of diatoms will be an indi- cation of nutrient enrichment, as is generally observed in upwelled waters. j The species composition of the diatom commu- " nity is of considerable interest, given the observa- tions of Longhurst et al. (1967). They reported that Coscinodiscus species, especially C. eccentricus, were important dominants of the upwelling com- J munities in June and August 1964 near Mag- " dalena Bay, lying south of the present survey area. Blooms of this genus are of exceptional in- terest. Coscinodiscus, a priori, is not generally expected to occur in great abundance pelagically in unmodified coastal and oceanic water masses. ■ This is generally confirmed by worldwide observa- 1 tions, as reported in the extensive literature on phytoplankton surveys. The periodic, enormous spring blooms of Coscinodiscus concinnus in the North Sea are noteworthy and puzzling (Gr0ntved 1952). This interest in local species composition is sus- tained, given the remarkable occurrence and abundance of the red crab, Pleuroncodes planipes. 46 SMAYDA: NET PHYTOPLANKTON IN UPWELLING WATERS (Longhurst 1968) in these waters. Although it is omnivorous, while herbivorous it grazes on phyto- plankton cells >25 jum (Longhurst et al. 1967), i.e., the size class of Coscinodiscus . Indeed, these authors report active grazing on this genus under experimental conditions, and confirmed during the present study (unpubl.). Therefore, is an abundant Coscinodiscus community significant causally to Pleuroncodes , whose occurrence is a major biotic characteristic of the Baja California upwelling system? Some calculations will be made to evaluate this relationship, and to exam- ine the other questions posed in the Introduction. The maximum observed abundance of all Coscinodiscus (Brenneckella) spp. was 2,243 cells/liter; the mean abundance for all stations in the upper 50 m was 458 cells/liter (Table 4). This meager abundance contrasts with a mean of 4.3 x 10^ cells/liter reported for Coscinodiscus eccentricus by Longhurst et al. (1967). In their study, this concentration represented only 8% of the total community, which was dominated by several Nitzschia species. Coscinodiscus cells of <20 Mm diameter were also not present in bloom concentrations in the present material. Therefore, unlike in Magdalena Bay, this genus was not im- portant numerically, at least during the initial stages of upwelling in the survey area. It remains obscure whether a regional patchi- ness characterizes the abundance of Coscino- discus during upwelling along the coast of Baja California, as for Coscinodiscus asteromphalus in the Gulf of California (Round 1967). Allen and Cupp referred repeatedly to such patchiness in other species in these waters. It is also possible that the Coscinodiscus bloom reported by Long- hurst et al. represents a later state in a species succession. Finally, it might have represented an episodic bloom in response to local, unique factors, rather than reflect a general regional or suc- cessional phenomenon. Nonetheless, the reported summer abundance of Coscinodiscus eccentricus during upwelling in 1964 remains intriguing. The dynamics of Coscinodiscus populations in these waters warrant further study. The dominant (non-setose) species numerically in the >20-/jm fraction was Lauderia annulata, although blooms of Schroederella delicatula and Thalassiosira rotula characterized individual sta- tions (Table 2). The total Coscinodiscus (Brenneckella) spp. represented only about 10% of the mean population numerically, but this rep- resented 33% of the mean carbon; corresponding values (or Lauderia annulata are 47% and 28%, respectively. Thus, although Coscinodiscus was not as abundant as in the Longhurst et al. survey it dominated the >20-jum biomass fraction during MESCAL II. The percent of the total phytoplankton com- munity represented by the >20-jum fraction can be established indirectly from chlorophyll deter- minations made at 10 of the stations for which quantitative >20-/^m phytoplankton counts were also made. The mean concentration (based on 5 depths) in the upper 20 m was 3.46 Mg Chi a/liter. This depth is near the compensation depth; chlorophyll determinations were not made at depths greater than this 1% level. The significant decrease in mean phytoplankton abundance be- tween 20 and 30 m was pointed out previously (Table 4). The mean carbon content of the non- setose fraction >20 jum in the upper 20 m is 16.4 M g/liter. Longhurst et al. (1967) give a mean carbon/ chlorophyll a ratio of 258:1 for their material. This is exceptionally high, and contrasts with a mean (n = 17) of 110:1 characterizing the com- munity dominated by Gonyaulax polyedra during the 1972 MESCAL I survey (Walsh et al. 1974). A mean ratio of 40:1 characterized diatom- dominated communities found throughout the euphotic zone in the Peru Current (Lorenzen 1968). Applying this conversion factor yields a mean carbon content of 138 m g C/liter in the upper 20 m in the present survey. If a similar carbon/chlorophyll ratio characterizes the >20-Mm fraction (it may differ with cell size), then this size group (exclusive of setose species) contributes at least 12% of the viable phytoplankton carbon in the euphotic zone. Lauderia annulata and the Coscinodiscus (Brenneckella) species each contrib- ute 3.5%. The non-setose component of this size grouping would appear to represent only a modest portion of the phytoplankton biomass in the euphotic zone. However, significant diel varia- tions in this component occur, which indicate a high turnover rate. The fluxes and kinetics of this response are considered elsewhere (Smayda in press b). Longhurst et al. (1967) estimated that the graz- ing rate of Pleuroncodes on phytoplankton was 540 liters/day per animal. Its mean abundance during MESCAL II was 1 animal/m^ (Whitledge, pers. commun.), threefold greater than that dur- ing Longhurst and coworkers' study. The total phytoplankton population in the upper 20 m was 47 FISHERY BULLETIN: VOL. 73, NO. 1 276 mg/m^, assuming a dry weight : carbon ratio of 2. Smith et al. (footnote 2) report a mean caloric content of 1,699 cal for Pleuroncodes during MESCAL II, and cite a caloric value of 3,814 cal/g dry wt for diatoms. From these data, a daily caloric ingestion of phytoplankton of 568 cal/m^ within the euphotic zone is calculated, which represents 33% of the total caloric content of the crab. Coscinodiscus would contribute only 1.2% of this daily caloric ingestion and the non-setose com- ponent of the >20-jum size fraction 49c, based on their contributions of 3.5 and 12% , respectively, to the phytoplankton standing stock in the upper 20 m. Even at the maximum growth rates of 3 divisions/day for Coscinodiscus observed during the survey (Smayda in press b) this genus would provide a negligible fraction of the daily caloric intake estimated (or Pleuroncodes. This suggests that the Coscinodiscus population could not then support the Pleuroncodes population; other food sources were necessary. Smith et al. (footnote 2) demonstrated that the respiration rate (as oxycaloric equivalents) of Pleuroncodes is only 3% of the ingestion rates cal- culated using the grazing rate proposed by Long- hurst et al. (1967). Other calculations made by them support their notion that the grazing rate of 540 liters/day is too high, and partly accounts for the discrepancy between rates. Other factors which might contribute to the apparent feeding inefficiency of Pleuroncodes would be high energy losses as fecal material. Longhurst et al. observed the copious production of fecal material packed with Coscinodiscus. While the magnitude of this waste production during MESCAL II can not yet be evaluated, the relative rates of deposition of frustules and organic matter to the sediments when contained in fecal pellets and as free cells can be put into perspective. The sinking rates (n = 24) of fecal pellets pro- duced by freshly collected crabs, and determined on board ship (unpubl.), ranged from 61 to 144 m/h. These rates exceed by 1 to 4 orders of mag- nitude those calculated (Smayda 1970) for the dif- ferent sizes of Coscinodiscus encountered, and that (5.2 m/hr) estimated (Smayda 1969) for the mean zooplankton fecal pellet size (320,000 lum^) collected routinely in the >20-jum fraction. Thus, while Coscinodiscus apparently contributed only a negligible fraction of the daily caloric ingestion of Pleuroncodes, the latter's ingestion and void- ance in fecal material of this genus and other heavily silicified diatoms >20 nm represent a means of exceptionally rapid deposition onto the sea floor. The mean carbon content of 138 /jg/liter during the initial stages of upwelling compares with a mean standing stock of 566 /ug C/liter at 20 sta- tions reported for this region during the Gonyaulax polyedra bloom in March 1972 (from Table 1 in Walsh et al. 1974). The mean carbon content ranged from 23 to 100 /Ug/liter at three stations sampled over a 5-mo period off La Jolla, Calif. (Eppley et al. 1970). The mean concentra- tion during upwelling south of the survey region during June 1964 ranged from 48 yug C/liter (from C/Chl a of 40: 1) to 308 y.g C/liter using data given by Longhurst et al. (1967). However, the data are too limited as yet for any meaningful comparison of regional or seasonal variations in apparent pro- ductivity in these coastal waters. They also indi- cated that the net plankton was usually more abundant in April (upwelling) between Punta Ab- reojos and Punta Eugenia, i.e., in the present sur- vey area (Figure 1). However, quantitative data are needed to confirm this. ACKNOWLEDGMENTS This research was supported by National Sci- ence Foundation Grant GX 33502 as part of the IDOE Coastal Upwelling Ecosystem Analysis program. I wish to express my thanks to Terry f Whitlege, Cruise Leader during this portion of the investigation, and to other members of the scientific party on board then, including James Kelley and John Walsh for helping to make this an informative cruise. Blanche Coyne typed the manuscript and drafted the figures. LITERATURE CITED Allen, W. E. 1924. Observations on surface distribution of marine diatoms of lower California in 1922. Ecology 5:389-392. 1934. Marine plankton diatoms of lower California in 1931. Bot. Gaz. 95:485-492. 1938. The Templeton Crocker Expedition to the Gulf of California in 1935 — the phytoplankton. Trans. Am. Microsc. Soc. 57:328-335. 1945. Vernal distribution of marine plankton diatoms offshore in southern California in 1940. Bull. Scripps Inst. Oceanogr., Univ. Calif 5:335-369. Balech, E. 1960. The changes in the phytoplankton population off the California coast. Calif Coop. Oceanic Fish. Invest. Rep. 7:127-132. Blackburn, M. 1969. Conditions related to upwelling which determine 48 SMAYDA: NET PHYTOPLANKTON IN UPWELLING WATERS distribution of tropical tunas off western Baja California. U.S. Fish Wildl. Serv., Fish. Bull. 68:147-176. Cupp, E. E. 1930. Quantitative studies of miscellaneous series of surface catches of marine diatoms and dinoflagellates taken between Seattle and the Canal Zone from 1924 to 1928. Trans. Am. Microsc. Soc. 49:238-245. 1934. Analysis of marine plankton diatom collections taken from the Canal Zone to California during March, 1933. Trans. Am. Microsc. Soc. 53:22-29. Cupp, E. E., AND W. E. Allen. 1938. Plankton diatoms of the Gulf of California obtained by Allan Hancock Pacific Expedition of 1937. Allan Hancock Found. Pac. Exped. 3:61-99. DuRBiN, E. G., R. W. Krawiec, and T. J. Smayda. In press. Seasonal studies on the relative importance of different size fractions of phytoplankton in Narragansett Bay. Mar. Biol. (Berl.) Eppley, R. W., F. M. H. Reid, and J. D. H. Strickland. 1970. The ecology of the plankton off La Jolla, California, in the period April through September, 1967. Part III. Estimates of phytoplankton crop size, growth rate, and primary production. Bull. Scripps Inst. Oceanogr., Univ. Calif. 17:33-42. Gaarder, K. R., and G. R. Hasle. 1961. On the assumed symbiosis between diatoms and coccolithophorids in Brenneckella. Nytt Mag. Bot. 9:145-149. Gilbert, J. Y., and W. E. Allen. 1943. The phytoplankton of the Gulf of California obtained by the "E. W. SCRIPPS" in 1939 and 1940. J. Mar. Res. 5:89-110. Gr0ntved, J. 1952. Investigations on the phytoplankton in the southern North Sea in May 1947. [Dan. summ.] Medd. Komm. Dan. Fisk. Havunders., Ser. Plankton 5(5): 1-49. LoEBLiCH, A. R., Ill, W. W. Wight, and W. M. Darley. 1968. A unique colonial marine centric diatom Coeno- biodiscus muriformis gen. et sp. nov. J. Phycol. 4:23-29. LONGHURST, A. R. 1968. The biology of mass occurrences of galatheid crustaceans and their utilization as a fisheries resource. FAO (Food Agric. Organ. U.N.) Fish. Rep. 57:95-110. LONGHURST, A. R., C. J. LORENZEN, AND W. H. ThOMAS. 1967. The role of pelagic crabs in the grazing of phyto- plankton off Baja California. Ecology 48:190-200. LORENZEN, C. J. 1968. Carbon/chlorophyll relationships in an upwelling area. Limnol. Oceanogr. 13:202-204. Round, F. E. 1967. The phytoplankton of the Gulf of California. Part I. Its composition, distribution and contribution to the sediments. J. Exp. Mar. Biol. Ecol. 1:76-97. 1968. The phytoplankton of the Gulf of California. Part II. The distribution of phytoplanktonic diatoms in cores. J. Exp. Mar. Biol. Ecol. 2:64-86. 1972. Some observations on colonies and ultrastructure of the frustule of Coenobiodiscus muriformis and its transfer to Planktoniella. J. Phycol. 8:222-231. Schiller, J. 1937. Dinoflagellatae (Peridinieae) Zweiter Teil. Raben- horst Kryptogamen-Flora 10(3), 590 p. Schrader, H. J. 1972. Thalassiosira partheneia, eine neue Gallertlager bildende zentrale Diatomee. Meteor Forsch.-Ergebnisse, Reihe D 10:58-64. Smayda, T. J. 1963. A quantitative analysis of the phytoplankton of the Gulf of Panama. I. Results of the regional phytoplankton surveys during July and November, 1957 and March, 1958. Bull. Inter- Am. Trop. Tuna Comm. 7:191-253. 1966. A quantitative analysis of the phytoplankton of the Gulf of Panama. HI. General ecological conditions and the phytoplankton dynamics at 8°45'N, 79°23'W from November 1954 to May 1957. Bull. Inter-Am. Trop. Tuna Comm. 11:354-612. 1969. Some measurements of the sinking rate of fecal pellets. Limnol. Oceanogr. 14:621-625. 1970. The suspension and sinking of phs^toplankton in the sea. Oceanogr. Mar. Biol. Annu. Rev. 8:353-414. In press a. Phased cell division in natural populations of the marine diatom Ditylum brightwelli , and the pos- sible significance of diel phytoplankton behavior in the sea. Deep-Sea Res. In press b. Dynamics of aCoscinodiscus population during two days in an upwelling area. II. Influence of growth rates, sinking rates and grazing on diel variations. Limnol. Oceanogr. Strathmann, R. R. 1967. Estimating the organic carbon content of phyto- plankton from cell volume or plasma volume. Limnol. Oceanogr. 12:411-418. VON Stosch, H. a. 1964. Zum Problem der sexuellen Fortpflanzung in der Peridineengattung Ceratium. [Engl, abstr.] Helgo- lander wiss. Meeresunters. 10:140-152. Walsh, J. J., J. C. Kelley, T. E. Whitledge, J. J. MacIssac, and S. a. Huntsman. 1974. Spin-up of the Baja California upwelling eco- system. Limnol. Oceanogr. 19:553-572. 49 FISHERY BULLETIN: VOL. 73, NO. 1 Appendix Table 1. — List of phytoplankton taxa identified to species found in net tows and in >20-Mni size fraction. BACILLARIOPHYCEAE Actinocyclus octonarius Ehrenberg Actinoptychus undulatus (Bailey) Ralfs Asterionella japonica Castracane Asterolampra marylandica Ehrenberg Asteromphalus heptactis (Br6bisson) Ralfs Bacteriastrum hyalinum Lauder Biddulphia mobiliensis Bailey Biddulphia cf. sinensis Greville Cerataulina pelagica (Cleve) Hendey Chaetoceros affinis Lauder Ch. cf. costatus Pavillard Ch. curvisetus Cleve Ch. debilis Cleve Ch. decipiens Cleve Ch. didymus Ehrenberg Ch. peruvianas Brightwell Ch. socialis Lauder Ch. subsecundus (Grunow) Hustedt Corethron pelagicum Brun Coscinodiscus cf. asteromphalus Ehrenberg C. centralis var. pacifica Gran et Angst C. concinnus W. Smith C. curvatulus Grunow C. eccentricus Ehrenberg C. granii Gough C. perforatus var. pavillardi (Forti) Hustedt C. radiatus Ehrenberg Coscinodiscus (Brenneckella) eccentricus (Lohmann) Gaarder et Hasle Ditylum brightwelli (West) Grunow/ cf. Ethmodiscus rex (Rattray) Hendey Eucampia cornuta (Cleve) Grunow Guinardia flaccida (Castracane) Peragallo Hemidiscus cuneiformis Wallich Lauderia annulata Cleve Leptocylindrus danicus Cleve BACILLARIOPHYCEAE— Cont. Lithodesmium undulatum Ehrenberg Parana sulcata (Ehrenberg) Cleve Planktoniella muriformis (Loeblich III, Wight et Darley) Round Planktoniella sol (Wallich) Schutt Rhizosolenia alata Brightwell R. bergoni H. Peragallo R. calcar avis M Schultze R. delicatula Cleve R. imbricata Mar. shrubsolei (Cleve) Schroder R. robustum Norman R. stolterfothii H. Peragallo Roperia tessellata (Roper) Grunow Shroederella delicatula (Peragallo) Pavillard Skeletonema costatum (Greville) Cleve Stephanopyxis turris (Greville) Ralfs Thalassionema nitzschioides (Grunow) Hustedt Thalassiosira rotula Meunier Thalassiothrix frauenfeldii Grunow T. longissinna Cleve et Grunow T. mediterranea var. pacifica Cupp DINOPHYCEAE Ceratium furca (Ehrenberg) Claparede et Lachmann Ceratium fusus (Ehrenberg) Dujardin Dinophysis miles Cleve Gonyaulax cf. polyedra Stein Noctiluca scintillans (Macartney) Kofoid et Swezy Pyrocystis cf. lunula SchiJtt Pyrophacus horologicum Stein CHRYSOPHYCEAE Distephanus speculum (Ehrenberg) Haeckel HAPTOPHYCEAE Phaeocystis poucheti (Hariot) Lagerheim PRASINOPHYCEAE cf. Halosphaera viridis Schmitz J 50 OPTIMUM ECONOMIC YIELD OF AN INTERNATIONALLY UTILIZED COMMON PROPERTY RESOURCE* Lee G. Anderson'^ ABSTRACT The exploitation of a common property resource, specifically a fishery, by nationals of two countries is discussed using a simple general equilibrium analysis. The interdependence of their production possibility curves is used to describe the open-access equilibrium yield, local maximum economic yields, and a true international maximum economic yield. Finally a complete description of the conditions necessary for this international maximum economic yield and why they are different from those in a national fishery is presented. The purpose of this paper is to analyse, using a simple general equilibrium model, the problem of the allocation of resources where common prop- erty or open access exists for some of them. The common property or open-access resource will be a fish stock. The economics of fisheries has been quite extensively developed. See for example Gor- don (1954), Scott (1955), Crutchfield and Zellner (1962), Turvey (1964), Crutchfield (1965), Christy and Scott (1965), Smith (1969), Copes (1970), Scott and Southey (1970), Gould (1972), Southey (1972), and Anderson (1973). The present paper follows Scott and Southey and uses a production possibil- ity (PP) curve model which takes into direct ac- count all the resources of the economy and not just the fishery. This change in focus is especially use- ful for analysing economic aspects of international use of common property resources, a problem that has long been recognized but which has received very little treatment to date. The following quote from Christy and Scott (1965:223) summarizes the problem fairly well: "Two countries contemplating the same fishery may rightly make different choices about the intensity and combination of fishing activities .... These different valuations are ulti- mately the result of the obstacles to the movement of factors from one economy to another. More directly, they result from differences in population, national income, and tastes. It is a commonplace of the theory of comparative costs that the same industry may use a different technique in each country, de- pending on the structure of wages and prices in each place. But 'This study was sponsored by the University of Miami's Sea Grant Institutional Program which is part of the National Sea Grant Program administered by the National Oceanic and At- mospheric Administration of the U.S. Department of Commerce. ^Present address: College of Marine Studies, University of Delaware, Newark, DE 19711. it has never, to our knowledge, been pointed out that the ocean is the main locale where these structures clash . . . .Of course, it is possible to exaggerate these discrepancies. Forces outside the fisheries tend to bring the national wage and price struc- ture into line, through the movement of goods and the sale of services. And within the fishery itself the increasing inter- national trade in this equipment, all tend to press toward a uniform set of labor-capital-fish price-ratios." The model presented will allow a more formal analysis of these and other problems. The first section of the paper describes a one country model of the economics of fisheries from a general equilibrium point of view. Results identi- cal to the earlier works are derived as a starting point for discussion. The second section expands the model to consider two nations both having access to the same fish stock and describes the conditions necessary for an international open- access equilibrium yield, for local maximum economic yields (MEY), and for a true interna- tional ME Y. The third section describes the condi- tions for an international MEY in more detail and shows the ways in which the countries can go about achieving them. Throughout the analysis is static. Consider a country with a specified amount of resources, a given technology, and exclusive use (either through default or international law) of a fish stock. Using its resources, it can either pro- duce manufactured goods (M) or fishing effort (E) which can be applied to the fish stock to catch fish. Let the implicit function for the PP curve between M and E be: Manuscript accepted May 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. 51 FISHERY BULLETIN: VOL. 73, NO. 1 G {E, M) = 0. (1) Assume that it is quasi-concave so that there will be a concave transformation curve between E and M. Let the sustained yield curve of the fish stock (i.e. the production function) be expressed as:^ F{E) ^ aE - bE^ (2) Using this equation assumes that the fish stock vidll always be in a biologic equilibrium. F will increase until E is equal to -^ and vdll thereafter Zo decrease. F will be zero whenE = 0 and whenE = ^. As long as the maximum amount of £ possible is greater than ^ but less than-?-, then the PP curve for M and F will be similar to the solid one in Figure 1. (Ignore for the moment the dotted one.) The slope of the curve is: dF dF dE G, dM dE m =-^-- 2^^) ^ (3) where Gj is the derivative of G with respect to its first argument, etc. Fish output will be at a max- imum when E equals ^, not when all of the re- sources are used in the production of E. As long as the marginal productivity of E in fishing is negative, the PP curve will have a positive slope. Switching resources out of effort and into manu- facturing will actually increase both F and M. Where E's marginal productivity in F is positive, the PP curve will have its normal negative slope. Because both -^ and (a - 2bE) increase as M increases (i.e. as E decreases), the curve wall be concave to the origin. Also assume that there is a linearly homogeneous social utility function of the form U = U (F, M). (4) As pointed out in the literature cited above (see especially Turvey 1964 and Scott and Southey 1970), as long as no one regulates entry into the fishing industry, profit maximizing individuals will continue to produce or buy E as long as the 'The sustained yield curve is the relationship between the amount of effort expended and the amount of fish that will be captured period after period. The particular expression here follows Schaefer (1957). Although other expressions have been discussed recently (see the papers by Southey and Gould cited above), Expression ( 1) is descriptive enough to capture the essen- tials of the argument. OM Figure 1. — The solid concave curve is the production possibility curve and the set of convex curves are the community indiffer- ence curves. Open-access equilibrium will occur at B, maximum sustainable yield at H, and maximum economic yield at D. In the two country model, a decrease in fishing effort in the other country will shift the production possibility curve to the dotted one. value of the average catch per unit of £^ is greater than the price of effort. The effects of this are as follows. If £■ and M are produced in pure competi- tion,then- ^ = % t^E dM Equilibrium will occur in the open-access fishery when P^-^ equals P^ ; that is when the average return to effort equals its cost. [Smith (1969) has shown that vmder certain cir- cumstances, the fishery wdll not reach an equilib- rium. For the moment let us ignore this possibility although its effects will be discussed briefly below.] It can be shown therefore that with an open-access fishery and pure competition in the production of E and M, producers will arrange their production such that for any given price ratio the following condition will hold: M FIE dMIdE (5) Maximum consumer welfare occurs where the slope of the social indifference curve is equal to the price ratio. That is where M Therefore a general equilibrium in the production and the consumption sectors of the economy will occur when 52 ANDERSON: OPTIMUM ECONOMIC YIELD U2 M FIE Pj, dMIdE (6) Conditions for the maximization of social welfare, however, are: M dFldE ^ dF_ dMIdE " dU (7) dF An expression for — — - is given in (3) and dM {FIE)l{dMldE) can be expressed as: The ratio {FIE)l(dMldE) = -ia FIE m^ (8) will increase in absolute size as dMIdE M increases, and because of the assumption that the maximum E is less than alb, it will always be negative, even when the slope of the PP curve is positive. It can be seen that when they are both negative, this ratio will be larger in absolute size than the slope of the PP curve at that point; i.e. it will have a steeper slope. The small lines on the PP FIE dMIdE at curve in Figure 1 represent the ratio that point. In terms of Figure 1, open-access equilibrium will occur at point B where the slope of the indif- ference curve as it intersects the PP curve is equal to the ratio of FIE at that point.^ The social dMIdE optimum is at point D where the indifference curve is just tangent to the PP curve. The common property or open-access equilibrium will always be to the left of the optimal point; therefore with open access, too many resources will be allocated to F under the market system. It is even possible that the market equilibrium will occur in the posi- tive sloped segment of the PP curve. By way of comparing the present analysis with the standard one, point H on Figure 1 is the point of maximum sustained yield for a fishery and point D is the MEY. The latter point has less fish but more manufactured goods than the former (and may even have less fish than the point where the unregulated fishery wall operate). At MEY, ■•As Scott and Southey (1970) point out, if there are increasing returns to scale and if the social utility function is not linearly homogeneous, it is possible that there may be multiple equilib- ria. I have ignored that complication for purposes of this paper. however, no fish is produced unless its value is greater than its opportunity cost. Although MEY in the traditional literature refers to a specified amount of fish production, it assumes that the resources not in fishing are used efficiently in the production of other goods. Describing the model in terms of a PP curve makes this explicit. Through proper regulation, the country can move to MEY. This could involve a ceiling on the amount of fishing effort allowed or the granting of property rights to the fishery to certain individu- als. The former has been tried but usually by means of decreasing efficiency rather than by shifting resources to other types of production, and the latter can lead to monopoly or oligopoly unless the property rights are distributed widely or there are other fish stocks that can provide the neces- sary competition. If the government only allows a units of effort, where a is less than the open-access amount of effort, and then distributes the rights to this number of units among a large enough group such that there is still pure competition in the market for both effort and fish, these people will be earn- ing a rent per period, R, of PpF(a) - P^a where Fia) is the amount of fish caught by a units of effort. Unless reductions in effort have perverse effects on price, average catch, or cost of effort, this rent will be positive. See Anderson (1973:513). The optimal amount of effort is where the total amount of rent is a maximum (Christy and Scott 1965:8). By using the standard mathematical pro- cedure it can be shown that the first order condi- tion for 7? to be a maximum is: p dF _p Under the above assumptions, the open-access problem of the fishery has been solved in a way that keeps pure competition in the production of M and £. Therefore -P^^IP^ is equal to dEldM and so maximization of the rent of the fishery will guarantee that M dFldE dF Pp dMIdE dM ^^^ This will mean that the conditions for the maximi- zation of social welfare, expressed in (7) above, will hold. Therefore a policy that maximizes the rent from the fishery also maximizes social wel- fare. In summary, a country with exclusive rights to an open-access fishery wall operate inefficiently as 53 FISHERY BULLETIN: VOL. 73, NO. 1 long as there is no regulation of fishing effort. This will be because as long as the average returns to fishing are greater than the price of effort, private decision makers will continue to demand E. Also since E andF are directly related, there is always a direct relationship between Pg and Pp . II Now to turn to the case of more than one country exploiting the same fish stock, analysis of this is made very difficult by a variety of intriguing prob- lems. For instance, technology may be so different in the two countries that it is very hard to find a common measure of fishing effort, tastes may be such that one country prefers small fish while the other prefers large ones and yet the sustained yield curve is dependent on the size of catch, each country may be using other criteria for harvesting the fish; for example, one may look at it as a place to put unemployed labor, or as a source of earning foreign exchange. For purposes of discussion these intricacies will not be considered. Assume that two countries, country X and coun- try Y, both with specified production capacities (G^ (Ex, M^) = 0 and G^ {Ey, My) = 0) and lin- early homogeneous community welfare functions (U^ = U^ (Fx, Mx) and U^ = U^ (Fy, My) ,are the exclusive users of a fish stock with the sus- tainable yield curve (2) above. Since a unit of effort in country X, (Ex), is identical to one in Y, (Ey), the sustained yield curve can be expressed as: F(Ey,Ex) - aiEx + Ey) - h{Ex + Eyf- As before the total catch from the fishery will reach a maximum when E^ plus Ey is equal to hx and will fall to zero if total effort gets as large a asr- o The catch of one country v^ll be in proportion to its effort in relation to total effort, therefore: [• F^ (E^ ,E^) E, Ex + E-\ aiEx +Ey) - biEx + Ey) This can be simplified to: F^ wdll reach a maximum when Ex equals and will fall to zero if it gets as large as a bE, 26 a -bE, The equation for Fy is analogous. The amount of fish that country X can catch using a specified amount ofE^ depends upon how much Ey country Y is producing and using. Simi- larly the catch of country Y depends upon the amount of Ex used by country X. Therefore, the shape and position of each country's PP curve for F and M is dependent upon the amount of E the other country uses. Let the two PP curves in Fig- ure 1 be two possible ones for country X. The solid one is for the larger level of Ey Note that the lower curve gets further away from the higher one P\E as My decreases. This is because -— ^ , the vertical ^ dEy displacement of the curve due to a change in effort in country Y, is equal to -bE^. Therefore, the higher the level of Ex , that is the lower the level of M^, the greater will be the vertical displacement. The maximum amount of F-^ will be at a higher amount of Mx^ (a lower amount of F^, ) because F^ is a maximum when Ex is equal to — ^-r — -. 2.0 Using this two country model let us consider the implications of three types of exploitation: 1) open access in both countries, 2) local MEY in both countries, and 3) a true international MEY. From the above description, it can be seen that the shape and position of the PP curve for M andF in each country is dependent upon the level of effort used in the other. Therefore the open-access free market equilibrium in each country will de- pend upon the level of effort used in the other. The mathematical condition for an international open-access equilibrium is the following set of simultaneous equations: X Country X ul Country Y — ^ = dMy/dEy F ,E Yl Y dMy/dEy (11a) (lib) aEx — bEx bE^Ey (10) This simply states that the open-access condition for each country (see Equation (6)) must hold in both simultaneously. In terms of Figure 1, each country must be operating at a point such as B. 54 ANDERSON: OPTIMUM ECONOMIC YIELD Note however, that in country X, average catch (F^ lE^ ) is a function of both Ex and Ey . Therefore an equilibrium in country X can be reached only for a given level o^Ey, (i.e. for a given PP curve). Similarly an equilibrium in country Y is possible only for a given level of £'y . Therefore an interna- tional equilibrium is possible only at that combination(s) of E^ and Ey where Equations (11a) and (lib) both hold simultaneously. If free international trade between these coun- tries is possible, the price ratios in both countries will be equalized, and so at the equilibrium, the marginal rates of substitution^ — ijwill also be equal. Therefore the following condition will hold: m Uo ^xl^x Uf Uj dM^ldEx dMv/dE. (12) Graphically the international trade case can be interpreted as follows. For a given level of E produced in the other country, each country will produce at that point on the PP curve where the trade price ratio is equal to ^ '^ . It will then dM/dE trade along the price ratio line until welfare is maximized. Consider a country that would oper- ate under autarky at point B in Figure 1. Under our assumptions the location of the PP curve is related to the amount of E being produced in the p other country. If trade opens up with a lower _M , the production point will move to A, but the con- sumption point will be at C because of imports of M and exports of F. From this it can be concluded that for each level of E produced in the other p country, a decrease in -^, i.e. a relative increase "f in Pp, will increase the amount of E produced locally. p As a sidelight notice that the decrease in -^ actually decreased the welfare of the fish export- ing country described in Figure 1. Trade allowed for a further misallocation of resources due to an expanding market for fish to such an extent that welfare fell. Of course, if the price line through A intersected the indifference curve through B, then welfare would have been increased in spite of the harmful effects. To be precise it should be noted that in the general equilibrium analysis, the amount of E produced by the other country will fall in most cases which will shift the PP curve out and may cause welfare to increase enough to over- come the initial loss. On the other hand, increases p in p^ brought about by trade will improve the F allocation of resources and always increase wel- fare initially; however, the increase in E in the other country will have the opposite effect on wel- fare. So whether the country exports or imports fish, changes in the terms of trade may decrease welfare depending upon the direction and mag- nitudes of the changes caused by these two factors. Equation (7) above states the condition for the maximization of social welfare (i.e. MEY) in the one country case. With free international trade, if both countries attempt to maximize welfare given the level of effort used in the other country, the condition for an international equilibrium is: U2 _ Ui _ dF^/dEx _ bFy/bEy jjx jjY dMxIdEx dMyldEy The last two terms can be simplified to 9Fv bMx (13) and respectively. These will be recognized as the ■dMy slopes of the PP curves of the two countries. What this condition states is that for a local MEY, the marginal rate of substitution between M andF in each country must equal each other and they must also equal the internal marginal rate of transfor- mation between M and F given the level of effort in the other country. In terms of Figure 1, each country will be operating at a point such as D, where the slope of the social indifference curve is equal to the slope of the existing PP curve. Notice that in equation (13), -^rr^ and -r^ are both par- A Y tially determined by the level of effort in the other country, so that here again the equilibrium com- bination of Ex and Ey must be simultaneously determined. One main purpose of this paper is to describe the necessary condition for an international MEY. It is important to note at this time that they are different from Equation (13), the conditions of local ME Y's given the level of effort in the other country. Since the level of effort in each country affects the PP curve, and hence potential welfare, in both countries, the maximizing conditions 55 must take this into account. With free inter- national trade, these conditions are:^ ul dF^ dFy dEx ' dEx dFy ^ dFx dEy dEy ul dMx dMy (14) dEx dE, *This condition can be derived in the following manner. With international trade, the community welfare fiinctions become U X _ U^ [F;f (£;f , £y ) + F^ , Afx + Mf] and U^ =U^' [Fy(Ey,Ey)-Fj,My -Mt] where Fj and Mj are the amounts of F and M respectively that are traded. If we wish to maximize the welfare of country X subject to a specified amount in country Y and to the productive capacities, we get the following Lagrangian function. L =r/^ + X,([/^ - t/^) + KiG^iExMx) +X3G^(£y,My). The first order conditions for a maximum (using the normal notation for derivatives) are: (a) 3L _ ,rX dFx . -v jjY ^Fy . /jX _ 0 (b) (0 a^y- =U^+ XaG^ - 0 dL 3Fv = ttX U dFx 1 3Fv ^if^rll; + Xgcr =0 (d) fiT- = A,[/r + A.G,^ =0 dsr^ (e) (f) dL dL dL dFr ^•^2 ■3'-'2 f/f + XiU\ = 0 U^ + XiU2 = 0. Note that Conditions (a) and (c) show that a change in the level of effort in one country has a direct effect on the level of welfare on the other. For this reason the Pareto conditions for an interna- tional optimum are different than in the standard case. Solving (e) for X 1 substituting that expression in (a) and then dividing (b) by (a) yields Ul dFy dFy Gf Similarly substituting the value of Xi into (c) and (d) and then dividing (d) by (c) yields Ul u^ dFy ^ dF^ dEy hEy G\ [/2 ul Since from (e) and (f) it can be shown that ^ = "" — , and by G ^ dM G ^ dM ^ ' ^ ' definition = ,„^ and = je^ — , it can be shown that gI °*'X Gi °^'^ Condition (12) holds. FISHERY BULLETIN: VOL. 73, NO. 1 Alternatively this condition can be written as: f/r Ul (-) dFx + ( + ) dFy (-) bFy + ( + ) dFx Ul U^ dM^ dMx ^^Y ^My (14') Expression (14) is useful for comparisons with the open-access free market international equilib- rium conditions in (12) and with the local MEY condition in (13), while Expression (14') is useful for tying the analysis to the PP curve. In words these conditions state that the margin- al rate of substitution for M and F and a special type of marginal rate of transformation (MRT) in both countries must equal each other. The margin- al rate of transformation is special in that it con- siders the effect on fish production in both coun- tries, of a change in manufacturing in only one. To be more precise a "socially optimal" interna- tional policy should guarantee that neither coun- try expand their fishing effort unless the value of the extra yield, regardless of who catches it, is equal to the value of the extra M that must be fore- gone. That is country X should compare the oppor- tunity value of producing effort with its effect on local catch ( J^) and with its effect on country Y's catch (—^) . The same restriction must be placed on country Y's fishing industry also. It is important to stress at this point that these international MEY conditions were derived by maximizing the level of welfare in one country while specifying a certain level in the other. That is, an initial distribution of the fishery is essential before the maximizing conditions for an interna- tional MEY can be utilized. This same condition will hold at many combinations of ^^^ ^^id Ey de- pending upon how the wealth of fishery is distrib- uted. This is one of the major differences between a national MEY and an international MEY. The importance of the beginning distribution will be discussed in greater detail in Section EI. It can be shown from the equations for Fx and 9^x , ^Fy , dFy , dFx , ,, ^ equals ^r^ + W^ and that Fy that ^^ + X dEx dE^ dE, F^IE^ equals FylEy. Therefore in both the open- access equilibrium (Condition 12) and at any true international optimum point (Condition 14), dM^ldEx must equal dMy /dEy. That is, the real cost of producing fishing effort will be the same in both countries. The difference is that only in the 56 ANDERSON: OPTIMUM ECONOMIC YIELD latter is the proper amount of it produced. The equalizing mechanism in both cases is the trade in fish which is indirect trade in effort. Figure 2 depicts the international MEY situa- tion in terms of the PP curve of both countries. Expression (14') says that the absolute value of the slope of the indifference curves in both coun- tries (~ T7^) must be less than the absolute value of the slope of their existing PP curves at the point of operation (—-). That is at the equilibrium point, the slope of the indifference curve must be less steep than the slope of the PP curve. Therefore the slope of the price ratio line must also be less steep than the slope of the PP curve. What this means is that both countries must produce less fish than they would under normal free market conditions given the relative cost of producing F and M. The reason for this is that they must take into account the effect of their output levels on the production of fish in the other country. In the diagram the regulated price ratio common to both countries is represented by the two straight lines. Country X, producing at point A and consuming at point B, is importing Mj, units of M and exporting F7. units of F. Country Y, producing at point A' and consum- ing at point B', is doing the reverse. Since at the equilibrium, producers in both countries are bas- ing the production decision on the same price ratio, and since dMy dM dE X dE - , there will be no F„A COUNTRY X COUNTRY Y Figure 2. — In the two country case, the international maximum economic yield can be represented by the countries producing at A and A' and consuming at B and B', the difference being made up by international trade. The exact relationship between the slope of the indifference curves and the production possibility curves is expressed in Equations (14) and (14'). balance of payments problem; i.e. the value ofF traded will equal the value of M traded. Two technical points regarding this diagram should be pointed out. First, since there are inter- national interdependencies involved, operation at the international MEY requires government reg- ulation. Some form of taxes or other means of control will be necessary in each country to keep producers operating where the price ratio to con- sumers, as represented by the slope of the indiffer- ence curve, is different than the ratio of marginal costs of production, as represented by the slope of the PP curve. Second, it may seem strange that country X, the importer offish, is consuming at a point inside its existing PP curve. (If the indiffer- ence curve for country Y through point B' inter- sects the PP curve, that country will also be oper- ating at a point where its welfare is not as large as it might be given its existing PP curve.) Would it not be to its advantage to stop trading and ex- pand its own fishing by moving up its PP curve? In answering this question it must be remembered that the only reason country X's PP curve is as high as it is, is that country Y has reduced its level of effort. Only if country Y were foolish enough to keep its level of effort the same regard- less of country X's behavior would the latter bene- fit from an increase in effort. It would gain wel- fare while country Y would lose. This discussion points out, however, that proper management of international fisheries will be difficult to en- force because one or both of the countries involved will be motivated to increase effort from the op- timal point. So far three distinguishable points on each PP curve can be identified: the open-access equilib- rium point (where the slope of the indifference curve, or the international price line, as it inter- sects the PP curve equals F/E ;, i.e. point B in dM/dE Figure 1); the local MEY optima given the level of effort in the other country (where the slope of the indifference curve or the international price line is equal to — — , i.e. point D in Figure 1); and the point where the country contributes to an international MEY given the level of F produced abroad, i.e. at point A or A' in Figure 2. With regard to the latter, only if both countries are operating in this fash- ion, is it a true international MEY, where the value of the net increase in fish production by the marginal unit of effort, regardless of its origin. 57 FISHERY BULLETIN: VOL. 73, NO. 1 is just equal to the value of the resultant decrease in the production of M. As a sidelight it is interesting to note that if one country unilaterally adopts a local optimum regu- lation policy given the level of effort in the other country, at the new equilibrium it will be using less effort and in most cases the other country v^ll react to this by increasing their level of effort. Therefore, while the decrease in effort will in- crease its level of welfare (it vn\l move from point B to point D in Figure 1), the increase in effort by the other country will shift the PP curve toward the origin, and this vdll reduce the gains. It is even possible that the shift of the PP curve could be large enough that at the new equilibrium the country actually loses welfare. This has interesting implications for cases where international cooperation in fisheries man- agement does not exist. National regulation policies must be derived taking into account the reaction of other countries to specific actions. Each country wdll have to know how the other will react to a change in its level of effort. Taking this into account, it should only reduce its own effort (i.e. transfer resources from producing effort into the production of M) as long as the resultant increase in welfare is greater than the decline due to any possible increase in foreign fishing.^ If these reac- tions are not known, the determination of the proper regulation program will require some sort of game theory approach. In conclusion it should be pointed out that sim- ply because it is possible to list the conditions that are necessary for a certain type of equilibrium to exist does not mean that it will in fact exist. As Smith (1969) has pointed out, a fishery will reach a bionomic equilibrium only if certain relationships exist between the growth rate of the fish stock and the rate at which effort enters and leaves the *In formal mathematical terms the country must maximize welfare subject to its production constraint knowing that the equilibrium level of effort in the other country is a function of its own effort. The proper Lagrangian for country X and its first order conditions are: L, = U'^IFx{Ex,Ey{Ex)),Mx] + X^G'^iExMx) dLi 3£x Y , dry dry dEv Y = U2 + ^lG2 = 0. The first order condition with respect to Ex takes into account the total effect on the amount offish caught by a change in effort. There is the direct change in catch and the indirect effect caused by a change in the level of effort in country Y. fishery (either because of market forces or reg- ulatory decree). As pointed out earlier, however, the present analysis is static and will ignore these complications. Ill It will prove useful to view the problem from a different angle. There are two countries each with its own productive capacity and preference func- tion, and between them they share an open-access fishery. Given this information, it is possible to construct a welfare possibility curve for the two countries (Figure 3). Any point on the curve is the mgiximum amount of welfare that can be obtained for one country at the level of welfare specified for the other country given the productive capacities of both countries and the sustained yield curve of the fishery. At any point on the curve. Condition (14) holds. Therefore, at each point there is an international MEY from the fishery since in all cases the value of the last fish caught vdll be worth its opportunity cost. As is well known, there is no way of choosing one point on the curve from another. To digress a moment, if there were no open- access resources or other market imperfections, the two countries through market-directed pro- duction and trade will end up at a point on that possibility curve. If they each operated indepen- dently, they could obtain a certain amount of wel- fare, say the amounts represented by point A. Under free market conditions, each would be motivated to change its output combination and then trade such that both would be better off at a point such as B. Point B is not inherently superior to any other point on the curve. It is merely the point where given the productive capacities and the preferences of the two countries, they will op- erate under the conditions of a free international market. At that point no country can be made better off v^thout making the other one worse off. If for some reason there was a redistribution of productive capacity, the final equilibrium would still be on the curve but at a different point than B. Now to turn back to the case of the open-access fishery, if neither country exploits the fishery and they do not engage in trade, then operating inde- pendently, each would be able to obtain a certain amount of welfare. Again let this point be rep- resented by A in Figure 3. If free trade is intro- duced and if both countries begin to exploit the fishery taking into account the effect of their effort 58 ANDERSON: OPTIMUM ECONOMIC YIELD i>Wx Figure 3. — Each point on the curve represents a distribution of the fishery where one country cannot be made better off without hurting the other. B represents the point where it is distributed on the basis of abiUty to harvest fish. C represents the distribu- tion that is obtained by open-access exploitation. While both countries can benefit from changes from this point, note that in this case a move to the "ability" distribution at B represents a decrease in the welfare of country Y. on the catch in the other country, a point such as B on the possibility curve will be reached. The wealth from the fishery will have been distributed between the two countries on their ability to pro- duce the effort to harvest it. In fact, if the cost of effort was always less in one country, then at the MEY point, that country would be doing all the fishing and gaining all the wealth from the fish stock. The other country would gain from trade in goods but not from the fishery itself There is noth- ing inherently superior about point B, however. There does not appear to be a moral argument that one country deserves the wealth from an interna- tional common property resource simply because it has a comparative advantage in the ability to capture it. Under open-access conditions, the two countries will operate somewhere inside the welfare possi- bility curve, say at point C. This point is analogous to the solution of Equations (11a) and (lib). It is possible for both countries to increase their wel- fare by moving to a point such as D. Just how these gains can be obtained is discussed in detail below. But for now notice that in the case depicted here, if the countries are forced to move to point B (i.e. the point where the wealth from the fishery is distrib- uted on the basis of ability to produce effort), country Y will suffer a decrease in welfare. This will not always be the case but will depend upon the position of C relative to that of B. The point to be made from all this is that dis- tribution is a critical part of determining the makeup on an international MEY. It is important to separate who obtains the wealth from the fishery from who harvests the fish. When the two are linked together, economic efficiency can be obtained only if the fishery is distributed accord- ing to ability to harvest. Under these conditions, therefore, one of the countries may suffer a de- crease in welfare in the process of obtaining an international MEY. However if distribution and harvesting can be separated, an international MEY can be obtained using any criterion for dis- tribution. Further, one can be obtained whereby both countries will improve their welfare from that at the open-access equilibrium. The remainder of this paper will discuss a pro- cess for reaching an international MEY making explicit the distributional problem and its rela- tionship with Condition (14). Let us consider how two countries that are operating at a point such as C in Figure 3 can move to an international MEY at a point such as D. Such a move would entail up to four mutually inderdependent types of trades be- tween the two countries, including trade in mutual changes in fishing effort (essentially trades that alter, to the mutual advantage of both countries, the property rights to the fishery from those established by the rule of capture in the open-access fishery), trade in fishing effort or rights to fish when one country has the right to fish but the other can produce effort with less cost, and trade in the produced goods F and M. The first of these trades establishes a distribution of the fishery, and the rest insure that Condition (14) will hold for that distribution. These trades are interdependent since any trade can alter demand conditions if the gains are large relative to wealth. Each of these trades will be discussed separately so as to clarify the concepts involved. It should be remembered however, that the theoretical max- imum advantage from international cooperation can not be achieved unless the trades are consid- ered simultaneously. First let us consider the potential for mutual gain from trade in mutual changes in fishing ef- fort. Assume that two countries have reached an international open-access equilibrium with coun- 59 FISHERY BULLETIN: VOL. 73, NO. 1 try X producing Ex\ units of effort and country Y producing Eyi units. (To be completely general this combination of effort can also be thought of as the one that both countries agree to use as an initial bargaining point.) Assume that under these conditions country X is operating at point A in Figure 4a. At that point, which is on social indifference curve/j , there is a specified amount of Ey (which determines the shape and position of X's PP curve) and Ex (which determines the posi- tion on the curve) being produced. There are other combinations of Ex and Ey that will cause X to operate on/j however. For example, liEy remains the same and Ex is reduced (i.e. resources are shifted from the production of effort to manufac- turing) such that there is a movement to point B, the level of social welfare will not change.'' Smal- ler reductions o^Ex that are matched by increases vn.Ey will leave welfare unchanged if the increase in Ey shifts the PP curve down such that the country is still operating on /j. Similarly, in- creases in Ex , or reductions by more than is neces- sary to shift the country to point B, will result in constant welfare if there is a simultaneous reduc- tion in Ey large enough to shift the PP curve up by the appropriate amount. This information can be more meaningfully dis- played in terms of the property right indifference curves (PRI curves) in Figure 4b. The axis repre- sent allowable levels of Ex and Ey. These allow- able levels are essentially property rights to the annual harvest that the specified amount ofE will catch. They are labeled PRx and PRy, but when there is no trade in effort, then Ex equals PRx and Ey equals PRy. Point A' represents the interna- tional open-access equilibrium point. That is, Eyi is the level of effort in country Y that will cause country X to be operating on the PP curve in Figure 4a, and Exi is the amount of effort in coun- try X that will cause it to operate at point A on that curve. Every other point in the diagram rep- resents a different combination of effort in each country and, in effect, represents a distribution of the fishery. Point A' is the distribution of the property rights by the rule of capture. Movements to the left represent reductions in the amount of allowable effort for country X, and downward 'Throughout it is assumed that there is free mobility of re- sources between fishing and manufacturing. As has been cor- rectly pointed out in the past, this is not always the case. Rather there is a time lag of perhaps as much as a generation involved. This fact should be considered when making practical applica- tions of the model. (Eyi) M, X(E.) Figure 4. — The property right indifference (PRI) curves for each country follow directly from the relationship between their pro- duction possibility curves and indifference curves. movements represent a reduction for country Y. PRIxi is that collection of bundles ofP/?Y andPRy where country X is operating on social indiffer- ence curve 1 1 . Increases in PRx (movements to the right) will only result in a constant welfare if it is matched by reductions in PRy. Small reductions in PRx with PRy remaining unchanged, will nor- mally increase welfare, and so for welfare to re- main constant, PRy must increase. As reductions in PRx get larger, however, welfare will remain constant only if there are reductions in both PRx and PRy. Similarly, PRIx2 and PRIx3 are combi- nations ofPRx and PRy where the level of welfare is the same as along 1 2 and /g, respectively.^ It 1 ^The curves will be concave from below. For reductions in allowable levels of effort, the greater the reduction, the greater is the increase in Fx that is necessary to keep welfare constant, and at the same time, the effect of decreases in the allowable 60 ANDERSON: OPTIMUM ECONOMIC YIELD follows then that any distribution of property rights to the fishery represented by a point inside the area delineated by PRIxi will lead to an im- provement in welfare in country X over that which is obtained at the international open-access equilibrium. Note that because of the shape of the curve, welfare in country X can actually be in- creased in some cases where its allowable level of effort decreases while that for the other country goes up. This is possible because at the open-access equilibrium, country X can gain from switching some resources from producing effort to producing the other good and, up to a point, these gains are possible even if country Y increases effort. (Points A, B, C, D, F, and G are analogous to A', B', C, D', F', and G'.) The reader should be aware by now of the similarity between these curves and trade in- difference curves in international trade theory. Before using these curves in the analysis of the problem at hand, however, a few more points are in order. The short line through 1 2 at F is meant to represent the slope of the PP curve if Ex remains constant and Ey decreases so that country X is operating at F. A decrease in PRy will cause the slope of X's PP curve to decrease at every level of Mx-^ As pictured here it has decreased from a positive to a negative. If it decreases such that it is steeper than the social indifference curve at that point, then the PRI curve will look like PRIx4- That is, the PRI curve will not have a negatively sloped segment to the left of the open-access equilibrium amount of effort for country X. This means that reductions in the allowable level of effort in country X, with the amount in country Y held constant, will always result in a reduction in welfare for country X. Along the same line if coun- level of effort in country Y onF^ decreases asE^ increases (i.e. dFx dKy = -bEx). Therefore greater reductions in PRy will be necessary to compensate for equal reductions in PRx as the amount of Ex is reduced from the international equilibrium level. For increases in PRx , the greater the increase the smaller is the marginal increeise in fish caught and yet the greater must be the increase in catch in order to keep welfare constant. Therefore greater reductions in PRy will be necessary to compensate for equal increases in PRx as the amount ofE^ is increased from the international equilibrium level. 'dMZ G^ = - (a - 2bEx - bEy) and so \dMx/ _ dEv 91 Therefore, as Ey decreases, the slope will decrease. try X pursues a local maximizing policy (i.e. it operates at point G in Figure 4a), the interna- tional equilibrium will be at point G' in Figure 4b. This means that under no circumstances will country X be better off if it unilaterally decreases its allowable effort and it will always be worse off if country Y increases its level of effort. This is not the case if the international equilibrium is at point A'. Figure 5a is similar to Figure 4b except that PRI curves for country Y have been added. PRIyi has the same meaning for country Y as does PRI xi for country X and is constructed in an identical fash- ion. Any distribution of property rights represented by a point inside the area delineated by PRIyi would result in an increase in the welfare of coun- try Y. It follows then that any combination that is in the area common to both PRIxi andPiJ/y^ (see hatched area of Figure 5a) will increase the wel- fare of both countries over that achieved by the open-access "law of capture" distribution of the rights to the fishery. Note again that it is possible for both countries to be better off in some cases where the trade involves a reduction in property rights in one country and yet an increase in the other. ^(EY)^' PRv ♦ '(EY) PRIy PRI XI PRX,EX) Figure 5. — The area common to the initial property right indif- ference (PRI) curves of both countries represents those distribu- tions of the fishery where both countries will be better off than at the open-access equilibrium. In some special cases, there is no such area (see b). 61 FISHERY BULLETIN: VOL. 73, NO. 1 It is also possible that in some cases there may be no changes in both Ex and Ey that will benefit both countries. If both countries adopt a local op- timum regulation policy, the PRI's will be of the general shape of those depicted in Figure 5b. In this case, there have to be mutual reductions in order for either country to gain, but as pictured here, there are no mutual reductions that will benefit both countries. If the governments have the power to control the level of effort in their countries, then it is possible for both of them to increase their welfare by each agreeing to a change in the property right dis- tribution such that the new combination lies within the area described. And further gains are possible if the PRI's for the countries are not tangent at the new point. In other words, given that the equations for the PRI's are of the form Wpfi = Wpfi (Pi?;^,P/?y), further gains are possible unless (15) dPRx dWpl dPRx aw^Pfl m/^ dPR^ dPR, that is, unless the slopes of the PRI curves are equal. Formally this says that the ratio of the change in welfare in country X due to a change in property rights in country X and to a change in rights in country Y must be equal to the ratio of the change in welfare in country Y due to a change in rights in country X and in country Y. This can be rewritten in terms of the earlier notation as: dFr + dM, Ui ^E, U2 dE dFy Ul dE, (15') U,' dEy rjy aFv , rry dM a^v dEy The change in welfare in either country due to a change in its allowable effort is equal to the change in welfare due to a change in F times the change in F due to a change in allowable effort plus the change in welfare due to a change in M times the amount of M that must be given up to produce the extra allowable effort. The change in welfare in the other country is simply the change in welfare due to a change inF times the change in F due to a change in allowable effort in the first country. Where the final trading position will be and hence what the exact gain to each country is can- not be accurately determined in advance. It de- pends however upon the international free market equilibrium distribution of the property rights to the fishery which determine the position of the PRI's, the trading ability of the two countries, the extent of the knowledge concerning each other's PRI's, and the number and particular composition of any small trades that lead up the final equilib- rium. It would be possible to construct off'er curves from the PRI's similar to the ones used in interna- tional trade theory, but since trade in mutual changes in property rights will necessitate inter- governmental negotiations and since they will, more than likely, take place on a lump-sum basis, the equilibrium determined by their intersection would be of doubtful significance. To summarize this discussion let us consider point J in Figure 5a, which is one possible final trading position. Notice that it is not possible to redistribute the property rights from that point v^dthout forcing one of the countries to suffer a loss in welfare; that is, there are no further changes in the distribution of the property rights that will be mutually beneficial. This is one of the conditions that must hold for an MEY of an international fishery. It determines the amount of fish that should be caught and the distribution of the rights to catch it. An important point to remember how- ever is that this condition vdll not guarantee that the fish are caught at the lowest possible cost, and yet this is a very important aspect of MEY. Let us now consider the potential for mutual gains from trade in actual property rights or in fishing effort. Such trade is not possible unless the rights to fish have been formalized either at the open-access equilibrium or at some other mutu- ally agreed upon point. Again it should be remem- bered that this is only one type of trade, and the degree to which each country is willing to engage in it depends to some extent upon the makeup of the other trades. Just because a country has the right to fish does not mean that it should necessarily produce the effort to catch the fish. For instance, if the oppor- tunity cost of producing effort is cheaper inX, then both countries can gain if X expands the produc- tion of effort and then sells the increase to Y, who must make a corresponding reduction in its pro- duction of effort. If the price of effort for these international sales is between that in each coun- try, both will be able to gain. Country X will gain because it is getting more for the effort that it cost to produce. Country Y will gain because it can buy 62 ANDERSON: OPTIMUM ECONOMIC YIELD effort cheaper than it can produce it at home. These mutually beneficial trades can continue until the opportunity cost of producing effort is the same in both countries, i.e. until: dMy dMy dEx dE^ (16) The same thing could be accomplished by X purchasing rights to apply effort from Y until the MRT's for E and M are equal. Assume for simplic- ity that Pp = Pp . Initially the price for a right to use one unit of effort would have to be somewhat above the rent the right-holder in Y would earn by doing the fishing himself. {Ry = Pp^^ - -Pj , where a in this case is the total of the allowable efforts from both countries.) People in X will be able to pay more than that since P^ is less than P J . In trade equilibrium the prices of fish and effort are the same in both countries, and therefore the rents in both countries will be identical and no further gains from trade are possible. While the above will not change the amount of fish produced, it will make sure that effort is being produced at a minimum cost. The savings can be used to produce more of the manufactured good which can be distributed such that both countries are better off. Now that two of the possible types of trade have been discussed, it will prove worthwhile to show exactly how they can be interrelated. Trade in E or in fishing rights may have an effect on the bargaining for the distribution of property rights. To see this, assume that after such bargainings country X is at point D in Figure 1 , • dMy . , , dMy ^^ 4a and at that point -W^ is less than -jy- . If it produces q more units of £ but sells them to Y who reduces its production of £ by the same amount, the PP curve will not change. Initially X will op- erate somewhere horizontally to the left of point D because it had to give up units of M^ to get the extra units o^E. Y will be willing to pay sufficient units of M toX such that it will ultimately operate somewhere horizontally to the right of D and will therefore show an increase in welfare. Therefore at point D' in Figure 4b, which represents the rights to fish and not the actual amount of £■ pro- duced in each country, the welfare of X will in- crease. By similar analysis it can be shown that if trade is possible, Y will always be at a higher level of welfare at D' also. This means that the PRI's of both X and Y will change shape and position. Therefore more than likely there will be the possi- bility of further mutually beneficial trades in the distribution of fishing rights. The final type of trade to consider is trade in the final products M andF . If the relative prices are different in the two countries, mutually beneficial trades can be arranged. These trades can continue to be mutually beneficial until the marginal rate of substitution in both countries is the same, i.e. until: U^. u u-^ t/,r (17) These trades will be affected by trades in E and also in changes in the allocation of the property rights. On a practical note it must be admitted that few countries will be willing to let their international trade policy in all goods be dictated by their fishery management program. Therefore it is un- realistic to assume that they will drop all restric- tions on international trade on this account. This means that even after the rights to fish have been distributed, there are four things that can be traded: fish, manufactured goods, effort, and rights to fish. Because the prices of the last two are directly related to those of the first two, the relative demands for M and F will determine the equilibrium set of prices. It is impossible to pre- dict, however, just what the actual trade bundle will be. For instance, nothing in the model allows us to predict whetherX will export effort or import fishing rights if it has a comparative advantage in producing effort. The outcome of that, however, will affect its exports or imports of F. Although the exact makeup of the international MEY position cannot be described, Conditions (15), (16), and (17) must hold simultaneously for it to be in effect. (Condition (15) sets a distribution from which no further mutual gains are possible, and Conditions (16) and (17) guarantee that Con- dition (14) above will hold for that distribution.) That is all potential mutual gains (where a mutual gain could consist of one country being made better off and the other remaining the same) by (1) altering the distribution of the rights to use effort, (2) trading in actual rights or in effort itself, or (3) trading in final goods, have been achieved. This point (say at point D in Figure 3) is a Pareto point that can be reached by mutually advanta- 63 FISHERY BULLETIN: VOL. 73, NO. 1 geous trades between the two countries given their initial positions which include their produc- tive capacity and the rights to the fishery that they have obtained by the right of capture. At this point there will be an MEY to the fishery. The proper amount offish will be harvested and at the lowest cost possible. But since there is nothing sacred about these initial positions, point D is not inherently superior to any other point on the curve. If the world order somehow alters their initial positions, for instance, by saying that since y is a poor country it should be able to expand its effort and X should do the opposite, the same types of trades will still be possible, and they will lead to a point on the curve that is more advan- tageous to country X than was point D. This point would also be an MEY given the distribution of productive capacity and of the wealth of the fish- ery. The distribution of the rights to the fishery is very important in determining the MEY of the fishery. Let us consider some of the practical im- plications of this discussion. First, before an inter- national fishery can be optimally managed, the wealth from it must be distributed. The exact makeup of the distribution is not important, but it is possible, in most cases, to find a distribution whereby both countries are better off than at the initial bargaining point. The rights to the fish- ery should be transferable if the country owning them is to receive the maximum possible benefit. This way, it can sell the rights or hire effort from other countries to utilize them if it does not have a comparative advantage in producing effort. Therefore, unless the upcoming Law of the Sea Conference can agree to some sort of distribution of the wealth of the fishery and make allowances for possible trades in the makeup of the distribu- tion bundle and also in fishing rights and effort, there is little hope for economically rational management of international fisheries. The results of this two country, one fishery model can be expanded in a fairly straightforward fashion to a situation where there are many coun- tries that simultaneously exploit several different fisheries. An international open-access equilib- rium will occur when, in each country, the average returns from fishing the various stocks are equal to the average cost of providing effort. The dis- tribution of the wealth from the fisheries will de- pend on the ability of each of the countries to produce the effort that is most efficient for a par- ticular fishery. The more efficient producers will capture a larger share of the fisheries. If perfect international trade in fish products is not possible, then the distribution of the fishery by the "rule of capture" wall also depend upon the tastes of the countries. A country that has the potential to har- vest a certain type offish very efficiently but has little desire for the product and cannot use it in international trade wa 11 not exploit that stock very extensively. The usefulness of unilateral regulation in this situation will probably be less than in the two country case. Any reduction in effort will more than likely be met by an increase from one of the other countries. Therefore, while the country will show an increase in the amount of other products it can produce, it is entirely possible that the value of its total production will fall due to the decrease in catch. Proper international regulation must take into account the effect that effort from one country will have on the yields to other countries exploiting the same stocks. With this consideration in mind, each country can benefit from some program of reallocation of the rights to the fish stocks from that which exists under open access. To achieve the maximum potential benefits, this program should include the possibility of trade in effort, fishing rights, and final products. The existence of many countries wall of course make it much more difficult to specify the set of redistributions that would be beneficial to all concerned and even more difficult to get the countries to agree to one combi- nation within that set. A major problem with in- ternational regulation is that allocational re- quirements are just as important as economic efficiency requirements. But given a mutually agreed upon allocation (i.e. a certain allowable level of effort in each country for all fisheries), the efficiency requirements can be met. The prob- lem is to get agreement on a distribution plan with many different countries involved. SUMMARY AND CONCLUSIONS In the first section of the paper the general equilibrium model was used to derive the familiar result that in an open-access fishery too many resources will be allocated to the production of fishing effort. Using this model it is possible to explicitly take into account the lost production of other goods. In the second section the general equilibrium model was expanded to include two countries exploiting the same open-access fishery. The amount of effort used in one country will af- 64 ANDERSON: OPTIMUM ECONOMIC YIELD feet the production possibilities in the other by changing the catch per unit of effort. Therefore, there is a direct technical relationship between the two countries. An international open-access equilibrium will exist when the average return to effort is equal to the marginal cost of providing it. (Whether or not such an equilibrium will ever be reached is another question.) The international optimum is where the marginal increase in the value of the fish caught (regardless of the country in which it is landed) is equal to the marginal cost of producing the last unit of effort in both coun- tries. Using this model, two interesting points can be made. First, under open access, what are nor- mally considered to be improvements in the terms of trade, for either the exporter or the importer of fish, can in some circumstances lead to a decrease in welfare. Also attempts at unilateral manage- ment can lead to decreases in welfare depending on the way in which the other country's fishing industry reacts. Proper regulation policies should directly take these things into account. The topic of the third section was the necessary conditions for an MEY of an international fishery. The discussion with its implicit assumptions of governments that are willing and able to negotiate in an open and far ranging manner at zero cost, free trade in all goods, regulation methods that are not at the expense of efficiency, a physically independent fish stock that is only available to two countries, showed if negotiation is possible that an international MEY can be reached. This point will be the MEY of the fishery. (Even if the assumption about the possibility of free trade in final goods is dropped, the analysis of trade concerning the distribution of property rights to the fishery and trade in rights or effort is still valid. Therefore, even if there are different price and cost structures in the two countries, there is a basis for selecting a second best total amount and composition of fishing effort.) It is also pointed out that there are many points that satisfy the conditions of an international MEY and that the distribution of the rights to the fishery (especially where the wealth from the fishery is large relative to the productive capacities of the countries) and, to a lesser extent, the differences in negotiating ability have an ef- fect on which one will apply at any point in time. (There will not be one point that can be called MEY as in the case of a national fishery.) This is important because fishery negotiations typically work in the reverse. They try to find some op- timum total amount of effort that should be ap- plied and then they divide it in some equitable fashion, but it is impossible to choose an optimum amount unless the distribution has already been determined. With regard to the argument that the underde- veloped countries should be granted preferential treatment in the distribution of the ocean's living resources, the model points out that if this is ac- cepted, it does not mean that they should necessar- ily do the fishing. Rather, if they do not have a comparative advantage in the production of fishing effort, they would be better off by either selling their rights to the fish or by hiring fishing effort from other countries. In conclusion this paper has formalized the analysis of the problems of international fisheries management that earlier writers only briefly dis- cussed. To their list of problems of different prices, taste, and cost structures, it adds the eff"ect that the distribution of the wealth of the fishery itself can have on the final outcome. It presents the three conditions for an MEY of an internationally utilized fishery. More generally the conditions guarantee the proper production bundle of all goods and its optimal distribution given the pro- ductive capacity of the countries and of the fishery and the distribution of wealth. Although the discussion has been in terms of a fishery, the analysis could be expanded to other common property resources, such as air and watersheds, deep-sea mineral sources, etc. by taking proper consideration of the various physi- cal characteristics of the resource involved. ACKNOWLEDGMENTS The author is appreciative of helpful comments on an earlier draft of this paper from Gardner Brown and an anonymous referee but retains full responsibility for the contents. LITERATURE CITED Anderson, L. G. 1973. Optimum economic yield of a fishery given a vari- able price of output. J. Fish. Res. Board Can. 30:509-518. Christy, F. T., Jr., and A. Scott. 1965. The common wealth in ocean fisheries; some prob- lems of growth and economic allocation. Johns Hopkins Press, Baltimore, 281 p. Copes, P. 1970. The backward-bending supply curve of the fishing industry. Scot. J. Polit. Econ. 17:69-77. 65 FISHERY BULLETIN: VOL. 73, NO. 1 Crutchfield, J. A. (editor). 1965. The fisheries; problems in resource management. Univ. Wash. Press, Seattle, 136 p. Crutchfield, J., and A. Zellner. 1962. Economic aspects of the Pacific halibut fishery. U.S. Fish Wildl. Serv., Fish. Ind. Res. 1:1-173. Gordon, H. S. 1954. The economic theory of a common-property re- source: The fishery. J. Polit. Econ. 62:124-142. Gould, J. R. 1972. Extinction of a fishery by commercial exploitation: A note. J. Polit. Econ. 80:1031-1038. Schaefer, M. B. 1957. Some considerations of population dynamics and economics in relation to the management of the commer- cial marine fisheries. J. Fish. Res. Board Can. 14:669-681. Scott, A. 1955. The fishery: The objectives of sole ownership. J. Polit. Econ. 63:116-124. Scott, A., and C. Southey. 1970. The problem of achieving efficient regulation of a fishery. In A. D. Scott (editor). Economics of fisheries management. Univ. B.C., Inst. Anim. Resour. Ecol., Van- couver, 115 p. Smith, V. L. 1969. On models of commercial fishing. J. Polit. Econ. 77:181-198. Southey, C. 1972. Policy prescriptions in bionomic models: The case of the fishery. J. Polit. Econ. 80:769-775. TURVEY, R. 1964. Optimization and suboptimization in fishery regula- tion. Am. Econ. Rev. 54(l):64-76. I I 66 I IMPACT OF THERMAL EFFLUENT FROM A STEAM-ELECTRIC STATION ON A MARSHLAND NURSERY AREA DURING THE HOT SEASON William E. S. Carri and James T. Giesel^ ABSTRACT Seine samples of fishes were collected during the hot season from three similar marshland creeks situated at various distances from a steam-electric station near Jacksonville, Fla. Thermal effluent from the electric station is discharged directly into one creek and enters a second creek on the initial stage of each rising tide. The third creek remained at ambient temperature. Fishes collected in the samples were analyzed for species composition and for density and biomass per unit area. A total of 48 species belonging to 23 families were identified. Thirty-seven species were collected at least once in the ambient temperature creek whereas 30 species were collected in the creek receiving the maxi- mum amount of thermal effluent. Twenty species appearing in the samples are categorized as utilizable species because they are used by man either as food or for various fishery products. Specimens of all utilizable species were juveniles. In the thermally affected creeks, both the numbers and the biomass per unit area of juveniles of utilizable species were 3- to 10-fold smaller than those obtained in collections from the ambient temperature creek. When data for the entire hot season are considered, the creek receiving the largest input of thermal effluent supported a population of fishes having approximately 19% of its numbers and 32% of its biomass composed of juveniles of utilizable species. In contrast, the ambient temperature creek supported a population having approximately 73% of its numbers and 83% of its biomass composed of such species. Whereas juveniles of two species of mullet (Mugil curema and M. cephalus) accounted for the majority of the utilizable fishes using the thermally affected creeks as a nursery area, large numbers of juveniles of at least five additional utilizable species occupied the ambient tempera- ture creek. These species were as follows (in order of decreasing abundance): tidewater silverside, Menidia beryllina; spot, Leiostomus xanthurus; Atlantic menhaden, Brevoortia tyrannus; silver perch, Bairdiella chrysura; and Atlantic thread herring, Opisthonema ogUnum. There is no longer any doubt that estuarine areas play a vital role in the life cycles of the majority of species of finfish and shellfish that are harvested annually in coastal fisheries. The role of estuaries as nursery areas for both sport and commercial species is now well documented (Skud and Wilson 1960; Smith et al. 1966; Sykes and Finucane 1966; Carr and Adams 1973; others). The majority of sport and commercial species must inhabit es- tuarine areas during at least part of their life cycles. Most frequently it is the early juvenile stages that exhibit the most pronounced estuarine dependence. Thermal additions from power plants are con- sidered to pose a potentially serious threat to valu- able estuarine habitats. Krenkel and Parker (1969) have estimated that the amount of water required for condenser cooling by power plants in *C. V. Whitney Marine Laboratory of University of Florida at Marineland, Route 1, Box 121, St. Augustine, FL 32084 ^Department of Zoology, University of Florida, Gainesville, FL 32601 this country will increase from 50 trillion gallons per year in 1968 to 100 trillion gallons per year by 1980. This latter amount represents approxi- mately one-fifth of the total land runoff in the contiguous United States. The immense volumes of water required for cooling by power plants are most readily obtained by building these plants adjacent to estuaries or in other coastal locations. The fact that estuarine areas "are among the most productive natural ecosystems in the world" (Schelske and Odum 1962) raises the question as to whether meeting the increasing needs for elec- tricity by our growing population is best satisfied by using estuarine areas as the receiving waters for ever increasing discharges of thermal effluents. Although a large literature exists concerning various biological facets of "thermal pollution" (reviews are provided by Naylor 1965; Wurtz and Renn 1965; Krenkel and Parker 1969; Jensen et al. 1969; Coutant 1970, 1971; Sylvester 1972; others) we are aware of no published studies that Manuscript accepted May 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. 67 FISHERY BULLETIN: VOL. 73. NO. 1 attempted to measure in situ the impact of ther- mal additions upon the capacity of an estuarine habitat to continue functioning as a viable nur- sery area, particularly for species of sport and commercial significance. Nugent (1970) provided one of the most complete studies on the effects of a thermal effluent on the estuarine macrofauna in the vicinity of a power station south of Miami, Fla. Nugent concluded that there were both beneficial and harmful effects attributable to the thermal additions but that the overall impact was "detri- mental to many of the economically valuable ani- mals of the waterway." Nugent found that during the hot summer months the heated effluent de- creased the number of fishes present in the dis- charge area and also contributed to the death of certain organisms. However, the methods used in this study for the collection of fishes (gill nets, traps, and hoop nets) are unsuitable for the collec- tion of many juvenile specimens and are some- what inappropriate for estimates of density and standing crop. Grimes (1971) and Grimes and Mountain (1971) studied the effects of a thermal effluent upon marine fishes in the vicinity of a power station near Crystal River, Fla. Their major conclusions were that the natural seasonal abun- dance and the diversity of fishes were slightly altered by fishes being attracted into the heated area during late fall and early winter and by being repulsed during the summer. However the collect- ing methods and the station locations used in this study make the data difficult to assess in terms of the impact of the thermal effluent on the nursery area capacity of the affected area. The current study was designed to evaluate in quantitative terms the impact that the discharge of a thermal effluent by a steam electric station had upon the capacity of an estuarine habitat to continue functioning as a nursery area during the hot season. This study was conducted in a marsh- land area to the northeast of Jacksonville, Fla. The data were obtained by analyzing the contents of seine samples taken from shallow-water sta- tions located in three marshland creeks situated in the vicinity of the power station. METHODS Description of Study Area San Carlos Creek, a small marshland creek draining into the St. Johns River, receives the discharge of thermal effluent from the Northside Generating Station (NGS) operated for the city of Jacksonville by the Jacksonville Electric Author- ity (see Figure 1). The NGS is situated in a rela- tively undeveloped marshland area to the north- east of Jacksonville approximately 10 miles west of the juncture of the St. Johns River with the Atlantic Ocean. Currently the NGS has two, of an anticipated three, oil-fired steam-electric units on line. Units 1 and 2 of the NGS (550-MW generat- ing capacity) discharge approximately 280,000 gallons/min of thermal effluent directly into San Carlos Creek via outfalls situated 150 ft apart. The completion of Unit 3 (550 MW) in 1976 will result in the discharge of an additional 280,000 gallons/min of thermal effluent into this same creek. Cooling water for the NGS enters via a flume from the St. Johns River and the heated effluent is discharged into San Carlos Creek at a point approximately 0.75 mile upstream from the river. San Carlos Creek and two other physically simi- lar creeks located adjacent to the site described above were used for the collection of fishes de- scribed in the current study. San Carlos Creek not only receives directly the thermal effluent from Figure 1. — Study area showing location of Northside Generat- ing Station and marshland creeks adjacent to St. Johns River north of Jacksonville, Fla (see inset). Locations of sampling stations in San Carlos Creek, Nichols Creek, and Browns Creek are indicated by numbers. Juncture of river and ocean is situated about 10 miles to the east. 68 CARR and GIESEL: IMPACT OF THERMAL EFFLUENT the NGS but on each rising tide this effluent is backed up by tidal action and much of it is retained within the confines of this creek. At this time the thermal effects extend to the uppermost reaches of the creek. A second creek, Nichols Creek (see Fig- ure 1), receives an injection of thermal effluent during the initial stage of each rising tide. Nichols Creek converges with San Carlos Creek just prior to the juncture of both with the river. A major branch of a third creek, Browns Creek, is situated approximately 1 mile east of San Carlos Creek and is completely beyond the zone of thermal influence produced by the power plant. Browns Creek served as a "control" creek that remained at am- bient temperature. The three creeks are physically quite similar in terms of their size, depth, and contiguous marsh- land and upland areas. The substrate in each con- sists primarily of soft black mud rich in organic material. Scattered bars of a firmer sand- mud composition are present. Each creek is lined on either side with marsh grasses, primarily Spartina alterniflora and J uncus roemarianus. No submerged sea grass or other attached mac- rophytes are present in the creek beds themselves. The major variables affecting differentially the habitats of the three creeks are the thermal effluent, the chemical agents used in the cleaning of condenser tubes, and the clearing and altera- tion of the landscape necessary for the construc- tion of the power plant. Sampling Stations Three sampling stations were established in each of the three creeks (see Figure 1). One of the stations in each creek was situated near the creek mouth whereas the other two were situated at appropriate distances upstream. The station sites in each creek were selected such that two stations were situated at the sites of juncture of small ad- joining creeklets and the third at the edge of a bar. During the hot season of 1973, samples were taken from all nine stations during June and July and from seven of the stations in September. Collecting Methods Fishes were collected at all stations with a bag seine (50 x 6 ft) constructed of 3/8-inch stretch mesh netting. The dimensions of the area seined in each sample were measured with a steel tape at the time the sample was taken. Areas sampled at the stations varied somewhat according to the particular configuration of each seine haul; these areas ranged from 102 to 403 m^ per station. Dur- ing each sampling period, all seine hauls were made during the day on consecutive days within 1.5 h of low tide. Specimens were preserved im- mediately in 20% Formalin^-seawater and later washed with tap water and stored in 75% iso- propyl alcohol. Determinations of biomass are based on weights of preserved specimens. Inver- tebrates obtained in the samples were also re- tained for future analysis. A Beckman electrodeless induction salinometer was used to obtain measurements of temperature and salinity. Quantitative core samples and plankton sam- ples were also taken but their analyses are incom- plete and they are not reported here. Presentation of Data To minimize the number of tables and figures necessary for the presentation of data, analyses of the samples taken from the three stations in each creek have been pooled for each monthly collec- tion. Although this method prohibits comparisons of variations between individual stations within a particular creek, this procedure provides a more direct means of analyzing and comparing the overall population structure within each creek. RESULTS Table 1 presents temperature and salinity measurements taken from San Carlos Creek, Nichols Creek, and Browns Creek during the study period. The temperature data from Browns Creek, which is beyond the range of thermal influence produced by the power station, can be used as a measure of the daytime ambient tem- perature regime for a creek in this area. During the June sampling period, the average recorded temperature of water discharged by Units 1 and 2 of the NGS into San Carlos Creek was from 5.6° to 7.7°C above the average temperatures recorded at the three stations in Browns Creek. During July, San Carlos Creek received water from the power station that averaged 8.0° to 8.9°C higher than the averages recorded in Browns Creek. During Sep- tember, this differential increased to 9.1° to 10.8°C. The highest temperature that we recorded ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 69 FISHERY BULLETIN: VOL. 73, NO. 1 Table 1. — Daytime measurements of temperature and salinity recorded during the hot season of 1973 from three creeks in the vicinity of the Northside Generating Station, Jacksonville, Fla. Outfall s m San Carlos C reek N chols Creek Brow/n Creek San Carlo! Unit 1 5 Creek Item Stn. 1 Stn. 2 Stn. 3 Stn. 1 Stn 2 Stn. 3 Stn 1 Stn. 2 Stn. 3 Unit 2 23-27 June: Temperature. °C: Maximum 35.9 35.0 36.4 31.1 32.6 34.6 30.5 30.5 '27.9 36.5 37.8 Minimum 30.3 31.0 30.0 27.8 28.2 27.6 26.7 26.9 31.8 33.9 Average 33.1 33.1 33.5 29.4 29.9 30.2 28.5 28.5 34.1 36.2 Salinity, "/oo: Maximum 30.6 30.1 29.0 27.0 26.6 25.6 27.5 26.2 '21.7 31.0 31.0 Minimum 20.7 24.7 18.2 18.9 18.3 18.7 18.3 17.7 23.3 24.2 Average 25.9 27.3 25.6 21.9 21.9 21.7 22.6 22,3 26.5 26.8 25-26 July; Temperature, °C: Maximum 38.0 38.2 392 37.5 37.3 35.8 31.2 30.6 31.6 39.7 39.4 Minimum 30.4 33.9 32.8 30.1 30.4 31.6 29.2 29.7 29.8 36.9 38.5 Average 35.3 35.7 35.5 32.3 32.4 32.9 30.2 30.2 30.5 38.5 39.1 Salinity, %o: Maximum 31.2 29.8 31.0 31.5 31.3 29.3 27.5 27.3 27.2 30.8 30.8 Minimum 23.3 26.2 26.6 23.5 23.6 24.1 25.5 21.8 21.7 25.7 25.2 Average 28.1 28.9 28.8 27.4 27.6 26.0 26.5 25.2 25.1 29.2 29.2 19-20 September: Temperature. °C: Maximum 37.3 37.7 37.9 36.2 2 35.0 29.7 29.6 27.5 37.6 38.4 Minimum 30.6 32.3 32.1 28.5 27.9 27.3 26.5 27.0 37.4 37.8 Average 34.3 36.2 36.3 31.5 31.4 28.4 27.9 27.3 37.5 38.1 Salinity, "/oo: Maximum 26.4 24.3 24.3 24.8 2 24.7 27.2 26.2 26.4 23.2 24.0 Minimum 19.5 22.1 22.7 18.1 21.6 21.9 20.8 19.6 20.6 21.0 Average 22.0 23.2 23.5 22.4 23.4 24.6 22.7 23.0 22.0 22.5 'Only one recording. ^Not recorded. was 39.7°C taken at the outfall of Unit 2. The data shown in Table 1 suggest that the thermal regime present in Nichols Creek was somewhat inter- mediate between that of the other two creeks. However this is not entirely the case. During the initial phase of each rising tide, tidal action causes the injection of heated effluent from the power plant up the entire length of Nichols Creek. On 19 September, we recorded this injection as it reached and later passed Station 1 in this creek (see Figure 2). The highest temperature recorded at Station 1 during this day was 36.2°C. The pas- sage time of this injection of hot water was approx- imately 2 h with temperatures greater than 34°C lasting approximately 1 h. Only a slight drop in temperature was apparent when the hot water reached Station 3 situated approximately 0.6 mile away (see Figure 2). Hence, whereas the minimum and average temperatures in Nichols Creek are more similar to those of Browns Creek than to those in San Carlos Creek, the maximum temperatures in Nichols Creek are more similar to those in San Carlos Creek. Consequently, at the onset of each rising tide (twice daily), organisms living in Nichols Creek are subjected to a period of 1- or 2-h duration during which the water temper- ature is markedly above ambient and almost as high as that in San Carlos Creek. Table 2 provides a list of the species of fishes collected from the three creeks during the hot sea- son of 1973. A total of 48 species belonging to 23 families were collected. Aside from the Cy- prinodontidae and certain of the Gerreidae and o " 33.0 a TIDE TIMES. , . , . . 1 ■ ' LO* TIDE - B-.4 3 AM y^ - HIGH TIDE 3.13PM [ A ^ \ ^STOP - 1 STATION 1 \ / STATION 3 / ^ / ^STOP j / ■ STABI^ START.^ TIME OF DAY Figure 2. — Recordings of water temperature taken in Nichols Creek on the initial stage of rising tide, 19 September 1973. Appearance of thermal effluent from the power plant is indicated by the sudden increases in temperature at Stations 1 and 3. 70 CARR and GlESEL: IMPACT OF THERMAL EFFLUENT Table 2. — List of fishes collected during the hot season of 1973 from three creeks in the vicinity of the Northside Generating Station, Jacksonville, Fla. Family Scientific name Common name' Class^ Species utilized^ Elopidae Clupeidae Engraulidae Synodontidae Ariidae Batrachoididae Belonidae Cyprinodontidae Poeciliidae Atherinidae Syngnathidae Carangidae Lutjanidae Gerreldae Sparidae Sciaenidae Ephippidae Mugilidae Gobiidae Triglidae Bctliidae Cynoglossidae Tetraodontidae Elops saurus Brevoortia tyrannus" Opisthonema oglinum Anchoa hepsetus Anchoa mitchilli Synodus toetens Arius fells Opsanus tau Strongylura marina Cyprlnodon variegatus Fundulus grandis Fundulus heteroclitus Fundulus majalis^ Gambusia afflnis Poecilla latiplnna Menldia beryllina Syngnathus florldae Caranx hippos Chloroscombrus chrysurus Selene vomer Trachlnotus falcatus Lutjanus griseus DIapterus ollsthostomus Euclnostomous argenteus Euclnostomous gula Gerres cinereus Archosargus probatocephalus Lagodon rhomboldes Balrdlella chrysura Cynosclon nebulosus Leiostomus xanthurus MIcropogon undulatus Pogonlas cromis Sclaenops ocellata Chaetodipterus faber Mugll cephalus Mugll curema Goblonellus boleosoma Gobionellus hastatus Goblonellus smaragdus Goblosoma bosci MIcrogoblus gulosus Prionotus tribulus CItharlchthys spllopterus Etropus crossotus Parallchthys lethostlgma Symphurus plaglusa Sphoeroldes nephelus Ladyfisli Atlantic menhaden Atlantic thread herring Striped anchovy Bay anchovy Inshore lizardfish Sea catfish Oyster toadfish Atlantic needlefish Sheepshead minnow Gulf killifish Mummichog Striped killifish Mosquitofish Sailfin molly Tidewater silverside Dusky pipefish Crevalle jack Atlantic bumper Lookdown Permit Gray snapper Irish pompano Spotfin mojarra Silver jenny Yellov/fin mojarra Sheepshead Pinfish Silver perch Spotted seatrout Spot Atlantic croaker Black drum Red drum Atlantic spadefish Striped mullet White mullet Darter goby Sharptail goby Emerald goby Naked goby Clown goby Bighead searobin Bay whiff Fringed flounder Southern flounder Blackcheek tonguefish Southern puffer J X J X J X J J J J J J J-A J-A J-A J-A J-A J-A J X J J X J J J X J X J J-A J J X J X J X J X J X J X J X J X J X J X J X J X J J-A J J J-A J J J J X J J 'Common names recommended by Bailey (1970) are used. 2J = juvenile; A = adult. 'Species utilized refers to either sport species or to species cited by Lyies (1969:463-487) as being used by man for food or related fishery products. *Some of these specimens may have been B. smithll and/or hybrids of Brevoortia smithll as described by Dahlberg (1970). Since they were all juvenile specimens the major characters given by Dahlberg for distinguishing between the three possibilities were extremely difficult to apply with certainty. ^According to Carter R. Gilbert (pers. commun.) of the Florida State Museum, some of these specimens may have been F. simllis. The taxonomic status of the two species on the northeast coast of Florida is somewhat uncertain. Gobiidae, all of the specimens were juveniles that were using the creeks as a nursery area. Twenty of the species are utilized directly by man, i.e., are species used for food and/or related fishery prod- ucts including sport species and bait fishes. Sub- sequent references to "utilizable species" refer to those used by man as defined above. Tables 3-5 provide monthly summaries of the numbers of individuals and the estimated den- sities of all fish species collected from the three creeks. Of the 48 species obtained in one or more collections, 30 were collected at least once in San Carlos Creek, 23 were collected in Nichols Creek, and 37 appeared in Browns Creek. Four species, Cyprinodon variegatus and three species of gobies {Gobionellus smaragdus, Gobiosomo bosci, and Microgobius gulosus), were collected only in San Carlos Creek thereby suggesting that they preferred the high temperature regime af- forded there. However, the three species of gobies appeared only in the June samples. Eleven other species of temperature tolerant fishes were pres- ent in San Carlos Creek at densities that were either as great, or greater, than the densities pres- ent in the ambient temperature creek. These species were as follows (in order of decreasing 71 FISHERY BULLETIN: VOL. 73, NO. 1 Table 3. Collections of fishes from three stations on San Carlos Creek during the hot season of 1973. Area seined in June was 706 m*, in July was 563 m^, and in September was 563^. Total area seined was 1,832 m^. June collection July collection Septen- bar collection Total collections Family Species No. No./lOO m2 No. No./lOO m2 No. No./lOO m2 No. No./lOO m2 Elopidae Elops saurus 3 0.4 — — 1 02 4 0.2 Clupeidae Brevoortia tyrannus — — 4 0.7 — — 4 0,2 Opisthonema oglinum — — 6 1.1 — — 6 0.3 Engraulidae Anchoa hepsetus 1 0.1 — — — — 1 0.05 Belonidae Strongylura marina 1 0.1 — — — — 1 0.05 Cyprinodontldae Cyprinodon variegatus 10 1.4 4 0.7 12 2.1 26 1.4 Fundulus grandis 62 8.8 161 28.6 179 31.6 402 21.9 Fundulus heteroclitus 192 27.2 669 118.8 405 72.1 1266 69.2 Fundulus majalis 15 2.1 44 7.8 50 8.9 109 5.9 Poeciliidae Gambusia afflnis — — 5 0,9 2 0.4 7 0.4 Poecilia latipinna 40 5.7 44 7.8 31 5.5 115 6.3 Atherinidae Menidia beryllina 2 0.3 — — 5 0.7 7 0.4 Lutjanidae Lutjanus griseus — — — — 4 0.7 4 0.2 Gerreidae Diapterus olisthostomus — — — — 1 0.2 1 0.05 Eucinostomous argenteus 137 19.4 36 6.4 693 123.1 866 47.3 Eucinostomous gula — — — — 3 0.5 3 0.2 Gerres cinereus 1 0.1 2 0.4 3 0.5 6 0.3 Sparldae Lagodon rhomboides 1 0.1 — — — — 1 0.05 Sciaenidae Cynoscion nebulosus 1 0.1 — — 2 0.4 3 0.2 Leiostomus xanthurus 44 6.2 — — — — 44 2.4 Pogonias cromis 4 0.6 — — — — 4 0.2 Sciaenops ocellata 1 0.1 — — — — 1 0.05 Mugilidae Mugil cephalus 40 5.7 13 2.3 1 0.2 54 2.9 Mugil curema 211 29.9 292 51.9 42 7.5 545 29.7 Gobiidae Gobionellus boleosoma 5 0.7 — — — — 5 0.3 Gobionellus hastatus 12 1.7 2 0.4 — — 14 0.8 Gobionellus smaragdus 2 0.3 — — — — 2 0.1 Gobiosoma bosci 1 0.1 — — — — 1 0.05 Microgobius gulosus 3 0.4 — — — — 3 0.2 Bothidae Citharichthys spilopterus Total 3 792 0.4 111.9 — — — — 3 3,508 0.2 1,282 227.8 1.434 254.6 191.5 Total utilizable species 308 43.5 317 56.4 58 10.2 683 37.1 abundance): Fundulus heteroclitus, Eucinosto- mous argenteus, Mugil curema, F. grandis, Poecilia latipinna, F. majalis, Gobionellus has- tatus, Gambusia af finis, Gerres cinereus, Elops saurus, and Lutjanus griseus. Eleven species, Synodus foetens, Opsanus tau, Syngnathus floridae, Chloroscombrus chrysurus , Selene vomer, Micropogon undulatus, Chaeto- dipterus faber, Prionotus tribulus, Etropus crossotus , Paralichthys lethostigma, and Table 4. — Collections of fishes from three stations on Nichols Creek during the hot season of 1973. Area seined in June was 819 m^, in July was 815 m^, and in September 714 m^. Total area seined was 2,348 m^. Species June No. collection No./lOO m2 July collection September collection' No. No./lOO m2 Tota No. collections Family No. No./lOO m2 No./lOO m2 Clupeidae Brevoortia tyrannus 1 0.1 — — — 1 0.04 Engraulidae Anchoa mitchilli 36 4.4 — — — — 36 1.5 Ariidae Arius felis — — 1 0.1 — — 1 0.04 Cyprinodontldae Fundulus grandis 77 9.4 55 6.7 — — 132 5.6 Fundulus heteroclitus 346 42.2 766 94.0 10 1.4 1122 47.8 Fundulus majalis 4 0.5 4 0.5 — — 8 0.3 Poeciliidae Poecilia latipinna — — 12 1.5 — — 12 0.5 Atherinidae Menidia beryllina 13 1.6 11 1.3 102 14.3 126 5.4 Carangidae Caranx hippos — — — — 2 0.3 2 0.1 Trachinotus falcatus — — 1 0.1 — — 0.04 Gerreidae Diapterus olisthostomus 1 0.1 — — — — 0.04 Eucinostomous argenteus 168 20.5 246 30.2 103 14.4 517 22.0 Eucinostomous gula — — 49 6.0 12 1.7 61 2.6 Sparidae Archosargus probatocephalus — — 1 0.1 — — 0.04 Lagodon rhomboides — — 1 0.1 — — 0.04 Sciaenidae Bairdiella chrysura — — — — 1 0,1 0.04 Leiostomus xanthurus 105 12.8 47 5.8 6 0.8 158 6.7 Pogonias cromis — — 1 0.1 — — 0.04 Mugilidae Mugil cephalus 121 14.8 7 0.9 1 0.1 129 5.5 Mugil curema 653 79.7 183 22.5 35 4.9 871 37.1 Gobiidae Gobionellus boleosoma 1 0.1 — — — — 0.04 Gobionellus hastatus 1 0.1 — — — — 0.04 Bothidae Citharichythys spHopterus 8 1.0 — — 1 0.1 9 0.4 Cynoglossidae Symphurus plagiusa Total 2 0.2 — — — 38.1 2 0.1 1,537 187.5 1,385 1699 273 3,195 136.0 Total utilizable species 893 109.0 252 30.9 147 20.5 1,292 55.0 'Station 2 not sampled in September due to mechanical problems. 72 CARR and GIESEL: IMPACT OF THERMAL EFFLUENT Table 5. — Collections of fishes from three stations on Browns Creek during the hot season of 1973. Area seined in June was 685 m*, in July was 676 m^, and in September was 285 m^. Total area seined was 1,646 m^. Species June collection July collection Septem No. bar collection' No./lOO m2 Total collections Family No. No./lOO m2 No. No./lOO m2 No. No./lOO m2 Elopidae Elops saurus — — 1 0.1 — — 1 0.06 Clupeidae Brevoortia tyrannus 551 80.4 420 62.1 — — 971 59.0 Opisthonema oglinum — — 103 15.2 — — 103 6.3 Engraulidae Anchoa hepsetus 21 3.1 — — 1 0.4 22 1.3 Anchoa mitchilli 265 38.7 377 55.8 16 5.6 658 40.0 Synodontidae Synodus foetens 3 0.4 — — 1 0.4 4 0.2 Batrachoididae Opsanus tau 1 0.1 — — 1 0.4 2 0.1 Belonidae Strongylura marina 1 0.1 — — — — 1 0.06 Cyprinodonfidae Fundulus grandis 2 0.3 3 0.4 1 0.4 6 0.4 Fundulus heteroclitus 717 105.0 342 50.6 9 3.5 1,068 64.9 Fundulus majalis 1 0.1 — — — — 1 0.06 Poeciliidae Gambusia affinis 2 0.3 4 0.6 6 2.1 12 0.7 Atherinidae Menidia beryllina 222 32.4 2,288 338.5 80 28.1 2,590 157.4 Syngnathidae Syngnathus floridae 1 0.1 — — 1 0.4 2 0.1 Carangidae Caranx hippos 1 0.1 1 0.1 1 0.4 3 0.2 Chloroscombrus chrysurus — — — — 2 0.7 2 0.1 Selene vomer 2 0.3 1 0.1 — — 3 0.2 Lutjanidae Lutjanus griseus — — 1 0.1 4 1.4 5 0.3 Gerreidae Eucinostomous argenteus 14 2.0 79 11.7 44 15.4 137 8.3 Eucinostomous gula 2 0.3 44 6.5 2 0.7 48 2.9 Gerres cinereus — — 1 0.1 — — 1 0.06 Sparidae Lagodon rhomboides 10 1.5 1 0.1 — — 11 0.7 Sciaenidae Bairdiella chrysura 77 11.2 38 5.6 4 1.4 119 7.2 Cynoscion nebulosus — — — — 1 0.4 1 0.06 Leiostomus xanthurus 912 133.0 134 19.8 25 8.8 1,071 65.1 Micropogon undulatus 9 1.3 2 0.3 — — 11 0.7 Sciaenops ocellata 4 0.6 — — — — 4 0.2 Ephippidae Chaetodipterus faber — — 2 0.3 — — 2 0.1 Mugilidae Mugil cephalus 61 8.9 31 4.6 1 0.4 93 5.7 Mugil curema 238 34.7 206 30.5 21 7.4 465 28.3 Gobiidae Gobionellus boleosoma 1 0.1 1 0.1 — — 2 0.1 Triglidae Prionotus tribulus — — 1 0.1 — — 1 0.06 Bothidae Citharichthys spilopterus 5 0.7 4 0.6 — — 9 0.6 Etropus crossotus — — — — 1 0.4 1 0.06 Paralichthys lethostigma 1 0.1 — — — — 1 0.06 Cynoglossidae Symphurus plagiusa 1 0.1 13 1.9 — — 14 0.9 Tetraodontidae Sphoeroides nephelus Total 2 3,127 0.3 456.2 — — 2 224 0.7 79.4 4 7,449 0.2 4,098 605.8 452.7 Total utilizable species 2,086 304.2 3,229 477.0 137 48.3 5,452 331.4 'Station 1 not sampled in September due to mechanical problems. Sphoeroides nephelus, were collected only in Browns Creek and not in either of the thermally affected creeks. However, none of the species listed above made a major contribution to the total den- sity of fishes in this ambient temperature creek. Among the other species entirely absent from San Carlos collections, only two species, Anchoa mitchilli and Bairdiella chrysura, made a significant contribution to the fish density in Browns Creek. When considered alone, the differ- ences cited above might be construed to suggest that the nursery capacity of thermally affected San Carlos Creek is not markedly different from that of Browns Creek which functions at ambient temperature. However, a critical comparison of the densities, the biomasses, and the population structure of the fishes in the three creeks reveal some important differences that are described below. Figure 3 illustrates the relative densities of both total fishes and utilizable fishes as they ap- peared in the monthly samples. The figure shows that the following three major differences existed between the populations present in the thermally affected creeks (San Carlos and Nichols) and the population present in the ambient temperature creek (Browns): 1. In June and July the density of total fishes was highest in the ambient temperature creek. 2. Throughout the entire study period the den- sity of utilizable species was markedly higher in the ambient temperature creek. 3. In the ambient temperature creek, the ma- jority of the population consisted of juveniles of utilizable species, whereas in the ther- mally affected creeks the majority consisted of species not utilized by man. In June and July the estimated density of total 73 FISHERY BULLETIN: VOL. 73, NO. 1 .UTILIZABLE SPECIES / (39.0%)/ ,/. UTILIZABLE SPECIES /(24.8 %) feg^WH^ ANJHgA/ie; UTILIZABLE^ /. SPECIES ( 58J%) '^■z UTILIZABLE SPECIES ). of the canals was lower than at the surface in any one sampling period. A definite thermocline was noted in January and February with the most inland stations exhibiting the greatest differences between surface and bottom temperatures. The greatest difference was at Station 5 in February when the bottom was 4.0°C lower than the surface. In the previous year's study, the greatest differ- ence was at Station 4 (February 1971) when the bottom was 1.8°C lower than at the surface (Lin- dall et al. 1973). SALINITY Surface and bottom salinities at the control sta- tion ranged from 19.1 to 28.0''/oo during the study and were nearly identical in any one sampling period (Figure 3). The greatest difference was in May when the bottom was 0. 7*^/00 lower than the surface. Surface salinities at canal stations were similar to those at the control station, ranging from 19.1 to 28.5%o. With few exceptions, how- ever, salinity at the bottom of the canals was higher than at the surface in any one sampling period. The greatest difference was at Station 3 in October when thd bottom was 4.5%o higher than the surface. OCT. NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP Figure 2. — Monthly water temperature at the surface and bot- tom of all hydrologic stations, October 1971-November 1972. Stratification of salinity was also noted in the previous year's study (Lindall et al. 1973). Differ- ences between surface and bottom were not as pronounced during most of that study because drought conditions prevailed throughout most of the year. Heavy rains in August 1971 ended the drought. Thus, greater differences between sur- face and bottom salinities (as much as 15%o) were recorded in the previous study than in the present study. OXYGEN Dissolved oxygen levels at each station are shown in Figure 4. Only at the control station were surface and bottom values similar, differing no more than 0.3 ml/liter in any one sampling period. At this station the lowest observed con- centration was 2.2 ml/liter (July 1972). Surface oxygen values in the canals ranged from 2.4 to 6.2 ml/liter and were similar to those at the control station throughout the year. Oxygen at the bottom 82 LINDALL, FABLE, and COLLINS: CONDITIONS OF UPLAND CANALS STATION 1 iCONTROLi 30 26 22 18 30K STATION 2 26 22 SURFACE n^ BOTTOM JJ^ M 1 1 1 18 30 O ,0 26 22 — 18 Z 30 26 22 18 30 26 22 18 30 26 22 18 1 1 1 I 1 1 1 STATION 3 1 1 ■1 1 1 1 1 1 1 STATION 4 1 1 J 1 1 1 1 J. STATION 5 1 1 1 nri 1 STATION 6 1 1 1 - ri 1 1 1 1 1 1 1 1 OCT NOV DEC JAN FEB MAIi APR MAT JUN JUL AUG SEP 6 4 2 0 6 4 2 s° Z 4 UJ go Q 6 a 0 6 4 2 0 6 4 2 0 STATION 1 ICONTROLi SURFACE [^BOTTOM 1 1 1 1 STATION 2 BO, ^ - STATION 3 1 STATION 4 STATION 5 T STATION 6 1 1 OCT NOV DEC JAN FEB MAR APR MAY JUN JUl AUG SEP Figure 3. ^Monthly salinity at the surface and bottom of all hydrologic stations, October 1971 -November 1972. of the canals was always less than at the surface with the single exception of Station 6 in June. Moreover, about 50% of the bottom samples taken throughout the year at stations farthest from the bayou (Stations 3-5) contained less than 2.0 ml/liter of oxygen; several were anoxic or nearly so. At Station 6, closest to the bayou, oxygen levels were never observed to be less than 2.1 ml/liter. Trent et al. (1972) also reported oxygen depletion at inland portions of housing development canals in Galveston Bay, Tex., during the summer. Results of the previous year's study showed se- vere oxygen depletion in the canals during the summer months following a red tide (caused by Gymnodinium breve) outbreak (Lindall et al. 1973). In that study decaying fish killed by the red tide placed additional oxygen demand on the sys- tem and precluded the determination of the extent to which dissolved oxygen would have been de- pressed in the absence of red tide. In the present study, no red tide occurred, but oxygen was again Figure 4. — Monthly dissolved oxygen at the surface and bottom of all hydrologic stations, October 1971-November 1972. depleted at the bottom of the most inland stations in the canals during the summer. In fact, low dis- solved oxygen occurred more frequently and over a longer period of time (October 1971 and May through September 1972) than in the previous year. FISHES AND MACROINVERTEBRATES Thirty-eight species and 9,502 individuals of vertebrates and invertebrates were collected in the canals during the year (Table 1). Of the 38 species, 34 were finfish, 1 was the diamondback terrapin, Malaclemys terrapin, and 3 were com- mercially important invertebrates (blue crab, Callinectes sapidus; pink shrimp, Penaeus duorarum; and brief squid, Lolliguncula brevis). Fourteen of the 34 species of finfish did not occur in the previous year's catch. These 14 species, how- ever, made up less than 1% of the total catch. 83 FISHERY BULLETIN: VOL. 73, NO. 1 Table 1 — Monthly occurrence and number of individuals of vertebrates and invertebrates collected with otter trawl at all stations from October 1971 through September 1972. Total Species Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug, Sept. No. % Vertebrates: Anchoa mitchilli 429 1.360 3.368 1,516 296 360 0 1,005 42 6 22 449 8,853 93.2 Anchoa hepsetus 0 0 0 0 0 0 0 208 1 0 1 0 210 2.3 Bairdiella chrysura 0 0 19 0 0 0 4 29 41 1 1 0 95 1.1 Chaetodipterus faber 0 0 9 0 0 0 1 1 16 5 1 1 34 05 Syngnathus scovelli^ 0 0 0 0 4 1 0 7 5 0 0 0 17 0.2 Lagodon rhomboides 0 0 0 0 0 2 6 3 3 0 0 0 14 0.1 Gobiosoma bosci 0 1 1 0 2 4 0 2 3 0 0 0 13 0.1 Pogonias cromis 0 0 1 1 0 2 8 0 0 0 1 0 13 0.1 Lucania parva^ 0 0 1 1 2 2 0 4 0 0 0 0 10 0.1 Diapterus plumierl^ 0 0 4 0 2 0 0 0 0 0 2 1 9 0.1 Eucinostomus argenteus 0 0 7 1 0 0 0 1 0 0 0 0 9 0.1 Menticirrhus americanus 0 1 0 0 0 0 0 5 0 2 1 0 9 0.1 Cynoscion arenarius 0 0 0 0 0 0 0 3 0 1 2 2 8 0.1 Orthopristis chrysoptera 0 0 0 0 0 0 4 1 0 0 0 0 5 0.1 Eucinostomus gula 0 0 0 0 0 0 0 0 4 0 0 0 4 0.1 Trinectes maculatus^- 0 0 4 0 0 0 0 0 0 0 0 0 4 0.1 Cynoscion nebulosus 0 0 0 0 0 0 0 3 0 0 0 0 3 0.0 Microgobius gulosus 0 0 0 0 0 0 0 0 0 1 1 1 3 0.0 Opisthonema oglinum 0 0 0 0 0 0 1 2 0 0 0 0 3 0.0 Sciaenops ocellata 0 0 1 1 1 0 0 0 0 0 0 0 3 0.0 Sphoeroides nephelus 0 2 0 0 1 0 0 0 0 0 0 0 3 0.0 Chilomycterus schoepfi^ 0 0 0 0 0 0 0 1 1 0 0 0 2 0.0 Epinephelus itajara^ 0 0 0 1 0 0 0 0 0 1 0 0 2 0.0 Gobiosoma robustum'' 0 0 0 0 0 1 0 1 0 0 0 0 2 0.0 Hippocampus erectus^ 0 0 0 0 1 0 0 0 0 0 1 0 2 00 Leiostomus xanthurus 0 0 0 0 0 2 0 0 0 0 0 0 2 00 Malaclemys terrapin 0 0 0 0 0 0 0 0 1 0 0 1 2 0.0 Monacanthus hispidus'' 0 0 0 0 0 1 0 1 0 0 0 0 2 0.0 Syngnathus louisianae'' 0 0 0 0 0 0 0 1 0 1 0 0 2 0.0 Archosargus probatocephalus 0 0 0 0 0 0 0 0 1 0 0 0 1 0.0 Arius felis 0 0 0 0 0 0 0 0 0 0 1 0 1 0.0 Elops saurus^ 0 0 0 0 1 0 0 0 0 0 0 0 1 0.0 Harengula pensacolae^ 0 0 0 0 0 0 0 1 0 0 0 0 1 0.0 Hippocampus zosterae^ 0 0 0 0 0 0 0 0 0 1 0 0 1 0.0 Lactophrys quadricornis^ 0 0 0 0 0 0 0 0 0 1 0 0 1 0.0 Invertebrates: Lolliguncula brevis 0 0 1 0 1 11 7 17 21 13 0 15 86 0.9 Callinectes sapidus 0 4 11 8 5 4 10 2 3 0 1 1 49 0.5 Penaeus duorarum 0 1 5 6 1 0 0 1 1 3 3 2 23 0.2 Total species 1 6 13 8 12 11 8 22 14 12 13 9 38 Total individuals 429 1,369 3,432 1,535 317 390 41 1,299 143 36 38 473 9.502 100.0 'Did not occur in catches from August 1970 through August 1971 (Lindall et al. 1973). The four species of finfish caught in greatest abundance represented 97% of the total number of specimens (Table 1). They were the bay anchovy, Anchoa mitchilli; striped anchovy, A. hepsetus; silver perch, Bairdiella chrysura; and Atlantic spadefish, Chaetodipterus faber. The bay anchovy alone accounted for more than 93% of the total number caught. In the previous year's study (Lindall et al. 1973) the four dominant species of fish, representing 92% of the catch, were bay anchovy (7,557 indi- viduals— 72%); spotfin mojarra, Eucinostomus argenteus (921 individuals — 8.8%); spot, Leiostomus xanthurus (821 individuals — 7.8%); silver jenny, Eucinostomus gula (372 individuals — 3.5%). The latter three species combined consisted of only 15 individuals in the present study and made up only 0.2% of the catch (Table 1). Each of these three species is a bottom feeder (Darnell 1958; Springer and Wood- burn 1960; Carr and Adams 1973), and the pro- longed period of low dissolved oxygen at the bot- tom of the canals probably accounted for the 99% reduction in their numbers. The brief squid was the most abundant inver- tebrate (54% of all invertebrates collected) and made up about 1% of all animals collected during the year. Based on the previous year's catch, the number of squid in the canal system declined by about 78% , while the total numbers of pink shrimp and blue crab remained about the same. Of the 38 species collected during the year, most occurred at Station 4 (28 species), followed by Sta- tion 1 (21 species). Station 3 (18 species), and Sta- tion 2 (14 species). Compared with the previous year, the number of species collected at Stations 1 and 4 were about the same, but those at Stations 2 and 3 declined markedly (30% and 50% respec- 84 LINDALL, FABLE, and COLLINS; CONDITIONS OF UPLAND CANALS lively). We were not surprised to find fewer species at Stations 2 and 3, because these stations are farthest from the bayou and were most affected by the critically low oxygen levels. As evidence, catches at the four trawl stations during the sum- mer period of low dissolved oxygen (July- August) are compared in Figure 5. The vast majority of species and individuals occurred nearest the bayou (Stations 1 and 4) during this period of stress. 2 3 STATIONS Figure 5. — Number of species and individuals caught at each trawl station during the summer period of low dissolved oxygen (June through August 1972). CONCLUSIONS The upland canal system known as Tanglewood Estates is poorly designed with respect to provid- ing year-round, quality habitat for estuarine species offish and shellfish. Apparently caused by prolonged periods of low dissolved oxygen at the bottom of the canals, the numbers of squid {Lolliguncula breuis) and three species of finfish (Eucinostomus argenteus , E. gula, and Leiostomus xanthurus) were drastically reduced in the second year of the system's existence. We believe that the ability of the canal system to pro- vide adequate oxygen for respiration of bottom- dwelling fishes is becoming progressively worse. The main causative factors are: 1) lack of water exchange with the adjacent bayou, 2) water depths greater than the depth of the photic zone, thus preventing photosynthesis by benthic flora, and 3) continuing accumulation of decomposing soft sediments (Hall and Lindall'*). The major advantages of upland canal develop- ment, as opposed to bayfill development, are that bay bottom is not adversely altered and water circulation patterns are not altered significantly. In fact, estuarine area is increased. However, as long as land developers continue to design upland canals with dead ends and excessive depths, ox- ygen depletion and the resulting impoverishment of fauna on or near the bottom can be expected to be a recurring problem in summer months. LITERATURE CITED Barada, W., and W. M. Partington. 1972. Report of investigation of the environmental effects of private waterfront canals. Environ. Inf. Cent. Fla. Conserv. Found., Winter Park, Fla., 63 p. Carr, W. E. S., and C. a. Adams. 1973. Food habits of juvenile marine fishes occupying seagrass beds in the estuarine zone near Crystal River, Florida. Trans. Am. Fish. Soc. 102:511-540. Darnell, R. M. 1958. Food habits of fishes and larger invertebrates of Lake Pontchartrain, Louisiana, an estuarine community. Publ. Inst. Mar. Sci., Univ. Tex. 5:353-416. LiNDALL, W. N., Jr., J. R. Hall, and C. H. Saloman. 1973. Fishes, macroinvertebrates, and hydrological condi- tions of upland canals in Tampa Bay, Florida. Fish. Bull., U.S. 71:155-163. McNuLTY, J. K., W. N. Lindall, Jr., and J. E. Sykes. 1972. Cooperative Gulf of Mexico estuarine inventory and study, Florida: Phase I, area description. U.S. Dep. Com- mer., NOAA Tech. Rep. NMFS CIRC-368, 126 p. Springer, V. G., and K. D. Woodburn. 1960. An ecological study of the fishes of the Tampa Bay area. Fla. State Board Conserv., Mar. Lab., Prof. Pap. Ser. 1, 104 p. Trent, W. L., E. J. Pullen, and D. Moore. 1972. Waterfront housing developments: their effect on the ecology of a Texas estuarine area. In M. Ruivo (editor). Marine pollution and sea life, p. 411-417. Fishing News (Books) Ltd., Lond. ■'Hall, J. R., and W. N. Lindall, Jr. Benthic macroinvertebrates and sedimentology of upland canals in Old Tampa Bay, Fla. Unpubl. manuscr., 121 p. Gulf Coastal Fisheries Center, Na- tional Marine Fisheries Service, Panama City, FL 32401. 85 A REEVALUATION OF THE COMBINED EFFECTS OF TEMPERATURE AND SALINITY ON SURVIVAL AND GROWTH OF BIVALVE LARVAE USING RESPONSE SURFACE TECHNIQUES R. Gregory Lough^ ABSTRACT The combined effects of temperature and salinity on larval survival and growth of Crassostrea virginica, Mercenaria mercenaria, and Mulinia lateralis as reported in the literature were critically examined using response surface techniques. The late veliger larvae generally have a greater tolerance to both temperature and salinity than the developing embryos. Each species shows its own characteristic change in temperature-salinity tolerance as it develops and approaches the range normally tolerated by the adults as it matures. Maximum growth of the veliger larvae required higher temperatures and somewhat higher salinities than maximum survival. Dif- ferences in temperature-salinity ranges estimated for maximum survival and growth were significantly different for all three species. In each case growth showed a significant temperature- salinity interaction. Response surface plots are given for early larval survival and late veliger survival and growth. Inferences of tolerance studies are made to the fields of pollution and aquaculture. Recent studies of the combined effects of tem- perature and salinity on early development of bivalve larvae have been done by Davis and Calabrese (1964) for Crassostrea virginica and Mercenaria mercenaria, Brenko and Calabrese (1969) for Mytilus edulis, Calabrese (1969) for Mulinia lateralis, Lough and Gonor (1971, 1973a, b) for Adula californiensis , and Goodwin (1973) for Panope generosa. However, only Lough and Gonor (1973a, b) have critically examined the effects of temperature and salinity on bivalve larval life by multiple regression analyses and the fitting of response surfaces to survival, growth, and respiration of early and late stage larvae. The use and evaluation of this response surface technique in marine ecology has been reviewed in detail by Alderdice (1972). This technique not only facilitates the prediction of an organism's response to a wide range of untested conditions but also visually represents any change in its response at various stages of development. The experimental data from the above mentioned species have been critically analyzed by response surface techniques to reevaluate the combined effects of temperature and salinity on larval survival and growth. The results for Crassostrea virginica, Mercenaria mercenaria, and Mulinia lateralis are given in this paper. 'School of Oceanography, Oregon State University, Corvallis, OR 97331. METHODS The mathematical model used in the analyses was of the form: y = 6o + 6i ^T) + 62 (S) + 63 {T"") + 64 {S^) + b^{T X S) where Y = percentage survival or growth 60 = a constant T = linear effect of temperature S = linear effect of salinity T^ = quadratic effect of temperature S^ = quadratic effect of salinity T X S = interaction effect between tempera- ture and salinity The coefficients in the model (6's) were esti- mated by a stepwise multiple regression com- puter program contained in the Oregon State University Statistical Program Library. F-levels were set equal to zero to enter and remove var- iables. This allowed all variables to come into the equation by a forward selection process, their order of insertion determined by using the par- tial correlation coefficient as a measure of their importance. The contribution a variable makes in reducing the variance of the equation can also be considered by looking at the various values given as the program proceeds. One of the more useful is the square of the multiple correlation Manuscript accepted February 1974. FISHERY BULLETIN: VOL. 73, NO, 1, 1975. 86 LOUGH: TEMPERATURE-SALINITY EFFECTS ON BIVALVE LARVAE coefficient, R^, defined as the sum of squares due to regression divided by the total sum of squares corrected for the mean. It is often stated as a percentage, lOOR^. The larger R^ is, the better the fitted equation explains the variation in the data. Values ofi?^ can be compared at each stage of the regression program. A ^-test also is made indicating the equality of the individual regres- sion coefficients to zero and their level of significance. The calculated regression coefficients from a particular equation were fitted by computer to a full quadratic equation in temperature and salinity in order to print a contour diagram of the response surface. The computer program was instructed to print 20% contour intervals, wide enough to exclude the approximate ± 10% experi- mental error reported by the authors. Tempera- ture and salinity scales on all plots were set to range from 0 to 40 in order to facilitate response comparison and to allow the overall form of the surface to be visualized. Contours extrapolated beyond the experimental data are given as dotted lines. Analysis of covariance methods, as used in Lough and Gonor {1973a, b), were used to test the significance of the difference between the estimated polynomials for early and late larval survival and between late survival and growth. RESULTS Crassostrea virginica Davis and Calabrese (1964) first reared the larvae for 2 days at six levels of temperature and nine levels of salinity to study the effect of these factors on early development, or the period from fertilization to approximately the veliger stage. To learn what effect these same combinations of temperature and salinity had on late larval development, larvae were initially reared from eggs for 2 days at normal seawater conditions (24.0°C, 27.5%o) and then transferred at the veliger stage to the experimental condi- tions. Tables of the multiple regression analyses are given in the Appendix and will not be referred to in this section. Survival to 2 days of develop- ment was affected most by the linear and quad- ratic effects of salinity and the linear effect of temperature. Maximum survival of the 2-day-old larvae (80% survival contour) was estimated to occur at temperature and salinity conditions be- tween 19° and 30.5°C and 19 and 30%o (Figure 1), which is in good agreement with the experimental results. The analysis of survival of 10-day-old larvae, after 8 days of rearing at experimental conditions, indicated that the linear and quadratic effects of temperature and the quadratic effect of salinity significantly affected survival. Maximum survival after 8 days (60% survival contour) was estimated to occur above 21°C and between 8 and 30.5"/oo (Figure 2). The 10-day-old larvae showed a tolerance to much higher temperature and a wider salinity range than the 2-day-old larvae. Analysis of covariance showed a significant dif- ference (1% level) between the 2- and 8-day survival polynomials further substantiating that the range of temperatures and salinities tolerated by the late veliger larvae were significantly different than that of the early embryos. Growth of the larvae during 8 days was af- fected most by the interacting effect of tempera- ture and salinity and the quadratic effect of salinity. Maximum growth (100% response con- tour) was estimated to occur at temperatures and salinities above 19°C and 33%o (Figure 3). There was a significant difference (1% level) between the polynomials estimated for 8-day 40 30 UJ cr H < K UJ 20 lO- 10 20 30 SALINITY ( %o ) 40 Figure 1. — Response surface estimation of percent survival of Crassostrea virginica larvae after 2 days of develop- ment at experimental temperature and salinity combinations given in Davis and Calabrese (1964). 87 FISHERY BULLETIN: VOL. 73, NO. 1 40 < 40 10 20 30 40 SALINITY ( %„ ) Figure 2. — Response surface estimation of percent survival of Crassostrea virginica veliger larvae after 8 days of development at experimental temperature and salinity com- binations given in Davis and Calabrese (1964). IT 3 < 30 20- 100 ,80' y" 4 -^ A y y .y y ^ y y / / 10 20 30 S ALIN I TY (%„) 40 Figure 3. — Response surface estimation of percent growth of Crassostrea virginica veliger larvae after 8 days of development at experimental temperature and salinity com- binations given in Davis and Calabrese (1964). survival and growth indicating a significantly higher salinity range is required for optimum growth than is required for optimum survival. An analysis of combined 8-day-survival and growth data indicated that the linear effect of temperature, the interacting effect of tempera- ture and salinity, and the quadratic effect of salinity were the more important factors ex- plaining the data. Optimum (80^ contour) tem- perature and salinity conditions for maximizing both larval survival and growth was estimated at above 30°C and between 18 and 35%o. Mercenaria mercenaria The same experimental design with the ex- ception of nine levels of temperature and six levels of salinity was used by Davis and Calabrese (1964) to study the larval tolerance of this species. Survival to 2 days of development, or from fertilization to veliger stage larvae, was affected most by the quadratic effects of salinity and temperature, and the interacting effect of tem- perature and salinity. Tbe response surface for 2-day-old larvae clearly shows the skewed con- tours resulting from the interaction effect (Figure 4). Maximum survival to 2 days of development (100% survival contour) was esti- mated to occur at temperatures and salinities above 7.2°C and 28%o. Their experimental data show maximum survival between 17.5° and 30°C at a salinity of 27%o. I- < UJ 0- 40 30- 20 10 \ V ^ ^ ^ 0^ ^. s^ \ -iv- 10 20 30 SALINITY ( %„ ) 40 Figure 4. — Response surface estimation of percent survival of Mercenaria mercenaria larvae after 2 days of develop- ment at experimental temperature and salinity combinations given in Davis and Calabrese (1964). 88 LOUGH: TEMPERATURE-SALINITY EFFECTS ON BIVALVE LARVAE 40n 30 a: < 20 lO- 10 20 SALIN I TY (%. 30 40 Figure 5. — Response surface estimation of percent survival of Mercenaria mercenaria veliger larvae after 10 days of development at experimental temperature and salinity combi- nations given in Davis and Calabrese (1964). Late larval survival after 10 days of rearing at the experimental conditions indicated that the linear and quadratic effects of salinity and the interacting effect of temperature and salinity 40 30 3 < 20 • lO- 20 SALIN I TY (%„ 30 40 Figure 6. — Response surface estimation of percent growth of Mercenaria mercenaria veliger larvae after 10 days of development at experimental temperature and salinity com- binations given in Davis and Calabrese (1964). were the more important factors affecting sur- vival. Maximum survival of these 12-day-old larvae (80% survival contour) was estimated to occur between temperatures and salinities of 19° and 29.5°C and 21 and 29%o (Figure 5). Although the late larvae had a much narrower temperature tolerance than the developing em- bryos, the late larvae showed a significantly greater tolerance to low salinity. This difference in tolerance of these two life stages was further substantiated by the fact that there was a sig- nificant difference (1% level) between the 2- and 10-day survival polynomials. Growth of the larvae during the 10-day ex- perimental period was most affected by the inter- acting effect of temperature and salinity and by the linear and quadratic effects of tempera- ture, and the linear effect of salinity. Maximum growth (80% contour) was estimated to occur at temperatures and salinities between 22.5° and 36.5°C and 21.5 and 30o/oo (Figure 6). There was a significant difference (1% level) between the polynomials estimated for 10-day survival and growth indicating that the higher tempera- tures and salinities required for optimum growth are significantly different than those conditions estimated for optimum survival. Larval survival and growth estimated by these techniques above the experimental temperature and salinity of 32.5°C and 27.0%o are questionable. Higher temperature and salinity levels need to be added to the experimental design to more carefully define the response surface. The combined 10-day survival and growth analysis indicated that they were affected by all of the variables of temperature and salinity, but by salinity more than by temperature. Optimum temperature and salinity conditions (80% contour) for maximizing both larval survival and growth to 12 days was estimated at 21.5° to 33°C and 22 to 31%o. Mulinia lateralis Six levels each of temperature and salinity were used to investigate the tolerances of early and late development of this species by Calabrese (1969) in the same manner as used for the other species. Survival of the early embryos for 2 days under the experimental conditions was affected by all the variables except the interacting effect of 89 FISHERY BULLETIN: VOL. 73, NO. 1 temperature and salinity. Maximum survival of the 2-day-old larvae (807f contour) was estimated to occur at temperatures between 18.5° and 24.5°C and salinities between 22 and 28.5%o (Figure 7). The analysis of survival after 6 to 8 days of rearing beyond the veliger stage indicated that the linear and quadratic effects of temperature and the interacting effect of temperature and salinity were the more important variables affecting survival. Response surface estimation predicted 80% survival at temperatures between 8.5° and 26.5°C and salinities above 12o/oo (Figure 8). A significant difference (1% level) was cal- culated by the analysis of covariance for the 2- and 6- to 8-day survival polynomials. The veliger larvae showed a much greater tolerance to low temperatures and a wider range of salini- ties than the early embryos. Growth of the veliger larvae was most affected by the interacting effect of temperature and salinity, the quadratic effect of salinity, and the linear effect of temperature. Maximum growth (60% contour) was estimated to occur at tempera- tures between 18° and 38°C and salinities above 16.5%o (Figure 9). The axis of the growth con- tours are observed to lie diagonal to the factor axes showing the effect of the temperature- 40 3 < a: a. 5 30 20 10 ' ■ t ■ 1 _ -'- ■^ "IT" — ■ — _ ,^ -0' __ , . _ — - — — - ^-_ _ ^ ^ ^-" .20 ^^■''' """ ^ ~~ ^ - y 40 ,y--^ ^^^^^ ■"^ -^ ,^ y y y -' / y'' ^ / y^ " / ^ "^ ^ .^° X ^ \ / / / \ 1 / / ' '■ ' / / / / / / ' 80 J \ \ \ v /■ N \ S^ \ ^^ \ ^^^^^ ^- -s -^ ^^^ V ■ — s ^ — _^ — -^ ^ -V ^ ^ ^ ^-.. "- -- ^ __ ^ ^^'^ __ t^ -i' ■ 1 7 ~ ^, 1 1 1 ■ ^' 10 20 30 SALINI TY (%„) 40 Figure 8. — Response surface estimation of percent survival of Mulinia lateralis veliger larvae after 6 to 8 days of development at experimental temperature and salinity com- binations given in Calabrese (1969). salinity interaction. There was a significant (1% level) difference between the polynomials estimated for the 6- to 8-day survival and growth indicating that the higher temperatures required ClJ tr I- < LiJ Q. 2 40 30 tr < 20- 10- 10 20 30 SALINI TY (%„) 40 Figure 7. — Response surface estimation of percent survival of Mulinia lateralis larvae after 2 days of development at experimental temperature and salinity combinations given in Calabrese (1969). Figure 9. — Response surface estimation of percent growth of Mulinia lateralis veliger larvae after 6 to 8 days of development at experimental temperature and salinity com- binations given in Calabrese (1969). 90 LOUGH: TEMPERATURE-SALINITY EFFECTS ON BIVALVE LARVAE for optimum growi:h were significantly different than those required for optimum survival. Analysis of the combined 6- to 8-day survival and growth indicated that the interacting effect of temperature and salinity and the linear effect of temperature were the more important variables explaining the data, although only 30.4% of the variance was explained by the combined polynomial. Optimum temperature and salinity conditions (80% contour) for maximizing both larval survival and grovd;h to 8 to 10 days of age were predicted between 20° and 26°C and 23 and 32%o. DISCUSSION Despite the fact that the adults of the three species studied are euryhaline to varying degrees, their early embryos and larvae have a com- paratively narrow salinity range. Early larvae of Mercenaria mercenaria appear to be much more tolerant to high temperatures than the other two species, but require essentially oceanic salinities. The older larvae, having been reared from fertilization to the veliger stage at optimum conditions, now appear to have a generally greater tolerance to both temperature and salin- ity. The late larvae of C. virginica appear to tolerate a higher temperature range than the early larvae while Mulinia lateralis late larvae seem to tolerate best temperatures at the lower end of its range. Late larvae of Mercenaria mercenaria are able to tolerate low salinities somewhat better than the early larvae, but their temperature range is quite restricted. The observed progressive change in their temperature- salinity tolerance with time approaches the range normally tolerated by the adults. This same progressive change was observed for the larvae of Adula californiensis by Lough and Gonor (1973a, b). The range of temperature-salinity conditions estimated for maximum growth was significantly different from that estimated for maximum sur- vival of the same late stage larvae. Maximum predicted growi;h occurred at higher temperatures and at somewhat higher salinities than those for maximum survival for all three species studied. All three species showed a significant temperature-salinity interaction effect for growth. Growth, classically, is positively corre- lated with temperature up to some limit; how- ever, the role of salinity appears to complicate the temperature effect. The combining of late larval survival and growth to maximize both responses seems in- tuitively pleasing as one would expect a compro- mise situation in nature. An organism probably can operate most effectively when it is in a set of environmental conditions which maximize all its biological responses. It has been shown by Lough and Gonor (1973a, b) that temperatures for maximum growth response may be an ab- normal stress environment which ultimately results in high mortality. Similarly, low tempera- tures may be suitable for larval survival but not necessarily highly productive for recruit- ment and growth to the adult population. Al- though the optimum temperatures and salinities usually can be estimated from the raw data, the statistical techniques used in this study allow one to define and interpret an organism's response to a matrix of environmental factors and to determine whether the response(s) between stages of development or sampling intervals is significantly different. INFERENCES Tolerance studies of various stages or at various times in the life history of an organism are especially important to pollution studies. Dif- ferent stages of crab larvae have been shown to have different temperature-salinity tolerances of ecological significance (Costlow et al. 1960, 1962, 1966). This study demonstrates that dif- ferent periods in the life of bivalve larvae also differ in their tolerance to temperature and salinity. The determination of water quality standards based on only one stage in the life of an organism is not realistic. All stages of development are important, particularly when the synergistic effect of a pollutant is studied. Davis and Hidu (1969) found it was necessary to evaluate the effects of pesticides on all stages of clam and oyster larvae as their tolerances are markedly different. The field of aquaculture also may benefit from these tolerance studies. Based on this study a long-term experimental program should be under- taken to maximize both survival and growth recognizing that different stages of an organism may have different optimum conditions. Possibly, 91 FISHERY BULLETIN: VOL. 73, NO. 1 larvae should be reared at one set of conditions from fertilization to veliger stage and then transferred to another set of conditions for the late stages. Juvenile clams may have yet another set of optimum conditions different than those of the late larval stage. The larvae and the adults are tw^o distinct morphological and physiological organisms and occupy distinctly different ecologi- cal environments. Recent work by Costlow and Bookhout (1971) on the cyclic effect of temperatures on the larval development of an estuarine mud crab, Rhithro- panopeus harrisii, emphasizes the need for more research on the fluctuating environmental var- iables that normally occur in nature. The possible stimulating or inhibiting effect of fluctuating temperatures on bivalve larval survival and grow^th in relation to both pollution and aqua- culture should be investigated in the future. ACKNOWLEDGMENT This research was supported in part by the National Oceanic and Atmospheric Administra- tion (maintained by the U.S. Department of Commerce) Institutional Sea Grant 2-35187. LITERATURE CITED Alderdice, D. F. 1972. Factor combinations. Responses of marine poikilo- therms to environmental factors acting in concert. In 0. Kinne (editor), Marine ecology, Vol. 1, Part 3, p. 1659-1722. Wiley-Interscience, Lond. Brenko, M. Hrs., and a. Calabrese. 1969. The combined effects of salinity and temperature on larvae of the mussel Mytilus edulis. Mar. Biol. (Berl.) 4:224-226. Calabrese, A. 1969. Individual and combined effects of salinity and temperature on embryos and larvae of the coot clam, Mulinia lateralis (Say). Biol. Bull. (Woods Hole) 137: 417-428. Costlow, J. D., Jr., and C. G. Bookhout. 1971. The effect of cyclic temperatures on larval development in the mud-crab Rhithropanopeus harrisii. In D. J. Crisp (editor). Fourth European Marine Biology Symposium, p. 211-220. Cambridge Univ. Press, Lond. Costlow, J. D., Jr., C. G. Bookhout, and R. Monroe. 1960. The effect of salinity and temp>erature on larval development of Sesarma cinereum (Bosc) reared in the laboratory. Biol. Bull. (Woods Hole) 118:183-202. 1962. Salinity-temperature effects on the larval develop- ment of the crab Panopeus herbstii Milne-Edwards, reared in the laboratory. Physiol. Zool. 35:79-93. 1966. Studies on the larval development of the crab, Rhithropanopeus harrisii (Gould). I. The effect of salinity and temperature on larval development. Physiol. Zool. 39:81-100. Davis, H. C, and A. Calabrese. 1964. Combined effects of temperature and salinity on development of eggs and growth of larvae of M. mercenaria and C. virginica. U.S. Fish Wildl. Serv., Fish. Bull. 63:643-655. Davis, H. C, and H. Hidu. 1969. Effects of pesticides on embryonic development of clams and oysters and on survival and growth of the larvae. U.S. Fish Wildl. Serv., Fish. Bull. 67:393-404. Goodwin, L. 1973. Effects of salinity and temperature on embryos of the Geoduck clam iPanope generosa Gould). Proc. Natl. Shellfish. Assoc. 63:93-95. Lough, R. G., and J. J. Gonor. 1971. Early embryonic stages of Adula californiensis (Pelecypoda: Mytilidae) and the effect of temperature and salinity on developmental rate. Mar. Biol. (Berl.) 8:118-125. 1973a. A. response-surface approach to the combined effects of temperature and salinity on the larval develop- ment of Adula californiensis (Pelecypoda: Mytilidae). L Survival and growth of three and fifteen-day old larvae. Mar. Biol. (Berl.) 22:241-250. 1973b. A response-surface approach to the combined effects of temperature and salinity on the larval development of Adula californiensis (Pelecypoda: Mytili- dae). II. Long-term larval survival and growth in relation to respiration. Mar. Biol. (Berl.) 22:295-305. 92 LOUGH: TEMPERATURE-SALINITY EFFECTS ON BIVALVE LARVAE APPENDIX Appenbix Table 1. — Multiple regression analyses of Crassostrea virginica larvae. Step Level of Level of number Variable R2 F-level df significance Coefficient 7-value significance 2-day survival 1 S 0.661 77.857 (1,40) 1% 32.8617 8.007 1% 2 S2 .793 24.818 (2.39) 1% -0.6234 6.706 1% 3 Tz .795 .356 (3,38) N.S. -0.5195 5.715 1% 4 7 .889 31 397 (4,37) 1% 27.7755 5.821 1% 5 7 X S Constant .894 1.875 (5,36) N.S. -0.0971 -643.9149 1.369 N.S. 8-day survival 1 r 0.426 38.600 (1.52) 1% 10.221 2 992 1% 2 72 .493 6.781 (2,51) 1% -0.2006 3.008 1% 3 7 xS .501 .778 (3,50) N.S. 0.0996 2.492 5% 4 S2 .625 1.612 (4,49) N.S. -0.1455 3.687 1% 5 s Constant .643 2.430 (5,48) 8-day g N.S. rowth 2.6621 -104.389 1.559 N.S. 1 T X S 0.642 93.200 (1.52) 1% 0.2450 6.645 1% 2 S2 .907 144.539 (2,51) 1% -0.2329 6.401 1% 3 s .918 7.139 (3,50) 1% 4.1512 2.636 5% 4 7 .919 .317 (4,49) N.S. 7.5246 2.389 5% 5 Constant .927 5.365 (5,48) 8-day survival 1% and growth -0.1425 -152.2672 2.316 5% 1 7 0.431 80.437 (1,106) 1% 8.8727 2 182 5% 2 7 . S .528 21.559 (2,105) 1% 0.1723 3.620 1% 3 S2 .602 19.110 (3,104) 1% -0.1892 4028 1% 4 72 .619 4.587 (4,103) 1% -0.1715 2.160 5% 5 S Constant .629 2.807 (5,102) 5% 3.4066 -128.3283 1.676 N.S. Appendix Table 2. — Multiple regression analyses of Mercenaria mercenaria larvae. Step Level of Level of number Variable fl2 F-level df sig nificance Coefficient f-value significance 2-day s survival 1 S2 0.561 43.513 (1.34) 1% 0.2439 0.611 N.S. 2 7 X S .581 1.555 (2.33) N.S. 0.3947 2.345 5% 3 72 .640 5.223 (3,32) 1% -0.1219 2.039 N.S. 4 T .648 .678 (4,31) N.S. -2.7229 0.819 N.S. 5 S Constant .653 .449 (5,30) 10-day surviva N.S. 1 -12.2188 110.3864 0.670 N.S. 1 S 0.488 49.560 (1.52) 1% 15.6884 4.021 1% 2 S2 .594 13.307 (2,51) 1% -0.4142 4.570 1% 3 72 .609 1.894 (3,50) N.S. -0.2630 4.916 1% 4 7 X S 732 22.546 (4,49) 1% 0.2111 3.295 1% 5 7 Constant .769 7.591 (5,48) 10-day growth 1% 7.4766 -201.8315 2.755 1% 1 7 X S 0.631 88.900 (1.52) 1% 0.2438 4.532 1% 2 72 .739 21.109 (2,51) 1% -0.3305 7.262 1% 3 7 .829 26.270 (3,50) 1% 12.3631 5.363 1% 4 S2 .841 3.706 (4.49) 5% -0.3702 4.835 1% 5 s Constant .885 18.518 (5,48) 1% 14.0885 -288.6339 4.303 1% 10-day survival and growth 1 S 0.463 91.215 (1,106) 1% 14.5902 4.532 1% 2 S2 .535 16.411 (2,105) 1% -0.3876 5.164 1% 3 7 X S .590 13.881 (3,104) 1% 0 2316 4.378 1% 4 72 .685 31.236 (4,103) 1% -0.2987 6.719 1% 5 7 Constant .736 19.522 (5,102) 1% 9.9556 -243.0117 4.418 1% 93 FISHERY BULLETIN; VOL. 73, NO. 1 Appendix Table 3. — Multiple regression analyses of Mulinia lateralis larvae. Step Level of Level of number Variable fl2 f-level df significance Coefficient f-value significance 2-clay survival 1 S 0.156 6269 (1.34) 5% 14,4237 4,949 1% 2 S2 .390 12.705 (2.33) 1% -0,2942 4,916 1% 3 T X S .421 1.708 (3,32) N.S. 0.0284 0,554 N.S. 4 P .478 3.359 (4.31) 5% -0.3256 5.440 1% 5 r Constant .709 23769 (5.30) 6- to 8-day 1% survival 13 1265 -240.8807 4.875 1% 1 -n 0.353 18.540 (1.34) 1% -0.1976 7.190 1% 2 T .627 24.193 (2,33) 1% 6.0749 4.914 1% 3 T X S .716 10.002 (3,32) 1% 0,0307 1.305 N.S. 4 S2 .724 .898 (4,31) N.S. -0.0781 2.843 1% 5 S Constant .760 7.013 (5,30) 6- to 8-day 1% growth 3.5437 -2.8961 2.648 5% 1 T X S 0.498 33.698 (1.34) 1% 0.0993 2.646 5% 2 S2 .605 8.979 (2,33) 1% -0.1258 2.867 1% 3 P .641 3.220 (3,32) 5% -0.2120 4.833 1% 4 r .765 16.222 (4,31) 1% 8.9352 4,528 1% 5 s Constant .796 4,584 (5,30) 1% 4.5735 -113.4013 2,141 5% 6- to 8-day survival and growth 1 T X S 0.102 7.963 (1.70) 1% 0.0650 1.438 N.S. 2 ■n .120 1.408 (2,69) N.S. -0.2048 3.876 1% 3 T .262 13,046 (3,68) 1% 7.5051 2.377 1% 4 S2 .278 1.491 (4,67) N.S. -0.1020 1.930 N.S. 5 S Constant .304 2.489 (5,66) 5% 4,0586 -58,1487 1.578 N.S. 94 SWIM-BLADDER STATE AND STRUCTURE IN RELATION TO BEHAVIOR AND MODE OF LIFE IN STROMATEOID FISHES Michael H. Horn^ ABSTRACT Fourteen of the 15 genera of stromateoid fishes possess a relatively small (0.6-3.4% of body volume), euphysoclistous swim bladder which forms early in life (3-5 mm standard length) and regresses in all genera except pwssiblyA'^omeMs before maturity (150-200 mm standard length) is reached. The organ is thus an almost exclusive characteristic of the juveniles which occupy surface layers (upper 100-150 m) in coastal and oceanic waters. The gas gland of the swim bladder consists of cuboidal to irregularly shaped cells 6-46 m m in greatest dimension. The retia mirabilia range in length from 0.4 to 2.0 mm and in diameter of individual capillaries from 4 to 10 fi m. The area of the gas gland and the length of the retia relative to the size of the swim-bladder lumen are great compared to the same in other epipelagic fishes and are similar to those of deeper living, mesopelagic fishes. The relatively large gas-secreting complex is considered to be an adaptation for rapid and efficient gas secretion in maintaining hydrostatic equilibrium as the juvenile fishes swim in the surface layers, frequently in association with floating objects, where pressure changes are greatest with depth. Swim-bladder loss accompanies changes in behavior and mode of life and is part of the transition from the juvenile to the adult stage of life. Hovering and high maneuverability as principal components of locomotor behavior in juveniles give way to continuous swimming in adults which are generally independent of floating objects and occupy a greater depth range. The relative length of the paired fins changes with age and varies among the species. Peprilus triacanthus and P. simillimus , negatively buoyant, active swimmers, have long pectoral fins as adults whereas Schedophilus medusophagus , a neutrally or nearly neutrally buoyant, slow-moving fish, has short pectoral fins. Both P. simillimus and S. medusophagus have high levels of lipid which may serve to replace the swim bladder in a buoyancy function when the fishes are adults. The swim bladder (or gas bladder), a gas-filled organ unique to bony fishes, has its greatest func- tional significance as a hydrostatic device, i.e., one that provides neutral or nearly neutral buoyancy to the fish. It is one of the most plastic of vertebrate organs (Marshall 1960) and occurs in a great di- versity of fishes from a variety of habitats. The swim bladder is not necessary for life as it is ab- sent in many fishes, but according to Fange ( 1 966) about one-half of the 20,000 existing species have it as adults and even more as larvae or juveniles. The organ, owing to its diversity of form and wide- spread occurrence, should reflect in its presence or absence and structure the behavior and mode of life of the fishes possessing it. In stromateoid fishes, the swim bladder regresses in 13, possibly 14, of the 15 genera, and the regression seems to be associated with other morphological changes and changes in mode of life (Horn 1970a). 'Department of Biology, California State University, Fuller- ton, CA 92634. Manuscript accepted Meirch 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. The swim bladder of stromateoid fishes has re- ceived little mention in the literature partly due to its absence or reduced state in adults. Goode and Bean (1895) and Jordan and Evermann (1896) stated that the organ was "usually absent" in the Stromateidae, and the former as well as Grey (1955) reported its absence in the Tetragonuridae. Fowler (1936) in his treatment of several stromateoid genera indicated that the swim blad- der was "present or absent." Goode and Bean (1895) stated that it was present in Nomeus as did Haedrich (1967) for Ariomma. Based upon an ex- amination of approximately one-half of the species in the group, I have found the organ to be present at some stage (larval and/or juvenile) in the life of all stromateoid genera except Pampus . Even in Pampus it may be present at an early stage since larvae or small juveniles (< 20 mm SL, standard length) were not studied. The perciform suborder Stromateoidei con- sists of 6 families, 15 genera, and about 60 species (Haedrich and Horn 1972) and is characterized by toothed saccular outgrowths in the gullet and by 95 FISHERY BULLETIN: VOL. 73, NO. 1 small teeth approximately unilateral in the jaws (Haedrich 1967). The larvae and juveniles occur mainly in the surface layers of the ocean and are frequently associated with animate or inanimate floating objects. The adults, ranging in maximum size from about 30 to 120 cm, form a diverse group of temperate and tropical species which variously occupy a wide range of depths in coastal and oceanic waters (including mesopelagic and bathy- pelagic levels). The Centrolophidae (six genera) are either coastal or oceanic, the Stromateidae (three genera) are coastal, the Amarsipidae (one genus), Nomeidae (three genera), and Tet- ragonuridae (one genus) are oceanic, and the Ariommidae (one genus) are benthopelagic on the continental shelf and slope. The purposes of the present paper are to 1) describe the morphology and histology of the stromateoid' swim bladder, 2) compare the dimensions and capabilities of the stromateoid swim bladder with that of other fishes of similar and different habitats, and 3) discuss the relationship of swim-bladder state and structure to the behavior and mode of life of stromateoids based upon the results of the present and other studies. MATERIALS AND METHODS The majority of specimens examined (Table 1) for swim-bladder structure and other morphologi- cal detail were preserved although fresh or frozen material of several species was studied. Observa- tions on the behavior of certain species were made and are briefly described in appropriate sections of the paper. Specimens in addition to those personally col- lected were obtained from the following institu- tions: British Museum (Natural History), London; Museum of Comparative Zoology, Harvard Uni- versity, Cambridge, Mass.; Institute of Oceano- graphic Sciences, Wormley, England; Natural History Museum of Los Angeles County, Los Angeles, Calif.; Scripps Institution of Oceanog- raphy, La Jolla, Calif.; Southwest Fisheries Center, National Marine Fisheries Service, NOAA, La Jolla; Woods Hole Oceanographic Institution, Woods Hole, Mass.; and Zoological Museum, Copenhagen, Denmark. Swim-bladder dimensions were measured with an ocular micrometer in either a dissecting or compound microscope or, in large specimens, with 96 Table 1. — Stromateoid specimens examined for swim bladder and other morphological characteristics. Number of Size range Family and species specimens (mm SL) Centrolophidae: Hyperoglyphe antarctica 4 23.4- 34.9 Hyperoglyphe perciformis 2 35.8. 47.7 Schedophilus huttoni 1 22.9 Schedophilus maculatus 2 70.2, 77.5 Schedophilus medusophagus 7 10.4-285.0 Centrolophus maoricus 3 15.1-127.8 Centrolophus niger 2 124.0, 231.0 Icichthys lockingtoni 20 3.5-268.0 Seriolella punctata 2 132.0, 162.6 Seriolella violacea 3 12.8- 84.0 Psenopsis cyanea 2 94.2. 104.2 Stromateidae: Stromateus brasiliensis 7 75.7-167.3 Stromateus fiatola 9 12.5- 93.5 Stromateus stellatus 5 17.5- 99.4 Peprilus burti 3 57.4- 95.1 Peprilus paru 10 28.6-123.0 Peprilus simillimus 17 2.0-135.0 Peprilus triacanthus 10 12.0-120.4 Pampus argenteus 7 24.6- 49.0 Pampus chinensis 6 25.4- 67.3 Amarsipidae: Amarsipus carlsbergi 3 22.0- 67.5 Nomeidae: Cubiceps caeruleus 1 18.5 Cubiceps carinatus 1 8.5 Cubiceps gracilis 10 16.5-330.0 Nomeus gronovii 13 11.6-142.7 Psenes arafurensis 2 17.0, 18.0 Psenes cyanophrys 17 9.1-120.0 Psenes maculatus 1 33.6 Psenes pellucidus 2 26.9, 34.8 Unidentified (probably Psenes) 5 3.4- 12.3 Tetragonuridae: Tetragonurus cuvieri 13 4.0-242.0 Ariommidae: Ariomma bondi 7 20.9-124.3 Ariomma indica 1 597 Ariomma melanum 2 each 134.1 Ariomma regulus 3 123.3-150.0 Ariomma sp 1 16.5 (either /^. bondi or A. melanum) Total number of specimens 204 dial calipers. Swim-bladder and body volumes were determined by displacement and/or, for the former, calculated on the assumption that the bladder was a prolate spheroid (u = 4/3TTab^, where a and b are the major and minor semiaxes (see Capen 1967)). Volume measurements were made from swim bladders that were in most cases well expanded. Ten percent was allowed for shrinkage of preserved material. Transverse or longitudinal serial sections of the swim bladder of 13 genera and species were cut at S-jum thickness and stained with haemalum and eosin. Buoyancy determinations were made by weigh- ing each fish in air and in water of known temper- ature and salinity. Results are expressed as the percentage of the air weight that each fish weighed in seawater. 1 HORN: SWIM-BLADDER STATE AND STRUCTURE RESULTS Swim-Bladder Structure The stromateoid swim bladder is of the physoclistous, two-chambered type usually found in perciform fishes (Horn 1970a) (Figure 1). The delicate, thin-walled sac lies in the upper part of the body cavity above the gut and below the kidney and is closely invested by the dorsal peritoneum. A muscular diaphragm (not always visible) divides the bladder into anterior and posterior chambers (Figures 1, 2), the latter of which serves a gas-resorbing function (a euphysoclistous condition). The gas gland, as- sociated with the anterior chamber, typically forms a U-shape and may be single or divided into two or more lobes (Figure 1). Cells composing the gland are cuboidal to irregular in shape and usually in two or more layers (Figures 2-5). Some cells appear to be either syncytial or of the giant type found widely distributed in marine euphysoc- lists (Fange 1953) and in some deep-sea fishes (Marshall 1960). The retia mirabilia are unipolar, Figure 1. — Ventral view of the swim bladder of 11 species of stroma teoids (all drawn to same scale), gg, gas gland (slightly flattened and expanded); rm, rete mirabile; ra and rv, retial artery and retial vein; ac, anterior chamber; d, diaphragm; pc, posterior chamber. A, Ariomma bondi, 24.2 mm SL; B,Centrolophus maoricus, 15.1 mmSL; C, Tetragonurus cuvieri , 28.8 mmSL; D, Seriolella violacea , 12.8 mm SL; E,Cubiceps gracilis, 30.5 mm SL; F, Schedophilus medusophagus , 17.4 mm SL (lateral and ventral view); G, Psenes cyanophrys ,11.5 mm SL; H, Nomeus gronovii ,26.4 mm SL; I, Stromateus fiatola , 34. 1 mm SL; J, Icichthys lockingtoni ,16.3 mm SL; K, Hyperoglyphe antarctica, 34.9 mm SL. 97 FISHERY BULLETIN: VOL. 73, NO. 1 vvv .:^, I Figure 2. — Sagittal section of the anterior chamber of the swim bladder of Ariomma bondi. d, diaphragm; gg, gas gland; rm, rete mirabile. (From same specimen as Figure lA.) f Figure 3. — Sagittal section of the gas secreting complex of the swim bladder of Cubiceps gracilis . gg, gas gland; rm, rete mirabile. (From same specimen as Figure IB.) 98 HORN: SWIM-BLADDER STATE AND STRUCTURE Figure 4.— Gas gland cells and retial capillaries of the swimbladder oi Cubiceps gracilis. Arrow points to retial capillary between gas gland cells. (From same specimen as Figures IE and 3.) Figure 5. — Gas gland of Peprilus triacanthus, 16.5 mm SL, showing arrangement of cells. (Transverse section.) 99 FISHERY BULLETIN: VOL. 73, NO. 1 i.e., the artery and vein subdivide to form parallel capillaries which enter the gas gland (Figures 3, 4, 6). Retial orientation varies from a position parallel to the long axis of the swim bladder to one that is perpendicular (Figure 1). Size characteris- tics of swim-bladder components are given in Table 2. Distinctive features of the swim bladder of the six stromateoid families are described below. Centrolophidae Swdm-bladder volume in the fishes examined varied from less than 1% of body volume in Schedophilus to greater than 3% in Hyperoglyphe. Size (greatest dimension) of the gas gland cells ranged from 10-17 ^m in Hyperoglyphe and Seriolella to 20-40 /uni in Schedophilus. Retial length ranged from 1.1 mm in Seriolella to 2.0 mm Figure 6. — Transverse section through the rete mirabile of the swim bladder oi Tetragonurus cuvieri. (From same specimen as Figure IC.) Table 2. — Size characteristics of the swim-bladder lumen, gas gland, and rete mirabile in 12 species of stromateoids. Rete mirabile Lumen Gas gland Total Total Size L X W Vol Vol Area Cell size Capillary Length number of capillary Species (mm SL) (mm) (mm^) (%) (mm^) (/um) diam (pm) (mm) capillaries length (m) Hyperoglyphe antarctica 34.9 9.8 X 2.7 37.0 3.4 5.4 10-17 7-10 1.3-2.0 2,000 2.6 -4.0 Schedophilus medusophagus 17.4 2.1 X 0.8 — — 1.8 20-40 8-10 1.3 800 1.04 Icichthys lockingtoni 16.3 2.4 X 1.2 2.0 3.0 2.1 15-20 6-8 1.4 1,000 1.4 Seriolella violacea 12.8 3.5 X 1.1 — — 1.0 10-17 4-8 1.1-1.2 1,200 1.3 -1.4 Stromateus fiatola 34.1 9.1 X 2.0 — __ 2.9 10-20 6-8 1.8-2.0 1,500 2.7 -3.0 Peprilus triacanthus 16.5 2.5 X 1.0 3.0 2.3 0.8 6-10 5-6 0.8 600 0.48 Amarsipus carlsbergi 22.0 2.0 X 0.8 0.7 — 1.0 8-20 4-8 0.7-0.9 1,100 0.77-0.99 Cubiceps gracilis 30.5 6.6 X 2.3 21.0 3.3 8.7 25-40 8-10 1.1-1.9 3,000 3.5 -5.7 Nomeus gronovii 26.4 6.7 X 1.0 3.5 0.7 1.6 — — 0.4-0.6 — — 27.4 6.8 X 1.3 — — 2.5 10-30 5-6 0.5-0.6 2,000 1.0 -1.2 Psenes cyanophrys 14.1 3.7 X 1.2 2.5 2.1 0.7 — 1.0 — . — 18.6 5.1 X 1.6 — — 6.0 25-40 5-6 0.8-0.9 1,500 1.2 -1.35 Tetragonurus cuvieri 28.8 3.8 X 1.0 2.0 0.6 0.6 8-20 4-10 1.3 1,000 1.3 Ariomma bondi 23.3 6.5 X 1.3 7.0 2.9 3.6 — 0.9 — — 24.2 5 0 X 1.4 5.0 1.7 3.7 20-46 8 0.6-0.9 2,500 1.5 -225 100 HORN: SWIM-BLADDER STATE AND STRUCTURE in Hyperoglyphe. Retial orientation was generally parallel to the long axis of the bladder, and the retial bundle either remained single anteriorly as in Icichthys (Figure IJ) or variously branched into smaller bundles perpendicular to the long axis as in Hyperoglyphe (Figure IK). The rete bundle of Schedophilus had a sharp turn near the posterior end producing a sigmoid outline (Figure IF). Swim-bladder shape which depends to a large de- gree upon secretory and absorptive states varied from elongate with a large posterior chamber to short and bulbous with either a small posterior chamber or no posterior chamber visible. Stromateidae The organ was similar in structure and shape to that of the Centrolophidae. The gas gland cells of Peprilus triacanthus were small (6-10 /jm) and were arranged in loops and rings (Figure 5). In one P. triacanthus (16.5 mm SL) the retial blood ves- sels formed a single bundle posteriorly which ex- panded anteriorly over the gas gland whereas in two somewhat larger juveniles (22.2 and 33.9 mm SL) the retia were more nearly perpendicular to the long axis of the bladder and consisted of 7 or 8 distinct branches. Amarsipidae The swim bladder was similar to that of the Centrolophidae. The rete originated posteriorly as a single bundle and divided anteriorly into 7 or 8 distinct branches before entering the gas gland. Nomeidae Swim-bladder volume ranged from 0.7% in Nomeus to 3.3% of body volume in Cubiceps and gas gland size from 10-30 ;u m in Nomeus to 25-40 /L( m in Cubiceps and Psenes. Retial length varied from 0.4 mm in Nomeus to 1.9 mm in Cubiceps. The retia were divided into several branches and in position were more nearly perpendicular than parallel to the long axis of the bladder (Figure IE, H). Small juvenile Psenes cyanophrys (9.1 and 14.1 mm SL) had retia almost parallel to the long axis of the bladder (Figure IG) whereas larger juveniles (e.g., 60.8 mm SL) tended to have retia which were more nearly perpendicular to the long axis and more highly branched. The pattern, seen also in Peprilus triacanthus , may be part of the re- gression process that the swim bladder undergoes. Tetragonuridae The sac was small (0.6% of body volume) and elongate. The retial bundle was parallel to the long axis of the bladder and, as in Schedophilus medusophagus , had an S-shaped turn near the posterior end (Figure IC). The gas gland was rela- tively small and located at the anterior end of the lumen. Ariommidae The swim bladder was relatively large (up to 2.9% of body volume) and elongate. The gas gland cells were in the upper range of size (20-46 ;u m) among the stromateoids examined, and the retia were broad, fanlike and perpendicular to the long axis of the bladder (Figure lA). Size at Swim-Bladder Inflation The swim bladder becomes functional early in the life of stromateoids. Whether the larval fishes gulp air at the surface or whether gas is secreted to initially fill the bladder was not determined. Examination of larvae of four genera indicated that the organ in one species was almost completely developed at 3.0 mm SL and in three others at slightly larger sizes. Specimens of Peprilus simillimus as small as 3.0 mm SL had what appeared to be a fully developed swim bladder whereas in smaller individuals, e.g., 2.7 mm SL, the sac was inflated but the gas gland and retia were incomplete. The bladder was absent in a fish of 2.4 mm SL. Aseriesof larvae, 3.4-5.0 mm SL, of an unidentified species of Psenes had an inflated swim bladder, and larvae ofTetragonurus cuvieri as small as 4.0 mm SL had a gas-filled sac which was visible through the semitransparent body wall. Individuals of Icichthys lockingtoni, 5.0 and 7.5 mm SL, had an inflated sac and an apparently fully developed gas gland and retial complement. Swim-Bladder Regression The swim bladder regresses, becomes nonfunc- tional, and finally disappears before the adult stage is reached in all stromateoid genera except Pampus in which the organ is apparently absent and possibly Nomeus in which the largest individual examined (142.7 mm SL) had a functional swim bladder. The regression is a 101 FISHERY BULLETIN: VOL. 73, NO. 1 Table 3. ^-Surface area of the gas gland (mm^) and length of the rete mirabile (mm) relative to swim-bladder volume (mm^ or ml) or dimension (length x width, in mm) in the European eel, certain shallow-sea and deep-sea fishes, and in 12 species of stromateoids. Total capillary length (m) = retial length x number of retial capillaries. Retial lengths of stromateoids are means of individuals for each species. Species European eel' Anguilla anguilla Shallow-sea (epipelagic):' Cypsilurus cyanopterus Danichthys rondeletll Exocoetus volitans Petalichthys capensis Hyporhamphus sp. Scomberesox saurus Gadus minutus Capros aper Deep-sea:' Gonostoma denudatum Pollichthys mauli Bonapartia pedaliota Vinciguerria attenuata Vinciguerria nimbaria Argyropelecus aculeatus Polyipnus laternatus Astronesthes niger Astronesthes similis Myctophum punctatum Benthosema suborbitale Lampanyctus guntheri Diaphus rafinesquei Melamphaes megalops Stephanoberyx monae Chiasmodon niger Sfromateoids: Hyperoglyphe antarctica Schedophilus medusophagus Icichthys lockingtoni Seriolella violacea Stromateus fiatola Peprilus triacanthus Amarsipus carlsbergi Cubiceps gracilis Nomeus gronovii Psenes cyanoplirys Tetragonurus cuvieri Ariomma bondi 'Data from Marshall (1960). Gas gland area Retial len gth Total capillary Size X 1,000/swlm-bladder X 1,000/swim- ■bladder length/swim-bladder (mm SL) volume (mm3) dimension vol ume (ml) — — — 30 290.0 8 0.5 214.0 8 1.3 — 159.0 6 5 — — — 1.5 — 112.0 40 8 2 13 18 — — — — 81.0 76 43.0 — 125 — 67.0 — 71 — 43.5 140 34 — — — — 150 23.0 140 38 50 36.0 — 143 100 41.0 250 111 — 104.0 170 — — 71.0 200 40 20 24.0 500 — — 53.0 330 142 21 250 — 56.0 60 30 83.5 — 63 — 104.0 250 71 — 34.9 140 63 89 17.4 — 1,000 — 16.3 1,000 500 700 12.8 — 333 — 34.1 — 100 — 16.5 250 333 167 22.0 — 500 — 30.5 500 100 214 26.4 500 77 — 11.5 1.000 333 — 28.8 330 333 650 23.3 500 111 380 gradual process which makes difficult the determination of the exact time of loss of function. Several stages are recognizable in the process although they vary in appearance, and both the stages and the overall regression vary in duration among and within species. Estimated ranges of fish size during which regression occurs in nine stromateoid species are given in Table 4. Early in the regression the gas gland contracts and thickens and the sac begins to decrease in size. Later the gas gland and retia mirabilia become atrophied as the cells and capillaries lose integrity (Figure 7). A yellowish-white material, possibly lipid, frequently invests the gas gland. Finally, the swim-bladder wall is resorbed, and the gas secreting and absorbing complexes become indis- tinct. A large stromateoid (> 100-200 mm SL, see Table 4) may have either a small irregularly shaped mass of yellowish-white material lying in the dorsal mesentery (Figure 7) as the only remnant of the swim bladder or no visible trace at all of the organ. DISCUSSION Relative Dimensions and Capabilities of the Swim Bladder Volume Mean percentage volumes were relatively small, 0.6-3.4% (Table 2), and generally below the 3.1-5.7 range for the swim bladder calculated by Alexander (1966) to be necessary for neutral buoyancy in seawater. A number of mid-water fishes also have swim bladders of low volume 102 HORN: SWIM-BLADDER STATE AND STRUCTURE Table 4. — Size ranges during which swim-bladder regression occurs in nine species of stromateoids and during which the same species have been observed in association with animate (mainly coelenterate) or inanimate floating objects. Former ranges derived from data of present study and latter from sources listed. Size during Size during regression association Associated species Species (mm SL) (mm SL, FL, or TL)' or object Source^ Centrolophus niger 50-75 30-40 XL Rhizostoma pulmo Mansueti 1963 103-477 SL Mola mola Mackay 1972 Icichthys lockingtoni 40-65 55 TL Pelagia noctiluca Mansueti 1963 16.3 SL Pelagia noctiluca Specimen label (MCZ) 180.5 TL Pelagia noctiluca Mansueti 1963 Stromateus fiatola 50-75 1 0-40 TL Rhizostoma pulmo Mansueti 1963 (including S. lasciatus) 1 0-40 TL Cotylorhiza tuberculata Mansueti 1963 127 TL Unidentified medusa Mansueti 1963 12.5-28.2 SL Cassiopeia carbonica Specimen label (ZMC) Peprilus triacanthus 75-100 10-20 TL Chrysaora quinquerirrha Mansueti 1963 51-64 TL Cyanea capillata Mansueti 1963 50-73 TL Unidentified medusa Mansueti 1963 Peprilus paru (including 60-100 18-69 TL Chrysaora quinquecirrha Mansueti 1963 P. alepidotus) 13TL Chrysaora quinquecirrha Mansueti 1963 147 TL Unidentified medusa Mansueti 1963 28.6 SL Unidentified medusa Specimen label (BMNH 1956.11.12.12) Cubiceps gracilis 40-75 42 SL Unidentified medusa Specimen label (BMNH 76.6.21-2) Nomeus gronovii Swim bladder 20 FL Drifting raft Gooding and Magnuson 1967 present 51-76 TL Physalia pelagica Mansueti 1963 at si 50 SL 127-152 TL Stomolophus meleagris Mansueti 1963 142.7 SL Physalia pelagica NIO specimen, 1 Stn. 6688-3 HMS Discovery II Psenes cyanophrys 110-130 15-124 FL Drifting raft Gooding and Magnuson 1967 (including P. paciticus) 10-133 SL Flotsam Hunter and Mitcfiell 1967 Tetragonurus cuvieri 40-60 34 SL Unidentified medusa Specimen label (BMNH 76.6.21.23) 'SL = standard lengtfi, FL = fork lengtfi. TL = total length. ^MCZ = Museum of Comparative Zoology, Harvard University. ZMC = Zoological Museum, Copentiagen. BMNH = British Museum (Natural History), London. (Capen 1967; Kleckner and Gibbs 1972) and even a relatively small gas-filled sac provides some de- gree of buoyancy which may be significant de- pending upon what other lift or buoyancy devices are utilized. Larval and juvenile stromateoids, the stages which have the organ, have a different mode of life (see below) and are in some species at least probably less dense than adults. Only a 1% reduction in specific gravity of a fish lowers the required percentage volume for neutral buoyancy from 3.1%, the lower value in Alexander's (1966) calculated range (which was based upon specific gravities of adults), to 2.2% (Horn 1970a). Thus, even a small swim bladder would be an important contribution to buoyancy. Data on specific gravi- ties of young stromateoids which might help to ex- plain the range of percentage volumes found with- in the group are not yet available. Gas Gland The area of the gas gland relative to swim-blad- der volume is similar to that in a number of deep- sea fishes and much greater than that of a series of epipelagic or shallow-sea ones (Table 3). Marshall (1960) stated that the large gas gland of deep-sea fishes, especially vertical migrators, may be an adaptation for rapid gas secretion as the fish de- scends. Even though juvenile stromateoids occur only in the epipelagial, the adaptive significance of a large gas gland would be the same for them as for deep vertical migrators since stromateoids range over depths in the upper 100-150 m where pressure changes are greatest (e.g., the pressure at 10 m is 2 X that at the surface). Maintaining association with animate floating objects as many stromateoids do requires that fishes range, even if slowly, over depths in the surface layers and in so doing secrete gas during descents if the hydrostat- ic advantage of the swim bladder is to be effected. Thus, the main selective value of the large gas gland may be for making fine adjustments to buoyancy. At least some of the epipelagic fishes listed in Table 3 have a narrow vertical range near the surface and would not require as large a gas gland. The size and structure of the gas gland cells vary widely among stromateoids, a common situation in both shallow- water (Woodland 1911; Fange 1953) and deep-sea fishes (Marshall 1960). Cells measured in stromateoids ranged from 6 to 46 /j m (Table 2), although some other cells in a few species appeared to be multinucleate and syncy- tial or similar to the giant cells (50-150 /jm) de- scribed by Fange (1953) and Marshall (1960). The gland consisted of relatively large cells in a 103 FISHERY BULLETIN: VOL. 73, NO. 1 Figure 7. — Transverse section through the regressed swim bladder of Ariomma indica, 59.7 mm SL. Arrow (1) points to regressed rete mirabile, arrow (2) to regressed gas gland. complex, multilayered arrangement as in Ariom- ma bondi (Figure 2), or, less frequently, of small cells arranged in circles or loops as in Peprilus triacanthus (Figure 5). The functional significance of cells of either different sizes or arrangements is poorly understood. Rete Mirabile Retial length in stromateoids ranged from 0.4 to 2.0 mm (Table 2), similar to the 0.75 to 2.0-mm range listed by Marshall (1972) for upper mesopelagic (200-600 m) fishes. The ratio of retial length to swim-bladder dimension (length x width) (Table 3) as an approximation of relative development is high in stromateoids and similar to that of Marshall's (1960) deep-sea group which includes some vertical migrators. The stromateoid ratio is much higher than that of other epipelagic fishes and demonstrates that the retia, as with the gas gland, which together form the gas-secreting complex, are relatively well developed in stromateoids. In addition, the total length of the retial capillaries (retial length x number of retial capillaries) in relation to swdm-bladder volume is similar to or exceeds that for the eel, Anguilla anguilla, and certain deep-sea fishes (Table 3). Marshall (1972) pointed out that the only flex- ible adaptation to increase the surface available for countercurrent gas exchange is an increase in length of the retial capillaries. An increase in length will not only lead to increased gas ex- change but also slow the rate of bloodflow and so further enhance the efficiency of exchange (Mar- shall 1960). Marshall (1972) showed that the deeper the living space of a fish the greater the absolute length of the retia. On the basis of the pattern of retial length and depth of living de- scribed by Marshall, the predicted depth zone for larval and juvenile stromateoids would be the upper mesopelagial. Besides length, the diameter of the retial capil- laries of stromateoids shows a somewhat greater similarity to that of deep-sea fishes than to other epipelagic fishes. Stromateoids have capillary bores of 4-10 iim (Table 2) whereas epipelagic fishes listed by Marshall (1960) had diameters of greater than 10 ^/m. Deep-sea fishes with large erythrocytes have retial capillaries 7-18 jum in diameter and those with small, nonnucleated erythrocytes such as Maurolicus and Vinciguerria have retial capillaries 2-10 /jm in diameter (Mar- shall 1972). The smaller the diameter the greater the efficiency of gaseous exchange (Marshall 104 HORN: SWIM-BLADDER STATE AND STRUCTURE 1960), although decreased bore is an adaptation Hmited in most fish species by the size of the eryth- rocytes. Swim-Bladder Regression in Relation to Behavior and Mode of Life Data from the present study and other sources on depth distribution, association with floating objects, and locomotion and buoyancy make it pos- sible to formulate a general outline of the changes in behavior and mode of life that accompany the regression of the swim bladder and which are part of the transition from the juvenile to the adult state. Depth Distribution Adult stromateoids generally occupy a wide range of depths either over the continental shelf or in the open ocean, whereas larvae and juveniles of all or nearly all species occur in the surface layers (mainly the upper 100 m) (Haedrich 1967, 1969; Horn 1970a). Larval nomeids (Psenes and Cubiceps) are important constituents of the epipelagic fauna; this is known especially for the eastern tropical Pacific (Ahlstrom 1971, 1972). The Centrolophidae and Tetragonuridae were listed by Ahlstrom (1969) as two of the principal families of deep-sea fishes which had larvae in the surface layers of the California Current region. The stromateid, Peprilus simillimus, occurs mainly in the upper 50 m of coastal waters off California and Baja California (Ahlstrom 1959), and ariommid larvae and juveniles apparently live in the surface layers although the adults are benthopelagic (Horn 1972). Thus, swim-bladder loss occurs as the fishes increase their range of vertical distribution. Association with Floating Objects Beginning at a small size (^ 10 mm SL) and usually ending before maturity is reached (^200 mm SL), stromateoids commonly associate with a wide variety of animate and inanimate floating objects (Mansueti 1963; Haedrich 1967; Horn 1970a) and during the period that the swim blad- der is functional (Table 4). The associations are not obligatory but rather, as Mansueti (1963) de- scribed them, temporary ecological phenomena in which the objects (e.g., jellyfishes or fiotsam) are essentially passive hosts and the fishes active op- portunists. Scyphozoan medusae of several genera form a major group of associates particularly of stromateids and to some extent of centrolophids, nomeids, and tetragonurids (Mansueti 1963). The nomeid, Nomeus gronovii, and the Portuguese man-of-war, Physalia, form the apparently most intimate and enduring of "fish-jellyfish" associa- tions. Certain stromateoids have also been found inside pelagic ascidians (Grey 1955), beneath the ocean sunfish, Mola (Mackay 1972) and beneath floating plants such as Sargassum (Haedrich 1967). Several species occur beneath flotsam, and the nomeid, Psenes cyanophrys , is one of the more abundant fishes under drifting objects (Hunter and Mitchell 1967, 1968; Gooding and Magnuson 1967). Drift associations are not well understood but probably provide one or more ecological ad- vantages such as food, protection, or visual stimuli. Juvenile stromateoids in their coloration and maneuverability are well adapted for life beneath floating objects, especially coelenterates. Young fish typically have a banded, mottled, or blotched pattern whereas adults are generally uniform in color or are dark above and pale below. The dura- tion of the juvenile color pattern is similar to the period when the fishes are associated with floating objects, and the patterns according to Haedrich (1967) serve as protective coloration beneath the shifting shadows of objects, especially jellyfishes. Nomeus which retains its association with floating objects longer than most or all other stromateoids also retains its mottled color pattern in the largest specimens known. Maneuverability and avoidance by the fish ap- pear to be of primary importance in all or most stromateoid-coelenterate associations (Mansueti 1963; Horn 1970a). Peprilus triacanthus placed in tanks with a medusa, Chrysaora quinquecirrha, gradually increased the amount of time spent near the jellyfish and after 72 h remained within a 4-cm distance of the bell at least 75% of the time (Horn unpubl. data). Hovering and rapid turning were significant parts of the locomotor behavior of the fish in avoiding the tentacles of the medusa. Con- tact of the skin of the fish by the tentacles resulted in nematocyst firing as evidenced by the clinging of the tentacles to the fish's body causing the fish to rapidly swim away. In a two-way feeding relation- ship, P. triacanthus frequently nibbled at the manubrium and tentacles of the medusa, while weakened or otherwise slow-moving fish were captured and ingested by the medusa. 105 FISHERY BULLETIN: VOL. 73, NO. 1 Although Lane (1960) reported that Nomeus can survive doses ofPhysalia toxin as much as 10 times the amount that would kill other fishes of the same size and type, Nomeus is stung if forced into contact with the tentacles (Lane 1960) and can be killed if touched by the tentacles according to Zahl (1952). Maul (1964) found that Schedophilus (= Mupus) ovalis also suffered large weals on the body from nematocysts when in contact with Physalia and that safety for the fish must be due in part to its ability to avoid con- tact with the tentacles. Mansueti (1963) concluded that in all fish-jellyfish associations the former skillfully maneuver between tentacles and gen- erally avoid being stung but that contacts are inevitable. Locomotion and buoyancy The differences in locomotor behavior found be- tween juvenile and adult stromateoids that have been observed illustrate the importance of ma- neuverability for juveniles and correspond to swim-bladder loss and increased independence of floating objects as maturity is reached. The paired fins are important locomotor devices among stromateoids. The pectoral fins are moved in a rotary manner for maintaining position in juve- niles of Peprilus triacanthus and Schedophilus medusophagus when hovering beneath floating objects (pers. obs.) and sculled for effecting con- tinuous swimming at less than maximum speeds in these species (Horn 1970b, unpubl. obs.) and in other stromateoids such as Cubiceps gracilis (Fig- ure 8). I have observed adults of both P. triacan- thus and P. simillimus in public aquaria and calculated that the pectorals are used at least 80-909?^ of the time as a main propulsive force at cruising speeds. The pelvic fins which may be absent (all stromateid species except one) or small (as in certain centrolophids) are well devel- oped in juveniles of certain species. Pelvics are large in Nomeus and apparently important for increasing maneuverability and enhancing pro- tective coloration for a fish living among the tenta- cles of Physalia. The relative length of the paired fins changes with age and varies among the species (Haedrich 1967; Horn 1970b). Extremes are represented by P. triacanthus and S. medusophagus (Figures 9, 10). In P. triacanthus (which lacks pelvic fins) the relative length of the pectoral fin increases rapidly until the fish reaches about 75-80 mm Figure 8. — Cubiceps gracilis, 68 mm SL, swimming in plastic container and using pectoral fins as principal locomotor force. Swim bladder of this fish partially regressed (see Table 4). Specimen captured at the surface in the North Atlantic. SL beyond which the fin length ceases to increase (Figure 9). This fish size is in the range of that when the swim bladder regresses and the fish deserts its coelenterate host (Table 4). Individuals of P. triacanthus greater than 75-80 mm SL are negatively buoyant (see below) and swim continu- ously using mainly the long pectorals which also generate dynamic lift. In S. medusophagus the relative length of the paired fins decreases with age (Figure 10), a pattern opposite that of P. triacanthus. The swim bladder regresses in a size range of about 40-60 mm SL corresponding closely to the size interval during which the marked change in paired fin length occurs and apparently during which the fish deserts its coelenterate host 106 HORN: SWIM-BLADDER STATE AND STRUCTURE ■^ • • ' • • 35 • • • • • •Ik • • • !• • • • • •• • •*• • • • .• ; • • • • C3 •j. • • • ••4 • Z 0 • • • -1 • • I* • • • • • • o oc 25 • •• • - 2 • t z ^ • 10 PECTORAL • PELVIC A • • _RP_, AP — t ▲ ▲ 100 200 STANDARD LENGTH mm 300 400 Figure 10. — Scatter diagram of pectoral fin and pelvic fin lengths as a percentage of standard length in Schedophilus medusophagus. RP and AP as in Figure 9. 107 FISHERY BULLETIN: VOL. 73, NO. 1 (Mansueti 1963). Unlike P. triacanthus , S . medu- sophagus becomes neutrally buoyant or nearly so (see below) and has a poorly ossified skeleton and soft musculature (Bone and Brook 1973). Adult S. medusophagus swim slowly and continuously in near anguilliform manner and with only a minor part of the propulsive force provided by the small pectorals (pers. obs.). Because of the fish's low density, little or no lift is required from locomotor activity. Changes in the level of buoyancy and in the nature of the buoyancy mechanism may coincide with swim-bladder loss and other changes occur- ring as stromateoids mature although the data are as yet insufficient to permit conclusions to be reached. Peprilus triacanthus and a closely re- lated species, P. simillimus, are negatively buoyant as adults (weight in water 1.4-2.3% of weight in air) (unpubl. data). Juvenile .S. medusophagus (85-200 mm SL) are slightly nega- tively buoyant (Bone and Brook 1973) whereas a larger (285 mm SL) specimen was found to be neutrally buoyant in surface seawater (unpubl. data). Large amounts of lipid have been found in adults of both P. simillimus andS. medusophagus especially in the skull and spine (Lee et al. in press). Bone and Brook (1973) found relatively low amounts of lipid in juvenile S. medusophagus , an indication that lipid content may increase with size in this species. Lipids may serve to partially replace the swim bladder in a buoyancy function as the organ regresses in P. simillimus and S. medusophagus, two morphologically and ecologi- cally dissimilar stromateoids. Peprilus simil- limus, an active, continuous swimmer with long pectoral fins, has a relatively well ossified skele- ton, firm musculature, and is negatively buoyant, whereas S. medusophagus, a slow moving, con- tinuous swimmer with short pectoral fins, has poorly ossified bones and soft, loosely packed muscles, and approaches or attains neutral buoy- ancy. Increased lipid content as a buoyancy replace- ment for the swim bladder would be advantageous for P. simillimus, S. medusophagus , and probably other stromateoids that range over the upper sev- eral hundred meters of the water column since the low coefficient of compressibility of lipid compared to gas reduces the stress of pressure changes with depth. Nevenzel et al. (1969) pointed out the ad- vantage of lipid for a vertically migrating mid- water fish, and Butler and Pearcy (1972) discov- ered that in two such species, the myctophids Stenobrachius leucopsarus and Diaphus theta, swim bladder-to-body volumes were inversely re- lated to body size and lipid content indicating that lipids assume the primary buoyancy function as the swim bladder regresses vdth age. An addi- tional advantage of stored lipid, especially tri- glycerides, may be as an energy source (Lee et al. in press). Bone (1973) has suggested that verti- cally migrating myctophids can be grouped into functional types based on swim-bladder state, lipid content, density, and size of the pectoral fins. Stromateoids are not classed as a principal group of vertical migrators partly because of their rela- tive rarity, but many species do have a broad ver- tical range. With more data, it may be possible to divide stromateoids into functional groups accord- ing to characteristics similar to those listed by Bone (1973) for myctophids. ACKNOWLEDGMENTS A number of people have allowed me to examine and in certain cases dissect specimens in their care. My thanks go to N. B. Marshall and Alwyne C. Wheeler, British Museum (Natural History); Richard L. Haedrich, Woods Hole Oceanographic Institution; Julian Badcock, Institute of Oceano- graphic Sciences, England; Elbert H. Ahlstrom and Elaine Sandknop, Southwest Fisheries Center, National Marine Fisheries Service, NOAA; J0rgen Nielsen, Zoological Museum, Uni- versity of Copenhagen; Robert J. Lavenberg, Natural History Museum of Los Angeles County; and Richard H. Rosenblatt, Scripps Institution of Oceanography. I sincerely appreciate the efforts of N. B. Mar- shall (now at Queen Mary College, University of London) who provided working space and facilities at the British Museum (Natural His- tory), passed along a great deal of information on swim bladders, and read the manuscript. The re- viewers gave valuable comments for improvement of the manuscript. Paula K. McKenzie made the drawing of Fig- ure 1. Financial support for this study was provided in part by a NATO postdoctoral fellowship awarded through the National Science Foundation and held at the British Museum (Natural History) and in part by a Sigma Xi Grant-in- Aid of Research and by a Faculty Research Grant awarded by California State University, Fullerton. 108 HORN: SWIM-BLADDER STATE AND STRUCTURE LITERATURE CITED Ahlstrom, E. H. 1959. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. U.S. Fish. Wildl. Serv., Fish. Bull. 60:107-146. 1969. Mesopelagic and bathypelagic fishes in the Califor- nia Current region. Calif Coop. Oceanic Fish. Invest. Rep. 13:39-44. 1971. Kinds and abundance offish larvae in the eastern tropical Pacific, based on collections made on EAS- TROPAC I. Fish. Bull., U.S. 69:3-77. 1972. Kinds and abundance of fish larvae in the eastern tropical Pacific on the second multivessel EASTROPAC survey, and observations on the annual cycle of larval abundance. Fish. Bull., U.S. 70:1153-1242. Alexander, R. McN. 1966. Physical aspects of swimbladder function. Biol. Rev. (Camb.) 41:141-176. Bone, Q. 1973. A note on the buoyancy of some lantern-fishes (Myc- tophoidei). J. Mar. Biol. Assoc. U.K. 53:619-633. Bone, Q., and C. E. R. Brook. 1973. On Schedophilus medusophagus (Pisces: Stroma- teoidei). J. Mar. Biol. Assoc. U.K. 53:753-761. Butler, J. L., and W. G. Pearcy. 1972. Swimbladder morphology and specific gravity of myctophids off Oregon. J. Fish. Res. Board Can. 29:1145-1150. Capen, R. L. 1967. Swimbladder morphology of some mesopelagic fishes in relation to sound scattering. U.S. Navy Electron. Lab., Res. Rep. 1447, 25 p. Fange, R. 1953 . The mechanisms of gas transport in the euphy soclist swimbladder. Acta Physiol. Scand. 30, Suppl. 110, 133 p. 1966. Physiology of the swimbladder. Physiol. Rev. 46:299-322. Fowler, H. W. 1936. The marine fishes of West Africa, based on the col- lection of the American Museum Congo Expedition, 1909—1915, Part II. Bull. Am. Mus. Nat. Hist. 70:609-1493. Goode, G. B., and T. H. Bean. 1895. Oceanic ichthyology. U.S. Natl. Mus., Spec. Bull. 2, 553 p. Gooding, R. M., and J. J. Magnuson. 1967. Ecological significance of a drifting object to pelagic fishes. Pac. Sci. 21:486-497. Grey, M. 1955. The fishes of the genus Tetragonurus Risso. Dana Rep., Carlsberg Found. 41:1-75. Haedrich, R. L. 1967. The stromateoid fishes: systematics and a classification. Bull. Mus. Comp. Zool. 135:31-139. 1969. A new family of aberrant stromateoid fishes from the equatorial Indo-Pacific. Dana Rep., Carlsberg Found. 76:1-13. Haedrich, R. L., and M. H. Horn. 1972. A key to the stromateoid fishes. 2nd ed. WHOI (Woods Hole Oceanogr. Inst.) Tech. Rep. 72-15, 46 p. Horn, M. H. 1970a. The swimbladder as a juvenile organ in stromateoid fishes. Breviora 359, 9 p. 1970b. Systematics and biology of the stromateid fishes of the genus Pepn/us. Bull. Mus. Comp. Zool. 140:165-261. 1972. Systematic status and Eispects of the ecology of elon- gate ariommid fishes (suborder Stromateoidei) in the At- lantic. Bull. Mar. Sci. 22:537-558. Hunter, J. R., and C. T. Mitchell. 1967. Association of fishes with flotsam in the offshore waters of Central America. U.S. Fish Wildl. Serv., Fish. Bull. 66:13-29. 1968. Field experiments on the attraction of pelagic fish to floating objects. J. Cons. 31:427-434. Jordan, D. S., and B. W. Evermann. 1896. The fishes of North and Middle America. Part I. U.S. Natl. Mus., Bull. 47, 1240 p. Kleckner, R. C, and R. H. Gibbs, Jr. 1972. Swimbladder structure of Mediterranean midwater fishes and a method of comparing swimbladder data with acoustic profiles. Mediterr. Biol. Stud., Final Rep. 1:230-281. Smithson. Inst., Wash. Lane, C. E. 1960. The Portuguese man-of-war. Sci. Am. 202:158-168. Lee, R. F., C. F. Phleger, and M. H. Horn. In press. Composition of lipid stores in fish bones: possible function in neutral buoyancy. Comp. Biochem. Physiol. Mackay, K. T. 1972. Further records of the stromateoid fish Centrolophus niger from the northwestern Atlantic, with comments on body proportions and behavior. Copeia 1972:185-187. Mansueti, R. 1963. Symbiotic behavior between small fishes and jellyfishes, with new data on that between the stromateid, Peprilus alepidotus, and the scypho medusa, Chrysaora quinquecirrha. Copeia 1963:40-80. Marshall, N. B. 1960. Swimbladder structure of deep-sea fishes in relation to their systematics and biology. Discovery Rep. 31:1-121. 1972. Swimbladder organization and depth ranges of deep-sea teleosts. Soc. Exp. Biol., Symp. 26:261-272. Maul, G. E. 1964. Observations on young live Mupus maculatus (Giinther) and Mupus ovalis (Valenciennes). Copeia 1964: 93-97. Nevenzel, J. D., W. Rodegker, J. S. Robinson, and M. Kayama. 1969. The lipids of some lantern fishes (Family Myc- tophidae). Comp. Biochem. Physiol. 31:25-36. Woodland, W. N. F. 1911. On the structure and function of the gas glands and retia mirabilia associated with the gas bladder of some teleostean fishes, with notes on the teleost pancreas. Proc. Zool. Soc. Lond. 1911(l):183-248. Zahl, p. a. 1952. Man-of-war fleet attacks Bimini. Natl. Geogr. Mag. 101:185-212. 109 DISTRIBUTION AND RELATIVE ABUNDANCE OF SEVEN SPECIES OF SKATES (PISCES: RAJIDAE) WHICH OCCUR BETWEEN NOVA SCOTIA AND CAPE HATTERAS^ John D. McEachran^ and J. A. Musick^ ABSTRACT Data collected during eight groundfish surveys of the area from Nova Scotia to Cape Hatteras, North Carolina, and during five seasonal surveys of Chesapeake Bight were used to define the distribution and relative abundance of Raja eglanteria, R. garmani, R. laevis, R. erinacea, R. ocellata, R. senta, and R. radiata. Ancillary distributional data for the area from the Straits of Florida to Cape Hatteras and the areas off northern Nova Scotia and the Gulf of St. Lawrence were used qualitatively to extend the distributional study. Raja eglanteria is a Carolinian species abundant north of Cape Hatteras only during the warmer months. Raja garmani, a skate of the outer continental shelf and upjjer slope, consists of two populations which have different temperature preferences. Raja laevis is the most wide- spread species studied and does not appear to be as abundant as the other skates in any region of the study. Raja erinacea, a Virginian to boreal species, occurs from southern Nova Scotia to Cape Hatteras in shallow water but is present at depths down to 384 m. Raja ocellata is a Virginian to boreal species distributed similarly toR. erinacea except that the former is widespread in the Gulf of St. Lawrence and off northern Nova Scotia. Raja senta, a boreal species, fre- quently occurs on the northern offshore banks of Nova Scotia and at temperatures as low as -1.3°C. Raja radiata is a boreal to arctic species. Raja erinacea and R. ocellata are sympatric over the greater part of their ranges as are R. senta and R. radiata. The two pairs of species have complementary distributions. Raja ocellata has slightly lower temperature preferences than R. erinacea, and R. radiata is more widespread and has wider temperature tolerances than R. senta. The genus Raja is represented by R. eglan- teria, R. garmani, R. laevis, R. erinacea, R. ocellata, R. senta and R. radiata along the continental shelf of North America between Nova Scotia and Cape Hatteras, NC. Notes on the occurrence and distribution of these species have been summarized by Bigelow and Schroeder (1953, 1954), Leim and Scott (1966), and McEachran (1973); however, most of this infor- mation is based on scattered regional studies. The present study presents data gathered during comprehensive groundfish surveys of the area from Nova Scotia to Cape Hatteras and defines the distribution and relative abundance of each species, as well as cooccurrence among species. MATERIALS AND METHODS Data used in this study were divided into two categories: 1) quantitative data used to deter- 'Contribution No. 651 Virginia Institute of Marine Science. ^Department of Wildlife and Fisheries Sciences, Texas A & M University, College Station, TX 77843. 'Virginia Institute of Marine Science, Gloucester Point, VA 23062. mine relative abundance of the skates, and 2) qualitative data used only to determine the temperature, depth, and geographical ranges of the skates. Data supplied by National Marine Fisheries Service (NMFS) Biological Laboratory at Woods Hole, Mass. (now Northeast Fisheries Center) and by the Virginia Institute of Marine Science (VIMS) at Gloucester Point, Va. were used to determine relative abundance. The former data consisted of eight groundfish surveys of the con- tinental shelf (27-366 m) from LaHave Bank, off southeastern Nova Scotia, and the Gulf of Maine to Cape Hatteras. A total of 2,247 stations were made during the winters of 1968-70, the summer of 1969, and the autumns of 1967-70 (Table 1) by the RV Albatross IV, except that part of 70-06 was conducted by the RV Delaware II. The survey area was divided into 58 strata according to depth and geographical area, and three or more stations were randomly selected within each stratum per cruise (Figure 1) (Gross- lein 1969). A No. 36 Yankee trawl equipped with a cod end liner of 0.25-inch bar mesh and Manuscript accepted March 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. 110 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES Table 1. — Groundfish surveys conducted by the Biological Laboratory of the National Marine Fisheries Service at Woods Hole, Mass (now Northeast Fisheries Center). No. of Cruise Dates Season stations 67-21 17 0ct.- 9 Dec. 1967 Autumn 271 68-03 4 Mar. - 16 May 1968 Winter 262 68-17 10 Oct.- 26 Nov. 1968 Autumn 275 69-02 5 Mar. - 10 Apr. 1969 Winter 266 69-08 14 July- 18 Aug. 1969 Summer 267 69-11 8 Oct - 23 Nov. 1969 Autumn 276 70-03 12 Mar. - 29 Apr. 1970 Winter 289 70-01 ISOct- 20 Nov. 1970 Autumn 341 Total 2,247 18 inch rollers on the ground rope was towed at 3.5 knots for 0.5 h at each station. Distance of tow averaged 1.75 miles. Prior to data analysis, the 58 sampling strata were grouped into five ecological subareas accord- ing to hydrography and substratum. Schopf and Colton (1966) stated that the southern Nova Scotian shelf, Gulf of Maine, and Georges Bank have different bottom temperatures and faunal assemblages. Although Georges Bank and Nan- tucket shoals (northern section of the mid- Atlantic Bight) have similar bottom tempera- tures and faunal assemblages (Schopf and Colton 1966), the area extending from Georges Bank to Cape Hatteras was subdivided because of its great size. The southern section of the mid- Atlantic Bight consisted of strata 61 to 76; the northern mid-Atlantic Bight was composed of strata 1 to 12 and 25; Georges Bank was made up of strata 13 to 23; the Gulf of Maine included strata 24, 26 to 30, and 36 to 40; and the Nova Scotian shelf consisted of strata 31 to 35, 41, and 42. All four depth zones (27-55, 56-110, 111-183, 184-366 m) were sampled in the first three subareas; the three deeper zones were surveyed in the Gulf of Maine; and only two zones, 56-110 and 111-183 m, were sampled on the Nova Scotian shelf. Preliminary examination of the skate data indi- cated contagion as Taylor (1953) and Roessler (1965) had demonstrated for trawl catch data in general. A logarithmic transformation tends to normalize contagious distributions (Pereyra et al. 1967), so skate counts were transformed to log {X + 1). Transformed values were used to deter- mine the geometric mean numbers (indices of abundance) of skates per stratum per cruise. The indices of abundance were weighted by dividing them by the area of the strata to correct DEPTH ZONES (meters) n * 55 i 1 56-110 ^ III- 183 ■■ >I83 Figure 1. — Strata sampled by the RV Albatross IV and Delaware II, 1967-70. Strata numbers 43-60 were not included in the surveys. Ill FISHERY BULLETIN: VOL. 73, NO. 1 for areal differences between strata. Area of the strata are listed in Table 2. Indices of abundance for all stations, within temperature intervals of 1°C for each of the five ecological subareas, were calculated for each species. Indices were not weighted. Length fre- quencies were calculated for strata sets corre- sponding to each of the four depth intervals (27-55, 56-110, 111-183, 184-366 m) within each of the subareas. This analysis gave the per- centages that each 3-cm length increment con- tributed to the total catch of a species within each of the strata sets of the subareas. Hurlbert's (1969) index of association was used to determine the level of cooccurrence based on presence and absence of two species at the same stations. Species pairs with a significant positive index were compared by the product moment correlation (simple correlation coefficient) to determine if the two species were positively or negatively related by numbers. The correlation indices were computed from transformed abun- dance values [log iX + 1)] at stations where the two species cooccurred. According to Hurlbert (1969), a negative correlation, showing an inverse relationship in numbers of individuals between the species, may indicate that the two species compete for the same resources. The VIMS data included five seasonal ground- fish surveys of the Chesapeake Bight (lat. 38° 43'N to 35°13'N) in 9 to 274 m during the four seasons of 1967 and winter of 1968. This area was divided into grids of lat. 15' by long. 12.5'. A 1-h tow was made in each grid per survey with an Atlantic western trawl without rollers (Musick and McEachran 1972). The Chesapeake Bight was divided into two areas, one north and one south of the Virginia Capes (lat. 37°N) for data analysis. Index of abundance (geometric mean) was computed for each of the species {R. eglanteria, R. garmani, R. erinacea, and R. ocellata) captured during the VIMS survey, by depth zone (0-18, 19-37, 38-73, 74-110, 111-165, 166-274 m) and by tem- perature intervals of 1°C, north and south of lat. 37°N separately. This index was not weighted by area of the depth zone. The qualitative data were obtained from the NMFS Exploratory Fishing and Gear Research Base at Pascagoula, Miss, (now Southeast Fish- eries Center, Pascagoula Laboratory) for the area from the Straits of Florida to Cape Hatteras, Table 2. — Area of sampling strata. Stratum Area Stratum Area Stratum Area number (mi2) number (mi^) number (mi2) 1 2,516 21 424 40 578 2 2.078 22 454 41 3,752 3 566 23 1,016 42 589 4 188 24 2.569 61 1.318 5 1,475 25 390 62 243 6 2,554 26 1.014 63 86 7 514 27 720 64 60 8 230 28 2.249 65 2,832 9 1,522 29 3,245 66 555 10 2.722 30 619 67 86 11 622 31 2,185 68 52 12 176 32 712 69 2,433 13 2,374 33 816 70 1.024 14 656 34 1,766 71 281 15 230 35 1,097 72 105 16 2,980 36 4,069 73 2.145 17 360 37 2,108 74 1,273 18 172 38 2,560 75 139 19 2.454 39 730 76 60 20 1,221 and from the Fisheries Research Board of Canada Biological Station at St. Andrews, New Bruns- wick for the area off northeastern Nova Scotia, including Banquereau, Sable Island Bank, West- ern Bank, and the Gulf of St. Lawrence. Dis- tributional data from south of Cape Hatteras were collected from 1961 to 1968, and data from northeastern Nova Scotia and the Gulf of St. Lawrence were collected from 1960 to 1970. Several vessels and different types of trawling gear were used. Small specimens of/?, erinacea and/?, ocellata are difficult to distinguish (McEachran and Musick 1973), and field personnel often mis- identified them. Records of species not verified by the authors were evaluated with discretion. Records were not used when the correct species could not be determined. RESULTS AND DISCUSSION Seasonal bottom isotherms were plotted from the RW Albatross IV surveys of 1969 because this was the only year that included a summer cruise, and the winter and autumn temperature profiles appeared typical. Temperatures were lowest during the winter survey, and isotherms in the mid-Atlantic Bight tended to parallel the coast line (Figure 2), as stated by Bigelow (1933). During the summer cruise a mass of cold water, surrounded by warmer water, extended south- ward almost to the Virginia Capes, a condition previously described by Bigelow (1933). Tempera- tures were warmest during the autumn survey. 112 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES CAPE HATTERAS / CAPE HATTERAS WINTER SUMMER CAPE '^^ HATTERAS Figure 2. — Bottom isotherms plotted from measurements taken during winter, summer, and autumn 1969 surveys of the RV Albatross IV. Temperatures are in degrees Celsius. 113 FISHERY BULLETIN: VOL. 73. NO. 1 The summer survey was conducted during July and August, and the autumn survey during October and November (Table 1). Waters of inter- mediate depths of both the mid-Atlantic Bight and Georges Bank reach their maximum tempera- tures in October (Bigelow 1927, 1933; Schopf and Colton 1966). Length frequencies by strata sets revealed that small to large specimens of each species were found together and all length sizes w^ere pooled for the distributional analyses of each species. Small specimens of R. erinacea and R. ocellata were seldom captured. The young of these two species may lie outside the sampling region or may be less vulnerable to the gear used. Richards et al. (1963) also noted the absence of young R. erinacea on the fishing grounds of Block Island and Long Island sounds where the larger individuals were abundant. Charts showing the distribution by strata, and histograms shovdng the distribution by tempera- ture were illustrated for the Albatross IV cruises of 1969. Only the four most abundant species {R. erinacea, R. ocellata, R. senta, and/?, radiata) were included. Distribution by temperature and depth zones was illustrated for two species {R. eglanteria and R. garmani) captured during the VIMS survey of the Chesapeake Bight. Raja eglanteria Raja eglanteria was captured from the southern section of the mid- Atlantic Bight to about midway along the eastern coast of Florida. A few indi- viduals were taken in the southern section of the mid-Atlantic Bight on all Albatross IV cruises, except for summer 1969 and winter 1970. During the VIMS survey of the Chesapeake Bight R. eglanteria was more abundant in shoal water during the spring and summer than during the autumn and winter (Figure 3) and was more abundant in the Chesapeake Bight during the summer and autumn than in the winter and spring. It was captured between 5° and 26°C in the Chesapeake Bight and was most abundant between 9° and 20°C (Figure 4). South of Cape Hatteras it was taken from 9° to 27°C. Over its entire range, it was most abundant at depths less than 111 m. Raja eglanteria was captured only at 9 of the 676 stations which were in water deeper than 110 m. It was taken at 5 of the 43 deeper stations during the VIMS survey but at only 4 of the 633 deeper stations south of Cape Hatteras, thus it has a greater tendency to inhabit deeper water in the northern part of its range. Raja eglanteria, a Carolinian species in the sense of Johnson (1934) and Hedgpeth (1957), occurs north of Cape Hatteras all year but is abundant there only during the warmer months. Bigelow and Schroeder (1953) stated that it is most abundant from the sublitoral zone to about 55 m. However, Edwards et al. (1962) captured it in 280 and 329 m off Cape May, N.J. during the wdnter. In autumn, R. eglanteria leaves the embayments and shallow areas of the mid- Atlantic Bight (Bigelow and Schroeder 1953; Schwartz 1961; Massman 1962; Fitz and Daiber 1963; Schaefer 1967) and moves offshore and southward. Raja eglanteria was not captured in the mid-Atlantic Bight during the summer Albatross IV cruise probably because the species is concentrated then at depths less than 27 m. Apparently many individuals that summer in the southern part of the Chesapeake Bight move around Cape Hatteras during the autumn or early winter. The individuals south of Cape Hatteras inhabit slightly warmer water as suggested by Bigelow and Schroeder (1953) and do not appear to move into deeper water during the winter. Dahlberg and Odum (1970) reported that this species is resident year-round in Georgia estuaries. Raja garmani Raja garmani was captured in deep water from Nantucket Shoals to the Straits of Florida. Between Nantucket Shoals and Cape Hatteras it was most abundant in the southern section of Chesapeake Bight. Over the Chesapeake Bight it was found between 33 and 196 m and gen- erally deeper than 73 m (Figure 5), and appeared to move shoreward in the summer. In the Chesa- peake Bight R. garmani was captured at tem- peratures of 6° to 17°C and was most abundant bewteen 9° and 13°C (Figure 6). Between Cape Hatteras and Georgia it was found in 66 to 123 m at 17°C; off Georgia and northern Florida it was captured in 77 to 155 m at 11° to 19°C. From northern Florida to the Straits of Florida it occurred in 99 to 366 m at 17°C, and all but one of the captures were in 183 to 366 m. 114 McEACHRAN and MUSICK; DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES 2-T WINTER 1 96 7 2 7 1 4 12 4 6 1 1 T 2 T 2 16 SPRING 1967 Nort h 37° N CO 7 o (Tt to I 00 to in ? CM I U) 00 !o I at to 1^ oo to IT) <3- u> M 14 8 15 3 I 3 South 37° N 2 -I 2 -I SUMMER 1967 UJ < 3 m < o X UJ o 13 25 North 37° N AUTUMN 1967 2 T 13 7 II 5 II 00 ro 01 to I CO ro in u> 'J- CM t 00 lO a> ro 1^ 00 in ^ N CVI 4 ■ 2 12 15 South 37° N II 12 4 Wl NTER 1968 North 37° N to 01 4 2 13 4 to g s 00 fO ? = CM t Raja eglanteria 2 -I 10 Sou th 37° N 6 DEPTH IN METERS Figure 3. — Index of abundance (geometric mean) of Raja eglanteria captured in Chesapeake Bight during each cruise within each depth stratum. Data collected north and south of lat. 37°N were analyzed separately. The fraction over each bar is the ratio of the number of stations at which the species was captured to the total number of stations in each stratum. Whole numbers represent the number of stations in the strata in which no specimens were captured. 115 FISHERY BULLETIN: VOL. 73, NO. 1 WINTER 1967 UJ o z < o z m < X Ui a 2-1 SPRl N G 1967 2 '2 7eA7l North 37° N =1?1 -I 1 1 15 20 25 C 3 'U South 37° N —I AUTUMN 1967 North 37° N I 2 —1 ^ ' — I 1 1 5 10 15 20 25 C y Qt South 37°N 2 -I WINTER 1968 i.6 -' 4 J 545 Da z-* '^''UV North 37°N T — 1 1 1 10 15 20 25 C South 37°N TEMPERATURE Raja eglanteria Figure 4. — Index of abundance 'geometric mean) of Raja eglanteria captured in the Chesapeake Bight during each cruise within temperature intervals of TC. Data collected north and south of lat. 37°N were analyzed separately. See Figure 3 for explanation of fractions and whole numbers. 116 McEACHRAN and NRTSICK: DISTRIBUTION AND RELATI\'E ABUNDANCE OF SKATES Uf o z < z m < o X Ui o 2 -I WINTER 1967 6 JUL CO o to 00 to ? |4 J. 4 _^ SUM ME R 19 6 7 — o 2 - 4 -> 25 rO ff> 00 ro ? r If) to 2_ 3 n in North 37° N (0 ro "" 1 ' 1 1 L 5 10 15 20 2B C J I I 4 6 South 37° N bl u z < z s < X UJ o SUMMER I 967 ^sin r ~5 10 I. I 26 2 01 I 3 I TJ 3 -I r 20 25 Ji I 4 7 13 2-, WINTER 1968 43 S 4 S46 1 5 4^ AUTUMN 1967 5 T North 37° N 2 I 2 699 2 I -I r — ' — I ' 1 0 5 10 15 20 25C J 1 , J—, 1 1 North 37° N 2-' ■'■4tL, ' 10 Ts 20 ?5 C I I L 21 3 In 3 South 37° N TEMPERATURE |U South 37°N Raja garmani Figure 6. — Index of abundance (geometric mean) of Raja garmani captured in the Chesapeake Bight during each cruise within temperature intervals of 1°C. Data collected north and south of lat. 37°N were analyzed separately. See Figure 3 for explanation effractions and whole numbers. 118 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES Raja garmani probably does not occur regu- larly on the eastern slope of Georges Bank, con- trary to Schroeder (1955), because no specimens were captured there during the Albatross IV cruises. The depth and temperature ranges of 51 to 494 m and 5.3° to 15°C given by Bigelow and Schroeder (1953) are close to those for the area north of Cape Hatteras in the present study. It has more limited depth range and is found in warmer water in the southern part of its range than in the northern part as stated by Bigelow and Schroeder (1953). Staiger (1970) stated that it is found between the 119- and 366-m isobaths on Pourtales Terrace, and north of Pourtales Terrace it occurs in 183 m up the coast of Florida. This species appears to have separate popu- lations, one north and the other south of Cape Hatteras. North of Cape Hatteras mature speci- mens are 335 mm TL (McEachran 1970) to 439 mm TL and south of Cape Hatteras they are mature between 250 and 314 mm TL. The dif- ferences in temperature ranges north and south of Cape Hatteras may be due to differences in physiological requirements of the two populations. Raja laevis Raja laevis was captured from the Gulf of St. Lawrence, along the northeastern coast and off- shore banks of Nova Scotia, to the northeastern coast of Florida. During the Albatross IV cruises it was taken from the Nova Scotian shelf to the southern section of the mid-Atlantic Bight and was most frequently taken in the northern sec- tion of the mid-Atlantic Bight, the eastern part of Georges Bank, eastern Gulf of Maine, and the Nova Scotian shelf. No specimens were ob- tained from the western Gulf of Maine. Seasonal changes in abundance were not evident. In the Gulf of St. Lawrence, i?. laevis was found in 315 m at 4.7°C. Off northeastern Nova Scotia it was caught at depths of 24 to 375 m at 1.2° to 10.7°C. Depths and temperatures at capture for the area from southern Nova Scotia to Cape Hatteras ranged from 38 to 351 m and 3° to 20°C. Raja laevis was caught in 302 to 368 m off northeastern Florida. Raja laevis is the most widespread of the species studied, but too few were taken during this study to elaborate on its distribution. Bigelow and Schroeder (1953) stated that this species occurs from the tidemark to about 750 m at 1.2° to 20°C. The southern limit of its range remains in doubt because of the apparent con- fusion of this species with R. floridana which has been captured from Cape Lookout, N.C. to Dry Tortugas, Fla. (Bigelow and Schroeder 1968). Raja floridana is very similar to R. laevis (Bige- low and Schroeder 1962) and the specimens used to describe R. floridana came from some of the same stations at which Bullis and Thompson (1965) listed R. laevis. The senior author has examined the specimens identified as R. laevis at the United States National Museum and University of Miami School of Marine and At- mospheric Sciences, and all of those from south of Cape Hatteras have proven to be i?. floridana. Also R. laevis does not occur in the species lists of Struhsaker (1969) or Staiger (1970). Thus it is likely that many or all of the records ofR. laevis from south of Cape Hatteras refer toR. floridana. Raja erinacea Raja erinacea was recorded from the Gulf of St. Lawrence; off Cape Breton, Nova Scotia; Western Bank; and two specimens were posi- tively identified from Sable Island Bank. It was the most abundant species captured on Georges Bank and in the northern section of the mid- Atlantic Bight. It was rarely taken in the western Gulf of Maine (Figure 7). Raja erinacea was most abundant in Chesapeake Bight during the winter and those that remained there during the summer moved into deeper water. Throughout its range, R. erinacea was gener- ally caught at depths less than 111 m, but was occasionally taken at depths greater than 183 m, especially in the northern section of the mid- Atlantic Bight and on Georges Bank where it occurred as deep as 329 m. Edwards et al. (1962) captured R. erinacea as deep as 384 m off New Jersey, thus the species is not as restricted to shallow water as stated by Bigelow and Schroeder ( 1954) who reported that 159 m was the maximum depth of the species. Temperatures at depth of capture ranged from 2° to 19°C with most captures occurring between 2° and 15°C. The recorded temperature range of the species is 1.2°C (Tyler 1971) to 21°C (Bigelow and Schroeder 1953). Raja erinacea is a Virginian to boreal species whose center of abundance is in the northern section of the mid- Atlantic Bight and on Georges Bank. Only in these areas was it found year- round over almost the entire range of tempera- tures recorded for the areas (Figures 8-10). In ^y 119 FISHERY BULLETIN: VOL. 73, NO. 1 CAPE HATTERAS / CAPE HATTERAS CAPE HATTERAS INDEX ABUNDANCE I[IIII1<0 24 □ O 25-0.99 CM) I 00-2 49 >2 50 Raja erinacea 1969 — WINTER 14° 36° Raja erinacea 1969 — SUMMER Raja erinacea 1969 — AUTUMN 38° 40° 42° Figure 7. — Index of abundance (geometric mean) of Raja erinacea captured by sampling strata during the winter, summer, and autumn 1969 cruise of the RW Albatross IV . 120 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES < Q 2 Z> CD < X UJ o SOUTHERN MID-ATLANTIC B I GHT 3 -I 2 - NORTHERN MID-ATLANTIC BIGHT 2 50 34° J^ Raja ocellata 1969 — WINTER Raja ocellata 1969 — SUMMER Raja ocellata 1969 — AUTUMN 38° 40° 42 o ^ ^ ^ Figure 11. — Index of abundance (geometric mean) of Raja ocellata captured by sampling strata during the winter, summer, and autumn 1969 cruise of the RV Albatross IV. 124 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES SOUTHERN MID -ATLANTIC BIGHT 3 —I UJ o z < Q Z OD < X UJ o z 2 — 0- 1^ NORTHERN MID-ATLANTIC BIGHT I 14 4 6 1 I \ 10 4_ 6 3 ■Si 15 C 0 6 7 4 GEORGES BANK 9 22 n3 0 2 I 3 2 10 15 C 0 r 10 15 C u. o o z < X Q UJ z Q 3 Z 03 - < 0 0 GULF OF MAINE I 9 II 3nii7n2 T" 10 15 20 C NOVA SCOTIAN SHELF 5 3 I 3 II 2 3 0 20 -C TEMPERATURE Figure 12. — Index of abundance (geometric mean) of Raja ocellata captured in each subarea during winter 1969 within temperature intervals of 1°C. See Figure 3 for explanation of fractions and whole numbers. Kaja radiata Raja radiata was the most abundant skate en- countered in the Gulf of St. Lawrence, off north- eastern and southeastern Nova Scotia, and in the Gulf of Maine. It was widespread along the eastern and northwestern slopes of Georges Bank (Figure 19). Raja radiata occurred between 27 and 439 m but was most abundant between 111 and 366 m. Bigelow and Schroeder (1953) listed a depth range of 18 to 896 m for this species in the western Atlantic. Temperatures at which it was captured ranged from -1.3° to 14°C. The previously recorded temperature range was -1.4°C (Backus, 1957) to 10°C (Bigelow and Schroeder 1953). Raja radiata is a boreal to arctic species whose center of abundance in the western Atlantic extends northward from the Gulf of Maine probably as far as the Gulf of St. Lawrence. It 125 2 -I o z < o z m < o X o z SOUTHERN MID-ATLANTIC BIGHT T 5 3543 7 6653 I NORTHERN MID-ATLANTIC BIGHT 10 15 20 C 0 FISHERY BULLETIN: VOL. 73, NO. 1 GEORGES BANK 2 9147I 6 5 2 0 0 1 in i^ 10 15 20 C 0 10 15 C I -1 oz < xQ ujZ q3 zm - < GULF OF MAINE 5 16 24 7 2 _, NOVA SCOT I AN SHELF 2 3 5 I 1 1 10 10 15 20 C 10 15 20 C TEMPER AT URE Figure 13.— Index of abundance (geometric mean) of Raja ocellata captured in each subarea during summer 1969 within temperature intervals of 1°C. See Figure 3 for explanation of fractions and whole numbers. UJ o z < Q z m < X UJ Q 2-1 I - SOUTHERN MID-ATLANTIC BIGHT 256 10 I 31 126532 NORTHERN MID-ATLANTIC BIGHT 15 20 C 0 I 3 I3["l0 6 1 0 I GEORGES BANK .5. h n 10 15 20 C 0 10 15 C I -, 00 z ii — CD < GULF OF MAINE 19 4 I0r-il26 NOVA SCOTI AN SHELF 4 3 15 sH I 5 10 15 20 C 0 5 10 15 20 C TEMPERATURE Figure 14. — Index of abundance (geometric mean) oi Raja ocellata captured in each subarea during autumn 1969 within tempera- ture intervals of TC. See Figure 3 for explanation effractions and whole numbers. 126 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES y • 78' CAPE HATTERAS .^^i CAPE HATTERAS /74'= CAPE HATTERAS INDEX ABUNDANCE l<0 24 r lO 25-0 99 11131.00-2 49 >2 50 ^ A_ 38° 40° Raja senta 1969 — WINTER Raja senta 1969 — SUMMER Raja senta 1969 — AUTUMN 42° Figure 15. — Index of abundance (geometric mean) of Raja senta captured by sampling strata during the winter, summer, and autumn 1969 cruise of the RV Albatross IV. 127 FISHERY BULLETIN: VOL. 73, NO. 1 2-1 I - UJ o z < Q Z 3 m < NORTHERN MID-ATLANTIC BIGHT 17 8 6t— lO I 2 4 3 6 7 4 GEORGES BANK * I 1 Hi 3 0 1 3 2 X UJ a 2 -\ GULF OF MAINE I - 5_ T NOVA SCOTIAN SHELF 5 1 I 3[— ' |2 3 10 15 20 C 10 I 15 20 c TEMPERATURE Figure 16. — Index of abundance (geometric mean) of Raja senta captured in each subarea during winter 1969 within tempera- ture intervals of 1°C. See Figure 3 for explanation of fractions and whole numbers. was found over almost the entire temperature range in the Gulf of Maine and off southeastern Nova Scotia. (Figures 20-22). INTERSPECIFIC RELATIONSHIPS Five of the species cooccurred significantly writh one or more of the other species (Table 3). Raja laeuis was associated with bothi?. erinacea and/?, ocellata for half or more oi the Albatross IV cruises. Raja erinacea andi?. ocellata cooccurred significantly during all of the survey cruises and were positively associated by abundance. The product moment coefficients for the Albatross IV winter, summer, and autumn cruises of 1969 were: r = 0.656, 0.471, and 0.640. Percent of the variation in j' associated with x was: 43%, 22%, and 41% respectively. The slopes of all three regressions were significant at the 1% probability level. No reason was apparent for the low corre- lation obtained during the summer cruise. Raja senta and R. radiata had the highest coefficient of association, and these two species were often negatively associated with R. erinacea and R. ocellata. Raja senta and R. radiata were not correlated by numbers; the coefficients for the Albatross IV winter, summer, and autumn cruises of 1969 were: 0.310, 0.081, and 0.283. Only a 128 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES 2-1 I - iij o z < o z m < NORTHERN MID-ATLANTIC BIGHT 9 14 7965200 I GEORGES BANK 3 2 96654 113 2 T X Ul a 2 -, GULF OF MAINE NOVA SCOTIAN SHELF ^ z 10 2 3 4, JiT DiJ- 20 C TEMPERATURE 10 15 20 C Figure 17. — Index of abundance (geometric mean) of Raja senta captured in each subarea during summer 1969 within temperature intervals of 1°C. See Figure 3 for explanation of fractions and whole numbers. small part of the variance could be assigned to the correlation, and the slopes were not significant at the 5% probability level. Raja erinacea and/?, ocellata are predominantly found at depths less than 111 m in areas w^hich, according to Uchupi (1963) are covered v^fith sand or gravel. They have similar responses to seasonal temperature changes. In the southern periphery of their ranges they move southward during the colder months of the year and off- shore and northward during the warmer months of the year. Within their centers of abundance, neither species undergoes a seasonal migration, each being able to tolerate the seasonal tempera- ture extreme. Raja ocellata appears to have a slightly lower temperature preference as sug- gested by the difference in latitudinal distribu- tion of the species. The apparent rareness of the species pair in the Gulf of Maine may be due to insufficient sampling. The shallowest depth zone (27-55 m) was not sampled during the Albatross IV cruises. Although the species have similar habitat requirements their positive corre- lation by numbers suggests that they are not competing for the same resources. Also a study of the food habits of the two species indicates that R. erinacea feeds largely on epifaunal organisms, and/?, ocellata predominately selects infaunal organisms (McEachran 1973). Raja laeuis is found in the same areas as the 129 FISHERY BULLETIN: VOL. 73, NO. 1 2 -I I - UJ o < a 3 ffl < NORTHERN B I MID- H T ATLANTIC 2 13 9 106 7 3 0 I GEORGES BANK 2 3 4 3 8 3 5 7 5 X o 2-T 1 - GULF OF MAINE z .4 3I9 1 n 3.55 NOVA SCOTIAN SHELF li 4 i Hi I m 15 20 C 15 20 C TEMPERATURE Figure 18. — Index of abundance (geometric mean) of Raja senta captured in each subarea during autumn 1969 within temperature intervals of 1°C. See Figure 3 for explanation of fractions and whole numbers. above species pair but has wider substratum and depth tolerance. Its low abundance may in part be explained by its considerably larger maximum size (Bigelow and Schroeder 1953) which makes it less available to the sampling gear. The distribution of the R. senta-R. radiata species pair complements that of the/?, erinacea- R. ocellata species pair. The former is found predominately in areas which, according to Uchupi (1963), were covered with sandy silt to silt and clay. They are taken over a narrower and lower temperature range than R. erinacea- R. ocellata and generally occur below 110 m. In the southern periphery of their ranges they are limited to a narrow band on the continental slope where the waters are thermally stable (Bigelow, 1933). Neither species appears to make seasonal movements. Raja radiata appears to have a wider temperature range and a lower temperature preference, and it is the more abun- dant of the two. The low abundance of/?, senta may explain the lack of a positive or negative correlation by numbers between the species. SUMMARY Below the geographical, temperature and depth distribution of each species, based on literature 130 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES CAPE HATTER AS / CAPE HATTERAS /^ CAPE HATTERAS INDEX OF ABUNDANCE l<0 24 □ O 25-099 EH 1.00-2 49 ^ >2 50 34° 36" Raja radiata 1969 — WINTER Raja radiata 1969 — SUMMER Raja radiata 1969 — AUTUMN 38° 40° 42° X. Figure 19. — Index of abundance (geometric mean) of Raja radiata captured by sampling strata during the winter, summer, and autumn 1969 cruise of the RV Albatross IV . 131 FISHERY BULLETIN: VOL. 73, NO. 1 2-1 UJ o z < o < u. o X UJ a NORTHERN B MID-ATLANT G HT IC GEORGES BANK 3. 17 8 6 0 12 4 3 6 7 4 2 -I I - 0- GULF OF MAINE 0 "T 5 I 10 15 20 C 0 TEMPERATURE Figure 20. — Index of abundance (geometric mean) of Raja radiata captured in each subarea during winter 1969 within temperature intervals of 1°C. See Figure 3 for explanation of fractions and whole numbers. reports (Bigelow and Schroeder 1953; Leim and Scott 1966; and McEachran 1973) and findings in the present study, are summarized. Raja eglanteria is found from Long Island to northern Mexico but is rare off southern Florida. It occurs from the shore zone to 329 m at 5° to 27°C, but is most abundant between the shore zone and 111 m at 9° to 20°C. Raja garmani occurs from the offing of Nan- tucket Shoals to the Dry Tortugas, Fla. North of Cape Hatteras, N.C., it is found in 37 to 366 m at 6° to 17°C, and south of there it occurs from 66 to 366 m at 11° to 19°C. Raja laevis extends from the southern New- foundland banks and the Gulf of St. Lawrence south to North Carolina. It is found from shore to 750 m at 1.2° to 20°C. Raja erinacea regularly occurs from southern Nova Scotia to Cape Hatteras. It is found between shore and 384 m at 2° to 21°C but is most abun- dant in water shallower than 111 m at 2° to 15°C. Raja ocellata is found from the Newfound- land banks and southern Gulf of St. Lawrence to Cape Hatteras. It occurs from shore to 371 m at -1.2° to 19°C but is most abundant in water shallower than 111 m at 2° to 15°C. Raja senta occurs from the southern Newfound- land banks and the Gulf of St. Lawrence to South Carolina. It occurs from 31 to 974 m at -1.3° to 14°C but is most abundant below 110 m at 2° to 10°C. Raja radiata extends from Labrador, west Greenland, Hudson Bay, Grand Banks, and Gulf of St. Lawrence to South Carolina. It occurs from 132 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES 2-r I - UJ u z < o z GO < NORTHERN MID-ATLANTIC 1 5 9i4 7|— leRz 0 0 I GEORGES BANK 2 3 4 5 2 6 2 9 2_ II 6 5 4n32 X o 2 -f GULF OF MAINE I I T NOVA SCOTIAN SHELF 2? IzT 10 15 20 C 6 16 T6 2 2 3 5 I 5 10 "T 20 C TEMPERATURE Figure 21. — Index of abundance (geometric mean) of Raja radiata captured in each subarea during summer 1969 within temperature intervals of 1°C. See Figure 3 for explanation of fractions and whole numbers. 18 to 996 m at - 1.4° to 14°C but is most abundant below 110 m at 2° to 10°C. Raja erinacea and R. ocellata are sympatric species with very similar habitat requirements. Raja ocellata has slightly lower temperature preferences than R. erinacea and occurs farther to the north than the latter. Raja senta and R. radiata are sympatric species; R. radiata has wider temperature range and is more widespread than R. senta. ACKNOWLEDGMENTS We are very grateful to the Northeast Fish- eries Center Woods Hole Laboratory, NMFS, NOAA; the Fisheries Research Board of Canada Biological Station at St. Andrews, New Bruns- wick; and the Southeast Fisheries Center Pas- cagoula Laboratory, NMFS, NOAA for furnishing data for this study. The two former institutions also permitted the senior author to take part in their groundfish surveys. The Northeast Fisheries Center Woods Hole Laboratory, NMFS, NOAA gave access to their computer programs and computer facilities to summarize data. The VIMS survey of Chesapeake Bight was supported in part by the NMFS under P.L. 88-309, Project 3-5-D, Jackson Davis, Principle Investigator. Special thanks is given to Marvin Grosslein of the Northeast Fisheries Center Woods Hole 133 FISHEKY BULLETIN: VOL. 73, NO. 1 2n NORTHERN MID -ATL ANTIC I - ui o < o z CD < X UJ a 2 "—19 10 6 7 3 0 I GEORGES 5 3 BANK 7 GULF OF MAINE 2 -I I - U 5 2.19 I NOVA SCOTIAN SHELF fl4 10 15 20 C 10 15 20 C TEMPERATURE Figure 22. — Index of abundance (geometric mean) of Raja radiata captured in each subarea during autumn 1969 within temperature intervals of TC. See Figure 3 for explanation of fractions and whole numbers. Laboratory, NMFS, NO A A for his cooperation during all phases of this study. John B. Colton, Jr., also of the Northeast Fisheries Center Woods Hole Laboratory, NMFS, NOAA supervised the construction of the isotherm charts. The following VIMS staff members and students contributed greatly to this study: Mark E. Chit- tenden and George C. Grant reviewed the manu- script; Russel L. Bradley and Kay Stubblefield did the drafting; Ken Thornberry did the photo- graphic work; and Charles Wenner, Linda Mercer, Ken Able, Doug Markle, and Jim Weaver assisted with data collection. LITERATURE CITED Backus, R. H. 1957. The fishes of Labrador. Bull. Am. Mus. Nat. Hist. 113(4):279-337. BiGELOW, H. B. 1927. Physical oceanography of the Gulf of Maine. U.S. Bur. Fish., Bull. 40:511-1027. 1933. Studies of the waters on the continental shelf, Cape Cod to Chesapeake Bay. I. The cycle of temperature. Pap. Phys. Oceanogr. Meteorol., Mass. Inst. Technol. and Woods Hole Oceanogr. Inst. 2(4), 135 p. BlGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the western North Atlantic. Part 2. Sawfishes, guitarfishes, skates and rays [and] chimae- roids. Mem. Sears Found. Mar. Res., Yale Univ. 1, 588 p. 134 McEACHRAN and MUSICK: DISTRIBUTION AND RELATIVE ABUNDANCE OF SKATES Table 3. — Coefficients of interspecific association for Raja ocellata, R. erinacea, R. senta, R. radiata, and R. laevis. Cruise and species R. ocellata R. erinacea R. senta R. radiata Cruise 67-21: R. erinacea 0.61" — — — R. senta -0.02 -0.71 — — R. radiata -0.28 -0.53" 0.60" — R. laevis -0.02 0.00 0.00 0.00 Cruise 68-03: R. erinacea 0.67" — — — R. senta 0.00 0.00 — — R. radiata -0.04 -0.32 0.84" — R. laevis 0.25- 0.53" 0.00 0.27 Cruise 68-17: R. erinacea 0.52" — — R. senta 0.00 0.00 — R. radiata -0.85" -0.54" 0.78" R. laevis 0.14 0.45 0,00 0.00 Cruise 69-02: R. erinacea 0.63" — — R. senta -0.35 -0.34 — R. radiata -0.31 -0.23 0.95" R. laevis 0.54" 0.36 -0.01 -0.03 Cruise 69-08: R. erinacea 0.71 — — — R. senta 0.00 -0.62" — — R. radiata -0.56- . -0.42- 0.75" — R. laevis 0.72" 0.84" -0.09 -0.02 Cruise 69-11: R. erinacea 0.57" — — — R. senta -0.70 -0.85- — — R. radiata -0.21 -0.54" 1.00" — R. laevis 0.48 0.79" 0.00 0.34 Cruise 70-03: R. erinacea 0.53 — R. senta -0.01 -0.42 — — R. radiata -0.12 -0.38 1.00" — R. laevis 0,13 0.47- -0.09 0.01 Cruise 70-06: R. erinacea 0.41" — — — R. senta -0.82" -0.84" — — R. radiata -0.61" -0.44" 0.80" — R. laevis 0.01 0.01 0.00 0.51 ■Significant at the 0.05 probability level. "Significant at the 0.01 probability level. 1954. Deep water elasmobranchs and chimaeroids from the northwestern Atlantic slope. Bull. Mus. Comp. Zool. Harvard Coll. 112:38-87. 1962. New and little known batoid fishes from the western Atlantic. Bull. Mus. Comp. Zool. Harvard Coll. 128:159-244. 1968. Additional notes on batoid fishes from the western Atlantic. Breviora 281, 23 p. BuLLis, H. R., Jr., and J. R. Thompson. 1965. Collections by the exploratory fishing vessels Oregon, Silver Bay, Combat, and Pelican made during 1956 to 1960 in the southwestern North Atlantic. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 510, 130 p. Dahlberg, M. D., and E. P. Odum. 1970. Annual cycles of species occurrence, abundance, and diversity in Georgia estuarine fish populations. Am. Midi. Nat. 83:382-392. Edwards, R. L., R. Livingstone, Jr., and P. E. Hamer. 1962. Winter water temperatures and an annotated list of fishes — Nantucket Shoals to Cape Hatteras, Albatross III Cruise no. 126. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 397, 31 p. FiTz, E. S., Jr., and F. C. Daiber. 1963. An introduction to the biology of Raja eglanteria Bosc 1802 and Raja erinacea Mitchill 1825 as they occur in Delaware Bay. Bull. Bingham Oceanogr. Collect., Yale Univ. 18(3):69-97. Grosslein, M. D. 1969. Groundfish survey program of BCF Woods Hole. Commer. Fish. Rev. 31(8-9):22-30. Hedgpeth, J. W. 1957. Marine biogeography. In J. W. Hedgpeth (editor). Treatise on marine ecology and paleoecology. Vol. I. Ecology. Geol. Soc. Am., Mem. 67:359-382. Hurlbert, S. H. 1969. A coefficient of interspecific association. Ecology 50:1-9. Johnson, C. W. 1934. List of marine mollusca of the Atlantic coast from Labrador to Texas. Proc. Boston Soc. Nat. Hist. 40:1-204. Leim, a. H., and W. B. Scott. 1966. Fishes of the Atlantic Coast of Canada. Fish. Res. Board Can., Bull. 155, 485 p. 135 FISHERY BULLETIN: VOL. 73, NO. 1 Massman, W. H. 1962. Water temperatures, salinities, and fishes col- lected during trawl surveys of Chesapeake Bay and York and Pamunkey Rivers. 1956-1959. Va. Inst. Mar. Sci., Spec. Sci. Rep. 27, 51 p. McEachran, J. D. 1970. Egg capsules and reproductive biology of the skate Raja garmani (Pisces: Rajidae). Copeia 1970:197-199. 1973. Biology of seven species of skates (Pisces: Rajidae). Ph.D. Thesis, Coll. William and Mary, Williamsburg, Va. McEachran, J. D., and J. A. Musick. 1973. Characters for distinguishing between immature specimens of the sibling species. Raja erinacea and Raja ocellata (Pisces: Rajidae). Copeia 1973:238-250. Merriman, D., Y. H. Olsen, S. B. Wheatland, and L. H. Calhoun. 1953. Addendum to Raja erinacea. In Fishes of the western North Atlantic. Part 2. Sawfishes, guitarfishes, skates and rays [and] chimaeroids, p. 187-194. Mem. Sears Found. Mar. Res., Yale Univ. 1. Musick, J. A., and J. D. McEachran. 1972. Autumn and winter occurrence of decapod crusta- ceans in Chesapeake Bight, U.S.A. Crustaceana 22: 190- 200. Pereyra, W. T,, H. Heyamoto, and R. R. Simpson. 1967. Relative catching efficiency of a 70-foot semiballoon shrimp trawl and a 94-foot eastern fish trawl. U.S. Fish Wildl. Serv., Fish. Ind. Res. 4:49-71. Richards, S. W. 1963. The demersal fish population of Long Island Sound. I. Species composition and relative abundance in two localities, 1956-57. Bull. Bingham Oceanogr. Collect., Yale Univ. 18(2):5-31. Richards, S. W., D. Merriman, and L. H. Calhoun. 1963. Studies on the marine resources of southern New England. IX. The biology of the little skate. Raja erinacea Mitchill. Bull. Bingham Oceanogr. Collect., Yale Univ. 18(3):5-67. Roessler, M. 1965. An analysis of the variability of fish populations taken by otter trawl in Biscayne Bay, Florida. Trans. Am. Fish. Soc. 94:311-318. SCHAEFER, R. H. 1967. Species composition, size and seasonal abundance of fish in the surf waters of Long Island. N.Y. Fish Game J. 14:1-46. ScHOPF, T. J. M., AND J. B. Colton, Jr. 1966. Bottom temperature and faunal provinces: Conti- nental shelf from Hudson Canyon to Nova Scotia. [Abstr.] Biol. Bull. (Woods Hole) 131:406. SCHROEDER, W. C. 1955. Report on the results of exploratory otter-trawling along the continental shelf and slope between Nova Scotia and Virginia during the summers of 1952 and 1953. Deep-Sea Res., Suppl. Vol. 3:358-372. Schwartz, F. J. 1961. Fishes of Chincoteague and Sinepuxent Bays. Am. Midi. Nat. 65:384-408. Staiger, J. C. 1970. The distribution of the benthic fishes found below two hundred meters in the Straits of Florida. Ph.D. Thesis, Univ. Miami, 245 p. Struhsaker, p. 1969. Demersal fish resources: Composition, distribution, and commercial potential of the Continental Shelf stocks off Southeastern United States. U.S. Fish Wildl. Serv., Fish. Ind. Res. 4:261-300. Taylor, C. C. 1953. Nature of variability in trawl catches. U.S. Fish Wildl. Serv., Fish. Bull. 54:145-166. Tyler, A. V. 1971. Periodic and resident components in communities of Atlantic fishes. J. Fish. Res. Board Can. 28:935-946. 1972. Surges of winter flounder, Pseudopleuronectes americanus, into the intertidal zone. J. Fish. Res. Board Can. 28:1727-1732. UcHUPi, E. 1963. Sediments on the continental margin off eastern United States. U.S. Geol. Surv. Prof. Pap. 475-C:C132- C137. 136 THE GENERAL FEEDING ECOLOGY OF POSTLARVAL FISHES IN THE NEWPORT RIVER ESTUARY^ Martin A. Kjelson, David S. Peters, Gordon W. Thayer, and George N. Johnson^ ABSTRACT Food preferences, feeding intensity and chronology, evacuation rates, and daily rations were determined for postlarval stages of Atlantic menhaden, Brevoortia tyrannus (25-32 mm); pinfish, Lagodon rhomboides (16-20 mm); and spot, Leiostomus xanthurus (17-24 mm). Four copepod taxa, Centropages. Temora, Acartia, and Harpacticoida, made up 76-99%' of the total gut contents. Postlarval feeding intensity was greatest during early daylight hours. Postlarval menhaden lost an estimated 60% of their orginal gut contents due to the stress of handling and capture. Similar stress caused no food loss in either postlarval pinfish or spot. Gastrointestinal evacuation of copepods and Artemia nauplii were described by linear regression. Evacuation rates varied directly with the amount of food in the gut. Rate constants were used in conjunction with infor- mation on the chronology of gut contents to determine daily rations. Daily ration estimates as a percent of the fish's wet body weight were: menhaden, 4.9%; pinfish, 3.5%; spot, 4.3% and 9.0%. The ration estimates for spot in terms of calories per fish per day were similar to the metabolic needs estimated from oxygen consumption measurements but were lower than the estimates from oxygen consumption for menhaden and pinfish. Larval and postlarval fish are significant con- sumers in aquatic ecosystems, yet our knowledge of their feeding habits and daily food consump- tion is incomplete. This paper deals with the general feeding ecology of the postlarval stages of three common estuarine fishes. Four major aspects are discussed. These include 1) food preferences, 2) feeding intensity and chronology, 3) evacuation rate, and 4) daily ration. Postlarval Atlantic menhaden, Brevoortia tyrannus; pinfish, Lagodon rhomboides; and spot, Leiostomus xanthurus, were collected during March of 1972 and 1973 from the Newport River estuary, Carteret County, N.C. The fish (hereafter referred to as larvae) were taken near Pivers Island, approximately 2.5 km inside the Beaufort Inlet. Pinfish and spot were collected using a seine and dip nets, while menhaden were captured in a channel net (Lewis et al. 1970) and with dip nets. One additional group of samples was collected in bongo nets. Most fish were frozen immediately following capture, thus stopping their digestive processes. The only exceptions to preservations by freezing were the bongo net 'This research was supported through a cooperative agree- ment between the National Marine Fisheries Service and the U.S. Atomic Energy Commission. ''Atlantic Estuarine Fisheries Center, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. samples which were placed in 5% Formalin.^ Food preferences were determined by examin- ing the contents of entire digestive tracts. The gut contents from 120 fish of each species col- lected throughout the day were combined and individual food items identified, counted, and measured. Copepodite and adult copepods com- posed 99-100% (by both number and volume) of the identifiable food items in the digestive tracts. The average-sized copepod fed upon by each larval species was determined by measuring 100 cope- pods chosen from the combined digestive tract contents of all larvae collected in a 24-h period. Diel periodicity of digestive tract contents indi- cated the intensity and chronology of feeding by the larvae. Twenty fish of each species were collected at 4-h intervals for 24 consecutive hours. Larval evacuation rates for copepods and for Artemia salina were determined from laboratory experiments performed at 15°-17°C and 25-30%o; conditions which typify larval collection sites during March. Copepod evacuation was deter- mined by collecting larvae from the estuary, placing them in food-free seawater tanks, and observing the decrease in their gut contents through time. At the time of initial capture Manuscript accepted March 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 137 FISHERY BULLETIN: VOL. 73, NO. 1 and every 3 or 4 h thereafter, at least 10 fish were killed and the copepods present were counted. Evacuation of newly hatched Artemia nauplii was measured by allowing unfed larvae to feed to satiation on high prey densities (0.3 to 3.0 nauplii/cm^), placing them in a food-free environment, and then periodically removing larvae for determination of remaining gut con- tents. Sampling of both copepod- and Artemia -fed fish continued periodically from the time of feeding until more than one-half of the fish had empty tracts. Linear regression equations of log-transformed data were used to describe the evacuation process (Peters et al. 1974). The equations were of the form: logioC = A + Bt where C = 1 + the mean number of copepods or Artemia present in the gut t = time A +B = regression terms Instantaneous evacuation rates were calculated from the equation — = 2.303 EC (Peters and Kjelson in press). Daily rations were calculated using information on diel periodicity of gut content and instantan- eous evacuation rates. Previous calculations of fish rations (Bajkov 1935; Seaburg and Moyle 1964) have assumed a constant evacuation rate, but more recent data (Tyler 1970; Elliot 1972) indicate that digestive rate changes with the quantity of food in the digestive tract. Our method of calculating daily ration (Peters and Kjelson in press) accounts for changes in evacuation rate which accompany diel changes of feeding intensity. To calculate the rations, we first determined an average instantaneous evac- uation rate (in copepods per hour) for each of the 4-h sampling periods in the diel cycle. This average rate was the geometric mean of the instantaneous evacuation rates at the beginning and end of the period. Since each period lasted 4 h, the estimate of food evacuated during the period was four times the average instantaneous hourly evacuation rate. The total food evacuated per day was achieved by summing the six 4-h evacuation estimates, and is an estimate of the daily ration, because average ingestion rate must equal the rate at which material leaves the gut whether by assimilation or expulsion. Daily rations were calculated initially as copepods per fish per day and then transformed to percent of the larval body weight and calories per fish per day. Dry weights of ingested copepods were estimated from the length-weight rela- tionship: W = 6.274L - 0.153 where W = dry weight in micrograms L = copepod length, based on Heinle's (1966) data for all stages of Acartia tonsa Copepod dry weights were converted to wet weights using a factor of 9.1 based upon our measurements of the wet/dry ratio for zooplank- ton, and were compared to wet weights of the fish to compute the daily ration as a percent of live body weight. Daily caloric intake was computed using our estimation of 0.555 cal/mg wet weight of an average size copepod during March, based on microbomb calorimeter mea- surements of mixed estuarine zooplankton (Thayer et al. 1974). FOOD PREFERENCES The larvae we collected were feeding primarily upon copepods, a common food source for both freshwater and marine fish larvae (Werner 1969; May 1970). Copepods composed 99% (by volume and number) of the gut contents of larval spot, pinfish, and menhaden (Table 1). Four copepod taxa (Centropages, Temora, Acartia, and Harpac- ticoida) were dominant. Diatoms, amphipods, barnacle larvae, crab zoea, and ostracods, al- though present in some larvae, were rare. Table 1. — Relative (percent) composition by number of the major taxa in the total gut contents of three species of larval fish. Lar val species Taxa Pinfish Spot Menhaden Harpactlcoida 32 32 22 Centropages 28 28 40 Temora 3 21 6 Acartia 13 8 30 Other copepods 23 10 1 Other organisms 1 1 1 Total 100 100 100 138 KJELSON ET AL.: FEEDING ECOLOGY OF POSTLARVAL FISHES Prey size is an important factor in determining the individuals selected by planktivorous fish (Ivlev 1961; Brooks and Dodson 1965; Kjelson 1971). The larval fish we studied appeared to restrict the majority of their feeding to items of a size ranging between 300 and 1,200 jum. Our observations of the mean length of ingested copepods showed that the larger menhaden larvae (26-31 mm, x = 29 mm TL) ingested 750-fj.m copepods with an estimated copepod wet weight of 0.04 mg, while the smaller spot (17- 22 mm, x = 19 mm TL) and pinfish larvae (16- 20 mm, X = 18 mm TL) fed upon 600-/jm copepods with an estimated wet weight of 0.03 mg. Small zooplankters such as copepod nauplii, barnacle larvae, or small adult copepods such as Oithona (all present in the plankton tows) were rarely found in gut contents. Copepods larger than 1.2 mm were in the plankton, but were rarely consumed. Perhaps copepods were the only food items of the appropriate size present in sufficient abundance. Had we collected smaller larvae, it is possible food preferences may have been for smaller food items such as copepod nauplii and copepodites and adults of small-sized species as well as phytoplankton. May (1970) stressed the fact that larval fish require progressively larger prey as they grow. However, since larvae smaller than the size we collected are rarely found in the Newport River estuary, we feel our data indicates that smaller planktonic forms are rela- tively unimportant to the larval fish studied in this estuary. Thayer et al. (1974) found that as a yearly average, copepods represented 81% of the zoo- plankton numbers and 85% of the zooplankton biomass retained by a No. 10 mesh plankton net. Since larval fishes enter the Newport River estuary during winter and spring, the con- sumption of copepods by these three larval species may, in part, explain the decrease in copepod abundance observed by Thayer et al. (1974) dur- ing this period. They noted that the four copepod taxa utilized by these larvae decreased from a mean of 81% of the copepod biomass during March 1970 and 1971 to a mean of 48% of the biomass during the summer. FEEDING CHRONOLOGY AND INTENSITY All three larval fishes had the highest food content in their digestive tracts during daylight hours (Figures 1-3). Periodicity of gastrointestinal contents indicates that each population begins feeding near dawn and reaches a maximum gut fullness near midday. The rapid single increase in the gut content of the three species indicates they have one major burst of feeding activity per day (Figures 1-3). Other studies (Blaxter 1965; Schumann 1965; Braum 1967; June and Carlson 1971) have shown that larval fish generally do not have food within their digestive tracts when captured at night, suggesting that larval fish do not feed at low light intensities. Considerable variation was observed in the amounts of food present in larval guts (Figures 1-3). The variation is probably due to differences in prey abundance or capture and handling techniques, although other factors such as fish size and copepod size may also be important. During our 24 h sampling the variation in numbers of copepods in individual fish was high at some times and low at others. The ratio of the standard error of the estimate to the mean varied from 4 to 48% for spot with a mean of 21%; for menhaden the ratio varied from 0 to 100% with a mean of 40%' ; and for pinfish it varied from 0 to 100% with a mean of 43%. Spot larvae •- 40 o t 35 2 30 a. s \0 1973 BONGO NET COLLECTION 1972 HAUL SEINE DIP NET COLLECTION 1973 HAUL SEINE DIP NET ., COLLECTION f \ 1 \ 1 \ 1 \ J \ 1 \ / \ / \ / \ / \ / V / \ / ^ / » / \ / \ / \ / » ^ ; A \ A \ / / \ ^ ' / \ ^ ^-4 / \ \ --"""'''X / / ^— — V^^ \ ' / ^ \ ^ / \ ^ \ / ^ \ ^^^t^^"^""^ —■"*"""*"•- \ • "" ' "*^- — ._ , . __ . ^ 0400 0800 1200 1600 2000 2400 0400 TrME OF DAY Figure 1. — Variation in diel cycle of gastrointestinal contents in postlarval spot. 139 FISHERY BULLETIN: VOL. 73, NO. 1 Z 4 UJ O < X z < > ; 3 o o s Z 1973 BONGO NET COLLECTION 1972 CHANNEL NET DIP NET COLLECTION 1973 CHANNEL NET DIP NET COLLECTION 0400 0800 1200 1600 2000 2400 0400 TIME OF DAY Figure 2. — Variation in diel cycle of gastrointestinal contents in postlarval Atlantic menhaden. 40 > 30 a 2 25 O u i 3 Z Z < 20 1$ 10 1973 BONGO NET COLLECTION 1972 HAUL SEINE DIP NET COLLECTION 1973 HAUL SEINE DIP NET COLLECTION f^ I I ( 1 I I I 1 I I 0400 OeOO 1200 1600 2000 2400 0400 TIMt OF DAY Figure 3. — Variation in diel cycle of gastrointestinal contents in postlarval pinfish. (18-24 mm, x = 21.5 mm) collected by dip net at one location and between 0830 and 1030 h over a 4-day period had little variation in their mean gut contents. Spot larvae collected 2 April averaged 25.3 copepods/fish (SE = 2.3), on 3 April, 21.3 (SE = 2.0), and on 5 April, 26.3 (SE = 3.7). The similarity of food quantity in the larval digestive tracts suggests that prey abundance may have remained relatively constant over the 4-day period thus allowing the fish to consume similar amounts of food. ESTIMATES OF LARVAL GUT CAPACITIES Laboratory feeding experiments were con- ducted at 15°-16°C to estimate the maximum gut capacity of the larvae. The fish were fed high densities of Arte mia nauplii until their digestive tracts were completely packed from esophagus to anus. Menhaden (28-32 mm, x = 30 mm TL) fed for 20 min on a concentration ofArtemia nauplii, 3 nauplii/cm^, had an average of 145 nauplii/fish in their digestive tracts (SE = 9.6). Spot larvae (19-23 mm, x = 21 mm TL) fed for 15 min on 0.3 nauplius/cm^ had an average of 89 (SE = 7.0) nauplii/fish; and pinfish ( 16-20 mm,x = 18 mm TL) fed for 1 h on 0.3 nauplius/cm^, had an average of 75 (SE = 15.1) nauplii/fish. By comparing individual Artemia and copepods of the four major taxa side by side under a microscope, we estimated that the volume of two 450-/um Artemia nauplii were equivalent to that of one 650- jum copepod. Using 0.5 as a conversion factor, we calculated maximum gut capacities in terms of copepods. Menhaden larvae of 30 mm have a gut capacity of 72 copepods, 21-mm spot a gut capacity of 44 copepods, and 18-mm pinfish a gut capacity of 37 copepods. These estimates of gut capacity were comparable to the maximum numbers of copepods observed in the digestive tracts of larval fish collected in the estuary for spot (36.5 copepods/fish) (Figure 1) and pinfish (35.3 copepods/fish) (Figure 3), but not for men- haden (5.2 copepods/fish) (Figure 2). This large difference between gut capacity and observed gut contents suggests that menhaden larvae either feed very little under natural conditions, and never approach the estimated maximum gut capacity or capture and/or handling causes them to regurgitate or defecate causing inaccuracy in our estimate of natural gut content. To test 140 KJELSON ET AL.: FEEDING ECOLOGY OF POSTLARVAL FISHES the latter possibility, we performed a variety of experiments to determine if handling and capture technique influences the quantities of food observed in the larval gut. EFFECTS OF SAMPLING TECHNIQUE ON GUT CONTENT Larvae of all three species were first collected by a 3-m channel net with an attached live box. Captured fish were counted, identified, divided into two groups, and transferred (underwater) into separate containers. One group of fish was anesthetized with 0.12 g/liter MS-222 (tricaine methanesulfonate) and then dissected, while the other group was transferred carefully into the posterior end of a 20-cm bongo net (keeping them underwater throughout transfer). The net was towed for 5 min, and after retrieval the larvae were removed, identified, and counted to assure that none were lost and that no new larvae were captured. The fish were then dissected to determine the number of copepods present in their guts. Menhaden lost 68% of their gut contents when exposed to the stress of bongo tows, whereas gut contents of larval spot and pinfish before and after the bongo tow did not differ statistically (Table 2). The amounts of food present in all three species of larvae collected at the Beaufort Inlet in the 24-h bongo samples were lower than the food quantities observed in larvae collected by the other techniques inside the estuary (Figures 1-3). Thus, factors other than the stress of capture may be responsible for the low gut contents in larvae collected in the bongo nets: 1) copepod abundance may have been lower at the Beaufort Inlet sampling site than further inside the estuary, 2) the use of Formalin (restricted to bongo samples) to kill and preserve the larvae may have caused defecation of the copepods prior to analysis (June and Carlson ( 1971) showed that larval menhaden when placed in Formalin had violent spasms accompanied by Table 2. — The effect of bongo net tow stress upon the amount of food observed in larval menhaden, pinfish, and spot. Capture technique Menhaden' Pinfish^ Spot^ Channel net Channel net + bongo net tow Mean number copepods/fish ± one SE 7.4 ±2 1.3 ±0.6 6.4 ±4.4 2.4 ±0.7 0,8 ±0.3 7.0 ±2 defecation), and 3) larvae collected in midchannel by bongo nets may not be feeding as actively since they are exposed to a greater tidal current (perhaps the protected inshore waters of the estuary may allow the larvae to feed more effi- ciently and result in fish with greater numbers of prey in their digestive tracts). No significant differences were observed in the food contents of spot larvae collected by routine seining and those collected by seining with a more gentle sampling technique (Table 3). In routine seining, the larvae were picked out of the seine as it lay on the shore and placed in a bucket of ice water. The gentle sampling tech- nique consisted of surrounding a body of water with the seine and then concentrating the larvae, taking care that fish were not forced against the net. Once concentrated, the larvae were dipped out of the water in a bucket and anesthe- tized with MS-222. The results of our sampling experiments indicate that routine field sampling techniques used to collect spot and pinfish larvae probably caused little loss of food from the digestive tracts. EFFECTS OF HANDLING TECHNIQUE ON GUT CONTENT In the laboratory, handling stress did not reduce the food quantities present in larval spot and pinfish, but did reduce the amount of food remain- ing in larval menhaden (Table 4). Two groups Table 3. — Comparison of food quantities, mean number of copepods per fish ± one SE, present in larval spot (22-33 mm, X = 27 mm) collected by haul seine using rough and gentle handling techniques. Ten fish were collected per sample. Date Gentle Rough April 2 April 3 84.5 ± 7.4 69.7 ± 5.9 78.6: 66.9: 5.3 5.5 Table 4. — The effects of handling on the retention of Artemia nauplii in digestive tracts of larval Atlantic menhaden, pin- fish, and spot. Rough handling is approximately equivalent to field capture by dip net and haul seine. 22 larvae. 18 larvae. 5 larvae. Species (Range mm) Experiment 1 Experi ment 2 Gentle Rough Gentle Rough — Mean number ± one SE — Menhaden' 71 ± 15 29 ± 10 1 45 ± 10 76± 11 (28-32) Pinfish^ 37 ± 4 34 ± 5 35 ± 9 43 ± 6 (16-20) Spot2 51 ± 5 47 ± 5 89 ± 7 92± 10 (19-23) 'n = 18 larvae per sample. 2n = 36 larvae per sample. 141 FISHERY BULLETIN: VOL. 73, NO. 1 of unfed larvae of each species were offered identical concentrations of Artemia nauplii. One group was handled roughly to represent the physical stress associated with field capture, while the other group was handled gently. The roughly handled fish were chased around the tank with a dip net for 10 to 30 s, captured with the net, allowed to suffocate in air, and then dissected. After feeding, the other fish were anesthetized by carefully adding an aqueous solution of MS-222 to the tank and then were dissected immediately to determine the numbers of nauplii in their digestive tracts. The roughly handled menhaden had only 40 to 52% of the Artemia numbers present in the guts of the gently handled menhaden (Table 4). The loss of food in menhaden larvae probably was due to the stress-related defecation or regurgitation and thus, may explain the consistently low quantities of food observed in larval menhaden captured in the estuary. Roughly handled pinfish and spot larvae showed no significant decrease in gut contents (Table 4). The curved digestive tract of larval spot and pinfish may prevent rapid passage of food, while the straight tubelike gut of menhaden may permit easy loss of food. This gut shape difference may account for the dif- ferences we observed. A separate experiment was conducted to deter- mine if the technique used to kill menhaden larvae in the handling experiments (exposure to air and suffocation versus anesthesia with MS-222) influenced the amount of food remaining in the gut. No difference was found. Fish killed by suffocation had a mean of 19 Artemia nauplii/ fish (SE = 4.7), while fish anesthetized with MS-222 had a mean of 20 Artemia nauplii/fish (SE = 4.2). EVACUATION RATES Estimated regression coefficients for the equa- tions describing the evacuation of copepods and Artemia nauplii are provided in Tables 5 and 6. Certain factors may alter the reliability of our estimates of evacuation rate under natural estuarine conditions. Bias may result from the temperature difference between estuarine waters from which fish were captured (14°-15°C) and the aquaria temperature during evacuation experiments (16°-17°C). The effect of a 2° tem- perature change on evacuation rate of larvae Table 5. — Linear regressions describing evacuation of copepods in Atlantic menhaden, pinfish, and spot larvae. Y = A + Bt where Y = logio (1 -i- mean number of copepods per larva) and t = hours since feeding, n = the number of data points. Species Mean TL (Range mm) A B n r2 Tempera- ture CO Menhaden 29 1.14 -0.17 3 0,98 16 Pinfish (27-31) 17 0.94 -0,10 3 0.86 16 Pinfish (15-20) 16 0.68 -0.08 4 098 17 Spot (13-19) 20 (17-23) 0,91 -0.10 5 0 98 17 is unknown, although a similar change signi- ficantly increases the evacuation rates in some juvenile marine fish (Peters and Kjelson in press). Although our regression model could probably be improved, the r^ values (Tables 5, 6) indi- cate the model is reasonable. Initial analysis included data collected until all the fish were empty. This resulted in nonlinearity near the end of evacuation due to bias near the end of evacuation period where more weight was given to the slower evacuating fish. Thus, by including in the regression analysis data from only those samples in which at least half of the larvae contained some food, this bias was decreased and the linear regression model appeared to represent larvae evacuation adequately. INFLUENCE OF HANDLING AND CAPTURE ON EVACUATION Evacuation experiments using Artemia nauplii were performed to determine if handling and capture influenced the rate of evacuation. Each Table 6. — Linear regressions describing evacuation of Artemia nauplii in Atlantic menhaden, pinfish, and spot larvae under varied handling conditions. Y = A + Bt where Y = log jq (1 + mean number of Artemia per larva) and t = hours since feeding, n = the number of data points. Species MeanTL (Range mm) A e n r2 Handling condition Tempera- ture CO Menhaden 29 2.36 -0.28 5 0.96 Gentle 15 Menhaden (27-32) 29 2.04 -0.34 3 0,86 Rough 15 Pinfish (27-32) 16 1.64 -0.26 4 0.97 Gentle 16 Pinfish (14-18) 16 1.73 -0.28 3 0.92 Rough 16 Spot (14-17) 20 2,12 -0.19 5 0.94 Gentle 16 Spot (18-23) 20 (18-23) 2.11 -0.18 5 0.95 Rough 16 142 KJELSON ET AL.: FEEDING ECOLOGY OF POSTLARVAL FISHES of the three larval fish species were fed concen- trated amounts oi Artemia (> 0.3 nauplius/cm^) and allowed to feed until their digestive tracts were full. Each species then was transferred to food-free containers and separated into two groups, one handled roughly and another handled gently. Fish were sampled immediately and every 2 h thereafter. The rough treatment was similar to that used to study the influence of handling on gut content. The gently handled fish were sampled by dipping them carefully out of the tank with a beaker and anesthetizing them prior to dissection. The similarities of the regression coefficients (Table 6) for fish of the same species under the two treatments indicate that evacuation rates were not affected by rough treatment. The higher B value for roughly handled menhaden was not significantly dif- ferent. Thus, our use of laboratory evacuation data to represent the normal evacuation in nature appears reasonable. The regression coefficients ior Artemia nauplii evacuation were larger (B values ranging from -0.18 to -0.34) than those for copepod evacua- tion (5 values ranging from -0.08 to -0.17) for all three species (Tables 5, 6). This was expected since the Artemia nauplii were estimated to be only one-half the volume of copepods ingested by the larvae. Food quality may also affect evacuation rate. Rosenthal and Hempel (1970) working with herring larvae found that Artemia nauplii were not digested as completely as copepods. We also observed that copepods become transparent in the posterior gut, whereas Artemia nauplii re- mained opaque. The variation in the numbers of prey per larva between individual menhaden and pinfish larvae increased with each successive sampling period (0, 2, 4, 6, and 8 h after feeding stopped), but fluctuated in spot. The increasing variation in menhaden and pinfish may be explained by differences in individual evacuation rates. Food densities and gut capacities were relatively con- stant for the individual larva and thus, the initial numbers of prey per larva were similar. Varied individual evacuation rates would influence the amounts present in the tracts of the fish sampled at later times and therefore increase the varia- tion. Individual fish may have significantly dif- ferent evacuation rate constants as has been shown for juvenile pinfish (Peters and Hoss 1974). DAILY RATIONS The estimated daily rations for the three larval fish species varied between 3.5 and 9.0% of the mean wet weight of the fish or from 38 to 99 copepods/fishday. The daily ration estimate for menhaden larvae (Table 7) was corrected by a factor of 2.5 to account for the fact that men- haden larvae lose approximately 60-68% of their gut contents during capture and subsequent handlings (Tables 2, 4). Since pinfish and spot larvae did not lose food from their gut when put under the stress, no correction factor was used. Two estimates of daily ration, based upon both the 1972 and 1973 haul seine-dip net collections (Figure 1), are provided for spot larvae (Table 7). The two spot rations (4.3% and 9.0% of the body weight) differ considerably, probably due to dif- ferences in food availability. Measurements of larval metabolic expenditures based on O2 consumption (D. E. Hoss and W. F. Hettler, Jr., Atlantic Estuarine Fisheries Center, Table 7. — Daily rations calculated from feeding studies and O2 con- sumption measurements at 15°-17°C for larval Atlantic menhaden, pinfish, and spot in the Newport River estuary, N.C. Species (range mm) Mean larvae wet weight (mg) Number copepods/ fish day Percent of body weight Calories/ fishday Calories/fish day estimated from02 consumption'. 2 1972. Menhaden 43 53 4.9 1.18 3.0 (27-32) Pinfish 32 38 3.5 0.63 1.2 (16-20) Spot 33 47 4.3 0.78 1.2 (17-23) 1973: Spot (17-23) 33 99 9.0 1.65 1.2 'From Hettler an d Hoss. unpubl. data. 23.38 cal/mg O2. 143 FISHERY BULLETIN: VOL. 73, NO. 1 National Marine Fisheries Service, NOAA, pers. commun.) are higher than three of our four larval ration estimates (Table 7). Our menhaden ration of 1.18 calories/fish day was 4(y7c of the 3.0 calculated from Hoss and Hettler's measurements of respiration rate, indicating that for this species our estimate is probably low. Our pinfish ration was also lower, being 527c of that calculated from the O2 consumption method. Our menhaden estimate is highly dependent on a very tentative factor used to adjust for handling effects. More accurate measurement of this conversion factor would probably provide better correlation with metabolic costs. Our 1972 spot ration was 66% of maintenance needs and may be indicative (as with pinfish and menhaden) of natural food shortages or environmental conditions not optimal for feeding on the dates of collection in 1972. The 1973 spot ration was nearly twice that estimated from O2 consumption measure- ments and provides sufficient energy for general metabolism and growth. We must consider our larval ration estimates as tentative in light of the high variability in the ration estimates for spot. This variation is due to differences in natural gut content, possibly as a result of differences in food availability on the sampling dates. Extensive sampling under the varied environmental conditions and zooplankton abundances and repeated evacua- tion rate measurements will provide us with more accurate estimates of their daily ration. ACKNOWLEDGMENTS We wish to express our sincere appreciation to Ronald L. Garner and Jerry D. Watson for their technical assistance during the entire study. LITERATURE CITED Bajkov, a. D. 1935. How to estimate the daily food consumption of fish under natural conditions. Trans. Am. Fish. Soc. 65:288-289. Blaxter, J. H. S. 1965. The feeding of herring larvae and their ecology in relation to feeding. Calif. Coop. Oceanic Fish. Invest., Rep. 10:79-88. Braum, E. 1967. The survival of fish larvae with reference to their feeding behaviour and the food supply. In S. D. Gerking (editor), The biological basis of freshwater fish produc- tion, p. 113-131. Blackwell Sci. Publ., Oxford. Brooks, J. L., and S..I. Dodson. 1965. Predation, body size, and composition of plankton. Science (Wash., DC.) 150:28-35. Elliott, J. M. 1972. Rates of gastric evacuation in brown trout, Salmo trutta L. Freshwater Biol. 2:1-18. Heinle, D. R. 1966. Production of a calanoid copepod, Acartia tonsa in the Patuxent River estuary. Chesapeake Sci. 7:59-74. IVLEV, V. S. 1961. Experimental ecology of the feeding of fishes. (Translated from Russian by D. Scott.) Yale Univ. Press, New Haven, 302 p. June, F. C, and F. T Carlson. 1971. Food of young Atlantic menhaden, Brevoortia tyrannus, in relation to metamorphosis. Fish. Bull., U.S. 68:493-512. Kjelson, M. a. 1971. Selective predation by a freshwater planktivore, the threadfin shad, Dorosoma petenense. Ph.D. Thesis, Univ. California, Davis, 123 p. Lewis, R. M., W. F. Hettler, Jr., E. P. H. Wilkens, and G. N. Johnson. 1970. A channel net for catching larval fishes. Chesa- peake Sci. 11:196-197. May, R. C. 1970. Feeding larval marine fishes in the laboratory: A review. Calif. Coop. Oceanic Fish. Invest., Rep. 14:76-83. Peters, D. S., and D. E. Hoss. 1974. A radioisotopic method of measuring food evacua- tion time in fish. Trans. Am. Fish. Soc. 103:626-629. Peters, D. S., and M. A. Kjelson. In press. Consumption and utilization of food by various postlarval and juvenile North Carolina estuarine fishes. Proc. 2nd Int. Estuarine Res. Conf., Oct. 15-18, 1973, Myrtle Beach, S.C. Peters, D. S., M. A. Kjelson, and M. T. Boyd. 1974. The effect of temperature on digestion rate in the pinfish, Lagodon rhomboides; spot, Leostomus xanthurus; and silverside, Menidia menidia. Proc. 26th Annu. Conf. Southeast. Assoc. Game Fish Comm., p. 637-643. Rosenthal, H., and G. Hempel. 1970. Experimental studies in feeding and food require- ments of herring larvae iClupea harengus L.). In J. H. Steele (editor). Marine food chains, p. 344-364. Univ. Calif Press, Berkeley. Schumann, G. O. 1965. Some aspects of behavior in clupeid larvae. Calif. Coop. Oceanic Fish. Invest., Rep. 10:71-78. Seaburg, K. G., and J. B. Moyle. 1964. Feeding habits, digestive rates, and growth of some Minnesota warmwater fishes. Trans. Am. Fish. Soc. 93:269-285. Thayer, G. W., D. E. Hoss, M. A. Kjelson, W. F. Hettler, Jr., AND M. W. LaCroix. 1974. Biomass of zooplankton in the Newport River estuary and the influence of postlarval fishes. Chesapeake Sci. 15:9-16. Tyler, A. V. 1970. Rates of gastric emptying in young cod. J. Fish. Res. Board Can. 27:1177-1189. Werner, R. G. 1969. Ecology of limnetic bluegill (Lepomis macrochirus) fry in Crane Lake, Indiana. Am. Midi. Nat. 81:164-181. 144 THE LARVAL DEVELOPMENT OF PACIFIC EUPHAUSIA GIBBOIDES (EUPHAUSIACEA) Margaret D. Knight^ ABSTRACT The larval development of Euphausia gibboides is described and illustrated, including nauplius stages I and II, metanauplius stage, calyptopis stages I-III, and furcilia stages I- VI; dominant and variant forms, with respect to reduction in number of terminal telson spines, were found in furcilia IV- VI. Identification of developmental stages was substantiated by the study of a series of juveniles oiE. gibboides, the largest of which had characters of both the furcilia phase and the adult. Larvae were studied in plankton samples from several areas within the range of the species in the Pacific Ocean; variation in size of calyptopes in different areas is described. Euphausia gibboides Ortmann is a relatively large euphausiid of the temperate and tropical Pacific. It is closely related to E. sanzoi Torelli and E. fallax Hansen and with them forms a "Euphausia gibboides group" (Brinton 1962). In the North Pacific, E. gibboides is found in the transition zone between lat. 30° and 45°N and extending southward to about lat. 20°N in the east where it is considered a major species of the California Current system; in the South Pacific it occurs in the eastern equatorial zone. Eu- phausia sanzoi has been found in the Red Sea and western Indian Ocean, and E. fallax in the west- ern tropical Pacific. The distributions of these species are discussed by Brinton (1962, 1967a, b, 1973), Brinton and Gopalakrishnan (1973), Roger (1967), and Mauchline and Fisher (1969). The distribution of the larvae of £. gibboides in the California Current is shown by Brinton (1967a, b, 1973). Hansen (1911) divided the species of the genus Euphausia Dana into four groups with respect to armature of carapace and abdomen; of these, groups A and D were considered to be "well sepa- rated" but groups B and C "somewhat badly de- fined." Group C, the largest of the four, contains 12 of the 32 species now recognized in the genus: E. mucronata, E. paragibba, E. pseudogibba, E. hemigibba, E. gibba, E. lamilligera, E. distin- guenda, E. sibogae, E. gibboides, E. fallax, E. sanzoi, and E. vallentini (E. aluae and E. con- suelae, both considered difficult to evaluate are 'Scripps Institution of Oceanography, University of Cali- fornia, San Diego, P.O. Box 109, La Jolla, CA 92037. Manuscript accepted March 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. not included) (Boden et al. 1955). An early fur- cilia of E. hemigibba (Lebour 1949) and a late furcilia of E. distinguenda (Hansen 1912) have been identified, but a series of developmental stages has been described for only one of the group C species, E. vallentini (John 1936). In his investigation of the adults and larvae of the southern species of Euphausia, John has shown the affinity of E. vallentini and certain species of group B with which it may now be associated (Mauchline and Fisher 1969). Studies of the lar- vae of additional species should aid not only in identification of planktonic forms but also in definition of specific relationships within the genus. The present paper provides descriptions of the developmental stages of Euphausia gibboides; it is part of a larger study whose purpose is to identify and describe larvae of the three species of the "Euphausia gibboides Group" and to com- pare the larval morphology of these closely related forms. METHODS AND MATERIALS Larvae ofE. gibboides were obtained from pre- served plankton samples in the Marine Inverte- brate Collections of the Scripps Institution of Oceanography. They were sorted from net hauls, taken with the standard CalCOFI (California Cooperative Oceanic Fisheries Investigations) 1-m net (Ahlstrom 1954), which were known to contain larvae and juveniles of the species. The positions of these tows are given in Table 1; station data for the samples are given by Snyder 145 FISHERY BULLETIN: VOL. 73, NO. 1 Table 1.— The area, station number, and position of samples from which larvae of £. gibboides were obtained. Position Area Cruise Station (Lat., Long.) North Pacific: Eastern CalCOFI 6304 60.140 34 65.0'N, 129°19.5W CaiCOFI 6304 70.90 34°53.0N, 125°13.0'W CalCOFI 6304 70.100 34^33.0'N, 125°13.0'W CalCOFI 6304 110.70 28°36.0'N, 118°18.0'W CalCOFI 6304 117.90 26°47.5N. 118°50.0'W CalCOFI 6304 120.120 25°12.5'N. 120=22. 5'W CalCOFI 6304 133.80 24°14.5'N, 116°17.5'W CalCOFI 6307 117.80 27°07.5N, 118°06.0W Western Transpac 56A - B 41=49.0'N, 166°38.6'E Transpac 76A 39°56.4'N. 143°38.5'E Equatorial Pacific: Eastern Shellback 187 r39.5'N, 92°05.0'W Shellback 188 r06.5'N, 93°14.5'W and Fleminger ( 1965) and in University of Cali- fornia Data Reports (Scripps Institution of Oceanography 1964a, b). The larvae were grouped by developmental phase, measured, and dissected for detailed study of appendages. The identification of eggs, nauplii, and metanauplius is based on their relative abundance in samples in which calyptopes and furcilia of E. gibboides were clearly the domi- nant euphausiid larvae. Identification of calyp- topis and furciUa stages, based on morphology, distribution, and relative abundance with juve- niles and adults of .E. gibboides, was substan- tiated by the study of a series of juvenile forms the largest of which had characters of both the furcilia phase and the adult. The identification of calyptopis I was confirmed by rearing after the manuscript had been accepted for publication. A gravid female of £^. gibboides, caught in a mid-water trawl collection at lat. 27°35.5'N, long. 115°52.0'W, deposited her eggs soon after capture and larvae which hatched from the eggs were cultured through the first four developmental stages. I am indebted to Edward Brinton and Annie Townsend who under- took the rearing study of E. gibboides aboard RV Alexander Agassiz during Leg I of Scripps Institution of Oceanography Expedition Krill, May-June 1974. Reviews of the literature dealing with the larval development of the Euphausiacea and discussions of their larval phases are given by Mauchline and Fisher { 1969) and Gopalakrishnan (1973). The nomenclature used in the descrip- tion of £. gibboides is modified from Sars (1885) as follows. Nauplius phase (two stages): Body oval, unsegmented, without compound eyes; 3 pairs of limbs present, antennulae uniramous, antennae and mandibles biramous and natatory. Metanauplius phase (one stage): Body unseg- mented, with carapace; only 2 pairs of limbs present (antennulae and antennae); mandi- bles, maxillules, maxillae, and maxillipeds (first thoracic legs) present as bud-like prominences. Calyptopis phase (three stages): Body divided into two principal sections; abdomen becomes segmented; thoracic segments develop but are much compressed; compound eyes im- perfectly developed, immobile and covered by hood-like expansion of carapace; man- dibles, maxillae, and maxillipeds distinct and functional; thoracic legs posterior to first leg and pleopods not present; uropods develop. Furcilia phase (variable number of stages): Compound eyes more fully developed, mobile, and projecting beyond sides of carapace; antennae at first retaining original natatory structure, later transformed to adult form wdth scale and developing flagellum; legs and pleopods develop; method of locomotion thus changes as setose pleopods replace modi- fied antennae for swimming; photophores develop; terminal telson spines become reduced in number, last furcilia stage with 1 terminal telson spine and 3 posterolateral spines. Juvenile phase: Begins when telson has 2 posterolateral and 1 terminal telson spines, the adult number. Individuals were straightened on a glass slide in a drop of preservative for measurement with an ocular micrometer. Measurements of develop- mental phases were as follows. Egg: Diameter of capsule and width of peri- vitelline space measured only in specimens with undeveloped embryos. Nauplius: Length between midpoints of anterior and posterior margins; width at widest point. Metanauplius: Length between midpoints of anterior margin of rostral hood and posterior margin of abdomen; width of rostral hood at widest point; width of body at widest point posterior to rostral hood; measurements exclude spinose fringe on rostral hood and telson spines. 146 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES Calyptopis: Total length between midpoints of anterior margin of carapace and posterior margin of telson; carapace length from center of anterior margin to distal point on posterior margin excluding dorsal spine; carapace width at widest point on anterolateral margins; measurements exclude spinose fringe of carapace and telson spines. Furcilia: Total length between midpoints of anterior margin of carapace and posterior margin of telson, the carapace measurement excludes spines until median spine appears and then is made from tip of spine, the telson measurement excludes spines until develop- ment of 1 terminal spine in last stage and then is taken from tip of spine; carapace length from posterior margin of orbit to distal point on posterior margin excluding spine in furcilia I; rostrum width at widest point proximal to eyestalks, excluding spines; eye height on cornea between upper and lower lobes measured in lateral view. Juvenile: Total length as in last furcilia stage. The range, mean (x), and standard deviation (SD) of each measurement with number of speci- mens measured in) is given in Tables 5-8. Larvae were placed in glycerine for dissection. The description of setation and form of appen- dages is based on dissection of at least 10 speci- mens of each developmental stage. The common form of each appendage is figured; when the setation varies within a stage, the number of appendages v^dth each setation observed is given in parentheses behind the number of setae. Only changes in setation or structure from the preced- ing stage are noted. Drawings were prepared with a Wild M-20 microscope^ equipped with drawing attachment. RESULTS Developmental Stages The following larval forms of E. gibhoides were found: nauplius phase, stages I, II; meta- nauplius phase, one stage; calyptopis phase, stages I-III; furcilia phase, stages I-VI. There was no variation in the number of stages in nauplius, metanauplius, and calyptopis phases or in the first half of the furcilia phase in which stages are defined by the pattern of pleopod development. In the later furcilia stages, usually characterized by the sequential reduction in number of terminal telson spines, dominant and variant forms were found. The features used to differentiate furcilia in the initial sorting were: number and position of setose and non- setose pleopods, form of antenna, number of terminal telson spines, total length, and relative abundance. The furcilia identified are listed in Table 2. When representatives of each stage were dis- sected and studied in more detail, two forms of the furcilia with 3 terminal telson spines were found; one was the dominant furcilia V and the other an advanced form which was comparable in size and development to the furcilia with 1 terminal telson spine. There also were two forms of furcilia with 2 terminal telson spines; the smallest was equivalent to furcilia V and the largest to furcilia VI. The relatively large furcilia with 2 and 3 telson spines considered to be var- iants of furcilia VI lacked the 2nd (middle) pair of posterolateral spines on the telson of the next instar developing beneath the cuticle and pre- sumably would be classified as juvenile after the Table 2. — The furcilia identified during initial survey. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Form of Pairs of pleopods No. terminal Stage antenna Non -setose Setose telson spines 1 natatory 1 0 7 II natatory 3 1 7 III natatory 1 4 7 IV dominant natatory 0 5 5 variant natatory 0 5 7 variant natatory 0 5 6 variant natatory 0 5 4 V dominant juvenile 0 5 3 variant juvenile 0 5 5 variant juvenile 0 5 4 variant juvenile 0 5 2 VI juvenile 0 5 1 Table 3. — Some of the characters used to group variant forms of furcilia stages IV- VI. Character Stage IV t Stage V stage VI No. terminal telson spines; Dominant 5 3 1 Variant 7, 6, 4 5,4,2 3, 2 Antenna; Form natatory juvenile juvenile Right mandible: Dentate process + + - near incisor teeth Maxillule: Pseudexopod bud - - + Pleopod 5: Endopod setae 1 2 4 147 FISHERY BULLETIN: VOL. 73. NO 1 next molt. A few of the details which helped to clarify the relationship between furcilia with variant forms are noted in Table 3. The total number of dominant and variant forms of furcilia IV- VI in samples examined and the percentage of each form within these stages is given in Table 4. The potential variation in reduction of the number of terminal telson spines was estimated by counting the number of spines developing on the telson of the next instar when possible. The range observed in each form is shown in Table 4. John (1936) in describing larvae of species of Euphausia from the Southern Ocean noted that the "furcilia stages recognized by the number of terminal spines on the telson are not such natural groups as those recognized by the char- acter and number of the pleopods"; this appears to be true as well for E. gibboides. As observed in other species (Mauchline and Fisher 1969), there is a general correlation between size of furcilia and the number of terminal telson spines (Tables 7, 8). As the larvae become larger, on the average, the number of spines usually de- creases, and stages may be characterized by size, number of spines, and relative abundance. Furcilia identified by a vdder range of develop- mental details, however, seem to be grouped more naturally. Description of Stages Nauplius I (Figure lA) Body egg-shaped, with 3 pairs of appendages. Antennule (Figure 6A) uniramous, unseg- Table 4. — The number of dominant and variant forms of furcilia IV, V, and VI observed, the percentage of each form within stage, and the variation in number of terminal telson spines on developing telson of next instar among individuals of each form. Stage No. terminal telson spines No. larvae %of stage No. terminal telson spines in next Instar IV dominant variant variant variant 5 7 6 4 242 9 17 5 88.6 3.3 6.2 1.8 5, 4, 3. or 2 5 5. 4. or 3 3 or 2 V dominant variant variant variant 3 5 4 2 122 11 14 8 787 7.1 9.0 5.2 3 or 1 3 3, 2, or 1 1 VI dominant variant variant 1 3 2 78 21 19 66.1 17.8 16.1 1 1 1 mented, with 1 seta and 2 small spines termi- nally, and 1 small subterminal spine. Antenna (Figure 7A) biramous, unsegmented; exopod with 4 setae and tiny tooth distally; endopod with 2 setae and small spine terminally and 1 subterminal seta. Mandible (Figure 7G) biramous, unsegmented; endopod and exopod each with 3 setae. Nauplius II (Figure IB) Body longer, with 2 pairs posterior spines, outer pair very small. Antennule (Figure 6B) with 2 setae and 1 spine terminally, and a small subterminal spine. Antenna (Figure 7B) with 5 setae and some- times a rudimentary 6th seta on exopod. Endopod with 3 setae and a small spine terminally, and 1 subterminal seta. Mandible as in nauplius I. Metanauplius (Figure IC, D) Carapace produced into wide rostral hood fringed with marginal spines; anterior margin with 3 or 4 relatively long pairs interspersed; posterolateral lobes curved ventrally around body; dorsal crest prominent, without spines. Abdomen short, posterior margin with median indentation and 5 pairs of spines; 3rd pair relatively long bear- ing setules, other pairs small and fused with telson, one or both of inner pair sometimes rudimentary. There are only 2 pairs of func- tional appendages. Antennule (Figure 6C) with 2 setae, 1 aesthe- tasc (sensory seta), and 1 spine terminally and a small subterminal spine. Antennal exopod and endopod (Figure 7C) articulated with basal segment which may show incipient segmentation. Exopod with 6 setae on 5 small distal segments; terminal segment, too small to be visible in figure, bears 2 setae. Endopod with 4 setae and 2 small spines distally and 1 subterminal seta on inner margin; rudi- ment of proximal 2nd marginal seta sometimes present. Mandibles, maxillules, maxillae, and maxilli- peds present as rudimentary buds. Calyptopis I (Figure 2A-C) Carapace with distinctive broad rostral hood fringed with small marginal spines; lateral 148 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES 0.1 mm Figure 1. — Nauplius I: A, dorsal view. Nauplius II: B, dorsal view. Metanauplius: C, dorsal view; D, lateral view. margins constricted behind eyes; posterior margin produced into strong dorsal spine; dorsal crest prominent. Compound eyes widen as they develop during stage, striated body of photophore visible. Thoracic segments may be visible; abdomen unsegmented. Antennule (Figure 6D) 2-segmented, basal segment with 2 dorsal setae, 1 medial seta, and medial spine on distal margin; small terminal segment with 2 aesthetascs, 3 setae, 1 strong medial spine and tiny spine. Antenna (Figure 7D) with 2-segmented proto- pod. Exopod wdth 7 setae on 5 distal segments; terminal segment wdth 3 setae, subterminal seg- ments with 1 seta each. Endopod with 4 terminal setae and 2 setae on inner margin, the proximal marginal seta may be rudimentary; in addition to setules, distal marginal seta and 2 terminal setae bear small spinules and 3rd terminal seta armed with proximal row of comblike setules. This setation remains unchanged until furcilia V. Mandibles (Figure 7H) asymmetrical; both with narrow plate near pars molaris and tuft of setae at base of plate; right mandible with dentate process near incisor teeth; when mandibles close dentate process bends inward toward mouth, the lower plates overlap. Conical anterolateral process and small prominent lateral knob pres- ent; lateral knob disappears in furcilia I, and anterolateral process decreases in size gradually up to late furcilia stages. Maxillule (Figure 8A) with 6(1) or 7(20) setae 149 FISHERY BULLETIN: VOL. 73, NO. 1 0 5 rnm Figure 2.— Calyptopis I: A, dorsal view; B, lateral view; C, development of eyes within stage. Calyptopis II: D, dorsal view. Calyptopis III: E, dorsal view. on coxal endite, 1 of 2 large setae distinctively armed distally with strong triangular spines rather than setules; basal endite with 3 spines armed with spinules. Endopod 2-segmented, terminal segment with 3 and proximal segment with 2 setae. Exopod a small lobe with 4 plumose setae. Setation of endopod does not change until furcilia V and that of exopod does not change throughout larval development. Maxilla (Figure 8H) with 5 setose lobes on inner 150 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES margin (proximal 2 considered coxal and distal 3 basal although segmentation is unclear); seta- tion of medial lobes 1-5 progressing distally is 8-4-4-4-3; 2 setae on lobe 1 and 1 seta on lobes 2-4 situated submarginally on posterior face. Endopod 1-segmented with 3 setae; exopod repre- sented by 1 plumose seta on lateral margin. There is no change in setation in calyptopis phase. Maxilliped (Figure 9A) usually with 5 setae on coxa, 4 marginal and 1 (sometimes absent) on posterior face, 4(3) or 5( 17) setae were observed. Basis Math 6 setae. Endopod 2-segmented; termi- nal segment with 4 and proximal segment with 3 setae; 1 distal seta on basis and 1 on proximal segment of endopod situated submarginally on posterior face; 1 marginal seta on basis and 1 on first segment of endopod relatively short and stout with tiny marginal spinules. Exopod with 4 terminal setae and 1 proximal seta near in- distinct articulation with basis. Fine marginal hairs present as figured. Telson with 1 pair of lateral spines, 3 pairs of posterolateral spines and 6 terminal spines, posterolateral spine 3 (inner) slightly longer than central posterolateral spine 2; terminal spines and posterolateral spine 3 armed with spinules on lateral margins, lateral spines and posterolateral spines 1 and 2 with spinules on inner margins only. Calyptopis II (Figure 2D) Calyptopis III (Figure 2E) Carapace with rudiment of small denticle on posterolateral margin present in furcilia I (Figure 4A). Abdomen with 6 segments; 6th segment, now separate from telson, with pair of biramous uropods. Antennule (Figure 6F) with 3-segmented peduncle, basal segment produced distally into strong lateral spine extending to or slightly beyond tip of inner ramus, inner margin of spine setose. Peduncle segments 1-3 with 1-2-2 plumose setae on inner margins; segment 3 with dorsal lobe bearing 3 setae on distal margin; basal segment with 1 large lateral seta at base of spine. Inner flagellum about two-thirds length of outer flagellum and may have 3rd terminal seta; other- wise setation of rami unchanged. Mandible armature (Figure 71) unchanged. Maxillule with 7 setae on coxal endite and 5 spines on basal endite; no variation observed. Maxilla usually unchanged, with setation of 8-4-4-4-3 on lobes 1-5; lobe 3 varied with 3(1) or 4(20) setae and lobe 5 with 2(1) or 3(20) setae. Maxilliped (Figure 9B) usually with 6 setae on coxa, 5(3) or 6(16) setae were observed. Uropod (Figure IIQ) biramous; protopod with ventral spine above endopod; exopod with strong posterolateral spine, 2 small spines and 2 setae distally; endopod incompletely articulated with protopod, bearing 1 spine and 2 setae distally and 1 small subterminal dorsally projecting seta. Telson armature unchanged. Broad rostral hood of carapace with more pronounced inward curve between eyes; dorsal crest less prominent. Abdomen with 5 segments. Antennule (Figure 6E) biramous. Peduncle unsegmented but may be constricted with seg- mentation of calyptopis III visible beneath cuticle; distal margin with 3 dorsal setae and inner margin with 1 seta. Outer ramus with 2 aesthe- tascs, 1-3 setae and 2-4 spines terminally; inner ramus short, with 2 setae and 1-4 spines. Maxillule (Figure 8B) with 6(3) or 7(15) setae on coxal endite; basal endite with 5 spines. Maxilliped with 4(1) or 5(19) setae on coxa. Telson, with addition of small median spine, armed with 7 terminal spines; posterolateral spine 2 now longest; lateral and posterolateral spines with relatively large dorsal spinule slightly more than halfway to tip. Furcilia I (Figures 3A, 4A) Eyes large, stalked and moveable, with 3-lobed appearance due to arrangement of ommatidia and concentrations of pigment as well as con- strictions in cornea; lower lobe largest and most distinctly defined; convex middle lobe especially contributes to characteristic shape of eye. Cara- pace emarginate behind eyes; rostrum broad, blunt, fringed with small spines; posterior margin produced into dorsal spine; posterolateral margins with denticle; dorsal crest near midlength. First segment of abdomen with pair of non-setose pleopods; developing photophore between pleo- pods sometimes with faint pigment. Small anal spine present. Antennule (Figure 6G) with lateral spine of peduncle segment 1 extending to distal margin 151 FISHERY BULLETIN: VOL. 73, NO. 1 05 mm I 1 Figure 3.— Dorsal view: A, furcilia I; B, furcilia II; C, furcilia III; D, furcilia IV; E, furcilia V; F, furcilia VI. of segment 3; spine with 5 pairs of setae spaced along inner margin and small setae between; peduncle segments 1 and 2 each with 2 plumose setae on inner margin and small dorsal setae; segment 3 with 3 setae on inner margin, a 4th slightly ventral seta on distal margin, and 4 setae on dorsal lobe. Flagella usually of equal length; outer ramus with 1 aesthetasc at about midlength of inner margin; terminal setation of flagella apparently unchanged but too frequently broken to determine. In subsequent furciliar stages num- bers of setae on dorsal surface increase but 152 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES 5 mm Figure 4.— Lateral view: A, furcilia I; B, furcilia II; C, furcilia III; D, furcilia IV. 153 FISHERY BULLETIN: VOL. 73, NO. 1 number of plumose setae on medial margin re- main the same; the lateral spine on segment 1 gradually decreases in length. Maxillule with 6(2) or 7(20) setae on coxal endite; basal endite (Figure 8C) with 6(1) or 7(21) spines. Maxilla (Figure 81) usually with setation of 8-4-5-4-3 on inner lobes 1-5; lobe 3 now bears 5 setae; lobe 1 variable, with 7(1) or 8(20) setae. Maxilliped with 5 setae on terminal segment of endopod (Figure 9C); coxa with 5(3) or 6(18) setae. Leg 2 (Figure lOA) present, rudimentary; bud of leg 3 sometimes visible. Pleopod (Figure IIL) non-setose and unseg- mented, or with incipient segmentation and bud of endopod. Uropod (Figure IIR) with 6 plumose setae on exopod; endopod articulated with protopod, bear- ing 6 marginal plumose setae and 3-5 small dorsal setae. Telson (Figure 12 A) with posterolateral spine 2 relatively longer. Furcilia II (Figures 3B, 4B) Rostrum of carapace a little narrower, with smaller marginal spines; posterior margin with- out dorsal spine; lobes of eye more defined (Figure 6K). Abdomen with 1 pair setose and 3 pairs non-setose pleopods on segments 1-4 respectively; photophore on segment 1 pigmented and functional, developing photophore on seg- ment 4 sometimes with faint pigment. Antennule (Figure 6H) with 5 setae on dorsal lobe of peduncle segment 3 one of which projects dorsally; this setation, with dorsally oriented seta becoming longer and stronger, is found in subse- quent furcilia stages. Flagella now approximately as long as 3rd segment of peduncle. Maxillule (Figure 8D) with 7(1) or 8(23) setae on coxal endite; basal endite with 7 marginal spines and often, in 16 of 24 appendages, with small seta on proximal margin. Maxilla usually with setation of 8-4-5-5-3; lobe 3 variable with 5(23) or 6(1) setae and lobe 4 (Figure 8J) with 4(3) or 5(21) setae. Maxilliped usually with 6 setae on terminal segment of endopod (Figure 9D), 5(3) or 6(21) setae were observed; coxa with 5(1) or 6(23) setae. Leg 2 (Figure lOB) with endopod bearing 2 terminal setae and unsegmented or with 2 or 3 weakly defined segments; exopod rudimentary, without setae; gill bilobed; developing photo- phore on coxa sometimes with faint pigment. Leg 3 (Figure lOH) rudimentary, or with bud of exopod and gill. Bud of leg 4 may be present. Setose pleopod 1 (Figure IIM) with 6 plumose setae on exopod, small endopod with single seta and median hook; non-setose pleopods 2-4 as in furcilia L Uropod (Figure US) with 8(21) or 9(1) setae on exopod, endopod with 7 marginal and 11 or 12 dorsal setae. This is the last stage in which numbers of setae can be counted; in preserved specimens the marginal setae are too frequently broken to attempt enumeration. Telson (Figure 12B) narrower, posterolateral spine 3 wider basally. Furcilia III (Figures 3C, 4C) Carapace with rostrum narrowing, anterior marginal spines may be very small remnants. Abdomen with 4 pairs setose and 1 pair non- setose pleopods on segments 1-4 respectively; photophores on segments 1 and 4 pigmented and functional; developing photophore on segment 2 sometimes with faint pigment. Antennular flagella (Figure 61) almost twice as long as peduncle segment 3 and may be 2- segmented; outer flagellum with 2 aesthetascs on inner margin one of which bifurcates distally. Mandible with anterolateral process about one- half as long as that figured for calyptopis IIL Maxillule with 8 setae on coxal endite; basal endite (Figure 8E) with 7(1), 8(4), or 9(17) spines on medial margin and 1 small seta on proximal margin. Maxilla usually with setation of 8-4-6-5-3; lobe 3 (Figure 8K) now with 5(2) or 6(20) setae; lobe 5 variable with 2(1) or 3(20) setae. Maxilliped with 5(3) or 6(16) setae on coxa and 6 on terminal segment of endopod. Leg 2 (Figure IOC) with endopod 5-segmented, articulation with basis indistinct, setation variable, terminal segment with more than 2 setae; exopod with 0(7), 1(5), or 2(5) setae; gill bilobed; photophore pigmented and functional. Leg 3 (Figure 101) with endopod unsegmented or with a few (less than 5) weakly defined segments, setation variable, distal segment usually with 2 terminal setae, 2(21) or 3(1) setae were observed; exopod rudimentary, without setae; gill bilobed. 154 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES Leg 4 (Figure ION) rudimentary, with bud of exopod and small bilobed or simple bud of gill; endopod usually without terminal setae, 0(22) or 1(2) seta were observed. Leg 5 present as bud. Leg 7 rudimentary with gill bud and developing photophore. Setation of pleopods on abdominal segments 1-4 as follows: pleopod 1 (Figure UN) — endopod 2, exopod 6(8), 7(11), or 8(1); pleopods 2-4 — endopod 1, exopod 6. Non-setose pleopod of segment 5 as in furcilia L Endopod of pleopod 1 with appendix interna, a small medial lobe with tiny hooks. Telson (Figure 12C) narrower; posterolateral spine 3 quite broad, inner margin smooth except for 1 or 2 tiny distal spinules near larger dorsal spinule. Five terminal spines of furcilia IV may often be seen beneath integument. Furcilia IV (Figures 3D, 4D) Rostrum of carapace usually with smooth margin, there may be tiny remnants of marginal spines but no median spine. Abdomen with 5 pairs of setose pleopods; photophores on segments 1, 2, and 4 pigmented and functional; developing photophore on segment 3 sometimes with faint pigment. Antennular flagella (Figure 6J) with about 6 or 7 segments, segmentation usually indistinct; outer flagellum with 3 aesthetascs, 1 proximal to pair on medial margin, 1 aesthetasc no longer bifurcate. Maxillule with 8(12) or 9(10) setae on coxal endite (Figure 8F); basal endite with 8(1) or 9(21) marginal spines and 1 seta on proximal margin. Maxilla usually unchanged, with setation of 8-4-6-5-3; lobe 4 variable with 5(20) or 6(2) setae. Maxilliped (Figure 9E) with 5(4) or 6(18) setae on coxa; basis with 6(9) or 7(10) setae. Endopod becoming 3-segmented as small terminal segment forms wdth setation of 3-2-4 for segments 1-3. Leg 2 (Figure lOD) with endopod larger, more setose, and becoming geniculate with terminal 3 segments reflexed as in adult; exopod with 4(3) or 5(6) setae (seldom intact); gill bilobed. Leg 3 (Figure lOJ) with endopod 5-segmented, some- times slightly reflexed, articulation with basis indistinct, setation variable, terminal segment with more than 2 setae; exopod with 2(1), 3(1), or 4(12) setae; gill bilobed. Leg 4 (Figure lOO) endopod with few (less than 5) weakly delineated segments, terminal segment with 2 setae, other setation variable; exopod usually without setae, 1 of 15 appendages examined with 1 seta; gill bilobed. Leg 5 (Figure llA) rudimentary with bud of exopod and bilobed or simple bud of gill. Bud of leg 6 present. Leg 7 (Figure 111) with gill bilobed or with small bud of 3rd lobe, photophore may have pigment. Leg 8 (Figure 1 IF) represented by bilobed or trilobed gill. Pleopod setation as follows: pleopod 1 (Figure 110) — endopod 3(2) or 4(17), exopod 8; pleopod 2 — endopod 2, exopod 7(1) or 8(19); pleopod 3 — endopod 2, exopod 7(3) or 8(15); pleopod 4 — endopod 2, exopod 7(8) or 8(9); pleopod 5 — endo- pod 1, exopod 6. Telson (Figure 12D) with 5 terminal spines, the 3 terminal spines of next instar often visible beneath cuticle. VARIANT FORMS. — A small furcilia IV with 6 telson spines was less mature in that leg 4 had 1 terminal seta and exopods of pleopods 3 and 4 had only 6 setae. A furcilia IV with 4 telson spines showed bud of 3rd lobe of gill on leg 2. Furcilia V (Figures 3E, 5A) Rostrum usually wdth smooth margin, there may be a very small median spine. Photophores on abdominal segments 1-4 now pigmented. Antennule with lateral spine of peduncle seg- ment 1 extending to about midpoint of segment 3, none of the specimens available had flagella intact. Antenna (Figure 7E) transformed, no longer natatory; basal segment with distolateral spine; endopod with 8 segments, 3 peduncular and 5 flagellar, division of terminal segment not always distinct; exopod (scale) with 13 or 14 plumose marginal setae. Mandible (Figure 7J) with anterolateral pro- cess now considerably reduced in size. Maxillule with 8(1) or 9(22) setae on coxal endite; basal endite with 9(21) or 10(2) marginal spines and 1(21) or 2(2) small setae on proximal margin. Endopod with segmentation weak or indistinct; in 6 of 21 appendages examined 1- segmented v^dth 1 seta on lateral margin as figured for furcilia VI (Figure 8G). Maxilla usually with setation of 8-4/5-6-6-3; lobe 2 (Figure 8L) variable with 4(15) or 5(9) setae and lobe 4 with 4(1), 5(4), or 6(19) setae. 155 FISHERY BULLETIN; VOL. 73. NO. 1 Figure 5.— Lateral view: A, furcilia V; B, furcilia VI. Maxilliped (Figure 9F) with increasingly var- iable setation; coxa with 5(2), 6(12), 7(9), or 8(1) setae; basis with 6(2), 7(9), or 8(12) setae. Endo- pod lengthened, usually with 3 segments; 4 of 20 appendages examined with 4 or 5 segments indicated, distal segments weakly delineated; setations of 3-2-4, 3-1-1-4, and 3-2-0-1-4, pro- gressing distally, were observed. Leg 2 (Figure lOE) with dactyl of endopod becoming modified; exopod with 6 setae; gill usually with bud of 3rd lobe. Leg 3 (Figure lOK) with endopod reflexed, longer and more setose; exopod with 5(8) or 6(8) setae; gill with bud or sizeable rudiment of 3rd lobe. Leg 4 (Figure lOP) with endopod 5-segmented, articulation with basis never clear, setation variable, terminal seg- ment with more than 2 setae; exopod with 4(16) or 5(1) setae; gill bilobed. Leg 5 (Figure IIB) with endopod unsegmented or weakly segmented with less than 5 segments, with 1(9) or 2(13) terminal setae and sometimes a few marginal setae; exopod with 0(22) or 1(1) seta; gill bilobed. Leg 6 (Figure IID) rudimentary with gill bud, may be slightly bifid. Leg 7 (Figure IIJ) with pigmented photophore; gill sometimes wdth bud of 3rd or 4th lobes. Leg 8 (Figure llG) ramified, with varying numbers of lobes. Setation of pleopods as follows: pleopod 1 — endopod 4, exopod 8( 11), 9(5), or 10(2); pleopod 2 — endopod 4, exopod 8(7), 9(11), or 10(1); pleopod 3 — endopod 4, exopod 8(15) or 9(5); pleopod 4 — endopod 3(1) or 4(19), exopod 8; pleopod 5 — endopod 2, exopod 6(1), 7(17), or 8(5). Telson (Figure 12E) narrow, with 3 terminal spines, 3rd pair of posterolateral spines lengthen- ing relatively; single terminal spine of final furcilia may often be seen beneath cuticle. VARIANT FORM. — A small furcilia V with 5 telson spines had antennal flagellum of about 5 segments, endopod of leg 5 without terminal setae, and endopod of pleopods 3-5 with 2, 2, and 156 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES 1 setae respectively, one of the endopods of pleo- pod 5 had rudiment of 2nd seta. Furcilia VI (Figures 3F, 5B) Rostrum usually with small median spine. Antennule with lateral spine of peduncle seg- ment 1 about as long as segment 2; outer flagellum may have additional proximal aesthetasc; one apparently intact inner flagellum with 10 segments. Antennal scale (Figure 7F) with approximately 15 or 16 setae; one intact flagellum with 3 peduncular and 11 flagellar segments. Right mandible (Figure 7K) now without dentate process near incisor teeth; toothed plates relatively smaller; rudimentary palp may begin to increase in size. Maxillule (Figure 8G) with 9(15) or 10(9) setae on coxal endite; basal endite with 9(4) or 10(20) marginal spines and 1(8) or 2(16) small setae on proximal margin. Endopod of 1 segment with 1 proximal lateral seta; terminal and medial seta- tion unchanged. Coxa now with rudiment of pseudexopod. Maxilla (Figure 8M) usually with setation of 8-6-6-6-3; lobe 2 variable with 4(2), 5(4), or 6(18) setae. Exopod represented by 1 (22) or 2(2) setae; endopod rounder. Maxilliped (Figure 9G) modifying to adult form; coxa with 5-9 setae, long seta on posterior face no longer present; basis with 8(22) or 9(2) setae. Endopod of 5 segments with variable setation; articulation with basis not clear. Exopod still with 4 terminal setae. Leg 2 (Figure lOF, G) with endopod more setose, dactyl with a few more "cleaning" comb setae; exopod with 6 setae; gill trilobed. Leg 3 (Figure lOL, M) with long terminal setae on dactyl of endopod; exopod with 6 setae; gill with bud of 3rd lobe or trilobed. Leg 4 (Figure lOQ) with endopod reflexed; exopod with 5(1) or 6(14) setae; gill with bud or larger rudiment of 3rd lobe. Leg 5 (Figure llC) with endopod 5- segmented and setation variable, terminal seg- ment with more than 2 setae; exopod with 1(2), 2(4), 3(2), or 4(10) setae; gill trilobed. Leg 6 (Figure HE) with endopod unsegmented and non- setose; exopod without setae; gill trilobed; exopod and gill may be rudimentary. Legs 7 (Figure IIK) and 8 (Figure IIH) with increasing num- ber of gill lobes. Pleopods with setation as follows: pleopod 1 (Figure HP) — endopod 4(5), 5(10), or 6(7), exopod 9(7) or 10(12); pleopod 2 — endopod 4(14), 5(7), or 6(2), exopod 9(2), 10(13), or 11(3); pleopod 3 — endopod 4(17), 5(1), or 6(3), exopod 9(2), 10(13), or 11(1); pleopod 4 — endopod 4(21) or 6(2), exopod 9(10) or 10(7); pleopod 5 — endopod 3(1) or 4(20), exopod 8. Telson (Figure 12F) quite slender with 1 termi- nal spine and 3 pairs posterolateral spines; posterolateral spine 2 was missing on one side in 5 of 12 larvae dissected. Developing telson of next instar is without spine 2 on either side. VARIANT FORMS. — In furcilia VI with 2 and 3 telson spines, basis of maxilliped sometimes with 7 setae and exopod of leg 2 with 6, 7, or 8 setae. Once in furcilia with 3 telson spines, lobe 3 of maxilla with 7 setae and right mandible with tiny remnant of dentate process. Measurements The eggs assumed to be those of E. gibboides have a relatively wide perivitelline space. The measurements, in millimeters, of 100 eggs from one sample (6304- 110. 70) are: diameter of capsule, range = 0.61-0.75, x = 0.69, SD = 0.03; peri- vitelline space, range = 0.13-0.19, x = 0.16, SD = 0.01. The measurements of developmental stages are given in Tables 5-8. The growth factor (mean length in stage divided by mean length in pre- ceding stage) for dominant forms is as follows: Growth Growth Stage factor Stage factor Furcilia I 1.23 Nauplius II 1.04 Furcilia II 1.17 Metanauplius 1.08 Furcilia III 1.14 Calyptopis I 2.03 Furcilia IV 1.10 Calyptopis II 1.55 Furcilia V 1.10 Calyptopis III 1.39 Furcilia VI 1.12 There was variation in size of comparable developmental stages between the different areas from which samples were studied. The lengths of calyptopis stages are compared in Table 9. The larvae sampled in April 1963 (Cruise 6304) in the eastern North Pacific became larger, on the average, during the calyptopis phase in the 157 FISHERY BULLETIN: VOL. 73, NO. 1 Figure 6.— Antennule: A, nauplius I; B, nauplius II; C, metanauplius; D, calyptopis I; E, calyptopis II; F, calyptopis III; G, furcilia I; H, furcilia II; I, furcilia III; J, furcilia IV. Eye: K, furcilia II. 158 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES Figure 7.— Antenna: A, nauplius I; B, nauplius II; C, metanauplius; D, calyptopis I; E, furcilia V; F, furcilia VI. Mandibles: G, nauplius I; H, calyptopis I; I, calyptopis III; J, furcilia V; K, furcilia VI. 159 FISHERY BULLETIN: VOL. 73, NO. 1 Figure 8. — Maxillule: A, calyptopis I; B, calyptopis II; C, furcilia I, basal endite; D, furcilia II; E, furcilia III, basal endite; F, furcilia IV, coxal endite; G, furcilia VI. Maxilla: H, calyptopis I; I, furcilia I; J, furcilia II, lobe 4; K, furcilia III, lobe 3; L, furcilia V, lobe 2; M, furcilia VI. 160 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES Figure 9. — Maxilliped (leg 1): A, calyptopis I; B, calyptopis III; C, furcilia I, endopod; D, fxircilia II, endopod; E, furcilia IV; F, furcilia V; G, furcilia VI. 161 FISHERY BULLETIN: VOL. 73, NO. 1 Figure 10.— Thoracic legs. Leg 2: A, furcilia I; B, furcilia II; C, furcilia IE, D, furcilia IV; E, furcilia V; F, furcilia VI; G, dactyl, furcilia VI. Leg 3: H, furcilia II; I, furcilia III; J, furcilia IV; K, furcilia V; L, furcilia VI; M, dactyl, furcilia VI. Leg 4: N, furcilia III; O, furcilia IV; P, furcilia V; Q, furcilia VI. 162 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES Figure 11.— Thoracic leg 5: A, furcilia IV; B, furcilia V; C, furcilia VI. Leg 6: D, furcilia V; E, furcilia VI. Leg 7: I, furcilia IV; J, furcilia V; K, furcilia VI. Leg 8: F, furcilia IV; G, furcilia V: H, furcilia VI. Pleopod 1: L, furcilia I; M, furcilia II; N, furcilia III; O, furcilia IV; P, furcilia VI. Uropods: Q, calyptopis III; R, furcilia I; S, furcilia II. more northern areas. In the sample from August 1963 (6306-117.80), the sizes of developmental stages were similar to those found in the same general area in the spring. There is insufficient information at this time to consider the effects of environmental conditions on the rate of larval growth and development in E. gibboides, but similar variation has been observed in other species of euphausiids (Einarsson 1945; Mauch- hne 1965). The range and mean of carapace width in calyptopis stages expressed as percent of cara- pace length is given in Table 10 as the propor- tional anterolateral expansion of carapace appears to be a useful character for identifica- tion of E. gibboides. Comparison by area shows 163 FISHERY BULLETIN: VOL. 73, NO. 1 0.1 mm I 1 Figure 12.— Telson: A, furcilia I; B, furcilia II; C, furcilia III; D, furcilia IV; E, furcilia V; F, furcilia VI. that the average ratio tends to increase in northern and western Pacific samples. Juveniles There was a good series of related juvenile euphausiids in the net haul from station 6304- 117.90. Sixty-four were measured and examined in some detail. The smaller juveniles had the distinctive 3-lobed eye described for furcilia stages ofE. gihboides while some of the larger individ- uals had the characteristic eye as well as a small dorsal spine on the posterior margin of the 3rd segment of the abdomen and a small dorsal lappet with triangular pointed tip on the margin of the 1st segment of the antennule. The abdominal spine and shape of rudimentary lappet together with the relatively large eye identify the juve- niles as E. gibboides (Boden et al. 1955). The shape of the eye provides continuity with the 164 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES larvae described and confirms their identification. The juveniles examined ranged from 5.3 to 8.2 mm in total length and a dorsal spine ap- peared on the 3rd segment of the abdomen at a length of 6.8 mm. A 3-lobed eye was found in a 7.2-mm individual. At 7.0 mm, the con- striction between the upper and middle lobes of the eye may disappear; the lower large lobe re- mains well defined and, although the pigment Table 5. — Measurements of nauplius and metanauplius stages. still appears darker in the lateral position of the middle lobe, the eye becomes increasingly 2- lobed in appearance. The antennule may have a small lappet in 6.3-mm individuals, but the pointed triangular tip was not seen in animals less than 7.0 mm in length, and then it was not always directed outward as in the adult. The lobe and keel of the 2nd and 3rd antennular segments respectively were not developed. A rostral spine may be missing or very small in the early Total length Width of body Width of rostral hood Table 8.— Measurements of furcilia stages IV-VI. stage (mm) (mm) (mm) Total Carapace Eye Nauplius 1: length length height Range 0.48-0.53 0.29-0.32 — Stage (mm) (mm) (mm) X 0.51 0.30 SO 0.02 001 Furcilia IV: n 12 12 dominant — 5 ts (telson spines) Nauplius II: Range 4.20-4.93 1.01-1 19 0.32-0.38 Range 0.51-056 0.28-0.32 X 4.53 1.10 0.34 X 0.53 0.30 SD 0.15 0.04 0.02 SD 0.01 0.01 n 58 56 59 n 12 12 variant — 7ts Metanauplius: Range 4.34-4.85 1.05-1.17 0.32-0.36 Range 0.53-0.61 0.28-0.36 0.38-0.45 X 4.63 1.11 0.34 X 0.57 0.32 0.42 SD 0.16 0.05 0.01 SD 0.02 0.02 0.02 n 7 5 7 n 38 38 38 variant — Range 6ts 4.16-4.79 1.03-1.17 0.30-0.36 X 4.47 1.08 0.34 Table 6. — Measurements of calyptopis stages. SD 0.20 0.05 0.02 n variant — 11 10 10 Total Carapace Carapace 4ts length wicjth length Range 4.36-4.61 1.07-1.15 0.34-0.36 Stage (mm) (mm) (mm) X SD 4.51 0.12 1.10 0.03 0.34 0.01 Calyptopis 1: n 5 4 5 Range 1.09-1.27 0.59-0.71 0.69-0.79 Furcilia V: X 1.16 0.64 0.72 dominant — 3 ts SD 0.03 0.03 0.02 Range 4.61-5.41 1.11-1.31 0.36-0.40 n 124 124 124 X 4.98 1.20 0.37 Calyptopis II: SD 0.21 0.06 0.02 Range 1.66-1.98 0.65-0.87 0.79-0.93 n 46 43 46 X 1.80 0.76 0.86 variant — 5ts SD 0.07 0.05 0.03 Range 4.57-5.25 1.09-1.27 0.36-0.40 n 158 158 157 X 4.87 1.17 0.37 Calyptopis III: SD 0.24 0.05 0.01 Range 2.34-2.71 0.75-0.99 0.89-1.09 n 10 9 9 X 2.51 0.86 1.00 variant — 4ts SD 0.09 0.06 0.05 Range 4.65-5.33 1.11-1.31 0.34-0.38 n 149 149 148 X 4.96 0.20 1.19 0.06 0.37 0.01 SD n 13 11 13 Table 7 . — Measurements of furcilia stages MIL variant — 2ts Range X 5.01-5.29 5.19 1.17-1.29 1.25 0.36-0.40 Total Carapace Rostrum Eye 0.38 length length width height SD 0.09 0.04 0.02 stage (mm) (mm) (mm) (mm) n Furcilia VI: 7 6 8 Furcilia 1: dominant — 1 ts Range 2.85-3.37 0.75-0.85 0.51-0.65 0.24-0.28 Range 5.13-5.90 1.17-1.45 0.38-0.44 X 3.09 0.80 0.58 0.25 5.58 1.33 0.40 SD 0.10 0.02 0.03 0.01 SD 0.21 0.06 0.02 n 123 104 107 109 n 36 35 35 Furcilia II: Range 3.19-3.94 0.79-097 — 0.26-0.32 variant — Range X 3 ts 5.13-5.78 1.19-1.43 0.38-0.40 X 3.61 0.88 0.29 5.40 1.28 0.40 SD 0.14 0.04 0.01 SD 0.17 0.05 0.01 n 104 98 103 n 15 14 15 Furcilia III: Range 3.80-4.57 0.91-1.09 — 0.30-0.34 variant — Range x 2 ts 5.17-5.78 1.23-1.41 0.38-0.42 X 4.10 1.02 0.32 5.46 1.30 0.40 SD n 0.13 143 0.04 82 0.01 83 SD n 0.21 11 0.05 10 0.01 7 165 FISHERY BULLETIN: VOL. 73. NO. 1 Tablk 9. — Variation in total length of calyptopis stages between the different areas from which samples were studied. Calyptop is! Calyptopis II Calyptop sill Sample Range X SD n Range X SD n Range X SD n Equatorial Pacific: Eastern Shellback 187 + 188 1.11-1.17 1.14 0.02 9 1.70-1.94 1.82 0.07 13 2.36-2.63 2.49 0.07 21 North Pacific: Eastern 6304-133.80 1.11-1.21 1.16 0.03 19 1 66-1.82 1.76 0.04 20 2.36-2.55 2.43 0.05 20 6304-120.120 1.09-1.19 1.15 0.02 20 1.66-1.80 1.72 0.04 20 2.34-2.50 2.40 0.05 9 6304-117.90 1.11-1.21 1.15 0.03 20 1.68-1.88 1.77 0.05 20 236-2.55 2.45 0.05 20 6304-110.70 1.09-1.19 1.16 0.03 20 1.72-1.86 1.80 0.04 20 2.34-2.67 2.51 0.09 20 6304-70.90 + 100 1.09-1.21 1.15 0.04 21 1.72-1.92 1.82 0.05 20 246-2.69 2.56 0.08 11 6304-60.140 1.13-1.27 1.19 0.04 11 1.82-1.98 1.88 0.05 19 255-2.71 2.62 0.05 15 6307-1 17.80 1.09-1.17 1.13 0.02 10 1.72-1.86 1.80 0.05 10 2.40-2.59 2.50 0.06 10 Western Transpac 56A + B — — — — 1.80-1.86 1.84 0.02 6 2.46-2.71 2.58 0.08 11 Transpac 76A 1.11-1.23 1.18 0.06 4 1.68-1.90 1.81 0.06 20 2.44-2.65 2.55 0.06 20 juveniles; it is well developed in larger indi- viduals but never more than one-half the length of the eyestalk in specimens examined. DISCUSSION The larvae of many species of the genus Euphausia have not been studied but, although preliminary, it may be useful to note ways in which E. gibboides larvae differ from related identified forms. The described larvae of Eu- phausia which have features such as armature of carapace or telson similar to those of E. gibboides during some phase of development belong to the following species: Group A E. brevis — (Gurney 1942) E. krohnii — (Sars 1885; Lebour 1926; Frost 1934) E. diomediae — (Ponomareva 1969) E. eximia — (author unpubl.) Group B E.pacifica — (Boden 1950; Banse and Komaki 19663; author unpubl.) Group D E. longirostris — (Tattersall 1924; John 1936) E. spinifera — (Tattersall 1924; John 1936; Sheard 1953) Euphausia sp. (Ruud 1932; Lebour 1949; Boden 1955) A metanauplius with marginal fringe of spines on the rostral hood of the carapace is found in E. brevis, E. krohnii, E. eximia, E. diomediae, E. pacifica, and Ruud's E. sp. as well as in E. gibboides. The metanauplius figured by Ruud differs from the others, however, in that the ^Banse, K., and Y. Komaki. 1966. Studies of Euphausiidae (Crustacea) off the Washington and Oregon coasts. Annual Re- port to NSF(Natl. Sci. Found.), Grant No. GB-3360, 6 p. Unpubl. Table 10. — Carapace width expressed as percent of carapace length in calyptopis stages of £. gibboides (the number measured is given in Table 9). Calyptop is 1 Calyptop IS II Calyptop s III Sample Range X Range X Range X Equatorial Pacific: Eastern Shellback 187 + 188 86 1-91.4 89.3 85.7-92.9 90.1 83.7-91.8 87.7 North Pacific: Eastern 6304-133.80 83.3-91.4 87.4 83.3-90.0 86.0 80.0-857 83.1 6304-120.120 83.3-91.2 87.1 82.1-90.0 85.5 80.4-88.6 83.5 6304-117.90 853-94.4 88.1 82.5-90.7 86.4 79.6-87.5 83.4 6304-110.70 82.8-91.4 86.8 82.9-90.5 86.5 80.0-87.8 83.8 6304-70.90 + 100 86.1-94.4 90.1 860-97.7 91.6 80.0-94.0 86.0 6304-60.140 86.5-91.9 90.1 88.9-95.3 91.0 86.8-92.4 88.8 6307-117.80 87.9-91.4 88.9 87 8-905 89.3 90 0-86.3 83.3 Western Transpac 56A + B — — 90.5-97.6 94.4 88.2-95.8 90.4 Transpac 76A 83.8-91.7 878 85.7-93.0 89.6 84.6-906 88.0 166 KNIGHT: DEVELOPMENT OF EUPHAUSIA GIBBOIDES entire margin of carapace, not only the rostral hood, is spinose. Euphausia brevis, E. krohnii, and E. eximia, unlike E. gibboides, have two small dorsal spines on the carapace; E. dio- mediae, the only other species of Group A identi- fied has instead a "sharp eminence" which, as figured (Ponomareva 1969, Figure Ic), is con- siderably higher and sharper than the dorsal prominence of E. gibboides. The metanauplius o{ E. pacifica has a dorsal crest more like that ofE. gibboides but may prove, with further study, to be consistently smaller; 25 specimens mea- sured from one location by the author ranged from 0.44 to 0.48 mm in total length with an average of 0.46 mm. A metanauplius with fringed rostral hood and two small dorsal spines is figured by Boden (1955) as one of the larval stages o{E. lucens. It appears, however, that the larvae are those of another species of the genus (Bary 1956), and the form of the metanauplius suggests that it might belong to a species of Group A Euphausia. Calyptopis stages with spinose anterior margin of carapace are found in all of the species listed above excepts, pacifica. The calyptopes of Group A species may be easily distinguished from those of E. gibboides by relative width of carapace; they do not have the anterolateral expansion over the eyes. The carapace of the two species of Group D is wide but, unlike E. gibboides, with a very high peaked dorsal crest. Also, the entire margin of the carapace of E. longirostris is spinose, the first calyptopis is not described but presumably it does not differ from calyptopes II and III in this respect. The third calyptopis of Lebour's E. sp. (1949, Figure 4, 3-4) resembles E. gibboides in width of carapace, but the lateral margins of the carapace are spinose. The carapace of the calyptopis I described by Boden (1955, Figure 12) is expanded anterolaterally, but it appears to be proportionally longer than the carapace oiE. gibboides. The relative lengths of the posterolateral spines of the telson also differ; the 3rd posterolateral spine is relatively short; as figured it is no longer than the terminal spines. The second and third calyptopes of this species have relatively narrow carapaces. The most useful character for the identifica- tion of furcilia stages oiE. gibboides is the rela- tively large 3-lobed eye; wddth of rostral plate and form of pleopods and telson may be helpful as well in differentiating furcilia with spinose anterior margin of carapace. Furcilia ofE. gib- boides may be separated from those of Group A Euphausia as follows: Furcilia with 1 pair of non-setose pleopods — the rostral plate appears to be of greater vddth \nE. gibboides; Furcilia with both setose and non-setose pleo- pods — in Group AEuphausia there is usually only one form and it has 1 setose plus 4 non- setose pairs of pleopods on abdominal seg- ments 1-5 respectively (Sheard (1953) reports numerous variants in the furciliar develop- ment of a species identified as E. recurua), E. gibboides has two forms, 1 setose plus 3 non-setose and 4 setose plus 1 non-setose pair of pleopods; Furcilia with 5 pairs of setose pleopods — the inner margin of the 3rd (inner) posterolateral spine of the telson is smooth except for tiny distal spinules in larvae of £■. gibboides and spinose in larvae of Group A. A single character is sufficient to separate furcilia of E. gibboides from those of E. longi- rostris and E. spinifera; both Group D species have a dorsal spine on segment 3 of the abdomen beginning in furcilia I. The furcilia with 1 pair of non-setose pleopods figured asE. sp. by Lebour (1949, Figure 4, 5-6) differs from E. gibboides in relative length of posterolateral spines 2 and 3 of the telson; as drawn they are almost equal in length. The telson of the second furcilia which, like E. gibboides, has 1 setose and 3 non-setose pairs of pleopods is not figured, and details of the two forms are not described. The first furcilia figured by Boden (1955, Figure 15) also differs from E. gibboides in length of posterolateral spines of the telson; the 2nd pair are almost the same length as the 3rd pair and only a little longer than the 1st pair. The second furcilia of the species has 1 setose and 4 non-setose pairs of pleopods as found in species of Group A Euphausia. ACKNOWLEDGMENTS I am grateful to E. Brinton for encourage- ment and assistance as well as review of the manuscript. The work was supported by the Marine Life Research Program, the Scripps Institution of Oceanography's component of the California Cooperative Oceanic Fisheries In- vestigations, a project sponsored by the Marine 167 FISHERY BULLETIN: VOL. 73, NO. 1 Research Committee of the State of California, and by the Oceanography Section, National Science Foundation, NSF Grant GA-31783. LITERATURE CITED Ahlstrom, E. H. 1954. Distribution and abundance of egg and larval populations of the Pacific sardine. U.S. Fish Wild!. Serv., Fish. Bull. 56:83-140. Bary, B. M. 1956. Notes on ecology, systematics, and development of some Mysidacea and Euphausiacea (Crustacea) from New Zealand. Pac. Sci. 10:431-467. BODEN, B. P. 1950. The post-naupliar stages of the crustacean Euphausia pacifica. Trans. Am. Microsc. See. 69:373-386. 1955. Euphausiacea of the Benguela Current. First survey, R.R.S. "William Scoresby", March 1950. Dis- covery Rep. 27:337-376. BoDEN, B. p., M. W. Johnson, and E. Brinton. 1955. The Euphausiacea (Crustacea) of the North Pacific. Bull. Scripps Inst. Oceanogr., Univ. Calif. 6:287-400. Brinton, E. 1962. The distribution of Pacific euphausiids. Bull. Scripps Inst. Oceanogr., Univ. Calif. 8:51-269. 1967a. Vertical migration and avoidance capability of euphausiids in the California Current. Limnol. Oceanogr. 12:451-483. 1967b. Distributional atlas of Euphausiacea (Crustacea) in the California Current region, Part I. Calif. Coop. Oceanic Fish. Invest., Atlas 5, 275 p. 1973. Distributional atlas of Euphausiacea (Crustacea) in the California Current region, Part II. Calif. Coop. Oceanic Fish. Invest., Atlas 18, 336 p. Brinton, E., and K. Gopalakrishnan. 1973. The distribution of Indian Ocean Euphausiids. Ecol. Stud., Anal. Synth. 3:357-382. Einarsson, H. 1945. Euphausiacea. 1. North Atlantic species. Dana Rep. Carlsberg Found. 27, 185 p. Frost, W. E. 1934. The occurrence and development of Euphausia krohnii off the south-west coast of Ireland. Proc. R. Irish Acad.,Sec.B, 42:17-40. Gopalakrishnan, K. 1973. Developmental and growth studies of the euphau- siid Nematoscelis difficilis (Crustacea) based on rearing. Bull. Scripps Inst. Oceanogr., Univ. Calif. 20:1-39. GURNEY, R. 1942. Larvae of decapod Crustacea. Ray Soc. Publ. 129, Ray Society, Lond., 306 p. Hansen, H. J. 1911. The genera and species of the order Euphau- siacea, with account of remarkable variation. Bull. Inst. Oceanogr. Monaco 210:1-54. 1912. The Schizopoda. Reports on the scientific results of the expedition to the tropical Pacific, in charge of Alexander Agassiz, by the U.S. Fish Commission steamer "Albatross", from August, 1899, to March, 1900, Commander Jefferson F. Mosier, U.S.N., Com- manding. Parts XVI and XXVII. Mem. Mus. Comp. Zool. (Harvard) 35(4):175-296. John, D. D. 1936. The southern species of the genus Euphausia. Discovery Rep. 14:193-324. Lebour, M. V. 1926. On some larval euphausiids from the Mediter- ranean in the neighbourhood of Alexandria, Egypt, collected by Mr. F. S. Russell. Proc. Zool. Soc. Lond. 1926:765-776. 1950. Some euphausids from Bermuda. Proc. Zool. Soc. Lond. 119:823-837. Mauchline, J. 1965. The larval development of the euphausiid, Thysa- noessa raschii (M. Sars). Crustaceana 9:31-40. Mauchline, J., and L. R. Fisher. 1969. The biology of euphausiids. Adv. Mar. Biol. 7:1-454. Ponomareva, L. a. 1969. Investigations on some tropical euphausiid species of the Indian Ocean. Mar. Biol. (Berl.) 3:81-86. Roger, C. 1967. Note on the distribution of Euphausia eximia and E. gibboides in the equatorial Pacific. Pac. Sci. 21:429-430. RuuD, J. T. 1932. On the biology of southern Euphausiidae. Hval- radets Skr. 2, 105 p. Sars, G. O. 1885. Report on the Schizopoda collected by H.M.S. Challenger during the years 1873-76. Rep. Sci. Res. Voyage H.M.S. Challenger 13(37), 228 p. Scripps Institution of Oceanography. 1964a. Physical and chemical data. CCOFI Cruise 6304, 9 April-24 May 1963, CCOFI Cruise 6306, 25-26 June 1963, and USCG Station November, 12 May-2 June 1963. SIO (Scripps Inst. Oceanogr., Univ. Calif.) Ref 64-13. Data Rep., 130 p. 1964b. Physical and chemical data report. CCOFI Cruise 6307, 10 July-8 August 1963 and CCOFI Cruise 6309, 3-29 September 1963. SIO (Scripps Inst. Oceanogr., Univ. Calif.) Ref 64-18. Data Rep., 163 p. Sheard, K. 1953. Taxonomy, distribution and development of the Euphausiacea (Crustacea). B.A.N.Z. (Br. Aust. N.Z.) Ant- arct. Res. Exped. Rep., Ser. B, 8(1): 1-72. Snyder, H. G., and A. Fleminger. 1965. A catalogue of zooplankton samples in the marine invertebrate collections of Scripps Institution of Oceanog- raphy. SIO (Scripps Inst. Oceanogr., Univ. Calif.) Ref. 65-14A, 140 p. Tattersal, W. M. 1924. Crustacea. Part VIII — Euphausiacea. Br. Antarct. ("Terra Nova") Exped., 1910, Zool. 8:1-36. 168 NEW RECORDS OF ELLOBIOPSIDAE (PROTISTA (INCERTAE SEDIS)) FROM THE NORTH PACIFIC WITH A DESCRIPTION OF THALASSOMYCES ALBATROSSI N.SP., A PARASITE OF THE MYSID STILOMYSIS MAJOR Bruce L. Wing^ ABSTRACT Ten species of ellobiopsids are currently known to occur in the North Pacific Ocean — three on mysids and seven on other crustaceans. Thalassomyces boschmai parasitizes mysids of genera Acanthomysis, Neomysis, and Meterythrops from the coastal waters of Alaska, British Columbia, and Washington. Thalassomyces albatrossi n.sp. is described as a parasite of Stilomysis major from Korea. Thalassomyces fasciatus parasitizes the pelagic mysids Gnathophausia ingens and G. gracilis from Baja California and southern California. Thalassomyces marsupii parasitizes the hyperiid amphipods Parathemisto pacifica and P. libellula and the lysianassid amphipod Cypho- caris challengeri in the northeastern Pacific. Thalassomyces fagei parasitizes euphausiids of the genera Euphausia and Thysanoessa in the northeastern Pacific from the southern Chukchi Sea to southern California, and occurs off the coast of Japan in the western Pacific. Thalassomyces capillosus parasitizes the decapod shrimp Pasiphaea pacifica in the northeastern Pacific from Alaska to Oregon, while Thalassomyces californiensis parasitizes Pasiphaea emarginata from central California. An eighth species of Thalassomyces parasitizing pasiphaeid shrimp from Baja California remains undescribed. Ellobiopsis chattoni parasitizes the calanoid copepods Metridia longa and Pseudocalanus minutus in the coastal waters of southeastern Alaska. Ellobiocystis caridarum is found frequently on the mouth parts oi Pasiphaea pacifica from southeastern Alaska. An epibiont closely resembling Ellobiocystis caridarum has been found on the benthic gammarid amphipod Rhachotropis helleri from Auke Bay, Alaska. Where sufficient data are available, notes on variability, seasonal occurrence, and effects on the hosts are presented for each species of ellobiopsid. The family Ellobiopsidae (Protista {incertae sedis)) is a heterogeneous group of parasites and epibionts found on various crustaceans (mostly planktonic) and on the benthic polychaete worm Nephthys ciliata Miiller. The Ellobiopsidae have been classified at various times as protistans, colorless algae, fungi, or protozoans. The recent work of Gait and Whisler (1970) suggests includ- ing the parasitic ellobiopsids among the dino- flagellates. The parasitic ellobiopsids are multinucleate protistans with reproductive structures out- side the host and absorptive portions inside. The reproductive structures often resemble a large mold; consequently, much of the descrip- tive terminology of ellobiopsids is mycological. The reproductive parts of an ellobiopsid (Figure 1) consist of a short primary stalk passing from 'Northwest Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 155, Auke Bay, AK 99821. DISTAL CONOMERE— w 7 ___ CONOMERE BEGINNING /^fiSr TO FORM SPORES // CONOMERE ABOUT fj F TO RELEASE SPORES PROXIMAL H j 7 ../^-"^^ CONOMERE ^ / 1 x-:;^^^-^] NEWLY FORMING. ^^!\0> y^^-^^ - TROPHOMERE TROPHOMERES C^"^^ {^ \ -*\ — PRIMARY STALK \ 1 i .^ „. SIEVE PLj^TE — ^ ABSORPTIVE FILAMENTS -D ORGAN OF FIXATION Manuscript accepted April 1974. FISHERY BULLETIN: VOL. 73. NO. 1, 1975. Figure 1. — Schematic of an ellobiopsid (Thalassomyces sp.) the organ of fixation through the cuticle of the host and one or more trophomeres which branch from the primary stalk. The trophomeres in turn bear one or more gonomeres at their distal end. 169 / FISHERY BULLETIN: VOL. 73, NO. 1 The mature distal gonomeres further subdivide to produce motile biflagellate spores (Gait and Whisler 1970). The internal portion of a parasitic ellobiopsid, the organ of fixation, may be compact (like a bulb or taproot — Figure 1) or may be branching (rhizomorphous). The compact forms bear ridged sieve plates from which extend fine protoplasmic filaments. These filaments are believed to absorb nutrients from the host. The internal portions are difficult to observe without staining and sec- tioning techniques and have not been used much for taxonomic purposes. The nonparasitic epibionts of the genus Ello- biocystis Coutiere do not have an internal organ of fixation but attach directly to the host's cuticle. These epibionts superficially resemble single trophomeres of the parasitic ellobiopsids but usually have a single gonomere. They are small and attached singly or in clusters to the mouth parts of various shrimps, mysids, and amphipods. Only the morphology of the Ellobiocystis spp. has been described. The inclusion of Ellobiocystis spp. in the Ellobiopsidae is very questionable. Other than their morphology, little is known about the ellobiopsids of the genus Thalassomyces or their effects on hosts. The development of reproductive spores has been described (Gait and Whisler 1970); however, the mode of infection, time required to mature, and true incidence of infection remain subjects of speculation. Some of the parasitic ellobiopsids sterilize the hosts and probably exert some control on the molting cycle of crustacean hosts. Undoubtedly, the parasites draw heavily on the metabolic resources of the hosts, which conceivably would increase mortality and decrease reproduction in the host populations. Ellobiopsids were first recognized as a compo- nent of the northeastern Pacific fauna when Mc- Cauley (1962) recorded Thalassomyces capillosus (Page) as a parasite of the shrimp Pasiphaea pacifica Rathbun. Since then eight additional ellobiopsids have been recognized in zooplank- ton collections from the eastern North Pacific. Only two species of ellobiopsids have been re- ported from the western North Pacific. In the fol- lowing discussions for each species, I summarize observations, some new and some from the litera- ture, on the occurrence ^nd hosts of the North Pacific ellobiopsids. I list synonymies only for references to material from the North Pacific. For convenience only, I have treated those ello- biopsids found on mysids first and those found on amphipods, euphausiids, shrimp, and copepods second. ARTIFICIAL KEY TO ELLOBIOPSIDS FOUND ON MYSIDS The published keys to the ellobiopsids (Kane 1964; Collard 1966) do not give complete coverage to the ellobiopsids found on mysids. Kane's key is restricted to genera of ellobiopsids, and Collard's key covers only 9 of the 11 known species of Thalassomyces. Identification of the known species of Ellobiocystis, and Ellobiopsis and most Thalassomyces is possible by reference to the summaries by Boschma (1949, 1957, 1959). The following key supplements Collard's but treats only the ellobiopsids found on mysids. Two of the species, T. nouveli (Hoenigman 1954) and T. niezabitowskii (Hoenigman 1960), are known only from the Mediterranean Sea; and Ellobio- cystis caridarum (Coutiere) while not known from North Pacific mysids is found on Antarctic mysids (Boschma 1949, 1959) and has been found on the North Pacific decapod shrimp Pasiphaea pacifica. 1. No root system of attachment to host; attached to oral appendages. Mature parasite consists of single trophomere with one or more gonomeres Ellobiocystis caridarum. (Coutiere) Root system of attachment to host; not attached to oral appendages. Mature parasite with many trophomeres branching from stalk(s) . .Thalassomyces (2) 2. Parasite attached to ventral surface of first abdominal segment. Long pendu- lous umbellate trophomeres; usually only one gonomere per trophomere (length of ma.ture gonomere 1.5 to over 2 times the diameter) . . .T. fasciatus (Fage) Site of attachment usually dorsal thoracic, but variable. Trophomeres not pendu- lous; usually more than one gonomere per trophomere (2-4) (3) 3. Mature gonomeres flattened spheres. Mean length-diameter ratio less than 1 T. nouveli (Hoenigman) Mature gonomeres globular to oval. Mean length-diameter ratio greater than 1 . . . .(4) 170 WING: ELLOBIOPSIDAE FROM NORTH PACIFIC 4. Primary stalks widely spaced when more than one per host. Mature gonomere length-diameter ratio 0.7-1.9 (mean about 1.2) (5) Primary stalks closely spaced when more than one per host. Mature gonomere length-diameter ratio 1.5-2.4 (mean about 2) T. albatrossi n.sp. 5. Mature terminal gonomere shape ellip- soid, with the distal end the same size as the proximal end, to spherical T. boschmai (Nouvel) Mature terminal gonomere shape ovoid, with the distal end smaller than the proximal end, to spherical T. niezabitowskii (Hoenigman) ELLOBIOPSIDS OF NORTH PACIFIC MYSIDS Thalassomyces boschmai (Nouvel 1954) Thalassomyces sp. — Wing (1965), Hoffman and Yancey (1966), Thorne (1968). Thalassomyces boschmai — Gait and Whisler (1970), Vader (1973b). Ellobiopsids of the genus Thalassomyces have been observed on Mysidae from Alaska (Wing 1965; Hoffman and Yancey 1966); from Puget Sound (Thorne 1968); and from southern British Columbia (J. Gait, Friday Harbor Laboratory, University of Washington, Friday Harbor, WA 98250, pers. commun.). New collections of Alaska mysids (Table 1) plus supplementary material from Puget Sound enabled me to identify these ellobiopsids as T. boschmai. Characteristics of T. boschmai The identification of Thalasso?nyces spp. para- sitizing mysids is based on external portions so variable that for definitive identifications, several characters must be examined. The external characters used to identify a species are the total size or height of the parasite, length of tropho- meres, number of trophomeres per primary stalk, number of gonomeres per trophomere, and size and shape of gonomeres. The number of primary stalks and the site of attachment are also useful characteristics. Differences between specimens from different localities may be associated with Table 1. — Records of Thalassomyces boschmai found on mysids in Alaska, 1963-67. Area, collection number, and species of host Number of mysids with T. boschmai Number of T. boschmai Little Port Walter' AB66-243 Acanthomysis pseudomacropsis Neomysis kadiakensis AB66-244 4 70+ 4 80+ Acanthomysis pseudomacropsis Neomysis l 1 . 1 Figure 1. — Numbers and statistics of vertebrae and scales of the 10 species of western Atlantic Synodontidae. Vertebrae: range — horizontal line; mean — vertical line; standard deviation, one on each side of the mean — open rectangle; standard error — two on each side of the mean — shaded rectangle. Scales: range — cross- hatched bar; mean — small triangle. Synodus saurus from the eastern Atlantic (5 specimens) had a range in vertebrae of 55-59, greater than the western Atlantic (14 specimens) vertebrae range of 56-58. Trachinocephalus myops had varying but similar vertebral ranges in small samples encompassing its extensive geographic range. Area n Range U.S.— Brazil 11 55.4 54-57 Nigeria 1 55.0 55 Philippines 7 53.1 52-54 Hawaii 5 54.8 54-55 Abnormalities in vertebral structure (speci- mens not included in the tables or figure) occurred in 11 of 317 western Atlantic speci- mens examined. In six, pairs of vertebrae were shortened with irregular and expanded ossifica- tions at their adjoining ends. In five, a single centrum in the caudal region was elongated and had two neural spines (in two), two hemal spines (in two), or double neural and hemal spines (in one). Scales Pored lateral-line scales in our samples of the 10 western Atlantic species range from 43 to 63 (Table 1). We have not confirmed any higher or lower values for these or other species of the family. Ranges in scale complements that we have confirmed for specimens from the west- ern Atlantic, with clarification where these ranges differ from those given by Anderson et al. (1966a), are: Synodus foetens 57-64; the range of 56-65 given by Anderson et al. was in error, as determined by our reexamination of the material originally reported. Synodus saurus 56-60; the range of 55-62 given by Anderson et al. included a low count for an eastern Atlantic specimen and a published but un- substantiated high count. Synodus synodus 54-59. Synodus intermedius 47-51; the range of 45-52 given by Anderson et al. was in error, as determined by our reexamination of the material originally reported. Synodus poeyi 43-48. Trachinocephalus myops 53-59; the 204 ANDERSON, GEHRINGER, and BERRY: NUMBERS OF VERTEBRAE AND LATERAL-LINE SCALES range of 51-61 reported by Anderson et al. was based on previously published records from other geographic areas. Saurida brasiliensis 43-49; the range of 40-50 reported in Anderson et al. was based on a low count previously published and currently unconfirmable and on a high count that we have since confirmed in a specimen from the eastern Atlantic. Saurida normani 51-56. Saurida suspicio 52-54; a high count of 56 previously published has not been confirmed by us. Saurida caribbaea 51-60; examination of additional specimens has enlarged the range of 54-60 given by Anderson et al. Many of the specimens used in the confirma- tions above are not included in Table 1, because corresponding vertebral counts were not made. Bilateral symmetry in scale numbers char- acterized one-half to three-quarters of the speci- mens of each species. In the total sample, 62fFc were bilaterally symmetrical. Asymmetry appears to be random, 20% having more scales on the left side and 18% having more scales on the right side. Asymmetry was of only one scale difference in all species, except in our largest species sample. In Sy. foetens, which also has the greatest number of scales, of 118 specimens 3 had two more scales on one side than the other, 52 had one more scale on one side than the other, and 63 were bilaterally symmetrical. Correlations Frequency distributions of numbers of verte- brae and associated numbers of pored lateral- line scales are shown for the two species for which we examined the largest number of speci- mens, Sy. foetens (Table 2) and Sy. intermedius (Table 3). The trend of positive correlation is apparent from visual inspection of both tables. The coefficients of correlation (Table 1) docu- ment the positive nature of the correlation, Sy. foetens (r = 0.86) and Sy. intermedius (r = 0.76) (Table 1). The same kinds of data for the other eight species are given below, with number of verte- brae separated by a hyphen from the number of scales and followed in parentheses by the frequency for that combination: Synodus saurus, vertebrae 56-58 scales (2), 57-58(1), 57-59(4), 57-60(1), 58-58(7), 58-59(8), 58-60(5). Synodus synodus, 55-55(3), 55-56(7), Table 2. ^Frequency distribiutions of numbers of vertebrae and pored lateral-line scales in 118 Synodus foetens . Vertebrae Scales 56 57 58 59 60 61 62 63 — — 10 25 2 62 — — — 4 57 15 2 61 — 4 15 21 4 — 60 — 1 19 14 2 — — 59 4 15 5 2 2 — — 58 4 7 — 1 — — — 57 — 1 ~ ~ ~ Table 3. — Frequency distributions of numbers of vertebrae and pored lateral-line scales in 85 Synodus intermedius . Vert sbrae Scales 47 48 49 50 51 1 9 2 50 — 5 37 2 49 2 44 16 — 48 26 21 — — 47 2 3 — — 56-55(1), 56-56(1), 56-57(2), 57-55(1), 57-56(2), 57-57(2), 57-58(1). Synodus poeyi, 44-43(4), 44-44(2), 44-45(3), 44-46(1), 45-43(1), 45-44(2), 45-45(1), 45-46(2), 46-44(1), 46-45(2), 46-46(2), 46-47(9), 47-48(4), 48-48(4). Trachinocephalus myops, 54-56(3), 54-57(1), 55-54(1), 55-55(2), 55-56(3), 55-57(2), 56-56(1), 56-57(5), 56-58(2), 57-56(1), 57-57(1). Saurida brasiliensis, 46- 47(4), 46-48(6), 47-49(6), 48-48(3), 48-49(1). Saurida normani, 49-52(3), 49-53(1), 50-53(2), 51-53(1), 51-54(3), 51-55(4), 52-54(2), 52-55(5), 52-56(1). Saurida suspicio, 49-52(1), 49-53(1), 51-52(7), 51-53(3), 52-53(7), 52-54(3). Saurida car/66aea, 48-51(2), 49-51(3), 50-52(2), 52-53(3), 52-54(1), 54-55(1), 54-56(1), 54-58(2), 55-57(2), 55-58(3), 55-59(5), 56-57(3), 56-58(4), 56-59(1), 57-59(2), 58-60(2). The correlation coefficients of the samples for all species are positive, ranging from 0.96 for Sa. caribbaea to 0.20 for Sy. saurus (Table 1). The species with the larger number of specimens (19 to 118) generally had the higher correlation coefficients {r 0.76 to 0.96). Of the species with a lesser number of specimens (11 to 14), one had a high positive value (0.84), and the others were low (0.20 to 0.52). We suspect that the relatively low value of positive correlation for five of the species is due to the small and somewhat hetero- geneous samples used for these species. Statistics describing the samples of vertebrae and scales for each species (from Table 1) are illustrated in Figure 1. The nature of positive 205 FISHERY BULLETIN: VOL. 73, NO. 1 correlation of vertebrae and scales for the 10 species is apparent in this figure. The ratio of scales to vertebrae is nearly 1:1 for the 10 species, but in each species the total number of scales averages slightly more than the total number of vertebrae (50% or more of the scale counts in any species are greater than the vertebral counts). In species of Saurida the number of scales averages from one to three more than the number of vertebrae and ranges from an equal number of each to four more scales than vertebrae. In species of Synodus and in Trachinocephalus the number of scales averages one or two more than the number of vertebrae and ranges from two fewer to three more scales than vertebrae. Of 118 Sy. foetens 10% had three more scales than vertebrae, 57% had two more scales, 27% had one more, 5% had an equal number, and 1% had one less scale than vertebrae. The positional relationship of scales to verte- brae in lateral aspect was investigated. In a Sy. foetens with 62 pored lateral-line scales on each side and 60 vertebrae, pins were inserted at the posterior margins of certain numbered lateral-line scales on the left side, and the specimen was X-rayed. The first scale was lateral to the junction of the 4th and 5th centra, the 30th scale was lateral to the junction of the 32nd and 33rd centra, the 60th scale was lateral to the last centrum, and the last scale was lateral to the posterior ends of the hypural bones and overlapping anterior ends of the median caudal-fin rays. Similarly, in a Sy. intermedius with 49 scales on each side and 48 vertebrae the first scale was lateral to the 4th centrum, the 47th scale was lateral to the 48th centrum, and the last scale was lateral to the posterior ends of the hypural bones. DATA ON TYPE SPECIMENS AT USNM Counts of vertebrae and pored lateral-line scales on type specimens of 12 nominal species of Synodontidae in the U.S. National Museum of Natural History are recorded here for use in future studies. The four data items following the collection number for each type specimen are. in sequence, number of vertebrae, number of left-side scales, number of right-side scales, and standard length in millimeters: Synodus binotatus Schultz, holotype USNM 140801, 53-54-ca. 54-86.5. Synodus cinereus Hildebrand, holotype USNM 53079, 57-58- 59-112. Synodus englemani Schultz, holotype USNM 140815, 59-60-60-104. Synodus ever- manni Jordan and Bollman, one of 11 syntypes USNM 41144, 47-48-48-142. Synodus jenkinsi Jordan and Bollman, holotype USNM 41171, 60-ca. 60-61-282. Synodus lacertinus Gilbert, holotype USNM 44300, 61-63-62-129. Synodus marchenae Hildebrand, holotype USNM 120111, QQ-Q2-Q2-b0.b. Synodus sechuraemide- brand, holotype USNM 127829, 57-58-58-130. Synodus simulans Garman, paratype USNM 153607, 60-ca. 62-ca. 61-ca. 45. Synodus ulae Schultz, holotype USNM 52671, 64-ca. 63-ca. 63-177. Saurida eso Jordan and Herre, holotype USNM 57847, 59-62-61-290. Saurida normani Longley, holotype USNM 107330, 52-52-53-320. In these type specimens the ratio of nearly 1:1 for number of vertebrae and scales suggests a positive correlation of these two variates in the species that they represent. ACKNOWLEDGMENTS We are grateful to Ernest A. Lachner, James E. Bohlke, and Carter R. Gilbert for providing specimens for study and to the late James A. Peters for assistance with the time-share com- puter at the Smithsonian Institution. LITERATURE CITED Anderson, W. W., J. W. Gehringer, and F. H. Berry. 1966a. Family Synodontidae. Lizardfishes. In Fishes of the western North Atlantic. Part Five, p. 30-102. Mem. Sears Found. Mar. Res., Yale Univ. 1. 1966b. Field guide to the Synodontidae (lizardfishes) of the western Atlantic Ocean. U.S. Fish Wildl. Serv., Circ. 245, 12 p. GiBBS, R. H., Jr. 1959. A synopsis of the postlarvae of Western Atlantic lizard-fishes (Synodontidae). Copeia 1959:232-236. Norman, J. R. 1935. A revision of the lizard-fishes of the genera Synodus, Trachinocephalus, and Saurida. Proc. Zool. Soc. Lond. 1935:99-135. 206 ACUTE TOXICITY OF AMMONIA TO SEVERAL DEVELOPMENTAL STAGES OF RAINBOW TROUT, SALMO GAIRDNERI Stanley D. Rice' and Robert M. Stokes^ ABSTRACT Median tolerance limits derived from 24-h bioassays demonstrated that fertilized eggs and alevins of rainbow trout, Salmo gairdneri, were not vulnerable to 3.58 ppm un-ionized ammonia at 10°C (pH 8.3). At the end of yolk absorption, rainbow trout fry increased in susceptibility dramatically; their median tolerance limit values were about 0.072 ppm, the same as for adult trout. Fertilization of eggs was not prevented in un-ionized ammonia solutions up to 1.79 ppm, the highest exposure tested. Much information is available on the toxicity of ammonia to juvenile and adult trout, but the paucity of information on the toxicity of ammonia to fertilized eggs and larvae of teleosts is sur- prising since these life stages are often assumed to be relatively sensitive. Several studies have examined ammonia toxicity to adult trout (Lloyd 1961; Ball 1967; Wilson et al. 1969), including the effects of increased ammonia toxicity to trout at lower oxygen levels (Downing and Merkens 1955) and decreased toxicity at higher carbon dioxide levels (Lloyd and Herbert 1960). Tem- perature, oxygen, pH, carbon dioxide, and bicar- bonate alkalinity influence the toxicity of am- monia and are discussed in a report by the European Inland Fisheries Advisory Commission (1970). Exposure of juvenile or adult salmonids to ammonia has been associated with decreased growth (Brockway 1950; Burrows 1964; Larmo- yeux and Piper 1973), gill damage (Burrows 1964; Reichenbach-Klinke 1967), and other sublethal physiological effects (Reichenbach-Klinke 1967; Fromm and Gillette 1968; Lloyd and Orr 1969), and similar effects may occur with salmonid eggs and alevins. Exposure to ammonia has also been associated with increased incidence of disease in juvenile and adult salmonids (Burrows 1964; Larmoyeux and Piper 1973) and in salmonid alevins (Wolf 1957). The only study of toxicity of ammonia to eggs and larvae (Penaz 1965) involved three stages 'Department of Biological Science, Kent State University, Kent, OH 44242; present address: Northwest Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 155, Auke Bay, AK 99821. ^Department of Biological Science, Kent State University, Kent, OH 44242. Manuscript accepted May 1974. FISHERY BULLETIN: VOL. 73, NO. 1, 1975. of eggs and two stages of yolk fry of Salmo trutta. Penaz observed an increase in sensitivity of the eggs with age to brief (120 min) exposures to ammonia at pH 8 and temperatures of 5.68° to 3.56°C. A similar pattern was observed with longer exposures (10 h) of newly hatched and 12-day-old alevins to ammonia at pH 8 and temperatures of 11° and 16.9°C. The early eggs were resistant to the highest dose he tested — 50 mg/liter of un-ionized ammonia. These data suggest changes in sensitivity with development, but the changes in lengths of exposure and temperature make it difficult to compare dif- ferences between eggs and alevins. We used a series of bioassays to determine the stage of development at which eggs and larvae of rainbow trout, Salmo gairdneri, were most susceptible to acute ammonia toxicity. Such information is needed to establish realistic limits for survival of eggs and larvae in both natural and hatchery environments. Knowledge of con- centrations of ammonia that may limit survival is particularly important in hatchery operations where it is advantageous to maintain the greatest density of fish and eggs per unit water flow. MATERIALS AND METHODS Freshly fertilized rainbow trout eggs were obtained from Bowden National Fish Hatchery, W.V., (courtesy of the U.S. Bureau of Sport Fisheries and Wildlife) and transported to the laboratory within 6 h. About 2,000 of the eggs were poured under water into 4-inch-square trays with nylon screen bottoms at about 25 to 35 eggs per tray. For incubation the trays were put into a 10°C water 207 FISHERY BULLETIN: VOL. 73, NO. 1 bath that recycled through both charcoal and a gravel bacterial filter at a rate of 3 gallons/min. Ammonia levels were measured periodically and never attained 0.1 ppm. All ammonia analyses w^ere made by separating ammonia by diffusion (Conway and Cooke 1939) and followed by nesslerization of the separated ammonia. For the ammonia bioassays, the small trays were transferred directly to the experimental medium. By conducting the bioassays in the same trays in which the eggs or larvae were incubated, we did not have to pipette them to other con- tainers— a process that might have injured them. Two series of duplicated ammonia toxicity bio- assays were conducted according to standard pro- cedures outlined by DoudorofF et al. (1951) and results were expressed as 24-h median tolerance limits (24-h TLm)^. The bioassays were conducted every 4 to 7 days from fertilization to the com- pletion of yolk sac absorption. Toxicity of am- monia to adult rainbow trout (length 7-9 inches) was also measured with static bioassays (12 fish per concentration tested, 1 fish per 10-liter aquar- ium) at the same water temperature and pH used with the eggs and larvae. All bioassays were conducted in aged tap water (total hardness 5.94 ppm as calcium carbonate at pH 7.8) adjusted to pH 8.3 with tris buffer (final concentration 0.05 M). Ammonia, in the form of ammonia sulfate, was added to arrive at the various test concentrations. The resulting conditions made the ammonia toxicity assays more severe than would normally be encoun- tered because the toxicity of ammonia increases as pH increases due to the conversion of ionized NH^ into the un-ionized NHg form. At 10°C and pH 8.3, 3.58% of the ammonia in water is un- ionized, considerably more than the 0.19% un- ionized ammonia at pH 7 (Trussell 1972). Since the un-ionized form of ammonia has been identified as the toxic form, we report our results in units of un-ionized ammonia rather than total ammonia. Several water quality parameters were mea- sured at the beginning and end of the bioassays, since changes could affect the results. Ammonia levels never dropped below 93% of the initial bioassay concentrations during the course of the 24-h experiments. Very low levels of un-ionized ammonia (0.011 ppm) were detected in the con- trol exposures after 24 h. The tris buffer prevented any changes in pH from occurring during the 24-h tests. Dissolved oxygen remained above 91% saturation in the shallow egg-alevein bio- assay containers and above 88% saturation in the adult bioassays (measured with YSI oxygen probe).'* Carbon dioxide was not measured in any of the bioassays. We tested the influence of ammonia on egg fertilization and viability during the water- hardening stage by exposing some eggs to am- monia at Bowden Hatchery on the day our experi- mental eggs were collected. Approximately 200 to 300 eggs from one female were stripped into each of several pans containing tris buffered water (pH 8.3, temperature 8°-10°C), some with added ammonia at concentrations up to 1.79 mg/liter of un-ionized ammonia. Milt from at least two young males was stripped into each pan of water and eggs 15 to 30 s later. Buss and Corl (1966) determined that fertilization must be completed within the first 1 or 2 min because the sperm are viable in water for only a few seconds. By replacing the ammonia solutions with fresh water in one-half of the pans after 2 or 3 min of ammonia exposure and in the remaining pans after 1 h, we hoped to separate the effects of ammonia on fertilization per se from the effects on the viability of the fertilized eggs during the water-hardening stage of the first hour. The effects were measured by determining the per- centage of eggs that hatched. RESULTS AND DISCUSSION Neither fertilized eggs, embryos, nor alevins (embryo after hatching) were susceptible to a 24-h exposure of un-ionized ammonia (3.58 mg/liter) until about the 50th day of development (Figure 1). At that time, susceptibility increased dramati- cally and continued to increase until most of the yolk was absorbed (when alevins became fry). The median tolerance limits (24-h TLm) for 85- day-old fry were 0.068 mg/liter, slightly less than the 0.097 mg/liter value we observed for adult trout; in the bioassays for both the fry and the adults, temperature was 10°C and pH was 8.3. Buss and Corl (1966) found that the viability of eggs of brook trout, Salvelinus fontinalis, and '24-h TLm = the concentration resulting in 50% survival after 24-h exposures. ■•YSI = Yellow Springs l.astrument Company, Inc., Yellow Springs, Ohio. Reference to trade name does not imply endorse- ment by the National Marine Fisheries Service, NOAA. 208 RICE and STOKES: TOXICITY OF AMMONIA TO RAINBOW TROUT 1.1 S 1.0 _ ALL TLm VALUES \ ^ LU -TRrATrr? tiiam \ ^ K _l 3.58 MC, LITER FROM \ 5 °-^ " 0 TO 52 DAYS \ ° 5 \ Q \ z < 0.8 ■ Q \ " z O \ io.7 - a. LU \ S Q- \ < O \ Q 0.6 - z \ LU M I \ u \ g 0 5 - 1- < \ I \ z \ => o.^ - \ O ^ ii:o.3 _ \^ I \ 3- \ £-0.2 - \ -I ^s^ 1- V. 0.1 _ ^^^4--^ 1 n ' 1 ] 1 1 ' 1 1 1 1 10 20 ■ECC- 30 ao 50 — ALEVINS- ACE IN DAYS 60 fry- Figure 1. — Twenty-four-hour median tolerance limits (TLm) of un-ionized ammonia to eggs and alevins of rainbow trout ( 10°C, pH 8.3). Points indicate mean of two bioassays; bars indicate the range. Adult trout 24-h TLm was 0.097 mg/liter (10°C, pH 8.3). brown trout, Salmo trutta, drops significantly after 15 s in water — in our experiments this was about the minimum time lapse between stripping eggs into the water and introduction of sperm. Because we did not control the time lapse before sperm introduction precisely enough, we cannot evaluate any subtle effects of ammonia on prevention of fertilization. It was obvious, however, that high ammonia concentrations did not cause complete loss of eggs or sperm (Table 1) because more than half of the eggs were ferti- lized at all ammonia exposures. No obvious dif- ferences in the percentages of eggs that hatched were noticed between ammonia exposures of 2 or 3 min and 1 h, even at the highest concen- trations of un-ionized ammonia (1.79 mg/liter) we tested. The fertilization and water-hardening stages are similar to later stages (before 50 days of development) in their relative insensitivity to ammonia when compared with older fry with absorbed yolks (after about 60 days of develop- ment). Our observations of great resistance of eggs and alevins of rainbow trout to ammonia toxicity are consistent with results of ammonia toxicity studies of Penaz (1965) and with other studies of other toxicants. Trout eggs and sac fry were only slightly susceptible to endrin at concentra- tions that seriously affected adults (Wenger 1973). Burdick et al. (1964) observed that a high pro- portion of lake trout, Salvelinus namaycush, fry from normal appearing eggs containing 2.95 ppm DDT or more died. The sensitive fry died at the completion of yolk absorption when feeding would normally begin. Eggs of "common trout" were less susceptible to anionic detergent toxicity (sodium alkylsulphate) than alevins, whose sen- sitivity continued to increase for 6 wk (Wurtz- Arlet 1959). Eggs of two salmonids were about one-tenth as sensitive to a commercial formula- tion of rotenone and derivatives as fry at the same temperature (Garrison 1968). A study of zinc toxicity by Skidmore (1965) showed that eggs of zebrafish, Brachydanio rerio, were relatively less susceptible than newly hatched fish. It appears then that eggs and developing em- bryos are resistant to several toxicants, including ammonia. One obvious explanation for the resis- Table 1. — Effect of ammonia on fertilization Concentration of un-ionized ammonia Percentage' 2-3 min hatch at ex Dosure of 1 h 0 mg/liter 0.0358 mg/liter 1.79 mg/liter 66.8 74.3 58.2 68.4 70.1 68.8 'Percentage hatch of each group of 250 eggs. 209 FISHERY BULLETIN: VOL. 73. NO 1 tance may be the protection afforded the embryo by the surrounding egg membranes which sep- arate the internal from the external environ- ment. However, in a second study of zinc toxicity to zebrafish embryos, Skidmore (1966) found no evidence of protection of the embryo by the egg membranes. He found that embryos with ruptured outer membranes actually survived longer in a zinc sulphate solution than embryos of the same age with an intact membrane. If the outer egg membrane impermeability were a major factor in preventing ammonia toxicity, all alevins would be instantly vulnerable at hatching. No sudden susceptibility to toxicants in newly hatched fish was observed in this study or in several others. We can see no satisfactory explanation for the observed high resistance to ammonia and other toxicants during early developmental stages of teleosts. The higher resistance of sac fry than eggs to toxicants indicates that the egg mem- branes are not always protective barriers and that the explanation is more complex. In our study, the susceptibility to ammonia developed during the transition from alevin to fry, toward the end of yolk absorption. This transi- tion, although gradual, is probably more of a physiological change than the changes that occur at hatching. The newly hatched alevins are more "embryo" than "juvenile." They normally reside in the incubation gravels, have few voluntary responses to changes in their environment, and continue to develop by catabolizing their yolk. As the alevin develops and becomes prepared for emergence, susceptibility to some toxicants in- creases. The alevins are now more juvenile than embryo, even to the point of preemergent feeding as concluded by Dill (1967) for sockeye salmon alevins. Our results indicate that rainbow trout embryos and alevins are safer from ammonia toxicity than are older salmonids (Burrows 1964; Larmoyeux and Piper 1973). A dramatic increase in the excretion of ammonia (Rice and Stokes in press) and sensitivity to ammonia appears to begin about the time the fry complete absorption of their yolk. Chronic exposure to ammonia would prob- ably exert its greatest effects beginning at this stage also. ACKNOWLEDGMENTS We appreciate the aid of the staff of the Auke Bay Fisheries Laboratory in the preparation of this paper and the Bureau of Sport Fisheries and Wildlife Bowden National Fish Hatchery for providing the trout eggs. LITERATURE CITED Ball, I. R. 1967. The relative susceptibilities of some species of fresh-water fish to poisons — I. Ammonia. Water Res. 1:767-775. Brockway, D. R. 1950. Metabolic products and their effects. Prog. Fish- Cult. 12:127-129. BuRDiCK, G. E., E. J. Harris, H. J. Dean, T. M. Walker, J. Skea, and D. Colby. 1964. The accumulation of DDT in lake trout and the effect on reproduction. Trans. Am. Fish. Soc. 93:127-136. Burrows, R. E. 1964. Effects of accumulated excretory products on hatchery-reared salmonids. U.S. Bur. Sport Fish. Wildl., Res. Rep. 66, 12 p. Buss, K., AND K. G. Corl. 1966. The viability of trout germ cells immersed in water. Prog. Fish-Cult. 28:152-153. Conway, E. J., and R. Cooke. 1939. Blood ammonia. Biochem. J. 33:457-478. Dill, L. M. 1967. Studies on the early feeding of sockeye salmon alevins. Can. Fish Cult. 39:23-34. DouDOROFF, p., B. G. Anderson, G. E. Burdick, P. S. Galtsoff, W. B. Hart, R. Patrick, E. R. Strong, E. W. SURBER, AND W. M. VaN HORN. 1951. Bio-assay methods for the evaluation of acute toxicity of industrial wastes to fish. Sewage Ind. Wastes 23:1380-1397. Downing, K. M., and J. C. Merkens. 1955. The influence of dissolved-oxygen concentration on the toxicity of un-ionized ammonia to rainbow trout {Salmo gairdnerii Richardson). Ann. Appl. Biol. 43: 243-246. European Inland Fisheries Advisory Commission. 1970. Water quality criteria for European freshwater fish. Report on ammonia and inland fisheries. FAO (Food Agric. Organ. U.N.), EIFAC (Eur. Inland Fish. Advis. Comm.) Tech. Pap. 11, 12 p. Fromm, p. O., and J. R. Gillette. 1968. Effect of ambient ammonia on blood ammonia and nitrogen excretion of rainbow trout {Salmo gaird- neri). Comp. Biochem. Physiol. 26:887-896. Garrison, R. L. 1968. The toxicity of Pro-Noxfish to salmonid eggs and fry. Prog. Fish-Cult. 30:35-38. Larmoyeux, J. D., and R. G. Piper. 1973. Effects of water reuse on rainbow trout in hatch- eries. Prog. Fish-Cult. 35:2-8. Lloyd, R. 1961. The toxicity of ammonia to rainbow trout (Salmo gairdnerii Richardson). Water Waste Treat. J. 8:278-279. Lloyd, R., and D. W. M. Herbert. 1960. The influence of carbon dioxide on the toxicity of un-ionized ammonia to rainbow trout (Salmo gaird- nerii Richardson). Ann. Appl. Biol. 48:399-404. 210 Lloyd, R., and L. D. Ore. 1969. The diuretic response by rainbow trout to sub- lethal concentrations of ammonia. Water Res. 3:335-344. Penaz, M. 1965. Influence of ammonia on eggs and spawns of stream trout, Salmo trutta M. Fario. Zool. Listy, Folia Zool. 14:47-53. [Translated by and available from Foreign Fisheries (Translations), U.S. Dep. Commer., Wash., D.C.] Reichenbach-Klintke, H. H. 1967. Untersuchungen liber die einwirkung des am- moniakgehalts auf den fischorganismus (Research concerning the effect of ammonia content on the fish organism) Arch. Fischereiwiss. 17:122-132. [Translated by Agence Tunisienne de Puglic-Relations, Tunis, Tunisia; available U.S. Dep. Commer., Natl. Mar. Fish. Serv., Wash., D.C., as TT71-55453.] Rice, S. D., and R. M. Stokes. In press. Metabolism of nitrogenous wastes in the eggs and alevins of rainbow trout, Salmo gairdneri. In Proc. International Symposium of the Early Life History of Fish. Oban, Argyll, Scotl. Skidmore, J. F. 1965. Resistance to zinc sulphate of the zebrafish (Brachydanio rerio Hamilton-Buchanan) at different phases of its life history. Ann. Appl. Biol. 56:47-53. 1966. Resistance to zinc sulphate of zebrafish (Brachy- danio rerio) embryos after removal or rupture of the outer egg membrane. J. Fish. Res. Board Can. 23: 1037-1041. Trussell, R. p. 1972. The percent un-ionized ammonia in aqueous ammonia solutions at different pH levels and tempera- tures. J. Fish. Res. Board Can. 29:1505-1507. Wenger, D. p. 1973. The effects of endrin on the developmental stages of the rainbow trout, Salmo gairdneri. M.S. Thesis, Kent State Univ., Kent, Ohio. Wilson, R. P., R. O. Anderson, and R. A. Bloomfield. 1969. Ammonia toxicity in selected fishes. Comp. Biochem. Physiol. 28:107-118. Wolf, K. 1957. Experimental induction of blue-sac disease. Trans. Am. Fish. Soc. 86:61-70. Wur'k-Arlet, J. 1959. Toxicite des detergents anioniques vis-a-vis des alevins de Truite commune (The toxicity of anionic detergents towards the alevins of common trout). Bull. Fr. Piscic. 32:41-45. [Saw abstr. only.] 211 NOTES ADDITIONAL EVIDENCE SUBSTANTIATING EXISTENCE OF NORTHERN SUBPOPULATION OF NORTHERN ANCHOVY, ENGRAULIS MORDAX The northern anchovy, Engraulis mordax (Girard), ranges from Queen Charlotte Islands, British Columbia, to Cape San Lucas, lower Baja California. A study of variations in meristic characters (McHugh 1951) and genetic studies using serum transferrins (Vrooman and Smith 1971) generally support the hypothesis that three distinct subpopulations exist within this species' total geographic range. The dividing lines between subpopulations apparently occur at Point Conception, Calif, (delineating the north- ern and central elements), and at Cedros Island, central Baja California (delineating the central and southern elements). Extensive spawning activity by the central and southern subpopulations is evidenced from the results of comprehensive egg and larvae surveys conducted since 1951 by the California Cooperative Oceanic Fisheries Investigations (Baxter 1967). Although these surveys suggest that the time-space distributions of spawning effort by these two subpopulations tend to overlap, evidently each achieves enough reproductive iso- lation to generate genetic differences between serum transferrins. Apparently, then, the central and southern subpopulations are capable of independently producing their own recruitment. Until recently, the evidence for independent spawning by the northern subpopulation was not extensive. Ahlstrom ( 1968) noted that, in 1949 and 1950, anchovy larvae were found in moderate abundance off the Oregon coast. LeBrasseur (1970) indicated that a small number of larvae were taken in a 1958 survey of Queen Charlotte Sound, British Columbia. Waldron^ stated that no eggs or larvae were taken in incidental samples off the Washington-Oregon coast in 1966 but that a few anchovy larvae were obtained during a comprehensive survey in the spring of 1967. Such meager results might lead one to believe that the few larvae observed in northern waters were merely the result of incidental spawning activity. A conclusion might then be made that the northern subpopulation does not indepen- dently produce its own recruitment but relies instead upon an influx of anchovies from the two southern subpopulations. In 1969, however, Richardson (1973) encoun- tered such extensive numbers of anchovy larvae during a May-October survey of larval fishes off the Oregon coast (lat. 42°00'-46°30'N, coast- line— long.l29°30'W) that the above conclusions seemed to be refuted. Her results indicated the presence of a spawning stock of anchovies as- sociated with the warm, near-surface waters of the Columbia River plume. Moreover, the peak of spawning seemed to be correlated with that period in summer when warm plume water (>14°C) was a dominant oceanographic feature. Evidence from Length-Frequency Distributions An analysis of age- and length-frequency distributions played a major role in determining stock structure for the Pacific sardine, Sardinops sagax. A similar analysis of length-frequency distributions was undertaken for the northern anchovy. The following review outlines the rationale and criteria applied in the sardine analysis and adapted for this study. Early sardine investigators at first hypothe- sized that three subpopulations composed this species' total west coast population^. However, in addition to evidence of only sporadic spawn- ing activity (Ahlstrom 1954), age- and length- frequency distributions obtained from the so- called northern subpopulation failed to reveal the presence of the most recently produced age-groups, i.e., the O's, I's, and 2's (Harry 1948). These ages, however, were often observed in samples from the central and southern sub- populations. The consistent absence of O's from northern samples presumably confirmed a lack 'Waldron, K. D., Northwest Fisheries Center, National Marine Fisheries Service, NOAA, Seattle, Wash., pers. commun. 1971. ^A northern subpopulation supposedly ranged from British Columbia southward to central California, while central and southern subpopulations resided respectively off southern California ana lower Baja California. 212 of independent spawning activity by that sub- population of sardines (Harry 1949). The con- sistent presence of O's, of course, would have indicated that independent spawning had occurred. Figure 1 presents the length-frequency dis- tributions analyzed in this study. These anchovy lengths were obtained from four exploratory fishing surveys conducted from Cape Flattery, Wash., to Yaquina Bay, Oreg., during 1966-67. These surveys occasionally encountered schools of anchovies containing small fish which became gilled in the meshes of the survey gear (Figure 2). It was speculated that these small anchovies were the result of recent spawning activity in the Washington-Oregon area. To analyze these length distributions, one must first know the range of lengths associated with individuals belonging to age-group 0. Clark and Phillips ( 1952) indicated that 0-age anchovies begin entering the southern California live-bait fishery at lengths ranging from 5 to 9 cm. Miller (1955) stated that 0-age fish begin enter- ing the southern California commercial fishery at 8.5-9.0 cm. Tillman (1972) concluded that anchovies are at least 6 mo old when they enter the commercially exploitable population at 9 cm. Figure 3 presents the ranges, medians, and median quartiles of the lengths of anchovy larvae obtained by Richardson (1973). These indicate that, in the northern subpopulation, 0-age anchovies approach 6 cm after 5 mo of growth. Thus, a length range of 0-9 cm should define those anchovies which resulted from spawning, at least, during the past 6 mo. This range was used to define the 0-age component in all length-frequency distributions. Applying this criterion, the bar graphs of Figure 1 indicate that 0-age anchovies indeed were present in the northern subpopulation during the years surveyed. Lengths less than 4 cm were not found, but the 4-9 cm range composed, respectively, 11.6, 19.8, 87.0, and 39.0% of these four length-frequency distribu- tions. The results shown for November-December 1966 and February 1967 are particularly striking, having major modes located respectively at 6 and 5.5 cm. These latter two distributions result from the facts that anchovies tend to school by size and that the later 1966 and early 1967 surveys primarily encountered schools of small fish. Therefore, following the rationale discussed at the beginning of this section, the presence of such juveniles would tend to confirm the 20 20 10 a. January 1966 N = I270 tu b. April 1966 N=4I9 J m. 8 10 12 14 16 18 LENGTH (cm) C.November -December 1966 N = 364 V/A Juvenile I I Adult d. February 1967 N=4699 ^^^vy^ 70 8.5 10.0 11.5 130 LENGTH (0.5cm) TTTrT>. 145 16.0 Figure 1.— Composite length-frequency distributions of juvenile and adult northern anchovy sampled off Washington-Oregon during 1966-67. 213 Figure 2. — Juvenile northern anchovy gilled in the meshes (%-% inch) of a mid-water trawl off the Washington-Oregon coast. occurrence of independent spawning activity by the northern subpopulation of anchovies. Discussion Figure 1 gives the results of surveys which took place during the winter or spring. Since the northern subpopulation apparently spawns during the summer, then these figures indicate that spawning occurred during the summer which preceded each survey period. In other words, Figure la and b indicate that spawning occurred during the summer of 1965, resulting in recruit- ment of 0-age fish during January-April 1966. Moreover, Figure Ic and d indicate that spawn- ing occurred in the summer of 1966, resulting in recruitment during November 1966-February 1967. Thus, according to Richardson's data and this analysis, independent spawning by the northern subpopulation seems to have occurred June July- August August September October Figure 3. — Ranges, medians, and median quartiles of lengths of northern anchovy larvae obtained during May-October 1969 off Oregon (Richardson 1973). 214 quite regularly rather than incidentally (occur- ring at least in 1965, 1966, and 1969). Consequently, it is concluded that the presence of anchovies in northern waters does not repre- sent a mere expansion of this species' geographic range — an expansion that well might have ac- companied its recent fivefold increase in total population size. The previously mentioned genetic and meristic evidence, the results of recent larvae surveys, and the above length- frequency analysis would all seem to refute such a conclusion. Moreover, since this subpopulation was the mainstay of a substantive fishery for live bait during the 1940's (Pruter 1966), it seems to have been a persistent feature of the Washington- Oregon coast even before the dramatic expansion of the anchovy biomass which followed the demise of the sardine. Thus the weight of evidence seems to indicate that the northern subpopula- tion of anchovies is one of three independent population elements, all of which are capable of spawning and producing their own recruitment. Literature Cited adjacent ocean waters. U.S. Fish Wildl. Serv., Fish. Ind. Res. 3(3): 17-68. Richardson, S. L. 1973. Abundance and distribution of larval fishes in waters off Oregon, May-October 1969, with special emphasis on the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 71:697-711. Tillman, M. F. 1972. The economic consequences of alternative systems; a simulation study of the fishery for northern anchovy, Engraulis mordax Girard. Ph.D. Thesis, Univ. Washing- ton, Seattle, 227 p. Vrooman, a. M., and p. E. Smith. 1971. Biomass of the subpopulations of northern anchovy Engraulis mordax Girard. Calif. Coop. Ocean. Fish. Invest., Rep. 15:49-51. Michael F. Tillman Northwest Fisheries Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, WA 98112 COMMENT. INTRODUCTION OF CODWM IN NEW ENGLAND WATERS Ahlstrom, E. H. 1954. Distribution and abundance of egg and larval populations of the Pacific sardine. U.S. Fish Wildl. Serv., Fish. Bull. 56:83-140. 1968. What might be gained from an oceanwide survey of fish eggs and larvae in various seasons. Calif. Coop. Ocean. Fish. Invest., Rep. 12:64-67. Baxter, J. L. 1967. Summary of biological information on the northern anchovy Engraulis mordax Girard. Calif. Coop. Ocean. Fish. Invest., Rep. 11:110-116. Clark, F. N., and J. B. Phillips. 1952. The northern anchovy (Engraulis mordax mordax) in the California fishery. Calif. Fish Game 38:189-207. Harry, G. Y., Jr. 1948. Oregon pilchard fishery. Oreg. Fish Comm., Res. Briefs 1(2): 10-15. 1949. The pilchard situation in Oregon. Oreg. Fish Comm., Res. Briefs 2(2):17-22. LeBrasseur, R. 1970. Larval fish species collected in zooplankton samples from the northeastern Pacific Ocean, 1956-1959. Fish. Res. Board Can., Tech. Rep. 175, 47 p. McHuGH, J. L. 1951. Meristic variations and populations of northern anchovy (Engraulis mordax mordax). Bull. Scripps Inst. Oceanogr., Univ. Calif. 6:123-160. Miller, D. J. 1955. Studies relating to the validity of the scale method for age determination of the northern anchovy (Engraulis mordax). Calif. Dep. Fish Game., Fish Bull. 101:6-36. Pruter, A. T. 1966. Commercial fisheries of the Columbia River and Genus Codium is one of the most common forms of seaweed found in almost every latitude but, until recently, has been absent from the east coast of North America. Codium attaches to rocks, pilings, old molluscan shells, and also shells of living oysters, scallops, and mussels. This algae has a number of common names, such as spaghetti grass, staghorn, deadman's fingers, and Japanese weed. It grows rapidly and often becomes so dense that it sometimes creates undesirable conditions on cultivated and natural shellfish beds, as well as in some other environ- ments. At times it becomes buoyant enough to float and to carry along with it mollusks, to the shells of which it is attached. Mass mortalities of Codium are usually followed by quick decom- position, creating adverse conditions that result in the death of mollusks and other bottom forms. No Codium was known to exist in New England waters until approximately the end of the 1950's, when the first specimens of Codium fragile were reported from several aquatic areas adjacent to Long Island. Since then it has become established in the waters of New England, spreading as far north as the State of Maine. According to a recent article (Quinn 1971) "It is now a dominant sea- weed in the waters of Eastern Long Island and 215 can be found from Barnegat Bay, N.J., to Booth- bay Harbor, Me." Because of its wide distribution in the new environment, Codium now causes serious impact on local ecology and also creates serious problems on shellfish beds. There is some question, naturally, as to when the first introduction of this algae occurred and how this somewhat un- desirable "immigrant" was brought into our eastern waters. Quinn (1971) quotes Mueller, who, apparently without any evidence, specu- lates that "It was imported on the backs of oysters from Europe and Japan." Since I am responsible for the introduction of the European oyster, Ostrea edulis, into the waters of New England (Loosanoff 1951, 1955), I wish to comment on this matter. The European oysters were brought to Long Island Sound in October 1949, when I was the Director of the United States Bureau of Com- mercial Fisheries Biological Laboratory at Mil- ford, Conn. The shipment was comprised of approximately 2 bushels of the mollusks, ranging in age from 1 to 3 yr. They were shipped in a vegetable compartment of a large refrigerator on a Holland-American Line passenger ship and spent about 13 days in transit. The introduction of O. edulis was made in accordance with the decision reached after my consultations with members of the shellfish industry, as well as with leading marine biolo- gists of that period, including Paul S. Galtsoff of the United States Bureau of Commercial Fisheries and Thurlow Nelson of Rutgers Uni- versity. Federal authorities approved the impor- tation and the Director of the State of Maine Sea and Shore Fisheries, who was extremely interested in planting European oysters into those waters, gave me a small sum of money to pay for that shipment. The latter fact, obviously, discredits Mueller's statement, quoted by Quinn, that "The oysters were removed from Milford and Woods Hole without permission and intro- duced into local waters." In introducing European oysters it was our desire to establish a second commercial species of bivalves in the waters of Maine. At that time only one mollusk, the soft-shell clam, Mya arenaria, was commercially utilized in that region. However, because of extremely heavy mor- tality among the Mya in the mid-1940's, this species became almost extinct for a period of several years. As a result, many shore com- munities which depended upon soft-shell clam fisheries were deprived of the chief means of their livelihood. Therefore, it seemed logical to me that a second shellfishery should be de- veloped in those waters, namely that of O. edulis. If successful such a development would enhance the economy of the region. Ostrea edulis was chosen for the cold waters of Maine because, in addition to its high quality as human food, it is able to propagate at a considerably lower temperature than the American oyster, Cras- sostrea virginica. In bringing the oysters from Europe, I dealt with my friend, Peter Korringa, who is now Director of the Netherlands Institute for Fishery Investigations. At that time he was already considered one of the world's leading shellfish experts. Being fully aware of the possibility of introducing undesirable exotic species which might accompany the European oyster, our group of American biologists, as well as Korringa, decided to take precautionary measures con- sidered sufficient to prevent such an occurrence. The problem was discussed at great length in correspondence between Korringa and myself, and I still have in my files several of Korringa's letters attesting to this exchange. For example, in his letter of March 1949, Korringa wrote "I can kill any germs in the shell by disinfecting the consignment before shipment." In May of the same year he wrote again "I will disinfect very carefully every oyster we ship you with the chemicals we find satisfactory to that end." In his recent letter to me, dated 27 November 1973, Korringa wrote as follows: "I suggested to treat the oysters by bathing them in a mercury solu- tion, using the organic fungicide we used on large scale against infection with shell disease. This kills hundred procent all organisms on the outside of the shell which cannot withdraw in a hermetically closed shell. You see from my cor- respondence that I have treated the oysters with this disinfectant before shipping them. Therefore I feel sure that Codium fragile cannot have been introduced in the American Atlantic waters with our oysters." When the oysters arrived at Milford, they were again carefully examined, washed with fresh water, and dipped in a weak solution of copper salt. At that time, however, we were not con- cerned as much with the introduction oi Codium 216 as we were afraid of bringing along a highly destructive fungus causing so-called "oyster shell-disease." Because of Korringa's assurance, however, we were quite certain we would elimi- nate this and any similar dangers. The small shipment of European oysters was later divided into two parts, one taken to the U.S. Bureau of Commercial Fisheries Laboratory at Boothbay Harbor, Maine, where some of the oysters were suspended off the dock, the other part kept in Milford Harbor, remaining there under the observation of my associates and me for several years. Not in a single instance did I, other members of Milford Laboratory, or John B. Glude, Director of the U.S. Bureau of Com- mercial Fisheries Laboratory at Boothbay Harbor, or his colleagues, notice or report to me the presence ofCodium on the oyster shells. There- fore, considering the chemical treatment that was given the oysters before they were placed in open American waters and because of the results of our long-term observations of these oysters at both Milford and Boothbay Harbor, it is im- probable that Cocfium was brought into the waters of New England "on the backs of the European oysters." There are several much more plausible ex- planations as to the way Codium was intro- duced to our Atlantic coast. In my opinion, it was brought into our waters during World War IL At that time, to avoid being torpedoed by German submarines along the open Long Island coast, many freighters coming from Europe to the port of New York traveled through the well- protected inside passage — Long Island Sound. At times, these vessels were so numerous that many of them had to be anchored in Long Island Sound for several weeks before they could be unloaded at New York piers. I was then engaged in the study of plankton of Long Island Sound — in relation to propagation of oysters — running, sometimes, 14-h sampling series from a small boat. Several of our collecting stations were then located on the Bridgeport and New Haven oyster beds where seed oysters were dredged each fall and planted on cultivated beds of Long Island, Rhode Island, Massachusetts, and even Maine (Loosanoflf 1966). To avoid the wind and heavy wave action we would usually position our boat on the lee side of anchored freighters. Often we were so close to those vessels that we could converse with members of their crews. Many of these ships were of European registry and, because of the war, most of them were not able to undergo proper bottom cleaning for several years. As a result of this neglect, the ship bottoms were covered with heavy layers of marine fouling organisms. Sometimes such layers, as had been reported by Woods Hole investigators, were as much as 2, or even 3, feet thick (Woods Hole Oceanographic Institution 1952). The fouling mass was composed of many forms, including mussels, tunicates, and, no doubt, a variety of other organisms. The Codium was also present and sometimes clearly visible. While the freighters were riding at anchor, frequently large chunks of the fouling mass broke off and fell to the bottom of the Sound. We witnessed this phenomenon on numerous occasions. Thus, it appears logical that C. fragile gained entrance into the waters of eastern United States from the bottoms of European freighters during World War II. This possibility, however, seems to be ignored; the blame is placed instead, directly or indirectly, on a small, properly handled shipment of European oysters which was brought from Holland to Milford in 1949 (Quinn 1971). It may be mentioned, in conclusion, that, as originally planned, the European oysters planted in Boothbay Harbor not only survived in the new environment but reproduced under a new set of ecological conditions and became firmly estab- lished within a large area (Welch 1966). There- fore, these excellent "immigrants" may soon become the second commercial shellfish crop of Maine. Secondly, Codium, although a nuisance and a highly undesirable invader in some respects, and for introduction of which we claim no "credit," may be a welcome addition to localized biosystems by providing extensive, rich-in-food, protective nursery areas to the advanced larval stage and juveniles of many fishes and of such important species of commercial invertebrates as the American lobster, //omarj/s americanus, and the blue crab, Callinectes sapidus. I wish to thank John B. Glude for reading this manuscript and offering constructive suggestions. Literature Cited LOOSANOFF, V. L. 1951. European oyster, O. edulis, in the waters of the United States. (Abstr.) Anat. Rec. 111:542. 217 1955. The European oyster in American waters. Science Woods Hole Oceanographic Institution. (Wash., D.C.) 121:119-121. 1952. Marine fouling and its prevention. U.S. Nav. Inst., 1966. Time and intensity of setting of the oyster, Annapolis, 388 p. Crassostrea virginica, in Long Island Sound. Biol. Bull. (Woods Hole) 130:211-227. QuiNN, M. Victor L. Loosanoff 1971. A farm that's more briny than rustic. Newsday, Garden City, N.Y. July 26:16. Welch W. R. Pacific Marine Station 1966. The European oyster, Ostrea edulis, in Maine. University of the Pacific Proc. Natl. Shellfish. Assoc. 54:7-39. Dillon Beach, CA 94929 218 ERRATA Fishery Bulletin, Vol. 72, No. 4 Kennedy, V. S., W. H. Roosenburg, M. Castagna, and J. A. Mihursky, "Mercenaria mercenaria (Mollusca: Bivalvia): Temperature-time relationships for survival of embryos and larvae," p. 1160-1166. 1) Page 1161, right column, line 13, correct line to read: (25 mm) in an 8 x 11 matrix (see Figure 1, 2) Page 1163, the figure legends for Figures 2 and 3 were switched in printing. The legend for Figure 2 should be under the second drawing in the left column and that for Figure 3 should be under the drawing in the right column. INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the following instructions. These are not absolute requirements, of course, but desiderata. CONTENT OF MANUSCRIPT The title page should give only the title of the paper, the author's name, his affiliation, and mailing address, including Zip code. The abstract should not exceed one double- spaced page. In the text, Fisheiij Bulletin style, for the most part, follows that of the Style Manual for Biological Journals. Fish names follow the style of the American Fisheries Society Special Pub- lication No. 6, A List of Common and Scientific Names of Fishes from the U)iited States and Canada, Third Edition, 1970. The Meniam- Webster Third New Intey^national Dictionai^ is used as the authority for correct spelling and word division. Text footnotes should be typed separately from the text. Figures and tables, with their legends and headings, should be self-explanatory, not requir- ing reference to the text. Their placement should be indicated in the right-hand margin of the manuscript. Preferably figures should be reduced by pho- tography to 5% inches (for single-column fig- ures, allowing for 50% reduction in printing), or to" 12 inches (for double-column figures). The maximum height, for either width, is 14 inches. Photographs should be printed on glossy paper. Do not send original drawings to the Scien- tific Editor; if they, rather than the photo- graphic reductions, are needed by the printer, the Scientific Publications Staff will request them. Each table should start on a separate page. Consistency in headings and format is desirable. Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be numbered sequentially in arable numerals. To avoid confusion with powers, they should be placed to the left of numerals. Acknowledgments, if included, are placed at the end of the text. Literature is cited in the text as: Lynn and Reid (1968) or (Lynn and Reid 1968). All papers referred to in the text should be listed alphabetically by the senior author's surname under the heading "Literature Cited." Only the author's surname and initials are required in the literature cited. The accuracy of the lit- erature cited is the responsibility of the author. Abbreviations of names of periodicals and serials should conform to Biological Abstracts List of Serials with Title Abbreviations. (Chemical Ab- stracts also uses this system, which was devel- oped by the American Standards Association.) Common abbreviations and symbols, such as mm, m, g, ml, mg, °C (for Celsius), %, "/oo and so forth, should be used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbreviations. We prefer that measurements be given in metric units; other equivalent units may be given in parentheses. FORM OF THE MANUSCRIPT The original of the manuscript should be typed, double-spaced, on white bond paper. Please triple space above headings. We would rather receive good duplicated copies of manuscripts than carbon copies. The sequence of the material should be: TITLE PAGE ABSTRACT TEXT LITERATURE CITED APPENDIX TEXT FOOTNOTES TABLES (Each table should be numbered with an arable numeral and heading provided) LIST OF FIGURES (entire figure legends) FIGURES (Each figure should be numbered with an arable numeral ; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. Bruce B. Collette, Scientific Editor Fishery Bulletin National Marine Fisheries Service Systematics Laboratory National Museum of Natural History Washington, DC 20560 Fifty separates will be supplied to an author free of charge and 100 supplied to his organiza- tion. No covers will be supplied. Contents — continued GILMARTIN, MALVERN, and NOELIA REVELANTE. The concentration of mer- cury, copper, nickel, silver, cadmium, and lead in the northern Adriatic anchovy, Engraulis encrasicholus , and sardine, Sardina pilchardus 193 ANDERSON, WILLIAM W., JACK W. GEHRINGER, and FREDERICK H. BERRY. The correlation between numbers of vertebrae and lateral-line scales in western Atlantic lizardfishes (Synodontidae) 202 RICE, STANLEY D., and ROBERT M. STOKES. Acute toxicity of ammonia to several developmental stages of rainbow trout, Salmo gairdneri 207 Notes TILLMAN, MICHAEL F. Additional evidence substantiating existence of northern subpopulation of northern anchovy, Engraulis mordax 212 LOOSANOFF, VICTOR L. Comment. Introduction of Codium in New England waters 215 ^^t^^O^Co, Fishery Bulletin ^^ATES O^ ^ National Oceanic and Atmospheric Ajdoiiri)Strat^>af J*i^tiDnaHMerine|E,i5heriies I LIBRARY Service 1 9 ]s7S T Vol. 73, No. 2 L, ^g:,i:irr!!L!!^' April 1 975 GRAHAM, JEFFREY B. Heat exchange in the yellowfin tuna, Thunnns albacares, and skipjack tuna, Katsuwonus pelamis, and the adaptive significance of elevated body temperatures in scombrid fishes 219 DAYTON, PAUL K. Experimental studies of algal canopy interactions in a sea otter-dominated kelp community at Amchitka Island, Alaska 230 REEVE, M. R., and L. D. BAKER. Production of two planktonic carnivores (chae- tognath and ctenophore) in South Florida inshore vi^aters 238 MAY, ROBERT C. Effects of acclimation on the temperature and salinity tolerance of the yolk-sac larvae of Bairdiella icistia (Pisces: Sciaenidae) 249 AGNELLO, RICHARD J., and LAWRENCE P. DONNELLEY. The interaction of economic, biological, and legal forces in the Middle Atlantic oyster industry 256 WARNER, ROBERT R. The reproductive biology of the protogynous hermaphrodite Pimelometopon pulchrum (Pisces: Labridae) 262 JOHNSON, ROBERT KARL, and MICHAEL A. BARNETT. An inverse correlation between meristic characters and food supply in mid-water fishes: evidence and possible explanations 284 ROUBAL, WILLIAM T., and TRACY K. COLLIER. Spin-labeling techniques for studying mode of action of petroleum hydrocarbons on marine organisms 299 JOHNSON, JAMES H., and DOUGLAS R. McLAIN. Teleconnections between northeastern Pacific Ocean and the Gulf of Mexico and northwestern Atlantic Ocean 306 KENDALL, ARTHUR W., JR., and JOHN W. REINTJES. Geographic and hydrographic distribution of Atlantic menhaden eggs and larvae along the Middle Atlantic coast from RV Dolphin cruises, 1965-66 317 DARK, THOMAS A. Age and growth of Pacific hake, Merluccius productus 336 THOMAS, ALLAN E. Evaluation of the return of adult chinook salmon to the Aber- nathy incubation channel 356 SERFLING, STEVEN A., and RICHARD F. FORD. Ecological studies of the puerulus larval stage of the California spiny lobster, Panulirus interruptus 360 LANGE, G. D., and A. C. HURLEY. A theoretical treatment of unstructured food webs 378 BLACKBURN, MAURICE, and FRANCIS WILLIAMS. Distribution and ecology of skipjack tuna, Katsuwonus pelamis, in an offshore area of the eastern tropical Pacific Ocean 382 (Continued on back cover) V. / Seattle, Washington U.S. DEPARTMENTOFCOMMERCE NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Adminisirator NATIONALMARINE FISHERIES SERVICE Robert W. Schoning, Director Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 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. Reuben Lasker Scientific Editor, Fishery Bulletin National Marine Fisheries Service Southwest Fisheries Center La Jolla, California 92037 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. William H. Bayliff Inter-American Tropical Tuna Commission Dr. Daniel M. Cohen National Marine Fisheries Service Dr. Howard M. Feder University of Alaska Mr. John E. Fitch California Department of Fish and Game Dr. Marvin D. Grosslein National Marine Fisheries Service Dr. J. Frank Hebard National Marine Fisheries Service Dr. John R. Hunter National Marine Fisheries Service Dr. Arthur S. Merrill National Marine Fisheries Service Dr. Virgil J. Norton University of Rhode Island Mr. Alonzo T. Pruter National Marine Fisheries Service Dr. Theodore R. Rice National Marine Fisheries Service Dr. Brian J. Rothschild National Marine Fisheries Service Mr. Maurice E. Stansby National Marine Fisheries Service Dr. Maynard A. Steinberg National Marine Fisheries Service Dr. Roland L. Wigley National Marine Fisheries Service Kiyoshi G. Fukano, Managing Editor The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget through May 31, 1977. Fishery Bulletin CONTENTS Vol.73, No. 2 April 1975 GRAHAM, JEFFREY B. Heat exchange in the yellowfin tuna, Thunnus albacares, and skipjack tuna, Katsuwonus pelamis, and the adaptive significance of elevated body temperatures in scombrid fishes 219 DAYTON, PAUL K. Experimental studies of algal canopy interactions in a sea otter-dominated kelp community at Amchitka Island, Alaska 230 REEVE, M. R., and L. D. BAKER. Production of two planktonic carnivores (chae- tognath and ctenophore) in South Florida inshore waters 238 MAY, ROBERT C. Effects of acclimation on the temperature and salinity tolerance of the yolk-sac larvae of Bairdiella icistia (Pisces: Sciaenidae) 249 AGNELLO, RICHARD J., and LAWRENCE P. DONNELLEY. The interaction of economic, biological, and legal forces in the Middle Atlantic oyster industry 256 WARNER, ROBERT R. The reproductive biology of the protogynous hermaphrodite Pimelometopon pulchrum (Pisces: Labridae) 262 JOHNSON, ROBERT KARL, and MICHAEL A. BARNETT. An inverse correlation between meristic characters and food supply in mid-water fishes: evidence and possible explanations 284 ROUBAL, WILLIAM T., and TRACY K. COLLIER. Spin-labeling techniques for studying mode of action of petroleum hydrocarbons on marine organisms 299 JOHNSON, JAMES H., and DOUGLAS R. McLAIN. Teleconnections between northeastern Pacific Ocean and the Gulf of Mexico and northwestern Atlantic Ocean 306 KENDALL, ARTHUR W., JR., and JOHN W. REINTJES. Geographic and hydrographic distribution of Atlantic menhaden eggs and larvae along the Middle Atlantic coast from RV Dolphin cruises, 1965-66 317 DARK, THOMAS A. Age and growth of Pacific hake, Merluccius productus 336 THOMAS, ALLAN E. Evaluation of the return of adult chinook salmon to the Aber- nathy incubation channel 356 SERFLING, STEVEN A., and RICHARD F. FORD. Ecological studies of the puerulus larval stage of the California spiny lobster, Panulirus interruptus .... 360 LANGE, G. D., and A. C. HURLEY. A theoretical treatment of unstructured food webs 378 BLACKBURN, MAURICE, and FRANCIS WILLIAMS. Distribution and ecology of skipjack tuna, Katsuwonus pelamis, in an offshore area of the eastern tropical Pacific Ocean 382 (Continued on next page) Seattle, Washington For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402-Subscription price: $11.80 per year ($2.95 additional for foreign mailing). Cost per single issue - $2.95. Contents-continued CAIN, THOMAS D. Reproduction and recruitment of the brackish water clam Rangia cuneata in the James River, Virginia 412 SMIGIELSKI, ALPHONSE S. Hormonal-induced ovulation of the winter flounder, Pseudopleuronectes americanus 431 Notes GWINN, SHARON, and WILLIAM F. PERRIN. Distribution of melanin in the color pattern of Delphinus delphis (Cetacea; Delphinidae) 439 HAUSER, WILLIAM J. Occurrence of two Congridae leptocephali in an estuary . . 444 WILLIAMS, P. M., and K. J. ROBERTSON. Chlorinated hydrocarbons in sea-surface films and subsurface waters at nearshore stations and in the North Central Pacific Gyre 445 MIGHELL, JAMES L., and JAMES R. DANGEL. Hatching survival of hybrids of Oncorhynchus masou with Salmo gairdneri and with North American species of Oncorhynchus 447 SHELDON, WILLIAM W., and ROBERT L. DOW. Trap contributions to losses in the American lobster fishery 449 The National Marine Fisheries Service (NMFS) does not approve, rec- ommend or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. HEAT EXCHANGE IN THE YELLOWFIN TUNA, THUNNUS ALBACARES, AND SKIPJACK TUNA, KATSUWONUS PELAMIS, AND THE ADAPTIVE SIGNIFICANCE OF ELEVATED BODY TEMPERATURES IN SCOMBRID FISHES Jeffrey B. Graham' ABSTRACT Thunnus albacares and Katsuwonus pelamis are warm-bodied fish and use retia mirabilia as counter- current heat exchangers. Both species have four sets of lateral exchangers, two epaxial and two hypaxial, each consisting of a large cutaneous artery and vein and rete. Katsuwonus pelamis has a central exchanger, located within the haemal arch, which consists of the dorsal aorta, the posterior cardinal vein, and a large vertical rete. The central heat exchanger in T. albacares, while also in the haemal arch, is simpler, consisting of two small "wing-shaped" retia on either side of the dorsal aorta and cardinal vein. The adaptive significance of the specialization for heat conservation is discussed. Body temperatures, thermal profiles, and the natural histories of difl'erent warm-bodied species are compared, and warm fishes are contrasted with scombrids that do not conserve heat. The skipjack tunas, Euthynnus and Katsuwonus, have well-developed central heat exchangers and are much warmer than T. albacares. Higher body temperatures in skipjacks seems related to their requirement for a higher basal swimming speed and their faster burst speed. Comparisons on the basis of existing knowledge about the two phyletic groups of Thunnus reveal few differences in swimming ability or factors related to locomotion. The bluefin group, consisting of T. thynnus, T. maccoyii, and T. alalunga, however does contrast with the yellowfin group (T. albacares, T. aflanticus, and T. tonggol) by maintaining generally higher body temperature differentials, having incomplete vertebral circulation through the absence of a posterior cardinal vein, and occurring at higher latitudes. Scombrids (mackerels, bonitos, and tunas) are pelagic, oceanic fishes that are highly adapted for continuous swimming. Some of the more advanced scombrids (principally frigate mackerels, Auxis; skipjack tunas, Euthynnus and Katsuwonus; and tunas, Thunnus) have evolved the capacity to conserve heat generated by the continuous met- abolic activity of their swimming muscle and thus maintain body temperatures that are warmer than ambient seawater (Carey et al. 1971; Carey 1973). There has been convergent evolution for this specialization in mackerel sharks (Isuridae) a highly active, continually swimming group (Carey and Teal 1969a). Warm-bodied fish retain heat by using retia mirabilia (= wonderful network) as counter- current vascular heat exchangers. The principal advantage of a high and fairly constant body temperature is facilitation of continuous swim- 'Smithsonian Tropical Research Institute, P.O. Box 2072, Balboa, Canal Zone. Manuscript accepted August 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. ming by increasing the frequency of muscular contractions, thus increasing available swimming power (Carey et al. 1971). Also, warm-bodied fish probably achieve a marked independence from environmental temperature permitting them to make rapid vertical and latitudinal migrations vdthout the necessity of thermal acclimation. In their extensive review of warm-bodied fish, Carey et al. (1971) described two types of heat exchanger, lateral and central. Lateral heat exchangers (Figure 1) are present in many warm- bodied species but are best developed in the genus Thunnus where they consist of four sets of longi- tudinal subcutaneous arteries and veins (two epaxial and two hypaxial), each with adjoining layers of retial vessels that penetrate the red muscle near the midplane (Gibbs and Collette 1967; Carey et al. 1971). Large, highly developed central heat exchangers (Figure 1) are found in Euthynnus, Katsuwonus, and Auxis. These are located below the vertebral column, in the haemal arch, and consist of a large vertical rete formed from branches of the dorsal aorta and the posterior 219 FISHERY BULLETIN: VOL. 73, NO. 2 E. /meatus K. pelamis T. albacares T. thynnus Figure 1. -Transverse sections of four warm-bodiea species showing the position of central and lateral retia mirabilia (r) that function as vascular heat exchangers. The major blood vessels supplying retia are: dorsal aorta (da), posterior cardinal vein (pcv), cutaneous arteries (ca), and veins (cv). Veins are shown with larger diameters and thinner walls. Red muscle distribution (shaded areas) is also depicted. Noted that the position of cutaneous arteries and veins in Euthynnun lineatuit is reversed compared to that in other species and that only an epaxial pair is present. Also, Thunnus thynnus does not have a posterior cardinal vein. Frigate mackerels (Auxis) are not shown but are very similar to Euthynnus. Data contained in this figure are from various sources cited in the text. 220 GRAHAM: HEAT EXCHANGE IN SCOMBRIDS cardinal vein (Kishinouye 1923; Godsil 1954; Carey et al. 1971; Carey 1973; Graham 1973). Kishinouye (1923: 377; see discussion of Neothunnus, a synonym of T. albacares) described a special subspinal vascular plexus or "kurochiai" in the yellowfin tuna, T. albacares (Bonnaterre), and recent studies have indicated that this struc- ture is a central heat exchanger (Carey et al. 1971; Carey 1973). The kurochiai has not been fully described, nor has the relationship between it and T. albacares' well-developed lateral heat exchangers been considered. Body temperatures of fresh-caught and swimming yellowfin tuna are known to be less than those of skipjack tunas and some other tuna when measured under similar conditions (Barrett and Hester 1964; Carey and Teal 1969b; Stevens and Fry 1971; Carey 1973), but where heat is distributed in the body (thermal profiles) has not been determined for either T. albacares or the skipjack tuna K. pelamis (Lin- naeus). The purpose of this study is to investigate the relationship between body temperature and the types of heat exchanger in T. albacares. The pat- terns observed for this species and K. pelamis are compared with those of other warm-bodied fish. Body temperatures and thermal profiles of fresh- caught T. albacares and K. pelamis are reported, and their central heat exchangers are described. The general structure and circulation pattern of these species' heat exchangers are compared with those of the bluefin tuna T. thynnus, and other skipjack tunas, Euthynnus, and are discussed in terms of their relation to differences in body temperature, morphology, swimming capability, and the natural history of these species. Studies of this type may enable us to understand why there are different kinds of heat exchangers and how these evolved. MATERIALS AND METHODS Eleven T. albacares (360 to 700 mm fork length; weight, 1 to 5 kg) and four K. pelamis (500 to 600 mm, 3 to 4 kg) were caught by surface trolling in the Gulf of Panama and brought on board within 30 to 90 s of being hooked. Red and white muscle temperatures of these specimens were immedi- ately taken with a fast-reading hypodermic therm- istor probe (Yellow Springs Instrument No. 513)- ^Reference to trade names does not imply endorsement by the Na- tional Marine Fisheries Service, NOAA. that had been calibrated against a mercury ther- mometer. Measurements were made deep (near the vertebrae), midway from the vertebrae to the skin, and subcutaneously at several positions along the fishes' lateral midplane, from the oper- culum to the tail, in order to determine the relative contribution of the lateral and central heat exchangers to heat distribution. All temperatures were rounded to the nearest 0.5°C. Body temperatures of shaded fish remained fairly con- stant during the first 10 min after capture, and all measurements were made within this time. A criticism that has been directed against the measurement and interpretation of temperature data from fresh-caught fish is that burst swim- ming to catch a troll lure, or frenzied swimming, together with struggling once hooked may increase body temperatures above typical values. This probably has some validity, but the effects of struggling and handling seem generally overrat- ed. With telemetry, Carey (1973) has shown that free-swimming T. thynnus have body tempera- tures very similar to captured fish. Also, Barrett and Hester (1964) found that immediately cap- tured yellowfin tuna and those that had been tethered for a few minutes had similar tempera- tures. Large frozen T. albacares (800 to 1,400 mm, 8 to 42 kg) were obtained from a commercial fishing vessel, and a range of sizes was dissected to de- termine red muscle distribution, the position, size, and structure of the heat exchangers, and the dimensions of retial vessels. Specimens of K. pelamis were also dissected, and measurements were made. RESULTS Body Temperatures Average deep, intermediate, and subcutaneous red muscle temperatures of eight T. albacares (caught in surface water of 28.5°C) were 30.5°, 30.5°, and 29.5°C. Three specimens caught in 30°C water had average deep body temperatures of 32.5°C. Elevated temperatures in T. albacares oc- cur along the body from the pectoral fins to as far as the third or fourth finlet. The warm region also extends laterally through a large portion of the red muscle. Highest body temperatures were always found in the red and white muscle along and near the lateral midplane of the body. Katsuwonus pelamis is warmer than T. albacares, and its warm 221 FISHERY BULLETIN: VOL. 73, NO. 2 region extends laterally to just below the skin. The average deep, intermediate, and subcutaneous red muscle temperatures of four K. pelamis (caught in 28.5°C surface water) were 35°, 35°, and 33°C. Deep white muscle temperatures in these fish averaged 34° C; brain temperatures were 33° C. The temperatures reported here for T. albacares and K. pelamis are in good agreement with those found for these species by other investigators (Barrett and Hester 1964; Stevens and Fry 1971). Heat-Exchaneer Structure and Red-Muscle Distribution Thunnus albacares The distribution and structure of the lateral heat exchangers found for T. albacares in this study agree fully with those described by Gibbs and Collette (1967) and are summarized here with new notes on variations related to size. Epaxial and hypaxial arteries and veins subdivide from their respective trunks at about vertebrae no. 10 and extend along the body to about two-thirds of the way from the second dorsal fin to the tail (ver- tebrae no. 29 or 30) where they are rejoined by a commissure. One row of retial vessels originates from the lateral edge of each artery and vein, and this is consistent with the observations of Kishinouye (1923, as Neothunnus). Thunnus al- bacares' lateral retia are long and strongly curved towards the center of the body. Retial curvature was not observed in specimens smaller than 3 kg. Cutaneous vessel diameters increase dramatically with increased size, ranging from 0.5 to 1.0 mm (artery and vein) in a 1.1-kg specimen to 6.0 to 8.0 mm in a 42-kg fish. Retial vessels ranged from 0.05 to 0.1 mm in diameter. The central heat exchanger in T. albacares ex- tends from the first to the second dorsal fins (ver- tebrae no. 8 or 9 to 20) and is situated immediately below the vertebrae in the haemal arch. This structure is composed of the dorsal aorta, the posterior cardinal vein, and their small vessels that form two "wing-shaped" retia (Figure 2). Diameters of the dorsal aorta and posterior car- dinal vein only increase slightly with increasing size, ranging from 1.5 to 3.0 mm in a 2.7-kg fish to 3.5 to 4.0 mm in a 42-kg specimen. This contrasts markedly with the large weight-related change in the diameters of the lateral blood vessels. The central retia originate as thick bundles in the haemal arch, then extend supralaterally and pass through vertebral foramina into the red muscle. In the muscle these vessels flatten into broad con- tinuous sheets of alternating veins and arteries (0.1 to 0.2 mm in diameter) that are only one layer thick (Figure 2). This layer penetrates far into the muscle, from 18 mm in a 2.7-kg fish to 40 mm in a 42-kg fish. Red muscle in T. albacares appears in thin bands along each side of the fish at the level of the ver- tebrae (Figure 2). Only red fibers from the hypa- xial muscles actually reach the vertebrae, but epaxial and hypaxial muscle both extend well toward the fishes' side. Longitudinally, red muscle extends from behind the transverse septum (ver- tebrae no. 6 or 7) to as far as the fifth finlet (ver- tebrae no. 28 or 29) and is fairly uniform in thick- ness and shape (cf. Kishinouye 1923, Plate XVII, as Neothunnus). As was found for E. lineatus (Graham 1973) and, as would be expected, there is good agreement in the lineal distribution of red muscle and the lateral and central heat exchangers of T. albacares. Katsuwonus pelamis Except for its higher position in the body, the central exchanger of K. pelamis (Figure 3) is very similar to that of Euthynnus and Auxis, consist- ing of the closely associated dorsal aorta and posterior cardinal vein and a thick vertical rete, all in the haemal arch (Kishinouye 1923; Godsil 1954; Graham 1973). Just posterior to the pectoral fins in a 580-mm (about 4 kg) specimen, the following vessel diameters were measured: dorsal aorta, 2.0 mm; posterior cardinal vein, 4.0 mm; retial vessels, 0.05 to 0.1 mm. At its center (Figure 3), vessels in the central rete of this fish were 8.0 mm long. Lateral heat exchangers are better developed in K. pelamis than in either Euthynnus or Auxis (Figure 1). Both epaxial and hypaxial sets of cu- taneous vessels, with retia, are present, but they are further apart than in T. albacares (Figure 1), reflecting the laterally thicker wedge of red muscle in K. pelamis (see below). The cutaneous vessels are smaller than in Thunnus. The most developed retial vessels occur anteriorally but are variable in their position, length, and the direction they penetrate red muscle (cf. Godsil and Byers 1944, Figure 15). Red muscle in K. pelamis is thicker than in T. albacares but does not appear to extend as far into the tail. In a transverse section (Figure 3), both 222 GRAHAM: HEAT EXCHANGE IN SCOMBRIDS Figure 2.-Central heat exchanger of Thunnus albacares. (Top right, scale = 6.5 cm): Transverse sections show- ing the position of red muscle (rm) and the central heat exchanger (che). (Top left, scale = 1.5 cm): Transverse section of the che showing the dorsal aorta (da), posterior cardinal vein (pcv), and retia (r). (Middle right, scale = 2 cm): A close view of the che showing the da, pcv, and two wing- shaped retia that proceed supra- laterally from the vessels. (Middle left, scale = 1.2 cm): A ventrolateral view of the da, pcv, and the sheet of vessels (v) outside the haemal arch that penetrate red muscle. (Bottom, scale = 2.0 cm): Ventrolateral view showing the che on the left and the thin sheet of vessels in red muscle. hypaxial and epaxial red muscle reach the ver- tebrae. Longitudinally, shape as well as thickness of red muscle varies at different points (cf. Kishinouye 1923, Plate XVII). Generally, red muscle in K. pelamis appears to have more ligaments than T. albacares. In both species the myomeres are continuous through red and white muscle (Figures 2, 3), but red and white muscle are easily distinguished and separate with slight teasing. 223 FISHERY BULLETIN: VOL. 73, NO. 2 Figure 3.-Central heat exchanger of Katsuwonus pelamis. (Top right, scale = 5.6 cm): Transverse section show- ing the position of the central heat exchanger (che) and red muscle (rm). (Top left, scale = 1.1 cm): Close view of the rete (r), the dorsal aorta (da), and the posterior cardinal vein (pcv) in the haemal arch. (Bottom, scale = 1.9 cm): Red muscle and the central heat exchanger. COMPARISON OF KATSUWONUS, EUTHYNNUS, AUXIS, AND T. ALB AC ARES Differences in Heat Exchangers Central heat-exchanger differences can be summarized as follows: Katsuwonus, Euthynnus, and Auxis have only a single vertical rete w^hereas T. albacares has two much smaller retia. Thunnus albacares' central exchanger is immediately below the vertebral centrum (Figures 1 and 2) while in K. pelamis it is lower, about midway between the vertebrae and the coelomic cavity, and in Euthyn- nus and Auxis it is quite low, occurring just above the coelom (Kishinouye 1923; Godsil 1954; Graham 1973). In E. lineatus and E. alletteratus, and K. pelamis that I have examined, and in Auxis (God- sil 1954), the dorsal aorta is actually embedded in the dorsal side of the posterior cardinal vein and is surrounded by a vast network of retial vessels which in effect bathes the aorta in venous blood. This structure has been interpreted as allowing the rete to occupy a full arc over the vessels, thus maximizing its heat-exchanging area (Graham 1973). Both K. pelamis and T. albacares have two pairs of lateral exchangers. Katsuwonus has two 224 GRAHAM: HEAT EXCHANGE IN SCOMBRIDS somewhat variable rows of retial vessels in each lateral exchanger while T. albacares only has one. Euthynnus and Auxis (Figure 1) have only a small pair of epaxial heat exchangers. Thermal Profiles of Fish with Central Exchangers Lateral midplane thermal profiles of T. al- bacares and K. pelamis, taken in the red muscle just posterior to the pectoral fins, illustrate general differences in thermal profiles and body temperatures between these species, E. lineatus, and T. thynnus (Figure 4). Katsuwonus and Euthynnus have much warmer core temperatures than T. albacares, but warmest temperatures in Euthynnus are restricted to a fairly narrow zone around the vertebral column. Euthynnus' profile therefore seems related to its poorly developed lateral exchangers (also red muscle is very thick in the center and thinner laterally, Figure 1) and, based on structural similarities, this type of ther- mal profile would be predicted for Auxis. Kat- suwonus on the other hand, with its lateral exchangers has heat widely distributed across its body. Thunnus albacares, with a small central exchanger and well-developed lateral exchangers has a widely distributed warm region although it is much cooler than K. pelamis and E. lineatus (Figure 4). The dimensions of T. albacares' cu- taneous vessels increase at a much faster rate with increased body weight than do the dorsal aorta and posterior cardinal vein, and in larger fish a greater proportion of blood flow would occur through lateral vessels which might change the thermal profile. COMPARISONS WITHIN THE GENUS THUNNUS Heat Exchangers and Thermal Profiles in T. albacares and T. thynnus Comparative studies of the vascular anatomy of Thunnus show different levels of structural complexity in the heat-exchanging systems (Kishinouye 1923; Godsil and Byers 1944; Gibbs and Collette 1967) which relate to thermal profiles and body temperatures. In T. thynnus, lateral heat exchangers are used solely (Carey and Teal 1966). Two rows of retial vessels emanate from each cu- o o a. E 36 -I 35 34 33 « o 32 n ■o 30 29 - Water temperature 29 C T. albacares — I center edge Relative distance through body Figure 4. -Lateral midplane thermal profiles from the center (near the vertebrae) to the edge (subcutaneous) of red muscle in four species of warm-bodied fish. (Data for Thunnus thynnus were provided by F. G. Carey, that for Euthynnus lineatus are from Graham 1973). taneous artery and vein (only one row occurs in T. albacares), and these extend axially for a long distance (Carey et al. 1971). Reliance upon cu- taneous circulation is so extensive in T. thynnus that the dorsal aorta is reduced in diameter and the posterior cardinal vein is absent. In warm water T. thynnus is about the same temperature as T. albacares, but its thermal profile (Figure 4) reflects the exclusive presence of lateral heat exchangers in that warmest temperatures are found in the middle of the muscle, while the center of the fish is cooler (Carey et al. 1971). (Again, thermal profiles probably change with body size.) Anatomical Features and Phyletic Groupings Related to the Presence or Absence of Complete Vertebral Circulation In their comprehensive study of the genus Thunnus, Gibbs and Collette (1967) recognized seven species which, on the basis of 18 characters, were separated into two phyletic groups: the bluefin tuna group, T. thynnus, T. alalunga, and T. 225 FISHERY BULLETIN: VOL. 73, NO. 2 maccoyii; and the yellowfin tuna group, T. al- bacares, T. atlanticus, and T. tonggol. (The seventh species, T. obesus, has traits in common w^ith both groups and vv^ill be discussed later.) Several of the characters used (Gibbs and Collette 1967, Table 4) to distinguish these groups are related to the presence or absence of complete vertebral circula- tion (both a dorsal aorta and posterior cardinal vein present). The yellow^fin tuna group has a posterior cardinal vein, the bluefin tuna group does not. Another striking difference is the presence of large striations and vascular cones on the livers of fish in the bluefin tuna group. The importance of this is discussed below^. There are several structural modifications in the vertebrae of the yellowfin tuna group which per- mit the passage of more or larger blood vessels through the haemal arch. Prezygapophyses arise far more ventrad on the haemal arch, post- zygapophyses are longer, and the inferior foramina are larger (Gibbs and Collette 1967, Figures 10-13). In describing these vertebrae, Gibbs and Collette (1967:80) remarked that the development of the vertebral openings and processes in the yellowfin tuna group is almost as complex as that in Auxis, Euthynnus, and Kat- suwonus. The presence of complete vertebral cir- culation and appropriate modifications in the ver- tebral column suggested to me that other species in the yellowfin tuna group, in addition to T. al- hacares, may have central heat exchangers. I have examined a preserved section of vertebral column from T. atlanticus (collected in the Gulf of Mexico and sent to me by F. G. Carey) and T. tonggol (obtained by G. Sharp) both of which have a cen- tral exchanger like that of T. albacares. ADAPTIVE SIGNIFICANCE OF DIFFERENT HEAT EXCHANGERS, BODY TEMPERATURES, AND THERMAL PROFILES Heat exchangers in T. albacares differ from those of K. pelamis and E. lineatus, and among these three species, there are marked differences in body temperatures and thermal profiles (Figure 4). Thunnus albacares and T. thynnus also have different heat exchangers, different body temperatures, as well as different thermal profiles depending on body size. Are there morphological features related to locomotion, or ecological fac- tors, such as geographical distribution patterns or feeding behavior, that would explain thermal and anatomical differences between K. pelamis or E. lineatus and T. albacares or between species of Thunnusl Comparisons Within the Genus Thunnus The morphologies and locomotion of T. thynnus and T. albacares have not been compared. There are some data; however it is diffuse and mostly anecdotal, and it does not suggest functional differences in these two species or in the bluefin and yellowfin tuna groups of Thunnus. If species in the yellowfin and bluefin tuna groups are compared on the basis of existing body-temperature data (cf . Carey et al. 1971, Table 1), it is apparent that species in the yellowfin tuna group have lower relative temperatures than those in the bluefin tuna group. Ambient water temperatures are not the same for these different species, and only a general comparison is possible. Still, these differences agree with the known differences in T. albacares and T. thynnus (Carey and Teal 1969b; Carey 1973) and are suggestive of a general trend of body-temperature differences that might in turn reflect a significant functional difference between the two taxonomic groups. A feature in the natural history of species in the yellowfin and bluefin tuna groups that clearly separates them, and relates to their anatomical and temperature differences as well, is the water temperature that they normally inhabit. Thunnus maccoyii and T. alalunga of the bluefin tuna group occur only in cool water while T. thynnus, because of its thermoregulatory ability, is wide ranging and may occur in waters from 6° to 30°C but seems most common in the range 16° to 22° C (Gibbs and Collette 1967; Carey and Teal 1969b). Of the yellowfin tuna group, T. albacares usually occurs from 20° to 28°C (Schaeffer et al. 1963), and both T. tonggol and T. atlanticus are strictly tropical species (Gibbs and Collette 1967). Several facts suggest that incomplete vertebral circulation in the bluefin group is a specialization for living in cooler water and that central heat exchangers are a primitive character related to the occurrence of the yellowfin tuna group in tropical waters. First, central heat exchangers, being re- stricted to within the haemal arch, are, of necessity small and therefore have limited heat-exchanging capacity. Thus, in cool water, and, given that red muscle is large and located at varying distances away from the vertebrae, a small central heat 226 GRAHAM: HEAT EXCHANGE IN SCOMBRIDS exchanger may prove insufficient to maintain a warm temperature. Carey and Teal (1966, 1969b) pointed out the obvious insulative value of having a large lateral heat exchanger between the warm muscle and cool water. Also, in cool water it may not be efficient for heat conservation to pump a large volume of cool blood (from the gills) into the center of the body via the dorsal aorta, and this might explain why the dorsal aorta in T. thynnus is small. Indeed, the lower core temperature found in T. thynnus may result from the small volume of unheated blood that does flow through the dorsal aorta. Another vascular specialization that ap- pears directly related to the cool-water distribu- tion of the bluefin tuna group is the presence of vascular bundles on their livers which enables these fish to warm their viscera, thus facilitating digestion in cooler water. A consideration of the bigeye tuna, T. obesus, substantiates the idea that central heat exchangers and ultimately complete vertebral cir- culation are lost as tuna species evolve into cooler habitats. Although T. obesus and T. albacares have practically the same latitudinal distributions (Gibbs and Collette 1967), the former occurs in deeper and therefore cooler water (Kishinouye 1923:390 as Parathunnus mebachi, a synonym for T. obesus). This aspect of the distribution of T. obesus thus makes it intermediate, in terms of its thermal habitat, to that of the bluefin and yellowfin tuna groups. Thunnus obesus is also morphologically intermediate to the bluefin and yellowfin tuna groups of Thunnus. It has complete vertebral circulation and vascular bundles on its Hver (Gibbs and Collette 1967) yet, F. G. Carey (pers. commun.) who has extensively studied this species reports that it does not have a central heat exchanger. With respect to body temperatures, thermal profiles, and the structure of its lateral heat exchangers, T. obesus closely resembles T. thynnus (Carey and Teal 1966). Thus for the bigeye tuna, which in terms of adapting to cool water appears to be at an intermediate position between the yellowfin and bluefin tuna groups, a central heat exchanger is not present although complete vertebral circulation persists. With re- spect to the latter, however, and perhaps under- scoring the de-emphasis of vertebral circulation, it is relevant to point out that although T. obesus does have a posterior cardinal vein, Godsil and Byers (1944:114) describe it as "relatively small" and note that it fuses anteriorly with the right cutaneous vein. Elevated Body Temperatures and Locomotion in Skipjack Tunas and T. albacares Studies of scombrid locomotion (Fierstine and Walters 1968; Magnuson 1970, 1973) suggest that elevated body temperature in skipjacks, while related to their requirement for a faster typical (basal) speed, primarily contributes to their higher burst swimming speed. Magnuson (1970, 1973) pointed out that scombrids are negatively buoyant and that the skipjack tunas, which lack a gas bladder, are even more negatively buoyant than is T. albacares. To compensate for this, and to maintain hydrostatic equilibrium, skipjack tunas must swim more rapidly. Magnuson has argued that the need for a faster basal speed correlates well with a sig- nificantly higher amount of red muscle found in skipjack tunas (about 8% of body weight in Kat- suwonus and Euthynnus, compared with 7.4% in T. albacares of the same size) and with their slightly greater amounts of blood hemoglobin (Magnuson 1973, Table 7). The amount of red muscle of course bears an important relationship to body temperature. In warm-bodied fish, retia supply blood to red muscle which is highly aerobic. Red muscle is the principal organ used for basal swimming (Rayner and Keenan 1967), and therefore it is the principal site of thermogenesis. (White muscle mainly functions in burst swimming.) Thus skipjack tunas, to maintain a high basal speed, have a large mass of red muscle, and it could be logically concluded that to augment power output, the capacity to conserve heat and keep swimming muscles warm has evolved in skipjack tunas. The difficulty with this idea however is that other scombrids such as the Pacific bonito, Sarda chiliensis, have minimum speed requirements as high as those of the skip- jack tunas (Magnuson 1973), but are not warm- bodied, nor do they have high hemoglobin levels or large amounts of red muscle. This obviously in- dicates that elevated body temperatures and high amounts of hemoglobin and red muscle in the skipjack tunas, while contributing to the sustenance of a high basal speed, must have other functions as well. Further comparison of Sarda with Euthynnus provides valuable insight to the significance of elevated body temperature to burst swimming. Sarda velox and E. lineatus (Figure 5) attain about the same size and are morphologically 227 FISHERY BULLETIN: VOL. 73, NO. 2 Figure 5.— The bonita, Sarda vehx, (top) and the skipjack tuna, Euthynnus lineatus (bottom). Note differences in body shape and pectoral and caudal fin size and shape. similar. These species also have similar distribu- tions and, in the Gulf of Panama, they occur in the same areas and eat similar prey (crustaceans, squid, and small fishes; pers. obs.) although Sarda has a bigger mouth and large teeth. The different mouths and other differences suggest that the sv^^imming capability of these species are also different. Sarda has a smaller pectoral fin (Magnuson 1973; Figure 5, this paper) and a low^er caudal fin aspect ratio (Fierstine and Walters 1968, Table 7). Its red muscle is not as well developed as that in Euthynnus (Fierstine and Walters 1968:17), and Sarda has much less blood hemoglobin (Klaw^e et al. 1963). Finally, a very striking difference exists in the maximum burst speeds of E. affinia and S. chiliensis (Magnuson 1973, Table 6). In fact, the three vi^arm-bodied species listed by Magnuson (Table 6), all have burst speeds nearly double those of 5. chiliensis, suggesting that elevated body temperatures, coupled with morphological adaptations, greatly increase the maximum swimming speed. The principal contribution of high body temperature to burst swimming is probably the maintenance of a thermal profile that warms large portions of white muscle. For Katsuwonus, Euthynnus, and T. albacares, which are all tropical species, there are differences in several structures related to locomotion such as caudal fin aspect ratio and the amount, distribu- tion, and shape of red muscle (Fierstine and Walters 1968). It is reasonable to assume that these differences, combined with elevated body temperature, must confer different capabilities for acceleration, maneuverability, and sustained swimming on different species. One difficulty with the data presently available however is that T. albacares grows to be much larger than skipjack tunas, and allometric growth is known or thought to occur in several locomotion-related structures (see discussions by Gibbs and CoUette 1967; Mag- nuson 1973). Without quantitative data on growth patterns of these features, their contribution to locomotion cannot be fully evaluated. ACKNOWLEDGMENTS This study was supported by the Smithsonian 228 GRAHAM: HEAT EXCHANGE IN SCOMBRIDS Tropical Research Institute (STRI) and all field work was done on board the STRI Research Ves- sels Tethys and Stenella. My special thanks are extended to Lawerence G. Abele, Florida State University, Francis G. Carey, Woods Hole Oceanographic Institution, Bruce B. Collette, Na- tional Marine Fisheries Service, Robert H. Gibbs, Jr., Smithsonian Institution, and Robert R. Warner, STRI, who critically reviewed this manuscript and made substantial suggestions for its improvement. I also thank M. May, William Neill, John L. Roberts, Richard H. Rosenblatt, Sherry Steffel, and Gary Sharp for stimulating discussions of this work. Finally, I thank Fred S. Robison for his technical assistance and Panama Agencies, S. A. for their help in obtaining specimens from commercial vessels. LITERATURE CITED Barrett, I., and F. J. Hester. 1964. Body temperature of yellowfin and skipjack tunas in relation to sea surface temperature. Nature (Lond.) 203:96-97. Carey, F. G. 1973. Fishes with warm bodies. Sci. Am. 228(2):36-44. Carey, F. G., and J. M. Teal. 1966. Heat conservation in tuna fish muscle. Proc. Natl. Acad. Sci. U.S.A. 56:1464-1469. 1969a. Mako and porbeagle: Warm-bodied sharks. Comp. Biochem. Physiol. 28:199-204. 1969b. Regulation of body temperature by the bluefin tuna. Comp. Biochem. Physiol. 28:205-213. 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. FlERSTINE, H. L., and V. WALTERS. 1968. Studies in locomotion and anatomy of scombroid fishes. Mem. South. Calif. Acad. Sci. 6:1-29. 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. Godsil, H. C, and R. D. Byers. 1944. A systematic study of the Pacific tunas. Calif. Div. Fish Game, Fish. Bull. 60, 131 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. KiSHINOUYE, K. 1923. Contributions to the comparative study of the so- called scombroid fishes. J. Coll. Agric, Imp. Univ. Tokyo 8:293-475. Klawe, W. L., I. Barrett, and B. M. H. Klawe. 1963. Hemoglobin content of the blood of six species of scombroid fishes. Nature (Lond.) 198:96. Magnuson, J. J. 1970. Hydrostatic equilibrium of Euthynnus affinis, a pelagic teleost without a gas bladder. Copeia 1970:56-85. 1973. Comparative study of adaptations for continuous swimming and hydrostatic equilibrium of scombroid and xiphoid fishes. Fish. Bull, U.S. 71:337-356. Rayner, M. D., and M. J. Keenan. 1967. Role of red and white muscles in the swimming of the skipjack tuna. Nature (Lond.) 214:392-393. Schaeffer, M. B., G. C. Broadhead, and C. J. Orange. 1963. Synopsis on the biology of yellowfin tuna Thunnus (Neothunnus) albacares (Bonnaterre) 1788 (Pacific Ocean). FAO (Food Agric. Organ. U.N.) Fish. Rep. 6:538-561. Stevens, E. D., and F. E. J. Fry. 1971. Brain and muscle temperatures in ocean caught and captive skipjack tuna. Comp. Biochem. Physiol. 38A:203-211. 229 EXPERIMENTAL STUDIES OF ALGAL CANOPY INTERACTIONS IN A SEA OTTER-DOMINATED KELP COMMUNITY AT AMCHITKA ISLAND, ALASKA Paul K. Dayton' ABSTRACT Studies on the results of competitive interactions between three kelp canopy guilds were conducted in a community in which herbivorous invertebrates have been largely removed from shallow water (approximately 20 m) by sea otters. Small sea urchins observed in the haptera of kelps all disappeared following the canopy removal, suggesting that the canopy itself offers a modest refuge from their predators. Experiments prove that the largest alga, Alariafistulosa, behaves as a fugitive species with respect to Laminaria and Agarum species in spite of the structural dominance of a floating canopy. Vegetative regeneration may give Laminaria longipes an advantage over other Laminaria species, Alaria, and presumably Agarum cribrosum following disturbances in very shallow water {<5 m). Laminaria species suppress Agarum growth (and recruitment) in moderate depths (5-20 m) where either Laminaria or Agarum suppresses growth of red algal turf beneath them, and where both Laminaria and Agarum must be removed to allow recruitment and growth of Alaria fistulosa. Although kelps were observed to depths of 30 m, their lower distribution appears primarily limited by sea urchin grazing. Few natural communities are so influenced by one population as is the nearshore marine community dominated by the sea otter, Enhydra lutris Linn. The nearshore community at Amchitka Island, Alaska, is especially interesting in this regard because for almost 40 yr it has had a sizable sea otter population. This population has been at or near its carrying capacity for at least 20 yr (Kenyon 1969; Estes and Smith 1973), and is thus one of the few localities where the sea otter can be found in a natural balance with the rest of its community. The sea otter exerts its powerful influence in shallow water, where its predation on diverse kinds of invertebrates is remarkably efficient. In addition to drastically reducing populations of motile herbivores (McLean 1962; Ebert 1968; Lowry and Pearse 1973; Estes and Palmisano 1974), the sea otters eat many sessile animals and may release the algae from potential space competition with many potentially compe- titively important species such as the bivalves Mytilus edulis, Modiolus modiolus, and Pododes- mus macroschisma, and the barnacles Balanus spp. The algal community at Amchitka Island, then, offers unusual opportunities to evaluate al- ' Scripps Institution of Oceanography, P.O. Box 1529, La Jolla, CA 92037. gal-algal interactions in the natural absence of herbivores and animal space competitors. Such interactions might suggest important competitive components of the algal "niches." The sublittoral association of perennial algae at Amchitka has four separate canopies (Figure 1). Alariafistulosa P. et R. is a conspicuous kelp with long floating fronds that form a canopy on the surface (Kibbe 1915). The thickest Alaria canopy is usually found in relatively shallow (< 5 m) water. The second canopy level is composed of the following stipitate Laminaria species: L. groenlandica Rosenvinge, L. dentigera Kjellman, L. yezoensis Miyabe, and L. longipes Bory. This canopy can be found from the intertidal to depths of approximately 20 m. The third canopy is usually composed of Agarum cribrosum Bory with short stripes and large broad fronds lying prostrate on the substratum. This canopy of prostrate kelp oc- curs between 10 and 20 m. Finally there is a turf composed of numerous species of red algae and occasional clumps of green algae, especially Codium ritteri Setch. et Gardn. and Cladophora spp. The fact that the canopies tend to occupy nonoverlapping patches in shallow (< 10 m) water suggests that there are competitive interactions between the species comprising the canopies. This paper discusses tests of a series of hypotheses Manuscript accepted June 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 230 DAYTON: EXPERIMENTAL STUDIES OF ALGAL CANOPY TESTS: —Along effect - Lominorio spp effect -Ability of L; lonqipes to regrow in disturbed area, particularly in relation to ability of Laminoria spp Figure l.-Drawing of the kelp canopies at three different depths. Laminaria spp. refers to the large and very similar stipitate L. groenlandica, L. dentigera, and L. yezoensia which seem to occupy broadly overlapping depth profiles but form identical canopies because the stipe lengths and frond sizes are very similar. Diagrams of the experimental design testing hypotheses about the competitive effects between canopies is included for the two manipulated areas. about the competitive effects these canopies have on each other, the role of physical disturbance in canopy composition, and a gradient of herbivore pressures in deeper waters, where the sea otter foraging becomes less efficient. METHODS This research was done in July 1971 and April 1972 in a small bay between the remains of the old Constantino jetty and Kirilof Point on the Bering^ Sea. A total of 34 dives were made during the study. There were two study sites, a nearshore shallow (< 5 m) area beside an old quarry and a deeper (>7 m) reef about 150 m offshore. Immediately offshore in the shallow area there is a very heavy summer canopy of Alaria mixed with a dense growth of annual brown algae such as Cymathere triplicata (P. et R.) J. Ag., Desmares- tia intermedia P. et R., and numerous species of red algae representing such genera as Ptilota, Hypophyllum, etc. Offshore from this dense algal band, but still in the shallow area, are distinct patches of Alaria with thick canopies floating on the surface and patches of a very solid secondary Laminaria canopy. There are two Laminaria growth forms in the more shallow (< 5 m) area: L. groenlandica, L. dentigera, and L. yezoensis are solitary plants with one heavy 50-150 cm stipe per plant; L. longipes has thin multiple 20-40 cm stipes from a single rhizomelike holdfast (Markham 1968, 1972). The third prostrate canopy is represented in shallow water by scattered individuals of the heavy brown alga Thalassiophyllum clathrus (Gmelin) P. et R. The deeper offshore reef has a scattered and relatively thin (0-20%) canopy of Alaria and in the more shallow (7-12 m) levels a very thick canopy cover of Laminaria spp. With increasing depth the Alaria density decreases and the Laminaria is gradually replaced by Agarum cribrosum which forms the third prostrate canopy. The experimental sites were chosen on the basis of distinct patches of the respective canopies to be manipulated and on the ease of shore access and relocation. Pruning shears were used to clear areas by cutting the stipes just above the holdfasts. In every case an immediately adjacent area was monitored as a control. 231 FISHERY BULLETIN: VOL. 73, NO. 2 Methods of estimating percent canopy cover varied. The Alaria canopies represent visual es- timates. The 100% covers were very thick and in these cases the floating stipes seemed to form an almost impenetrable wall in the water column. A few photographs taken of the Alaria canopy in areas where it had less than 100% cover suggest that the visual estimates in these locations were conservative. The other percent cover estimates were made with the aid of 0.25 or 0. 16 m" quadrats which, in larger areas, were placed haphazardly, and in restricted experimental areas were placed systematically in such a way that the entire experimental area was sampled. The actual measurements were usually taken planimetrically from photographs as defined earlier (Dayton 1971). There were a number of cases in which visual estimates were used because of camera malfunction, running out of film, etc. I have com- pared such visual estimates with planimeter measurements and found that they are usually within 5% and always within 10% of each other (Dayton 1971, 1975). The data are presented as means because the actual sample numbers varied (but except where stated, were never fewer than 10); the variance is given as standard error. RESULTS Shallow Area This area is covered with an extremely thick growth of algae and is generally characterized by a conspicuous absence of herbivores (Estes and Palmisano 1974). I was surprised to find sea urchins^ among the Laminaria (especially L. lon- gipes) haptera and holdfasts upon removing the canopies for the experiments discussed below. The sea urchins may exist in these sheltered refuges Opinions are divided whether the Amchitka sea urchin is Strongylocentrotus drobachiensia orS. polyacanthus. because the canopy is both very dense and rela- tively close (25-35 cm) to the substratum, thus seriously reducing the foraging efficiencies of their visual predators. This sea urchin-refuge hypothesis was supported by the observation that the sea urchins remained untouched in both clear- ings from 3 and 6 July through 8 July, but all were gone on 9 July. I suspect that they were taken by a sea otter that found the cleared patches, as one was observed foraging in the vicinity on the morning of 9 July. However, predation by the common eider, Somateria mollissima (Williamson and Emison 1969), and emigration are other pos- sible explanations. At any rate, the small size (< 15 mm) and scarcity of these sea urchins do not seriously affect the contention that the herbivores have largely been eliminated from this area. The elimination of the grazing pressures makes the competition-based hypotheses discussed below more meaningful. Hypothesis I The Alaria fistulosa canopy excludes Laminaria spp. This hypothesis was tested (a) by cutting Alaria from several rocks and observing whether Laminaria recruited in the absence of Alaria and (b) by cutting Laminaria and observ- ing potential Alaria recruitment. Alaria and probably Laminaria spp. were fertile at the time of the cutting. Significantly more Laminaria recruitment into Alaria clearings than into uncleared controls would support the hypothesis, whereas significantly more Alaria recruitment into Laminaria clearings than into the control would negate the hypothesis and suggest the truth of the converse hypothesis, that Alaria behaves as an opportunistic or fugitive species (Dayton 1973, 1975) in the presence of competition with the competitively dominant Laminaria spp. The results of such clearings at a depth of 5 m (done 3 and 4 July 1971) are presented in Table 1. The Table 1. -Effects of canopies of Alaria fistulosa and Laminaria spp. on each other and on the cover of red algae in the nearshore experimental area (25 m-) at 3-5 m depth. The data are presented as percent cover with the variance presented as the 95% confidence interval about the mean. Data presented without variance were visual estimates. Control no. 1 suffered heavy algal loss from winter storms. The mean density of A. fistulosa in the April 1972 Laminaria removal experiment was 14.7 ( ± 1.1, SE) in ten 100 cm- quadrats. Alaria removal Laminaria removal Control no. 1 Control no. 2 Canopy species July 71 April 72 July 71 April 72 July 71 April 72 July 71 April 72 Alaria fistulosa Laminaria spp. Red algal turf '75 35.7 ± 15.0 40.4 ± 10.7 20.3 ± 20.0 39.2 ± 12.1 45.6 ± 13.6 5 187.2 ± 7.9 15.3 ±8.6 100 0 45.5 ± 4.6 45 100 nOO ± 0 25.3 ± 20.0 10.2 ±5.8 40.2 ±12.0 10 100 ±0 5.4 ±5.3 5 100 ±0 15.8 ± 7.9 'Signifies that the canopy was experimentally removed. ^Canopy ripped out during winter storms. 232 DAYTON: EXPERIMENTAL STUDIES OF ALGAL CANOPY Alaria forming a 75% canopy were removed from a 25 m^ area and no significant change was ob- served in the Laminaria or red algal turf canopies by April 1972. But the removal of an 87% cover of Laminaria produced dramatic (5-100%) increases in the Alaria cover and a significant (P< 0.001) increase in the red algal turf covers (<-test run on data normalized with an arcsine transformation). The 100% Laminaria cover in Control no. 1 suf- fered heavy damage when two large boulders, rolled about by winter storms, reduced Laminaria densities and resulted in significant increases in recruitment of Alaria and red algal turf covers (P<0.01). In addition to the extremely heavy Alaria recruitment in the Laminaria removal areas, there were also patches of Rhodymenia palmata (L.) Greville, Ptilota spp., Desmarestia spp., Cymathere triplicata, Chaetomorpha melagonium (Weber et Mohr) Jutz., and Coilodesme spp. No significant changes were ob- served in Control no. 2. To a certain extent these observations could be explained by a very slow growth rate of Laminaria spp. But certainly the hypothesis that Alaria dominates in competition over Laminaria was negated, and these data strongly support the conclusion that despite the expected competitive advantage gained by form- ing a surface canopy, Alaria fistulosa is not a competitive dominant, but a fugitive species colonizing areas released from competition with the dominant Laminaria canopy. Hypothesis II The rhizoidal growth pattern of Laminaria longipes allows an efficient recovery following a disturbance (Markham 1968). The hypothesis sug- gests that the removal of an L. longipes canopy results in the area being succeeded by its own extensive vegetative regrowth, in contrast to the invasion of many individuals of fugitive species seen following the removal of a mixed species canopy of Laminaria groenlandica, L. yezoensis, and L. dentigera. This hypothesis was tested by cutting the stipes near the holdfasts of a 100% cover of L. longipes from a 10 m^ patch at a depth of 3 m on 7 July 1971. Fifteen V4 m" quadrats observed after the 100% canopy was removed showed the following mean substratum covers: 57% (± 4.9, SE) L. longipes holdfasts, 7% {± 1.8, SE) sponges and compound tunicates, and 22% (± 5.2, SE) coralline algae, mainly Clathromorphum spp. They also showed mean Va m^ densities of the sea urchin, Strongylocentrotus sp., of 17.5 {+_ 3.8, SE) and the asteroid, Leptasterias aleutica, of 1.0 (±. 0.3, SE). Spores of the three other Laminaria species and of Alaria were potentially available from many plants on rocks on three sides of the clearing. By April 1972, the clearing had been completely recolonized by L. longipes, despite the proximity of large plants of the other species. The recovery was so complete that the clearing could only be recog- nized after a long search located a few "land- marks" (sponges, compound tunicates, and a Laminaria yezoensis holdfast with the stipe cut by pruning shears) photographed the previous year. This strongly supports the hypothesis that the rhizoidal growth pattern of L. longipes is an ef- fective adaptation for the recovery of its canopy following a disturbance and is in marked contrast to the heavy Alaria recruitment following the removal of a nearby Laminaria spp. canopy. I was unable to test the obvious hypothesis that this capacity for vegetative growth gives L. longipes an advantage over the other Laminaria spp. in a disturbed area, but loses a competitive advantage in less disturbed areas because the other Laminaria species have a higher, more effective canopy. Offshore Area An exploratory dive was made on the deeper offshore reef to investigate the relationship between sea urchin densities and the various algal canopies. Samples were taken from haphazardly placed V4 m^ quadrats. Five samples taken in the 12-15 m range showed means of 44% (± 23.3, SE) cover of Laminaria spp. and 62% (± 15.7, SE) cover of Agarum crihrosum, and a mean density of 11.2 (± 3.8, SE) sea urchins per V4 ml In the 15-21 m depth range five samples provided means of 36% (± 13.0, SE) canopy cover of Laminaria and 80% (± 4.9, SE) canopy cover of Agarum with a mean sea urchin density of 6.4 (± 3.2, SE) per Vi ml Few identifiable foliose algae were seen below 21 m, but there was a high mean sea urchin density of 30.4 ( +. 3.7, SE) per V4 m-. In these deeper areas there was almost a complete substratum cover of the encrusting coralline algae Clathromorphum spp. and the green alga, Codium ritteri. Only four Alaria plants were encountered in these 17 samples; all were growing from the top portion of one Laminaria stipe at 11 m. 233 FISHERY BULLETIN: VOL. 73, NO. 2 On 11 and 12 July 1971, a study site was chosen and the data in Figure 2 labelled July 1971 were collected. The differences between these data and those given in the preceding paragraph give an idea of the variation in this area. The inverse relationship between the percent cover of Laminaria and Agarum, in which the Laniinaria decreases and the Agarum increases with depth and sea urchin density, suggests that in shallow water Laminaria competition suppresses the growth of Agarum, but that Agarum, which has been demonstrated to be highly distasteful to Strong ylocenirotus drobachiensis (Vadas 1968), is Lominano spp i\ Agarum cribrosum o^A 1 ^Aprll 1972 ,'T-4 .9ar?/wi canopy cover at that time shown in Figure 2B is nearly complete only at those depths at which there is reduced Laminaria coverage and relatively low sea urchin den- sity. After removal of Laminaria, the Agarum canopy increased dramatically at the shallower depths. The increase of red algal cover after Laminaria removal is shown in Figure 2C. Variance is presented as the 95% confidence interval around the mean. more successful in the presence of a moderate density of grazers. Finally, Agarum itself may also have an important competitive effect against Alaria and the foliose red algal turf. Grazing pressure and limiting light conditions probably cause the severe reduction of foliose algae in deeper water. These data demonstrating high densities of sea urchins at depths below 20 m agree with the observations of Barr (1971), Estes and Smith (1973), and Estes and Palmisano (1974). This suggests that sea otters at Amchitka do not forage effectively below 18-20 m. That the experimental area could not be con- tinuously monitored meant that it was not possible to manipulate the sea urchin density, but compet- itive effects of the algae at this depth were readily testable by selective removal of algal species. Hypothesis III The presence of Laminaria spp. has no effect on other algae. This hypothesis was tested by remov- ing a 2-m wide strip of Laminaria from the area where the data in Figure 2A were collected. The hypothesis was negated as both Agarum and the foliose red algae canopies significantly increased their covers (Figure 2B, C). The spectacular increase in the cover of the Agarum canopy cer- tainly resulted partially from growth of the fronds; however, samples taken in April 1971 and repeated in July 1972 at approximately the same spots along the experimental Laminaria removal strip, showed that the mean Agarum density increased significantly from 4.1 (± 0.6, SE; ten Vi m^ samples) plants to 15.6 plants per Va m^ (was calculated from ten 1/16 m- samples with a mean of 3.9; +_ 0.4 SE). The increase in canopy cover of the red algal turf was less spectacular, but a one- tailed Wilcoxon matched-pairs signed-ranks test of mean percent canopy cover at all depths con- sidered shows a significant (P<0.005) general increase after the Laminaria were removed, this despite the fact that April may be early in the season for red algal growth. Thus the Laminaria canopy in the presence of an Agarum canopy has an important effect on other algal species. Hypothesis IV The Agarum cribrosum canopy alone has no ef- fect on the other algae. This hypothesis was tested by clearing 45-85% covers of Agarum from 4 m- 234 DAYTON: EXPERIMENTAL STUDIES OF ALGAL CANOPY plots at 9.1- and 16.8-m depths in July 1971. In both cases a 100% canopy of Laminaria persisted throughout the experiment. A slight recovery of the Agarum population was observed the follow- ing April (Table 2), but no significant differences were observed in the numbers or percent cover of the other species. Thus there is, at present, no reason to negate the hypothesis. Hypothesis V The Agarum cribrosum canopy in the absence of the Laminaria canopy has an important effect on the other species of algae. This hypothesis was tested by removing both Agarum and Laminaria canopies from 4 m- plots at 9.1- and 16.8-m depths. These clearings were then compared to those in the adjacent Laminaria-on\y removal ex- periments at the same depths (Figure 2C). A strict interpretation of this comparison suggests that either a Laminaria or Agarum canopy or both is sufficient to prevent an increase of red algal turf cover because there is, at those two particular depths, no significant increase of red algal turf in either the Laminaria-on\y or Agarum-on\y removal experiments (Figure 2C, Table 2). This interpretation is equivocal, however, as Hypothesis III demonstrated a slight but sig- nificant Laminaria effect on the red algal turf. There is no equivocation regarding the effect of the combined Laminaria and Agarum canopies on the red algal turf which increased from 7 to 49% at 9.1 m and 1 to 38% at 16.8 m (Table 2). These are much more dramatic increases than were observed in the Laminaria-on\y removal areas and con- vincingly argue for a strong Agarum effect in the absence of Laminaria. Some of the red algae in this experiment were Ptilota asplenoides (Esper) C. Ag., Laingia aleutica Wynne, Hypophyllum ruprechtianum Zinova, Constantinea rosa- marina (Gmelin) P. et R., Pantoneura juergensii (J. Ag.) Kylin, Cirrulicarpus gmelini (Grunow) Tokida et Masaki, Turnerella sp., Callophyllis flabellulata Harvey, and Nienburgia prolifera Wynne. The most impressive effect of the Agarum canopy in the absence of Laminaria was its inhibition of Alaria recruitment. In each of the two quadrats from which both Laminaria and Agarum were removed, the Alaria cover, consist- ing of a heavy recruitment of juvenile plants, increased from 0 to 100% canopy cover (Table 2). The Alaria response was particularly impressive because the dense Alaria recruitment completely filled, but was perfectly contained within, the Agarum-and-Laminaria removal patches. The mean density increased from 0 to 22.8 Alaria plants per 1/16 m- (± 3.5, SE). In contrast to this result in the Agarum-a,nd- Laminaria removal area, there was no Alaria recruitment in the rather extensive area from which Laminaria alone was removed (Figure 2). This result also contrasts sharply with those of the shallow Laminaria removal experiments (Table 1), in which no Agarum canopy level existed. An ad- jacent control was monitored for each experimen- tal clearing; no changes were observed in any of the controls. The above comparisons demonstrate that both the secondary Laminaria canopy and the tertiary Agarum canopy individually can significantly reduce the recruitment of Alaria, the species which forms the primary surface canopy. Further evidence of the intense competition in the deeper area where both understory canopies exist is provided by the observation that, of 100 Alaria plants surveyed, 79 were utilizing secondary sub- strata with their holdfasts attached high on Laminaria stipes (Figure 1). Table 2.— Effects of Agarum cribrosum and combined Agarum-Laminaria spp. canopies on each other, red algal turf, and Alaria fistuloita at 9.1-m and 16.8-m depths in the offshore study site. Each experimental clearing area was 4 m-. The data are presented as percent cover with the variance presented as the 95% confidence interval about the mean; data presented without variance are visual estimates. Depth: 9.1 m Depth: 16.8 m Canopy species Agarum (or July 71 ily) removal April 72 Agarum and Laminaria removal July 71 April 72 Agarum (on July 71 ly) removal April 72 Agarum and Laminaria removal July 71 April 72 Laminaria Agarum Red algal turf Alaria 100 ±0 165.3 ± 23.4 11.5 ± 12.9 0 100 ±0 11.5 ± 10.2 8.4 ± 10.7 0 1100 ±0 0 145.5 ±16.1 25.8 ±11.9 7.0 ± 4.9 49.2 ±14.0 0 100 ±0 100 ±0 185.2 ± 33.4 2.1 ± 5.6 0 100 ±0 17.5 ±7.1 0 0 i100±0 0 177.4 ± 12.7 11.5 ± 4.0 1.2 ± 4.0 37.5 ±10.2 0 100 ±0 iSignifies that the canopy was experimentally removed. 235 FISHERY BULLETIN: VOL. 73, NO. 2 DISCUSSION The pattern emerging from these and other (McLean 1962; Lowry and Pearse 1973; Estes and Palmisano 1974) studies of sea otter-dominated communities is that by consuming the populations of invertebrate herbivores, the sea otter has an extremely important role in maintaining the structure of shallow algal communities. In this study, high densities of sea urchins are found below 18-20 m, suggesting that this depth is the lower limit of effective sea otter foraging in this area. It is interesting to note that this depth is much more shallow than the 30-fathom profile speculated by Kenyon (1969). In addition, this seems to be a much more shallow limit to efficient foraging than is exhibited by the California population of sea otters, as I have seen evidence of their foraging to at least 30 m in the Carmel Bay region. Strong competitive interactions between species of benthic algae appear well expressed in the shallow nearshore waters of the Aleutian Islands which have sea otters. The shallower (3-5 m) waters, subject to severe storm disturbance, are functionally dominated by Laminaria species. When the larger Laminaria spp. (L. groenlandica, L. dentigera, and L. yezoensis) are removed, either experimentally or by natural storm disturbance, their space is quickly utilized by Alaria fistulosa. In contrast, the rhizomelike holdfast with multiple meristems of L. longipes appears to be an effective adaptation to disturbance, as it allowed quick regrowth of stipes and fronds after their experimental removal. In deeper water (12-20 m), where there are many sea urchins, Agarum cribrosum is one of the dominant algal species. Agarum, however, loses in competition for light to solid canopies of Laminaria spp., which have erect stipes supporting their fronds above the nearly prostrate Agarum. When freed from Laminaria competition, Agarum significantly increases its cover and abundance. When both Laminaria and Agarum are removed, there is a bloom of red algal turf and of Alaria fistulosa. These tests of competition-based hypotheses are probably valid despite the various depth-related changes in the physical environment because each was compared to immediately adjacent controls. It is interesting to note that despite having po- tentially long-lived individuals and the competi- tively superior adaptation of a floating canopy. Alaria fistulosa behaves as a fugitive species with its densest distribution in the highly disturbed immediate offshore area, occurring farther offshore only in areas where two understory canopy levels are removed or by growing on Laminaria stipes. This is surprising because quite the opposite situation seems to exist in the southern California kelp community, where Macrocystis pyrifera forms a heavy surface canopy which may inhibit the growth of the un- derstory species (North and Shaef f er 1964; Dayton unpubl. data). Although Alaria was observed in depths of over 25 m, its lower distribution appears to be restricted primarily by sea urchin grazing. Other research (Estes and Palmisano 1974; Palmisano in prep.) contrasts the nearshore and intertidal communities of Amchitka with nearby otter-free islands and convincingly demonstrates the powerful role the sea otters have in structuring the nearshore community. This paper has experimentally demonstrated competitive trends between different canopy guilds in an algal com- munity which contains an unusually high number (four) of Laminaria species which have semirigid stipes. It is tempting to speculate an evolutionary hjrpothesis in which the sea otters reduce the her- bivore pressure and thus allow a competitive differentiation of niches of these large stipitate kelps. Such hypothetical evolutionary thought has the common and serious flaw of ignoring the roles of extinct species, many of which may have left large and important "vacant niches" (such as those left by the mammal extinctions of the late Pleis- tocene discussed in Martin and Wright 1967). This problem is particularly acute in the Bering Sea, as Steller in 1751 (reference in Card et al. 1972) reported the giant sea cow, Hydrodamalis gigas (Zimmermann 1780), eating algae in the nearshore and tidal beaches of the Komandorskiye Islands. The large populations reported by Steller and various Russian and German sailors of this huge (ca. 10 tons, Scheffer 1973) kelp-eating (Stejmeger 1936) sirenian surely had important consequences to the kelp populations that weaken any present day speculation of the evolutionary consequences of kelp competition. It may be reasonable, however, to pose the hypothesis that by consuming invertebrate herbivores, particularly sea urchins, the sea otter was indirectly responsible for the high productivity of large algae necessary to maintain the sea cow populations. Such an hypothesis is supported by the overlap of the otter 236 DAYTON: EXPERIMENTAL STUDIES OF ALGAL CANOPY and sea cow populations in the Pleistocene (Jones 1967; Kenyon 1969; and Gard et al. 1972). This relationship is nicely diagrammed in Scheffer's (1973) touching story of the last day of the sea cow. ACKNOWLEDGMENTS I thank P. A. Lebednik for identifying all the algae, for making all logistic and diving arrangements, extensive diving help, and con- tinuing editorial assistance. J. F. Palmisano and J. A. Estes also made helpful editorial suggestions and assisted the diving program. I thank V. Currie and B. McNames for expert typing assistance and R. B. Searles, D. Rivera, L. Dayton, M. Neushel, and J. Pearse for helpful suggestions which have improved the manuscript. I am very greatful to J. Isakson, C. O'Clair, C. Simenstad, G. Tutmark, and M. Wynne for logistic, diving, and taxonomic as- sistance. All logistic support was provided under AEC contract AT (26-l)-171 to Battelle Memorial Institute, Columbus Laboratories. Publication was supported by NSF grants GA-30877 and GV-32511, Sea Grant GH-112, and the Marine Life Research group at Scripps Institution of Oceanography. LITERATURE CITED Barr, L. 1971. Studies of populations of sea urchins, Strongylocen- trotus sp., in relation to underground nuclear testing at Amchitka Island, Alaska. Bioscience 21:614-617. Dayton, P. K. 1971. Competition, disturbance, and community organiza- tion: The provision and subsequent utilization of space in a rocky intertidal community. Ecol. Monogr. 41:351-389. 1973. Dispersion, dispersal, and persistence of the annual intertidal alga, Postelsia palmaeformis ruprecht. Ecology 54:433-438. 1975. Experimental evaluation of ecological dominance in a rocky intertidal algal community. Ecol. Monogr. 45. Ebert, E. E. 1968. A food habits study of the southern sea otter, Enhydra lutris nereis. Calif. Fish Game 54:33-42. Estes, J. A., and J. F. Palmisano. 1974. Sea otters: Their role in structuring neashore com- munities. Science (Wash., D.C.) 185:1058-1060. Estes, J. A., and N. S. Smith. 1973. Research on the sea otter, Amchitka Island, Alaska. Amchitka Bio-environmental Program, Final Rep., U.S. At. Energy Comm., 68 p. Card, L. M., Jr., G. E. Lewis, and F. C. Whitmore, Jr. 1972. Steller's sea cow in pleistocene interglacial beach deposits on Amchitka, Aleutian Islands. Geol. Soc. Am., Bull. 83, p. 867-870. Jones, R. E. 1967. A Hydrodamalis skull fragment from Monterey Bay, California. J. Mammal. 48:143-144. Kenyon, K. W. 1969. The sea otter in the eastern Pacific Ocean. Bur. Sports Fish. Wildl., North Am. Fauna, 68, 352 p. KiBBE, A. L. 1915. Some points in the structure of Alaria fistulosa. Puget Sound Mar. Stn. Publ. 1:43-57. LowRY, L. F., AND J. S. Pearse. 1973. Abalones and sea urchins in an area inhabited by sea otters. Mar. Biol. (Berl.) 23:213-219. Markham, J. W. 1968. Studies on the haptera of Laminaria sinclairii (Har- vey) Farlow, Anderson et Eaton. Syesis 1:125-131. 1972. Distribution and taxonomy of Laminaria sinclairii and L. longipes (Phaeophyceae, Laminariales). Phycologia 11:147-157. Martin, P. S., and H. E. Wright, Jr. 1967. Pleistocene extinctions: The search for a cause. Yale Univ. Press, New Haven, 455 p. McLean, J. H. 1962. Sublittoral ecology of kelp beds of the open coast area near Carmel, California. Biol. Bull. (Woods Hole) 122:95-114. North, W. J., and M. B. Shaeffer. 1964. An investigation of the effects of discharged wastes on kelp. Calif. State Water Qual. Control Board Publ. 26, 126 p. SCHEFFER, V. B. 1973. The last days of the sea cow. Smithsonian 3(10):64-67. Stejneger, L. 1936. Georg Wilhelm Steller, the pioneer of Alaskan natural history. Harvard Univ. Press, Camb., 623 p. Vadas, R. L. 1968. The ecology of Agarum and the kelp bed communi- ty. Ph.D. Thesis, Univ. Washington, Seattle, 306 p. Williamson, F. S. L., and W. B. Emison. 1969. Studies of the avifauna on Amchitka Island, Alaska. Battelle Mem. Inst. 171-25 Annu. Prog. Rep., June 1968-July 1969. 237 PRODUCTION OF TWO PLANKTONIC CARNIVORES (CHAETOGNATH AND CTENOPHORE) IN SOUTH FLORIDA INSHORE WATERS' M. R. Reeve and L. D. Bakers ABSTRACT Seasonal changes in biomass and production of two planktonic carnivores, Sagitfa hispida Conant and Mnemiopsis mccradyi Mayer, were followed in a subtropical inshore marine environment. Production was estimated as the product of mean daily biomass (calculated from the sampled biomass and computed mortality rates) and daily growth rate. The latter was determined from laboratory culture experiments at three temperatures. Seasonal fluctuations of ctenophore biomass and production were much greater than those of chaetognaths. Mean daily production in milligram carbon per square meter was 2.00 and 4.80 for Sagitta in Card Sound and Biscay ne Bay respectively, and 1.01 for Mnem iopsis in Biscayne Bay. The ctenophore was absent from Card Sound, possibly because the zooplankton standing crop was an order of magnitude lower than in Biscayne Bay (excluding ctenophores). Average produc- tion/biomass ratios were 0.31 for Sagitta and 0.12 for Mnemiopsis. Most production data for zooplankton are re- stricted to the herbivorous copepods in temperate and cold w^aters (see review of MuUin 1969; Mullin and Brooks 1970; Riley 1972). Estimates for car- nivores are very few and include Sagitta elegans (McLaren 1969; Zo 1969; Sameoto 1971) and Pleurobrachia bachei (Hirota 1974). As pointed out by Mullin (1969) there is no simple technique for the measurement of produc- tion of natural populations of zooplankton com- parable to the relatively routine '*C uptake method for the determination of primary production by phytoplankton. Unlike the phytoplankton, which share a common characteristic of a single trophic level, zooplankton extend over at least two trophic levels, and an individual species may vary its trophic status on the basis of food availability or life history stage. In addition, zooplankton range in size from 20ju.m or less to 20 cm or more and have widely differing growth and reproduction rates. Attempts to measure total zooplankton production have been made, especially where a single species dominates the population over a period (e.g., Gushing and Vucetic 1963) or where a single group (such as copepods) dominates and is treated as a unit (e.g., Riley 1972), and most recently by relat- ing respiration to temperature and body weight ' Contribution from the Rosenstiel School of Marine and At- mospheric Science, University of Miami, Miami, FL 33149. ' Division of Biology and Living Resources, Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL 33149. . y , , and applying these data to the plankton biomass of the Kuroshio (Ikeda and Motoda in press). The data reported below are based on the in- dividual species approach, using experimentally determined growth rates to compute production from environmental biomass estimates for two planktonic carnivores, widely separated phylogenetically but dependent upon the same source of food. STUDY SITE AND SAMPLING METHODS The study area consisted of Biscayne Bay and Card Sound which form part of an extensive sys- tem of shallow, warm, semiestuarine, and semienclosed interconnected water bodies typical of the coastal region of a large part of Florida. Zooplankton sampling programs were conducted at 4 stations on 28 dates throughout 1971 in Card Sound and at 11 stations on 26 dates from October 1970 to February 1972 in central Biscayne Bay. Detailed reports of these programs were given by Reeve and Cosper (1973) and Baker (1973), respec- tively. In both locations, surface tows were made with a metered, Vz-m mouth diameter net of 200-/im mesh. In addition, a similar net of 64-jU,m mesh was used in Card Sound. In Biscayne Bay, a 1-m, 705- ju,m mesh net with a 14-liter flexible, vinyl cod end was employed to collect ctenophores. It was not used routinely in Card Sound because ctenophores Manuscript accepted June 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 238 REEVE and BAKER: PRODUCTION OF A CHAETOGNATH AND CTENOPHORE were not encountered. Zooplankton were thus collected from both locations using two nets (which were towed simultaneously), one of which (the 200-ju,m mesh) was common to both locations. An extensive series of samples was collected in Card Sound to check on the adequacy of the 64- and 200-ju.m mesh Va-m mouth diameter nets in sampling the entire size range of the population of Sagitta hispida. In comparisons between a 64- and 35-ju.m mesh, the size-frequency distribution of the population was not significantly different. Ab- solute numbers often differed, but this was at- tributable to the rapid clogging of 35-/xm mesh, which rendered flowmeter readings unreliable, and was why this mesh was not used routinely. The 64-^m mesh net, which filtered less than 50% of the volume of water of the 200-/xm mesh in the same time, collected fewer of the larger size chaetog- naths than the 200-]u,m mesh, indicating that a greater proportion of the larger animals were avoiding the smaller meshed net. Comparative tests between the 200-iU,m V2-m diameter net and a 200-/im 1-m net (which filtered 3 times more water) did not indicate that the larger net caught either a larger absolute number, or a higher percentage, of the larger size classes per volume filtered. These data are available by writing to the first author. It appeared, therefore, that the two standard V2-m nets utilized in the sampling program quantita- tively collected the entire size range of this species in the surface water. Vertical distribution of S. hispida Conant in the 3-m water column was investigated on six dates during the year using both towed nets and a pump as described by Reeve and Cosper (1973). There was considerable variability in vertical distribu- tion between sampling dates, due in part to variability in incident radiation and water tur- bidity, but it was estimated that the numbers per cubic meter from surface hauls should be mul- tiplied by a factor of 1.54 to obtain a mean water column density per cubic meter in the 3-m deep water column. This factor was very close to the 1.45 calculated for the plankton as a whole, by Reeve and Cosper (1973). As noted previously (Reeve and Walter 1972), S. hispida has the ability to attach itself to substrates in the laboratory and lays its eggs on surfaces in clumps. It does not attach significantly until near maturity and even then, most of the population is usually to be found swimming in the water column in aquaria. We believe that the biomass estimates of our plankton samples were not biassed downwards due to this behavioral pattern, as eggs are usually laid at night while the plankton samples were taken dur- ing the day, and the vertical sample series gave no indication of a higher proportion of older animals nearer the bottom. On the other hand, comparisons of the size-frequency distribution of a population sampled with a towed net and with an Okelmann sledge lightly skimmed across the bottom, which is an effective means of sampling the benthic Spadella, usually yielded a few mature individuals in the larger size classes which were absent from the net. The sledge, however, only provided a qualitative sample and it was not possible to ad- just the biomass of Table 1 to take these few animals into account. Our biomass estimates are, therefore, slightly underestimated on this ac- count. No estimates of egg numbers were made, since Sagitta hispida does not deposit them in the water column, but attaches them to objects on the bottom. Ctenophores presented different sampling problems. Lobate ctenophores, such as the genus Mnemiopsis, tend to break up easily in nets and are rapidly disintegrated in the usual fixatives. Baker (1973) reported that transference of in- dividual, newly hatched larvae by pipette from one beaker to another would result in the disap- pearance without a trace of over 90% of these 200-/im diameter animals. It was futile, therefore, to attempt to assess the numbers of eggs or the smallest larvae from net tows, and probably some of size class A (0.8-4.4 mm) were also fragmented beyond recognition. Even so, the pattern of dis- tribution of biomass between the size classes (Ta- ble 1) suggests that the fraction contributed by the smallest unsampled or inadequately sampled members of the population is small. It may be presumed that animals in the larger size classes were not avoiding nets, since Mnemiopsis is a weak swimmer with no rapid escape behavior, and hence were sampled adequately. No feasible method was devised of making tows near the bot- tom of this shallow water column with a 1-m mouth diameter net, and pumps were impractical for sampling ctenophores. The only indication we have that Mnemiopsis does not exhibit any marked vertical layering are observations by scuba. Analysis of Samples The chaetognaths of the preserved samples (all of which belonged to the species S. hispida) were 239 FISHERY BULLETIN: VOL. 73, NO. 2 E 3 10 O S * O > 0) CO c « IB N c -^ "I ™ T3 0) ra A o o d o o o CM CM CO d CO m ■o- CD CO O C7) »- O) to 0> CD r~- ID CO o o o o •-' d d d d O CO CD t^ to O O O O O ''^. o d d d Tt '- CO CO CO CM CM CM d o o o V to o o CO CM O O IT) CO 1- r^ .- -r^ d CO in r^ to d d o o o d d d V V V O 00 U5 O) T- IT) f-- CD d> -r^ d d d to t>- •>a- ■ o r^ CM T- CO 1- lO lO lO T- T- ^ CM CO "- CO d d d d CO in o ■t iC u^ ^° '^ ^ ^ ^ CM CO ^ CO d d d d to o in m ^ ° CO CM CO -^ d d d in in o ■- in cj> tj) in in ^ CM CM CO ■^ <3i CO CM f- m in CO CM CO ■^ .- o o O CO d d d V o o o d d d V V V CO in in 1- in CO CM CO -^ CM ' d d d E ■- to r- ^ to T- C\J CM CO CM CM CO lull p a) m y T3 O (u m _ m O g « C U lu (O Q TO -^ O « "■ to ^*"^ t: o < Q D. to O ^ •- o Q. O O p 15 counted and measured in the laboratory. The Card Sound samples from the four stations were pooled for each net on each sampling date in proportion to the filtered volumes they represented (Reeve and Cosper 1973). Aliquots of each pooled sample were taken such that they contained between 50 and 100 organisms. The total body length of each animal was measured (see Reeve 1970). The entire sample was examined for mature animals. The lengths were tabulated in 1-mm preserved length size classes (see next section for conversion to live length). Since two values were obtained for each size class from the Card Sound samples (i.e., one for each mesh size) the larger number was taken as the correct one, on the assumption that the smaller value was due either to avoidance by larger animals of the 64-/Am mesh, or escape of the smaller animals through the 200-/i,m mesh. The pooled 200-/Am Biscayne Bay samples were treated similarly. The numbers of S. hispida in Biscayne Bay were estimated by adjusting the numbers in each size class in the 200-/im net to total number on the basis of ratios computed for the 64- and 200-/xm counts from Card Sound. Analysis of ctenophore samples from the 1-m net presented special difficulties, because there was no known satisfactory method of preservation of lobate ctenophores. Following Miller (1970) analysis was performed on deck immediately after recovery of the net (see Baker 1973). The contents of the cod end were emptied into a stack of wire sieves of arbitrarily chosen decreasing mesh sizes (25-, 12.5-, 6.25-, 3.0-, and 0.7-mm mesh openings) immersed in seawater. The ctenophores from each sieve, except the smallest, were transferred to a graduated cylinder and the total volume of or- ganisms retained by each sieve measured. The average volume per individual retained in each sieve was determined either by counting the total number of animals in each sieve or, in the case of the larger animals, by direct volume displacement of randomly selected individual ctenophores. It was impractical to follow this routine with the smallest animals (0.7-mm sieve) since their total volume was too small to be measured accurately. Instead, they were resuspended in seawater, transferred to plastic bags, and returned to the laboratory where they were counted. No attempt was made to assess the number and hence produc- tion of ctenophores smaller than 0.7 mm in diameter. 240 REEVE and BAKER: PRODUCTION OF A CHAETOGNATH AND CTENOPHORE Conversion of Raw Data to Other Units The shrinkage in length of S. hispida with For- mahn' preservation was estimated by measuring over 100 live animals from a freshly caught 200-ju.m mesh sample, and repeating this 10 and 420 days following preservation of that collection in a 5% formaldehyde solution buffered with meth- enamine, which was the standard preservative for all plankton samples. The degree of shrinkage was judged by the extent of the downward shift in the peak of the length/frequency histogram. Half the total shrinkage (12.5% of the original length) oc- curred within the first 10 days. Assuming a linear rate of shrinkage after day 10, and preservation time of the samples before analysis varying from 1 to 9 mo, the degree of shrinkage was computed to be 20% with a range of + 3.5%. This mean estimate was used to adjust size classes from preserved to live length. Live length was converted to dry weight using the relationship obtained from a linear regression analysis of more than 40 separate weight deter- minations of animals over their entire size range. Animals to be weighed were rinsed in isotonic ammonium formate and dried at 60°C. The ash- free (i.e., organic) dry weight was previously de- termined to be 90.7% of the dry weight (Reeve et al. 1970). The mean carbon and nitrogen content of S. hispida was determined by a Perkin-Elmer elemental analyzer to be 44.9% with a standard error of ± 1.0% and 11.9% + 0.2% of the ash-free dry weight from 23 separate estimations over its entire size range. The raw biomass units for ctenophores were obtained in terms of live volume. Over 100 separate determinations of animals over their entire size range were made for wet (drained), dry (at 60°C), and ash (at 500°C) weights. Live volume was approximately numerically equal to wet weight (1.000 ml = 0.958 +. 0.002 g standard error). Dry weight was 4.43% + 0.40% of wet weight and ash-free dry weight was 21.90% ± 0.15% of dry weight. Eighteen separate determinations of carbon and nitrogen content of Mnemiopsis mccradyi Mayer were made which yielded unusually low values 8.72% + 0.06% and 2.32% + 0.07% of the ash-free dry weight of carbon and nitrogen re- spectively. A value of 44.9% carbon was reported for Sagitta (above), and Curl (1962) quoted values for various planktonic crustaceans between 44 and 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 52%. Even his value for Mnemiopsis sp. was con- siderably higher at 20.6%. Hirota (1974) assumed a 50% carbon content of organic weight for Pleurobrachia bachei in his calculations, because analysis by wet combustion with acid dichromate was unsuccessful due to problems with chloride ion interference (J. Hirota, pers. commun.). We considered the possibility that our analyses were also yielding incorrect results and tested three possible sources of error: a) interference in the analysis by the unusually large amount of inorganic salts present in the ctenophore tissue, b) errors of dry weight determination, and c) errors of ash weight determination. Mixtures of bovine serum albumin (5-15%) and sodium chloride did not reduce the theoretical yield of carbon when com- busted in the elemental analyzer. Since, however, the dried ctenophore material was a more intima- tely bound complex of organic and inorganic sub- stances, which might be more resistant to complete combustion, potassium persulfate was added to promote complete oxidization (see Strickland and Parsons 1968). No increase in car- bon yield was achieved by this method. The reliability of dry and ash weight determinations affects the reliability of the carbon value since the numbers so obtained are used in its computation. The possibility of any significant loss of organic matter during drying at 60°C was checked by performing carbon analyses on freeze-dried material. The previously derived mean value remained unchanged. Finally, ash weights were determined at a temperature 100°C lower than previously. Slightly higher ash weights resulted, which in turn slightly increased the computed carbon level to 10.3% of the ash-free dry weight. Since any significant source of error in this deter- mination has so far eluded us, we report produc- tion values below for ctenophores and chaetog- naths in terms of ash-free dry weight for direct comparison and in terms of the analyzed carbon. Mullin (pers. commun.), on the basis of un- published observations, suggested that the weight lost on ashing may be largely "bound" water, and that in Pleurobrachia bachei, at least, only about 12% of the ash-free dry weight is organic matter. This suggests that comparisons based on carbon content are more valid than those based on "or- ganic" or ash free-dry weight. Growth Rates Growth rates of populations of the ctenophore 241 and chaetognath were determined in the labora- tory using larvae hatched from wild adults ac- cording to methods detailed by Reeve and Walter (1972) for S. hispida and Baker and Reeve (1974) for M. mccradyi. Three separate popula- tions of the chaetognath and two of the ctenophore were grown at each of three temperatures (21°, 26°, 31°C), which corresponded to the mean monthly minimum, annual mean, and mean monthly maximum temperatures (to the nearest 1°C) off the laboratory dock in Biscayne Bay over 11 yr (unpubl. records). Food was provided in the form of naturally occurring zooplankton of suit- able size (see previously cited information on cul- ture technique), consisting mostly of the copepods Acartia tonsa and Paracalanus parvus, main- tained at a level such that no more than 50% were grazed down over 24 h. Growth rates were measured as length increase to avoid sacrificing any members of the popula- tions and the data converted to ash-free dry weight as previously described. Total length of Mnemiopsis was measured from the aboral to the oral pole (or tip of the oral lobes in adults) as described in detail by Baker (1973). FISHERY BULLETIN: VOL. 73, NO. 2 Production Calculation The method of calculating production was that employed by Mullin and Brooks (1970) and Hirota (1974), where for each size class an exponential coefficient of daily growth (G) and mortality (M) is obtained from laboratory growth rate and field size-frequency data. The growth coefficients were computed from the slope of the line relating the logarithm of increase in ash-free dry weight and age (Figure 1) follow- ing Crisp (1971). Data from each rearing experiment were combined for each species at the specified temperature. For S. hispida, the semilogarithmic relationship is linear over most of its size range until growth levels off at maturity (Reeve and Walter 1972). The termination of the linear (i.e., constant exponential growth) phase was arbitrarily set at 20, 25, and 30 days at 31°, 26°, and 21° C respectively, and the slope of the line calculated by linear regression analysis. At each temperature, slopes at two points beyond the linear phase were required, and these were derived by extrapolation on the basis of the remaining data points and other (unpubl.) data on lengths of I r 01 >- Q 001 UJ UJ tr I to < 0 001 10 20 30 AGE IN DAYS 40 50 100 10 ct Q X < 01 0 01 MNEMIOPSIS 10 20 30 AGE IN DAYS 40 50 Figure L— Growth rate of Sagitta and Mnemiopsis at three different temperatures. 242 REEVE and BAKER: PRODUCTION OF A CHAETOGNATH AND CTENOPHORE animals older than those surviving in these experiments. The potential errors in such a procedure are minimal, because the coefficients are tending towards zero and the biomass involved in the two largest size classes is only a small percen- tage of the total. The Mnemiopsis growth curves were treated differently because their slopes decreased progressively with age. In order to facilitate com- putation of the required slopes, the curves were divided into segments, the junctions of which were assigned by visual inspection to be at 3- and 50-mg ash-free dry weight. The slopes of the individual segments A, B, and C were individually calculated from the population mean points within them by linear regression analysis. Unlike S. hispida, where growth rate is proportional to temperature between 21° and 31° C, M. mccradyi grows faster at 26°C than at either end of the range. Survival was poor at 31°C, populations dying out by the 25th day. Since no points exist from which to compute a slope for segment C at 31°C, it was taken to be the same as that for the 21°C experiments, since segments A and B at the two temperatures are almost identical. Sampling dates were divided into three groups on the basis of the proximity of the ambient water temperature to 21°, 26°, and 31°C so that growth coefficients derived for these temperatures could be applied to the standing stock data. Similarly, mean mortality coefficients were derived for the three temperature ranges by averaging the numbers of animals in each size class over the sampling dates in each temperature range. These mean numbers were used to obtain mean ratios of Y/X (as did MuUin and Brooks 1970) where X and Y are the numbers of the earlier and later of two successive size classes. This ratio, and the duration of development in each of the two successive size classes, enables calculation of the exponential coefficient of daily mortality between the two size classes using computer-generated tables. We recognize that this procedure is an approximation which probably oversimplifies actual conditions by making unproven assumptions regarding con- stancy of mortality rate with time and between adjacent size classes, yielding a single value for m rather than a measure of its possible range (see Fager 1973). The duration of development in each size class at each temperature range was estimated from the arbitrarily defined limits of each size class and the laboratory growth rate data. Net production of a size class on a given sampling date, taking into account animals which die before the end of the day, is the product of the mean biomass and the daily exponential coefficient of growth for that temperature range. The day is assumed to start ^t_the time of sampling, and the mean biomass {WN) of that size class over the subsequent 24 h is obtained by application of the relationship given by Mullin and Brooks (1970) which utilizes the initial biomass, growth, and mortality coefficients. The initial biomass {WN in ash-free dry weight) is the product of sampled numbers (AO and mean ash-free dry weight {W) of an individual organism of that size class. Summing the production values for each size class provides an estimate of the total net production of the population on that day. No attempt was made to estimate egg production in either species. Net production was determined for chaetog- naths of the Card Sound population only; values quoted below for the Biscayne Bay population are estimated by applying the mean population production /biomass ratio for Card Sound to es- timated total biomass in Biscayne Bay. An es- timate of annual production is obtained by taking each sampling date as the midpoint of each sampling period, summing the product of daily production and number of days in that sampling period, and summing the total production for each sampling period and adjusting for 365 days. In the ctenophore population, which was sampled for 17 mo, and passed through two biomass peaks which Baker (1973) considered to be an annual winter event (Table 2), two values were computed (see Table 3), one for 365 days from the beginning and one for 365 days up to the end of the sampling program. Results and Discussion Seasonal Changes Summaries of the population dynamics and production data are contained in Tables 1 and 2, computed as detailed above from tabulations by sampling date and size class. Figure 1 contains the laboratory growth rate data. The standing stock and production data are summarized in Tables 1 and 2, and are derived from the Card Sound population of Sagitta and the Biscayne Bay population of Mnemiopsis, since these populations had been the most effectively sampled. For each size class (Table 1) averaged over the entire 243 FISHERY BULLETIN: VOL. 73, NO. 2 Table 2.-Suinmary of biomass and production data by sampling data averaged over all size classes. Sagi if/a Mnemiopsis Biomass Production Biomass Production mg ash-free dry mg ash-free dry mg ash-free dry mg ash-free dry Date vA/m' wt/mVday P/B Date wt/m' wt/m'/day P/B 1/06/71 3.82 095 .25 10/12/70 18 02 264 15 1/23/71 2.86 0.62 .22 10/23/70 135.59 12.12 .09 2/06/71 2.94 0.74 25 11/20/70 314.31 27.80 .09 2/16/71 7.09 1.44 .20 12/15/70 81 97 692 .09 3/05/71 3.68 0.94 26 1/15/71 18.31 2.09 .11 3/19/71 1.56 0.40 26 2/15/71 2356 2.35 .10 4/02/71 4.89 1.28 .26 3/12/71 5.46 067 .12 4/16/71 4.10 0.96 .23 4/08/71 20.34 1.63 .08 4/30/71 11.98 3.11 26 5/07/71 17.77 1.49 .08 5/14/71 5.53 1.77 .32 6/03/71 0.22 0.04 .18 5/28/71 2.50 0.69 .28 7/01/71 0.87 0.12 .14 6/11/71 4.46 1.54 ,35 7/26/71 017 0.01 .06 6/25/71 1.24 0.46 .37 8/25/71 — — — 7/09/71 1.58 0.60 38 9/17/71 0.06 <0,01 .06 7/23/71 0.10 0.04 40 9/30/71 0.58 0.09 .16 8/06/71 0.56 0.23 .41 10/14/71 1.29 0.14 .11 8/20/71 0.25 0.10 .40 10/28/71 3.27 0.64 .20 9/03/71 0.48 019 40 11/12/71 8.86 1.30 .15 9/09/71 0.62 0.24 39 11/24/71 6.80 1,07 .16 9/14/71 6.03 2.26 .37 12/02/71 56,55 6 14 .11 9/21/71 2.00 0.66 .33 12/20/71 20.89 2.16 .10 9/28/71 3.47 1.24 .36 1/06/72 200.35 15.66 .08 10/14/71 4.85 1.67 .34 1/21/72 103.59 7.38 .07 10/26/71 6.06 2.23 .37 2/03/72 769 0.75 .10 11/09/71 0.99 0.37 .37 11/23/71 3.93 0.97 .25 12/07/71 5.21 1.27 .23 12/15/71 1.16 0.30 .26 sampling period, the mean numbers, live length (or volume for Mnemiopsis), ash-free dry w^eight per organism, and ash-free dry v^^eight per size class are tabulated. The mean net daily production of each size class (per cubic meter) averaged over the entire period using the information on daily rates of growth and mortality and the average duration of each size class (over the three temperatures) is also provided. Seasonal changes in production reflected those of biomass generally as indicated in the produc- tion/biomass ratios (Table 2), which varied between 0.20 and 0.41 (mean, 0.31) for Sagitta and 0.059 and 0.20 (mean, 0.12) for Mnemiopsis. The ratios were highest in Sagitta in the summer when growth rates were maximum, but biomass and production was at its lowest. In Mnemiopsis, which also exhibited minimum summer biomass and production levels, the ratio tended to be low rela- tively, as was growth rate. This summer low point of biomass and production is a confirmation of the experience of some nine seasons of observation by the first author and is characteristic of the 200-;u,m net plankton of Card Sound and Biscayne Bay as a whole (for a discussion of which, see reviews of Reeve and Cosper 1973 and Reeve in press). Throughout the rest of the year the chaetognath biomass of Card Sound and Biscayne Bay fluc- tuated much less widely than that of the ctenophores in Biscayne Bay. The biomass of Sagitta ranged (excluding July-September) between 1- and 12-mg ash-free dry wt/m' in Card Sound and an estimated 2- and 20-mg ash-free dry weight in Biscayne Bay, whereas for ctenophores the range was 0.2 to 314 mg/m\ The mean annual biomass of Sagitta in Card Sound and Biscayne Bay was 3.36- and 8.04-mg ash-free dry wt/m' and for Mnemiopsis in Biscayne Bay was 25.2- to 42.5- mg ash-free dry wt/m' (reckoning 12 mo from the date of the first sample or 12 mo prior to the date of the last sample). The range of net daily production rate in terms of ash-free dry weight for Sagitta at the surface from Card Sound and Mnemiopsis from Biscayne Bay was 0.04 to 3.1 and < 0.01 to 27.8/m', respec- tively. Table 3 contains production estimates on an annual basis computed for surface and average water column (for Sagitta only) as ash-free dry weight and carbon per cubic meter. For Mnemiopsis, carbon production is computed both on the basis of the carbon content of Sagitta and the experimentally determined carbon content for Mnemiopsis. Daily production estimates per cubic meter and per square meter are also computed in terms of experimentally determined carbon con- tent. As noted above, the two values in each case for Mnemiopsis incorporate successive annual production peaks. 244 REEVE and BAKER: PRODUCTION OF A CHAETOGNATH AND CTENOPHORE Table 3.-Mean annual and daily production of the Sagitta populations of Card Sound and Biscayne Bay and Mnemiopsis population of Biscayne Bay. Annual production Daily pn mg ash-free dry wt/m' mgC/m- Surface Average water Carbon column 44.9% Carbon by analysis aduction mgC/m' mgC/m' Sagitta Card Sound Biscayne Bay Mnemiopsis Biscayne Bay 357 855 695/1,409 542 1,200 244 584 312/633 244 584 60.6/123 0.67 1.60 0.17/0.34 2.00 4.80 0.50/1.01 Details of methods for the calculation of production for populations with continuous breed- ing occur in Winberg (1971) and Crisp (1971). They are essentially similar to the method used here and by Mullin and Brooks (1970) and Hirota (1974) ex- cept that no adjustment is made to the sampled biomass (PFAO to compute the mean biomass {WN) during the 24 h immediately following the taking of the sample. This additional step, which we also performed, requires considerable extra effort (depending on the number of size classes and sampling dates involved) as well as access to com- puter services. In these warm waters, however, where growth and mortality rates may be less variable than in regions of more pronounced seasonality, the increase in W tends to cancel out the decrease in A^, the difference between WN and WN for Sagitta and Mnemiopsis being less than 10% (93 and 108%, respectively). Mortality coefficients tended to increase progressively with age in Sagitta and with increasing temperature. These environmental ob- servations correspond to the conditions of labora- tory cultures with respect to temperature, but in cultures young animals tend to die off more rapidly than juveniles and immature animals (Reeve and Walter 1972). A variety of interacting factors, including differences in predation pres- sure and food adequacy, may be responsible. In the ctenophore population the pattern of mortality is uniformally low except in size class B which corresponds to the time of change from ten- taculate larva to lobate adult. The unmeasurable mortality of size group A can be partly attributed to sampling inefficiency, though this was shown not to be the case for Sagitta (see above). Problems of Measuring Growth Rate In animals such as copepods, with life history stages marked by recognizable and abrupt changes (i.e., molts), division of the cycle into parts may be accomplished on the basis of some biologically meaningful criteria. Both chaetog- naths and ctenophores exhibit more gradual transformation from newly hatched larva to ma- ture adult, and size class separation is based on arbitrary limitations such as preserved length or sieve size. The only real validity of the particular size classes used here is that they represent a progression from the youngest to the oldest animals. Factors such as variability of size of animals of the same age at different temperatures and imprecision of raw measurements (larger ctenophores may pass through a mesh slightly smaller than their diameter by their own weight deforming their shape) tend to blur the sharpness of the line separating one size class from the next. The arbitrary choice of size classes resulted in large variations in the durations of development of each size class. In Sagitta the mean duration (i.e., averaged over the three experimental temperatures) of the initial size class was 12 days, shortening to 2 days as length increased rapidly, and increasing to 9 in the last size class as a final length was approached in the adult. In Mnemiop- sis size class durations proved to be even more erratic (see Table 1). On the basis of the definitions used by Reeve (1970) for S. hispida, the larval, juvenile, imma- ture, and mature stages correspond approximately to size classes A and B, C and D, E and F, and G and H, respectively. For M. mccradyi the ten- taculate larva extends to size class C and the first eggs are also produced by size class C animals (29 mm and larger). The most satisfactory way to determine growth and mortality in a population is to follow the increase in size and decrease in numbers of a cohort of the population over successive sampling dates by inspection of size-frequency histograms (Winberg 1971; Crisp 1971). In warmer waters, although biomass may fluctuate widely, breeding 245 FISHERY BULLETIN: VOL. 73, NO. 2 tends to extend over most or all of the year and distinct cohorts can rarely be identified. Grov^^th rate, therefore, was measured in the laboratory, and in as large a volume as practical (30-70 liters). No attempt was made to simulate natural food levels. There were various reasons for this. Mean annual zooplankton concentrations of the 200-yu,m mesh, which is the food source of older Sagitta and Mnemiopsis (Reeve and Walter 1972; Baker 1973), were of the order of magnitude of 1 organism/liter, an impractically low concentra- tion to work with in these volumes. It is certain that any environmental concentration estimated from a net tow is an average of several small-scale patches of higher and lower density. We have some information from direct observation by scuba (unpubl. data) that patch densities at least an order of magnitude greater occur, as well as in- formation (also unpubl. data) that both Sagitta and Mnemiopsis can ingest food several times faster following a period of starvation than they do under conditions of a constant supply of food. Sagitta is capable, under certain conditions, of in- gesting within 1 to 2 min all the food it consumes in 24 h under conditions of continuous abundant food supply. Despite the fact that feeding habits and en- vironmental food concentrations are poorly un- derstood at present, it is clear that for carnivorous zooplankton, at least, maintaining a continuous supply of food at mean environmental concentra- tions in small-scale experimental conditions, would be as artificial as maintaining a continuous abundant supply, even though there must ob- viously be a relationship between total food supply and production in the environment. The latter method does provide a standard (i.e., maximum) growth rate. When better data become available on the interrelationships of feeding, food supply, and growth rate, the production estimates com- puted on that basis can be revised downward. At present, there is little information available to even guess to what extent these growth rates and hence production estimates are overestimations. Hirota (1974) reported surprisingly little difference in growth rates of Pleurobrachia in experiments at food concentrations ranging between 1 and 350 )u.gC /liter, but pointed out that in the 70-m' tank in which the low food concentra- tion occurred, food organisms were not uniformly distributed because some species were concen- trated at the surface during the day. In Card Sound and Biscayne Bay the mean annual con- centration of food from the 200-/xm net (the size range fed to adult Sagitta and postlarval Mnemiopsis in our experiments) was 0.8 and 8.1 /AgC /liter. Taking into account all organisms down to a 20-)u,m retaining mesh those figures would be increased by a factor of 5 (Reeve and Cosper 1973). Production Comparisons Sameoto (1971) obtained a value for the net production of S. elegans in Nova Scotia waters (ranging in temperature approximately from 0.5° to 14°C) of 200 mgC/m^ per yr in a 50-m water column, and McLaren (1969) reported a similar range of values for this species from Ogac Lake on Baffin Island (49-196 and 318). Those authors es- timated production/biomass ratios between 1.0 and 2.1 on an annual basis. These figures compare with annual net production of S. hispida in Card Sound and estimated in Biscayne Bay of 730 and 1,750 mgC/m" per yr and production/biomass ratio of 109 on an annual basis. With a mean annual biomass two orders of magnitude lower, therefore, S. hispida in Card Sound exceeds the net produc- tion of 5. elegans in St. Margaret's Bay, Nova Scotia by virtue of its rapid growth rate and short generation time. The disparity would be even greater on a cubic meter basis because Card Sound is comparatively shallow. Hirota (1974) quoted a value for net annual production of the ctenophore Pleurobrachia bachei in waters off California (ranging in temperature approximately from 12.5° to 20°C) of 5,415 mg ash-free dry weight/m- per yr, and a daily production/biomass ratio of 0.02. These figures compare with an annual net production of M. mccradyi in Biscayne Bay of 2,086 to 4,227 mg ash-free dry weight/m- per yr and a production/ biomass ratio of 0.12. As in the previous com- parison, annual production of different species in different regions is surprisingly similar on a water column (square meter) basis. The growth rate of M. mccradyi, however, is some 5 times faster, and its production is supported by a water column depth of 3 m rather than in excess of 40 m in the case of Pleurobrachia bachei. The 10-fold difference in the mean annual standing stock of 200-ju,m mesh zooplankton between Card Sound and Biscayne Bay (and in phytoplankton pigment) is probably a reflection of the poor water exchange and limited land drainage into Card Sound as compared with Bis- J cayne Bay (Reeve and Cosper 1973). These 246 REEVE and BAKER: PRODUCTION OF A CHAETOGNATH AND CTENOPHORE differences in plankton biomass are accompanied by differences in biomass for both Sagitta and Mnemiopsis. In surface net tows from the 200-yum mesh, the biomass of the chaetognath in Biscayne Bay is 2.4 times that in Card Sound. The ctenophore is totally absent from Card Sound (ex- cept for rare isolated individuals). Baker (1973), relating stations with low plankton standing stock to low ctenophore levels in Biscayne Bay, sug- gested that the Card Sound plankton could not support a ctenophore population. Since the waters of Card Sound are contiguous with those of Biscayne Bay to the north, and neri- tic waters to the east, where ctenophores are often abundant, the phenomenon of their exclusion from Card Sound can hardly be a physical one. A pos- sibility is that chaetognaths are more efficient in collecting food at lower densities than are ctenophores. It is of interest that the seasonal variations of biomass and production of the ctenophore popula- tions both in Biscayne Bay and off California are extreme, to the extent that in both cases the months of peak production account for about two- thirds of the annual total. In the case of S. hispida this value is about one-fifth. There is probably some correlation between this extreme population instability of M. mccradyi and the suggestion above that its absence from Card Sound is related to its inefficiency in collecting food at low concen- trations compared to S. hispida. The dry weight of other zooplankton from the 200-ju.m mesh net in Card Sound (Reeve and Cosper 1973) never ex- ceeds the minimum value in central Biscayne Bay (Baker 1973). It is possible to get a rough estimate of the relationship between production of S. hispida and M. mccradyi and the rest of the zooplankton by utilizing the standing stock data for that period in the two reports referred to immediately above. The mean annual dry weight of zooplankton (excluding ctenophores and corrected for detritus) was 2.02 and 5.28 mg/m' in the 200- and 64-/xm mesh net respectively in Card Sound and 19.8 mg/m^ in the 200-ju,m mesh net in central Biscayne Bay. Assuming the ratio between 64- and 200- mesh plankton in Card Sound is applicable to Bis- cayne Bay, and the ash- free dry weight is the same percentage of dry weight as determined for S. hispida, the mean annual ash-free dry weight in Card Sound and central Biscayne Bay was 6.62 and 64.9 mg/m^ respectively. Since it appears that even the youngest larvae of Mnemiopsis and Sagitta do not utilize food organisms much smaller than those retained by the 64-/im mesh, and since neither carnivore appears to be able to utilize other sources of potential food such as detritus or phytoplankton (Reeve and Walter 1972; Baker and Reeve 1974), the plankton biomass quoted above is the only source of nutrition for these carnivores. If a production /biomass ratio the same as that de- termined for S. hispida is applied to these biomass figures, the net production available to these car- nivores is 2.05- and 20.1-mg ash-free dry weight/m' per day. Since these figures are for surface waters, they may be related to the equivalent values for Sagitta and Mnemiopsis derived earlier. The daily net production of S. his- pida in Card Sound and Biscayne Bay is then 47.7 and 11.7% of the production of potential food in those areas. For M. mccradyi in Biscayne Bay it was 9.5 or 19.2% depending on which production peak was included (see above). The total percen- tage for the two species is then 47.7 for Card Sound and 21.2 or 30.9 for central Biscayne Bay. If the ratio of production to food ingested is taken to be 50% on the basis that immature animals are re- sponsible for most of the production, and would have higher growth efficiencies than the 30-40% range for adults quoted by Reeve (1972), then the chaetognaths in Card Sound appear to utilize all the rest of the zooplankton above 64 ^m. For Bis- cayne Bay, the chaetognaths and ctenophores together utilize between 40 and 60% of the avail- able food. As explained earlier, these are overes- timated because the growth rates were maximum growth rates, but they do support the contention that there is little potential food reserve in Card Sound for other carnivores, and that Sagitta is more efficient in competing for the available sup- ply. This is in agreement with the fact that in Card Sound, its population was as high as 42% of that in central Biscayne Bay, while for larger decapod larvae, fish larvae, and ctenophores (the other major first-order plankton carnivores) the values were approximately 25, 25 (see Reeve in press), and 0%. ACKNOWLEDGMENTS We are grateful to Michael M. Mullin and Jed Hirota for reading this manuscript and for the support of National Science Foundation Grant GA-28522X (Biological Oceanography). 247 FISHERY BULLETIN: VOL. 73, NO. 2 LITERATURE CITED Baker, L. D. 1973. The ecology of the ctenophore Mnemiopsis mccradyi Mayer, in Biscayne Bay, Florida. Univ. Miami, Rosen- stiel School Mar. Atmos. Sci., Tech. Rep. UM- RSMAS-73016. Baker, L. D., and M. R. Reeve. 1974. Laboratory culture of a lobate ctenophore with notes on feeding and fecundity. Mar. Biol. (Berl.) 26:57-62. Crisp, D. J. 1971. Energy flow measurements. In N. A. Holme and A. D. Mclntyre (editors), Methods for the study of marine benthos, IBP (Int. Biol. Program.) Handbook 16, p. 197- 279. International Biological Programme, bond. Curl, H. C. 1962. Analyses of carbon in marine plankton organisms. J. Mar. Res. 20:181-188. CUSHING, D. H., AND T. VUCETIC. 1963. Studies on a Calanus patch. III. The quantity of food eaten by Calanus finmarchicus. J. Mar. Biol. Assoc. U.K. 43:349-371. Fager, E. W. 1973. Estimation of mortality coefficients from field samples of zooplankton. Limnol. Oceanogr. 18:297-300. HiROTA, J. 1974. Quantitative natural history of Pleurohranchia bachei in La Jolla Bight. Fish. Bull, U.S. 72:295-335. IKEDA, T., AND S. MOTODA. In press. An approach to the estimation of zooplankton production in the Kuroshio and adjacent region. Proc. Marine Science Special Symposium, Hong Kong, Dec. 1973. McLaren, I. A. 1969. Population and production ecology of zooplankton in Ogac Lake, a landlocked fiord on Baffin Island. J. Fish. Res. Board Can. 26:1485-1559. Miller, R. J. 1970. Distribution and energetics of an estuarine population of the ctenophore, Mnemiopsis leidyi. Ph.D. Thesis, North Carolina State Univ., Raleigh, 85 p. MULLIN, M. M. 1969. Production of zooplankton in the ocean: The present status and problems. Oceanogr. Mar. Biol., Annu. Rev. 7:293-314. MuLLiN, M. M., AND E. R. Brooks. 1970. The ecology of the plankton off La Jolla, California, in the period April through September, 1967. Part VII. Production of the planktonic copepod, Calanus hel- golandicus. Bull. Scripps Inst. Oceanogr., Univ. Calif. 17:89-103. Reeve, M. R. 1970. Complete cycle of development of a pelagic chaetog- nath in culture. Nature (Lond.) 227:381. 1972. Pelagic invertebrates. Int. Encycl. Food Nutr. 18:587-612. Reeve, M. R. In press. The ecological significance of zooplankton in the shallow subtropical waters of South Florida. Proc. 2nd Int. Conf. Adv. Estuarine Res. Reeve, M. R., and E. Cosper. 1973. The plankton and other seston in Card Sound, South Florida, in 1971. Univ. Miami, Rosenstiel School Mar. Atmos. Sci., Tech. Rep. UM-RSMAS-73007. Reeve, M. R., J. E. G. Raymont, and J. K. B. Raymont. 1970. Seasonal biochemical composition and energy sources of Sagitta hispida. Mar. Biol. (Berl.) 6:357-364. Reeve, M. R., and M. A. Walter. 1972. Conditions of culture, food-size selection and the ef- fects of temperature and salinity on growth rate and generation time in Sagitta hispida Conant. J. Exp. Mar. Biol. Ecol. 9:191-200. Riley, G. A. 1972. Patterns of production in marine ecosystems. In J. A. Wiens (editor). Ecosystem structure and function. Proceedings of the 3rd Annual Biology CoUociuium. Oregon State Univ. Press, Corvallis. Sameoto, D. D. 1971. Life history, ecological production, and an empirical mathematical model of the population of Sagitta elegans in St. Margaret's Bay, Nova Scotia. J. Fish. Res. Board Can. 28:971-985. Strickland, J. D. H., and T. R. Parsons. 1968. A practical handbook of seawater analysis. Fish. Res. Board Can., Bull. 167, 311 p. Winberg, G. G. 1971. Methods for the estimation of production of aquatic animals. Academic Press, Lond., 175 p. Zo, Z. 1969. Observations on the natural population of Sagitta elegans in Bedford Basin, Nova Scotia. M. A. Thesis, Dalhousie Univ., Halifax, N.S. 248 EFFECTS OF ACCLIMATION ON THE TEMPERATURE AND SALINITY TOLERANCE OF THE YOLK-SAC LARVAE OF BAIRDIELLA ICISTIA (PISCES: SCIAENIDAE)i Robert C. May' ABSTRACT Eggs of the bairdiella, Bairdiella icistia, were fertilized and incubated in various combinations of temperature and salinity, and the salinity and upper thermal tolerances of the yolk-sac larvae were determined. The upper thermal tolerance was enhanced by acclimation to high temperatures and low salinities. Acclimation to low salinities enhanced the lower salinity tolerance of larvae at 24 h after exposure to test conditions, but an acclimation effect on the upper salinity tolerance was not apparent until 48 h after exposure. Yolk-sac bairdiella larvae are more tolerant than the embryonic stages and less tolerant than adults to extremes of temperature and salinity. Techniques for inducing gonadal maturation and spawning under laboratory conditions are well developed for the bairdiella, Bairdiella icistia (Jordan and Gilbert), a sciaenid fish native to the Gulf of California and now present in the Salton Sea (Haydock 1971; May 1975). Hence bairdiella eggs and larvae are extremely favorable material for studying various facets of early development in a marine fish, and detailed information on the effects of temperature and salinity on fertiliza- tion, embryonic development, and hatching in this species has already been presented (May 1975). The present paper is concerned with the effects of acclimation on the tolerance of yolk-sac bairdiella larvae to temperature and salinity. Acclimation has been defined as "the process of bringing the animal to a steady state by setting one or more of the conditions to which it is exposed for an appropriate time before a given test (Fry 1971:14)." In the case of yolk-sac larvae of tropical fish species which develop very rapidly, the term acclimation has a somewhat special meaning, since it necessarily refers to the conditions obtaining during embryonic development. Virtually no studies of acclimation in this context have here- tofore been published. Although salinity has been shown to affect the upper thermal tolerance of 'Based on a portion of a dissertation submitted in partial sat- isfaction of the requirements for the Ph.D. degree at the University of California at San Diego, Scripps Institution of Oceanography. ^Hawaii Institute of Marine Biology, P.O. Box 1346, Kaneohe, HI 96744. adult fish (e.g., Garside and Jordan 1968), no com- parable work has been reported for fish larvae. This paper investigates the upper thermal tolerance of newly hatched bairdiella larvae and the modifying influence of acclimation, i.e., the influence of temperature and salinity during embryonic development. Since there is little likelihood that bairdiella larvae would encounter lower lethal temperatures in nature (May 1975), their lower thermal tolerance is not considered here. In addition to upper thermal tolerance, the upper and lower salinity tolerance of larval bair- diella and the effect of the acclimation salinity are also considered in this paper. This information, together with results on embryonic tolerances described earlier (May 1975) and available infor- mation concerning adult tolerances, should lead to a conclusion as to which stage in the life history of bairdiella is the most sensitive to temperature and salinity. METHODS General Bairdiella eggs were obtained from fish which had been held in normal seawater i^2t^/oo) and induced to mature and spawn in the laboratory, as described previously (May 1975). Eggs were ar- tifically fertilized at specified temperatures (within ±0.2°C) and salinities (±0.5''/oo) and maintained under the same conditions until hatching in specially designed incubators (May Manuscript accepted June 1974 FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 249 FISHERY BULLETIN: VOL. 73, NO. 2 1975). These conditions (which remained constant from fertilization to hatching, plus the period of time between hatching and transfer to the test conditions) constituted the conditions of acclima- tion. Larvae were not fed during the experiments. Test salinities were prepared by dilution with deionized water from a stock solution of 60%, which had been made by adding artificial sea salts to seawater (May 1975). Upper Thermal Tolerance Larvae were acclimated to temperatures of 21°, 24°, 27°, and 30°C, and to salinities of from 15 to 45^/00 (Table 1), covering the ranges of these two factors within which successful embryonic development can take place (May 1975). Since developmental rates were more rapid at the higher temperatures (May 1975), the period of acclima- tion (fertilization to transfer to test vials) was shorter at the higher acclimation temperatures. The median tolerance limit (TLm)^ of yolk-sac larvae to high temperatures was determined by the method of Doudoroff (1942). Larvae were transferred directly from the acclimation condi- tions to a series of 25-ml capped vials maintained in the dark at a series of high temperatures in a thermal gradient block (Thomas et al. 1963) within 5 to 10 h after hatching, at which time they were 1.8 to 2.0 mm in length. The highest test tempera- ture was 36°C, and between five and eight test temperatures, 1.5°C apart, were used depending on the acclimation temperature; the salinities in the test vials were the same as the acclimation salinities. Approximately 10 larvae were placed in each vial, and the test temperatures did not vary by more than ±0.1°C. Antibiotics were added to the water in the vials (May 1975), and the survival of larvae under optimal conditions in these vials 'The term "median tolerance limit" and the symbol TLm are recommended in "Standard Methods" (American Public Health Association 1971). was comparable to that in larger containers. The number of larvae surviving at each test tempera- ture was recorded at 0.5, 1, 3, 6, 12, 24, 48, and 72 h after transfer, and for each time the TLm-the temperature at which just 50% of the larvae sur- vived the given time interval-was estimated by graphical interpolation as described by Doudoroff (1942). At each observation, larvae which showed no movement were removed from the vials by pipette and examined under a dissection micro- scope. If the heart was not beating and the larva was opaque, the larva was considered dead and was discarded; live larvae were returned to the vials. In one instance (see Results) moribund lar- vae were found and counted as dead. Salinity Tolerance Larvae were acclimated to salinities of 15, 20, 25, 30, 35, and 40'*/oo, and their upper and lower salinity TLm's were determined by transferring larvae directly to a series of 25-ml vials containing test salinities ranging from 0 (deionized water) to 58°/oo. Between five and seven larvae were trans- ferred to each vial within 8 h after hatching. Because of limited material, only the upper TLm was determined for larvae from an incubation salinity of 40*'/oo, and only four larvae from this salinity were available for each vial. The temperature during fertilization, incubation, and testing was maintained at 24 ± 0.2°C by thermo- statically controlled water baths, and vials were kept under continuous room light of low intensity (May 1975). The number of larvae surviving at each salinity was recorded 24, 48, and 72 h after transfer, and the upper and lower TLm's were estimated graphically for each time interval as in the case of thermal tolerance. Larvae were con- sidered dead on the basis of the same criteria used in the study of thermal tolerance. RESULTS Table 1. -Acclimation conditions for larvae used in determina- tion of heat tolerances. Acclimation temperat ure(°C) C/oo) 21 24 27 30 15 X X 20 X X 25 X X 30 X X 35 X X 40 X 45 X Upper Thermal Tolerance The upper TLm dropped with increasing time intervals (Figures 1-4), and there was a leveling off of the time-temperature curves in the lower salinities as time increased. Most of the time- temperature curves have been separated by eye-fit lines into two major segments, the horizontal seg- ment defining the "incipient lethal temperature" 250 MAY: EFFECTS OF ACCLIMATION 36 o ^ 32 LiJ a. 30 26 • 15 %o o 25%, A 36 %o i 45 %o -~--i- -4— A ?"- I 3 6 12 24 DURATION OF EXPOSURE (hours) 48 72 Figure l.-Heat tolerance of larval bairdiella acclimated to 21°C in various salinities. The upper median tolerance limits (TLm) are plotted for various durations of exposure. The time scale is logarithmic, and lines were fitted by eye. (Brett 1956), but in several curves there is a suggestion of an early plateau during the first few hours of exposure to the test conditions (Figures 1, 3, 4). At any given time the TLm was usually higher in the lower salinities. Since the highest test temperature used in the experiments was only 36°C, at acclimation temperatures above 21 °C the 50% mortality point was usually not reached until 3 or more hours after exposure to the test condi- tions. Survival was very poor among larvae from eggs maintained at 30°C (a temperature highly stressful to eggs-May 1975) in 30°/oo, and at 24 h, survival in this group was below 50% at all test temperatures. For purposes of comparison, the o o E _i H (E LiJ a. Q- 3 36 1 1 I 1 1 I • 15%. 0 25%o 1 * - - 35%. \ • ^^^ X. ^ "-^ 34 : A • ^--.^ ^^^^ 32 — 0 ^v ^^^.^^^ - — 0 ^O -v^_ ^ - ■v \ Ov \ 30 - "O— 'k 28 1 1 1 1 1 1 1 3 6 12 24 DURATION OF EXPOSURE (hours) 48 72 Figure 3.-Heat tolerance of larval bairdiella acclimated to 27°C in various salinties. 24-h upper TLm has been plotted against salinity for various acclimation temperatures (Figure 5); in this graph the increase in TLm at lower salini- ties is clear, as is the general increase in TLm effected by higher acclimation temperatures. In a salinity of 15^/00, all larvae alive at 12 and 24 h in the 21°C acclimation group were moribund in test temperatures of 30°C and higher, i.e., they were contorted and totally immobile and unresponsive to touch, although their hearts were beating and they were not opaque. These larvae have been considered dead for the purpose of data presen- tation; if considered alive, they would raise the calculated 12-h upper TLm from 29.5° to 32.1°C (Figures 1, 5). At the salinity of normal seawater, the 24-h upper TLm of larval bairdiella lies 36 334 E LU Q. 13 32 30 - 28 I 1 I 1 1 1 1 « 20%. Vx • 30%. \ \ X 40%. \ \ \ \ - ^ — — ^^^\ """'"^^'-^ — "-^» ^~~~^^-e « Xv. \ ~ \ ""» 9 — \ \ 1 1 1 1 1 1 1 I 3 6 12 24 48 72 DURATION OF EXPOSURE (hours) Figure 2.-Heat tolerance of larval bairdiella acclimated to 24°C in various salinities. 36 1 1 c 1 1 c * 1 20%. 3 0%. 1 G34 0_^ - e — E _l H LU 0. _ ^ X - 3 * 30 _ - 28 1 1 1 1 1 1 I 3 6 12 24 48 72 DURATION OF EXPOSURE (hours) Figure 4.— Heat tolerance of larval bairdiella acclimated to 30°C in two salinities. 251 FISHERY BULLETIN: VOL. 73, NO. 2 1 1 1 1 1 1 • 2|»C I A 27°c A o 32 — ^\ Q^ 0 24°C A sec - O \^x e \ _) \ N, H il \ \ ,^- liJ Q- 30 - — 3 • . - "^ tr x 3 29 - — O X ^ CM 28 1 1 1 1 1 1 1 20 25 30 35 SALINITY (7oo) 40 45 Figure 5. -Twenty- four hour upper median thermal tolerance limits (TLm) at various salinities for larvae acclimated to 21°, 24°, 27°, and 30°C. between 29° and 31°C, depending on the acclima- tion temperature. Larvae could resist higher temperatures for shorter periods of time. Salinity Tolerance The 24 h upper TLm for salinity v^^as not greatly affected by the acclimation salinity and ranged from 43 to iS.b^/oo, but the 24-h lower TLm was appreciably higher among larvae incubated at higher salinities (Figure 6). The lower TLm's (24 h) 15 20 25 30 35 40 ACCLIMATION SALINITY (%o) Figure 7.-Upper and lower median tolerance limits (TLm) of salinity for a 48-h exposure. for larvae acclimated to 15 and 20^/oo were 4.2 and S.5^/oo, respectively, whereas those for larvae acclimated to 30 and 35°/oo were 10 and le.S^/oo, respectively. The major difference between the 24-h TLm's and those for 48 and 72 h is the progressive lowering of the upper TLm for larvae acclimated to low salinities (Figures 7, 8). The upper TLm for larvae from 15"/ oo shifted from 46.2V 00 at 24 h, to 36.6Voo at 48 h, and to 30.0V oo at 72 h; between 24 and 72 h, the upper TLm re- mained the same (43Voo) for larvae acclimated to 40V 00, and decreased only from 48 to 46.2"/ oo for 15 20 25 30 35 40 ACCLIMATION SALINITY (%o) Figure 6.-Upper and lower median tolerance limits (TLm) of salinity for a 24-h exposure. Larvae were acclimated to various salinities at 24°C. 5 2 0 25 ACCLIMATION 30 35 SALINITY (%o) Figure 8.-Upper and lower median tolerance limits (TLm) of salinity for a 72-h exposure. 252 MAY: EFFECTS OF ACCLIMATION those acclimated to 357oo. There was little change in the lower TLm's between 24 and 48 h, but a slight rise occurred between 48 and 72 h in all but the 30"/ 00 acclimation group. DISCUSSION Fry et al. (1946) define the "zone of tolerance" as the range of any environmental factor within which an animal can live indefinitely, and the "zone of resistance" as the range within which the animal can live for only a finite period of time, depending on the level of the factor. The zone of tolerance is bounded by the upper and lower "in- cipient lethal levels." In work on the upper thermal tolerance of fishes, the incipient lethal level is defined by an abrupt flattening of the time- temperature curve at a temperature below which less than 50% of the exposed individuals succumb (Brett 1956). Some of the curves generated in the present study (Figures 1-4) suggest that the in- cipient lethal level has been reached, but curves from the higher salinities lack a horizontal seg- ment. This points up a difficulty in working with early larvae: at 24°C the larval yolk supply is 95% consumed by about 40 h after hatching (May 1974), and this occurs even sooner at higher tempera- tures. The 48- and 72-h TLm's therefore apply to starving larvae. Unlike adult fish, larvae which hatch from pelagic eggs are extremely sensitive to food deprivation (e.g., Lasker et al. 1970) and begin dying of starvation soon after yolk absorp- tion if food is not provided for them, and unfed bairdiella larvae die sooner at high temperatures and salinities (May 1975). Therefore, prolonging these tests would not have helped in defining the upper incipient lethal temperature for larvae in the higher salinities-the TLm would simply con- tinue to fall. Even at the lower salinities, the TLm would decline after a sufficient period of time; the curves for a salinity of 35'*/oo (Figures 1, 3) show how a flat segment is reached, only to be followed by another drop in TLm. A further difficulty in estimating tolerance limits for warmwater larvae is that these larvae develop morphologically at an extremely rapid rate and are very different or- ganisms 1 or 2 days after hatching than they are at hatching. Newly hatched bairdiella are poorly developed and rather inactive (May 1975), whereas by 45 h after hatching (at 24°C) they have acquired functional eyes and an open mouth and are quite active. In this situation, consideration of the TLm at a more or less arbitrary time after exposure to the test conditions, such as 24 h, is at least a useful approach for comparative purposes. Larval bairdiella are more sensitive to high temperatures when the salinity is also high, as are bairdiella gametes and developing embryos (May 1975). This adds further weight to the suggestion (May 1975) that in nature, eggs spawned late in the season at high temperatures will have a reduced chance of contributing recruits to the population when natural salinities rise as they are doing in the Salton Sea. The survival of bairdiella larvae in the Salton Sea would be significantly reduced at temperatures above 31°C, and temperature data from the Salton Sea (May 1975) indicate that some larvae could be exposed to thermal stress of this level or greater. The highest TLm is reached in W/m, the lowest salinity in which larvae were tested and the nearest to being isosmotic with larval body fluids. Older fishes of various species are also most tolerant of high temperatures in isosmotic or nearly isosmotic salinities (Aral et al. 1963; Strawn and Dunn 1967; Garside and Jordan 1968; Simmons 1971). The add- ed burden of osmotic work seems to reduce the ability of both larval and adult fish to tolerate extremely high temperatures. It is clear that acclimation can alter the tolerance of early bairdiella larvae to both temperature and salinity, even though the rapid developmental rate of bairdiella eggs restricts the period of acclimation to between 20 and 40 h (the time between fertilization and transfer to test conditions, which is a function of incubation temperature). Incubation of bairdiella eggs at higher temperatures produces larvae with a higher upper thermal TLm. However, increasing the acclimation temperature from 27° to 30°C does not increase the upper TLm, even though the TLm's are generally above 30°C. Hence the lethal levels determined for an acclimation temperature of 27°C may represent "ultimate" incipient lethal temperatures (Fry et al. 1946), but here again one must consider the unique problems of working with early larvae. If the effect of thermal acclimation on the tolerance of yolk-sac larvae is to be studied, acclimation must take place during embryonic development, but the embryos are more sensitive to temperature than are the larvae to which they give rise (cf . May 1975). A temperature of 30°C is extremely stressful for developing eggs, and the larvae produced at this temperature sur- vive poorly, a trait magnified at higher salinities. 253 FISHERY BULLETIN: VOL. 73, NO. 2 The response of these larvae to elevated tempera- tures is therefore not a true reflection of thermal "acclimation," as the term is generally used, but is more a reflection of thermal stress during sensi- tive periods of morphogenesis. In an analogous way, salinity stress on embryos during acclimation at 4(y/f» probably accounts for the observation that the larvae have a reduced upper TLm for salinity when compared with larvae acclimated to 30 and SBVoo. Thermal acclimation has also been shown to af- fect the thermal tolerance of larval herring (Blaxter 1960), menhaden (Lewis 1965), and sal- monids (Bishai 1960; Iwai 1962), although only Blaxter's study utilized larvae which hatched from eggs maintained at the acclimation temperature. The mechanisms involved in thermal acclimation during early development have never been inves- tigated, but the present results for bairdiella sug- gest that they must be activated quite rapidly, within a day or two at most. A similarly rapid rate of acclimation to warm temperatures has been found in older fish (Brett 1970; Allen and Strawn 1971), so that a similar mechanism may be operat- ing in both cases. Factors involved in setting thermal tolerance limits in fishes are little under- stood (Fry 1967), but thermal inactivation of en- zymes has been suggested as a possible mechanism (Hochachka and Somero 1971). HoUiday and Blaxter (1960) found that the salinity prior to hatching had a limited effect on the salinity tolerance of larval herring. This effect was more pronounced in the present experiments with larval bairdiella, but there was a delay in the appearance of the acclimation response to high salinities. The upper TLm (salinity) was similar for all acclimation salinities 24 h after initial ex- posure to the test conditions, but at 48 h the larvae acclimated to high salinities had a higher TLm than those from low salinities (a very slight in- dication of the same phenomenon can be discerned in the results of Holliday and Blaxter 1960). This observation is difficult to explain, especially in view of the rudimentary state of our knowledge of larval osmoregulatory mechanisms; perhaps it is related to the opening of the mouth between 35 and 45 h after hatching (May 1974), which could expose the internal larval tissues more directly to the ambient salinity. Incubation at low salinities enables larvae to tolerate much lower salinities than larvae incubated in more saline water. Again, it is difficult to speculate on how this effect might be mediated. Early larvae of Bairdiella icistia are more tolerant than the embryonic stages and less tolerant than adults to extremes of temperature and salinity. Very few bairdiella eggs develop normally at 30°C (May 1975), and 15 to i(P/(x> is the approximate salinity range for normal fertiliza- tion and embryonic development. In contrast to the eggs, 50% of the newly hatched larvae are capable of withstanding temperatures between 30° and 33°C for 24 h or longer, except at the lowest acclimation temperature and highest salinity; and with proper acclimation, larvae can tolerate salinities ranging from about 4 to iS'^/oo for 24 h, or 5 to 45^/00 for 72 h. Juvenile and adult bairdiella must tolerate temperatures ranging from 10° to 34° or 35°C in the Salton Sea (Carpelan 1961). These fish have been found in freshwater (R. G. Hulquist, California Department of Fish and Game, pers. commun.) and can tolerate Salton Sea water with a salinity of 52.5''/oo for 96 h after direct transfer from ordinary Salton Sea water (approximately SS'^/oo), and 5S^/oo for over a week after gradual acclimation (Hanson 1970). The early larvae of some other species have also been shown to be more tolerant of temperature and salinity than their eggs. McCauley (1963) reports that prolarvae of the sea lamprey, Petromyzon marinus, are considerably more tolerant of high temperatures than are the eggs, and data presented by Holliday (1965) show that newly hatched herring, Clupea harengus; plaice, Pleuronectes platessa; and Atlantic cod, Gadus morhua, larvae are more tolerant of both high and low salinities than are their respective eggs. However, in the case of the herring and plaice, further larval development and metamorphosis are accompanied by a decrease in salinity tolerance (Holliday 1965), a pattern quite different from that found in bairdiella. ACKNOWLEDGMENTS I thank Reuben Lasker for his advice and material aid during this work. The University of California Institute of Marine Resources and the Southwest Fisheries Center, National Marine Fisheries Service, NOAA provided financial sup- port. LITERATURE CITED Allen, K. 0., and K. Strawn. 1971. Rate of acclimation of juvenile channel catfish, Ic- 254 MAY: EFFECTS OF ACCLIMATION talurus punctatus, to high temperatures. Trans. Am. Fish. Soc. 100:665-671. Arai, M. N., E. T. Cox, and F. E. J. Fry. 1963. An effect of dilutions of seawater on the lethal temperature of the guppy. Can. J. Zool. 41:1011-1015. BiSHAI, H. M. 1960. Upper lethal temperatures for larval salmonids. J. Cons. 25:129-133. Blaxter, J. H. S. 1960. The effect of extremes of temperature on herring larvae. J. Mar. Biol. Assoc. U.K. 39:605-608. Brett, J. R. 1956. Some principles in the thermal requirements of fishes. Q. Rev. Biol. 31:75-87. 1970. Fishes. Functional responses. In 0. Kinne (editor). Marine ecology, Vol. 1, Part 1, p. 515-560. Wiley-Inter- science, Lond. Carpelan, L. H. 1961. Physical and chemical characteristics. In B. W. Walker (editor), The ecology of the Salton Sea, California, in relation to the sport-fishery, p. 17-32. Calif. Dep. Fish Game, Fish Bull. 113. DOUDOROFF, P. 1942. The resistance and acclimatization of marine fishes to temperature changes. I. Experiments with Girella nigricans (Ayres). Biol. Bull. (Woods Hole) 83:219-244. Fry, F. E. J. 1967. Responses of vertebrate poikilotherms to tempera- ture. In A. H. Rose (editor), Thermobiology, p. 375- 409. Academic Press, Lond. 1971. The effect of environmental factors on the physiology of fish. In W. S. Hoar and D. J. Randall (editors), Fish physiology. Vol. 6, p. 1-98. Academic Press, N.Y. Fry, F. E. J., J. S. Hart, and K. F. Walker. 1946. Lethal temperature relations for a sample of young speckeled trout, Salvelinus fontinalis. Univ. Toronto Stud., Biol. Ser. 66:9-35. Garside, E. T., and C. M. Jordan. 1968. Upper lethal temperatures at various levels of salinity in the euryhaline cyprinodontids Fundulus heteroclitus and F. diaphanus after isosmotic acclimation. J. Fish. Res. Board Can. 25:2717-2720. Hanson, J. A. 1970. Salinity tolerances for Salton Sea fishes. Resour. Agency Calif., Dep. Fish Game, Inland Fish. Admin. Rep. 70-2, 8 p. Haydock, L 1971. Gonad maturation and hormone-induced spawning of the Gulf croaker, Bairdiella icistia. Fish. Bull., U.S. 69:157-180. Hochachka, p. W., and G. N. Somero. 1971. Biochemical adaptation to the environment. In W. S. Hoar and D. J. Randall (editors). Fish physiology. Vol. 6, p. 99-156. Academic Press, N.Y. HOLLIDAY, F. G. T. 1965. Osmoregulation in marine teleost eggs and lar- vae. Calif. Coop. Oceanic Fish. Invest. Rep. 10:89-95. HOLLIDAY, F. G. T., AND J. H. S. BLAXTER. 1960. The effects of salinity on the developing eggs and larvae of the herring. J. Mar. Biol. Assoc. U.K. 39:591-603. IWAI, T. 1962. Studies on the Plecoglossus altivelis problems: Embryology and histophysiology of digestive and osmo- regulatory organs. Bull. Misaki Mar. Biol. Inst, Kyoto Univ. 2, 101 p. Lasker, R., H. M. Feder, G. H. Theilacker, and R. C. May. 1970. Feeding, growth, and survival of Engraulis mordax larvae reared in the laboratory. Mar. Biol. (Berl.) 5:345-353. Lewis, R. M. 1965. The effect of minimum temperature on the survival of larval Atlantic menhaden, Brevoortia tyrannus. Trans. Am. Fish. Soc. 94:409-412. McCauley, R. W. 1963. Lethal temperatures of the developmental stages of the sea lamprey, Petromyzon marinus L. J. Fish. Res. Board Can. 20:483-490. May, R. C. 1974. Effects of temperature and salinity on yolk utilization in Bairdiella icistia (Jordan & Gilbert) (Pisces: Sciaenidae). J. Exp. Mar. Biol. Ecol. 16:213-225. 1975. Effects of temperature and salinity on fertilization, embryonic development, and hatching in Bairdiella icis- tia (Pisces: Sciaenidae), and the effect of parental salinity acclimation on embryonic and larval salinity tolerance. Fish. Bull., U.S. 73:1-22. Simmons, H. B. 1971. Thermal resistance and acclimation at various salini- ties in the sheephead minnow (Cyprinodon variegatus Lacepede). Texas A&M Univ., Sea Grant Publ. TAMU- SG-71-205, 41 p. Strawn, K., and J. E. Dunn. 1967. Resistance of Texas salt- and freshwater-marsh fishes to heat death at various salinities. Tex. J. Sci. 19:57-76. Thomas, W. H., H. L. Scotten, and J. S. Bradshaw. 1963. Thermal gradient incubators for small aquatic or- ganisms. Limnol. Oceanogr. 8:357-360. 255 THE INTERACTION OF ECONOMIC, BIOLOGICAL, AND LEGAL FORCES IN THE MIDDLE ATLANTIC OYSTER INDUSTRY Richard J. Agnello and Lawrence P. Donnelley' ABSTRACT Economic, environmental, and legal forces are contributing factors in the decline of the Middle Atlantic oyster industry. This paper determines the interactions and importance of these forces by quantifying and integrating some of the relevant variables into a supply and demand model of the oyster industry. The statistical results yield significant and e.xpected parameter values with useful information on price and income demand elasticities. Also implications of common property legal frameworks on resource utilization are revealed. The main conclusions are that efforts to rehabilitate the industry by cleaning up pollution, replacing cultch, and encouraging legal private property rights may have large social values. The historically important Middle Atlantic oyster industry is currently recognized as having many of the symptoms of a declining industry. Economic, biological, and legal forces are con- tributing causal factors in the fishery's decline. This paper attempts to integrate some of these variables into an estimable supply and demand model explaining oyster price and output movements over time for the region. Economic and biological variables are directly included in the model while the legal dimension is focused on in- directly by comparing empirical results for data generated from different common property structures. The multidimensional approach reveals infor- mation on price and income elasticities, substitut- ability relationships, and the effect a common property regulatory framework has on resource overutilization and depletion. The regional orien- tation taken in this paper, rather than a national or international focus, provides a departure from much of the traditional fishery analysis and ena- bles the effects of alternative property right structures between states to be observed. - The impact property rights structure has on economic efficiency and biological growth has been discussed widely in the theoretical literature of fishery modelling.' Less however has been written 'Department of Economics, College of Business and Economics, University of Delaware, Newark, DE 19711. ^Recent empirical work confined primarily to economic factors and directed towards species and regions different from our analysis includes Bell (1968), Doll (1972), and Waugh and Norton (1969). 'Much has been written in fishery economics on the effects of biological stock depletion due to common property legal struc- on empirical analyses of the effects property rights have on economic and biological efficiency.^ Consensus among the discussants is that common property leads to overexploitation of fish stocks and perhaps extinction of a species. Common property right systems result in less efficient resource allocation than private right systems since the former do not ensure that the total costs of an individual harvester's exploitation of the resource are borne fully by him. Private property internalizes the costs of the harvester's actions thereby forcing the producer to bear not only all of the costs of his actions but also to capture all of the benefits. MIDDLE ATLANTIC OYSTER FISHERY The American Eastern oyster represents the resource base of both the Gulf and Atlantic coasts oyster industries. Following a brief mobile larval stage, the oyster connects permanently to a firm subaqueous material such as rock or shell deposits. tures. Some of the earlier treatments that are still widely referred to can be found in Gordon (1954) and Scott (1955). For more recent theoretical analyses see FuUcnbaum et al. (1972), and Smith (1969). In the context of this literature, common property means that any member of a community has the right to harvest the fish stock. "Notable exceptions in the fishery literature where the effects of legal ownership frameworks have been quantified include Bell (1972) and O'Rourke (1971). For example Bell estimates the redundant effort employed in the American lobster fishery which is subject to common property. He concludes that approximately 50% of current fishing effort is required to achieve economic efficiency. Also the authors, in an unpublished paper (Efficiency and Property Rights in the U.S. Oyster Industry, 1974), estimate that in 1969 oystermens' income would have increased by almost 50% if all coastal states had relied on leasing of oyster beds. Manuscript accepted August 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 256 AGNELLO and DONNELLEY: INTERACTION OF FORCES Its habitat is the intermediate sahnity waters of the seacoast's intertidal zone and of inland rivers and bays. Water current, temperature, and biological productivity, in addition to salinity, are determinants of the resource productivity of a given parcel of subaqueous land. The property right structure characterizing oyster grounds varies widely among states. Courts have granted rights to subaqueous land to the people of each state for their own common use. State legislatures exercise these rights. The federal government has been granted the right to a 3-mile coastal zone and Congress in turn has ceded back to the states land and resource use rights within this zone."" States have responded in similar ways to the exercise of their rights to the oyster resource. In general, natural oyster beds have been set aside as a common fishery for state residents,*^ whereas other submerged land parcels are available for private leasing. However, great variation among the states exists in the proportion of area and quality of land set aside for public or common use versus private use depending on how broadly administrators define the term "natural oyster bed."^ An examination of the proportion of oyster catch by weight on private grounds to total catch by state reveals ratios ranging from a maximum of 1 to 0 for certain states in recent years. Within the Middle Atlantic region the two states with property rights in Delaware Bay (i.e., Delaware and New Jersey) can be characterized as essentially private property states, whereas Maryland and Virginia, which share the Chesapeake Bay, have significantly lower private to total catch ratios.** Private property rights in oystering tend to promote efficiency in several ways. First, exclusive user rights provide incentives for firms to pursue a policy of investing in cultch and maintaining it at ^See Power (1970) for a detailed description of court decisions involving rights to submerged land. 'Legislative codes usually prohibit nonresidents from entering the industry. See Power (1970:216-223) for a discussion of the constitutionality of these restrictions. 'Maryland, for example, classifies a natural bed as one such that the natural growth of oysters ". . . is of such abundance that during the preceding five years the public has resorted to them for livelihood," Power (1970:220). Courts reportedly view one in- dividual declaration of one day's work in a 5 yr period as sufficient evidence of the existence of a natural bed. Most states employ a less restrictive definition for a natural bed. 'It is useful to note that private property rights may be in- stitutionally arranged in a multitude of ways. The usual manner of leasing subaqueous lands to a relatively few individuals is by no means the only way of introducing private property, and in fact is often objected to as prejudicial to individual freedom. A more acceptable arrangement pointed out by a reviewer may be for states to assume control of beds and issue permits to harvest a given quantity of oysters. a desired level as influenced by market conditions. Second, congestion and overexploitation of the oyster resource is unlikely to occur since there is no pressing need to harvest quickly so as to not lose the resource's benefits. Finally, a communal property structure tends to lower efficiency by requiring the use of obsolete technology in order to prevent depletion of the resource stock. Inefficient technology often takes the form of obsolete capital regulated into use by legislative codes. In general, states relying on common property right struc- tures tend to impose greater restrictions on the use of capital than private property states.^ Between 1947 and 1968 the annual U.S. domestic oyster harvest declined from 63.1 to 55.6 million pounds. Imports increased from an insignificant 111 thousand pounds to 15.5 million pounds during the same years. Accounting for inventory changes, total consumption of oyster meat consequently expanded by 8.2 million pounds. Concurrently, both ex-vessel and wholesale prices rose. Between 1950 and 1968 ex-vessel prices rose 38% and wholesale prices rose 89%."* Significant regional differences in oyster catch trends characterize the post-World War II period. In general, the Gulf region has increased its land- ings while landings in the Middle Atlantic region (defined to include the states of New York, New Jersey, Delaware, Maryland, and Virginia) have declined by 45%. Delaware and New Jersey har- vests especially have fallen dramatically, no doubt in large part due to disease which affected stocks beginning in 1958. It is during this period of both relative and absolute decline that we shall es- timate the underlying factors explaining changes in quantities and prices for the Middle Atlantic oyster industry. ' MODEL Economic variables such as prices and quantities are generally explained by economists through the use of supply and demand models. Prices and quantities are determined through the equilibra- tion of supply and demand forces which incor- porate the effects of various predetermined 'For example, in the predominantely common property right state of Maryland power dredging is prohibited in the harvest of oysters. Consequently any dredging takes place through the use of sail-powered craft called skipjacks, the newest of which is around 50 yr old. "All data presented in this section are from Fhherij Statistics of the United States, Bureau of Commercial Fisheries. National retail price data are not readily available. 257 FISHERY BULLETIN: VOL. 73, NO. 2 (independent) variables on the endogenous (dependent) variables price and quantity. A general application of this methodology to any commodity would specify that quantity supplied (Qs) is dependent on input cost factors, prices of goods related in supply, and price of the product. Quantity demand (Q^ is usually hypothesized to depend on current price, income, and prices of related goods in demand. When specifying such a model for oyster markets several modifications to the supply specification of the above general framework are employed. Although from supply theory input factors include technological, environmental, and biological variables, data limitations restrict the inclusion to a single biological factor, the MSX disease." It is hypothesized that the protozoan oyster parasite commonly referred to as the MSX disease has a negative impact on the oyster in- dustry during the period of analysis. Also since no strong relationships between the production of oysters and other goods is readily apparent, we omit prices of related goods from the oyster supply relationship. The last modification of the supply relationship for oysters and probably the most unique feature of the model is the hypothesis that quantity supplied is a function of price lagged 1 yr rather than current price. As in the case of agricultural commodities, quantity supplied of oysters can be considered a function of past price and natural phenomena and hence fixed in the short run. In the fishery case a fixed supply is usually based on the presence of lags in generating fishing effort (e.g. securing capital and making occupational choices). Lagged price can be expected to positively influence current fishing effort.'^ In fishery es- timation with annual data however, the assump- tion of such long lags in adjusting fishing effort may be inappropriate and the inclusion of lagged "Little technological change has occurred during the period of analysis due in part to state regulation mandating old tech- nologies as a conservation device. Environmental factors such as pollution and siltation have doubtless had a negative impact on oyster supplies, but unfortunately little systematic and consis- tent information is available through time. '^e note that although lagged price may positively affect current effort the total effect on current harvest (i.e. supply) depends on what effect lagged price has on the current biological stock of the resource. For reasons explained below the net affect of lagged price on current supply might actually be negative if stocks have been reduced to tne point of depletion. For an example of the short-run supply assumption (i.e., supply is independent of current price) in the fishery area, see Bell (1968). It should be noted that Bell's empirical work is quite successful using monthly data. price as a determinant of effort will likely be a weak determinant of supply. An additional rationale is therefore necessary for including price lagged 1 yr as a determinant of supply. We hypothesize that lagged price has a negative impact on current supply due to a deple- tion effect. In fishery production not only current effort, but current biological stock determines current production. If current biological stock is negatively related to past effort, then variables explaining past effort may bear a strong relation to current supply. Price lagged 1 yr (or alterna- tively distributed lags of past prices if data were plentiful) may be closely related to past effort. Accordingly we hypothesize that price lagged 1 yr is a proxy for past levels of effort and thus nega- tively related to current biological stock. Also if the negative relationship that lagged price has on current biological stock is strong enough to offset the positive relationship lagged price has on current fishing effort, the net impact of lagged price on current supply might be negative. Furthermore we expect the negative depletion effect of lagged price to be stronger in fisheries subject to common property. Thus lagged price is likely to have a slight positive effect on supply in fisheries subject to private property due to the positive effort effect. For fisheries subject to common property the negative depletion effect is expected to offset the positive effort effect yield- ing a negative relationship between lagged price and quantity supplied. The structural equations and equilibrium condi- tion of the fixed supply model applied to oyster markets are written below with expected parameter signs appearing above each explana- tory variable. -1- SupplyQ, = S(MSX,P,.i) Demand Q^ = D, (P, I, P,) Equilibrium conditions Q, = Qd = Q where MSX, Pf .i, P, I, and P^ are the MSX disease, price lagged 1 yr, current price, income, and the price of a related good, respectively. In the supply equation a negative relationship with MSX is hypothesized a priori, and lagged price may be positive or negative depending on the intensity of depletion. In demand, current price is expected to have a negative effect according to the law of 258 AGNELLO and DONNELLEY: INTERACTION OF FORCES demand, the income effect is positive if oysters are a normal good, and the related good (poultry) is most likely a substitute for oysters. Since the fixed supply assumption removes the simultaneity from the model, both the supply and demand functions can be interpreted as reduced forms, and estimated directly by ordinary least squares regression techniques without regard to problems of identification. The demand function is solved for its only endogenous variable, price, and estimated in this form. The reduced form equa- tions estimated later in the paper are written below. SupplyQ = S(MSX,P,_i) + + - Demand P = D2{I,P,,Q) where Q is predetermined. In addition to conclusions concerning parameter signs and elasticity values, implications of property right structures are revealed in the model. It is hypothesized that in states relying more heavily on common property rather than private leasing of subaqueous beds, the depletion effect should be greater. The lagged price variable in supply should thus have a negative coefficient value for common property right areas. DATA Regression analyses are performed on the above model for the Middle Atlantic and Delaware Bay regions. The Middle Atlantic region is defined to include the five contiguous coastal states of Vir- ginia, Maryland, Delaware, New Jersey, and New York. This region includes the productive subaqueous resources of Chesapeake Bay, Delaware Bay, and part of Long Island Sound. The Delaware Bay area includes the states of Delaware and New Jersey only. We anticipate depletion to be a more important factor in the regional analyses which are dominated by the high production levels of Maryland and Virginia. These Chesapeake Bay states (especially Maryland) rely to a greater extent on common property than do the Delaware Bay states. The latter states allow much more extensive private leasing, and hence supply a greater proportion of their oysters from private leaseholds which may be expected a priori to be less subject to overfishing with the ensuing depletion. Time series annual data on quantities landed (in pounds) and implicit prices are obtained from Fishery Statistics of the United States (1940-1970) compiled by the Bureau of Commercial Fisheries of the U.S. Department of Commerce and the U.S. Department of the Interior. Data on the price of a related commodity in demand (i.e., the price of poultry) and personal income are obtained from the Bureau of Labor Statistics (U.S. Department of Labor) and the Survey of Current Business (U.S. Department of Commerce), respectively. The biological variable representing the MSX disease included in the oyster supply function of the model was obtained from site sampling of oysters in the Delaware River." The regions of consumption and production are not identical in that the consump- tion area includes a somewhat larger area. Precise definitions of the variables used are given below. Q Quantity per capita of oyster landings (measured as pounds per person) for Delaware Bay includes Delaware and New Jersey, and regional quantities include the five states of Delaware, New Jersey, Maryland, Virginia, and New York. Popula- tion refers to the seven-state region includ- ing New York, Pennsylvania, New Jersey, Delaware, Maryland, District of Columbia, and Virginia. P Price of oysters measured in dollars per pound (meat weight). Pf _i Price of oysters lagged 1 yr. MSX Biological variable referring to a pro- tozoan oyster parasite commonly called the MSX disease. / Personal income per capita (deflated by the Consumer Price Index for all items) for the seven-state region including New York, Pennsylvania, New Jersey, Delaware, Maryland, District of Columbia, and Vir- ginia. P^f^ National average price per pound for chickens (live weight). EMPIRICAL RESULTS We now turn to a brief discussion of the detailed findings, and conclude with some general remarks and policy implications. Tables 1 and 2 present the "The average annual prevalence of the MSX disease in a test sample of oysters in the Delaware River was obtained from H. Haslcin of Rutgers University. These percentages were zero before 1957 and exceeded 50% in some later years. 259 FISHERY BULLETIN: VOL. 73, NO. 2 Table l.-Middle Atlantic supply and demand regressions. Predetermined variables Statistics! R2 Elasticities^ Price Endogenous Equation variable Constant MSX n P,,. P,., QJ DW Income Supply 0' 1.733 -0.581 (-3.14)* -1.202 -(6.29)* 0.85 0.48 Supply In Q' -0.136 -1.001 (-3.60)* -0.319 -(3.05)* 0.68 0.39 Demand P 0.802 0.00002 0.010 -0.507 0.54 1.0 (0.20) (4.12)* (-3.14)* 0.76 0.1 Demand In P" -14.110 1.582 0.425 -0.421 0.53 2.4 (3.28)* (2.59)* (-1.50) 0.64 3.8 1 R^ and DW refer to the unadjusted coefficient of determination and the Durbin-Watson statistic for autocorrela- tion, respectively. 2 The formulae used in calculating price and income elasticities are ~^}^ • q and ^ '^ , respectively. In the re- gressions not utilizing logarithms mean values of variables are used to fix the point elasticities. 5 Quantity and Income are measured in per capita form. ■^All variables are measured in natural logarithms except tVlSX. * Refers to statistical significance at the 0.05 level. Table 2.-Delaware Bay supply and demand regressions. Predetermined variables Endogenous Equation variable Constant h^SX Q5 Statistics' DW Elasticities^ Price Income Supply Supply Demand Demand Q3 InO* P InP* 0.223 -1.718 0.165 -18.290 -0.409 (-8.05)* -4.317 (-5.99)* 0.003 (0.22) -0.155 (-0.77) 0.0003 (3.18)* 2.090 (5.04)* 0.008 (2.43)* 0.407 (2.51)* -1.169 (-2.42)* -0.170 (-2.51)* 0.70 0.77 0.66 0.98 0.55 1.06 0.66 1.24 3.4 4.1 5.9 12.3 iff2 and DW refer to the unadjusted coefficient of determination and the Durbln-Watson statistic for autocorrela- tion, respectively. 2The formulae used in calculating price and income elasticities are =^^ • q and "gf 'q, respectively. In the re- gressions not utilizing logarithms mean values of variables are used to fix the point elasticities. 3 Quantity and Income are measured in per capita form. ■•All variables are measured in natural logarithms except f\/ISX. * Refers to statistical significance at the 0.05 level. empirical results of the supply and demand model applied to oyster data for the years 1940 to 1970 in the Middle Atlantic and Delaware Bay regions. Numerical estimates of the coefficients along with t values in parentheses are obtained through the use of either linear or log linear regression analysis and ordinary least squares as the method of estimation. In general, parameters have expected signs and are significant for at least the 0.05 level. The coefficient of determination is reasonably high in most regressions indicating that the included predetermined variables explain a large fraction of the variation in the endogenous variables. Since the linear and logarithmic equation forms do not differ greatly there is no evidence of nonlineari- ties. In the supply equations the MSX variable displays a strong negative impact on quantity, and is more significant (i.e., larger t values) in the Delaware Bay regressions. Biological evidence in- dicates that the disease had a greater impact on Delaware Bay production than on Chesapeake Bay production although Virginia was hard hit during much of the period of analysis. Lagged price has a negative and highly sig- nificant impact on supply for the Middle Atlantic region indicating that the depletion effect dominates the effort effect where common property prevails. In contrast the Delaware Bay results indicate that lagged price is not a sig- nificant determinant of supply in a private property right structure. In the price (implicit demand) equations oysters display a significant positive income response in most regressions.'^ Oysters thus appear to be a normal good whose demand is likely to grow as consumer income rises over time. Since the price of chickens is a positive determinant of demand in all regressions, the relationship between the two commodities is one of substitution. The negative coefficient for quantity supports the law of "Preliminary cross sectional analyses conducted by National Marine Fisheries Service Economic Research Laboratory in- dicate much weaker income effects (and possibly negative) for oysters. 260 AGNELLO and DONNELLEY: INTERACTION OF FORCES demand indicating demand for oysters to be price-responsive. In order to determine meaningfully how re- sponsive quantity demanded is to price and income changes, it is useful to investigate the elasticities implied by the statistical results. Tables 1 and 2 indicate high elasticities in both the Middle Atlantic and Delaware Bay regions implying that oysters are price elastic and normal with respect to consumer income responses. If supplies were to increase in the future, one would expect increasing revenues for the oyster industry.'' Similarly we might expect consumer demand for oysters to increase by larger percentages than real personal income in the future. Efforts to rehabilitate the oyster industry by cleaning up water pollution, discouraging overfishing, and replacing oyster cultch may thus have large social values. Although the statistical results do lend support to the model, they are certainly not without difficulties. The time series problem of positive serial correlation is present throughout, thus de- tracting from the reliability of the results. The Durbin-Watson statistics in general indicate either positive autocorrelation or indeterminancy for the Middle Atlantic and Delaware Bay regions respectively using a two-tailed test at the 0.05 level of significance.'* An additional problem im- pairing both estimation and prediction is struc- tural change with data over a long time period. Parameters therefore may not remain constant with time series data. Also variables omitted from the model may have caused shifts in the functions over time. All of these problems make prediction hazardous and definitive conclusions should await further testing based on new data sets. CONCLUSIONS In general the statistical results support the model of supply and demand forces in the Middle Atlantic oyster industry. Estimates are generated on income and price elasticities of demand and lend optimism to the current rehabilitation efforts directed toward the oyster industry. The MSX disease has clearly had a debilitating effect, '^It has been reported by the Delaware State Department of Natural Resources and Environmental Control that oyster spat count recently have been the highest in several years indicating augmented supplies to be highly probable in the future. "When first differences are used to remove serial correlation, R' and t values fall to unacceptably low levels although serial correlation is removed. however, and must be solved as a condition of suc- cessful industry recovery. The common property characteristics of the in- dustry have also harmed the industry's progress. There exists evidence of overfishing in common property states, and hence less than optimal exploitation of the natural resource stocks. The results indicate that depletion is a much more serious problem for the Chesapeake Bay states than for the Delaware Bay states where private leasing of subaqueous lands is more prevalent. However, the reverse is true concerning the MSX disease characteristics of the regions. ACKNOWLEDGMENTS Research for the paper was funded under the Sea Grant Program of the National Oceanic and Atmospheric Administration, U.S. Department of Commerce. We wish to express our thanks to the anonymous reviewers for many helpful comments on an earlier version of this paper. Any errors remaining are, of course, entirely our respon- sibility. LITERATURE CITED Bell, F. W. 1968. The Pope and the price of fish. Am. Econ. Rev. 58:1346-1350. 1972. Technological externalities and common-property resources: An empirical study of the U.S. northern lobster fishery. J. Polit. Econ. 80:148-158. Doll, J. P. 1972. An econometric analysis of shrimp ex-vessel prices, 1950-1968. Am. J. Agric. Econ. 54:431-440. FULLENBAUM, R. F., E. W. CARLSON, AND F. W. BELL. 1972. On models of commercial fishing: A defense of the traditional literature. J. Polit. Econ. 80:761-768. Gordon, H. S. 1954. The economic theory of a common property resource: The fishery. J. Polit. Econ. 62:124-142. O'ROURKE, D. 1971. Economic potential of the California trawl fishery. Am. J. Agric. Econ. 53:583-592. Power, G. 1970. More about oysters than you wanted to know. Md. Law Rev. 30:199-225. Scott, A. D. 1955. The fishery: The objectives of sole ownership. J. Polit. Econ. 63:116-124. Smith, V. L. 1969. On models of commercial fishing. J. Polit. Econ. 77:181-198. Waugh, F. v., and V. J. Norton. 1969. Some analyses of fish prices. Univ. R.I., Agric. Exp. Stn. Bull. 401. 261 THE REPRODUCTIVE BIOLOGY OF THE PROTOGYNOUS HERMAPHRODITE PIMELOMETOPON PULCHRUM (PISCES: LABRIDAE) Robert R. Warner' ABSTRACT Pimelometopon pukhrum, California sheephead, a labrid fish of the eastern Pacific Ocean, was collected the year round at Catalina Island, Calif., and comparative material was taken at Guadalupe Island, Mexico. Individuals at Guadalupe were dwarfed relative to those at Catalina. Pimelometopon pukhrum is a protogynous hermaphrodite, the ovarian elements undergoing massive degeneration as sperma- togenic crypts proliferate in the gonads of transitional individuals. Sexual changes occur between breeding seasons. Individuals from both populations mature as females at age four; most of those at Catalina function as females for 4 yr and then change sex, at a length of around 310 mm. Sexual transformation occurs earlier on the average at Guadalupe; most individuals are male by age seven. In both populations, more rapidly growing fishes apparently change sex sooner than other individuals of the same age, and fishes that grow slowly may not change sex at all. Spawning appears to occur in July, August, and September in the Catalina population. Individuals probably spawn several times in a breeding season. The weight of active, prespawning ovaries increases at a rate approximately propor- tional to the third power of the length of the fish. Ovary weight increases in a linear fashion with age in the Catalina population. The rate of increase with age would be less in the Guadalupe population due to dwarfing. The three coloration phases of P. pulchrum are described, two of which are found in adult individuals. The uniform coloration is made up mostly of mature female and immature fishes. About 5% of the mature uniform individuals were males at Catalina, and about 12% at Guadalupe. The bicolored phase is made up exclusively of males and late transitional individuals. Data from field transects revealed that there were about five uniform individuals to every bicolored male. Based on an estimated yearly survival rate of about 0.7, the mature sex ratio at Catalina was approximately two females for every male. The ratio at Guadalupe was closer to three females for every two males, due in part to the earlier sex changes seen there. Sequential hermaphroditism, a phenomenon characterized by an individual changing from one sex to another at some point in its life history, is widespread in teleost fishes (Atz 1964; Reinboth 1970). In some species, individuals change from male to female (protandry) and in others the sit- uation is the reverse (protogyny). Most of the published information on the life histories of sequentially hermaphroditic species has dealt with the distribution of the sexes with size, sometimes correlated with a histological investigation of the gonads (Atz 1964; Reinboth 1970). However, in order to interpret the full implications of the sexual patterns seen in sequential hermaphrodites, data on the age dis- tribution, age-specific fecundity, and the sexual transformation schedule of the population are needed (Warner in press). 'Scripps Institution of Oceanography, University of Califor- nia, San Diego, La JoUa, CA 92037; present address, Smithsonian Tropical Research Institute, P.O. Box 2072, Balboa, Canal Zone. There are a few protogynous fish species for which the information is nearly complete. For example, Moe (1969) provided excellent data on the life history pattern, gonadal transformation, and survival rate of the serranid Epinephelus morio. Natural sex reversal in the synbranchid Monopterus albus has been extensively studied both in the field and laboratory (Liem 1963, 1968; Chan 1971), but little is known about its age- specific fecundity and survival pattern. A similar situation exists for the labrid Corisjulis (Reinboth 1957, 1962; Roede 1966), where again we lack infor- mation on the demography of the population. Among the Labridae, perhaps the most complete information exists on the seven Caribbean species of the genera Thalassoma, Halichoeres, and Hemipteronotus studied by Roede (1972). An un- fortunate limitation was placed on Roede's work by the tropical location, which precluded age de- termination from growth rings on scales or otoliths. Manuscript accepted July 1974. FISHERY BULLETIN; VOL. 73, NO. 2, 1975. 262 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM Pimelometopon pulchrum (Ayres), the Califor- nia sheephead, is a labrid of the subfamily Bodianinae. It is confined to temperate waters, ranging from Monterey Bay, Calif., to Cabo San Lucas at the tip of Baja California, Mexico (Miller and Lea 1972). Individuals can reach a large size (over 800 mm standard length [SL]) and are com- monly found off southern California along rocky shores at depths betv^een 5 and 50 m. In this report, it is demonstrated that P. pulchrum, like many other labrids, is a protogynous her- maphrodite. In addition, data are presented on age and growth, on the distribution of the sexes in relation to color, size, and age, and on the observed patterns of fecundity and survival. The study embraces two widely separated populations, chosen to reflect how differences in the demography of the population might lead to the observed differences in the schedule of sexual transformation (discussed in Warner in press). MATERIALS AND METHODS Source of Materials and Times of Sampling Pimelometopon pulchrum was taken by means of a hand spear while either skin diving or using scuba. The main collecting area was at Fisher- man's Cove on the northeast shore of Santa Cat- alina Island, Calif., near the University of Southern California Marine Station (lat. 33°27'N, long. 118°29'W). A total of 341 individuals of P. pulchrum were processed from samples taken the year round at monthly intervals. Collections began in December 1969 and continued, with occasional gaps, until July 1971; monthly samples were between 20 and 30 individuals. The other area sampled in this study was at Guadalupe Island, Mexico, located approximately 200 km west of Punta Baja, Baja California. Collections were made along the protected east side of the island, concentrating on an area 3 km from the southern tip known as Lobster Camp (lat. 29°01'N, long. 118°14'W). Year-round sampling at Guadalupe Island was not possible, and the 130 individuals taken there were from three expedi- tions, January 1970 (16 specimens), April 1970 (53 specimens), and May 1971 (61 specimens). Supplemental collections were made at La JoUa, Calif., including a sample of large individuals from a spearfishing meet on 19 July 1970. The standard length of each fish was measured. and its coloration noted. Several dorsal spines were removed and frozen, and the gonads were fixed in bouin's fluid. Age Determination Methods Age determination by counting annular marks on the otoliths or scales was precluded in P. pulchrum. The otoliths are extremely small and difficult to locate, and the central portions of nearly all the scales were either clear or irregularly banded, indicating regeneration. The bones and spines of P. pulchrum did show regular markings, and younger fish could be suc- cessfully aged by counting the marks on either the bones (opercula or cranial ridges) or the dorsal spines. However, the proximal portions of the bones tended to thicken and obscure the earlier marks on older California sheephead and only dorsal spine annuli could be used for age deter- mination. Dorsal spines were prepared as follows: the flesh was removed by means of a household enzyme product (Ossian 1970) and the spines were air dried. The classical methods of decalcification and/or thin sectioning (e.g., Cuerrier 1951) were not used. Instead, a high-speed grinding tool with a thin abrasive disc was used to cut cleanly through the spine at a point just distal to the swollen portion of the base. The spinous portion was then thrust through an opaque light shield so that only the cut base protruded. A strong microscope light was directed to the lower portion of the shield so that the only light visible on the other side then came through the projecting base of the spine. The hyaline layers of the spine transmit much more light and the illuminated pattern, resembling tree rings, is easily seen in a dissecting microscope. The second dorsal spine was used for primary counts; the ring patterns on the other spines were identical, and were used to verify counts for in- dividuals. Counts on each spine were made by two people and were used in the analysis of growth only when they agreed. False rings, probably caused by abnormal growth conditions, were identifiable in young individuals by their proximity to other annuli and their tendency to be incomplete. Rarely, older fish showed a marked degenera- tion of the central portion of the spine, which became hollow and oil-filled, making age deter- mination from spines impossible. 263 FISHERY BULLETIN: VOL. 73, NO. 2 A series of measurements of 100 spines was made with an ocular micrometer at a magnifica- tion of 30 X. At tiiis magnification, one ocular micrometer unit equals 0.033 mm. The radius was measured at midspine on a line perpendicular to and beginning at the indentation axis. Distances from the center of the spine to each annulus were recorded for back calculation of length, along the radius line. Finally, the distance from the spine margin to the outermost annulus was measured for determination of the time of annulus forma- tion. Methods of Reproductive Biology In the laboratory, each gonad was blotted dry, weighed, and a segment of one lobe was dehydrated in alcohol and embedded in paraffin. Slides were prepared of cross-sections of the lobe, cut at thicknesses of 5, 10, and 25 ^m; thicker sec- tions have less tendency to collapse and were made to ensure that the overall configuration of the cross-section could be observed. Sections were stained with ehrlich's hematoxylin and eosin. Each gonad was classified according to sex and state of development. Assignment of a develop- mental class depended on the predominate stage of gametogenesis seen in the gonad. The division of gametogenic stages is as follows: Oogenesis was divided into five stages, follow- ing criteria detailed for a variety of species by Kraft and Peters (1963). Smith (1965), and Moe (1969). Stage 1. Very small (15-30 /xm in diameter) oocytes with a large nucleus, single nucleolus, and a relatively small amount of basophilic cytoplasm. Stage 2. (30-50 ju,m) Previtellogenic oocytes with a strongly basophilic cytoplasm and multiple nucleoli around the nucleus margin. Stage 3. (150-300 [xm) Vitellogenesis begins with the deposition of yolk vesicles in the less darkly staining cytoplasm. A thin zona radiata can be seen in late stage 3. Stage 4. (280-450 /^m) Cytoplasm filled with yolk vesicles and globules; the zona radiata well developed and strongly acidophilic. Stage 5. (450-1,050 ixm) Mature or nearly ma- ture oocytes, uniform in appearance due to the coalescence of the yolk globules. The nucleus is eccentric and the zona radiata is thin and non- striated. These oocytes are often extremely irregular in outline and Roede (1972), who noted the same irregularity in mature eggs of other labrids, probably correctly attributed this to dis- tortion during fixation and staining. This stage was seldom seen, but several specimens were seen with eggs in the ovarian lumen and stage 5 oocytes still within the follicle. Spermatogenesis occurs in small crypts, in which all the cells are at the same stage. The development and appearance of the sperma- togonia, primary spermatocytes, secondary sper- matocytes, spermatids, and mature sperm follows very closely the descriptions given by Hyder (1969) for Tilapia and by Moe (1969) for Epinephelus morio, and will not be repeated here. The gonadal development classes, intended to parallel those of Moe (1969) and Smith (1965), were designated as follows: Class 1. Immature female. Stages 1 and 2 oocytes present, atretic or brown bodies (Chan et al. 1967) absent. The ovarian lamellae are pressed closely together and the lumen is small. Class 2. Resting mature female. Oocyte stages 1, 2, and 3 present, with stage 2 predominating. Atretic bodies are usually present. Class 3. Active mature female. Oocyte stages 3 and 4 predominate in the lamellae. In late class 3, stage 5 oocytes are also present. Class 4. Postspawning female. Ovary is disrupted, with many empty follicles in the lamellae. Some degenerating stages 4 and 5 oocytes are usually found in the lamellae and lumen, respectively. Class 5. Transitional. Seminiferous crypts begin to proliferate in the lamellae, but some stage 2 oocytes can be seen. These oocytes degenerate and decrease in number as spermatogenic activity begins to dominate the gonad. Class 6. Inactive male. Crypts containing primary and secondary spermatocytes pre- dominate; few spermatids and mature sperm are seen. Class 7. Active male. Spermatids and tailed sperm increase in abundance until, in the ripe phase, sperm are densely packed in the collecting ducts and many crypts have coalesced. Class 8. Postspawning male. Ducts are still ex- panded, but few sperm can be seen in them. Many new crypts containing spermatogonia are present. This apparently is a short-lived stage that rapidly gives way to the resting (class 6) testis. Fecundity determinations were made by count- 264 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM ing yolky oocytes. A thick cross-section of the ovary was cut from near the middle of the lobe, weighed, and then agitated to dislodge as many oocytes as possible from the ovarian lamellae. Oocytes remaining in the lamellae were teased out so that a complete count could be made. An es- timate of the number of yolky oocytes per gram of ovary could then be made directly from the sample. The total number of eggs in the ovary was then approximated by multiplying by the total weight of the ovary. Relative abundance of coloration types was es- timated directly from field observations. To eliminate the effects of any differential depth distribution, visual transects were either run per- pendicular to depth contours or were compiled from a series of equal length runs parallel to suc- cessive contours. Transects were approximately 50 m in length. The number of California sheephead in each color phase was recorded. It was assumed that both coloration types are equally visible, and this is probably valid. California sheephead are not secretive when adults; only juveniles tend to remain close to cover. Larger males appear warier than other individuals, but still remain in sight. The problem in observing California sheephead is not in avoidance, but inquisitiveness. Occasionally transects had to be aborted because of the ten- dency of Pimelometopon to follow the diver. RESULTS Age and Growth Van Oosten (1929) set forth criteria for the ac- ceptance of annuli on scales or bones as yearly marks. These criteria apply equally well to spines, and are as follows: (1) The spine must remain con- stant in identity and grow proportionally with the fish. (2) Only one mark must be formed each year. (3) The body lengths calculated by using prior an- nuli on the spine (back-calculated lengths) should agree with the actual lengths of younger age groups. The criteria will be discussed in order. (1) Dorsal spines were certainly constantly identifiable in all individuals of P. pulchrum examined. The relationship of spine radius to standard length (Figure 1) is satisfactorily expressed in a linear fashion (r = 0.787) and there is no apparent indication of allometric growth of the spine, at least for fishes of lengths greater than 130 mm. Much of the scatter in the data is due to variability in the location of the cut made across the tapering spine. (2) The increment of distance from the last an- nulus to the outer edge of the spine should increase with the time since the formation of that mark. If one mark is formed each year at a particular time, the average marginal increment should drop to near zero at the time of annulus formation, then steadily increase for the rest of the year. This pattern is shown (Figure 2) for 77 California sheephead from Catalina taken throughout the year. Successive age groups did not differ in time of annulus formation, so the data are combined for all aged fish. The distinct hyaline bands appeared to be formed in June and July, at the beginning of the period of warming water in the Catalina area (Quast 1968). Formation of growth marks has been found to occur in other inshore California fishes at a similar time (Joseph 1962; Norris 1963; Clarke 1970). Ring formation also overlaps with the ini- tiation of reproductive activity, although egg production and spawning continue well into Sep- tember (see below). (3) Lengths of P. pulchrum at previous ages were calculated by a modified direct-propor- >j-r\j • > 310 - . • / • • • • / ■ / • 280 - • 7 *• - • • / e 250 _ • • / • / I - • ;•./• • o - t / z • / • uj 220 — / • • _) >••/••* • Q ••/ • < •• /.: • Q • /• z 190 < », • •/• • t- m • '/ ** 160 • / • • • / Y=2I8. 38+7.92 (x-21. 61) 130 ~ / / r = 0.787 100^ T- / ^ 1 1 1 / 1 1 1 1 1 i i ^— O 6 12 18 24 30 36 SPINE RADIUS (ocular micrometer units) Figure 1.— The relationship of dorsal spine radius to standard length for 117 specimens of Pimelometopon pulchrum from Cat- alina Island. One ocular micrometer unit equals 0.033 mm at 30 X . 265 FISHERY BULLETIN: VOL. 73, NO. 2 t- i2 5 bJ c > UJ Q) rr o a> E <> k- _l o < F ^ O D CE Z3 < o ^ o ^ 2 0 J-F M-A M-J BIMONTHLY J-A S-0 INTERVAL N-D Figure 2.-Mean marginal increments for six bimonthly inter- vals from 78 specimens of Pimelometopon pulchrum from Cat- alina Island. Sample sizes are shown for each interval, and 95% confidence limits for the mean are drawn on either side of each point. Table 1 shows the back-calculated lengths for 100 California sheephead from Catalina over eight age groups, derived from spine radius measurements. The means from back calculation are also given in Figure 3 for comparison with empirical data. The mean standard lengths for each age (Figure 3) demonstrate good agreement with the back-calculated data. There appears to be a slight slowing of growth after the fourth year in the Catalina California sheephead population. This may reflect the onset of a diversion of a significant amount of energy into egg production, since most 4-yr-old fish examined were mature females (see below). A second period of more rapid growth is suggested after the seventh year, at an age where many of the Catalina California sheephead are beginning to transform from female to male. There is no evidence for a decrease in the rate of growth up to age 13, where the average standard length is 470 mm. Pimelometopon pulchrum is quite capable of growing larger than this, and some individuals Table 1. -Back-calculated lengths for age groups 1 through 8 of Pimelometopon pulchrum from Catalina Island. Age Mean length of subsample Mean length of total sample Back-calculated len gths (mm) for ages group N (mm) (mm) 1 2 3 4 6 6 7 8 1 8 100 116 97 2 16 158 155 100 127 3 22 198 197 106 139 168 4 17 231 238 124 152 178 200 5 11 246 245 130 164 185 203 225 6 11 286 272 129 159 191 212 230 251 7 7 289 294 128 163 190 225 247 274 295 8 8 359 368 145 183 220 254 279 299 320 343 Overall means of calculated lengths 117 150 184 214 242 272 308 343 Number of ind ivic uals 100 92 76 54 37 26 15 8 tionality method given by Rounsefell and Everhart (1953) as follows: L'-C L-C ^ S where L = length of the fish at the time the spine was obtained, L' = length at the time a particular annulus was formed, S = total length of the spine radius, and S' = length along the spine radius to the annulus in question. The term C is a factor used to correct for the length obtained before the spine was formed, and is estimated by the inter- cept of the length axis on a fish length versus spine radius plot (Figure 1). In the case of the Catalina California sheephead population, C was equal to 47.2 mm. have very long lifespans. Fitch and Lavenberg (1971) mention a 32-inch (815-mm) male aged at 53 yr, and an 8.3-kg female, no length given, that was 30 yr old. Although exact age determination becomes difficult for large and old individuals, it is occasionally possible. The largest California sheephead encountered in this study were a 592 mm SL male, 20 yr of age, and a 538 mm SL male which had lived 18 yr. Size-age distributions can vary for different locations. In a sample taken by the California Department of Fish and Game at a spearfishing meet at San Pedro, Calif., on 28 March 1971, the mean standard length for males was 661 mm (range 545-745 mm) and for females was 450 mm (range 294-656 mm). The pattern at Guadalupe Island is different from that at Catalina (Figure 4). While the sample 266 WAKNEK: KKFKUUUCTIVE BIOLOGY OF PIMEWMETOPON PULCHRUM 480 E E^ X »- Q < 400 17 0 1 0 16 Totals 341 111 153 12 65 Table 3.-Frequeney of sexual types in each 20-mm size class for the Guadalupe Island population of Pimelometopon pulchrum. Standard length (mm) Number of fish Immature Mature female Transitional Mature Male <100 100-119 120-139 140-159 160-179 180-199 200-219 220-239 240-259 260-279 280-299 300-319 320-339 340-359 360-379 380-399 >400 5 5 0 5 5 0 3 3 0 8 1 7 12 0 9 15 0 10 16 0 5 11 0 2 16 0 4 10 0 2 6 0 1 7 0 1 7 0 1 2 0 0 3 0 0 1 0 0 3 0 0 0 0 0 0 3 5 10 8 10 6 5 6 6 2 3 1 3 Totals 130 14 42 68 largest sizes. Transitionals were found in inter- mediate sizes, in numbers which varied seasonally (see below). At Catalina, most California sheephead mature at standard lengths between 190 and 230 mm. Sexual transformation occurs over a broader size range beginning at 250 mm, with a peak of ac- tivity apparently occurring at standard lengths between 310 and 330 mm. The dwarfing phenomenon found in the Guadalupe population is again evident (Table 3, Figure 8). Maturity begins at a length near 140 mm, and the majority of individuals are male by a length of 210 mm. Peak transformation activity appears to occur in the population in fishes rang- ing from 190 to 230 mm in standard length. The actual time courses for all these events become evident when the relative frequencies of 100 80 60 40 20 0 100 80 60 - 40 - 20 CATALINA U- GUADALUPE ( ^^ " \ / r^ 60 - \/ >) — -^ / ■ v 40 - J \ " / 1 >* ' 20 - ' ■> / / \ .,.o — o... 0 1 i- »^ 1 •--- 1 A — a immature GUADALUPE 0-— O T Q— <3 transitional 100 .A... ' 1 ^ . . o" 80 - ■ ' \ y^^^^^ 1 / • 1 y^ 60 - \l \ •^ b """S/^ 40 - 1 \ /-""^^X 1 > / > 20 - 1 V^"^^- ■'' 0 1 •/-' , '-^'' \'' \/ ^ 3 4 5 6 7 AGE (YEARS) 10* Figure 9.— Proportions of sexual types in each year class of Pimelometopon pulchrum from Catalina Island (top) and Guadalupe Island (bottom). The last age grouping consists of all fishes 10 or more years old. In both populations, sexual maturity begins in the fourth year of life for virtually all members. By using age groupings, the skewness introduced by the dwarfing at Guadalupe is removed, and differences in the transformation activity time schedule in the two populations are revealed. At Guadalupe, males are present in essentially the same abundance as females in age classes 5 and 6, and strongly predominate at age 7 and thereafter. Therefore the majority of California sheephead at Guadalupe Island spend no more than 2 or 3 yr as functional females. Transformation generally occurs later in the Catalina population. Most individuals are func- tional females for at least 4 yr, and males predominate only after age 8. Distribution of Gonad Development Classes with Time The active state of gonads may be determined directly through histological examination or in- ferred from the appearance and the size of gonad (gonad indices). The seasonal distribution of mature gonad 14 12 ;Class5 ; t. .:■: JFMAMJJ ASOND MONTH Figure lO.-Number of individuals of Pimelometopon pulchrum in each mature gonadal development class from monthly samples taken at Catalina Island. development classes for 166 California sheephead from Catalina is shown in Figure 10. As expected, immature fishes, which are not shown in the Figure, occur throughout the year. Resting stage females (class 2) were encountered from August through May, and predominated from October to April. Active females (class 3) were present May 271 FISHERY BULLETIN: VOL. 73, NO. 2 through September. Late class 3 gonads were seen in July, August, and September, and most spawn- ing activity probably takes place in these months. Females with postspawning ovaries (class 4) were captured in low numbers from August to early October. Class 4 appears short-lived, quickly receding into a resting class (class 2) or transi- tional (class 5) phase. Transitional individuals were found only from October to March at Catalina. Those taken in Oc- tober and November were all in the early stages of transformation, with many stage 2 oocytes and a few scattered spermatogenic crypts in evidence. Transitionals captured in February at Catalina or in May at Guadalupe were more advanced, with few stage 2 oocytes and spermatogenic crypts dominating the gonad. Most males at Catalina were inactive (class 6) from October through April, closely paralleling the period seen for females. Active gonads predominated in samples from fish taken in May through September. Again the pattern suggests that spawning activity takes place from August through early October. Further support for designating this period as the spawning season comes from the gonad indices (Figure 11) of Catalina females caught in different months. These reflect a similar pattern seen in the analysis of gonad development states. After a quiescent period from October through April, the ovaries begin to increase in size until a maximum is reached in June and July. Spawning reduces the average index steadily from then until 90 90 70 E E 60 o 50 z o 40 30 20 10 - 1 1 5 1 - / / L ~ / \ e ; I / rill \ 3 10 1 1 1 \ September. The resting value is then seen again, remaining constant through the winter. The index used was gonad weight scaled to compensate for different lengths of individuals. When only the mature females less than 310 mm in standard length are included in the analysis, the relationship between gonad weight and standard weight is sufficiently linear (Pearson correlation coefficient = 0.845 [P < 0.001] on 24 individuals caught in June) that the use of the following for- mula is justified: Gonad index = (9°"ad weight in grams) (100) (standard length in mm) The size range used includes the great majority of reproductive females at Catalina. An analysis of the spawning season of P. pulchrum at Guadalupe Island was not possible due to the lack of year-round sampling. Multiple Spawning and Fecundity Two or more distinct groups of ripening oocytes were usually apparent in the ovaries of P. pulchrum examined in June and July. The size distribution of yolky oocytes in an ovarian cross- section (Figure 12) from a female, 244 mm SL, captured at Catalina in mid-July, shows that there is one group of eggs ready to be spawned, and that two other distinct groups are undergoing vi- tellogenesis. This type of successive maturation of several groups of oocytes is termed asynchrony, and is characteristic of species that have com- paratively long breeding seasons and multiple spawnings by individuals within each season (Yamamoto and Yamazaki 1961). J J MONTH 4 6 8 10 12 14 16 18 OOCYTE DIAMETER (ocular micrometer units) Figure 11. -Average gonad indices for monthly samples of ma- ture female Pimelometopon pulchrum., all of standard lengths less than 300 mm. Sample sizes are shown above the bracketed lines, which are the 95% confidence limits of the mean. Figure 12.-Size distribution of yolky oocytes in an ovarian cross-section of a 244 mm SL female of Pimelometopon pulchrum captured 22 July at Catalina Island. The oocytes are also clas- sified according to their degree of development. 272 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM Further evidence for multiple spawning is seen in the ovaries from some females captured in August and early September. There were a few mature eggs free in the lumen and numerous empty follicles in the lamellae, both indications of recent spawning. At the same time, another group of vitellogenic oocytes were observed developing in the lamellae and these would presumably have been spawned at a later time. As Yamamoto and Yamazaki (1961) point out, the presence of multiple spawning complicates any determination of the number of eggs produced each year by an individual fish. Es- timates can be made from an analysis over time of frequencies of egg diameters, such as that done by Clark (1934) for Sardinops caerula. Such analyses require a large sample over the mature size range and this was not available for P. pulchrum. Counts of the yolky oocytes in subsamples of ovaries made for California sheephead females captured in July (Table 5) are probably overes- timates of the number of eggs spawned during the Table 5.-Estimates of the total number of vitellogenic oocytes and density of those oocytes in the ovaries of Pimelometopon pulchrum captured at Catalina Island in July 1970. Standard Weight of length Date of ovary (mm) capture (g) Estimated Estimated number of number of yoll ^30 o - • •• • • / • • / • / / • • / • • / / • • • •/ / • 20 • • • / • / / • • • • • •• W=l.3lxlO"^ L^-^^ 10 • • • • • r = 0.787 0 yji > 1 , 1 . 1 1 1 , 1 , 1 , 1 . 1 1 1 20 22 24 26 28 30 32 34 36 38 STANDARD LENGTH (cm) Figure 13.— Ovary weight versus standard length for females of Pimelometopon pulchrum captured in June and July at Catalina Island. 273 FISHERY BULLETIN: VOL. 73, NO. 2 length conforms to a simple cube law relationship, which would be expected if gonad weight remains some constant proportion of the total weight. The exponent 2.95 was determined by a least-squares regression fit to a logarithmic transformation of the data. The confidence limits around the regres- sion line become increasingly large with higher values of W and L, and the curve should not be used for extrapolations beyond the range of the data. Ovary weights in relation to age are shown in Figure 14. There are few data for the older age classes, but there is a definite positive correlation of the fecundity and the age of the individual. Coloration, Sex, and Field Distribution of Coloration Types The California sheephead is found in three main color phases (Crozier 1966), all of which are closely correlated with sexual state. For the first year, P. pulchrum has juvenile coloration, a gold or salmon body color with black spots on the anal fin, the anterior and posterior portions of the dorsal fins, and on the caudal peduncle, and with a silver lateral stripe extending from the eye to the caudal fin. Crozier (1966) stated that the initial body color was gold, and this was gradually replaced by the reddish adult shade. The juvenile coloration was seldom seen in individuals over 100 mm SL, and has never been found in sexually mature individuals. The most common color pattern of P. pulchrum is a uniform rose or salmon color, covering the entire body with the exception of the chin, which is usually white in mature individuals. The median and pelvic fins are darker than the body, ranging from dusky red to black. The pectoral fins usually match body color. Uniform coloration may be ob- scured by a melanistic condition which causes the entire body to appear brown. This occurs in vary- ing degrees, making the fish appear almost black in extreme cases. Eleven percent of the uniformly colored fish captured at Catalina were designated melanistic, as were one-third of all uniform types at Guadalupe. A uniform coloration is characteristic of imma- ture fish as well as mature females. Histological analysis indicated, however, that the relationship was not perfect. At Catalina, 3.5% of the in- dividuals designated uniform in color were dis- covered to have male gonads. When only sexually mature uniformly colored California sheephead were tallied, 5.1% were male. These males ranged 70 60 50 E iD 40 UJ >- a: §30 o 20 10 0 W=8.88L- 26.25 r = 0.802 VA- 4 5 6 7 8 9 AGE (years) Figure 14.-0vary weight versus age for females of Pimelomet- opon pulchrum captured in June and July at Catalina Island. in length from 245 to 315 mm. and were captured in June, July, and December. Examination of a large series of gonad cross-sections from these in- dividuals revealed a few stage 2 oocytes within the lamellae from two captured in July. This can be taken as evidence for a recent sex change. All the others had gonads of completely normal male ap- pearance. Field notes revealed that all of these individuals were melanistic, and 70% were recorded as having some male external characteristics, such as a small nuchal hump or slight differential darkenings of either the head or tail regions (see below). This suggests the possibility that some of these in- dividuals may have been incorrectly typed due to the ground color being obscured by melanin. 274 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM Males in uniform coloration are more frequent in the Guadalupe population. A total of 6 out of 57 (10.5%) uniformly colored individuals vi^ere males. Elimination of immature fish from the count raises the figure to 12.3% males. Four of the six were melanistic, one of these with slight male char- acteristics. The other two individuals possessed normal uniform coloration with no darkening. In the third color phase the head region, includ- ing the opercle, is dark brown or black. The chin remains white, and the midsection retains the reddish hue of the uniform type. The caudal por- tion, beginning approximately on a line connect- ing the initial soft rays of the dorsal fin with the anterior limit of the anal fin, is also dark brown or black. The median fins and pelvics remain generally dark in color, and the pectorals may acquire a dark band at their tips. This coloration is found exclusively in males and some transitionals (see below). It is usually ac- companied by two other male secondary sexual features common in the Bodianinae, the nuchal hump and filamentous extensions of the median fins. The hump appears to increase in relative size as the male gets larger, making the head appear increasingly angular in profile. No individual with this bicolored pattern was found to have func- tional ovaries. During the breeding season, the pattern serves as an excellent indicator of a func- tional male. Individuals classified as transitional varied in coloration. Of 11 transforming California sheephead for which coloration records exist, 3 were scored uniform in color and 2 as bicolored. The remaining 6 were recorded as intermediate in coloration, usually involving a slight darkening of the head, caudal region, or both. The three uniform individuals were classed as early transitionals (large amounts of stage 2 oocytes still in the gonad); the two bicolored fishes were classed as late transitionals (only a few degenerating oocytes in the gonad cross-sections). The distribution of uniform and bicolored types in field populations, determined by visual tran- sects (Table 6) shows that bicolored males are present in remarkably similar proportions in both localities, occurring in a ratio of about 5.5 uniform individuals to every bicolored individual. Con- fidence limits for estimating the proportion of bicolored individuals were calculated from a binomial distribution, n = 216 and 407 for Cat- alina Island and Guadalupe Island, respectively (Dixon and Massey 1969). Table 6.-Numbers of coloration types in two populations of Pimelometopon pulchrum, determined by a series of visual tran- sects. No. of No. of Proportion p of No. of uniform bicolored bicolored types Locality transects type type and 95% conf. limit Catalina Island 70 183 33 0.153 ±0.048 Guadalupe Island 93 343 64 0.157 ±0.035 DISCUSSION Anatomical Features of the Gonad and Sexual Transformation The ovary of P. pulchrum is essentially identical with that of the labrid Coris julis, which was studied in detail by Reinboth (1962). Reinboth, however, did distinguish between the testes of those C. julis born as males (primary males) and those that become males through sex reversal (secondary males). In the former, the testis ap- pears rather solid and flattened, and sperm are transported by means of a single vas deferens in each lobe. The secondary male has a testis like that described here for P. pulchrum. The two types differ in the structure of the vas deferens posterior to the gonadal lobes, which surrounds the old oviduct in secondary males, but is a simple tube in primary males (Reinboth 1970). When primary and secondary males are present in a single species, Reinboth (1970) termed the species diandric. When only secondary males are present, the species is termed monandric. To Reinboth's (1970) list of monandric species {Labrus turdus, L. merula, L. bergytta, Hemipteronotus novacula, and possibly L. bimaculatus) we may add P. pulchrum. Other labrid species have been studied without regard for the primary-secondary male phenomenon (Atz 1964; Reinboth 1970), and cannot be categorized with certainty as monandric or diandric. The transition from a functional ovary to a tes- tis has been described in detail for labrid fishes in both naturally occurring situations (Reinboth 1962; Sordi 1962; Okada 1962; Roede 1972) and under the influence of hormone administration (Reinboth 1962, 1963; Roede 1972). These reports are essentially in agreement with the present ob- servations on P. pulchrum. There is no evidence of synchronous hermaphroditism (Atz 1964:147) in the Labridae, but Robertson (1972) found sperma- togenic crypts in the ovaries of 28 of 29 females of 275 FISHERY BULLETIN: VOL. 73, NO. 2 Labroides dimidiatus, and 15 of these had crypts with sperm or spermatids. Thus, the possibility of encountering a synchronously hermaphroditic labrid species should not be ruled out. Transformation Schedule Pimelometopon pulchrum fits the general labrid pattern of size and sex distribution. No males were found smaller than 230 mm SL at Catalina Island, and none smaller than 150 mm at Guadalupe. The longer size classes (above 350 mm and 230 mm at Catalina and Guadalupe, respectively) contain mostly males. Data for labrid sexuality are usually in the form of size frequency distributions of males and females within a species (Atz 1964; Remade 1970; Roede 1972). Some of these studies have been con- founded by the presence of two distinct color pat- terns, the investigators wrongly assuming strict sexual dichromatism (see below). An additional complication is the possibility of two different types of males being present in some species, usually with different life histories and behavior (Reinboth 1970). Both of these problems are eliminated by histological examinations of the gonad, which also reveals the presence of inter- sexual or transitional individuals. The absence of males from the smaller size classes at least suggests protogyny. However, similar patterns can also result from samples of a species exhibiting differential growth rates for the sexes (e.g., see Strasburg 1970, for weight and sex distributions of blue marlin, Makaira nigricans), and this should be taken into con- sideration. Fourteen of the fifteen labrid species either reviewed or dealt with originally by Roede (1972) had similar patterns of length-sex distribution. Females predominated in the smaller size classes, males in the larger. The proportion of males in the smaller sizes (usually associated with a particular color pattern; see below) varied from practically none in some species of Halichoeres and Hemipteronotus up to nearly 30% for Stethojulis strigivenfer (Randall 1955). Males became increasingly common as length increased and the longest size classes consisted almost exclusively of males. Species of the genus Symphodus (Crenilabrus) appeared to exhibit a different pattern, with nearly equal numbers of males and females in the small size classes (Soljan 1930a, b). Remade (1970) believed that sex reversal is a rare phenomenon in this genus. Age determinations allow several more inferences about the sexual life history of a species. When few or no young males can be found, there is strong evidence for protogyny since the possibility of differential growth is eliminated. The rate of transformation in different age classes can be estimated, and this provides an idea of how long the average individual spends in different sexual phases. Finally, by comparing the distribution of sex versus length with sex versus the age of the individual, it may be possible to assign a more critical role to one or the other as a causative factor for sex reversal. The age at first maturity (4 yr) does not differ for P. pulchrum at Catalina and Guadalupe Islands, but the distribution of sexual transfor- mation with age differs markedly. Most in- dividuals in both populations function at least 1 yr as females. At Guadalupe, many change sex after 1 yr, and most are males within 3 yr after matur- ing. Most sex reversals occur at Catalina between the seventh and eighth year, 4 yr after maturing as a female. Some individuals remain female for shorter or longer periods of time. The oldest female encountered in this study was 17 yr old. The dwarfing phenomenon at Guadalupe, which should bring about a slower rate of increase of fecundity with age, would have enough effect to decrease the optimum age of transformation in that population when compared to the Catalina population. This will be discussed in detail else- where (Warner in press). Lonnberg and Gustafson (1937) determined the ages of a series of specimens of Labrus bimacula- tus (as L. ossifagus) which they correlated with sexual state. They found that sex reversal occurred in individuals from age seven onward, and was associated with a color change from red to blue- striped. Females were found in diminishing numbers up to age 18, mostly confined to the red phase. Males in the red phase ranged from around 3 to 7 yr old; blue-striped males got as old as 25. Other protogynous teleosts which have been investigated regarding age of transformation show a variety of patterns in the distribution of sex reversal over their life-span. Liem (1963) demonstrated that sex transformation in the synbranchid rice eel Monopterus albus occurs mainly when individuals reach about 30 mo of age (about 35 cm in length). Few fishes in nature deviated from this pattern. Liem (1963) was able 276 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM to induce earlier transformations by starving the individuals. Moe (1969) has carefully worked out the age distribution of sex reversal in the serranid Epinephelus morio, and found a rather smooth period of transition from female to male over at least 5 yr (ages 5 to 10), at a rate of about 15% of the individuals in a year class reversing per year. In a less comprehensive survey, McErlean and Smith (1964) estimated that transformation oc- curred at age 10 or 11 in Mycteroperca microlepis (Serranidae), and speculated that the age of the fish had more effect on sex reversal than the length. To determine the effect of individual size on sex transformation in P. pulchrum, mean lengths of males and females in each age class were com- pared (Figure 15). If length is closely related to sex reversal, one would expect males to be larger than females of the same age, and this was found in both the Catalina and Guadalupe populations. In every age grouping where sample size permitted statistical analysis, males tended to be larger than females. Five of the seven groups tested (one-sid- 350 - E E -300 X I- d z LlI Q < 250 Q Z 200 150 Jr 6UADALUPE IS. AGE (YEARS) Figure 15.-Mean lengths for successive age classes of males and females of Pimelomefopon pulchrum. Sample sizes are shown for each point, and standard error brackets are given when samples are large enough. For clarity, standard error brackets for males point to the right, and brackets for females point to the left. ed ^-test for difference in means) were sig- nificantly different at the 5% level or less, and the remaining two were significant at the 10% level. An assessment of the effect of age on sex reversal was made in similar fashion, comparing the mean ages of males and females in successive size groupings (Figure 16). If sex reversal were closely related to the age of an individual, then males would tend to be older than females in a given length group. The relationship between age and sex is less strong (Figure 16). Sample sizes are not large, and the range of ages encountered in a sample is small relative to measured length values, so fewer significant results might be expected. Only one size group was found where males were significantly older than the females. However, the existence of several large negative t values in groups where the age of females is greater than that of males supports the idea that size is more important than age in effecting sex change. The high average age of females in the larger size groupings of both populations (Figure 16) was not expected, and suggests that the individual growth rate may also be involved in sex reversal. The large separation between male and female mean ages begins with the 300-mm size grouping at Catalina, and the 200-mm group at Guadalupe. Inspection of Figure 8 reveals that at about these lengths, the proportions of males and females un- dergo an abrupt shift. A large percentage of the individuals in the populations apparently reverse sex at these sizes. Furthermore, the difference in ages between males and females of lengths above these "critical" sizes appears to be significant. The mean age of females larger than 300 mm at Cat- alina is 7.9, a full year older than males in the same size range (^31 = 1.51, P<0.10). Similarly, females larger than 200 mm at Guadalupe have a mean age of 9.5 yr, and males at that size range average 7.0 yr (^32 = 2.80, P< 0.001). Thus the females that pass through the "critical" lengths without changing sex appear to be those individuals with relatively low rates of growth, suggesting both that slow growing individuals tend to be refrac- tory to sex change, and that fishes with high rates of growth change sex more readily. The data of Figure 15 support this idea, as males are faster growing (larger) members of each age class. A check of the back-calculation information revealed that the growth rates of the large females were consistently low throughout their lifetime, and those of the small males had been high relative to other members of the age class. 277 FISHERY BULLETIN: VOL. 73, NO. 2 10 en en >^ LU < r CATALINA IS. Figure 16.-Mean ages for suc- cessive 20 mm standard length groupings of male and female Pimelometopon pulchrum. Sample sizes and standard errors are shown as in Figure 15. 170 190 210 230 250 270 290 310 330 350 370 STANDARD LENGTH (20mm groups) The best picture, then, that can be drawn from the present information is that rapidly growing individuals may transform sooner than other fishes of the same age. The bulk of the population, growing at the average rate, eventually reaches a "critical" size where most of them change sex. Fishes that grow slowly may not change sex at all. The Breeding Season, Multiple Spawning, and Fecundity Breder and Rosen (1966) have summarized the information available on the spawning seasons of labrids. In temperate species, most activity occurs over a period of approximately 3 mo, most com- monly in April, May, and June. The Catalina population of P. pulchrum is exceptional in this case, since spawning occurs from August to Oc- tober. The two other wrasses commonly found at Catalina Island also spawn later in the year than other labrids. Oxyjulis californica spawns from May until October (Bolin 1930), and Halichoeres semicinctus probably spawns in late June, July, and August (D. R. Diener, pers. commun.). The relatively late spawning seasons of these species may be caused by upwelling along the southern California coast which usually persists well into June or July, resulting in a delay of inshore water warming until that time (Quast 1968). Multiple spawning has not often been con- sidered in studies of labrid breeding seasons. Roede (1972) stated that labrids have "continuous, successive spawning cycles," and based this view upon the presence of many vitellogenic oocyte stages in mature ovaries of the seven species she investigated. She contended there is no resting stage of the ovary, but a series of year-round spawnings. At all times of the year she was able to find ovaries with several stages of developing oocytes as well as the stage 2 recruitment stock. This clearly is not the case in P. pulchrum, where the winter-resting ovary contains virtually no signs of vitellogenesis. Active ovaries of the California sheephead strongly resemble those pic- tured by Roede (1972, plates II and III) for Halichoeres and Hemipteronotus. The successive spawnings within a restricted season indicated for P. pulchrum may then be a curtailed version of a 278 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM year-round condition in its presumably tropical ancestor, representing an adaptation to the fluc- tuations of food availability characteristic of temperate regions. The size-specific increases in fecundity seen in Catalina P. pulchrum are, of course, common in most long-lived fishes. Many of the Guadalupe females vv^ere not sexually active when the sample v^^as taken and no fecundity data are available. However, it can be predicted that the average fecundity of Guadalupe Island individuals will increase much more slowly with age than that of individuals from Catalina, due to the low growth rate of the Guadalupe individuals discussed in an earlier section. If the active ovary weight increases with size in a fashion similar to that seen at Catalina (Figure 13), the average ovary weight for a 4-yr-old fish at Guadalupe would be approximately 8 g. Age class 4 California sheephead at Catalina had ovaries with an average weight of 13.13 g. The difference increases with age. Six- and eight-year-old individuals at Guadalupe should have ovaries weighing 9 and 15 g respectively. Weights for the same ages at Cat- alina were 23.1 g and 53.5 g. In the Catalina population, there may be an abrupt increase in the fecundity of fishes remain- ing female after age seven; this is the age where most sexual transformations occur (compare Figures 9 and 14). If such an increase does exist, it may be an indication of compensation by those remaining females for the relative gain in age- specific reproductive potential experienced by in- dividuals that do change sex. A more complete discussion of relative male and female age-specific fecundities can be found elsewhere (Warner in press). the Labridae, and extensive sampling is usually needed before the relationship between sex and coloration can be accurately described. Many labrid species exhibit a number of color phases, and these have often been attributed to sexual dimorphism or to differences between im- matures and adults. Roede (1972) has reviewed a number of cases where such an interpretation was incorrect, being based on casual observation or small samples. Apparently there is no strict dis- tribution of sex with color in the Labridae and the only generalization possible is that females tend to strongly predominate in the "first adult" (Roede 1972) colors, and the terminal-phase coloration is made up almost exclusively of males. In most species investigated, males make up 10 to 35% of the first adult-colored individuals (Roede 1972). In Gomphosus varius (Strasburg and Hiatt 1957), Halichoeres maculipinna, H. garnoti, and Hemipteronotus martinicensis (Roede 1972), no males are found in the initial color phase. In con- trast, Soljan (1930a, b) found that 48% of the Symphodus (Crenilabrus) ocellatus examined in the first adult phase were males. The terminal-phase coloration appears to be much more closely restricted as to sex. Of 14 labrid species exhibiting color phases mentioned by Roede (1972), the terminal phase consisted exclusively of males in all but two {Halichoeres garnoti and H. bivittatus). When other coloration classes are described, intermediate between the initial and terminal phases, the proportions of males and females in them are also intermediate. Roede (1972) notes that where color changes are more gradual, as in H. garnoti and H. bivittatus, the relationship between size and sex is the least exact. The Relationship of Color and Sex Pimelometopon pulchrum appears to follow the general labrid coloration pattern quite closely, with a preponderance of females and immatures in the initial uniform color phase, and the terminal bicolored phase containing only males. Thus the designation of the uniform phase as the "female" coloration and the bicolored phase as the "male" coloration (Jordan and Evermann 1898; Fitch and Lavenberg 1971; Miller and Lea 1972) is more or less correct, especially when immatures are included under the uniform designation (Barnhart 1936; Roedel 1948). Dichromatism, however, is not necessarily an indication of sexual dimorphism in Sex Ratio and an Estimate of Survival Roede (1972) believed that her collections were true random samples of populations and was able to estimate the sex ratio in the seven labrid species she investigated. There were two to four times as many females as males in all but one species {Hemipteronotus splendens), which had an equal sex ratio. The samples of P. pulchrum were not considered random and direct sex ratio estimates could not be made. Field transects at Catalina and Guadalupe islands yielded a ratio of about 5.5 uniformly colored individuals to every bicolored male. To es- timate the sex ratios of mature individuals, the 279 FISHERY BULLETIN: VOL. 73, NO. 2 proportion of immatures and males in the uniform group must be known, and this requires some knowledge of mortality rates. A rough estimate of mortality can be made from the transect data and the known color composition of each age. The yearly survival rate is calculated using a modification of a simple fisheries estimate (Ricker 1958). The rate is assumed to be constant, and can be estimated as: s = N, where A'^ is the number of individuals in a par- ticular age class in a sample. Where a large number of age classes are available, one can weight the classes according to their abundance and separate two or more ages from the numera- tor and denominator, giving, for example: N^ + N, + ... + N,_, For the Catalina population of P. pulchrum, the formula used was: In this first approximation we assume that bicolored fishes are all 8 or more years old, and younger fish (ages 2 through 7) are uniform. The decision to use age 7 as the dividing point comes from Figure 9, where between ages 7 and 8 the proportion of females drops to a low level and the males become predominant. From Catalina transect data, s is estimated by: s6 = _33 and s = 0.735. 183 The transect ratio can then be adjusted to com- pensate for bicolored individuals younger than age 8 and uniform individuals older than age 7 by us- ing proportions derived from Table 4, and each age's contribution to the numerator or denomina- tor can be weighted according to the first estimate of survival derived above. The new estimate of survival from the adjusted transect ratio is not very different from the original, s = 0.71. A similar estimate for the Guadalupe Island population, assuming in this case that the uniform individuals are ages 2 through 7 (see Figure 9) and adjusting as before, is s = 0.69. Mature sex ratios can now be estimated. Using 0.7 as the yearly survival rate, about 36% of the uniform individuals seen at Catalina should be mature, and approximately 5% of those individuals would be male. The ratio of mature males to ma- ture females from field transects would then be derived as: 33 + (183 X 0.36 x 0.05) 36 183 X 0.36 X 0.95 =l = "-s' or about two females for every mature male. For Guadalupe, about a third (34%) of the uniform individuals should be mature, and 90% of these would be female. The sex ratio at Guadalupe would then best estimated as: 64 -I- (343 X 0.34 x 0.1) 76 343 X 0.34 X 0.9 105 = 0.72 or approximately three females for every two males. An artifact of protogynous hermaphroditism is the concentration of females in the younger ages. Thus, the observed sex ratio depends on when the animals change sex, and upon the mortality oc- curring from year to year. Mortality causes sex ratios to be biased towards females and these become even more biased the greater the average age of transformation is in the population. This effect can be seen by comparing the estimated sex ratio of the Guadalupe population (0.72), where most females change sex within 3 yr after ma- turity, with that of Catalina (0.57), where trans- formation is relatively delayed. The deviations of sex ratio from unity seen here should not be taken as contradictions of the theories put forth on the adaptiveness of the 1:1 ratio (Fisher 1930; Bodmer and Edwards 1960; Kalmus and Smith 1960), as these were developed for nonhermaphroditic species, and sought to equalize the lifetime reproductive potentials for males and females. In sequential hermaphrodites, the same individual functions as both male and female at sometime in its life, and the question becomes one of changing sex at the proper time to maximize the individual's reproductive potential (Warner in press). SUMMARY Year-round sampling of a population of the California sheephead, Pimelometopon pulchrum, was carried out at Catalina Island, Calif., and comparative material was collected from a population at Guadalupe Island, Mexico. Age determinations indicate individuals in the Guadalupe population are dwarfed relative to 280 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM those at Catalina. The growth rate is lower for Guadalupe fishes and in both populations there may be a slowing of growth at the onset of ma- turity, as well as an increase in the growth rate after sexual transformation. Pimelometopon pulchrum is a protogynous her- maphrodite. During the sex change from female to male, the ovary degenerates and spermatogenic crypts dominate the gonad. The basic structure of the gonad remains ovarian however, with lamellae protruding into a central lumen. Sperm transport is through a series of ducts on the periphery of the gonad and oviduct. Catalina California sheephead attain sexual maturity at age 4, at a standard length of about 200 mm. Most function as females for 4 yr and then change sex, at a length of about 310 mm. Some individuals may transform earlier or later, or not at all. The Guadalupe population also matures at age 4, at a length of about 140 mm. But transfor- mation occurs at an earlier age, with most in- dividuals becoming males by age 7. Peak trans- formation activity occurs in fishes between 190 and 230 mm SL at Guadalupe. Gonad development states and gonad indices of Catalina California sheephead suggest that spawning occurs in July, August, and September and that sexual transformation occurs in the winter months between breeding seasons. Spawning probably takes place a number of times in a single breeding season, which complicates the determination of the actual number of eggs produced by a female each year. Ovary weight, however, can give a good indication of relative age and size-specific fecundities, since egg density does not appear to vary with fish length. The ovary weight of P. pulchrum increases exponentially with length and linearly with the age of the individual in the Catalina population. At Guadalupe, the average fecundity probably increases more slowly with age when compared to Catalina, due to the low average rate of growth. Pimelometopon pulchrum has three color phases. Juvenile coloration occurs in individuals usually less than a year old and smaller than 100 mm in length, and never in sexually mature in- dividuals. The uniform coloration is found in immatures and mature females. Melanization may obscure the ground coloration, but it appears that about 5% of the mature uniform individuals were males at Catalina, and 12% at Guadalupe. Bicolored fishes are exclusively males or late transitionals and usually have a nuchal hump and filamentous extensions of the median fins. Field observations indicate that there are about 5.5 uniformly colored individuals to every bicolored male at both Catalina and Guadalupe. Individual size appears to have a greater effect on the sex change than does age, and rapidly growing fishes may change sex sooner than slow growing individuals of the same age, which may not change sex at all. With the assumption of constant, age-indepen- dent mortality, the annual survival rate at both Catalina and Guadalupe was estimated as about 0.7, as judged from the field transect data. The mature sex ratio at Catalina was approximately two females for every male. At Guadalupe the ratio was closer to three females for every two males, due in part to the earlier sexual transformation seen there. ACKNOWLEDGMENTS I would like to particularly thank R. H. Rosenblatt and E. W. Fager for their early and continued encouragement and advice. I am also grateful to J. T. Enright, P. K. Dayton, and J. B. Graham for their critical comments on an earlier draft. Special thanks go to Isabel Downs for her help in many phases of this project. Finally, I would like to thank D. R. Diener for many stimulating discussions on the fascinating field of hermaphroditism in fishes. LITERATURE CITED Atz.J.W. 1964. Intersexuality in fishes. In C. N. Armstrong and A. J. Marshall (editors), Intersexuality in vertebrates includ- ing man, p. 145-232. Academic Press., Lond. Barnhart, p. S. 1936. Marine fishes of southern California. Univ. Calif. Press, Berkeley, 209 p. BODMER, W. F., AND A. W. F. EDWARDS. 1960. Natural selection and the sex ratio. Ann. Hum. Genet. 24:239-244. BoLiN, R. L. 1930. Embryonic development of the labrid fish Oxyjulis califomicus Gunther. Copeia 1930(4): 122- 128. Breder, C. M., Jr., and D. E. Rosen. 1966. Modesof reproduction in fishes. Natural History Press, Garden City, N.Y., 941 p. Chan, S. T. H. 1971. Natural sex reversal in vertebrates. Philos. Trans. R. Soc. Lond., Ser. B, Biol. Sci. 259:59-71. Chan, S. T. H., A. Wright, and J. G. Phillips. 1967. The atretic structures in the gonads of the rice-field 281 FISHERY BULLETIN: VOL. 73, NO. 2 eel {Manopterus albus) during natural sex-reversal. J. Zool. (Lond.) 153:527-539. Clark, F. N. 1934. Maturity of the California sardine {Sardina caeruka), determined by ova diameter measurements. Calif. Dep. Fish Game, Fish Bull. 42, 49 p. Clarke,!. A. 1970. Territorial behavior and population dynamics of a pomacentrid fish, the garibaldi, Hypsypops ruhicun- da. Ecol. Monogr. 40:189-212. Crozier, G. F., Jr. 1%6. Features of carotenoid metabolism in growth and sexual maturation of the labrid fish, Pimelometopon pulchrum (Ayres). Ph.D. Thesis, Univ. California, San Diego, 93 p. CUERRIER, J. P. 1951. The use of pectoral fin rays for determining age of sturgeon and other species of fish. Can. Fish Cult. 11:10-18. Dixon, W. J., and F. J. Massey, Jr. 1969. Introduction to statistical analysis. 3rd ed. McGraw- Hill, N.Y., 638 p. Fisher, R. A. 1930. The genetical theory of natural selection. The Clarendon Press, Oxford, 272 p. Fitch, J. E., and R. J. Lavenberg. 1971. Marine food and game fishes of California. Univ. Calif. Press, Berkeley, 179 p. Hyder, M. 1969. Histological studies on the testis of Tilapia leucosticta and other species of the genus Tilapia (Pisces: Teleos- tei). Trans. Am. Microsc. Soc. 88:211-231. Jordan, D. S., and B. W. Evermann. 1898. The fishes of North and Middle America. Bull. U.S. Natl. Mus. 47, Part 2:1241-2183. Joseph, D. C. 1962. Growth characteristics of two southern California surfishes, the California corbina and spotfin croaker, family Sciaenidae. Calif. Fish Game, Fish Bull. 119, 54 p. Kalmus H., and C. a. B. Smith. 1960. Evolutionary origin of sexual differentiation and the sex-ratio. Nature (Lond.) 186:1004-1006. Kraft, A. N., and H. M. Peters. 1963. Vergleichende Studien uber die Oogenese in der Gattung Tilapia (Cichlidae, Teleostei). Z. Zellforsch. Mikrosk. Anat. 61:434-485. LiEM, K. F. 1963. Sex reversal as a natural process in the synbranchiform fish, Monopterus albus. Copeia 1%3:303-312. 1%8. Geographical and taxonomic variation in the pattern of natural sex reversal in the teleost fish order Synbranchiformes. J. Zool. (Lond.) 156:225-238. LONNEBERG, E., AND G. GUSTAFSON. 1937. Contributions to the life history of the striped wrasse, Lahrus ossifagus Lin. Ark. Zool. 29A(7):1-16. McErlean, A. J., AND C. L. Smith. 1964. The age of sexual succession in the protogynous her- maphrodite Mycteroperca microlepis. Trans. Am. Fish. Soc. 93:301-302. Miller, D. J., and R. N. Lea. 1972. Guide to the coastal marine fishes of California. Calif. Dep. Fish Game, Bull. 157, 235 p. MOE, M. A.,Jr. 1969. Biology of the red grouper, Epinephelus morio (Valenciennes), from the eastern Gulf of Mexico. Fla. Dep. Nat. Resour. Mar. Res. Lab., Prof. Pap. 10, 95 p. NORRIS, K. S. 1963. The functions of temperature in the ecology of the percoid fish Girella nigricans (Ayres). Ecol. Monogr. 33:23-62. Okada,Y. K. 1962. Sex reversal in the Japanese wrasse, Halichoeres poecilopterus. Proc. Jap. Acad. Sci. 38:508-513. OSSIAN, C. R. 1970. Preparation of disarticulated skeletons using en- zyme-based laundry "pre-soakers." Copeia 1970:199-200. Quast, J. C. 1968. 4. Some physical aspects of the inshore environment, particularly as it affects kelp-bed fishes. In W. J. North and C. L. Hubbs (editors). Utilization of kelp-bed resources in southern California. Calif. Dep. Fish Game, Bull. 139:25-34. Randall, J. E. 1955. Stethojulis renardi, the adult male of the labrid fish Stethojulis strigiventer. Copeia 1955:237. Reinboth, R. 1957. Sur la sexualite' du t^l6ost6en Coris julis (L.). C. R. Acad. Sci. Paris 245:1662-1665. 1962. Morphologische und funktionelle Zweigesch- lechtlichkeit bei marinen Teleostiern (Serranidae, Sparidae, Centracanthidae, Labridae). Zool. Jahrb. Abt. Zool. Physiol. Tiere 69:405-480. 1963. Experimentell induzierter Geschlechtswechsel bei Fischen. Verb. Deutsch. Zool. Ges. Munchen 27:67-73. 1970. Intersexuality in fishes. Mem. Soc. Endocrinol. (Hor- mones and the Environment) 18:515-543. Remacle, C. 1970. Contribution a I'^tude de la sexualite chez certain Labridae et Sparidae. Bull. Inst. R. Sci. Belg. 46(35): 1- 13. RiCKER, W. E. 1958. Handbook of computations for biological statistics of fish populations. Fish Res. Board Can., Bull. 119, 300 p. Robertson, D. R. 1972. Social control of sex reversal in a coral-reef fish. Science (Wash., D.C.) 177:1007-1009. Roede, M. J. 1966. Notes on the labrid fish Coris julis (Linnaeus, 1758) with emphasis on dichromatism and sex. Vie Milieu A 17:1317-1333. 1972. Color as related to size, sex, and behavior in seven Caribbean labrid fish species (Genera Thalassoma, Halichoeres, and Hemipterinotus). Studies on the Fauna of Curasao and other Caribbean Islands 138:1-264. ROEDEL, P. M. 1948. Common marine fishes of California. Calif. Dep. Fish Game, Fish Bull. 68, 150 p. ROUNSEFELL, G. A., AND W. H. EVERHART. 1953. Fisheries science, its methods and applications. John Wiley and Sons, N.Y., 444 p. Smith, C. L. 1965. The patterns of sexuality and the classification of serranid fishes. Am. Mus. Novit. 2207, 20 p. SOUAN, T. 1930a. Nestbau eines adriatischen Lippfisches-Crenilabrus ocellatus Forsk. Z. Morphol. Oekol. Tiere 17:145-153. 1930b. Die Fortpflanzung und das Wachstum von Crenilabrus ocellatus Forsk., einem Lippfisch des Mit- telmeeres. Z. Wiss. Zool. 137:150-174. 282 WARNER: REPRODUCTIVE BIOLOGY OF PIMELOMETOPON PULCHRUM SORDI, M. 1962. Ermafroditismo proteroginoco in Labrus turdus L. e in L. merula L. Monitore Zool. Ital. 69(3-4):69-89. Strasburg, D. W. 1970. A report on the billfishes of the central Pacific Ocean. Bull. Mar. Sci. 20:575-604. Strasburg, D. W., and R. W. Hiatt. 1957. Sexual dimorphism in the labrid fish genus Gomphosus. Pac. Sci. 11:133-134. Van Oosten, J. 1929. Life history of the lake herring (Leucichtys artedi Le Seuer) of Lake Huron as revealed by its scales, with a critique of the scale method. Bull. U.S. Bur. Fish. 44:265-428. Warner, R. R. In press. The adaptive significance of sequential her- maphroditism in animals. Am. Nat. 109(965). Williams, G. C. 1966. Adaptation and natural selection: A critique of some current evolutionary thought. Princeton Univ. Press, Princeton, 307 p. Yamamoto, K., and F. Yamazaki. 1961. Rhythm of development in the oocyte of the gold-fish, Carassius auratus. Bull. Fac. Fish., Hokkaido Univ. 12:93-110. 283 AN INVERSE CORRELATION BETWEEN MERISTIC CHARACTERS AND FOOD SUPPLY IN MID-WATER FISHES: EVIDENCE AND POSSIBLE EXPLANATIONS Robert Karl Johnson' and Michael A. Barnett^ ABSTRACT In five species of mid-water fishes, Chauliodus sloani, Diphphos taenia, Pollichthys mauli, Vin- ciguerria lucetia, and V. nimbaria, the central values of meristic counts (anal fin rays, vertebrae, longitudinal photophore rows) and three measures of biological productivity (phosphate-phosphorus concentration, net primary production, zooplankton standing stocks) are correlated negatively. For the species and areas studied the meristic variation observed cannot be related to temperature, salinity, dissolved oxygen, or any other physical or chemical factor known to affect meristic variation in fishes. It is hypothesized that this relationship between meristic counts and measures of food availability involves differences in egg size, fecundity, size at hatching, and size at comparable stages of larval development between populations in different areas, and that these differences in turn reflect adap- tations to low food densities in areas of low productivity and higher predator densities in areas of higher productivity. Meristic characters have been widely used in studies of fish populations and species. Unlike body proportions or coloration, meristic characters are fixed usually at or before metamorphosis and remain constant throughout the life of an in- dividual. Variation in meristic characters stems from both genetic variation between populations and species, and from environmental variation, which, within genetically controlled limits, can directly affect the number of parts formed in developing embryos and larvae. Recent reviews of factors known to affect meristic characters in fishes include Barlow (1961), Blaxter (1969), Gar- side (1966), and Fowler (1970). An inverse relationship between vertebral and/ or other meristic counts and water temperature at the time of early development has been demonstrated in numerous studies (see above review articles). Experimental studies have shown that in many cases the effect of temperature upon meristic characters occurs within a restricted period of time, the so-called sensitive period, and that variations in temperature before and after this period have no effect (Hempel and Blaxter 1961). The sensitive period may vary with 'Division of Fishes, Field Museum of Natural History, Roosevelt Road at Lake Shore Drive, Chicago, IL 60605. ^Scripps Institution of Oceanography, University of Califor- nia, San Diego, La JoUa, CA 92037. different structures with the result that the tim- ing, magnitude, and in some cases the direction of response of different structures to temperature variation differs among different species (Fowler 1970). Hubbs (1926), Barlow (1961), and others, have suggested that the relationship between meristic counts and temperature involves differential ef- fects of temperature on rate of growth versus rate of differentiation, with the result that accelerated growth is associated with a shortening of the sen- sitive period, resulting in the laying down of fewer parts. The conclusion is that conditions retarding growth rates are associated with elevated meristic counts, conditions accelerating growth rates are associated with lowered meristic counts. This explanation has been extended to factors other than temperature known to affect meristic characters in fishes: dissolved oxygen concentra- tion (Alderdice et al. 1958), salinity (Forrester and Alderdice 1966; Blackburn 1967), carbon dioxide concentration, light intensity, exposure to X-rays, etc. (see Fowler 1970). In 1972, we reported a significant negative correlation between certain meristic counts in Diphphos taenia and three measures of food sup- ply: net primary production, phosphate- phosphorus concentration, and zooplankton standing stocks (Johnson and Barnett 1972). To our knowledge, this was the first suggestion of a Manuscript accepted July 1974 FISHERY BULLETIN; VOL. 73, NO. 2, 1975 284 JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY possible relationship between meristic counts and measures of food supply. We did not offer any explanation for this relationship in the earlier report. In the present paper we extend our infor- mation on D. taenia to the Atlantic Ocean, present corroborative evidence for the relationship between meristic counts and food supply based on four other species of mid-water fishes, and attempt to show that the relationship for the species and areas studied is with food supply and not temperature, salinity, or dissolved oxygen. We hypothesize that this relationship between meris- tic counts and food supply reflects differences in egg size, fecundity, size at hatching, and size at comparable stages of larval development between populations in different areas, and that these differences represent adaptations to low food densities in areas of low productivity and higher predator densities in areas of higher productivity. METHODS Collection and Analysis of Data Methods of taking counts follow those of Grey (1964), Morrow (1964), and Johnson (1970). Pho- tophore rows in a generalized stomiatoid fish are illustrated in Morrow (1964: Figure 73), but our nomenclature for segments of photophore rows follows that of Johnson (1970). All vertebral centra were counted including the compound element supporting the parhypural and hypurals (Weitz- man 1967). Standard statistical texts have been used as reference material (especially Tate and Clelland 1957; Sokal and Rohlf 1969). Agreement between sets of ranks is assessed via the tau coefficient of correlation or Kendall's coefficient of concordance, W (see Tate and Clelland 1957). Localities Studied We have studied specimens from eight areas (Figure 1): 1) the eastern tropical Pacific (ETP) off Mexico; 2) the central North Pacific (CNP) off the Hawaiian Islands; 3) the central equatorial Pacific (CEP) at long. 145° to 150°W; 4) the western equatorial Pacific (WEP) around long. 170°E; 5) the Philippine Sea (PS); 6) the South China Sea (SCS); 7) the Gulf of Guinea (GG); and 8) the cen- tral North Atlantic (CNA) including the Sargasso Sea. All of these areas are tropical oceanic habitats and represent a wide range of physical and biological features. Measures of Biological Productivity The measures used to assess relative richness of food supply are phosphate-phosphorus concentra- tion, net primary production, and zooplankton standing stocks. These three \ariables are highly intercorrelated (Gushing 1971). These measures were chosen because there are published attempts at contouring values of these variables on an oceanwide basis and because values for them are commonly reported in more regionally oriented studies. Despite many problems in both sampling and interpretation associated with attempts to con- tour values of biological variables on an oceanwide basis and to integrate values based on a limited number of measurements over a full year, we were forced to accept such attempts as the principal basis for ranking our eight study areas with re- spect to the three measures of food supply. Where possible we relied on synoptic studies presenting contours on an oceanwide basis: net primary production (as mg-C/m^ per day, Koblentz-Mishke et al. 1970; as g-C/m" per year, Ebeling 1962 based on Fleming and Laevastu 1956), zooplankton con- centration (as parts/ 10-' by volume in the upper 150 m of the Pacific Ocean, Reid 1962), and phosphate-phosphorus concentration (asjug- at/liter contoured at 100 m in the Pacific Ocean, Reid 1962). Where these studies did not cover several of our study areas, we used regional studies (SCS: Angot, Steemann Nielsen in Wyrtki 1961; Sorokin 1973; GG: Raymont 1963, Corcoran and Mahnken 1969, Kinzer 1969, Zeitschel 1969, Riley 1972; CNA: Menzel and Ryther 1961, Raymont 1963, Corcoran and Mahnken 1969, Zeitschel 1969, Riley 1972). We compared the contours or values for each of the three measures of productivity over all eight study areas and then ranked the eight areas with respect to each other for each measure (Table 1). As expected (Gushing 1971), the ranks for the three measures over the eight areas are highly concordant (IFgg = 0.85, P <0.01, concordance coefficient corrected for ties, see Tate and Clelland 1957). This highly significant concordance increases our confidence in this approach to ranking the eight areas with respect to produc- tivity and allows summation of the ranks of the three measures of food supply over each area, yielding a rank-sum. We then ranked this rank- sum and obtained the following rank-index order of productivity, from highest to lowest: 1) eastern 285 FISHERY BULLETIN: VOL. 73, NO. 2 in O in 0) ^1 ^ .1 % \ O, l* '^ .- fc. o — » *-» 5 't3 I. O to '-' J= « o ^ -C fc II ^ 286 JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY Table l.-Computation of rank-sum index of productivity. This table was produced by reproducing the contours or values presented by each of the authors cited in the text for each of the three measures of productivity-net primary production (NET), zooplankton concentration (ZOO), and phosphate phosphorus concentration (P04-P)-within the geographic Hmits of each of the eight areas and then ranking the eight areas with respect to one another for each measure. Area NET rank ZOO rank PO,-P rank Sum of ranks Rank-sum ETP 1 1 1 3 1 GG 2.5 3.5 2 8 2 SOS 2.5 '3.5 4.5 10.5 3 CEP 4 5 3 12 4 WEP 7 2 4.5 13.5 5 PS 5 6 7.5 18.5 6 CNA 6 7 6 19 7 CNP 8 8 7.5 23.5 8 w. 0.85, P<0.01 'Data from Brinton (pars, commun.) tropical Pacific, 2) Gulf of Guinea, 3) South China Sea, 4) central equatorial Pacific, 5) western equa- torial Pacific, 6) Philippine Sea, 7) central North Atlantic, and 8) central North Pacific. In es- tablishing the relationship between meristic counts and productivity we have compared central values of meristic counts with this rank-index value for productivity. RESULTS Diplophos taenia Giinther Diplophos taenia, a circumtropical mesopelagic gonostomatid, is the only species included in this study to occur in all eight study areas. Results for counts of anal fin rays, LLP photophores, and IPVALA photophores are illustrated in Figure 2. In nearly all cases counts are highest in areas of lowest productivity, lowest in areas of highest productivity, and intermediate in areas of inter- mediate productivity. Agreement between mean values for photophore row segments in terms of rank order by area, e.g. all four segments in the IC row, is highly significant {W^^^ = 0.81, P <0.01, Table 2), as is the agreement between mean values for anal fin rays, LLP photophores, and IPVALA photophores (W^^ = 0.94, P< 0.01, Table 3). This concordance allows computation of a rank-sum in- dex for mean values of meristic counts (Table 3). There exists no significant correlation between the observed meristic variation and temperature over the six Pacific areas (taug = 0.13, P >0.20, Table 4). Temperature data was taken from a chart of temperature at 100 m in the Pacific Ocean SCS-11 GG CEP WEP PS CNA CNP -J I 1 L. ANAL ETP SOS GG CEP WEP PS CNA-13 30 5 3 -39 8 8 _1 I I I I I I I I l_ 94 9 LLP ETP-68 SOS -1 0 GG CEP I CNA -20 CNP-29 =:_L _I I I I L_ -J I J L. -I I ] I U -J I l_ 134 136 138 140 142 144 146 146 ISO 152 154 156 IPVALA Figure 2.-Diplophos taenia. Comparison of mean (vertical line), 95% confidence limits for the mean (closed bar), one standard deviation on either side of the mean (open bar plus closed bar), and range (horizontal line) for anal fin rays (top), LLP pho- tophores (middle), and IPVALA photophores (bottom) for specimens from eight study areas (Figure 1). Numbers on or- dinate are number of specimens examined. Table 2.-Agreement between segments of IC row of pho- tophores in Diplophos taenia. Values are given as mean)rank. Number of specimens (n) is given for counts of IP and represent the minimum number of specimens counted for each character for each area. Sum of Area n IP PV VAV AC ranks ETP 74 15.7)8 24.9)8 15.1)8 42.0)8 32 GG 10 16.6)6 27.1)2.5 15.9)5 42.9)7 20.5 SOS 10 16.4)7 25.6)7 16.0)4 44.7)6 24 CEP 50 16.9)5 26.2)6 15.8)6 45.18)4 21 WEP 12 17.0)4 26.4)5 15.7)7 45.17)5 21 PS 18 17.1)3 26.9)4 16.3)2 46.67)3 12 CNA 25 17.4)1 27.5)1 16.2)3 46.72)2 7 CNP 29 17.3)2 27.1)2.5 16.6)1 47.5)1 6.5 w. ,8=0.812, P<0.01 287 FISHERY BULLETIN: VOL. 73, NO. 2 Table Z.—Diplophos taenia, computation of rank-sum index of meristic counts and comparison with rank-sum index of produc- tivity (from Table 1). Values are given as mean (number of specimens). Ana Mean 1 fin Rani LLP IPVALA k Mean Rank Sum of ranks Rank- sum Productivity Area k Mean Ran rank-sum ETP 57.0 8 86.5 8 137.5 8 24 8 1 GG (72) 59.5 7 (30) 91.3 5 (68) 143.3 7 19 6.5 2 SCS (10) 62.3 6 (5) 90.8 7 (10) 143.6 6 19 6.5 3 CEP (11) 62.4 5 (3) 91.2 6 (10) 145.54 5 16 5 4 WEP (49) 63.2 4 (39) 92.2 4 (37) 145.75 4 12 4 5 PS (12) 65.4 1 (8) 94.8 2 (12) 149.2 3 6 2 6 CNA (14) 63.8 3 (8) 94.2 3 (16) 150.2 2 8 3 7 CNP (25) 64.7 (29) 2 (13) 95.7 (22) 1 (20) 151.7 (29) 1 4 1 8 W3 ,3 — 0,942, P<0.01 TaUj = -0.893, P <0,01 Table A.—Diplophos taenia. Comparison of rank-sum index of meristic counts (from Table 3) with temperature at 100 m ranked over the six Pacific areas. Temperature data taken from Brinton (1962). Rank Area Counts Temperature CNP PS WEP CEP SCS ETP 5 3.5 1 2 3.5 6 Tau. = +0.13, P>0.20 given by Brinton (1962). The 100-m depth was chosen arbitrarily, but the conclusion holds if sur- face temperatures, whether summer or winter, are chosen. Meristic counts for D. taenia are lowest in specimens from the eastern tropical Pacific where temperature values are also the lowest. This is exactly opposite to the result expected if temperature were involved in determining the meristic variation observed over the six areas. In fact the data show no relationship between meristic counts and temperature for these areas. Values of salinity in the open ocean are far too conservative to be involved in determining the observed variation (Hubbs 1925; Sverdrup et al. 1942; Barlow 1961; Blackburn 1967). Although the eastern tropical Pacific is well known for a marked oxygen minimum layer (Brandhorst 1959), and oxygen concentration variation may affect the development of meristic characters (Alderdice et al. 1958; Garside 1959, 1966), in all eight areas oxygen is essentially saturated in the wind-mixed surface layer where the larvae and probably the eggs of D. taenia occur. The low counts (relative to other areas) of specimens of D. taenia from the eastern tropical Pacific run counter to what might be expected if dissolved oxygen concentrations were involved in determining the observed meris- tic variation. The rank-sum indices of meristic counts and productivity are significantly and negatively correlated (tau g = -0.893, P < 0.01, Table 3). Pollichthys mauli (Poll) Pollichthys mauli ranges from the western North Atlantic to the Philippine Sea. We have examined specimens of this species from the Gulf of Guinea, and the Philippine and South China seas. Results for IPVALA photophore counts (Ta- ble 5) parallel the results for Diplophos taenia; the counts from Philippine Sea specimens are sig- nificantly higher than counts from specimens from the South China Sea and Gulf of Guinea. Table b.-Pollichthys mauli, IPVALA photophores. Area 68 69 70 71 72 73 74 75 76 77 78 n Mean ± 95% limits GG 175614- __24 SCS -- 1 7 2 1 __11 PS -----113212 10 70.46 ± 0.635 71.27 ±0.528 75.70 ±1.170 Vinciguerria lucetia Garman Vinciguerria lucetia is endemic to the eastern Pacific (Ahlstrom and Counts 1958; Craddock and Mead 1970; Gorbunova 1972). The work of Ahl- strom and Counts (1958) has made the early life history of V. lucetia the best known of any of the species included in this report. We have not examined any specimens of V. lucetia in connec- tion with this work, but the following results of the study of this species by Ahlstrom and Counts (1958) seem to be particularly relevant to this paper: 1) In V. lucetia the total number of myo- tomes is formed in late-stage eggs, prior to hatching. 2) Metamorphosis in V. lucetia is marked by a period of rapid change in body proportions without a marked change in standard length. The completion of metamorphosis is sig- naled by the complete development of all pho- tophores, including the late-forming photophores of the posterior VAV and mid- AC segments. 3) Metamorphosis occurs at a smaller size south of lat. 25°N than north of lat. 27.5°N, with meta- 288 JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY morphosis at an intermediate size in specimens from lat. 25° to 27.5°N. Mean values of vertebral and IPVALA counts are lowest in specimens taken from areas where metamorphosis occurs at a smaller size. A delay in vertebral ossification is found in specimens from areas where meta- morphosis occurs at a larger size. 4) Ahlstrom and Counts (1958) report a north to south cline in mean values for IPVALA photophore counts and relate this to temperature. An east to west cline is also suggested by their data (Figure 3), with mean IPVALA counts lowest near the American con- tinent and increasing with distance offshore. Values for productivity measures in the eastern tropical Pacific tend to fall off with increasing distance from land (Reid 1962; Koblentz-Mishke et al. 1970). If variation in photophore numbers in V. lucetia is related to variation in productivity, we would expect mean IPVALA counts at a given latitude to be lower near the continent and higher with increasing distance from land. Data from Ahlstrom and Counts (1958) confirm this expecta- tion (Figure 3) for all but two latitudinal transects. Along these two transects, one just to the north of the equator, the second centered at about lat. 12°N, mean values obtained for IPVALA counts do not change or actually decrease to the 130° 120° 110° 100° 90° 80' W westward, an apparent contradiction of our hypothesis. However, these two zonal transects fall along zonal areas of high or elevated produc- tivity far to the westward of the American con- tinent, and this is true for net primary production, zooplankton standing stocks, or, as illustrated (Figure 3), phosphate-phosphorus concentration. Williams (1972) relates these zonal belts of elevat- ed productivity to the divergence systems at the equator and at the North Equatorial Counter- current-North Equatorial Current boundary. Williams (1972) states that the zonal band at lat. 10° to 12°N is best shown by data for zooplankton stocks, but we note that this band is quite ap- parent for phosphate-phosphorus concentration (Reid 1962). Thus the apparently discrepant values of mean IPVALA counts from specimens of V. lucetia taken along these two zonal transects in fact tend to further corroborate the hypothesized inverse relationship between meristic counts and productivity. Vinciguerria nimharia (Jordan and Williams) Vinciguerria nimharia is nearly circumtropical in distribution but does not occur in the Medi- PO4-P jjg at/I XIOO S>250 11200-250 P1 00 - 200 \S\ 50 - 1 00 []] 25-50 D <25 Figure 2,.— Vinciguerria lucetia. Left: IPVALA photophores; values given are means based on five or more specimens taken at each locality (see text for additional explanation, data from Ahlstrom and Counts 1958). Right: phosphate phosphorus data simplified from Reid (1962). 289 FISHERY BULLETIN: VOL. 73, NO. 2 terranean Sea nor in the eastern tropical Pacific (Ahlstrom and Counts 1958; Craddock and Mead 1970; Gorbunova 1972). The development of larvae of V. nimharia is apparently quite similar to the development of larvae of V. lucetia (Ahlstrom and Counts 1958; Silas and George 1971). We have examined specimens of V. nimharia from all of our study areas except the eastern tropical Pacific (where V.nimbaria is replaced by V. lucetia) and the western equatorial Pacific (no material available). Counts of IPVALA photophores for specimens of V. nimharia are given in Table 6. Mean values for specimens from the South China Sea, central equatorial Pacific, Philippine Sea, central North Atlantic, and central North Pacific agree in per- fect rank-order with the rank-sum index of meristic counts for specimens of Diplophos taenia from these five areas (Table 3). The mean value of counts of IPVALA photophores of Gulf of Guinea specimens does not fit this trend, it is too high. All of our material of V. nimharia from the Gulf of Guinea came from a single collection at the University of Miami (UMML 21902, lat. 0°54' to 1°05'N, long. 4°53' to 4°51'E, 23-24 May 1965). We have neither additional material of nor informa- tion on V. nimharia from the Gulf of Guinea, and, for the present, we are unable to explain these anomalous results. The values obtained for specimens of V. nim- haria from other study areas support our hypothesis of an inverse relationship between meristic counts and productivity. This is true for both IPVALA photophore (Table 6) and vertebral (Table 7) counts. Counts for Arabian Sea (AS, Table 6) specimens of V. nimharia are taken from Silas and George (1971). They studied specimens taken off the Malabar Coast of India and found larvae of V. nimharia to be most abundant along the edge of the continental shelf from Mangalore to south of Cochin. Gushing (1971) discusses the strong up- welling system occurring along this coast during the period of the Northeast Monsoon, and notes Table 1 .— Vincigxierria nimharia, vertebrae. Area 39 40 41 42 n Mean ± 95% lim ts SCS 6 5 _ — 11 39.45 ±0.351 CEP 1 23 7 — 31 40.19 ±0.175 PS — 3 12 8 23 41.22 ± 0.290 CNP - - 6 21 27 41.78 ± 0.168 that high values of productivity occur in this area over at least half of the year and are associated with the upwelling system. Silas and George (1971) found V. nimharia larvae to be most abundant during the upwelling season. Values for produc- tivity measures in this area given by Gushing (1971) approach values for the eastern tropical Pacific, are certainly larger than values for the Philippine Sea, central North Atlantic, and central North Pacific, and probably significantly larger than values for the central equatorial Pacific and South China Sea. We therefore expected values for meristic counts of specimens of V. nimharia from off the Mangalore Coast to be the lowest of any of these six areas. They are (Table 6). Chauliodus sloani Bloch and Schneider Chauliodus sloani occurs in tropical and temperate waters from the North Atlantic to the eastern Pacific, although throughout large oceanic areas it is replaced by other species of Chauliodus. The remaining six species of Chauliodus, includ- ing the recently described C. vasnetzovi Novikova, are limited to smaller areas, each entirely within one ocean basin (Morrow 1961; Gibbs and Hurwitz 1967; Novikova 1972). We have examined specimens of C. sloani only from our Philippine Sea and central North Pacific study areas, but data from other sources (Ege 1948; Blache 1964; Gibbs and Hurwitz 1967) have made it possible to compare our results for C. sloani with counts for this species from other areas, and with counts for the closely related species C. pammelas Alcock and C schmidti Ege (Table 8). IC photophore counts of C sloani from central gyral areas (CNP, CNA, PS) are higher than counts from specimens taken in the South Table 6.-Vinciguerria nimharia. IPVALA photophore counts. AS = study area of Silas and George (1971) along Malabar Coast of India in the Arabian Sea, data taken from their study. Counts presented as the average between right and left sizes of each specimen. Area 64 64.5 65 65.5 66 66.5 67 67.5 68 68.5 69 69.5 70 70.5 71 71.5 72 72.5 73 n Mean ± 95% limits AS GG SCS CEP PS CNA CNP 2 - 1 - 10 2 1 4 14 24 1 6 26 2 4 20 1 1 37 4 3 83 2 18 36 1 15 16 1 - - 9 65.56 ± 1.023 20 70.43 ± 0.380 24 66.46 ±0.378 32 67.98 ± 0.302 257 69.60 ±0.126 12 69.12 ± 0.623 43 70.34 ± 0.157 290 JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY Table 8.— IC photophore variation in three species of Chaxdiodus. (NIO, northern Indian Ocean, TAA, TAB, areas of eastern tropical Atlantic discussed in text). Species Area 58 59 60 61 62 63 64 65 66 67 68 69 n Mean ± 95% limits C.pammelas^ NIO 1 4 12 4 —---____ 21 59.90 ± 0.350 C. schmidtn TAA -- 17 32 14 ------ 54 62.09 it 0.186 C.schmidtP TAB ---3 16 14 7 5----45 62.89 ± 0.335 C.sloann SOS - - - - 1 16 69 91 25 2 - - 204 64.32 ±0.117 C.sloani PS _______ 3244- 13 66.69 ±0.714 C.sloann GNA -------- 363 1 13 67.15 ± 0.543 C.sloani CNR --- — - — _ 2 97 2 — 20 66.45 ± 0.386 iQibbs and Hurwitz 1967. 2Blache 1964. 3Ege 1948. China Sea. This agrees with results for other species discussed in this paper. The only character diagnostically separating C. schmidti from C. sloani is the lower number of serial photophores in C. schmidti (Morrow 1961; Blache 1964). Similarly Gibbs and Hurwitz (1967) concluded that the only characters separating C. pammelas from C. sloani were lower meristic counts (IC, VAV, vertebrae) in C. pammelas and greater development of the gill filaments in C. pammelas, with filaments both longer and with a greater number of lamellae per side. Gibbs and Hurwitz (1967) noted that the greater gill filament development of C. pammelas is correlated with a well-marked oxygen minimum layer in the northern Indian Ocean habitat of this species. Gill filament length may vary intraspecifically in some wide-ranging mid-water fish species (Johnson 1974). Both C. schmidti, inhabiting the eastern tropical Atlantic, and C. pammelas, inhabiting the northern Indian Ocean, are limited to areas of high biological productivity (Ryther and Menzel 1965; Gibbs and Hurwitz 1967; Gushing 1971). Both are distinguished from C. sloani by lower counts of serial photophores (and vertebrae in C. pam- melas), and essentially only by these lower counts. In both cases the lower counts apparently agree with our hypothesized relationship between meristic counts and productivity. The counts for C. schmidti are from specimens taken in two areas: TAA, along the west African coast from lat. 03°56' to 18°22'N, to the west and north of our Gulf of Guinea study area; and TAB, along the west African coast from lat. 01°20' to 17°53'S, to the south of our Gulf of Guinea study area (Ege 1948; Blache 1964). The counts for C pammelas are from specimens taken between lat. 08° and 14°N, long. 58° to 66°E, in the Arabian Sea (Gibbs and Hur- witz 1967). In view of other results presented in this paper, particularly those for Diplophos taenia, we sug- gest that a reexamination of the status of both C. pammelas and C. schmidti, with additional study of meristic variation in C. sloani throughout the range of this species, are in order. Results of the Antipodes Transect An essentially experimental opportunity to test the hypothesized relationship between meristic counts and productivity was afforded by fishes taken by the Antipodes Expedition of the Scripps Institution of Oceanography in 1970. On this ex- pedition 22 mid-water trawl collections were taken in the Philippine Sea and six mid-water trawl collections were taken in the South China Sea (Figure 4). Because of the 2,000 m or more sill depth separating these geographically contiguous areas, the upper water mass in both areas is the same and differences in physical parameters are minimal. Although the South China Sea is poorly known, there is little doubt that at least nearshore areas or areas over shelves of the South China Sea are substantially more productive than offshore areas in the Philippine Sea (Wyrtki, 1961; Sorokin 1973). We predicted: 1) that values of meristic counts for species occurring in both the South China Sea and Philippine Sea would be lower in specimens from the South China Sea, and 2) that values of meristic counts for species occurring in the Philippine Sea would be lower in specimens taken near land and increase with increasing distance from shore. In all four cases thus far examined (Figure 4) where differences exist in values of meristic counts from the two areas, the counts are sig- nificantly lower in specimens from the South China Sea. This supports our first prediction. Vinciguerria nimbaria was the only species taken in sufficient abundance to allow a test of our 291 Ii0° -iO" 140° 150° FISHERY BULLETIN: VOL. 73, NO. 2 PS scs PS scs PS scs -16 -10 -10 -11 PS scs -1 D taenia P maul 257 ■24 V nimbaria 6&46I IPVALA _L J , L J L I 1 L J_ J_ J I L -3 -2 -1 -13 -204 7 8 9 IC 10 64.63 Figure 4.-Left: Antipodes Expedition station positions. Right: Comparison of IPVALA and IC photophore counts between specimens from South China Sea (SCS) and Philippine Sea (PS). Data presented as in Figure 2 except that mean values for South China Sea material of all four species have been set equal to zero and all other statistics are plotted as deviations from this zero point. Data for Chauliodus sloani from SCS are from Ege (1948); the rest is original data. second prediction. In Table 9 mean IPVALA counts for Philippine Sea specimens of V. nim- baria are tested for relationship with distance of site of collection from land (Japan, Ryukyu Islands, Luzon, but not Bonin or Volcano islands). While the highest mean counts were found in specimens from the 4 stations most distant from land, the data show no relationship between mean counts and distance from land (tau^ = -0.273, P >0.20). Mean IPVALA values for specimens from each of the 11 pairs of Philippine Sea stations are significantly higher than the mean IPVALA values for specimens from the South China Sea. DISCUSSION One fact and two assumptions are prerequisite to our discussion of the possible explanations for the relationship between meristic counts and measures of food supply. The fact: in Diplophos taenia, Vinciguerria lucetia, and V. nimbaria, the values of the meristic characters we have studied are fixed at or before metamorphosis (Ahlstrom, pers. commun., Ahlstrom and Counts 1958, Silas and George 1971). This is probably also true for Pollichthys mauli and Chauliodus sloani. This means that any explanation involves factors operating on eggs and /or larvae. The assumptions: 1) that the meristic variation observed is not the result of selection for certain absolute values of the meristic counts, and 2) that the same basic mechanism underlies the variation in counts for all five species in the area studied (in this discussion we ignore results for specimens of V. nimbaria from the Gulf of Guinea). There are four possibilities: 1) that the observed variation is ecophenotypic, i.e. nongenetic modification of the phenotype resulting from the effects of differing food availability conditions upon early growth and development of meristic characters; 2) that the observed variation is a by- product and indicative of genetic differences between populations in these eight areas, and that these differences reflect differing selective pres- sures resulting from differing conditions for early growth; 3) that the observed variation is a com- bination of ecophenotypy and genetic differences; and 4) that the real explanation is none of these, that a causal relationship between meristic characters and productivity does not exist, and that we have overlooked the real meaning of our results. We are unable to deal with the third pos- 292 JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY Table 9.-IPVALA photophore counts of Vinciguerria nimbaria from the Philippine and South China seas. Antipodes station positions given in Figure 4. Distance from shore (Philippine Sea stations only) given in rank order from nearest to land to most distant offshore. Rr rank of mean, R,. = rank of distance offshore. Antipodes stations 65 65.5 66 66.5 67 67.5 68 68.5 69 69.5 70 70.5 71 71.5 72 72.5 n Mean ± 95% limits R. 1, 2 3,4 5, 6 8 10 11, 12 13, 14 15, 16 17, 18 19, 20 21, 22 - 1 2 — 1 1 3 1 1 — 3 4 3 2 4 3 3 — 2 2 2 1 12 8 1 — 1 — 2 4 2 10 3 3 2 1 3 - 1 1 2 — 1 1 — — 1 1 2 3 4 1 — — 1 — 1 — 1 2 1 — 5 6 2 1 6 3 1 — 1 4 — 3 3 1 1 1 7 8 5 20 36 11 4 3 1 - - 1 10 22 28 22 10 6 13 6 26 14 100 69.80 ± 0.895 69.47 ± 0.456 69.84 ±0.308 69.95 ± 0.374 70.30 : 69.17: 69.31 : 69.83 : 69.06 : 69.54 : 69.59 : ; 0.420 : 0.930 0.531 : 1.124 : 0.440 : 0.604 0.209 4 8 3 2 1 10 9 5 11 7 6 4 5 8 11 10 9 7 6 3 2 1 Philippine Sea, mean ^= 69.60 ± 0.126 Tau,, = -0.273 P>0.20 23, 24 115 1 25, 26 1-11 27 28 4 2 2 1 n 66.06 ± 0.524 66.92 ± 1.345 66.56 ±0.524 3 - - - - - 1 - - - - 6 South China Sea, mean = 66.46 : ± 0.378 sibility and ignore the fourth possibility in our subsequent discussion. We believe that the observed meristic variation is the result of genetic differences between populations and not the result of an ecophenotypic effect of food availability conditions on develop- ment of meristic characters. We present evidence available to support this belief, but we note that this evidence is not conclusive. A statement of the ecophenotypic explanation is easily made. The meristic variation observed could result if the effect of low food densities upon the development of meristic characters parallels the effect of low temperature, retarding growth rates more than differentiation rates, and lengthening the period of determination of meristic characters. Because the effect of low food availability upon egg maturation appears to be a reduction of egg number and not egg size (Anokhina 1960; Blaxter 1969), any ecophenotypic effect of low food density upon meristic characters would have to operate between the onset of feeding and metamorphosis. Riley (1966) and Blaxter (1969), among others, have found for the species they have studied that the time to reach metamorphosis may be sig- nificantly increased by decreasing the density of food. Therefore an indispensable condition of the ecophenotypic explanation is that for the species studied, the final values of meristic counts are de- termined after the onset of feeding. If so, the meristic variation observed might result from a concordant increase in the length of the period of determination of meristic characters with a delay in time to reach metamorphosis in larvae from areas of lower productivity. Three facts resulting from the study of the development of the eggs and larvae of Vin- ciguerria lucetia by Ahlstrom and Counts (1958) appear to support the ecophenotypic explanation: 1) Ahlstrom and Counts did find a direct rela- tionship between size at metamorphosis (no developmental time scale is available for any of the species studied) and numbers of longitudinal photophores and vertebrae; 2) vertebral ossifica- tion and photophore formation in V. lucetia occur in larvae 11 mm SL or more in size, well after yolk-sac absorption and presumably after the on- set of feeding; and 3) the distances between samples of V. lucetia utilized in construction of Figure 3 are small, much less in most cases than the distances between the eight study areas for the other species discussed in this paper. Yet the results for V. lucetia along the east to west tran- sect lines apparently agree with results for the other mid-water species. We find it difficult to believe that the results for V. lucetia are explainable in terms of genetically distinct populations distributed along these inshore to offshore transects. Three lines of evidence appear to contradict the ecophenotypic explanation in favor of the explanation hypothesizing that the observed meristic variation is the result of genetic differences between populations. (1) in Vin- cigusrria lucetia the total number of myomeres are formed in late stage eggs (Ahlstrom and Counts 1958). Since the number of myomeres, vertebra! counts, and longitudinal photophore row counts are usually highly intercorrelated, the ecophenotypic explanation appears to be invalid in 293 FISHERY BULLETIN: VOL. 73, NO. 2 this case. (2) The data for Vinciguerria nimharia may indicate the existance of separable popula- tions in our different study areas. This is sug- gested by the results for the Antipodes transect (Table 9) in which is found no clear evidence for an onshore to offshore trend toward higher IPVALA counts, despite the fact that the productivity measures are higher inshore and decrease (rapidly) to seaward (Reid 1962; Koblentz-Mishke et al. 1970). Mean values of IPVALA photophore counts for specimens from each of the 11 pairs of Philippine Sea stations are significantly higher than the mean value for South China Sea specimens. This may suggest that genetically dis- tinct, separable populations of V. nimharia are found in each area. Gill raker counts for V. nim- haria (Table 10) apparently support this sugges- tion in that counts of gill rakers are discordant with counts of IPVALA photophores (Table 6) and vertebrae (Table 7). For the four Pacific areas the counts of vertebrae and IPVALA photophores for V. nimharia agree in perfect rank-order with the IPVALA photophore counts for D. taenia (Table 10). That this is not true for gill raker counts may indicate the existence of separable populations of V. nimharia in the South China Sea, central equatorial Pacific, and the North Pacific central gyral areas (Philippine Sea, central North Pacific). (3) The ecophenotypic explanation implies that in areas of low productivity elevated meristic counts result from retardation of growth and that this retardation is the result of the average survivor being underfed compared to larvae in areas of higher productivity. As year class strength in pelagic fish populations is probably largely deter- mined in early stages of larval life and not by the total number of eggs produced or mortality during Table \0.— Vinciguerria nimharia, comparison of gill raker counts with vertebral and longitudinal photophore row counts. v. nimbaria, total gill rakers on first gill arch. Area 17 18 19 20 21 22 23 24 25 26 n Mean±95% limits GG ______ 3 6 10 1 20 24.45 ± 0.386 SOS 4 7 3i______i5 18.07 ±0.489 CEP - _ _ 4 13 13 1 - - - 31 21.35 ±0.277 PS - 2 26 54 31 2 - - - - 115 20.04 ± 0.147 CNA -4 8---_--_12 18.67 ±0.313 CNP -- 3 6 3----- 12 20.00 ± 0.469 V. nimbaria and Diplophos taenia, comparison of counts, given as mean (rank. Vinciguerria nimbaria Diplophos taenia Area Gill rakers IPVALA Vertebrae IPVALA SCS 18.1 (1 66.5 (1 39.4 (1 143.6 (1 CEP 21.4(4 68.0(2 40.2(2 145.5(2 PS 20.0(2.5 69.6(3 41.2(3 149.2(3 CNP 20.0(2.5 70.3(4 41.8(4 151.7(4 advanced prerecruit stages (Hempel 1965), it seems likely that selection would strongly favor any mechanisms that tended to protect the larvae of mid-water fishes occurring in areas of low productivity against starvation. The possible materials on which this selection might operate and the possible consequences on meristic characters form the basis for a second explanation of the observed meristic variation, that it is the by-product of genetic differences between separable populations in areas of low and high productivity. Hempel (1965), Blaxter (1965), and others, con- cerned mainly with pelagic clupeoid fishes, have developed strong evidence that under normal cir- cumstances the main restriction on the success of a year class occurs within a short period of larval life, the critical period of Hjort (1914, 1926) and others (e.g. Marr 1956; Schumann 1965). Selection has apparently resulted in adaptive mechanisms tending to balance the two main dangers to larval survival: the danger of starvation and the danger of predation (Blaxter and Hempel 1963; Hempel 1965). Blaxter (1965), Hunter (1972), and others, have shown that at the onset of feeding, just before or at the time of yolk-sac absorption, surprisingly small differences in size can significantly affect the probability of larval survival. Hunter (1972) has shown for northern anchovy Engraulis mor- dax Girard, larvae that slight increments in size are associated with highly increased searching abilities, highly increased success of attempted feeding acts, and vastly diminished minimum prey density requirements for survival. Blaxter (1965) discusses the significance of the greater spectrum of particle sizes available to larger larvae in terms of increased diversity of available prey organisms. Similar findings have been reported for other fish larvae (e.g. Arthur 1956; Einsele 1965). Size at hatching, at least for Atlantic herring, Clupea harengus Linnaeus, is a direct function of egg size; larger larvae hatch from larger eggs. Fecundity is inversely proportional to average egg size (Baxter 1959; Blaxter and Hempel 1963; Blaxter 1969). We believe that the meristic variation between populations occurring in areas differing in productivity values is the result of adaptations in- volving the adjustment of egg and larval size to the productivity regime. We believe that these adaptations reflect differences between areas of low and high productivity in the relative impor- 294 JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY tance of two principal dangers to larval survival: starvation versus predation. We hypothesize that selection on mid-water fish populations inhabiting areas of low productivity has favored mechanisms tending to offset the danger of larval starvation, and that these populations will exhibit: 1) larger average egg size, 2) lower fecundity, and 3) larger average larval size at hatching and at comparable stages of development than populations living in areas of higher productivity. Advantages that might accrue to larger larvae in areas of lower produc- tivity include increased mobility, a wider possible search volume, increased diversity of potential prey organisms, and a longer period of sur- vivorship solely on yolk reserves. The danger of starvation is presumably lower in areas of higher productivity but the danger of predation, resulting from presumed higher densi- ties of potential predators on fish larvae, may be greater. Here selection may have favored increased fecundity tending to offset the danger of increased predation on larvae. We believe that in areas of higher productivity populations of mid-water fishes will exhibit: 1) smaller average egg size, 2) increased fecundity, and 3) smaller average larval size at hatching and at comparable stages of development than populations living in areas of lower productivity. In developing this hypothesis we have largely followed Hempel's (1965) explanation for varia- tions in egg size and fecundity between popula- tions of herring in the eastern North Atlantic and North Sea. We note that there exists no evidence for increased predation pressure on larval popula- tions of mid-water fishes in areas of higher productivity. It is possible to retain the main fea- tures of our hypothesis without including preda- tion pressure by relating variation in fecundity and egg size solely to food density requirements. By definition, selection will favor maximizing reproductive output, thereby favoring fewer larger larvae in areas of low productivity where the danger of larval starvation is greater, and favoring higher fecundity (with the concomitant of smaller eggs and larvae) in areas where the danger of starvation is lessened. There exists limited available evidence to sup- port these predictions. Ahlstrom and Counts (1958) showed that egg size and size at hatching in Vin- ciguerria lucetia are directly related, the smallest larvae (at a defined stage of development) are found in areas where average egg diameters are least. They also showed that mean values of ver- tebral and IPVALA photophore counts are lowest in those areas where egg and larval size is least and where metamorphosis occurs earliest (i.e. at smallest size). Although no small larvae were available for this study, we were able to compare development in prejuvenile specimens of V. nim- baria from the central North Pacific with specimens from the central equatorial Pacific. In V. nimbaria the last four VAL photophores are late-forming, are laid down serially from anterior to posterior, usually the left member of a pair of VAL photophores develops just before the right, and the number yet to develop can be determined uniquely from the one to one correspondence with the posterior photophores in the V AV segment. In Figure 5 standard lengths of all available prejuvenile specimens of V. nimbaria from the central North Pacific and central equatorial Pacific are plotted against the number of VAL pho- tophores left to appear. If this character can be used as an index to comparable stages of develop- ment, then at comparable stages of development the larvae from the area of lower productivity are the larger, as predicted. We believe that the correlation between meris- tics and productivity results from a correlation between meristics and egg size, and that egg size and, hence, size at hatching are genetically deter- mined features reflecting adaptation to produc- tivity conditions. A number of authors have stated or suggested that such a correlation exists (Ahl- strom and Counts 1958; Lindsey 1958, 1961; Garside mm 18 - ■ /= NP 17 - -^ O/CEP 16 - -r .y ©/ o 15 14 ^ ■ ■ o 8 O O 8 13 - 12 u 4-4 4-3 3-3 La 3-2 rva( 2-2 2-1 1-1 1-0 Figure 5.— Size of larvae of Vinciguerria nimbaria from the central North Pacific (CNP) and the central equatorial Pacific (CEP) at comparable stages of development. Ordinate: standard length in millimeters. Abscissa: number of VAL photophores yet to form (determined from VAV count) on each side of each specimen. Lines fitted by eye. 295 FISHERY BULLETIN: VOL. 73, NO. 2 and Fry 1959). Lindsey and AH (1971) have recently argued against this suggested rela- tionship, despite the fact that their data showed a direct relationship betw^een the number of anal fin rays and egg size in the medaka, Oryzias latipes. Blaxter and Hempel (1963) found no relationship between incubation time and egg size in herring, but did find a positive correlation between time to yolk sac absorption and egg size. If this correlation is true for the mid- water fishes considered in this report, if the correlation continues beyond the point of yolk sac absorption, and if the meristic characters in question are determined after hatching, it might result in a longer period of de- termination of these characters in larvae from larger eggs. We lack essential developmental and ecological information to complete our hypothesis. We know little or nothing for most mid- water species about age and size at first spawning, number of spawn- ings per female, fecundity, seasonality of reproduction, course of larval development, or factors actually determining survivorship of lar- vae. The answer to the question of mechanism awaits the comparison of these population parameters between populations in areas of high and low productivity. ACKNOWLEDGMENTS We thank the following individuals and institu- tions for the loan of valuable specimens: R. Lavenberg, Natural History Museum of Los An- geles County (LACM), Los Angeles; E. Ahlstrom, National Marine Fisheries Service (NMFS-LJ), La Jolla, Calif.; P. Struhsaker, National Marine Fisheries Service (NMFS-H), Honolulu; P. Four- manoir. Office de la Recherche Scientifique et Technique Outre-Mer (ORSTOM), Noumea, New Caledonia; R. Rosenblatt, D. Dockins, J. Copp, Scripps Institution of Oceanography (SIO), La Jolla; T. Clarke, Hawaii Institute of Marine Biology, University of Hawaii (UH), Kaneohe, Hawaii; C. R. Robins, University of Miami (UMML), Miami; B. Nafpaktitis and R. McGinnis, University of Southern California (USC), Los An- geles; R. H. Gibbs, Jr., Division of Fishes, National Museum of Natural History (USNM), Washing- ton, D.C.; R. Backus, J. Craddock, R. Haedrich, Woods Hole Oceanographic Institution (WHOI), Woods Hole, Mass. We thank the following individuals for infor- mation from their own research and for aiding the completion of this project: E. Ahlstrom, B. Collette, D. Cohen, C. L. Hubbs, J. Kethley, G. Krefft, H. Marx, R. Rosenblatt. We thank the Division of Photography, Field Museum of Na- tural History, for aid in preparing the figures. This paper is based in part on the results of the An- tipodes and Styx Expeditions of the Scripps Insti- tution of Oceanography. This work was supported, in part, by NSF Grant GB 7596 to R. H. Rosenblatt and W. Newman. We are grateful to T. Poulson for critically reading the manuscript and offering valuable suggestions for its improvement. We are particarly indebted to E. H. Ahlstrom and R. H. Rosenblatt for their advice, criticism, and en- couragement throughout the development of this research. MATERIAL EXAMINED Diplophos taenia. The material examined of this species is listed in Johnson and Barnett 1972. Supplementary station list of materials examined for a study of meristic variation in Diplophos taenia Guenther. Ref . Ser. Scripps Inst. Oceanogr. 72-4, 1-8 (unpublished manuscript available from the Library, Scripps Institution of Oceanography, La Jolla, Calif). Pollichthys mauli. GG: 24 (28.5-46.9), UMML 22881 (1), UMML 24132 (1), UMML 24237 (1), UMML 24266 (3), UMML 24658 (6), UMML 27884 (6), UMML 27929 (5), UMML 28159 (1). SCS: 11 (21.1-31.1); SIO 61-744 (1), SIO 69-20, (10); PS: 10 (33.8-49.9); SIO 70-308 (1), SIO 70-309 (1), SIO 70-334 (2), SIO 70-337 (2), SIO 70-340 (4). Vinciguerria attenuata. SCS: 71 (13.0-37.8); SIO 70-341 (5), SIO 70-343 (5), SIO 70-344 (52), SIO 70-345 (5), SIO 70-346 (3), SIO 70-347 (1). PS: 71 (13.0-28.1); SIO 70-308 (12), SIO 70-309, (10), SIO 70-310 (6), SIO 70-311 (12), SIO 70-314 (11), SIO 70-318 (1), SIO 70-333 (15), SIO 70-334, (1), SIO 70-337 (2). Vinciguerria nimbaria. GG: 20 (21.0-37.5); UMML 21902 (20). SCS: 35 (11.7-32.0); SIO 70-341 (4), SIO 70-343 (5), SIO 70-344 (10), SIO 70-345 (5), SIO 70-346 (5), SIO 70-347 (6). PS: 729 (11.6-39.9); SIO 70-306 (63), SIO 70-308 (6), SIO 70-309 (18), SIO 1 70-310 (23), SIO 70-311 (29), SIO 70-314 (45), SIO 70-318 (52), SIO 70-326 (7), SIO 70-327 (3), SIO 70-328 (12), SIO 70-329 (22), SIO 70-331 (2), SIO 70-332 (11), SIO 70-333 (173), SIO 70-334 (15), SIO 70-336 (6), SIO 70-337 (14), SIO 70-339 (2), SIO 70-340 (226). CEP: FMNH (Field Museum of Na- tural History) 77100 32 (16.9-36.0). CNA: USNM, 296 JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY Ocean Acre Stations, all material from ca. lat. 32- 32.5°N., long 64°W. 12(19.1-35.8); 12-17C (1), 12- 18A (2), 12-18B (2), 12-28B (1), 12-34C (1), 12-35C (1), 12-62 (1), 12-80 (1), 12-81 (1), 12-86 (1). CNP: 49 (14.1-31.0); UH 69-11-5 (49). Chauliodus sloani. PS: 14 (42.2-214.5); SIO 70-306 (3), SIO 70-311 (2), SIO 70-326 (1), SIO 70-334 (8). CNP: 20, SIO 71-301 (1), SIO 71-307 (1), SIO 71-309 (5), SIO 71-373 (1), SIO 72-11 (2), SIO 72-16 (1), SIO 72-22 (1), SIO 72-25 (2), SIO 73-142 (1), SIO 73-149 (1), SIO 73-155 (1), SIO 73-158 (2), SIO 73-159 (1). LITERATURED CITED Ahlstrom, E. H., and R. C. Counts. 1958. Development and distribution of Vinciguerria lucetia and related species in the Eastern Pacific. U.S. Fish Wildl. Serv., Fish. Bull. 58:363-416. Alderdice, D. F., W. p. Wickett, and J. R. Brett. 1958. Some effects of temporary exposure to low dissolved oxygen levels on Pacific salmon eggs. J. Fish. Res. Board Can. 15:229-250. Anokhina, L. E. 1960. Interrelations between fecundity, variability of size of the eggs, and fatness of mother fish in Onega herring. Dokl. Akad. Nauk SSSR 133(4):960-963. Arthur, D. K. 1956. The particulate food and the food resources of the larvae of three pelagic fishes, especially the Pacific sar- dine, Sardinops caerulea (Girard). Ph.D. Thesis, Univ. California, 231 p. Barlow, G. W. 1961. Causes and significance of morphological variation in fishes. Syst. Zool. 10:105-117. Baxter, I. G. 1959. Fecundities of winter-spring and summer-autumn herring spawners. J. Cons. 25:73-80. Blache, J. 1964. Sur la validite de Chauliodus schmidti Ege 1948 {Pisces, Teleostei, Clupeiformi, Stoniiatoidei, Chauliodidae) espece characteristique de I'Atlantique oriental. Cah. O.R.S.T.O.M. (Off. Rech. Sci. Tech. Outre- Mer) Ser Oceanogr. II(l):33-44. Blackburn, M. 1967. Synopsis of biological information on the Australian anchovy Engraulis australis (White). Calif. Coop. Oceanic Fish. Invest. Rep. 11:34-43. Blaxter, J. H.S. 1965. The feeding of herring larvae and their ecology in relation to feeding. Calif. Coop. Oceanic Fish. Invest. Rep. 10:79-88. 1969. Development: Eggs and larvae. In W. S. Hoar and D. J. Randall (editors). Fish physiology 3:177-252. Academic Press, N.Y. Blaxter, J. H. S., and G. Hempel. 1963. The influence of egg size on herring larvae {Clupea harengus L.). J. Cons. 28:211-240. Brandhorst, W. 1959. Nitrification and denitrification in the Eastern Tropical North Pacific. J. Cons. 25:3-20. Brinton, E. 1962. The distribution of Pacific euphausiids. Bull. Scripps Inst. Oceanogr., Univ. Calif. 8:51-269. Corcoran, E. F., and C. V. W. Mahnken. 1969. Productivity of the tropical Atlantic Ocean. Proc. Symp. Oceanogr. Fish. Resour. Trop. Atl., Rev. Pap. Contrib. (UNESCO), Abidjan 1966, p. 57-68. Craddock, J. E., and G. W. Mead. 1970. Midwater fishes from the eastern South Pacific Ocean. Scientific Results of the SE Pacific. Exped. Anton Bruun, Rep. 3, 46 p. Gushing, D. H. 1971. Upwelling and the production of fish. Adv. Mar. Biol. 9:255-334. Ebeling, a. W. 1962. Melamphaidae. I. Systematics and zoogeography of the species in the bathypelagic fish genus Melamphaes Gunther. Dana Rep., Carlsberg Found. 58, 164 p. Ege, V. 1948. Chauliodus Schn., bathypelagic genus of fishes. A systematic, phylogenetic and geographical study. Dana Rep., Carlsberg Found. 31, 148 p. Einsele, W. 1965. Problems of fish-larvae survival in nature and the rearing of economically important middle European freshwater fishes. Calif. Coop. Oceanic Fish. Invest. Rep. 10:24-30. Fleming, R. H., and T. Laevastu. 1956. The influence of hydrographic conditions on the behavior of fish. (A preliminary literature survey.) FAO (Food Agric. Organ. U.N.) Fish. Bull. 9:181-196. Forrester, C. R., and D. F. Alderdice. 1966. Effects of salinity and temperature on embryonic development of the Pacific cod (Gadus macroce- phalus). J. Fish. Res. Board Can. 23:319-340. Fowler, J. A. 1970. Control of vertebral number in teleosts-an embryological problem. Q. Rev. Biol. 45:148-167. Garside, E. T. 1959. Some effects of oxygen in relation to temperature on the development of lake trout embryos. Can. J. Zool. 37:689-698. 1966. Developmental rate and vertebral number in salm- onids. J. Fish. Res. Board Can. 23:1537-1551. Garside, E. T., and F. E. J. Fry. 1959. A possible relationship between yolk size and differentiation in trout embryos. Can. J. Zool. 37:383-386. Gibbs, R. H., Jr., and B. A. Hurwitz. 1967. Systematics and zoogeography of the stomiatoid fishes, Chauliodus pammelas and C. sloani, of the Indian Ocean. Copeia 1967:798-805. Gorbunova, N. N. 1972. Systematics, distribution and biology of the fishes of the genus Vinciguerria (Pisces, Gonostomatidae). (In Russ., Engl, summ.) Akad. Nauk SSSR, Tr. Inst. Okeanol. 93:70-109. Grey, M. 1964. Family Gonostomatidae. In Y. H. Olsen (editor). Fishes of the western North Atlantic, Part 4, p. 78- 240. Sears Found. Mar. Res., Mem. I. Hempel, G. 1965. On the importance of larval survival for the population dynamics of marine food fish. Calif. Coop. Oceanic Fish. Invest. Rep. 10:13-23. 297 FISHERY BULLETIN: VOL. 73, NO. 2 Hempel, G., and J. H. S. Blaxter. 1961. The experimental modification of meristic characters in herring {Clupea harengus L.). J. Cons. 26:336-346. Hjort.J. 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Cons. Perm. Int. Explor. Mer., Rapp. P.-V. Reum. 20:1-228. 1926. Fluctuations in the year classes of important food fishes. J. Cons. 1:5-38. HUBBS, C. L. 1925. Racial and seasonal variation in the Pacific herring, California sardine and California anchovy. Calif. Dep. Fish Game, Fish. Bull. 8, 22 p. 1926. The structural consequences of modifications of the developmental rate in fishes considered in reference to certain problems of evolution. Am. Nat. 60:57-81. Hunter, J. R. 1972. Swimming and feeding behavior of larval anchovy, Engraulis mordax. Fish. Bull., U.S. 70:821-838. Johnson, R. K. 1970. A new species of Diplophos (Salmoniformes: Gono- stomatidae) from the western Pacific. Copeia 1970:437-443. 1974. A revision of the alepisauroid family Scope- larchidae (Pisces, Myctophiformes). Fieldiana: Zool. 66, 249 p. Johnson, R. K., and M. A. Barnett. 1972. Geographic meristic variation in Diplophos taenia Giinther (Salmoniformes: Gonostomatidae). Deep-Sea Res. 19:813-821. KiNZER, J. 1969. Quantitative distribution of zooplankton in surface waters of the Gulf of Guinea during August and Sep- tember 1963. Proc. Symp. Oceanogr. Fish. Resour. Trop. Atl., Rev. Pap. Contrib. (UNESCO), Abidjan 1966, p. 231-240. KOBLENTZ-MlSHKE.O. J., V. V. VOLKOVINSKY, AND J. G. KaBANOVA. 1970. Plankton primary production of the world ocean. In W. S. Wooster (editor), Scientific exploration of the South Pacific, p. 183-193. Natl. Acad. Sci., Wash., D.C. LiNDSEY, C. C. 1958. Modification of meristic characters by light duration in kokanee, Oncorhynchus nerka. Copeia 1958:134-136. 1961. The bearing of experimental meristic studies on racial analyses of fish populations. Proc. IX Pac. Sci. Congr. 10:54-58. LiNDSEY, C. C, AND M. Y. AlI. 1971. An experiment with medaka, Oryzias latipes. and a critique of the hypothesis that teleost egg size controls vertebral count. J. Fish. Res. Board Can. 28:1235-1240. Marr, J.C. 1956. The "critical period" in the early life history of marine fishes. J. Cons. 21:160-170. MENZEL, D. W., AND J. H. Ryther. 1%1. Zooplankton in the Sargasso Sea off Bermuda and its relation to organic production. J. Cons. 26:250-258. Morrow, J. E. 1961. Taxonomy of the deep sea fishes of the genus Chauliodus. Bull. Mus. Comp. Zool. (Harvard Univ.) 125:249-294. 1964. Family Chauliodontidae. In Y. H. Olsen (editor). Fishes of the western North Atlantic, Part 4, p. 290- 310. Sears Found. Mar. Res., Mem. I. NOVIKOVA, N. S. 1972. A new species of the genus Chauliodus (Pisces, Chauliodontidae) from the Southeastern Pacific. J. Ichthyol. 12:34-41. Raymont, J. E. G. 1963. Plankton and productivity in the oceans. Pergamon Press, Oxford, 660 p. Reid, J. L., Jr. 1962. On circulation, phosphate-phosphorus content, and zooplankton volumes in the upper part of the Pacific Ocean. Limnol. Oceanogr. 7:287-306. Riley, G. A. 1972. Patterns of production in marine ecosystems. /« J. A. Wiens (editor). Ecosystem structure and function, p. 91- 110. Oregon State Univ., Corvallis. Riley, J. D. 1966. Marine fish culture in Britain. VII. Plaice (Pleuronectes platessa L.) post-larval feeding on Artemia salina L. nauplii and the effects of varying feeding levels. J. Cons. 30:204-221. Ryther, J. H., and D. W. Menzel. 1965. On the production, composition, and distribution of organic matter in the Western Arabian Sea. Deep-Sea Res. 12:199-209. Schumann, G. 0. 1965. Some aspects of behavior in clupeid larvae. Calif. Coop. Oceanic Fish. Invest. Rep. 10:71-78. Silas, E. G., and K. C. George. 1971. On larval and postlarval development and distribution of the mesopelagic fish Vinciguerria nimharia (Jordan and Williams) off the western coast of India and the Laccadive Sea. J. Mar. Biol. Assoc. India 11:218-250. SoKAL, R. R., and F. J. Rohlf. 1969. Biometry, the principles and practice of statistics in biological research. W. H. Freeman & Co., San Francisco, 776 p. Sorokin, Y. I. 1973. Data on biological productivity of the Western tropical Pacific Ocean. Mar. Biol. (Berl.) 20:177-196. Sverdrup, H. U., M. W. Johnson, and R. H. Fleming. 1942. The oceans, their physics, chemistry, and general biology. Prentice-Hall, N.Y., 1087 p. Tate, M. W., and R. C. Clelland. 1957. Nonparametric and shortcut statistics in the social, biological, and medical sciences. Interstate Printers and Publishers, Inc., Danville, 111., 171 p. Weitzman, S. H. 1967. The origin of the stomiatoid fishes with comments on the classification of salmoniform fishes. Copeia 1967:507-540. Williams, F. 1972. Consideration of three proposed models of the migra- tion of young skipjack tuna (Katsuwonus pelamix) into the eastern Pacific Ocean. Fish. Bull., U.S. 70:741-762. Wyrtki, K. 1961. Physical oceanography of the Southeast Asian wa- ters. NAGA Rep. 2, 195 p. Zeitschel, B. 1969. Productivity and microbiomass in the tropical Atlan- tic in relation to the hydrographical conditions (with emphasis on the eastern area). Proc. Symp. Oceanogr. Fish. Resour. Trop. Atl., Rev. Pap. Contrib. (UNESCO), Abidjan 1966, p. 69-83. 298 SPIN-LABELING TECHNIQUES FOR STUDYING MODE OF ACTION OF PETROLEUM HYDROCARBONS ON MARINE ORGANISMS William T. Roubal' and Tracy K. Collier^ ABSTRACT Spin-labeling studies of membrane-contaminant interaction are being conducted by biochemists at the Northwest Fisheries Center in Seattle, Washington. The aim of these studies is to gain a better understanding of the mode of action of hydrocarbon contaminants at the molecular level. Basic spin-labeling theory together with experimental results are presented and discussed. Spin-labeling holds great promise not only for environmental studies but also for drug research, toxicology, and pharmacology as well. The interaction between contaminants and living systems commences when contaminants combine with so-called active sites in living tissue. Active sites are varied in nature, but often they are groups of molecules assembled in a special fashion such as those which comprise membranes and as- sociated enzymes or other biopolymers. Although the exact nature of contaminant-host interaction may not be known in each and every case, experimental data from biochemical /bio- physical studies allow us to draw certain conclusions about interactions. With detailed investigations, collected data may even allow us to draw a fairly accurate picture concerning the molecular basis of physiological changes which contaminants are able to induce. Admittedly, investigations such as these are difficult to perform. The molecular complexity of living systems defies ready characterization, and it is even more difficult to relate alterations in molecular organization to the subsequent physiological changes wrought by this con- taminant-host interaction. Recent years have seen an upsurge of interest in membranes and how membrane structure is modified when invaded by such things as drugs and insecticides. The reasons are several: membranes house cells, control the influx and efl!iux of nutrients and metabolites; membranes 'Northwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA, 2725 Montlake Boulevard East, Seattle, WA 98112. ^Northwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA, 2725 Montlake Boulevard East, Seattle, WA 98112, and The College of Fisheries, University of Washington, Seattle, WA 98195. Manuscript accepted July 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. control and form the basis for the transmission of electrical signals (nerve impulses) when nerve receptor sites are stimulated. Any major altera- tion to normal membrane structure may be ex- pected to play some role in animal physiology, especially if the contaminated membrane is associated with neural function or other viable life processes. Membranes consist of a sandwich of phospholipids, sterols, and proteins; individual membranes are microscopically thin. The thinness and complexity of membranes make their study most difficult. One way of characterizing membranes is by measuring their electrical properties (conductance, resistance, capacitance, etc.). However, with respect to contaminant-host interaction, it is more desirable to be able to deduce structural features such as lipid fluidity, protein-lipid interaction, and arrangements of constituents in dynamic tissue preparations. This, then, precludes the use of electron microscopy or other methods which are incompatible with main- taining tissue in a viable unchanged condition. THE THEME OF THIS REPORT- SPIN-LABELING Several years ago, it was observed that certain free radical derivatives of fatty acids could be in- troduced into membrane preparations without unduly disturbing the natural arrangement of native membrane constituents (Jost et al. 1971). Furthermore, it was shown that these free radicals would associate with and align in membranes 299 FISHERY BULLETIN; VOL. 73, NO. 2 much the same way as do the nonradical natural fatty acids present in membranes (Libertini et al. 1969; Hubbell and McConnell 1969; Schreier-Muc- cillo et al. 1973). By using appropriate instrumen- tation, it was found that these radicals could be used as submicroscopic probes (or labels) for investigating membrane structure. This has become known as spin-labeling, and forms the underlying theme of this paper. This report describes our work on contaminant-host interac- tion in fingerling salmon. To show how these studies were performed, we need to arm ourselves with some basic background information. Let us review briefly some very important points con- cerning radicals, electrons, and nuclei. THE FOUNDATION OF SPIN-LABELING Free Radicals Many free radicals are known or have been isolated. Most, as we know, are quite reactive chemically, and unless conditions conducive to their formation and stabilization (trapping) are maintained, radicals normally disappear once formed. Radical reactivity stems from the fact MAGNET ON, ELECTRONS FREE IN SPACE >- o LJ z UJ $a MAGNET OFF 0 O- ENERGY LEVEL DIFFERENCE Figure 1. -Resonance condition for isolated electrons. In the absence of external field, energy levels are indistinguishable. When a magnetic field is applied, two levels (and only two levels, by quantum mechanical restrictions) are populated by electrons. By employing X-band microwaves (9 GHz) with the proper energy {hv = gpH; see for instance, Roubal 1972), flipping between levels occur and the absorption of microwave energy is observed. that radicals, by definition, are molecules which contain one or more unpaired electrons. Two radical partners normally pair together to yield end products with the normal complement of two electrons per chemical bond. This is the usual covalent bond and is characteristic of organic compounds. There is one class of free radicals, the nitroxides, which are stable under many of the usual labora- tory conditions. Nitroxide stability derives from resonance and other contributing factors, but we need not discuss these here. Many nitroxides are relatively easy to synthesize, provided the neces- sary starting intermediates (some of which are rare) are at hand. The important point to be made is the fact that nitroxides can be used to characterize biological systems. Nitroxides so used are called spin-labels; spin from the fact that it is the unpaired elec- tron(s) (which is/are spinning) which forms the basis for the label or probe. Spin-labeling might just as conveniently be called spin-probing. Spinning Electrons and Their Magnetic Properties All electrons are in a state of motion; they all spin about on their axis. Spinning electrons are therefore moving charges of electricity. Thus electrons are magnetic. Spinning electrons therefore are influenced by an external magnetic field such as produced by a solenoid or elec- tromagnet. When a sample of free radicals is placed between the poles of an electromagnet, the spin of the electron is described as clockwise or counterclockwise and is depicted in Figure 1. Of immediate consequence is the fact that one spin condition is more stable than the other, and is so indicated by the reference to the energy of the system as shown. Although one population level is more stable than the other, the temperature of the system is always great enough to insure that the higher level contains just about as many electrons as the lower level. This is something akin to plac- ing two bar magnets end to end. If they are aligned N-S N-S, we know from everyday experience that the interaction will be attractive and stable. If, on the other hand, we try to force them N-S S-N, we know again from experience that this is an unstable situation and requires an expenditure of energy (heat in the case of elec- trons, physical in the case of magnets) to maintain 300 ROUBAL and COLLIER: SPIN-LABELING TECHNIQUES them aligned in this fashion. Apart from this, however, the analogy extends no further, for in the realm of electrons and nuclear phenomena, quantum mechanical postulates hold and the com- mon experience of our everyday world does not pertain. The important point we want to make here is that we can induce an electron with one spin rota- tion to flip over and assume the alternate rotation. Energy is required to do this. To achieve flipping, the sample is irradiated (while between the poles of the electromagnet) by microwaves with a frequency of 9 gigahertz (GHz). Thus we see that what we are really talking about is just another type of spectroscopic method. The presence of free radicals is detected by measuring the loss of microwave energy. The actual instrument used for such studies is called the electron paramagnetic resonance (EPR) spectrometer (also called the electron spin resonance/ESR/spectrometer). At this point we must consider other factors which contribute to the total magnetism of the system. Remember that the unpaired electron is not merely floating freely about in space. It belongs to a molecule. In fact it is coupled to a nucleus, and in the case of nitroxides, to a nitrogen nucleus, which is itself a magnetic entity. Thus in the presence of an external field, the nuclear magnetism can couple with the external field and alter the magnetism immediate to the electron. We must now consider this situation, called hyperfine splitting (hfs). Hyperfine Splitting (hfs) Simply stated, it is found that the nitrogen magnetism can add to, subtract from, or be orthogonal (no interaction) to the external field. This is depicted in Figure 2. Note that the situation of Figure 1 is now modified. The original resonance (flipping) condi- tion is broken down into three resonances (hfs). The flipping from one level to the other is depicted by the double ended arrows A, B, and C. Certain restrictions are placed on flipping, and we find that only those shown are allowed. All levels are equally populated, and an actual EPR spectrum of a ni- troxide in dilute solution is shown in Figure 3. Spin-Label Spectra While the spectrum of Figure 3 tells us most conclusively that we are dealing with a nitroxide. MMNET ON ♦ NITNOeEN NUCLEUS CONTRItUTON MABNET ON HA6NET OFF/ -MAGNETIC FIELDS ADD -NO PARTICIPATION 8Y NITROOEN -MAGNETIC FIELDS SUBTRACT Figure 2.— Resonance condition for the case of one electron interacting with a nitrogen-N" nucleus. Due to quantum mechanical restrictions, only those transitions shown are allowed. it serves no further purpose other than a possible quantitation of the amount of radical (number of spins) present. If all we ever measured were three sharp, hyperfine lines, we could not use nitroxides as spin-labels. Fortunately, when a nitroxide is placed in an actual biological system, the hyperfine lines are modified both in shape, intensity (relative height), and in spacing. These spectral modifications are environment dependent, and it is this dependency which makes nitroxides valuable as probes for characterizing biological systems. -13 Ogauss — -13 Ogauss - J Center line at 3379 gauss MAGNITIC FIELD STKEnGTH - Figure 3.-EPR spectrum of the nitroxide compound potassium peroxylamine disulfonate (Freemy's salt) (lO-'M) in water at room temperature. Three hyperfine lines of equal intensity, and spaced 13 gauss apart, distinguish nitroxides in water. 301 FISHERY BULLETIN: VOL. 73, NO. 2 It is sufficient to simply state that the unpaired electron is localized on nitrogen and this localiza- tion resides primarily in a p-orbital on nitrogen. p-Orbitals are dumbbell-shaped electron density regions in space, and the magnetic properties as- sociated with an electron in such an orbital depends on the so-called tumbling frequency of the electron (how fast the p-orbital assumes a random distribution of orientations while part of the nitroxide in a biological system). An example of two limiting situations is shown in Figure 4. Spectrum 4A was recorded for a nitroxide in liquid glycerol, while 4B is for the frozen solution. Spec- trum 4A tells us that moderate fast tumbling prevails, while 4B shows tumbling to be essentially quenched. The reader is referred to the literature (Roubal 1972) for a more thorough discussion on the dependency of spectral characteristics on tumbling frequency (label mobility). Intermediate mobilities are characterized by a family of spectra. Examples are to be found in the recent spin-labeling study of a hapten combining site of trout antibody by Roubal et al. (1974). Using appropriate mathematical manipulations of recorded spectra, one can measure tumbling frequencies with accuracy. These frequencies together with other derived data provide quanti- tative characterization of labeling studies. MEMBRANES AND MEMBRANE PROPERTIES The importance of membranes in biological roles of living systems cannot be overemphasized. Membranes, as mentioned, are responsible for Figure 4.-Label II (see text) in glycerol. A. In liquid glycerol at room temperature. B. Frozen in glycerol at liquid nitrogen temperature. Label mobility (effect of environment) is calculated by measurements of line widths, heights, intensity ratios, and spacing. neural function. Membranes participate in ion- binding and in governing tissue permeability. As- sociated with membranes, especially mi- tochondrial membranes, are a variety of enzymes. Cytochrome oxidase and other electron transport enzymes are membraneous in nature. Membrane- bound enzymes require the proper conditions such as lipid fluidity, proper phase transition tempera- tures, and lipid-protein interactions for their function. The participation of membranes in neural control is well documented. Neural membranes contain molecular size pores which mediate sodium /potassium transport. The exact nature of these pores has not been delineated completely, but several lines of evidence suggest that pores consist of a cagelike arrangement of protein which spans the membrane from the inner to the outer surface. Membranes are considered to be the basis of life itself. Membranes consist principally of proteins and lipids. Carbohydrates comprise 0-10% of the membrane mass. Lipids account for about 40% of the mass, and the balance is protein. Membranes are a matrix of lipids and proteins arranged in a bimolecular leaflet (Singer 1972; Green 1972), illustrated in Figure 5. The little circles represent phospholipid headgroups (choline, ethanolamine, serine, phosphatidic acid, etc.) while proteins are indicated by the larger "islands." Interspersed with the fatty acid tails (squiggly lines) are sterols and lesser tissue lipid components. Typical membranes are about 100 A thick. Membrane lipids are amphiphilic-provided FiGURK 5. -Membrane bilayer leaflet. Small circles represent phospholipid headgroups. Squiggly lines are fatty acid tails. Large islands represent membrane protein. From Singer 1972. (Courtesy of S. J. Singer and New York Academy of Sciences.) 302 ROUBAL and COLLIER: SPIN-LABELING TECHNIQUES with polar headgroups and nonpolar tails of car- bon-hydrogen chains. Thermodynamically, and for other reasons as well, the greatest stability for membrane structure results when membrane con- stituents are arranged as shown; polar surfaces are exposed to aqueous environments, and the fatty acid tails are tucked away out of contact with water. QUESTIONS ABOUT CONTAMINANT-HOST INTERACTION We would like to answer the following questions concerning membrane-contaminant interaction: a) are certain regions of membranes affected or is the whole membrane affected when invaded by contaminant? b) if the effect is localized, where is the localization? c) are there differences in membrane perturbations when treated on the one hand by paraffins, and aromatics on the other? d) can the differences, if they exist, be related to anything presently known about the toxicology of any of these contaminants? EXPERIMENTAL METHODS AND RESULTS The study was performed in two steps. First, an in vivo feeding study was undertaken. Here we used spin-labeled hydrocarbons (Roubal 1974) and fed them to fish. Second, an in vitro study was employed using excised tissue. In order to restrict our investigations of membranes (in vitro) to specified regions, the series of labels I, II, and III (Roubal 1974a) (Figure 6) were synthesized by the EPR group of this Center. The positively .charged quartemary nitrogen of label I directs this portion of the label to the polar membrane surface, which in turn insures that the nitroxide nitrogen is sit- uated at or very near to the membrane surface. The carboxyl groups of labels II and III direct the carboxyl end of these amphiphiles to the membrane surface, but now the nitroxide nitrogen Ues some 12-15 A below the membrane surface in II, and deep into the hydrophobic membrane interior in III. Thus we can "look" at the membrane's surface, subsurface, and interior. In the feeding study, spin-labeled hydrocarbons were incorporated into fish food and fed to coho salmon, Oncorhynchus kisutch, fingerlings in a 2- day feeding study. Within an hour, or even less, after onset of intake of food by fish, the blood showed EPR activity. Using radiotracers, we have shown this activity to be associated with blood lipoproteins and albumins (Roubal 1974b), with lipoproteins making the greatest contribution to hydrocarbon transport in blood. After an induction period of about a day, blood- associated labels (Roubal 1974b) slowly transfer to neural tissue and flesh. Weight for weight, the greatest concentrations are to be found in the spinal cord, lateral line nerve bundles, and brain. The nature of the EPR line shapes indicated that the invasion of hydrocarbon is site selective. All labeled paraffins appeared to intercalate with membrane in such a way that the nitroxide mobility is little impeded. In direct contrast to this, mobility of aromatics appeared to decrease. This in vivo study suggests that paraffins associate with molecularly fluid portions of membrane fatty acids, while aromatics associate with the more structured and rigid regions of membranes. In order to clarify these possible differences, an in vitro study was undertaken using labels I, II, and III (Roubal 1974a). Neural tissue from untreated fish was carefully excised and placed in cold, 0.1 M phosphate buffer, pH 7.4 at 4°C, isotonic in NaCl. A sonicated dis- persion of label, complexed to bovine serum al- bumin was added and allowed to transfer to neural membrane (overnight at 4°C). Membranes were then inserted directly into the EPR spectrometer, and the spectra were recorded immediately. After ar CH^CH^IjiN N, N- Dimethyl- 4- dodecyl-4-tempoyl-annmonium bromide n. m. 0^ ^/v-o CH^-fCH^IjgC - (CH2)g COOH "7- Nitroxide" stearic acid ri: O ^N-O "12- Nitroxide" stearic acid Figure 6.-Lipid-intercalating spin-labels. Label I reports on surface conditions. Label II reports on subsurface conditions. Label III reports on interior conditions. 303 FISHERY BULLETIN: VOL. 73, NO. 2 the spectrum for one sample was obtained, the tissue was returned to a large volume of fresh, cold buffer to which test hydrocarbon (in separate tests) was incorporated via sonication. Final hydrocarbon concentration in buffer was 15-25 ppm. Hydrocarbons included benzene, toluene, ethyl benzene, hexane, heptane, octadedecane, and cyclohexane. Actual uptake of paraffin hydrocar- bon by tissue membrane, as measured by gas- liquid chromatography (GLC), was on the order of 1 ppm. Aromatics were present in higher amounts (5-10 ppm). Tissue was exposed to hydrocarbon in buffer for 1 h. At the end of this period, tissues were withdrawn, rinsed well, and the EPR spectra were re-recorded. A comparison of the in vitro spectra for controls with those same samples after hydrocarbon treat- ment provided evidence that a differentiation in binding sites for paraffins and aromatic com- pounds does indeed exist (Figure 7). DISCUSSION We can explain rather easily the preference of paraffins seeking the interior of membranes- paraffins are nonpolar, very soluble in neutral hydrocarbons, such as those which comprise the hydrogen-carbon chains (or tails) of phospholipid fatty acids. Hence, thermodynamically, system stability is enhanced by mutual interaction of paraffin hydrocarbon with lipid tails. Aromatics, on the other hand, are unique, for in addition to their ready solubility in many organic environments, aromatic compounds are fashioned from conjugated double bond systems with pi- electron unsaturation. These factors give aromat- ics the ability to form quasi-chemical complexes with other molecules which can act in electron ac- ceptor-donor roles. The surface of the membrane contains many different sites, both polar, nonpolar, and electron- interactive. It appears that some of these sites contain the necessary properties which make binding of aromatics possible. A charge-transfer mechanism (Kier 1971) may direct aromatics away from the membrane interior to the surface. These site preferences for paraffins and aromat- ics may account in part for the differences we observe for retention of these substances in living tissue. For instance, in other studies (Roubal 1974b) we have shown via GLC, spin-labels, and radiotracers that paraffins are retained in living tissue for long periods of time; aromatics are not. What is more, paraffins are relatively nontoxic, while aromatics generally are quite toxic, even at low levels. The molecular basis for physiological phenomena is associated in a very direct way with important membrane properties. Ion-binding, lipid protein (enzyme) interaction, lipid phase- transition temperatures, and lipid fluidity are all involved in one way or another. Membrane disor- LABEL I LABEL II Figure 7.-Spin-label spectra of aromatic-treated and paraffin- treated coho salmon spinal cord (SC). SC + Label I ( + treatment). SC + Label II (+ treatment). SC + Label III (+ treatment). Changes in the A/R (line height) ratio, narrowing of peak widths (W), and shifts in distance D to lesser values provide data for characterizing influence of treatments (Roubal 1972; Roubal 1974a). C, spinal cord control. AR, treatment by aromatics. PAR, treatment by paraffins. 304 ROUBAL and COLLIER: SPIN-LABELING TECHNIQUES ganization is considered to involve alterations in these properties. Accordingly, on invasion of membrane primarily by aromatics, surface per- turbations as indicated by spin-labeling spectra reflect changes in ion-binding properties of phospholipid headgroups, enzyme activity, and permeability changes. From the standpoint of membranes, paraffins are tolerated up to a point, by shunting them into internal reaches of membrane, away from active metabolic processes. Our spin-labeling studies show lipid fluidity to be altered when tissue is exposed to parafl!ins. These changes are rather diffuse, however, and not associated with any one portion of the membrane interior. Such alterations however, could be operative in altering ion trans- port. We contend, therefore, that aromatic- membrane interaction is of paramount concern. This is especially true when behavioral/ physiological patterns are to be explained. Addi- tional insight into these areas will necessitate further biophysical studies— both spin-labeling and broad-based electrophysiological studies. FUTURE OUTLOOK ON SPIN-LABELING Spin-labeling was first described only as recently as about 1968. Since then, a vast array of labels have been described. New instrumentation has evolved, and the technique has grown from a tool of limited application to one of major impor- tance. For many biochemical and biophysical studies, the technique stands prominently above other methods. Applied to drug studies, phar- macology, immunology, cancer research, en- zymology, and protein structure studies, spin- labeling promises to play an ever growing role. Environmentalists, biologists, zoologists, and food scientists now apply this tool to their studies. The future of spin-labeling is bright indeed. LITERATURE CITED Green, D. E. 1972. Membrane proteins: A perspective. Ann. N.Y. Acad. Sci. 195:150-172. HUBBELL, W. L., AND H. M. McCONNELL. 1969. Motion of steroid spin labels in membranes. Proc. Natl. Acad. Sci. 63:16-22. JOST, P., L. J. LiBERTINI, AND V. C. HERBERT. 1971. Lipid spin labels in lecithin multilayers. A study of motion along fatty acid chains. J. Mol. Biol. 59:77-98. KiER, L. B. 1971. Molecular orbital theory in drug research. Academic Press, N.Y., p. 137-161. LiBERTINI, L. J., A. S. Waggoner, P. C. Jost, and 0. H. Griffith. 1969. Orientation of lipid spin labels in lecithin mul- tilayers. Proc. Natl. Acad. Sci. 64:13-19. ROUBAL, W. T. 1972. Spin-labeling with nitroxide compounds. A new approach to the in vivo and in vitro study of lipid-protein interaction. In R. T. Holman (editor), Progress in the chemistry of fats and other lipids, 13:61-86. Pergamon Press, Long. 1974a. In vivo and in vitro spin-labeling studies of pollu- tant-host interaction. In R. Haque and F. J. Biros (edi- tors), Mass spectrometry and NMR spectroscopy in pes- ticide chemistry, p. 305-324. Plenum Press, N.Y. 1974b. Spin-labeling of living tissue. In E. J. Vernberg (editor), Pollution and physiology of marine organisms, p. 367-379. Academic Press, N.Y. RouBAL, W. T., H. M. Etlinger, and H. 0. Hodgins. 1974. Spin-label studies of a hapten-combining site in rain- bow trout antibody. J. Immunol. 113:309-315. Schreier-Muccillo, S., D. Marsh, H. Dugas, H. Schneider, and L C. P. Smith. 1973. A spin probe study of the influence of cholesterol on motion and orientation of phospholipids in oriented mul- tibilayers and vesicles. Chem. Phys. Lipids 10:11-27. Singer, S. J. 1972. A fluid lipid-globular protein mosaic model of membrane structure. Ann. N.Y. Acad. Sci. 195:16-23. 305 TELECONNECTIONS BETWEEN NORTHEASTERN PACIFIC OCEAN AND THE GULF OF MEXICO AND NORTHWESTERN ATLANTIC OCEAN James H. Johnson and Douglas R. McLain' ABSTRACT Anomalous large-scale air-sea interactions that took place over the Gulf of Mexico and the Atlantic Ocean off the southeastern coast of the United States in the winter of 1957-58 caused a change of sea-surface temperatures from near average values to cold anomalies of nearly 3.0°C in some regions. Evidence suggests that the mechanism for the anomalous change in sea temperatures derived from frequent outbreaks of cold, North American continental air flowing over the Gulf and ocean waters with consequent anomalously high evaporation and sensible heat exchange. A contributing factor may have been divergence of surface waters with associated upwelling in regions of high cyclonic activity. These severe outbreaks of cold continental air over the eastern seaboard may be related to air-sea interactions in the Pacific, thousands of miles distant. It is not clear what the full consequences of these events are to fisheries. The evidence which is available suggests that they are significant and warrant continued investigation. In recent years large-scale air-sea interactions have captured the interest of a number of oceanographers and meteorologists. Motivated by the desire to develop and improve extended me- teorological forecasts, Namias (1959, 1963, 1972, and in numerous other articles) has been one of the foremost investigators in studying these interac- tions. Most of his work has centered in the temperate latitudes of the northern hemisphere. Bjerknes (1966a, b, 1969) has studied large-scale interactions in the tropical Pacific Ocean, especially, processes associated with the El Nino phenomena. He has shown a relationship of fluc- tuations in the atmospheric Hadley circulation to large-scale anomalies of ocean temperatures. He suggests that an anomalously high heat supply in the equatorial Pacific, characterized by high equa- torial ocean temperatures, intensifies the Hadley circulation providing more than normal flux of angular momentum to the mid-latitude belt of westerly winds, thus affecting the weather over the North American continent. He suggests that regular monitoring of sea temperatures in the eastern tropical Pacific is indispensable in long- range weather forecasting for North America. Also motivated to develop improvements in long-range weather forecasting, Quinn (1972) has 'Pacific Environmental Group, c/o Fleet Numerical Weather Central, National Marine Fisheries Service, NOAA, Monterey, CA 93940. shown that large-scale interactions over the equa- torial Pacific may affect the weather over the continental United States. He suggests that ex- tensive dry-zone developments in the equatorial zone, which are associated with low sea-surface temperatures, may contribute to ridge develop- ment in the upper air circulation over the Northeast Pacific and, conversely, wet-zone developments (high equatorial sea temperatures) are associated with trough formation. Further- more, he describes the effects downstream over the United States following development of these troughs and ridges and implies that now it may be possible, if these extreme developments are de- tected early enough, to make long-range weather predictions over North America. Rowntree (1972), recently carrying out com- puter model studies, has confirmed that a sea temperature maximum in the tropical eastern Pacific leads to a persistent trough in the mid-lat- itude flow to the north. Regions of the ocean particularly responsive to the overlying atmosphere lie to the east of large continental land masses. Jacobs (1951), Manabe (1957), Wyrtki (1966), and Hishida and Nishiyama (1969) have shown that in wintertime the temperate western Pacific Ocean loses large amounts of heat through evaporation and sensible heat exchange because of the overflow of cold, dry Siberian air masses. Parker (1968) pointed out a similar effect in the winter of 1966 in the northwestern Gulf of Mexico. Manuscript accepted June 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 806 JOHNSON and McLAIN: TELECONNECTIONS BETWEEN OCEANS This paper extends the findings of several of the authors mentioned above by describing the development of large-scale sea-surface tempera- ture anomalies in the winter of 1957-58 in the Gulf of Mexico and off the U.S. eastern seaboard and suggests that these anomalies are associated with ridges and troughs in the upper air circula- tion. These, in turn, are associated with air-sea interactions in the Pacific. It further describes conditions in other winters where similar situa- tions developed and suggests that changes in abundance and distribution of some fish popula- tions both in the Pacific and Atlantic may be caused by events of this nature. DATA SOURCES Data used in heat budget calculations and in the description of sea-surface temperatures and anomalies were obtained from Tape Data Family 11 obtained by Fleet Numerical Weather Central and the Navy Oceanographic Oflfice from the Na- tional Climatic Center, Asheville, N.C. This file contains merchant ship weather reports, some of which date back to 1854. Computer programs were developed to compute monthly averages by year of sea-surface temperature and heat exchange by 5° blocks. Numerous studies have been made to determine the accuracy of sea-surface temperatures reported by merchant vessels in their weather reports. Saur (1963) found that in the Pacific the average injec- tion temperature bias from that of a bucket temperature taken at the surface was about -l-1.2°F, and Franceschini (1955) noted similar results in the Gulf of Mexico. The latter suggested, however, that commercial vessel reports could be used for practical purposes such as meteorological and oceanographic research and for forecasting air mass modification over the Gulf. Because un- certainties remain concerning the relation between temperature at the sea surface and that at the injection intake depth of merchant vessels, no attempt was made in this study to apply a correction. OBSERVATIONAL EVIDENCE OF ANOMALOUS AIR-SEA EVENTS The Anomalous Winter of 1957-58 Charts of the height of the 700-mb (millibar) pressure surface (on the average about 10,000 feet above the surface of the earth in temperate lati- tudes) are particularly significant in air-sea interaction studies because the mass circulation of the atmosphere can be inferred from them includ- ing areas of cyclogenesis and movement of storms. Furthermore, mean 700-mb heights over the ocean are highly correlated with 700-mb temperatures, which in turn are measures of the vertical stability in the atmosphere over the ocean. For example, low temperatures at the 700-mb level, associated with negative 700-mb height anomalies, indicate instability in the atmosphere and loss of heat from the ocean through convective processes in the at- mosphere. O'Connor (1958) reported that the heights of the 700-mb surfaces during the winter of 1957-58 were very unusual. He reported the following three ab- normalities in the 700-mb circulation in January 1958 (Figure 1) which persisted into February: 1. A trough in the east-central Pacific with 700-mb heights 550 feet below normal 700 miles south of Kodiak, Alaska, 2. A block in the Davis Straight where the 700-mb heights averaged 520 feet above normal and, 3. A trough in the southeastern United States (having 700-mb heights about 300 feet below normal) accompanied by strong northerly surface flow. The trough along the eastern seaboard is par- ticularly important. It suggests greater than nor- mal penetration of cold continental air masses over the southeastern United States and adjacent waters. Indeed, the winter of 1957-58 will be remembered as one of the most severe of the cen- tury. Data taken at National Weather Service stations in the southeastern United States show slightly higher than normal air temperatures in November 1957 (Table 1). A small negative anomaly developed in December which increased substantially in January 1958 and reached a maximum in February. Monthly deviations of air temperature ranged up to 5.8°C below normal at stations along the northeast coast of the Gulf of Mexico. Large negative sea-surface temperature anomalies also prevailed in February 1958 and were greatest in the northeastern Gulf in the same general area where shore station air temperature anomalies were highest (Figure 2). Stearns' (1964, 307 FISHERY BULLETIN: VOL. 73, NO. 2 Figure l.-700-mb heights and departures from normal, tens of feet, January 1958. Major departures are enclosed in boxes. Table l.-Monthly average air temperatures (°C) and departures from average (°C), November 1957-March 1958 at selected National Weather Service Stations.' Novem Temper- ber1957 Depar- Decem ber 1957 Janua ry 1958 Februa ry 1958 Marc h 1958 Temper- Depar- Temper- Depar- Temper- Depar- Temper- Depar- Station ature ture ature ture ature ture ature ture ature ture New York, N.Y. 9.6 + 1.6 4.8 +2.7 0.0 -0.4 -2.6 -2.8 4,5 -0.3 Wilmington, Del. 8.1 +0.6 3.3 + 1.6 -0.7 -1.4 -2.6 -3.6 3.9 -1.9 Wilmington, N.C. 13.4 +0.2 9.6 +0.4 4.9 -3.8 4.9 -4.3 9.5 -3.1 Charleston, S.C. 15.4 +0.6 10.9 -0.3 7.0 -4.1 6.5 -5.0 11.8 -2.7 Savannah, Ga. 15.0 +0.8 10.3 -0.6 6.6 -4.3 6.8 -5.1 12.2 -2.6 Jacksonville, Fla. 18.5 +2.1 12.1 -1.5 9.1 -4.1 9.0 -5.2 15.1 -1.7 Miami Beach, Fla. 24.1 +0.8 19.9 -1.7 17.7 -3.4 16.6 -4.6 22.0 -0.3 Key West, Fla. 25.2 + 1.6 20.3 -1.8 18.5 -3.0 17.5 -4.5 21.3 -1.8 Fort Myers, Fla. 22.0 + 1.3 16.5 -2.1 13.9 -4.2 13.1 -5.6 18.8 -1.5 Pensacola, Fla. 16.2 +0.6 11.9 -0.9 8.3 -3.9 7.6 -5.7 14.3 -1.4 Mobile, Ala. 15.2 +0.4 11.2 -0.7 7.3 -4.2 6.9 -5.8 13.6 -1.6 New Orleans, La. 16.5 +0.5 12.9 -0.5 9.8 -3.5 9.1 -4.9 15.3 -1.7 Galveston, Tex. 17.2 -0.3 14.9 + 1.3 10.7 -1.7 10.4 -3.6 14.3 -2.3 Corpus ChristI, Tex. 17.3 -0.9 16.4 + 1.5 12.8 -1.1 13.9 -1.8 15.7 -2,8 iSource: Local Climatological Data reports, U.S. Department of Commerce, National Weather Service. 308 JOHNSON and McLAIN: TELECONNECTIONS BETWEEN OCEANS ^0.8 0.9 ^ ciJ + +'^ + + + ==2^ 0.3 0.3 0.21.0 1.4 - , . - . _ + 1 .4'0.8 0.4 0.40.1 0.4 2.01.0 0.4 0.20.0 0.1 ,61.7 1.2 2.,51.51.2 0.9i 0.6P.6!d.4 -- - -i. __4. + OVIO.6 0.4-2;51,00.6 0,2 0,30.3 0,2 I ■ ,■■ ,^ - -f + _i+ + + + 0.6 0.10.10.1 0.4 0.50.1 0.6 + + + + + + + 0.41.01 .3 0.7 0.60.9 0.3 T '■'■■■' I " ■- f ■ ■" 110° 100* 90° 80' 70 60 50 Figure 2.-Sea-surface temperature anomalies (°C) Gulf of Mexico and western Atlantic Ocean, February 1958. Shaded area colder than 20-yr (1948-67) mean. 1965) analysis of sea temperatures at coastal sta- tions confirm the large winter sea temperature anomaly in coastal areas of southeastern United States. The extent of the anomaly was large, in- deed, extending from below the Yucatan Straits northward throughout the Gulf to lat. 40°N off the eastern seaboard and to over a thousand miles offshore. Important in this study is the anomalous change in sea temperatures in the winter of 1957-58 (Figure 3). The change from November to February shows that much higher than average cooling occurred over a broad expanse of the ocean during this winter. For example, the 1948-67 average change in sea-surface temperature in winter (November average - February average) is /rr^^ (^ 1 .n.6 <=C7 - -'^+ + + cc^ 0.6O.9 0.9 1.S1.6 - / - . - + + 2.61 .50.1 1.0 0.60.6 2.20.50.3 0.1 0.10.2 1.3 1.0:1.4 2.91.50.90.5 0.30.50.5 - - : - + + on, v. 2-0, 7, 2.50.8 0.60.3 0.1 0^30^3 0.9 0.40.2 0.00.1 0.5 0.21.0 1+ + + + + + + 0.10.6 1.00.3 0.4 0.70.1 I o I o 100 90 80 Vo° 60 50 Figure 3.— Anomalous sea-surface temperature change (°C) November 1957 to February 1958. Shaded area indicates anomalous cooling. about -2.7°C in the area lat. 25° to 30°N, long. 80° to 90°W. In the winter of 1957-58, however, the change in this area was -4.7°C-an anomalous change of -2.0°C. Cause of Sea-Surface Temperature Change in the Winter of 1957-58 A variety of processes can cause changes in sea-surface temperatures. Some of these are horizontal and vertical advection and heat exchange at the air-sea interface. In studies in the eastern Pacific Ocean, Clark (1972) has found that horizontal advective heat transfer processes in winter and spring have a greater effect in causing anomalous sea-surface temperature change than nonadvective processes. The latter have a greater effect in summer and fall. The area that he studied, however, was not subject to influence of a large continental land mass interacting with atmospheric circulation up- stream from where heat exchange processes were calculated. Horizontal advection into the Gulf of Mexico is mainly through the Yucatan Straits. Sea-surface temperature anomalies in these Straits ranged from -0.2°C to -H 0.4°C (Table 2) for a year preced- ing the maximum development in February 1958 of the severe cold anomalies in the Gulf of Mexico and along the eastern seaboard. This clearly sug- gests that horizontal advection was not the cause of the development of the cold anomalies. Klein (1958) noted that the intensification of blocking in the 700-mb circulation in the north- west Atlantic in the winter of 1957-58 was as- sociated with frequent outbreaks of cold air in the Table 2. -Sea-surface temperatures and anomalies (°C) in 1° square (long. 85° to 86°W and lat. 21° to 22°N) in Yucatan Straits. Date Sea temperature °C Anomaly °C 1957: January 26.2 February 26.6 March 26.4 April 27.0 May 28.0 June 28.7 July 29.3 August 29.7 September 29.5 October 28.8 November 28.0 December 27.5 1958: January 26.2 February 25.6 -0.2 +0.3 -0.1 +0.1 +0.2 0.0 +0.1 0.0 -0.1 -0.2 0.0 +0.4 -0.2 -0.7 309 FISHERY BULLETIN: VOL. 73, NO. 2 eastern two-thirds of the United States. The con- trast between this cold air and the air heated by the Gulf of Mexico and western Atlantic waters led to baroclinic deepening of coastal storms. In these regions of cyclonic activity, vertical advec- tion may have caused some cooling through diver- gence of surface water and consequent local up- welling. Leipper (1967) has shown that hurricane Hilda, passing over the Gulf of Mexico in the early fall of 1964, indeed did cause significant cooling of surface waters through upwelling processes. However, the large area covered by anomalously low sea-surface temperatures in the winter of 1957-58, and other factors such as the frequent cold outbreaks over the entire eastern seaboard and the fact that the sea temperature anomalies ap- peared to occur over large areas contemporaneous with the overflow of cold air, suggests that the high rate of sea surface cooling was due to anomalously high loss of heat through evaporation and conduction of sensible heat. To test this supposition formulae for calculation of the heat exchange at the air-sea interface have been employed. Laevastu (1965), Seckel (1962) and many others have provided these formulae. In this study the procedures for calculation of the energy exchange as presented by Johnson et al. (1965) are used. It is not the intent here to review the ac- curacy of the various techniques for estimating air-sea energy exchange. Because of the possible inaccuracies of input data and of the formulae, the exchange values should be viewed with caution and should be considered only as relative indices of the magnitude of energy flux at the air-sea inter- face. They appear, however, to be sufl'iciently ac- curate to permit detection of large-scale seasonal and nonseasonal variations. The equation for the heat exchange at the air- sea interface is Qt = Qi- Qr-Qb- Qe- Qh where: Qt ~ Net heat gained or lost at the sea surface Qj = Incident solar radiation corrected for cloud cover Qfi = Reflected radiation Qg = Back radiation Qe = Evaporation Qff = Conduction of sensible heat. 1948 +1.0 1950 1952 1954 1956 YEAR 1958 1960 1962 1964 1966 0.0 e U < X u Ul UJ u oe M -1.0 — n — / \ / \ / \ •SEA-SURFACE TEMPERATURE CHANGE -NET HEAT EXCHANGE +100 -100 -200 < CM o >^ CO — O -I < -300 z I o X < UJ I -400 -500 1950 1952 1954 1956 1958 YEAR 1960 1962 1964 1966 Figure 4.-Relation between net heat exchange at air-sea boundary and sea-surface temperature change winters of 1948-67 in the area bounded by long. 80° to 90°W and lat. 25° to 30°N. I 310 JOHNSON and McLAIN: TELECONNECTIONS BETWEEN OCEANS The relation between Q-p and sea-surface temperature change in the winters of 1948-67 in the northern Gulf of Mexico is significant (Figure 4). Particularly noteworthy is the large net heat loss in the winter 1957-58 accompanied by the large sea-surface temperature change. The decrease in Q/ reaching the sea surface accounted for about 11% of the anomalous heat loss and increased Q^ about 12%. Q^. and Qf^, however, clearly stand out as significantly more important than the other heat budget elements in contribut- ing to the large net loss of heat. Q^; contributed 57% of the anomalous heat loss and Q//, 20%. Data are not available that indicate the depth to which the sea temperature anomaly extended. One might speculate that it should extend throughout the mixed layer which on the average in the area of study is about 80 m in winter (Robinson 1973). For the three months December 1957-February 1958, the anomalous heat loss in the Gulf of Mexico in the area of lat. 25° to 30°N, long. 80° to 90°W was approximately 19,000 cal/cm-. This should reflect an anomalous change in water temperature of about -2.4°C throughout the mixed layer which is very close to the anomalous -2.0°C change ob- served at the surface. Though the immediate cause of the development of the cold anomaly appears to have been the flow of cold continental air over the Gulf and western Atlantic, the more general cause may have been due to large-scale interactions over the North Pacific Ocean. A large positive sea temperature anomaly appeared in the eastern Pacific in late 1957 and persisted throughout 1958 (Sette and Isaacs 1960) (Eber 1971). Namias (1959) explains that the contrast of anomalous warm ocean temperatures to the east of cold ocean tempera- tures, which was the situation in 1957 in the east- em Pacific, provided abnormal feedback in heat exchange processes to the atmosphere which provided the additional baroclinicity upon which cyclones could feed. This cyclogenesis helped maintain the deep Pacific trough south of Kodiak Island which was abnormally intense by the late fall of 1957. Downstream from this area of ac- tivity, a responsive ridge developed in the western United States (evident from Figure 1) and a deep trough along the Atlantic seaboard. This distribu- tion is also consistent with the statistical findings of O'Connor (1969) who noted that when an anomalously deep trough forms in the 700-mb cir- culation in the east central Pacific, the chances for a trough off the eastern seaboard are high. Air-Sea Interactions in Other Years It is tempting to argue that events such as the 1957-58 occurrence described above occur so sel- dom that it is not worth the effort to study them and their effects on fisheries. Study of years when extreme conditions prevail, however, provides hints of the processes that are occurring in the natural system in other years. Definitive findings through study of more normal years are often diflficult to obtain because processes involved may be obscured by the subtle interactions of a number of factors. The interaction of the type described above may not be as infrequent as one might believe. A sit- uation similar to the winter of 1957-58 seems to have occurred in the winter of 1939-40. O'Connor (1958) noted very cold air temperatures along the eastern seaboard in the winter of 1939-40 which were as intense as those in 1957-58. Although sea temperature records are sparse for winter months (December through February) of 1939-40, what data are available show an extremely cold and widespread anomaly in the Gulf of Mexico (Figure 5) and along the eastern seaboard where temperature anomalies of 5° and 6°C below nor- mal were observed in February 1940. Further analysis of these large-scale air-sea interactions suggest a possible relation between equatorial Pacific Ocean temperatures and those in the Gulf of Mexico. Bjerknes (1969) has shown that in the winters 1957-58, 1963-64, and 1965-66 high sea-surface temperatures prevailed in the eastern tropical Pacific Ocean. This is characteris- tic of El Nino years which are best known for the invasion of warm water off the Peruvian coast and effects on the anchoveta populations there. It is known also that in the winter of 1939-40 a severe El Nino was present in the eastern Pacific. In all of these winters cold sea-surface temperature anomalies prevailed in the Gulf of Mexico (Figure 5). There appears, then, to be a relation between sea temperature anomalies in the equatorial Pacific and anomalies in the Gulf of Mexico, that is, negative sea-surface temperature anomalies in the Gulf and western Atlantic in some situations may be related to positive sea temperature anomalies in the eastern equatorial Pacific through processes described by various authors mentioned previously and those described in this paper. 311 1939-40 FISHERY BULLETIN: VOL. 73, NO. 2 1957-58 lOO 1963-64 1965-66 lOO" 95' 90 85' 80 100 95 90 85 Figure 5.-Sea-surface temperature anomalies (°C) in the Gulf of Mexico for selected winters. 80" During the recent 1972-73 El Nino, however, a situation occurred where this relation did not exist. A trough did not develop off the eastern seaboard until late winter and was short-lived in nature. During most of the winter it was situated over the central United States. Consequently, the flow of cold, continental air over the Gulf of Mexico, especially in the eastern Gulf, was not as intense as in previous El Nino years, and thus not as much cooling of surface waters occurred in the Gulf of Mexico. A situation opposite to the cold winters along the eastern seaboard described above occurred in the winter 1948-49. Very little cooling occurred in the waters of the Gulf, and the net heat exchange atthe air-sea interface likewise was small (Figure 4). The sea-surface temperature anomaly patterns in the winter of 1948-49 compared to the winter 1957-58 are remarkably opposite in sign and mag- nitude (Figure 6). Whereas, cold sea temperature anomalies prevailed in the latter winter off the southeastern United States, widespread warm anomalies were present in the winter 1948-49. In this winter a distinct ridge developed over the eastern United States in the 700-mb heights, a trough over the western U.S., and another ridge in the northeast Pacific (Figure 7) which is consistent with the hypothesis given by Quinn (1972) and by the findings of Namias. Namias (1972) suggested that the 1957-58 winter marked the beginning of a new climatic regime in the northern hemisphere. He shows, for example, that the winter mean air temperature at Atlanta, Ga. for the decade 1948-57 was about 5°F higher than the following decade. A trend is also noted in decadal differences of sea temperatures in the Gulf of Mexico. The 1948-57 decade average of February sea temperature was about 1.0°C higher than the 1958-67 February average. Somewhat lower sea temperatures generally prevailed along the entire eastern seaboard in the 1958-67 decade compared to the one preceding. Sea temperatures and circulation off the U.S. west coast also showed a climatic change. Huang (1972) calculated that southward transport in the general area of the California Current from San Diego to long. 150°W in the period 1958-69 was less than in the previous decade. He showed further that the California Current annual sea tempera- tures were as much as 1.4°C above normal in the 312 JOHNSON and McLAIN: TELECONNECTIONS BETWEEN OCEANS 60' lA 50'^ 40 e 30H 2(? Itf 3. 22. 4 0.33. 5 1 .30. 71. 03. 5 FEBRUARY 1949 + + +- --_:- 1.00.9 0.30.7!o.90.8 1.71.3 + + + +'-'--. 2.42.0 2.00.3 0.90.7 0.91 .4 + + + + - - - 3.23.7 2.31.1 0.01.0 1.91.5 + + + - -- _ -- 2. 0 1. 3 0.6: 0.3 0.50. 9 1.91.1; 0.1.-;^ __ O.f'b.7 ,::^ ^2.0 0.90.2 0.3 + /+ + + + 1.62.5 1.0 0.40.9 0.3 + + + + + 0.4 0.0 0.91.00.9' 1.5' ,0.9:T.-e\ J 1.3 1.5 2.01.50.4 0.2 0.00. - - i .^,< \ \ + + + + . _ ■ - : _ 1.21.1 0.7 0.30.1 0 0.4 !• 3 0.6; 0.8 1.0 0.80.7 " ' " ' " ■ 1.0 0.2 1.0, •0. I \ 1-1 1.1 1.41.00.2 0,4,0.20,7:0.6 1.0 ' 1.7 0.9 1.0 0.10.30.4 0.4 0.2:0.8 0.0 + ++---..--- 1.2 1.20.5 0.7 0.71 on «;o >^0.&0.4:n.3 o 'o 'o 'o 'o 'o 'o 'o 'o 'e 'o 'e 160 150 140 130 120 110 100 90 80 70 60 50 60' 50 TT" 7y 40- 30- 2tf 10 - 1.2 0.& + I + ! + I + i + 0.0 0.8 1.20.9 0.80.8 FEBRUARY 1958 ++ ++ +>+ D. 20. 3 0.10.4 0.71 .2 1.2 1.0 - - - + + + + 0.8 0.5 0.40.3 0.10.4 0.8 1.0 - , - . + + + + 0.6 1.0 1.00.6 0.60.9 1 .4 1 .8 _,--,+ + + + + 1.2 1.0 0.30.110.30.7 1.22.02.0r\ - + + , + + _, + '-',+, J^ 0.8O.9 _ <=r7 + ^+ v+ + + "^^ 0.30.3 0.2 1.01.4 1.40.80.4 0. 40. 10. 4 ,2.0.1.00.4:0.2:0.0,0.1 0.8,0.6; 0.?0. 2 0.60.7 0.70.21.9 ?v8!\ ; 1 ^ 1 .71 .2 2.5i I'si '20*9 o'§io"6b'4 + + + + +■-■: +'., 0 10 2 0.7100908 " ' 2,2 2,21,9. O'J :Q.60.4 '2.5; l.00.60,.,2. iO;3IO. 30.2 0.2 0.5 " 0.6 1.01.5 0.80.91.2 0.50.5 0.20.0^ O.frQ.l 0.10.10.4 0.5 0.10.6 -+ ++ + + + + + + + 0-g ",9-5; ' p.O 0.2^0.6 0.4-1.4 0.4^ 1.01.3p.7 0.6^Q.9n..3 160' 150" 140" I30° 120* IIO" 100* 90° 80* 70° 60° 50" Figure 6. -Sea-surface temperature anomalies (°C) February 1949 and February 1958 in eastern Pacific and western Atlantic oceans. Shaded areas colder than 20-yr (1948-67) mean. 1958-69 period and the seasonal temperature departure in winter was greater than 2°C in some places. Apparently, this was caused by less advec- tion of cold subarctic water southward. DISCUSSION The authors have attempted to show that development of sea-surface temperature anomalies in the Gulf of Mexico and along the U.S. eastern seaboard in wintertime may be related to the origin of the overlying air masses. In situa- tions where trough development occurs in the upper air circulation over the eastern United States, cold continental air from North America is likely to flow over the Gulf and waters off the eastern seaboard causing excessive loss of heat through evaporation and conduction of sensible heat. Conversely, in situations where a ridge develops, warm air masses predominate, and loss of heat from the ocean is retarded. In fact, as Figure 4 indicates, the water in some winters may cool very little. Furthermore, an attempt has been made to show that the development of these 313 FISHERY BULLETIN: VOL. 73, NO. 2 Figure 7.-700-mb heights and departures from normal, tens of feet, January 1949. Major departures are enclosed in boxes. troughs and ridges in the upper air circulation may have been related to air-sea interaction processes in the Pacific Ocean. It is interesting to speculate on the physical consequences following the development of a large-scale cold sea-surface temperature anomaly in the v^^inter of 1957-58 in the Gulf of Mexico and v^estern Atlantic. It is quite possible that further investigation vi^ill show differences of flow in the Gulf Stream. The fact that the anomaly developed in winter suggests that the water to the depth of the thermocline may be affected. A large cold mass of water of this type alters the density dis- tribution over a large area and thus alters the surface circulation. The authors have noted an ap- parent drift of the anomaly away from the U.S. east coast northeastward until the early summer of 1958 when the surface anomalies were ob- scured, although it may be quite possible and even likely that the deeper waters were still colder than usual. The effects on Europe of this cold water mass after a period of eastward drift can only be speculated at this time. The fact that the anomaly could be traced for a time following its formation suggests an interesting possibility for further research. The biological consequences of such large-scale air-sea interactions are even more complex. The effect of the warm sea temperature pool in the eastern Pacific in 1957 and 1958 on fish populations is well documented by Radovich (1961). Southern species were found much farther north than nor- mal in the temperate northeastern Pacific ap- parently in response to the warm sea tempera- tures. In the western Atlantic and Gulf the effects of the cold anomaly were less evident although there is a suggestion from the work of Williams (1969) that the shrimp populations were affected. 314 JOHNSON and McLAIN: TELECONNECTIONS BETWEEN OCEANS Williams believes that catches of penaeid shrimp in the southeastern United States fluctuate in such a way as to suggest dependence on coastal temperatures. His studies show an apparent as- sociation of good catches with warm years and poor catches with cold years. The shrimp season following the cold winter of 1957-58 was par- ticularly poor in several areas. His indices of cold and warm years are derived from coastal air temperatures which probably are a reasonable in- dicator of sea temperature variation in estuarine and coastal waters. Williams believes it might be possible to use winter coastal air temperatures as predictors for the succeeding year's catch when betterdefinition and measure of fishing effort are available. These hints of biological consequences of large- scale air-sea interactions point out possibilities for future research. Clearly, investigations of this nature will require a cooperative effort among meteorologists, oceanographers, and fishery biologists. ACKNOWLEDGMENTS We wish to thank J. Bjerknes, J. Namias, and W. H. Quinn for reviewing our manuscript and for providing much encouragement and many useful suggestions. LITERATURE CITED Bjerknes, J. 1966a. A possible response of the atmospheric Hadley cir- culation to equatorial anomalies of ocean tempera- tures. Tellus 18:820-828. 1966b. Survey of El Nino 1957-58 in its relation to tropical Pacific meteorology. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull. 12:25-86. 1969. Atmospheric teleconnections from the equatorial Pacific. U.S. Dep. Commer., Mon. Weather Rev. 97:163-172. Clark, N. E. 1972. Specification of sea surface temperature anomaly patterns in the eastern North Pacific. J. Phys. Oceanogr. 2:391-404. Eber, L. E. 1971. Characteristics of sea-surface temperature anomalies. Fish. Bull., U.S. 69:345-355. Franceschini, G. a. 1955. Reliability of commercial vessel reports of sea surface temperatures in the Gulf of Mexico. Bull. Mar. Sci. Gulf Caribb. 5:42-51. HiSHIDA, K., AND K. NiSHIYAMA. 1969. On the variation of heat exchange and evaporation at the sea surface in the Western North Pacific Ocean. J. Oceanogr. Soc. Jap. 25:1-9. HUANG,J.C. K. 1972. Recent decadal variation in the California Current System. J. Phys. Oceanogr. 2:382-390. Jacobs, W. C. 1951. The energy exchange between sea and atmosphere and some of its consequences. Bull. Scripps Inst. Oceanogr. Univ. Calif. 6:27-122. Johnson, J. H., G. A. Flittner, and M. W. Cline. 1965. Automatic data processing program for marine synoptic radio weather reports. U.S. Fish Wildl Serv., Spec. Sci. Rep. Fish. 503, 74 p. Klein, W. H. 1958. The weather and circulation of February 1958: A month with an expanded circumpolar vortex of record intensity. U.S. Dep. Commer., Mon. Weather Rev. 86:60-70. Laevastu, T. 1965. Daily heat exchange in the North Pacific, its relations to weather and its oceanographic consequences. Soc. Sci. Fenn. Commentat. Phys.-Math. 31(2), 53 p. Leipper, D. F. 1967. Observed ocean conditions and hurricane Hilda, 1964. J. Atmos. Sci. 24:182-196. Manabe, S. 1957. On the modification of air-mass over the Japan Sea when the outburst of cold air predominates. J. Meteorol. Soc. Jap., Ser. 11,35:311-325. Namias, J. 1959. Recent seasonal interactions between North Pacific waters and the overlying atmospheric circulation. J. Geophys. Res. 64:631-646. 1963. Large-scale air-sea interactions over the north Pacific from summer 1962 through the subsequent winter. J. Geophys. Res. 68:6171-6186. 1972. Large-scale and long-term fluctuations in some at- mospheric and oceanic variables. In. D. Dyrssen and D. Jagner (editors). The changing chemistry of the oceans, p. 27-48. Nobel Symp. 20. O'Connor, J. F. 1958. The weather and circulation of January 1958: Low index with record cold in southeastern United States. U.S. Dep. Commer., Mon. Weather Rev. 86:11-18. 1969. Hemispheric teleconnections of mean circulation anomalies at 700 millibars. U.S. Dep. Commer., ESSA Tech. Rep. WB 10, 103 p. Parker, C. A. 1968. The effect of a cold-air outbreak on the continental shelf water of the Northwestern Gulf of Mexico. Dep. Oceanogr., Tex. A & M Univ., Ref. 68-3T, 89 p. Quinn, W. H. 1972. Large-scale air-sea interactions and long-range forecasting. In The 2nd International Ocean Develop- ment Conference, October 5-7, 1972, Keidanren Kaikan, Tokyo. Preprints Vol. 1:226-254. Radovich, J. 1961. Relationships of some marine organisms of the Northeast Pacific to water temperatures particularly during 1957 through 1959. Calif. Dep. Fish. Game, Fish Bull. 112, 62 p. Robinson, M. K. 1973. Atlas of monthly mean sea surface and subsurface temperature and depth of the top of the thermocline Gulf of Mexico and Caribbean Sea. Scripps Inst. Oceanogr., Univ. Calif., Ref. 73-8, 12 p., 93 fig. 315 ROWNTREE, P. R. 1972. The influence of tropical east Pacific Ocean tempera- tures on the atmosphere. Q. J. R. Meteorol. Soc. 98:290-321. Saur, J. F. T. 1963. A study of the quality of sea water temperatures reported in logs of ships' weather observations. J. Appl. Meteorol. 2:417-425. Seckel, G. R. 1962. Atlas of the oceanographic climate of the Hawaiian Islands region. U.S. Fish Wildl. Serv., Fish. Bull. 61:371-427. Sette, 0. E., AND J. D. Isaacs. 1%0. Editors' summary of the symposium. In Symposium on "The changing Pacific Ocean in 1957 and 1958." Calif. FISHERY BULLETIN: VOL. 73, NO. 2 Coop. Oceanic Fish. Invest. Rep. 7:211-217. Stearns, F. 1964. Monthly sea-surface temperature anomaly graphs for Atlantic coast stations. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 491, 2 p., 2 fig. 1965. Sea-surface temperature anomaly study of records from Atlantic coast stations. J. Geophys. Res. 70:283-296. Williams, A. B. 1969. Penaeid shrimp catch and heat summation, an ap- parent relationship. FAO (Food Agric. Organ. U.N.) Fish. Rep. 57:643-656. Wyrtki, K. 1966. Seasonal variation of heat exchange and surface temperature in the North Pacific Ocean. Hawaii Inst. Geophys. Univ. Hawaii, 66-3, 7 p., 72 figs. 316 GEOGRAPHIC AND HYDROGRAPHIC DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE ALONG THE MIDDLE ATLANTIC COAST FROM RV DOLPHIN CRUISES, 1965-66 Arthur W. Kendall, Jr.' and John W. Reintjes- ABSTRACT Atlantic menhaden, Brevoortia tyrannus, eggs and larvae were collected during eight ichthyoplankton cruises of the RV Dolphin from December 1965 to December 1966. On each cruise tows were made with a Gulf V plankton net at 92 stations along 14 transects from the coast to the edge of the continental shelf from Martha's Vineyard, Mass., to Cape Lookout, N.C. Larvae resulting from a protracted spawning season were taken throughout the year. Eggs were taken over the middle of the shelf in fall. Seasonal shifts in geographic pattern of larvae indicated spawning started in summer off New Jersey and New York, became widespread in the Middle Atlantic Bight in fall, and continued into winter off North Carolina. Larvae were equally distributed in shallow (0-15 m) and deep (18-33 m) tows during night and day. Larvae occurred over a water temperature range from 0° to 25°C and a salinity range of 29 to 36"/ 00. Seasonal distribution of larvae suggests some of the annual variation in year classes may be due to cold-related mortality of larvae entering middle Atlantic estuaries in late fall. Along the Atlantic coast spawning and early development of many fishes occur in the ocean. The distribution of early stages of most coastal species is inadequately known. Personnel of the Sandy Hook Marine Laboratory designed a program to determine the spawning times and localities of migratory coastal fishes through a series of cruises off the Atlantic coast from Martha's Vineyard, Mass., to Cape Lookout, N.C. From December 1965 to December 1966 eight sur- vey cruises were conducted to collect fish eggs, larvae, and juveniles. These cruises, together with data on juvenile and adult distribution, provided information on the oceanic life history of most of the commercial and sport fishes of the region. Atlantic menhaden, Brevoortia tyrannus (La- trobe), an important commercial and forage fish, was among the species collected during this sur- vey. The early life history of menhaden has puzzled scientists since early accounts by Baird (1873) and Goode (1879). The eggs and larvae were first described by Kuntz and Radcliffe (1917). After early development at sea, the larvae enter estuaries along the coast where they me- tamorphose into juveniles. June and Chamberlin (1959) concisely review the estuarine stage of menhaden. 'Middle Atlantic Coastal Fisheries Center Sandy Hook Laboratory, National Marine Fisheries Service, NOAA, Highlands, NJ 07732. 'Atlantic Estuarine Fisheries Center, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. The seasonal cycle of menhaden spawning has been inferred from ovary studies by Higham and Nicholson (1964). From Maine to eastern Long Island ovarian development starts in May and reaches a peak in October. The seasonal occurrence of sexually mature fish begins in New Jersey and Delaware in April, continues sporadically in summer, and reaches a peak in October. Around Cape Hatteras maturing fish were taken mainly in late fall; some are found as late as March. Atlantic menhaden eggs and larvae have been collected offshore and in estuarine waters along the Atlantic coast (Table 1). These collections show a pattern similar to that found by Higham and Nicholson (1964). Spawning off New England oc- curs in late spring and early summer and again in early fall. Off the middle Atlantic coast eggs and larvae are found in late fall and in spring. Off North Carolina young occur in winter and spring. Inlet and estuarine studies have collected larval menhaden as they emigrate from the ocean (Table 1). The time of entry varies considerably along the coast and from year to year in the same estuarine areas. In some years entry occurs in late fall, before lowest temperatures are reached. In other years entry occurs primarily as temperatures are warming in early spring. Larvae in the estuaries are apparently killed if winter temperatures fall below 3°C (Reintjes and Pacheco 1966). Emigra- tion from the estuaries varies from late August in the north to January in the south, with some Manuscript accepted August 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 317 FISHERY BULLETIN: VOL. 73, NO. 2 Table 1. -Collections of menhaden eggs and larvae, east coast of United States. Sampling period Reference Years Months Open ocean Marak and Colton 1961 1953 Mar.-June Marak, Colton, and Foster 1962 1955 Feb. -May Marak, Colton, Foster, and 1956 Feb. -June Miller 1962 Reintjes 1961 1953-54 all Massmann et al. 1962 1959-60 all Reintjes 1969 1966 Dec. Sampling area Occurrences Bays and open sounds Bigelow and Schroeder 1953 Kuntz and Radcliffe 1917 Hildebrand and Schroeder 1928 1912-22 Pearson 1941 1929 May-Oct. 1930 Apr.-Dec. 1931 Jan. -Mar. Perlmufter 1939 1938 May-Oct. Merriman and Sclar 1952 1943-46 all Wheatland 1956 1952-53 all Richards 1959 1954-55 all Deubler 1958 1955-57 Dec. -Apr Herman 1963 1957-58 all Croker 1965 1960-61 all Dove! 1971 1963-67 all Georges Bank-Gulf of Maine Georges Bank-Gulf of Maine Georges Bank-Gulf of Maine Cape Hatteras-Florlda Off Chesapeake Bay Off North Carolina Gulf of Maine Near Woods Hole, Mass. Chesapeake Bay Lower Chesapeake Bay Lower Chesapeake Bay Lower Chesapeake Bay Around Long Island Block Island Sound Long Island Sound Long Island Sound Bogue Sound, N.C. Narragansett Bay Near Sandy Hook, N.J. Chesapeake Bay June— 1 egg, off Martha's Vineyard May— eggs, 130 km off Nantucket June— larvae, off Woods Hole Dec. -Feb. —eggs Dec. -Mar. —larvae, off North Carolina Nov. -Apr. —larvae Dec— eggs, patch of several thousands Oct. 1900— young fry, Casco Bay Oct. 1915— eggs and larvae, Nantucket Sound Aug.— eggs, off Gay Head July— larvae, Woods Hole Harbor Jan. -May— large larvae May— larvae Apr.— larvae May-Oct.— eggs May-Sept.— larvae Fall-larvae (see Wheatland 1956) June-Oct.— eggs June and Sept. -Dec. —larvae May-Oct.— eggs June-July and Sept. -Dec. —larvae Winter and spring— larvae, common May-Aug. and Oct.— eggs June, July and Oct. -Feb. (most in Oct.)- larvae May-June— eggs Nov. -Dec. —larvae Spring and summer near Solomons, Md. eggs Mar.-June and Nov.— larvae, upper Chesapeake Bay Inlets Reintjes and Pacheco 1966 1955-61 de Sylva et al. 1962 1956-58 Tagatz and Dudley 1961 1957-60 Lunz 1965 1964 Lewis and Mann 1971 1966-68 Smaller estuaries Warfel and Merriman 1944 1942-43 Massmann et al. 1954 1950-52 Pacheco and Grant 1965 ■> Reintjes and Pacheco 1966 > 19=5-61 Tagatz and Dudley 1961 1957-60 Pearcy and Richards 1962 1959-60 Dovel 1967 1965-66 Wilkens and Lewis 1971 1967-69 Sept. -June Indian River Inlet, Del. all all Jan. -Mar. Nov. -Apr. all Mar.-Oct. Sept. -June all all all all Indian River Inlet, Del.' Beaufort Inlet, N.C. Several South Carolina inlets Bogue and Beaufort inlets, N.C. Nov. -Apr.— larvae Oct. -May— larvae, month of peak occurrence varied with year from Dec.-Feb. Nov.-May— larvae Jan. and Apr.-May— larvae Feb. -Mar.— larvae Morris Cove, Conn. Five Virginia tidal rivers White Creek, Indian River, Del. Neuse River, N.C. Mystic River, Conn. Magothy River, Md. White Oak River, N.C. July-Nov.— larvae Apr.-May— larvae, brackish water Nov. -June, most Feb. -June— larvae Jan. -June and Sept.— larvae June-July— larvae Mar.— larvae Jan. -Apr.— larvae 'Larvae emigrating from ocean. juveniles overwintering in southern estuaries. Fish range from 50 to 160 mm w^hen they leave the estuaries (June and Chamberlin 1959). Recently Mansueti and Hardy (1967) amplified the descriptions of young menhaden and Reintjes (1969) reviewed the biology of the species. This paper describes the occurrences of menhaden eggs and larvae in the Dolphin surveys and relates these findings to daylight, tempera- ture, salinity, and depth, and to our understanding of menhaden early life history. PROCEDURES The eight ichthyoplankton cruises were made at approximately 6-wk intervals from December 1965 to December 1966 (Clark et al. 1969). On each cruise 92 stations were located along 14 transects from Martha's Vineyard to Cape Lookout (Figure 1). Stations were closely spaced inshore (9.3 km) along the transects and farther apart (27.8 km) offshore. The order that the transects were run and the time during the 24-h workday that stations were oc- cupied varied among the cruises. 318 KENDALL and REINTJES: DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE Figure l.-RV Dolphin survey, 1965-66. Locations of transects and collecting stations. Gulf V plankton samplers with 0.4-m openings and 0.52-mm wire mesh were towed for 30 min at 5 knots (2.6 m/s) in step-oblique tows at each sta- tion. Two nets were fished simultaneously at six 3-m depth intervals, from separate warps, one shallow (0-15 m) and one deep (18-33 m), where water depth permitted. In shallower water fewer depth intervals were sampled for longer time periods. Plankton samples were preserved in 4% buffered formaldehyde solution and brought to shore for sorting. Supplementary data collected at each station included water temperature and salinity with depth. A scaled-down Cobb mid- water trawl was towed at about half of the sta- tions to collect juvenile fishes. Fish eggs and larvae were separated from the plankton in the laboratory. Clupeoid larvae were distinguished by their slender body, long gut, sparse pigment, and short dorsal and anal fin bases (Table 2), and separated from other larvae. Several other fishes whose long slender larvae resemble clupeoids occur in the area but they differ in certain features. Lizardfish larvae have a large finfold, an adipose fin later in development, and a row of 6 to 12 paired dark patches ventrally along the body (Anderson et al. 1966). Other elon- gate salmoniform larvae generally have an adipose fin, photophores often, and oval or stalked eyes. Sand lance larvae have little pigment, but the dorsal and anal fins extend most of the length of the body (Norcross et al. 1961). Blennioid larvae have a short gut, with the anus forward of midlength of the body, and long dorsal and anal fin bases. Other larvae are excluded by myomere counts which fall outside the range for clupeoids (38-55) or by other distinctive characters not found on clupeoid larvae. Table 2.-Distinguishing features of Atlantic coast clupeoid larvae. Character Distinguishing features Body shape Fin positions Meristic counts Pigmentation General Body slender and elongate, the greatest body depth less than 20% of total length. Anus in posterior third of body. Dorsal — single, short, about two-thirds way back along body; no adipose fin. Anal — posterior to at least part of dorsal fin, not confluent with caudal fin. Pelvics — abdominal, at about midlength on the body. Myomeres (vertebrae) — 38-55. Dorsal fin rays — 9-22. Anal fin rays - 10-30. Principal caudal fin rays — 10 -|- 9. Little pigment except ventrally. Ventral pigment — small spots on throat, along the gut, and at base of caudal fin. Eyes round, in orbits, not stalked. Gut straight with annular folding of the intestine. 319 Identification of menhaden eggs and larvae among tiie four other clupeid genera and two engrauHd genera along the North American east coast was facilitated by published illustrations and descriptions (Kuntz and Radcliffe 1917; Mansueti and Hardy 1967; Houde and Fore 1973) and reported spawning areas and times (Reintjes 1961; Higham and Nicholson 1964). Menhaden larvae collected north of Cape Lookout are presumed to be Brevoortia tyrannus, since B. smithi occurs mainly farther south and spawns inshore (Reintjes 1962). The areas of larval occurrence of menhaden overlap Atlantic herring, Clupea harengus harengus, in the north, and round her- ring, Etrumeus teres; Spanish sardine, Sardinella anchovia; and Atlantic thread herring, Opisthonema oglinum, in the south. Clupeoid larvae less than 8 mm are difficult to distinguish because the median fins and other characters are not formed. However, pigmenta- tion, body shape, and gut length are helpful in small larvae. Among clupeoid larvae the stage of development at a particular size and area of cap- ture are helpful. For each comparable stage of larval development, menhaden are larger than anchovies, Spanish sardine, and Atlantic thread herring, and smaller than Atlantic herring. Only round herring show the same development with size as menhaden but are easily distinguished by the relative length of the snout at all sizes. Menhaden were counted and total length measurements of larvae less than 12 mm long were made to the nearest 0.1 mm with an ocular micrometer in a dissecting microscope; those longer than 12 mm were measured with dividers and a steel rule or dial calipers to the nearest 0.5 mm. Samples containing more than 50 fish were usually randomly subsampled before measuring. To get a subsample of 25-50 larvae, the sample was floated in formaldehyde solution in a 150-mm petri dish scribed into quarters with two diameters; a random half or quarter of the dish was chosen for measuring. The process was repeated with the chosen fraction with samples of more than 200 specimens. When larvae showed slight damage, such as broken caudal rays, measurements were adjusted to approximate total length. Larvae identifiable, but too mutilated to be measured, were counted. For geographic distribution analysis, numbers of larvae at each station were adjusted to a standard tow as in Smith (1973) and Fahay (1974). FISHERY BULLETIN: VOL. 73, NO. 2 RESULTS Geographic Distribution of Larvae Menhaden larvae are more widely distributed than any other clupeoid in the northwest Atlantic. Larvae have been reported from Maine to Mexico in oceanic, estuarine, and fresh waters. South of Cape Hatteras Atlantic menhaden spawn in the cooler months (October to June), while along the Middle Atlantic states they spawn during the warmer months (June to November). Seasonal oc- currences of larvae followed the annual north- south migration of adults. Small larvae were first collected in June near Delaware Bay. By October larvae were present from southern New England to Chesapeake Bay. In late fall they were found from New York to Cape Lookout. During the winter and spring larger larvae were present from Chesapeake Bay to Cape Lookout. Small menhaden larvae were collected throughout the year except in late spring, in- dicating nearly continual spawning (Figure 2). Menhaden hatch at about 3 mm (Mansueti and Hardy 1967), but we caught few less than 5 mm. This was probably due in part to our inability to identify with certainty small larvae and in part to their fragile, slender form which caused them to be extruded from our nets. The largest larvae we collected were 30 mm, about the size at which they enter estuaries and transform into juveniles (Lewis et al. 1972). It is not reasonable to attempt to determine a growth rate for menhaden larvae from our data. Small menhaden less than 8 mm were collected from June through February. Thus, spawning took place in our sampling area for 6 mo of the year. This protracted spawning season and probable geographic movements of larvae preclude the possibility of determining growth rate from this survey. Data associated with collections of menhaden larvae are presented in the Appendix Table. Their occurrences are illustrated in Figures 4-11. From the irregular horizontal and vertical distributions and highly nonnormal catch frequency curve (Figure 3), it appears that the larvae are very unevenly and probably patchily distributed. The location and density of patches may be related to local currents and water conditions on a smaller scale than we sampled. This discussion follows the sequence of spawning, i.e. starting in late spring 320 KENDALL and REINTJES: DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE 2000 1800 16O0 1400- 1200 1000 U1 QC 800 m S. D Z 600 400 200 45 40 35 30 25 20 15 10 5 JUNE, 1966 N= 14 F j p q -0- AUG., 1966 N = 8 OCT., 1966 N=5420 t- ~. NOV., 1966 N = 308 n [u 10 T r 15 20 25 30 160 140 120 100 80 60 40 20 70 60 50 40 30 20 10 7 6 5 4 3 2 1 4 3 2 1 DEC, 1965 N = 914 JAN.-FEB., 1966 N-305 APR., 1966 N = 34 Jl MAY, 1966 N=3 ~r 5 10 LENGTHS OF lARVAE ( m m,T L) IS 20 25 30 Figure 2.— Length-frequencies of menhaden larvae from the RV Dolphin survey, 1965-66. 1966. The cruises, however, began in December 1965. Thus, the cruises from December 1965 to May 1966 cover one spav^^ning season and those from June 1966 to November-December 1966 cover the foUovi^ing season. 50-1 40- < 30- O 5 20- Z lo- in late June a few larvae were collected nearshore off Delaware Bay (Figure 4). These larvae were small (7.8-14.4 mm), indicating they had been spawned recently. Spawning may have occurred in Delaware Bay since the larvae we llllliilllill til ii^iilniii I 4" LliJ_ _UJI 5 10 50 100 NUMBERS OF LARVAE PER HAUL (LOG SCALE] 500 —1 1000 5000 Figure 3.-Frequency distribution of menhaden larval catches per haul. 321 FISHERY BULLETIN: VOL. 73, NO. 2 caught were concentrated close to its mouth. The distribution of the collections indicates spawning was not widespread. Stations north of these collections were sampled about 10 days earlier, possibly accounting for the limited distribution we observed. The absence of larvae to the north in these earlier collections would indicate spawning had recently started. Temperatures in June in the area of capture ranged from 15° to 19°C, and salinity varied from 30.3 to 32. IVoo (Figure 4). Hydrographic conditions within these ranges oc- curred widely along the coast during this cruise (Clark et al. 1969). During our cruise in August there was evidence of limited spawning nearshore (Figure 5). A few small larvae, 5.6-10.5 mm, were taken off Long Island and New Jersey. Perhaps the earlier spawning in June was so limited that with disper- sal during growth, insufficient larvae survived to be taken in our August sampling. Temperatures in the area of capture in August were warmer than in June, 18° to 22°C at the surface, and the seasonal thermocline was well developed about 15 m below the surface (Figure 5). Temperatures below the thermocline were less than 10°C. Larvae were collected only in the shallow net, indicating they were in the warm water above the thermocline. Salinity in the area of capture ranged from 30.1 to 31.0''/oo. In October 5,420 larvae were collected that ranged from 3.5 to 18.5 mm (Figure 2). The length-frequency distribution was skewed to the left (mean 7.3 mm; mode 6.5 mm). Larvae were more widespread and in greater concentrations than during any other cruise (Figure 6). They oc- curred from Martha's Vineyard to Currituck Beach, N.C., and were abundant from Long Island to Maryland, near the middle of the shelf. They occurred nearshore from southern New England to New Jersey and near Chesapeake Bay. The fish Figure 4.-Distribution and abundance of menhaden larvae in the June cruise. Figure 5.-Distribution and abundance of menhaden larvae in the August cruise. 322 KENDALL and REINTJES: DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE ATLANTIC MENHADEN LARVAE CRUISE D-«6-12 SEPT. 28 - OCT. 20, 1956 were 8-12 mm in the northern part of the sampling range, from New York north. Fish in a broader size range, 4-12 mm, were present off New Jersey. Nearly all fish south of New Jersey were smaller, 4-8 mm. In October the thermocline was breaking down but still present over much of the area, and surface temperatures were about 3°C cooler than in August (Figure 6). Salinity values were about the same as in August, mostly between 30.5 and 32.0''/oo, except near the mouth of Chesapeake Bay where they dropped to 28.0''/<». The cruise in late fall 1966 (Figure 7) shows a distribution pattern quite similar to the cruise in December 1965 (Figure 8). During both cruises, larvae occurred mostly nearshore from Long Island to North Carolina. They were most abun- dant near Cape Hatteras, where they were taken to the edge of the shelf. In November 1966 there was a bimodal size dis- tribution with one peak at 8 mm and the other at 20 mm (Figure 2). There were few larvae between 14 and 18 mm. In December 1965 the peak at 8 mm is similar to that in late fall 1966, but the second peak at 20 mm is not seen (Figure 2). Possibly this difference is due to year-to-year variation in spawning pattern in the area studied. In the November and December cruises, small fish were found south of Delaware Bay and tended to occur at least 12 km offshore. Larger fish oc- curred mainly north of Chesapeake Bay and mostly within 15 km of shore. In transition areas between north and south and inshore and offshore areas, fish in a wide length range occurred at the same station and bimodal length-frequency curves were seen. This may indicate that spawning occurs in waves, and as the larvae grow they disperse from the area where they were spawned. By late fall the thermocline was gone and sur- face isotherms roughly paralleled the coastline (Figures 7, 8). Larvae were taken over a wide range of temperature, from 7° to 25°C. Most collections were in water between 10° and 20°C. Salinity varied considerably between the two late fall cruises. In 1965, several patches of low saline water, less than 30''/oo, were found mostly near the shore (Clark et al. 1969). However, in 1966 salinity throughout the area was greater than3lVoo,except immediately outside Chesapeake Bay. The dis- tribution of larvae is quite similar between these Figure 6.-Distribution and abundance of menhaden larvae and distribution of menhaden eggs in the October cruise. 323 FISHERY BULLETIN: VOL. 73, NO. 2 two cruises in spite of these differences in salinity. In February menhaden larv^ae were taken from Virginia to Cape Lookout (Figure 9). They prob- ably occurred farther south than we sampled since they were most abundant on our southernmost transect. Most larvae were taken between Cape Hatteras and Cape Lookout. Lengths presented a fairly symmetrical distribution with a peak at 13 mm (Figure 2). There is indication again that the larger fish were taken nearer to shore and farther north than the smaller fish. Few fish shorter than 8 mm were seen at this time, and the maximum length was 29 mm. Apparently, at about this size menhaden have either entered estuaries or can avoid our nets. In winter the temperature was 4°C at 11 of the 23 stations where menhaden were taken; it was less than 3°C at 4 stations (Figure 9). Water at these stations was practically isothermal with depth (Clark et al. 1969). During April larvae occurred in approximately the same areas as in February (Figure 10). Fewer, larger larvae were taken between Chesapeake Bay and Ocracoke Inlet, N.C., than earlier. Larvae north of Cape Hatteras ranged from 21 to 29 mm. Off Ocracoke Inlet larvae were 11-19 mm. The larger larvae north of Cape Hatteras were taken nearshore; those farther south extended 37 km offshore. A bimodal length-frequency curve had peaks centered at 15 and 24 mm, although in- sufficient numbers of fish were collected to deter- mine the statistical significance (Figure 2). By April, temperatures in the areas of capture were warmer than winter temperatures, above 8°C, and mostly between 10° and 12° (Figure 10). Salinity ranged from 30.3 to 35.6"/ «,. In May a few large (25-26 mm) larvae were taken off Virginia, inshore near Chesapeake Bay (Figure 11). These were apparently remnants of the early winter spawning, the rest of the larvae having already entered estuaries. Water in the areas of capture had warmed to 13° to 17°C (Figure 11). Salinity was generally lower, from 28.4 to 31.2'Vm, due to spring freshening nearshore (Clark et al. 1969). Temperature-Salinity Relations The catches of menhaden larvae during all cruises in the shallow net were compared Figure 7.-Distribution and abundance of menhaden larvae and distribution of menhaden eggs in the November-December 1966 cruise. ATLANTIC MENHADEN LARVAE CRUISE D-64-14 NOV. 9 -DEC. 4, 1966 324 KENDALL and REINTJES: DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE 73* ATLANTIC MENHADEN LARVAE CRUISE D-65-4 DEC. 3-15, 1965 34 76* graphically with observed surface temperatures and salinities. Surface observations were thought adequate since most larvae were collected when hydrographic conditions were nearly uniform with depth and menhaden larvae were scarce below the thermocline. Mean temperatures and salinities within the sampling depth range were also com- pared with catches and showed patterns similar to those discussed here. Observed surface temperatures varied from -1° to 28°C (Figure 12). At each of nine whole-degree intervals between 6° and 19°C, more than 30 sta- tions were occupied. More than 40 stations were occupied at 10° and 14°C. Menhaden occurred at stations when temperatures were between 0° and 25° C. The curve of positive stations (those where menhaden were taken) was similar in shape and range to that of total stations. The numbers of larvae taken at each temperature were plotted on a log scale. This plot was slightly skewed to the right, with modal catch at 18°C. Catches of over 100 larvae were made at temperatures from 9.3° to 20.5°C, with most between 15.8° and 18.5° C. Surface salinity varied from 23 to 38''/oo, with a mode at 31o/oo (Figure 13). Positive stations oc- curred over the entire range of salinities, with a mode at 30"/oo. The larval catch curve, on a log scale, is similar in shape to the total station curve, with a mode at 31"/oo. At stations with salinities between 30 and 36«/oo, a total of at least 200 larvae were taken within each part-per- thousand inter- val. Diel- Vertical Comparisons of Larval Catches Comparisons were made of the catches of larvae in shallow and deep tows made during night and day. These comparisons are on the basis of the volume of water sampled, which was assumed to be constant among the tows. The use of parametric statistics was precluded by the highly nonnormal catch frequency curve (Figure 3). Of the 172 tows with menhaden larvae, 48 contained only 1 larva, and 11 tows contained more than 100 larvae. Of the 11 tows with more than 100 larvae, 7 were taken in shallow tows during daylight. Altogether menhaden occurred in 85 daylight tows and 87 nighttime tows (Table 3). The distribution of catches was not significantly different with time of day (chi-square test; P > 0.50). Day and night Figure 8.-Distribution and abundance of menhaden larvae in the December 1965 cruise. 325 FISHERY BULLETIN: VOL. 73, NO. 2 Figure 9.-Distribution and abundance of menhaden larvae in the February cruise. Figure 11.— Distribution and abundance of menhaden larvae in the May cruise. Figure 10. -Distribution and abundance of menhaden larvae in the April cruise. 50 40- "- 30 s Z 20 ! I ■ n II 1 1 I » LARVAE POSITIVE STATION TOTAL STATION .-'• .N Mh\ •fl K I 1 1 i i U A j:^ 10,000 1000 100 10 z c 3 % o o > 10 15 TEMPERATURE 20 25 30 Figure 12.-Relation between surface temperature to menhaden larval catch and sampling effort. tows were combined to compare the distribution of catches by the shallow and deep nets. By the shallow net, menhaden were taken in 138 tows and by the deep net in 34 tows (Table 3). The distribu- tion of catches was not significantly different in the two nets {P >0.50). Comparisons of size of larvae between day and night tows with the shallow and deep nets showed no significant differences. 326 KENDALL and REINTJES: DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE Eggs We found menhaden eggs at only six stations, five from the October cruise and one from the late fall cruise in 1966 (Figures 6, 7). All but one sample contained less than 100 developing eggs. The ex- ceptional sample was from 85 km off Delaware Bay in October where about 2,000 eggs were taken. Precise counts were not possible due to the poor state of preservation of the samples when they were examined. Eggs were collected in areas where small larvae were taken. Apparently menhaden spawn as large schools producing dense patches of eggs (Reintjes 1969). During the short incubation time (48 h), these patches do not become dispersed. Thus the distribution of menhaden eggs at sea is probably more uneven than that of larvae. Chance was a dominant factor in catching menhaden eggs so distributed in our survey. Mid-Water Trawl Catches Sampling by mid- water trawl during the cruises collected a few larval and adult menhaden. Sampling effort was good in areas where the catches were made, so they probably reflect the actual geographic distribution of menhaden sub- ject to capture by this type of sampling (Clark et al. 1969). A few large larvae, 21-37 mm, were taken in August off the Chesapeake Bay area. Age-0 fish, 89-177 mm FL (fork length) (Reintjes 1969), oc- curred close to shore from southern New England to Chesapeake Bay in late fall 1966. These proba- bly represent young fish migrating south after spending the summer in estuaries (June and Chamberlin 1959). Other catches included two large fish, 305 and 361 mm FL, close to shore off southern New Jersey in May, and several age-0 fish off Oregon Inlet, N.C., in June. DISCUSSION Much speculation has surrounded the distribu- LARVAE 200 — TOTAL STATION / — 5 150 / O , h~ 1 < h- 1 ift 1 u. 100 f o 1 a: 1 ta f 1 1 1 I "^ s s 3 \ 2 iO . 1 1 ♦ \ / "••*i^ |V y / XI V 0 Kl .^ i-.4:^ . 1 1 1 1 1 t *f - 10,000 1000 -100 -1 10 -1 s n o o n > 25 30 SALINITY(S^o) 35 Figure 13.-Relation between surface salinity to menhaden lar- val catch and sampling effort. tion of early stages of menhaden. Spawning times and places have been inferred from examination of gonads of adults (McHugh et al. 1959; Higham and Nicholson 1964) and nearshore and estuarine samples of larvae and juveniles (e.g. Richards 1959; Sutherland 1963; Pacheco and Grant 1965). Few studies have actually taken menhaden eggs and larvae to determine more directly the area of spawning (Reintjes 1961; Massmann et al. 1962). Controversy has concerned whether menhaden spawn in Chesapeake Bay (Hildebrand and Schroeder 1928) and whether there are two separate populations along the east coast, one spawning in spring and one in fall (Nicholson 1972). Annual variation in time of spawning and entry of larvae into estuaries may account for some of the confusion, since most studies have been short-termed and in a relatively small por- tion of the range of menhaden. Caution needs to be exercised in analyzing the present data since they were collected during a single year and do not encompass the entire range of spawning of menhaden (Reintjes 1969). During summer larvae were taken from our inshore sta- tions to our farthest offshore station and at our most northerly station. The possibility of spawn- ing within estuaries is indicated by the presence of Table 3.-Diel and depth distribution of menhaden larval catches and mean lengths. Day Night Day and Night Shallow Deep Larvae/tow Shallow Deep Both Shallow Deep Both All 3er of to ws - 1-2 29 9 38 34 2 36 63 11 74 3-4 7 3 10 7 4 11 14 7 21 5-20 18 3 21 18 6 24 36 9 45 >21 15 1 16 10 6 16 25 7 32 Total tows 69 16 85 69 18 87 138 34 172 Mean larvae /tow 68.2 6.4 56.0 19.1 48.2 25.6 43.0 27.2 39.8 Mean larval length (r nm) 8.3 8.8 8.3 9.0 6.9 8.2 8.5 7.1 8.3 327 FISHERY BULLETIN: VOL. 73, NO. 2 small larv^ae near their mouths. In winter they were found at our southernmost stations. Therefore, spawning could have taken place inshore and offshore of our stations and farther north and south of our sampling. Our results on area of spawning confirm the conclusions of Higham and Nicholson (1964) and Nicholson (1972) that menhaden spawn during both their northward spring migration and their southward fall migration. We conclude that spawning apparently continues in winter in the south, based on our catches around Cape Lookout. Midsummer spawning may have occurred north of our sampling area, or may have been inhibited in 1966 by water cooler than usual. Harrison et al. (1967) postulated that bottom drift of waters off Chesapeake Bay influence the success of year classes of menhaden. During years when bottom drift was weak and southwesterly, poor year classes occurred. However, our data and that of Massmann et al. (1962) do not indicate a preference for bottom waters by larger menhaden larvae. Larvae entering the estuaries are found throughout the water column (Lewis and Mann 1971), and later, in the estuaries, they are found primarily in surface waters (Massmann et al. 1954). Possibly the factors affecting bottom drift also affect the success of year classes, but in an indirect way. Reintjes and Pacheco (1966) reported on inlet and nursery area collections in Indian River, Del., made over a 6-yr period (1955-61). Among the years studied, the peak in larval abundance at the inlet occurred in all months from December through March. Larvae were taken at the inlet from September through June in most years. It appeared that when temperatures at the inlet dropped to 3°C, larvae in the area were killed. In four seasons, when large catches of larvae were made at the inlet between December and February and temperature later dropped below 3°C, larvae were scarce or absent in upstream nursery areas. However, larvae taken in years when temperature remained above 3°C and those taken after the critical low temperature period were later represented in collections upstream. Due to the extreme year-to-year variability in time of entry at the inlet at Indian River, it is difficult to relate our catches to these data. The few small larvae taken near Delaware Bay in June would probably have entered the estuary in July, a month when Reintjes and Pacheco (1966) reported none. We made large catches in this area in Oc- tober. Larvae were also present offshore in November and December. In February, larvae were not taken north of Virginia and water temperature close to shore between Chesapeake and Delaware bays was less than 0°C. From these data, it would appear that 1965-66 was dissimilar to any year studied by Reintjes and Pacheco (1966) with regard to menhaden larvae near Delaware Bay. Presumably during 1965-66 larvae could have been taken in abundance in late fall but would have been killed by cold water in February. Sub- sequently no larvae would have entered in winter or spring, but some might have appeared in early summer. Menhaden larvae were taken off Chesapeake Bay in waters 1° to 15°C by Massmann et al. (1962). Herman (1963) reported them in Narragansett Bay from 1° to 22°C. The larvae we collected at temperatures below 3°C did not ap- pear decomposed as would be expected had they been dead when captured. Lewis (1965, 1966) has studied the effects of subjecting menhaden larvae to low temperatures and a range of salinities in the laboratory. He has shown that moderate salinities (10-20''/i)o) enhance survival at low temperature as do lowered acclimation tempera- tures. At 7°C acclimation temperature, the lowest he used, and 2° to 4°C test temperatures, 50% mortality occurred in about 40 h at 24"/oo. Salinities in our areas of capture were 31 to SS^/w-higher than those tested by Lewis (1966). If temperatures below 3°C in estuarine waters kill many menhaden (Reintjes and Pacheco 1966), it may be advantageous for larvae to remain in the ocean where temperature changes are more gradual. Menhaden entering estuaries are usually larvae less than 30 mm long, and in the estuary they rapidly transform by 38 mm (Lewis et al. 1972). Reintjes and Pacheco (1966) reported a few trans- forming specimens (greater than 34 mm) entering Indian River in May. That they are dependent on estuarine conditions for transformation is sup- ported by the absence of transforming specimens and juveniles in our collections. SUMMARY AND CONCLUSIONS A total of 7,006 menhaden larvae were collected in 172, 0.5-h Gulf V plankton tows. The larvae were taken on eight cruises along the middle Atlantic coast throughout the year. Eggs were taken in a few tows in the fall. The catch distribution was nonnormal, with 48 tows catching only 1 fish and 1 328 KENDALL and REINTJES: DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE tow catching 2,553 fish. No differences in larval catches or size v^ere found in shallow and deep tows or during day or night. Small larvae, less than 8 mm, were taken from June through February, indicating a protracted spawning period. However, there was a seasonal shift in area of spawning. In late spring and summer limited spawning was occurring off New Jersey and New York. By early fall spawning was widespread from southern New England to Virginia. By late fall and early winter spawning was limited to areas between Delaware and North Carolina. Larger larvae were taken in the north in late fall and south of Delaware Bay in winter and spring. Larvae occurred over a wide range of tempera- ture, from 0° to 25°C. Several were taken in waters cooler than the 3°C limit found lethal in laboratory tests and inferred to be limiting in inlet sampling studies. Most larvae were collected at temperatures between 15° and 20°C. There was an inverse relationship between temperature and size of larvae. Salinity seemed to have little influence on the distribution of larvae. They occurred at practically every salinity encountered and the frequency of salinities closely resembled the frequency of posi- tive tows. Our findings are similar to recent investigations of early life history of Atlantic menhaden based on inlet sampling of larvae, gonad studies, and scale annulus formation. It appears that spawning and early development at sea take place over a long period in a given coastal area, and the larvae resulting from this spawning may reach the inlets over a long seasonal time. In years with mild winters, successful immigration to estuarine areas may occur before the winter temperature minimum. However, under more severe condi- tions, when winter temperatures in the estuaries fall below 3°C, successful immigration may occur only as temperatures are increasing in spring because larvae entering estuaries during the fall may not survive the winter. This could account for some of the annual variation in year-class strength of Atlantic menhaden. LITERATURE CITED Anderson, W. W., J. W. Gehringer, and F. H. Berry. 1966. Field guide to the Synodontidae (lizardfishes) of the western Atlantic Ocean. U.S. Fish Wildl. Serv., Circ. 245, 12 p. Baird, S. F. 1873. Testimony in regard to the present condition of the fisheries, taken in 1871. In Report on the condition of the sea fisheries of the south coast of New England in 1871 and 1872, p. 7-72. Rep. U.S. Comm. Fish Fish. 1871-1872, Part I. BiGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Clark, J. R., W. G. Smith, A. W. Kendall, Jr., and M. P.. Fahay. 1969. Studies of estuarine dependence of Atlantic coastal fishes. Data report I: Northern section. Cape Cod to Cape Lookout. R. V. Dolphin cruises 1965-66: Zooplankton volumes, midwater trawl collections, temperatures and salinities. U.S. Bur. Sport Fish. Wildl., Tech. Pap. 28, 132 p. Croker, R. a. 1965. Planktonic fish eggs and larvae of Sandy Hook es- tuary. Chesapeake Sci. 6:92-95. de Sylva, D. p., F. a. Kalber, and C. N. Shuster, Jr. 1962. Fishes and ecological conditions in the shore zone of the Delaware River estuary, with notes on other species collected in deeper waters. Univ. Del. Mar. Lab., Inf. Ser. Publ. 5, 164 p. Deubler, E. E., Jr. 1958. A comparative study of the postlarvae of three flounders (Paralichthys) in North Carolina. Copeia 1958:112-116. Dovel, W. L. 1967. Fish eggs and larvae of the Magothy River, Maryland. Chesapeake Sci. 8:125-129. 1971. Fish eggs and larvae of the upper Chesapeake Bay. Nat. Res. Inst, Univ. Md., Spec. Rep. 4, 71 p. Fahay, M. P. 1974. Occurrence of silver hake, Merluccius bilinearis, eggs and larvae along the middle Atlantic continental shelf during 1966. Fish. Bull., U.S. 72:813-834. GooDE,G. B. 1879. The natural and economical history of the American menhaden. Rep. U.S. Comm. Fish Fish. 1877, Append. A, Part 1, p. 1-529. Harrison, W., J. J. Norcross, N. A. Pore, and E. M. Stanley. 1967. Circulation of shelf waters off Chesapeake Bight. Surface and bottom drift of Continental Shelf waters between Cape Henlopen, Delaware, and Cape Hatteras, North Carolina June 1963-December 1964. U.S. Dep. Commer., ESSA Prof. Pap. 3, 82 p. Herman, S. S. 1963. Planktonic fish eggs and larvae of Narragansett Bay. Limnol. Oceanogr. 8:103-109. HiGHAM, J. R., AND W. R. NICHOLSON. 1964. Sexual maturation and spawning of Atlantic menhaden. U.S. Fish WMl Serv., Fish. Bull. 63:255-271. HiLDEBRAND, S. F., AND W. C. SCHROEDER. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43(l):l-366. HouDE, E. D., AND P. L. Fore. 1973. Guide to identity of eggs and larvae of some Gulf of Mexico clupeid fishes. Fla. Dep. Nat. Resour., Mar. Res. Lab., Leafl. Ser. 4, Part 1(23), 14 p. June, F. C, and J. L. Chamberlin. 1959. The role of the estuary in the life history and biology of Atlantic menhaden. Proc. Gulf Caribb. Fish. Inst., nth Annu. Sess., p. 41-45. KuNTZ, A., and L. Radcliffe. 1917. Notes on the embryology and larval development of 329 FISHERY BULLETIN: VOL. 73, NO. 2 twelve teleostean fishes. Bull. U.S. Bur. Fish. 35:87-134. Lewis, R. M. 1965. The effect of minimum temperature on the survival of lan'al Atlantic menhaden, Brevoortia tyrannwi. Trans. Am. Fish. See. 94:409-412. 1966. Effects of salinity and temperature on survival and development of larval Atlantic menhaden, Brevoortia tyrannus. Trans. Am. Fish. See. 95:423-426. Lewis, R. M., and W. C. Mann. 1971. Occurrence and abundance of larval Atlantic menhaden, Brevoortia tyrannus, at two North Carolina inlets with notes on associated species. Trans. Am. Fish. Soc. 100:296-301. Lewis, R. M., E. P. H. Wilkens, and H. R. Gordy. 1972. A description of young Atlantic menhaden, Brevoortia tyrannus, in the White Oak River estuary. North Carolina. Fish. Bull, U.S. 70:115-118. LUNZ, R. 1965. Annual report 1963-1964. Bears Bluff Lab., Contrib. 41, 10 p. McHuGH, J. L., R. T. Oglesby, and A. L. Pacheco. 1959. Length, weight, and age composition of the menhaden catch in Virginia waters. Limnol. Oceanogr. 4:145-162. Mansueti, a. J., AND J. D. Hardy, Jr. 1967. Development of fishes of the Chesapeake Bay region. An atlas of egg, larval, and juvenile stages. Part L Nat. Res. Inst.. Univ. Md., 202 p. Marak, R. R., and J. B. CoLTON, Jr. 1961. Distribution of fish eggs and larvae, temperature, and salinity in the Georges Bank-Gulf of Maine area, 1953. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 398, 61 P- Marak, R. R., J. B. Colton, Jr., and D. B. Foster. 1962. Distribution of fish eggs and larvae, temperature, and salinity in the Georges Bank-Gulf of Maine area, 1955. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 411, 66 P- Marak, R. R., J. B. Colton, Jr., D. B. Foster, and D. Miller. 1962. Distribution of fish eggs and larvae, temperature, and salinity in the Georges Bank-Gulf of Maine area, 1956. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 412, 95 P- Massmann, W. H., E. C. Ladd, and H. N. McCutcheon. 1954. Postlarvae and young of menhaden {Brevoortia tyrannus) in brackish and fresh waters of Vir- ginia. Copeia 1954:19-23. Massman, W. H., J. J. NoRCROss, and E. B. Joseph. 1962. Atlantic menhaden larvae in Virginia coastal wa- ters. Chesapeake Sci. 3:42-45. Merriman, D., and R. C. Sclar. 1952. The pelagic fish eggs and larvae of Block Island Sound. In Hydrographic and biological studies of Block Island Sound, p. 165-219. Bull. Bingham Oceanogr. Collect., Yale Univ. 13(3). Nicholson, W. R. 1972. Population structure and movements of Atlantic menhaden, Brevoortia tyrannus, as inferred from back calculated length frequencies. Chesapeake Sci. 13:161-174. Norcross, J. J., W. H. Massmann, and E. B. Joseph. 1%1. Investigations of inner continental shelf waters off lower Chesapeake Bay. Part II. Sand lance larvae, Am- modytes americanus. Chesapeake Sci. 2:49-59. Pacheco, A. L., and G. C. Grant. 1965. Studies of the early life history of Atlantic menhaden in estuarine nurseries. Part I. -Seasonal occurrence of juvenile menhaden and other small fishes in a tributary creek of Indian River, Delaware, 1957-58. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 504, 32 p. Pearcy, W. G., and S. W. Richards. 1962. Distribution and ecology of fishes of the Mystic River estuary, Connecticut. Ecology 43:248-259. Pearson, J. C. 1941. The young of some marine fishes taken in lower Chesapeake Bay, Virginia, with special reference to the grey sea trout Cynoscion regalis (Bloch). U.S. Fish Wildl. Serv., Fish. Bull. 50:79-102. Perlmutter, a. 1939. Section I. An ecological survey of young fish and eggs identified from tow-net collections. In A biological sur- vey of the salt waters of Long Island, 1938, Part II, p. 11-71. N.Y. State Conserv. Dep., Suppl. 28th Anna. Rep., 1938, Salt-water Surv. 15. Reintjes, J. W. 1961. Menhaden eggs and larvae from M/V Theodore N. Gill cruises, South Atlantic coast of the United States, 1953- 54. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 393, 7 p. 1962. Development of eggs and yolk-sac larvae of yellowfin menhaden. U.S. Fish Wildl. Serv., Fish. Bull. 62:93-102. 1969. Synopsis of biological data on the Atlantic menhaden, Brevoortia tyrannus. U.S. Fish Wildl. Serv., Circ. 320, 30 P- Reintjes, J. W., and A. L. Pacheco. 1966. The relation of menhaden to estuaries. In R. F. Smith, A. H. Swartz, and W. H. Massmann (editors), A sym- posium on estuarine fisheries, p. 50-58. Am. Fish. Soc, Spec. Publ. 3. Richards, S. W. 1959. Pelagic fish eggs and larvae of Long Island Sound. In Oceanography of Long Island Sound, p. 95-124. Bull. Bingham Oceanogr. Collect., Yale Univ. 17(1). Smith, W. G. 1973. The distribution of summer flounder, Paralichihys dentatus, eggs and larvae on the continental shelf between Cape Cod and Cape Lookout, 1965-66. Fish. Bull., U.S. 71:527-548. Sutherland, D. F. ■ 1963. Variation in vertebral numbers of juvenile Atlantic menhaden. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 435, 21 p. Tagatz, M. E., and D. L. Dudley. 1961. Occurrence of marine fishes at four shore habitats near Beaufort, N.C., 1957-60. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 390, 19 p. Warfel, H. E., and D. Merriman. 1944. Studies on the marine resources of southern New England. I. An analysis of the fish population of the shore zone. Bull. Bingham Oceanogr. Collect., Yale Univ. 9(2), 91 p. Wheatland, S. B. 1956. Pelagic fish eggs and larvae. In Oceanography of Long Island Sound, 1952-1954, p. 234-314. Bull. Bingham Oceanogr. Collect., Yale Univ. 15. Wilkens, E. P. H., and R. M. Lewis. 1971. Abundance and distribution of young Atlantic menhaden, Brevoortia tyrannus, in the White Oak River estuary. North Carolina. Fish. Bull., U.S. 69:783-789. 330 KENDALL and REINTJES: DISTRIBUTION OF ATLANTIC MENHADEN EGGS AND LARVAE ci 3 o X 111 o So >»■ xo • "■ "S o ^ O 111 ro < — O 10 i '" tritr in in ^ m o fo rg <0 <«' CO •H r- o ^'^ o> o> f*> <<• t fO ft ft (t ft * ♦ It ft ft — o« rg (\j rg cu ^ o og rg ««i o rg fn w u <>! rg d <»■ nT ■»• rg f) fri m pD ^ «r * ft ft (^ «r -T * (t («1 fr( O O * • t • •f tr, ^ • t • ..j 09 • t fn ^ t • (Ti in CO in t t • • 1^ o* f*) in • t t • CO -^ o> t • t • rg .4 rg • • • 34.7 35.5 35.8 36.6 36.1 36.3 36.3 t r- ct • ft 00 CO m CO rg ft ft in CO 1 • • t t t < '4- in « 0>o> co«»"o o.^ coco oorgrg ftfri^fvj ft>*-ft ^ (»i>t^ ft rgrgrg rgrjft ftf<» tjrg ititftft ftititft (tfttt ft (tftft S ^^ g ai a >- I III III II II I t I I I I I I III I I I I I I I I I I I I I I III ci — c ^ OUl IXJ UJujlU LLujUJ UJUJ UJUJ UJlLlujUJ OIUJUJUJ UJLUUJ lUujUJUJUJUjUJ UJUJUJUJIJLIUJUJ UJlUlu ^ iriu z z ^ z z z z z z z z zzzz zzz z zzz zz zzzzz zzzzzzz zz z ujo z zzz zzz z z z z zzzz zzzz zzz zzzzzzz zzzzzzz zzz s •-«= .a • +j y~ m (co^in o^r'-| mrg p>r- oj^oco ecircTft r-fsjrg ir"0^-oao0'^<-i ft-Teo-o-otJO ftfnin S ♦ o t ♦ cc 00 r^r^oo r-r-c> ct>o oo 'J0"'0 o^-^"-* ^-rgrg (sjfMrgrorvjOo >j-o^t^r-r~r~-h- tr<7*'inoi'^tn rgrg^f O U cc • I'f t«t t, •• itat •,•« t>> •••ti<« •>•,•(• •§• M o 3 00 h-cooo r-r-o> oo oo «rooo orgrvj— ^rorg rgrgtNjrnfoftfn ^oo^-oooo fttriin V •^ (/) r-i-< mt —• ^1-1,.^ rt — _ii-i ^.H^ _iWr4<-i.^fMrg ^r'>-l^g(^Jrgr^g rgrgrg S Oi 05 ec z fv r-o-if (omrg arg irr* ^JinOflo n w o^ tr\ -f er, t^ •* ^ uj CD r~-r-co r^r-^ ao oo 00"^o o •-' i-' •-' -Jisiro tMrjfvjftrgrtM sj-oof-ooooo rt»jin ,^ < s 1-1 ^1-1 "-^»^^ ^^_i^ .^^_i ,-i^,_i.-i^f>gfvj i-i^^rgi-^rMivj rgrgrg »• (V c e\j cr r- r^ c> o 0^ 00 o c in in .^ rj * r- en — o >-i ^ .*■ 00 in -o a- i'^ ■* rj ft .*■ r- CD on r^ r^ o o o o o o 1—4 o ^ o o rj .^ .-1 •H ro M rg rg ryj i-< I-* -J ft 1— t rvJ M st fr. • • • ^ rg t • •»■ O f^ . • • O • on f .H in • • f "^ in '^ ^ iTi CT* * f" -t t • t j3 a. UJ < ^J^ ILi O 1* -a < "a « oc . . . .-.-- ...... .. ^ « CO r-t~-uj f^f^c o-o oo crooo o-<-ji-i <-if\jrg r^rgrgftpgfriM ^oBf^o^h-oo" rgO^m (3 « rt.-1-H ^t-<^rt-<^-H,-i_)i-i^i-i^i-i_irgfvj_i^--'fvj '§ X to a:>-— 1^ t-iccfi r-f--o oca fiin a-£ir~->f! in'-'-nin ocmin fiooOgD-ooo a-l-minincoco >i-'*rg S2 UJQ-r rj ^Hr-*fo ,-(^4fvj ^rg fMfsi t-f^^rg ogftftft '-•mft ■-••-trvjfsJrgaa ^^rgro-^'^rgrg r^ftco ~ »- Uj — -- — 2 < Ci * :* 0) IZ I < eo e ooc coc c<-^ oo ccoo oooir croc cooc-^oco occcood r>z>o 5 _j r z z z z z z z z zzzz zzzzz z z z z z S a u. y- f^ or-r-ir r^f<"0 rgg; r^f^ oOfvf^ ocoof^ Cr^jf^J coorr'inintnu" iroo,--i,-ic frinir. r^rrirj-.^rrft oir.i-rv)(\.oo r^frf<^ C c»-»-u f^ -j-ira r^ot— t<^ ••* m r~i aoo''^ c.a-^>o o~'--< aj(^ir.J>*-f\'rj oor-iu^mcra vog:m ^ ^(/)K»- rg ,— ^^ — rtC r-jrj o—i —•-loO oooo —ifvrg -<_4.-i,-i»j^-j Ooooooo ^ ^ ^ O. o >» ujirsrg fMf\/rg rvjcsifvi fvjfvi rgfvj rgr^rjfv rjrdrvjrg (\jf\jrg fsifMrgf\]rj(\jf\j rvrgrofvirgfNjfvj fMrorg mm l^•-^t^^o^O'JCT■^^-'rg-J•-lrJrgOO O* 4^ t^ <\t r*^ pM#-4 ^Hi-H r^ r~* ,-<<— 1^^ ^J (V^rfi— Ir-J r^<— • ^^i-t^^-rf^Hr-l— 4 -^i-^i— I S *X<-J -^ ~t cpvj rgo- f~- >C r^«r^ —i ft-j-^in inoc inc-0--^ocrft f^fnf.. ^ O i^' a CO oo m IT, m— *fci c fvjomC" fi^r* rgfvj>^sfir>fnir. f^oooo It-ZS" -H CM ^ ^ ^ ^ —I -) ^-.pj <-<^ ^ « UJ a. j^»- f^ r^fMC or~o o^rf^ —■a-' >}-u>,-(r^ »-»o^- rvf'ir >^fM^ce*C^lr^- ftfv'rom— -fvo fvrvjin > Pi ^ ^ • Si _j 1/1 ^ ^,-j^ (\-^^ j>fv tvfvj ^in^M — Crv cvro t-ir>-r^' -'f^f^fn -j— . crC—'ir— ^^^ ir<-c>r-'f^O orfvr^, g< « ^ t- -J r- 0-. in o -J rr, f^.' f-- <* T 3;>_— If .fiCr>fir ir. l^-i^^^ ^l^nJ- rr. Cu-, ir-*irf«- -riririrf^irf iTfrin C o 7' f~ (\ ^ ^ ^- ^•-'p-fMp— »-r ,-1— 'r>j— -r^ ,-t^-,-.f*^,--fr--fr^- H U- — I III III II 11 till I t I I III I I I I I I I I I I I I I I III c <" r c cr oco cC oc" cooo occrc oocr coocaCe oc"oc~cccer OcO u I/- t .r Z) t J- IT. 331 FISHERY BULLETIN: VOL. 73, NO. 2 UJ >• X zo ^ O UJ < •.> u to z < CO <*> o — ^ -o o • * • • (^ ^ r- >0 K f^ h- ■f 0 to \ri -^ in m CO rg tn lA "^ -< M Ifl ^ ^ ^ fO -< «M * m (»i m -- -« ^J ^ (^ t^ tf\ t^ -4 0 (M in if( (^ fO ^ ^ fO m w «n lA in in fO (^ f^ 0 Cl ^ o « o r\j •rf n m m in -< rg *■ (T> fn ffi "" »< rg m 1*1 (H <^ fo ••4 m rg ui in i*> rn (^ f*> (*^ m n m m m m d CI CO f*^ f^ ro <0 (M • • • a> eg « o> 1 • • 1 o» >»• « t t • « f» m 0 • • • • fo r~ .O «i CI • t • t • in rg CO a in -r 1 t > • t • ^ rg rg O -- rg ^ d »*» CI (*1 •^rg-^ --^rg^ -jrgrginm c\ o c^ ^ in in CI ci ci c> CI c> w CI — o CI o o CI >4 «4 O f^ o c* t o CI u CI •«■ •M O F> « • • • • t rg rg rg pg in O fO CI fO c> « ^ rg o^ c> c^ ci c^ c^ X — z lu •- o o O UJ Z lU Of oe UJ o X UJ »- o I I UJ UJ z z o o z z I I I I UJ UJ UJ UJ UJ z z z z z o 0000 Z Z 2 2 Z I I I UJ UJ UJ Z Z 2 o o o 2 2 2 I I I I UJ UJ UJ o 2 2 2 Z 0000 Z 2 Z OC I I I I I UJ UJ UJ UJ UJ 2 2 2 2 2 C3 o o o o z z z zz o I I I UJ UJ UJ V UJ Ol 2 Z Z < 2 Z o o o UJ a o zz z-xzz — r UJ UJ 2 Z o o Z 2 o 2 o o I I ^ UJ O U UJ < 2 2 2 2 UJ O O CJ o X 2 ct QC Z y- y- « * O m e 3 r- < in in — ^ rg 0 «>• -< Cl U' 0 0 "J" Cl r- P~ •H ^ <-■ m in <« « 00 00 >t ■4- rr\ its >*• f~ 0^ 0 in 0- r- >t .t in 9-4 ^iri ^ f>- a r^ 0 0 -H rg rg ^ rg O r-^^ff^cr^n • t • tttf* O &■ "^ Or-ia>^0D i-t r-t r* ,-> ml (Ti r- ,0 rgoorgrgO » • • tt*t* o o — rg — Mrg>-< f* w^ v-^ ^ri^4i-ifHirg UJ ec z 3 < t- UJ < z: ec a z UJ * « UJ (J < oc « 'O in in * in r~ rg ^ f^ Cl 00 ff- 0 <0 -< Cl Cl >c in •»• Cl CD 0 CD 0 <-> Cl -< — d •* rg Cl rg ^4 Cl Cl fo vO in CD CD 0^ 00 ^c t-- - 0 t o ^ 0 0 O' U ..4 Cl -< -4 (Ti .r rg Cl Cl ^4 Cl Cl Cl f^ in cc o r>- tr h- 0 o> -< rg rJ in in in • in m ^ •»• • • • • 0 Cl rg • • • c r^ o> M • t t • IP 0- f^ rj m in rg --_. 0 * -H --1 Cl •*■ ccimfM mr-^-CTCo ■o r^ a ^ o er «-< ro i-i •*• ^^ in c^oocioo • • • tctf* O O ^H o»-*^^o^^ — — < — .-I— 4,-i»4rg ■t r- >o rgooinoo t t • ••tt« o o •-' rgi-irgOi-< ^4 m-4 md ,-t f^ f^ m^ r^ o> c^ • t CD O -e Cl • • a- -J ^ r* ^ ^ ^ O '-' a o- o ■-" '^ rg I UJ a I- UJ < Q 3 X -< g3 ro rg ^ ininrgo o^ciin ciOOo rg »-rgrgin — rg^t i-iroMCi (7"-4-crMfv c ^^rgQs^O' — M i-i rgMrgrg^j- I 2 e c 1- (_) K->- > KI-KI- >->V >- ^ >- i- KI-KKI- >->•>>>•>- II < IIII <<< CJ D < V >- 1- I- I- I- < < I I I I C O C O C e' zz z z QC U- K O 0" Cl 3 < s: 1/5 o m .*• O *- >-' u. rg o •— H 1/1 I- — 00 — 1 cc/CDrgoo -"^m ci.orvjcc ciciprc^ •— 'fvjrg frr^rg»-- f-iOcic f^ccrj »£)f^coo •J-Oocco^ f^^«J'c^crco drg<-ioo cisj-"^r<^>l'>l" O'-^romir r^rg*-^ooooo m 00 in rg fM Cl CO in >o o "^ o •-I fr\ — cr 00 IT o in IT ,}• ^ rg O f^ !• c^ >o O- a 0 .J- ♦ t • • t • • 1 • • • • • • • • • • • « • « UJ rg 0> * 0- on r^ 0 r- c Cl or o- r- Cl r- c ■C m *H 0 in • ►- a H • t • • • « • • • • * • • • ff f • t • • * 0 C 0- Cl r- in r^ h- IT •3- r-- -^ t^ IT or Cl ^ >c r- f^ «o » 2 X i-t w* <-4 _i •-« •-J 1-4 ^H — I—* *-4 ,^ « UJ X ^ 2 «.« >-l IT m er< Cl 0 c Cl >c 0 •c on ■0 (^ Cl cr h- cr in cr frf 03 -0 ^ m U. «i t • • • • • • • • • • • • • t • • t « ■ • t t • • < u. -- in r- >c h- r^ r- rg C f^ r" ea >n r^ 0 r- c- ^-^ •-4 ■* f^ ~j- M 0 0 > s <\ -J r ^-^ r- o^ ^^ (V f-* t-- (^ w^ r^ ,— ' p>j f~i V. <— M o-l r-' v* --I a «i • ^ :s rg — 9^ C^ ^ t- — g? •^ --4 Cl M (V rg vC IT 0 in •c c a on c •1- »-i in n. « UJ —4 « o: X « UJ • * tD 3 « T _J » 3 < rj ^ r-* c- ^ — — ^C ■4- — C" r n' r •£ IT 0- in ■c Cl 0 eo in — IT. C « 2 t- -*• f—i 0 -t M * c 1-^ * 1— 0 0 1 T bd ■»- 1- ^ ^ c IT IT r J C 0 5 •- '-^ •~' f-4 Cl ,^ ^- f- •M Cl •— > — Cl ■< »- Uj c — 1 1 c e 1 C 1 1 1 1 0 0 f^ 0 1 0 1 1 c er 1 0 ^ 1 0 ^ 1 1 1 c 1 0 1 en 1 0 1 c 1 0 1 0 1 1 0 or H w^ w^ x 11. 0 l^ t •-H z y- ^. o X ■*• ~i- >!■ o coo CP o .-I -4 r-i rg c CJ »- >c >r >f 5 will Ceo I o •4- -l- >J- .J- ^l• OOOCO rg rg rg og rj 0> «*■ r- h* vC rv rg »-< ^^ p^ 0^ -J Cl •J3 (— ) 0 >J- K • • t • • m 0 >t ^^ Ht 2. rg rg *-* •H •—1 2. «^ ■* Cl 0 r- r\i c 0 m 0 • • • t • • t • t C r^ IT (V in *t \r tr C" rg r^i ri n f-< f-" «— ' ^-1 •— ' Cl —■ c- ^ -' fr, -J C >!• -< IT %f in .* u" ^ -J r, ^ 1 I I I I O c o or o f- ry, r- r «N N K r~ 0> < 1 • t f • t • • • • t lU O CO O M o o O o o o o >■ z m fvi «n •«> fO CO m w w 1^ w t o «M f^ >»• »>l 0> « r\ oo 00 p- o z o • t t • t • t • • • t ■>» -< o o CM o o o u o o ^ o UJ m (*> 10 m •0 m «*» fO m m PI < ^ u> no z ^ ff. r- o «0 (*l f-4 « <0 « 0> < t f • • t • f t • • • K eo ir\ ^ N o s; % o o o o N « «^ ^^ m a> pg 0" f Z z z •4 -4 PJ rt UJ a «^ 1 1 1 «■• 1 1 1 1 1 1 1 1 z (U d ^ ^ in in ^ .*■ tm o f>4 9^ _l U o 111 j^ u u UJ UJ UJ UJ U >£ o u s lU < z z z zz z z< z K K UJ O o o o o o D UJ O Q lU (9 X K ee z z z z cc 3 ec tc XIU H h- H \- >- W o «/> 1-0 f) - m <^ m m < t • • • • t • • • • f >- UJ ^ \t\ ^ - 0- 8? OP \r\ in m 00 fP in —4 « oc • • t • • • t • t t • * tt\ PI PJ IT 00 00 f- tr fP •*• o * I ^H f-H ^-t 1-4 1-1 f-4 1^ *-4 -H PJ oc t- <>*■ ^ a 00 «>^ * f- «*■ ^ o a 00 PvJ p- UJ o. r —1 "-1 <— 1 z v-J »-4 (SJ z: PO -.4 IN/ m m K UJ vc ^- •ta> < o X • H- O >- >■ h- >- V >• >- K >- K >- X ^ < < I o o o o o (J O O C o e t^ t^ o ^K »-i »— ■ — ^ z Z z z »- ,^ ^ „ a: UJ K IT r- o> K IT fcl r- H c- Pi r^ o m S < X" (/) o ^ (-4 (/) O rP f-H wo 4- >t m r t-4 c t- UJ m IT *-< UJ r- o; a UJ P>J rp o -H in t- i/i K — o o o - 00 tr tr « ?■ S" Z Z » UJ s- z Z -1 z o o O m^ K in O r U' S (\ (NJ f^ i-« «-4 or 0 C t' ly c -5 -5 H o UJ U- u c cc c (_; t_ O f\j4-pj »

i*moJ^ otPM^ •M«4^^0>-'0-'>' '^ r^ r* O r-» fr\ ff\ tfi p^pip^pipip^fnpifp pii^pipnpi fnmpi mpiinPiP»p^p>p^p)Pi pimpipipi pvipio «^P^p^<»'f>•o>^•om «^ooo>o .M_i>d 0'^^<-<000'^-' 0^4-iOi-4 P1P1P1 P1P>PlP>P^P>p>PlP0 P1P1P1P1P1 » m o<(^pJpJpJPJj'.*>»'«t pjpimpipi — Ill iiiililii iiiii o inif>0>0"^.*inin f> pjPJCvjPjpjPJPl 4Hpjrjpj~ l-KHKI- Kt-H>-K l/> l/)(/1l/1l/1l/) 4-i p-p-oooooo-»'-*-4-i crmm-^^ pjpjo pjpjo^o-mooP'r-tr 4-4OO0'0» -<.-l'-4>-4.-l ^4-4^ mpipi -4»f^* inininminininmin ino >«-pi-»- <*'Pi«e-*inintn.*-in in.o^inPi p^O^pi —j^t^^o^^mPJO'fNJ ^*^pjoo ^p^-i- inpi^inininin^m inc>j3in opjpj inp-r-ff'>toocr(r-j p-p>i-^h-^ oo4-'4-4i-- in^4.-4oooo s: piPi<»- pip^.*-.tPiPi (-•->- WKl-K>V>->->- XI < IIII<<< >- >- >- < < < < < c cj D a o Z Z -ZT rvi t^ p- P^ (T —' U C P- P5 a o o —4 —' o — — — -^ h- cC^in P*P-'4'»^^»OODOOO i/i 004-4 plp^P1p^oOP"p1l^ UJ pjpj>o cr-o^oopipi-j— ip- »- OOO — r-4PJPJ^-j.-4-HO JDSTOOO OOOOOOoOO ooooo ^ .^f— '1-4 •-4-^--l,H-^»-4^*»-44-4 -H4-4r-4»-4 — CDinmin ^■*>r'*'«T>4-t-»' p^pip^pipi -H -H4-44-4 t-4,_4--4*Mi-4--4r^-Hl— 4 f-.4-H-.4-.4-^ z z w O .t (^ ^-pJ^--Jp^c<^J^^ lrleoeo4-4-JO^cC^ 1-4 -W ,-4 ^^ -^ -^ C^^O^Oi-Jp-P^C • ••••••t >c. pi.*->*'p-aa->c PJ Pki Pv.1 ^ • • • a Cn vO PJ ^" -^4-4^-4-^ O r^ fn >}- • • • • m p- Cf in tr ^ c p-, m-jiro-O'r'iO'c 4-4^0—' — 4<*-4-' inf^pf^ffPJO^P^-J ITP'.toro -H P! P ^ IT >-j in c ^ — 4 oc t-4 (M P; (M C >*• ■- -^ iTsS-m m^?'l^Pl^■pl^Pl^ y •- i\i >- — p' ,^ p ,-4 p-, ^ p- -' w III I I I I I I I I I OeoC OocCffOtrCcr-O •a c t\ <\: -t fvj(\;P;n^^trin- .-4 p-. IT ir vt IT p^ -H ^ p •- p^ I I I 1 I o o or C or fj P fr < >t C (^ I.- l_ c^ 333 c o X FISHERY BULLETIN: VOL. 73, NO. 2 < «" (J u) z ^irNGDooorsjrvjirii-i (naooo* pjm •- lllllllll llllllll ilillli llllll III II II mUJOlU-'LU^^l^O UJUJOOOUOO UJ^^OO^^ UJlUtfOldC) lUUJUJ UJUJ ^9^ Z22zz<-^-h•>o>o>o a> cuooooooco" crc^tr -jpc mi-' fni-io^r\j>o^.*'ir c^■<•<^J>c■*'<^J>*■'^ O^ofi^rgoo fnfvjOC'-*-tr ^^-< >co oc !■ c-fvjfvJrvjfMrominif* ■-'-t-(_'-l-l-t- >>(-!-»-»-(->- t-t-l-^-l-VV KV>-VV>- >>->- VV >-V IIIIIIXlX <*IIIXXII XXIIX<< I< < ►- Ol < r a UJ e. *■ UJ UJ (J >- z «I « a: * * X a K UJ Ol \- O) < o 3 • 1- C T Z c c- •- o — K or OJ 3: - 1- (jO 1- IT ^ oa 0 o O UJ ir>XOOoOOOOOO OOOOOOOC OOOOOOO OOoOCO OOO OO OO (vjr-rteccoOir\f^o ^-frlOlf^^•0(MU^ o-«»'i'^M^in^ ir(»i-*'r^rri r" o •<•>*■ r-r- 4f ••••••••• t»*tt**i ••#••*# t*t«* • • •• •• ^ Uj oooosjOf\ji^f'i coOm^'^'^«lf^JO >c>-<t roK 5^'< Zj ^sr-Hr^r^o><(<"ir\ •coo^O'f^O-o hooooo--cr r-r-Mar^ o a \r \r inin ^KC K tft****** #••«•••• ,t*«*«*t ctttt • • ft •• * Z T _H -- ^ UJ 7 IXJ z r*^f<> ro fn frifrifn rornrrtr^^rn fO(^(<)f*^f*^ fornix r<^ rr\ t^ ff\ r^ (rifOfO r^f^f^ (r)f<^f*> I — 0 00 -< .t f4 r* fH ^ O 00 ^ t ir\ m r- »j O * 00 m CI WJ "^ <\J (»\ n fo pg fv (sj d m w ci o ci m CM CO «r '^ tf^ (vj CO ^ cr >0 o -r m CI OO O •-< "^ j n po fri fn (*^ m m (M rg fT» ft) * C^ rn (t^ ci CJ IN CM CI ci m CI CI CI >r «»• •»• CI c> c^ c^ o *^ m ci CI C> CI CI Ct Ul CI CI CI >r m m CI CI d a o • • f • ■»• 00 m t • • >* m in*r- o o -C •* "^ 0>eD>cciin wCJff> ^n<^)^-^-eo iMf ^minm.*- inin^ ^*■•9•^•J■^t <*-inin >tinr- >oooo~ Ol a. z a->-< -*• >o inoo f~ ^ v\ ^ ^ eu^fvjnjr- m— 'cr ir^-r^mr^ Oi-'O' ooom *r^-j- 3< ••• ,•... .• K UJ CO o- nj c^-*-^ Nj-.^inin-*' >i-ininsj->*- inm* ^-s-^^t^ >j-in«* ^*■ln^- oco ^^-rjo C- f-* T.j- mirmin~5" >ririn>*'ir mmm, ■*•>!•*•*•*■ ^mm ^mf^ >oeocr UJ U) ^ •-> ^— oin in^rvioci i^oocrom -joo> in>-K-*in moot oocin •fjsc « OC t • « CO c nj f<-,m,j- •J-^mm-*" *in>*-^<' mmN*- ^<*'^-*-^ -j-m^ ^.j-f^ *orc QCK-^ or^ «^ ^^ ccOsO of^>cCC Omcnj**" oonj mo>*'cr^ ^^f-^rvi 'OO'— ' m^o UJQ.2: (Md — -< ,-ifvJ-i--fvj -JPJrvj CC PsJtMrO «M(M<*' MINjCl h- U. — -^ < O l-O ZK K t- K»-|- >-t-t-l-t- >->->->-l- t-t_>- >->.VI-|_ >>->- >->K >->>- XZ 3X X X XXX *■ 'm.^oc.^-ci «On.flO (\c<^Jmm ec^r- oOc -i-r^c s-'o i-i^— Cr-C"nj moch-m' orsjm oocom--' ecr-r^ m"J"m ccec l-l/lK-w OO -- (M OOo — »-i,-i--rv. -^ — OOC OC"-" OO^CO oco ,-ii—rsj -joo U-I^IfV^J fV '* ^ <-* ^ r-l-^<>4>4W »<.^^,— -4 --^^ _<^rt_H'' ^ — _| ^--_| ^^^ (_.0 --I— -' -^ ^-<_. ^_d_j^^ ^^^_j^ rf„^ —_- — ^^ ^^_ ^-J_< ^—1^ c mm mmcoo o^- mo m co o<*'^ ^-j 4/f •• ••••• •• •• t •• ••• t« _ om f- cr cr^r ^•m^com o^"^-mo u^>cm ^mom-^ Cf^c r-oo" ^-oro u, < .• • t • .,.,t ••• ••• ••• ••• < u n-- ^ c -"Om ojscrmo' mmmmp- ooc mrv>c^m om--' cooc cCr-- >3^ (Mfvn.--n'n— 'fv-— ^- ^~ ^^ r-* ^ ,^ i-',— ,-*-'^^,— 1«-- r-* a. < • J ^^ ^-t-i (M m «--cnj mmm*-"^^ •-»-jfv;n.'^- »-i.-4^ r^^^^-r^, ,^ mn'^ cNO t-Mstmi *' lf\ •-* »^ c" ^^ ^^ * UJ * ecTL « UJ * ec * r _i g •~i-<^ ■-••-in-nj^ ,-i^>c r-^i-in.^ m.n;— i mr^o ^>j-m a pq 2i_^ >cm ^ >c >c^m «ccmm^J•m vC^-*' ■^ <: -c c (vm >- -" --fviC f-'(\ic.j-m »-c^mK n«c^ -ifMc.tm "-iimC —m^ ^(mc ^ £v5C i_«J C u. u-U-u- cccoc ixxll "5-!-: i;s<:je:s^ic _J-J_J TIX ZZZ 335 AGE AND GROWTH OF PACIFIC HAKE, MERLUCCIUS PRODUCTUS Thomas A. Dark' ABSTRACT The age and growth of Pacific hake, Merluccius productus, collected off California, Oregon, and Washington in 1964-69, were studied. The age determination procedure was examined and considered to provide valid ages. Several sources of variation in the age structure of the population were given cursory examination. Relative size of the year class and sampling area (average age tends to increase with latitude) contribute substantially to the variation of the population age composition while sex and sampling season have lesser effects. Growth in length is rapid during the first 3 yr after which it slows and approaches an asymptote in the oldest ages, 10-13 yr. Females have a faster rate of growth than males and tend to survive 2 or 3 yr longer, to age 13. Growth in length can be adequately described by the von Bertalanffy growth equations: /, = 56.29 (l-e"^'^^ ('C2U>) f^^ ^^^^^ ^^^ /^ ^ gj 23 (i_e-0-30 (^-O.Oi)^ j^^. fgj^^les. Year class variation in growth rate was detected by back-calculation, using the body length (F) -otolith radius {X) relationship Y = 18.78957 - 3.79065X -1- 0.67490^^^ - 0.01836Jf'. Growth in weight was determined by use of the length-weight equations: log VF = - 1.45990 + 2.55618 log L for males and \og W = - 1.68944 -I- 2.69509 log L for females. Males attain an average weight of about 1,211 g by age 11 and females reach an average weight of 1,374 g by age 13. Annual instantaneous growth rates in weight were computed and were found to decrease most during the fourth year for both sexes and very little growth occurred after the sixth year for males or after the ninth year for females. The Pacific hake, Merluccius productus, is a com- mon gadid fish that ranges from the Gulf of CaHfornia to the Gulf of Alaska (Hart 1973) but is most abundant from Baja California to southern British Columbia (Alverson et al. 1964). There is apparently a single population offshore and another in Puget Sound, Wash. (Utter and Hod- gins 1971). The Puget Sound population supports only a small fishery and is not considered in this report. Feeding adult hake are usually found over the continental shelf and exhibit pronounced diel movement. During the day they are most com- monly found in compact schools near the seabed, but as darkness approaches the schools rise and become more loosely structured. During their spawning period mature hake are more pelagic in behavior than during the rest of the year. They apparently spawn at intermediate depths in water 1,000 m deep or more and demonstrate little diel movement (Nelson 1967). Spawning occurs from January through April off northern Mexico and southern California (Ahlstrom and Counts 1955). ' Northwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA, 2725 Montlake Blvd. E., Seattle, WA 98112. Eggs and larvae are pelagic and are found mostly near the thermocline at depths of about 45 to 100 m. It is not clear at what age juvenile hake leave their pelagic phase and become more closely as- sociated with the seabed. One-year-old hake are found in inshore waters off southern California, associated at times with schools of northern anchovy, Engraulis mordax (Dark et al. 1970). Hake, 1 to 3 yr old, are taken in shrimp trawls along the Oregon and California coasts (Morgan and Gates 1961). Pacific hake less than 4 yr old are rarely found north of Oregon. Most 4- to 13-yr- old hake mature and are found feeding off the coasts of Oregon, Washington, and southern Bri- tish Columbia during the spring and summer. By early winter only a small portion of the summer population remains in these areas. Temporal and areal distribution of the various life history stages suggest that adult Pacific hake undertake extensive annual migrations along the west coast of North America (Alverson and Larkins 1969). Most adult hake seem to move northward along the coasts of California, Oregon, and Washington in early spring on a feeding migration as far north as central Vancouver Island. In late fall the adults begin a return Manuscript accepted August 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 336 DARK: AGE AND GROWTH OF PACIFIC HAKE migration to the south which terminates in the spawning area off southern California and Mexico. The eggs and larvae drift onto the con- tinental shelf and the young inhabit the waters of California and Oregon as l-to-3 yr-olds. Some 3- yr-olds and most 4-yr-olds become sexually mature (Best 1963) and are recruited to the adult popula- tion. Pacific hake were landed in small quantities in California ports as early as 1879. California land- ings from 1916 to 1951 varied from about 0.2 to 90 metric tons. An animal food fishery developed in 1952 creating an increased demand for low value species and hake landings increased to about 590 metric tons in 1956. From 1956 to 1968, landings averaged about 200 metric tons annually. Prior to 1965 Washington and Oregon fishermen did not purposefully fish for hake and, in fact, considered them a nuisance species. In 1965 a small fishery was initiated under the guidance of the U.S. Bureau of Commercial Fisheries (BCF)- to examine the feasibility of efficiently harvesting Pacific hake off Washington and Oregon. Four vessels began fishing commercially in 1966 because favorable results were obtained from the feasibility study and a new fish reduction plant had begun operations at Aberdeen, Wash. During the same year a large Soviet trawl fleet appeared off the Washington-Oregon coast fishing for rockfish, Sebastes spp., and Pacific hake. Competi- tion from the Soviet fleet was so severe that it seriously threatened the existence of the U.S. fishery. Negotiations between the United States and the Soviet Union in February 1967 resulted in an agreement which restricted the size of the Soviet fishing area off the southern Washington coast. The U.S. fleet was more successful in 1967 because of a reduction of Soviet competition and increased efficiency of U.S. vessels. This greater efficiency stemmed from the increased experience of fishermen, improved fishing gear, and greater scouting capability. The total U.S. catch in 1967 was 8,381 metric tons (catch-per-unit-effort [CPUE] = 4.5 metric tons/h), as compared to a total catch of 1,694 metric tons (CPUE = 3.0 me- tric tons/h) in 1966 (Nelson 1970). Soviet catches in 1966 and 1967 were about 136,050 and 170,590 me- tric tons, respectively. Since 1967 the Soviets have continued to fish for Pacific hake and annual catches have averaged about 140,000 metric tons. The U.S. fishery was discontinued in 1968 when the Presently, the National Marine Fisheries Service. reduction industry, facing a depressed fish meal market, was unable to give vessel owners prices that were competitive with those offered by the shrimp and groundfish processors (Pereyra and Richards 1970). An agreement pertaining to the joint exploita- tion of groundfish in the northeast Pacific Ocean was negotiated between the United States and the Soviet Union in 1967 and renegotiated in 1969 and 1971. Scientific meetings have been held annually to discuss problems of mutual concern such as as- sessing the size of the Pacific hake population, de- termining the effects of the fishery, and estimat- ing rates of growth, mortality, and maximum sustainable yield. Recommendations resulting from the scientific meetings provide a basis for modification of the bilateral fishery agree- ment—which can be done every 2 yr. Initial growth estimates were based on preliminary data but served to provide essential real time estimates of maximum sustainable yield. Subsequently, additional data have been collected allowing for refinement of early growth estimates. The objectives of the present study were to provide new estimates of the growth rate of Pacific hake and to examine some of the potential sources of variation. An analysis was made of the reliability of the age determination method used since age information is basic to growth studies. The variability in the age structure of the Pacific hake population was also examined since age composition is frequently used to evaluate relative year class strengths, mortality rates, and the ef- fects of fishing. SAMPLING Collection Methods Biological data were collected from two sources: "commercial" samples from the commercial fishery and "research" samples taken aboard research vessels. Commercial samples were taken mainly in 1966-67 when a U.S. hake fishery was conducted off the Washington coast during May-September. A sampling station was established at the reduc- tion plant in Aberdeen, Wash., where essentially all hake taken off Washington and Oregon were landed. An attempt was made to man the station every other week and to sample the catches as they were unloaded at the plant. Irregular landing schedules, especially in 1966, resulted in sporadic 337 sampling. In some sampling weeks, landings were made during 3 or 4 days while in other weeks there were no landings. Landings were sampled as fish moved from the vessel to the plant over a conveyor system. Approximately 200 specimens were collected for each sample. Specimens were first dissected to determine the sex, then were measured (to the nearest centimeter) from the snout to the fork of the tail. An otolith was removed for age determination and, when time allowed, whole specimens were weighed to the nearest decagram. To simplify the collection of otoliths, most otolith samples were stratified by 1-cm body length intervals. Otoliths were taken from five males and five females in each length interval until the sample was exhausted. Although average length- at-age information can be taken directly from such stratified samples, randomization was neces- sary to obtain unbiased estimates of age composi- tion. Research vessel samples were collected from 1964 through 1969 and were both stratified and random. Most research vessel samples were processed at sea for the same biological data as those taken from commercial samples. When weights were taken a small hand-held steelyard was used which provided more consistent readings than did spring scales. The research vessel samples used herein are geographically and temporally restricted simply because it was beyond the capability of a single vessel to conduct more extensive sampling. The samples taken at the reduction plant in 1966-67 provided the best temporal coverage over a season, but their areal distribution was restricted mainly to fishing grounds off Grays and Willapa Harbors, Wash. Sample Representativeness Whereas the nature of available samples places constraints on some aspects of the following study, the large number and size of samples collected over several years and over a large part of the species' geographical range render them valuable in examining the reliability of earlier estimates of growth (Best 1963; Tillman 1968). Also, some of the more conspicuous variations in both growth rates and age composition can be examined. Figure 1 gives a general representation of the distribution of sampling effort in 1964-69. The adult portion of the population (4- to 13-yr-olds) 338 FISHERY BULLETIN: VOL. 73, NO. 2 occurring off the coast of Washington during the summer was the most intensively sampled, especially in 1966-67 when commercial samples were taken. Research vessels sampled adults off Washington, Oregon, and California, mostly dur- ing the summers of 1965-67. Juveniles (1- to 3-yr- olds) were sampled only sporadically and much less intensively. In the winters of 1965 and 1968, research vessels searching off southern California and northern Mexico for spawning hake obtained some samples of 1- and 2-yr-old specimens. Very few 3-yr-old hake were captured, probably because there was relatively little sampling effort in areas where they were likely to be most abundant. Commercial vessels fishing for Pacific hake off Washington used Cobb pelagic and BCF univer- sal trawls (Johnson and High 1970). The stretched mesh size varied from 5.1 to 7.6 cm in the trawl bodies and cod ends. Research vessels used the 50« 130° 45' 40' 35' Figure 1. -Distribution of sampling effort for Pacific hake, 1964-69. (Darker stipling infers more intensive sampling.) DARK: AGE AND GROWTH OF PACIFIC HAKE same trawls, but with 3.8-cm liners in the cod ends much of the time. Liners were used to determine the availability of the youngest age groups. Research and commercial samples were never taken in such a manner that length frequencies could be compared to isolate the effects, if any, of the 3.8-cm liner. But there was probably no sig- nificant selection for fish length without the liner since similar unlined trawls used in the Puget Sound hake fishery apparently retain all fish of 35 cm or greater.^ Because very few hake off Washington were as small as 35 cm, sampling gear differences were not considered to be a significant source of sampling error. Therefore research and commercial samples taken in 1966-67 were com- bined at times to increase sample sizes and to improve temporal and areal sampling coverage. DETERMINATION OF AGE COMPOSITION Aging Technique A prerequisite to any growth study is a method for reliably determining the age of individual fish. European investigators (Birtwistle and Lewis 1925; Hickling 1933; Bagenal 1954) found the otolith to be the most useful structure in deter- mining the age of the European hake, Merluccius nierliiccius. Bigelow and Schroeder (1953) arrived at the same conclusion while studying the silver hake, M. bilinearis. Apparently Best (1963) was the first to age Pacific hake. He found that from the standpoint of availability otoliths were superior to scales as most scales were absent on trawl-caught specimens. Our collection of otoliths was standardized in an effort to control sampling variation. Samplers at- tempted to always collect the otolith from the right side of the head to avoid any confounding effects due to possible otolith asymmetry. If the right otolith was damaged during the extraction process, the left otolith was accepted as an alter- nate (5-10% of all samples). Otoliths were thoroughly cleaned and preserved in a solution of 10-30% ethyl alcohol. Occasionally otoliths with a uniform chalky appearance were encountered and were cleared by dipping them in a weak solution of hydrochloric acid. This practice was followed with ■'Larkins, H. A., H. H. Shippen, and K. D. Waldron. Features of a northern Puget Sound hake population. Unpubl. manuscr. Northwest Fish Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, Wash. care to prevent the dissolution of annuli at the otolith edge. Otoliths were placed in a petri dish with the bottom painted black, illuminated with a reflected light, and read under a dissecting microscope at a magnification of 6.6 x. Each otolith was read by two readers and if the ages did not agree, as was the case in 25-40% of the otoliths processed, the otolith was examined by a third reader. The best estimate of age was taken as the age agreed upon by any two readers. When all three readers disagreed (about 5% of the readings), the middle reading was used. If one reader could not make a determination and agreement could not be reached by the other two, the otolith was con- sidered unreadable. Generally there was a 3-5% rejection rate. The majority of the hake otoliths were collected during the summer (May-September). Assuming that the past winter was represented by the last (most recent) translucent zone, the age was taken to be simply the total number of translucent zones on the otolith. The few winter (February-March) samples collected were composed mainly of fish completing their first or second year of life. The same aging criteria cited above were applied to winter samples, except those otoliths without a translucent zone were assigned to age "1" instead of "0." This was done on the premise that the translucent zone would have been deposited shortly after the sample was taken, since young of the year would not have been captured by the sampling gear. Validity of Aging Technique Because the use of otoliths in aging Pacific hake had not been completely evaluated, some attention was devoted to determining the reliability of the procedure. Graham (1929) gives three indirect methods of evaluating the use of scales and otoliths for age determination: 1) agreement with Petersen's (1895) method; 2) seasonal changes in scale or otolith margins; and 3) observation of a strong year class over a period of years. The Petersen's (1895) method, which compares the relative abundance of age groups as deter- mined by length distribution with age groups as determined by analysis of scales, otoliths, or other structures, is generally only applicable to the first three or four age groups. For Pacific hake, the length distributions of the age groups overlap ex- tensively after age 3, restricting the use of the 339 FISHERY BULLETIN: VOL. 73, NO. 2 method to the first three age groups. Because few 2- and 3-yr-old hake were present in the samples, Petersen's (1895) method could not be effectively applied. According to Graham's (1929) second method, observations of the development of translucent and opaque zones on the perimeter of the otolith (based on data from a population that had been sampled periodically during a year) may provide an indication of the frequency with which the zones are formed. A single occurrence of a par- ticular zone during a year would provide an annual mark which may be suitable for age determina- tion. In European hake (Hickling 1933) opaque zones have been associated with good physical condition and growth whereas translucent zones have been associated with a lesser physical well- being and retardation or cessation of growth. Poor condition can result from a decrease in the food supply, the onset of maturation and spawning, or both. For the present study a special effort was made to record the zone type on the edge of all hake otoliths collected during 1967. Samples were taken during March, April, May, June, and August. Otoliths from a sample collected in November 1969 were added to the above spring and summer samples for data on the winter appearance of the edge. It appears that there are long and overlap- ping periods when zones are deposited because most samples had some otoliths with opaque mar- gins and others with translucent margins. One exception is the small, March 1967 sample that contains only 1-yr-olds. Even though opaque edges are plainly recognizable in the young age groups, all otoliths in this sample had translucent mar- gins. The persistent occurrence during the summer (the apparent growing season) of otoliths with a translucent edge may be at least partly because the newly deposited opaque material is not always detectable due to the thinness of the edge and the resulting transparency. Figure 2 demonstrates that the frequency of opaque edges decreases rapidly with the age of the fish. This is almost certainly a bias resulting from the increased difficulty in distinguishing the zone type on the edge of the otolith as the fish becomes older. Opaque bands on the otoliths of young fish (1-4 yr) growing at a relatively fast rate are wide, dense, and readily distinguishable. As growth slows in older specimens, new opaque zones become narrower, and are not always apparent 100 80 60 40 20 UJ 100 o < 80 I- 2 60 ui a 40 UJ Q. 20 0 100 80 ' 60 40 20 MARCH 1967 I 2 J 4 5 6 7 8 9 10 II 12 MAY 1967 APRIL 1967 H-1 I 2 3 4 5 6 7 8 9 10 II 12 JUNE 1967 tk 2 34567 89 10 II 12 AUGUST 1967 ti I 2 3 4 5 6 7 8 9 10 II 12 NOVEMBER 1969 ETL 123456789 10 II 12 123456789 10 ii 12 AGE ( years ) Figure 2.-Percentage of otoliths with opaque edges by age group (ages actually observed in bold print). until late in the growing season or until bordered by a new translucent zone. Because the zone type on the edge of the otolith is related to the age of the fish, a comparison of otoliths could be misleading if the sample age compositions vary to a large extent. Whereas the age compositions of the 1967 samples were similar, the age composition of the 1969 sample was no- ticeably different (Figure 3). To avoid the effects of advanced fish age on a reader's ability to ac- curately judge opaque zones at the otolith edge type, samples of 6-yr-olds taken in April- November and one sample of 1-yr-olds taken in March were compared in Figure 4. The graph suggests that Pacific hake start to deposit opaque material around April. The time of deposition may vary with age, but other age groups were not present in numbers adequate for comparison. The incidence of otoliths with opaque edges increased steadily through August when it peaked at about 72%. A dramatic decrease in otoliths with opaque edges occurred in the November sample. The foregoing analysis indicates that the physical well-being of Pacific hake improves in early spring 340 DARK: AGE AND GROWTH OF PACIFIC HAKE 80r 1967 Samples \ 1969 Sample AGE (year) Figure 3.-Age composition of samples selected for analysis of zone type on the otolith edge. 100 90 80 70 60 50 40 30 20 10 (1561 MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Figure 4.-Percentage of otoliths with opaque edges by month of collection, 1967-69. (Sample size occurs in parentheses.) (indicated by the appearance of the opaque zone in the otolith edge). Growth occurs throughout the summer until sometime after August. The onset of maturation and the spawning migration in late fall probably place additional demands on the energies of the animal, resulting in slowed growth which is reflected in the appearance of the translucent zone at the otolith edge. Although observations are not available during all months, the unimodal characteristic of the curve in Figure 4 strongly suggests that one opaque and one translucent zone are deposited each year and that the zones provide reliable an- nual marks. However, comprehensive monthly sampling is required to confirm unimodality. The third method suggested by Graham (1929) is based on the rationale that if a predominant year class enters a population of fish and if the aging technique is reasonably reliable, then the year class should be observed progressing normally through the population age structure for several years. Fortunately the 1961 year class of Pacific hake was extremely strong when it was partially recruited (designated as 4-yr-olds) to the adult hake population in 1965. In that year the 4-yr-olds comprised 15% of the adult population off Washington, while in other years from 1964 to 1969, 4-yr-olds contributed only 0-2%. In 1966 sampling indicated that over 50% of the adult population off Washington was composed of 5-yr- old hake. The dominance of the 1961 year class evidently continued through 1967, when over 70% of the population was 6-yr-olds; 1968 when about 64% was 7-yr-olds; and 1969 when about 24% was 8-yr-olds (Figure 5). In contrast, from 1964 to 1968, 8-yr-olds comprised only 2.4-13.7% of the popula- tion. This movement of the 1961 year class through the population age structure is accepted as addi- tional evidence that the translucent zone represents a single annual mark. Randomization of Age Samples As noted previously some otolith samples used herein were collected aboard research vessels and are considered to be random, but the majority were commercial samples in which otoliths were 341 FISHERY BULLETIN: VOL. 73, NO. 2 60 50 40 30 20 I 0 0 70 60 Ld ^ 50 I- z 40 UJ o 30 LiJ a. 20 I 0 0 70 60 50 40 30 20 10 0 1965 m^ 1966 1967 1968 ■4=1=1 969 n ■3' '5 7' 9 II AGE (years) 1 — r- FiGURE 5.-Age composition of Pacific hake taken off Washing- ton during May-August, 1965-69. (Shaded bar denotes the 1961 year class.) collected in a stratified manner. Before the com- mercial samples could render representative age compositions, they had to be randomized. This was accomplished by constructing an age-length key for each sample. The percentage age frequency per length interval observed in stratified sub- samples was applied to the length frequency dis- tribution of the entire sample. VARIABILITY IN AGE COMPOSITION Several potential sources of variation in age composition were analyzed using research and commercial samples. Age composition is used for evaluation of the effects of fishing, estimation of recruitment, growth, and mortality. Therefore the major sources of sampling variation should be identified. The effects of annual, seasonal, lati- tudinal, and sexual variation are considered in this section. Annual Variation Annual variation in age composition was studied using samples collected off the Washing- ton coast during May-September 1965-69. It has been assumed (Tillman 1968; Nelson and Larkins 1970) that Pacific hake are fully recruited to the fishery at age 5. If this assumption is valid and if recruitment and mortality rates are constant then one would expect, from a relatively unexploited population, a typical catch curve with the 5-yr-olds most numerous and the succeeding ages decreas- ing at a rate equal to the rate of natural mortality. This pattern was apparent in 1965 (Figure 5). The partially recruited 4-yr-olds were not as numerous as the 5-yr-olds which predominated. From age 5 there was a progressive decrease in the relative abundance of succeeding age groups until only a few 13-yr-olds remained. By examining the age compositions in subsequent years, it became ob- vious that the 4-yr-old age group in 1965 was con- siderably larger than usual. This was the first in- dication that the 1961 year class was unusually large. In 1966 the 1961 year class (5-yr-olds) was probably fully recruited and strongly dominated the age structure. The relative abundance of the incoming 1962 year class (4-yr-olds) was much smaller than the 1961 year class in the 1965 samples. The 1961 year class can be followed through the age composition in 1967-69. In 1969 the 1961 year class lost its dominance to the 1962 year class, but still produced extraordinarily large numbers of 8-yr-olds. Apparently the 1965-66 year classes were smaller than those observed previously since in the 1969 sample no 4-yr-olds were observed. Obviously annual variation in age composition does occur in Pacific hake and is at least partly due to varying levels of recruitment of incoming year classes. Seasonal Variation In 1966-67 regular sampling of Pacific hake from off the southern coast (lat. 46°00'-46°59'N) of 342 DARK: AGE AND GROWTH OF PACIFIC HAKE Washington occurred throughout the fishing season. These samples were combined as "early" (collected in May-June), "middle" (July), and "late" (August-October) samples for the purpose of identifying any gross seasonal changes in age composition (Figure 6). The 1966 early sample contains somewhat fewer 5-yr-olds and more 6- yr-olds than either the middle or late samples which are very similar. For all practical purposes, the age compositions of the three 1967 samples are identical. There is little evidence in these com- parisons to suggest that there is significant change in age composition during the time the Pacific hake are present in commercial quantities off southern Washington. The paucity of samples appropriate for further comparisons precludes as- sessment of seasonal age composition variation which may occur at other times and places. Latitudinal Variation Variation occurs in the age composition of Pacific hake samples taken from different por- tions of the latitudinal distribution (Nelson 1967; Tillman 1968). Table 1 is adapted from Tillman's table 14 to show the percentage age composition < 60 50 40 30 20 10 0 90 80 ^ 70 o S 60 50 40 30 20 lOh 0 966 967 Summer — Early -- Middle — Late —I 1 1 1 1 — 4 5 6 7 8 9 AGE (years) 10 II 12 Figure 6.-Age composition of Pacific hake collected off southern Washington (lat. 46°00'- 46°59'N) by early, middle, and late summer periods. 343 FISHERY BULLETIN: VOL. 73, NO. 2 Table L— Age composition (percentage) of hake taicen off Washington, Oregon, and California in 1965 by research vessels (adapted from Tillman 1968). Washington Oregon California Age Sexes Sexes Sexes (years) Male Female combined Male Female combined Male Female combined 1 — — _ — — 0.2 1.5 0.6 2 — — — — — — 1.8 4.0 2.4 3 — — — — — — 3.8 6.9 4.7 4 13.0 22.5 17.7 26.8 46.4 37.2 44.4 38.2 42.6 5 28.9 20.6 24.7 37.0 27.6 32.0 29.8 25.1 28.4 6 27.2 23.8 25.1 22.8 11.9 17.0 8.1 10.6 8.8 7 13.5 5.2 8.5 4.1 4.6 4.4 5.1 5.5 5.2 8 12.9 10.6 11.5 9.1 4.3 6.6 3.5 5.1 4.0 9 3.0 10.3 7.4 0.1 3.9 2.1 2.7 2.5 2.6 10 1.4 4.9 3.5 — 0.6 0.3 0.4 0.6 0.4 11 — 1.4 0.9 — 0.5 0.3 0.1 — 0.1 12 — 0.6 0.4 — — — 0.1 — 0.0 13 — - - - 0.1 0.1 - - - Sample size 1,474 2,196 3,670 810 914 1,724 1,586 670 2,256 50r — California — Oregon — Washington T 1 1 1 1 I 4 5 6 7 8 9 AGE (years) Figure l.—Age composition of Pacific hake collected off California, Oregon, and Washington in 1965. for research vessel samples taken off Washington, Oregon, and California in 1965. These samples were used because they provided the best geographic coverage and were taken with similar trawl gear, equipped with small-mesh cod end liners capable of capturing juvenile hake. The age composition of the California sample probably does not truly reflect the relative abundance of the age groups. The southern distribution and abun- dance of adult hake was a primary consideration at the time of sampling and, because the very young hake normally are not found associated with the adults, they probably were not taken in proportion to their abundance. Figure 7 shows, however, that samples taken off California included 1-, 2-, and 3-yr-olds, which did not occur in Washington and Oregon samples. Washington and Oregon samples composed of fish 4-yr-old and older were considered to be representative of the population in those areas. There is a smaller per- 344 DARK: AGE AND GROWTH OF PACIFIC HAKE centage of 4-yr-olds and a greater percentage of 5- and 6-yr-olds in the Oregon sample than in the California sample. Similarly, the Washington sample contains a much smaller percentage of 4- yr-olds and a greater percentage of nearly all the older age groups than does the Oregon sample. It cannot be clearly demonstrated whether there are in fact fewer 4-yr-olds in the hake population off Washington or whether additional older specimens are recruited, depressing the relative abundance of the 4-yr-olds. However, since 3-yr- old hake are not recruited to the Washington fishery and it is difficult to rationalize the sudden occurrence of additional large numbers of older, adult fish not found off Oregon and California, the most likely event is that the 4-yr-olds are only partially recruited off Washington and are not as numerous as they are off Oregon. In May, July, and October 1965 and in August 1966, samples were available from off the Washington coast from lat. 46°00'N to 48°59'N. For each year the samples were grouped by half degree intervals of latitude and compared in Figure 8 to determine if changes in age composi- tion among smaller spatial units than used in the foregoing discussion could be detected. The 1965 samples were collected during a period of several months but, as indicated in a previous section, seasonal (spring-fall) variation should not be a significant factor. Although there is considerable random variation in the relative abundance of an age group among areas, there appear to be some LATITUDE 46»00;- 46''29' 50 r 40 46»30'- 47''00'- 48" 00'- 48-30' 46»59' 47''29' 48° 29' 48''59' 30 20- '^ 10 CD < 2 0 LiJ o 60 tr UJ Q. 50 40 - 30 20 10 1965 ta^ a PI- LL e 10 4 t^--^ 6 e 10 4 6 Ul_ 1966 rm , 1 8 10 4 6 8 10 4 6 e 10 AGE (years) Figure 8.-Age composition of Pacific hake taken at various latitudes off Washing- ton in 1965-66. 345 FISHERY BULLETIN; VOL. 73, NO. 2 trends associated with latitude. Generally, the relative abundance of 4- and 5-yr-olds tended to decrease as sampling progressed from south to north while the relative abundance of 7- to 10-yr- olds tended to increase. This is consistent with obsen'ations over a larger sampling area. In summary, 1-, 2-, and 3-yr-old Pacific hake are rarely encountered north of California and, as sampling progresses northward along the Washington coast, the younger age groups (4- and 5-yr-olds) contribute less to the population while the relative abundance of the older age groups (7- to 10-yr-olds) increases. Such latitudinal stratification of age groups, apparently occurring to some extent even within 3 degrees of latitude, compounds the problem of representatively sampling the age composition of not only the entire hake population, but also the com.mercially available portion of the population. This spatial variation should be considered when comparing annual changes in age composition. Sexual Variation Pacific hake samples taken off Washington in 1965-69 were used to examine the variation in age composition between male and female com- ponents. The percentage of each sex by age group was calculated (Table 2) and plotted in Figure 9. The age compositions for males and females are similar in all years. The greatest difference oc- curred in 1965 when females contributed a larger proportion of the 4-yr-old group. In all years, a greater percentage of the females occurred in the older age groups (8 yr and older). These results correspond well with earlier observations (Best 1963; Tillman 1968) that females live longer and may be the sole survivors by age 11 and 12. Therefore sampling that is highly selective for sex would provide biased estimates of the relative abundance of older age groups. GROWTH OF PACIFIC HAKE Growth in Length To determine the general shape of the growth curve for Pacific hake, average lengths-at-age by sex were computed using combined data from samples taken off Washington, Oregon, and California during 1965-69 (Table 3). Growth of both sexes is quite rapid during the first 3 yr, then slows abruptly (Figure 10). Deceleration of the growth rate probably is not as pronounced as indicated. Latitudinal stratification of ages discussed previously is probably a result of stratification by size; therefore, mainly the larger members of the younger age groups would be recruited to the Washington-Oregon area where most samples were collected. Best (1963), for ins- tance, computed the average lengths for small samples of hake (age groups 4-13 consisted of females only) taken off northern California and the 3- to 6-yr-olds were somewhat shorter than those in the Washington-Oregon samples (Figure 10). The most accurate growth curve for 2- to 6- yr-olds would probably fall somewhere between that based on Best's (1963) data and the curves generated from my data. Asymptotic growth has been shown for Cape hake, M. capeiisis (Botha 1969); silver hake (Fritz 1962); and Pacific hake in Puget Sound (see footnote 3). For the purpose of this study it is assumed that the growth curve for the length of coastal Pacific hake is also asympto- tic. Sexual differences in the size at age occur in most hake species (Hickling 1933; Hart 1948; Fritz 1962; Botha 1969), and the Pacific hake is no ex- ception (Figure 10). Females are noticeably longer by age 4, but may be longer even at an earlier age. Larkins et al. (see footnote 3) reported that female Pacific hake of the Puget Sound population are slightly longer than males even as 1-yr-olds. This cannot be demonstrated for coastal hake because sex information was not available for 1-yr-olds and so few 2- to 3-yr-olds were collected that one cannot accept the estimated mean lengths by sex with confidence. Although sexual differences in growth exist, these differences are not large. The maximum difference in mean lengths is 3.12 cm occurring at 11 yr of age (Table 3). Year class variation with respect to growth was examined by two means: 1) the comparison of age-length data collected in 1965-69, and 2) the back-calculation and comparison of growth rates for five year classes. Mean body lengths by age for the 1956-62 year classes are found in Table 4 and compared in Figure 11. Only samples taken off the Washington coast were used so that successive year classes could be compared while minimizing any spatial effects. These growth curves generally fall into two sets. Individuals of the 1956-59 year classes were considerably larger at age than the members of the 1961-()4 year classes. The growth curve for the 1960 year class falls between the two sets. 346 DARK: AGE AND GROWTH OF PACIFIC HAKE Table 2.-Age composition (percentage) of male and female hake taken each year off Washington, 1965-69. Age 1965 1966 1967 968 1969 (years) Male Female Male Female Male Female Male Female Male Female 2 _ _ 0.1 — 0.1 0.1 3 — — 0.1 0.1 1.4 1.2 — — — — 4 13.0 22.5 0.3 0.3 1.6 2.6 1.4 2.5 0.1 0.2 5 28.9 20.6 44.2 46.0 5.2 4.5 1.9 2.9 6.9 9.0 6 27.2 23.8 33.1 23.7 62.1 53.6 11.5 7.0 16.6 18.5 7 13.5 5.2 14.5 15.9 22.2 19.7 68.6 60.3 55.7 41.3 8 12.9 10.6 3.1 6.7 5.4 10.2 13.3 17.2 19.7 28.2 9 3.0 10.3 3.2 3.3 1.2 4.8 2.3 4.8 1.0 2.5 10 1.4 4.9 0.7 1.8 0.6 2.2 0.9 3.2 — 0.1 11 — 1,4 0.7 1.5 0.1 0.9 0.2 1.4 — 0.1 12 — 0.6 — 0.5 0.1 0.2 — 0.4 — — 13 — - - 0.1 - - - 0.3 - - Sample size 1,474 2,196 1,355 1,724 1,195 1,432 1,047 1,546 1,009 811 Figure 9.-Age composition of male and female Pacific hake taken off Washington, 1965-69. 347 FISHERY BULLETIN: VOL. 73, NO. 2 Table 3.-Average body length at various ages for male and female hake taken off California, Oregon, and Washington dur- ing 1965-69. Female Male Age Sample Mean Sample Mean (years) size length (cm) size length (cm) 1.0 385 '15.40 385 '15.40 2.0 36 28.03 28 26.93 3.3 17 41.18 13 42.23 4.3 135 46.20 83 44.59 5.3 750 48.23 628 47.63 6.3 1,073 50.26 1,134 49.67 7.3 1,459 51.82 1,761 50.87 8.3 626 54.27 432 52.30 9.3 199 56.98 93 54.77 10.3 97 58.93 21 56.43 11.3 44 59.00 8 55.88 12.3 11 60,91 — — 13.3 6 ^4,453 61.83 - — 24,201 'Assigned value based on the mean of all 1-yr-old hake; sex determinations were not available for 1-yr-olds. 2Does not include the unsexed 1-yr-old group While there is little variation apparent within year class groups, length at age may vary by as much as 4 cm betv^een year class groups. These data suggest that some variation in growth can occur among year classes, but the irregular sampling of all ages, particularly the youngest, precludes construction of complete growth curves by year class using length-at-age data. Therefore a back-calculation technique was utilized to reconstruct the growth curves for several year classes as a means of further examining year-class variation. Back calculation of body lengths was based on an otolith radius-body length relationship derived from hake samples collected in 1966-68. Approximately 10 otoliths (5 male and 5 female) per 1-cm body length interval were selected for 70r 60 50 E o X I- Z UJ o 40 - 30- 20 10 . '- — Moles California, Oregon a Washington samples combined. California sample (ages 4-13 females only) Females 2 3.3 4.3 5.3 6.3 7.3 8.3 9.3 10.3 IL3 12.3 13.3 AGE (years) Figure lO.-Average fork lengths at various ages for Pacific hake collected off California, Oregon, and Washington combined, and from California alone (Best 1963). Table 4.-Mean body length (cm) at various ages (in years) for 1956-62 year classes of hake. 1956 1957 1958 1959 1960 1961 1962 Age Length Age Length Age Length Age Length Age Length Age Length Age Length — — _ — — — — _ _ _ _ _ 3.0 35.6 — — — — — — — — — — 4.3 44.9 4.5 47.8 — — — — — — 5.7 52.5 5.3 48.9 5.5 48.2 5.3 48.7 — — — — 6.7 53.9 6.3 52.6 6.5 51.6 6.3 49.2 6.3 49.7 — — 7.7 54.2 7.3 53.9 7.5 54.8 7.3 53.2 7.3 50.6 7.5 50.6 8.7 54.9 8.3 55.2 8.5 56.1 8.3 55.9 8.3 54.1 8.5 51.8 — — 9.3 57.4 9.5 56.8 9.3 56.5 9.3 56.6 9.5 54.2 — — — — 10.5 57.8 10.3 59.7 10,3 58.2 10.5 57.0 — — — — — — 11.3 58.7 11.3 58.0 11.5 60.0 — — — — — — — — 12.3 60.0 - - - - - - - - - - - - 348 DARK: AGE AND GROWTH OF PACIFIC HAKE 62 60 58 56 E 54 u ""' 52 I o 50 z tu 48 _j 46 §441- CD 421- 40 38 361- (1958) (1964)/ <• (1963) '/ (1956) — I— 3 — 1 1 1 1 1 1 1 1 1 5 6 7 8 9 10 II 12 13 AGE ( years ) Figure 11. -Average fork lengths at various ages by year class (in parentheses) for Pacific hake taken off Washington, 1964-69. analysis. A sample of 10 was not attained in those centimeter groups near the extremes of the length distribution nor in those length groups where 10 samples were taken but not all otoliths were readable. Fork lengths represented ranged from 11 to 68 cm. To facilitate the measuring all otoliths were photographed and enlargements made so that the prints were 16 times the size of the otoliths. Measurements were taken directly from each print. A midpoint was determined on the photograph of each otolith by measuring the dis- tance from the anterior edge to the posterior edge of the first translucent zone and halving that measurement (Figure 12). Measurements were made from this midpoint to the anterior margin of each annulus and to the anterior edge of the otolith. When annuli were not clearly defined, measurements were not made. The mean body length per centimeter of otolith radius was computed and plotted in Figure 13. A curve constructed from the individual observa- tions (n = 370) is superimposed on the means. The curve was constructed from a third degree polynomial equation of the form Y = 18.78957 - 3.79065X + 0.67490Z^ - 0.01836X\ where Y = es- timated body length and X = otolith radius. This equation provided the best fit (smallest residual sum of squares and mean square) of the several functions examined (Table 5). The correlation Anterior edge 7 th annulus 6th onnulus 5th annulus 4 th annulus 3rd annulus 2 nd onnulus I St onnulus Midpoint 2nd year radius, etc. 1st year radius Figure 12.-Pacific hake otolith showing the calculated midpoint and measurement intervals from the midpoint to the anterior margins of successive annuli. 349 FISHERY BULLETIN: VOL. 73, NO. 2 70r 60 t 50 I t- o 40 z UJ > o o m 30 20 10 40 60 80 100 120 140 OTOLITH RADIUS ( mm) 160 180 Figure 13.-A scattergram of mean body lengths per millimeter of otolith radius (as taken from a 16 x photographic enlargement) and superimposed curve of the function used to back-calculate growth. some evidence in Table 7 that the radii of the first one or two annual marks increase as the total age increases, only otoliths from 7-yr-olds were used to avoid any possible age-related effects. Older specimens could not be used because they were not available in sufficient numbers and younger specimens had a growth history which was too brief. Fifteen otoliths of each sex for each of the 5 yr were selected. The desired sample size (30 per year class) was seldom obtained because many of the otoliths had deteriorated in storage so that opaque and hyaline zones could not be distin- guished on the photographs. Back-calculated lengths are presented in Table 8, and a comparison of back-calculated growth curves for the five year classes is made in Figure 15. It appears that year class variation was not great among the 1957-60 year classes, but that the members of the 1961 year class were on the average noticeably smaller at age than members of the other year classes. This latter observation Table 5.-Functions examined for the best fit of otolith radius-body length data, corresponding residual sums of squares, and mean squares. Function Residual sum of squares Mean square y = a + bx 3,865 10.44 y = a + fax + cx2 3,700 10.03 y = a + bx + cx2 + dx^ 3,146 8.55 y ^= ab' 7,611 20.57 y =: e^ xb 3,768 10.18 between estimated and observed values is 0.9860, so an improved fit was not attempted. The Y in- tercept is at 18.78957 cm indicating that the func- tion does not adequately represent the otolith radius-body length relationship in fish less than 1 yr of age. Therefore the back calculation of body lengths beyond the range of the data fitted is ob- viously not meaningful. The mean total otolith radii by age were used to back calculate lengths at age which are presented in Table 6 and compared in Figure 14 with ob- served lengths at age from combined 1965-69 samples. The atypical 1961 year class was excluded in this comparison. In spite of a certain amount of variation in the back-calculated curve (probably induced by the small sample size), the two curves correspond very well. The lengths-at-age for five year classes (1957-61) were compared by back calculating the lengths of approximately equal numbers of males and females from each year class. Because there is Table 6.-Mean otolith radii and back-calculated body lengths at various ages for Pacific hake. Age (years) 1 2 3 4 5 6 7 8 9 10 11 12 13 Mean otolith radii (cm) Calculated body lengths (cm) 6.08 9.06 12.53 13.50 13.60 14.43 14.80 16.44 15.79 17.29 17.74 18.30 17.90 16.57 26.19 41.13 45.44 45.88 49.45 51.00 57.30 54.92 60.11 61.44 62.92 61.88 Figure 14.-A comparison of growth curves as constructed from observed and backcalculated fork lengths at various ages. 350 DARK: AGE AND GROWTH OF PACIFIC HAKE Table 7.-Mean radii (cm) of photographed otolith annuli by total age group. Annuli Total age (years) 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th 13th 1 5.96 — _ _ _ _ _ _ _ 2 5.85 8.87 — — _ _ _ 3 7.10 10.13 12.53 — _ — _ 4 6.34 9.73 11.66 13.13 — _ — 5 6.57 10.03 11.50 12.87 13.57 — _ _ 6 6.38 9.73 11.51 12.79 13.79 14.32 — — _ 7 6.16 9.59 11.60 12.97 13.89 14.37 14.71 — — _ — 8 6.18 9.86 12.02 13.63 14.90 15.61 16.15 16.50 — 9 6.33 9.80 11.56 12.89 14.05 14.53 14.94 15.39 15.66 — 10 6.44 9.92 11.65 13.39 14.57 15.66 16.21 16.73 17.14 17.50 — 11 6.60 9.84 1 1 .38 13.08 14.42 15.52 16.00 16.54 16.94 17.36 17.66 — 12 6.75 10.85 12.75 14.50 15.50 16.10 16.60 17.60 17.90 18.10 18.40 18.60 — 13 7.15 10.45 12.20 13.40 14.35 14.95 15.35 15.75 16.20 16.55 16.90 17.25 17.65 Table 8.— Back-calculated body length (cm) at various ages for the 1957-61 year classes of Pacific hake. Year class and (in parei ntheses) sample size Age 1957 1958 1959 1960 1961 (years) (27) (15) (22) (27) (30) 1 19.02 18.64 16.45 16.85 16.95 2 31.16 31.01 30.36 29.92 28.50 3 38.71 39.05 38.65 37.11 37.06 4 44.61 40.66 46.09 46.12 43.58 5 49.31 50.14 50.92 50.38 47.60 6 52.28 50.98 53.74 52.65 49.88 7 54.10 54.61 55.51 54.20 51.48 corroborates the slow growth rate of the 1961 year class suggested by the length-age data presented previously. The 1961 year class was an extremely large year class numerically, exceeding by far the size of other year classes of record. Perhaps den- sity-dependent growth was operative in this ins- tance. The size of the 1961 year class may have been so large that the competition for food and space noticeably restricted the growth of in- dividuals. It cannot be ascertained whether the members of smaller year classes might undergo density-dependent growth to a lesser extent or if the phenomenon is triggered only by unusually large year classes. The von Bertalanffy growth equation is com- monly used to describe asymptotic growth. It is used herein because it fits the data well and is readily incorporated in certain yield models. Von Bertalanffy's equation is /, =/oo 1-e^^' <'-'o))^ where 1, = body length at time t;l oo = estimated average maximum body length; k= rate of growth; t^ — theoretical age when growth conforms to the von Bertalanffy equation. Since sexual differences in growth characteristics exist in Pacific hake, separate curves were fitted to male and female length data. Average length at age data from Ta- ble 3 were used to compute the growth curves. In this instance the growth of the 1961 year class was considered atypical and it was excluded from the analysis. A computer program' utilizing the method of Stevens (1951) was used to compute the constants for the von Bertalanffy equation. The resulting equations are: /, = 56.29 (l-e-o-39 ('-0.20)) for males and /, = 61.23 (l-e-o-^oc-o.oi)) for females. By comparison, Tillman's (1968) estimates of A: were 0.41 for males and 0.19 for females. He reported that treatment of Best's (1963) data also yielded 0.19 for females. In the present study /oo for both sexes are reasonable estimates of average maximum body lengths. Males as long as 66 cm and females as long as 80 cm have been observed. Growth of Pacific hake is adequately represented by the von Bertalanffy equations as the curves fit the observed lengths-at-age very well (Figure 16). The curves are nearly superimposed until about age 5 when they begin to diverge and continue to do so with age. Although sex-specific growth rates are apparent, the growth rate of the entire Pacific hake population may be best represented by an equation based on data with the sexes combined. The growth equation with sexes combined is /, — 60.85 (1-e -0-30(^-003)). Growth in Weight In 1964-69 length and weight data were collect- ed on 2,417 male and 3,117 female Pacific hake taken from Washington to southern California. Lengths and weights were taken from fresh fish as they were unloaded from vessels at the processing plant and from research samples at sea. Fork lengths were taken to the nearest centimeter and weights to the nearest decagram. The majority of 'Program developed by George Hirschhorn, Northwest Fisheries Center, National Marine Fisheries Service, NOAA, Seattle, Wash. 351 FISHERY BULLETIN: VOL. 73, NO. 2 60,- e40 30 20 Figure 15.— Back-calculated growth curves for 1957-61 year classes. (Only otoliths from 7-yr- old specimens were used.) 10 3.3 AGE 43 ( years ! 5.3 6.3 7.3 observations were made from mature specimens in commercial samples, and therefore hake less than about 40 cm long were inadequately Femoie represented or not represented at all. The com- mercial data were fitted using the length-weight equation W = aL^, where W = weight in grams, L = length in centimeters, and a and b are cons- tants. In linear form this equation becomes log W = log a + b (log L). The length-weight rela- tionships were calculated to be log W = -1.45990 -I- 2.55618 log L for males and log W = -1.68944 -I- 2.69509 log L for females (Figure 17). Typically the exponent in the length-weight equation for most fusiform fishes approximates 3, implying isometric growth. On the basis of data 3.3 4.3 5.3 6.3 73 8 3 9.3 103 11.3 12.3 13.3 AGE ( yeors ) Figure 16. -Von Bertalanffy growth curves for male and female Pacific hake superimposed on mean body lengths at various ages, (o = females; + = males.) 352 DARK: AGE AND GROWTH OF PACIFIC HAKE 2,000 r 1,800 1.600 1,400 1.200 5 1,000 800 600 400 200 o Female log W = -1.68944 + 2.69509 log L + Mole logW=-l. 45990 + 2.55618 log L Female Male — I — 20 — I — 30 40 50 LENGTH (cm) 60 70 Figure 17.-Length-weight curves for male and female Pacific hake superimposed on mean body weights (n > 5). from 58 female hake from California ports, Best (1963) estimated the exponent of the length- weight equation to be 3.0668. On the chance that the hake in commercial samples had lost weight between capture and delivery through the loss of body fluids, another pair of equations was fit to less extensive length-weight data taken aboard research vessels off Washington from freshly caught specimens. Student's ^-test was used to test the null hypothesis that regression slopes cal- culated from research and commercial samples did not differ significantly from 3. All tests were sig- nificant at the 1% level (Table 9) and the hypothesis was rejected. A possible explanation is that because most specimens are in an immediate postspawning state as they arrive off Washington and Oregon, their weight relative to their length is less than it might be later in the year. Growth in weight for males and females was calculated from the length-weight equations (Figure 18). These curves suggest that Pacific hake gain weight at an increasing rate until they are 3 yr old. After age 3 the rate of growth in weight decreases and remains roughly constant until death. By age 3, males have grown to approxima- tely 50% of their total weight at 11 yr of age, whereas females by age 3 have attained about 40% of their total weight at 11 yr of age. The growth rates are sex specific and the curves begin to diverge noticeably between ages 3 and 4. At 11 yr of age females weigh on the average about 200 g Table 9.— Results of f-tests to determine if slopes of length- weight regressions calculated from commercial and research samples differ significantly from 3. All comparisons were sig- nificant at the 1% level. Sample type Sum of deviations from mean (2x2) Variance (Sxy2) Slope {b) Sample size t Commercial: Male Female Research: Male Female 6.4064 10.5651 2.6224 4.2177 0.0049 0.0032 0.0154 0.0075 2.55618 2.69509 2.63189 2.65436 2,417 3,117 432 587 16.02 17.62 4.80 8.20 353 FISHERY BULLETIN: VOL. 73, NO. 2 Figure 18.-Average body weight at various ages for Pacific hake collected off California, Oregon, and Washington, 1965-69. more than males. For each sex, Table 10 gives average weight at age values, annual grovv^th rates, and annual instantaneous growth rates, as- suming exponential growth in weight. SUMMARY AND CONCLUSIONS Biological data from samples of Pacific hake taken in 1964-69 off the coasts of Washington, Oregon, and California were utilized to study the age and growth of the species. Table lO.-Mean weight, annual increase in weight, and instan- taneous growth rate at various ages for male and female hake taken off California, Oregon, and Washington, 1964-69. Male Female Body Annual Instant. Body Annual Instant. Age weight increase growth weight increase growth (years) (g) (%) rate (g) (%) rate 1 37.6 32.4 318 1.43 400 1.61 2 157.0 162.9 216 1.15 183 1.04 3.3 496.0 459.5 15 0.14 36 0.31 4.3 570.0 626.5 18 0.17 13 0.12 5.3 674.6 703.5 12 0.11 12 0.11 6.3 751.0 786.2 6 0.06 8 0.08 7.3 798.2 853.6 7 0.07 13 0.12 8.3 856.8 966.8 13 0.12 14 0.13 9.3 964.1 1,102.5 8 0.08 9 0.09 10.3 1,040.6 1,207.2 0 0.00 0.3 0.00 11.3 1,014.8 1,211.1 — — 8 0.08 12.3 - 1,319.6 — — 4 0.04 13.3 — 1,374.0 — — — — The method of age determination from annuli on otoliths was e)tamined, and all evidence sug- gests that the method provided reliable age data. Several sources of variation in the age structure of the population were considered. The relative size of newly recruited year classes varied sub- stantially, creating noticeably annual variation in the age composition. There was no detectable seasonal variation in the age composition of the hake population found off Washington from spring through fall. Latitudinal stratification of hake by age (known to occur over large geographical areas) was further examined, and some variation in age composition was found even among 1/2 degree intervals of latitude off the Washington coast. The relative abundance of the 4- and 5-yr-olds decreased as sampling progressed northward from the mouth of the Columbia River to the Strait of Juan de Fuca, while the relative abundance of 7- to 10-yr-olds increased. This strat- ification of ages by latitude supports the theory that there is a northward migration of hake in the early spring along the Pacific coast of North America with the older (larger) individuals tend- ing to migrate farthest. There is little variation in age composition due to sex, except that the longer lived females tend to predominate from 8 to 10 yr of age and are usually the sole survivors at 11-13 yr of age. Pacific hake grow rapidly in length during their first 3 yr after which growth slows and becomes asymptotic. At about 4 yr of age, females grow noticeably faster and by age 11 may average 3.12 cm longer than males. Individual males may reach 66 cm, while some females may reach 80 cm in length. Year class variation in growth rates was detected by analysis of age-length data and back calculation of growth from otoliths. The equation Y = 18.78957 - 3.79065A' + 0.67490A^ - 0.01836A' was used to describe the relationship of body length (7) and otolith radius (X). The extraordinarily large 1961 year class grew at a substantially slower rate than the 1957-60 year classes. This difference possibly is indicative of density-dependent growth. Growth in length can be expressed adequately by the von Ber- talanffy growth equations: If = 56.29 ( 1-^ -0.39 (/ -0.20)) for males, I, = 61.23 (l-(-'''^«"-ooi') for females, and /, - 60.85 (l-e -0-30 ((-0.03)) f^^ ^^^ gg^es combined. 354 DARK: AGE AND GROWTH OF PACIFIC HAKE Growth in weight was determined by applying the length-weight equations: log W = -1.45990 + 2.55618 log L for males, and log W = -1.68944 + 2.69509 log L for females to average length-at-age data. By age 3, males have grown to about 50% of their total weight at age 11, and females to about 40% of their total weight at age 11. At 11 yr of age females weigh on the average 200 g more than males. Males attain an average weight of about 1,015 g by age 11 and females reach an average weight of 1,374 g by 13 yr of age. ACKNOWLEDGMENTS The author wishes to express his gratitude to George Hirschhorn, John J. LaLanne, and Michael F. Tillman for their valuable assistance in the preparation of this paper and to Sandra Guilbert, Ruth Mandapat, and Robert Loghry for providing age determinations, otolith measurements, and associated information essential to the study. LITERATURE CITED Ahlstrom, E. H., and R. C. Counts. 1955. Eggs and larvae of the Pacific hake Merluccius productus. U.S. Fish Wildl. Serv., Fish. Bull. 56:295-329. Alverson, D. L., and H. A. Larkins. 1969. Status of knowledge of the Pacific hake resource. Calif. Coop. Oceanic Fish. Invest., Rep. 13, p. 24-31. Alverson, D. L., A. T. Pruter, and L. L. Ronholt. 1964. A study of demersal fishes and fisheries of the northeastern Pacific Ocean. H. R. MacMillan Lectures in Fisheries, Inst. Fish., Univ. B.C., 190 p. Bagenal, T. B. 1954. The growth rate of the hake, Merluccius merluccius (L.), in the Clyde and other Scottish sea areas. J. Mar. Biol. Assoc. U.K. 33:69-95. Best, E. A. 1963. Contribution to the biology of the Pacific hake, Merluccius productus (Ayres). Calif. Coop. Oceanic Fish. Invest, Rep. 9, p. 51-56. BiGELOW, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Ser\'., Fish. Bull. 53, 577 p. BiRTWISTLE, W., AND H. M. LEWIS. 1925. Hake investigations. Lancashire Sea-Fish. Lab. Rep. 33, p. 36-56. Botha, L. 1969. The growth of the Cape hake Merluccius capensis. S. Afr. Div. Sea Fish., Invest. Rep. 82, 9 p. Dark, T. A., H. H. Shippen, and K. D. Waldron. 1970. Pacific ocean perch & hake studied off west coast. Commer. Fish. Rev. 32(3):25-30. Fritz, R. L. 1962. Silver hake. U.S. Fish Wildl. Serv., Fish. Leafl. 538, 7 p. Graham, M., and J. R. Lumby. 1929. Studies of age-determination in fish. Part I. A study of the growth-rate of codling (Gadus callarias L.) on the inner herring-trawling ground. G. B. Minist. Agric, Fish. Food, Fish. Invest. Ser. 2, 11(2), 50 p. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can., Bull. 180, 740 p. Hart, T. J. 1948. The distribution and biology of hake. Biol. Rev. (Camb.) 23:62-80. Hickling, C. F. 1933. The natural history of the hake. Part IV. Age deter- minations and the growth rate. G. B. Minist. Agric, Fish. Food, Fish. Invest. Ser. 2, 13(2), 120 p. Johnson, L. J., and W. L. High. 1970. Midwater trawling equipment and fishing technique for capturing hake off the coast of Washington and Oregon. U.S. Fish Wildl. Serv., Circ. 332:77-101. Morgan, A. R., and D. E. Gates. 1961. A cooperative study of shrimp and incidental fish catches taken in shrimp fishing gear in California and Oregon, 1958. Pac. Mar. Fish. Comm., Bull. 5, p. 85-106. Nelson, M. 0. 1967. Availability of Pacific hake (Merluccius productus) related to the harvesting process. FAO (Food Agric. Or- gan. U.N.) conference on fish behavior in relation to fishing techniques and tactics, Bergen, Norway, 19-27/10/67, Experience Pap. FR:FB/67/E/34, 26 p. 1970. Pacific hake fishery in Washington and Oregon coastal waters. U.S. Fish Wildl. Serv., Circ. 332:43-52. Nelson, M. 0., and H. A. Larkins. 1970. Distribution and biology of Pacific hake: A synop- sis. U.S. Fish Wildl. Serv., Circ. 332:23-33. Pereyra, W. T., and J. A. Richards. 1970. Economic aspects of the 1967 offshore Pacific hake fishery. U.S. Fish Wildl. Serv., Circ. 332:103-119. Petersen, C. G. J. 1895. Eine Methode zur Bestimmung des Alters und Wuchses der Fische. Mitth. Deutsch. Seefisch.-Ver. 11:226-235. Stevens, W. L. 1951. Asymptotic regression. Biometrics 7:247-267. Tillman, M. F. 1968. Tentative recommendations for management of the coastal fishery for Pacific hake, Merluccius productus (Ayres), based on a simulation study of the effects of fishing upon a virgin population. M.S. Thesis, Univ. Washington, Seattle, 197 p. Utter, F. M., and H. 0. Hodgins. 1971. Biochemical polymorphisms in the Pacific hake (Merluccius productus). Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer. 161:87-89. 355 EVALUATION OF THE RETURN OF ADULT CHINOOK SALMON TO THE ABERNATHY INCUBATION CHANNEL Allan E. Thomas' ABSTRACT Adult returns of progeny of the 1964 year class of chinook salmon, Oncorhynchus tshawytscha, were determined for the Abernathy incubation channel, natural production, and hatchery sources. A total of 4,620,600 fry were released from the channel into Abernathy Creek (state of Washington) as unmarked fry. Natural production in the creek was estimated from spawning ground counts and fyke net sampling of migrants at 16,700 fry, or 0.36% of the total unmarked fish. A total of 557,649 hatchery fish were marked by feeding tetracycline, and 161,579 were both fin-clipped and fed tetracycline. All 2-, 3-, and 4-yr-old adult fish returning to the hatchery holding pond were examined for fin clips and fluores- cent bands on the vertebrae. Returns from hatchery sources were 506 fish or 0.070%. Potential egg production was 865,000, or 76% of the original 1,142,604 eggs. Returns from the channel totaled 733 fish or 0.016%. Egg potential was 2,050,000, or 35% of the original 5,888,048 eggs. Returns attributed to Abernathy Creek totaled only three fish and egg potential was 0.2% of the original 2,880,000 eggs. Survival of this year class was considerably below the 9-yr average of 0.118%. Intuitively, costs favor incubation channel production over hatchery production. However, additional studies are needed to determine if contributions to the fisheries and survivals are comparable in years of better year class survival. Spawning and incubation channels for salmon produce higher survivals of downstream migrants than do natural spawning grounds. Studies by Gangmark and Broad (1956), Lister and Walker (1966), Thomas and Shelton (1968), and others have shown that channels with flow and sediment con- trol devices can produce many times as many migrant-sized fish as are produced by the parent stream. Little information is available on the per- centage of fry produced in channels that return as adults. Such data are needed to determine the value of incubation channels for supporting and maintaining salmon runs. In the present study, I compare the survival to the adult stage of chinook salmon, Oncorhynchus tshawytscha, from three sources: the Abernathy incubation channel (near Longview, Wash.); the Salmon-Cultural Laboratory hatchery adjacent to the channel; and natural spawning in Abernathy Creek. MATERIALS AND METHODS Problems were encountered in selecting the techniques to use to distinguish returning adults 'U.S. Fish and Wildlife Service, Fish Farming Experimental Station, Stuttgart, AR 72160. from the various sources. Fin-clipping presented two problems. First, fin-clipping of fry produced in the channel was undesirable mainly because of the small size of the fry (average weight, <0.5 g). The chance of injury or incomplete removal of fins would be high. Fin-clipping of hatchery fish has been shown to reduce adult returns by 43.3% (Weber and Wahle 1969). Second, the daily numbers of fry migrating varied widely (from a few thousand to over 120,000), depending upon water conditions. Fin-clipping at a practical rate would require that many fish be held beyond their normal migration date. Marking channel and creek migrants with tet- racycline would not be acceptable. Although marking fish in this manner using techniques developed by Weber and Ridgway (1962) had no detrimental effect on adult survival (Weber and Wahle 1969), the drug is normally administered in the food. The artificial feeding of channel fish might alter their characteristics and invalidate the survival evaluations. Attempts to mark small fish by immersion in solutions of tetracycline and 4% dimethyl sulfoxide have shown some favorable results (Richard C. Johnsen, pers. commun.), but the technique has not been suflficiently developed for reliable, large-scale marking. For these reasons migrant fry from the incuba- tion channel and Abernathy Creek were not Manuscript accepted July 1974. FISHERY BULLETIN:VOL. 73, NO. 2, 1975. 356 THOMAS: EVALUATION OF RETURN OF CHINOOK SALMON marked. Rather, the numbers migrating down- stream were determined and the returns were compared with returns from marked hatchery fish stocked in the creek. It is assumed that fish migrating from the channel and Abernathy Creek would have the same rates of survival. Table 1 summarizes the number, size, time of migration, and mark of each of the fish sources involved in the evaluation. Table L-Summary of the number, size, time of migration, and mark of the various fish sources included in the evaluation. Origin of Size Time of migrants Number (g/fish) migration Mark Abernathy Creek 16,700 '0.5 111/64-4/65 none Incubation channel 4,620,600 0.5 12/64-4/65 none Hatchery 61,000 8.2 5/65 2-OTC2 445,500 4.8 5/65 2-OTC 51,149 16.8 8/65 3-OTC 161,579 16.8 8/65 3-OTC + fin clips 'Approximate size and time of migration based upon fyke net sampling. ^Oxytetracycline. Migrants from the Incubation Channel A total of 4,620,600 unmarked fry were released from the channel during the 1964-65 season. This number represents a 78.5% survival from the 5,888,100 eyed eggs planted. If the mortality of green eggs is also considered, the survival to migrant stage was 75.0%. A description of tech- niques for planting eggs and counting fry was given by Thomas and Shelton (1968). Migrants from Abernathy Creek Numbers of female chinook salmon which spawned in Abernathy Creek during the 1964 spawning season were estimated at 576 from spawning ground counts. On the basis of an average of 5,000 eggs per female, an estimated 2,880,000 eggs were deposited in the creek. Fry migrating downstream from this area were sampled with a fyke net, a method demonstrated to be reliable by Tait et al (1962). Recaptures from known numbers of marked fish released upstream indicated that the net sampled 13.3% of the migrating fish; the calculated survival of migrants from natural spawning was 16,700, or 0.58%. Flooding and superimposition of eggs during spawning are probably responsible for this low survival. These fish were unmarked and returning adults could not be distinguished from those which originated in the incubation channel. Chinook salmon spawning area in Abernathy Creek extends for about 1 mile below the hatchery weir. Returns from natural spawners would not necessarily enter the hatchery holding pond. Releases of Hatchery Fingerlings All fish released from the Salmon-Cultural Laboratory hatchery were marked in some manner for later identification. Initially, all fish were marked in mid-April 1965 by feeding tet- racycline after the technique developed by Weber and Ridgway (1962). All fish in a 100-fish sample examined in late April showed fluorescent bands on their vertebrae. Fish released into Abernathy Creek in May had two bands and those in August, three bands. These second and third marks were poor because the fish fed little, presumably because of the high level of drugs in the diet. Total hatchery fish released with only tetracycline marking totaled 557,649. Two groups of fish used in a nutrition experiment received double fin clips, in addition to the tetracycline marks. These two groups served as controls in that the returning adults would be easily recognizable from the clipped fins, and a check of their vertebrae would indicate the per- sistence of the fluorescence. The fin-clipped groups, which numbered 161,579, were released in mid- August 1965. Treatment of Adult Fish All adult chinook salmon returning to the hold- ing pond at the Salmon-Cultural Laboratory were examined during 1966-68. All fish returning as 2- yr-old jack salmon and as 3- and 4-yr-old adults were examined. Past records of age classification indicated that numbers of 5- and 6-yr-old adults in the Abernathy Creek spawning run are insig- nificant. Fish that returned to Abernathy Creek but not to the hatchery holding pond were not included in the evaluations. No attempt was made to sample the sport or commercial fisheries or to search the adjacent streams and hatcheries for strays. The evaluation is based only upon returns to the Salmon-Cultural Laboratory holding pond. No correction was made for strays from other hatcheries or streams that might enter the holding pond and be counted as survivors from the channel since their number would be insignificant (Worlund et al. 1969). All adult fish were measured and the sex was 357 recorded. Scales were taken for age determina- tion. Age determinations of fish returning to the hatchery in previous years had indicated the size range for 2-yr-old jack salmon, but 3- and 4-yr-old fish could not be separated by size. During the 1966 spawning season, a vertebra was removed from all fish smaller than the maximum length previously found for 2-yr-old fish, for determination of tet- racycline marks. During the 1967 and 1968 spawning seasons, a vertebra was also removed from all fish in the size range of previous 3- and 4-yr-old fish. Vertebrae were scanned under ultraviolet light by the technique described by Weber and Ridgway (1962). Vertebrae with tetracycline marks were classified as hatchery fish. The age of fish without vertebra marks was determined by examination of the scales. All unmarked fish of the correct age were considered to be recoveries from either the channel or creek. Tetracycline bands were visible in the vertebrae of all fin-clipped salmon. Adult returns from two sources— the hatchery or Abernathy incubation channel and Abernathy Creek— were determined by measuring and aging fish and by identifying those with fin clips and tetracycline marks on vertabrae. Potential egg production was calculated on the basis of 5,000 eggs per female, the average number found in Chinook salmon returning to the hatchery in previous years. RESULTS FISHERY BULLETIN: VOL. 73, NO. 2 only 20 returned. Disease and parasite problems encountered during the summer rearing season probably contributed to the low survival. Fish released before the warm-water season appeared to have better survival. The total survival was 0.070%. Adult Returns from Abernathy Incubation Channel and Creek The percentage returns of fish from the Aber- nathy incubation channel (0.016%) was considera- bly lower than that from the hatchery (Table 3). The higher ratio of female fish, however, resulted in a relatively higher number of eggs per return- ing fish. The numbers of returning fish that originated in Abernathy Creek were insignificant. If the sur- vival rate after migration is assumed to be iden- tical for channel and creek migrants, only 0.36% of the 736 returning unmarked adults— or about 3 fish— were from the creek. Table 3.-Number of male and female adult chinook salmon re- turns from 4,620,600 fry released at the Abernathy incubation channel, 1965. Age at return (years) Males Females Potential egg production (thousands) 2 3 4 Total 16 202 107 325 0 220 188 408 0 1,100 940 2,040 Adult Returns from Hatchery- Reared Fingerlings Table 2 presents adult return data for both 1) the fin-clipped plus tetracycline-marked group and 2) the tetracycline-marked groups. Survival of fish that had been marked with both tetracycline and fin clip was low; of 161,579 fingerlings released, Table 2.-Number of male and female adult chinook salmon re- turns from 719,228' hatchery-reared fingerlings released from the Salmon-Cultural Laboratory hatchery, 1965. Age at return (years) Males Females Potential egg production (thousands) 2 3 4 Total 209 90 34 333 0 96 77 173 0 480 385 865 'Of this total, 557,649 fish were released with tetracycline bands and 161,579 with double fin clips and tetracycline bands. DISCUSSION Survival was low from chinook salmon of the 1964 year class released from most Columbia River hatcheries, as they were in this experiment. Reasons for the poor survival are largely un- known. Although the evaluation of survival from several year classes would have been desirable, comparisons of survival from the different sources of young fish provide information on the relative survival of channel-reared and hatchery-reared fish, as well as an insight into the potential sur- vival from an incubation channel. Ideally, sufficient adults should return to a salmon hatchery to provide 100% or more of the original egg supply. Assuming about 5,000 eggs per female and a 50:50 ratio of males to females, a chinook hatchery is self-sustaining when the return is about 0.045%. 358 THOMAS: EVALUATION OF RETURN OF CHINOOK SALMON Table 4.-Potential egg production of adult chinook salmon returns in relation to original egg numbers from which they were produced. Return of adult salmon Salmon eggs Source of downs! rpam Oriqlnal P°'ential production from adults number Number Percent of (thousands) (thousands) original number migrants Number Percent Salmon-Cultural Laboratory hatchery Abernathy incubation channel Abernathy Creek 506 0.070 773 0.016 3 0.016 1,143 5,888 2,880 865 76 2,040 35 7 0.2 Potential egg replacement of adult returns from the various sources is shown in Table 4. The adult return from the hatchery totaled 0.070%; however, more than 40% of the adult fish returned as 2-yr- old males. Only about 76% of the original eggs were replaced for this brood year. Although the percentage survival of hatchery fish was more than four times that of channel fish, the advantage in egg replacement was less than twofold. Even so, a channel could not operate long with a less than 35% replacement of eggs. If adults from natural spawning are assumed to enter the holding pond at an equal ratio with those from the channel, possible egg replacement from the creek source was only about 7,000. The return of chinook salmon as 2-, 3-, and 4-yr- old, and sometimes as 5- and 6-yr-old fish, ensures overlapping of brood stocks. Consequently returns from year classes with poor survival are mixed with returns from other year classes, to help en- sure that the hatcheries and streams may still have adequate egg supplies. Production of salmon by incubation channels is usually evaluated only by the number of out- migrants produced and the total number of adults that return. The origins and ages of adults are seldom investigated. Only relative comparisons can be made of sur- vival from channel and hatchery releases due to variables such as time of release, size at release, and nutritional background. The channel had the sizable advantages of having very low costs for rearing the fry and no cost for food. A 9-yr average of adult returns to the hatchery was 0.118%. Survival to adults from the 1963 year class of fish released by the hatchery was about 0.39%. There were no major differences in fish diet, times of stocking, or other known factors that might improve survival of this year class over that of the 1964 year class. If it is assumed that sur- vivals of fish from the channel follow the same trends as do those from the hatchery, the more than fivefold greater hatchery returns for the 1963 year class would be reflected in channel production returns. This would provide more than sufficient egg production replacement. However, additional studies are needed to confirm this survival as- sumption. ACKNOWLEDGMENT I thank the staff of the Salmon-Cultural Laboratory for their help in conducting this experiment, and especially Laurie G. Fowler and Joe L. Banks, for their help in identifying marked salmon vertebrae and the Oregon Fish Commis- sion's mark processing center which made the age analysis from collected scales. LITERATURE CITED Gangmark, H. a., and R. D. Broad. 1956. Further observations on stream survival of king salm- on spawn. Calif. Fish Game 42:37-49. Lister, D. B., and C. E. Walker. 1966. The effect of flow control on freshwater survival of chum, coho and chinook salmon in the Big Qualicum River. Can. Fish Cult. 37:3-25. TAIT, H. D., J. L. HOUT, AND F. V. Thorsteinson. 1962. An evaluation of fyke trapping as a means of indexing salmon escapements in turbid streams. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 428, 18 p. Thomas, A. E., and J. M. Shelton. 1968. Operation of Abernathy Channel for incubation of salmon eggs. U.S. Bur. Sport Fish. Wildl., Tech. Pap. 23, 19 p. Weber, D. D., and G. J. Ridgway. 1962. The deposition of tetracycline drugs in bones and scales of fish and its possible use for marking. Prog. Fish-Cult. 24:150-155. Weber, D., and R. J. Wahle. 1969. Effect of finclipping on survival of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Board Can. 26:1263-1271. Worlund, D. D., R. J. Wahle, and P. D. Zimmer. 1969. Contribution of Columbia River hatcheries to harvest of fall chinook salmon (Oncorhynchia^ tshawytscha). U.S. Fish Wildl. Serv., Fish. Bull. 67:361-391. 359 ECOLOGICAL STUDIES OF THE PUERULUS LARVAL STAGE OF THE CALIFORNIA SPINY LOBSTER, PANULIRUS INTERRUPTUS' Steven A. Serfung and Richard F. Ford' ABSTRACT Ecological and related behavioral studies of the puerulus larval stage of the California spiny lobster, Panulirus interruptua, involved the development and use of artificial and natural seaweed habitat traps, special paired neuston nets, and underwater night-lights for collecting and observing pueruli in nature. The results obtained indicate that pueruli first enter the coastal waters off San Diego, Calif., during May, and continue to appear regularly through September, but with no apparent relationship to lunar or temperature cycles. Pueruli exhibit a strong attraction to floating habitat traps containing the surfgrass Phyllogpadix torreyi and to bright lights at night. The results also suggest that the puerulus is a surface-swimming, pelagic form which may actively seek out specific nearshore areas for settlement, and thereby serve the important function of returning larval stages to areas suitable for demersal life of the young juveniles. Previous studies of the California spiny lobster, Panulirus interruptus (Randall), and other spiny lobster species have concentrated primarily on either the adult or phyllosoma larval stages, leav- ing the biology of the intermediate, yet very im- portant, puerulus and juvenile stages relatively unknown. This is particularly true for P. in- terruptus. Prior to the present study, essentially nothing was known about the ecological requirements or behavior of the puerulus or early juvenile stages, which together represent a period of 2-3 yr in the life history of this species. This is an unfortunate situation for an animal as heavily exploited by a commercial fishery as the California spiny lobster because, as Thorson (1950) has in- dicated, the abundance of any adult population is primarily dependent on the recruitment, survival, and growth of its larval and juvenile stages. Con- sequently, attempts to improve fishery yields through a better understanding of adult behavior and population ecology alone provide only a partial and temporary solution to the problem. Increasing fishing pressure on the steadily declining stocks of P. interruptus and other spiny lobster species urges more than an academic interest in the ecological requirements of their puerulus and juvenile stages. On the basis of this 'Contribution No. 3 from the San Diego State University Center for Marine Studies. 'Center for Marine Studies, San Diego State University, San Diego, CA 92182. information, for example, it may be possible to protect their natural habitats or to develop supplementary artificial habitats for them in na- ture. Techniques for culturing these stages under artificial conditions, as a means of supplementing natural stocks, must also be given serious con- sideration for the following reasons. Evidence concerning P. interruptus (Johnson 1956, 1960, 1971) and other palinurid species (see, for example, Chittleborough and Thomas 1969) suggests that a majority of the phyllosoma and puerulus larvae may be lost from the population due to their long larval life (5-10 mo), during which they may be swept hundreds of kilometers offshore. Because of their small size, the surviving postpuerulus and early juvenile stages probably experience much higher predation mortality rates than do the adults (Winget 1968). In addition, evidence from limited studies of other spiny lob- ster species, including P. argus, P. longipes cygnus, and Jasus edwardsii, suggests that the nursery grounds of their postpuerulus and early juvenile stages are located in protected bays or estuaries (Sheard 1949; Lewis et al. 1952; Witham et al. 1964, 1968; Sweat 1968; C. B. Kensler, pers. commun.). If such estuarine nursery grounds are required by the early benthic stages of P. interruptus, reduc- tion of this type of natural habitat as the result of commercial developments and water pollution may create a "weak link" in the life history of this species, at least within part of its geographical range. Manu.script accepted .June 1974. FISHERY BULLETIN; VOL. 73, NO. 2, 1975. 360 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS Aquaculture of P. interruptus or other palinurid lobsters, starting with the egg, has been attempt- ed by several investigators as a means of bypass- ing what appears to be an inefficient recruitment process in nature. Yet all attempts to culture the phyllosoma larval stages from the egg through to the puerulus stage have been unsuccessful (for example, see Dexter 1972). This approach does not appear to be feasible at the present time, par- ticularly on a mass culture basis, due to the long duration (5-10 mo in nature) and the poorly un- derstood requirements of the delicate phyllosoma larval stages. In contrast, the succeeding puerulus and juvenile stages of spiny lobsters respond well to laboratory culture (Kensler 1967; Witham et al. 1968; Serfling and Ford in press a, b). Thus, if the puerulus stage of P. Interruptus proves to be abundant and relatively easy to collect in large numbers, particularly in certain offshore waters where these individuals might otherwise be swept away from coastal nursery grounds, it may be ad- vantageous to supplement natural stocks by collecting pueruli and holding them in mass cul- ture under controlled, optimal conditions through at least the early juvenile stages. These pueruli could either be reared directly to a marketable size under artificial conditions, or returned to the ocean, after passing the potentially critical early juvenile stages, for final growth and maturation. This approach to spiny lobster culture has also been proposed for P. argus by Ingle and Witham (1969). The lack of ecological information on the puerulus stage of any spiny lobster species is due largely to the fact that investigators usually have been unable to collect more than a few individuals, despite extensive sampling in areas where they might be expected to occur. Consequently, the central question precluding further research has concerned the mode of life of the puerulus stage, i.e., whether it is primarily a benthic or pelagic form. Evidence presented in the literature thus far supports about equally each side of the issue. Investigators who have failed to collect more than a few isolated pueruli, even after years of exten- sive sampling by conventional methods (Lewis et al. 1952; Johnson 1956, 1960, 1971; Harada 1957; Saisho 1966; Sims and Ingle 1966; Lazarus 1967), have concluded that the puerulus stage of their respective species is either primarily benthic, or concentrated in unknown pelagic areas. On the other hand, Gurney (1942), Sheard (1949), George and Cawthorn (1962), Chittleborough and Thomas (1969), and Chittleborough (1970) were able to collect small numbers of individuals in net hauls far out to sea, and thus suggested that the puerulus is a free-swimming stage. Harada (1957) and Johnson (1960) reported collecting a few pueruli which were attracted to bright lights over shallow water at night, but both authors apparently believed they were lured from the bottom. Yet Lindberg (1955) reported that University of Hawaii personnel collected several pueruli of a Panulirus species which were at- tracted to surface night-lights in water several hundred meters deep, thereby eliminating the possibility that they were attracted from the bot- tom. Witham et al. (1968), Sweat (1968), PhiUips (1972), and we have successfully attracted large numbers of pueruli to artificial habitat traps floating at the surface, thereby offering the first strong evidence that the puerulus is primarily a pelagic, surface-dwelling form. Preliminary observations made originally by J. C. Van 01st (pers. commun.) in the spring of 1968, demonstrated that puerulus larvae of P. in- terruptus occurring in local coastal waters could be attracted to bright lights suspended under- water near the surface at night. This discovery prompted the following investigation into several aspects of puerulus behavior and ecology. The specific objectives of this study were: 1) to deter- mine the mode of life of the puerulus stage of this species, e.g., whether it is primarily benthic or pelagic; and, having done so, 2) to develop and apply suitable field sampling techniques to study the general dynamics of puerulus recruitment with regard to seasonal, lunar, and daily periodicity, area of settlement, specific habitat preferences, and migratory behavior; and 3) to es- timate the general abundance and possibility of collecting large numbers of this stage for use in aquaculture. Closely related studies were also conducted concurrently to investigate the habitat preferences and the natural growth rates of the early juvenile stages in nature, as well as the growth and survival of the juvenile stages at elevated temperatures in recirculating culture systems (Serfling 1972; Serfling and Ford in press a,b). MATERIALS AND METHODS The failure of standard sampling methods used by previous investigators to collect the puerulus 361 FISHERY BULLETIN: VOL. 73, NO. 2 stage prompted the development of novel equip- ment and techniques. These included underwater night-lights of high intensity, floating artificial and natural seaweed habitat traps, and a special paired neuston net, which are described in the following sections. Underwater Night-Lights Most of the night-light observations and collec- tions were made from an adjustable ladder plat- form near the end of the Scripps Institution pier, at the position shown in Figure 1, under a variety of environmental conditions. A standard motion picture projection lamp of either 500 or 1,000 W was waterproofed at the electrical connection with silicone sealant. The light was mounted on a pole and lowered 30-50 cm below the surface. This unit illuminated a spherical area 5-8 m in diameter, depending on water turbidity. When suspended in this way it provided much greater illumination than when suspended just above the surface. Pueruli attracted to the light were removed with a dip net and maintained alive for further studies. PACIFIC OCEAN HABITAT POSITIONS -SHORE UNLIGHTEO ^'V-' '■■»- T< y SCRIPPS INST PIER ■300M- NIGHT-LIGHT -^ OBSERVATION AREA AREA ILLUMINATED' BY PIER LIGHTS Figure l.-Standard positions of puerulus habitat traps placed on the Scripps Institution pier, and the position from which night-light observations and collections of pueruli were made. several models of the "Witham" habitat design used by Witham et al. (1968) for collecting P. argus pueruli (Figure 2C). This trap was constructed of a synthetic fibrous material labeled "Conservation Web #200," manufactured by Minnesota Manufacturing and Mining Company.' The traps we developed and tested consisted of wood frames which contained either synthetic materials, e.g., burlap, foam rubber, nylon mop, and plastic screen and netting, or fresh natural seaweeds, held within either plastic cage sides (Figure 2A, B), or within net bags or wire screen (Figure 2D, E). Seaweed Habitat Traps Pueruli placed in aquaria quickly settled and remained hidden in a variety of intertidal rock and plant substrates, but showed a preference for the surf grass Phyllospadix torreyi. They would con- tinue to cling to any of these substrates, even when removed from the water. In an attempt to take advantage of this behavior, a variety of con- tainers (Figure 2) for holding Phyllospadix and various species of red algae were floated under the lighted end of the Scripps Institution pier at the positions shown in Figure 1. The pier lights (three 200 W incandescent flood lamps mounted approximately 15 m above the water) apparently attracted pueruli to the pier, and the seaweed habitat traps provided convenient refuges in which they settled. The traps were retrieved and examined for pueruli every 4-6 days, providing a standard sampling system which could be operat- ed continuously. The same types of habitat traps were also maintained on buoyed lines at several coastal locations in water 3-30 m deep, as discussed in a later section. A variety of natural and artificial seaweed habitat trap designs, shown in Figure 2, were tested initially to develop a type most suitable for attracting pueruli. Included in the evaluation were ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. WITHAM DOUBLE FRAME SINGLE FRAME SCRKKN Figure 2.— Floating habitat trap designs employed in sampling puerulus larvae of Panulirus interruptus. 362 [ XLi/l/VU-Ti UO Suspension of the Habitat Traps The strong surf and current conditions prevailing around the pier pilings required the development of a special system for suspending the habitat traps, as shown in Figure 3. This sys- tem allowed the trap to rise and fall during the tidal cycle, which has a vertical range of approximately 2.5 m, and in response to wave mo- tion, thereby always remaining at the surface. The counterweight maintained a taut line and counteracted drag on the habitat trap caused by currents. No bottom anchor was necessary, thus eliminating problems of retrieval and fouling on kelp, and allowing the traps to be raised and lowered easily for examination. The traps also were held away from the pier pilings by the pole extension in order to prevent damage and en- tanglement. Paired Neuston Nets The results of night-lighting operations, dis- cussed in a later section, suggested that the puerulus stage of P. interruptus swims within a few centimeters of the ocean surface. Thus, plankton or other nets towed below or near the surface probably would fail to catch this stage. Even surface tows made with a net breaking the surface might not be successful if it were trailed directly behind a vessel, because the "snow plow" action of the vessel might effectively push surface organisms away from the path of the net. Thus, paired neuston nets which could be mounted laterally on a small boat were developed to allow unobstructed surface water sampling, as shown in Figure 4. The presumed rarity of the puerulus stage required that these nets be of relatively large mesh size (5.0 mm) to allow reasonably fast towing speeds with a small boat (approximately 3 knots) and the filtration of large volumes of water. The short length and lateral position of the nets allowed their cod ends to be removed and emptied periodically without hauling the entire net from the water. The nets were towed with approximate- ly 30% of the frontal area above water to ensure continual filtration of the top few centimeters of surface water, even during periods of low surface waves. The average frontal net area maintained submerged was approximately 0.7 square meter per net. Only rough estimates were made of the volumes and surface areas of water filtered during initial trials conducted during this study, as these Figure 3.-The method employed in suspending habitat traps from the Scripps Institution pier. were primarily concerned with evaluation of the sampler and with qualitative rather than quanti- tative results. The estimates of the water volume and surface area sampled were based on average boat speed and duration of the tow. RESULTS Night-Light Observations Night-light observations were first conducted briefly in May 1968, and then resumed for a 2-mo period in September 1968. Night-light observa- tions were also conducted again in the spring and summer of 1969, primarily to supplement infor- mation from the habitat traps also maintained during this period. The results of these observa- tions, together with pertinent environmental da- ta, are summarized in Table 1. During calm eve- nings in September 1968, pueruli were collected from the Scripps Institution pier at an average rate of approximately 4 per hour, with the highest rate reaching 12 pueruli collected within 90 min (8 per hour). No pueruli appeared during night- lighting conducted in October 1968, and further attempts were curtailed due to the onset of rough winter surf conditions which made the operation difficult. Night-lighting activities also were con- ducted at other San Diego localities during the period September-October 1968. These were: Quivira Basin in outer Mission Bay, Shelter Island in outer San Diego Bay, and the Imperial Beach pier. No pueruli were observed at any of these 363 I'lbMJ'^Kr BULLtTlM: \VL. li,NV.Z Figure 4.-Paired neuston nets employed in sampling the pueruius larval stage of Panulirus interruptus, showing the method used to deploy them from a small boat. Table l.-Results of night-light sampling for pueruli at the Scripps Institution pier and other San Diego localities. Date Number collected Location Time of sampling Moon phase Water temp. (°C) Sampling period (min) Suitability of conditions' 1968: May 15 0 Mission Bay 2000-2130 Full 16.0 45 Good May 20 1 Scripps 2000-2130 Last Quarter 15.7 50 Fair May 25 0 Mission Bay 1930-2030 New 16.4 60 Good Aug. 28 9 Scripps 2130-2250 First Quarter 18.6 60 Good Sept. 4 0 Imperial Beacti 2030-2200 Full 20.7 80 Good Sept. 7 4 Scripps 2200-2330 Full 20.0 40 Fair Sept. 10 12 Scripps 2100-2230 Full-Last Quarter 20.5 90 . Excellent Sept. 16 3 Scripps 2000-2045 Last Quarter 20.7 40 Good Sept. 26 4 Scripps 2000-2100 New— First Quarter 19.5 40 Fair Oct. 2 0 Scripps 2230-2330 First Quarter— Full 18.3 50 Good Oct. 10 0 Shelter Id. 2130-2250 Full— Last Quarter 18.5 80 Good Oct. 20 0 Scripps 2200-2300 New 17.3 40 Good Oct. 30 0 Scripps 2330-2430 First Quarter 16.8 30 Fair 1969: Apr. 5 0 Scripps 2030-2130 Full-Last Quarter 14.8 15 Poor Apr. 20 0 Scripps 2130-2250 New— First Quarter 15.5 50 Good May 7 0 Scripps 2030-2130 Last Quarter 15.9 20 Poor May 15 0 Scripps 2130-2230 New 16.2 60 Good May 23 6 Scripps 2145-2345 First Quarter 17.3 90 Good June 4 3 Scripps 2200-2300 Last Quarter 18.1 30 Fair July 10 14 Scripps 2100-2330 Last Quarter— New 20.5 90 Excellent Aug. 7 6 Scripps 2130-2250 Last Quarter 21.5 70 Good Sept. 12 4 Scripps 2210-2250 New 18.5 60 Fair Sept. 22 0 Scripps 2030-2120 Full 18.0 20 Fair Oct. 11 0 Scripps 2100-2230 New 18.0 50 Good Oct. 20 0 Scripps 2000-2130 First Quarter 17.2 30 Fair 'The suitability of conditions for collecting pueruli was determined subjectively from genera velocity, wave height, and water clarity during the collecting period. I observations of current direction and 364 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS localities and work there was discontinued. Night-lighting observations during 1969 at the Scripps Institution pier provided a more complete picture of the apparent seasonal occurrence of pueruli, which first appeared in late May and were not observed after 4 September 1969. These initial night-light studies allowed several interesting new observations on puerulus behavior. Previous investigators who reported that pueruli were occasionally attracted to night- lights (Harada 1957; Johnson 1960) presumed that the pueruli were lured off the bottom from rocky habitats where they had settled. However, this apparently was not the situation at the Scripps Institution pier for the following reasons. The natural bottom substrate adjacent to the Scripps Institution pier for a radius of at least 500 m is entirely surf-washed sand. There is a small intertidal and subtidal rocky area approximately 500 m to the north, and an extensive rocky shoreline begins approximately 2 km to the south. While the pier pilings might afford a suitable habitat, we observed no pueruli or juveniles in several careful daytime and nighttime examina- tion of the pilings, using scuba, during periods when there was active puerulus settlement in the habitat traps. In addition, direct observations of the free-swimming pueruli indicated that many individuals approached the light from a direction opposite that of the pilings. Perhaps most sig- nificant was the observation that the pueruli were seen swimming in only the top few centimeters of water; no individuals were ever seen approaching the light from a greater depth. This opportunity to observe the free-swimming puerulus stage of P. interruptus thus demon- strated that, at least under these conditions, it is a surface-dwelling organism which occurs only in the top few centimeters of water. Consequently, the inconsistency or failure of standard plankton net tows to collect pueruli of this and other species probably is not due to the presumed benthic habits of this stage, but rather to improper sampling techniques. Only nets extending above the sur- face, and streamed parallel to the vessel's course with an unobstructed path, appear to be suitable. The likelihood of net avoidance accounting for previous failures seems minimal, as the pueruli of this species are relatively slow moving and easily captured in a hand held dip net, at least near a night-light. Puerulus swimming speed in nature was measured by timing the passage of an individual between two lines suspended from a 2 m long, horizontally oriented pole, as it approached the underwater light. The mean swimming speed of six different pueruli under these conditions was 8 cm/s (range 6 to 9 cm/s). Individuals appeared to maintain this speed continuously, unless dis- turbed. In response to disturbance, they would spread their antennae and legs, sink slowly 5-20 cm, and then resume normal surface cruising with antennae held together and legs withdrawn. Illustrations of these swimming and sinking pos- tures are shown in Figure 5. SINKING POSTURE SWIMMING POSTURE Figure 5.— Sinking and swimming postures of the puerulus lar- val stage of Panulirus interruptus as observed under a night- light. Puerulus Pigmentation in Relation to Date of Settlement Newly settled pueruli, or those collected during night-lighting, were always completely trans- parent. Pueruli held in open and closed system aquaria provided with Phyllospadix or other in- tertidal flora and fauna almost immediately began acquiring pigmentation, and this continued at a rapid rate. At temperatures of 18-20°C, pueruli would change from complete transparency to nearly complete pigmentation within 9-10 days, and would then moult into the first postpuerulus stage. This occurred consistently among over 50 in- dividuals we held in aquaria and observed systematically. Thus, the degree of pigmentation provided a reliable method of determining, within 1-2 days, the date of settlement. This technique 365 FISHERY BULLETIN: VOL. 73, NO. 2 provided a more accurate means of evaluating possible correlations between puerulus recruit- ment and changing environmental factors than by merely recording the date on which the habitat trap was examined, as done by Witham et al. (1968) and Sweat (1968). In his later study, Parker (1972) provided a more detailed description of the pigmentation and transformation of the puerulus and postpuerulus stages of P. interruptus. Laboratory observations indicated that once the puerulus settled in a suitable substrate, its trans- formation into a postpuerulus proceeded without delay. However, a few individuals which were held in containers without fresh Phyllospadix, algae, or rocks containing epifauna, remained as trans- parent pueruli for 2-3 wk until death, suggesting strongly that the presence of such suitable habitat features is necessary to induce transformation to the postpuerulus stage. After settlement, pueruli became photonegative and were never again seen swimming. In order to test the related question of whether or not a settled puerulus might desert a habitat trap upon its removal from the water, and thus produce unreliable trap catch data, one seaweed frame habitat trap and one Witham trap containing known numbers of pueruli were returned to the water for 5-10 min on two different occasions, and then retrieved. No individuals left either the Witham or seaweed traps. This, and the laboratory observations described above, indicate that the number of individuals found in a habitat trap probably is a reliable estimate of the number that settled there. Comparison of Habitat Trap Types Habitat traps were mounted either singly or in groups of two or three from one main support bar in order that at least six different types could be tested simultaneously from the three available Scripps Institution pier positions (position A, B, or C, as shown in Figure 1). The exact pier position and combinations of traps were alternated from time to time, as indicated in Table 2, in order to compensate for possible variations in unknown factors influencing puerulus settlement, such as current direction, eddies near the pilings, and differences in daytime or nighttime light intensi- ties. The results of comparative habitat trap design studies conducted during the summer of 1969 are summarized in Table 2. These results in- dicate that all habitat traps containing natural seaweed proved markedly superior to the Witham artificial habitat trap for this species (Figure 2C). Only 5 pueruli were collected by two Witham habitat traps {x = 2.5 per trap), while 97 pueruli were collected by four seaweed habitats (x = 24.2 Table 2.-Results of a comparison of the catch effectiveness of habitat traps suspended off the Scripps Institution pier. The catches of pueruli per trap are also recorded in terms of the specific location of the trap at the pier (a, b, or c). Positions A and B were lighted, while position C was unlighted, as shown in Figure 1. For descriptions of the habitat trap designs, see Figure 2 and text. raps ned i9 Duration of sampling period (days) Number of pueruli collected in fiabitat traps Date t Seaweed fiabitat traps Witfiam #1 habitat traps #2 exami 196 Double frame Single frame Bag Screen Total collected June 14 6 _i 2a — — Oa — 2 20 6 — — — — — — — 27 7 — — — — Oa — 0 July 2 4 0 7a — — 1 a — 8 5 3 3b 1 a — — Oa — 4 9 4 3 b 1 a — — Oa — 4 13 4 2 b 1 a — — Oa — 3 17 4 3 b Oa — — Oa — 3 23 6 3b 1 a — — 1 b — 5 28 5 1 b 1 a — — Ob — 2 Aug. 4 6 23 b 2a _ — 1 b _ 26 6 2 2b 1 a — — 0 b — 3 10 4 1 a 4b 0 c Oc Oc Ob 5 16 6 3a 8b 1 c Oc 0 c 0 b 12 27 11 2a 12b 1 c 1 c Oc 1 b 17 Sept. 4 8 1 a 1 b 0 c Oc 0 c 1 b 3 12 8 Oc Oc 2b 1 b Oc Ob 3 19 7 Oc 1 c 1 b 0 b 0 c Ob 2 27 8 Oc Oc Ob Ob 0 c Ob 0 Oct. 3 6 Oc Oc Ob Ob 0 c Ob 0 Total 47 43 5 2 3 2 102 'Habitat trap not in position at tfiis time. 366 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS per trap), when all were in continuous use during the same 90-day time period from June through September 1969 (Table 2). Comparison of these catch figures by a chi-square test for equality in- dicates that the mean catch of the seaweed traps was significantly higher than that of the Witham traps (P < 0.05). The Witham habitat trap was not effective, either when new or moderately fouled with a variety of sessile organisms, yet the natural seaweed habitat traps appeared to catch well regardless of the condition of the plant material. Variations in the abundance of pueruli throughout the summer and differences in settlement in the different habitat positions, as discussed in subsequent sections, did not allow specific comparisons to be made between results of the habitat trap designs listed in Table 2. Yet general comparisons indicate that there were no major differences in catch between these trap designs that are not attributable to the environ- mental causes discussed in subsequent sections below. However, the nylon bag habitat trap (Figure 2E) proved to be the best in terms of cost, ease of use, and durability. It also appeared to catch as well as the other types, although tests were begun too late in the final puerulus settlement season during which we sampled to verify this. Thus, the bag trap design is recom- mended for future studies of this nature, and was used by Parker (1972) in a later study. Habitat Trap Success in Relation to Position on the Pier A comparison of the catch results of the "dou- ble" and "single" frame seaweed habitat traps, as shown in Table 2, indicates that more pueruli were consistently collected from the B than the A posi- tion on the Scripps Institution pier (Figure 1), when either type of trap was placed there. A com- bined total of 65 larvae was collected in the B position, versus only 24 in the A position by these two traps during the same 82-day period from 2 July to 19 September. A comparison of these catch figures by a chi-square test for equality indicates that the catch at position B was significantly greater than at position A (P < 0.05). No explanation for this difference is apparent at the present time, but it seems to indicate that there may be subtle environmental effects which influence puerulus settlement in habitat traps to a greater extent than do variations in trap design. Significance of Nocturnal Illumination of Traps in the Attraction of Pueruli The strongly positive phototactic response exhibited by the puerulus stage during night- lighting observations suggested that nocturnal illumination may play an important role in the successful operation of habitat traps. To evaluate this, additional traps were maintained in an unlighted area of the Scripps Institution pier (area C in Figure 1). The results, summarized in Table 3, indicate that during the period from 10 August to 29 Sep- tember 1969, when each trap design was main- tained at both the lighted (positions A and B) and nonlighted (position C) pier locations, all traps caught more larvae in the lighted positions. A total of 38 larvae were collected over 132 "trap-days" (1 trap in place for a 24-h period = 1 trap-day) in the lighted position, versus only 4 per 132 trap-days in the nonlighted position. Comparison of these total catch values by means of a chi-square test for equality indicates that the value for the lighted positions was significantly greater than for the unlighted one {P < 0.05). This suggests that noc- Table 3.-A comparison of the number of pueruli caught in lighted and nonlighted habitat traps. For detailed descriptions of trap types, and lighted and nonlighted pier positions, see text and Figures 1 and 2. Lighted (positions A and B) Nonlighted (position C) Type of habitat trap No. individuals collected per trap Sampling period ' (days)' Catch per 20 trap-days No. individuals collected per trap Sampling period (days)' 'Number of trap days during the period of 10 August to 29 September only (see Table 1). ^Habitat trap not in place. Catch per 20 trap-days Double frame 7 29 4.8 0 15 0 Single frame 25 29 17.2 1 15 1.32 Bag 3 15 4.0 2 29 1.38 Screen 1 15 1.32 1 29 0.68 Witham "a" _2 - - 0 44 0 Witham "b" 2 44 0.48 — — Mean catch per 20 trap-days 5.4 0.68 367 FISHERY BULLETIN: VOL. 73, NO. 2 turnal illumination plays a major role in the suc- cessful use of these habitat traps for pueruli of P. interrupt us. The fact that no pueruli were caught in the unlighted habitat traps maintained offshore, as discussed in the next section, provides additional strong support for this conclusion. Habitat Traps Maintained Offshore Unlighted habitat traps of the "bag" and "screen" designs were anchored in water 3- to 15- m deep in the San Diego area off Point Loma, La Jolla, and off the northeast shore of Catalina Island, Calif., at the positions shown in Figure 6a, b. This was done to monitor the recruitment of pueruli at other locations along the San Diego coastline, and on the leeward side of a large offshore island, as well as to determine the im- portance of lights in attracting pueruli by com- paring these unlighted traps with those placed under the Scripps Institution pier lights. The traps placed offshore were checked and refilled with fresh Phyllospadix and red algae every 10-15 days during August and September 1969. Many were lost, apparently because of entanglement with kelp and boat propellers. However, the large number maintained successfully failed to collect any pueruli, despite the fact that puerulus settlement at the Scripps Institution pier was relatively high during this same period. The four traps maintained at Catalina Island were situated directly above a shallow rocky area, approximately 2-6 m in depth, where we had observed many small first year juvenile lobsters the previous winter. We presumed that the presence of these young juveniles indicated that this was an area of high puerulus recruitment. Thus, failure of the unlighted traps to collect pueruli at this location was particularly surprising. The failure of these offshore habitat traps may have been due not only to the lack of nocturnal illumination, but also to the presence of large quantities of Phyllospadix and algal flotsam in the areas where they were maintained. The probability of a puerulus settling in a seaweed habitat trap under such conditions would be small, considering the large volume of seemingly equally suitable patches of floating surfgrass and algae present in these areas. In contrast, relatively few masses of plant flotsam of this kind were observed in the area around the Scripps Institution pier. In an attempt to evaluate this, many clumps of floating Phyllospadix were collected and examined nr — f ^^^. Bird Rock Isthmus Raef (a) (b) SCRIPPS INSTITUTION PlER LA JOLLA PACIFIC OCEAN MISSION BAY MISSION BAY CHANNEL Figure 6.-Locations of sampling areas off northeastern Ca- talina Island (a) and San Diego (b) for puerulus and early juvenile Panulirus interruptus. for pueruli while deploying and checking the offshore habitat traps. No pueruli were found in this manner, but only a relatively small percen- tage of this material was examined relative to that present, thus leaving the question open to further investigation. Neuston Net Sampling Semiquantitative sampling with the paired 368 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS neuston nets was initiated late in the study, dur- ing the period 3 September to 12 October 1969 (Table 4). The results of the habitat trap collec- tions (Table 2) suggest that these towing opera- tions were conducted during a period of very low puerulus abundance, yet two pueruli were collect- ed during two separate towing efforts. Unfor- tunately, it was not possible to complete the development of these nets and begin sampling operations during the summer period of peak puerulus abundance. Concurrent studies of the distribution and abundance of the phyllosoma larval stage of P. interruptus in the same inshore areas, using standard 1 m and larger conical nets in surface and oblique towing patterns, failed to catch any pueruli (W. E. Hazen and J. H. Rutherford, pers. commun.). In addition, considering that Johnson (1956, 1960) obtained only a few pueruli during 7 yr of extensive sampling with surface and oblique hauls of a 1-m conical net, our preliminary results suggest that use of this neuston net system probably is a more effective sampling technique and should be thoroughly evaluated in future studies. These results also strengthen the ar- gument that the puerulus is pelagic, occurs at the surface, and is a temporary member of the neuston community. Relationship of Puerulus Settlement to Environmental Factors Efforts to relate the influx of pueruli in major environmental factors, such as temperature, salinity, and lunar phase, were complicated by lack of information concerning other unmea-sured or unknown variables which may have subtle effects on the settlement of pueruli in the habitat traps. These may include specific wave characteristics, current velocity and direction, and water tur- bidity. Therefore, the apparent low abundance of pueruli during some periods might be due to their failure to settle in the habitat traps, rather than their absence from coastal waters. Variations in surface salinity levels of water masses along the open coast when our sampling was conducted were very slight. Consequently, the influence of this factor probably was negligible and showed no ob- vious relationship to puerulus settlement, based on data available for the area of the Scripps Institu- tion pier. Variations in wave heights and surf conditions occurring at the Scripps Institution pier during the summer also were relatively small. Table 4.-Results of puerulus sampling with paired neuston nets. Surface Water area volume San Diego Date sampled sampled sampling Number 1969 Time (m^)i (m')i localities collected 3 Sept. 1400-1500 800 400 Mission Bay 0 10 Sept. 2000-2100 800 400 Mission Bay Channel 0 10 Sept. 2130-2230 800 400 Mission Bay Channel 1 20 Sept. 1800-1840 500 250 Off Pacific Beach 0 20 Sept. 1900-1930 300 150 Off Pacific Beach 1 7 Oct. 2000-2100 800 400 Off Pacific Beach 0 12 Oct. 1200-1330 1,000 500 Mission Bay Channel 0 'Estimated from average boat speed and duration of tow. The range in height of swells recorded at the pier was 0.3-0.9 m (1-3 feet), and showed no obvious pattern in relation to puerulus settlement. However, the following general relationships between puerulus settlement and environmental conditions seem apparent. Seasonal Periodicity The results of night-lighting during 1969 (Table 1), and habitat trap sampling conducted in 1969 (Table 2) are presented together in Figure 7. These data indicate that the pueruli of P. interruptus began to appear in nearshore San Diego waters during late May, and occurred there continuously until mid-September, apparently reaching their greatest abundance during the first week of August. The habitat traps could not be maintained during rough winter surf conditions, and oc- casional night-light efforts during the winter were unsuccessful, so we were not able to establish conclusively that pueruli are absent from nearshore waters between October and May. However, evidence presented by Johnson (1960) on the seasonal periodicity of the later phyllosoma stages of P. interruptus, which are abundant only during the period from January to June, suggests that there would be no major influx of pueruli during the winter months, although a small number of individuals might be present throughout the year. Relationship to Water Temperature A comparison between surface water tempera- tures and puerulus trap catches obtained at the Scripps Institution pier indicates that the 369 FISHERY BULLETIN: VOL. 73, NO. 2 lunar phase 0])«([0])»eO])« d o D • d O D • a o D surface temperature puerull collected by night-light (no per hr ) 0 0 4 6 10 5 4 0 pueruli settlement in seaweed traps no traps present -:■,■.--,- vV.^.JlpiUi J u M biMiti May 1969 June July Aug Sept Oct Figure 7.— Relationship of puerulus settlement to lunar, temperature, and seasonal cycles. The habitat trap catches are presented as catch per unit effort, and represent the total combined catch of two seaweed traps (primarily the single and double frame designs, Figure 2) maintained in the lighted pier positions (a and b). The date of original settlement for each puerulus was estimated within 1 to 2 days by its degree of pigmentation. Temperature data are from Scripps Institution pier records. prevailing influx of pueruli corresponds with the seasonal period of highest temperatures, as shown in Figure 7. The peaks of puerulus settlement oc- curring during the first weeks of July and August also appear to correspond with periods of rising water temperatures. However, another possible interpretation of the settlement patterns shown in Figure 7 is that the low abundance of pueruli ob- served during the period of 10-25 July may have been caused by an influx of cold water originating from local submarine canyon upwelling, which oc- curs periodically. Such a cold upwelling water mass would not be expected to contain any pueruli if the surface-swimming habits attributed to this stage are correct, and could effectively prevent pueruli in warmer water masses from entering the pier area we sampled. Lunar Perdicity Studies by Witham et al. (1968) and Sweat (1968) of P. argus pueruli in Florida and by Phillips (1972) of P. longipes cygnus pueruli in Australia deter- mined that puerulus settlement was highest dur- ing the new moon phase and did not occur at all during full moon periods. The abundance of P. in- terruptus pueruli, on the other hand, did not show any evident relationship to lunar phases. On two occasions during the first parts of July and August, there were apparent peaks in puerulus abundance during the full moon phase, as shown by the habitat trap results in Figure 7. Night-light and habitat trap collections and their relationship to lunar phases are summarized in Table 5. These data indicate that there were no lunar periods when pueruli of P. interruptus were not present and did not settle, and thus that they apparently do not respond to lunar cues, or at least not in the same manner as do species studied elsewhere. Witham et al. (1968), Sweat (1968), and Phiflips (1972) were unable to determine whether the ab- sence of pueruli from their habitat traps during full moon periods was due to their absence from the area or their avoidance of the traps. The fact that pueruli of P. interruptus were present and settled in our habitat traps directly under the bright lights of the Scripps Institution pier in- Table 5.— The relationship between lunar phases and puerulus abundance, as determined by habitat trap and night-light results. Total catches, given in detailed form in Tables 1 and 2, are presented below as catch per unit effort. Lunar phase (eight equal Habitat trap Night-light divisions) (mean catch/hour)i (mean catch/hour)^ Full moon (1) 5.0 3.0 (2) 2.5 4.0 Last quarter (3) 1.7 4.0 (4) 1.7 10.0 New moon (5) 2.0 1.0 (6) 1.5 3.0 First quarter (7) 1.8 6.5 (8) 1.0 0.0 'Based on catches of the single and double frame traps only. ^Periods of poor conditions (i.e., unsuitable for observing pueruli) were excluded from calculation of the mean. 370 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS dicates that, at least for this species, illumination is not an inhibitory factor. In fact, it may serve as a stimulus for settlement. In this regard it is important to note that per- sistent stratus overcast in the southern California coastal zone during the summer months results from upwelled water coursing southward from Point Conception. Thus, normal background illumination from the moon and stars generally is eliminated, although sky glow is present near major coastal cities. Consequently, at our sampling site the prevailing overcast and noctur- nal illumination from the Scripps Institution pier lights could have masked any lunar effect, result- ing in what we observed, puerulus settlement in traps during all moon phases. Duration of the Puerulus Stage Essentially nothing is known about the duration of the puerulus stage of any spiny lobster species. Sheard (1949) suggested that the puerulus of the western Australian spiny lobster, P. longipes cyg- nus, lasts for 2-3 wk, but offered no supporting evidence. April was consistently the period of greatest abundance for the last phyllosoma stage (stage 11) \ 1 Phyllosoma \2 Stage No. \3 / \ '.4 \ ^. 6 V7 ^ 9 \ ESTIMAIED DATE OF METAMORPHOSIS TO PUERULUS ■. 10 \ i<- 2j mo.- Puerulus settlement- Aug Sept Oct Nov Dec Jan Feb Mar Apni May June July Aug Sept Figure 8.— Comparison of seasonal occurrence and dates of greatest abundance of the phyllosoma and puerulus larval stages. Data on the phyllosoma stages are from Johnson (1960). of P. interruptus, based on evidence obtained by Johnson (1960) in extensive sampling during the period 1949-1957, as summarized in Figures 8 and 9. Most of the late stage phyllosoma larvae he 'Srfn Francisco Phyllosonia Larvae : 1 - No. Stage 10 1~ No. Stage 1 1 .'20 Kn- (?C0 miles) Figure 9.-Offshore distribution of the late stage phyllosoma larvae (stages 10 and 11) of Panulirus interruptus during the period from 1949 to 1955, as reported by Johnson (1960). collected were taken at stations 160-320 km (100-200 miles) offshore (Figure 9). Our data in- dicate that the greatest abundance of pueruli in coastal waters during 1969 was in early August, approximately 3V2 mo after this peak in the abun- dance of stage 11 phyllosomes. Assuming that the duration of the eleventh phyllosoma stage extends for a period of 1 mo, and that shoreward migration is accomplished by the puerulus, rather than by the last phyllosoma larval stage (see subsequent dis- cussion), then this suggests that the puerulus stage of P. interruptus may have an average duration of approximately 2V2 mo. This timing is represented diagrammatically in Figure 8. Thus, a period of 2-3 mo appears to be a reasonable es- timate for the average duration of the puerulus stage. 371 FISHERY BULLETIN: VOL. 73, NO. 2 DISCUSSION Mode of Life of the Puerulus Stage Johnson (1960) and Lindberg (1955) speculated that the puerulus of P. interruptua is a benthic form, largely because it occurs so infrequently in standard net tows. However, the results of this study strongly suggest that it is a pelagic form, and in particular a member of the surface- dwelling neuston. Several lines of evidence sup- port this conclusion. Our observations of many free-swimming pueruli which were attracted to night-lights revealed that the pueruli always swam in the top few centimeters of surface water, and no individuals were ever observed moving toward the light from deeper water. Pueruli were readily collected in habitat traps floating at the surface. However, the occurrence of pueruli at greater depths cannot be ruled out on the basis of habitat trap evidence alone because the traps were not maintained at greater depths. Although preliminary and very limited, our surface sampling with neuston nets yielded a much higher catch per unit effort than did standard near-sur- face and oblique tow sampling methods employed by previous investigators. The puerulus stage of P. interruptus also has specialized physical characteristics which suggest that it is adapted to a pelagic existence. These include: 1) heavily setose pleopods and a streamlined body for efficient, forward swimming; 2) a transparent body completely devoid of pig- mentation; and 3) extremely long, delicate anten- nae which appear unsuitable for a benthic exis- tence because they quickly break off upon settlement. In addition, at the time the puerulus moults into the first postpuerulus stage, it loses the setose pleopods and acquires stronger antennae, an expanded cephalothorax, and walking ability typical of later demersal stages. In fact, pueruli we collected and held in aquaria were never observed walking upon the substrate, instead only clinging to algae or crevices in rocks and shells. A more detailed description of this transformation process from puerulus to postpuerulus was developed in a subsequent study by Parker (1972). Both Witham et al. (1968) and Phillips (1972) reported that, in comparative tests between habi- tat traps maintained on the surface and traps anchored 1-4 m below the surface, all but a few pueruli of P. argus and P. longipes cygnus were collected in those at the surface. Phillips collected only 1 out of 38 pueruli in traps at 4-m depth, and Witham collected 12 from one trap anchored on the bottom in water 1 m deep, as compared to an average of 27 pueruli in surface traps located nearby. Sweat (1968) utilized a multiple plankton net array for studying the depth preference of P. ar- gus pueruli in the Florida Keys area. This system consisted of three conical plankton nets mounted from a bridge across a shallow channel connecting Florida Bay with the Atlantic Ocean. The nets were suspended at the surface, at mid-depth (1 m), and on the bottom (2.3 m), and were operated for 2-h periods within the new moon phase during evening flood tides. The largest proportion of the pueruli were collected by the mid-depth net (116 at the surface, 418 at 1 m, 61 at 2.3 m depth). However, this may have been due to the particular conditions in the channel, including its shallow depth, water turbulence, and the possibility that the pueruli were in the process of settlement at the time of sampling. In general, these observations by Sweat (1968), Witham et al. (1968), and Phillips (1972) agree with ours on P. interruptus. The pueruli of these species seem to be restricted primarily to the ocean sur- face, at least just prior to settlement. The Functional Significance of the Puerulus Stage On the basis of morphology and our observations of its behavior, the puerulus stage of P. interrup- tus appears to be well adapted for directed, for- ward swimming, as shown in Figure 5, rather than for the passive, pelagic existence apparently exhibited by all phyllosoma larval stages of spiny lobsters. What, then, is the purpose of this directed swimming? A possible answer is suggested by considering the early life history of the scyllarid lobster, Scyllarus americanus, which also has a phyllosoma larval stage but no transitional puerulus stage, moulting instead directly into a nonswimming, benthic juvenile (Robertson 1968). Some other scyllarid species have a transitional form called a puerulus or, more properly, a nesto stage. The phyllosoma larval period of S. americanus is quite short (4-5 wk), thereby allow- ing the larvae a much greater chance of remaining near the shallow coastal areas suitable for later demersal life. In contrast, the phyllosomes of palinurid lobsters and other species of scyllarids may be carried several hundred kilometers out to 372 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS sea during their typical 5-10 mo larval existence. Obviously, the larvae must return not only to the coastal area, but also to very shallow nearshore zones if they are to transform and become es- tablished as demersal juveniles. Previous investigators (Lindberg 1955; Johnson 1956, 1960, 1971; Saisho 1966; Sims and Ingle 1966; Lazarus 1967; Chittleborough and Thomas 1969; Chittleborough 1970) have postulated for several species that this recruitment in nearshore waters takes place during the phyllosoma stage, possibly through the action of countercurrents, upwelling, and eddies. Several studies, particularly those of Johnson (1960) and Chittleborough (1970), have shown clearly that large numbers of late stage phyllosoma larvae do remain well within the coast- al areas in which the earlier larval stages occur. Evidence presented by Johnson (1960) also has shown that hydrography plays a major role in re- taining a supply of late stage phyllosomes, and presumably pueruli, within reasonable distances from the coast. The presence, within one net haul, of several different phyllosome stages that must have been produced months apart, and the presence of late stage phyllosomes at the same locality, is good evidence that mixing processes and retaining eddies prevent wholesale flushing of these larvae from the coastal area. However, the late stage phyllosomes seem to be concentrated primarily in areas 50-250 km offshore, and the specific mechanism of onshore movement and recruitment has not been demonstrated. If there were major active or passive movements of late stage phyllosomes toward shore, then one would expect to find relatively large numbers of them in shallow, inshore waters as well, or at least a trend in this direction. Most of the sampling reported by Johnson (1956, 1960) was conducted in waters 8 or more kilometers ( > 5 miles) offshore. However, studies of the distribu- tion and abundance of phyllosoma larvae of P. in- terruptus in San Diego waters much closer inshore by W. E. Hazen and J. H. Rutherford (pers. com- mun.) during the summer months of 1969-1970, employing surface and oblique net tows, failed to collect any individuals older than stage 2. This suggests that the later stages occur either many kilometers offshore, as observed by Johnson (1956, 1960), or are concentrated in unknown areas. For lack of evidence to the contrary, it has also been suggested that these phyllosoma larvae oc- curring far offshore are lost to the population and must perish. For similar reasons, Lindberg (1955) and Johnson (1960) have speculated that when these phyllosomes moult into the puerulus stage, the puerulus quickly settles to the bottom while still in deep water. Presumably, some of these are then able to migrate onshore to the shallow coastal nursery areas as benthic puerulus or postpuerulus forms. However, in light of our behavioral observations on both pueruli and juveniles, another more likely explanation is that the puerulus stage may have evolved specifically for this purpose of recruit- ment. These observations suggest to us that the puerulus is a transitional, pelagic stage specifically adapted for directional swimming, and that it is capable of returning by active means to nearshore nursery areas suitable for settlement, thereby fulfilling the key role in recruitment. During this process, the surface-swimming and associated positive phototactic behavior of the puerulus stage probably aid it in locating the shallow nearshore areas which our related observations suggest are required as nursery grounds for the early juvenile stages. Exactly how important a role the puerulus stage plays in the recruitment of the demersal popula- tion probably depends on three factors: 1) the degree of "assistance" contributed previously by the late phyllosoma stages which may move ac- tively, e.g., by vertical migrations, into shoreward-directed currents or eddies; 2) how ef- fective the puerulus is in travelling over long dis- tances, with regard to both swimming speed and endurance; and 3) how well the puerulus can navigate, considering the fact that aimless wan- dering or swimming away from the coast would markedly reduce its probability of survival. It seems reasonable to expect that the puerulus can swim nearly continuously, as other nektonic crustaceans, such as euphausiids, apparently do. If so, our estimate of an 8 cm /sec average swimming speed for the puerulus indicates it has the poten- tial to travel approximately 7 km (4.3 miles) per day, or about 500 km (350 miles) during the period of 80 days estimated as the approximate average duration of this stage. Thus, the 160-320 km (100-200 mile) distance from shore at which John- son (1960) found most late stage P. interruptus phyllosomes (Figure 9) could be within the basic swimming capabilities of the puerulus stage, even if part of the time was spent swimming against surface currents or in an inactive state. The current patterns off the southern California and Baja California coasts are complex, and have 373 FISHERY BULLETIN: VOL. 73, NO. 2 been studied extensively (see, for example, John- son 1956, 1960; Wyllie 1966). During the summer months the California Current system has a generally southward trend, but displays retaining eddies and a net northward onshore drift in the northern range of P. interruptus larval distribu- tion near the southern California coast and ad- jacent Channel Islands. There is a net southerly or offshore drift for water masses off most of Baja California. Even the apparent swimming capabilities of the puerulus stage probably would not allow it to move against these strong, offshore surface currents, particularly because an object in the surface waters would be vectored at a 45° angle, or westwardly, to the wind and current forces. As a specific example of this problem, review of mean geostrophic flow at the surface off Califor- nia and Baja California for the months of June- September during the typical period 1950-1964 (Wyllie 1966) reveals that there was a net surface transport southward from Point Conception to Cabo San Lucas offshore from approximately 80-320 km (50-200 miles). Northward flow near shore during these summer periods occurred only in the Southern California Bight (San Diego to Ventura), while net offshore transport apparently occurred from Bahia San Quintin south to Cabo San Lucas as a precursor to the California Current Extension. Thus, it appears very likely that a majority of the late stage phyllosoma and puerulus larvae in the surface layers in this region, more than about 40-95 km (25-60 miles) offshore, depending on variations in the current system, were swept seaward by geostrophic flows which averaged greater than 46 cm /sec (0.9 knots) during this typical 15-yr interval. Such individuals undoubt- edly are lost to the population. On the other hand, individuals present closer to shore, or in retaining eddies near the Southern California Channel Islands and the shallow Bahia Sebastian Viz- caino-Isla Cedros area (Johnson 1960), are within distances and ocean surface conditions which would allow their nearshore recruitment by directed swimming of the puerulus stage. The Significance of Phyllospadix in the Settlement of the Puerulus Stage The strong preference by the puerulus stage for the habitat traps containing Phyllospadix torreyi, as compared to generally similar synthetic material in the Witham traps, seems particularly significant in view of the fact that both Serfling (1972) and Parker (1972) discovered numerous early juvenile stages primarily in areas which had thick growths of this surfgrass. Comparative evaluations of habitat traps filled with Phyllo- spadix and other substrates, such as giant kelp {Macrocystis pyrifera) fronds and holdfasts, the eelgrass {Zostera marina), and MytHns clumps, might prove useful as a means of improving collection success with the traps. Preference tests involving various substrates typical of different nearshore habitats might also suggest other areas of natural puerulus settlement. In this regard, however, substrate preference tests of the puerulus and postpuerulus stages conducted in the laboratory by Parker (1972) suggest that these stages favor Phyllo- spadix over Macrocystis, Zostera, several species of red algae, sand, and rock. Evaluation of the Natural Seaweed and Artificial Habitat Traps In comparative tests conducted by Witham et al. (1968), their Witham habitat trap proved somewhat more successful than two other seemingly poor refuges, a tire and a shingle (102 versus 77 and 57 pueruli collected, respectively, over a 10-mo period). Phillips (1972), studying P. longipes cygnus in Australia, found that his ar- tificial seaweed habitat trap design collected more pueruli than a modified Witham trap, but only by a factor of approximately two. In contrast, the results of our comparative evaluations (Table 2) indicate that the average number of pueruli caught by the two lighted natural seaweed habitat traps (single and double seaweed frame) was 47 per trap, while on lighted Witham artificial sub- strate trap caught only 5 pueruli over the same time period and at the same location. This sug- gests that the natural seaweed trap design was approximately nine times more effective than the Witham trap. Secondly, the lighted habitat traps, regardless of the design, caught more pueruli than did nonlighted ones at the Scripps Institution pier during the same time period (33 and 7 pueruli re- spectively), suggested that a lighted trap was approximately four to five times more effective than an unlighted one. Thus, the system developed i in this study, utilizing a combination of both noc- turnal illumination and natural seaweed, clearly is j much more effective in collecting pueruli of P. in- " 374 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS terruptus than the nonlighted artificial habitat system utilized by Witham et al. (1968). This suggests that natural seaweed habitat traps filled with native flora characteristic of juvenile habitats, in combination with nocturnal illumination, could prove to be a more successful means of sampling the pueruli of other spiny lob- ster species as well. If so, use of this modified sampling technique might indicate that the abun- dances of P. argus pueruli in Florida and those of P. longipes cygnus in Australia actually are much greater than previously estimated by Witham et al. (1968), Sweat (1968), and Phillips (1972). Implications for Aquaculture and Fishery Management If the small numbers of pueruli captured during this study are representative of puerulus availability throughout the geographic range of P. interruptus, large-scale collecting of this stage for purposes of aquaculture and restocking is not feasible and probably could not be justified. However, other locations, particularly those closer to the center of adult and larval concentrations, such as the Bahia Sebastian Vizcaino-Isla Cedros area off Baja California (Figure 9), should be investigated as potential sites for such large-scale collecting operations, as well as for purposes of locating the primary areas of puerulus settle- ment. It also seems reasonable that the habitat trap collecting system developed in this study, if stan- dardized and employed on a wider geographic scale, could prove useful for monitoring fluctua- tions in year class recruitment of pueruli, and thereby provide a means of predicting fluctuations in the size of the demersal population in following years. For example, an extension of our study by Parker (1972) during the years 1970-71 indicates that puerulus settlement at the Scripps Institution pier was much less than we observed during the same months in 1969. If this reduced recruitment was representative of a wider geographic area, then the size of the adult population available to the commercial fishery within the succeeding 5-8 yr might be expected to show corresponding changes. SUMMARY Basic ecological and behavioral information was obtained about the recruitment process, habitat preferences, and general abundance of the puerulus larval stage of Panulirus interruptus. Pueruli of P. interruptus exhibit a strong posi- tive phototactic response, and could be lured to a bright underwater night-light from the surface water surrounding the Scripps Institution of Oceanography pier. Direct observations of free swimming pueruli by this method demonstrated that this stage is typically pelagic rather than benthic, and swims at the surface in a continuous and directed manner. Estimates of swimming speed were obtained. The surface swimming behavior of this stage indicates that it probably can be properly sampled quantitatively only by large nets towed horizon- tally at the surface. This may explain why few pueruli have been taken by other conventional sampling methods. Paired neuston nets were developed specifically for this purpose and pueruli seemed to be sampled effectively in this manner during preliminary evaluations. Puerulus larvae also were collected effectively in floating habitat traps containing the surfgrass, Phyllospadix torreyi. A variety of natural seaweed habitat trap designs were tested, and all appeared to be about equally effective in collecting pueruli; however, a nylon bag habitat trap proved best in terms of cost and durability. All natural seaweed habitat traps were markedly superior in collecting pueruli compared to the Witham habitat trap design, formed of synthetic fibrous material. Habitat traps maintained under the lighted end of Scripps Institution pier collected many more pueruli than those not subject to such artificial illumination. The failure of habitat traps placed offshore to collect any pueruli may have been due to the availability of abundant seaweed flotsam in the areas where they were maintained, as well as lack of artificial illumination. Both the presence of intertidal plants (particularly Phyllospadix) and nocturnal illumination appear to play significant roles in the settlement of puerulus larvae in habi- tat traps. The results of night-lighting and habitat trap sampling indicate that off San Diego, Calif., the seasonal influx and settlement of puerulus larvae is continuous, beginning in May and ending in September. Estimates based on a comparison of the peak periods of abundance for pueruli and the preced- ing final phyllosoma larval stage suggest that the puerulus stage of P. interruptus has a duration of approximately 2V2 mo. This is followed by 375 FISHERY BULLETIN: VOL. 73, NO. 2 settlement in shallow water and transformation to a benthic, postpuerulus form. Based on observations of its surface swimming behavior and capabilities, preference for plant covered substrates, settlement behavior, and morphology, the puerulus appears to be a transi- tional, pelagic stage specifically adapted for directional swimming, whose function is to return from offshore by active means to nearshore areas suitable for settlement. Thus, it probably occupies the key role in recruitment. The seemingly low abundance of pueruli in the southern California areas sampled suggests that it would not be practical or beneficial to attempt large collections there for purposes of aquaculture and restocking. However, other locations, includ- ing those closer to the center of the geographical range, should be investigated as potential large scale collecting sites. Employed in a standardized manner, the habitat trap system developed in this study could prove useful in locating primary areas of puerulus settlement and in monitoring fluctuations in year class recruitment. ACKNOWLEDGMENTS This work is the result of research sponsored by the National Science Foundation Office of Sea Grant Programs, under Grant No. GH-36 and by the NOAA Office of Sea Grant Programs, Depart- ment of Commerce, under Grant No. USDC 2-35208. The U.S. Government is authorized to produce and distribute reprints for governmental purposes, notwithstanding any copyright notation that may appear herein. We wish to thank Mar- garet Knight of the Scripps Institution of Oceanography (SIO) for providing space in the SIO experimental aquarium. We also wish to ac- knowledge the assistance of Martin W. Johnson of the Scripps Institution of Oceanography and David A. Farris and Glenn A. Flittner of San Diego State University in reviewing the manuscript. LITERATURE CITED Chittleborough, R. G. 1970. Studies on recruitment in the Western Australian rock lobster Panulirus longipes cygnus George: Density and natural mortality of juveniles. Aust. J. Mar. Fresh- water Res. 21:131-48. Chittleborough, R. C., and L. R. Thomas. 1969. Larval ecology of the Western Australian marine crayfish, with notes upon other panulirid larvae from the eastern Indian Ocean. Aust. J. Mar. Freshwater Res. 20:199-223. Dexter, D. M. 1972. Moulting and growth in laboratory reared phyllosomes of the California spiny lobster, Panilirus interruptus. Calif. Fish Game 58:107-115. George, R. W., and P. Cawthorn. 1962. Investigations on the phyllosoma larvae of the West- ern Australian crayfish. Rep. W. Aust. Mus. 1962:1-12. Gurney, R. 1942. Larvae of decapod Crustacea. Ray Soc. Publ. 129, Ray Society, Lond., 306 p. Harada, E. 1957. Ecological observations on the Japanese spiny lobster, Panulirus japonicus (Von Siebold), in its larval and adult life. Publ. Seto Mar. Biol. Lab. 6:99-120. Ingle, R. M., and R. Witham. 1969. Biological considerations in spiny lobster cul- ture. Proc. Gulf Caribb. Fish. Inst., 21st Annu. Sess., p. 158-162. Johnson, M. W. 1956. The larval development of the California spiny lob- ster, Panuliris interruptus (Randall), with notes on Panuliris gracilis Streets. Proc. Calif. Acad. Sci., Ser. 4, 29:1-19. 1960. Production and distribution of larvae of the spiny lobster Panulirus interruptus (Randall) with records on P. gracilus Streets. Bull. Scripps Inst. Oceanogr., Univ. Calif. 7:413-461. 1971. The palinurid and scyllarid lobster larvae of the tropical eastern Pacific and their distribution as related to the prevailing hydrography. Bull. Scripps Inst. Oceanogr., Univ. Calif. 19:1-22. Kensler, C. B. 1967. Notes on laboratory rearing of juvenile spiny lobsters, Jasus edwardsii (Hutton) (Crustacea: Decapoda: Palinu- ridae). N.Z.J. Mar. Freshwater Res. 1:71-75. Lazarus, B. L 1967. The occurrence of phyllosomata off the Cape with particular reference to Jasus lalandii. Invest. Rep., Div. SeaFish.S.Afr. 63:1-38. Lewis, J. B., H. B. Moore, and W. Babis. 1952. The post-larval stages of the spiny lobster Panulirus argus. Bull. Mar. Sci. Gulf Caribb. 2:324-337. Lindberg, R. G. 1955. Growth, population dynamics, and field behavior in the spiny lobster, Panulirus interruptus (Randall). Univ. Calif. Publ. Zool. 59:157-248. Parker, K. P. 1972. Recruitment and behavior of puerulus larvae and juveniles of the California spiny lobster, Panulirus in- terruptus. Master's Thesis, San Diego State Univ., 91 p. Phillips, B. F. 1972. A semi-quantitative collector of the Puerulus larvae of the western rock lobster Panulirus longipes cygnus George (Decapoda, Palinuridea). Crustaceana 22:147-154. Robertson, P. B. 1968. The complete larval development of the sand lobster, Scyllarus americanus (Smith), (Decapoda, Scyllaridae) in the laboratory, with notes on larvae from the plank- ton. Bull. Mar. Sci. 18:294-342. Saisho, K. 1966. Studies on the phyllosoma larvae with reference to the 376 SERFLING and FORD: ECOLOGICAL STUDIES OF PANULIRUS INTERRUPTUS oceanographical conditions. [In Jap., Engl, abstr.] Mem. Fac. Fish., Kagoshima Univ. 15:177-239. Serfling, S. a. 1972. Recruitment, habitat preference, and growth of the puerulus and juvenile stages of the California spiny lob- ster, Panulirus interruptus (Randall). Master's Thesis, San Diego State Univ., 124 p. Serfling, S. A., and R. F. Ford. In press a. Natural growth and age-size relationships in juvenile stages of the California spiny lobster, Panulirus interruptus. Fish. Bull., U.S. In press b. Laboratory culture of juvenile stages of the California spiny lobster, Panulirus interruptus (Ran- dall), at elevated temperatures. Aquaculture. Sheard, K. 1949. The marine crayfishes (spiny lobsters), family Palinuridae, of Western Australia with particular reference to the fishery on the Western Australian crayfish (Panulirus longipes). Aust. C.S.I.R.O. (Com- monw. Sci. Ind. Res. Organ.), Bull. 247, 45 p. Sims, H. W., Jr., and R. M. Ingle. 1966. Caribbean recruitment of Florida's spiny lobster population. Q. J. Fla. Acad. Sci. 29:207-242. Sweat, D. E. 1968. Growth and tagging studies on Panulirus argus (La- treille) in the Florida Keys. Fla. Board Conserv., Tech. Ser. 57, 30 p. Thorson, G. 1950. Reproductive and larval ecology of marine bottom invertebrates. Biol. Rev. (Camb.) 25:1-45. WiNGET, R. R. 1968. Trophic relationships and metabolic energy budget of the California spiny lobster, Panulirus interruptus (Randall). Master's Thesis, San Diego State Univ., 232 p. Witham, R., R. M. Ingle, and E. A. Joyce, Jr. 1968. Physiological and ecological studies of Panulirus ar- gus from the St. Lucie estuary. Fla. Board Conserv., Tech. Ser. 53, 31 p. Witham, R., R. M. Ingle, and H. W. Sims, Jr. 1964. Notes on postlarvae of Panulirus argus. Q. J. Fla. Acad. Sci. 27:289-297. Wyllie, J. G. 1966. Geostrophic flow of the California Current at the sur- face and at 200 meters. Calif. Coop. Oceanic Fish. Invest. Atlas 4, 288 p. 377 A THEORETICAL TREATMENT OF UNSTRUCTURED FOOD WEBS G. D. Lange and a. C. Hurley' ABSTRACT In a recent paper, Isaacs has proposed a model for an unstructured food web in which the interconnec- tions are so diverse that all heterotrophs in the system can be treated as if they were at the same average trophic position. This paper recasts the original model in terms of a 3 x 3 matrix using three empirical constants. In this form, the model can be easily generalized to one having nine constants and reflecting a more realistic view of the interactions among levels of a community. Recent papers by Isaacs (1972, 1973) proposed an alternative to trophic level schemes for represent- ing interactions among species. He termed this an unstructured food web and proposed a "matrix"^ technique (Isaacs 1972) for evaluating the equilibrium distribution of energy (or matter) which would result from these interactions. In this paper we propose an alternative formulation of Isaacs' model which utilizes classical matrix and operator techniques. SERIES APPROACH Isaacs' model was originally proposed to account for Young's data (Young 1970) from the Gulf of California which indicated that cesium was not found concentrated in ratios one would expect from a simple food chain. Isaacs assumes that the principal interconnections in the marine food web are so diverse that all heterotrophs in the system (from microorganisms to vertebrates) can be treated as if they derived their food from a com- mon source that is only coarsely differentiated. Therefore, the heterotrophs can all be treated as if they were at the same average trophic position. In this unstructured food web, Isaacs visualizes four levels of matter or energy: 1) source, 2) living tis- sue, 3) nonliving but retrievable matter, and 4) irretrievable matter. The source is assumed to be phytoplankton which is added to the system at a 'Department of Neurosciences, School of Medicine, and Marine Neurobiology Unit, Scripps Institution of Oceanography, University of California San Diego, La Jolla, CA 92037. 'We have enclosed Isaacs' use of the word "matrix" in quotes because he has used the word in a common rather than in the standard mathematical sense. When the word appears without quotes in this text we are using it in the standard sense of a rectangular array of elements which operates on column vectors from the left to produce new column vectors. Manuscript accepted June 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 378 constant rate. The living matter consists of all heterotrophs, while the dead retrievable matter may consist of such sources of carbon as organic detritus or dissolved organic matter. The irre- trievable component is that matter (or energy) which is forever lost to the system through such processes as respiratory combustion or mineralization. The "unstructured" nature of the food web comes from a set of coefficients which represent movement of material between these groups. The transitions are not in a trophic level line. Rather, groups two and three interact bilaterally and groups three and four can receive from other levels bypassing intermediates. Isaacs calculates the final steady state values for the total living and dead material by summing two infinite series. To obtain these series, he in- troduces a "matrix" which is designed to aid in the formulation of each of the terms. The series take the form: M\ = \-{K, + K,) K, M", = Mf,K^ + Mo K^ {K^ ■¥ K^) + M,K,[K,{K, + K,) + K,{K, + K,)] M,K, 1 - (^1 + ^3) where Mq = increment of initial input periodically introduced into the LANut. and HUKL,tL,i: : uwsi Kuiji UKiiu I'uuu wttis system at intervals equal to the time taken by one average step in the food web, M'f = total quantity of material in living tissue (level two), M"f = total in nonliving recoverable material (level three), K^ = a coefficient of conversion of matter (or energy) in food into living tis- sue, K2 = acoefficientof conversion of matter (or energy) in food into irretriev- able form (e.g., by respiratory com- bustion or mineralization), and K^ = a coefficient of conversion of matter (or energy) in food into nonliving but retrievable form (e.g., organic detritus or dissolved organic mat- ter). Restrictions on coefficients are: K^ + K2 + Ks= 1, 000 In Isaacs' terms Cj = Mq and the limiting values for the second and third compartments are M\ and M"t respectively. Therefore M\ =M,K,/K2 M'; = MoK,/K2 which is exactly Isaacs' result. For the nine constant model, there is also always a steady state distribution of matter in the sys- tem. By finding the eigenvector corresponding to an eigenvalue of one, we can obtain the following steady state values of M', (total quantity of material in living matter) and M',' (total in nonliving recoverable material) in terms of a con- stant input Mq-. {k,-l){k,-l)-k,k, ' {k,-\){k^-\)-k^k^ Trophic Level Equations In addition to values for total amounts of living and retrievable dead matter, Isaacs develops equations for general trophic levels. His equations can be generated by our approach if our original matrix is broken down into component parts and then applied to the steady state vector. For example, let us consider Isaacs' case (Isaacs 1973) of a subset of trophic levels which are complete and mutually exclusive. He considers strict herbivores, detrital feeders, and full predators to be such a subset. Our original matrix A can be written in the following way A = A s + i? + ^// + ^Z) + ^i where As + r = 0 0 ^^3^3^3> ^0 0 0 ,0 0 0 '0 0 0 A^ = I 0 0/^1 ^0 0 0 /o 0 0' Ap = I 0 K^ 0 \0 0 0, -^s+ij - matrix responsible for the biomass in source and the retrievable dead matter, = matrix responsible for biomass in her- bivores, = matrix responsible for biomass in detrital feeders, and Ap = matrix responsible for biomass in preda- tors. To obtain the potential biomass for each of the trophic levels, we take the appropriate matrix times the steady state vector. Thus, the equation for the potential biomass of herbivores is obtained from ^H ^D AfjUi 0 0 0 A',0 0 0 0 0 Mn = \M,K, 380 LANGE and HURLEY: UNSTRUCTURED FOOD WEBS Similarly, for detrital feeders and for predators ApU^ = All of Isaacs' other equations for trophic level potential biomasses or fluxes can be obtained in a similar manner. Equations for the potential biomass of trophic levels can also be calculated for the generalized model. This is done in a manner similar to that described in the previous section. Strict herbivores (feeding on source): Mm =M^k,. Omnivores (feeding on source, living and re- trievable dead): M = M I -^l(^8-l) + ^2^7 '\ik,-l){k,-l)-k,k, = k^MQ + k^ M\ + k^M'l. Particle feeders (feeding on source and re- trievable dead): M^=Mo k,{k,-\){k^-l) + k^k,{\-k,) {k,-l){k,-l)-k,k, = kiMQ + kr; M'l . Detrital feeders (feeding on retrievable dead): M, = M. -k^k-jk^ + k^krj + k^k^krj ~lk,-l){k,~l)-k,k, = k^ M'[. Full predators (feeding on living): Mp = M, -k^k^ {k^-\) + k^ k2krj « > (Z:4-l)(^8-l)-^5^7 = k,M', Nonherbivorous omnivores (feeding on living and retrievable dead): M = M /~"'4"'i"'8 + k^k^ + k2k'j + k-^k^k^j {k,-l){k,-l)-k,k, = k^M'f + kjM';. ACKNOWLEDGMENTS This work w^as partially supported by Public Health Service Grant NS-09342. We would like to thank the following people for reading the manuscript and making valuable comments: J. D. Isaacs, G. Wick, J. Enright, P. Hartline, and M. Mullin. Thanks are also due E. Venrick who brought the problem to our attention and the other students in NS 242 who patiently allowed us to work out the model as a classroom example. LITERATURE CITED Isaacs, J. D. 1972. Unstructured marine food webs and "pollutant analogues." Fish. Bull., U.S. 70:1053-1059. 1973. Potential trophic biomasses and trace-substance con- centrations in unstructured marine food webs. Mar. Biol. (Berl.) 22:97-104. Young, D. R. 1970. The distribution of cesium, rubidium, and potassium in the quasi-marine ecosystem of the Salton Sea. Ph.D. Thesis, Univ. California, San Diego, 234 p. 381 DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA, KATSUWONUS PEL AMIS, IN AN OFFSHORE AREA OF THE EASTERN TROPICAL PACIFIC OCEAN Maurice Blackburn' and Francis Williams^ ABSTRACT Distributions of skipjacit tuna, Katsuwonus pelamis, were studied in the offshore eastern tropical Pacific between lat. 15°N and 5°S, long. 115° and 125°W, during two cruises in 1970 and 1971. Another cruise was made there with different methods in 1969. All cruises were between October and April. Various environmental properties were measured. Catches of skipjack included fish smaller and larger than those generally taken near the American coast. This is consistent with previous hypotheses that mature adults and their larvae generally occur far offshore, whereas adolescents are generally coastal, in the eastern Pacific. The juveniles arriving near the coast and the older fish leaving it evidently cross the studied area on migrations from and to the spawning regions. In 1970 and 1971 skipjack > 45 cm were most abundant in the equatorial upwelling and at the northern boundary of the North Equatorial Countercurrent, and scarce in the Countercurrent. Correlation coefllicients between skipjack > 45 cm and skipjack forage in 1970 were positive and significant by the usual criteria, but the significance may in part be disputable because many other correlations involving skipjack were nonsignificant. The apparent significance was lost when juvenile skipjack ( < 45 cm) were included with the larger ones. Juveniles may have different relations to environment. The 1971 data were scanty and yielded no significant correlations between skipjack and forage. On the 1969 cruise forage was not studied. Skipjack were abundant in the Countercurrent, but at a prespawning stage, whereas postspawners predominated on the other cruises. Other studies suggest that skipjack larvae require relatively high temperatures, which occurred only in the Countercurrent on the 1969 cruise. Skipjack may be distributed according to the environmental requirements of their larvae when spawning and according to their own feeding requirements when not spawning. Williams (1971) described plans for a series of cruises in two offshore areas of the eastern tropical Pacific Ocean. This report deals with results of two cruises made in 1970 and 1971 in one of the areas, bounded by lat. 15°N-5°S and long. 115°-125°W (Figure 1). The cruises were initiated by the National Marine Fisheries Service, South- west Fisheries Center, and the Scripps Tuna Oceanography Research (STOR) Program, Insti- tute of Marine Resources, University of Califor- nia. They were designed to investigate on a seasonal basis the occurrence and relative abun- dance of skipjack tuna, Katsuwonus pelamis, in relation to environmental conditions. Coverage of 'Scripps Institution of Oceanography, Institute of Marine Resources, University of California, San Diego, P.O. Box 1529, La Jolla, CA 92037. ^Scripps Institution of Oceanography, Institute of Marine Resources, University of California, San Diego, P.O. Box 1529, La Jolla, CA 92037; present address: Rosenstiel School of Marine and Atmospheric Science, University of Miami, 10 Rickenbacker Causeway, Miami, FL 33149. Figure l.-Area of eastern tropical Pacific Ocean under inves- tigation. Manuscript accepted June 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 382 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA offshore areas by the U.S. surface tuna fishery has been very limited, especially for skipjack. Coverage by the Japanese subsurface tuna fishery has been greater, but still poor for skipjack (Miyake 1968). The regulation of yellowfin tuna, Thunnus albacares, in the eastern Pacific is ex- pected to increase the need for information on skipjack in the offshore waters. Work in coastal waters has shown that adult skipjack are most numerous when temperature is in the range 20° to 29°C (Williams 1970) and standing stock of skipjack forage is high (Black- burn 1965, 1969). Sea surface temperatures are in the suitable range for most of the year throughout the offshore eastern tropical Pacific (Wyrtki 1964; La Violette and Seim 1969; Love 1971b, 1972a, b, in prep.). Thus distribution of forage seemed likely to be a main factor in determining distribution of adult skipjack in offshore waters. The distribution of forage in that region was described from data of EASTROPAC Expedition (1967-68) by Black- burn and Laurs (1972), who expected that adult skipjack would prove to be distributed in the same way. One of the purposes of the present study was to test that expectation. The forage concentra- tions on EASTROPAC were characteristically high in certain zones of latitude and low in others, broadly corresponding to distributions of phy- toplankton, zooplankton, and total micronekton (Love 1970, 1971a, in prep.; Blackburn et al. 1970; Owen and Zeitzschel 1970a). Evidence, summarized by Williams (1972), in- dicates that most exploited skipjack in the eastern Pacific have a spawning origin in the central Pacific west of long. 130°W. It also suggests that the majority of the skipjack enter the present fishery, which is concentrated near tropical American coasts, during only 1 yr of their life history. They are then relatively small (average length about 50 to 55 cm, average weight 3 to 3.5 kg) and sexually immature. Thus it is probable that migration pathways exist for skipjack, both below and above these sizes, across the offshore eastern tropical Pacific. Such pathways might occur at particular lati- tudes since forage concentrations and many other ocean conditions, including surface currents, are zonally oriented. On this basis Williams (1972) presented three qualitative models of the migra- tion of young (recruit) skipjack from the central Pacific into the eastern Pacific fishing areas. The data from the present cruises may be useful for testing and modification of the models, and incor- poration with other models of skipjack movement such as that of Seckel (1972). PLAN OF THE INVESTIGATIONS The inward skipjack migration models of Williams (1972) assumed that the routes were principally zonal across the offshore eastern Pacific. Mechanisms and timing of the migrations are probably dependent on oceanographic condi- tions and events in this region. The strategy for the present investigations was to have latitudinal sampling of fish and environmental parameters in meridional areas considered critical to the migra- tions. The area discussed in this paper includes the meridian of 119°W. It is important because the surface North Equatorial Countercurrent nor- mally becomes intermittent or absent east of this meridian from January to May. Each cruise had two parts (Williams 1971). Part I was a rapid meridional transect of the area along long. 119°W to monitor ocean conditions and com- pare them with previous data. These results showed the positions of zonal surface current boundaries and forage bands, and hence the lati- tudinal zones that were to be fished for skipjack. In Part II detailed fisheries operations were carried out in the selected zones to a standardized plan based on a "unit area" of 2° latitude by 2° longitude, together with supporting environmen- tal observations. Fishing was by multiple trolling during daylight. Figure 2 shows schematically the track and scheduled observations. The work time for each unit area, including entry and exit, was 96 h. Coverage in a zone of latitude could consist of any multiple of unit areas or fractions thereof (quadrants or 1° x 1° areas). The first cruise in November-December 1970 utilized the vessels Townsend Cromwell (Cruise C 51) (R. Uchida, Chief Scientist) and David Starr Jordan (J 57) (F. Williams, Chief Scientist). The second cruise in March-April 1971, was made with only the Jordan (J 60) (M. Blackburn, Chief Scientist). The Jordan 60 cruise was severely cur- tailed due to illness of a crew member. On the first cruise the Part I transect was completed by the Cromwell; and data were sent by radio to the Jor- dan; subsequent Part II operations were carried out by both vessels. This paper also discusses data from a cruise made by National Marine Fisheries Service, Hawaii, to the same area in October-November 383 FISHKKY BULLtril.N. VUL. /a. INU. Z TRACK *ND OeSERVATONS FOR 2«X Z'UNIT AREA WVESTI6ATI0NS START - TROLLIMG. DAYLIGHT (6-l/t tti — — RU pMVE, night FULL SPEED (11-1/2 kO • STO/NISKIN ORNANSENCSCOm) O XBT A UICR0NEKT0N.5 XSNET . ZDOPLANKTON.CALCOFI liiiai/2ni " NEUSTON D MIWKATER TRAWL (UNIVERSAL TRAWL) Figure 2.-Track and observations scheduled for each 2° x 2° unit area during Part II operations. [Dawn trawl haul was eliminated on cruise Jordan 60]. 1969, in more detail than the previous cruise report by Hida (1970). METHODS This section deals with methods used on the 1970 and 1971 cruises. The Bissett-Berman' Salinity- Temperature-Depth probe (STD) and Sippican expendable bathythermograph (XBT) were used for measuring temperature and salinity. A Niskin 12-bottle rosette sampler coupled with the STD was used to collect water samples for salinity and ox>'gen. The STD system had digital (magnetic tape data logger) and analog chart outputs, as did the XBT system (punched paper tape and analog chart). STD /Niskin casts were normally made to 500 m and XBT drops to 450 m. Nansen bottle casts, for calibration of the STD system, were made at the start and finish of the Part I transect and at internals during Part II fishing operations. On the Part I transect, STD /Niskin stations for temperature, salinity, and oxygen were made every 6 h, with one or more XBT drops between STD stations. In Part II operations, three STD stations were made at night in each 2° x 2° unit area, and five XBT drops between dawn and dusk on fishing tracks (Figure 2). Processing of the 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. physical oceanographic data was as described by Taft and Miller (1970). Depth of the mixed layer was derived directly from analog charts of XBT and STD systems according to criteria of Owen (1970a). Dissolved oxygen content of water samples from the Niskin sampler was determined, and da- ta processed, as indicated by Owen (1970b). Discrete surface samples for chlorophyll a were taken at approximately noon and midnight and processed as described by Owen and Zeitzschel (1970b). Zooplankton hauls were made and samples processed by the methods used on EASTROPAC Expedition (Laurs 1970). Oblique hauls from 200 m to surface were made with nets of 50 cm and 1 m mouth diameter in a paired frame. A wire angle of 45° was maintained during the haul at a speed about 1.5 to 2 knots. In Part II operations, one daylight haul was made during each of 4 days of fishing in a 2° x 2° unit area (see Figure 2), and three night hauls were made during the 4-day period. No hauls were made near dawn or dusk. Data were expressed in displacement volume in milHHters per 1,000 m\ Micronekton was sampled with a net 1.5 m square at the mouth, in oblique hauls from 200 m to surface at a ship speed of 5 knots (Blackburn 1968, 1970; Blackburn et al. 1970). During the Part I transect, micronekton hauls were made at approximately 12-h intervals following STD casts, one during daylight and one at night. In Part II operations, day and night hauls were made with the same frequency as for zooplankton (see Figure 2). Processing of the samples and estimation of volume of water strained was as discussed by the same authors, and total micronekton was expressed as displacement volume in milliliters per 1,000 m\ A variable and generally large proportion of this micronekton consisted of or- ganisms that skipjack are known or likely to eat (skipjack forage) in the eastern tropical Pacific. The micronekton catches were therefore sorted into forage and nonforage organisms (Blackburn and Laurs 1972). Forage organisms were all crus- taceans, all cephalopods, all epipelagic fish and Vinciguerria. Nonforage organisms were all mesopelagic fish except Vinciguerria and all lep- tocephali. The trolling gear used to catch skipjack and other fish was similar to that used in the albacore fishery off the U.S. west coast (Yoshida 1966). Feather jigs were fished with nylon traces and 384 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA nylon parachute cord lines. The Cromwell could fish only 4 lines from the stern, compared with 11 fished by the Jordan, 8 from outriggers and 3 from the stern. On the Jordan the 4 lines on each out- rigger were connected to a set of hydraulic power gurdies for rapid hauling. Charts and sections were contoured by hand ex- cept those of dissolved oxygen content which were prepared by computer and Calcomp plotter. Gear and methods used for neuston and midwater trawl samples are not discussed because the results are not utilized in this report. The same applies to observ^ations on birds, fish schools, and marine mammals. RESULTS OF THE EXPERIMENTAL FISHING Based on the position of the surface current boundaries and distributions of temperature and skipjack forage derived from Part I operations, the latitudinal zones investigated in Part II fishing operations were similar on the 1970 and 1971 cruises: 12°-14°N, 9°-ll°N, 6°-8°N, 3°-5°N, 1°-3°N, and 2°-4°S. Figures 3 and 4 show the cruise tracks during Part II fishing operations in the above-mentioned zones. They also show approximate positions of surface current bound- aries, which were obtained from data on ther- mocline topography. The total fishing effort (number of line-hours) in or immediately adjacent to the study area was very much higher on the November-December 1970 cruise than in March-April 1971, because of the curtailment of the latter cruise (Table 1). The catch by species on each cruise is given in Table 2, which shows the number boarded and kept, tagged and released, and lost but identified, and the overall size range of each species. Skipjack was ob\iously the dominant species on each cruise. DISTRIBUTION AND RELATIVE ABUNDANCE OF SKIPJACK AND OTHER TUNA This section deals with results from the 1970 and 1971 cruises. Relative abundance of skipjack was calculated in terms of catch per line-hour on track. Catch equals number boarded, tagged, and lost but identified. Fish taken when the vessel circled following an initial strike, or when chumming with live ancho\y, are not included. Troll catches of skipjack made on track and separated, arbi- trarily, by > 10 min are considered to have come from separate schools or aggregations of fish, and an index of schools encountered per hour of trolling has been derived. There are highly sig- nificant positive correlations between catch /line- hour and schools/hour for each cruise (r = +0.901 for data of Table 3 and -1-0.716 for Table 4, both significant at the I'^c level). Schools/hour is a more consenative indicator of relative abundance of skipjack than catch/line-hour because each Table 2.-Sumniar>- of fish catch by species: Cruises Jordan ol-Cromirell 51, November-December 1970, and Jordan 60, March-April 1971. Table 1. -Total fishing effort in the study area. Line-hours fished Period Cruise no. in unit areas on passage Nov. -Dec. 1970 Nov.-Dec. 1970 Mar.-Apr. 1971 Jordan 57 2,746-1 3,213 131 Cromwell 51 467 i Jordan 60 2,088 255 Jordan 57-Cromwell 51 Jordan 60 Lost but Size range Lost but Size range Species Boarded 114 Tagged 67 identified (cm) Boarded Tagged identified (cm) Skipjack '-1 18 32.5- 71.0 61 8 49 41.8-70.0 (Katsuwonus pelamis) Yellowfin tuna 30 4 5 26.2-111.5 2 — — 33.5-44.7 (Thunnus albacares) Frigate mackerel 2 — — 30.2- 31.0 1 — — 34.6 (Auxis thazard) Unidentified tuna — — 12 — — — — — Wahoo 4 — 1 52.0-110.0 — — — — (Acanthocybium solandri) Shortbill spearflsh — — 1 — — — — — (Tetrapturus angustirostris) Dolphin 8 — — 20.5-114.5 6 - ~5 34.0-46.3 (Coryphaena hippurus) Pompano dolphin 4 — 2 26.5- — — — — (C. equiselis) -^50.0 Rainbow runner — — — — 1 — — 81.5 (Elagatis bipinnulatus) Unidentified fish (lost) - - 10 - - 10 - 385 FISHERY BULLETIN: VOL. 73, NO. 2 I25°W I20''W I5*W IS'N. lO'N S'N 0-. S'S IS'N (yN .5'N Figure 3.-Cruise tracks, Part II operations, Jordan 51-Cromwell 51, November-December 1970. Thickened lines indicate daylight fishing tracks. Approximate boundaries of surface currents are indicated. I25»W IZO-W 386 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA IS^N ICN- -lO'N 5»N- 0»- 5'S 5°N 5»N Figure 4.-Cruise tracks, Part II operations, Jordan 60, March-April 197L Thickened lines indicate daylight fishing tracks. Approximate bound- aries of surface currents are indicated. 5»S IZO'W 387 FISHERY BULLETIN: VOL. 73, NO. 2 T.'VBLE 3.-Relative abundance of troll-caught skipjack and other tuna. Cruises Jordan ol-Cromwell 51, November-December 1970. Zone latitude Current system Catch/line-hour Schools/hour Skipjack Yellowfin All tuna skipjack 13'-14 = N NEC 0.072 0 0.072 0.10 12--13'N NEC 0.038 0.057 0.101 0.17 10=-11 N NEC 0.064 0.022 0.099 0.17 9°-10=N NECC 0.014 0 0.014 0.08 7'-B'U NECC 0.075 0.004 0.079 0.10 6 -7 N SEC 0.009 0 0.037 0.05 4"-5°N SEC 0.100 0 0.100 0.10 3--4=N SEC 0.130 0.016 0.146 0.25 2"-3=N SEC 0.015 0 0.015 0.06 1=-2-~N SEC 0.354 0 0.354 0.40 0='40'-1-Ni SEC 0.208 0 0.208 0.48 2=-3 = S SEC 0.096 0 0.096 0.13 3 -4 S SEC 0.028 0 0.028 0.10 'Data on this line are from fishing on passage, <5Q line-h. Table 4.— Relative abundance of troll-caught skipjack and other tuna, Cruise Jordan 60, March-April 1971. Data in square brackets are from fishing on passage and are included in totals to the left. Catch/line-h( Dur Zone Current Schools/ hour latitude system Skipjack Yellowfin All tuna skipjack [16-17 N NEC 0.064 0 0.064 0.32] 13^-14 N NEC 0.008 0.008 0.016 0.09 12'-13 = N NEC 0.054 0 0.054 0.08 10'-11=N NEC 0.107 0 0.107 0.46 9-10 = N NEC 0.134 0 0.134 0.17 7"-8 = N NEC 0 0 0 0 6°-7=N NECC 0.031 0 0.031 0.09 -4°-5°N NECC/SEC- -0.042 0.004-0.046 n r»" 3°-4°N SEC 0 0 0 0 2=-3=N SEC 0.023 [0.019] 0 0.023 [0.019] 0.12 [0.10] 1°-2=N SEC 0.067 [0.064] 0 0.067 [0.064] 0.17 [0.17] 2'-3'S SEC 0.062 0 0.063 0.25 3'-4=S SEC 0.039 0 0.029 0.16 aggregation is given equal weighting irrespective of the number of fish caught. The index does not reflect the size of the aggregations, which on both cruises were considered to be relatively small. No large surface schools of skipjack or other tuna were seen in the study area. Tables 3 and 4 show the relative abundance of skipjack, yellowfin tuna and total tuna as catch /line-hour and skipjack as schools/hour for 1° latitudinal zones fished in Part II operations. Approximate boundaries of surface current sys- tems are also indicated: NEC, NECC, and SEC mean North Equatorial Current, North Equatorial Countercurrent, and South Equatorial Current. In November-December 1970 the highest level of relative abundance of skipjack was between lat. 0°40' and 2°N, with a secondary maximum at lat 3° to 5°N and other high levels north of the NECC and south of the Equator. Within the NECC at lat. 7° to 10°N, relative abundance in terms of catch rate was variable, but generally low in terms of schools encountered. With the added contribution of yellowfin and other tuna, there was a marked maximum north of the NECC. In March-April 1971, overall relative abundance was much lower, and the principal maximum was situated north of the NECC, between lat. 9° and 11°N. Secondary maxima occurred at lat. 1° to 3°N and south of the Equator. The relative abundance of skipjack in the NECC was again low. In Figures 5, 6, 7, and 8, daily values of relative abundance (catch rates and schools) are plotted and contoured. The results are more difficult to interpret in this form, but there are some general agreements with the zonally averaged data. Off- track Catches of Skipjack Off-track troll catches of skipjack were made with the use of anchovy, Engraulis mordax, live bait on Jordan 57. On five occasions schools of skipjack were chummed with live bait, following initial jig strikes (4) or fish sighting (1). On two of these occasions the chumming and circling of the vessel produced a substantial additional catch. Use of live bait on cruises of this type is advantageous in order to increase sample size of fish. Distribution of Skipjack by Time of Day Percentages of the total numbers of skipjack schools encountered on track in each 1-h period have been calculated and are given in Table 5. Schools are defined as above. Some 60-min periods included station time (i.e., not fishing), and the numbers of schools per unit time have been ad- justed. Variability in occurrence is considerable between cruises for 1-h periods. However, when presented by 2-h periods, the temporal occurrence of skipjack shows remarkable similarity on the two cruises. Surprisingly, fewest aggregations were encountered before 1000 h. Aggregations were encountered most frequently between 1200 and 1500 h, and again, as expected, in the predusk period, 1700 to 1759 h. Biological Characteristics of Skipjack and Other Tuna Size of Skipjack Measurements of fish length (tip of snout to tip of median caudal fin rays) were made to the nearest millimeter on all fish. Table 6 shows the 388 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA 125° * I20°W h5*W IS«N. CATCH/LINE- HR. iOn*. 5«N 5*S .Km •_„ — ■— • — • — ' — • — • — • .; ! JORDAN CROHWELL .eru S'U ys Hyw Figure 5.— Relative abundance of skipjack in catch/line-hour, cruises Jordan bl-Cromwell 51, November-December 1970. percent of skipjack in three broad size categories for the two cruises. The most significant feature is that 13.3% of skipjack were < 45 cm in November-December 1970, as against 2.9% on the next cruise. The fish < 45 cm were not distributed over a large part of the area, as fish >45 cm were, in November-December 1970. Of the small fish, 16 (85%) were from areas north of the NECC, lat. 10° to 14°N, and the remaining 3 (15%) from south of the NECC, lat. 0°30' to 4°N; none were found in the NECC or south of the equator. Table 7 shows small skipjack ( < 45 cm) as percent of total in the lati- tudinal zones north of 10°N. It appears that the I5'N lO'N IIS'W _1 L_ CATCH/LINE-HR. <0.05 0.05 - 0 I 0 I - 0.2 02-0.4 >0.4 5*N- 0'- 5'S -lO'N I5'N -5*N 5'S I20*W Figure 6.— Relative abundance of skipjack in catch/line-hour, cruise Jordan 60, March-April 1971. Table 5.— Percent' of total number of skipjack schools encoun- tered on track, by 1-h periods (all fishing days combined): Cruises Jordan bl-Cromivell 51, November-December 1970, and Jordan 60, March- April 1971. Time period Jordan 57-Cromwell 51 Jordan 60 Start2-0659 0700-0759 3.7%'» 8.6 J 12.3% 12.2%-» 0 ' 12.2% 0800-0859 2.5 9.4 3.6 9.4 0900-0959 6.9 5.8 1000-1059 1100-1159 8.8 7.2 16.0 10.6 5.6 16.2 1200-1259 1300-1359 14.9 6.3 21.2 2.9 17.9 20.2 1400-1459 1500-1559 7.5 7.3 14.8 15.8 5.8 21.6 1600-1659 1700-1759 7.3 19.0 26.3 5.4 14.4 19.8 Total fishing days 28 21 'Adjusted data, see text p. 388. ^Mean start 0556 h and 0539 h on Jordan 57-Cromwell 51 and Jordan 60 respectively. 389 FISHERY BULLETIN: VOL. 73, NO. 2 125° W i2crw lio** IS'N lont. 5*N. 0*. S-S SCHOOLS /HR <0I =^ 01-02 02-04 >0.4 JORDAN CROUWCLL 125* V» (TN KTH snt .0* .S'S Figure 7.— Relative abundance of skipjack in schools/hour, cruises Jordan 57-Cromwell 51, November-December 1970. fish < 45 cm were largely segregated geograph- ically from the medium and large ones. Figure 9 indicates a distinct separation in age as well. Small skipjack were very scarce in March-April 1971. Figure 9 shows the percent length-frequency distributions of skipjack by 2-cm classes. In November- December 1970, the principal mode was at 58 cm, with a minor mode at 36 cm and perhaps another at 48 cm. The fish of modal size 36 cm could be about 14 to 15 mo old, if one accepts the growth rates for juveniles indicated by Yoshida (1971) and Joseph and Calkins (1969: from tagging data, averaged). This would suggest a spawning origin I5*N lO'N 5'N 5'S SCHOOLS/HR. <0.l 0I-02 O2-0.4 >0.4 ^ I5*N lO'N 9*N 5'S 120'W II5*W Figure 8.— Relative abundance of skipjack in schools /hour, cruise Jordan 60, March-April 1971. in the northern summer. Because no definite in- formation is available on skipjack growth rates beyond this size, the age represented by the 58-cm mode is uncertain, but probably it is 2 to 3 yr. Whether the possible mode at 48 cm represents fish 1 yr or 6 mo between the other two modes is not known. If there were 6-mo difference, it would signify a spawning origin in the southern summer. In March- April 1971, there is a single wide mode composed of fish >48 cm (peak 56 to 60 cm). The fish had about the same size distribution in all areas. Skipjack in the eastern Pacific coastal fishery ranged from about 3 to 3.5 kg, with mean and modal lengths 50 to 55 cm, in the years 1955-71 (Miyake 1968; Inter-American Tropical Tuna 390 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA Table 6.-Skipjack in size categories as percent of total, Cruises Jordan 57-Cromwell 51, November-December 1970 and Jordan 60, March-April 1971. Size Percent of total (cm) Jordan 57-Cromwell 51 Jordan 60 <45 45-60 >60 13.3 62.2 24.5 2.9 55.1 42.0 Total skipjack 143 70 Table 7.-Small skipjack (<45 cm) as percent of total in lati- tudinal zones north of 10°N, Cruises Jordan ol-Cromwell 51, November-December 1970. Zone Size (cm) No. of fish latitude Catego ry Range Mean Percent 12=-14'N 10'-11°N <45 >45 <45 >45 32.5-39.8 48.8-56.5 33.4-34.7 45.0-65.0 35.5 54.2 34.2 58.0 13 4 3 19 76.5 23.5 13.6 86.4 >- o z UJ o 25 20 15 10 5 0 25 20 15 10 5 0 JORDAN -57 AND CROMWELL -51 NOVEMBER - DECEMBER 1970 n = l43 JTT-^ffHJ" Pn-ri I I I I I I I ' I 1 I I I JORDAN -60 MARCH -APRIL 1971 n= 70 a I I I I I I I I' I ■ I 'I ■! ■ I ■! "I'l "ri "Ti "I "I 'Ti I ■! I I I I I I I ' I 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 86 90 2 cm FL CLASSES Figure 9.-Skipjack percent-length-frequency distribution in the study area. Smoothed curves are from 3-figure moving averages. Stated length indicates midpoint of class. Commission 1966, 1972). The principal component of the Hawaiian catch usually consists of fish of modal sizes > 60 cm (U.S. Bureau of Commercial Fisheries 1963; Rothschild 1965; Higgins 1966). The principal modes of skipjack caught in the study area, 56 to 60 cm on the two cruises, are intermediate between those in the eastern Pacific coastal fishery and the Hawaiian fishery. Purse seine samples of skipjack obtained in and near the study area in 1970 and 1971, which are mentioned later, had mean sizes from 58 to 61 cm. Size of Other Tuna Only 34 yellowfin tuna were boarded (39 caught) on the November-December 1970 cruise, and 2 in March-April 1971. None were taken south of lat. 3°N. Yellowfin ranged from 26 to 112 cm, with mean lengths for different aggregations ranging from 28.5 to 42.3 cm. The majority (68%) were < 45 cm. These small yellowfin were mainly (83%) from the same latitudinal zone (10°-14°N) as the small skipjack. On three occasions schools of mixed small tunas-skipjack, yellowfin, and frigate mackerel {Auxis: 30 to 31 cm)-were sampled north of lat. 10°N. Small tuna of different species may occur together because of similar environmental and food requirements, and behavior. Sex and Maturity of Skipjack Sex ratios of skipjack for the two cruises were as follows: November-December 1970: Total 143, Sexed 72 Males 24; females 33; indeterminate 15 Ratio: Males to females, 1:1.4 March- April 1971: Total 72, sexed 59 Males 24; females 35 Ratio: Males to females, 1:1.5 Gonad maturity was determined macroscopically in the field, and stages were classified as follows: Immature, virgin Roughly Immature, resting equivalent Maturing to indicated Spent stages in Spent-recovering Orange (1961) 1-S 1 2 5-A 5-B The number of gonads in each stage by size of fish are given in Tables 8 and 9. In November-December the smallest spent or spent recovering female skipjack was 46.5 cm. In March-April the corresponding size for females was 49.8 cm, and for males 48.3 cm. These data support other evidence that first maturity in female skipjack in the Pacific is reached between 40 and 45 cm (Orange 1961; Waldron 1963; Kawasaki 1965). A relatively large number of recently spawned fish, i.e., with spent and spent-recovering gonads, was taken on each cruise: 30% of the females and 7% of the males in November-December 1970, and 391 FISHERY BULLETIN: VOL. 73, NO. 2 Table 8.-Number of skipjack gonads in each maturity stage by size of fish, Cruises Jordan 57-Cromwell 51, November- December 1970. Imm ature Mat uring Spent Spent Size class Virgin Resting recovering (cm) M F M F M F M F M F <40 ? 15 ? 40-49.9 1 5 1 1 50-59.9 5 12 1 1 2 6 60-69.9 11 2 2 >70 2 Totals: males (M) 24; females (F) 28; Indeterminate (<40 cm) 15. Table 9.-Number of skipjack gonads in each maturity stage by size of fish, Cruise Jordan 60, March-April 1971. Immature Maturing M F Spent M F Spent Size class (cm) Virgin Resting M F M F recovering M F <40 40-49.9 50-59.9 60-69.9 >70 2 1 1 2 1 10 5 1 13 8 Totals: males (M) 24; females (F) 34. 63% and 67%, respectively, in March-April 1971 (Tables 8, 9). Most others were at the immature resting stage. The spent-recovering fish appeared to be at a more advanced stage of recovery in March- April than in November-December 1970. Data on other skipjack ovaries taken from or near the study area in 1970-71 are available from samples of purse seine-caught fish examined by the Inter-American Tropical Tuna Commission (C. L. Petersen, pers. commun.). Gonad indices were calculated by them according to Orange (1961), and data are given in Table 10. The principal interest lies in the samples (mean lengths >57 cm) from long. 110°W westwards. It is assumed that gonad indices < 15.1 indicate immature (virgin or resting) or spent-recovering ovaries, 15.1 to 45 indicate maturing ovaries, and > 45 indicate mature or ripe ovaries. The oc- currence of 35% of skipjack ovaries with indices > 45 in the middle of the study area in September 1970 may correspond to the 36% of ovaries in spent condition in fish > 50 cm in the same area in November-December 1970 (Table 8). The sample from lat. 5°N, long. 110°W in April 1971 (Table 10) showed similar ovarian states to those in skipjack > 50 cm caught in March-April 1971 (Table 9): about 75% immature or spent-recovering, and 25% maturing, in each case. Sex and Maturity of Yellowfin In November-December 1970, 24 specimens of yellowfin under 50 cm were classified as immature, sex indeterminate; of 5 yellowfin from 50 to 60 cm, 3 were immature female, 1 was immature male, and 1 was indeterminate. OBSERVATIONS ON THE ENVIRONMENT IN RELATION TO SKIPJACK Temperature The data from Part I operations on the 1970 and 1971 cruises were used to construct temperature Figure lO.-Temperature (°C) section from lat. 15°N to 5°S along long. 119°W, cruise Cromwell 51, 1-7 November 1970. Table lO.-Gonad indices of female skipjack caught by purse seine in northeastern tropical Pacific, 1970-71. Date of Approximate position Gonad indices Samolc ^'"^^ *'^"^' capture Lat. '^N Long. °W % <15.1 %15.1-45 % >45 no. Range Mean 23 August 1970 8 September 1970 April 1971 30 April 1971 May 1971 135 120 110 90 90 48.9 2.5 76.0 57.1 32.0 46.8 62.5 24.0 42.9 60.0 4.3 35.0 0 0 8.0 50 40 50 49 59 53.5-63.8 55.0-62.3 54.3-67.6 53.3-68.7 57.8-70.0 59.9 57.8 60.5 62.3 64.5 392 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA 200 m Figure 11. -Temperature (°C) section from lat. 15°N to 5°S along long. 119°W, cruise Jordan 60, 5-11 March 1971. I25°W IS^N. 10^. 5»N. 5'S <22 <22 • JOnOAN » CROMWELL >23 li5«W I25'W I20°W tsru KTH S*N .9-S Hsrw I5°N IIS'W lO'N 5'N 5-S "' if <26 I20*W "T 1 1 1 1 r- II5*W I5*N lO'N ■S'N •0* 5*S Figure 13.— Surface temperature (°C) during Part II operations, cruise Jordan 60, March- April 1971. sections (Figures 10, 11). The temperature dis- tributions were generally similar to those ob- served on the same meridian at the same time of year during EASTROPAC Expedition (M. Tsuchiya, pers. commun.). Surface temperatures suitable for skipjack, i.e., between 20° and 29°C, occurred at all latitudes on both cruises. The distribution of surface isotherms (assumed synopticity) during Part II (fishing) operations on both cruises is shown in Figures 12 and 13. These figures have been compared, using overlays, with all charts of relative abundance of skipjack Figure 12.-Surface temperature (°C) during Part II operations, cruises Jordan bl-Cromwell 51, November-December 1970. 393 FISHERY BULLETIN: VOL. 73, NO. 2 (Figures 5, 6, 7, 8). The only apparent relationship between surface temperature and skipjack abun- dance appears to be the high abundance in the western part of the area of strong temperature gradient at lat. 1° to 3°N in November-December 1970. Mixed Layer Depth The depths (meters) of the bottom of the mixed layer are contoured at 20-m intervals in Figures 14 and 15 for Part II of the cruises. Figures 10 and 11 show the depths on Part I. Even though data are zonally discontinuous in parts of Figures 14 and 125° W 120** II5*W 19^ 9*H 0». 9^ .en ":u 20 ~Q' .KTU '2°-Cs^. >60 • JOROaN • CROHWILL 60 40 <46/ ' ' ^^^-^^0 izvw .9*»( 9*S HTW Figure 14.-Depth (m) of upper mixed layer, Part II operations, cruises Jordan bl-Cromwell 51, November-December 1970. 15, they are generally consistent with those of Cromwell (1958), Wyrtki (1964), and Love (1971b, 1972a, b, in prep.: EASTROPAC data). In November-December 1970, the mixed layer depth was shallow ( < 40 m) north of lat. 9°N, but south of there increased rapidly to > 100 m in the region of lat. 5°N. This ridge and trough are to be expected at the approximate northern and southern boundaries of the surface NECC. The gradient of change from this trough southwards to another ridge was particularly intense in the east- ern edge of the area around lat. 4°N. In March- April 1971, the mixed layer was very shallow over most of the area surveyed south of lat. 10°N, becoming < 10 m in the region lat. 3° to 4°N. The even depth of the mixed layer from lat. 4° to 10°N IS'N lO'N 5'N 5'S 40 40 40 >40 I5*N 5*N -0« 5'S I20'W II5*W Figure 15.-Depth (m) of upper mixed layer, Part II operations, cruise Jordan 60, March-April 1971. 394 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA indicates the extreme weakness of the surface NECC at this time. The charts of mixed layer depth have been com- pared by overlay with those of relative abundance of skipjack (Figures 5, 6, 7, 8). Areas of high rela- tive abundance in November-December were generally close to ridges ( < 40 m) in the mixed layer depth. There were exceptions, some of which are less obvious when utilizing schools/hour. There is no such trend in the March- April data. Oxygen Oxygen content (milliters/liter) was measured to 500 m, and sections showing oxygen distribu- tions along the Part I transects of both cruises are presented in Figures 16 and 17. Sampling was also carried out in Part II fishing operations but for various reasons was more restricted than planned. Figure 16 shows a strong oxycline (2 to 4 ml /liter) throughout the section of November 1971 parallel to the thermocline (Figure 10). In view of the probability of the lethal level of oxygen for skipjack being about 2.4 to 2.8 ml/liter (Anonymous 1973; R. Lasker pers. commun.), the depth of the 2.5- ml /liter isopleth may delimit the region of the water column suitable for skipjack. Its shallowest level in November-December 1970 (Figure 16) was about 50 m. In March-April 1971, (Figure 17) the oxycline was weaker and much reduced just north of the equator. The 2.5-ml/liter isopleth was on the average a little shallower than in November-December. The strong thermo- oxycline over most of the area on both cruises probably represented a bottom limit to vertical movement of skipjack, but even so at least the top 50 m of water were available to the fish. The ridge in the oxycUne at about lat. 10°N ap- pears to be the western extension of the upper edge of the low oxygen-content water mass stretching out from the coast of Central America. (Tsuchiya 1968, Figure 7a, oxygen on the gr = 400 cl/T surface, which there is at < 100 m). Currents The positions of the boundaries between the North Equatorial Current (NEC), the surface North Equatorial Countercurrent (NECC), and the South Equatorial Current (SEC) have been in- dicated schematically in Figures 3 and 4. They were based on the slope of the thermocline during Part I and II operations. In November-December 1970 the surface NECC was confined to a narrow band between lat. 7° and 10°N, with a geostrophic flow of < 1 knot (44 cm/s). On the second cruise in March-April 1971 the surface NECC was located between lat. 4° and 8°N during Part I operations, but a short time later in Part II it had narrowed to between lat. 4°30' and 7°N. The geostrophic flow was lower than on the previous cruise, < 0.5 knot (< 25 cm/s), except at the southern boundary, where the subsurface NECC may have surfaced and geostrophic flow was about 1.25 knots (65 cm/s). Average current charts for the area (Wyrtki 1965) show the surface NECC absent east of long. 120°W at this time of year. EASTROPAC data (Love 1971b, 1972a; M. Tsuchiya, pers. com- mun.) show the surface current absent at this meridian and time in 1967, but present in 1968. Data from XBT records of the return passage of the Cromwell from the study area to Hawaii in November 1970 have been used to construct a diagonal temperature section from lat. 3°N, long. 124°20'W to lat. 16°30'N, long. 146°06'W (Figure 18). From the slope of the thermocline, the approximate boundaries of the surface NECC along the transect are defined as lat. 6° to 8°N (southern boundary) and lat. 10° to 11°N (northern boundary). Chlorophyll In November-December 1970, the range of sur- face chlorophyll a values was 0.03 to 0.22 mg/m^ with a maximum ( > 0.20) between lat. 0°30' and 2°00'S. A small area with chlorophyll values > 0.20 also existed at about lat. 2°30'N, long. 119°W. An area of low chlorophyll ( < 0.05) was located at lat. 9° to 10°N, long. 117° to 119°W. During March-April 1971, surface chlorophyll ranged from 0.03 to at least 0.25 and probably to about 0.40 mg/ml Maxima ( > 0.20) occurred from lat. 9° to 11°N and lat. 13° to 14°N, and a minimum ( < 0.05) occurred from lat. 5° to 7°30'N, all east of long. 118°W. Zooplankton All four sets of zooplankton data (1-m and 0.5-m nets, day and night) from Part II operations show broadly similar distributions for the same cruise on contour charts, and it is unnecessary to show them for both nets. The 0.5-m net catches probably give a better representation of the standing stock of small herbivores than the 1-m net catches, and 395 FISHERY BULLETIN: VOL. 73, NO. 2 LATITUDE 5N 100 S 200 I I- 0. iij a 300 300 400 500 Figure 16.— Dissolved oxygen content (ml/liter), Part I transect, cruise Cromwell 51, 1-7 November 1970. their distributions are charted here (Figures 19 to 22). Catches by the 1-m net are more likely to be related to skipjack tuna than those by the 0.5-m net, because more skipjack forage organisms occur in them. These relationships are investigated sta- tistically in the next section. The main features of the day and night zooplankton distributions in November-December 1970 are a maximum at about lat. 1° to 2°N, a minimum in the extreme north of the area, and a secondary minimum at about lat. 3° to 5°N. Maxima occurred in March-April 1971 at about lat. 3° to 6°N and 9° to 10°N; elsewhere at that period the catches were moderate with no conspicuous spatial minima except in the extreme north. Com- parisons between Figures 19 to 22, and Figures 5 to 8, made by overlay, show no obvious relation between distributions of zooplankton and skip- jack. Skipjack Forage (and Zooplankton in Part) One of the objectives of the cruises was to see if availability of offshore skipjack varied with the forage, as presumed by Blackburn (1965, 1969) and Blackburn and Laurs (1972); or with zooplankton, as suggested by Schaefer (1961); or with the arithmetic product of forage and zooplankton, as suggested by Riley (1963). Charts of skipjack forage (in milliters/ 1,000 m'), both day and night data, are shown in Figures 23 and 24 for both parts of the November-December 1970 cruise, and similarly in Figures 25 and 26 for the March- April 1971 cruise. Corresponding charts of total micronekton were very similar and are not given here. Data are available at Southwest Fisheries Center, National Marine Fisheries Ser- vice, La JoUa, Calif. Measurements of day concentration of forage were more numerous than those of night concen- tration. Thus Figures 23 and 25 show more detail than Figures 24 and 26. Blackburn and Laurs (1972) showed that day and night distributions of forage were broadly similar in the area on a given cruise as far as locations of maxima and minima were concerned, although the night concentrations were about 10 times higher than day concentra- tions. Contours for day and night distributions were therefore drawn to agree with each other as 396 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA LATITUDE 5N en cr LlI I- UJ I H Q. Figure 17.— Dissolved oxygen content (ml/liter), Part I transect, cruise Jordan 60, 5-11 March 1971. Om 50 m b°N IO°N I5°N 100 m h- ISOtn - 200 m - 250 m 300 m 350 m 10 10 5"^ lO'N IS'N Figure 18. -Temperature (°C) section from lat. 3°N, long. 124°20'W, to lat. 16°30'N, long. 146°06'W, cruise Cromwell 51, November 21-26, 1970. much as possible without violating the data. Fewer data were available for the March-April cruise than for the November-December one, for reasons given elsewhere. According to Blackburn and Laurs (1972), the most conspicuous and consistent feature of day and night forage distributions in the study area is a zonally oriented maximum between lat. 0° and 5°N, which is probably associated with the equa- torial upwelling as explained by King (1958). Figures 23 and 24 show this feature, and Figures 25 and 26 show one which is probably the same although it does not appear to be zonally oriented. Stations at the eastern end of the maximum in Figures 25 and 26 were occupied 13 days after the last stations at the western end were occupied, and the maximum could have moved north in the in- terval. Other maxima and minima in Figures 23 to 26 are smaller and less consistent in location between cruises. The data from these micronekton net hauls, es- pecially those for the March- April cruise which are very sparse, may not give a complete picture of the distribution of skipjack forage. On the March- 397 FISHERY BULLETIN; VOL. 73, NO. 2 125° * 120** io"v» I20»W II5-W IS-N. KTH 5'N O*. ys ZOOPLANKTON ( ml/ 1000 m^ ) <50 50-100 100-200 >200 .KTH • JOROtN o cnouwCLi. aru S*N .9*S Hsrw Figure 19.-Day standing stock of zooplankton by 0.5-m net (ml/1,000 m'), cruises Jordan bl -Cromwell 51, November- December 1970. April cruise large catches of skipjack forage were obtained by midwater trawl as well as by micronekton net in the area of the maximum between lat. 0° and 5°N, and one was obtained also at about lat. 10°N where the micronekton net in- dicated a rather low concentration. This result from a single trawl haul thus indicates an area of rich forage of unknown extent at about lat. 10°N, which the sparse data from the micronekton net hauls do not show. The forage may have been patchy in this area. The expected resemblance between the spatial distributions of forage and skipjack is not strongly lO'N 5*N- 5'S ZOOPLANKTON ( ml/ 1000 m3 ) <50 50-100 100-200 200-400 >400 I6'N 9*N 5*S II5*W Figure 20.— Day standing stock of zooplankton by 0.5-m net (ml/ 1,000 m'), cruise J&rdan 60, March-April 1971. evident when Figures 23 to 26 are compared with charts of skipjack availability on the same cruises (Figures 5 to 8). However, Table 3 shows that mean catch per line-hour and mean number of schools per hour for all fishing days in a zone of latitude was highest in November-December from lat. 1° to 5°N, where the principal forage maximum was located (Figures 23 and 24). Table 4 shows that the same indices were highest in March- April from lat. 9° to 11°N, where Figures 25 and 26 show no forage maximum, although one may have existed there as explained above. A secondary skipjack maximum occurred in March- April in the equatorial region, south of lat. 3°N, where a forage maximum was present. Attempts were made to correlate forage con- 398 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA 126° W I20*W IS-N. II5*V» ZOOPLANKTQN ( ml/ 1000 m3 ) 10^. 200 9*N aru .urn 8*N • JORDAN O CROMWII.L s-s 125* W I20*W .9*S Figure 21.-Night standing stock of zooplankton by 0.5-m net (ml/1,000 m"), cruises Jordan 5T-Cromwell 51, November- December 1970. centration and related measurements with skip- jack availability for all data from the study area. Table 11 gives the results of the tests using con- centrations of day forage, day zooplankton, and their arithmetic product. Table 12 gives similar results using night concentrations. All variables were transformed to logarithms in order to bring distributions closer to normal, before correlation coefficients were calculated. A distinction is made between all skipjack and large skipjack; the latter excludes skipjack < 45 cm, which seem to be a separate age-group and exhibited some segrega- tion from the other skipjack in space and time. I5*N I20"W II5«W -1 — 1 1 ] i_ ZOOPLANKTON ( ml/ 1000 m3 ) lO'N- <25 25-50 50-100 (00-200 >200 5'N 5'S- I5*N lO'N 5*N 120'* 5*S II5*W Figure 22.-Night standing stock of zooplankton by 0.5-m net (ml/ 1,000 m'), cruise Jordan 60, March-April 1971. Zooplankton data are from hauls of the 1-m net only, as explained above. None of the 72 correlations in Table 11, involv- ing day concentrations of forage and zooplankton with skipjack, are significant. On the other hand 4 of the 48 correlations in Table 12, involving night concentrations of forage and zooplankton with skipjack, are significant by the usual criteria and positive: two coefficients are above the 5% level of probability and two are above the 1% level. They refer only to availability of large November- December skipjack measured as catch per line- hour or schools per hour, in relation to night forage and to the product of night forage and night zooplankton, with both variables averaged over 399 FISHERY BULLETIN: VOL. 73, NO. 2 125° W I2(rw II9*W 125° W I20*W I5°N y* ■ IO°N 5°N 125' W 5°S II5°W Figure 23.-Day standing stock of skipjack forage (ml/ 1,000 m') (combined Part I and II data), cruises Jordan bl-Cromwell 51, November-December 1970. zones about 2° of latitude wide. There is no difference in the significance of coefficients depending on whether forage or forage x zooplankton was the variable, and no significant coefficients are obtained with zooplankton alone. Table 13 gives the data that yielded the sig- nificant correlations between large skipjack and forage. The correlation coefficients are +0.947 with catch per line-hour and + 0.886 with number of schools per hour, significant at the 0.5 and 2.0% probability levels respectively. The corresponding Spearman rank correlation coefficients are + 0.952 and +0.905, both significant at the 5% level. No other grouping of 2°-latitude zones would have IS'N. ■lb°N I0°N 125' W IZO'W Figure 24.-Night standing stock of skipjack forage (ml/ 1,000 m') (combined Part I and II data), cruises Jordan bl-Cromwell 51, November-December 1970. given so many zones with so much data in each (see Figure 3). The significance of the four correlation coefficients in Table 12 has been disputed because of the much larger number of nonsignificant correlations in Tables 11 and 12 combined. It has also been pointed out that the two coefficients in- volving forage x zooplankton are not independent of the two coefficients involving forage alone. In our following comments we ignore all coefficients with forage x zooplankton, whether apparently significant or otherwise. We then have two possi- bly significant coefficients in a total of 80 for Ta- bles 11 and 12, i.e., one in 40. From the previous paragraph, there is a chance of about one in 50 that 400 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA I20°W II5"W _i 1 1 1_ FORAGE ml/ lOOOm^ IO°N- < 25 .25-5 5 -10 >l 0 5'N- 5"S I20*W Figure 25.-Day standing stock of skipjack forage (ml/1,000 m') (combined Part I and II data), cruise Jordan 60, March-April 197L I20°W 15-N I5"N II5"W lO'N lO'N 5*N 5'N FORAGE ml/ 1000 m^ <25 25-50 50-10.0 >I00 5«S IB'N IO*N ■5*N -4-5*S I20»W Figure 26.-Night standing stock of skipjack forage (ml/ 1,000 m^) (combined Part I and II data), cruise Jordan 60, March-April 1971. the lower coefficient could have been so high with no correlation, and a corresponding chance of only about one in 200 for the higher coefficient. There seems no reason to presume nonsignificance, at least for the higher coefficient. Significant correlations were not obtained for any data from the March-April cruise. Only two degrees of freedom are available to test the sig- nificance of correlations corresponding to the sig- nificant correlations in the November-December data. In any case the meager forage data for March- April seem not to have adequately depicted the actual conditions, as noted above. Relations of skipjack to forage are discussed later. CRUISE 116 OF RV CHARLES H. GILBERT, OCTOBER-NOVEMBER 1969 In October-November 1969, the RV Charles H. Gilbert of the Honolulu Laboratory of the National Marine Fisheries Service made a cruise to collect samples of skipjack and other tunas for a sub- population study. The main fishing operations were located in the same area as the combined cruise Jordan bl-Cromwell 51, though approx- imately 1 yr earlier. The fishing and environmen- tal data are analyzed here in more detail than by Hida (1970). Figure 27 shows the cruise track of 401 FISHERY BULLETIN: VOL. 73, NO. 2 Table 11. -Measures of daytime availability of skipjack correlated with day concentrations (ml/l.OOO m') of skipjack forage, zooplankton, and their arithmetic product. All tests were made with the variables transformed to logarithms. Numbers are pairs of observations from which correlation coefficients were calculated. No coefficients were significant. "AH" means all skipjack; "large" means skipjack >45 cm. Skipjack Data Forage Zooplankton availability Period grouping' (F) (Z) F X Z Catch/line-hour All Large Schools/hour All Large Nov. -Dec. A 35 33 33 B 11 10 10 C 6 6 6 Mar. -Apr. A 16 16 16 B 4 4 4 C 6 6 6 Nov.-Dec. A 35 33 33 B 11 10 10 C 6 6 6 Mar. -Apr. A 16 16 16 B 4 4 4 C 6 6 6 Nov.-Dec. A 35 33 33 B 11 10 10 C 6 6 6 Mar.-Apr. A 16 16 16 B 4 4 4 C 6 6 6 Nov.-Dec. A 35 33 33 B 11 10 10 C 6 6 6 Mar.-Apr. A 16 16 16 B 4 4 4 C 6 6 6 'A, Numbers to the right of this letter are numbers of 1° X 1° quadrants. For each quad- rant skipjack availability is based on all observations for the day, and F and Z are each based on a single observation made in the quadrant during the day. B. Numbers to the right of this letter are numbers of 1° zonal row/s of quadrants with > 2 quadrants per row. For each row skipjack availability, F and Z are means of the data for the individual quadrants. C. Numbers to the right of this letter are numbers of 2' zonal rows of quadrants. For each row skipjack availability, F and Z are means of the data for the individual quadrants. Table 12.-Measures of daytime availability of skipjack correlated with night concentra- tions (ml/1,000 m^) of skipjack forage, zooplankton, and their arithmetic product. All tests were made with the variables transformed to logarithms. Numbers are pairs of observa- tions from which correlation coefficients were calculated. Significant correlations occurred as shown by * (5% level of probability) or ** (1% level) and were positive. "AH" means all skipjack; "large" means skipjack > 45 cm. Skipjack Data Forage Zooplankton availability Period grouping' (F) (Z) F X Z Catch/line-hour All Large Schools/hour All Large Nov.-Dec. A 25 25 25 B 6 6 6 I^ar.-Apr. A 11 11 11 B 4 4 4 Nov.-Dec. A 25 25 25 B 6** 6 6** Mar.-Apr. A 11 11 11 B 4 4 4 Nov.-Dec. A 25 25 25 B 6 6 6 Mar.-Apr. A 11 11 11 B 4 4 4 Nov.-Dec. A 25 25 25 B 6* 6 6* Mar.-Apr. A 11 11 11 B 4 4 4 'A. Numbers to the right of this letter are numbers of night stations, one station per night, at which observations of F and Z were made. F, Z, and F X Z for a night station were paired for correlation purposes with the mean of skipjack availability for the daytime periods im- mediately before and after the night on which the station was occupied, in closely adjacent r X 1° quadrants. B. Numbers to the right of this letter are numbers of 2° zonal rows of quadrants and night stations. For each row skipjack availability is the mean of the data for the individual quadrants, and F and Z are the means of the data for the individual night stations. 402 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA Table 13.-Means of night skipjack forage and availability of large skipjack for 2° zonal rows of quadrants and night stations in November-December 1970, on which significant correlations in Table 12 are based. lecw I40OW 120^~^ 1 1 1 1 1 20-N lecw wcw I20''W Figure 27.-Track of cruise Charles H. Gilbert 116, October- November 1969. Noon positions are indicated. Area of present investigations is outlined by dashed lines. NECC is surface North Equatorial Countercurrent. Figure 28.-Temperature (°C) section from lat. 19°55'N, long. 156°36'W to lat. 4°08'S, long. 133°40'W, cruise Gilbert 116, 3-13 October 1969. surface currents. There was relatively little fishing effort ( < 50 line-hours) in some latitudinal zones, which makes it difficult to define limits of maxima of skipjack abundance. However, the area from lat. 6° to 8°10'N is a maximum, with indices com- parable with the highest ones in Table 3. Moder- ately high indices are seen at lat. 5° to 6°S and in the area lat. 0°-6°N, alternating with areas fished for less than 50 line-hours. Catch indices of other tunas were moderately high in the two zones where they occurred. Relative abundance of skipjack as schools /hour in the area is given in Table 16 in two ways: firstly, as in previous sections of this report based on troll catches, and secondly, based on schools encoun- tered during trolling and pole-and-line fishing. Usually, pole-and-line fishing was carried out sub- sequent to a jig strike, but occasionally not. The 403 FISHERY BULLETIN: VOL. 73, NO. 2 Om 50 m 100 m 150 m 200 m 250 m - 6°N S-N I0°N I2°N I4°N I6°N I8°N 20°^ Figure 29.-Temperature (°C) sections from lat. 8°05'S to 5°21'N along long. 117° to 120°W, and from lat. 5°45'N, long. 120°14'W to lat. 21°03'N, long. 157° 14'W, cruise Gilbert 116, 19 October-7 November 1969. Table 14.-Tota] fishing effort Cruise Gilbert 116, October 1969. Locality Line-hours fished Study area Outward track Inward track 963 Total 1,851 Table 15.-Relative abundance (catch /line-hour) of troll-caught skipjack and other tuna in the study area. Cruise Gilbert 116, October 1969. YF and BE mean yellowfin and bigeye. * means <501ine-h. Zone latitude Current system Skipjack catch/ line-hour Other tuna catch/line-hour T'-S^IO'N NECC 6=-7° NECC 5°-6° SEC 4°-5° SEC 3'-4° SEC 2°-3° SEC 1°-2° SEC 0=-1°N SEC 0°-1°S SEC 1°-2° SEC 2°-3° SEC SM" SEC 4°-5° SEC 5°-6° SEC 6°-7° SEC 7°-8 S SEC 0.240 0.342 0* 0.065 0* 0.087 0* 0.095 0* 0* 0* 0* 0 0.097 0.017 0.017 0 0 0 0 0 0.138 (YF) 0 0 0 0 0 0 0.093 (YF or BE) 0 0 0 correlation between catch/line-hour and schools/hour was significant at the 1% level (r = 0.908, data of Tables 15 and 16), as on the cruises made in 1970 and 1971. The covariance analysis in Table 17 shows that the regressions of schools/hour on catch/line-hour for the three cruises did not differ significantly in slope, but differed in elevation. Table 16.-Relative abundance (schools/hour) of skipjack in the study area. Cruise Gilbert 116, October 1969. *means < 9 h of observations Schools/hour Zone latitude Current system Based on troll catch Based on total catch (troll plus pole-and-line) 7°-8°10'N NECC 6='-7° NECC 5°-6° SEC 4°-5° SEC 3°-4° SEC 2°-3° SEC 1'-2° SEC 0='-1°N SEC 0°-1°S SEC 1°-2° SEC 2"-3° SEC 3°-4° SEC 4°-5° SEC 5°-6° SEC 6°-7° SEC 7°-8°S SEC 0.42 0.50 0* 0.31 0* 0.21 0* 0.34 0* 0* 0* 0* 0 0.15 0.05 0.08 0.42 0.50 0* 0.39 0* 0.26 0* 0.34 0 0* 0* 0* 0.22 0.15 0.05 0.08 Table 16 shows the maximum at lat. 6° to 8°10'N, as in Table 15. Inclusion of the pole-and- hne data increases the indices in other zones, i.e. lat. 4° to 5°N, 2° to 3°N, and 4° to 5°S. The range of the indices for schools /hour is about the same as for the other cruises. Considerable catches of skipjack and other tuna were made on passage to and from the area. Data on relative abundance (catch/line-hour) are given in Table 18. Results should be used with care as fishing effort never exceeded 50 line-hours per day. There were moderate to high catch rates between lat. 4°45' and 8°30'N (long. 148° to 138°W), with indices at about the same level as at lat. 6° to 8°10'N in the study area (Table 15). Catch indices for other tuna were low. 404 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA Table 17.-Analysis of covariance: schools/hour (7) on catch /line-hour (X) for the 1969, 1970, and 1971 cruises, data of Tables 3, 4, 15, and 16. Year df Sx^ 2xy Sy^ b df Srf2 I^.S. 1969 7 0.0905 0.1144 0.1754 1.264 6 0.0308 0.0051 1970 11 0.0936 0.0904 0.1060 0.965 10 0.0187 0.0019 1971 11 0.0187 0.0402 0.1684 2.147 10 0.0822 0.0082 Within 26 0.1317 0.0051 Reg. Coet. 2 0.0223 0.0112 Common 29 0.2028 0.2450 0.4498 28 0.1540 0.0055 Adj. Means 2 0.0410 0.0205 Total 31 0.2286 0.2807 0.5394 30 0.1950 F (slope) = 2.196, nonsignificant. F (elevation) = 3.727, significant at 5%. Table 18.-Relative abundance (catch/!ine-hour) of troll-caught skipjack and other tuna, on track to and from the study area. Cruise Gilbert 116, October-November 1969. YF and BE mean yellowfin and bigeye. Approximate position Fishing effort (line-hour) (start and finish of Latitude day's trolling) Longitude Current system Catch/li ine-hour Date Skipjack Other tuna Outward track: 4 Oct. 16°15' -17°30'N 153° -154°15'W NEC 24 0 0 5 14° -15°15'N 150°30'-151°45'W NEC 24 0.043 0.043 (YF) 6 12' -13°15'N 148°30'-149°30'W NEC 23 0 0 7 9°30' -10°45'N 147° -147°45'W NEC/NECC 23 0.088 0 8 7°15' - 8°30'N 144°30'-145°45'W NECC 23 0.216 0 9 4°45' - 6°15'N 142°15'-143°30'W NECC/SEC 24 0.336 0 10 2°15' - 3°30'N 139°45'-141°W SEC 24 0.083 0.042 (YF) 11 1°15'N- 0°15'S 137°15'-138°15'W SEC 24 0.083 0 12 1°30' - 2°30'S 134°45'-136°W SEC 24 0 0 13 3°45' - 5°S 133° -133°45'W SEC 48 0.042 0 14 5°S 130°15'-131°45'W SEC 50 0.040 0 15 5°S 127°15'-128°45'W SEC 48 0.063 0 Inward track: 30 Oct. 8° - 9°N 127°15'-129°W NECC 48 0 0 31 9° -10°N 130°45'-132°45'W NEC 48 0 0 1 Nov. 10°45' -11°45'N 134°30'-136°30'W NEC 48 0.021 0 2 12°30' -13°30'N 138°15'-140°15'W NEC 48 0.042 0.063 (BE) 3 14°30' -15°15'N 142° -144°W NEC 48 0 0 4 16° -17°N 145°45'-147°45'W NEC 48 0 0 5 17°30' -18°30'N 149°30'-151°30'W NEC 48 0 0 Pole-and-Line Fishing Hida (1970) recorded 109 schools of fish sighted during cruise Gilbert 116: 27 skipjack, 1 bigeye tuna, 1 yellowfin tuna, 2 mixed tunas (skipjack- bigeye-yellowfin), 13 of nontuna species, and 65 unidentified. On the outward track and in the southern part of the study area (from lat. 4°27'S, long. 133°22'W, to lat. 6°25'S, long. 117°47'W: 13 to 20 October 1969), at least 13 tuna schools were sighted or discovered by jig strikes. They were chummed, but did not respond to live bait. From 21 to 27 October in the study area (between lat. 4°09'S, long. 117°47'W and lat. 5°04'N, long. 118°50'W), at least 11 schools of tuna were chummed with live bait and of these 6 were suc- cessfully fished by pole-and-line (these schools were sighted and pursued, not discovered by jig strikes). Details are given in Table 19. All schools fished successfully by pole-and-line were located in the South Equatorial Current. These results are of special interest because they were obtained by a commercial fishing method. Trolling is not a com- mercial fishing method for skipjack in U.S. fisheries. Size of Skipjack and Other Tuna Table 20 shows the percent of skipjack in three broad size categories by fishing method and area. Only one skipjack (45 cm) was taken on the inward track to Honolulu. In the study area the percent- ages in size groups are similar for the two fishing methods. The percentage of fish < 45 cm, namely 8 to 11%, is much the same as that in the area in November-December 1970. The smallest skipjack measured 34 cm. The largest skipjack (mean lengths > 75 cm) were taken in the extreme south of the area. Elsewhere mean sizes of skipjack ranged from 46 to 67 cm, with two exceptions: 34 cm (trolling) and 40 cm (pole-and-line). On the outward track to the area, troll-caught skipjack 405 FISHERY BULLETIN: VOL. 73, NO. 2 Table 19.-Details of successful live-bait pole-and-line fishing of tuna schools in the study area, Cruise Gilbert 116, October 1969. SJ, YF, and BE mean skipjack, yellowfin, and bigeye. Time Approx imate successful No. Approximate School data! No. birds DOSit on fishing commenced Species No. taken Size (cm) mea- sured weic ht (lb) Date Lat. Long. Range Mean Range (Mean) Oct. 21 _ 15 birds 4°S, 118°W 07102 SJ 13 37-52 40 13 — (3) YF 8 37-43 39 8 — (3) BE 7 37-43 41 7 - (3) 24 — Flock, 50 2°N, 119°W 12203 SJ 213 64-70 68 50 15-18 (17) 26 Breezer Flock, 150 4°N, 119°W 1125* SJ 519 41-56 47 60 3-8 (5) YF 28 40-58 47 28 4-8 (4) BE 13 41-46 43 13 3-4 (4) Boiler Flock, 500 4°N, 119°W 17073 BE 97 50-81 69 33 10-20 (-) 27 Boiler Flock, 300 4°50'N 118°25'W 06502.3 SJ 49 47-67 59 49 — (10) Breezer Flock, 250 5"N, 118°49'W 10313 SJ 110 48-61 52 54 — (-) 1 After Scott (1969). 2 Poor biting school. 3 School abandoned when sample complete. * Very large school still around ship at 1500 h. Table 20. -Skipjack by size categories as percent of total, and by fishing method and area, Cruise Gilbert 116, October 1969. Size Study area Outward track to study area (cm) Troll Pole-and- ine Troll <45 45-59.9 >60 8.0 52.0 40.0 11.2 57.0 31.8 80.0 15.0 5.0 Total skipjack 25 223 20 ranged from 29 to 78 cm, mean 41 cm. Although very few were caught, the high percentage of fish < 45 cm is of interest. In the study area all small fish ( < 45 cm) were from areas of the South Equatorial Current. In the area west of long. 125°W, small fish were found in all three current systems (NEC, NECC and SEC), approximately from lat. 15°15'N, long. 151°45'W to lat. 0°25'S, long. 137°15'W. Skipjack percent length frequency distribution by 2-cm classes is given in Figure 30 for the study area in October 1969. There appear to be three >- o z LlI Z) o LU a: 25 p 20- 15 - 10- 5 - n = 247 til I I I 1 I I I I I I I I I "I "I "I"! Tl ■ I ■ |"i'|"I I I 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 2cm FL CLASSES Figure 30.-Skipjack percent length frequency distribution (study area only), cruise Gilbert 116, October 1969. Smoothed curves are from 3-figure moving average. Stated length indicates midpoint of class. modes at 46, 56, and 66 cm. Inclusion of data from the outward track would increase the probability of another mode at 36 to 38 cm. The November- December 1970 length data for the study area were similar and showed modes at 36, 48, and 58 cm; only the 66-cm mode was absent (Figure 9). The similarity suggests that the modes of 1969 (October-November) and 1970 (November- December) represent age-classes. On the outward track yellowfin were small, mean lengths 37 and 32 cm, and on the return track bigeye had a mean length of 57 cm (Table 18). In the study area mean lengths of yellowfin ranged from 39 to 47 cm and those of bigeye from 41 to 69 cm (Table 19). Sex and Maturity of Skipjack and Other Tuna The sex ratio of skipjack in the study area was males to females 1:0.89 (w = 249), and on the outward track 1:1 (n = 20). Tuna gonads taken on cruise Gilbert 116 were recorded as immature, maturing, mature, or spent and can be roughly compared with those for the other two cruises. The number of skipjack gonads in each maturity stage by size of fish is given in Tables 21 and 22 for the study area and outward track. Apart from three females, all immature fish (19% of total in the study area and 80% of total on outward track) were <50 cm. The principal difference between fish caught on this cruise and the other two is the virtual absence of spent fish. Most fish ( > 74%) were classed as maturing, and 406 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA this could indicate that they were southern spawners. All troll-caught yellowfin tuna were immature, sex indeterminate, as were the bigeye except for one immature male of 59 cm. DISCUSSION AND CONCLUSIONS Gulland (1971) reviewed research findings which indicate that skipjack is the most abundant tuna in the Pacific, except possibly for frigate mackerel which is a small and presently valueless species. Our results and those of Hida (1970) show that each of the three cruises yielded many more skip- jack than all other species of fish combined, including nontunas and unidentified fish (Tables 2, 15, 18, 19). On each cruise some skipjack were ob- tained in almost every part of the area in which fishing was done. Occurrences of other tunas (yellowfin, bigeye, and frigate mackerel) were much fewer and more localized (Tables 2, 3, 4, 15, 18, 19). Our results also support the general hypothesis of Rothschild (1965), Williams (1972), and others that skipjack migrate as juveniles from central Pacific spawning areas towards the American coast, spend part of their adolescent life near the coast, and then return to the central Pacific. The present study area lies between the spawning areas and the coast. Thus one would expect the skipjack in that area to include, at times, in- dividuals both smaller and larger than those Table 21. -Number of skipjack gonads in each maturity stage by size of fish, study area, Cruise Gilbert 116, October-November 1969. Size class Immature Mat M uring F Mature (cm) M F M F 30-39.9 2 6 1 40-49.9 16 21 21 4 50-59.9 3 48 38 3 6 60-69.9 38 34 1 2 70-79.9 1 2 2 Totals: males (M) 132, females (F) 117. Table 22. -Number of skipjack gonads in each maturity stage by size of fish, outward track, Cruise Gilbert 116, October 1969. Size class Im mati ure Maturing Mature Spent (cm) M F M F M F M F 20-29.9 2 30-39.9 3 4 40-49.9 4 3 50-59.9 1 1 1 60-69.9 70-79.9 1 Totals: males (M) 10; females (F) 10. typical of coastal waters. We have demonstrated their occurrence (Tables 6, 7, 20; Figures 9, 30). Matsumoto (1966), Ueyanagi (1969), and Love (1970, 1971a, in prep.: EASTROPAC data) indicate that skipjack larvae are rare east of long. 130° W in the tropical Pacific, but increase rapidly west of that meridian. Thus our study area is close to a spawning region. One would then expect some of the large skipjack in the study area to have ma- turing, spent, or spent-recovering gonads at times, and this condition was found (Tables 8, 9, 10, 21, 22). The occurrence of spent-recovering and rest- ing gonads in November-December suggest prin- cipally northern summer spawning, especially since no spawned-out fish were taken south of lat. 3°N. The presence of spent-recovering fish in March-April perhaps indicates southern summer spawners (northern winter); however, the oc- currence of skipjack with maturing gonads at this time may also signify northern summer spawners. Fish taken on the two cruises may be of two spawning groups, northern and southern (see Williams 1972). The juvenile skipjack ( < 45 cm) constituted a small proportion, 13% or less on each cruise, of the total skipjack caught in the study area. Their dis- tribution varied spatially and temporally (Tables 6, 7, 20; text p. 389). On the cruise of October- November 1969 they were found only in the South Equatorial Current, but the North Equatorial Current was not sampled. In November-December 1970 they were found principally in the North Equatorial Current and sparsely in the South Equatorial Current. In March-April 1971 they were very scarce or absent in all parts of the study area. West of long. 125°W in October-November 1969, some juveniles were taken in the North Equatorial Countercurrent as well as in the other two currents. It does not seem possible from these data to make a choice among any of the three models of coastward migration of juveniles proposed by Williams (1972), or to eliminate any of them from consideration. Data from other periods of the year are desirable. From previous studies by Williams (1970), adult skipjack (> 45 cm) were expected to occur in waters of surface temperature 20° to 29°C, but not preferentially at particular temperatures within that range. All waters of the study area had such temperatures on all cruises (Figures 10, 11, 12, 13, 28, 29). Thus they were all suitable for skipjack as far as temperature was concerned, and skipjack occurred to some extent in most of them (Figures 407 FISHERY BULLETIN: VOL. 73, NO. 2 5, 6, 7, 8; Tables 15, 16, 19). No relation appears between skipjack distribution and particular temperatures within the 20° to 29°C range. Blackburn and Laurs (1972) expected adult skipjack to be distributed like their forage in offshore areas where all surface temperatures are suitable. This was because Blackburn (1969) found such a relation for skipjack in waters of suitable temperature near the coast, and Magnuson (1969) found that skipjack eat the equivalent of 15% of their body weight per day when fed to saturation. Thus adult skipjack would probably be most numerous in the latitudinally oriented zones of abundant forage which occur offshore, with forage concentrations comparable to those in coastal waters, during their westward movement from the coast to the spawning areas (Blackburn and Laurs 1972). They would probably migrate slowly through forage-rich zones or areas and quickly through those poor in forage, and thus be more abundant per unit area in the forage-rich situa- tions. Blackburn and Laurs (1972) showed from EAS- TROPAC data that the richest and most persis- tent zones of skipjack forage in our study area occurred a few degrees north and sometimes south of the equatorial upwelling. They also recognized a less conspicuous zonal forage maximum near the northern boundary of the North Equatorial Coun- tercurrent, probably associated with high biological production over the shoal pycnocline. Data from the November-December cruise show the expected maxima of forage and skipjack near the Equator, but do not clearly show a maximum of either on the north side of the Countercurrent (Figures 23, 24; Table 3). Tables 12 and 13 show two statistically significant positive correlations between availability of large skipjack and their forage on the same cruise, although only for night concentrations of forage and for data averaged over a 2° zone of latitude. As mentioned earlier the actual significance may be disputable for the lower of these correlation coefficients, but not for the higher one, taking the total number of correlations in Tables 11 and 12 into account. Correlations between skipjack and day forage were not sig- nificant (Table 11). Skipjack probably do much of their feeding in the daytime (Nakamura 1962) although forage is much scarcer in the upper water layers by day than by night. Thus the lack of relation between skip- jack and day forage may seem surprising. One could however interpret these results as follows. Spatial distributions of day and night forage broadly coincide (Blackburn and Laurs 1972) because they are determined by the same physicochemical and basic biological features of the environment. Skipjack tend to occur in broad zones where both kinds of forage are initially abundant, for reasons suggested above. Within these zones they aggregate in the richer patches of day forage and eat them down, whereby their relation with the day forage will be sometimes direct and sometimes inverse. If they eat the much more abundant night forage they probably do not so frequently reduce it to a point at which the relation becomes inverse. The significant November-December correla- tions become nonsignificant when data for skip- jack < 45 cm are included (Table 12). Thus juvenile skipjack may be distributed in relation to a different kind of forage, or possibly to other environmental properties excluding forage. Blackburn and Laurs (1972) made no statement about ecology of juveniles. The relatively sparse data for the March-April cruise of 1971 show a forage maximum near the Equator but not clearly elsewhere (Figures 25 and 26), although one may nevertheless have been present near the northern edge of the Counter- current, as mentioned previously. The principal maximum of skipjack in March-April 1971 was located slightly north of the North Equatorial Countercurrent, and there was a secondary maximum near the Equator (Table 4). Data on skipjack and forage yielded no significant correla- tions. They were probably too sparse to do so (Ta- ble 12). Tables 3 and 4 show that skipjack were less abundant in the North Equatorial Countercurrent than in either of the adjacent currents, on both the 1970 and 1971 cruises. This was expected because neither Blackburn and Laurs (1972) nor we found much forage in the Countercurrent. However skipjack availability was much higher in the Countercurrent (lat. 6° to 8°N) than in the South Equatorial Current on the 1969 cruise (Tables 15, 16). Forage data are lacking for the cruise, but it is not likely that forage was highly abundant in the North Equatorial Countercurrent. We note that the large skipjack taken in October-November 1969 had sexually maturing or mature gonads (Table 21), whereas most of those taken on the other cruises had spent, spent-recovering, or rest- ing gonads (Tables 8, 9). Possibly the October- November fish were close to spawning, and thus 408 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA becoming distributed in accordance with the requirements of their larvae. There is evidence that skipjack larvae occur only at sea tempera- tures from 23° to 31°C and are most common at about 29° to 30°C (Inter-American Tropical Tuna Commission 1971). Maximum surface tempera- tures in the study area in October-November 1969 were between 27° and 28° C and occurred from about lat. 4° to 11°N (Figure 29). Thus the Coun- tercurrent waters at lat. 6° to 8°N could have been particularly suitable for the survival of skipjack larvae, and the parent fish may have been becom- ing distributed accordingly. We found no direct relations between skipjack and mixed layer depth, dissolved oxygen, surface currents, chlorophyll, or zooplankton although some of these properties and features should have indirect effects on skipjack through their effects on temperature and forage. Some of them could also have direct effects upon larval or juvenile skipjack. Significant correlations between skip- jack and zooplankton were not found (Tables 11, 12). Significant correlations between large skip- jack and forage are also significant between skip- jack and the arithmetic product of forage and zooplankton, but not between skipjack and zooplankton alone. Although this paper has contributed to our knowledge of the distribution and relative abun- dance of skipjack in the offshore eastern tropical Pacific, where little information was previously available, the prospects for commercial fishing remain unknown. Our simple experimental fishing procedures served to identify zones of maximum occurrence of skipjack, but commercial trials will be needed to show if those zones can be exploited profitably. Ideally there should be trials by live- bait boats as well as purse seiners, in view of the fact that live-bait fishing gave good results on an experimental scale during the 1969 cruise. Our da- ta and interpretations should be useful as a guide to those who make these tests. ACKNOWLEDGMENTS We are grateful to associates in the Scripps Tuna Oceanography Research Program, the Inter-American Tropical Tuna Commission, and the Southwest Fisheries Center Honolulu and La Jolla Laboratories, National Marine Fisheries Ser- vice, NOAA, who participated in the cruises and processing of the data. Additionally, we thank the Director of the Honolulu Laboratory, National Marine Fisheries Service, for supplying the original data collected on cruise 116 of the RV Charles Gilbert. Thanks are also due to T. S. Hida, J. D. Isaacs, R. W. Owen, and P. E. Smith who reviewed the manuscript. The work was part of the Scripps Tuna Oceanography Research (STOR) Program of the Institute of Marine Resources, University of California. It was supported by the National Marine Fisheries Service under Contracts 14-17-0007-989, 14-17-0001-2311, and N208-0047-72 (N) with the Institute of Marine Resources. LITERATURE CITED Anonymous. 1973. Further experiments on lethal oxygen levels in skip- jack tuna confirm earlier data. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Southwest Fish. Cent, Tuna Newsl. 12:7. Blackburn, M. 1965. Oceanography and the ecology of tunas. Oceanogr. Mar. Biol., Annu. Rev. 3:299-322. 1968. Micronekton of the eastern tropical Pacific Ocean: Family composition, distribution, abundance, and rela- tions to tuna. U.S. Fish Wildl. Serv., Fish. Bull. 67:71-115. 1969. Conditions related to upwelling which determine dis- tribution of tropical tunas off western Baja Califor- nia. U.S. Fish Wildl. Serv., Fish. Bull. 68:147-176. 1970. Collection and processing of data: Micronekton. In C. M. Love (editor), EASTROPAC atlas. Vol. 4. Biological and nutrient chemistry data from principal participating ships, first and second monitor cruises, April-July 1967, p. 10- n. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. Blackburn, M., and R. M. Laurs. 1972. Distribution of forage of skipjack tuna (Euthynnus pelamis) in the eastern tropical Pacific. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-649, 16 p. Blackburn, M., R. M. Laurs, R. W. Owen, and B. Zeitzschel. 1970. Seasonal and areal changes in standing stocks of phy- toplankton, zooplankton and micronekton in the eastern tropical Pacific. Mar. Biol. (Berl.) 7:14-31. Cromwell, T. 1958. Thermocline topography, horizontal currents and "ridging" in the Eastern Tropical Pacific. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull. 3:133-164. Gulland, J. A. (editor). 1971. The fish resources of the ocean. Fishing News (Books) Ltd., West Byfleet, Engl., 255 p. Hida, T. S. 1970. Surface tuna schools located and fished in equatorial eastern Pacific. Commer. Fish. Rev. 32(4):34-37. HiGGINS, B. E. 1966. Sizes of albacore and bigeye, yellowfin, and skipjack tunas in the major fisheries of the Pacific Ocean. In T. A. Manar (editor), Proc. Governor's Conf. Cent. Pac. Fish. Resour., State of Hawaii, p. 169-195. Inter-American Tropical Tuna Commission. 1966. Annual report of the Inter-American Tropical Tuna Commission for 1965, 106 p. [In Engl, and Span.] 409 FISHERY BULLETIN: VOL. 73, NO. 2 1971. Annual report of the Inter-American Tropical Tuna Commission for 1970, 127 p. [In Engl, and Span.] 1972. Annual report of the Inter-American Tropical Tuna Commission for 1971, 129 p. [In Engl, and Span.] Joseph, J., and T. P. Calkins. 1969. Population dynamics of the skipjack tuna {Kat- suwonus pelamis) of the eastern Pacific Ocean. [In Engl. and Span.] Inter-Am. Trop. Tuna Comm., Bull. 13:1-273. Kawasaki, T. 1965. Ecology and dynamics of the skipjack population. I. Resources and fishing conditions. [In Jap.] Study Ser. Jap. Fish. Resour. Conserv. Assoc. 8-1:1-48. Engl, transl. by M. P. Miyake, Inter-Am. Trop. Tuna Comm., 1967). King, J. E. 1958. Variation in abundance of zooplankton and forage organisms in the central Pacific in respect to the equa- torial upwelling. Proc. 9th Pac. Sci. Congr. 16:98-107. Laurs, R. M. 1970. Collection and processing of the data: Zooplankton and fish larvae. In C. M. Love (editor), EASTROPAC atlas. Vol. 4. Biological and nutrient chemistry data from principle participating ships, first and second monitor cruises, April-July 1967, p. 10. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. LaVIOLETTE, P. E., AND S. E. Seim. 1969. Monthly charts of the mean, minimum and maximum sea surface temperature of the North Pacific Ocean. Spec. Publ. 123, Nav. Oceanogr. Off., Wash., D.C., 62 p. Love, C. M. (editor). 1970. EASTROPAC atlas. Vol. 4. Biological and nutrient chemistry data from principal participating ships, first and second monitor cruises, April-July 1967. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. 1971a. EASTROPAC atlas. Vol. 2. Biological and nutrient chemistry data from principal participating ships, first survey cruise, February-March 1967. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. 1971b. EASTROPAC atlas. Vol. 3. Physical oceanographic and meteorological data from principal participating ships, first and second monitor cruises, April-July 1%7. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. 1972a. EASTROPAC atlas. Vol. 1. Physical oceanographic and meteorological data from principal participating ships, first survey cruise, February-March 1967. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. 1972b. EASTROPAC atlas, Vol. 5. Physical oceanographic and meteorological data from principal participating ships, second survey cruise, August-September 1%7. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. Magnuson, J. J. 1969. Digestion and food consumption by skipjack tuna (Katsuwonus pelamis). Trans. Am. Fish. Soc. 98:379-392. Matsumoto, W. M. 1966. Distribution and abundance of tuna larvae in the Pacific Ocean. In T. A. Manar (editor), Proc. Governor's Conf. Cent. Pac. Fish. Resour., State of Hawaii, p. 221-230. MlVAKE, M. P. 1968. Distribution of skipjack in the Pacific Ocean, based on records of incidental catches by the Japanese longline tuna fishery. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull 12:509-608. Nakamura, E. L. 1962. Observations on the behavior of skipjack tuna, Enthynnus pelamis, in captivity. Copeia 1962:499-505. Orange, C. J. 1961. Spawning of yellowfin tuna and skipjack in the east- ern tropical Pacific, as inferred from studies of gonad development. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull. 5:457-526. Owen, R. W.,JR. 1970a. Collection and processing of the data: Thickness of the upper mixed layer. In C. M. Love (editor), EAS- TROPAC atlas, Vol. 4. Biological and nutrient chemistry data from principal participating ships, first and second monitor cruises, April-July 1967, p. 7. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. 1970b. Collection and processing of the data: Dissolved oxygen. In C. M. Love (editor), EASTROPAC atlas, Vol. 4. Biological and nutrient chemistry data from principal participating ships, first and second monitor cruises, April-July 1967, p. 7. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. Owen, R. W., and B. Zeitzschel. 1970a. Phytoplankton production: Seasonal change in the oceanic eastern tropical Pacific. Mar. Biol. (Berl.) 7:32-36. 1970b. Collection and processing of the data: Phytoplankton standing stocks and production. In C. M. Love (editor), EASTROPAC atlas. Vol. 4. Biological and nutrient chemistry data from principal participating ships, first and second monitor cruises, April-July 1967, p. 9-10. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. Riley, G. A. 1963. Theory of food-chain relations in the ocean. In M. N. Hill (editor). The sea: Ideas and observations on progress in the study of the seas. Vol. 2, p. 438-463. Interscience Publishers, N.Y. Rothschild, B. J. 1965. Hypotheses on the origin of exploited skipjack tuna (Katsuwonus pelamis) in the eastern and central Pacific Ocean. U.S. Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 512, 20 p. Schaefer, M. B. 1961. Tuna oceanography programs in the tropical Central and Eastern Pacific. Calif. Coop. Oceanic Fish. Invest. Rep. 8:41-44. Scott, J. M. 1969. Tuna schooling terminology. Calif. Fish Game 55:136-140. Seckel, G. R. 1972. Hawaiian-caught skipjack tuna and their physical environment. Fish. Bull., U.S. 70:763-787. Taft, B. a., and F. R. Miller. 1970. Collection and processing of the data: Temperature, salinity, and derived quantities. In C. M. Love (editor), EASTROPAC atlas, Vol. 4. Biological and nutrient chemistry data from principal participating ships, first and second monitor cruises, April-July 1967, p. 6-7. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Circ. 330. TSUCHIYA, M. 1968. Upper waters of the intertropical Pacific Ocean. Johns Hopkins Oceanogr. Stud. 4, 50 p. Ueyanagi, S. 1969. Observations on the distribution of tuna larvae in the Indo-Pacific Ocean with emphasis on the delineation of the spawning areas of albacore, Thunnus alalunga. [In 410 BLACKBURN and WILLIAMS: DISTRIBUTION AND ECOLOGY OF SKIPJACK TUNA Jap., Engl, synop.] Bull. Far Seas Fish. Res. Lab. (Shimizu) 2:177-256. U.S. Bureau of Commercial Fisheries. 1%3. Skipjack - A world resource. U.S. Fish Wildl. Serv., Circ. 165, 28 p. Waldron, K. D. 1963. Synopsis of biological data on skipjack Katsuwonus pelamis (Linnaeus) 1758 (Pacific Ocean). FAO (Food Agric. Organ. U.N.) Fish. Rep. 6:695-748. Williams, F. 1970. Sea surface temperature and the distribution and ap- parent abundance of skipjack (Katsuwonus pelamis) in the eastern Pacific Ocean, 1951-1968. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull. 15:229-281. 1971. Current skipjack oceanography cruises in eastern tropical Pacific Ocean. Commer. Fish. Rev. 33(2):29-38. 1972. Consideration of three proposed models of the migra- tion of young skipjack tuna (Katsuwonus pelamis) into the eastern Pacific Ocean. Fish. Bull., U.S. 70:741-762. Wyrtki, K. 1964. The thermal structure of the eastern Pacific Ocean. Erganzungsh. Dtsch. Hydrog. Z., Reihe A 6, 84 p. 1965. Surface currents of the eastern tropical Pacific Ocean. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm., Bull. 9:269-304. YOSHIDA, H.O. 1966. Tuna fishing vessels, gear, and techniques in the Pacific Ocean. In T. A. Manar (editor), Proc. Governor's Conf. Cent. Pac. Fish. Resour., State of Hawaii, p. 67-89. 1971. The early life history of skipjack tuna, Katsuwonus pelamis, in the Pacific Ocean. Fish. Bull., U.S. 69:545-554. 411 REPRODUCTION AND RECRUITMENT OF THE BRACKISH WATER CLAM RANGIA CUNEATA IN THE JAMES RIVER, VIRGINIA'^ Thomas D. Cain^ ABSTRACT Reproduction and recruitment of the brackish water clam Rangia cuneata were investigated in the James River, Va., from February 1970 to January 1972. Histological examinations of gonads were made, newly set clams were collected, and temperature and salinity measurements were taken from three populations living in different salinity regimes. Gametogenesis began in April and ripe gonads were found from May to late November with no inactive period. From observations of set abundance, two periods of spawning were determined: one in early through midsummer, coinciding with the beginning of spawning as determined from gonadal examinations; and a second and longer period in late fall and early winter, with an increased percen- tage of partially spawned and spent clams. Gametogenesis ceased in December through March as residual gametes were cytolyzed. Sex was not detected during this last phase. More females than males were found in the upstream (lower salinity) populations. Temperature was important in initiating gametogenesis in the spring and midsummer. Spawning correlated best with changes in salinity to approximately 5 "/oo. Over its estuarine range, salinity has a controlling effect on Rangia spawning and recruitment. Seasonal reduction in input of freshwater (increased seawater intrusion) is needed to induce spawning and recruitment in upstream populations. Best recruitment occurred to the middle of the habitat range which has an annual salinity change from fresh to 5"/oo. A southern species of clam, Rangia cuneata (Gray) has in the last 15 yr extended its range into Chesapeake Bay estuaries (Hopkins and Andrews 1970). This clam occupies an otherwise "open niche" in the oligohaline region of these estuaries. Although species diversity is usually low, there may be large numbers of individuals of species adapted to this environment. In the upper James River estuary, Rangia accounts for nearly 95% of the benthic biomass. Rangia is important both ecologically and com- mercially (Hopkins 1970). It provides a substantial food source for several species of fish and crabs (Darnell 1958), and waterfowl (Wass and Wright 1969). It is ecologically significant because it con- verts detritus into biomass that can be utilized by these organisms (Odum and Copeland 1969). Not only is R. cuneata a species for which low 'Contribution No. 582 from the Virginia Institute of Marine Science, Gloucester Point, VA 23062. ■This paper is part of a dissertation submitted to the Univer- sity of Virginia at Charlottesville, in partial fulfillment of the requirements for a Ph.D. degree. The research was funded by the Virginia Electric and Power Company. 'Division of Technical Review, U.S. Nuclear Regulatory Com- mission, Washington, D.C. 20555. salinity, l-15''/oo, is optimal; it is also a species which evidently cannot maintain a population outside this range (Hopkins 1970). That Rangia thrives in a zone unfavorable for most animals indicates it has some unusual adaptations. Despite its abundance in favorable environments and long history on the Gulf Coast, this clam has received little attention. The study reported here concerns the reproduc- tive cycle and recruitment of Rangia cuneata in the James River. The major objectives were to: (a) study the gametogenic cycle of Rangia from his- tological sections; (b) determine differences in gametogenesis or spawning of clams over the species range in the estuary; (c) investigate, from analysis of field data, the influence of temperature and salinity on initiation of gametogenesis and spawning; (d) corroborate gametogenic findings by collecting newly set clams; and (e) determine the duration of the larval period and differences in set abundance in the estuary. Fairbanks (1963) studied the spawning cycle of Rangia in Lake Pontchartrain, La., but it is known that physiologically different races of bivalves can occur at different latitudes (Loosanof f 1969). Cain Manuscript accepted May 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. 412 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA (1973, 1974) reported on the laboratory spawning of Rangia and the combined effects of sahnity and temperature on embryos and larvae. The impending action (at the time the work was initiated) of discharge of waste heat into the Rangia community by the Virginia Electric and Power Company's (VEPCO) Nuclear Generating Station at Surry, Va., was a further impetus to the study of the reproductive cycle, especially those factors that initiate gametogenesis and spawning. DESCRIPTIONS OF THE STUDY AREA The study area in the James River is a transition region between freshwater and salt water. The area has a seasonably variable salinity that ranges between about 0 and 15°/oo, depending on the volume of freshwater input. In the spring, high river flow covers most of this region (except sta- tion A) with freshwater (Figure 1). Occasionally, in late summer and fall the study area may exhibit measurable salinity as far upstream as station C. The mean annual discharge of the James River is approximately 212 mVs (7,500 cfs). Figure l.-Location of sampling stations for Rangia cuneata in the James River, Va. (N = nuclear generating station). Field and hydraulic model studies of the James River estuary have shown a two-layer density flow pattern, in which the deep, more saline water has a net upstream flow and the surface, fresher water has a net downstream flow. The net sediment transport of the two-layer section averaged over many tidal cycles is upstream (Pritchard 1952). The transition section of the James River is characterized by high natural turbidity and sedimentation from the flocculation of river-borne sediments. The distribution of bottom sediment types in the James River estuary has been studied by Nichols (1972). His survey indicates silt-clay subs- trates at stations A, B, C, and D and sand subs- trates at locations As, Bs, Cs, and Ds. The distribution of Rangia in the James River was found to be approximately from nautical mile 25 to 55 above the river mouth (Figure 2). The downriver extent of its range overlaps the habitat of typical estuarine organisms such as the oyster; at the upriver limit Rangia is associated with completely freshwater forms such as freshwater mussels. In the oligohaline portion of its range it is typically associated with the polychaetes Scolecolepides viridis and Laeonereis culveri; the crustaceans Cyathura polita, Corophium lacustre, a,nd Gammarus sp.; and the bivalves Macoma balthica, Brachidontes recurvus, and Congeria leucophaeta. METHODS AND MATERIALS The reproductive cycle was investigated by collecting clams at stations A, B, and C (Figure 1). Station A was near the downstream range of the clam. Station B was 18.5 km above station A. Sta- tion C was located (18.5 km above station B) near the mouth of the Chickahominy River in order to include part of the clam's range which seldom experiences salinity changes. All stations were located at approximately the same depth (3-4 m). The clams used in this study were collected from a predominantly silt-clay substrate. Although Tenore et al. (1968) indicated that such sediments were detrimental to Rangia, the clams at the various stations appeared to be thriving over the 2-yr study. Beginning in February 1970 approximately 20 clams, 30-40 mm long, were collected at stations A and B using a modified oyster dredge. Attempts were made to collect clams at these stations every 413 FISHERY BULLETIN: VOL. 73, NO. 2 Figure 2.-The distribution of Rangia cuneata in the James River, Va. Seg- ments are at 5-nautical mile intervals. 2 wk, but bad weather and boat failures oc- casionally delayed this to 3 wk. Beginning 22 Sep- tember 1970 collections at station C commenced. Collections at all stations were terminated in January 1972. In the laboratory these clams were measured, weighed, shucked, and the gonads dissected out and placed in a solution of alcohol. Formalin,^ and acetic acid (AFA) for fixation. Gonad tissues were sectioned at 7-10 /xm with a rotary microtome, stained with Delafield's hematoxylin, and coun- terstained with eosin. Gonad tissue stage of development was determined following the scheme of Ropes (1968) who categorized the seasonal gametogenic cycle of Spisula solidissima as: early active, late active, mature, partially spawned, or spent. Similar stages of development were first described by Ropes and Stickney (1965) for Mya arenaria and have subsequently been used for two other members of the family Mac- tridae; Mulinia lateralis (Calabrese 1970), and Tresus capax (Machell and De Martini 1971). The number of clams in each category, regardless of their sex, was recorded for each sample. The sex ratio of clams from each station was calculated and a chi-square test used to establish goodness of fit to a 1:1 ratio. During June 1970, clams collected at stations A and B were placed in four groups (5-10, 11-20, 21-30, and 41-50 mm). These clams were sectioned and stained to determine the size at which they contain reproductive products. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Set Collectors Collectors used to determine the time and in- tensity of setting were placed at stations A, B, C, and D and inshore from these areas in shallow, sandy areas (Figure 1). These stations were designated as As, Bs, Cs, and Ds. Stations A, As, B, Bs, and Ds were examined from June 1970 and C, Cs, and D from September 1970, at approximately 2-wk intervals until January 1972. The set collector was a plastic gallon jar with an 8.7 cm diameter mouth. The mouth was covered with a plastic 5 mm mesh to prevent the entry of predators. The jar rested on the bottom, fastened to a concrete block. Water flowed across the mouth of the jar and suspended sediments, detritus, and me- tamorphosing clams settled to the bottom. In the field, the contents of the jar were washed through a 0.174 mm mesh screen. In the laboratory each sample was elutriated to remove most of the de- tritus (Cofl^n and Welch 1964). The material remaining after elutriation was examined for clams under a dissecting microscope. All bivalves were counted and some were measured. Set of bivalve species other than Rangia was also iden- tified and counted. Environmental Data Water samples for salinity, dissolved oxygen, and temperature measurements were taken whenever biological collections were made. A Kemmerer bottle was used to obtain bottom water samples at the deep stations. Samples at shallow 414 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA stations were collected about 0.3 m below the sur- face. Temperature measurements were taken im- mediately with a stem thermometer. Bottom temperatures recorded at VEPCO instrument towers 1 and 6 (near stations A and B, respec- tively) were also used in this study. Salinity samples were analyzed in the laboratory with a Beckman RS7B induction salinometer. Dissolved oxygen samples were fixed immediately after collection and analyzed in the laboratory by a modified Winkler method. Freshwater input was compiled from records taken at gauging stations on the James River near Richmond, the Appomattox River near Matoaca, and the Chickahominy River near Providence Forge. Combined, these three rivers are the major sources of freshwater for the James River in the study region. RESULTS Histological Study of the Reproductive Cycle A histological basis for classifying the gonadal condition was used because the external ap- pearance of gonads did not accurately reflect phase of development. The appearance of gonads of both sexes is superficially the same during each phase. No evidence of gonadal parasitism was found in any of the tissue sections, nor were hermaphrodi- tic individuals found. There was little difference in the time of initia- tion of gametogenesis and ripening between the sexes so the number of males and females in each stage was combined for analyses. Figures 3 to 6 show the phases in the development of the female and male gonads. Station A The reproductive cycle of clams at station A (Figure 7) was more complex than at the other stations. From early February to late March 1970 most clams were in the spent phase, although a few male clams contained ripe sperm with sperm balls. In early April 1970, 40% of the sample were in the early active phase. By May, 40% were ripe, with 10% partially spawned. From May through September clams were found in all gonadal phases. Evidently some spawning and rematuration oc- curred during the summer months. In early Oc- tober 1970, all clams examined were ripe. The volume of eggs and sperm at the second ripening was much greater than that in the early spring and summer. Partially spawned clams were numerous at the end of October and by mid-November 85% of the sample were partially spawned or spent. Throughout the rest of the winter most clams were in the spent stage, although some males retained sperm and slight gonadal activity was noted in some females. The reproductive cycle for 1971 was basically the same as the previous year. In early June 1971, 65% of the clams were ripe. Spawning was indicated during the next 2-wk period because 60% were spent or partially spent. The fall spawning season was very similar to that of 1970, with 95% ripe by late September. Spawning was completed by early November and, again, some ripe males were ob- served during the winter. Station B Clams in the spent or inactive phase were found from February to early April 1970 (Figure 8). Some males still contained sperm in various stages of cytolysis. By late April 1970, half of those collected had begun gametogenesis, resulting in 80% being ripe by early June 1970. Clams remained in the ripe phase throughout the summer with some spawning occurring during July. During August there was a second development, resulting in all clams observed being ripe in early Sep- tember 1970. Spawning commenced in early Oc- tober and was completed by mid-November 1970. Immediately after spawning some clams were in the early active phase, but development did not proceed further during the winter. The reproductive cycle for 1971 was similar to that of 1970. Gametogenesis commenced in early May. Fewer ripe clams were observed during the summer months than in 1970. The second cycle began in early July and by early September all clams were ripe. Spawning began in late October and was completed by late November. As found at station A, the second seasonal cycle for station B was more intensive; more ripe clams were found and their gonads contained far more sperm and eggs. Spawning was more intensive during the fall, with gonads progressing from ripe to spent in a month. 415 FISHERY BULLCTIN: VOL. 73. NO. 2 r^f^'T^I ,1 -i' 19k Figure 3.-A, section of Rangia V*C "f- .VAV ovary in the early active phase of oogenesis ( x 120); B, ovary in early active phase ( x 500); C, late active phase ( x 120); D, ripe ovary ( x 120). , ovary in 416 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA » i '#.1 y -*>' . i €> x^ ^'j,'-. •^^■*:>::x ■ \ *< fe ^ •i? "^•"iTvr i * .# r % vr ©* i: ' lb / ,11 o. •^ t % >i- • •% A' .*. yiX* .0 */»' Figure 4.— A, Rangia ovary in the partially spawned phase ( x 120); B, partially spawned ovary ( x 500); C, spent ovary with few ova retained ( x 120); D, cytolysis of unspawned eggs ( x 120). 417 FISHERY BULLETIN: VOL. 73, NO. 2 . -jY.'.,^.^. Bi»i^^-?-^jt FIGURE 5.- A. Section of testis of Rangia in early active phase of spermatogenesis ( x 120); B, testis in late active phase-note sperm in center ( x 500); C, ripe male ( x 120); D, ripe male ( X 500). 418 CAIN; REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA r©. **» ^-•'V *v ¥^.> D y .v^ Figure 6.-A, Section of testis of Rangia in partially spawned phase ( x 120); B, testis with retained sperm and sperm balls ( x 120); C, testis with sperm balls ( x 500); D, spent testis with few sperm retained ( x 120). 419 FISHERY BULLETIN: VOL. 73, NO. 2 1970 OS 100 10- 1971 - , , n — ^ "t^" 1 1 1 1 1 r — — I r*- 1 — FEB ' MAR ' APR ' MAY ' JUN ' JUL ' AUO ' SEP ' OCT ' NOV ' DEC JAN FEB MAR APR MAY JUN JUL AUO SEP OCT NOV OCC JAN DSPCNT QpARTIALLr SPENT ■riPE OQ LATE ACTIVE E3 EARLY ACTIVE Figure 7.-Gonadal phases and setting of Rangia at station A in relation to salinity from 1970 to 197L The length of each shaded area represents the percentage frequency of clams in each category. Station C The reproductive cycle at station C was similar to that at station B during 1970. However, earlier spawning was indicated by 88% of the clams being either partially spawned or spent in early October (Figure 9). Clams remained in the spent phase until early May 1971, when gametogenesis began. Few ripe clams were found at this upstream sta- tion during the late spring and summer. Clams in the early active phase were found during July and by the end of August all clams were ripe. From September to November 1971 clams were predominantly ripe, but there was no spawning. Cytolysis of the eggs began in November, result- ing in a spawned-out appearance (Figure 4D). No spawning occurred at station C during 1971, in marked contrast to stations A and B where the fall spawning was intense. Sex Ratio The data were divided into summer-fall and winter-spring seasons, because many clams con- tained no discernible gonads during the winter I 970 FEB MAR APR MAY JUN JUL AUO MP OCT MOV DEC ' JAN ' FEB ' MAR ' APR ' MAY ' JUN JUL AU8 SEP OCT NOV OCC JAN □ spent Q partially spent Bripe on) late active Oearlt active Figure 8. -Gonadal phases and setting of Rangia at station B in relation to salinity from 1970 to 1971. The length of each shaded area represents the percentage frequency of clams in each category. 420 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA >. 15-, '*^° CO Figure 9.-Gonadal phases and set- ting of Rangia at station C in relation to salinity from 1970 to 1971. The length of each shaded area represents the percentage frequency of clams in each category. 0 1000- 19 7 1 u. « ^dX 100- ^4 ^_E OCT NOV DEC JAn'*"fEB ' MAR ' APR ' MAY ' JUN ' JUL DSPENT ^PARTIALLY SPENT BriPE H) L ATE ACT I VE □ EARLY ACTIVE and early spring. During summer and fall the gonads of most clams could be recognized (Table 1). The ratio of females to males at station A was not significantly different from 1:1. Females predominated at stations B, D, and C during the summer and fall months of 1970 and 1971. When the clams of non-determinable sex are added to the male group, there are still significantly more females. Of the clams collected in late June 1970 at sta- tion A, none measuring 5-10 mm contained gonads. In the second group 50% showed signs of gametogenesis and had recognizable sex products; 70% of the third group had discernible gonads; and most clams in the fourth group contained gonads. Most small clams were males, but too few were examined to test the significance of sex ratios. Larval Setting The number of Rangia clams setting at each station is shown in Figures 7 to 14. This number Table 1.— The ratio of females to males and the number of clams with non-determinable (ND) gonads at each station. Summer and fall seasons were tested with the ND clams added to the male group. Seasons Station A F ND M Station B Station D ND M F ND M Station C F ND M 166 9 70' Summer and fall 1970 140 0 153ns 209 3 57** Winter and spring 61 53 83 46 128 35 13 44 13 29 118 66 Summer and fall 1971 127 7 102ns 178 22 34** 41 5 11** 138 38 54** ns = not significant. ** ^ highly significant. does not provide an estimate of survival because predators were excluded by the screen cover. The number of set clams at both deep and shallow sta- tions by season is shown in Table 2. At stations A and As a small number ( < 7) set during late July and August 1970. In December 1970 through March 1971 set clams were common but not very abundant at these two downriver stations. Only one setting period (29 individuals) was recorded during the summer months at station As. Larvae began setting again during the third week of December 1971 with 70 at station A. Collectors at both stations continued receiving set until sampling was terminated on 18 January 1972. Setting at stations B and Bs was sporadic dur- ing the spring and summer of 1970, with no more than four clams in any jar. Setting began at sta- tion B in mid-November 1970 and continued there until late March 1971. A maximum of 42 in- dividuals was collected on 8 February 1971. Set clams were found on only one sampling date dur- ing the next summer. Setting in the fall began in late October 1971, with very large numbers during late November and early December. Setting at station D was similar to that at sta- tion B except more individuals were found during the fall and winter of 1970-71. One hundred eighty-five were collected in the station D bottle in mid-December 1971. Only data beginning in September 1970 were available from station C. As noted in the tissue sections, spawning had commenced earlier. Set- ting from early September 1970 to March 1971 was very heavy, with 257 clams collected at this station 421 < 1000 -I u. o £1" 10 03 ~^ T T FISHERY BULLETIN: VOL. 73, NO. 2 A, S ' 0 ' N ^ P^*- I T I 1 ml A ' M ' J ' J ' A ' S ' 0 ' N ' D ' J ' J'ASONDJ'F'M Figure 10.— Number of Rangia set and salinity at station Ag from July 1970 to January 1972. > 15 t- z 10 _l 5 in 0 ^ 1000 UJ V) b."* 100 Oo u CC" UJ Q> 10 :e- 3 I - P np, ^rpt^ ■ i B. T r — I r T"' — r^ — I 1 A S 0 N D J F *y * ,^. T 1 1 1""" I'"' — I 1 r MAMJJASON Figure 11. -Number of Rangia set and salinity at station Bg from July 1970 to January 1972. S 0 N'DJF'M'A'MJJ'AS OND'J Figure 12.-Number of Rangia set at station D in relation to salinity for September 1970 to January 1972. 422 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA 's '^g^^ D ' J '"f"' M ' a ' M ' J ' J — I 1 1 — A S 0 N JASON Figure 13.-Nuinber of Rangia set and salinity at station Dg from July 1970 to January 1972. D J o >- < H lOOOn iij V) }i:?ICX)H O o u 3 "^ 10- FiGURE 14.-Number of Rangia set and salinity at station Cg from July 1970 to January 1972. Table 2.-Average number of Rangia set (per collector) at both deep and shallow stations by season. Station Season A B D C Summer (1970) Fall-winter Summer (1971) Fall-winter 0.7 3.6 1.6 6.2 0.7 6.1 0.2 44.9 0.7 25.5 0.4 15.4 74.3 0.1 2.5 by early October 1970. This was the highest number collected at any time or place during the study. No set was collected at stations C or Cs from late March to late December 1971. The fall setting was very small with only 5 collected at station C and 18 at station Cs during December 1971 and January 1972. Rangia set ranged in length from 230 to 500 /^m, but averaged about 300 ju,m. Larger individuals were generally collected at the shallow stations but may have been members of an earlier set washed in by wave action. The setting patterns of the other bivalves are presented in Figure 15. Station A received more set of all three species than the other stations because of its proximity to their adult populations. Brachidontes recurvus and Macoma mitchelli were collected farther upriver than M. balthica. The first two species were common at stations A, B, and D. Brachidontes recurvus and M. mitchelli are evidently more tolerant of low salinities as they were the only set found during the low salinity conditions of fall 1971. All three species have a nearly year long spawning season with minor spring peaks and a major peak in the fall. The three most common organisms found on the bottles were Rhithropanopeus harrisii, Callinectes sapidus, and the blenny, Chasmodes bosquianus. These three potential set predators were typically found at stations A, B, and D during the fall and winter months. During high salinity periods, R. harrisii was found at station C. Hydrographic Data Freshwater input levels were usually high in late winter and spring, declining to low levels in late summer and fall (Figure 16). Flow during the 423 FISHERY BULLETIN: VOL. 73, NO. 2 Brachidontti recurvus Macoma balthica Macomc mitchtlli Figure 15.-Setting patterns of other bivalves at stations A, B, and D from September 1970 to January 1972. Data combined at deep and shallow sta- tions. fall of 1970 was very low, with about 28.32 mVs (1,000 cfs) input in September and October. This low input allowed measurable-salinity water to extend as far upstream as mile 45. Input was high and variable during the winter and early spring of 1971. The peak for the 2-yr period of 2,747 mVs (97,000 cfs) was recorded on 1 June 1971. The summer input was fairly low but quite variable and river flow during September and October was considerably higher than the previous fall. The salinity was rarely measurable at station B 100,000 T -^ 10,000 - o 3 1000- throughout the fall and winter months of 1971. The annual temperature pattern for station B (Figure 17) is representative of the study area. Within relatively narrow limits, all deep stations exhibited similar temperature profiles. Lowest temperatures were measured during late January and early February, followed by a relatively smooth increase to a maximum of 29°C during early August. A period of stable high tempera- tures was recorded from June through September. Dissolved oxygen concentrations in this region 1970 Figure 16.— Freshwater input into the James River at 4-day intervals. The data include input from the Appomattox and Chickahominy Rivers. 424 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA 30 n Figure 17.-Bottom water tempera- ture at station B in the James River, Va. Values taken from VEPCO in- strument tower (#6) and bottom water samples. FEB ' MAR ' APR ' MAY JUN ' JUL ' AUG ' SEP ' OCT ' NOV ' DEC ' JAN ' showed the normal seasonal variation, with the highest concentrations in the early part of the year during the low-water-temperature period. Low values (6-8 mg /liter) were recorded for deep stations during the summer months. Relationship of the Reproductive Cycle to Environmental Data At station A, gametogenesis started when the water temperature was near 10°C in the spring of 1970 (Figure 7). Ripe clams were first observed when the water temperature was 16°C and spawning was first noted when the salinity was between 3 and S'/oo. Gametogenesis and spawning occurred through the summer period of high salinities and high temperatures. The fall spawn- ing started at the highest salinities of the year with a definite major spawning after the salinity dropped IQP/oo (Figure 7). The temperature during the fall spawning was 13°-15°C. Set first appeared the second week in November and occurred throughout the winter months. Fall and winter sets generally were accompanied by temperatures below 10°C and occasionally to 1°C. Game- togenesis was in progress again by the time the water temperature reached 15°C. No set clams were collected during the summer of 1971, although the histological sections showed all stages of development. Salinities and tempera- tures approximated those of the previous year, although salinity was more variable. Renewed gametogenesis coincided with the highest temperature of the summer and rising salinities. Some spawning took place with the declining salinity, but the major spawning occurred near 5Voo. Gametogenesis at station B started almost 2 wk later than at station A in the spring of 1970, with the temperature above 13°C (Figure 8). Ripe clams were similarly noted nearly 3 wk later than at station A. The salinity during the summer months was near 5''/oo and gradually rising. There was lit- tle agreement between spawning times as noted in the histological sections and setting in the collectors during the summer. The progression of gonads in the fall from ripe to spent was clearer and more defined at station B than at station A. Spawning commenced when the salinity reached the yearly high and peaked when the salinity fell rapidly from 15 to l^/oo. Setting took place 2 wk later and continued into the winter. Temperatures during spawning ranged from 22° to 12°C. Fewer clams were in the ripe phase throughout the summer of 1971 when the water was nearly fresh. More clams became ripe as the salinity increased, but few set were collected all summer at this sta- tion. Fall spawning commenced at 22°C and 6^/00 and continued until the salinity reached approximately zero and the temperature dropped to 17°C. Some set were collected immediately after salinity decline and setting continued into January 1972. Spawning was completed at station C by the end of October 1970, after salinity had fluctuated from 0 to 5''/oo for the previous 2 mo. Setting began at temperatures above 25°C and continued throughout the winter at low water temperatures and low salinities (Figure 9). The salinity at sta- tion C remained below V/oo until the termination of the study in January 1972. There was very little spawning and setting at station C during the low salinity period even though gametogenesis took place normally. 425 FISHERY BULLETIN: VOL. 73, NO. 2 DISCUSSION The histological examination of gonads in- dicates several generalizations about the reproductive cycle of R. cuneata. Gametogenesis began in early April and continued throughout the summer months. Ripe gametes v^^ere observed from May to late November. A slight spawning peak w^as noted during the summer, but a major spawn occurred in the fall. This was probably not a second cycle because gametogenesis in most cases had not terminated during the summer; instead, it appeared to be a continuation of gamete develop- ment at an increased rate. The gametogenic cycle of Rangia in high and low salinities was basically the same. Temperature appeared to be the more important stimulus in initiating gametogenesis in the spring and summer. A temperature of approximately 15°C coincided with initiation of gametogenesis at all stations. Gametogenesis in clams from freshwater occurred at a slower rate during the spring and summer with more clams in the spent phase than at the other localities. The reproductive cycle as determined in this study is similar to that reported for Rangia in other geographic areas. Fairbanks (1963) indicat- ed that in Lake Pontchartrain, La., ripe clams could be found during March, April, May, and in the late summer and fall. A prolonged spawning season would seem reasonable as the rise in water temperature to 15°C in the spring is nearly 2 mo earlier at that location than in the James River. In addition, the drop in water temperature to 15° C in the fall is later. He also indicated a postspawning recovery phase during midsummer. The very high temperature (near 33°C) during the summer may have inhibited gametogenesis, but in the James River population a renewed surge occurred at the high midsummer temperatures of 28°-29°C. Tenore (1970) studying the macrobenthos of the Pamlico River, N.C., found Rangia containing mature gametes only in the fall. This observation was probably based on visual inspection of the viscera. The spring and summer ripening may have been missed because the gonads are not nearly as distended and colored as in the fall ripening. Pfitzenmcyer and Drobeck (1964) collected Rangia in August and September from the Potomac River. Clams at this time contained mature gametes, indicating that spawning was imminent. The correlation of the environmental data to gonadal conditions suggests that temperature and salinity are important factors in spawning. Salinity, however, was more important than temperature. Clams upstream at station C spawned in fall 1970 following a 5"/oo rise in salinity, but failed to spawn in 1971 when the salinity remained low. Spawning at station B was apparently related to salinity decreases. The correlation of salinity to spawning was not as clear at station A. The salinity variation at this station was very large over a tidal cycle and may have prevented complete synchrony of spawning. Cain (1973) found that a salinity change was necessary for Rangia spawning in the laboratory. Spawning was accomplished by placing ripe clams from low salinities ( < iVoo) into 5Voo, 28°C flow- ing water. Evidently clams in upstream areas require a rise in salinity to spawn, while down- stream populations require a reduction in salinity from the 10 to 15"/oo levels at which they live. There is little additional evidence in the litera- ture on the importance of salinity to Rangia spawning. Fairbanks (1963)could not induce spawn- ing. Chanley (1965) induced spawning at 15 ''/oo by rapidly increasing the water temperature 7°C and adding sperm stripped from a ripe male. However, spawning was poor and he did not study the survival of the eggs. Such strong stimuli may cause premature release of immature eggs, with subsequent poor fertilization and survival. The only data suggesting the importance of salinity to Rangia spawning is that from the Bureau of Sport Fisheries and Wildlife (1965) in connection with a study on the food habits of ducks in Back Bay, Va., and Currituck Sound, N.C. These two bodies of water normally have a salinity of less than l^/oo. During that study, a storm forced ocean water into the bay and raised the salinity to about 4.5''/(».This intrusion of salt water must have caused spawning and successful setting since the following year nearly 9% of the diet of dabbling and diving ducks consisted of small Rangia, an amount estimated to be 83,000 lb (dry weight) for the year. During the previous 3 yr no Rangia were consumed by these ducks. Two periods of setting occurred in the James River. The first was in early and midsummer, which coincided with the beginning of spawning as inferred from tissue sections. The second period was much longer, with a greater number of collected set, and took place in the late fall and winter, coinciding with the increased percentage 426 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA of partially spawned and spent clams, as identified by histological preparations. The second peak of spawning appeared to be the major one for the normal reproductive period (Table 1). Fairbanks (1963) found set (>0.375Aim) from October to April. A longer and more intense setting period was found in the area with the more variable salinity. Tenore (1970) collected set in bottom grabs only during the fall and winter months. The spring and summer spawning was either so light in these areas that no set were found or an abun- dance of predators at this time quickly consumed the sparse set. Sex Ratio There are at least two possible explanations for the unusual sex ratios in Rangia. Rangia may be protandric with a higher ratio of females to males in the older stages. If so, the clams at stations B, D, and C would have had to be older than those at station A; however, there were no consistent differences in the lengths of the clams at the various stations. This does not preclude the pos- sibility that the ages of clams at the stations are different, but masked by varying growth rates resulting from substrate effects or nutrient levels. The second possibility is that the environmental conditions upriver differ enough from those downriver at station A to affect the sex ratios. Changes in environmental factors could affect either juveniles or adults during the undifferen- tiated period. There is no proof of this type of sex alteration. None of the other mactrids studied, M. lateralis, S. solidissima, or T. capax have been found to have a ratio other than 1:1. Relationship of Larval Studies to Setting Cain (1973), on the basis of laboratory results, indicated that best survival and growth of larvae would be expected in the summer. Higher temperatures and generally high salinities are expected to provide for rapid growth to setting. But setting is very poor in summer (Table 1). A number of factors may account for this: clams are in all stages of gametogenesis, the gonads of ripe clams are not as full, and even though the stimulus to spawn may exist there is probably poor synchrony of spawning. In the James River, fall and winter were the times of greatest setting activity. Fall spawning occurred as the temperature dropped from 29° to 15° C. This temperature range would provide good survival of eggs to straight-hinge larvae, but the larvae are exposed to declining fall temperatures. Larvae at low temperatures (and low salinity) should survive well, but grow slowly. Con- sequently, set in the late fall and winter come from a fall spawning after the delay in growth and meta- morphosis expected from low temperatures. Some set in the jars could have come from previously set, slow-growing clams washed in by turbulence. This set, which is fairly active, tends to crawl over the bottom by use of the muscular foot and therefore is affected by currents (Carriker 1961). Distribution and Recruitment of Rangia in the James River Rangia extends to nautical mile 60 in the James River. In the upper reaches of its distribution it has been in freshwater for the last 4 or 5 yr. Since the embryos and early straight-hinge larvae can- not tolerate freshwater (Cain 1973) and a salinity rise is needed to induce spawning, there must be little recruitment to this population. No set or small clams have been found in this area, which raises the similar question of how Rangia spread into this region initially. The upriver population consisted of clams ranging in length from 53 to 63 mm in the spring of 1971. Using the von Ber- talanffy growth curve constructed for Rangia by Wolfe and Petteway (1968), these clams were es- timated to be from 5 to 7 yr old. The water records for the James River basin were analyzed from 1963 to 1966 (Anonymous 1966, 1970b). These records show yearly lows in the late summer and fall when the input dropped below 22.66 mVs (800 cf s), and in 1966 the input dropped to an average of 13.59 mVs (480 cfs) during the first half of Sep- tember. These very low flow periods allowed measurable-salinity water to extend 63.5 miles upstream in December 1965 (Brehmer and Hal- tiwanger 1966). To reach these upstream areas larvae must be transported in the more saline bottom water which has a net upstream movement (Pritchard 1952; Nichols 1972). Although the mechanism of such transport has not been deduced for Rangia, work done on the eastern oyster may indicate some possible mechanisms. Wood and Hargis (1971), who studied the lower James estuary, found a definite upstream transport of oyster larvae and 427 FISHERY BULLETIN: VOL. 73, NO. 2 also of small coal particles. They indicated that the larvae move upstream by selectively swimming in more saline water associated with the flood tide. They further indicated that Korringa's (1952) idea of passive transport could not be denied, as the coal particles also had a net upstream motion. Rangia could be carried into upstream areas both by selective swimming or by passive transport under low flow conditions, and a series of dry years would allow set to progressively move upriver. Set from one year should be able to spawn the next year and certainly within 2 yr. Gonads occurred in clams 14 mm long, a length easily reached by the end of the first year (Fairbanks 1963; Wolfe and Petteway 1968). Rapid early growth and a rela- tively short larval life (above 20°C) should allow for the fast spread of set into areas uninhabited by adults. As Rangia has an 8-yr average life span (Fairbanks 1963) and a maximum life of 14 yr (Wolfe and Petteway 1968), recruitment to the population could occur at fairly long intervals. The Virginia Division of Water Resources (Anonymous 1970a) statistically predicts that low flows of less than 1,000 cfs for 7 days will occur at 5-yr intervals. This situation could allow minimum recruitment to maintain the upstream population, assuming good survival of set and adults in this region. The downstream extent of Rangia could be de- termined by the adults approaching their high- salinity limit. This may not be the case as the larvae can survive higher salinities than normally occur at the downstream limit— and the adults may also do so. The downstream extent likely represents a multifactor barrier involving biological competition for space and food, and increased predation of the set. Recruitment to the downstream population was at a low level but more regular, with more set collected in the summer months than at the up- stream stations (Table 1). Probably the best recruitment would be expected in the population at station B near the lower middle of the habitat range. The fall set there was high and fairly con- sistent over the 2-yr period. Averaged over many years, this segment would likely receive more set, as this part of the estuary usually has an annual salinity change from fresh to S^/oo. General Discussion The adults can utilize the high levels of detritus in this oligohaline sector (Darnell 1958) and con- vert it into growth and reproductive materials. Rangia is ripe for at least 7 mo of the year so it can spawn whenever favorable changes in salinity allow successful reproduction. Although adults are euryhaline, embryos are much more sensitive. Spawning at a salinity near 5°/oo allows for the survival of the sensitive stages to the more eury- topic later larval stages. The increased tolerance of the larvae permits good survival during its more stressful pelagic existence. The planktonic existence of Rangia larvae is greatly extended by low temperatures. Thorson (1946) indicated that prolonged low temperatures exposed larvae to increased mortality from disease, starvation, predation, dispersion, and en- vironmental stress. Rangia larvae evidently are well adapted to a prolonged exposure because many set were collected during the coldest months. This increase in dispersal may allow Rangia to consume the unexploited resources of the species-poor environment in low salinity. Increased dispersal may also provide genetic in- terchange between populations distributed over a relatively extended habitat in the estuary. Pfit- zenmeyer and Drobeck (1964) found the rate of increase of Rangia over a 4-yr period in the Po- tomac River to be very great. Pfitzenmeyer (1970) also recorded this clam in the upper Chesapeake Bay when its numbers increased from 0 to 10,000 per square meter in 2 yr. Spawning of Rangia apparently is controlled by salinity change. The mechanism of this control, however, has not been examined. The exogenous factor of salinity may activate an endogenous control system of osmoregulation and serve as a signal to induce synchronous spawning of the population. The concept of "critical salinity" reviewed by Khlebovich (1969) appears operative for Rangia which spawns near 5'Voo. Khlebovich concluded that the salinity range of 5-8''/oo is a faunal boundary defining the distribution of marine and freshwater species. Characteristic differences in physiological performances (adap- tation, growth, activity and, especially, os- moregulation) are revealed at this salinity range. The latitudinal distribution of Rangia is impor- tant because of its spread northward in the last decade. In the present study, gametogenesis and spawning were observed to occur over a wide range of temperatures. Although larval growth was best at high temperatures, survival and growth apparently take place even at low temperatures. Consequently, it appears that the 428 CAIN: REPRODUCTION AND RECRUITMENT OF RANGIA CUNEATA northward limit of Rangia is not controlled by low temperature effects on reproduction or larval tolerance. Reports on populations in upper Chesapeake Bay (Gallagher and Wells 1969; Pfit- zenmeyer 1970) infer large adult mortalities from low temperature and low salinity. The only mor- tality of adults (freshly gaped clams) in this study was seen in the dredge hauls during the winter and early spring. Long periods of low temperature and freshwater may combine to increase the mor- tality of adults and thereby limit the northward range. The possibility that widely separated popula- tions may belong to different physiological races cannot be excluded (Loosanoff 1969). Populations of marine animals exposed to different environ- ments within their geographical range may have different physiological properties (Sastry 1970), suggesting that more research should be conduct- ed on the tolerances of embryos, larvae, and adults from different geographical areas. The key to the welfare of a Rangia population over its normal distribution is probably not the physiology of the adult individual but successful reproduction and recruitment. Adults may live for years in habitats where reproduction is impossible. Spawning will not occur unless salinity changes, up from low salinity or down from high salinity. If spawning does occur, embryos and early larvae will have poor survival unless salinity is between 2 and lOy 00 (Cain 1973). Once the larvae have developed past the swimming stage and settled to the bottom as juvenile clams, salinity is probably not as critical (except in combination with low temperatures). The influence of salinity on reproduction and recruitment indicate that some changes in its en- vironment may restrict the distribution of Rangia in the estuary. Rangia's estuarine distribution is maintained by changes in salinity related to variations in the freshwater input. Any overall reduction in freshwater input or reductions in the seasonal variations of salinity will limit its range. These characteristics of this species should be considered before dams are constructed, fresh- water is diverted for other uses, or other changes in the hydrography of the estuary are approved. ACKNOWLEDGMENTS For critically reviewing the manuscript, I wish to thank the members of my graduate committee, Morris L. Brehmer, Dexter Haven, Joseph G. Loesch, and Clinton E. Parker. I am indebted to my major adviser, Marvin L. Wass, for his help during this study and for his careful review of the manuscript. I am especially grateful to my colleague, Richard Peddicord, for his helpful suggestions and assistance with the field work. The work done by the Histology Department of the Virginia Institute of Marine Science, especially Patsy Berry, on the gonadal sections is appreciated. I also thank Samuel Rivkin, for his advice and the Algae Department for providing algae food when needed. I am especially indebted to my wife, Diane, for her encouragement and help. LITERATURE CITED Anonymous. 1966. Water Resources Data for Virginia - Part I Surface Water Records. U.S. Geol. Surv. 1970a. James River Basin, Comprehensive Water Resources Plan. Va. Dep. Conserv. Econ. Dev., Div. Water Resour., Planning Bull. 215, Vol. III. 1970b. Surface Water Supply of the U.S. 1961-1965: Part II South Atlantic slope and eastern Gulf of Mexico basins. Vol. I Basins from James River to Savannah River. U.S. Geol. Surv. Pap. No. 1905. Brehmer, M. L., and S. Haltiw anger. 1966. A biological and chemical study of the tidal James River. Va. Inst. Mar. Sci., Final Rep. Fed. Water Pollut. Contr. Admin., Contract PH86-65-86, Nov. 1966, 32 p. Bureau of Sport Fisheries and Wildlife, North Carolina Wildlife Resources Commission, and Virginia Com- mission OF Game and Inland Fisheries. 1965. Back Bay - Currituck Sound data report. Vol. 1, 84 p. Cain, T. D. 1973. The combined effects of temperature and salinity on embryos and larvae of the clam Rangia cuneata. Mar. Biol. (Berl.) 21:1-6. 1974. Combined effects of changes in temperature and salinity on early stages of Rangia cuneata. Va. J. Sci. 25:30-31. Calabrese, a. 1970. Reproductive cycle of the Coot clam, Mulinia lateralis (Say) in Long Island Sound. Veliger 12:265-269. Carriker, M. R. 1961. Interrelation of functional morphology, behavior, and autecology in early stages of the bivalve Mercenaria mercenaria. J. Elisha Mitchell Sci. Soc. 77:168-241. Chanley, p. 1965. Larval development of the brackish water mactrid clam, Rangia cuneata. Chesapeake Sci. 6:209-213. Coffin, G. W., and W. R. Welch. 1964. A technique for separating small mollusks from bot- tom sediments. Proc. Natl. Shellfish Assoc. 53:175-180. Darnell, R. M. 1958. Food habits of fishes and larger invertebrates of Lake Pontchartrain, Louisiana, an estuarine communi- ty. Publ. Inst. Mar. Sci. Univ. Tex. 5:353-416. Fairbanks, L. D. 1963. Biodemographic studies of the clam Rangia cuneata Gray. Tulane Stud. Zool. 10:3-47. 429 FISHERY BULLETIN: VOL. 73, NO. 2 Gallagher, J. L., and H. W. Wells. 1969. Northern range extension and winter mortality of Rangia cuneata. Nautilus 83:22-26. Hopkins, S. H. 1970. Studies on brackish water clams of the genus Rangia in Texas. (Abstr.) Proc. Natl. Shellfish Assoc. 60:5-6. Hopkins, S. H., and J. D. Andrews. 1970. Rangia cuneata on the east coast: Thousand mile range extension, or resurgence? Science (Wash., D.C.) 167:868-869. Khlebovish, V. V. 1969. Aspects of animal evolution related to critical salinity and internal state. Mar. Biol. (Berl.) 2:338-345. Korringa, p. 1952. Recent advances in oyster biology. Q. Rev. Biol. 27:266-308, 339-365. LOOSANOFF, V. L. 1969. Maturation of gonads of oysters Crassostrea vir- ginica, of different geographical areas subjected to rela- tively low temperatures. Veliger 11:153-163. Machell, J. R., AND J. D. De Martinl 1971. An annual reproductive cycle of the gaper clam, Tresus capax (Gould), in south Humboldt Bay, Califor- nia. Calif. Fish Game 57:274-282. Nichols, M. M. 1972. Sedimentsof the James River estuary, Virginia. InB. W. Nelson (editor), Environmental framework of coastal plain estuaries, p. 169-212. Geol. Soc. Am. Mem. 133. OdUM, H. T., and B. J. COPELAND. 1969. A functional classification of the coastal ecological systems. In. H. Ti Odum, B. J. Copeland, and E. A. McMahan (editors), Coastal ecological systems of the United States, p. 9-86. Rep. Fed. Water Pollut. Control Admin., Wash., D.C. Pfitzenmeyer, H. T. 1970. Project C. Benthos. In Gross physical and biological effects of overboard spoil disposal in upper Chesapeake Bay, p. 26-38. Univ. Md. Nat. Resour. Inst., Spec. Rep. 3. Pfitzenmeyer, H. T., and K. G. Drobeck. 1964. The occurrence of the brackish water clam, Rangia cuneata, in the Potomac River, Maryland. Chesapeake Sci. 5:209-212. Pritchard, D. W. 1952. Salinity distribution and circulation in the Chesapeake Bay estuarine system. J. Mar. Res. 11:106-123. RoPES, J. W. 1968. Reproductive cycle of the surf clam Spisula solidis- sima, in offshore New Jersey. Biol. Bull. (Woods Hole) 135:349-365. ROPES, J. W., AND A. P. Stickney. 1965. Reproductive cycle of Mya arenaria_ in New England. Biol. Bull. (Woods Hole) 128:315-327. Sastry, a. N. 1970. Reproductive physiological variation in latitudinally separated populations of the bay scallop, Aequipecten irradians Lamarck. Biol. Bull. (Woods Hole) 138:56-65. Tenore, K. R. 1970. The macrobenthos of the Pamlico River estuary. North Carolina. Ph.D. Thesis. North Carolina State Univ., Raleigh, 121 p. Tenore, K. R., D. B. Horton, and T. W. Duke. 1968. Effects of bottom substrate on the brackish water bivalve Rangia cuneata. Chesapeake Sci. 9:238-248. Thorson, G. 1946. Reproduction and larval development of Danish marine bottom invertebrates, with special reference to the planktonic larvae in the sound (Oresund). Medd. Komm. Dan. Fisk.-og. Havunders. Ser. Plankton. 4(l):l-523. Wass, M. L., and T. D. Wright. 1969. Coastal wetlands of Virginia. Va. Inst. Mar. Sci., Spec. Rep. Appl. Mar. Sci. Ocean Eng. 10, 154 p. Wolfe, D. A., and E. N. Petteway. 1968. Growth of Rangia cuneata Gray. Chesapeake Sci. 9:99-102. Wood, L., and W. J. Hargis, Jr. 1971. Transport of bivalve larvae in a tidal estuary. In B.J. Crisp (editor). Fourth European Marine Biology Sym- posium, p. 29-44. Cambridge Univ. Press, Lond. 430 HORMONAL-INDUCED OVULATION OF THE WINTER FLOUNDER, PSEUDOPLEURONECTES AMERICANUS Alphonse S. Smigielski' ABSTRACT The response of winter flounder, Pseudopleuronectes americanus, to human chorionic gonadotropin (HCG), oxytocin, pregnant mare serum gonadotropin (PMSG), deoxycorticosterone (DOCA), and freeze-dried carp pituitary is described. HCG and PMS were successful in some instances in producing viable eggs and larvae while carp pituitary was successful in all instances. These were the first known successful attempts to induce maturation and spawning of winter flounder artificially in the laboratory. Larvae obtained from these hormonal-induced spawnings were normal in all respects and were reared in the laboratory through metamorphosis. Wild plankton obtained from Narragansett Bay, brine shrimp nauplii, and chopped clams were fed as food. The early life history of this flatfish can, for the first time, be completed under controlled laboratory conditions. The winter flounder, Pseudopleuronectes americanus (Walbaum), an important species in local New England commercial and sport fishing industries, occurs from Chesapeake Bay to the northern shore of the Gulf of St. Lawrence (Bigelow and Schroeder 1953). Winter flounder spawn from January through April in Rhode Island estuaries with a peak spawning period in February. Demersal eggs are produced which range from 0.74 to 0.85 mm in diameter. Investigations to evaluate methods of inducing spawning of winter flounder with the aid of hor- mones under controlled laboratory conditions were undertaken at the National Marine Fisheries Service, Northeast Fisheries Center, Narragan- sett Laboratory in 1970 and continued over a 3-yr period. The use of hormone injections for inducing spawning in other species of fish has been well documented by Pickford and Atz (1957). The in- duction of spawning in winter flounder in the laboratory provides a practical method of supply- ing viable eggs and larvae for physiological studies. By controlling water temperatures, pho- toperiods, and injecting hormones, the research time for this species can be extended. As far as is known, these were the first reported successful attempts to artificially mature and spawn the winter flounder in the laboratory. 'Northeast Fisheries Center, National Marine Fisheries Serv- ice, NOAA, Narragansett, RI 02882. Manuscript accepted June 1974. FISHERY BULLETIN: VOL. 73, NO. 2, 1975. MATERIALS AND METHODS Adult winter flounder were captured by otter trawling in Narragansett Bay in the autumn of 1970, 1971, and 1972, and were brought to the Narragansett Laboratory in a 380-liter live car equipped with an aerator. In the laboratory, the fish were held in 1,890-liter circular aquaria (1.2 m in diameter; water depth, 0.8 m). A continual sup- ply of filtered seawater was pumped to the aquaria from Narragansett Bay. After acclimating in the laboratory, the fish were segregated by sex, measured and weighed, and tagged with numbered plastic pennants secured through the caudal peduncle with a double barbed stainless steel wire. Winter flounder adults may be sexed easily by placing the fish on its back and running a hand down the white underside. Females are quite smooth while the males are very rough to the touch. Winter flounder lend them- selves quite readily to handling and have proven to be a durable flsh. No anesthesia was required prior to injecting hormones as the fish rarely struggled. An artificial photoperiod of 9 h light and 15 h darkness (9L:15D) simulating seasonal light con- ditions was maintained for all experiments by time clocks controlling four banks of 80-W cool white fluorescent lights suspended 1.2 m above the tanks. Since winter flounder fast during the spawning season, as many other fish do, their food regimen was not difficult to maintain. A varied diet of clams, squid, chopped menhaden, and silversides 431 FISHERY BULLETIN: VOL. 73, NO. 2 was provided. Very few females fed actively. However the males, a majority of which were running ripe, fed throughout the experiments. Hormone injections were carried out with 0-2 cc syringes fitted with 20 gauge, 3.85 cm (1.5 inch) needles. All injections were intramuscular. In- traperitoneal injections were ruled out for fear of injuring or killing the experimental fish. Injec- tions were made into the back muscle below the dorsal fin. Inserting and withdrawing the needle slowly aided in retaining most of the fluid in the fish (Figure 1). After injection, the flesh of the fish was massaged to diffuse the fluid into the muscles. A saline solution of isotonic sodium chloride was used as a carrier for all hormone injections except with deoxycorticosterone, which was mixed with sesame oil and injected as a slurry. Hormones tested in the various studies were: human chorionic gonadotropin (HCG), oxytocin. k Figure l.-Winter flounder receiving hormone injection. deoxycorticosterone (DOCA), pregnant mare serum gonadotropin (PMSG) and carp pituitary (freeze-dried powder). The criteria chosen to test the effectiveness of the hormones were gonad in- dex, spawning, fertilization of eggs, and hatching success. Hormones were prepared on the day of injection, and dosages were established by the weight of each individual fish. Running ripe fish were stripped by hand. Winter flounder spawn adhesive demersal eggs which form clumps under experimental conditions. Spawned and fertilized eggs were handled and treated according to the separation techniques of Smigielski and Arnold (1972). Human chorionic gonadotropin was the first hormone selected for evaluation. Stevens (1966) reported successful spawnings of striped bass with dosage levels of this hormone ranging from 31 through 403 lU /pound fish. Two hormone dosage levels and two time sequences of injecting were administered to three groups of test fish number- ing five per group at dosage levels of 150 IU/454 g fish injected daily, 300 IU/454 g fish injected daily, and 300 IU/454 g fish injected every other day. A fourth group served as controls. Water tempera- tures during the first testing trial ranged between 5° and 7°C with a mean of 6.2° C. A second series of experiments was initiated when the water temperature fell to 4°C. Water temperatures during these trials ranged between 3° and 5°C with a mean of 3.2°C. A test group numbering 15 fish was established and held in three 1,890-liter aquaria. Dosage levels of 150 IU/454 g fish were dispensed daily and adminis- tered over a time period of 12 days. Further experiments were initiated to evaluate oxytocin, DOCA, and PMSG at the two water temperature ranges of 6°-7.5°C and 3°-5°C. Mean water temperatures during the trials were 7.2° and 3.2°C. Because of variable results obtained with the hormone HCG in prior experiments, HCG was included in these trials for further testing. Three dosage levels of each hormone were evaluated and a total of five injections adminis- tered over a period of 10 days. The dosage levels were: oxytocin at 10, 20, and 40 IU/454 g fish; DOCA at 5, 10, and 20 mg/454 g fish; PMSG at 55, 110, and 220 IU/454 g fish; and HCG at 100, 200, and 400 IU/454 g fish. Female test fish were measured, weighed, tagged, and placed into four 1,890-liter circular tanks supplied with a continual supply of seawater. Each tank contained a total of 432 SMIGIELSKI: OVULATION OF WINTER FLOUNDER 16 fish. The three differing dosage levels of each test hormone were injected into groups consisting of four fish per group, and the remaining four fish in each tank served as uninjected controls. At the termination of the trial, test fish were sacrificed, ovaries examined, and gonosomatic index (GSI) levels recorded. A final series of experiments was initiated to evaluate the effectiveness of carp pituitary (freeze-dried powder) at the dosage levels of 5.0 mg and 0.5 mg/454 g fish injected daily. Two groups of test fish with each group consisting of three trial fish and three uninjected controls were established and held in two 1,890-liter tanks. Dur- ing the trials, water temperatures ranged between 1.5° and 3.5°C with a mean at 2.5°C. RESULTS AND DISCUSSION In the group of test fish that received HCG in- jections at 300 lU daily, one fish hydrated after a single injection (Table 1). Hydration, an increase in total body weight due to water uptake by the ovaries resulting in higher GSI levels, was rapid, and the following morning this fish was grossly bloated. Externally, hydrated fish may become slightly swollen or grossly bloated. The majority of the eggs stripped were opaque and misshapen, and approximately 5% of the total egg mass in the ovaries were viable. These eggs, numbering approximately 10,000, were fertilized but embryonic development ceased in the blastula stage and none survived to hatch. This rapid hydration may have been responsible for the poor quality of the eggs. A long development period for oocytes with a minimum of 2 or 3 yr from the time Table l.-Effects of human chorionic gonadotropin (HCG) on Pseudopleuronectes americanus, temperature range 5°-7°C (J 6.2°C). Photoperiod 9L/15D. Fertili- Hormone and No. of No. of Number Number zation Hatch dosage fish injections hydrated ovulated (%) (%) 300 lU HCG/ 454 g fish daily 1 1 1 1 15 0 300 lU HCG/ 454 g fish daily 4 12 0 0 — — 300 lU HCG/ 454 g fish every 48 h 5 7 1 21 95 0 150 lU HCG/ 454 g fish daily 5 12 2 0 — — Controls 6 0 1 0 - — 'Approximate. ^Ovulated ar Id spawned 10 days after last injection. oocytes become histologically recognizable until they are spawned has been suggested for winter flounder by Dunn and Tyler (1969). Our observa- tions in the past have noted that the natural hydration and ovulation process of winter flounder can occupy a lengthy period of time (unpubl. data). The remaining fish in this test group were refrac- tory after receiving a total of 12 injections. The group of five test fish that received injec- tions of 300 lU every other day (for a total of seven injections in 13 days) contained three fish which were refractory at the conclusion of the trial; one fish displayed signs of slight hydration, but did not ovulate. The remaining fish in the group ovulated and was spawned 10 days after the last injection. Approximately 95% of the eggs were fertilized but development ceased at the blastula stage and none survived to hatch. The group of test fish which received daily in- jections of 150 lU for 12 days contained three fish which were refractory and two which hydrated slightly but did not ovulate. Of the six fish that served as uninjected controls, five displayed no signs of hydration, and the remaining fish hydrated slightly but did not ovulate. At the termination of the trials, the fish were not sacrificed. The only criterion for evaluating the success of the hormone was obtaining viable eggs. It was reasoned that the relatively warm- water temperatures (5°-7°C) coupled with sus- pected low GSI levels inhibited the effectiveness of the hormones. Observations made in the inner parts of the Gulf of Maine, (Bigelow and Schroeder 1953) have shown that extensive spawning of winter flounder does not occur in water tempera- tures above 6°C. At the conclusion of the second series of experiments, four fish hydrated and ovulated after receiving from two to nine injections (Table 2). Fertilization was high for all fish but embryonic mortalities were high in the blastula and gastrula stages, and hatches were poor, ranging from 2 to 20%. Larvae obtained from these spawnings ap- peared to be normal in all respects and several were reared through metamorphosis. Three addi- tional test fish became grossly bloated after receiving a total of 10 injections in 10 days and hydrated to the point of dying. Membranous plugs formed in their oviducts and the eggs were water hardened. The formation of these plugs is not un- derstood as prior to receiving their last injection they were hydrating at a normal rate. Shehadeh and Ellis (1970) reported plugs forming in striped 433 FISHERY BULLETIN: VOL. 73, NO. 2 Table 2.— Effects of hormones on Pseudopleuronectes americanus. Temperature range 3°-5°C {i 3.2°C). Photoperiod 9L/15D. Symbols ( + = Did, 0 = Did not hydrate or ovulate). Total Initial Weight change Fertili- Hormone and Number of length body weight at termination GSP zation Hatch dosage injections (mm) (g) (% initial wt.) (% final wt.) Hydrated Ovulated (%) (%) 'HCG 150 IU/454 2 271 271 -3.32 1.29 + + 80 2 g fish daily 4 316 792 +0.76 1.08 + + 90 5 8 349 640 +0.63 1.24 + + 80 10 9 370 692 +0.58 1.15 + + 80 20 10 360 545 +3.67 32.74 + 30 — — 10 290 275 +5.45 30.34 + 30 — — 10 421 982 +2.75 30.13 + 30 — — 12 391 670 +0.60 16.02 + 0 — — 12 334 571 -0.35 8.08 0 0 — — 12 377 486 +0.62 7.57 0 0 — — 12 317 562 + 1.60 16.46 + 0 — — 12 342 454 +0.44 10.13 0 0 — — 12 280 352 + 1.42 15.13 + 0 — — 12 336 599 +0.33 9.48 0 0 — — 12 352 800 +0.50 20.15 + 0 — — Control 0 341 721 + 1.05 1.14 + + 85 75 0 330 540 + 1.11 19.60 + 0 — — 0 321 449 + 1.56 19.08 + 0 — — 0 337 497 + 1.01 18.73 + 0 — — 0 271 299 -0.67 10.24 0 0 — — 0 269 356 +2.25 12.09 0 0 — — 'Gonadosomatic index. ^Human chorionic gonadotropin. 3Plug formed, fish became grossly bloated and died, eggs were water hardened in ovaries. mullet, Mugil cephalus, while attempting induced spawning. The findings at the conclusion of the second series of experiments indicated that the injection of HCG, although more effective at the lower water temperatures, resulted in poor egg quality and egg survival, the formation of membranous plugs, and gross hydration causing death. Oxytocin at all three dosage levels tested had little effect when water temperatures were above 5°C. Three fish hydrated but none ovulated or contained matured eggs (Table 3). DOCA administered at the lower dosage levels of 5 and 10 mg/454 g fish and at water tempera- tures above 5°C, resulted in several fish hydrating but none ovulating. At the higher dosage level of 20 mg/454 g fish, two test fish hydrated, and although their GSI levels were high, no ovulation occurred and no mature eggs were present in their ovaries. PMSG at the three dosage levels of 55, 110, and 220 IU/454 g fish showed low activity when ad- ministered at water temperatures above 5°C. Although some fish had hydrated, none ovulated and GSI levels were low in all of the test groups. HCG was ineffective at the dosage levels of 100 and 200 lU. Dosage levels of 400 lU resulted in three fish hydrating at low GSI levels. Another fish ovulated and was stripped 3 days after the last injection. Egg fertilization was high and approximately 80% of the eggs hatched. The larvae obtained from this induced spawning appeared to be normal in all respects and several were reared through metamorphosis. The data gleaned from this trial at water temperatures in the 6°-7.5°C range substantiated results derived from earlier experiments, sug- gesting that water temperatures above 6°C inhibit maturation of winter flounder, and for the most part hormones are ineffective. The manner in which hormones exert their effects on fish is poorly understood, and dosage levels are probably meaningless as the largest fish may not neces- sarily be the most sexually mature and may differ in receptibility to hormone injections. Haydock (1971) has observed a temperature threshold of 17°C below which the gulf croaker, Bairdiella icistia, would not hydrate or ovulate. It is very probable that a similar temperature threshold exists for winter flounder above 6°C. Observations in the past at our laboratory have noted that gravid female flounder perished in water temperatures of 10°C, and ova from these fish were stunted and misshapen. Male flounder held under the same conditions suffered no ap- parent ill effects. In the lower temperature range tested (range 3°-5°C, mean 3.2°C), oxytocin produced slightly better results (Table 4). Three fish hydrated but none ovulated nor contained mature eggs. GSI levels were higher at the lower temperatures at all dosage levels tested. No abnormal hydration was 434 SMIGIELSKI: OVULATION OF WINTER FLOUNDER Table 3.-Effects of hormones on Pseudopleuronectes americanus given five injections over a 10-day period. All fish experienced a 9L:15D photoperiod, water temperature range 6°-7.5°C (x 7.2°C). Symbols ( + = Did, 0 = Did not hydrate or ovulate). Total Initial Fertili- Hormone and length body weight Weight change GSI' zation Hatch dosage (mm) (g) (% initial wt.) (% final wt.) Hydrated Ovulated (%) {%) Oxytocin 386 716 +3.91 13.87 + 0 10 IU/454 336 495 0 15.25 0 0 g fish daily 308 344 +0.87 17.87 0 0 333 504 -0.79 14.00 0 0 Oxytocin 323 453 +2.21 17.17 + 0 20 IU/454 305 350 -0.29 14.90 0 0 g fish 345 477 -1.68 9.81 0 0 370 662 -0.30 12.00 0 0 Oxytocin 306 409 +0.98 14.53 + 0 40 IU/454 338 377 -1.06 13.94 0 0 g fish 308 404 -0.50 8.83 0 0 349 471 0 16.99 0 0 Control 346 443 +0.45 13.82 0 0 296 327 + 1.22 13.14 0 0 302 389 +3.08 13.19 4- 0 370 634 + 1.58 16.61 0 0 2D0CA 5 mg/454 340 456 +2.63 1.39 0 0 g fish 345 549 +3.10 15.37 + 0 372 540 +5.74 14.97 + 0 299 310 +9.68 13.09 + 0 2D0CA 10 mg/454 287 270 + 13.70 11.24 + 0 g fish 283 272 +7.35 8.39 + 0 300 354 -0.28 12.75 + 0 345 525 +5.71 15.05 + 0 2D0CA 20 mg/454 392 726 +1.38 4.08 0 0 g fish 306 338 +4.73 7.20 0 0 376 726 +5.79 19.79 + 0 375 516 + 15.70 28.14 + 0 Control 295 327 + 1.22 12.24 0 0 323 408 +3.67 14.07 + 0 356 526 + 1.71 24.58 + 30 342 480 +2.92 17.21 + 0 "PMSG 55 IU/454 315 448 + 1.79 11.84 + 0 g fish 321 420 -0.24 9.07 0 0 381 652 +0.46 10.23 0 0 358 582 +6.19 16.34 + 0 "PMSG 110 IU/454 294 318 + 1.57 0.93 50 0 g fish 349 546 -0.92 12.29 0 0 349 219 +1.83 9.42 0 0 303 341 +4.69 13.45 + 0 "PMSG 220 IU/454 335 488 +2.46 18.00 + 0 gfish 321 389 -1.03 3.64 0 0 315 390 +5.64 13.23 + 0 342 480 -0.21 6.26 0 0 Control 357 474 +0.63 1.23 0 0 424 989 + 1.01 16.52 0 0 310 351 +1.14 0.70 0 0 265 263 +1.90 1.12 0 0 'HCG 100 IU/454 321 326 -1.23 6.83 0 0 g fish 323 455 -0.22 13.00 0 0 358 517 0 10.35 0 0 350 530 +1.89 12.50 + 0 (■HCG 200 IU/454 315 300 -0.33 13.55 0 0 g fish 316 345 +2.03 9.09 0 0 282 247 +4.05 11.67 0 0 292 275 +4.36 11.32 0 0 i-HCG 400 IU/454 337 549 -1.82 0.74 + + 98 80 gfish 330 323 +3.10 2.25 + 0 388 706 +4.39 8.07 + 0 330 391 +2.30 9.75 + 0 Control 301 309 -0.65 10.75 0 0 296 339 -0.88 14.14 0 0 282 346 -0.29 8.41 0 0 281 229 +6.99 18.57 + 30 'Gonadosomatic i index. ■♦Pregnant mare serum gonadotropin. ^Deoxycorticosterone. sSexually immature. ^Matured eggs in ovaries. 'Human chorionic gonadotropin. 435 FISHERY BULLETIN: VOL. 73, NO. 2 Table 4.— Effects of hormones on Pseudopleuronectes americanus given five injections over a 10-day period. All fish experienced a 9L:15D photoperiod, water temperature range 3°-5°C (x 3.2°C). Symbols ( + = Did, 0 = Did not hydrate or ovulate). Total Initial Fertili- Hormone and length body weight Weight change OS! zation Hatch dosage (mm) (g) (% initial wt.) (% final wt,) Hydrated Ovulated (%) (%) Oxytocin 378 578 -0.35 19.36 0 0 10 IU/454 306 321 -1.25 13.72 0 0 g fish 412 752 -3.32 20.50 0 0 340 488 +0.41 13,76 0 0 Oxytocin 335 481 -1,66 12,90 0 0 20 IU/454 373 654 +0.92 24.62 + 0 g fish 308 356 -1.69 16,29 0 0 391 774 -0.65 18,66 0 0 Oxytocin 305 352 +4.83 17.21 + 0 40 IU/454 284 290 -4.92 14.88 0 0 g fish 338 440 +3.30 20.34 + 0 352 478 -0.42 11.40 0 0 Control 292 327 -1.21 12.08 0 0 235 184 +0.54 22.55 + 20 316 360 0 17.92 + 0 300 399 -0.50 16.67 + 0 3D0CA 5 mg/454 320 430 +3.49 11.24 0 0 g fisfi 332 419 +0.48 16.79 + 0 360 619 +2.75 23.58 + 0 335 517 +0.77 23.03 + 0 3D0CA 10 mg/454 314 270 -3.70 10.96 0 0 g fish 374 649 +4.31 16.84 + 0 323 460 +4.35 24.79 + 20 271 225 +8.44 11.07 0 0 3D0CA 20 mg/454 333 528 +7.39 19.22 + 0 g fish 402 1,003 +9.07 32.45 + 20 337 468 +2.78 12.68 0 0 340 546 +2.01 23.07 + 0 Control 284 314 +5.37 11.78 0 0 295 279 +2.20 11.83 0 0 396 800 +2.30 13.25 0 0 347 579 +0.35 24.27 + 0 "PMSG 55 IU/454 334 693 +5.34 18,36 + 0 g fish 336 435 +2.30 18,71 + 0 425 1,105 -0.09 18.98 + 0 321 539 +7.98 25.17 + +5 »0 0 "PMSG 110 IU/454 327 467 +5.14 14.97 + 0 g fish 340 498 +1.20 14,58 + 0 324 411 -2.92 14.44 0 0 303 372 -4.30 13.34 0 0 "PMSG 220 IU/454 360 593 0 13.07 0 0 g fish 261 207 -8.70 20,11 0 +'^ 0 0 350 541 +0.74 17,61 + 0 437 1,173 +6.05 41.52 + +5 20 18 Control 295 334 -6.59 14.74 0 0 329 500 + 1.00 13.74 + 0 287 346 + 1.45 15.10 + 0 310 350 -1.71 1.16 + + 98 70 'HCG 100 IU/454 400 722 +2,08 11.13 0 0 g fish 343 462 -5.19 18.04 0 0 335 446 -6.95 12.29 0 0 402 812 +2.83 20.12 + 0 'HCG 200 IU/454 340 519 +0.39 19.67 + 0 g fish 352 500 -4.20 18.58 0 0 331 462 +6.93 23.99 + 0 290 320 +9.69 1,21 + + 98 80 'HCG 400 IU/454 285 298 -1.34 17,01 0 0 g fish 348 516 +5.04 18,91 + 0 330 513 +3.12 27,03 + +^ 15 0 296 336 -5.65 14,67 0 0 Control 362 249 + 1.20 14,48 0 0 324 548 +2.24 21,81 + 0 397 943 + 1.18 24,13 + 0 318 500 +4.82 13.10 0 0 'Gonadosomatic index. ^Matured ova in ovaries. ^Deoxycorticosterone, "Pregnant mare serum gonadotropin. spiug formed in oviduct, 'Eggs water hardened, ^Human chorionic gonadotropin. 436 SMIGIELSKI: OVULATION OF WINTER FLOUNDER noted nor did any plugs form in the test fish. A weight loss occurred in almost all of the test fish at the lower temperatures. Administering DOCA at lower temperatures resulted in GSI levels being higher in test fish at all three dosage levels. Several fish hydrated and although no ovulation occurred, some matured eggs were present in the ovaries of two test fish that received injections at the higher levels of 10 and 20 mg. No plugs were present in any of the test fish and no abnormal hydrations were ob- served. Tests with PMSG in the lower temperature range resulted in some fish at all three dosage levels hydrating but not spawning. Plugs formed in the oviducts in two of the fish that received dosage levels of 55 and 220 lU respectively. Eggs were water hardened in one and were nonfer- tilizable. Approximately 20% of the total eggs ob- tained from the second fish were fertilized. These eggs were normal in size and the majority of them developed and hatched. The larvae obtained from this induced spawning appeared to be normal in all respects and several were reared through metamorphosis. HCG administered at the level of 200 lU at lower temperatures produced ovulation in one fish, and 4 days after the last planned injection it was spawned. The eggs obtained were normal in size and appearance. Approximately 95% were fer- tilized and approximately 80% developed and hatched. The larvae obtained from this hormone- induced spawning appeared normal in all respects and many were reared through metamorphosis. One other fish that received the higher dosage level of 400 lU had a plug form in the oviduct and approximately 15% of the eggs present in the ovaries were mature. These eggs were spawned and fertilization was successful. All development halted at the blastula stage and none survived to hatch. The limited success that was obtained with the hormones oxytocin, DOCA, PMSG, and HCG ad- ministered alone is apparent. It is hoped future studies will evaluate the synergistic actions of various hormone combinations administered to winter flounder. Two test fish that received injections of carp pituitary at the level of 5.0 mg ovulated and were stripped (Table 5). The eggs obtained from these induced spawnings were in the normal size range of winter flounder eggs and approximately 90-95% were fertilized. The majority of them developed normally and approximately 85% hatched. The larvae appeared normal in all respects and many were reared through metamorphosis. At the dosage level of 0.5 mg, three fish ovulated and were spawned after receiving three to six injections. During the time of injecting these fish hydrated normally. The eggs obtained from these induced spawnings were normal in size and approximately 95% were fertilized. Their development was nor- mal and approximately 70-85% survived to hatch. Larvae obtained appeared normal in all respects and several were reared through metamorphosis. All of the uninjected controls at the conclusion of the trials displayed signs of hydrating. Four out of the six fish were sacrificed and GSI levels recorded. The remaining two fish that displayed the best signs of hydration were spared and Table 5.-Effects of carp pituitary on Pseudopleuronectes americanus. All fish experienced a 9L:15D photoperiod, water temperature range 1.5°-3.5°C (x 2:5°C). Symbols (+ = Did, 0 = Did not hydrate or ovulate). Total Initial Fertili- Hormone and No. of length body weight Weight change GSI< zation Hatch dosage injections (mm) (g) (% final wt.) {% final wt.) Hydrate Ovulate (%) {%) Carp pituitary 3 410 830 +1.20 1.11 + + 90 85 5.0 mg/454 3 415 1,069 +2.99 1.06 + + 95 85 g llsh 3 316 453 -0.04 1.60 20 0 Carp pituitary 3 325 509 +2.55 1.11 + + 95 75 0.5 mg/454 6 357 556 +2.16 1.10 + + 95 80 g fish 6 345 509 +3.54 1.32 + + 95 70 Control 0 360 494 +3.64 20.04 + 0 0 347 462 +4.11 21.21 + 0 0 339 510 +4.31 21.37 + 0 0 310 381 +6.04 20.47 + 0 0 298 321 +2.18 2.18 + +5 95 30 0 359 492 +2.64 1.11 + +' 95 90 'Gonadosomatlc index. ^Sexually immature. ^Ovulated and spawned 46 days after termination of testing trials. 437 FISHERY BULLETIN: VOL. 73, NO. 2 allowed to continue to develop without interrup- tion. After 46 days they ovulated and were stripped. The ability of carp pituitary as an aid in induc- ing spawning in winter flounder was dramatically shown in these trials. At least six additional weeks of research time was able to be realized when carp pituitary was administered in conjunction with low water temperatures (1.5°-2.5°C). It is hoped that future tests will be initiated to evaluate the effectiveness of various dosage levels and at warmer water temperatures. By controlling pho- toperiods and water temperatures and injecting carp pituitary it may be possible to extend the research time on the eggs and larvae of this species of flatfish in the laboratory for several ad- ditional months through induced spawnings. In conclusion, it would appear that water temperature may be the most critical factor in producing ovulation; and while the administering of hormones was effective in producing hydration, hydration alone is not suflScient to initiate ovula- tion. Low GSI levels below 12% in conjunction with water temperatures above 6°C resulted in the majority of test fish not hydrating regardless of the hormone administered. Hormone treatments had a range of effects depending on the degree of ovarian maturation; test fish in the later stages of development re- sponded while less mature fish developed higher GSI levels but did not ovulate and spawn. No evidence was discovered indicating that spawning induced by hormones produced abnormal larvae. ACKNOWLEDGMENTS The author wishes to express his appreciation to Hugh A. Poston of the Tunison Laboratory of Fish Nutrition and Geoffrey C. Laurence of the Na- tional Marine Fisheries Service, NOAA for their many helpful criticisms of the manuscript. LITERATURE CITED BiGELOW, H. B., AND W. C. SCHROEDER. 1953. Winter flounder Pseudopleuronectes americanus (Walbaum) 1792. In Fishes of the Gulf of Maine, p. 276- 283. U.S. Fish Wildl. Serv., Fish. Bull. 53. Dunn, R. S., and A. V. Tyler. 1969. Aspects of the anatomy of the winter flounder ovary with hypotheses on oocyte maturation time. J. Fish. Res. Board Can. 26:1943-1947. Haydock, I. 1971. Gonad maturation and hormone-induced spawning of the Gulf croaker, Bairdiella icistia. Fish. Bull., U.S. 69:157-180. Pickford, G. E., and J. W. Atz. 1957. The physiology of the pituitary gland of fishes. New York Zoological Society, N.Y., 613 p. Shehadeh, Z. H., and J. N. Ellis. 1970. Induced spawning of the striped mullet Mugil cephalus L. J. Fish. Biol. 2:355-360. Smigielski, a. S., and C. R. Arnold. 1972. Separating and incubating winter flounder eggs. Prog.-Fish Cult. 34:113. Stevens, R. E. 1966. Hormone-induced spawning of striped bass for reser- voir stocking. Prog.-Fish Cult. 28:19-28. 438 NOTES DISTRIBUTION OF MELANIN IN THE COLOR PATTERN OF DELPHINUS DELPHIS (CETACEA; DELPHINIDAE) Previous studies of cetacean pigmentation have been concerned with the description of color pat- terns and the possibilities for their evolutionary production and their adaptive significance. Mitchell (1970) identified four basic color patterns among the Delphinidae: saddled, as exemplified by some species of Stenella; spotted, as seen in Stenella plagiodon; striped, as seen in Stenella coeruleoalba; and crisscross, as seen in Delphinus del-phis. Naming the crisscross pattern as the most complex, Mitchell used it to establish a ter- minology for elements of the color patterns. One of his conclusions concerning the evolutionary development of the patterns was that the saddled pattern is most primitive, since it is closest to generalized countershading and because one may hypothetically derive the other three patterns from it by addition of certain features, emphasis of some features, and de-emphasis of others. Perrin (1972) compared the color pattern of a partially albinistic whitebelly spinner (S. lon- girostris) with that of a normally pigmented in- dividual and showed that the normal color pattern may be described in terms of two independently produced but interacting pigmentation systems or components, only one of which had developed in the partially albinistic individual. Using the two- component approach, he analyzed the color pat- terns of other delphinids, including Delphinus spp., and proposed pattern homologies among the species. He suggested that the more generalized of the two pigmentation systems involves the cape (terminology of Perrin 1970) and its accessory stripes, eye and gape marks, and dorsal fin and fluke colorations. This is overlaid by a second component system that he named the "dorsal overlay system." He proposed that partial overlapping of the two produces the four-part crisscross pattern in Delphinus. If two discrete interacting pigmentation sys- tems are involved in the color pattern of Delphinus, that fact should be evidenced in the microstructure of the skin. Previous histological study of Delphinus skin provides only a descrip- tion of the general microscopic anatomy. Stigl- bauer (1913) described the microstructure of the skin of Delphinus delphis in great detail, including the existence of large dermal papillae and epider- mal pegs with granular inclusions of pigment that he identified as melanin. His sample, however, was from the back of a single animal, and he did not have the opportunity to compare the distributions of pigment in different parts of the color pattern. Sokolov (1962) commented very briefly on the pig- mentation of two specimens of Delphinus, stating that the epidermis on the back and below the dor- sal fin was moderately pigmented, on the side of one animal was lightly pigmented, and along the side of the other animal and on the bellies of both was unpigmented. This paper reports the results of comparative microscopic examination of skin samples taken from various areas of the color pattern. Materials and Methods Skin samples, each about 5 cm square, were taken from various areas of the bodies of two animals as shown in Figure 1. The porpoise were collected at San Diego, Calif. One animal (field no. WFP 125, adult female, 176 cm, 61 kg) had been frozen for several months before dissection, and the other (WFP 221, adult male, 185 cm, 83 kg, Figure 2) was sampled about 1 h after death. One centimeter-square specimens were fixed in 10% Formalin,' and imbedded in paraffin. Sections were cut 8 ixm thick and stained with Schmorl's ferricyanide for malinin. This method stains melanins a dark blue or blue-green, while other epidermal and dermal tissue is stained light green. The prepared sections were examined under a light microscope. Pigment densities were scored at 125 diameters magnification. Results General microscopic anatomy of the skin of Delphinus is simple when compared with that of terrestrial mammals and fits the description of cetacean skin given by Simpson and Gardner 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 439 Figure 1. -Major pattern features (terminology of Mitchell 1970) and locations from which skin samples were taken: SF = spinal field (black), TP = thoracic patch (buff), FP = flank patch (gray), AF = abdominal field (white). (1972). The epidermis generally lacks hair (a few hairs are present on the snout in early develop- ment but disappear before or shortly after birth), cuticular keratin, and accessory glands. Its thick- ness varies, being greatest in the ventral region and least on the flippers, dorsal fin, and flukes. The superficial layer of the epidermis, about 10 cells thick, shows considerable flattening of cells. The epidermis consists almost entirely of polyhedral prickle cells, many of which show clear, distended infranuclear cytoplasm. This clear cytoplasm at first appears as holes in the tissue when it is viewed in section, but upon closer examination one may identify the nucleoli inside the clear areas. The epidermis interlocks with the dermis by in- terdigitation of epidermal rete pegs and dermal papillae (Figure 3), and the dermis grades into the subcutaneous blubber layer. The dermis varies in thickness and density and is composed largely of collagen fibers. The dermal papillae contain blood vessels, blood cells, and other loosely organized connective tissue. Melanin pigmentation is restricted to the epidermis and is consistently more concentrated at the edges of the epidermis than elsewhere. It is usually most concentrated around the bases of the dermal papillae (at the apices of the epidermal rete pegs) and extending in bands from the apices of the dermal papillae. The pigment has been clas- sified here into three groups for purposes of quan- tification. "Diffuse" pigment appeared simply as an area which stained darker green than the background and in which pigment grains could not be discerned even at 1,250 diameters magnifica- tion. In addition to diffuse pigment, there were granules ranging from less than 0.1 /xm to over 5 ju-m in diameter. Most were spherical or ellipsoid. Granules less than 5 /xm in greatest diameter were termed "small grains." These seemed to be ac- tually aggregations of even smaller granules. Those termed "large grains" were larger than 5 ixm in diameter and were so dense as to appear as single entities even at high power. Diffuse pig- ment was characteristically situated peripheral to the nuclei of the prickle cells. This was particularly evident in the central areas of the epidermis. Pig- ment granules were in general most concentrated at the edges of the epidermis and appeared to extend toward the surface in diffuse bands. In the more lightly colored skin specimens these bands extended only from the apices of the dermal papillae, but in more densely pigmented skin they were more numerous and tended to blend together, resulting in uniform density of pigment throughout the epidermis. The density of pigment was noted in three regions in each sample: around the bases and edges of the dermal papillae and in the bands described above (Table 1). The observations were each coded from 0 to 4 as follows: for diffuse pig- ment, 0 = none, 1 = very smaU amount, 2 = small amount, 3 = medium amount, 4 = large amount; for small grains, 0 = < 1 grain per mm^ 1 = 1-3 grains per mm^ 2 = 4-7 grains per mm^ 3 = 8-11 grains per mm^ 4 = > 12 grains per mm"; for large grains, 0 = 0 grains per 4 mm^ 1=1 grain per 4 mm^ 2 = 2-3 grains per 4 mm% 3 = 3-6 grains per 4 mm^ 4 = > 6 grains per 4 mm". Skin samples which appeared white were completely unpigmented or showed very small amounts of diffuse and/or small grains. The buff color characteristic of the thoracic patch (terminology of Mitchell 1970) was associated with either equal prominence of diffuse and granular pigment or 440 Figure 2.-Delphinus delphis: Specimen no. WFP 221. Figure 3. -Typical appearance of dermal-epidermal boundary of skin of Delphinus delphis (WFP 221 sample no. 10 in Table 1). Note concentration of pigment granules around bases of dermal papillae and in bands extending from ends of papillae. Mag- nification 1,250 diameters. greater prominence of diffuse pigment. In gray samples from the flank patch, large grains were present, but the small grains were most prominent. Samples of black skin from flippers and flukes showed the highest densities of large grains, which were more prominent than the small grains. The histological evidence (summarized in Table 2) supports the concept of interacting color pat- tern components. As indicated previously, the spinal field may be described as a result of the interaction of the pigmentation of the cape and the dorsal field. The buff colored thoracic patch (i.e., the cape) contains relatively high concentra- tions of diffuse pigment, while in the grey flank patch (i.e., the dorsal overlay) granular pigment is more prominent. The spinal field is characterized by a larger amount of diffuse pigment than is present in the flank patch, and a larger amount of granular pigment than is present in the thoracic patch. The high concentration of grains in the spinal field indicates a possible synergistic effect of the two component systems. Perrin (1972) suggested that the flukes are pig- mented only as part of the cape system. The skin of the flukes (sample no. 13) contained large amounts of all three types of pigment, indicating that pigmentation of the flukes involves both the cape system and the dorsal overlay system. The skin of the flipper (sample no. 8) also contained large amounts of all three types of pigment. The major difference between samples taken from WFP 125 (female) and those taken from WFP 221 (male) was that the epidermis was thicker (0.75-1.70 mm as opposed to 0.50-0.95 mm) in the latter animal. Thus a lower density of pig- ment was required to produce the same color. Also, the samples taken from the buff -colored area of WFP 221 showed conspicuously less diffuse pig- ment. 441 Table l.-Results of microscopic examination of pigment distribution in skin of two Deiphinus delphis. Codes explained in text. Pigment den sity (CO ins ided) Large grains Diffuse Small gra Sample no. Bases Edges Bands Bases Edges Bands Bases Edges Bands ! Color WFP 125$ 1 4 3 2 2 2 2 0 0 0 buff 2 4 3 1 4 4 2 4 3 3 black 3 3 3 1 4 3 2 4 3 2 black 4 3 3 2 2 2 1 1 0 0 buff 5 3 3 2 2 2 1 1 0 0 buff 6 3 3 1 2 2 1 0 0 0 buff 7 2 2 1 3 2 1 0 0 0 wfiite 8 3 2 1 4 3 3 4 3 2 black 9 2 2 0 1 1 0 0 0 0 white 10 4 3 0 3 3 2 2 1 0 gray 11 2 2 1 1 1 0 0 0 0 white 12 2 1 0 3 2 2 1 0 0 gray 13 4 3 2 4 3 2 4 3 2 black 14 3 3 1 4 2 2 3 2 1 black 15 3 3 2 2 2 1 0 0 0 buff 16 2 2 0 3 2 2 1 0 0 gray 17 2 2 1 1 1 0 0 0 0 white WFP 221 5 1 2 2 1 2 2 1 0 0 0 buff 2 2 2 2 4 4 3 3 3 3 black 3 3 2 2 4 4 2 3 2 2 black 4 1 1 1 2 2 2 0 0 0 buff 5 2 2 1 2 2 1 0 0 0 buff 6 2 2 1 2 2 0 0 0 0 buff 7 0 0 0 1 1 0 0 0 0 white 8 3 2 1 4 4 3 3 2 2 black 9 0 0 0 0 0 0 0 0 0 white 10 3 2 1 3 3 2 2 2 2 gray 11 1 1 0 0 0 0 0 0 0 white 12 0 0 0 2 2 1 0 0 0 gray 13 3 2 2 4 4 2 3 2 2 black 14 3 3 2 3 3 2 3 2 2 black 15 2 2 1 1 1 0 0 0 0 buff 16 2 1 1 2 2 1 1 1 1 gray 17 1 1 0 1 1 0 0 0 0 white Table 2.-Average pigment density values for four major fea- tures of the color pattern of Deiphinus delphis. Coded values in Table 1 are averaged over bases, edges, and bands, over all relevant samples, and over both animals. Number of samples included for each animal is in parentheses. Pigment dens ity Feature Diffuse Small grains Large grains Ventral field: white (4) Thoracic patch: buff (5) Flank patch: gray (2) Spinal field: black (4) 0.96 2.37 0.92 2.46 0.67 1.60 2.00 3.08 0 0.07 0.42 2.67 Discussion The color pattern of Deiphinus evidently con- sists of two overlapping and interacting pigmen- tation components that differ mainly in the den- sity and size of pigment particles. The pigment exhibits a continuum of grain size, with the very smallest particles being associated with a buff color. It seems likely that the diffuse pigment is also composed of particles too small for resolution with the light microscope. This evidence suggests that there may be a developmental progression in grain size from the unpigmented or truly white condition through buff and gray to black. This result is consistent with Mitchell's (1970) proposal of the "saddled" pattern as most primitive in recent cetaceans. Inhibition of pigment aggrega- tion could result in the high concentration of dif- fuse pigment which is typical of buff regions in Deiphinus. Fox (1953) and others have stated that there are different types of melanins, possibly characterized by different chemical compositions or structures, although none have been adequately described chemically, due to their insolubility and their general lack of adaptability to most physico- chemical methods. From the previously described evidence it may be suggested that the type of malanin which produces the buff color is of a com- position which does not allow its further polymerization, but may well favor its combina- tion with a protein, as phaeomelanin, instead of aggregation into granules. One might also hypothesize control by dispersing or concentrating 443 hormones, although this would be physiologically and developmentally more complex. In gray and black areas, aggregation of pigment seems to continue, and melanocytes migrate toward the surface from the base of the epidermis until dif- fuse pigment is largely replaced by granular pig- ment. The process is apparently stopped at some point, after which increase in thickness of the epidermis may result in a lower average density of pigment. Acknowledgments Benjamin Landing offered suggestions for his- tological treatment of samples, and Ernest Link did the actual histological preparation of sections. Photomicrography was done by Tedd Wells. Literature Cited Fox, D. L. 1953. Animal biochromes and structural colours. Camb. Univ. Press, Camb., Engl., 379 p. MrrcHELL, E. 1970. Pigmentation pattern evolution in delphinid ce- taceans: An essay in adaptive coloration. Can. J. Zool. 48:717-740. Perrin, W. F. 1970. Color pattern of the eastern Pacific spotted porpoise Stenella graffmani Lonnberg (Cetacea, Delphini- dae). Zoologica (N.Y.) 54(4): 135- 142. 1972. Color patterns of spinner porpoises (Stenella cf. S. longirostrix) of the eastern Pacific and Hawaii, with comments on delphinid pigmentation. Fish. Bull, U.S. 70:983-1003. Simpson, J. G., and M. B. Gardner. 1972. Comparative microscopic anatomy of selected marine mammals. In S. H. Ridgeway (editor). Mammals of the sea, biology and medicine, p. 298-418. Chas. C. Thomas, Springfield, 111. SOKOLOV, V. E. 1962. Structure of the tegument in cetaceans. Part II. [In Russ.] Nauchn. Dokl. Vysshei Shkoly, Biol. Nauki 3:45-55. Stiglbauer, R. 1913. Der histologische Bau der Delphinhant mit besonderer Beriicksichtigung der Pigmentierung. Sitzungs- ber. Kais. Akad. Wissensch. (Wien) Math.-Naturw. Kl. 122, H.II.Abt. 111:17-27. Sharon Gwinn Division of Natural Sciences University of California Santa Cruz, CA 9506U William F. Perrin Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NO A A La Jolla, CA 92037 OCCURRENCE OF TWO CONGRIDAE LEPTOCEPHALI IN AN ESTUARY During the night of 23-24 August 1971, I caught two congrid leptocephali in Montsweag Bay, part of the Sheepscot River-Back River estuary, Wis- casset, on the southern Maine coast. These larvae were identified as conger eel. Conger oceanicus (D. G. Smith, pers. commun., 15 April 1974). The es- tuary was described by Stickney (1959). Recksiek and McCleave (1973) provide additional informa- tion about the estuary and Montsweag Bay. The leptocephali were collected near their sampling station G3 (lat. 43°56'N, long. 69°42'W). Briefly, Montsweag Bay is a shallow (1 m at mean low water) and wide (2.4 km) basin, but it has a narrow channel (9 m deep at mean low water) through most of its length. Narrow openings at its northern and southern ends allow tidal flow. Mean tidal difference is approximately 3 m. Seasonally, water temperature extremes in Montsweag Bay range from 0.0° to 18.5°C. Salinity ranges from 7 to 30"/ 00. Gear used was essentially that described by Graham and Venno (1968). One larva (98 mm TL) was captured during the flooding tide 1 m below the surface; the other (91 mm TL) during the ebbing tide within 3 m of the bottom. Water depth at this location was approximately 9 m at mean low water. During this period, the average salinity was 26.0"/ oo and the average water temperature was 17.7°C. Conger eel adults and leptocephali have been reported from the Gulf of Maine (Bigelow and Schroeder 1953), but apparently most leptocephali are found in the western North Atlantic (Schmidt 1931). Conger eel leptocephali, however, have never been reported from such low-salinity water. Bigelow and Schroeder (1953) illustrated one 84 mm long from Chesapeake Bay, but they do not give the salinity at the collection site. They also state that conger eel leptocephali grow to 150-160 mm. Smith (pers. commun., 15 April 1974) com- mented that my specimens were beginning to metamorphose since the gut of each had shortened noticeably. Conger eel leptocephali apparently are able to tolerate this low-salinity water at least during metamorphosis. If conger eel leptocephali typically grow to the size reported by Bigelow and Schroeder (1953), then they must shrink tremendously in length during metamorphosis. My specimens probably shrank during storage, but probably not enough to account for that much size difference. 444 Acknowledgments David G. Smith, Marine Research, Inc., Woods Hole, Mass., identified the specimens. Equipment and facilities were furnished by the University of Maine, Orono, Maine. P. C. Jensen operated the RV Cypris. Research funds for this and associated research were provided by the Maine Yankee Atomic Pow'.jr Company. Literature Cited BiGELOW, H., AND W. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Graham, J. J., and P. M. W. Venno. 1968. Sampling larval herring from tidewaters with buoyed and anchored nets. J. Fish. Res. Board Can. 25:1169-1179. Recksiek, C, and J. D. McCleave. 1973. Distribution of pelagic fishes in the Sheepscot River— Back River estuary, Wiscasset, Maine. Trans. Am. Fish. Soc. 102:541-551. Schmidt, J. 1931. Eels and conger eels of the North Atlantic. Nature (Lond.) 128:602-604. Stickney, a. p. 1959. Ecology of the Sheepscot River estuary. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 309, 21 p. William J. Hauser University of California-Riverside Imperial Valley Field Station lOOJ, East Holton Road El Centra, CA 922i3 CHLORINATED HYDROCARBONS IN SEA-SURFACE FILMS AND SUBSURFACE WATERS AT NEARSHORE STATIONS AND IN THE NORTH CENTRAL PACIFIC GYRE Chlorinated hydrocarbons, DDT residues, and polychlorinated biphenyls (PCB's) entering the oceans via atmospheric transport, runoff, and outfalls (National Academy of Sciences 1971) may be concentrated in the lipid constituents (Garrett 1967; Duce et al. 1972) found in surface films. The chlorinated hydrocarbons can then enter marine food chains, most probably by association with particulate detritus and subsequent ingestion by filter- feeding organisms. Further concentration in higher trophic levels is well documented (see, for example, Harvey et al. 1971) and will not be dis- cussed here. There have been only two reported studies on the concentration of DDT residues and PCB's in surface films. Seba and Corcoran (1969) found high concentrations of p,p'DDT, p,p'DDE, o,pDDT, aldrin, and dieldrin in the surface microlayer collected at locations in Biscayne Bay, Fla., and 10 miles offshore in the Florida Strait. Duce et al. (1972) found that PCB's (but no DDT residues) were concentrated in surface films from Narragansett Bay, R.I. Seawater collected at 1-2 m in the California Current was analyzed for DDT residues (Cox 1971) and these results will be taken as subsurface water concentrations in California coastal waters. This note reports on the content of p,p'DDT, p,p'DDE, and PCB's in surface films collected at coastal stations off southern California and Mexico; and in surface films, subsurface waters, and particulate matter from the North Central Pacific Gyre (Table 1). Methodology All surface films were collected with a Monel' or stainless steel screen (Garrett 1965) into 2.5-liter glass bottles. The coastal samples (SIO 1-2; M 1-4) were poisoned with mercuric chloride. The Cato samples were filtered on shipboard through sol- vent-extracted and ignited GF/C glass-fiber filters. The filters were frozen in glass vials at -20°C, and the filtrate preserved with 75 ml of hexane. Subsurface samples were collected in 2.5- liter glass bottles 10-15 cm below the surface and treated as above. In all operations the surface films were collected from a skiff at least 0.5 mile upwind from the ship. All glassware, screens, filters, etc., were scrupulously freed of organic matter by ig- nition at 550°C, rinsing with double distilled sol- vents, or both. In the laboratory, the filtrates were acidified to pH 2 with distilled 6N HCl and extracted with three 60-ml portions of hexane. The hexane ex- tracts were dried by passage through anhydrous Na2S04, and then concentrated to 10-15 ml in a Kuderna-Danish evaporator. This extract was further reduced to 50 jul in vacuo, put onto an alumina microcolumn (McClure 1972), and eluted with 3.5 ml of hexane. The eluate was dried in vacuo and taken up in 50 /aI of isooctane. The filters were extracted in a soxhlet overnight with 20 ml of 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 445 Table l.-Sample locations, collection dates, and sea conditions. Collection Sample 1 number and description Sample location date Sea conditions iSIO-1 Surtace film Off SIO Pier 7-July-71 Calm. Moderate film SI 0-2 Surface film Off SIO Pier 7-July-71 Calm. Moderate film 2M-1 Surface film Lat. 21 30'N, long. 108'30'W Moutfi of Gulf of Calif. 23-Oct.-71 Calm. No visible film M-2 Surface film Lat. 19 10'N, long. 104^30'W 4 mi. off Manzanillo 25-Oct.-71 Calm. Well-developed film M-3 Surface film Lat. 17'37.2'N, long. 101°33'W Zihuatenejo Harbor 28-Oct.-71 Calm. Slight film M-4 Surface film Lat. 16°49'N, long. 99°53'W 1 mile off Acapuico 30-Oct.-71 Calm. Slight film Cato 1-4 Surface film filtrate and filter Subsurface water filtrate and filter Lat. 31°51.9'N, long. 127°25.0'W California Current 10-June-72 Calm. No visible film Cato I-A1 Surface film filtrate and filter Subsurface water filtrate and filter Lat. 30=59. 1'N, long. 155°24.0'W North Central Pacific Gyre 27-June-72 Choppy. No visible film Cato I-A2 Surface film filtrate and filter Subsurface water filtrate and filter Lat. 31°1.8'N, long. 155°25.8'W North Central Pacific Gyre 29-June-72 Choppy. No visible film 'Scripps Institution of Oceanography. ^Mexico. a 50:50 acetone-hexane solution, the extracts dried over anhydrous Na2S04, and dried in vacuo. The residues were taken up in 50 /xl of hexane and treated identically as the above water samples. Aliquots of the isooctane solutions were injected into a Hewlett-Packard model 5750B gas chroma- tograph equipped with a Ni^s electron capture detector and a 6-foot glass column packed with either 1.5% OV-17 and 1.95% QF-1 or 5% SP2401 on 100/200 mesh Supelcoport. Column temperature was 195°C. An alkaline (KOH, NaOH) precolumn for saponification of DDT residues was used to give confirmatory identification (Miller and Wells 1969). Retention times and peak heights were compared with standard mixtures of p,p'DDT, p.p'DDE, and Aroclor 1254. Reagent and appara- tus blanks were essentially zero, and reextraction of the water sample with hexane showed no addi- tional traces of DDT residues or PCB's. The minimum detectable amounts of p,p'DDT, p.p'DDE, and PCB's (as Aroclor 1254) are 5x10'- g, 3 X 10-'^ g, and 1 x lO"'" g, respectively. Results and Discussion In the North Central Pacific Gyre (Table 2), the concentration of DDT residues in surface films and subsurface waters was less than 0.03 ng/liter for all samples, while the PCB content was two orders of magnitude higher, and PCB's were al- ways present. A higher concentration of PCB's in the surface films as compared to the subsurface Table 2.-Concentrations of PCB's, p.p'DDT, and p,p'DDE in surface films and subsurface waters. All values in nanograms per liter. station Sample PCB's p,p' DDT P,P' DDE SIO-1 Surface film 11 15 0.4 SIO-2 Surface film 50 12 0.2 M-1 Surface film 90 i<0.02 i<0.01 M-2 Surface film 12 6.8 0.4 M-3 Surface film 24 8.3 1.8 M-4 Surface film 13 2.1 0.1 Cato 1-4 Surface film 4.9 0.3 0.1 filtrate and filter 3.1 0.1 <0.01 Subsurface water 1.6 <0.02 <0.01 filtrate and filter 0.9 <0.02 <0.01 Cato I-A1 Surface film 3.3 <0.02 <0.01 filtrate and filter 2.9 <0.02 <0.01 Subsurface water 3.6, <0.02 <0.01 filtrate and filter 1.1 <0.02 <0.01 Cato I-A2 Surface film 3.5 <0.02 <0.01 filtrate and filter 1.7 <0.02 <0.01 Subsurface water (contaminated) filtrate and filter 0.7 <0.02 <0.01 'These lower limits are greater than the absolute minimum amounts detectable due to sample splitting during analysis. waters was significant but not striking, largely because well-defined surface films were not present at the time of sampling. In the California and Mexican coastal waters, both PCB and DDT residue concentrations in the surface films were slightly higher (as would be expected) than in the 1-2 m subsurface waters of the nearshore California Current (Cox 1971). The available data in the literature pertaining to the chlorinated hydrocarbon content of surface films and subsurface waters is collected in Table 3. This table does not include a considerable amount of unpublished data taken in conjunction with outfall studies and pollution problems in general. 446 Table 3.-Comparison of this work with literature values for the concentration of PCB's and DDT residues in surface films and subsurface waters. All values in nanograms per liter, and are the sum of filtrate and filter from Table 2 where applicable. Location PCB's DDT residues Reference Subsurface Surface films waters Surface films Subsurface waters Seba and Corcoran (1969) Biscayne Bay Florida Strait NDi ND ND ND 185-13,710 70 <1 <1 Duce et al. (1972) Narragansett Bay 450-4,200 < 50-1 50 undetected undetected Cox (1971) Nearshore Cali- fornia Current ND ND ND 2.3-5.6 (1-2 m depths) This work California coastal Mexican coastal Offshore California Current North Central Pacific Gyre 11-50 12-90 8.0 5.2-6.2 ND ND 2.5 4.7 12.2-15.4 <0.03-11.2 0.4 <0.02 ND ND 0.1 <0.01 'Not determined. The primary aim of this work has been to es- tablish open ocean concentrations of PCB's and DDT residues in surface films and subsurface waters in oligotrophic regions of the ocean such as the North Central Pacific Gyre. The PCB content of open ocean waters are significantly lower rela- tive to inshore waters, and represent the first such numbers for an open ocean environment in the Northeast Pacific. Acknowledgments We thank H. Bezdek for collection of the surface film samples from the Mexican coastal waters (M 1-4). This work was supported by AEC Contract AT(11-1)GEN 10, P.A. 20. Miller, G. A., and C. E. Wells. 1969. Alkaline pre-column for use in gas chromatographic pesticide residue analysis. J. Assoc. Off. Anal. Chem. 52:548-553. National Academy of Sciences. 1971. Chlorinated hydrocarbons in the marine environ- ment. A report prepared by the panel on monitoring persistent pesticides in the marine environment of the Committee on Oceanography. Natl. Acad. Sci., Wash., D.C., 42 p. Seba, D. B., and E. F. Corcoran. 1969. Surface slicks as concentrators of pesticides in the marine environment. Pestic. Monit. J. 3:190-193. P. M. Williams K. J. Robertson Institute of Marine Resources University of California, San Diego La Jolla, CA 92037 Literature Cited Cox, J. L. 1971. DDT residues in seawater and particulate matter in the California Current system. Fish. Bull., U.S. 69:443-450. Duce, R. A., J. G. Quinn, C. E. Olney, S. R. Piotrowicz, B. J. Ray, and T. L. Wade. 1972. Enrichment of heavy metals and organic compounds in the surface microlayer of Narragansett Bay, Rhode Island. Science (Wash., D.C.) 176:161-163. Garrett, W. D. 1965. Collection of slick-forming materials from the sea surface. Limnol. Oceanogr. 10:602-605. 1967. The organic chemical composition of the ocean sur- face. Deep-Sea Res. 14:221-227. Harvey, G. R., V. T. Bowen, R. H. Backus, and G. D. Grice. 1972. Chlorinated hydrocarbons in open ocean Atlantic or- ganisms. Proc. Nobel Symp. 20:177-186. McClure, V. E. 1972. Precisely deactivated adsorbents applied to the separation of chlorinated hydrocarbons. J. Chromatogr. 70:168-170. HATCHING SURVIVAL OF HYBRIDS OF ONCORHYNCHUS MASOU WITH SALMO GAIRDNERI AND WITH NORTH AMERICAN SPECIES OF ONCORHYNCHUS The cherry salmon, Oncorhynchus masou, which is native only to Asian watersheds discharging into the northwestern Pacific Ocean, is a recent in- troduction to North America. While cherry salmon have been crossed with some Asian salmonids, in- formation on their ability to hybridize with North American salmonids has not been reported in the literature. The primary purpose of these experiments was to determine hatching survival of some interspecific crosses involving cherry salmon, leading to a sound basis for predicting their effects on indigenous salmonid species and their potential value in salmon management. 447 Materials and Methods From 10 October to 12 November 1972, hybridization experiments were carried out between 1-yr-old male cherry salmon parr from anadromous stock and female rainbow trout, Sal- mo gairdneri; pink salmon, 0. gorbuscha; chum salmon, 0. keta; coho salmon, 0. kisutch; sockeye salmon, 0. nerka; and chinook salmon, 0. tshawytscha. Our cherry salmon were reared at the Washington State Department of Fisheries' Minter Creek Hatchery from eyed eggs sent in 1971 by the Hokkaido Salmon Hatchery, Sapporo, Japan. Incubation facilities were located at the Northwest Fisheries Center, National Marine Fisheries Service, Seattle, Wash. The standard dry fertilization technique was used in conjunction with delayed fertilization techniques described by Poon and Johnson (1970). All fertilization took place within 3 h of collection, with the exception of pink salmon eggs (14 h). There were no apparent effects from delayed fertilization. Numbers of eggs incubated ranged from 1,700 to 8,400; sur- vival was based on the total eggs in each lot. Discussion Oshima (1957) reported that cherry salmon have successfully hybridized with redspot salmon, 0. rhodurus, for many years. Other than hybrids of cherry salmon with redspot or Asian pink salmon, hybrids of cherry salmon with other salmon and trout are rare or unreported (Schwartz 1972; Dangel et al. 1973). Results of our own experiments, as shown in Table 1, show that crosses of cherry salmon with chum, chinook, and pink salmon and with rainbow trout were highly successful, each yielding higher hatching percen- tages than their respective controls. The reason for this phenomenon is not presently understood but it does indicate an area for further research. Only crosses of coho and sockeye salmon with cherry salmon showed poorer survival than their controls (Table 1). It is interesting to note that though there was no hatch of the cherry x sockeye cross, virtually all of the eggs were fertilized and developed to notochord formation. Each of the successful hybrid crosses yielded surviving fry to a 1 g or larger size accounting for over 85% of the hatch, except for the rainbow and coho crosses where survival to this size was less than 10%. Literature Cited Dangel, J. R., P. T. Macy, and F. C. Withler. 1973. Annotated bibliography of interspecific hybridization of fishes of the subfamily Salmoninae. U.S. Dep. Commer., NOAA Tech. Memo. NMFS NWFC-1, 48 p. Oshima, M. 1957. Studies on the dimorphic salmons, Oncorhynchus masou (Brevoort) and Oncorhynchus rhodurus Jordan & McGregor, found in Japan and adjacent territories. Nire Shobo (Nire Book Company), Sapporo, Japan, 81 p. (Translated from Jap., 1972, Fish. Res. Board Can., Transl. Ser. 2104.) Poon, D. C, and A. K. Johnson. 1970. The effect of delayed fertilization on transported salmon eggs. Prog. Fish-Cult. 32:81-84. Sano, S., and H. Eguchi. 1936. Interspecific hybridization among salmonid fishes. Hokkaido Sakemasu Fukajo (Hokkaido Salmon- Trout Hatchery), 13 p. (Translated from Jap., 1968, Fish. Res. Board Can., Transl. Ser. 1164.) Schwartz, F. J. 1972. World literature to fish hybrids with an analysis by family, species, and hybrid. Publ. Gulf Coast Res. Lab. Mus. 3, 328 p. Table l.-Hatching success of eggs of hybrid crosses and controls. Species' Number of degree-days incubated Salmo gairdneri X O. masou Salmo gairdneri (control) Oncorhynchus gorbuscha X O. masou Oncorhynchus gorbuscha (control) Oncorhynchus keta X O. masou Oncorhynchus keta (control) Oncorhynchus kisutch X O. masou Oncorhynchus kisutch (control) Oncorhynchus nerka X O. masou Oncorhynchus nerka (control) Oncorhynchus tshawytscha X O. masou Oncorhynchus tshawytscha (control) 'Female listed first and male last. 396 321 512 593 436 504 300 333 660 642 426 486 Percentage hatched Previously reported results Reference 39.5 34.7 71.6 62.5 94.1 90.9 26.5 90.9 0.0 96.0 97.4 72.9 85% hatch Suzuki and Fukuda 1971a, b 37-46% hatch Smirnov 1969 77% hatch 0-96% hatch 0-69% hatch None 0% hatch 0-3.3% hatch None Sano and Eguchi 1936 Smirnov 1969 Terao and Hayashinaka 1961 Suzuki and Fukuda 1971a, b Terao and Hayashinaka 1961 448 Smirnov, a. I. 1969. Hybrids of Pacific salmon of the genus Oncorhynchus, characteristics of development and prospects of utiliza- tion. In B. I. Cherfas (editor), Genetika, selektsiya i gibridizatsiya ryb, p. 139-159. Izdatel'stvo "Nauka," Moscow. (Translated by Isr. Prog. Sci. Transl. 1972, p. 131-147 in B. I. Cherfas [editor], Genetics, selection, and hybridization of fish, avail. U.S. Dep. Commer., Natl. Tech. Inf. Serv., Springfield, Va. 22151 as TT 71-50112.) Suzuki, R., and Y. Fukuda. 1971a. Survival potential of Fj hybrids among salmonid fishes. Bull. Freshwater Fish. Res. Lab. (Tokyo) 21:69-83. 1971b. Growth and survival of Fi hybrids among salmonid fishes. Bull. Freshwater Fish. Res. Lab. (Tokyo) 21:117-138. Terao, T., and H. Hayashinaka. 1961. On the artificial hybridization among the salmonid fishes. L Sci. Rep. Hokkaido Fish Hatchery 16:51-62. (Translated from Jap., 1968, Fish. Res. Board Can., Transl. Ser. 1047.) James L. Mighell Northwest Fisheries Center National Marine Fisheries Service, NOAA 2725 Monflake Boulevard East Seattle, WA 98112 Alaska Regional Office National Marine Fisheries Service, NOAA P.O. Box 1668, Juneau, AK 99801 James R. Dangel TRAP CONTRIBUTIONS TO LOSSES IN THE AMERICAN LOBSTER FISHERY Studies to evaluate the impact of unbuoyed traps on American lobster, Homarus aviericanus, sur- vival were conducted in Maine waters from July 1971 to June 1973. Materials On 22 July 1971, 98 tagged lobsters of various legal and illegal sizes and both sexes were placed in 35 unbaited conventional square traps, with 30- mm lath spacing, without buoy lines, on the sea bottom near Jonesport, Maine, in depths ranging from about 10 to 20 m (Table 1). On 29 July 1971, four tagged lobsters were added to one trap from which the previous occupants had escaped by 24 July. The 84-m- study site, considered by fishermen not to be a good lobster habitat, having a muddy bottom and no rocks which could be utilized as cover, was purposely selected because its use would not interfere with commercial fishing and traps would be protected from storm damage. Methods Traps were checked on nine occasions before 15 October 1971, by scuba diving. When traps were checked by diving, it was possible to count the lobsters and observe evidence of cannibalism, but tagged lobsters could not readily be distinguished from others that entered the traps. In order to differentiate tagged from untagged lobsters, all traps were brought to the surface for more thorough examination. This practice was com- menced on 15 October 1971 and continued throughout the remaining period of the study. Traps were retrieved 16 times between 15 Oc- tober 1971 and 26 June 1973, making a total of 25 checks during the investigation. The length of time between observations of the 2-yr period ranged from 1 to 161 days, with a median interval of 13 days and a mean of 28 days. Observations were curtailed during the low temperature months because of the inactivity of lobsters in relatively shallow water. Results During the first summer-fall season, 43% of the tagged lobsters cannot be accounted for; 25% remained captive; 20% escaped and were recap- tured; and 12% were cannibalized. During the second summer-fall season, 126% recruitment oc- curred; 22% cannot be accounted for; 18% of both tagged and recruited lobsters were cannibalized; 55% remained in the traps; and 5% of tagged lob- sters escaped and were recaptured. A minimum 67 "wild" lobsters were recruited by the traps, of which 24 still remained captive when the study was terminated. Two tagged lobsters that departed their original traps entered other experimental traps which they in turn left before entering two of the commercial traps surrounding the study site. A tagged male lobster missing from trap no. 6 was caught in a conmiercial trap 0.4 km from the study area on 28 April 1973, after having remained in trap no. 6 for 22 mo and having moulted once in October 1971 from sublegal to legal size. Four traps failed to recruit any lobsters; 9 recruited one each; 13, two each; 6, three each; 2, four each; and 1, six. Only five traps recruited more lobsters than were initially placed in them, six recruited a like number. 449 CO eg CO c- c •-5 CO IN O •-5 £ o 3 cr 15 c o '■2 c > c o u 13 O) X! c 3 LC 00 > o CO »-. CM in eg CM in en o in CM CO CJ) CO ea — . CO CJ) eg eg CO eg eg eg o c Q. o^— ocMOO'-^cgt-o^ooo-'-oco^O'-ooegT-i-oO'-cgo^^egi- r^ eg 0'-'-'-o>-»-i-oooT-ooO'-oejT-ooooO'-T-oO'---0'-»-0'- co 0'-'-»-0'-r-coooo»-oooi-T-cgi-i-ooT-OT-egoO'-'-0'-^o»- Ot-i-»-0'-*-COOOOt-OOO^i-'-t-t-00'-Ot-CMOO'-j-0»-t-Ot- O'— '— t-O'-'-coooo^-ooOT-T-r-T-^— ooi— o-r-egoO'-^Oi— T-o*- co O'-'-'-O'-'-c0OOOt-OOOt-t-CSJi-^OO''-O-'-C0OO'-'-OCM>-O"- CO CM O'-'-'-O'-egcoooOT-ooO'-T-CM^i-Oi-'-o-r-cooO'-T-oegi-O'- co O'-'-^'-CMcgcoooO'-O'-OT-T-eMT-i-oo'-OT-cooo'-'-oegi-O'- oi-'-'f-oegegcoooo^OOoeg-i-egT-T-oo^-oegcooO'-'-oeg^OT- 0'-i-^oejegnooo-i-ooocgT-egcot-'-0'-oejcoO''-'-'-OT-i-0'- o^CNj^-ocNjejcO'-OT-'-oooegT-co co^T-o^-oejcoO'-cOr-O'-oO'- oo Oi-CMT-ococMcoooo^-ooocM'-ejCM'-coO'-O'-egOr-^-i-O'-oOi- co t- ---- -- ----.„^ '-T-T-T-ococMegeuT-egt-ooo'-i-cOT-,- CMi-t-T-egcMOr-T-x-O'-ooi- co '-'-^-i-oegegegi-'- "-r-T-ooT-ot-cvjo egoooi-oo^-t-i- oi-^-o^- 00t-t-0»-t-'-t-0t-0t-000-»-0t-0t- i-t-Ot-OOt-t-t-O'-OOt- c3) O'-i-CMOi-eg^T-'-'-'-egoooego^o^-i-T- O'-oOi-egr-^-i-ooeg en ^oegegeg-i-CMi-i-^'-7-egegoT-egOT-OT-i-CMO-'-oocMCMegT-T-oo'«J- o ■-ege>iT-TtT-T-ejco'^eg-r co <3) (/) '-egcOTfincor^cooJO'-egcO'^mcDKcooo^egco^incoN-cooOi-cMco'^ir) "Jn '-i-T-T-T-i-i-i-i-i-egCMCMegegcMCMegcgejcocococococo con o o )- CO ■* ■ 3 E o CT> CO CJ> •a a> « °' ™ □) o 2 h- N 'c c CO O 0) T3 CO o 3 ■»- o o •o -■ c o c m <0 CO n o ■o o (U c c oo o (0 ra CJ o 54 450 Discussion In the three most recent years, gross fishing effort in the Maine lobster fishery has, perhaps temporarily, stabilized at approximately 1.25 million traps, a 67% increase over the 0.75 million level of the preceding 12-yr period. Annual loss of traps has varied markedly since the mid-1940's. In major late summer-fall storm years, fishermen have reported losses of up to 100% in many fishing areas; at other times less than 10% in other areas. An average annual loss of 20 to 25% has been estimated from interviews v^ith fisher- men and counts made by departmental scientific and enforcement personnel of traps stranded in- tertidally by storms. This estimate would indicate that about 200,000 traps have been lost annually during the past decade from storms, accidents, or vandalism, with each trap containing an average of 3.1 lobsters (Dow 1961). Storm-lost traps are the most consistently damaged and when they are washed ashore, they usually contain dead lobsters. Cannibalism occurs principally from July to early November, coincident with the greatest concentration of traps, fishermen, and catch. Within this period, 70 to 75% of the annual catch is made, which for the last 30 yr has averaged 9,000 metric tons, consisting of approximately 18 million lobsters. Previous studies (Dow 1961, 1966) also demonstrated a 2V2-fold increase in the total number of lobsters entering traps of the summer- fall fishery in comparison with the winter and spring fisheries. During the 2-yr period of this investigation, in which only 12% of the traps used were sufficiently damaged by lobster chelipeds to permit escape, the annual Maine lobster catch was 7,670 metric tons, consisting of 14.2 million lobsters caught in 1.25 million traps. Between July and November when the peak of cannibalism occurs, 77% of the annual catch was made and consisted of 10.9 million lob- sters. Conclusions 1. Unbaited, unbuoyed traps continue to catch lobsters for an indefinite time, with most of the catch being made between June and September. 2. Cannibalism occurs during the summer and fall coincident with moult. 3. Approximately one-third or more of all lob- sters in or entering unbuoyed traps will be lost to the fishery from cannibalism or retention. Literature Cited Dow, R. L. 1961. Some factors influencing Maine lobster landings. Commer. Fish. Rev. 23(9): 1-11. 1966. Limitations on measurement of effort-yield in the Maine lobster fishery. Fishing News Int. 5(8):32-36. William W. Sheldon Robert L. Dow Maine Department of Marine Resources State House Augusta, ME OJ^SSO. 451 INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the following instructions. These are not absolute requirements, of course, but desiderata. CONTENT OF MANUSCRIPT The title page should give only the title of the paper, the author's name, his affiliation, and mailing address, including Zip code. The abstract should not exceed one double- spaced page. In the text. Fishery Bulletin style, for the most part, follows that of the Style Manual for Biological Journals. Fish names follow the style of the American Fisheries Society Special Pub- lication No. 6, A List of Common and Scientific Names of Fishes from the United States and Canada, Third Edition, 1970. The Merriam- Webster Third New International Dictionai^ is used as the authority for correct spelling and word division. Text footnotes should be typed separately from the text. Figures and tables, with their legends and headings, should be self-explanatory, not requir- ing reference to the text. Their placement should be indicated in the right-hand margin of the manuscript. Preferably figures should be reduced by pho- tography to 5% inches (for single-column fig- ures, allowing for 50% reduction in printing), or to 12 inches (for double-column figures). The maximum height, for either width, is 14 inches. Photographs should be printed on glossy paper. Do not send original drawings to the Scien- tific Editor; if they, rather than the photo- graphic reductions, are needed by the printer, the Scientific Publications Staff will request them. Each table should start on a separate page. Consistency in headings and format is desirable. Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be numbered sequentially in arabic numerals. To avoid confusion with powers, they should be placed to the left of numerals. Acknowledgments, if included, are placed at the end of the text. Literature is cited in the text as: Lynn and Reid (1968) or (Lynn and Reid 1968). All papers referred to in the text should be listed alphabetically by the senior author's surname under the heading "Literature Cited." Only the author's surname and initials are required in the literature cited. The accuracy of the lit- erature cited is the responsibility of the author. Abbreviations of names of periodicals and serials should conform to Biological Abstracts List of Se7-ials with Title Abbreviations. (Chemical Ab- stracts also uses this system, which was devel- oped by the American Standards Association.) Common abbreviations and symbols, such as mm, m, g, ml, mg, °C (for Celsius), %, "/oo and so forth, should be used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbreviations. We prefer that measurements be given in metric units; other equivalent units may be given in parentheses. FORM OF THE MANUSCRIPT The original of the manuscript should be typed, double-spaced, on white bond paper. Please triple space above headings. We would rather receive good duplicated copies of manuscripts than carbon copies. The sequence of the material should be: TITLE PAGE ABSTRACT TEXT LITERATURE CITED APPENDIX TEXT FOOTNOTES TABLES (Each table should be numbered with an arabic numeral and heading provided) LIST OF FIGURES (entire figure legends) FIGURES (Each figure should be numbered with an arabic numeral ; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. Bruce B. Collette, Scientific Editor Fishery Bulletin National Marine Fisheries Service Systematics Laboratory National Museum of Natural History Washington, DC 20560 Fifty separates will be supplied to an author free of charge and 100 supplied to his organiza- tion. No covers will be supplied. Contents — continued CAIN, THOMAS D. Reproduction and recruitment of the brackish water clam Rangia cuneata in the James River, Virginia 412 SMIGIELSKI, ALPHONSE S. Hormonal-induced ovulation of the winter flounder, Pseudopkuronectes americanus 431 Notes GWINN, SHARON, and WILLIAM F. PERRIN. Distribution of melanin in the color pattern of Delphinus delphis (Cetacea; Delphinidae) 439 HAUSER, WILLIAM J. Occurrence of two Congridae leptocephali in an estuary . . 444 WILLIAMS, P. M., and K. J. ROBERTSON. Chlorinated hydrocarbons in sea-surface films and subsurface waters at nearshore stations and in the North Central Pacific Gyre 445 MIGHELL, JAMES L., and JAMES R. DANGEL. Hatching survival of hybrids of Onc(yrhynchus masou with Salmo gairdneri and with North American species of Oncorhynchus • 447 SHELDON, WILLIAM W., and ROBERT L. DOW. Trap contributions to losses in the American lobster fishery 449 ^OV-VJT/O/v^ ^< °^^o^ ^^ATES O^ ^' Fishery Bulletin <^ National Oceanic and Atmospheric Administration • National Marine Fisheries Service I Marine ^• L \ i_jj i\ A K 'i' "N Vol. 73, No. 3 If f July 1975 LASKER, REUBEN. Field criteria for survival of anchovy larvae: The relation between inshore chlorophyll maximum layers and successful first feeding 453 PEREZ FARFANTE, ISABEL. Spermatophores and thelyca of the American white shrimps, genus Penaeus, subgenus Litopenaeua 463 CHITTENDEN, MARK E., JR. Dynamics of American shad, Alosa sapidissima, runs in the Delaware River 487 SILLIMAN, RALPH P. Selective and unselective exploitation of experimental populations of Tilapia mossambica 495 CLARK, ROBERT C, JR., and JOHN S. FINLEY. Uptake and loss of petroleum hydrocarbons by the mussel, Mytilus edulis, in laboratory experiments 508 COLLETTE, BRUCE B., and LABBISH N. CHAO. Systematics and morphology of the bonitos {Sarda) and their relatives (Scombridae, Sardini) 516 INGHAM, MERTON C. Velocity and transport of the Antilles Current northeast of the Bahama Islands 626 HARTMANN, A. RUCKER, and THOMAS A. CLARKE. The distribution of myc- tophid fishes across the central equatorial Pacific 633 PETERSON, WILLIAM T., and CHARLES B. MILLER. Year-to-year variations in the planktology of the Oregon upwelling zone 642 RYBOCK, J. T., H. F. HORTON, and J. L. FESSLER. Use of otoliths to separate juvenile steelhead trout from juvenile rainbow trout 654 HOUDE, EDWARD D., and L. J. SWANSON, JR. Description of eggs and larvae of yellowfin menhaden, Brevoortia smithi 660 Notes KIEFER, DALE A., and REUBEN LASKER. Two blooms of Gijmnodinium splen- dens, an unarmored dinoflagellate /< 675 SANDIFER, PAUL A., PAUL B. ZIELINSKI, and WALTER E. CASTRO. Enhanced survival of larval grass shrimp in dilute solutions of the synthetic polymer, polyethylene oxide 678 YONG, MARIAN Y. Y., and ROBERT A. SKILLMAN. A computer program for analysis of polymodal frequency distributions (ENORMSEP), FORTRAN IV /->. . 681 RICHARDSON, SALLY L., and DOUGLAS A. DeHART. Records of larval, trans- forming, and adult specimens of the quillfish, Ptilichthys goodei, from waters off Oregon 681 (Continued on back cover) Seattle, Washington U.S. DEPARTMENTOFCOMMERCE Rogers C. B. Morton, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Administrator NATIONALMARINE FISHERIESSERVICE Robert W. Schoning, Director Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 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. Bruce B. Collette Scientific Editor, Fishery Bulletin National Marine Fisheries Service Systematics Laboratory National Museum of Natural History Washington, DC 20560 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. WiUiam H. Bayliff Inter-American Tropical Tuna Commission Dr. Roger F. Cressey, Jr. U.S. National Museum Mr. John E. Fitch California Department of Fish and Game Dr. WilUam W. Fox, Jr. National Marine Fisheries Service Dr. Marvin D. Grosslein National Marine Fisheries Service Dr. Edward D. Houde University of Miami Dr. Merton C. Ingham National Marine Fisheries Service Dr. Reuben Lasker National Marine Fisheries Service Dr. Jay C. Quast National Marine Fisheries Service Dr. Paul J. Struhsaker National Marine Fisheries Service Dr. Austin Williams National Marine Fisheries Service Kiyoshi G. Fukano, Managing Editor The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget through May 31, 1977. Fishery Bulletin CONTENTS Vol. 73, No. 3 July 1975 LASKER, REUBEN. Field criteria for survival of anchovy larvae: The relation betw^een inshore chlorophyll maximum layers and successful first feeding 453 PEREZ FARE ANTE, ISABEL. Spermatophores and thelyca of the American white shrimps, genus Penaeus, subgenus Litopenaeus 463 CHITTENDEN, MARK E., JR. Dynamics of American shad, Alosa sapidissima, runs in the Delaw^are River 487 SILLIMAN, RALPH P. Selective and unselective exploitation of experimental populations of Tilapia mossambica 495 CLARK, ROBERT C, JR., and JOHN S. FINLEY. Uptake and loss of petroleum hydrocarbons by the mussel, Mytilus edulis, in laboratory experiments 508 COLLETTE, BRUCE B., and LABBISH N. CHAO. Systematics and morphology of the bonitos (Sarda) and their relatives (Scombridae, Sardini) 516 INGHAM, MERTON C. Velocity and transport of the Antilles Current northeast of the Bahama Islands 626 HARTMANN, A. RUCKER, and THOMAS A. CLARKE. The distribution of myc- tophid fishes across the central equatorial Pacific 633 PETERSON, WILLIAM T., and CHARLES B. MILLER. Year-to-year variations in the planktology of the Oregon upvv^elling zone 642 RYBOCK, J. T., H. F. HORTON, and J. L. FESSLER. Use of otoliths to separate juvenile steelhead trout from juvenile rainbow trout 654 HOUDE, EDWARD D., and L. J. SWANSON, JR. Description of eggs and larvae of yellowfin menhaden, Brevoortia smithi 660 Notes KIEFER, DALE A., and REUBEN LASKER. Two blooms of Gymnodinium splen- dens, an unarmored dinoflagellate 675 SANDIFER, PAUL A., PAUL B. ZIELINSKI, and WALTER E. CASTRO. Enhanced survival of larval grass shrimp in dilute solutions of the synthetic polymer, polyethylene oxide " '° YONG, MARIAN Y. Y., and ROBERT A. SKILLMAN. A computer program for analysis of polymodal frequency distributions (ENORMSEP), FORTRAN IV ... . 681 RICHARDSON, SALLY L., and DOUGLAS A. DeHART. Records of larval, trans- forming, and adult specimens of the quillfish, Ptilichthys goodei, from waters off Oregon ^^^ (Contimied on next page) Seattle, Washington For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402-Subscription price; $11.80 per year ($2.95 additional for foreign mailing). Cost per single issue - $2.95. Contents —continued SILLIMAN, RALPH P. Effect of crowding on stock and catch in Tilapia mossambica 685 WENNER, CHARLES A. The occurrence of elvers of Synaphobranchus affinis on the continental slope off North Carolina 687 PEARCY, WILLIAM G., DANIEL A. PANSHIN, and DONALD F. KEENE. Catches of albacore at different times of the day 691 Vol. 73, No. 2 was published on 30 April 1975. The National Marine Fisheries Service (NMFS) does not approve, rec- ommend or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. FIELD CRITERIA FOR SURVIVAL OF ANCHOVY LARVAE: THE RELATION BETWEEN INSHORE CHLOROPHYLL MAXIMUM LAYERS AND SUCCESSFUL FIRST FEEDING' Reuben Lasker^ ABSTRACT Northern anchovy larvae, Engraulis mordax, produced by laboratory-spawned fish, have been used to detect concentrations of larval fish food in situ along the California coast. First-feeding larval anchovies, whose development was controlled by temperature manipulation aboard ship, were placed in samples of Los Angeles Bight water taken from the surface and from chlorophyll maximum layers. Feeding by larvae in water from the surface was minimal in all experiments but extensive feeding occurred in water from the chlorophyll maximum layers when these contained phytoplankters having minimum diameters of approximately 40 ju,m and which occurred in densities of 20 to 40 particles/ml. In March and April 1974, the chlorophyll maximum layer along the California coast from Malibu to San Onofre (a distance of about 100 km) consisted chiefly of a bloom of the naked dinoflagellate Gym- nodinium splendens, a food organism known to support growth in anchovy larvae. Copepod nauplii and nonliving particles were never in high enough concentration or of the proper size to be eaten by the larvae. A storm which caused extensive mixing of the top 20 m of water obliterated the chlorophyll maximum layer and effectively destroyed this feeding ground of the larval anchovy. Probably the major problem confounding fishery scientists interested in rational management of fisheries is an inability to predict recruitment failure (Gulland 1973) despite the vast amount of laboratory and field work on food chain analysis leading to fish production which has occupied many workers in marine studies for the past two decades (Steele 1970). Gulland (1973) asks the most pressing question, "Can a study of stock and recruitment aid management decisions?" and in the same article answers "No." This pessimistic reply is given because, as he says, "there is no method which is likely to be generally successful, [because] the most promising depends on lengthy and costly collection of data, probably extending over a long period." The work reported in this paper suggests an approach which has not been previously used in fishery research as far as I am aware, and which, I believe, makes the answer Gulland has given somewhat less pessimistic than when he made it. However, it is generally agreed among fishery biologists that large spawning populations of fish 'MARMAP (Marine Resources Monitoring, Assessment, and Prediction) Contribution No. 17. Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, La Jolla, CA 92037. Manuscript accepted November 1974. FISHERY BULLETIN: VOL. 73, NO. 3, 1975. do not ensure subsequent large year classes, and conversely, small spawning populations oc- casionally give rise to exceptionally large classes (Hjort 1926). Hjort (1914) postulated that these variations in year class strength are probably due to differential mortality of the larvae. He believed, for example, that the larvae of the Nor- wegian herring, Clupea harengus, suffered huge mortalities resulting in small year classes when there was a lack of food for the first-feeding lar- vae. The attractiveness of this hypothesis has generated a number of laboratory studies (see, for example, Lasker et al. 1970) which have shown that the density of larval food must be higher than that usually found at sea in order to obtain even moderate larval growth and survival. For an in- depth discussion of the larval "critical period" as affected by food see May (1974), and for a review of laboratory attempts to rear fish larvae, refer to May (1971). The conclusion that the mean density of larval food organisms in the ocean is generally too low to support reasonable survival of fish lar- vae through metamorphosis, is also substantiated by data from field surveys (Beers and Stewart 1967, 1969). Thus, despite extensive efforts in quantitative marine food chain analysis, it is yet to be demonstrated whether, where, and to what extent there are rich feeding areas in the sea for larval fishes. 453 KISHKKY BULLhJTlN: VUL. 73, NU. 3 As a new approach to this problem it is the pur- pose of this study to show how laboratory-spawned fish larvae can be used to detect larval feeding grounds at sea and to point out some of the ways this technique might be used to provide the link between marine food chain research and stock and recruitment predictions in fisheries; the latter by determining what the environmental conditions at sea must be with respect to larval fish food to result in a good or bad survival year for particular species of fish larvae. Because it is essential to the general methodology of using larval fishes as assay or- ganisms for the fitness of seawater as larval fish feeding grounds, in the following I describe some background information on maturation and spawning of anchovies in the laboratory; the methods used for feeding larval anchovies; and the laboratory-determined criteria for feeding already known for the larva of this species. A description of the field work is then given, concluding with a discussion of the criteria which can be used to judge the fitness of the larval anchovy's environment. THE NORTHERN ANCHOVY In the California Current, the major pelagic fish population at present is that of the clupeoid Engraulis mordax, the northern anchovy. This species is found from British Columbia, Canada to Cape San Lucas, Baja California and extending west about 600 km. Although the anchovy has a protracted spawning season from December to August, about three-quarters of its spawning oc- curs in the winter and spring months of January, February, March, and April. The factors affecting larval mortality of clupeoids have been inves- tigated in a number of laboratories throughout the world (Holliday and Blaxter 1963) including my own. Recently, it has been possible to intensify laboratory research at the Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service because of the continuous availability of anchovy larvae. This has been made possible by inducing sexual maturation of adults in the laboratory resulting in daily spawning and fertilization of anchovy eggs throughout the year. Details of this maturation and spawning tech- nique are given by Leong (1971) and Lasker (in press). The availability of first-feeding larvae hatched from laboratory-spawned eggs has made possible the development of a technique whereby specific areas of the ocean could be examined for their potential as larval feeding grounds, and criteria established to characterize parts of the ocean as good or bad areas for larval survival. LABORATORY-DETERMINED CRITERL\ FOR SUCCESSFUL LARVAL ANCHOVY FEEDING A background of information on larval anchovy feeding is available from a number of studies and was used to guide this investigation. 1. Particle size at first feeding appears to be crit- ical. First-feeding anchovy larvae (standard length 3.5 mm) have small mouths and require a food particle about 50 ]u.m in diameter, although particles larger than 100 /j,m may be taken (Berner 1959). Smaller particles may not be visible to the larvae. Berner (1959) reported that anchovy larvae smaller than 4 mm long taken in plankton tows had eaten particles ranging in length from 24 to 186 /xm. However, over 70% of the food in their intestines was between 60 and SOjxm long. 2. The number of particles per unit volume in the anchovy larva's environment must be above a minimum concentration. O'Connell and Raymond (1970), using natural plankton as food, showed that the survival of first-feeding anchovy larvae was dependent, in their experiments, on the number of micronauplii per unit volume available to the lar- vae. Successful first feeding, as pointed out by Hunter (1972), also depends on a sufficiently high density of food particles to compensate for the low capture efficiency (about 10%) exhibited by anchovy larvae when they begin to feed. 3. The kind of food organism determines survival and growth. Lasker et al. (1970) fed a variety of phytoplankters and zooplankters to first-feeding anchovy larvae. Only one phytoplankter of those tested, Gymnodinium splendens, supported growth and gave relatively good results in survival experiments when compared with larvae fed nat- ural plankton. The rotifer Brachionus plicatilis, although not found in the anchovy's normal habi- tat, also could be used as a laboratory food for older anchovy larvae and a small proportion of first- feeding larvae (Theilacker and McMaster 1971; Hunter 1972). 4. The greater the concentration of food particles, the more frequent are the feeding strikes made by anchovy larvae; consequently the greater the suc- cess in capturing food. Although examination of field-caught anchovy larvae reveal very few with 454 LASKER: FIELD CRITERIA FOR SURVIVAL any food in their intestine (Arthur 1956; Berner 1959), this seems to be a result of rough handling due to capture in a plankton net and subsequent preservation with Formalin^ which causes almost all anchovy larvae to defecate (Kjelson et al. 1975). In the laboratory a high proportion of anchovy larvae which had never seen a food particle will strike at and ingest them provided there is a high enough concentration of the right size food par- ticles. Hunter and Thomas (1974) have shown that the rate of larval anchovy feeding increases with increasing food density. Despite the success of O'Connell and Raymond (1970) and Kramer and Zweifel (1970) who were able to rear anchovy larvae using micronauplii concentrated in wild plankton samples, the quan- tities needed by the larvae seemed inordinately higher than the concentration of nauplii reported by Beers and Stewart (1967) for the euphotic zone of the California Current. It was possible, of course, that concentrations of nauplii exist in dense aggregations. Beers and Stewart (1970a) reported that nauplii concentrate in or immedi- ately above chlorophyll maximum layers off the California coast. However, no concentrations have been identified which are high enough to support anchovy larvae (e.g., 1/ml). On the other hand, Lasker et al. (1970) showed that anchovy larvae would feed and grow on a diet of the dinoflagellate G. splendens. The fact that blooms of a variety of phytoplankters are known to occur in the Califor- nia Current, particularly in the spring, suggested that phytoplankton cells were more likely to provide the particle size and cell density essential to survival and growth of first-feeding anchovy larvae. For these reasons, inshore chlorophyll maximum layers were selected as possible fruitful areas to investigate for larval feeding. METHODS To determine areas in the sea where high con- centrations of living particles might be present, a pump was used to bring water on board from known depths. The hose from the pump was lowered below the surface by means of a metered winch. The water was pumped through a Turner fluorometer in the ship's laboratory to measure chlorophyll a and other fluorescing substances. Chlorophyll a was extracted from water samples "Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. taken at each station from different depths and the fluorescence profile adjusted to reflect only chlorophyll a (Kiefer and Lasker 1975). A school of approximately 700 sexually mature northern anchovies, maintained in the aquarium of the Southwest Fisheries Center, La Jolla Laboratory, produce about 1,000 fertilized eggs per day. The number of eggs can be increased by injections of gonadotrophins to stimulate massive spawning in individual fish (Leong 1971). With temperature control of development (Lasker 1964), larvae in first-feeding condition can be made available whenever desired. Thus, preliminary to a cruise, development of embryos and larvae can be accelerated or retarded by temperature manipulation to ensure that on each day of the cruise there will be at least several hundred larvae ready to feed. Prior to each of the cruises described here, fer- tilized eggs at the same stage of development were sorted into liter jars containing seawater previously filtered through a 5-/Am pore size Cuno Aquapure filter, and the jars were immersed in water baths at suitable temperatures. For trans- port to the ship insulated chests were used, and on board a temperature-controlled room was con- tinually adjusted between 13° and 18°C to insure that feeding larvae would be available on specific days. Recent studies have shown that newly feed- ing larvae have only about 2V2 days after the yolk sac is absorbed to get suflRcient food or they will die (Lasker et al. 1970). Experiments to determine if samples of seawater contained suitable food for anchovy lar- vae were done in cylindrical, 8-liter jars wrapped with dull black cardboard and set on black plastic. Above the jars a bank of four "daylight" fluores- cent lights were suspended which illuminated the surface of the jars at approximately 2,152 lux. When a sample of seawater was brought into the ship's laboratory it was permitted to warm slowly to room temperature. Larvae were added to 5 liters of seawater by pouring them from the incubation jars. Dilution of the 5 liters by the larval incuba- tion water was corrected by concentrating the contents of an additional liter of seawater to a few milliliters with fine mesh netting and by adding the concentrate to the whole. A gentle air stream was directed onto the sur- face of the water in each experimental vessel to ensure mixing of the water. Control experiments on shipboard with cultured organisms indicated that this had little effect on the larvae. 455 FISHERY BULLETIN: VOL. 73, NO. 3 Two cultured food organisms, G. splendens and B. plicatilis, were used aboard ship to determine if any particular batch of test larvae vi^as in good condition and would feed. Details on the culture of these organisms are given by Thomas et al. (1973) for G. splendens, and Theilacker and McMaster (1971) for B. plicatilis. Usually a control 5-liter seawater sample was seeded with one or the other cultured organism and a feeding experiment with larvae run concurrently with samples of natural seawater. Larvae were permitted to feed for 8 h then siphoned out and sucked down onto a membrane filter (pore size O.Sfim) using a vacuum pump. This rapid removal of larvae from the experimental containers and their fast immobilization on the filters prevented defecation. After air drying, microscopic examination of the transparent larvae permitted counts to be made of larvae which had been feeding and those which had not. The relative proportion of feeding to nonfeeding larvae was subsequently correlated with sizes of the food particles, chlorophyll content, species composition, and the number of particles available to the larvae. A 1-liter subsample of each seawater sample was preserved with Formalin (final concentration 5%) (Beers and Stewart 1970b). Later the particles in the preserved seawater were settled out and the species enumerated. The method of Utermdhl (as described by Lund et al. 1958) was used to concen- trate organisms from known volumes of preserved seawater, usually 100 ml. At least 100 cells larger than 20 /i.m in diameter were counted from each settled water sample. On a cruise of the NOAA RV David Starr Jor- dan, 18-21 March 1974, a 16-channel electronic particle counter with a 280-/u,m pore, Coulter Counter Model Ta, was used to determine the size distribution and numbers of sized particles in seawater samples used for feeding experiments. Only particles 20ju,m and larger were counted. Very good agreement was obtained between the elec- tronic counter particle counts and those obtained with the inverted microscope. A comparison is shown in Figure 1. The speed of counting and siz- ing particles makes the multichannel particle siz- ing instrument desirable for rapid field assess- ment of larval fish food organisms. Because the electronic counter was unavailable for two sub- sequent cruises. 8-12 and 22-23 April 1974, the results for these are given from microscope counts only. O iij < S £ llJ CM Hi (£ _l UJ O I- d I- 2 cr a. 100 NO. OF PARTICLES PER ML FROM INVERTED MICROSCOPE COUNTS ^ 20|jm DIAMETER Figure L-Comparison between instantaneous particle counts per milliliter taken with Model Ta Coulter Counter and mean particle count per milliliter observed through an inverted microscope. A 100-ml sample was concentrated and at least 100 particles of the dominant organism were counted in any visual field. Only particles larger than 20 fxm were counted. To ascertain if anchovy larvae were present at particular depths, plankton tows were taken with a 0.333-mm mesh, 0.5-m mouth diameter net. Because an opening and closing net was unavail- able for this work, an open net was dropped rapidly to the desired depth, towed for 10 min at a main- tained wire angle of 45°, then pulled up rapidly. The total proportion of time the net spent in water other than in the desired stratum never exceeded 5% of the total towing time. All larvae captured were measured, sorted to species, and counted. Larval counts were corrected to a comparative volume of 1,000 m^ (Kramer et al. 1972). A flow- meter in the mouth of the net provided a record of the volume of water filtered. The shipboard experiments were ordinariy done at temperatures of between 15° and 19°C, whereas concentrations of larval fish food organisms and anchovy larvae were occasionally found between 14° and 15°C. It was desirable, therefore, to de- termine the feeding response of first-feeding anchovy larvae at different environmental temperatures. Experiments to determine this were identical to those done at sea except that the concentration of the cultured organism G. splen- dens was varied from 5 to 200 cells/ml, and the water temperature was controlled within ±0.2°C. 456 LASKER: FIELD CRITERIA FOR SURVIVAL AREA OF STUDY This investigation was conducted in the Los Angeles Bight along the southern California coastline from Malibu to San Onofre, between 18 and 21 March 1974. All the stations occupied were over the 20-fathom line except for the Laguna Beach station which was over the 270-f athom con- tour. Table 1 gives the coordinate positions and Figure 2 shows the relative location of the sta- tions. The San Onofre station was reoccupied on 8 April 1974 to determine the persistence of the chlorophyll maximum layer which earlier had con- tained relatively large numbers of G. splendens. The station was occupied again on 10 April, after a violent wind storm and later on 22-23 April after a period of no storms. Figure 2.-Stations in the Los Angeles Bight. RESULTS AND DISCUSSION Shipboard Experiments with First-Feeding Anchovy Larvae Table 1 provides a summary of the results of feeding experiments with first-feeding anchovy larvae in water from the surface and from the chlorophyll maximum layer or from a depth of about 15 m if no clear chlorophyll maximum was observed. The dominant phytoplankter in the chlorophyll maximum layers was G. splendens. For details about G. splendens blooms from Baja California, Mexico, and the Los Angeles Bight see Kiefer and Lasker (1975); as reported earlier, Lasker et al. (1970) had demonstrated that anchovy larvae will grow when fed on G. splen- dens. Also given in Table 1 are the results of a feeding experiment at the San Onofre station on 8 April 1974, 18 days after a chlorophyll maximum layer containing G. splendens as the dominant phy- toplankter was found. The chlorophyll maximum layer was still present and heavily populated with G. splendens. A violent wind storm on 9 April obliterated the chlorophyll maximum layer and no G. splendens were seen at this station when it was reoccupied on 10 and 11 April. A comparison of chlorophyll a profiles taken before and after the 9 April storm is shown in Figure 3. The results of control feeding experiments are given in Table 2. SAN ONOFRE SS^'IS.S' N., II7''35.3' W. April 8, 1974 1500 -—April 10, 1974 1710 April II, 1974 0830 UJ Q 0 0.5 1.0 1.5 20 2.5 CHLOROPHYLL-a (pg/ liter) Figure 3.-Chlorophyll maximum layers before (8 April 1974) and after (10 and 11 April 1974) a violent wind storm near San Onofre, Calif. Some anchovy larvae capture a few particles in any concentration of 20- to 100-/Am particles over an 8-h period, but experience in the laboratory has shown that feeding on less than 1 particle/h will not sustain a first-feeding larva which becomes weak and dies. Thus, in Tables 1 and 2, two feeding categories are indicated: larvae observed with food organisms packed into the intestine, and 457 FISHERY BULLETIN: VOL. 73, NO. 3 Table 1.— Summary of results of 8-h on-board feeding experiments with first-feeding anchovy larvae. Gymnodinium splendens appeared in the chlorophyll maximum layers (chl. max.) from Malibu to San Onofre, a distance of approximately 130 km. The subsurface bloom of G. splendens at San Onofre persisted until 8 April 1974 (see no. 7 below). A storm on 9 April obliterated the maximum and evidently dispersed the G. splendens by wind mixing (see Figure 3). Date and time Total number particles/ml Feedi ng by anchovy larvae Percent of larvae with: A. Surf. temp. Number of B. Chl. max., temp. Position 23-37yU,m >37-299um diameter ^g Chl. a 1/4 to 1-8 parti- larvae per No. and depth (lat. N-long. W) diameter per liter full gut cles in gut experiment 1 20 March 1974, 1250 A. 15"C 34=00.8'-118''40.6' 14.2 '4.1 (< 1) n.o 2 15 69 B. 14.2 C, 12 m (Malibu) 37.3 38.0 (12) 1.8 23 46 93 2 20 March 1974, 1745 A. 15.2'C 33°52.5'-1 18=27.0' 23.1 6.1 (6) 0.4 0 11 94 B. 14.5 C, 13.5 m (Manhattan Beach) 29.8 19.7 (12) 1.3 0 18 49 3 21 March 1974, 0900 A. 15,8 = C 33'36.5'-118°04.3' 217.9 33,2 (<1) 0.3 0 12 42 B. 14.2"C, 16.5 m (Seal Beach) 53.8 352.0 (380) 42.0 22 24 104 4 21 March 1974, 1340 A. - 33=30.8'-1 17=50.3' 34.0 9.0 (0) 0.6 0 10 20 B. -, 15 m (Laguna Beach) 29.0 5.7 (0) 0.7 0 11 46 (no chl. max.) 5 21 March 1974, 1715 A. 15.5=C 33°26.3'-1 17=42.8' 18.7 5.9 (<1) 0.5 35 16 19 B. -, 15 m (Dana Point) 55.2 23.2 (5) 1.3 0 69 26 6 21 March 1974, 1900 A. 15.2 C 33°19.9'-117°35.3' 9.3 4.0 (<1) 0.2 0 8 49 B. -, 19.5 m (San Onofre) 42.4 47.7 (34) 2.3 9 25 32 7 8 April 1974, 1500 A. 17.1°C 33°19.9'-117°35.3' 5.7 9.1 (<1) 0.2 0 13 23 B. 14.8°C, 16 m (San Onofre) 14.0 81.3 (64) 2.3 9 40 58 8 10 April 1974, 1710 A. 14"C 33°19.4'-117°34.6' 8.4 14.1 (0) — 0 12 33 B. 13,5 = C, 14 m (San Onofre) 10.5 23.2 (0) 0.8 0 15 20 (no chl. max.) 9 11 April 1974, 0915 A. 13.0'C, 5 m 33=19.5'-117'34.6' 6.4 10.5 (0) 0.6 0 4 50 (no chl. max.) (San Onofre) i( ) = number of G. splendens per milliliter. ^Particles smaller than 20 [x,m may have contributed to the elevated chlorophyll a at this station. 3This 5% figure represents only one larva which filled its intestine V4 full. Table 2.-Controls for the experiments reported in Table 1. Surface water was seeded with Gymnodinium splendens or Brachionus plicatilis. In each instance the results showed that the larvae on shipboard were competent to feed. Feeding time was 8h. Species seeded Number particles/ml Feedi ng by anchovy larvae Date Percent of larvae with: ., Number of Va to 1-8 parti- larvae per full gut cles in gut experiment 20 March 1974 8 April 1974 10 April 1974 B. plicatilis G. splendens G. splendens 40 100 100 29 27 32 45 72 13 30 22 46 those with eight or fewer particles in the intestine after an 8-h feeding period. The largest proportion of larvae did not feed at all, a result common to laboratory experiments as well. The feeding in- tensity at Malibu, Seal Beach, and San Onofre is typical of first-feeding anchovy larvae in labora- tory experiments seeded with a like number of suitable size particles, e.g., G. splendens. The data presented in Table 1 show that the criteria for larval anchovy feeding determined by laboratory experiments are the same when freshly obtained seawater is tested as a source of larval anchovy food. Large numbers of particles smaller than 37 ^um in diameter did not stimulate feeding in anchovy larvae. This was particularly apparent at Seal Beach on 21 March; surface water having 218 particles/ml smaller than 37 /xm in diameter but with low chlorophyll a did not stimulate anchovy larvae to feed. Conversely, the bloom of G. splendens in the chlorophyll maximum layer produced heavy feeding larvae tested in shipboard experiments. Furthermore, even with particles 458 LASKER: FIELD CRITERIA FOR SURVIVAL having the right diameter for feeding, a minimum concentration of perhaps between 25 and 50 cells/ml was needed. The Effect of Temperature on Feeding The chlorophyll maximum layers along the coast were characterized by temperatures between 14° and 15°C, which is lower than the optimum temperature for feeding and growth in anchovy larvae (16°C). Because shipboard experiments were done at temperatures higher than those found in the chlorophyll maximum layers it was desirable to determine if the minimum particle count at which first-feeding larvae were stimu- lated to feed was in any way reduced by lower temperatures. Figure 4 illustrates the results of experiments which show that at 14°C, the food particle count must be higher than 20 cells /ml before significant feeding can occur over an 8-h period. At the higher temperature tested, 18°- 19°C, food particle counts of between 5 and 20 particles/ml may stimulate feeding. However, le-is-c _,-' 6 splandens CELLS PER MILLILITER Figure 4.-The effect on larval anchovy feeding of different concentrations of food at two temperatures. Each experiment began with 100 larval anchovies which were in first-feeding con- dition and was terminated after 8 h. See text for details. during the shipboard experiments, particle counts of 5-20 cells/ml did not stimulate larval feeding even at the higher temperatures (15°-19°C) of the ship's laboratory (Table 1). This discrepancy may be due to the different kinds of food particles available to the larvae, as well as to other factors related to a larva's inability to capture certain particles as opposed to others. For example, when Chaetoceros sp. was present in any of the samples, anchovy larvae did not feed on this phytoplankter, owing probably to the spiny nature of this chain- forming diatom, despite the considerable lengths (longer than 37ju,m) of the chains. In the Seal Beach surface sample taken on 21 March 1974, Chaetoceros sp. and other chain-forming diatoms made up over 30% of the longer than 37-jam category. This result was confirmed at a station off Imperial Beach, Calif, (lat. 32°34.0' N; long. 117°10.5' W) on 11 April where a dense bloom dominated by Thalas- siosira sp. (37 chains/ml) was found. Chlorophyll a at the surface was 4.8)u,g/liter and slightly higher, 5.1)u.g/liter, at a depth of 7 m. Feeding by anchovy larvae on this organism was virtually nil. Thus, the composition of the stock of phytoplankton appears to be an important factor in the initial feeding of anchovy larvae. The observations described above indicate that chlorophyll measurements alone, and the indica- tion of a strong chlorophyll maximum layer are not by themselves sufficient criteria for establishing the existence of good conditions for the feeding of anchovy larvae. A cruise of the RV David Starr Jordan was made back to the San Onofre station on 22 and 23 April 1974. A sharp chlorophyll maximum layer was discovered there once again and was found to extend seaward for at least 14 km (Figure 5), yet shipboard larval anchovy feed- ing was negative. Subsequent microscopic examination of the water from these layers in- dicated that cryptomonads of about 10 ixm in diameter dominated the samples in concentrations of 3,400-7,200 cells/ml, a size too small to be fed upon by anchovy larvae. 50 - CHLOROPHYLL-O ( |l« / lit*' 1 Figure 5.-Chlorophyll maximum layers off San Onofre, Calif., 22-23 April 1974. 459 FISHERY BULLETIN: VOL. 73, NO. 3 Vertical Distribution of Anchovy Larvae At San Onofre on 8 April the chlorophyll maximum layer was due, in part, to a high density of G. splendens. Larvae on board ship fed freely in the water from this layer. Judging from the larval feeding and the size and density of the food par- ticles, this chlorophyll maximum layer should have been an ideal place for first-feeding larvae to be found. To test this, plankton tows were made at three depths: within the chlorophyll maximum layer at 16 m, just above the layer at 10 m, and at the surface (see Figure 3). The standard length of anchovy larvae from these tows was measured, and the degree of eye pigmentation noted (full pigmentation indicating a visually competent larva). The results, given in Table 3, show that there was a distinct difference in vertical dis- tribution between the number of first-feeding larvae that could see as opposed to yolk-sac larvae which lack eye pigmentation and could not see. The surface water contained 2,100 anchovy larve 71,000 m\ Larvae without eye pigmentation outnumbered sighted larvae two to one. At the 10-m stratum above the chlorophyll maximum layer there were 40,000 anchovy larvae/ 1,000 m^ with larvae capable of seeing outnumbered by eight to one. In the chlorophyll maximum layer 4,900 anchovy larvae/ 1,000 m^ were present but larvae that could see were about as numerous as those which could not. Although it may be coin- cidental, the possibility that larvae were actively seeking out areas with food cannot be dismissed. Table 3. -Distribution of anciiovy eggs and larvae depths off San Onofre, Calif., 8 April 1974. at three Number of eggs or larvae per 1,000 mJ Stratum Larvae with Larvae with pigmented eyes unpigmented eyes Eggs Surface 10 m 16 m 756 4,610 2,004 1,344 35,804 2,941 67,200 151,965 60,034 Criteria for Successful First-Feeding by Anchovy Larvae It is evident from the data presented in this report that the following environmental criteria must be met before first-feeding anchovy larvae can feed successfully in the ocean. Phytoplankton aggregations with over 20 cells/ml must be available at the same time or within 2V2 days after the larvae are ready to feed. Individual phy- toplankton cells must be about 40]u,m in diameter. Successful feeding is dependent on food density so that the higher the concentration of cells, the bet- ter the feeding. Monotypic algal blooms are re- sponsible for some chlorophyll maximum layers off the California coast and first-feeding anchovy larvae were found to be living within them. Only some phytoplankters stimulate feeding and sup- port growth of anchovy larvae; for example G. splendens is known to support growth while anchovy larvae would not feed on Chaetoceros sp. or Thalassiosira sp., spiny and/or chain-forming diatoms. Finally, it could not be demonstrated that micronauplii or other microzooplankton contribute significantly to larval anchovy survival during the first week of larval life. Beers and Stewart (1967) reported that in December 1965, the inshore sta- tion off San Diego (their station I) contained a maximum of only 30 organisms /I iter in the 35 to 103-ju,m size class. Of these organisms, copepod nauplii and post-nauplii together numbered 7- 9/liter, two orders of magnitude lower than that required by anchovy larvae to survive, i.e., 1,000/liter (O'Connell and Raymond 1970). However, it is reasonable to assume that under special circumstances suitable concentrations of micronauplii might serve as a food source for first-feeding anchovy larvae. Nonliving particles larger than 37 jxm were not seen and may be in- significant in the nutrition of first-feeding anchovy larvae because of their low concentration in anchovy spawning areas. It is important to emphasize the transient na- ture of good feeding conditions. There was a large number of larvae present at depth at the San Onofre station on 8 April 1974, capable of taking advantage of the subsurface bloom of Gym- nodinium. Furthermore, spawning had been ex- tensive in the entire water column as indicated by the large number of anchovy eggs caught during the same tows (Table 3). Earlier I indicated that a wind storm obliterated the chlorophyll maximum layer at San Onofre on 9 April 1974, and that 8-h shipboard experiments showed the larvae were unable to capture enough food on 10 and 11 April to fill or partially fill their intestines. If my con- tention is correct, then a large proportion of the larvae which were present as eggs or yolk-sac lar- vae on 8 April were doomed to die from lack of food after the storm on 9 April because of the dilution and dispersion of suitable larval food organisms. Although this investigation was confined to first-feeding anchovy larvae, the technique 460 LASKER: FIELD CRITERIA FOR SURVIVAL described here, using laboratory-reared larvae for field tests of possible larval feeding areas, proba- bly can be extended to older larvae and other species as vi^ell. An on-board electronic counter can give a rapid first evaluation of particle size and cell numbers. Microscopic examination of subsamples can be used to verify the shipboard counts and give additional information on species composition of phytoplankters. Chlorophyll profiles can be analyzed routinely on most oceanographic vessels. With information on the biology of the larvae be- ing investigated, it may be possible to determine routinely the quality and extent of larval feeding grounds and v^^ith comprehensive temporal infor- mation on the food of the larvae, the degree of larval mortality due to inadequate food may be predictable. ACKNOWLEDGMENTS Thanks go to Eileen Setzler, Anne N. Dodson, Dale Kiefer, and the able crew of the NOAA RV David Starr Jordan for their invaluable assis- tance during this study, and to Richard Eppley for stimulating discussions which led to the examina- tion of chlorophyll maximum layers as possible feeding grounds for larval fish. Roderick K. Leong provided the fish larvae on demand; without his help this study would not have been possible. LITERATURE CITED Arthur, D. K. 1956. The particulate food and the food resources of the larvae of three pelagic fishes, especially the Pacific sar- dine, Sardinops caerulea (Girard). Ph.D. Thesis, Univ. California, Scripps Inst. Oceanogr., La Jolla, 231 p. Beers, J. R., and G. L. Stewart. 1967. Micro-zooplankton in the euphotic zone at five loca- tions across the California Current. J. Fish. Res. Board Can. 24:2053-2068. 1%9. Micro-zooplankton and its abundance relative to the larger zooplankton and other seston components. Mar. Biol. (Berl.) 4:182-189. 1970a. Numerical abundance and estimated biomass of microzooplankton. In J. D. H. Strickland (editor), The ecology of the plankton off La Jolla, California, in the period April through September, 1967, p. 67-87. Bull. Scripps Inst. Oceanogr. 17. 1970b. The preservation of acantharians in fixed plankton samples. Limnol. Oceanogr. 15:825-827. Berner, L., Jr. 1959. The food of the larvae of the northern anchovy Engraulis mordax. Bull. Inter-Am. Trop. Tuna Comm. 4:1-22. Blaxter, J. H. S., AND F. G. T. Holliday. 1963. The behaviour and physiology of herring and other clupeids. Adv. Mar. Biol. 1:261-393. Gulland, J. A. 1973. Can a study of stock and recruitment aid management decisions? Rapp. P.-V. Reun. Cons. Perm. Int. Explor Mer 164:368-372. Hjort, J. 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 20:1-228. 1926. Fluctuations in the year classes of important food fishes. J. Cons. 1:5-38. Hunter, J. R. 1972. Swimming and feeding behavior of larval anchovy Engraulis mordax. Fish. Bull., U.S. 70:821-838. Hunter, J. R., and G. L. Thomas. 1974. Effect of prey distribution and density on the searching and feeding behaviour of larval anchovy Engraulis mordax. In J. H. S. Blaxter (editor), the early life history of fish, p. 559-574. Springer- Verlag, Berl. Kiefer, D. A., and R. Lasker. 1975. Two blooms of Gymnodinium splendens, an unar- mored dinoflagellate. Fish. Bull., U.S. 73:675-678. Kjelson, M. a., D. S. Peters, G. W. Thayer, and G. N. Johnson. 1975. The general feeding ecology of postlarval fishes in the Newport River estuary. Fish. Bull., U.S. 73:137-144. Kramer, D., and J. R. Zweifel. 1970. Growth of anchovy larvae {Engraulis mordax Girard) in the laboratory as influenced by temperature. Calif. Coop. Oceanic Fish. Invest. Rep. 14:84-87. Kramer, D., M. J. Kaun, E. G. Stevens, J. R. Thrailkill, and J. R. Zweifel. 1972. Collecting and processing data on fish eggs and larvae in the California Current region. U.S. Dep. Commer., NOAA Tech. Rep. NMFS CIRC-370, 38 p. Lasker, R. 1964. An experimental study of the effect of temperature on the incubation time, development and growth of Pacific sardine embryos and larvae. Copeia 1964:399-405. In press. Induced maturation and spawning of marine fish at the Southwest Fisheries Center, La Jolla, Califor- nia. Proc. 5th Annu. Workshop, World Mariculture Soc. Lasker, R., H. M. Feder, G. H. Theilacker, and R. C. May. 1970. Feeding, growth and survival of Engraulis mordax larvae reared in the laboratory. Mar. Biol. (Berl.) 5:345-353. Leong, R. 1971. Induced spawning of the northern anchovy, Engraulis mordax Girard. Fish. Bull., U.S. 69:357-360. Lund, J. W. G., C. Kipling, and E. D. LeCren. 1958. The inverted microscope method of estimating algal numbers and the statistical basis of estimations by counting. Hydrobiologia 11:143-170. May, R. C. 1971. An annotated bibliography of attempts to rear the larvae of marine fishes in the laboratory. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-632, 24 p. 1974. Larval mortality in marine fishes and the critical period concept. In J. H. S. Blaxter (editor), The early life history of fish, p. 3-19. Springer-Verlag, Berl. O'CONNELL, C. P., AND L. P. RAYMOND. 1970. The effect of food density on survival and growth of early post yolk-sac larvae of the northern anchovy (Engraulis mordax Girard) in the laboratory. J. Exp. Mar. Biol. Ecol. 5:187-197. 461 FISHERY BULLETIN: VOL. 73, NO. 3 Steele, J. H. (editor). Thomas, W. H., A. N. Dodson, and C. A. Linden. 1970. Marine food chains. Univ. Calif. Press, Berkeley, 1973. Optimum light and temperature requirements for 552 p. Gymnodinium splendens, a larval fish food organism. Theilacker, G. H., and M. F. McMaster. Fish. Bull., U.S. 71:599-601. 1971. Mass culture of the rotifer Brachionus plicatilis and its evaluation as a food for larval anchovies. Mar. Biol. (Berl.) 10:183-188. 462 SPERMATOPHORES AND THELYCA OF THE AMERICAN WHITE SHRIMPS, GENUS PENAEUS, SUBGENUS LITOPENAEUS Isabel P^rez Farfante' ABSTRACT The spermaiophores of the five species of the American subgenus Litopenaeus of the genus Penaeus three in the Pacific-P. (L.) occidentalis, P. (L.) stylirostris, and P (L.) vannamei-^^nd two in the Atlantic-P. (L.) schmith and P. (L.) setiferus-are described in detail and illustrated. The sperma- tophore of P. vannamei uniquely lacks a wing and a lateral blade. That of P. stylirostris possesses a sac with overlapping walls, the free lateral margin not being attached to the underlying wall throughout most of its length; this spermatophore also exhibits the broadest wing, consisting of a rigid anterior region and a posterior membranous one. The spermatophore of P. occidentalis is the only one armed with an anterior lobe, a transverse anterior lamina, and a sclerotized flap. The spermatophores of the Atlantic species are very similar, both possessing moderately broad wings, large caudolateral flanges, and a lateral blade; however, in P. schmitti the blade is broad anteriorly, whereas in P. setiferus it is very narrow. During copulation, as the males deposit paired spermatophores on the females, the sperm masses are released through anterodorsal ruptures of the sperm sacs and become lodged on the thelycum, protected ventrally by the anterior part of the sacs. The open-type thelycum (sperm receptacle lacking),' characteristic of the members of Litopenaeus, is unique within the genus Penaeus. The thelyca are described in order to facilitate understanding how the compound spermatophores are held in place. The thelycum of each species exhibits at least one obvious typical feature by which it may be easily recognized: that of P. vannamei is provided with an inverted troughlike median protuberance on sternite XIII; in P. occidentalis it possesses densely set setae over most of sternite XIV; in P. styliros- tris it is armed with a strong, subpyramidal median protuberance on sternite XIV; in P. schmitti, on the other hand, it exhibits paired subparallel anterolateral ridges; and in P. setiferus paired crescentic anterolateral ridges are present which, although convergent, do not meet on the midline. In all but one species, thelycal concavities of sternite XIII serve to lodge the sperm masses which protrude from attached spermatophores; however, in P. occidentalis spoonlike coxal plates of the third pereopods receive the sperm masses. The five species of Penaeus, subgenus Li- topenaeus, commonly known as white shrimps, support some of the most intensive and valuable fisheries in American waters. Three species are limited to the eastern Pacific, Penaeus occiden- talis, P. stylirostris, and P. vannamei, and two occur in the western Atlantic, P. schmitti and P. setiferus. Mass rearing experiments to discover methods for artificial cultivation on a commercial scale are being conducted on all five species, and spermatophore-bearing or "impregnated" females are needed for this undertaking. Despite these ef- forts, and the considerable interest of biologists in the reproduction of these species (including mat- ing, spawning, and fertilization), descriptions of the spermatophores of only two of them and brief notes on a third are available. General features of the spermatophore of P. setiferus were presented by both Burkenroad (1934) and King (1948), and an 'Systematics Laboratory, National Marine Fisheries Service, NOAA, National Museum of Natural History, Washington, DC 20560. Manuscript accepted December 1974. FISHERY BULLETIN: VOL. 73, NO. 3, 1975. account of that of P. stylirostris was given by Cardenas Figueroa (1952). Subsequently, Ewald (1965) and Perez Farfante (1969) recorded a few observations on the spermatophore of P. schmitti. The spermatophores of the remaining two species, P. occidentalis and P. vannamei, have not been mentioned previously. Lack of information on spermatophores of the subgenus Lit&penaeus is due, at least in part, to the fact that impregnated females are not readily found (Weymouth et al. 1933; Burkenroad 1939; Heegaard 1953). In this exclusively American group, the females possess an open-type thelycum (Burkenroad 1934; Perez Farfante 1969), lacking a seminal receptacle and consisting, instead, of pro- tuberances, ridges, concavities, or grooves and, oc- casionally, lamellae on sternites XII to XIV to whichthe spermatophore is attached. The latter is thus exposed to the surrounding water and might be dislpdged during capture, as suggested by Burkenroad (1939), or retained for only a short period after copulation. In the females of all other 463 FISHERY BULLETIN: VOL. 73, NO. 3 subgenera of Penaeus, the thelycum exhibits a seminal receptacle where the spermatophores, deposited by the male, remain well protected until the time of spawning or until the succeeding molt. Among the extensive collections of the subgenus Litapenaeus that I have examined, no females of either P. occidentalis of P. vannamei carrying spermatophores were found. Four impregnated females of P. stylirostris and two of P. schmitti in the National Museum of Natural History, Smith- sonian Institution, constituted the only material at my disposal when this study was initiated. After several unfruitful attempts to collect sper- matophore-bearing females in various localities throughout the range of the Pacific species, I ob- tained such specimens of the three in the Gulf of Panama in March 1973. 1 caught an additional one of P. occidentalis in the same month off Buenaventura, Colombia. In the fall of 1974, Billy R. Drummond sent me three impregnated females of P. stylirostris which had been collected off Costa Rica, and recently Harold H. Webber brought me eight spermatophore-bearing females of P. stylirostris and five of P. vannamei from the same area. The study of the spermatophore of the Atlantic P. setiferus was based largely on one female carrying a complete spermatophore and three additional ones in which paired masses of sperm embraced by winglike processes were present on sternite XIII; these specimens had been caught in the waters of North Carolina, and were made available to me by Austin B. Williams. Recently, further observations were made on four impregnated females, two from Apalachicola Bay, Fla., and two from off Texas, given to me by William H. Clark, Jr. and Kenneth N. Baxter. The spermatophores of the various species are similar (herein described as when in position on the female), each consisting basically of a roughly semicylindrical hardened sperm sac enclosing a columnar sperm mass (spermatozoa within a vis- cous fluid) surrounded by a thick "sheath" (King 1948) of gelatinous substance (Figure lA, B). The sac usually bears an anterolateral aliform process, the wing, and is produced caudally or caudolaterally, in a flange. A lateral flap, variable in width and consistency, extends both along the sac and the flange, or only along the latter, usually attached to a firm, elongate blade, and a hardened but plastic dorsal plate is present on the posterodorsal surface of the spermatophore. Finally, a glutinous material always lies against the flange, adhering to the mesial side of the flap. These various accessories associated with the sperm sac presumably help to anchor the sperma- tophores to the female. The wall of the sperm sac consists of three somewhat distinct longitudinal regions: a thick, opaque ventral wall, a mostly thin and translucent lateral wall, and an entirely translucent dor- somesial wall. The heavy ventral wall is produced in a longitudinal mesial lapel, clearly delimited by a line along which the ventral and dorsomesial walls meet. The above terminology is consistently employed in the descriptions that follow. During copulation, two closely attached sper- matophores, which are here referred as the com- pound spermatophore, are transferred to the female (Figure IC). Immediately after expulsion from the paired terminal ampullae of the male, each spermatophore joins its mate firmly along the dorsomesial walls of the sacs, thus forcing the mesial lapels to project ventrally from the con- tiguous ventral walls (Figure ID). These walls, becoming strongly convex when the compound spermatophore is anchored to the thelycum, are responsible for the podlike appearance of the con- joined paired sacs, which constitute a median double structure referred to below as the geminate body. If the posterior part of the thorax of live mature males is compressed, the spermatophores are readily expelled through the gonopores situated mesially on the coxae of the fifth pereopods. It then may be observed that the spermatophores leave the ampullae with the anterior end foremost, the surface facing the sternum of the male being the same as that which, through a rotation, comes to lie against the thelycum. Consequently, as Kishinouye (1900) first indicated, the right and left spermatophores on the female originate in the corresponding right and left terminal ampullae of the male. Observations by Hudinaga (1942) on Penaeus (Marsupenaeus) japonicus Bate 1888, demonstrated that copulation takes place in a head to head position, the sternum of the male pressing against that of the female. Previously, Burkenroad (1934) and later King (1948) presented hypotheses concerning the transfer of the sper- matophores to the thelycum, taking into account the probable utilization of the petasma-ap- parently so well fitted to lodge the compound spermatophore— and the pereopods. No satisfac- tory explanation, however, has been advanced as to how a rotation of the spermatophore through 180° around its longitudinal axis, is accomplished. 464 PEREZ FARFANTE: SPERMATOPHORES OF AMERICAN WHITE SHRIMPS W ■^•^i Figure 1.— The spermatophores of Penaeus {Litopenaeus) schmitti and P. (L.) setiferus illustrating terms used in the descriptive accounts. A , Ventrolateral view of P. (L.) schmitti. B, Dorsomesial view of same. C, Ventral view of a compound spermatophore of P. (L.) setiferus attached to female. D, Cross section of the geminate body of same, immediately posterior to the wings (in preparation of the section the blades together with the torn contiguous portion of the lateral walls have become displaced laterally), b, blade; dp, dorsal plate; dw, dorsomesial wall;/, flap;/g', flange; gb, geminate body; gm, glutinous material; gp, gonopore; gs, gelatinous substance; Iw, lateral wall; ml, mesial lapel; sm, sperm mass; ss, sperm sac; vw, ventral wall; w, wing; XII, XIII, and XIV, sternites. 465 FISHERY BULLETIN: VOL. 73, NO. 3 The intent of this paper is to present detailed descriptions of the spermatophores of the five species of Litopenaeus. Except for the close resemblance of the spermatophores of P. schmitti and P. setiferus, all of them, although structurally similar, are quite different in appearance; therefore I have emphasized apparent homologies. An effort has been made to explain the manner in which the spermatozoa egress from the sperma- tophores to fertilize the eggs. The association of the components of attached spermatophores with the corresponding thelycum in each species is in- dicated. Finally, the role played by the coxal plates of the third through the fifth pairs of pereopods of the females in keeping the compound sperma- tophore attached to the thelycum is briefly dis- cussed. The material examined is indicated in the treatment of each species. The following abbreviations are used for repositories of the specimens: ANSP - Academy of Natural Sciences of Philadelphia; INIBP - Instituto Nacional de Figure 2.-Pena€Ms (Litopenaeiia) vannamei Boone. Compound spermatophore attached to female, 9 43 mm cl, off Panama Province, Panamd. Abbreviations as in Figure 1. Investigaciones Biologico-Pesqueras, Mexico; UNC-IMS - Institute of Marine Sciences, Univer- sity of North Carolina; USNM - National Museum of Natural History, Smithsonian Institution; and YPM - Peabody Museum of Natural History, Yale University. The carapace length (cl) is the linear distance between the orbital margin and the mid- posterior margin of the carapace. The illustrations have been made from preserved specimens; the accompanying scales are in millimeters. DESCRIPTIONS OF SPERMATOPHORES AND THELYCA The descriptive accounts of the Pacific species are ordered according to the relative complexity of the respective spermatophores (P. vannamei, P. occidentalis, and P. stylirostris), and are followed by those of the Atlantic species (P. schmitti and P. setiferus), the spermatophores of which are markedly similar. The spermatophores of each species are described both as attached to the females, where they invariably occur in pairs, and as they appear when removed from the terminal ampullae of males. Next, detailed accounts of the thelyca are presented, which emphasize the main features for the support of the component parts of the spermatophores. A list of the material examined is given, including the numbers of impregnated females. Finally, the geographic range of each species is indicated. Penaeus {Litopenaeus) vannamei Boone 1931 Figures 2-4 Spermatophore The compound spermatophore (Figure 2) con- sists of a slender geminate body lacking wings and blades, and bearing thick, broad, lateral flaps, and a pair of long, caudal flanges. Ventrally, each spermatophore (Figure 3A) exhibits a lateral furrow that roughly delimits a subovate anterior portion, bulging laterally, from an elongate, smoothly convex portion that extends to the flange; the thick, opaque ventral wall merges indistinctly with the lateral wall, which is mostly opaque and to which is loosely attached a conspicuous internal lamina; posteriorly, just before joining the flange, these walls turn strongly , dorsad along an oblique line forming the fundus of the sac. The broad, mantlelike flap projecting from the lateral wall is thick, fleshy ventrally, and 466 PEREZ FARFANTE: SPERMATOPHORES OF AMERICAN WHITE SHRIMPS A B C Figure 3.-Penaeus (Litopenaeus) vannamei. Left spermatophore dissected from terminal ampulla, < » - S- 2 - a ^ = S £ »-*■ A (0 t_ .E £ o) E " 1^5 "° 1^2 Si oc ojD tr a o E- (Q (D C M CO r (1) CO (r a ?S?.E I O CO CO >» — XI r- co o E - .°X ® S c E CO •- ™ 1" E ^ o o CD t ffl CO OC m - 0) <31 O) ^ CO - ;;! ^ 0] 3 CO £;H 2 3 2 3 o -I n c 5 o CO Z CO o c u CO c c 0} CO m v> O) c o x 5 o CO z ■D ri CO £ 2 5 m E 0) c c o CO ■D CO o •D CO o . >. •a 0, CO T3 C < 0 n — CO "D 0) i^ °^ CO o ■ - CD O < o .2 □) CO "O E - o ^ C (U > o ■° r CO < 2>-S. CO o <» CD — CD Qj 0) JD < c o CU V CO T3 W CO B 0)-o -J o (3) CD 0) - o 0)T3 c u O CD _1 O 2 « - o C3)T3 C =J O CO -I o D) c CO E W a (U a. Q. E E w CO .« S ra E c cu a E o o >. >^ 0) O ra ra £ o 0) a, ^ o> "S S) o t: o i- 5 J5 2i5 © ^ a. Q. E E M « o o o 6 <: o CO CO 3 CD ■S. (a cu c E o cu cow « S CD_ E 5 O CO CO a> oB ■a o t: o S'^ o a 3 a C JO c o o > o o "cow S S 7515 to 0] to ro It seems noteworthy that P. vannamei is the only species of the subgenus Litopenaeus in which a strongly developed median protuberance (projecting from sternite XIII) is present on the thelycum. Females of the remaining species of Litopenaeus lack a median protuberance, unlike all of the other species of Penaeus. In the former females, the midposterior part of sternite XIII bears instead a simple, relatively small knob-in P. occidentalis— or is produced into a shelf which overhangs sternite XIII. This shelf is horizontal in P. schmitti and P. setiferus, and subvertical in P. stylirostris. The spermatophore of P. occidentalis is con- siderably more elaborate than that of P. van- namei, although the sac is structurally similar in the two and simpler than that of P. stylirostris. The spermatophore of P. occidentalis, unlike that of P. vannamei, possesses a wing and, unlike all other species of Litopenaeus, bears an anterior lobe, and is produced caudally in a very short flange. Also, it bears the largest lateral blade to be found in any of them, the blade being divided in several sections and continuous with the typical stiffened ventromesial extension, ending in a rather flexible caudal lobe; that extension seems to correspond to the flap borne by spermatophores of the other species. In addition, the spermatophore of P. occidentalis possesses a unique transverse lamina on the anterior extremity of the sac, which is not intimately associated or even firmly at- tached to it or to the other contiguous components, i.e., the anterior lobe and the wing. In this sper- matophore, however, the flange is inconspicuous, consisting of a short rigid shelf at the posterior end of the sac. Finally, the compound sperma- tophore of P. occidentalis is affixed to the female farther anteriorly than in the other species, the anterior lobes extending over (ventrally) the coxae of the third pereopods to become attached to ster- nite XII; this brings the openings of the sacs to the scooplike coxal plates of the third pereopods, al- most directly opposite the female gonopores. The spermatophore of P. stylirostris differs from all the others chiefly in the structure of the sperm sac, which is largely formed by the dor- somesial wall. This wall, after extending laterally, is bent mesially in such a way as to reach, ven- trally, the base of the mesial flap; as a result, part of the ventral and lateral walls, apart from giving support to the sac, serve as a protecting shield. The spermatophore of P. stylirostris also possesses the broadest wing to be found within the subgenus. 484 fKKEZ FAKKANTl!;: ShKKMATOPHORES OF AMERICAN WHITE SHRIMPS and is armed with a blade, the anterior portion of which is directed both dorsomesially and ven- tromesially instead of laterally, in specimens removed from males. Along the sac, the lateral flap is almost as broad as that in P. vannamei, but thinner and not fleshy. Furthermore, in P. stylirostris, the paired dorsal plates affix only the posterior part of the spermatophore to the females, not directly supporting the midportion. The flanges become attached to the female almost perpendicular to the geminate body, instead of extending entirely caudad as in P. vannamei or somewhat caudad as in P. schmitti and P. se- tiferus. The spermatophores of P. schmitti and P. se- tiferus are almost identical. They may be distin- guished by the width of the lateral blade, the anterior portion of which is broad (except for a narrow portion at the base of the wing) in P. schmitti and very narrow in P. setiferus. Sperm sacs of both species are similar to those of P. van- namei and P. occidentalis; however, they bear wings, which are lacking in P. vannamei, and the wings are moderately large and scabrous, with various projections on the ventral surface, thus very different from the small wing with the mar- gins extensively folded found in P. occiden- talis.Also, the flanges extend considerably anteriad along the lateral walls of the sacs, whereas those in P. vannamei are virtually caudal and, unlike P. occidentalis, are produced posteriorly much beyond the sac instead of barely overreaching the fundus. Finally, flaps borne by the flanges are broad in the two Atlantic species, instead of narrow as in P. vannamei, and, although firm, are different in texture from the heavy sclerotized shelf sustained by the flange in P. oc- cidentalis. Despite the similarities of the various sacs, the mode of dehiscence varies -differences among the spermatophores are due more to the elements as- sociated with the sac than to the sac itself. It seems that in P. occidentalis, the sperm is released through anterior openings of the sacs which become applied to the coxal plates of the third pereopods, and are well protected by the anterior lobes of the spermatophores. In P. stylirostris as well as in P. schmitti and P. setiferus, it seems that the compound spermatophore attached to the female splits longitudinally into two parts, the geminate body breaks away leaving paired masses of sperm on the thelycum freely exposed to the surrounding water. In P. schmitti and P. setiferus. however, there are certain indications that the sperm reaches the water through a passageway between the wings and the body of the female, the geminate body persisting. On the basis of the available material of P. vannamei, there is no in- dication as to how the sperm escapes from the spermatophore. Understanding of the precise manner in which spermatozoa are freed from spermatophores in all of the species must await direct observations. ACKNOWLEDGMENTS I am deeply indebted to Meredith L. Jones, Leader, Marine Biota of Panama Project, Smith- sonian Institution, whose interest and efforts in providing equipment and assistance in the field contributed substantially to the success of my trip to Central America; Ira Rubinoff, Director, Smithsonian Tropical Research Institute, who arranged for shipboard explorations in the Gulf of Panama and who made available the facilities of the Institute at Balboa; and Richard A. Neal, Investigation Chief of Aquaculture Research Technology, Gulf Coastal Fisheries Center Galveston Laboratory, National Marine Fisheries Service, NOAA, for assistance in securing living white shrimp from off Texas. I am especially grateful to Alfredo Rizo and Ernestina Rizo for placing at my disposal the shrimp trawler San- tanderino, manned by a most efficient crew, and for their part in making my stay in Colombia so thoroughly enjoyable; Ernestina Rizo also cooperated in handling and preserving specimens aboard ship. Gratitude is also expressed to Billy R. Drummond and Harold H. Webber who went out of their way to make available to me recently collected and beautifully preserved impregnated females of two species from Costa Rican waters. It is a pleasure to extend an especial word of gratitude to Horton H. Hobbs, Jr., of the Smith- sonian Institution, whose carcinological background has been the source of many valuable suggestions. Thanks are due Fenner A. Chace, Jr., of the Smithsonian Institution; Austin B. Williams, of the Systematics Laboratory, National Marine Fisheries Service, NOAA; and Won Tack Yang, of the Institute of Marine Sciences, University of Miami, for their critical review of the manuscript. The patience and wholehearted cooperation of Maria M. Dieguez in executing the excellent illustrations are fully appreciated. Funding for collecting in Panamanian waters 485 FISHERY BULLETIN: VOL. 73, NO. 3 was provided by the Smithsonian Institution, En- vironmental Sciences Program. LITERATURE CITED BURKENROAD, M.D. 1934. The Penaeidea of Louisiana with a discussion of their world relationships. Bull. Am. Mus. Nat. Hist. 68:61- 143. 1939. Further observations on Penaeidae of the northern Gulf of Mexico. Bull. Bingham Oceanogr. Collect. 6(6):l-62. BURUKOVSKII, R. N. 1972. Some problems of the systematics and distribution of the shrimps of the genus Penaeus. [In Russ.] Fisheries Research in the Atlantic Ocean. Tr. Atlant. Nauchno- Issled. Inst. Rybn. Khoz. Okeanogr. 42:3-21. (Translated by Israel Program Sci. Transl., TT72-50101.) CARDENAS FiGUEROA, M. 1952. Descripci6n del espermat6foro de Penaeus stylirostris Stimpson. Rev. See. Mex. Hist. Nat. 13:9-16. Eldred, B. 1958. Observations on the structural development of the genitalia and the impregnation of the pink shrimp, Penaeus duorarum Burkenroad. Fla. State Board Con- serv. Mar. Lab., Tech. Ser. 23, 26 p. EWALD, J. J. 1965. Investigaciones sobre la biologfa del camar6n comer- cial en el occidente de Venezuela. Inst. Venez. Invest. Cient., Segundo Informe anual al Fondo de Inves- tigaciones Agropecuarias. 1-128, 137-147 p. Heegaard, p. E. 1953. Observations on spawning and larval history of the shrimp, Penaeus setiferus (L.). Publ. Inst. Mar. Sci., Univ. Tex. 3(1):73-105. Heldt, J. H. 1938a. De I'appareil genital des Penaeidae. Relations morphologiques entre spermatophore, thelycum et pe- tasma. Trav. Stn. Zool. Wimereux 13:349-358. 1938b. Le reproduction chez les crustaces decapodes de la famille des peneides. Ann. Inst. Oceanogr. 18:31-206. Hudinaga (Fujinaga), M. 1942. Reproduction, development and rearing of Penaeus japonicus Bate. Jap. J. Sci. 10:305-393. King, J. E. 1948. A study of the reproductive organs of the common marine shrimp, Penaeus setiferus (Linnaeus). Biol. Bull. (Woods Hole) 94:244-262. KiSHINOUYE, K. 1900. Japanese species of the genus Penaeus. J. Fish. Bur. Tokyo 8:1-29. Malek, S. R. a., and F. M. Bawab. 1974a. The formation of the spermatophore in Penaeus kerathurus (Forskal, 1775) (Decapoda, Penaeidae). I. The initial formation of a sperm mass. Crustaceana 26:273-285. 1974b. The formation of the spermatophore in Penaeus kerathurus (Forskal, 1775) (Decapoda, Penaeidae). II. The deposition of the main layers of the body and of the wing. Crustaceana 27:73-83. MOUCHET, S. 1931. Sur I'appareil genital male de Penaeus trisulcatus Leach. Bull. Soc. Zool. Fr. 56:458-467. PfiREZ FARFANTE, I. 1969. Western Atlantic shrimps of the genus Penaeus. U.S. Fish Wildl. Serv., Fish. Bull. 67:461-591. TiRMizi, N. M. 1958. A study of some developmental stages of the thelycum and its relation to the spermatophores in the prawn Penaeus japonicus Bate. Proc. Zool. Soc. Lond. 131:231-244. Weymouth, F. W., M. J. Lindner, and W. W. Anderson. 1933. Preliminary report on the life history of the common shrimp Penaeus setiferus (Linn.). Bull. U.S. Bur. Fish. 48:1-26. 486 DYNAMICS OF AMERICAN SHAD, ALOSA SAPIDISSIMA, RUNS IN THE DELAWARE RIVER' Mark E. Chittenden, Jb.^ ABSTRACT Adult American shad were collected by haul seining at regular intervals during the spring migrations from 1963 to 1965 and by collecting specimens from a fishkill in spring 1965. Supplementary sex and size composition data were obtained from summer rotenone sampling in 1961 and 1962. Males tended to precede females in the spring run. Annual sex compositions were strongly male dominated from 1961 to 1963 and were strongly female dominated in 1964 and 1965. Repeat spawners composed 2.6% of 729 fish examined from 1963 to 1965. Age I and II fish were absent or virtually absent from the run. Males migrated upstream primarily at age IV and females at age V. Large runs during the early 1960's were based on the 1958 and 1959 year classes as defined which were much larger than other year classes produced in the period 1957-64. Delaware River American shad runs now have little buffering against fluctuations in abundance because few year classes are successful, few age groups support the run, and there are essentially no repeat spawners. Larger runs in the Delaware since 1925 and probably somewhat earlier, in general, were apparently based on one large year class and essen- tially no repeat spawners. The American shad, Alosa sapidissima, formerly was one of the most abundant anadromous fishes in the United States where more than 50 million pounds were landed in 1896. In contrast, only 8 million pounds were caught in 1960 (Walburg and Nichols 1967). Much of this decline is due to the collapse of the Delaware River Basin fisheries which once supported larger landings of American shad than any other river system (Stevenson 1899). Estimates of the 1896 Delaware Basin catch are about 16.5 to 19.2 million pounds (Smith 1898; Sykes and Lehman 1957; Walburg and Nichols 1967)-about a third of the national total. Annual Delaware Basin catches about that time varied from 14 to 17 million pounds and were probably primarily fish of Delaware River origin (Chitten- den 1974). In contrast, annual Delaware Basin landings since 1920 have consistently been much less than 0.5 million pounds (Sykes and Lehman 1957; Chittenden 1974). Trends in abundance of Delaware Basin stocks have been described by Sykes and Lehman (1957) and Chittenden (1974), and causes of these fluc- tuations have been described by Ellis et al. (1947), Sykes and Lehman (1957), Walburg and Nichols 'Based on part of a dissertation submitted in partial fulfillment of the reauirements for PhD degree, Rutgers University, New BrunswicK, N.J. 'Department of Wildlife and Fisheries Sciences, Texas A & M University, College Station, TX 77840. Manuscript accepted October 1974. FISHERY BULLETIN: VOL. 73, NO. 3, 1975. (1967), and Chittenden (1969). The dominant fac- tor affecting abundance during the past 60 yr, at least, has been pollution near Philadelphia, Pa. Complete understanding of the causes of fluctua- tions in abundance, however, depends on detailed knowledge of the population dynamics of this species. Many workers have described certain as- pects of the population dynamics of American shad in other river systems. Little work has been published for the Delaware, in part because Delaware River American shad stocks have been so low that it has been difficult to collect large numbers of fish. The present paper describes data collected on sex, size, age, and repeat spawner composition; comparative magnitudes of American shad runs; and year-class strengths in the Delaware River during the late 1950's and the 1960's. MATERIALS AND METHODS Adult fish were collected 22.5 km (14 miles) above tidal water in the years 1963 to 1965 at the site of the Lewis Fishery in Lambertville, N.J., using a 76-mm (3-inch) stretch-mesh, 107-m (350- foot) long and 3.6-m (12-foot) deep (4.3 m = 14 feet in 1963) haul seine that was paid out from a boat and landed about 396 m (1,300 feet) downstream. Sampling occurred at 4-day intervals after a ran- domly selected date in 1963 but at fixed intervals 487 FISHERY BULLETIN: VOL. 73, NO. 3 twice weekly in later years. Sampling occurred from 5 April to 19 May in 1963, from 20 March to 18 May in 1964 and from 26 March to 7 May in 1965. In 1964 and 1965, at least, several collections were made both before and after the first and last American shad were captured, but a few fish were captured on the first and last sampling dates in 1963. My sampling in 1963 essentially ended when the run did at Lambertville. I captured 1 fish after 15 May, and the Lewis Fishery captured 5 fish after that date of a total catch of about 4,000 fish (Chittenden 1969). The number of seine hauls and time of sampling varied each day during 1963. After 1963 two seine hauls were made from 1100 to 1300 EST on each sampling day after the first American shad was landed. In spring 1965, low dissolved oxygen levels near Philadelphia blocked upstream passage of part of the spawning run, and few fish were captured at Lambertville (Chit- tenden 1%9). Hundreds of dead fish were collected during a fishkill seaward of Philadelphia in the area from Paulsboro, N.J., to Marcus Hook, Pa., during the period from 21 May to 10 June. Sex and size composition data for 1961 and 1962 were ob- tained from cooperative surveys (hereinafter referred to as the Tri-state Surveys) using rotenone during July and August by the states of New Jersey, New York, and Pennsylvania in con- junction with the U.S. Fish and Wildlife Service. I personally examined many of the American shad collected. Each fish collected after 1962 was measured and was sexed by examination of the gonads, and scales were taken from the midline of the left side below the dorsal fin following Cating (1953) or from the same area on the right side. Many fish collected near Marcus Hook during 1965 had lost all or nearly all their scales, so scales were taken where available on these fish. Scales selected for age determination were cleaned and were dry- mounted between glass slides. Aging was done with a microprojector using Cating's (1953) method which was verified by La Pointe (1958) and- Judy (1961). Scales were examined in the time sequence 1963, 1964, and 1965 for an initial deter- mination. Many of the large fish were known to have been misaged when the sequence was completed. Scales were next thoroughly mixed to remove bias due to knowledge of the collection year and were reread. The first two readings disagreed on 31 of 301 fish (10%) from 1963, 38 of 199 (19%) from 1964, and 3 of 184 (2%) from 1965. Of these, 32 disagreements were on females reas- signed from age-group V to VI (9 from 1963 and 23 from 1964). A third examination of the misaged scales agreed with the second on 67 of the 72 fish. The remaining 5 were discarded. RESULTS AND DISCUSSION Sex Composition of the Spawning Runs Many American shad were captured at Lam- bertville during 1963 (301 fish) and 1964 (199 fish). Figure 1 shows trends in the cumulative percent- age of males as these spawning runs progressed. Trends were similar during each year although the 1963 run contained a much higher percentage of males than the 1964 run. The cumulative propor- tion of males decreased as the run progressed in agreement with the observations or statements of many workers including Stevenson (1899), Prince (1907), Leach (1925), Hildebrand and Schroeder (1928), Nichols and Tagatz (1960), and Walburg and Nichols (1967). Annual sex compositions of the runs from 1961 to 1965 are presented in Table 1 with 95% con- fidence limits for the proportion of males based upon normal approximations except for 1962 when a Poisson approximation to the binomial was used (Cochran 1953). Annual sex composition varied greatly from 1961 to 1965. It was male dominated from 1961 to 1963 and female dominated thereafter. > 60 (7) X = 1963 (92) (49) X X • (125) X (185) '*I2Z9I (296) (301)1302) X X X " = 1964 (21) "''is?) (116)(127) • • (150)ll82l„98„ig9) Figure 1. -Trends in the cumulative percentage of male American shad. Numbers in parentheses represent cumulative numbers of fish captured. 488 CHITTENDEN: DYNAMICS OF AMERICAN SHAD RUNS Table 1.- Annual sex compositions of American shad runs, 1961-65. Year Sample size Proportion male (p) 95% CL about pi 1961 198 0.86 0.81-0.94 1962 220 0.99 0.96-1.00 1963 302 0.62 0.57-0.67 1964 199 0.38 0.31-0.45 1965 Lambertville 23 0.52 0.32-0.72 1965 Tidal area 190 0.23 0.17-0.29 'Confidence limit for proportion of males. Size Composition Mean fork lengths, numbers of fish, 95% con- fidence limits for means, and 99% limits for in- dividuals collected at Lambertville or the tidal area and during the Tri-state Surveys are sum- marized in Table 2. Data were originally expressed in total or fork length depending on year of collection. Means and confidence limits were transformed to fork lengths in inches using the regression of fork on total length presented by Chittenden (1969) and were then transformed to metric units. Females from Marcus Hook and Lambertville did not appear significantly different in size and were pooled. Males from Marcus Hook were significantly smaller {t = 4.67) than those from Lambertville, and they were not combined. Fish were smaller in 1961 and 1962 than in later years, and this probably reflects the presence of very young individuals of the strong 1958 and 1959 year classes (see section "Comparative Year-Class Strengths"). Females averaged about 50 mm (2 inches) longer than males each year. Confidence limits for individuals indicate that there were few females longer than 545 mm (21.5 inches) and few males longer than 505 mm (19.9 inches). Maximum observed sizes were 559 and 495 mm (22.0 and 19.5 inches) for females and males, respectively. Small fish did not migrate upstream. The smallest female was about 64 mm (2.5 inches) longer than the smallest male each year. There were few females shorter than 416 mm (16.4 inches) in 1963, 1964, and 1965, the smallest being 406 mm (16.0 inches). The smallest female observed was 376 mm (14.8 inches) and 391 mm (15.4 inches) in 1961 and 1962, respectively. Confidence limits indicate that there were few males shorter than about 337 mm (13.2 inches) from 1962 to 1965. There were few males shorter than 289 mm (11.4 inches) in 1961, the smallest fish measured being 269 mm (10.6 inches). Age and Repeat Spawner Comsposition There were only 2.6% repeat spawners in 729 fish examined from 1963 to 1965. Annual occurrences were 1 of 299 fish (0.3%) in 1963, 3 of 199 (1.5%) in 1964, and 15 of 231 (6.5%) in 1965. Confidence limits for the annual percentages are 0.0 to 1.9% (1963) and 0.03 to 4.4% (1964) based upon Poisson approximations and 3.5 to 9.5% (1965) based upon the normal approximation. Table 3 summarizes male and female age-class structures of the 1963 and 1964 runs at Lambert- ville, pooled data for 1965 from Lambertville and Marcus Hook, and data pooled over all years. Fluctuations in age composition due to variable year-class recruitment are apparent, but certain general features of the age composition stand out. Table 2. -Summary of fork lengths of American shad, 1961-65. Lengths are in millimeters (top rows) and inches (bottom rows). Males Females Year n Mean (x) 95% CL X 299% CLx n Mean (x) 195% CL X 299% CL X 1961 170 401 394-407 289-512 28 452 441-463 371-533 15.8 15.5-16.0 11.4-20.1 17.8 17.4-18.2 14.6-21.0 1962 217 415 414-415 362-467 3 451 322-579 — 16.3 16.3-16.3 14.2-18.4 17.8 12.7-22.8 — 1963 186 428 425-431 371-485 115 472 468-476 416-529 16.9 16.7-17.0 14.6-19.1 18.6 18.4-18.8 16.4-20.8 1964 76 427 420-434 350-504 122 480 477-484 428-533 16.8 16.6-17.1 13.8-19.8 18.9 18.8-19.1 16.9-21.0 1965 12 436 417-455 342-530 11 483 472-495 431-536 Lambertville 17.2 16.4-17.9 13.5-20.9 19.0 18.6-19.5 17.0-21.1 1965 43 417 408-426 337-498 147 484 480-488 422-546 Tidal area 16.4 16.1-16.8 13.2-19.6 19.1 18.9-19.2 16.6-21.5 1965 Combined — '~~ = 158 484 19.1 481-488 18.9-19.2 423-545 16.7-21.5 'Confidence limit for mean fork lengtfi. ^Confidence limit for indivdual fish collected. 489 !• ISHKK Y BULLhTlN: VOL. 73, NO. 3 Table 3.-Age-class structures of American shad, 1963-65. 1963 1964 1965 Total Age n % n % n % n % Males II 0 0 1 1 0 0 1 0.00 III 4 2 1 1 3 6 8 2.58 IV 149 82 51 68 36 68 236 76.13 V 27 15 21 28 14 26 62 20.00 VI 2 1 1 1 -emales 0 0 3 0.01 IV 31 27 9 7 10 8 50 13.70 V 63 56 77 64 85 65 225 61.64 VI 19 17 35 29 34 26 88 24.11 VII 0 0 0 0 2 2 2 0.01 The following comments essentially apply to vir- gin fish because the percentage of repeat spawners was negligible. Age I and II males did not migrate upstream, although one age II male was captured in 1964. Few age III males migrated upstream, the per- centage based on pooled data over all years being less than 3%. Males migrated upstream primarily at age IV (76%), but age V (20%) was also impor- tant. Few males survived to age VI, and all were virgin. No males older than age VI were captured. No females younger than age IV or older than age VII were observed. Only two age VII females (less than 1% of the pooled total) were captured, and one of these was a virgin. Females were primarily age V (62%) when they first entered the fishery, but ages IV (14%) and VI (24%) were also important. Comparative Magnitudes of American Shad Runs, 1961-68 I here define magnitude of American shad runs as the numbers of adults reaching Lambertville on their upstream migration. The magnitude of runs as defined is influenced by year-class strength as later defined herein and by dissolved oxygen levels that the adults encounter during migration up- stream past the Philadelphia area (Chittenden 1969). Chittenden examined in detail the annual effects of oxygen upon passage of adults and young. In general, dissolved oxygen is sufficiently high that the earlier stages, at least, of the adult run successfully migrate upstream. Reasonably precise estimates of the compara- tive magnitudes of American shad runs in the 1960's are available from three sources of evidence: 1) catches of the Lewis Fishery at Lambertville presented by Chittenden (1974), 2) my own catches at Lambertville, and 3) sex compositions of the runs presented in Table 1. Lewis Fishery annual catch /seine haul records indicate the comparative magnitudes of runs in descending size order were: 1963 (56.1 fish), 1964 (18.3), 1962 (13.9), 1965 (6.6), 1967 (3.7) = 1961 (3.5), and 1966 (1.8) = 1968 (1.2). Values in parentheses represent catch/seine haul. My annual catch/seine haul at Lambertville for the time period between capture of the first and last fish was 17.4 fish in 1963, 9.0 in 1964, and 3.0 in 1965, a pattern in agreement with the Lewis Fishery records. My general impressions derived from angling experience, interviews with other anglers, and visual observations of the abundance of adult American shad during extensive float trips each year from 1961 to 1974 are in general agreement with the patterns suggested by the Lewis Fishery records. Runs were much smaller from 1966 to 1968 than from 1962 to 1964. The size of the run in 1965 (and possibly 1961 based on my general impres- sions) was intermediate between the sizes in these two periods. The relative magnitudes of runs estimated from sex composition data (Table 1) agree with patterns indicated by catch records at Lambertville. The sex ratio shift from 1963 to 1964 suggests passage through the fishery of a year class stronger than the one immediately following because males tend to enter the run a year earlier than females. This indicates the 1963 run was larger than that in 1964. The male proportion at Lambertville in 1965 (0.52) is biased towards the high side because the later part of the run was blocked by low dissolved oxygen near Philadelphia. The male run was much greater in 1963 than in 1964. Confidence limits for the male proportion of fish collected at Marcus Hook in 1965 were much lower than any single daily male proportion in 1963 (Figure 1) but were similar to the lowest daily male proportion in 1964. Therefore, the 1965 run was probably similar in sex composition to that of 1964. Females enter the run a year older than males. Unless females suffer a much lower annual mortality, similarity of sex compositions suggests that the 1965 run was smaller than the 1964 run-even before the fish reached the Philadelphia area. The high female proportion in 1965 suggests an even smaller run was due in 1966, another year when low oxygen levels blocked part of the run (Chittenden 1969). Dissolved oxygen was suflSciently high throughout the 1967 run to permit large numbers of fish to reach Lambertville (Chittenden 1969). The Lewis Fishery in 1967 did not make large catches to reflect 1968 males or females associated with the 490 CHITTENDEN; DYNAMICS OF AMERICAN SHAD RUNS 1966 males, suggesting that the 1966 and 1968 runs were small-even before the fish reached the Philadelphia area. The higher proportion of males in 1962 than in 1961 indicates that the 1962 run was larger than the 1961 run. Comparative Year-Class Strengths, 1956-64 I here define year-class strength as the numbers of young which exist at some constant point in time-say 1 January-after the entire year class has "passed" seaward through the grossly polluted Philadelphia area. Year-class strength is influenced by spawning success. However, the dominant factor, by far, in setting year-class strength in the Delaware River is the success with which the young pass seaward through the Philadelphia area in summer and fall (Chittenden 1969). Comparative year-class strengths can be es- timated from the age and sex compositions from 1963 to 1965 (Tables 1, 3) supported by estimates of comparative run sizes. Males usually first migrate upstream at age IV and females at age V, so that the size of a run chiefly reflects the strength of year classes produced 4 and 5 yr earlier assuming constant survival at sea. Therefore, American shad runs from 1961 to 1968 chiefly reflect year- class strengths from 1956 to 1964. The largest year classes as defined were produced from 1957 to 1960 because the largest runs were from 1962 to 1964. The 1963 American shad run was the largest in this period and primarily reflects the 1958 and 1959 year classes. I captured 90 age V fish at Lambert- ville in 1963 and 98 in 1964 with approximately equal effort, suggesting that the 1958 and 1959 year classes were similar. This agrees with com- parative magnitude estimates for the 1962 and 1963 runs. The 1963 run was based on two large year classes and was larger than the 1962 run which was based on one large and one smaller year class. The 1960 and 1959 year classes can be compared. I collected 180 age IV fish at Lambertville in 1963 but only 60 age IV's in 1964. This suggests that the 1959 and, by inference, 1958 year classes were much larger than that of 1960. This deduction is supported by the shift in sex composition from 1963 to 1964, which indicates the age V year class from 1959 was much stronger than the age IV year class from 1960. The 1963 run was based on two large year classes and was larger than the 1964 run which was based on one large and one smaller year class. The 1957 and 1958 year classes can be compared. At Lambertville, 36 and 21 age VI fish were cap- tured in 1964 and 1963, respectively, suggesting that the 1958 and, by inference, 195*9 year classes were stronger than that of 1957. This is supported by mean sizes of males collected during the Tri- state Sur\'eys of 1961 and 1962 (Table 2). The 1961 fish (mean = 401 mm = 15.8 inches) were sig- nificantly smaller than the 1962 fish (mean = 414 mm = 16.3 inches). The obsen-ed mean fork length of age IV males in 1963 and 1964 was about 422 mm (16.6 inches) (Chittenden 1969, table 17), suggest- ing that age III fish from 1958 were important in the 1961 run. The 1961 and 1960 year classes can be compared. Sex composition and catch data show that the 1964 run was smaller than the 1965 run. Age structures were almost identical, however, and this indicates that the 1961 year class was smaller than that of 1960. The strength of the 1962, 1963, and 1964 year classes as defined must have been small, probably being no larger than the 1961 year class, because the 1966 to 1968 runs were very small. In summary, the 1958 and 1959 year classes were extremely large in comparison to the year classes produced in the period 1961-64. The 1957 and 1960 year classes were much smaller than those of 1958 and 1959, but were larger than the 1962, 1963, and 1964 year classes. GENERAL DISCUSSION Delaware River American shad runs in the 1960's showed great shifts in the proportion of male fish. This variation was due, in large part, to fluctuations in year-class strength. The propor- tions in 1961 (0^86) and 1962 (0.99) are extreme. They are based upon summer collections and are biased if females tend to return seaward earlier than males do. However, these proportions also probably primarily reflect the temporary resur- gence of the Delaware River American shad runs in the early 1960's reported by Chittenden (1974). The magnitude of American shad runs in the Delaware River was at a very low level prior to 1961, increased in 1961 and 1962, and in 1963 the magnitude was such that it ranked among the largest runs in the last 45 >t or more. Because Delaware River males tend to undertake their first 491 FISHERY BULLETIN: VOL. 73, NO. 3 spawning migration at age IV and females at age V, the strongly male dominated runs of 1961 and 1962 reflect the entry of unusually large year classes into the fishery. The magnitude of the spawning runs declined greatly after 1963, and the shift towards female dominance reflects the failure of new year classes to be recruited. The strength of the 1962, 1963, and 1964 year classes as defined was small because the 1966 to 1968 runs were very small. Yet, from visual observations and collections of young during the summer, Chitten- den (1969) concluded that strong year classes were produced throughout the period 1962-66. I presented dissolved oxygen data for the periods when the young passed the Philadelphia area each year and ascribed the failure of these strong year classes to later recruit to the fishery to cata- strophic destruction of the young as they passed through the grossly polluted area. There was a negligible percentage of repeat spawners in the Delaware River American shad runs from 1963 to 1965, and there were no repeat spawners among 245 American shad collected from the 1961 run (J. Malcolm, pers. commun.). The existence of numerically few repeat spawners must have continued to be the case after 1965 because the runs from 1966 to 1968 were very small. Delaware River runs apparently have included few repeat spawners for at least some 25 yr because Sykes and Lehman (1957) reported that less than 2% of 423 fish captured in 1944, 1945, 1947, and 1952 were repeat spawners. Even the produc- tion of very large year classes appears to have little effect on the percentage of repeat spawners in the Delaware River. The percentage was only 6.5% in 1965, a year when the very strong 1958 and 1959 year classes should have increased the per- centage greatly rather than by only a few points. As Sykes and Lehman (1957) pointed out, the vir- tual absence of repeat spawners in the Delaware River is very unlike that in other middle-North Atlantic coast rivers where repeat spawner per- centages have been: St. Johns River, N.B., Canada, 22-81% (Leggett 1969); Connecticut River, 14-60% (Moss 1946; Nichols and Tagatz 1960; Leggett 1969); Hudson River, 51% (Talbot 1954); Susquehanna River, 37% (La Pointe 1958); Po- tomac River, 17% (Walburg and Sykes 1957); York River, 21-27% (Nichols and Massmann 1963; Leg- gett 1969); James River, 27%, (Walburg and Sykes 1957). Some of the higher percentages of repeat spawners in these rivers are undoubtedly biased towards the high side because they are based on collections of fish from hightly selective commercial gill nets. However, it appears that the Delaware River has a much lower percentage of repeat spawners. The Delaware River seems most like rivers south of Cape Hatteras, N.C., where few repeat spawners have been reported: Neuse River, less than 3% (La Pointe 1958; Walburg 1957); Edisto River, 0% (Walburg 1956); Ogeechee River, 0% (Sykes 1956); Altamaha River, 0% (Godwin and McBay 1967; Godwin 1968); St. Jones River, Fla., 0% (Walburg 1960; Leggett 1969). Age-class structures of American shad runs have been reported by many workers including Talbot (1954), Fredin (1954), Walburg (1956, 1957, 1960, 1961), Walburg and Sykes (1957), Sykes (1956), La Pointe (1958), Nichols and Tagatz (1960), Nichols and Massmann (1963), Godwin (1968), and Leggett (1969). Walburg and Nichols (1967) sum- marized the available information by stating that age IV and V fish make up the bulk of the catch in South Atlantic rivers and Chesapeake Bay tribu- taries while the catch is primarily composed of age IV to VII fish in middle Atlantic rivers. Delaware River runs in the 1960's were primarily supported by age IV and V fish, although age VI females were also important. Sykes and Lehman (1957) reported similar findings for fish collected in the 1940's and 1952 except that they gave more weight to age VI fish, possibly due to sampling with com- mercial gill nets. Because of the absence of repeat spawners, which tend to be older fish, few age-groups support American shad runs in the Delaware River. This is similar to the situation in South Atlantic rivers but unlike that in North Atlantic rivers. The apparent similarity of Delaware River American shad to those of southern rivers rather than to geographically more closely related fish from northern rivers is an artifact caused by man. Howell (1837) reported that American shad cap- tured in the Delaware River averaged about 3,175 g (7.00 pounds). In contrast, Chittenden (1969) reported that males averaged 1,107 g (2.44 pounds) with 95% confidence limits about the mean being 1,080-1,134 g (2.38-2.50 pounds) and females averaged 1,737 g (3.83 pounds) with the 95% con- fidence limits being 1,701-1,774 g (3.75-3.91 pounds). The older age-classes obviously present in Howell's time are now absent, primarily due to pollution and fishing activities. Sykes and Lehman (1957) and Chittenden (1969) attributed the virtual absence of repeat spawners to mortality of adults in low oxygen water near Philadelphia as they 492 CHITTENDEN: DYNAMICS OF AMERICAN SHAD RUNS migrate seaward after spawning. Added to this is a large postspawning mortality in nontidal water (Chittenden 1969). Fishing, in general, decreases the age of the stock exploited. The historical and recent effect of fishing on Delaware River stocks is not completely clear, but this was probably a much more important factor before 1910 when commercial landings were as high as 14 to 17 million pounds annually (Sykes and Lehman 1957; Chittenden 1974). White et al. (1969) tagging studies suggested that in recent years the fishing rate was probably 20%. The larger runs in the Delaware River in the early 1960's appear to have been primarily based upon one large year class except in 1963, when two large year classes were involved. Because few age-classes and only one year class support the run each year, Delaware River American shad stocks have little buffering against fluctuations in abun- dance due to adverse natural or man-made en- vironmental factors. Large fluctuations, in fact, do appear in the catch records of the Lewis Fishery since 1925 (Chittenden 1974, fig. 3). The Lewis Fishery records show large runs over a 1- or 2-yr period intermixed with stable periods when the run was of small magnitude. This is the pattern which would be expected when the fishery is sup- ported by one year class in which males tend to enter the fishery a year before the females and there are essentially no repeat spawners. Therefore, it appears probable that sine 1925, at least, larger runs in the Delaware River have been based upon one large year class and essentially no repeat spawners. ACKNOWLEDGMENTS For assistance in the field collections, I am deeply grateful to James Westman, James Hoff, John Harakal, Don Riemer, James Barker, Frank Bolton, Richard Coluntuno, Kenneth Compton, Richard Gross, Charles Masser, Robert Stewart, John Miletich, Sherman Hoyt, Leonard Schulman, John Musick, Michael Bender, James Gift, Charles Townsend, Ronald Bogaczk, and Kenneth Mar- cellus of or formerly of Rutgers University, or New Jersey Division of Fish and Game. Fred and William Lewis, Jr. generously gave permission to collect fish at the site of their fishery at Lambertville and frequently provided assist- ance in seining. Jess Malcolm, formerly of the U.S. Bureau of Sport Fisheries and Wildlife and of the Delaware River Basin Commission, provided information on the composition of the 1961 run. The U.S. Bureau of Sport Fisheries and Wildlife, New Jersey Division of Fish and Game, Pennsylvania Fish Commission, and New York Department of En- vironmental Conservation kindly permitted use of data collected during the Tri-state Surveys of the Delaware River. Financial support was provided, in part, by the Sport Fishing Institute, Delaware River Basin Commission, and U.S. Public Health Service. LITERATURE CITED Gating, J. P. 1953. Determining age of Atlantic shad from their scales. U.S. Fish Wildl. Serv., Fish. Bull. 54:187-199. Chittenden, M. E., Jr. 1969. Life history and ecology of the American shad, Alosa sapidissima, in the Delaware River. Ph.D. Thesis, Rutgers Univ., New Brunswick, N.J., 458 p. 1974. Trends in the abundance of American shad, Alosa sapidissima, in the Delaware River Basin. Chesapeake Sci. 15:96-103. Cochran, W. G. 1953. Sampling techniques. John Wiley & Sons, Inc., N.Y., 330 p. Ellis, M. M., B. A. Westfall, D. K. Meyer, and W. S. Platner. 1947. Water quality studies of the Delaware River with reference to shad migration. U.S. Fish Wildl. Serv., Spec. Sci. Rep. 38, 19 p. Fredin, R. A. 1954. Causes of fluctuations in abundance of Connecticut River shad. U.S. Fish Wildl. Serv., Fish. Bull. 54:247-259. Godwin, W. F. 1968. The shad fishery of the Altamaha River, Georgia. Ga. Game Fish Comm., Mar. Fish. Div., Contrib. Ser. 8, 39 p. Godwin, W. F., and L. G. McBay. 1967. Preliminary studies of the shad fishery of the Al- tamaha River, Georgia. Ga. Game Fish Comm., Mar. Fish. Div., Contrib. Ser. 2, 24 p. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of Chesapeake Bay. U.S. Bur. Fish., Bull. 43(l):l-366. Howell, S. 1837. Notice of the shad and shad fisheries of the river Delaware. Am. J. Sci. Arts 32:134-140. Judy, M. H. 1961. Validity of age determination from scales of marked American shad. U.S. Fish Wildl. Serv., Fish. Bull. 61:161-170. La Pointe, D. F. 1958. Age and growth of the American shad, from three Atlantic coast rivers. Trans. Am. Fish. Soc. 87:139-150. Leach, G. C. 1925. Artificial progagation of shad. Rep. U.S. Comm. Fish. 1924. Append. 8:459-486. (Doc. 981.) Leggett, W. C. 1969. A study of the reproductive potential of the American shad (Alosa sapidissima) in the Connecticut River, and of the possible effects of natural or man induced changes in 493 FISHERY BULLETIN: VOL. 73, NO. 3 the population structure of the species on its reproductive success. Conn. Res. Comm., Essex Mar. Lab., Essex, Conn., 72 p. Moss, D. D. 1946. Preliminary studies of the shad (Alosa sapidissima) catch in the lower Connecticut River, 1944. North Am. Wildl. Conf. Trans. 11:230-239. Nichols, P. R., and W. H. Massmann. 1963. Abundance, age, and fecundity of shad, York River, Va., 1953-59. U.S. Fish Wildl. Serv., Fish. Bull. 63:179-187. Nichols, P. R., and M. E. Tagatz. 1960. Creel census Connecticut River shad sport fishery, 1957-58, and estimate of catch, 1941-56. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 351, 12 p. Prince, E. E. 1907. The eggs and early life-history of the herring, gaspereau, shad and other clupeoids. Contrib. Can. Biol. 1902-1905:95-110. Smith, H. M. 1898. Report of the division of statistics and methods of the fisheries. U.S. Comm. Fish Fish., Part 23, Rep. Comm. 1897:CXXIV-CXLVI. Stevenson, C.H. 1899. The shad fisheries of the Atlantic coast of the United States. U.S. Comm. Fish Fish., Part 24, Rep. Comm. 1898:101-269. Sykes, J. E. 1956. Shad fishery of the Ogeechee River, Georgia, in 1954. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 191, lip. Sykes, J. E., and B. A. Lehman. 1957. Past and present Delaware River shad fishery and considerations for its future. U.S. Fish Wildl. Serv., Res. Rep. 46, 25 p. Talbot, G. B. 1954. Factors associated with fluctuations in abundance of Hudson River shad. U.S. Fish Wildl. Serv., Fish. BulL 56:373-413. Walburg, C. H. 1956. Commercial and sport shad fisheries of the Edisto River South Carolina, 1955. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 187, 9 p. 1957. Neuse River shad investigations, 1953. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 206, 13 p. 1960. Abundance and life history of shad, St. Johns River, Florida. U.S. Fish Wildl. Serv., Fish. Bull. 60:487-501. 1961. Natural mortality of American shad. Trans. Am. Fish. Soc. 90:228-230. Walburg, C. H., and P. R. Nichols. 1967. Biology and management of the American shad and status of the fisheries, Atlantic Coast of the United States, 1960. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 550, 105 p. Walburg, C. H., and J. E. Sykes. 1957. Shad fishery of Chesapeake Bay with special emphasis on the fishery of Virginia. U.S. Fish. Wildl. Serv., Res. Rep. 48, 26 p. White, R. L., J. T. Lane, and P. E. Hamer. 1969. Population and migration study of major anadromous fish. N.J. Dep. Conserv. Econ. Dev., Div. Fish Game, Bur. Fish., Misc. Rep. 3M, 6 p. 494 SELECTIVE AND UNSELECTIVE EXPLOITATION OF EXPERIMENTAL POPULATIONS OF TILAPIA MOSSAMBICA Ralph P. Silliman' ABSTRACT Two populations of Tilapia mossambica were grown under controlled conditions. After a period of growth and stabilization at about 10 kg and 200 fish, exploitation was started; about 50 fish of outside stock were added to each population to increase genetic variability. Initial exploitation at 10% (later 20%) per 2 mo encompassed all sizes above fry in the unselectively fished population. In the selectively fished population, exploitation was practiced only on fish that could not pass through 25-mm Oater 22-mm) vertical slots between glass bars. Recruitment was estimated from data of stock, mortality, and catch. Parabolas fitted to the stock- recruitment relation suggested greater recruitment in the selectively fished stock than in the unselec- tively fished one. Rectilinear thickness-length regressions were calculated for immature and male fish and separately for females. The exploitation-yield relation was assessed by fitting Fox surplus-yield models to both populations. These revealed a greater maximum sustainable yield in weight from the unselectively fished population than from the selectively fished one. Efficiency of food conversion was 29-36%. To test for genetic effect of selection, a group of 46 fish, matched as closely as possible in size and sex composition, was selected from each population. Growth in length over a period of 150 days was significantly greater among males from the unselectively fished population than among those from the selectively fished one. Growth in length of females was practically identical for both groups. Growth in total weight was distinctly greater for the group from the unselectively fished population than from the selectively fished one. As applied to commercial fisheries, the experimental results suggested fishing as wide a range of sizes as possible. If economic gains from selection are indicated, they should be balanced against possible costs in reduced yield and retarded growth rate. Controlled selective breeding for desirable at- tributes in plants and animals is a well-recognized technique in agriculture. This technique has also had limited application in fish culture, particularly with trout. Claimed achievements have included increased size and earlier age at maturity. Fishery biologists have speculated whether the reverse process, attainment of undesirable attributes, may have occurred in some fished populations because of inadvertent imposition of selection by the fishery. Although gill nets and trawls are perhaps the most obvious gear elements causing selection, the phenomenon is probably present to some extent with practically all fishing gears. It thus becomes a matter of considerable economic importance to determine if gear selectivity has adversely affected fish stocks. The general subject of selection for slow growth by fishing was briefly reviewed by Miller (1957). He adduced no data, however, and drew no firm 'Northwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA, Seattle, WA 98112; present address: 4135 Baker NW, Seattle, WA 98107. Manuscript accepted December 1974. FISHERif BULLETIN: VOL. 73, NO. 3, 1975. conclusions, merely noting that he knew of no in- stances where changed growth rates in fish could not be attributed to some cause other than genetic change. It seems entirely possible, nevertheless, that such a change could occur, if selection were of sufficient strength and continued during a sufficient number of generations. Such a pos- sibility is indicated by the success of artificial selection in altering quantitative characters in a wide variety of organisms. The purpose of the work reported herein was to test experimentally whether selective fishing could produce a genetic change in the growth pat- tern of the fish in the fished stock. This problem' was approached by growing two populations of Mozambique tilapia, Tilapia mossambica, under as nearly identical conditions as possible. One of these was then fished selectively from only those fish above a fixed thickness. The other was fished over the entire range of sizes, except the small "fry." A secondary purpose was to compare amount and size composition of the yields under selective 495 FISHERY BULLETIN: VOL. 73, NO. 3 and unselective fishing. To achieve this com- parison, records of weight and size composition were kept for each catch made during the experiments. MATERIALS AND METHODS For the experimental animal, Tilapia mossam- bica was chosen. This species is hardy and will grow readily in experimental tanks. It also is used widely in tropical pond culture and thus has some economic importance. Since it is a mouth breeder, handling and exploitation were done only at approximately 2-mo intervals. Tanks, feeding, etc. were as reported in Silliman (1970) and represented a modification of the methods of Uchida and King (1962). Briefly sum- marized, the procedures were to raise the two populations in hatchery-type troughs of 850-liter (225-gallon) capacity. Water condition was main- tained either by changing it biweekly or by a con- tinuous dribbling flow into the head of each trough plus bimonthly partial changing. Temperature was maintained at 80° ± 5°F or 26.7°C (weekly means). Illumination was by fluorescent light 12 h per day. Feeding schedules and water condition are detailed in Tables 1 and 2. Rectangular enclosures at the standpipe ends were separated from the rest of the troughs by plates with 3-mm holes through which the newly expectorated "fry" could escape, thus furnishing them refuges from cannibalism by the adults. After each counting, all fish in the refuges were placed in the main part of the tanks. Fishing was done at approximately 2-mo inter- vals by removing each nth fish for fishing rate 1/n {n was always an integer). For the selectively fished population, all fish were placed on one side of a grid consisting of 25-mm diameter vertical glass rods spaced 25 mm (22 mm in latter part of experiment) apart. All fish were provided an op- portunity to swim through the spaces between the rods; only those which could not do so were fished. In the unselectively fished population all sizes were fished except the fry (under 4-mm thick- ness). Counting was done simply by netting fish from one container to another. For weighing, fish were drained in a net and then placed in a previously weighed container of water. Fish weight was ob- tained by subtracting the tare from the total. All caught fish and the preexploitation stocks were measured for thickness and length. They were categorized as immature (where sex could not be determined by external inspection), male, and female. Sex determination was based on the characteristics set forth by St. Amant (1966). Length (total length to outermost tip of caudal fin) was measured on a board with millimeter scale and head block. Thickness was measured on the same device, plus a sliding block; the fish were held upright between the sliding block and the head block with firm pressure for the thickness measurement. Fish for the pretest measurements Table 1.— Food placed in tanks, grams. Trout pellets Tropica 1 fish food Date' Month Day of week Moist Dry A2 B2 Total 15 Aug. 1966 0.5-40.2 Sun. 30 10 4 5 49 to Mon. 40 10 4 10 64 6 Dec. 1969 Tues. 40 10 4 10 64 Wed. 40 10 4 10 64 Thurs. 40 10 4 10 64 Fri. 40 10 4 10 64 Sat. 40 10 4 10 64 Total 270 70 28 65 433 7 Dec. 1969 40.2-82.2 Sun. 30 10 5 49 to Mon. 40 10 10 64 5 June 1973 Tues. 40 10 10 64 Wed. 40 10 10 64 Thurs. 40 10 10 64 Fri. A.M. 40 10 10 64 Fri. P.M.3 40 10 10 64 Total 270 70 28 65 433 'Diet was varied Initially to achieve optimal reproduction and growth; it was stabilized at the listed amounts on 18 June 1967, month 10.6. ^Commercal makes of dry food. 'This feeding was combined with the Friday A.M. feeding in 37 out of 183 wk. and with the Sun- day feeding once. 496 SILLIMAN: EXPERIMENTAL POPULATIONS OF TILAPIA MOSSAMBICA Table 2.-Water condition on selected dates.' O, (ppm.) CO, (ppm.) pH Date Month Test2 Control 3 Testz Control Test^ ControP 1968: 9 Aug. 16 29 6 13 20 27 4 11 18 25 8 15 28 6 13 26 1969: 2 Jan. 17 Sept. 2 Oct. 9 23 30 1971: 25-26 Feb. 1972: 23 Mar. 1973: 29 May Sept. Oct. Nov. Dec. 24.3 24.5 24.9 25.2 25.4 25.7 25.9 26.1 26.4 26.6 26.8 27.3 27.5 27.9 28.2 28.4 28.8 29.1 37.6 38.1 38.3 38.7 39.0 54.9 67.7 81.9 5.2 4.6 4.8 3.6 3.6 4.0 4.0 4.0 3.8 3.0 3.0 3.6 3.4 4.2 3.8 3.4 3.0 3.4 5.2 5.4 5.0 5.0 5.4 4.6 3.4 3.4 5.0 4.4 4.8 3.8 3.6 4.4 4.0 4.4 4.0 3.8 4.0 4.0 4.0 4.4 4.0 4.0 3.4 3.6 4.8 5.4 4.8 4.8 5.0 4.4 "2.6 3.2 10 10 10 6.9 6.9 10 — 10 10 6.5 6.7 20 20 15 20 25 15 6.0 6.5 7.0 6.0 6.5 7.0 'With the exception noted in footnote 4, all values were within (or above for oxygen) the ranges stated to be suitable for warm- water fishes by Lewis (1963). These were: oxygen, 3-5 ppm.; car- bon dioxide, below 30 ppm.; pH, 5-9. ^Selectively fished. ^Unselectively fished. ■•Aeration was increased and O^ had risen to 3.8 ppm. by the next day. were anesthetized with MS-222 ^ (tricaine methanesulfonate) in a 1:2,500 solution. Caught fish were measured some time after removal. They were usually alive or freshly dead, and rigor mortis was rare. The group selected for growth study at the end of the experiment was measured alive without an anesthetic. COURSE OF POPULATIONS A single population was started on 15 August 1966, but this was divided into two populations as nearly equal as possible after 2 mo. A period of population growth then ensued (Table 3, Figures 1, 2). This growth was extensively discussed by Silliman (1970), who found growth in biomass of the two populations to be practically identical. He therefore combined the two populations for growth analysis. A Gompertz curve fitted to the total biomass of both populations had the formula ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Wi = 1.337 exp[2.85 - 2.85exp(-0.2 (^-3.6) )], where Wf is biomass in kilograms at the time t in months. The asymptote of this curve was 23.1 kg or 11.55 kg per population. An accidental interruption to population growth (Figures 1, 2) resulted from temporary relocation of the fish during refinishing of the laboratory floor. After recovery and re- equalization, the populations did not approach the asymptote predicted by the Gompertz curve but leveled off at about 10 kg each. Because the fish were descendants of an original three males and three females, I felt that in- 400 300 200 100- O a: 2 300- 200- 100- Control 10 20 30 40 50 60 MONTH 70 80 Figure l.-Population size and catch, numbers. Percentages in- dicate target exploitation rates. Test population was selectively fished; control, unselectively. Figure 2.-PopuIation size and catch, weight. Percentages in- dicate target exploitation rates. Test population was selectively fished; control, unselectively. 497 FISHERY BULLETIN: VOL. 73, NO. 3 Table 3.— Population and catch, Tilapia mossambica. Test (selectively fishied) population Control (unselective My fished) population Target expl. rate^ Num iber Weight (g) Nun iber Weight (g) Monthi Stock Catch Stock Catch Stock Catch Stock Catch 0.5 0.00 6 2.5 272 — — — 273 — — — 3.6 258 — 856 — 253 — 825 — 4.5 220 — 1,318 — 251 — 1,325 — 5.5 231 — 1,924 — 239 — 1,447 — 7.5 228 — 3,306 — 247 — 3,303 — 10.5 228 — 5,614 — 246 — 5,749 — 13.0 224 — 7,420 — 232 — 7,313 — 15.0 239 — 8,275 — 253 — 8,692 — 17.1 228 — 9,346 — 237 — 9,448 — 19.2 226 — 10,174 — 232 — 10,244 — 21.1 224 — 10,459 — 251 — 10,907 — 23.2 3221 — 39,859 — 3226 — 39,826 — 25.4 3156 — 38,567 — 3186 — 38,561 — 27.2 181 — 9,560 — 184 — 9,431 — 29.1 180 — 10,415 — 183 — 9,967 — 31.3 189 — 10,525 — 189 — 9,791 — 33.1 181 — 10,348 — 177 — 9,622 — 35.3 177 — 9,881 — 173 — 9,661 — 39.2 0.10 218 19 10,760 2,023 183 18 10,009 1,168 41.3 201 18 9,468 1,714 174 17 10,139 1,033 43.2 187 8 8,984 906 187 18 9,743 974 45.2 187 8 8,311 794 199 19 9,300 956 47.3 ••240 8 ^9,534 883 ■•217 22 "9,917 909 48.9 231 2 8,860 182 191 4 9,843 191 51.2 233 14 9,690 960 186 18 10,368 1,057 53.2 217 14 9,575 991 163 16 9,532 927 55.1 207 14 9,037 929 165 16 9,315 951 57.2 221 12 8,542 848 148 15 8,956 895 59.0 210 10 8,056 806 131 13 8,507 827 61.1 196 10 7,935 853 117 12 8,287 824 63.2 0.20 191 17 7,687 1,575 177 35 8,322 1,681 65.4 172 17 6,944 1,377 124 25 7,965 1,531 67.1 157 11 6,382 738 114 12 7,261 840 69.1 345 15 6,684 1,050 176 26 7,592 1,217 71.2 388 14 5,465 1,061 155 31 7,664 1,574 73.2 389 14 7,046 1,170 123 21 7,389 1,272 75.1 371 6 6,901 438 101 6 7,280 459 10 = 1 August 1966. ^Because of selection problems, actual rates varied considerably from these. In analyses, effective rates in terms of weight were used. 3Populations re-equalized after accidental mortality from temporary relocation of fish. "New fish added for genetic variability. sufficient genetic variability might exist in the populations to permit a genetic effect of selection. Since one of the major objectives of the experiment was to detect such an effect, I decided to add outside stock. During the 2-mo period preceding month 47.3 (Table 3), I added 45-47 (2 fish uncertainty due to counting difficulties) im- mature Tilapia mossambica from Arizona to each population. These fish were descendants of ones from Malacca. Exploitation was started at month 39.2 (Table 3, Figures 1, 2) at a conservative 10% per 2 mo. This exploitation period included 1.0 to 2.6 of the brood intervals reported by various authors (Kelly 1957, 30-40 days; Swingle 1960, 30-40 days; Uchida and King 1962, 23-61 days). Because of irregularities in recruitment, population numbers (Figure 1) reflected population responses less well than population biomasses (Figure 2). The latter, however, generally reflected the expected popula- tion decrease from imposition of the 10% exploi- tation rate. When the rate was increased to 20% at month 63.2 (Table 3, Figure 2), further declines in population biomasses occurred. BASIC RELATIONS Recruitment was estimated from changes (/?int) in stock number and data of mortality and catch (Table 4), using the approach of Silliman (1972). Because of variations in length of period between counts, values of R^^jwere adjusted to a standard 2-mo interval (Table 4). Observation of large numbers of fry in the refuges, followed by the later appearance of peaks of recruitment (Table 4), indicated that the "reproductive lag" was about 2 498 SILLIMAN: EXPERIMENTAL POPULATIONS OF TILAPIA MOSSAMBICA Table 4.-Recruitment and stock. i? INT = Pn + l-Pn + Mint + Cint, where Pis stock, INT is interval between counts n and n+l,Ris net changei , M is recorded mortality and C is catch, all in numbers. S is mean stock in kilograms for previous interval. Interval Interval length (months) Selective y fished population Unselective y fished population (months) "iNT ^'fl,NT S '^INT '«INT S 5.5- 7.5 2.0 -3 -3.0 1.4 + 9 + 9.0 1.1 7.5-10.5 3.0 -HI +0.7 2.6 + 1 +0.7 2.4 10.5-13.0 2.5 + 4 + 3.2 4.4 -13 -10.4 4.5 13.0-15.0 2.0 -1-15 + 15.0 6.5 + 21 + 21.0 6.5 15.0-17.1 2.1 -5 -4.8 7.8 -5 -4.8 8.0 17.1-19.2 2.1 -2 -1.9 8.8 -5 -4.8 9.0 19.2-21.1 1.9 0 .0 9.8 + 19 + 20.0 9.8 25.4-27.2 1.8 -t-25 + 27.8 3 -1 -1.1 3 27.2-29.1 1.9 -1 -1.1 9.1 -1 -1.1 9.0 29.1-31.3 2.2 -1-9 + 8.2 10.0 + 6 + 5.5 9.7 31.3-33.1 1.8 -6 -6.7 10.4 -12 -13.3 9.9 33.1-35.3 2.2 -2 -1.8 10.4 -2 -1.8 9.7 35.3-39.2 3.9 -1-44 + 22.6 10.1 + 16 + 8.2 9.6 39.2-41.3 2.1 + 7 + 6.7 10.8 + 10 + 9.5 9.8 41.3-43.2 1.9 +4 + 4.2 9.1 + 30 + 31.6 9.4 43.2-45.2 2.0 + 8 +8.0 8.4 + 30 + 30.0 9.4 45.2-47.3 2,1 -i+ie + 15.2 8.2 ■*-6 -5.7 9.0 47.3-48.9 1.6 + 2 + 2.5 8.5 0 .0 9.1 48.9-51.2 2.3 + 4 + 3.5 8.8 -1 -0.9 9.4 51.2-53.2 2.0 -1 -1.0 9.2 -3 -3.0 10.0 53.2-55.1 1.9 -1-5 + 5.3 9.2 + 18 + 18.9 9.4 55.1-57.2 2.1 -1-31 + 29.5 8.8 0 .0 9.0 57.2-59.0 1.8 +3 + 3.3 8.3 -1 -1.1 8.7 59.0-61.1 2.1 -2 -1.9 7.9 + 2 + 1.9 8.3 61.1-63.2 2.1 -t-5 + 4.8 7.6 + 73 + 69.5 8.0 63.2-65.4 2.2 -1 -0.9 7.4 -16 -14.5 7.9 65.4-67.1 1.7 + 3 + 3.5 6.5 + 15 + 17.6 7.3 67.1-69.1 2.0 + 199 + 199.0 6.0 + 75 + 75.0 6.8 69.1-71.2 2.1 -H62 + 59.0 6.2 + 6 +5.7 7.0 71.2-73.2 2.0 + 16 + 16.0 5.6 -1 -1.0 7.0 73.2-75.1 1.9 -2 -2.1 5.7 + 1 + 1.1 6.8 'Rf^f > 0 indicates recruitment of at least the indicated number of fish; R^^j < 0 indicates unrecorded mortality of at least the indicated number and R^f^-^ = 0 indicates either no recruit- ment and unrecorded mortality or the two exactly balanced. ^Adjusted to a standard 2-mo interval length. ^Indicated intervals omitted because of re-equalization of stocks. ''Exclusive of 46 new fish added for genetic variability. mo. Each value of i?iNTwas therefore compared with the mean stock (S) for the preceding 2-mo interval (Table 4). The stock-recruitment data were highly varia- ble and were, therefore, treated as group means based on 5-kg intervals of S, considering negative values of Ri^^ to be equal to zero. Although the data indicated no regular relation (Figure 3), they were fitted with parabolas to indicate central ten- dency, even though fits were poor. These, based on 30 pairs of observations each, were: Selectively fished stock R^+i = 6.163 S^ - 0.5209 S/ Unselectively fished stock Rn+i = 3.304 S^ - 0.2158 S^f, where N is number of the 2-mo interval, R is in -r 12 4 6 8 10 MEAN STOCK, INTERVAL N (kg) 14 16 Figure 3.-Stock and recruitment. Test popuhition was selec- tively fished; control, unselectively. Parabolas shown were fitted by least squares. 499 FISHERY BULLETIN: VOL. 73, NO. 3 numbers, and 5 is in kilograms. The somewhat higher maximum for the selectively fished stock may have resulted from its changed size composi- tion; it included fewer extremely large males than the unselectively fished population. Since selection reflected the ability of fish to escape through vertical slots between glass bars, thickness was the controlling dimension. Most of the fish-size data in this report are therefore in thickness. However, to reduce fish handling to a minimum, the growth measurements of live fish during the final part of the experiment were in lengths. Because of this, and because other biologists may wish to compare their length data with my thickness data, I calculated thickness- length relations. Measurements for the relation were from the caught fish, for which both thickness and length were recorded. Preliminary analysis showed that data for immature and male fish could be combined into a single rectilinear regression of length on thickness. The regression for females was also rectilinear but had a gentler slope, probably because of the distention of mature fish carrying eggs. It was therefore calculated separately. The regression equations, numbers of pairs of obser- vations in parentheses, and correlation coefficients were (fitting was by least squares): Immature and male L = 7.027 T (368) r = 0.987 r^ = 0.974. Female L = 26.89 -I- 5.037 T (207) r = 0.866 r^ = 0.750, where L is length and T is thickness, both in millimeters. The squared coefficients suggest that 97 and 75% of variations in length were associated with variations in thickness. RESULTS Exploitation and Response Before selective exploitation could be started, it was necessary to determine the selection point. To aid in this the thickness of all of the fish in both populations was measured at month 33.5. At this time the population to be selectively fished (pre- test) consisted of 85 males and 95 females-that to 25- Pret est i 20- 1 (1 l\ 1 ' /, 1 15- 10- 1 5- A A. / 1 / X to 0- C^ VVV^ ..N \ li- 1 1 1 1 1 ^ T 30-1 Female- 15 20 THICKNESS (mm) T 25 30 35 Figure 4.-Thickness frequencies at month 33.5. Test population was to be selectively fished; control, unselectively. be unselectively fished (precontrol), 77 and 98, respectively. The thickness frequencies (Figure 4) revealed a low point in the pretest population between males and females at about 25 mm. This was used as the initial selection point and it meant, of course, that the early catches were mostly males. It will be shown below, however, that later catches were composed of roughly equal propor- tions of males and females. Changes from exploitation, in addition to those described under "Course of Populations," were reflected in the size composition of the catches (Figures 5, 6). In the test population, the catches were roughly the target percentage of the selected group; the percentages in the control population were adjusted to take the same proportion of the entire population as taken in the test population. Percentages were by number at months 39.2 and 41.3, but it became evident that this procedure took too large a proportion of the biomass. At months 43.2-75.1, the percentages were by weight 500 SILLIMAN: EXPERIMENTAL POPULATIONS OF TILAPIA MOSSAMBICA 10- 5- T. mossambico 39.2 ,test ■/v - 59.0 1 5- 41.3 > A - 61.1 1 5- n 43.2 ^-u 63.2 1 5- 45.2 ^^N "■^ A . I t 0- 47.3 /Xyw^ "■' u O ^ 0= 5- UJ 3 m 2 0- 48.9 y-^ 69.1 1 A A =3 " z 5- 51.2 Ik - 71.2 1 5- 0- 53.2 '/\ 73.2 1 5- 0- 55.1 [^ ■ 75.1 1 iA 5- 0- 57.2 k 10 20 30 40 THICKNESS (mm) 10 20 30 THICKNESS (mm) 40 Figure 5.-Bimonthly thickness frequencies of catches, test (selectively fished) population. Vertical broken lines indicate selection points. Numbers in panels indicate months from start. (Table 3). Numbers of fish growing above the selection point rapidly diminished in the test population (Figure 5) so that insufficient numbers were available from which to catch the target percentage. To continue exploitation, it was necessary to lower the selection point to 22 mm as shown. With few exceptions, all fish caught from the test population were above the selection point. Catches from the control population were taken representatively over all sizes larger than fry and, therefore, represented the size composition of the stock above the fry size (Figure 6). Significant amounts of recruitment at months 43.2, 45.2, 55.1, 63.2, 67.1, and 69.1 (Table 4) appear as modes of small fish in the size frequencies, and the more prominent ones can be traced through succeeding frequencies. A summary of catch size frequencies (Figure 7) clearly reveals the differences between catches from test and control populations. It is evident that the selection device employed was almost completely effective. The appearance of roughly equal proportions of males and females in the test catches, after lowering the selection point, is also apparent. It can be seen that a selection curve was at work, such that fish were not fully retained until they had reached a thickness of about 2-4 mm above the selection point. The relation of yield to exploitation was as- sessed by fitting a Fox (1970) exponential surplus-yield model to data of catches and stock (Figure 8). The method of fitting used requires equilibrium yields. Although absolute equilibrium obviously was not attained, it was considered that the biomass and catch levels (Figure 2) at months 29-35 (zero exploitation), 59-61 (10% target rate), and 69-73 (20% target rate) represented sufficiently close approximations to equilibrium for fitting the model. The calculated maximum sustainable yield (1.39 kg per 2 mo) from the selectively fished test population was substantially lower than that for the unselectively fished control (2.36 kg). If we wish to consider a comparable commercial fishery, however, we might assume that only the fish above the selection point are commercially desirable. Catch thickness frequen- cies for the 22-mm selection point (Figure 7) showed that 97% of the fish in test catches were above the selection point as compared with 40% for the control catches. Although these data cannot be T. mossombico, control 10- 5- in g: 5- Ll_ o 0- 39.2 ^ /XnA 41.3 A >v /^ /U 43.2 45.2 AVsA AA 47.3 48.9 51.2 53.2 A^ A'^^^y^ 55.1 _cv_ Aa^^^^^s^ 59.0 61.1 A^ A A KL ^^AA\ aAA 6 5.4 JS^ 67.1 ^A^/^y\/\ y<. ,^_ 69.1 A. Ar^ A_ 71.2 . Aa.^ /VA ^^ 73.2 A^^V/^ /\ 75.1 57.2 A^ ^ .^A/v ^ y\/\ /S. —I — " — "-r" — — r— ■-■ 1 — 10 20 30 40 THICKNESS (mm) 0 10 20 30 40 THICKNESS(mm) Figure 6.— Bimonthly thickness frequencies of catches, control (unselectively fished) population. Numbers in panels indicate months from start. 501 FISHERY BULLETIN: VOL. 73. NO. 3 X 15- 10- 5- T. mossambicQ Control, month 39.2-48.9 20- 15- 10- 5- Test ^ 0-f o q: 15-1 UJ m 10- 3 5- z 0- Control, month 51.2-75.1 30- 25- 20- 15- 10- 5- 0-- Test immature- mole- female. 20 30 40 THICKNESS (mm) — I 50 FiGUKB T.-Sammary of catch thickness frequencies. Test population was fished selectively; control, unselectively. Vertical lines indicate selection points. converted to weights, it is certain that the com- parison for fish above the selection point would be more favorable to the test catches than was true for all sizes of fish. It is possible to calculate the efficiency of con- version of food to fish flesh under both types of fishing. The amount fed per 2-mo period was 3.75 kg (433 g per week from Table 1, for 8% wk). Maximum sustainable yields (MSY's) of 1.39 kg and 2.36 kg indicate 37 and 63% conversion efficiencies for selective and unselective fishing, respectively. Since the theoretical MSY's represent a considerable extrapolation (Figure 8), it is of interest to calculate from equilibrium yields actually attained during the experiment. The larg- est yields were under the 20% per 2-mo target rate control -| 1 •8 1.2 1.6 2.0 2.4 EFFECTIVE EFFORT control 12 BIOMASS (kg) FiGUBE 8.-Fitting of Fox (1970) model. Catch-per-unit-effort (CPUE) is considered proportional to biomass; effort is in arbi- trary units. Regression lines shown are least-squares fits. Target exploitation rates were 0, 10%, and 20% per 2 mo, left-to-right in upper panel, reversed in lower panel. for both populations. Effective exploitation rates and corresponding yields were: selectively fished, 17.1% and 1.09 kg; unselectively fished, 17.9% and 1.35 kg. These yields indicate 29 and 36% efficiency, respectively. The values are in fair agreement with the 33% calculated by Silliman (1970) for the initial growth of the populations. Genetic Response Knowledge of the number of generations in- volved is essential to any genetic experiment. Progression of the two most prominent thickness frequency modes (months 63.2 and 69.1) in the un- selectively fished population (Figure 6) gave an indication of the growth rate of the fish. Frequen- 502 SILLIMAN: EXPERIMENTAL POPULATIONS OF TILAPIA MOSSAMBICA cies of thickness for immature and mature fish (Figure 7) suggested a thickness at maturity of about 15-20 mm. The two prominent modes in the frequencies seemed to require about 4 mo (63.2-67.1, 69.1-73.2) to reach this size. To this must be added the 2-mo "reproductive lag" mentioned above under "Basic Relations," suggesting a generation length of about 6 mo. The 36-mo period of selective fishing would thus include about six potential generations. Because of irregularities in recruitment, however, the effective number of generations was less, and it was necessary to make an estimate. Such an estimate can be derived from the record of recruitment numbers (i^iNT* Table 4, Figure 9). An arbitrary criterion for significant recruitment was established, requiring at least 15 recruits per 2 mo. A generation was considered to be such a peak separated from the previous filial generation by a period of at least 6 mo (the parental generation for the test group had been fished selectively at month 40). Under the arbitrary criterion the estimated generations (Figure 9) during the selection period were only three for the selectively fished test and four for the unselectively fished control popula- tion. Experiments with other animals, such as those of Robertson with thorax length of Drosophila cited by Falconer (1960), have shown that measurable change in a size character can occur in as few as three generations of selection. To test whether genetic response to selection did Table 5.-Growth of selected groups of fish. Lengths are from snout to tip of taiL 200 160' 120- 80- 3 o lu (E 40- a: CD '-' 5 80- 40- Test Control J^^ 20 30 40 MONTH 50 80 Figure 9.-Recruitment numbers from ^jnt in Table 4, with negative values considered zero. Test population was selectively fished; control, unselectively. Male Fe male Total Mean Mean length length Group Dayi No. (mm) No. (mm) No. Wt. (g) Test 0 10 152.0 36 140.6 46 2,158 56 29 180.0 2.336 155.0 45 3,199 118 9 197.8 ■•34 165.0 43 3,794 150 9 207.2 333 169.8 42 4,078 Control 0 10 148.0 36 141.4 46 2,561 55 58 195.0 3.537 153.4 45 3,504 119 8 235.6 '35 164.4 43 4,455 151 8 253.1 '34 169.1 42 4,819 '0^5 January 1973. 20ne female misclassified as male on day 0. 30ne female died. ■•Two females died. 5Two females misclassified as males on day 0. 'Two females removed to match mortalities in test group. 'One female removed to match mortality in test group. occur, groups of 46 mature fish as similar in sex and length composition as possible were selected from test and control populations on 5 January 1973 at month 77.2 (Table 5). It was not possible to match these fish as closely as desired by total weight, and that of the control group exceeded that of the test group by 19%. These fish were fed the standard diet (Table 1) which furnished them, even at the end of the growth period, with 1.5% (test) or 1.3% (control) of body weight of food per day. This was 2.5 (test) or 2.1 (control) times as much as was received by the 10 kg preexploitation stocks. Any offspring that appeared were removed as soon as possible. Growth of the selected fish was measured by determining the lengths of individual fish and the total weight of each group at 55-56, 118-119, and 150-151 days after the start of the growth period (Table 5, Figures 10, 11). The length frequencies reveal the general correspondence of the groups at the beginning of the period, in addition to the expected more rapid growth of the males than the females in each group. They also reveal that the males in the unselectively fished control group grew more rapidly than those in the selectively fished test group. Growth was further studied by curv.es based on mean lengths and total weights of the selected groups. Gompertz curves fitted to lengths had the equation: Lt- = Lo exp[G - G exp{-gt')l where L is mean length in millimeters, t' is time in days, and G and g are empirical constants. This curve and all other Gompertz curves were fitted by 503 FISHERY BULLETIN: VOL. 73, NO. 3 1-5-73 1-5-73 Male - Female- 100 120 140 160 180 TOTAL LENGTH, mm T 1 200 220 240 Figure lO.-Length frequencies of group selected from selec- tively fished test population. Lengths are from snout to tip of tail. the analog computer method of Silliman (1967). Growth in length was essentially identical in the two groups for the females (Table 5, Figure 12), and a single curve was fitted. Constants are given in Table 6. For males, however, growth was sig- nificantly greater in the unselectively fished con- trol group than in the selectively fished test group. The sexual misclassification of one fish in the test group and two in the control group (Table 5) must be considered in relation to possible effects on the results. These fish were misclassified at the beginning of the growth period, when the fish were relatively small (chosen so to provide room for growth) and sex determination was difficult. As the fish grew and determination became easier, the errors were discovered and corrected. To test the effect of the errors it was assumed that they occurred in the manner most contrary to the conclusion adopted-that growth was greater among males in the control than in the test group. TOTAL LENGTH (mm) Figure ll.-Length frequencies of group selected from unselec- tively fished control population. Lengths are from snout to tip of tail. 260- 240 220- 200 180 P 140 Figure 12.-Gompertz curves fitted to mean lengths in group selected from selectively fished test population and unselectively fished control population. 504 SILLIMAN: EXPERIMENTAL POPULATIONS OF TILAPIA MOSSAMBICA Table 6.-Constants of Gompertz curves for growth of selected groups of fish. Popu- '■o Loo W„ Wcr^ Sex Variable lation (mm) (mm) (kg) (kg) 9 G d" Length Test 148 214 — 0.0150 0.367 Control 146 274 — — 0.0127 0.630 V Length Both 141 174 — — 0.0128 0.211 Both Weight Test — — 2.23 4.45 0.0130 0.692 Control — — 2.49 5.65 0.0105 0.819 Note: Lg = Length at time zero. Loo^ Asymptotic limit of length. W„ Weight at time zero. Wqo^ Asymptotic limit of weight. g and G ^ Empirical constants of the Gompertz curve. Thus it was assumed that at day 0 one of the two males at maximum length in the test group was misclassified as a female and, similarly, for the two smallest males in the control group. Resulting mean lengths in millimeters comparable to those for day 0 in Table 5 are: test male, 149.4; test female, 141.5; control male, 151.2; control female, 141.1. Percentage differences are 1.7, 0.6, 2.2, and 0.2, respectively. The means under the "most con- trary assumption" are indistinguishable from the values used on the scale of Figure 12. It is evident that substitution of the most contrary values would not change the conclusion of greater male growth in the controls. Gompertz curves were also fitted to biomasses of the two groups (Table 5, Figure 13). Here the equation was: W,= Woexp[G-Gexp(-gt')l where W is total weight in kilograms, t' is time in T mossombico , growth in weight Figure 13. -Gompertz curves fitted to total weights of groups from selectively fished test population and unselectively fished control population. Line branching upward from test curve in- dicates control curve moved over to same starting point as test curve. days, and G and g are empirical constants. Con- stants are given in Table 6. In weighings, fish were not separated as to sex, and only a single growth curve was available for each population. Because of the initial difference in total weight, the curve for the control population is shown moved over to the time when weight of the test group had grown to the initial weight of the control group. Even so treated, the control group exhibits markedly greater growth than the test group, reflecting greater growth of the males in it. This is even more striking when it is considered that the amount of food per weight of fish was somewhat less in the control than in the test group. The ob- served growth in biomass supports the conclusion of diminished genetic growth in males as a result of selective fishing. CONCLUSIONS Analyses presented above have revealed sig- nificant differences of responses to exploitation between the selectively fished test population and the unselectively fished control population. These differences were demonstrated both in catches obtained and in genetic growth patterns. Yield models fitted demonstrated marked differences in the catches obtained under selective and unselective fishing. It is clear that, with the particular populations studied and under the as- sumptions of stability made, weight of yield under unselective fishing was greater than that under selection. This yield included a large proportion of fish below the selection point, however. To the ex- tent that one may generalize from this experiment, it appears that unselective fishing would be preferable if maximum physical yield were the sole objective. If selection is required to secure fish that are of appropriate size for the market, the objective may be achieved only at the sacrifice of part of the weight of the catch. That three generations of selective fishing caused a change in the genetic growth pattern of males, resulting in slower growth than in the con- trols, seems certain from the results. It is neces- sary to explain, however, why a similar change did not occur in the females. This may have resulted from the phenomenon of epistasis. In this it is considered that a single gene may control the hormone which permits males to grow to larger ultimate size than females. Since females have less of this hormone than males, they are unable to 505 FISHERY BULLETIN: VOL. 73, NO. 3 express genotypic differences which otherwise might cause changed growth patterns. In other words, the degree of selection imposed did not work against females because they were unable to achieve extra large size in any event. This hypothesis cannot be tested with data from the present experiment. Fishing in the experiment was similar to that in a commercial fishery, with the vertical slots in the apparatus corresponding to the meshes of com- mercial gear. Results may, therefore, be of some significance in fishery management. They suggest that as wide a range of sizes as possible be included in the catch. An appropriate balance should be struck between the possible higher market value of large fish and the lesser yields that may be achieved under selection. Also, the possibility of a genetic change in growth pattern under selection should not be overlooked. SUMMARY 1. Two populations of Tilapia mossambica were grown with as nearly identical space, water condition, and food as possible. 2. After a period of initial growth each population stabilized at a weight of about 10 kg. Numbers were less stable at this time and ranged from 173 to 218 fish. 3. To increase genetic variability, 45-47 immature fish of Malacca descent were added to each population at month 47.3. 4. Exploitation was started at month 39.2, at 10% per 2 mo (1.0-2.6 brood intervals) and increased to 20% at month 63.2. Selective fishing was from fish which could not pass through 25-mm (later 22-mm) vertical slots between glass rods. Unselective fishing was from all fish except fry (under 4-mm thick- ness). 5. Recruitment was estimated from data of stock number, mortality, and catch. Reproductive lag was 2 mo. The stock-recruitment relations, roughly fitted with parabolas, suggested greater recruitment in the selectively fished stock than in the unselectively fished one. 6. Two rectilinear thickness-length relations were calculated, one for immature and male fish and another for females. 7. Catch thickness frequencies for the unselec- tively fish population revealed modes corre- sponding to peaks of recruitment. 8. Catch thickness frequencies for the selectively fished population, compared with those for the unselectively fished population, demonstrated that the device for selection at 25 and 22 mm was almost completely effective. 9. The exploitation-yield relation was assessed by fitting Fox exponential surplus-yield models to data from both populations. Fitted models indicated a higher maximum sustainable yield in weight for the unselectively fished popula- tion than for the selectively fished one. Eflficiency of food conversion was 29-36%. 10. Growth rates from catch thickness frequen- cies, together with the 2-mo reproductive lag, suggested a generation length of 6 mo. Recruitment records indicated three genera- tions under exploitation for the selectively fished population and four during the same period for the unselectively fished one. 11. To test for genetic effect of selection, a group of 46 fish was selected from each population. These were matched as closely as possible by size and sex composition and grown under previously established standard conditions. 12. Growth in length over a period of 150 days was significantly greater among males from the unselectively fished population than among males from the selectively fished one. Growth for females in the two groups was practically identical. 13. GroMd:h in total weight was distinctly greater for the group from the unselectively fished population than in that from the selectively fished one. 14. It was concluded that these experiments demonstrated diminished total yield and re- tarded male growth in the selectively fished population compared with the unselectively fished one. An hypothesis based on epistasis was advanced to explain lack of growth re- sponse among females. 15. As applied to commercial fisheries, the experimental results suggest fishing as wide a range of sizes as possible. If economic gains from selection are indicated, they should be balanced against possible costs in reduced total yield and retarded growth rate. ACKNOWLEDGMENTS I am grateful for the advice and encouragement rendered by Francis M. Fukuhara, Frederick M. 506 SILLIMAN: EXPERIMENTAL POPULATIONS OF TILAPIA MOSSAMBICA Utter, Harold 0. Hodgins, and Donald D. Worlund, all of the Northwest Fisheries Center, National Marine Fisheries Service, NOAA. The plan of the experiment was reviewed by them and by Joseph Felsenstein of the University of Washington, Seattle, Wash. It is pertinent to note, however, that some changes were made in the plan due to developments during the experiment. My special thanks are due Alban R. Essbach of the Arizona Game and Fish Department. He generously shipped stocks of fish on two different occasions. Experimental populations were maintained during various periods by George F. Slusser, Christopher E. Mathews, Martin G. Beam, Jimmy R. Chrnaoski, and Judy A. Trauth. I also thank George F. Slusser for furnishing the initial stock of fish. All of these persons are present or former members of the National Marine Fisheries Ser- vice, NOAA (formerly Bureau of Commercial Fisheries). LITERATURE CITED Falconer, D. S. 1960. Introduction to quantitative genetics. Ronald Press, N.Y., 365 p. Fox,W.W.,jR. 1970. An exponential surplus-yield model for optimizing exploited fish populations. Trans. Am. Fish. Soc. 99:80-88. Kelly, H. D. 1957. Preliminary studies on Tilapia mossamhica Peters relative to experimental pond culture. Proc. 10th Annu. Conf. Southeast Assoc. Game Fish Comm., p. 139-149. Lewis, W. M. 1963. Maintaining fishes for experimental and instructional purposes. South. 111. Univ. Press, Carbondale, 111., 100 p. Miller, R. B. 1957. Have the genetic patterns of fishes been altered by introductions or by selective fishing? J. Fish. Res. Board Can. 14:797-806. St. Amant, J. A. 1966. Progress report on the culture of Tilapia mossambica (Peters) hybrids in southern California. Resour. Agency Calif. Dep. Fish Game, Inland Fish. Adm. Rep. 66-9, 25 p. SiLLIMAN, R. P. 1967. Analog computer models of fish populations. U.S. Fish Wildl. Serv., Fish. Bull. 66:31-46. 1970. Growth in experimental populations of Tilapia mossambica. BioScience 20:1109-1110. 1972. Effect of crowding on relation between exploitation and yield in Tilapia macrocephala. Fish. Bull., U.S. 70:693-698. Swingle, H. S. 1960. Comparative evaluation of two tilapias as pondfishes in Alabama. Trans. Am. Fish. Soc. 89:142-148. UcHiDA, R. N., AND J. E. King. 1962. Tank culture of Tilapia. U.S. Fish Wildl. Serv., Fish. Bull. 62:21-52. 507 UPTAKE AND LOSS OF PETROLEUM HYDROCARBONS BY THE MUSSEL, MYTILUS EDULIS, IN LABORATORY EXPERIMENTS Robert C. Clark, Jr., and John S. Finley' ABSTRACT Petroleum paraffin hydrocarbons (W-C14H30 to n-Cs'jK'je) from No. 2 and No. 5 fuel oils were rapidly incorporated into the mussel, Mytilus edulis, in a laboratory system that simulated tides. The mussels were exposed to levels of petroleum hydrocarbons from a surface slick similar to those encountered in the environment after an oil spill. After 14 days in clean seawater, the mussels had lost most of the hydrocarbons from the fuel oils; however, detectable traces of the No. 2 fuel oil still remained after 35 days. Preliminary results from these laboratory studies confirm previous studies of pollutant uptake and loss following actual oil spills. Petroleum hydrocarbon uptake by the common bay mussel, Mytilus edulis, can be readily deter- mined in the laboratory with the analytical chemical methods of solvent extraction, liquid- solid, and gas-liquid chromatography. Mussels lend themselves well to such studies because of 1) their worldwide distribution and ready availability (Davies 1969; Becker et al. 1973); 2) the considerable amount of physiological base line data available (Field 1922); 3) their hardiness as an experimental test organism (GilfiUan 1973); 4) their convenient size, which is small enough to sample adequately and use in the laboratory experimentally but large enough for specific organ dissection (Lee, Sauerheber, and Benson 1972); 5) their position in the intertidal ecosystem as a major pathway for energy transfer utilizing phy- toplankton and debris (Ricketts and Calvin 1962); and 6) their known capacity for concentrating various pollutants from their environment (Gref- fard and Meury 1967; Modin 1969; Zitko 1970; Clark and Finley 1973a). Earlier studies of hydrocarbon uptake and its effects on marine organisms include those by Griffith (1970) who determined the toxicity of Arabian light crude oil and oil-dispersant mix- tures on mussels under tidal conditions, and Lee, Sauerheber, and Benson (1972) who used mineral oil and radio-labeled ['^C] heptadecane to study the laboratory uptake, body distribution, and loss of hydrocarbons in mussels. We previously reported on uptake of petroleum hydrocarbons by aquatic 'Northwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA, 2725 Montlake Boulevard East, Seattle, WA 98112. organisms from several oil spills (Clark and Finley 1973a). This paper reports the findings of a preliminary laboratory study using two refined petroleum products, a No. 2 fuel oil and a No. 5 fuel oil, in a laboratory system that simulates tidal ac- tion. EXPERIMENTAL METHODS A tidal aquarium for laboratory studies of the uptake and loss of petroleum by intertidal or- ganisms has been described (Clark and Finley in press) (Figure 1). This system consists of two aquaria set at a 25° angle to represent a beach surface. The first aquarium contained the or- ganisms being exposed to the pollutant, and the second aquarium, where all procedures were duplicated except for the pollutant exposure, con- tained control organisms. These control organisms served as the base line comparison for mortality studies and hydrocarbon analysis. The flood tide was simulated twice a day by pumping an artificial seawater medium (LaRoche et al. 1970) from a carboy using a timer-equipped, variable-speed pump. The ebb tide was accomplished by siphoning the seawater medium back into the carboys from beneath the surface oil slick in the test tank. Prior to exposure, the mussels (collected in an area dis- tant from known sources of petroleum pollution) were acclimated to the tidal system for 24 h following a previous 48- to 96-h conditioning in the laboratory in an aerated aquarium (Table 1). In practice, usually two sets of test organisms were used; one set was placed in the intertidal zone, held by glass rods placed horizontally across Manuscript accepted September 1974. FISHERY BULLETIN: VOL. 73, NO. 3, 1975. 508 - • CLARK and FINLEY: UPTAKE AND LOSS OF PETROLEUM HYDROCARBONS Figure l.-Tidal aquarium system showing the test and control tanks. the bottom of the sloping aquarium bottom, and exposed to the tidal sweep of the oil floating on the water surface. The second set was placed below the extreme lower level of the tidal sweep so that the organisms were continuously immersed in the medium where they were exposed to dissolved and emulsified fractions of the pollutant but never physically covered with oil. During the experiment, a known volume of oil was layered onto the 1,700 cm^ surface of the test aquarium at "high" tide. The system was then allowed to run through several complete tidal cycles in an attempt to simulate the conditions Table 1.— Data on size of mussels, number of specimens, and experimental conditions. No. 2 No. 5 Item fuel oil fuel oil Average shell length 56.7 mm 54.7 mm Number of specimens: Controls 36 16 Exposed — Surface slick 18 18 Exposed — Submerged 18 6 Exposure time 48 h 32 h Complete tidal cycles 4 2.7 Water/air temperature 2rc 20°-22°C Actual volume of oil applied 100 ml 97 ml Slicl< thickness calculated: ■■High tide" 0.55 mm 0.53 mm "Low tide" 1.53 mm 1.49 mm which might be found under a pier or along a beach following a significant oil spill. Two types of refined products were used in the uptake studies, a No. 2 heating fuel oil and a No. 5 heavy burner fuel oil. At the end of the exposure, the bulk oil was skimmed from the surface at "high" tide, and the organisms were immediately placed in aerated aquaria containing seawater medium. The water was changed daily and the mussels fed a clam- based diet three times a day (LaRoche et al. 1970). Groups of the two sets of mussels were removed from these aquaria for analysis 1, 7, 14, and 35 days after the end of the exposure experiments. The paraffin hydrocarbon analysis techniques for marine organisms have been described by Clark and Finley (1973b). Mussels were sampled and analyzed in groups of two to six individuals; analysis was run on the combined tissue and body fluids. No hydrocarbon content of the test seawater medium nor of the clean seawater was determined. Care was taken to minimize con- tamination of the shucked meats from oil that might have adhered to the shell, particularly in the early samples (1 and 7 days). All results have been reduced to "parts-per-million (ppm) dry extract- 509 FISHERY BULLETIN: VOL. 73, NO. 3 ed-weight" basis (lO-^g of w-parafRn hydrocarbon per gram of the sum of the dried residue plus the solvent extractables). RESULTS Mortality Studies Acute toxicity studies were not intended to be a major portion of this investigation since our paraffin hydrocarbon analysis techniques are used primarily on surviving organisms that have taken up oil pollutants at levels below that detectable by sight or smell. The percentage of cumulative mortality (Figures 2 and 3) shows an approximate doubling for mussels exposed to the No. 2 fuel oil compared with the mussels exposed to the No. 5 fuel oil, although the duration of exposure was also greater for the No. 2 fuel oil (46 h vs. 32 h). Both the slick-exposed and the submerged specimens in the No. 5 fuel oil had a slightly higher mortality than the controls, but these differences might not be significant because of the small number of or- ganisms used. The No. 2 fuel oil slick-exposed specimens, however, showed a mortality over twice that of the controls. The submerged specimens had a low initial mortality, but after 2 wk mortality had increased to the level of the slick-exposed specimens. These mortality data provide a comparison of the two petroleum pollutants and of the test and control groups but were not further utilized to compute median tolerance limits which were beyond the scope and objective of these preliminary qualitative studies. Griffith (1970) reported no mortalities in mussels exposed to aged crude oil for four tidal cycles over a total of 120 h. When the same mussels were placed in clean seawater, their rate of recovery was assessed by noting that byssal reattachment required 18 h for 50% of the exposed organisms but less than 12 h for the controls. No observations beyond 120 h (5 days) were reported. No. 2 Fuel Oil Studies Groups of mussels were collected 1, 7, 14, and 35 days after their removal to clean seawater medium. The n-paraffin hydrocarbon patterns for one set are presented as an example (Figure 4). If one assumes that the hydrocarbon pattern of the "controls" represents the natural or biogenic paraffin hydrocarbons and that the pattern of the "test" specimens represents the biogenic plus the pollutant, then by subtracting the former from the latter the resulting "residual" pattern might be expected to depict the pollution hydrocarbon pat- tern of the petroleum product tested. The "residual" paraffin patterns are shown for the No. 2 fuel oil bioassay study for mussels sampled 1 and 7 days (Figure 5) and 14 and 35 days (Figure 6) after removal from the pollutant. The shape of the residual paraffin patterns for all four sampling periods approximates that of the pollutant below ?t-C27H56 and above a residual content level of 0.050 ppm., although individual paraffin hydrocarbons may show variation from the smooth, nearly bell-shaped curve for the No. 2 fuel oil. The quantities of uptake and loss of n- paraffin hydrocarbons from the No. 2 fuel oil by the mussels are shown in Figure 7. The uptake after 50i Controls — Eiposed - Surface slick — Exposed- Submerged Test exposure 10 15 20 25 DAYS AFTER REMOVAL 30 35 Figure 2.-Cumulative mortalities of mussels for 35 days following a 48-h exposure to a No. 2 fuel oil. 50 40 30 Ui > 20 S lOi Controls Exposed - Surfoce slick - Exposed - Submerged r Test exposure 10 15 20 25 DAYS AFTER REMOVAL 30 Figure 3.-Cumulative mortalities of mussels for 35 following a 32-h exposure to a No. 5 fuel oil. 35 days 510 CLARK and FINLEY: UPTAKE AND LOSS OF PETROLEUM HYDROCARBONS 48-h exposure plus 1 day in clean medium was 110 ppm. for the slick-exposed specimens compared with only 29 ppm. for the submerged specimens. Within 7 days the residual content had dropped by nearly 75%, after which it declined at a slower rate but was still significantly above background (8 ppm.) after 35 days in the slick-exposed specimens. No. 5 Fuel Oil Studies The residual paraffin hydrocarbon pattern (Figure 8) for mussels exposed to a No. 5 fuel oil revealed a definite uptake of pollutant hydrocar- bons at the end of the 32-h exposure, and the specimens collected 7 days later from clean seawater medium contained less than 1 ppm. total residual paraffin hydrocarbons attributable to the pollutant (Figure 7). Gas-liquid chromatography of the saturated hydrocarbon fraction of the exposed mussels revealed a series of branched-chain hydrocarbons X CONTROLS EXPOSED -SURFACE SLICK EXPOSED -SUBMERGED UJ a. Q. 10 15 20 25 30 35 CARBON ATOMS /MOLECULE 40 from below C14 to C26, but of this series only pris- tane was quantified and included in the calcula- tions. Most unsaturated and aromatic compounds were separated from the saturated fraction at the silica gel/alumina column chromatography stage without further analysis. DISCUSSION AND CONCLUSIONS These experiments, although preliminary in nature, provide four basic conclusions: 1) mussels rapidly took up pollution hydrocarbons during ex- posure; 2) mussels lost pollution hydrocarbons when removed from the test tanks and held in clean seawater (depuration), but significant quan- tities of No. 2 fuel oil remained for 35 days; 3) the w-paraffin residual pattern (exposed levels minus control levels) established for the exposed mussels nearly duplicated the shape of the pollutant hydrocarbon pattern; and 4) these laboratory results confirmed analyses made on shellfish following actual oil spills in the marine environ- ment. O O 48-H AGED NO. 2 FUEL OIL X X EXPOSED TO SURFACE SLICK I DAY AFTER REMOVAL 9 ■o V >, O o a. a. a. Q- v> _i 2 — UJ o u UJ a. EXPOSED TO SURFACE SLICK 7 DAYS AFTER REMOVAL 15 20 25 30 35 CARBON ATOMS / MOLECULE 40 Figure 4.-ParafRn hydrocarbon patterns in mussels exposed to a No. 2 fuel oil: 1 day after removal. PiGUSE 5.-Residual paraffin patterns of mussels exposed to a No. 2 fuel oil: 1 and 7 days after removal. 511 FISHERY BULLETIN: VOL. 73, NO. 3 O 48-H AGED NO. 2 FUEL OIL EXPOSED TO SURFACE SLICK 14 DAYS AFTER REMOVAL EXPOSED TO SURFACE SLICK 35 DAYS AFTER REMOVAL 9 u o . O s z Q. a. a. a. CO _i 2 — uj o u bJ a. 15 20 25 30 35 CARBON ATOMS / MOLECULE 40 Figure 6. -Residual paraffin patterns for mussels exposed to a No. 2 fuel oil: 14 and 35 days after removal. The mussels rapidly took up a wide range of petroleum paraffin hydrocarbons. Lee, Sauerheber, and Benson (1972) had found this to be the case for mussels where [^O] heptadecane was detected within 15 min. The level of hydrocarbons we found in our mussels exposed to No. 2 fuel oil represents less than maximum up- take because after exposure they were held 24 h in clean seawater before sampling. The No. 5 fuel oil specimens were collected immediately after their 32-h exposure to the oil, and extreme care was exercised to avoid contaminating the tissue with any oil adhering to the outside of the shell when shucking prior to extraction. Griffith (1970) suggests that the attachment of the byssal threads by the mussel could be affected by the exposure to petroleum. The byssal attach- ment is made by means of a grooved foot, which is extended from the shell and placed in contact with the substratum. Glandular secretions of collagen mixed with phenolic protein run from the foot groove, become attached to the substratum, and during withdrawal of the foot undergo tanning by the action of polyphenoloxidase (Pujol 1967). It is No. 2 Fuel oil (48-h exposure ) 0 Surfoce slick 1 Subsurface No. 5 Fuel oil (32-h exposure) A Surfoce slick 10 15 20 25 DAYS AFTER REMOVAL 30 35 Figure 7.-w-Paraffin hydrocarbon uptake and loss of fuel oil by mussels. not certain whether the oil upsets this chemical process or inhibits the muscular actions of the foot necessary to anchor the byssal thread. When other mollusks such as the American oyster, Crassostrea virginica, were exposed to an — O 79-H AGED NO. 5 FUEL OIL --X EXPOSED TO SURFACE SLICK IMMEDIATELY AFTER REMOVAL EXPOSED TO SURFACE SLICK 7 DAYS AFTER REMOVAL >, O Q a. a. Q. Q. UJ o o UJ a. 15 20 25 30 35 CARBON ATOMS / MOLECULE 40 Figure 8.-Residual paraffin patterns for mussels exposed to a No. 5 fuel oil. 512 CLARK and FINLEY: UPTAKE AND LOSS OF PETROLEUM HYDROCARBONS oil-water mixture, as distinct from our nonagitat- ed surface oil slick under tidal conditions, they also showed rapid uptake. Anderson (1973) found up- take of No. 2 fuel oil to be greatest in the first 24 h with lower uptake at longer periods. By com- parison, he found uptake in clams, Rangia cunea- ta, reached a maximum in 72 h, and greater con- centrations in the clam tissue were reached than for oysters. Aromatic hydrocarbons showed a greater uptake in the moUusk tissue than satu- rated forms. Stegeman and Teal (1973), using a flow-through exposure system, showed that oysters initially took up a No. 2 fuel oil-water mixture in direct relation to the hydrocarbon concentration in the water up to at least 450iWg/liter, above which they remained closed. They also found an enrichment in aromatic fractions compared to the saturated fraction, and under long-term exposure (49 days) the latter fraction showed a progressive decrease in amount with increasing length of exposure. Under our conditions the No. 5 fuel oil-exposed specimens took up less w-paraffin hydrocarbons than the No. 2 fuel oil-exposed mussels, and the former also lost them faster. Both sets of specimens had depurated their residual hydrocar- bons by 75% within 1 week, but the No. 2 fuel oil-exposed mussels still contained detectable amounts at the end of 35 days. Lee, Sauerheber, and Benson (1972) found that mussels would discharge over 90% of the incorporated mineral oil after several days, a result they confirmed with labelled w-heptadecane. Anderson (1973) found that oysters lost 94% of the saturated hydrocarbon uptake but only 82% of their aromatics after 13 days when exposed to a No. 2 fuel oil; after 52 days no residual pollutant hydrocarbons were detected at the 0.5-ppm. level (wet weight). Exposure to South Louisiana crude oil showed detectable but low levels of saturates but no detectable aromatics after 27 days depura- tion. Stegeman and Teal (1973) also found a rapid loss of hydrocarbons from oysters exposed to No. 2 fuel oil-water mixture, but a persistent portion (34 ppm. wet weight above background) remained. These preliminary experiments did not provide results as to the mechanisms or pathways of pe- troleum hydrocarbon uptake by the contaminated mussels. The mechanisms of uptake and transport of pollutant hydrocarbons from the environment into organisms may have a very important effect on the degree to which the subsequent depuration is reversible. For instance, hydrocarbons trans- ported across gill membranes in solution or as emulsified droplets enter the bloodstream of fishes very rapidly and can be rapidly depleted on removal of the pollutant (Roubal 1974). On the other hand, hydrocarbons in food sources are resorbed at a different site, which, for the basking shark occurs in the spiral valve where they are transferred to the liver and remain highly persis- tent (Blumer 1967). The residual paraffin hydrocarbon patterns showed a strikingly similar pattern to the aged pollutant, yet the organisms appeared healthy and had no visible contamination or oily odor. Lee, Sauerheber, and Benson (1972) indicated that gas-liquid chromatograms of mineral oil in mus- sels were like the original mineral oil except for some loss of the short-chain paraffin hydrocarbons. Blumer et al.(1970) used gas chromatograms of oysters and scallops contaminated by a No. 2 fuel oil to show that the patterns had the same general features as the chromatogram of the fuel oil. Anderson (1973) found an %-paraffin hydrocar- bon pattern in clams similar to the high-aromatic No. 2 fuel oil; however, the maximum hydrocarbon in the clam (n-Cn'. approximately 1.6 times more than n-Cie and 1.9 times n-Cig ) was one carbon number higher than the pollutant maximum [n- Ci6 : 1.1-1.2 times the adjacent paraffins). Stegeman and Teal (1973) showed gas chroma- tograms of contaminated oysters having a patter'n similar to the pollutant and a maximum n-paraf- fin concentration of C is or C 18' '19- Our presentation of petroleum hydrocarbon up- take as evidenced by w-paraffin analyses is not a complete picture of petroleum contamination since it reflects only a portion of the total hydrocarbons and non-hydrocarbons in the oil. Al- so, the various hydrocarbon components of the pe- troleum are not necessarily available to the or- ganisms in the water in the same proportion as they exist in the original petroleum. Vaughan (1973) reported an enrichment of methyl- naphthalenes compared to w-C 12.20 paraffins of 15:1 in oysters exposed to a South Louisiana crude oil (50/ig/ml) in a seven-day bioassay experiment and an enrichment in the water extracts of about 3:1. Further, the w-paraffin content of petroleum pollutants in shellfish is often depleted with time relative to that of the source (Stegeman and Teal 1973). Therefore, while our values for pollutant uptake based on w-paraffin hydrocarbon analyses yield conservative estimates, they demonstrate that these experimentally simple methods can be 513 useful. Aromatic hydrocarbons, which are a biogenic rarity yet often a major component of petroleum and its refined products found to be rapidly taken up by marine organisms, may be rapidly lost from contaminated organisms (Lee, Sauerheber and Dobbs 1972; Stegeman and Teal 1973; Anderson 1973). Consequently, the utility of these compounds in marine pollution monitoring programs and in long-term bioassay experiments may be somewhat limited. We previously reported (Clark and Finley 1973b) uptake by mussels, M. edulis and M. califarnianus, of petroleum n-paraffins following oil spills in the marine environment. A No. 2 fuel oil spill resulted in considerably greater uptake (10 ppm. of n-Ci-j) than for mussels exposed to Navy Special fuel oil residue (nearly 1 ppm. of n.-C^-j); however, in both cases it was obvious that the ap- parently healthy mussels had acquired an n- paraffin hydrocarbon pattern like that of the pollutant. By creating an oil slick and a tidal sys- tem within an aquarium under laboratory condi- tions, it is possible to show that these earlier findings can be reproduced. These results add further support to the data of other investigators who used different approaches and analytical techniques. We did not analyze hydrocarbon levels in specific organs, conduct metabolic studies, or de- termine aromatic content. Stegeman and Teal (1973) and Anderson (1973) found that aromatic hydrocarbons were often enriched in oysters in preference to the n-paraffins. While we have given percentage loss of pollu- tant paraffin hydrocarbons in our presentation as well as actual concentration levels (Figure 7), the percentage value is very dependent on both the level of initial pollution exposure and on the lower limit of sensitivity of the experimental method for detecting pollutant uptake near biogenic background concentrations in marine organisms. Thus, one might have two sets of organisms showing similar concentrations of residual pollu- tant after considerable depuration but with a dramatically different percentage loss as a result of different exposure levels. The variation of uptake and loss of petroleum hydrocarbons in marine organisms most certainly is related to the magnitude of the exposure-the amount of the pollutant and the duration, as well as physical and chemical properties of the pollu- tant. FISHERY BULLETIN: VOL. 73, NO. 3 LITERATURE CITED Anderson, J. W. 1973. Uptake and depuration of specific hydrocarbons from oil by the bivalves Rangia cuneata and Crassostrea vir- ginica. Background papers for A Workshop on Inputs, Fates and Effects of Petroleum in the Marine Environ- ment, Airlie, Virginia, 21-25 May, Vol. II, p. 690-708. Becker, C. D., J. A. Lichatowich, M. J. Schneider, and J. A. Strand. 1973. Regional survey of marine biota for bioassay stan- dardization of oil and oil dispersant chemicals. Am. Pet. Inst., Res. Rep., Publ. 4167, 106 p. Blumer, M. 1967. Hydrocarbons in digestive tract and liver of a basking shark. Science (Wash., D.C.) 156:390-391. Blumer, M., G. Souza, and J. Sass. 1970. Hydrocarbon pollution of edible shellfish by an oil spill. Mar. Biol. (Berl.) 5:195-202. Clark, R. C, Jr., and J. S. Finley. 1973a. Paraffin hydrocarbon patterns in petroleum-polluted mussels. Mar. Pollut. Bull. 4:172-176. 1973b. Techniques for analysis of paraffin hydrocarbons and for interpretation of data to assess oil spill effects in aquatic organisms. Proceedings 1973 Joint Conference on Prevention and Control of Oil Spills, p. 161-172. Am. Pet. Inst., Wash., D.C. In press. Tidal aquarium for laboratory studies of environ- mental effects on marine organisms. Prog. Fish-Cult. Davies, G. 1969. Mussels as a world food resource. In F. E. Firth (edi- tor), The encyclopedia of marine resources, p. 421- 426. Van Nostrand Reinhold Co., N.Y. Field, I. A. 1922. Biology and economic value of the sea mussel Mytilus edulis. Bull. U.S. Bur. Fish. 38:127-259. GiLFILLAN, E. S. 1973. Effect of seawrater extracts of crude oil on carbon budgets in two species of mussels. Proceedings 1973 Joint Conference on Prevention and Control of Oil Spills, p. 691-695. Am. Pet. Inst., Wash., D.C. GREFFARD, J., AND J. MEURY. 1967. Note sur la pollution en rade de Toulon par les hydrocarbures carcerigenes. Cah. Oc6anogr., Bull. Inf. Com. Cent. Oc6anogr. Itud. Cotes 19:457-468. Griffith, D. de G. 1970. Toxicity of crude oil and detergents to two species of edible molluscs under artificial tidal conditions. FAO (Food Agric. Organ. U.N.) Tech. Conf. Mar. Pollut., Rome, December 1970, Pap. E-16, 12 p. LaRoche, G., R. Eisler, and C. M. Tarzwell. 1970. Bioassay procedures for oil and oil dispersant toxicity evaluation. J. Water Pollut. Control Fed. 42:1982-1989. Lee, R. F., R. Sauerheber, and A. A. Benson. 1973. Petroleum hydrocarbons: uptake and discharge by the marine mussel Mytilus edulis. Science (Wash., D.C.) 177:344-346. Lee, R. F., R. Sauerheber, and G. H. Dobbs. 1972. Uptake, metabolism and discharge of polycyclic aromatic hydrocarbons by marine fish. Mar. Biol. (Berl.) 17:201-208. MODIN, J. C. 1969. Residues in fish, wildlife, and estuaries. Chlorinated 514 CLARK and FINLEY: UPTAKE AND LOSS OF PETROLEUM HYDROCARBONS hydrocarbon pesticides in California bays and es- tuaries. Pestic. Monit. J. 3:1-7. Pujol, J. P. 1967. Formation of the byssus in the common mussel (My- tilus edulis L.). Nature (Lond.) 214:204-205. RiCKETTS, E. F., AND J. CaLVIN. 1962. Between Pacific tides. 3rd ed., revised. Stanford Univ. Press, Stanford, 516 p. ROUBAL, W. T. 1974. Spin-labeling of living tissue. In E. J. Vernberg (edi- tor). Pollution and physiology of marine organisms, p. 367- 379. Academic Press, N.Y. Stegeman, J. J., AND J. M. Teal. 1973. Accumulation, release and retention of petroleum hydrocarbons by the oyster Crassostrea virginica. Mar. Biol. (Berl.) 22:37-44. Vaughan, B. E. 1973. Effects of oil and chemically dispersed oil on selected marine biota— A laboratory study. Am. Pet. Inst., Publ. 4191, Wash., D.C., 105 p. ZiTKO, V. 1970. Determination of residual fuel oil contamination by aquatic animals. Bull. Environ. Contam. Toxicol. 5:559-564. N 515 SYSTEMATICS AND MORPHOLOGY OF THE BONITOS (SARD A) AND THEIR RELATIVES (SCOMBRIDAE, SARDINI)^ Bruce B. Collette^ and Labbish N. Chao^ ABSTRACT The bonitos constitute the scombrid tribe Sardini, consisting of eight species placed in five genera. They differ from the more primitive mackerels and Spanish mackerels in lacking a notch in the hypural plate and in having a bony lateral keel on the posterior caudal vertebrae. From the higher tunas, they differ in having the bony keel only incompletely developed and in lacking a specialized subcutaneous vascular system. The monotypic Australian endemic Cybiosarda elegans shares several characters with the monotypic eastern Atlantic endemic Orcynopsis unicolor (structure of the bony caudal keel; relative lengths of liver lobes; position and size of spleen) that distinguish them from Gymnosarda unicolor and the species of Sarda. Sarda contains four allopatric species, which differ from each other in such characters as numbers of fin rays, gill rakers, vertebrae, and teeth: the Atlantic S. sarda; the southeastern Australian S. australis; the tropical Indo-Pacific S. orientalis; and the eastern temperate Pacific S. chiliensis. The monotypic Indo-West Pacific reef species Gymnosarda unicolor is the only member of the Sardini that has a swim bladder and lacks intermuscular bones on the back of the skull. The monotypic Southern Ocean Allothunnus fallai differs from all other scombrids in having laterally extended prootic wings. It is more closely related to the bonitos than to any other scombrids. Allothunnus resembles the higher tunas in having a prootic pit but lacks the subcutaneous vascular system. Tables of meristic characters, diagrams of the soft anatomy, and drawings of most bones are included in the first part of the paper. The second part of the paper includes sections on synonymy, comparative diagnosis, types of nominal species, and distribution for each species. The purpose of this paper is to clarify the rela- tionships of the Sardini at the generic and specific level. This work is part of a continuing study of the systematics of the Scombridae. The methods used are similar to those used by Gibbs and Collette (1967) in a revision of Thunnus and rely heavily on the classic work of Kishinouye (1923) and Godsil (1954, 1955). The bonitos (Sarda) and their relatives form a tribe (Sardini) of the subfamily Scombrinae inter- mediate between the primitive mackerels (Scom- brini) and Spanish mackerels (Scomberomorini), and the more advanced tunas (Thunnini) (see Collette and Gibbs 1963a; Gibbs and Collette 1967). Five genera are treated in this paper. The status of the related monotypic genera Orcynopsis, Cybiosarda, and Gymnosarda has been unclear; for example, Fraser-Brunner (1950) placed Cybiosarda in the synonymy of Gymnosarda. The ■Contribution No. 529, Virginia Institute of Marine Science, Gloucester Point, VA 23062. ^vstematics Laboratory, National Marine Fisheries Service, NOAA, National Museum of Natural History, Washington, DC 20560. 'Virginia Institute of Marine Science, Gloucester Point, VA 23062. systematic position of the monotypic Allothunnus has been still more confused-whether it is closer to Thunnus (Fraser-Brunner 1950), to Sarda (Fitch and Craig 1964), or strikingly different from all other scombrids (Nakamura and Mori 1966). There has been no agreement on the number of species of Sarda. Fraser-Brunner (1950) recognized three species: chiliensis, orientalis, and sarda. Godsil (1955) believed that there were two basic groups of species— sarda-chiliensis and orientalis-velox. Some authors have considered S. australis as a valid species, others as a subspecies of S. chilien- sis. This project was initiated at the request of the FAO (Food and Agriculture Organization of the United Nations) panel of Experts for the Facili- tation of Tuna Research at its Fourth Session in La Jolla, Calif, in November 1971, and should be considered as a report from the Working Party on Tuna and Billfish Taxonomy. Bonitos, as a group, are one of the few underexploited groups of tunalike fishes; therefore, research on their systematics is a necessary predecessor of success- ful management. According to the FAO Yearbook of Fishery Statistics for 1972 (Food and Agriculture Or- Manuscript acx:epted January 1975. FISHERY BULLETIN: VOL. 73, NO. 3, 1975. 516 COLLETTE and CHAO: SYSTEMATICS AND MORPHOLOGY OF THE BONITOS (SARDINI) ganization of the United Nations 1973), the two species of bonitos that are presently of economic importance are Sarda chiliensis and S. sarda. Peruvian fishermen landed 54,000-73,000 metric tons per year of the southeast population of S. chiliensis in 1965-1972. Smaller catches by Chile and of the northeast Pacific population by Mexico and the United States made the total 65,000-94,000 metric tons per year during that period. Sarda sarda is fished particularly by Turkey in the Mediterranean and the Black Sea where 11,700-55,200 metric tons per year were landed in 1965-1972. Other catches of S. sarda by Spain, Portugal, Greece, Angola, Argentina, and Brazil made the total 25,000-65,000 metric tons per year in 1965-1972. Both the Japanese and the Koreans fish for S. orientaXis and there are smaller catches elsewhere throughout its range. Sarda australis comes into the markets in Sydney and probably elsewhere in southeastern Australia. In 1971, Morocco was reported to have landed 600 metric tons of Orcynopsis and we have seen Orcynopsis in the markets in Tunis. We have seen specimens of Cybiosarda in the Sydney fish market mixed with S. australis. The only commercial catch of Allothunnus was the 230 tons taken with purse seines off eastern Tasmania in June 1974 (Webb and Wolfe 1974). Gymnosarda occurs around coral reefs where it is taken by fishermen on hook and line. Emphasis was placed on obtaining fresh or frozen specimens from each population of each species for dissection. Standard counts and measurements were taken, color pattern was recorded, and a search was made for parasitic copepods. Results of the copepod study will be reported on later by Roger F. Cressey (United States National Museum, USNM). The viscera were examined and drawn in situ following removal of an oval portion of the ventral body wall. The viscera were then removed and drawings were made of the liver and other selected organs. The kidneys and anterior parts of the arterial system were then drawn. Counts of ribs and intermus- cular bones were made and the specimen was then skeletonized. Specimens were immersed in hot water to assist removal of the flesh. For morphometric comparisons, the base measurement used for fresh, frozen, and preserved specimens was millimeters fork length (mm FL). Skeletal material was measured in millimeters skeletal length, the distance from the anterior margin of the ethmoid to the posterior tip of the hypural plate, a distance somewhat shorter than fork length. Skulls were measured from the anterior margin of the ethmoid to the postero- ventral junction of the skull with the first ver- tebral centrum. This paper is divided into two major sections. The first part describes and illustrates the squamation, morphometry, meristic characters, soft anatomy, and osteology of the Sardini. The second part treats the genera and species sepa- rately including synonymy, diagnosis (based on characters from the first section), types of nominal species, geographical distribution, and, for some species, geographic variation. MATERIAL Abbreviations used for the institutions cited herein are as follows: AB - Northwest Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA, Auke Bay, Alaska. AMS - Australian Museum, Sydney. ANSP -Academy of Natural Sciences, Philadelphia, Pa. BMNH -British Museum (Natural History), London. BPBM -Bernice P. Bishop Museum, Honolulu, Hawaii. CAS - California Academy of Sciences, San Francisco. CBL -Chesapeake Biological Laboratory, Solomons, Md. CSIRO -CSIRO Marine Biological Laboratory, Cronulla, N.S.W., Australia. DM -Dominion Museum, Wellington, New Zealand. FMNH -Field Museum of Natural History, Chicago, 111. HUJ - Hebrew University, Jerusalem. LACM - Los Angeles County Museum of Natural History, Los Angeles, Calif. MACN - Museo Argentina de Ciencias Naturales, Buenos Aires. MCZ - Museum of Comparative Zoology, Har- vard. MNHN -Mus6um National d'Histoire Naturelle, Paris. MSNG - Museo di Storia Naturale, Genoa. MSUF -Museo de La Specola, Universita di Firenze, Florence. 517 FISHERY BULLETIN: VOL. 73, NO. 3 NHMV - Naturhistorisches Museum, Vienna. NMC -National Museum of Natural Sciences, Ottawa. QM -Queensland Museum, Brisbane. RMNH -Rijksmuseum van Natuurlijke Historie, Leiden. RUSI -J. L. B. Smith Institute of Ichthyology, Rhodes University, Grahamstown. SAM - South African Museum, Capetown. SFRS -Sea Fisheries Research Station, Haifa, Israel. SIO - Scripps Institution of Oceanography, La Jolla, Calif. SMF -Senckenberg Museum, Frankfurt-am- Main. TABL -Southeast Fisheries Center, National Marine Fisheries Service, NOAA (for- merly Tropical Atlantic Biological Laboratory), Miami, Fla. UBC -Institute of Fisheries, University of British Columbia, Vancouver. UCLA - University of California, Los Angeles. UMML -Rosenstiel School of Marine and At- mospheric Science, Miami, Fla. UMMZ -University of Michigan Museum of Zoology, Ann Arbor. USNM -United States National Museum, Washington, D.C. WAM - Western Australia Museum, Perth. WHOI -Woods Hole Oceanographic Institution, Woods Hole, Mass. ZMK -Zoological Museum, Copenhagen. ZMO -Zoological Museum, Oslo. The material examined is listed by general locality under four or five headings for each species (except Sarda chiliensis, S. orientalis, and S. sarda which are subdivided into two populations each). The numbers in each category are not addi- tive but are included to give some degree of con- fidence in the morphological data presented in the body of the paper. "Total specimens" is the total number of individuals examined whether preserved, dissected, or skeletons. "Measured and counted" includes specimens that were sub- sequently dissected and the preserved museum specimens used for detailed morphometric and meristic examination. "Counts only" are the addi- tional museum specimens used only for meristic examination. "Skeletons" refer to all the skeletal material examined, specimens that were dissected plus skeletal museum material. Allothunnus fallai. Total 8 specimens (451-787 mm FL). Dissected 4 (680-778). Tasmania (3); California CD- Measured and counted 7 (642-787). Tasmania (4); California (1); New Zealand (2). Skeletons 5 (451-778). Tasmania (3); California (1); South Africa (1). Cybiosarda elegans. Total 22 specimens (250-422 mm FL). Dissected 5 (355-422). Perth, Western Australia (1); Macleay River, New South Wales (4). Measured and counted 21 (250-422). New South Wales (11); E. Queensland (7, including holotype of Scomberomorus (Cybiosarda) elegans Whitley); Gulf of Carpentaria (1); Western Australia (2). Examined 1 (380). New South Wales. Skeletons 5 (355-422). Western Australia (1); New South Wales (4). Gymnosarda unicolor. Total 38 specimens (71.6-1,080 mm FL). Dissected 6 (522-787). Amirante Islands (2); Truk Islands, Caroline Islands (3); Bikini, Marshall Islands (1, partial). Measured and counted 31 (71.6-1,040). Red Sea (5, including holotype of Thynnus unicolor Riip- pell); Comoro Islands (1); Amirante Islands (2); Madagascar (1); New Britain (1); Solomon Islands (2); Gilbert Islands (1); Japan (3); Palau Islands (1); Caroline Islands (7); Marshall Islands (5); Society Islands (1); Marquesas Islands (1). Examined 1 (267-mm head of 1,080-mm specimen). Pitcairn Group. Skeletons 11 (about 625-1,013). Amirante Islands (2); Marshall Islands (4); Truk Islands (3); unknown locality (2). Orcynxypsis unicolor. Total 55 specimens (164-960 mm FL). Dissected 11 (332-645). Israel (5, partial); Tunisia (6, complete). Measured and counted 43 (164-960). Lebanon (12, 242-325); Israel (12, 285-735); Egypt (5, 164-280); Tunisia (7, 312-645); Nice (1, 553); Mauritania (2, 400-410); Senegal (2, 405-960); Norway (2, 565-570, types of Thynnus peregrinus Collett). Counts only 12 (417-950). Pizza (1, ca. 950); Elba (1, ca. 790); Gulf of Genoa (1, 670); Rimini, Adriatic (1, ca. 600); Egypt (6, 417-556); locality unknown (2). Skeletons 11 (332-645). Israel (5); Tunisia (6). 518 COLLETTE and CHAO: SYSTEMATICS AND MORPHOLOGY OF THE BONITOS (SARDINI) Sarda australis. Total 21 specimens (195-495 mm FL). Dissected 3 (360-495). New South Wales, Aus- tralia. Measured and counted 21 (195-495). Norfolk Island (1); New South Wales (20, including holo- type of Pelamys australis Macleay). Skeletons 3 (360-495). New South Wales. Sarda chiliensis— northeast Pacific. Total 91 specimens (207-643 mm FL). Dissected 4 (401-472). La Jolla, Calif. Measured and counted 24 (207-587). California: La Jolla; San Diego (including holotype of Pelamys lineolatus Girard); Los Angeles; San Pedro; Santa Barbara; Oceanside. Baja California: Coronados Island; Natividad Island; Cedros Island; Blanca Bay. Counts only 21 (220-625). Vancouver, British Columbia (1); Alaska (2); S. California (17); Revillagigedos Island (1). Skeletons 50 (310-643). S. California (including holotype of Sarda stockii (David)). Sarda chiliensis -southeast Pacific. Total 44 specimens (57.2-672 mm FL). Dissected 7 (437-571). Callao, Peru. Measured and counted 23 (94.1-672). Valparaiso, Chile (holotype of Pelamys chiliensis Cuvier). Arica Bay, Chile. Peru: Callao; Foca Island; San Lorenzo Island; Pachacamac Island; Guanape Island; San Gallan Island. Counts only 9 (57.2-636). Peru: San Lorenzo Island; San Gallan Island; Foca Island; Callao. Skeletons 18 (437-571). Callao, Peru. Sarda orientalis—lndo-West and central Pacific. Total 31 specimens (150-645 mm FL). Dissected 5 (341-500). Tokyo (3); Hawaii (2). Measured and counted 27 (150-645). South Africa (2); Seychelles Islands (1); Red Sea (2); Cochin, India (1); Western Australia (paratype of Sarda orientalis serventyi Whitley); China (4); Japan (12, including types of Pelamys orientalis Temminck and Schlegel); Hawaii (4). Counts only 3 (223-370). Muscat (2); Ceylon (1). Skeletons 6 (341-500). Muscat (1); Tokyo (3); Hawaii (2). Sarda orientalis— eastern Pacific. Total 21 specimens (354-447 mm FL). Dissected 4 (354-447). Navidad Bay, Mexico (1); Pinas Bay, Panama (2); Pearl Islands, Panama (1). Measured and counted 12 (354-447). Mexico (1); Panama (8, including holotype of Sarda velox Meek and Hildebrand); Galapagos Islands (2); Giilf of Guayaquil (1). Counts only 7 (429-650). Cabo San Lucas and Las Tres Marias Islands, Mexico (4); Galapagos Islands (3). Skeletons 6 (354-447). Mexico (2); Panama (3); unknown locality (1). Sarda sarda-western Atlantic. Total 86 specimens (118-637 mm FL). Dissected 2 (333). New Jersey (1); Miami, Fla. (1). Measured 29 (228-637). North America 12 (257-637): Massachusetts (4); New York (3); Chesapeake Bay (2); Florida (1); Cuba (1). Gulf of Mexico 5 (228-321): Florida (1); Texas (4). South America 12 (202-450): Gulf of Carioca, Venezuela (6); Brazil (5); Mar del Plata (1). Counts only 51 (118-572). North America 39 (118-572): Massachusetts (24); Rhode Island (3); New York (3); New Jersey (1); Chesapeake Bay (2); Maryland (3); Florida (1). Gulf of Mexico 11 (103-400): Florida (3); Mississippi delta (3); Texas (1). South America 5 (214-570): Venezuela (1); Brazil (3); Mar del Plata (1). Skeletons 9 (333-577). Massachusetts (2); Con- necticut (1); New York (1); New Jersey (1); Florida (1); exact locality unknown (3). Sarda sarda-eastern Atlantic. Total 62 specimens (104-680 mm FL). Dissected 5 (363-504). Azores (3); Tunisia (1); Gulf of Guinea (1). Measured and counted 30 (260-600). Atlantic Europe 9 (418-600): Norway (5); Spain (1); Azores (3). Mediterranean 8 (260-564). Black Sea 1 (550). Gulf of Guinea 10 (305-478). Port Elizabeth, South Africa 2 (447-517). Counts only 31 (104-680). Europe 2 (482-670). Mediterranean 18 (187-487). Black Sea 9 (104-680). Gulf of Guinea 2 (366-375). Skeletons 6 (363-504). Azores (3), Tunisia (1); Gulf of Guinea (2). ACKNOWLEDGMENTS For permission to examine specimens in their institutions, or for donating specimens to the USNM collections, we thank the following: Shelton P. Applegate (LACM); Maria Luisa Azzaroli (MSUF); Adam Ben-Tuvia (SFRS); Reeve M. Bailey (UMMZ); Marie-Louise Bauchot(MNHN); 519 FISHERY BULLETIN: VUL. 73, NO. 3 M. Boeseman (RMNH); James E. Bohlke (ANSP); J. Cadenat (formerly at Institut frangais d'Afrique Noir, Dakar, Senegal); Joseph F. Copp (SIO); Myvanwy Dick (MCZ); William N. Esch- meyer (CAS); Thomas H. Eraser (formerly at RUSI); Carl George (formerly at American University of Beirut); Charles Gruchy (NMC); Robert K. Johnson (FMNH); Paul Kahsbauer (NHMV); W. Klausewitz (SMF); Robert J. Laven- berg (LACM); Rogelio B. Lopez (MACN); John M. Mason, Jr. (WHOI); Frank J. Mather III (WHOI); R. J. McKay (formerly at WAM); J. Moreland (DM); Ian S. R. Munro (CSIRO); Eugene L. Nakamura (Gulf Coastal Fisheries Center Panama City Laboratory, NMFS); Th. Monod (formerly at Institut fran^ais d'Afrique Noir, Dakar, Senegal); J0rgen G. Nielsen (ZMK); G. Palmer (BMNH); John R. Paxton (AMS); Per Pethon (ZMO); Thomas Potthoff (TABL); Jay Quast (AB); Helen Randall (BPBM); William J. Richards (TABL); C. Richard Robins (UMML); Richard Rosenblatt (SIO); Pearl Sonoda (CAS); the late H. Steinitz (HUJ); Frank H. Talbot (AMS and, previously, SAM); Enrico Tortonese (MSNG); Boyd Walker (UCLA); Martin Wiley (CBL); and Norman J. Wilimovsky (UBC). Frozen material, vital to this project, was ob- tained through the much appreciated efforts of: Tokiharu Abe (University of Tokyo); Adam Ben- Tuvia (SFRS); R. Budd (Sydney Fish Market); Norma Chirichigno (Instituto del Mar del Peru, Callao); William P. Davis (formerly at Medi- terranean Marine Sorting Center, Khereddine, Tunisia); C. E. Dawson (Gulf Coast Research Laboratory, Ocean Springs, Miss.); John E. Fitch (California Department of Fish and Game); Jeffrey B. Graham (Smithsonian Tropical Research Institute, Balboa, Panama); Frank J. Hester (formerly at Southwest Fisheries Center Honolulu Laboratory, NMFS); Meredith Jones (USNM); G. L. Kesteven (CSIRO); W. L. Klawe (Inter-American Tropical Tuna Commission, La JoUa, Calif.); Leslie W. Knapp (Smithsonian Oceanographic Sorting Center); A. M. Olsen (CSIRO, Hobart, Tasmania); the late Al Pflueger (Miami); Robert V. Miller (NMFS, Washington, D.C.); Ira and Roberta Rubinoff (Smithsonian Tropical Research Institute, Balboa, Panama); Paul J. Struhsaker (Southwest Fisheries Center Honolulu Laboratory, NMFS); and Charles Wenner (Virginia Institute of Marine Science). Work at the Australian Museum was made pos- sible through the National Marine Fisheries Ser- vice and the Trustees of the Australian Museum, its Director, Frank H. Talbot, its Curator of Fishes, John R. Paxton, the Director of New South Wales State Fisheries, Donald D. Francois, and the chief Market Inspector, R. Budd. A trip to Tunisia to study Orcynopsis was arranged by William P. Davis, former Director of the Medi- terranean Marine Sorting Center, Khereddine, Tunisia. Jack Marquardt and his staff at the Smithsonian library were most helpful in finding and obtaining early or obscure references to boni- tos. The figures, which are an integral part of this paper, were drawn by Keiko Hiratsuka Moore. Radiographs were taken by George Clipper. Typ- ing, retyping, proofreading, xeroxing, and all the other necessary clerical work was done by Arleen McClain, Partheina Mackabee, Sara E. Collette, and Joseph Russo. Robert H. Gibbs, Jr., Steven Gray, Linda Pushee Mercer, and Joseph Russo as- sisted with some dissections. E. H. Ahlstrom and Thomas Potthoff provided valuable comments on the caudal complex section as did John E. Fitch on the otolith section. Drafts of the entire manuscript were reviewed by Daniel M. Cohen, W. L. Klawe, Izumi Nakamura, Thomas Potthoff, and Stanley H. Weitzman. KEY TO THE SPECIES OF SARDINI la. Jaw teeth tiny, 40-55 on each side of upper and lower jaws; gill rakers fine and numerous, total of 70-80 on first arch; body elongate, snout to second dorsal 610-654 thousandths of fork length; maxilla short, 354-379 thousandths of head length Allothunnus fallai Serventy lb. Jaw teeth larger and more prominent, 10-30 on each side of upper and lower jaws; total gill rakers on first arch 8-27; body less elongate, snout to second dorsal 481-610 thousandths of fork length; maxilla longer, 431-557 thousandths of head length 2 520 COLLETTE and CHAO: SYSTEMATICS AND MORPHOLOGY OF THE BONITOS (SARDINI) 2b 3a, 3b, 4a 2a. Five to ten narrow, dark, longitudinal stripes on upper part of body; no teeth on the tongue; spleen prominent in posterior third of body cavity in ventral view Sarda 3 Body either without stripes or with dark spots above the lateral line and longitudinal dark stripes below; two patches of teeth present on tongue; spleen either concealed or located in anterior third of body cavity in ventral view 6 Spines in first dorsal fin 2023; total vertebrae 50-55 S. sarda (Bloch) Spines in first dorsal fin 17-19; total vertebrae 43-46 4 Total gill rakers on first arch 8-13; supramaxilla narrow (see Figure 32e) S. orientalis (Temminck and Schlegel) 4b. Total gill rakers on first arch 19-27; supramaxilla wider (see Figure 32c-d) 5 5a. Total gill rakers on first arch 19-21; pectoral rays 25-27, modally 26; teeth sometimes present on vomer; length of first dorsal base 315-343 thousandths of fork length; maxilla 503-539 thousandths of head length S. australis (Macleay) 5b. Total gill rakers on first arch 23-27; pectoral rays 22-26, modally 24 or 25; teeth never present on vomer; length of first dorsal base 267-314 thousandths of fork length; maxilla 460-503 thousandths of head length S. chiliensis (Cuvier) 6a. Body with dark spots above lateral line and dark longitudinal stripes below (see Figure la); spines in first dorsal fin 16-18 Cybiosarda elegans (Whitley) 6b. Body without a prominent pattern of stripes or spots (see Figure 2); spines in first dorsal fin 12-15 7a. Pectoral rays 21-23; small conical teeth in jaws; total gill rakers on first arch usually 14 or more; interpelvic process bifid; spleen not visible in ventral view; laminae in olfactory rosette 25-28; interorbital width 239-310 thousandths of head length Orcynopsis unicolor (Geoffroy St. Hilaire) 7b. Pectoral rays 25-28; jaw teeth very large and conspicuous; total gill rakers on first arch usually 13 or fewer; interpelvic process single; spleen visible on right side of body cavity in ventral view; laminae in olfactory rosette 48-56; interorbital width 321-400 thousandths of head length Gymnosarda unicolor (Ruppell) PART 1. COMPARATIVE MORPHOLOGY The morphological characters useful for distin- guishing the species of bonitos and for evaluating their phylogenetic relationships are divided into six categories: color pattern, scales, morphometry, meristics, soft anatomy, and osteology. Color Pattern The most strikingly colored species of the Sar- dini, and perhaps the entire family Scombridae, is clearly Cybiosarda elegans (Figure la). The light venter has several stripes reminiscent of the skip- jack tuna, Katsuwonus pelamis (Linnaeus). The dorsum is covered with black spots over a deep blue background. The high first dorsal fin is jet black anteriorly and white posteriorly. The anal and second dorsal fins are yellow. Orcynopsis unicolor (Figure lb) has a high black first dorsal fin as in Cybiosarda, but there the similarity ends because adult Orcynopsis have only a faint mottled pattern that has been deliberately exaggerated in the figure. All species of Sarda (Figure Ic) have stripes along their backs but the number of stripes and their alignment varies both interspecifically and intraspecifically. Sarda australis has stripes on the venter as well as on the dorsum. Sarda also has a black first dorsal fin but it is lower and longer than in Cybiosarda and Orcynopsis. Gymnosarda unicolor (Figure 2a) is deep blue without any dis- tinct pattern; Allothunnus fallai (Figure 2b) also lacks distinctive markings. Color plates have been published of all the species of bonitos eoccept Allothunnus. Paintings of three Australian species by George Coates were published by Marshall (1964, 1966): Cybiosarda elegans (fig. 345), Sarda australis (fig. 348), and Gymnosarda unicolor (fig. 342). Color illustrations of Orcynopsis unicolor were published by Lozano y Rey (1952, pi. 41, fig. 2-800-mm adult and fig. 3-150-mm juvenile) and by Bini (1968:39). Sarda sarda was illustrated by La Monte (1945, pi. 8; 1952, pi. 17), Lozano y Rey (1952, pi. 39, fig. 4-500- mm adult), and Bini (1968:37). Walford (1937, pi. 521 FISHERY BULLETIN: VOL. 73, NO. 3 522 COLLETTE and CHAO: SYSTEMATICS AND MORPHOLOGY OF THE BONITOS (SARDINI) Figure l.-Diagramatic lateral views of three species of Sardini to show general pigment pattern, extent of corselet (coarse stippling), and parts of the body covered by smaller scales (fine stippling), a. Cybiosarda elegans, New South Wales, 337 mm FL, USNM 259407-F2. b. Orcynopsis unicolor, Tunisia, 312 mm FL, USNM 206526. c. Sarda sarda, Gulf of Mexico, 287 mm FL, USNM 118646. 38) includes color photographs of a northeastern Pacific Sarda chiliensis and an eastern Pacific S. orientalis. Scales In bonitos, the body scales are cycloid and usually small. Those on the corselet, along the fin bases, and along the lateral line are larger and more elongate. The predorsal and opercular scales are larger and are embedded under the skin. No scales are present on the snout, the interorbital area, or on the fins. Posterior to the corselet, the distribution of scales differs among the genera of bonitos (Figures 1, 2). Species of Sarda have their body completely covered with small scales except for the distal portion of the caudal keels (Figure Ic). Allothunnus has the dorsal half of the body covered with scales (Figure 2b), but they do not extend onto the caudal keels, although they do cover the base of the caudal fin. Serventy (1948) described the type of Allothunnus fallai as hav- ing its whole body covered with scales. But later authors, Talbot (1960), Olsen (1962), and Nakamura and Mori (1966), all indicated that the minute scales of Allothunnus are present only on a Figure 2.— Diagrammatic lateral views of two species of Sardini to show extent of corselet (coarse stippling) and parts of body covered by smaller scales (fine stippling), a. Gymnosarda unicolor, Tahiti, 446 mm FL, ANSP 93818. b. Allothunnus fallai, New Zealand, 642 mm FL. 523 FISHERY BULLETIN: VOL. 73, NO. 3 the dorsal half of the fish. A patch of scales is also present around the base of the pelvic fins. Cybiosarda has a band of scales dorsally extending along the entire midline (Figure la). Ventrally, scales are present around the base of the pelvic fins and a broad band of scales extends from the anal fin origin posterodorsally to the caudal peduncle. The peduncular region is entirely covered with scales except for the distal margin of the caudal keel. Orcynopsis (Figure lb) has fewer scales than Cybiosarda. The band of scales along the dorsal midline is narrower and ends at the dorsal finlets. Ventrally, Orcynopsis has scales around the bases of the pelvic and anal fins. The caudal peduncle is naked except for the caudal keel. Gymnosarda is completely naked posterior to the corselet except for the lateral line, dorsal fin base, and caudal keel (Figure 2a). The corselet, composed of enlarged scales, is well defined in the pectoral region of bonitos. It ex- tends from the dorsal end of the gill slit to the tip of the pectoral fin, except in Sarda and Allothun- nus. Anterior and ventral to the pectoral fin base, the scales are smaller than on other parts of the corselet in bonitos. In Sarda, an extra wing of the corselet extends dorsally toward the origin of the first dorsal fin. Allothunnus has the most exten- sive corselet, covering most of the area between the first dorsal fin base and the pectoral fin. distances between the snout and the origins of the anal and second dorsal fins. Cybiosarda and Or- cynopsis both have high first dorsal, second dorsal, and anal fins compared to other bonitos. Gym- nosarda has a differently shaped head than do other bonitos: the interorbital distance is much wider, the eyes are larger, the postorbital distance is shorter, and the distance between the origins of the pectoral and pelvic fins is much larger. In ad- dition, Allothunnus has large eyes and a very short snout and maxilla. Because of small sample size, restricted geographical distribution, or both, morphometric data were combined for each of four species: Cybiosarda elegans, Sarda australis, Gymnosarda unicolor, and Allothunnus fallai. Three popula- tions of Orcynopsis unicolor are compared: Israel, Lebanon, and Tunisia. The southeast Pacific population of Sarda chiliensis (nominal S. c. chiliensis) is compared with the northeast Pacific population {S. c. lineolata). The population of Sar- da orientalis in the eastern tropical Pacific (nominal S. o. velox) is compared with the only other suflficiently large sample, northwest Pacific. Three populations of S. sarda are compared: west- ern Atlantic, Mediterranean Sea (including the Black Sea), and the Gulf of Guinea. Meristic Characters Morphometric Characters Twenty-six measurements, in addition to fork length, were routinely made on all specimens des- tined to be dissected to insure that these data would be available if needed. Preserved material was also measured until an adequate sample was obtained. Measurements follow the methods of Marr and Schaefer (1949) as modified by Gibbs and Collette (1967). Morphometric characters can be used to separate genera, species, and populations within species. Tables showing the 26 characters as thousandths of fork length and 8 characters as thousandths of head length are presented in the systematic section of the paper. Most of the characters are best used at the species level; therefore, only a summary table of the means of proportions (Table 1) is presented in this section. Orcynopsis is short-bodied and short-headed. It has shorter snout-anal and snout-second dorsal distances than do the other bonitos. Cybiosarda is also relatively short-bodied. Allothunnus is the most elongate of the bonitos and has the greatest Countable structures are of special value systematically because they are relatively easy to record unambiguously and because they are easy to summarize in tabular fashion. Meristic characters that have proved valuable systemat- ically in the Sardini include numbers of fin rays (first dorsal spines, second dorsal rays, dorsal finlets, anal rays, anal finlets, pectoral rays), gill rakers, teeth (especially on the upper and lower jaws), vertebrae, and laminae in the olfactory rosettes. Olfactory laminae are discussed as the last section under soft anatomy. The other meris- tic characters are discussed in the relevant os- teological sections of the paper. Soft Anatomy The relative position, shape, and size of the various internal organs provide valuable diagnos- tic characters. Within the genus Sarda, these characters are useful at the species level. For pur- poses of discussion, the characters in the soft ana- tomy are divided into five sections: viscera, vas- 524 COLLETTE and CHAO: SYSTEMATICS AND MORPHOLOGY OF THE BONITOS (SARDINI) Table l.-Morphometric comparison of species and populations of Sardini. Means as thousandths of fork length or head length. Character <0 c R) o> ■2 u 0 CO Q. m 0 to — u, < P CO CO c CO 0 ra <0 ::k CO 0 3 0 c S. M .O) CD E CO E 0 -c: 5) 2 c CO 3 3 Cl. ■C ■c 0 Co 3 CI. 0 a 0 0 ■Q 3 ■C Q. CO 0 ^ 0 1 <0 .9. 5 5 QK O) c i| CB -5 li 0) CI c III 9) ^ O 05 03 CD I- 3 S 3 5 .0 '~ c s « £ -c 2 o Q O 3 o . ^ ■= TO (0 (0 O) E £ CO CO 3 3 ■c -s o -c: CJ CD CD O) CD < 1 636 Table 4.— Species of myctophids and number taken per tow for stations given in Table 2. 50 m 100 m 200 m 300 m 450 m 500 m Species Stn. 48 49 58 68 57 69 61 67 52 76 60 78 Electrona rissoi — — — — — — — 1 — — — — Hygophum proximum 1 2534 631 — 1216 Diogenichthys atlanticus 3 3 4 — — 1 1 — — 1 — 5 Myctophum aurolaternatum — — — — — 1 — — — — — — Myctophum asperum 3 — — 1 — — — — — — — — Myctophum nitidulum — 2 1 1 1 — — — 1 — — — Myctophum obtusirostrum 1 — — 1 — — 2 — — — 1 — Myctophum selenoides — — 3 — — — 1 — — — — — Myctophum spinosum — 2 2 — 1 1 1 — — — — — Symbolophorus evermanni 5 8983 — 12 — 1 — 1 Diaphus drachmani — — — — — — 1 — — 1 — — Diaphus elucens 7 17 1 — 1 — — 3 1 1 — — Diaphus Iragilis 5 6 122 — 22 — — — — Diaphus jenseni — — 56 5 2 4 2 1 — — — — Diaphus longleyi — — 31 6 4 10 1 1 4 — 3 2 Diaphus lucidus — — 4 4 — — — — — — — — Diaphus luetl.■§ o •« o 2 4) II OJ Q ^ 5S OS D O ( J 01 X ;iAj/-0N) 33NvaNnav o I in I X JC 6 o E ^ (/> (u ^^™ en o a. o k. C V u o en o X in X ' a O o a. a o .,- k_ o o < X a-> ^^^ r- m _ en" « OJ ^^ CO < o ■^ I- en ~~> — o o CD O o o o o o o o {,1^/ON) 3DNvaNng\/ iiAi/oN) 33NvaNnav o o 648 o X in X X < 1^ (T) ^> 1 5 a F ^^^ a> en 1 o — o — i*— w O a. o < a. a. o cn to CD < U3 jg 23 C .-I •2 i — o 2 < CO CD < U3 CD CO (D CT) (j.oi »£i^/ON) 3DNvaNnav (tW/ON) aoNVQNnav 649 Table 3.— Total relative density and frequency of occurrence of the copepod species taken within 18 km of the Oregon coast, during 1969, 1970, and 1971. Relative density is average number of individuals per cubic meter in samples in which the species occurred. The table entries represent the sums of relative density at each of four stations. A total of 33 samples were collected in 1969, 44 in 1970, and 40 in 1971. Total relative density Frequency Species' 1969 1970 1971 1969 1970 1971 Calanus sp. 1,475.2 482.5 436.1 32 44 40 C. tenuicornis 1.3 5.1 7.9 1 4 14 Eucalanus bungii 21.1 3.1 9.0 13 10 15 Paracalanus parvus 80.7 147.3 16.8 29 21 20 Pseudocalanus sp. 23,776.3 6,682.4 3,994.5 33 44 40 Microcalanus pusillus 2.2 18.5 1.8 4 17 2 Clausocalanus arcuicornis 0 1.4 4.0 0 3 7 C. pergens 20.2 5.9 6.6 5 5 9 Clausocalanus immatures 0 0.5 2.1 0 5 2 Ctenocalanus vanus 4.8 31.0 11.0 5 7 16 Aetideus pacificus 1.5 2.3 2.7 4 4 1 Gaidius immatures 2.5 3.4 3.7 3 3 2 Racovitzanus antarcticas 0.6 2.1 1.3 1 2 2 Scolecithricella minor 9.3 4.6 16.0 7 14 16 Metridia lucens 21.4 16.3 48.6 18 29 26 M. pacilica 6.7 2.9 3.7 2 5 6 Lucicutia llavicornis 0 0.4 2.0 0 1 9 Centropages abdominalis 371.8 686.2 110.7 29 42 23 Epilabidocera longipedata 2.9 10.8 0 5 6 0 Acartia clausii 1,178.1 6,045.1 414.9 31 37 29 A. longiremis 1,078.2 1,509.2 1,331.1 33 44 38 A. tonsa 130.7 27.5 0 31 19 0 Tortanus discaudatus 17.5 12.3 0 3 10 0 Oithona similis 369.9 275.0 416.2 32 42 40 0. spinirostris 19.9 16.0 55.5 19 28 31 Corycaeus anglicus 2.8 8.0 3.6 2 10 3 'The following species were found in less than five samples: Calanus plumchrus, Gaetanus immatures, Paraeuchaeta japonica immatures, Candacia bipinnata, Eurytemora thompsoni, Rhincalanus nasutus, Oncaea borealis, 0. tenella, O. media hymena, O. mediterranea, Sapphirina sp., and Microsetella sp. Table 4. -Total relative density and frequency of occurrence of other holoplanktonic taxa and meroplankton taken within 18 km of the coast during 1969, 1970 and 1971 upwelling seasons. Entries are sums of average abundances at each of four stations. Total relative density Frequency Species 1969 1970 1971 1969 1970 1971 Calanus nauplil 119.5 695.5 172.7 21 40 28 Other Copepod nauplii 43.1 68.1 52.3 10 20 20 Amphipods 8.5 18.5 15.7 5 15 14 Euphausiid nauplii 46.3 85.9 84.0 5 26 18 Euphausild calyptopis 13.3 14.5 17.2 4 17 11 Euphausiid furcilia 30.2 13.6 17.7 14 20 10 Thysanoessa spinilera 35.4 4.0 87.3 2 7 11 Evadne nordmanni 73.7 58.9 9.8 17 26 2 Podon leukarti 2.8 115.3 5.2 2 12 1 Pteropods 10.2 24.6 60.6 11 22 35 Chaetognaths 89.4 50.3 30.8 25 33 34 Oikopleura 69.2 85.7 66.5 11 15 21 Ctenophores 6.0 2.5 34.9 7 5 19 Scyphomedusae 22.9 70.9 22.8 13 28 22 Decapod shrimp mysis 142.7 52.6 45.3 16 24 22 Barnacle nauplii 59.3 168.3 231.4 8 32 23 Barnacle cypris 4.4 64.0 8.3 2 19 10 Polychaete post-trochophores 16.2 20.1 21.4 5 23 15 Bivalve veligers 170.5 258.9 68.3 20 40 27 Gastropod veligers 28.9 79.2 42.2 16 33 23 Hydromedusae 6.1 3.2 10.3 2 2 11 Unidentified annelid without parapodia 8,2 23.1 35.8 3 3 16 Pluteus 0 16.0 117.6 0 5 11 Large round eggs (fish) 36.8 25.0 17.8 11 13 12 Small round eggs 1870.1 168.7 226.1 10 28 25 Euphausiid eggs, early 55.0 686.1 449.6 11 29 24 Euphausiid eggs, late 70.0 57.5 39.6 2 16 14 Other fish eggs 19.1 35.1 34.3 12 18 18 'Biased by a single observation of 760 individuals/m^. 650 Only a few of the taxa in these tables had similar average abundances in each season. Some of the taxa can be assigned "good" and "bad" years on the basis of either their abundance or frequency of occurrence in samples. Others cannot be assigned with much confidence. Accordingly, on the basis of frequency of occurrence, 1971 was the "best" year for the following copepods: Calanus tenuicornis Dana, Clausocalanus arcuicornis Dana, Ctenocalanus vanus Giesbrecht s.l. and Lucicutia flavicornis Glaus. On the basis of abundance, the following categories can be added to the list: Me- tridia lucens Boeck, 0. spinirostris Glaus, the euphausiid Thysanoessa spinifera Holmes, the pteropod Limacina helicina (Phipps), ctenophores, hydromedusae, echinoderm pluteus larvae, and unidentified annelids. Using the same criteria, 1971 was the poorest year for the copepods Paracalanus parvus Glaus, Pseudocalanus, Ae- tideus pacificus Brodskii, Centropages ab- dominalis, Acartia clausii, A. tonsa Dana, Tor- tanus discaudatus Thompson and Scott, and Epilabidocera longipedata Sato ( = E. amphitrites McMurrich). Also poorly represented in 1971 were the cladoceran Evadne normanni (Loven) and bivalve mollusc veligers. A definite pattern emerges from the above classifications. All of the copepod species having their best year in 1971 are basically offshore, warmwater species that can always be found well off the Oregon coast (Peterson and Anderson 1966; Peterson 1972) and which seem to l>ave their highest abundances to the south (Fleminger 1967). Those species which had their poorest year in 1971 are all nearshore, coastal species. In fact, some of the neritic species which had their "best" years in 1969 or 1970 did not even occur in 1971 (Epilabidocera longipedata, T. discaudatus, and A. tonsa; see Figure 6). Fleminger (1967) listed Paracalanus parvus and A. tonsa as temperate- subtropical neritic, and Pseudocalanus, A. clausii, T. discaudatus, and E. longipedata as boreal- temperate neritic. Centropages mcmurrichi is also a boreal neritic species (Gameron 1957). The at- tribute shared by each of the animals having a "poor" year in 1971, is restriction to the neritic zone. Warmwater or cold-water affinities seem unimportant. Even though surface temperatures were much higher in 1971, two important animals with norther affinities and not narrowly restricted to the neritic zone, maintained the same level of abundance as they had in 1969 and 1970: A. lon- giremis and 0. similis. 100 50 O z ISO 100 _1L 150 100 50 150 100 50 ..Jl lij uL J A S 1969 NH- 10 NH- 5 NH-3 NH- I I 971 Figure 6.-Density of Acartia tonsa at NH 1, NH 3, NH 5, and NH 10 in the upwelling seasons of 1969, 1970, and 1971. There are important differences between the good upwelling years of 1969 and 1970. A number of taxa were very abundant only in 1970: the copepods Microcalanus pusillus Sars and A. clausii, Calanus nauplii, the cladoceran Podon leukarti (Sars), barnacle cypris, and gastropod veligers. The year 1969 was the better year for two dominant copepod species (Calanus sp. and Pseudocalanus sp.), for the copepods Eucalanus bungii Giesbrecht and A. tonsa, and for shrimp larvae. We do not fully understand these observa- tions. The greatest share of the taxa listed in Tables 3 and 4 seem to be either equally abundant over all three upwelling seasons (copepod nauplii, euphausiid life history stages, Oikopleura sp., polychaete post-trochophores, small round eggs, euphausiid eggs, and fish eggs), or have uncertain or no relationship to the upwelling seasons. These animals include the copepods Clausocalanus per- gens Ferran, Scolecithricella minor Brady, Me- tridia pacifica Brodskii, Racovitzanus antarcticus Giesbrecht s.l., Gaidius sp., and chaetognaths, barnacle nauplii, and scyphomedusae. DISCUSSION Upwelling along the Oregon coast was relatively weak in 1971. It was seldom strong enough to create the low temperature-high salinity condi- tions close to the beach that are characteristic of the process. This occurred because there were no 651 sustained periods of southward wind in 1971. There were only four "upwelling events": 2-10 May, 16-25 May, 20 June-2 July, and 10-24 July. Each of these events was unusual compared to those of other years in that the wind also had a substantial eastward component. There were also four storms from the southwest of the sort that characterize the winter period on the Oregon coast: 11-15 May, 16-23 June, 7-9 July, and 27-31 August. It is expected that under these conditions surface waters from offshore would have been more prevalent in the nearshore zone in 1971 than in other years. The composition of the plankton observed in 1971 within 18 km from shore is in agreement with that hypothesis. Comparison of onshore-offshore hydrographic sections for upwelling events and for intervals of low southward winds (Smith 1974) suggests an explanation for the fact that NH 10 does not show the same degree of year-to-year variations as sta- tions nearer shore. During the upwelling season all isopycnals below 15 m slope upward toward the shore at least as far seaward as 30 km. Upward sloping extends to the shore during upwelling events, but during lapses of the southward wind the isopycnals come to slope downward toward shore from 10 km seaward to the beach. Seaward of 10 km they continue to slope upward despite prolonged lapses. It seems likely that coastal up- welling only takes the form of pulsed events in this most inshore zone. Thus it is reasonable that the low frequency and amplitude of upwelling events in 1971 only had a pronounced effect on the planktology at stations less than 10 km from shore. On the other hand, Hubbard and Pearcy (1971) demonstrated marked changes in the species composition of the salp fauna off Oregon at dis- tances beyond 28 km from shore in 1963, another year of anomalously low coastal upwelling (Bakun 1973). The detailed relationship between inshore and offshore plankton changes as correlated with year-to-year weather variations cannot yet be deduced. The length, frequency, and spatial extent of the data set necessary to deal with this problem probably puts it beyond our reach. There is a suggestion in the data that intense upwelling events rather immediately result in high zooplankton abundance at the NH 1 and NH 3 stations. Huge population peaks on 10, 18, and 25 July 1969 and 23 June and 2 July 1970 were as- sociated with periods of intense upwelling. The high density found on 22 June 1969, however. followed a 41-day period of little or no north wind. It seems most likely that peak densities are simply reached at about the same time each year, namely late June and early July. We do not as yet know the relationship between copepod developmental schedules and the seasons in this area well enough to decide this issue with any certainty. Further analysis of our data as it bears on this point is planned. Summers of below average upwelling like 1971, together with the resultant reductions in primary and secondary production, probably have impor- tant effects upon nearshore fisheries. A statistical link exists between summer upwelling strength and Dungeness crab production (Peterson, 1973). A strong upwelling season results in a heavy crab catch iy2 yr later. The Dungeness crab catch for the 1972-1973 season was one of the lowest on record. Other fisheries seem to have been similarly af- fected. The Fish Commission of Oregon has documented 1971 as a poor growth year for the shrimp Pandalus jordani (Robert L. Demory, Oregon Fish Commission, Newport, Ore., pers. commun.), coho salmon, Oncorynchus kisutch (Paul H. Reed, Oregon Fish Commission, Newport, Ore., pers. commun.), and razor clams Siliqua pa- tula (C. Dale Snow, Oregon Fish Commission, Newport, Ore., pers. commun.). The shrimp data are mean carapace length of Age I animals from the Coos Bay, Ore. area (lat. 43°15'N) and are as follows: 1969, 16.45 cm; 1970, 16.76 cm; and 1971, 15.87 cm. Averaged dressed weights of coho salm- on were 2.59 kg in 1969, 3.41 kg in 1970, and 2.68 kg in 1971. Razor clam lengths averaged 75.4 mm in 1969, (no data for 1970), 69.4 mm in 1971, 78.5 mm in 1972, and 83.5 mm in 1973. Some of these data may be better interpreted in terms of good growth years. As shown by the wind data (Figure 3), 1970 had many more days of upwelling inducing winds than 1969. Primary and secondary produc- tion should have been greater in 1970. Both shrimp and coho salmon were larger in 1970. Unfortu- nately no razor clam data were taken in 1970, but data for other years support the conclusion that 1971 was a poor growth year. ACKNOWLEDGME NTS The sampling program for this study was start- ed by Jefferson J. Conor and William G. Pearcy who graciously allowed us to participate. R. Gregory Lough, Peter Rothlisberg, and others 652 helped with sampling. William Gilbert provided the wind data. The work was supported by NOAA (U.S. Department of Commerce) Sea Grant Insti- tutional Grant No. 04-3-158-4. LITERATURE CITED Bakun, a. 1973. Coastal upwelling indices, west coast of North America, 1946-71. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-671, 103 p. BOURKE, R. H. 1972. A study of the seasonal variation in temperature and salinity along the Oregon-Northern California Coast. Ph.D. Thesis, Oregon State Univ., Corvallis, 107 p. Burt, W. V., and B. Wyatt. 1964. Drift bottle observations of the Davidson Current off Oregon. In K. Yoshida (editor). Studies on oceanography, p. 156-165. Univ. Tokyo Press, Tokyo. Cameron, F. E. 1957. Some factors influencing the distribution of pelagic copepods in the Queen Charlotte Islands area. J. Fish. Res. Board Can. 14:165-202. Collins, C. A., C. N. K. Mooers, M. R. Stevenson, R. C. Smith, AND J. G. Pattullo. 1968. Direct current measurements in the frontal zone of a coastal upwelling region. J. Oceanogr. Soc. Jap. 24:295-306. Cross, F. A. 1964. Seasonal and geographical distribution of pelagic copepods in Oregon coastal waters. M.S. Thesis, Oregon State Univ., Corvallis, 73 p. Cross, F. A., and L. F. Small. 1967. Copepod indicators of surface water movements off the Oregon coast. Limnol. Oceanogr. 12:60-72. Fleminger, a. 1967. Distributional atlas of calanoid copepods in the California Current Region. Part 2. Calif. Coop. Oceanic Fish. Invest., Atlas 7, 213 p. Frost, B. W. 1974. Calanus marshallae a new species of calanoid copepod closely allied to the sibling species C.finmarchicus and C. glacialis. Mar. Biol. (Berl.) 26:77-99. Hebard, J. F. 1966. Distribution of Euphausiacea and Copepoda off Oregon in relation to oceanographic conditions. Ph.D. Thesis, Oregon State Univ., Corvallis, 85 p. Hubbard, L. T., and W. G. Pearcy. 1971. Geographic distribution and relative abundance of Salpidae off the Oregon coast. J. Fish. Res. Board Can. 28:1831-1836. Huyer, a. 1974. Observations of the coastal upwelling region off Oregon during 1972. Ph.D. Thesis, Oregon State Univ., Corvallis, 149 p. Laurs, R. M. 1967. Coastal upwelling and the ecology of lower trophic levels. Ph.D. Thesis, Oregon State Univ., Corvallis, 121 p. Lee, W. 1971. The copepods in a collection from the southern Oregon coast, 1963. M.S. Thesis, Oregon State Univ., Corvallis, 62 p. Pattullo, J., and W. Denner. 1965. Processes affecting seawater characteristics along the Oregon coast. Limnol. Oceanogr. 10:443-450. Peterson, W. K. 1972. Distribution of pelagic Copepoda off the coasts of Washington and Oregon during 1961 and 1962. In A. T. Pruter and D. L. Alverson (editors), The Columbia River Estuary and adjacent ocean waters, p. 313-343. Univ. Wash. Press, Seattle. Peterson, W. K., and G. C. Anderson. 1966. Net zooplankton data from the Northeast Pacific Ocean: Columbia River effluent area, 1961, 1962. Univ. Wash. Tech. Rep. 160, 225 p. Peterson, W. T. 1973. Upwelling indices and annual catches of Dungeness crab. Cancer magister, along the west coast of the United States. Fish. Bull., U.S. 71:902-910. PiLLSBURY, R. D. 1972. A description of hydrography, winds, and currents during the upwelling season near Newport, Oregon. Ph.D. Thesis, Oregon State Univ., Corvallis, 163 p. Smith, R. L. 1974. A description of current, wind, and sea level variations during coastal upwelling off the Oregon coast, July- August 1972. J. Geophys. Res. 79:435-443. 653 USE OF OTOLITHS TO SEPARATE JUVENILE STEELHEAD TROUT FROM JUVENILE RAINBOW TROUT^ J. T. Rybock^ H. F. H0RT0N^ AND J. L. Fessler' ABSTRACT Otolith nuclei were investigated as a means of separating juvenile steelhead trout, Salmo gairdneri, from juvenile rainbrow trout, S. gairdneri, in the lower Deschutes River, Oreg. An intensive recrea- tional fishery necessitated development of a technique for differentiation so that impact of the fishery on each race could be assessed independently. Investigations of adults and hatchery-reared young of S. gairdneri revealed that otolith nuclei of steelhead are significantly larger than those of rainbow, size of otolith nuclei does not change with growth of either race, and there are no differences in size of otolith nuclei due to sex or origin (wild vs. hatchery). Thus, size of otolith nuclei provides a means to differentiate effectively juvenile steelhead trout and juvenile rainbow trout regardless of sex or origin. Results also indicated that steelhead mature at a larger size than rainbow, egg size is directly related to body size of dam in both races, and size of otolith nuclei is likely determined by egg size. This paper reports on an investigation of growth characteristics of the sagittae, the largest of the otoliths, as a means to separate juvenile steelhead trout, Salmo gairdneri, from juvenile rainbow trout, S. gairdneri. The technology for such differentiation is presently lacking but is neces- sary for independent management of the two races in streams where they coexist. In the lower Deschutes River, Oreg., for example, the most in- tensive fishery for rainbow trout occurs during the first week in May when most steelhead smolts migrate; consequently, the catch may be composed of 22-80% juvenile steelhead (King 1966; Wagner and Haxton 1968). Precise knowledge of this catch composition at various locations and times would allow fisheries managers to manipulate fishing pressure so that most steelhead smolts escape capture during migration. Previously, otoliths have been used to differen- tiate stocks and races of salmonids. Kim (1963) found differences in the appearance and size of growth rings between spawning groups of sockeye salmon, Oncorhynchus nerka. The study most relevant to our investigation demonstrated that winter and summer races of steelhead trout can be separated on the basis of differences in size of the otolith nucleus (ON) (McKern et al. 1974). 'Technical Paper No. 3619, Oregon Agricultural Experiment Station, Corvallis, OR 97331. ■'Department of Fisheries and Wildlife, Oregon State Univer- sity, Corvallis, OR 97331. 'Research Division, Oregon Wildlife Commission, 303 Exten- sion Hall, Corvallis, OR 97331. The latter authors found that the otolith nucleus is formed early in steelhead trout embryos when all or a great part of nutrition comes from the yolk, and that ON size appears to be directly related to egg size. Also, egg size and fish length of sal- monids are often directly related (McFadden et al. 1965; Bulkley 1967; Galkina 1970), and steelhead trout are generally larger than rainbow trout at maturity. Therefore, we hypothesized that steelhead ON are suflficiently larger than rainbow ON to permit separation of juveniles of both races. We investigated this hypothesis via two series of observations. In the first, an indirect test of validity of the hypothesis, we compared measurements of ON of fish of known race and then determined whether these measurements changed with growth of fish or whether they were related to sex or origin. In the second series, we measured body size of adult fish, egg size of ripe dam, and ON size of fry hatched from these eggs, to determine if correlations between these varia- bles logically accounted for differences in ON sizes. All investigations were conducted on adults and hatchery-reared fingerlings of summer steelhead trout and resident rainbow trout cap- tured in 1971-73 from the lower Deschutes River. METHODS AND MATERIALS Study Area The study area was the lower 100 miles of the Manuscript accepted December 1974. FISHERY BULLETIN: VOL. 73, NO. 3, 1975. 654 Deschutes River in north central Oregon (Figure 1). The Deschutes River drains approximately 10,400 square miles, or nearly 11% of the land area of Oregon. Its western tributaries stem from the Cascade Mountains, while eastern tributaries drain Oregon's high plateau. Regulated river flows below Pelton Dam average from 3,000 to 7,100 cfs. Important sport fish in the area include resident rainbow trout, summer steelhead trout, and Chinook salmon, 0. tshawytscha, (Montgomery 1971). Collection of Samples Otoliths were obtained from adult (>200 mm fork length [FL]) rainbow and steelhead (n = 101) COLUMBIA RIVER LOCKIT PELTON DAM Figure l.-Map of study area on the lower 100 miles of the Deschutes River, Oreg. sampled during routine Oregon Wildlife Commis- sion creel censuses at Webb's access road (at Buck Hollow Creek) and near Maupin (Figure 1) during August and September 1971 and 1972. Otoliths were removed with a punch described by McKern and Horton (1970). Each fish was measured (FL) and scales (ca. 20) were removed from an area below the origin of the dorsal fin and just above the lateral line. Race was determined from a com- bination of coloration, relative size, and analysis of scales (Maher and Larkin 1954). In most cases sex was determined from jaw conformation and oper- cular coloration (steelhead only), and occasionally from fisherman's observations if the fish had been cleaned. To determine origin, we examined steelhead for hatchery marks; hatchery-reared rainbows were distinguished by worn or rounded fins, excessive number of missing scales, and other abnormalities. Other adult fish (w = 92) were collected by elec- trofishing near Maupin, below Pelton Dam, and in Bakeoven and Trout creeks (Figure 1) in April- June 1971 and August 1972. Each fish was measured (FL), and race, sex, and origin were de- termined as above. Otoliths were removed by dis- section. In January 1973, 52 steelhead fingerlings were obtained from the stock of Deschutes River steelhead reared at Wizard Falls Hatchery (Oregon Wildlife Commission) on the Metolius River. These fish represented a random assort- ment of the offspring of ca. 150 females captured below Pelton Dam. Fork lengths were measured, and otoliths were removed by dissection. To determine body lengths of mature steelhead trout and rainbow trout, specimens were obtained by electrofishing in the lower Deschutes River in 1972. Fork lengths were measured, and race was determined from hatchery marks or coloration (migrating summer steelhead are more silvery than resident rainbow). For determination of ova size, adult steelhead were captured in late winter 1972 by trapping below Pelton Dam and were held in tanks until ripe. Twenty-two females were measured (FL), and a sample of eggs (ca. 100) was collected from each fish, fertilized, and allowed to water harden 8-22 h. From 20 to 60 eggs from each pairing were then measured volumetrically (10"^ ml) in a 25-ml burette. Rainbow trout were captured in spring 1972 by electrofishing in the main stem of the Deschutes River. Male-female pairs were individually 655 spawned, and, after water hardening, the eggs were transported to a laboratory in Corvallis to be hatched. Shortly after arrival, 20 eggs from each of 13 matings were measured as above. Fork lengths were later determined from the frozen dams and sires. To obtain samples for determination of possible correlation between egg size and ON size of the hatched fry, we randomly selected 10 fingerlings from each of eight available matings of rainbow trout individually hatched and reared in Corvallis (above). Fork lengths were measured, and otoliths were removed by dissection. Storage and Treatment of Otoliths The enveloping membrane (sacculus) was removed from each otolith prior to storage. Ini- tially, otoliths were stored dry in coin envelopes before transfer to a clearing solution. Because this method led to breakage of otoliths, later samples were placed in a clearing solution immediately after removal from the fish. Otoliths were cleared from 1 to 21 mo before examination; there was no apparent relationship between clearing time and readableness of the otolith. Samples were initially cleared in methyl salicylate. Because some otoliths did not clear sufficiently, a 50:50 mixture of glycerin and water (McKern et al. 1974) was used for the remainder of the samples, but this solution tended to increase the opacity of the entire otolith. Neither burning these otoliths on an asbestos pad over a bunsen burner nor clearing the otoliths in oil of cloves increased contrast between the opaque and hyaline parts. Consequently, it was difficult to discern the nucleus. Improved readings were obtained by applying drops of HCl to the medial surface of otoliths preserved in glycerin and water; this resulted in a dissolution of the medial lobes, a consequent thin- ning of the otolith, and clearer definition of den- sity patterns. This method is quick (a few milliliters of HCl applied for 2-4 min for a large otolith) and is easily controlled by periodic inspec- tion of the otolith during treatment. Because the edges of the otolith are dissolved, this method should not be used when age determinations are required. Terminology and Examination of Otoliths When viewed under reflected light on a black 656 background, the ON of S. gairdneri is hyaline with a narrow opaque ring around the border (Figure 2). The metamorphic check is a narrow hyaline ring delineating the nucleus (Kim and Koo 1963). For examination, otoliths were placed lateral sur- face up on black Plexiglas^ depression plates, illuminated with a beam of light at 45° and pho- tographed on 35-mm film through a microscope at 50 X . Panatomic-X film (ASA 32) was used, and the negatives were enlarged to 4x5 or 5x7 inches onto grade 3 or 4 (high contrast) paper. A stage micrometer was also photographed and enlarged at the same magnifications so that otolith measurements could be determined from the pho- tographs. The length and width of the nucleus (Figure 2) was measured from the photographs by using a compass and the corresponding pho- tograph of the micrometer. RESULTS AND DISCUSSION Size of Otolith Nucleus The linear correlation between ON length and width was strong in both rainbow trout (r = 0.838) ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. \ METAMORPHIC / CHECK i I I- o t [^LENGTH B Figure 2.-Illustration of (A) otolith and (B) otolith nucleus of Salmo gairdneri, with notation of measurements and ter- minology used. and steelhead trout (r = 0.916). Neither seemed easier to read. Primarily due to problems of developing methodology, 189 ON (29% of 641 examined) were not sufficiently distinct to permit measurement. Usually, the hyaline center of the nucleus was visible, but the metamorphic check could not be distinguished. Because measurement of the larger dimension would likely be more precise than widths, ON length was used in the following analyses. The mean ON lengths of steelhead trout (0.354 mm; n = 114) and rainbow trout (0.245 mm; n = 145) differed significantly (P< 0.001). The length- frequency plot of these data (Figure 3A) demon- strates an overlap of lengths. Most unexpected in this plot are the nucleus lengths for steelhead less than 0.26 mm. These values occur in direct proportion to the values for rainbow; also, these steelhead are from Wizard Falls Hatchery, where both rainbow and steelhead are reared. Perhaps these fish are rainbow offspring which were inad- vertently mixed with steelhead during hatchery operations or hybrids of the two species. The length-frequency plot of steelhead ON excluding those from Wizard Falls Hatchery increased the normality of the histogram (Figure 3B); the nadir ta = RAINBOW TROUT (n=l45) O = STEELHEAD TROUT (n=ll4) 16 20 24 28 32 36 40 44 48 52 ^ = RAINBOW TROUT (n= 145) EZ3 = STEELHEAD TROUT (n=72) 16 20 24 28 32 36 40 44 48 52 LENGTH OF OTOLITH NUCLEUS (10"^ mm) Figure 3.-Length-frequency distribution of otolith nuclei of (A) rainbow trout and all steelhead trout and (B) rainbow trout and steelhead trout excluding those from Wizard Falls Hatchery. All fish. were captured from the lower Deschutes River, Oreg., 1971-73. at 0.46-0.48 mm is probably due to the small sample size of each interval. Therefore, although overlap of ON size of the two races occurs between 0.28 mm and 0.34 mm, the race of most juvenile S. gairdneri from the Deschutes River can be determined reliably on the basis of this measurement. The histogram for rainbow more closely approximates a normal distribution, probably the result of a larger sample size and of the many sources of variation operating within a more narrow size range of spawning fish. The ON length of one adult rainbow was 0.48 mm. Although no hatchery marks were noticed, scale characteristics suggested a hatchery origin; because we have ob- served that almost all hatchery-reared rainbow succumb to Ceratomyxa before reaching maturity in the lower Deschutes River, this may have been a nonmigratory steelhead. In general, though, data from our samples do not support the suggestion of Wagner and Haxton (1968) that there may be a great number of such nonmigrants in the Deschutes River. To determine whether size of ON changes dur- ing growth of fish, we regressed length of ON against FL of fish: For rainbow, r = 0.060. For all steelhead, r = 0.694; however, if Wizard Falls fish are excluded, r = 0.018. Even with this exclusion, a wide range of steelhead FL (504-762 mm) was tested; and if the relationship was strong it should have been noticeable in these data. Mean length of ON was 0.339 mm for all females and 0.317 mm for all males; they are not sig- nificantly different (P >0.20). Also, the data sug- gest no significant male-female difference within either race. Mean length of ON of wild steelhead (0.395 mm; n = 52) was not significantly different (P >0.20) from that of hatchery-reared steelhead excluding those from Wizard Falls Hatchery (0.405 mm; n = 20). A similar comparison between adult hatchery-reared and wild rainbow cannot be made since there are few, if any, adult hatchery-reared rainbow in the lower Deschutes River (as men- tioned earlier, hatchery fish released in spring succumb to Ceratomyxa by summer). Fish, Egg, and ON Size Relationships The lengths of rainbow trout and steelhead trout from the lower Deschutes River are dis- tributed into discrete size ranges (Figure 4). Although these fish are not necessarily ready to 657 250r RAINBOW TROUT (n= 1,852) t _1 I I 1 1 L. STEELHEAD TROUT (n-- 1.816) _1 I 1_ 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 84 FORK LENGTH (cm) Figure 4. -Length-frequency distribution of mature (>20 cm) rainbow trout and steelhead trout from the lower Deschutes River, Oreg., 1972. spawn, the data indicate low likelihood of sig- nificant overlap in length of mature rainbow and steelhead trout. The mean egg size of steelhead (0.0936 ml) was significantly greater (P< 0.001) than that of rain- bow (0.0727 ml). Also, mean egg size was strongly correlated with length of female (r = 0.829 and 0.791 for rainbow and steelhead trout, respec- tively) (Figure 5), although there was much variability of mean egg sizes between fish of a similar length and of egg sizes from any one female. For some fish, the largest egg was twice the size of the smallest. The above r values between body size of dam and egg size are higher than those reported in many other investigations. Scott (1962) measured FL and egg weight of rainbow trout and found no significant correlation. Considering the narrow range of FL (231-264 mm) and the great variability of egg size within length classes, his 18 16 i 14 b 3 12 5 10 O = RAINBOW TROUT y = 0.341 + O.OISx (j; = 0.829) m-- STEELHEAD TROUT y = -2.974 + 0.196? (r = 0.791) 1 F !i 34 38 42 46 50 54 58 62 66 70 74 78 82 86 FORK LENGTH (cm) Figure 5.-Means and ranges of egg size plotted against length of dam for rainbow trout and steelhead trout from the lower Deschutes River, Oreg., 1972. results are not surprising. Galkina (1970) found that length of rainbow trout, S. irideus, was not highly correlated with mean egg weight (r = 0.48). Although eggs of average size were found in all his females, the smallest eggs were obtained only from smaller females and vice versa. McFadden et al. (1965) found a higher correlation (r = 0.73) between egg size and length of brown trout, S. trutta. Blaxter (1969), Galkina (1970), and Lindsey and AH (1971) cited numerous authors who examined this relationship in many species of fish; although most authors reported a wide range of egg sizes in females of similar length, there is general agreement that a direct, and often high, correlation exists between egg size and dam size. The presence or absence of any correlation between egg size and ON size of the hatched fish could not be determined directly for the rainbow trout groups reared in Corvallis (egg size had been measured for only four of the eight females whose offspring were available, and we considered this sample size too small). However, since there was a high correlation between egg size and length of dam (r = 0.829), this latter measurement was regressed against ON size of offspring from the eight matings (Figure 6). The r value of 0.489 and the overlap of ranges indicate the relationship is not strong; however, it is a positive correlation. Also, the small sample size (8), the narrow range of dam lengths (295-415 mm), the variation of egg size within any dam, and the substitution of FL for egg size are factors which may have obscured the real extent of the relationship of fish size to ON size. In summary, we found that because steelhead ON are larger than rainbow ON, size of ON does not change with growth of either race, and correlations of ON size between dams and sires and between wild and hatchery-reared fish of either race are insignificant, ON size is an effec- tive means of differentiating juvenile steelhead trout and juvenile rainbow trout regardless of sex or origin. We also concluded that because steelhead trout are larger at maturity than rain- bow trout and because egg size is a direct function of body size, eggs of steelhead trout are larger than those of rainbow trout; and although we did not conclusively demonstrate that ON size is directly related to egg size, other evidence was offered to support the hypothesis that larger egg size was the mechanism responsible for larger ON size in steelhead trout as compared to rainbow trout. 658 „ 34 E E 32 30 28 26 24 - 22 F 20 uj 18 - - -9 y = 1.610 + 0.023X (1= 0.489) 1 30 32 34 36 38 FORK LENGTH (cm) 40 42 Figure 6.-Regression of length of rainbow trout dam from the Deschutes River on length of otolith nucleus of offspring cul- tured in Corvallis, Oreg., 1972. (Circles are means, and vertical bars are ranges of observation.) ACKNOWLEDGMENTS This project was funded jointly by the Oregon Agricultural Experiment Station, Oregon State University, and the Research Division of the Oregon Wildlife Commission. F. H. Sumner, Scale Analyst (retired), Oregon Wildlife Commission, advised on scale reading. H. H. Wagner, Chief of the Research Division, Oregon Wildlife Commis- sion, gave many useful suggestions for the research and reviev^^ed the manuscript. R. E. Millemann and W. H. Staeger, Department of Fisheries and Wildlife, Oregon State University, also critiqued the manuscript. LITERATURE CITED Blaxter, J. H.S. 1969. Development: eggs and larvae. In W. S. Hoar and D. J. Randall (editors), Fish physiology 3:117-252. Academic Press, N.Y. BULKLEY, R. V. 1967. Fecundity of steelhead trout, Salmo gairdneri, from Alsea River, Oregon. J. Fish. Res. Board Can. 24:917-926. Galkina, Z. I. 1970. Dependence of egg size on the size and age of female salmon [Salmo salar (L.)] and rainbow trout [Salmo irideus (Gib.)]. J. Ichthyol. 10:625-633. KiM.W.S. 1963. On the use of otoliths of red salmon for age and racial studies. M.S. Thesis, Univ. Washington, Seattle, 63 p. Kim, W. S., and T. S. Y. Koo. 1963. The use of otoliths for age determination in red salm- on Univ. Wash., Fish. Res. Inst., Coll. Fish., Contrib. 147:17-19. King, D. N. 1966. Deschutes River summer steelhead. Oreg. State Game Comm., Cent. Reg. Adm. Rep. 66-3, 41 p. LiNDSEY, C. C, AND M. Y. AlI. 1971. An experiment with medaka, Oryzias latipes, and a critique of the hypothesis that teleost egg size controls vertebral count. J. Fish. Res. Board Can. 28:1235-1240. MAHER, F. p., AND P. A. Larkin. 1954. Life history of the steelhead trout of the Chilliwack River, British Columbia. Trans. Am. Fish. Soc. 84:27-38. McFadden, J. T., E. L. Cooper, and J. K. Anderson. 1965. Some effects of environment on egg production in brown trout (Salmo trutta). Limnol. Oceanogr. 10:88-95. McKeRN, J. L., AND H. F. HORTON. 1970. A punch to facilitate the removal of salmonid otoliths. Calif. Fish Game 56:65-68. McKern, J. L., H. F. HoRTON, and K V. Koski. 1974. Development of steelhead trout {Salmo gairdneri) otoliths and their use for age analysis and for separating summer from winter races and wild from hatchery stocks. J. Fish. Res. Board Can. 31:1420-1426. Montgomery, M. 1971. The Deschutes fishery. Oreg. State Game Comm. Bull. 26(3):3-5, 7. Scott, D. P. 1962. Effect of food quantity on fecundity of rainbow trout, Salmo gairdneri. J. Fish. Res. Board Can. 19:715-731. Wagner, H., and J. Haxton. 1968. Observations on the migration disposition of summer steelhead reared at Oak Springs and Wizard Falls Hatcheries. Oreg. State Game Comm., Res. Div., Prog. Rep., 5 p. 659 DESCRIPTION OF EGGS AND LARVAE OF YELLOWFIN MENHADEN, BREVOORTIA SMITH I' Edward D. Houde- and L. J. Swanson, Jr.' ABSTRACT Development of yellowfin menhaden, Brevoortia smithi, is described from eggs and larvae reared in the laboratory. Eggs were collected during November 1972 in Biscayne Bay, Florida. Eight embryos and 66 larvae and juveniles ranging from 3.7 to 36.2 mm standard length were used to describe development. Mean egg diameter was 1.27 mm, mean oil globule diameter was 0.15 mm, and mean yolk diameter was 1.07 mm. The perivitelline space averaged 16% of the egg diameter. Length at hatching was about 3.0 mm standard length. Larvae were fed on zooplankton and grew about 0.45 mm per day from the 4th until the 20th day after hatching at 26°C. Morphology, meristics, osteology, and pigmentation are described. Transformation from larvae to juveniles apparently was completed at 20 to 23 mm standard length. During transformation full complements of fin rays were developed, the dorsal fin moved forward, the gut shortened, and the anal fin moved forward. Yellowfin menhaden larvae have some characteristics that serve to distinguish them from larvae of other clupeid genera occupying the same geographic range, and also have some characters that may be helpful to distinguish them from other species in the genus Brevoortia. The yellowfin menhaden, Brevoortia smithi Hidelbrand, is one of four species of Brevoortia that occur along the Atlantic and Gulf of Mexico coasts of the United States. The biology and sys- tematica of yellowfin menhaden were discussed in detail by Hildebrand (1963) and most recently by Dahlberg (1970). Dahlberg reported that B. smithi occurs from North Carolina to Louisiana. Atlantic and Gulf of Mexico populations exist, which ap- parently are distinct, and the species is uncommon south of West Palm Beach on the Florida Atlantic coast and north of Tampa Bay on the Florida west coast. Although common in parts of its range, yellowfin menhaden are not abundant enough to contribute substantially to commercial menhaden catches (Dahlberg 1970). Reproducing populations apparently are confined to coastal areas of the United States, but Levi (1973) reported some juveniles from the Bahamas. Hybrids of B. smithi X B. tyrannus on the Atlantic coast and B. smithi X B. patronus on the Gulf coast commonly occur (Turner 1969; Dahlberg 1970). 'This paper is a contribution from the Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL3314?. , y , . ■^Division of Biologv and Living Resources, Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, PL 33149. 'Division of Fisheries and Applied Estuarine Ecology, Rosen- stiel School of Marine and Atmospheric Science, University of Miami, Miami, FL 33149; present address: Conservation Consul- tants, Inc., P.O. Box 35, Palmetto, FL 33561. Reintjes (1962) artificially fertilized eggs of yellowfin menhaden from Indian River, Fla. He presented a series of photographs and described developing embryos and yolk-sac larvae. Hybrid embryos and yolk-sac larvae of yellowfin menhaden and Gulf menhaden, B. patronus, were produced artificially (Hettler 1968), and pho- tographs of these embryos and larvae were published. More recently, Hettler (1970a) reared some yellowfin menhaden larvae from artificially fertilized eggs to 14.9 mm. He illustrated larvae of 7.6 and 11.9 mm total length (TL). Despite the literature on yellowfin menhaden development, no complete series from egg through transformation of larvae to the juvenile stage is available, nor have detailed illustrations been published that would be helpful to distinguish yellowfin menhaden from other similar clupeid larvae. We have reared yellowfin menhaden from naturally spawned planktonic eggs to advanced juveniles and we describe development of these stages in this paper. Eggs and larvae of other Brevoortia species have been described, but only those of the Atlantic menhaden, Brevoortia tyrannus, are well known. Mansueti and Hardy (1967) have reviewed published information on Atlantic menhaden development. Suttkus (1956) described larvae of Gulf menhaden 18.9 mm and longer, but smaller specimens are undescribed. The eggs and larvae of finescale menhaden, Brevoortia gunteri, have not Manuscript accepted September 1974. FISHERY BULLETIN: VOL. 73, NO. 3, 1975. 660 been described. DeCiechomski (1968) has discussed occurrence and photographed eggs and yolk-sac larvae of Brevoortia aurea from Argentina. Using our rearing techniques (Houde 1973b), it was relatively simple to rear yellowfin menhaden from eggs to advanced juveniles. Presumably the other species, including Atlantic and Gulf menhaden, can be reared by similar methods if their eggs can be obtained. It should now be pos- sible to conduct experiments under laboratory conditions, testing environmental factors on development of eggs and larvae of these impor- tant commercial species. METHODS Collecting Eggs Naturally fertilized eggs from Biscayne Bay, Miami, Fla. were collected in 1-m diameter, 505- /xm mesh plankton nets suspended from the dock of the Rosenstiel School of Marine and At- mospheric Science on 3 November and 26 November 1972. Surface temperatures were 25.4° and 22.7°C on the respective collecting days and salinity was 32-33°/ oa Yellowfin menhaden eggs were sorted by pipette from the other plankton organisms. The eggs that were sorted were known to be yellowfin menhaden because some of the same type had been hatched and the larvae reared during rearing trials in 1971. A total of 90 eggs on 3 November and 170 eggs on 26 November were placed in rectangular tanks of 38-liter capacity for rearing. Rearing and Preserving Methods Rearing techniques were similar to those described by Houde (1973b) and Houde and Palko (1970). For the first 20 days of culture, temperature was controlled at 26° ± 1.0°C. Salinities ranged from 33.5 to 37.0°/ oo and light was provided by fluorescent fixtures at an intensity of 2,500 Ix. Zooplankton, consisting mostly of copepod nauplii and copepodites, collected in a 35-/xm mesh plank- ton net, was fed to larvae for the first 12 days; subsequently Artemia salina nauplii were fed in addition to zooplankton. A total of 8 embryos and 66 surviving larvae were preserved in 5% buffered Formalin^ during the culture period to provide the 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. series used to describe development. Specimens from 3.7 mm to 36.2 mm SL (standard length) were included in the developmental series (Table 1), but many juveniles continued to survive and were reared to lengths of 50-60 mm before experiments were terminated. Meristics and Morphometries Methods for counting and measuring are iden- tical to those used by Houde et al. (1974) for Harengula jaguana Poey larvae. Fin rays were counted in each of the developing fins of unstained larvae (Table 2). Myomeres were counted (Table 3) and examined in relation to the dorsal fin and anus. The following myomere counts were made: Total myomeres: all myomeres; does not include the triangular area preceding the first myoseptum. Preanus myomeres: number anterior to the anus. Postanus myomeres: number posterior to the anus. Predorsal myomeres: number anterior to the dorsal fin origin. Postdorsal-preanus myomeres: number between the posterior insertion of the dorsal fin and the anus. The following measurements were made (Table 1): total length, standard length, preanus length, predorsal length, prepelvic length, head length, snout length, eye diameter, and body depth. All references to lengths of larvae in text are to stan- dard length unless otherwise noted. Osteology Sequence of ossification was determined from 10 specimens ranging from 5.2 to 25.3 mm SL. They were cleared with trypsin and stained with alizarin, using the method of Taylor (1967). DESCRIPTION Embryos Eight fertilized eggs from the plankton collec- tions were preserved. The embryos were approximately at the midstage of development at the time of collection. Eggs were spherical, the 661 Table l.-Specimens of laboratory-reared Brevoortia smithi used to describe development. Measurements in millimeters. Total Standard Preanus Predorsal Prepelvic Head Snout Eye Body length length length length length length length diameter depth 4.0 3.7 3.1 — — 0.46 0.08 0.22 — 4.2 4.0 3.4 — — 0.60 0.11 0.22 0.38 4.2 4.1 3.3 — — 0.56 0.09 0.20 0.39 4.3 4.1 3.4 — — 0.52 0.10 0.24 — 4.3 4.2 3.4 — — 0.52 0.09 0.24 — 4.4 4.2 3.4 — — 0.56 0.10 0.19 0.40 4.4 4.2 3.4 — — 0.54 0.06 0.20 0.37 4.4 4.2 3.5 — — 0.60 0.08 0.26 — 4.4 4.2 3.6 — — 0.54 0.08 0.26 — 4.5 4.3 3.5 — — 0.56 0.07 0.22 0.39 4.6 4.4 3.5 — — 0.55 0.08 0.22 0.34 4.6 4.4 3.6 — — 0.56 0.08 0.20 0.40 4.6 4.4 3.6 — — 0.57 0.08 0.23 0.34 4.6 4.4 3.6 — — 0.60 0.08 0.22 0.38 4.6 4.4 3.6 — — 0.61 0.10 0.21 0.42 4.7 4.5 3.7 — — 0.56 0.06 0.24 — 4.7 4.5 3.8 — — 0.62 0.09 0.24 0.34 4.7 4.5 3.7 — — 0.56 0.07 0.24 0.40 4.8 4.6 3.8 — — 0.56 0.08 0.24 0.36 4.8 4.6 3.8 — — 0.55 0.08 0.22 0.36 4.8 4.6 3.7 — — 0.56 0.10 0.24 0.36 4.9 4.7 3.7 — — 0.60 0.09 0.24 0.42 5.4 5.2 4.3 — — 0.69 0.11 0.25 0.43 5.5 5.2 4.2 — — 0.74 0.16 0.24 0.43 5.9 5.7 4.8 — — 0.72 0.10 0.26 0.43 6.3 6.1 5.0 — — 0.86 0.16 0.26 0.48 6.7 6.4 5.5 — — 0.84 0.18 0.27 0.47 7.3 7.0 5.9 — — 0.93 0.18 0.28 0.50 7.5 7.3 6.1 5.0 — 0.88 0.16 0.34 0.58 7.5 7.2 6.0 4.9 — 0.82 0.12 0.26 0.50 7.7 7.4 6.1 5.0 — 1.04 0.26 0.32 0.56 7.7 7.5 6.2 5.1 — 1.08 0.18 0.36 0.65 7.8 7.4 6.3 4.8 — 1.14 0.20 0.26 0.64 8.8 8.6 7.2 5.7 — 1.28 0.22 0.36 0.72 8.8 8.6 7.1 5.5 — 1.42 0,30 0.42 0.70 8.9 8.6 7.1 5.8 — 1.04 0.23 0.28 0.60 8.9 8.6 7.2 5.0 — 1.16 0.28 0.32 0.56 9.3 8.8 7.6 5.8 — 1.42 0.32 0.40 0.74 9.6 9.2 7.8 6.1 — 1.20 0.34 0.38 0.72 10.0 *.5 8.0 6.2 — 1.36 0.22 0.42 0.73 10.2 9.6 8.1 6.3 — 1.56 0.36 0.42 0.72 10.6 10.2 8.6 6.9 — 1.48 0.32 0.42 0.84 10.7 9.9 8.4 6.4 — 1.68 0.30 0.46 0.80 10.8 10.0 8.4 6.5 — 1.64 0.30 0.44 0.80 11.0 10.1 8.5 6.4 — 1.72 0.34 0.48 0.86 11.1 10.3 8.7 6.8 — 1.80 0.38 0.48 0.92 11.2 10.5 8.9 6.7 — 1.64 0.34 0.48 0.80 11.3 10.3 8.7 6.6 — 1.80 0.34 0.48 0.88 11.8 10.9 9.3 6.8 — 2.04 0.40 0.59 0.94 12.3 11.2 9.S 7.0 — 2.24 0.52 0.64 1.04 12.8 11.3 9.3 6.8 5.5 2.14 0.47 0.60 1.00 13.6 12.1 10.3 7.9 — 2.32 0.47 0.60 1.12 13.7 12.3 10.6 7.7 5.8 2.32 0.48 0.76 1.18 14.2 12.6 10.3 7.7 6.1 2.76 0.56 0.84 1.42 15.2 13.5 10.9 8.4 6.6 3.00 0.66 0.80 1.48 15.6 13.8 11.4 8.4 6.2 2.68 0.52 0.82 1.54 18.0 15.5 12.5 9.2 7.9 4.00 0.83 1.32 2.60 18.0 15.9 13.0 9.2 6.9 3.12 0.68 0.90 1.80 19.1 16.2 12.6 9.2 7.9 4.12 0.86 1.24 2.66 19.5 16.9 13.0 9.3 8.8 4.83 1.00 1.50 3.42 21.1 17.8 13.3 9.4 9.1 5.25 1.20 1.56 3.92 21.2 18.0 14.2 9.5 9.1 5.17 1.25 1.62 4.25 27.1 22.7 16.8 11.7 11.8 7.46 1.67 2.50 6.58 30.1 25.3 18.3 12.5 13.2 7.92 1.90 2.50 7.50 41.2 33.3 23.9 15.7 17.4 11.10 2.31 3.42 10.60 43.9 36.2 26.7 18.0 20.0 12.05 2.74 3.42 11.96 chorion was thin and unsculptured, and they had a single yellowish oil globule. Egg diameters ranged from 1.21 to 1.34 mm (mean = 1.27 mm), and oil globule diameters ranged from 0.12 to 0.17 mm (mean = 0.15 mm). Yolk diameters ranged from 0.80 to 1.19 mm (mean = 1.07 mm) and were large relative to egg diameter. Our egg diameters and oil globule diameters are much like those reported 662 Table 2.-Some meristic characters of laboratory-reared specimens of Brevoortia smithi. Standard Procurrent length Days after Principal caudal rays Dorsal Anal Ppr^tnrfil Pelvic rays (mm) hatching caudal rays Dorsal Ventral rays rays rays 3.7-7.2 0-5 No elements present on 29 specimens in this size range. 7.3 5 — — 6 „__ 7.4 7 — — 7 __ 7.4 5 — 4 ^^ ,^_ 7.5 9 2 — 7 __ 8.6 6 2 — 6 ^_ .^_ 8.6 5 2 — 6 __ ^_ 8.6 7 4 — 8 ^__ ^_ 8.6 15 3 — — 12 .^— 8.8 9 18 — 14 7 __ 9.2 7 3 — — 9 ^_ _^ 9.5 8 17 — 11 __ 9.6 9 19 — 13 ? _^ 9.9 11 19 — — 15 10 10.0 10 19 — 14 10.1 16 19 — — 17 7 ..^ 10.2 7 17 — 13 10.3 13 19 — — 16 12 , 10.3 11 19 — — 14 8 10.5 11 19 — — 15 9 __ 10.9 18 19 — — 16 16 ^_ __ 11.2 18 19 3 2 18 14 «^ 11.3 15 19 2 1 18 18 __ 12.1 15 19 3 2 17 14 .^_ 12.3 20 19 3 2 19 18 12.6 29 19 4 3 18 20 6 13.5 24 19 5 3 18 19 6 13.8 24 19 4 3 19 19 6 15.5 43 19 7 6 20 21 13 7 15.9 •18 19 6 5 17 18 5 4 16.2 28 19 8 6 19 20 13 7 16.9 43 19 7 6 19 20 14 7 17.8 31 19 8 7 20 21 15 7 18.0 38 19 9 8 20 20 14 7 22.7 60 19 8 6 21 21 14 7 25.3 50 19 8 7 20 21 14 7 33.3 176 19 8 7 20 21 16 7 36.2 102 19 8 7 20 20 15 7 Table 3. -Distribution of myomeres relative to other body parts for Brevoortia smithi larvae. Length Preanus myomeres Postanus myomeres Predorsal myome res Postdorsal-preanus m Number of lyomeres class Number of Number of Number of (mm, SL) specimens Range Mean specimens Range Mean specimens Range Mean specimens Range Mean 3.7- 6.0 24 36-39 37.58 24 7- 9 7.96 6.1- 8.0 8 36-38 37.25 8 7- 9 8.12 4 27-29 28.00 4 5-6 5.25 8.1-10.0 10 35-37 36.20 10 8-11 9.70 10 24-28 26.20 10 4-5 4.60 10.1-12.0 8 33-37 35.38 8 10-13 10.75 8 22-25 24.00 8 3-5 4.00 12.1-14,0 5 33-35 33.40 5 11-13 12.20 5 21-22 21.40 5 3-4 3.40 14.1-16.0 2 30-33 31.50 2 13-15 14.00 2 19-21 20.00 2 2-4 3.00 16.1-18.0 4 29-31 29.75 4 15-16 15.50 4 18-22 19.50 4 2 2.00 >22.7 4 28-29 28.50 4 16-18 17.00 4 15-17 15.75 4 2-3 2.25 by Reintjes (1962) for yellowfin menhaden, but yolk diameters differ greatly. In his specimens the perivitelline space averaged more than 30% of the egg diameter, but in our specimens it averaged only 16%. The difference might be partly account- ed for by the incubation salinities. Salinities at Indian River where Reintjes did his study were relatively low, ranging from 20.5 to 27.2°/ oo, while those in Biscay ne Bay during our study exceeded 32°/ oa Hettler (1968) also reported narrow perivi- telline spaces in hybrid eggs from B. smithi x B. patronus that he artificially fertilized at the collection site where salinities were 33-34 Voo. Previtelline spaces in his embryos ranged from 7 to 17% of the egg diameters, and he attributed the narrow spaces to possible effects of high salinity. Yellowfin menhaden eggs have been collected by us from Biscayne Bay on many occasions in 1971 to 1973, and they always are characterized by a rela- tively narrow perivitelline space, which is unusual 663 for clupeid species that spawn in South Florida marine waters. Developing embryos resembled those described and photographically illustrated by Reintjes (1962). On our preserved specimens, tiny melanophores were present only on the dorsal surface of the developing embryo. The yolk mass was segmented, but this was observed with difficulty in preserved material. No pigment was observed on the yolk sac or oil globule of our preserved eggs. Spawning by yellowfin menhaden in Biscayne Bay occurred at least between November and February. Spawning by this species has not been reported previously in Biscayne Bay, but eggs were common during 1971-73. We do not know the time of day at which spawning took place or the total incubation time at 22° to 26°C. Hettler (1970b) collected planktonic eggs of yellowfin menhaden in the Indian River and observed that they spawned at dusk. Reintjes (1962) reported hatching in 46 h at 19°C. Since we have collected embryos only in a single stage of development when temperatures were above 22°C, incubation time might be 24 h or less at these temperatures. Eggs were collected between 0900 and 1400 h; hatching in our aquaria was complete before 2400 h. Description of Larvae Body Shape and Growth Larvae were about 3.0 mm SL at hatching. They averaged 4.2 mm SL at 20 h after hatching (Figure 2A) when the body axis had straightened. During the first 2 days after hatching they were similar to yellowfin menhaden larvae described by Reintjes (1962). Larvae resemble those of other clupeid fishes. They are elongate, rod-shaped larvae after the yolk has been absorbed. At transformation they become deeper bodied and more laterally compressed. Proportional measurements of larvae in relation to standard length are summarized in Table 4. Most larvae did not grow in length from the 1st until the 4th day after hatching at 26°C. Growth was rapid from the 4th until the 10th day (Figure 1), averaging about 0.80 mm/day. Growth was slower and more variable on subsequent days, but averaged about 0.45 mm /day from the 4th until the 20th day after hatching. Our 36.2-mm specimen was preserved 102 days after hatching, 18 15- ol2 z o z .< in 3- »»*, 10 15 20 25 30 DAYS AFTER HATCHING 35 40 45 Figure l.-Growth of laboratory-reared larvae of Brevoartia smithi. but growth rate of juveniles was almost certainly lower than might be expected under better feed- ing conditions because we did not maintain a careful feeding schedule after larvae transformed to the juvenile stage. Several juveniles from 50 to 60 mm were preserved when the experiments were terminated at 235 days after hatching. At 20°C, Hettler (1970a) reported growth rates of about 0.27 mm /day for yellowfin menhaden larvae that he reared for 27 days. The four smallest larvae that we preserved on the 4th and 5th days after hatching were smaller than larvae preserved on previous days (Figure 1). It is probable that those larvae had not begun to feed and that they were starving at the time of preservation. Yolk Absorption and Gut Differentiation The yolk sac is broadly ellipsoid in newly hatched larvae with the oil globule located ven- trally and just posterior to the middle of the yolk mass (Figure 2A; see also Reintjes 1962). The yolk was nearly absorbed in specimens preserved 1 day after hatching (Figure 2B). All visible yolk remains, including the oil globule, had disappeared by 60 h after hatching at 26°C. Many larvae were observed to begin feeding before all of the yolk had been absorbed. At that time the gut was a straight tube, but within 24 h (at about 5.0 to 5.5 mm SL) it had developed into distinct fore and hind sections, the latter characterized by the bands of muscle common to all clupeid larvae. Preanus Length Preanus length averaged 82 to 84% SL from hatching until larvae were 15.0 mm SL (Table 4). 664 Table 4.-Summary of the measurements of various body parts of Brevoortia smithi larvae and juveniles in relation to standard length. Tabulated measurements are means for measurements from 1-mm length classes. Length class Number of Preanus Predorsal Prepelvic Body Head Snout Eye diam- (mm, SL) specimens lengthiSL length:SL length:SL depth:SL lengthiSL lengthiSL eteriSL 3.1- 4.0 2 0.84 — 0.09 0.14 0.02 0.06 4.1- 5.0 20 0.82 — — 0.09 0.13 0.02 0.05 5.1- 6.0 3 0.82 — — 0.08 0.13 0.02 0.05 6.1- 7.0 3 0.84 — — 0.08 0.14 0.03 0.04 7.1- 8.0 5 0.84 0.67 — 0.08 0.13 0.03 0.04 8.1- 9.0 5 0.84 0.66 — 0.08 0.15 0.03 0.04 9.1-10.0 5 0.84 0.65 — 0.08 0.15 0.03 0.04 10.1-11.0 6 0.84 0.64 ,— 0.08 0.17 0.03 0.05 11.1-12.0 2 0.84 0.62 0.49 0.09 0.20 0.04 0.06 12.1-13.0 3 0.84 0.63 0.47 0.10 0.20 0.04 0.06 13.1-14.0 2 0.82 0.62 0.47 0.11 0.21 0.04 0.06 15.1-16.0 2 0.81 0.59 0.47 0.14 0.23 0.05 0.07 16.1-17.0 2 0.77 0.56 0.50 0.18 0.27 0.06 0.08 17.1-18.0 2 0.77 0.53 0.51 0.23 0.29 0.07 0.09 22.7 1 0.74 0.52 0.52 0.29 0.33 0.07 0.11 25.3 1 0.72 0.49 0.52 0.30 0.31 0.08 0.10 33.3 1 0.72 0.47 0.52 0.32 0.33 0.07 0.10 36.2 1 0.73 0.49 0.55 0.33 0.33 0.08 0.09 "^ ^ "^^^^^^^ZZ y* V y y T~ f-^-f ■■ , j-T-T^ " ' .' / /' ,'', -— sVdU {({<< •(TTTT mttG ~SXSS}M ^^E *,^*==^ o 1mm 222S I22I 1mm Figure 2.-4.2-mm SL (20 h posthatching) and 4.6-mm SL (41 h posthatching) larvae of Brevowtia smithi. As the gut shortened during transformation, preanus length was reduced to 77% SL at 16.0 to 18.0 mm and averaged 72 to 74% for juveniles 22.7 mm and longer. Head Length Head length averaged 13 to 14% SL for larvae from 3.0 to 8.0 mm (Table 4). It then increased gradually to 29% at 18.0 mm and stabilized at 31 to 33% SL for juvenile specimens. Hildebrand (1963) and Dahlberg (1970) recorded head lengths rang- ing from 29 to 32.5% SL for large juvenile and adult yellowfin menhaden. Eye Diameter Eye diameter averaged 6% SL on newly hatched larvae but gradually decreased to 4% SL at about 6.0 mm and was stable until larvae grew to 10.0 mm (Table 4). Eye diameter then increased to 9% SL at 18.0 mm and averaged 9 to 11% SL for juveniles 22.7 to 36.2 mm. A relative decrease in eye diameter must occur in older juveniles because 665 Hildebrand (1963) reported eye diameters ranging from 6.1 to 7.5% SL for specimens 91 mm and longer. Snout Length Snout length increased gradually throughout development. It averaged 2% SL at hatching and increased to 7 to 8% SL in our juveniles (Table 4). Snout lengths of large juvenile and adult yellowfin menhaden range from 6.8 to 8.0% SL (Hildebrand 1963). Body Depth Body depth, measured at the pectoral symphysis, averaged 8 to 9% SL from hatching until larvae were 12.0 mm (Table 4). A rapid increase in body depth then occurred; it was 23% SL at 18.0 mm and apparently was still increasing in juvenile specimens 22.7 mm and longer. Our 36.2-mm specimen had a body depth of 33% SL. Hildebrand (1963) did not measure body depth by the same method that we used, but he noted that yellowfin menhaden juveniles are deep bodied, more so than adults of this species. Predorsal Length Predorsal lengths were measured on larvae that had dorsal fin rays developing. No dorsal fin development was observed on any specimens less than 7.3 mm. Predorsal length decreased gradually from 67 to 62% SL for larvae from 7.3 to 14.0 mm and then decreased more rapidly for larger specimens (Table 4). It was 52% SL for our 22.7- mm specimen and ranged from 47 to 49% SL for larger individuals. The decrease in predorsal length from 67 to 62% for 7.3- to 14.0-mm larvae may be partly due to measuring predorsal length on specimens with incompletely developed dorsal fins. The marked decrease in predorsal length on larger larvae resulted from forward movement of the dorsal fin as larvae began to transform to the juvenile stage. Prepelvic Length Prepelvic lengths were measured on larvae that had pelvic fin buds or fins. From 11.1 to 16.0 mm, prepelvic lengths ranged from 47 to 49% SL (Table 4). In larger specimens, prepelvic lengths increased to 51% SL at 18.0 mm and to about 52% SL for 22.7- to 33.3-mm individuals. Our 36.2-mm specimen had a prepelvic length of 55% SL. Pelvic fins moved posteriorly during growth of yellowfin menhaden larvae, causing the observed increase in prepelvic length. Meristics Myomeres The number of myomeres in fishes corresponds approximately to the number of vertebrae. Dahl- berg (1970) reported 43 to 46 vertebrae (mean = 45) for juvenile and adult B. smithi. Myomeres can be counted on larvae before development of ver- tebrae and are a valuable meristic character. Total numbers of myomeres in our specimens ranged from 45 to 47 on the 63 individuals for which ac- curate counts were obtained. There was no correlation between the number of myomeres and standard length, indicating that the full complement was present at hatching. The frequencies of occurrence were as follows: Number of myomeres Frequency ^ ^6 1+7 26 32 5 The mean number of myomeres was 45.67 (Sj- = 0.1109). The distribution of myomeres in relation to other body parts can be useful for identifying clupeid larvae (Ahlstrom 1968). We examined the distribution of myomeres for yellowfin menhaden larvae relative to the dorsal fin and anus (Table 3). Preanus myomeres decreased from a mean number of 37.6 in newly hatched larvae to 28.5 in juveniles that were 22.7 mm and longer. Postanus myomeres increased accordingly, from 8.0 in newly hatched larvae to 17.0 for juveniles. Short- ening of the gut during development caused the change in distribution of preanus and postanus myomeres. Predorsal myomeres decreased in numbers as larvae developed, because the dorsal fin moved forward. Larvae of 6.1 to 8.0 mm had a mean number of 28.0 predorsal myomeres, but juveniles had only 15.8. Numbers of predorsal myomeres were variable for specimens within any given size class (Table 3). The number of postdor- sal-preanus myomeres decreased as larvae grew. Larvae 6.1 to 8.0 mm had a mean number of 5.3 postdorsal-preanus myomeres, but advanced lar- vae and juveniles always had 4 or fewer (usually 2 666 or 3). Larvae of Opisthonema oglinum and Harengula jaguana always had 5 or more post- dorsal-preanus myomeres during all developmen- tal stages (Richards et al. 1974; Houde et al. 1974). Fin Development A finfold surrounded the trunk and caudal area of newly hatched yellowfin menhaden larvae. Some parts of it remained along the ventral body margin until larvae were approximately 16.0 mm. Pectoral fin buds were already present when lar- vae hatched (Figure 2A), but the pelvic and median fins were not formed. Rays first appeared in fins in the following sequence: dorsal, caudal, anal, pelvics, and pectorals. Because rays first develop as cartilaginous structures, the size at which full complements were present was not necessarily the size at which all rays were ossified (Tables 2, 5). Although the beginning and comple- tion of fin ray development were best correlated with length of larvae, age was also a factor, especially for anal fin development (Table 2). Median fins had full complements of rays when larvae were 17.0 mm (Table 5). Dorsal rays first appeared at 7.3 mm, although an opaque area was present in the dorsal finfold, near the future dorsal fin, in some larvae as small as 6.4 mm. Full complements of 20 or 21 dorsal rays usually were attained when larvae were 15.5 to 17.0 mm. Prin- cipal caudal rays first appeared at 7.5 to 8.6 mm and the full complement of 19 principal rays was present when larvae were 9.6 to 10.3 mm. The no- tochord began to flex while principal caudal rays and other caudal fin structures were developing. Procurrent caudal rays began to develop at 11.2 mm, and full complements of 8 or 9 dorsal and 6 or 7 ventral rays were present at 16.2 to 17.8 mm. Anal rays first developed at 8.8 to 10.3 mm. A full complement of 20 or 21 rays was present on all larvae 16.2 mm or longer, although one specimen only 12.6 mm had 20 anal rays. Hildebrand (1963) and Miller and Jorgenson (1973) reported from 21 to 24 anal rays in yellowfin menhaden specimens longer than 72 mm that they examined. We examined six juveniles from our rearing experiment that were 40-50 mm in length. Two of these specimens had 22 anal rays, two had 21 rays, and two had 20 rays. Rays in paired fins began to develop later than in median fins. Pectoral fins without rays were present soon after hatching, but no rays developed until larvae had attained approximately 15.5 mm. A full complement of 14 to 16 pectoral rays was present on larvae 16.9 mm and longer. Pelvic fins appeared as tiny buds when larvae were 10.9 to 11.3 mm, but rays did not develop until larvae were about 13.0 mm. A full complement of 7 pelvic rays was attained at 15.5 to 16.2 mm. Scales and Scutes Scales and ventral scutes were observed in specimens 17.8 mm and longer. Scales first developed anterior to the dorsal fin and in the region of the caudal peduncle. Specimens 22.7 mm or longer were fully scaled. Ventral scutes first developed anterior to the pelvic fins when larvae were 16.9 mm. Full complements of 30 to 32 (18 to 20 anterior to the pelvic fins and 11 to 13 posterior to the pelvic fins) were present on specimens 22.7 mm and longer. These counts are the same as those given for adult B. smithi by Dahlberg (1970). Osteology Ten specimens of yellowfin menhaden were cleared and stained to determine sequence of development of skeletal structures. Ossification was similar to that described for larvae of Atlantic thread herring (Richards et al. 1974) and of scaled sardine (Houde et al. 1974). Consequently, Table 5.-Suminary of fin development sequence in larvae of Brevoortia smithi. Buds first appear Standard length (mm) Fin Rays first appear Full complement of rays Number of rays in fully developed fin Dorsal Caudal: Principal Procurrent Anal Pelvic Pectoral 10.9 to 11.3 <4.0 7.3 7.5 to 8.6 11.2 8.8 to 10.3 -v12.6 '^'IS.S mm 15.5 1Ox>^17.0 9.6 to 10.3 16.2 to 17.8 12.6 to 16.2 15.5 to 16.2 16.9 19 to 21 19 8 or 9 dorsal 6 or 7 ventral 20 or 21 7 14 to 16 'Rays were present at the tabulated lengths, but not necessarily ossified at those sizes. 667 osteology of yellowfin menhaden larvae is treated rather briefly in this paper. Ossification of most structures occurred at a smaller size in yellowfin menhaden than in either Atlantic thread herring or scaled sardines, but the sequence of develop- ment was similar in all of the species. No bones were ossified in our 5.2-mm specimen, but the cleithra were lightly stained in 6.1- and 7.2-mm specimens. Cleithra were well stained in a 7.4-mm specimen, but no other bones were ossified. Slight ossification of the maxillaries and den- taries, in addition to the cleithra, was observed in our 8.6-mm larva. At 10.5 mm, the caudal fin complex began to ossify, cranial bones were lightly stained, and 8 maxillary and 3 dentary teeth were present. Vertebral centra were beginning to stain at 12.3 mm; neural and hemal arches were developing, but were unstained. Dorsal fin rays were ossifying at 12.3 mm. Also, cranial bones were ossifying, the hyoid apparatus was stained, 11 teeth were present on the maxillaries, and 4 were present on the dentaries. At 16.2 mm, most of the major skeletal structures were at least partly ossified. Rays in median and paired fins were stained as were neural and hemal spines along the vertebral column. Premaxillaries and posterior supramaxillaries were ossified in this specimen. At 18.0 mm, the degree of stain uptake increased in most bones. Also, ribs were stained, anterior supramaxillaries were stained, and 16 maxillary teeth were present, but the dentary bore no teeth. Ossification was complete in our 25.3-mm specimen. A total of 25 maxillary teeth but no dentary teeth were present. Dentary teeth are a transient larval character in B. smithi. One large, erect tooth was present on the basihyal of our 18.0-mm specimen, and two were present on the 25.3-mm specimen. Basihyal teeth also were reported from Atlantic thread herring and scaled sardine larvae (Richards et al. 1974; Houde et al. 1974). The caudal fin complex of yellowfin menhaden developed much like that of scaled sardine, and we give a brief description here, using the ter- minology of Houde et al. (1974) in their description of scaled sardine. Some cartilaginous, principal caudal rays developed in specimens as small as 7.5 mm. Flexure of the notochord and appearance of cartilaginous hypural plate elements occurred at about 8.5 to 9.0 mm. Our 10.5-mm specimen had stain uptake in the proximal parts of the 19 prin- cipal caudal rays, and the first uroneural was slightly stained. Hypural elements were present but unstained. At 12.3 mm, the 19 principal caudal rays were fully stained, the second ural vertebra was stained as were the first and second uroneurals, and the parhypural was lightly stained. The hypurals were present but unstained as were two epural bones. Ossification was progressing in our 16.2-mm specimen. Both the first and second ural vertebrae were stained, the six hypurals were stained, the parhypural was well stained, and all three uroneurals were now stained. In addition to the 19 principal caudal rays, 8 dorsal procurrent caudal rays and 6 ventral procurrent caudal rays were stained. Two epurals were present on this specimen but were unstained. At 18.0 mm, all of the bones in the caudal fin area were at least partly ossified. The two epurals were now slightly stained, and 9 dorsal plus 8 ventral procurrent caudal rays were present and stained. The 25.3-mm specimen had all caudal fin bones ossified. The two epurals were well stained on this specimen, these bones being the last to ossify in the caudal fin complex. Pigmentation Melanophore distribution on yellowfin menhaden larvae is similar to other clupeid larvae, but there are some distinctive characteristics which may serve to distinguish them from other clupeid larvae with which they can occur. Melanophores were contracted on some specimens and expanded on others, accounting for some of the apparent variability among individuals. Our illustrated specimens (Figures 2-5) have pigment that is typical of most specimens of those lengths. Head Region Newly hatched yellowfin menhaden larvae have several tiny melanophores on the snout and a few over the brain. Within 1 day after hatching those melanophores have migrated or disappeared, because no pigment is present on the heads until larvae attain about 9.0 mm. The eyes became pig- mented at about 4.5 mm, at 1 day after hatching. Typical pigmentation on the pectoral symphysis and over the heart developed at 4.5 to 6.0 mm. One or two melanophores appeared on the pectoral symphysis after yolk absorption when larvae were 4.5 to 5.0 mm. Those melanophores developed into two distinct pairs by 7.0 mm. Either one or two melanophores developed over the heart at about 6.0 mm. A single melanophore occurred at the 668 1mm Figure 3.-7.0-mm SL (4 days posthatching) and 8.6-mm SL (6 days posthatching) larvae of Brevoartia smithi. Figure 4.-12.1-mm SL (15 days posthatching), 15.9-mm SL (18 days posthatching), and 17.8-mm SL (31 days posthatching) larvae of Brevoortia smithi. 669 5mm Figure 5.-22.7-mm SL (60 days posthatching) juvenile of Brevoortia smithi. pectoral fin base in some larvae as small as 4.5 mm and was present in all larvae at 7.0 mm. The pig- ment pattern associated with the pectoral symphysis, heart area, and pectoral fin base was retained until larvae were about 16.5 mm. At 9.0 mm, from one to three stellate melanophores had developed internally and could be seen through the otic capsules. A single, stellate melanophore frequently was present over the hindbrain at 10.0 mm, and some specimens had a small melanophore just posterior to and dorsal to the eye at that length. At about 12.0 mm, the pig- mentation on the head began to increase substan- tially. Melanophores appeared on both jaws, on the side of the head, and over both midbrain and hindbrain regions. The number of melanophores increased as larvae grew, and numerous stellate melanophores were present on the heads of the larvae at 16.5 mm. Melanophores were especially concentrated on the jaws and over the brain in larvae larger than 16.5 mm. Gut and Trunk Region Newly hatched yellowfin menhaden had a few tiny melanophores on the dorsal surface of the trunk along the forebody. Within 12 h of hatching, these melanophores apparently migrated ven- trally because they disappeared on the dorsal sur- face, but a series of malanophores was developing along the gut region of larvae. Paired series of melanophores along the gut margin, which are typical of clupeid larvae, were present on yellowfin menhaden larvae of 4.5 mm and longer. Distinct pairs, numbering from 8 to 16, developed along the dorsolateral margin of the foregut and less distinct pairs, numbering 9 to 14, occurred along the ventral margin of the hindgut. These series were clearly visible until larvae were about 17.0 mm; they were not present on specimens longer than 18.0 mm. From one to three stellate melanophores usually occurred near the anus along the dorsal surface of the gut on larvae longer than 4.5 mm. These were continuous with an internal series of melanophores that were visi- ble over the hindgut on most specimens longer than 8.5 mm. The number in this series increased from 6 to about 17 at 10.5 mm. Three or four melanophores were associated with the developing swim bladder in 10.0- to 12.0-mm larvae. A second internal series of melanophores was associated with developing vertebrae, but these were too in- distinct to count accurately. Pigment developed along the sides of the trunk of some yellowfin menhaden as small as 5.2 mm. From 1 to 3 stellate melanophores were present on some larvae between 5.5 and 7.0 mm, and this number usually increased to as many as 10 for 7.0- to 10.0-mm larvae. Some larvae up to 8.6 mm had no lateral pigment on the trunk, but most specimens, when examined closely, were observed to have these melanophores. The number of lateral melanophores increased greatly when larvae were between 10.0 and 12.0 mm; some specimens of those lengths had as many as 25 lateral melanophores. When larvae were 14.0 mm or longer, these melanophores became very numerous, and most were located above the lateral midline. A paired series of melanophores developed along the dorsal midline, both anterior and posterior to the dorsal fin, on specimens longer than 16.5 mm. There are melanophores associated with the developing fins. From 1 to 3 stellate melanophores 670 were present at the dorsal fin base of 9.5- to 11.0- mm larvae. They numbered from 3 to 8 at about 14.0 mm and then as many as 15 at 18.0 mm. Two or three stellate melanophores were present at the anal fin base on 11.0- to 12.0-mm larvae; these numbers increased rapidly to 11 or 12 at 18.0 mm. Numbers of melanophores at the dorsal and anal fin bases were variable among individuals of the same length. A single stellate melanophore developed at the pelvic fin base on specimens as small as 12.3 mm. Some tiny melanophores began to occur in the pectoral, dorsal, and anal fins at 16.2 mm. Caudal Region Newly hatched larvae have melanophores on both the dorsal and ventral sides of the notochord tip. Numbers on the dorsal tip range from one to two, while those on the ventral tip range from one to three. In addition, one or two melanophores are located along the ventral midline posterior to the anus. Pigment along the ventral tip of the no- tochord began to migrate into the caudal finfold at 7.0 to 7.2 mm. This pigment was associated with developing caudal rays in larger larvae. Pigment in the caudal fin increased rapidly when larvae exceeded 10.0 mm. Internal melanophores were first present in the hypural plate region on larvae 10.3 mm and longer. Larvae longer than 16.5 mm invariably had many melanophores among the rays of the caudal fin. Transformation Transformation of larvae to juveniles ap- parently was complete between 20 and 23 mm. Unfortunately, we preserved no specimens that were between 18.0 and 22.7 mm. However, our specimens of 17.8 and 18.0 mm were not complete- ly transformed while the 22.7-mm specimen had acquired nearly all of the juvenile characteristics. Larvae began transforming at about 14.0 mm. At that time, proportional measurements relating preanal length, predorsal length, body depth, head length, snout length, and eye diameter to standard length (Table 4) began to change rapidly. Except for body depth, which continued to increase, the proportional measurements became nearly con- stant at 22.7 mm. The distribution of myomeres relative to other body parts became stable for specimens 22.7 to 36.2 mm (Table 3). Fin rays were completely ossified at 18.0 mm (Table 5), but the epural bones of the caudal skeleton were still unossified at that size. Although some scales were present on our 17.8- and 18.0-mm specimens, the 22.7-mm specimen was the smallest that was fully scaled. The slender, rodlike shape of the larva was replaced by the deeper bodied, laterally compressed form of the juvenile between 17.8 and 22.7 mm. The silvery coloration of juveniles was present on our specimens from 22.7 to 36.2 mm. Transformation included the following features: forward movement of the dorsal fin; shortening of the gut; forward movement of the anal fin; and relative increases in head length, snout length, eye diameter, and body depth. The same features were noted for transforming larvae of Atlantic thread herring (Richards et al. 1974) and scaled sardine (Houde et al. 1974). COMPARISONS Eggs and larvae of the genus Brevoortia almost always can be distinguished from those of other clupeid genera that spawn in marine waters of the south Florida and Gulf of Mexico region. Members of the genera Alosa and Dorosoma have demersal eggs, unlike those of Brevoortia which are pelagic. Dorosoma spawns in fresh waters and Alosa in fresh or nearly fresh waters, so that occurrence of their larvae with those of Brevoortia is unlikely. Larvae of Alosa spp. have more total myomeres than any species of Brevoortia, and the genera can be easily distinguished. Dwarf herrings (Jenkin- sia spp.) might occur with B. smithi at the southern extreme of their range in Florida. Neither spawning habits nor eggs and larvae of Jenkinsia have been described, but the total myomeres of Jenkinsia did not exceed 42 (Miller and Jorgenson 1973), making it unlikely that lar- vae could be confused with Brevoortia which have higher myomere numbers {Brevoortia, 44-48). Since B. smithi may occur with either B. tyrannus or B. patronus and because hybrids are known (Dahlberg 1970), the specific identification of menhaden eggs and larvae from plankton collec- tions is still in doubt where the species' ranges overlap. Eggs of B. smithi are smaller than those of Harengula jaguana which range from 1.55 to 1.78 mm in diameter (Houde et al. 1974), and they can- not be confused with those of Etrumeus teres because eggs of that species have no oil globule. Our B. smithi eggs were similar to those of Opisthonema oglinum (Richards et al. 1974) and to 671 those of western Atlantic Sardinella sp. (Simpson and Gonzales 1967; Matsuura 1971; Houde and Fore 1973), but their average diameter was greater than for those species. Our B. smithi eggs ranged from 1.21 to 1.34 mm in diameter (mean = 1.27 mm) while those of 0. oglinum are 1.10 to 1.28 mm (mean = 1.19 mm) (Richards et al. 1974) and those of western Atlantic Sardinella sp. are 1.00 to 1.32 mm (mean = 1.12-1.18 mm) (Simpson and Gonzales 1967; Matsuura 1971; Houde and Fore 1973). Reintjes (1962) reported B. smithi eggs ranging from 1.15 to 1.48 mm, including both planktonic and artificially fertilized eggs. The eggs of 0. oglinum would not usually occur with those of B. smithi because Opisthonema spawns during spring and summer (Fuss et al 1969; Houde 1973a; Richards et al. 1974), while members of the genus Brevoortia are winter spawners (e.g.. Turner 1969; Dahlberg 1970) in waters of south Florida and the Gulf of Mexico. Sardinella eggs could conceivably occur with those of B. smithi, but Sardinella probably spawns farther offshore than does B. smithi. Eggs of B. tyrannus ap- parently are larger than those of B. smithi, the reported diameters ranging from 1.30 to 1.95 mm (Mansueti and Hardy 1967). Brevoortia patronus eggs usually are slightly smaller than B. smithi eggs, the diameters ranging from 1.04 to 1.30 mm (Houde and Fore 1973). Hybrid embryos from ar- tificial fertilization of B. smithi x B. patronus ranged from 1.05 to 1.18 mm (Hettler 1968). Larvae of Brevoortia spp. have some distinctive characters that serve to distinguish them from other clupeid larvae with which they may occur. Newly hatched larvae of B. smithi have been pho- tographed by Reintjes (1962), but these pho- tographs fail to show distinguishing characters of larvae in that size range. Hettler (1970a) present- ed illustrations of 7.6- and 11.9-mm TL larvae of B. smithi, but only the 11.9-mm specimen has some characteristics illustrated that help to identify it as a Brevoortia sp. Myomere numbers ranged from 45 to 47 in B. smithi, thus separating it from H. jaguana (42 or fewer) (Houde et al. 1974) and E. teres (48 or more) (Houde and Fore 1973). Total myomere counts of B. smithi overlap those of 0. oglinum (45 to 49), Sardinella sp. (45 to 47) (Houde and Fore 1973), and the other species of Brevoortia. Numbers of post- dorsal-preanus myomeres always were fewer than 5 in B. smithi larvae longer than 10 mm and never exceeded 6 in smaller larvae (usually 4 or 5). The other identified clupeid genera from this region. excepting Etrumeus, have 5 or more (usually 6 to 9) postdorsal-preanus myomeres in all length classes, thus serving to distinguish them from Brevoortia larvae. Pigmentation of newly hatched Brevoortia lar- vae apparently differs from that of other clupeid genera in its details. Recently hatched B. smithi larvae have pigment on both the dorsal and ventral sides of the notochord tip (Figures 2A-3B), distin- guishing them from other co-occurring clupeid genera, except for some specimens of Harengula (see Houde et al. 1974). Brevoortia tyrannus has pigmentation similar to B. smithi at the notochord tip (Mansueti and Hardy 1967), and we suspect that B. patronus also has this pigment characteristic based on our observations of Brevoortia larvae that were collected in the northern Gulf of Mexico, where B. patronus is known to spawn. Recently hatched larvae of 0. oglinum, Sardinella sp., and E. teres have pigment only on the ventral side of the notochord tip. Lateral pigmentation is present on B. smithi larvae as small as 5.2 mm-which is smaller than other clupeids found in their geographical range. At 10 to 12 mm, most of our B. smithi larvae had more than 5 melanophores on their sides, and some had as many as 25. No larvae of Harengula, Opisthonema, Sardinella, or Etrumeus that we have observed have had pigment on the sides until they were at least 15 mm in length. We do not know if B. tyrannus or B. patronus develop lateral pigmentation at sizes as small as B. smithi, but illustrations of B. tyrannus larvae (Mansueti and Hardy 1967) from 8 to 23 mm do not show any such pigment, nor is it mentioned in their text. Size at transformation varies among clupeid species. Brevoortia smithi had completed transfor- mation at about the same size as H. jaguana (Houde et al. 1974) and 0. oglinum (Richards et al. 1974), at lengths from 20 to 24 mm. However, other species of Brevoortia apparently do not transform until they are of larger size. Brevoortia tyrannus exceeds 30 mm before having a typical juvenile form (Mansueti and Hardy 1967), and the obser- vations and morphological data of Suttkus (1956) suggest that B. patronus does not transform until 28 mm or longer. It is possible that our tank-reared B. smithi transformed at smaller sizes than in na- ture, but the seemingly good growth rate and the selection of normal appearing larvae to describe development lead us to believe that B. smithi is transformed at approximately 22 mm. 672 ACKNOWLEDGMENTS We thank Walter Stepien and A. Keith Taniguchi for their assistance in rearing the yellowfin menhaden. Michael D. Dahlberg confirmed identification of B. smi^/it juveniles that we reared. Thomas Potthoff cleared and stained specimens that we used to describe osteology. Susan Stevens Suarez illustrated the yellowfin menhaden larvae. William J. Richards assisted in interpreting developmental osteology and cri- ticized an early draft of the manuscript. William F. Hettler, Jr. and John W. Reintjes also reviewed the manuscript and provided helpful suggestions. Partial financial support for this study was derived from NSF Biological Oceanography Grant GA-34138 and from NOAA Sea Grant 04-3-158-27 Sub. 3 to the University of Miami. LITERATURE CITED Ahlstrom, E. H. 1968. Book review of: Mansueti, A. J., and J. D. Hardy, Jr. 1967. Development of fishes of the Chesapeake Bay region, an atlas of egg, larval, and juvenile stages. Part I. Copeia 1968:648-651. Dahlberg, M. D. 1970. Atlantic and Gulf of Mexico menhadens, genus Brevoortia (Pisces: Clupeidae). Bull. Fla. State Mus. Biol. Sci. 15:91-162. DeCiechomski, J. D. 1968. Huevas y larvas de tres especies de peces marinos, Anchoa marinii, Brevoortia aurea y Prionotus nudigula de la zona de Mar del Plata. Bol. Inst. Biol. Mar. Mar del Plata 17:1-28. Fuss, C. M., Jr., J. A. Kelly, Jr., and K. W. Prest, Jr. 1969. Gulf thread herring; aspects of the developing fishery and biological research. Proc. Gulf Caribb. Fish. Inst., 21st Annu. Sess., p. 111-125. Hettler, W. F., Jr. 1968. Artificial fertilization among yellowfin and Gulf menhaden {Brevoortia) and their hybrid. Trans. Am. Fish. Soc. 97:119-123. 1970a. Rearing larvae of yellowfin menhaden, Brevoortia smithi. Copeia 1970:775-776. 1970b. Rearing menhaden. In Research in fiscal year 1969 at the Bureau of Commercial Fisheries Biological Laboratory, Beaufort, N.C., p. 24-26. U.S. Fish Wildl. Serv., Circ. 350. HiLDEBRAND, S. F. 1963. Family Clupeidae. In Fishes of the western North Atlantic. Part Three, p. 257-454. Mem. Sears Found. Mar. Res., Yale Univ. 1. HouDE, E. D. 1973a. Estimating abundance of sardine-like fishes from egg and larval surveys. Eastern Gulf of Mexico: preliminary report. Proc. Gulf Caribb. Fish. Inst., 25th Annu. Sess., p. 68-78. 1973b. Some recent advances and unsolved problems in the culture of marine fish larvae. Proc. World Mariculture Soc. 3:83-112. HouDE, E. D., AND p. L. Fore. 1973. Guide to identity of eggs and larvae of some Gulf of Mexico clupeid fishes. Fla. Dep. Nat. Resour., Mar. Res. Lab., Leafl. Ser. 4, Part 1, No. 23, 14 p. Houde, E. D., and B. J. Palko. 1970. Laboratory rearing of the clupeid fish Harengula pensacolae from fertilized eggs. Mar. Biol. (Berl.) 5:354-358. Houde, E. D., W. J. Richards, and V. P. Saksena. 1974. Description of eggs and larvae of scaled sardine, Harengula jaguana. Fish. Bull., U.S. 72:1106-1122. Levi, E. J. 1973. Juvenile yellowfin menhaden from the Bahama Islands. Trans. Am. Fish. Soc. 102:848-849. Mansueti, A. J., and J. D. Hardy, Jr. 1967. Development of fishes of the Chesapeake Bay region. An atlas of egg, larval and juvenile stages. Part I. Nat. Resour. Inst., Univ. Md., 202 p. Matsuura, Y. 1971. A study of the life history of Brazilian sardines, Sar- dinella aurita. I. Distribution and abundance of sardine eggs in the region of Ilha Grande, Rio de Janeiro. Bol. Inst. Oceanogr. (Sao Paulo) 20:33-60. Miller, G. L., and S. C. Jorgenson. 1973. Meristic characters of some marine fishes of the wes- tern Atlantic Ocean. Fish Bull., U.S. 71:301-312. Reintjes, J. W. 1962. Development of eggs and yolk-sac larvae of yellowfin menhaden. U.S. Fish Wildl. Serv., Fish. Bull. 62:93-102. Richards, W. J., R. V. Miller, and E. D. Houde. 1974. Egg and larval development of the Atlantic thread herring, Opisthonema oglinum. Fish. Bull., U.S. 72:1123-1136. Simpson, J. G., and G. Gonzalez. 1967. Some aspects of the early life history and environment of the sardine, Sardinella anchovia, in eastern Venezuela. Ser. Recursos Explor. Pesq. 1:38-93. SUTTKUS, R. D. 1956. Early life history of the largescale menhaden, Brevoortia patronus, in Louisiana. Trans. North Am. Wildl. Conf . 21:390-407. Taylor, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. U.S. Natl. Mus. 122(3596):1-17. Turner, W. R. 1969. Life history of menhadens in the eastern Gulf of Mexico. Trans. Am. Fish. Soc. 98:216-224. 673 NOTES TWO BLOOMS OF GYMNODINIUM SPLENDENS, AN UNARMORED DINOFLAGELLATE Little is known about the ecology and physiology of an unarmored dinoflagellate (30 to 50 /im), Gym- nodinium splendens Lebour, although feeding experiments have shown it to be an important food source for certain marine herbivores. Lasker et al. (1970) found that anchovy larvae may be reared the first week upon unialgal suspensions of G. splendens while Paffenhofer (1970, 1971) and Barnett (1974) showed it to be a preferred food for Calanus finmarchicus and larval stages of Labidocera trispinosa. Pokorny and Gold (1973) reported on cell ultrastructure of G. splendens, Sweeney (1954) observed vitamin B12 requirements, and Thomas et al. (1973) described optimal light and temperature requirements. In addition to these laboratory studies, Loftus et al. (1972) have noted G. splendens in a bloom of diverse dinoflagellate species in Chesapeake Bay. This note reports upon two field studies of blooms of G. splendens. The first observation was made in Coyote Bay of Bahia Concepcion, Gulf of California, where G. splendens was the dominant phytoplankter in March 1971. The second observa- tion was made in March 1974, when large concen- trations were observed in coastal waters of the Southern California Bight. In both occurrences G. splendens dominated the phytoplankton crop so that measurements of primary production and the chemical composition of suspended particles allowed a reasonable description of this species. Gymnodinium splendens in Coyote Bay Coyote Bay Oat. 26°43.0'N, long. 111°53.0'W) of Bahia Concepcion is well protected and shallow. Chemical and physical observations were made while the ship (RV E. B. Scripps) was at anchor in 20 m of water and included measurements of par- ticulate adenosine triphosphate (Holm-Hansen and Booth 1966) and chlorophyll (Yentsch and Menzel 1963; Holm-Hansen et al. 1965), micro- scopic examination of water samples preserved in 5% (V/V) buffered Formalin' (Utermohl 1958), 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. and determinations of primary production based upon rates of incorporation of radioactive carbon (Steemann Nielsen 1952). The depth distribution of phytoplankton was recorded at regular inter- vals by continuous vertical profiles to the bottom with both a submersible transmissometer (Petzold and Austin 1968) and a fluorometer attached to a hose and submersible pump (Lorenzen 1966; Kiefer and Austin 1974). The continuous profiles of in situ fluorescence were translated into chlorophyll a concentrations by frequent calibra- tion with discrete samples which were analyzed fluorometrically for chlorophyll and phaeophytin concentration (Kiefer 1973). The five phytoplankters which occurred together and were numerically most abundant in Coyote Bay were: G. splendens (l.OxlOVliter), Leptocylindrus danicus (3.4x lOVliter), Skele- tonema costatum (1.4 X lOVliter), Cerataulina bergonii (1.4x lO^/liter), and Thalassiothrix frauenfeldii (4.0 x lOVliter). Chlorophyll concen- tration varied with depth and ranged from 4.4 )Ltg/liter to 13 /tg/liter for numerous samplings of the 20-m water column. Figure 1 shows a time sequence of profiles of in situ fluorescence of chlorophyll. Profiles of light transmission displayed a similar stratified structure. The increase in depth of the upper chlorophyll maximum between 1845 and 2300 and the decrease in depth between 2300 and 0720 the following day indicated a diel migration of G. splendens. This suggestion was supported by the predominance of G. splendens in the maxima and by the improbability of physical factors such as advection or internal waves affecting such variations. Con- ditions were calm at the sea surface and the water column was isothermal with depth. The upper chlorophyll maximum (Figure 1) moved downward at sunset with a velocity of approximately 1.7 m/h. Such velocities are similar to those of other dinoflagellates. For example, a natural bloom of Ceratiumfurca occurring off the southern California coast was observed to migrate downward at 2 m/h and mass cultures of Cachonina niei and Gonyaulax polyedra displayed migratory rates of 1 to 2 m/h (Eppley et al. 1968). Our observations suggested that a portion of the G. splendens population moved upward between 2300 and 0400 the following day. Since sunrise was 675 CHLOROPHYLL a (/ng/ L) 0 4 8 12 T UJ Q CHLOROPHYLL a (/xg/L) 0 4 8 12 T 201- CHLOROPHYLL a (^g/L) 0 4 8 12 e 5 F 10 ^ 15 20 3/18/71 1-2300 T 1 IxJ Q CHLOROPHYLL a (fig/L) 0 4 8 12 — r 3/19/71 U 0400 10 15 20"- T T CHLOROPHYLL a (yu.g/L) 0 4 8 12 201- FiGURE 1.— Profiles of the concentration of chlorophyll o based upon fluorescence profiles. The upper layer of Gymnodinium splendens at 1230 h was concentrated at 6 m; by 1845 it had moved to 8 m and reached 15 m by 2300. Movement upward commenced at 0200 reaching 10 m by 0720. The lower layer remained relatively close to the bottom during this time. around 0600, this movement may originate from a biological clock rather than from a phototactic response. These observations partially conflict with laboratory studies of phototaxis in G. splen- dens (Forward 1974). He found that not only was the cell strongly phototactic but that the strength of the response was subject to a circadian rhythm, being strongest at the end of the entrained dark period. By assuming that G. splendens dominated the production as well as the standing crop of phy- toplankton in Coyote Bay, we obtain information on the steady state doubling time for the species. Water samples were collected from four depths, 0, 5, 10, and 18 m, inoculated with NaH^^COsi, and incubated from sunrise to sunset in situ. Primary production was then determined from rates of light-induced incorporation of '^C into particles removed by filtration. The water samples were al- so analyzed for concentrations of chlorophyll a and adenosine triphosphate (ATP). By multiplying the concentration of ATP by 250, we obtained an es- timate of "living-carbon" (Holm-Hansen and Booth 1966); this estimate allowed a crude deter- mination of doubling time it^ from the steady-state equation: h = In 2 C-ln2 where jLi is the specific growth rate and equal to the rate of carbon assimilation, AC/A^, divided by C, the concentration of cell carbon. On this basis, the doubling time for G. splendens at 0, 5, 10, and 18 m was 2.3, 2.6, 2.7, and 62 days, respectively (Table 1). These estimates of doubling time for a natural population of G. splendens may be compared with a maximal doubling time of 1.6 days for cells grown in the laboratory (Thomas et al. 1973). We also note that our estimates of the chlorophyll a concentration per cell yielded a value of approximately 100 pg/cell, typical for laboratory cultures (Bailey 1974). Table l.-Production, chlorophyll a, ATP, and doubling time for a Gymnodinium splendens bloom in Coyote Bay, Gulf of California. Chlorophyll a Doubling Depth Production concentration ATP time (m) ^u,g C/liter»day) ^ug/liter) (fig/liter) (days) 0 125 5.8 1.7 2.3 5 108 5.3 1.6 2.6 10 126 6.2 1.9 2.6 18 35 4.7 1.3 62 /* AC/A« Gymnodinium splendens in the Southern California Bight A second bloom of G. splendens was observed in March 1974, along the southern coast of California, 676 during a cruise on RV David Starr Jordan. Sta- tions in the sampling program extended along the 20-fathom contour from Malibu (lat. 34°00.8'N, long. 118°40.6'W) south to San Onofre (lat. 32°56.0'N, long. 117°17.4'W). Intermediate sta- tions included Manhattan Beach (lat. 33°52.5'N, long. 118°27.0'W), Seal Beach (lat. 33°36.5'N, long. 119°04.3'W), and Dana Point (lat. 33°26.3'N, long. 117°42.8'W). A sixth station was on the 270- fathom contour off Laguna Beach (lat. 33°30.8'N, long. 117°50.3'W). Continuous vertical profiles of in situ chlorophyll fluorescence were made to a depth of 35 m. Water from the outflow of the fluorometer was collected at the surface and within the fluorescence maximum. Three analyses were made on sub- samples of water from the two depths. First, the size distribution of suspended particles was im- mediately determined with an electronic particle counter (model Ta Coulter Counter). We ac- cumulated counts in the upper nine channels which gave us a frequency distribution for particles with equivalent diameters ranging from 20 to 128 /^m. Second, subsamples were preserved in 5% For- malin for species determination. Third, the chlorophyll a concentration in each subsample was determined fluorometrically for acetone extracts of filtered particles. Vertical profiles made at various times of the day and night at each of the six stations on the 20-fathom contour were characterized by a unimodal distribution of chlorophyll. The chlorophyll maximum varied little in depth from 15 to 20 m within a moderately developed ther- mocline, and was most often less than 4 m thick. At these stations, G. splendens contributed most of the phytoplankton crop within the maxima. However, in surface waters it contributed a much smaller fraction of the crop. The highest concen- tration (Figure 2) of G. splendens was within the well-defined maximum at Seal Beach where its concentration reached 4x10^ cells/liter (chlorophyll a = 42.0/xg/liter). The lowest concen- tration within a maximum was at Manhattan Beach, 1.2x10'' cells/liter (chlorophyll a = 1.3 /Ag/liter. The predominance of G. splendens in the chlorophyll maximum was also evident from the particle size distributions obtained with the elec- tronic particle counter. Within the maximum, particles with equivalent diameters between 36 and 57 fxm far outnumbered smaller- and larger- sized particles. In surface waters the smaller-sized CHLOROPHYLL a ( ^g/L ) 0 I? 24 36 48 T T" 1 1 1 1 1 SEAL BEACH STATION 3/21/74 0845 I I- CL CHLOROPHYLL a ( ^g / L ) 0 0.5 1.0 1.5 2.0 2.5 20 30 1 r MALIBU STATION 3/20/74 0935 Figure 2.-Profiles of the concentration of chlorophyll a based upon fluorescence profiles at two stations along the 20-fathom contour of the Southern California Bight. The subsurface maxima are predominantly composed of Gymnodinium splendens. particles outnumbered particles of the size of G. splendens. The other phytoplankton in surface waters included Ceratium furca and C. kofoidii, Dinophysis acuminata, and a species of Gyrodinium. Very few diatoms were present. It appeared that the bloom of G. splendens dis- sipated seaward since the subsurface chlorophyll maximum was poorly developed at the Laguna Beach station which was on the 270-fathom con- tour. Here the concentration of chlorophyll in the maximum was only 0.76)u,g/liter, while the concen- tration at the surface was 0.63 /ig/liter. In addi- tion, both the particle size distribution and microscopic counts indicated a more diverse as- semblage of dinoflagellate species at this station, with the unarmored dinoflagellate Cochlodinium catenatum being most abundant. Thus, the bloom of G. splendens appeared to be limited to nearshore waters, in a band extending as far as 100 km along the coast. This subsurface bloom was presumed to be a large food source for planktonic herbivores, but more field sampling is necessary to determine whether the bloom is a seasonal occurrence. In another paper, Lasker (1975) describes the feeding responses of anchovy larvae to these natural concentrations of G. splendens. Acknowledgments This work was supported, in part, by the Na- tional Science Foundation (GA-36511). We thank Eileen Setzler, Anne Dodson, and Freda Reid for 677 identification and enumeration of preserved phy- toplankton, and Tom Herman for measurements of primary production. Literature Cited Bailey, T. 1974. Chemical responses of two marine dinoflagellates to nitrate and phosphate limitation. M. A. Thesis, Univ. California, Santa Barbara. Barnett, a. M. 1974. The feeding ecology of an omnivorous neritic copepod, Labidocera trispinosa (Esterly). Ph.D. Thesis, Univ. California, San Diego, 215 p. Eppley, R. W., 0. Holm-Hansen, and J. D. H. Strickland. 1968. Some observations on the vertical migration of dinoflagellates. J. Phycol. 4:333-340. Forward, R. B., Jr. 1974. Phototaxis by the dinoflagellate Gymnodinium splen- dens Lebour. J. Protozool. 21:312-315. Holm-Hansen, 0., and C. R. Booth. 1966. The measurement of adenosine triphosphate in the ocean and its ecological significance. Limnol. Oceanogr. 11:510-519. Holm-Hansen, 0., C.J. Lorenzen, R.W. Holmes, and J. D. H. Strickland. 1965. Fluorometric determination of chlorophyll. J. Cons. 30:3-15. KlEFER, D. A. 1973. Fluorescence properties of natural phytoplankton populations. Mar. Biol. (Berl.) 22:263-269. KlEFER, D. A., AND R. W. AUSTIN. 1974. The effect of varying phytoplankton concentration on submarine light transmission in the Gulf of Califor- nia. Limnol. Oceanogr. 19:55-64. Lasker, R. 1975. Field criteria for survival of anchovy larvae: The rela- tion between inshore chlorophyll maximum layers and successful first feeding. Fish. Bull., U.S. 73:453-462. Lasker, R., H. M. Feder, G. H. Theilackek, and R. C. May. 1970. Feeding, growth, and survival of Engraulis mordax larvae reared in the laboratory. Mar. Biol. (Berl.) 5:345-353. LOFTUS, M. E., D. V. SuBBA Rao, and H. H. Seliger. 1972. Growth and dissipation of phytoplankton in Chesapeake Bay. L Response to a large pulse of rain- fall. Chesapeake Sci. 13:282-299. Lorenzen, C. J. 1966. A method for the continuous measurement of in vivo chlorophyll concentration. Deep-Sea Res. 13:223-227. Paffenhofer, G.-A. 1970. Cultivation of Calanus helgolandicus under controlled conditions. Helgolander wiss. Meeresunters. 20:346-359. 1971. Grazing and ingestion rates of nauplii, copepodids and adults of the marine planktonic copepod Calanus hel- golandicus. Mar. Biol. (Berl.) 11:286-298. Petzold, T. J., and R. W. Austin. 1968. An underwater transmissometer for ocean survey work. Scripps Inst. Oceanogr. Ref. 68-59, 5 p. Pokorny, K. S., and K. Gold. 1973. The morphological types of particulate inclusions in marine dinoflagellates. J. Phycol. 9:218-224. Steemann Nielsen, E. 1952. The use of radio-active carbon (C^'*) for measuring organic production in the sea. J. Cons. 18:117-140. Sweeney, B. M. 1954. Gymnodinium splendens, a marine dinoflagellate requiring vitamin Bi2- Am. J. Bot. 41:821-824. Thomas, W. H., A. N. Dodson, and C. A. Linden. 1973. Optimum light and temperature requirements for Gymnodinium splendens, a larval fish food organism. Fish. Bull., U.S. 71:599-601. Utermohl, H. 1958. Zur Vervollkommnung der quantitativen Phy- toplankton-Methodik. Int. Ver. Theor. Angew. Limnol. Verb. 17:47-71. Yentsch, C. S., and D. W. Menzel. 1963. A method for the determination of phytoplankton chlorophyll and phaeophytin by fluorescence. Deep-Sea Res. 10:221-231. Dale A. Kiefer Scripps Institution of Oceanography University of California La Jolla, CA 92037 Reuben Lasker Southwest Fisheries Center National Marine Fisheries Service, NOAA La Jolla, CA 92037 ENHANCED SURVIVAL OF LARVAL GRASS SHRIMP IN DILUTE SOLUTIONS OF THE SYNTHETIC POLYMER, POLYETHYLENE OXIDE' Small amounts of linear, high molecular weight synthetic polymers when added to liquids can sig- nificantly reduce frictional resistance in turbulent pipe and channel flow (Castro and Squire 1967; Peterson et al. 1974). These drag-reducing agents have potential for improving efficiency of sewer, water, and fire-fighting systems (Castro 1972); reducing friction around ships' hulls (Wade 1973); and perhaps increasing water flow and circulation in mariculture operations (Zielinski et al. in press). Such uses may result in the introduction of rela- tively large quantities of polymers into nearshore marine and estuarine waters or culture tanks. We report here experiments to evaluate effects of chronic exposure to polyethylene oxide, a very effective friction-reducing additive, on larvae of estuarine grass shrimp, Palaemonetes vulgaris and P. pugio. This polymer exhibits a very low 'Contribution No. 22 from the South Carolina Marine Resources Center. This work is a result of research sponsored by NOAA Office of Sea Grant, under Grant #NG-33-72. 678 degree of toxicity and is approved for food contact applications and as an additive to some foods (Smyth et al. 1970). Palaemonetes shrimp were chosen for this study because of their importance in estuarine food chains (Hedgpeth 1947; Welsh 1973), the ease with which their larvae may be cultured in the laboratory, the general similarity of their larvae to those of Macrobrachium shrimp being evaluated for commercial culture, and the known sensitivity of these carideans to a variety of toxic agents (Lowe et al. 1971; Hansen et al. 1973; Redmann 1973; Sandifer and Shealy 1974^). Methods and Materials Two experiments were conducted with P. vul- garis, one with P. pugio. Effects of three polyethylene oxide concentrations (25, 50, and 100 wppm— weight parts per million) were tested ver- sus controls in all experiments. Forty Palaemonetes vulgaris larvae were reared at each condition in experiments I (10 replicates of 4 animals each in 10.5-cm finger bowls) and II (4 replicates of 10 animals each in 19.1-cm bowls), while 72 P. pugio zoeae (4 replicates of 18 each) were maintained at each concentration in experiment III. The P. pugio larvae were isolated in compartments of covered plastic boxes. The culture containers were placed in a Percival Model I-35VL incubator* (Percival Manufacturing Co., Boone, Iowa) where temperature was held at approximately 25°C in experiment I and 28°C in the subsequent trials. A 14-h light - 10-h dark schedule was maintained in all studies. All animals were transferred to clean containers with fresh, filtered sea water (30°/ oo salinity) and fed newly hatched nauplii of Artemia salina daily. Fresh stock solutions (200 wppm) of polyethylene oxide (Polyox Coagulant, molecular weight approximately 5x10^ [Union Carbide Corp.]) in sea water were prepared every 3 or 4 days. Test solutions were prepared by diluting the stock with appropriate volumes of seawater. Results and Discussion Addition of small amounts of polyethylene oxide ^Sandifer, P. A., and M. H. Shealy, Jr. 1974. Some effects of mercury on survival and development of larval grass shrimp, Palaemonetes vulgaris (Say). (Unpubl. manuscr.) 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA; NOAA Office of Sea Grant; or the State of South Carolina. to the culture water significantly enhanced the survival of grass shrimp larvae in static water culture (Figure 1). The polymer affected neither the number of molts to the postlarval stage nor the size of postlarvae produced. However, a slight but definite trend toward increasing development time with increasing polyethylene oxide concen- tration was apparent in all experiments (Table 1). Stranding of larvae above the waterline on the walls of the culture containers was a significant cause of mortality in all control cultures, but addi- tion of >25-wppm polyethylene oxide virtually eliminated stranding deaths (Figure 1). This ef- fect was probably the result of the reduced surface tension and increased viscosity, lubricity, and stringiness of the treatment solutions. Of course, this type of effect would not be manifested in 100 90- 80- 70- 60- 50- 40 30 20- 10- 0 i I I Survival to Postlarva lijJH First Doy Mortolity ^ Stranding Mortality Other Mortality lOO-i SO- SO 70-1 60 c S 50- " 40 30-1 20 0 25 50 100 Polyethylene Oxide Concentratlon(wppm) b I 0 25 50 100 Polyethylene Oxide Concentration (wppm) 100 90-1 80 70 60- 50- 40 30H 20 10 0 I M. 0 25 50 100 Polyethylene Oxide Concentration (wppm) Figure l.-Percentage survival and mortality of Palaemonetes larvae reared in polymer and control solutions, (a) P. vulgaris experiment I, (b) P. vulgaris experiment II, (c) P. pugio. 679 Table 1.- Development of Palaemonetes larvae exposed to polyethylene oxide solutions (Mean with standard deviation). Species Pol yethylene oxide concentration 1 [wppm) Experiment 0 25 50 100 Development time (days) 1 Palaemonetes vulgaris 19.1 ± 1.5 20.9 ±2.0 21.3 ±2.5 22.1 ± 2.8 II P. vulgaris 14.3 ±0.9 14.6 ±0.8 14.9 ±0.9 15.6 ± 1.4 III P. pugio 15.5 ± 1.6 15.5 ±1.5 16.2 ±1.4 Molts to postlarva 16.2 ±2.1 III P. pugio 8.0 ± 0.8 8.1 ±0.9 8.5 ±0.8 Lengtti of postlarvae (mm) 8.3 ± 1.4 II P. vulgaris 6.5 ± 0.6 6.4 ±0.5 6.4 ±0.5 6.1 ±0.5 III P. pugio 6.6 ±0.6 6.6 ±0.6 6.6 ±0.6 6.6 ±0.5 natural waters, but it may appear in tank culture operations. First-day mortalities were significant only in the higher treatment concentrations when P. pugio larvae were reared in covered plastic boxes (Figure Ic). These deaths apparently were the result of oxygen depletion in the culture water caused by overfeeding and the relatively high biochemical oxygen demand of the polymer solu- tions (Wade 1973). Other mortalities totaled only 5.6 and 6.9% in the 50- and 100-wppm concentra- tions, respectively, after the first day. In all but one instance, larvae in the polymer solutions exhibited a marked reduction in other mortalities (i.e., "natural" deaths) over the con- trols. Thus, in addition to eliminating stranding, the polyethylene oxide somehow acted to reduce other causes of mortality. The reason for this beneficial effect is unknown, but it is unlikely to be nutritional since, in vertebrates at least, the polymer is poorly absorbed from the gut (Smyth et al. 1970). Further study is needed to examine the reasons for this effect and to evaluate the poten- tial of polyethylene oxide for use in mariculture operations. Acknowledgments We thank J. Williams for assistance in the laboratory and E. Myatt for preparing the figure. Literature Cited Castro, W. E. 1972. Reduction of flow friction with polymer additives. Water Resour. Res. Inst., Clemson Univ., ilep. 24:1-56. Castro, W. E., and W. Squire. 1967. The effect of polymer additives on transition in pipe flow. Appl. Sci. Res. 18:81-96. Hansen, D. J., S. C. Schimmel, and J. M. Keltner, Jr. 1973. Avoidance of pesticides by grass shrimp {Palaemonetes pugio). Bull. Environ. Contam. Toxicol. 9:129-133. Hedgpeth, J. W. 1947. River shrimps, interesting crustaceans about which little has been written. Prog. Fish-Cult. 9:181-184. Lowe, J. L, P. R. Parrish, A. J. Wilson, Jr., P. D. Wilson, and T. W. Duke. 1971. Effects of mirex on selected estuarine organisms. Trans. 36th North Am. Wildl. Nat. Resour. Conf., p. 171-186. Peterson, J. P., W. E. Castro, P. B. Zielinski, and W. F. Beckwith. 1974. Increased turbulent dispersion in high polymer drag reducing open channel flow. ASCE (Am. Soc. Civ. Eng.) Proc., Hydraul. Div. Hy6:773-785. Redmann, G. 1973. Studies on the toxicity of mirex to the estuarine grass shrimp, Palaemonetes pugio. Gulf Res. Rep. 4:272-277. Smyth, H. F., Jr., C. S. Weil, M. D. Woodside, J. H. Knaak, L. J. Sullivan, and C. P. Carpenter. 1970. Experimental toxicity of a high molecular weight poly (ethylene oxide). Toxicol. Appl. Pharmacol. 16:442-445. Wade, R. H. 1973. The pollution potential of drag reducing polymers. In N. D. Sylvester (editor). Drag reduction in polymer solu- tions, p. 87-90 Am. Inst. Chem. Eng., N.Y. Welsh, B. L. 1973. The grass shrimp, Palaemonetes pugio, as a major component of a salt marsh ecosystem. Ph.D. Thesis, Univ. Rhode Island, Kingston, 90 p. Zielinski, P. B., W. E. Castro, and P. A. Sandifer. In press. The evaluation and optimization of Macrobrachium shrimp larva tank designs and support systems. Proc. World Mariculture Soc. 5. Paul A. Sandifer Marine Resources Research Institute P.O. Box 12559 Charleston, SC 29^12 College of Engineering Clemson University Clemson, SC 29631 Paul B. Zielinski Walter E. Castro 680 A COMPUTER PROGRAM FOR ANALYSIS OF POLYMODAL FREQUENCY DISTRIBUTIONS (ENORMSEP), FORTRAN IV Program ENORMSEP (Extended Normal Separator Program) separates a polymodal frequency distribution into its component groups where aging studies have not been or cannot be performed. The program calculates preliminary estimates of the number of size groups and their points of overlap using probit analysis and polynomial regression techniques. These preliminary estimates are then entered into NORMSEP (Normal Separator Program) (Has- selblad 1966), used as a subroutine, in order to complete the analysis. Output data are generated both as listings and punched cards. Listings include at the option of the user: 1) table of values of the standardized normal distribution; 2) table of values of probabilities, standardized normal variables, and probits; 3) polynomial regressions and analyses of variance of probits; 4) table of residuals for the final regression; 5) table of roots corresponding to all regressions after taking second derivative; 6) tables for analyses for the separation of modes; 7) plots of observed and predicted values for the final regression; and 8) plot of the original frequency distribution. Punched card output includes the number of observed frequency distributions with their intervals and probits and regression coeflftcients for the polynomials. Input data require the observed size frequency together with values for identification and control purposes. No more than nine size groups may be separated because of limits on the eflftciency of parameter estimate in the polynomial regression. This computer program was developed on an IBM 360/651 computer' using release 20.7 MVT/HASP system at the Statistical and Com- puting Center at the University of Hawaii. This computer program is capable of processing mul- tiple sets of data. For a "typical" problem, the program takes about 1 min of central processing unit time and a total machine unit time of 1.5 min to run a single problem "individually." The requirement for core storage is 168K, where K is 1,024 bytes and where a byte is an address collec- tion consisting of eight binary bits or binary digits. A description of the program, including program listing as well as input and output for two examples, is available from the authors upon request. Literature Cited Hasselblad, v. 1966. Estimation of parameters for a mixture of normal distributions. Technometrics 8:431-444. Marian Y. Y. Yong Robert A. Skillman Southwest Fisheries Center Honolulu Laboratory National Marine Fisheries Service, NOAA Honolulu, HI 96812 'Reference to this particular computer system does not imply endorsement of the product by the National Marine Fisheries Service, NOAA, but is given to provide the reader with a base for determining the cost of performing jobs with the particular computer system at his disposal. RECORDS OF LARVAL, TRANSFORMING, AND ADULT SPECIMENS OF THE QUILLFISH, PTILICHTHYS GOOD EI, FROM WATERS OFF OREGON This report extends the southern range of the quillfish, Ptilichthys goodei Bean 1881, in the northeast Pacific to waters off the central coast of Oregon where larval, transforming, and adult specimens have been collected. The previously reported range of this species in the North Pacific was from the Okhotsk and Bering seas to northern Washington and Puget Sound (DeLacy et al. 1972; Quast and Hall 1972; Hart 1973). The life history of the quillfish is poorly understood and nothing is known of the early stages (Walker 1953; Makushok 1958; Grinols 1965; Hart 1973). Materials and Methods Three larvae (20.3, 24.7, 36.0 mm SL-standard length) and one transforming specimen (114 mm SL) of P. goodei came from plankton collections made with large-mouth (0.7 m) bongos having 0.571-mm mesh nets. Tows were made in a step- oblique or oblique manner from near the bottom or 150 m (at deeper stations) to the surface at a vessel speed of 2 knots. Tow times were 16 to 25 min. The specimens were fixed in 10% and stored in 5% buf- fered Formalin.' 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 681 Two adults (272, 309 mm SL) of P. goodei came from otter trawl (5-m headrope, 4-cm stretch mesh body, 1.25-cm mesh cod end liner) collections. The trawl was towed on the bottom for 15 min at a speed of 2 to 3 knots. The specimens were fixed in 10% Formalin and stored in 40% isopropyl alcohol. Body measurements were made as described by Hubbs and Lagler (1958). For larvae, standard length was measured from snout tip to notochord tip. The point of reference used to determine no- tochord tip in larvae of P. goodei is indicated by an arrow in Figure 1. It is the point near the end of the tail, at the posterior edge of the region having no pigment on the ventral margin. This point was determined from the tail tip of the 114-mm specimen which had been stained with alizarin red S. Meristic counts were made on unstained larval and transforming specimens and radiographs of the adults. Results Descriptions The larvae of P. goodei are characterized by their slender, elongate form; gut length (35-40% SL, decreasing with growth); myomere numbers, (55 to 57) -I- (170 to 174) = 225 to 229; and pigment pattern (Figure 1). Morphometries and meristics are in Table 1. Compared with adults, the larvae have a short snout (17-18% HL-head length) and no fleshy protrusion of the lower jaw. The mouth is oblique. Dorsal and anal fin rays are evident in the 36-mm specimen but the adult numbers have not been attained. The spines of the first dorsal and the rays of the second dorsal begin to form at the posterior and anterior ends of the fins respec- tively. Development then proceeds anteriorly in the first dorsal fin and posteriorly in the second dorsal. The anal fin rays begin to form slightly anterior to the center of the fin with development proceeding anteriorly and posteriorly. Pectoral fin rays were not formed and pelvic fins were absent in the size range examined. Pigmentation (Figure 1) on the three larval specimens is similar. Head pigmentation consists of that on the lower jaw, anterior part of the upper jaw, throat, and internally at the base of the hindbrain. Gut pigmentation is mostly restricted to the dorsal and ventral surface with some addi- tional melanophores scattered over the anterior region. The melanophores on the ventral gut sur- !i h3 CO I £ I •s I § I I I I H g i ■ 682 Table 1.— Morphometries (mm) and meristics of Ptilichthys goodei from Oregon waters. Mo rphometrics Meristics Snout length Eye diam- eter Head length Upper jaw length Lower jaw length Snout to anus Depth at pectoral base Depth at anus P 1 Bran- chios- tegals Myomeres SL TLI Dl Dl|2 A2 Left Right Preana 1 Postanal Total Larvae: 20.3 21.0 0.4 0.7 2.3 0.8 1.1 8.0 0.9 0.9 — — — — — 55 174 229 24.7 25.5 0.4 0.6 2.2 0.7 1.1 9.3 1.1 1.1 — — — — — — 55 170 225 36.0 36.9 0.5 0.7 2.7 0.9 1.1 12.5 1.1 1.1 40 '>.80'wll3 — — 5 57 170 227 Transformli ig: 114 117 1.6 1.1 6.1 1.8 2.6 32.2 2.0 2.0 83 148 179 13 13 5 55 172 Vertebrae 227 Adults: Abdominal CaudaP Total 272 276 5.2 2.8 17.0 3.8 5.0 81 5.5 7.0 83 144 180 13 13 5 53 174 227 309 313 6.3 3.0 19.0 4.2 6.0 91 6.0 7.7 88 142 181 13 13 5 53 174 227 'Total length is given for comparison with other publications, although the long caudal filament was not intact. ^Dorsal and anal fin ray counts include possible caudal fin elements dorsal and ventral to the fleshy caudal extension respectively. 30ne fused caudal vertebra found in each adult specimen was counted as one vertebra. face form a single row on the anterior one-fourth to one-half the length of the gut and a double row along the remaining length. Body pigmentation is concentrated dorsally and ventrally. From a dorsal view, the dorsal melanophores appear somewhat as a double row, one on each side of the dorsal midline, extending from a point over the middle of the gut to near the tail tip (arrow in Figure 1). These dorsal melanophores are larger than the ventral ones. Ventrally, a heavy concentration of melanophores lines the body margin from the hindgut to near the tail tip. Posterior to this ven- tral body pigment is a small unpigmented area. Posterior to the ventral unpigmented area (arrow in Figure 1) is the fleshy caudal extension characteristic of the species. Pigment on -this ex- tension is distinct from that on the rest of the body. It is scattered rather evenly dorsally and ventrally on the body and out on to the finfolds. The lateral midline of this area remains unpig- mented. Identification of the larvae was possible because of the link to the adults provided by the 1 14-mm SL transforming specimen captured in the same area. The 114-mm specimen has meristics (Table 2), hooked dorsal spines, and a fleshy protrusion of the lower jaw characteristic of adult P. goodei and pigmentation similar to the larvae described above. The fleshy caudal extension is more distinct than in the larvae. Gut length (28% SL) is propor- tionately shorter and snout length (26% HL) proportionately longer. Additional pigment occurs on the dorsal surface of the head posterior to the eye, on the snout, and in a line along the margin of the preopercle extending posteriorly from the angle of the lower jaw. The ventral gut melanophores are in a single row along the entire gut length. Body pigmentation is less pronounced than in the larvae but still distinct. Adults of P. goodei are characterized by their extremely elongate body, the absence of a distinct caudal fin, and the presence of a fleshy protrusion at the tip of the lower jaw. When alive, the two Oregon specimens were brightly colored. The body was light green dorsally shading to yellow ven- trally and orange on the throat. Two dark maroon, horizontal stripes were present laterally with maroon spots scattered over the entire dorsal sur- face. Several dashed maroon lines radiated posteriorly from the snout. A distinct maroon- colored, horizontal bar extended anteriorly from the margin of each eye half the distance to the snout tip. Morphometries and meristics appear in Table 1. Gut length (29% SL) is similar to that for the transforming specimen, but the snout length (31-33% HL) is greater. Both specimens exhibited a vertebral anomaly in which the centra of two adjacent vertebrae were fused to form a single element with two neural and two hemal spines. In the 309-mm SL specimen the 160th vertebra was fused and in the 272-mm SL specimen it was the 169th element. Occurrence Collection data for P. goodei from Oregon waters is presented in Table 2. All specimens came from waters off the central coast of Oregon between March and August. All but one was captured dur- ing daylight. All but one was taken in water greater than 120 m deep on the continental shelf 18 km or closer to shore where the bottom was primarily gray sand. 683 Table 2.— Collection data for Ptilichthys goodei from Oregon waters. Location Km from coast Coastal reference Time Bottom depth (m) Tow depth (m) Gear Bottom type' Surface temp SL (mm) Date Lat. N Long. W ( C) Larvae: 20.3 25 Mar. 1973 44^00' 124=22.1' 18 Siuslaw River 1526-1542 117 100-0 Bonoos Gray sand- 11.0 24.7 20 Apr. 1973 44=00' 124°22.1' 18 Siuslaw( River 0235-0300 109 75-0 Bongos green mud Gray sand- 10.4 36.0 14 May 1971 44=39.1' 124°17.7' 18 Newport 1621-1641 80 75-0 Bongos green mud Gray sand 11.2 Transforming: 114 29 June 1971 44°39.1' 124°52.7' 65 Newport 1128-1153 340 150-0 Bongos (Slope) 14.5 Adults: 272 7 Aug. 1973 309 3 July 1973 44^42' 44°45' 124°7' 124°14' 5 13 Moolach Beach Cape Foulweather 0945-1007 1010-1028 52 80 52 80 Otter trawl Otter trawl Gray sand Gray sand 9 10 iFrom USC&GS Charts No. 5702 and No. 5802. The larval and transforming specimens report- ed here are the only representatives of P. goodei found in 847 small-mouth (0.2 m) bongo and 413 large-mouth (0.7 m) bongo samples analyzed to date from waters off Oregon. The samples are part of an ongoing project to study seasonal and annual variations in distribution and abundance of larval fishes. Other studies of larval fishes off Oregon (Richardson 1973; Pearcy and Myers 1974) yielded no P. goodei. The two adult specimens are the only ones recovered from 23 trawl samples taken during the summer of 1973 in conjunction with an ecological baseline study of the nearshore region of the mid- Oregon coast in the vicinity of Yaquina Head. Although they were taken with a bottom trawl, it is possible that the specimens entered the net shortly before it was brought on board. Thus, they may have been in near surface water rather than on the bottom as their presence in trawl samples would suggest. Discussion Vertebral counts, (53 to 55) -I- (172 to 174) = 227, of the transforming and adult Oregon specimens are lower than those, (58 to 59) + (179 to 182) = 238 to 240, reported by Makushok (1958), presumably for Bering Sea specimens. The counts are also lower than those, 236 to 240, given by Hart (1973), presumably for British Columbia specimens. This could indicate clinal variation with the southern specimens having fewer ver- tebrae. On the other hand, the lower number of both abdominal and caudal vertebrae of the Oregon specimens could indicate the presence of an undescribed species. Additional Oregon specimens are needed to determine the range of variation in vertebral number and to compare with northern specimens to see if they are actually the same species. Reasons why quillfish have not previously been reported from Oregon waters are speculative. A partial explanation may be the lack of major sampling efforts in Oregon's coastal zone until recent years. Rarity (Makushok 1958; Hart 1973), inaccessibility, avoidance, and escapement also offer explanations. It is possible the adults bury themselves in the bottom (Makushok 1958) and are thus inaccessible to conventional types of gear. Behavior exhibited by one of the adults taken off Oregon suggests the quillfish may readily avoid and/or escape from trawl gear. Immediately after the trawl was brought aboard, the slender fish wriggled through the meshes of the net onto the deck of the vessel. It demonstrated great agility with snakelike undulations. The larvae may remain on or near the bottom, or they may spend all or part of the time in the neuston. Either sit- uation would make them inaccessible to most plankton gear. The large size of the larvae in- dicates good avoidance capabilities. Acknowledgments Thanks are extended to Sharon Roe, David Stein, and Elbert H. Ahlstrom for corroboration of the larval identification; to April G. McLean for preparing the radiographs; and to Carl Bond for reviewing and commenting on the manuscript. This research was supported by NO A A Institu- tional Sea Grant 04-3-158-4 and by the Eugene Water and Electric Board, Pacific Power and Light Company, and Portland General Electric Com- pany. 684 Literatured Cited DeLacy, a. C, B. S. Miller, and S. F. Borton. 1972. Checklist of Puget Sound fishes. Univ. Wash. Sea Grant Publ. 72-3, 43 p. Grinols, R. B. 1965. Check-list of the offshore marine fishes occurring in the northeastern Pacific Ocean, principally off the coasts of British Columbia, Washington, and Oregon. M.S. Thesis, Univ. Washington, Seattle, 217 p. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can., Bull. 180, 740 p. HuBBS, C. L., and K. F. Lagler. 1958. Fishes of the Greak Lakes region. Revised ed. Cranbrook Inst. Sci., Bull. 26, 213 p. Makushok, V. M. 1958. The morphology and classification of the northern blennioid fishes (Stichaeoidae, Blennioidei, Pisces). Proc. Zool. Inst., Akad. Nauk SSSR Tr. Zool. Inst. 25:3- 129. (Translated by A. R. Gosline, Ichthyol. Lab., U.S. Fish Wildl. Serv., 1959, 59 p.; U.S. Natl. Mus., Wash., D.C.) Pearcy, W. G., and S. S. Myers. 1974. Larval fishes of Yaquina Bay, Oregon: A nursery ground for marine fishes? Fish Bull., U.S. 72:201-213. QuAST, J. C, AND E. L. Hall. 1972. List of fishes of Alaska and adjacent waters with a guide to some of their literature. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-658, 47 p. Richardson, S. L. 1973. Abundance and distribution of larval fishes in waters off Oregon, May-October 1969, with special emphasis on the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 71:697-711. Walker, E. T. 1953. Records of uncommon fishes from Puget Sound. Copeia 1953:239. Sally L. Richardson Douglas A. DeHart School of Oceanography Oregon State University Carvallis, OR 97331 EFFECT OF CROWDING ON STOCK AND CATCH IN TILAPIA MOSSAMBICA In a previous report (Silliman 1972) I described the effect of crowding on the relation between exploitation and yield in Tilapia macrocephala. Subsequently I performed a similar experiment with T. mossambica. Since the results were somewhat different for the latter species and because of its wide use in pond culture, a brief report of the second experiment seems justifiable. Apparatus and Procedures Most of the procedures and apparatus were identical with those reported by Silliman (1972). Essentially the approach was to raise the popula- tions in two conventional aquariums, one (L) with a volume of 155.2 liters and the other (S) with 77.6 liters so that S had exactly one-half the capacity of L. Aeration was by airstones and illumination by overhead fluorescent lamps. Rectangular spaces at the ends of the aquariums were fenced off with rods placed 3 mm apart, providing refuges for the young. Further shelter was provided by floats with suspended cords and by fiber brush shelters. Covering part of the aquarium walls with black plastic furnished shaded areas for spawning. Water condition was maintained by filtration and weekly partial water changes. Water temperature was 24° + 2°C to month 5.7 and 30° ± 2°C thereafter. Feeding details are given in Table 1. Populations were counted and weighed at approximately 2-mo intervals. Since T. mossam- bica is a mouthbreeder, it was desirable not to handle the fish more often than this. The 2-mo period includes 1.0 to 2.6 of the brood intervals reported by various authors (Kelly 1957, 30-40 days; Swingle 1960, 30-40 days; Uchida and King 1962, 23-61 days). Exploitation consisted of removing each 10th fish. In weighing, fish were drained in a net and placed in a previously weighed container of water; fish weight was total weight less the tare. Table 1.— Food (in g) placed in tanks. Day of Trout pellets Tropical fish food week Moist Dry Ai B> Total Sun. 4.0 1.5 0.5 1.0 7.0 Men. 5.5 1.5 0.5 1.5 9.0 Tues. 5.5 1.5 0.5 1.5 9.0 Wed. 5.5 1.5 0.5 1.5 9.0 Thurs. 5.5 1.5 0.5 1.5 9.0 Fri. A.M. 5.5 1.5 0.5 1.5 9.0 Fri. P.M.2 5.5 1.5 0.5 1.5 9.0 Total 37.0 10.5 3.5 10.0 61.0 iCommercial makes of dry food. 2This was combined witti the Friday A.M. feeding in 35 out of 131 wk and with the Sunday feeding once. Results and Conclusions The two populations were started 10 July 1970 (Table 2, Figure 1). Recruitment (estimated from counts as in Silliman 1972) occurred after the temperature increase at month 5.7 and readjust- ment of the sex ratios at month 6.9 (Table 2). As was true for T. macrocephala, recruitment was greater in tank L (62) than in tank S (20). Some 685 Table 2.-Population and catch, Tilapia mossambica, in two sizes of tanks. Target exploitation rate was 10% per 2 mo. S— 77. 6-liter tank L — 155.2-liter tank Number Weight (g) Num ber Weight (g) Month' Stock Catch Stock Catch Stock Catch Stock Catch 0.3 211 211 as 316 — — — 316 — — — 4.1 "9 — 650 — "9 — 593 — 6.9 54 — 282 — 54 — 303 — 9.2 20 — 372 — 50 — 529 — 11.1 20 — 641 — 49 — 571 — 13.1 20 — 831 — 46 — 754 — 15.2 20 — 983 — 44 — 900 — 17.0 20 — 1,088 — 46 — 1,006 — 19.1 20 2 1,154 119 46 5 1,081 115 21.2 18 2 1,121 108 41 4 1,047 113 23.0 16 1 1,120 146 36 4 1,071 100 25.1 14 1 987 89 34 3 1,070 100 27.1 12 1 912 90 31 3 1,083 198 29.2 10 1 861 114 29 3 1,043 119 10 = 1 July 1970. ^Initial stocks were: S-6 Immatures, 2 males, 3 females; L-4 immatures, 4 males, 3 females. 3To each stock, 5 Immatures were added. ■•Stocks were adjusted to 3 immatures, 2 males, and 4 females each. sstocks readjusted to 1 male and 3 females eacti. Temperature was increased from 24° to 30°C at month 5.7. I 2n Figure l.-Course of biomass and catch. recruitment occurred throughout the experiment after month 6.9 in L but was limited to 2-mo period 6.9-9.2 in S. Exploitation began at month 19.1 for both populations, at a target rate of 10% per 2 mo. Because of the small numbers of fish in the populations, actual percentages removed (Table 2) varied considerably from 10%. Populations dif- fered in their response, S declining while L remained almost constant (Figure 1). Mean values of catch were S, 111 g; L, 124 g. Although the exploitation data were too few for firm conclusions, they suggest a greater yield from the larger tank, under the same catch rate and food amount. Here the response for T. mossambica was reversed from that found by Silliman (1972) for T. macrocephala. If significant, this difference may be due to the fact that T. mossanbica reaches larger ultimate size than T. macrocephala. The presence of a few large individuals in a population of small numbers (Table 2) could lead to a different response of the population to space available. Literature Cited Kelly, H. D. 1957. Preliminary studies on Tilapia mossambica Peters relative to experimental pond culture. Proc. 10th Annu. Conf. Southeast Assoc. Game Fish Comm., p. 139-149. Silliman, R. P. 1972. Effect of crowding on relation between exploitation and yield in Tilapia macrocephala. Fish. Bull., U.S. 70:693-698. Swingle, H.S. 1960. Comparative evaluation of two tilapias as pondfishes in Alabama. Trans. Am. Fish. Soc. 89:142-148. UcHiDA, R. N., AND J. E. King. 1962. Tank culture of Tilapia. U.S. Fish Wildl. Serv., Fish. Bull. 62:21-52. Ralph P. Silliman Northwest Fisheries Center National Marine Fisheries Service. NOAA Seattle, WA 98112 Present address: il35 Baker NW Seattle, WA 98107 686 THE OCCURRENCE OF ELVERS OF SYNAPHOBRANCHUS AFFINIS ON THE CONTINENTAL SLOPE OFF NORTH CAROLINA' Members of the family Synaphobranchidae are demersal eels which are widely distributed in the Atlantic and Indo-Pacific oceans (Castle 1964). Data from a June 1973 cruise in the Norfolk Canyon area indicate that they are an important part of the fish community in both numbers and biomass in depths around 1,000 m (Virginia Insti- tute of Marine Science unpubl. data). Bruun (1937) reported on the life histories and larval develop- ment of several synaphobranchids, and Castle (1964) listed synonomies in addition to keys to genera and species. Robins (1971) gave os- teological, meristic, and morphometric data and also discussed the life history and ecology (Robins 1968). Although Robins (1971) intensively examined 46 Synaphobranchus affinis Gunter 1877 greater than 193 mm in total length, the occurrence of elvers or unpigmented juveniles of this species is unre- ported. The purpose of this report is to provide a record of capture, meristic and morphometric da- ta, and some observations in food habits of S. affinis elvers. Materials and Methods A total of 89 elvers of S. affinis (Figure 1) were collected during a y2-h otter trawl haul from 1745 to 1815 h EST aboard the RV Eastward on 29 April 1973 at Eastward station 22039, lat. 34°03.2'N, long. 75°52.0'W at depths from 550 to 600 m. The gear used was a 30-foot shrimp trawl with a \4-inch stretch-mesh cod end liner. Total length of all specimens was measured to the nearest millimeter. A subsample of 40 elvers was taken for meristic analysis using a table of random numbers (Rohlf and Sokal 1969). Elvers were cleared in 2% potassium hydroxide, stained with alizarin red-S in 2% KOH, passed through a graded series of glycerine, and stored in 100% glycerine to which thymol was added. Three replicate counts were made of the following meristic characters: total vertebrae; dorsal, anal, caudal, and left and right pectoral fin rays; left and right branchiostegals. The presence and position of vertebral deformities (fused or partially fused 'Virginia Institute of Marine Science contribution no. 652. vertebral centra; extra, fused, or distorted neural or hemal spines) were noted and representative types were drawn with the aid of a camera lucida. To determine if osteological deformities might result in differential mortality of 5. affinis during later life, 40 additional specimens (^ total length = 220 mm, extremes 173-305 mm) collected on 13 June 1973 aboard the RV Columbus Iselin in 630 m of water with a 45-foot otter trawl at lat. 37°03.2'N, long. 74°34.1'W were examined. These fishes were X-rayed, vertebrae counted, and the presence or absence of deformities noted. Frequency of occurrence of deformities in elvers and larger fish were compared by X^ analysis (Sokal and Rohlf 1969). Morphometric measurements were taken following the method of Robins (1971) with either dividers and dial calipers or a binocular microscope fitted with a calibrated ocular micrometer. Results and Discussion Eels of the genus Synaphobranchus are characterized by confluent branchial apertures with a slitlike opening on the midline of the throat (Robins 1971). Synaphobranchid eels commonly encountered in trawls on the continental slope near Cape Hatteras, N.C. are Synaphobranchus kaupi, S. affinis, and Ryophis brunneus (Markle 1972; Virginia Institute of Marine Science unpubl. trawl records). Members of this group show vary- ing degrees of plasticity and overlap in morphometric characters but also show mean differences (Robins 1971). The specific identifica- tion of these elvers as S. affinis, therefore, was based on vertebral counts. Mean, 95% confidence interval, and the frequency distribution of vertebral counts are shown in Figure 2. Vertebral counts of the 40 specimens had extremes of 130 and 136 with a mode of 134. This is within the range of values for 44 S. affinis given by Robins (1971) (x = 133.1, extremes 128-139) and outside the range of the other species of synaphobranchids common to this region {S. kaupi: x = 148.0, extremes 146-150; Ryophis brunneus: x = 147.5, extremes 144-151). Means, 95% confidence intervals, and frequency distributions of other meristic characters are found in Figure 2. Paired t tests (Sokal and Rohlf 1969) showed no significant differences between the number of left and right pectoral fin rays (t = 1.00, df = 39) or the number of left and right branchiostegals {t = 0.42, df = 39). Dorsal and 687 >•"" 'i^]iinjiiif|iiii|ii[ii|Tiiifliii| ^ ' 1 2 3 SPEC. DATE. B i f¥ A ■m9^ ^^ifiiifliiiiniiiniijiriiini * 1 2 3 f ^ .rV 'Jfew Figure 1.-(A) Photograph of a series of Synaphobranchus affinis elvers. (B) close-up view of anterior region of one elver of S. affinis showing dark, bluish-black peritoneum and slight pigmentation above lateral line. 688 IZ9 130 131 132 133 134 135 136 137 13 14 15 16 17 18 IS VERTEBRAL COUNT LEFT PECTORAL FIN RAYS X ± 95 % C I = 2 72 8 ± 5 3 X 1 95% C I = 165 +0 2 ?77pL ^__ _, 230 240 250 260 270 2aO 290 300 310 14 15 16 17 18 19 20 DORSAL FIN RAYS CAUDAL FIN RAYS x± 95% C I = 244.3+ 5 I X ± 95% CI. = 15 9± 0 4 < < a: o o 300- 280 260- 240- 220 200 — r~ 220 T T" "T" 240 ANAL FIN RAYS —I — 260 0 86Xj + 62 6 r = 0 77 r^x 100 = 59% 280 Figure 3.-Regression relationship between dorsal and anal fin rays of Synaphobranchus affinis elvers, where y = dorsal fin rays, x = anal fin rays, r = correlation coefficient, and r^ x 100% = coefficient of determination. ANAL FIN RAYS LEFT BRANCHIOSTEGALS Figure 2.-Means, 95% confidence intervals (C.I.), and frequency distributions of various meristic characters of Synaphobranchus affinis elvers. anal fin rays showed a great degree of variability which is also characteristic of the American eel, Anguilla rostrata (Wenner 1972), and the snake eel, Pisodonophis cruentifer (Wenner unpubl. ob- servations). Dorsal fin rays had extremes of 243 and 309 whereas anal fin rays had extremes of 219 and 271. Plots of dorsal against anal fin rays sug- gested a correlation and therefore the linear regression equation, correlation coefficient, and coefficient of determination were calculated (Figure 3). Fifty-nine percent of the variation in the number of dorsal fin rays was associated with the number of anal fin rays. Mean length of the S. affinis elvers was 89 mm with extremes of 72 and 105 mm. Length- frequency distribution is found in Figure 4. Means, 95% confidence intervals, and extremes for morphometrical measurements are given in Table 1. All values fell within the ranges of those of S. affinis presented by Robins (1971). Osteological deformities associated with the vertebral column were found in 72.5% of the 40 elvers and in 60% of the larger S. affinis examined. In both instances, most deformities were in the caudal part of the vertebral column, generally in the last 30 vertebrae. Illustrations of some representative deformities are shown in Figure 5 X ± 95% CI = 84 9 + I 4mm 75 80 85 90 95 100 105 SIZE INTERVALS (mm) Figure 4.-Frequency distribution, mean, and 95% confidence interval (C.I.) of total length of Synaphobranchus affinis elvers. Table 1.— Summary of morphometric characters of elvers of Synaphobranchus affinis. C. I. refers to confidence interval. Morphometric character Mean ±95% C.I. Extremes n % total length Preanal length 27.7 ± 0.5 25.6-32.0 39 Predorsal length 30.0 ± 0.7 25.6-34.7 39 Head length 13.0 ±0.2 11.3-14.2 40 % head length Gape length 66.9 ± 1.1 58.3-74.5 40 Horizontal eye diameter 16.8 ±0.4 14.8-21.0 40 and a summary of the major types is in Table 2. X^ analysis showed that there was no significant difference between elvers and larger immature fish (X^ = 1.72, df = 39), suggesting that these deformities do not result in differential mortality in fishes possessing them. It is conjectural whether these specimens 689 B I mm I mm -HS I mm Figure 5.— Lateral views of various osteologicai deformities associated with the caudal vertebrae of Synaphobratwhus affinis. Dashed lines in figures represent areas of incomplete ossification. NS = neural spine; NA = neural arch; NC = neural canal; C = centrum; F = foramen; HA = hemal arch; HC = hemal canal; HS = hemal spine; TL = total length in millimeters. (A) Fused centra with a bony knob projecting laterally; TL = 92. (B) Two sets of hemal spines which are fused together forming an arch; TL = 102. (C) Hemal spines projecting anteriorad rather than posteriorad; TL = 9L (D) Extra unfused set of hemal spines with one set projecting anteriorad and one set posteriorad; TL = 79. represent the size at which the elvers descend to the bottom from the pelagic realm, where they pass their larval existence, because closing and mid-water nets were not used. Of the 40 cleared and stained specimens, 12 had material in the gut cavity. Three contained crus- tacean appendages whereas nine had discernable fish remains such as disarticulated vertebrae. One elver contained an intact gonostomatid which had been swallowed head first. Gonostomatids are mesopelagic but the elver could have ingested it while in the trawl. Acknowledgments This research was supported in part by NSF Grant GA-37561, J. A. Musick principal investiga- tor. The specimens were collected on cruise 73-1 of the RV Eastward supported by NSF Grant GA-27725. 690 I Table 2. -Summary of the types of vertebral anomalies in Synaphobranchiis affinis elvers expressed as percent of specimens with anomalies. The sum of the percentages is greater than 100% because some individuals had more than one type. Larger Item Elvers immature fish Sample size 29 24 Abnormal hemal spines 81.6 83.3 Abnormal neural spines 3.4 0.0 Fused or partially fused ve rteb rae 13.6 16.7 Multiple anomalies 57.8 16.7 I thank J. A. Musick of the Virginia Institute of Marine Science and D. M. Cohen of the Systema- tics Laboratory, National Marine Fisheries Ser- vice (NMFS), NOAA, National Museum of Natural History, Wash., D.C. for critical evaluation of the manuscript. B. B. Collette of the Systematics Laboratory, NMFS, NOAA, kindly provided X-ray facilities. R. Bradley illustrated the text and Ken Thornberry photographed the specimens. Thanks are also in order to my fellow students, Labbish Chao and D. Markle, for encouraging me to complete this project. Literature Cited Bruun, A. F. 1937. Contributions to the life histories of the deep sea eels: Synaphobranchidae. Dana-Rep., Carlsberg Found. 2(9):1-31. Castle, P. H.J. 1964. Deep-sea eels: Family Synaphobranchidae. Galathea Rep. 7:29-42. Markle, D. F. 1972. Benthic fish associations on the Continental Slope of the Mid-Atlantic Bight. M. A. Thesis, Coll. William Mary, Williamsburg, Va., 68 p. Robins, C. H. 1968. The comparative osteology and ecology of the synaphobranchid eels of the Straits of Florida. Ph.D. Thesis, Univ. Miami, Coral Gables, Fla., 149 p. 1971. The comparative morphology of the Synaphobranchid eels of the Straits of Florida. Proc. Acad. Nat. Sci. Phila. 123:153-204. ROHLF, F. J., AND R. R. SOKAL. 1969. Statistical tables. W. H. Freeman and Company, San Franc, 253 p. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. The principles and practice of statistics in biological research. W. H. Freeman and Company, San Franc, 776 p. Wenner, C. a. 1972. Aspects of the biology and systematics of the American eel, Anguilla rostrcta (Lesueur). M. A. Thesis, Coll. William Mary, Williamsburg, Va., 109 p. Charles A. Wenner Virginia Institute of Marine Science Gloucester Point, VA 23062 CATCHES OF ALBACORE AT DIFFERENT TIMES OF THE DAY The purpose of this study is to examine the hypothesis that diel variations occur in the catches of albacore by boats trolling surface jigs off Oregon. Although albacore fishermen talk of "morning bites" and "evening bites," no published data exist, to our knowledge, confirming these trends. Studies on the feeding habits of tunas, however, provide evidence for intense feeding activity dur- ing certain periods of the day. Based on the quan- tity of food in stomachs, Iverson (1962) concluded that major feeding periods of albacore occurred in early morning and late afternoon-evening. Similarly, Nakamura (1965) and Dragovich (1970) found evidence for morning and late afternoon peaks in the stomach fullness of skipjack and yellowfin tunas. Food consumption of captive skipjack was greatest between 0630 and 0830 h, and skipjack tuna in only one of three tanks fed intensively in late afternoon (Magnuson 1969). This was in agreement with Uda (1940) who reported that catches of skipjack tuna with pole and live bait peaked in early morning hours and were usually followed by successively lower peaks later in the day. Fishermen were solicited to record data on 1969 and 1970 albacore catches in special logbooks. Several entries per day were requested. The records of five boats fishing off Oregon during July, August, and September 1969 were used for this study. These five skippers kept detailed records averaging eight entries per day. In 1970 the records of 12 boats were used that recorded catches at least every 4 h during the fishing day for 20 July-2 August 1970. All boats were 45-60 feet in length. Average catches per boat were calculated for each hour fished, usually 0500-2200 h or 2300 PDT, for 3 mo in 1969 and 2 wk in 1970. When the inter- val between logged catches was greater than 1 h, the catch for the interval was divided by the number of hours fished, and this average number was distributed uniformly within the interval. Because the selected boats did not necessarily fish in the same locality or during the same days of the months, the data provide only an estimate of the general trends in hourly catches of albacore off Oregon. The catches of albacore versus hours of the day are shown in Figure 1. Chi-square tests of the 691 '10' '12' '14' '16' '18' '20' HOUR OF THE DAY (PDT) Figure 1. -Average catch of albacore per boat for each hour of the day from 0500 to 2200 for 3 mo in 1969 and for 20 July-2 August 1970. The numbers for hours of the day refer to the time at the beginning of the interval. heterogeneity of catches for all four of the data sets are significant (P<0.05), indicating that catch rates were not constant with time but varied throughout the day. Consistent trends for diel periodicity in catches, however, are not apparent. Only during 1970 was there an obvious trend for peak catch rates to occur early in the morning and in the evening. These peaks coincided with the local mean times of sunrise and sunset (about 0600 and 2100 PDT). The separate months of 1969 show a variable pattern with peaks occurring at different times during different months. There is no evidence for a pronounced early morning "bite," except perhaps 0800-0900 in July. Catch raies generally increased with time during the day. Most 1969 afternoon catch rates were above the median for each of the 3 mo, but they were not markedly higher in the evening as was found for 1970. The catches during 1970 were exceptionally high. Even the lowest 1970 catch rate was higher than any of the peak 1969 catches (Figure 1). The 1970 jig-boat season for albacore off Oregon was unusually short (most fish were caught between 22 and 29 July) and the fleet was localized in a small area (Pearcy 1973; Keene 1974). For those reasons, the spatiotemporal variability for the 1970 data is probably much less than for the separate months of 1969. 692 From our limited data we conclude that albacore feed throughout the day. Catch rates averaged over several weeks do not always indicate morning and evening periods of intense feeding. Obviously this does not preclude the occurrence of morning and/or evening bites in some areas on some days. Although stomach fullness and catch rates are both related to feeding behavior, they may provide different results on feeding periodicity because of the influence of such factors as changes in depth distributions, vulnerability to fishing methods, and the times required for accumulation and digestion of food in the fish's stomach. For example, the especially high rates of capture of albacore on surface jigs during sunrise-sunset periods of 1970 may be related to the vertical migration of Pacific saury, Cololabis saira, their major prey during this time (Pearcy 1973). Saury migrate to surface waters at sunset and descend into deeper waters at sunrise (Hotta and Odate 1956; Hughes and Gill 1970). Relatively low catch rates during the day may therefore be the result of either albacore pursuing saury into deeper water where albacore are less vulnerable to surface lures or to reduced feeding activity. Acknowledgments We thank the albacore fishermen who made this study possible. This research was part of the Oregon State University Sea Grant Program, supported by NOAA Office of Sea Grant, Depart- ment of Commerce under Grant No. 04-5-158-2. Literature Cited Dragovich, a. 1970. The food of sitipjack and yellowfin tunas in the Atlantic Ocean. Fish. Bull., U.S. 68:445-460. Hotta, H., and K. Odate. 1956. The food and feeding habits of the saury, Cololabis saira. [In Jap., Engl, abstr.] Bull. Tohoku Reg. Fish. Res. Lab. 7:60-69. Hughes, S. E., and C. D. Gill. 1970. Saury is promising 'new' fish on west coast. Natl. Fisherman 50(12): 12C-14C. Iverson, R.T. B. 1962. Food of albacore tuna, Thunnus germo (Lacepede), in the central and northeastern Pacific. U.S. Fish Wildl. Serv., Fish. Bull. 62:459-481. Keene, D. 1974. Tactics of Pacific northwest albacore fishermen-1968, 1969, 1970. Ph.D. Thesis, Oregon State Univ., Corvallis, 93 p. Magnuson, J. J. 1969. Digestion and food consumption by skipjack tuna (Katsuwonus pelamis). Trans. Am. Fish. Soc. 98:379-392. i Nakamura, E. L. Sci. Fish. 9(3):103-106. (Engl, transl. 1951, U.S. Fish 1965. Food and feeding habits of skipjack tuna {Kat- Wildl. Serv., Spec. Sci. Rep. Fish. 51:12-17.) suwonus pelamis) from the Marquesas and Tuamotu Islands. Trans. Am. Fish. Soc. 94:236-242. William G. Pearcy Pearcy, W. G. Daniel A. Panshin 1973. Albacore oceanography of f Oregon - 1970. Fish. Bull., Donald F. Keene U.S. 71:489-504. Uda, M. School of Oceanography 1940. The time and duration of angling and the catch of Oregon State University "Katsuo", Euthynnus vagans (Lesson). Bull. Jap. Soc. Cm-vallis, OR 97331 693 INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the following instructions. These are not absolute requirements, of course, but desiderata. CONTENT OF MANUSCRIPT The title page should give only the title of the paper, the author's name, his affiliation, and mailing address, including Zip code. The abstract should not exceed one double- spaced page. In the text, Fishery Bulletin style, for the most part, follows that of the Style Manual for Biological Journals. Fish names follow the style of the American Fisheries Society Special Pub- lication No. 6, A List of Common and Scientific Names of Fishes from the United States and Canada, Third Edition, 1970. The Merriam- Webster Third New International Dictionary is used as the authority for correct spelling and word division. Text footnotes should be typed separately from the text. Figures and tables, with their legends and headings, should be self-explanatory, not requir- ing reference to the text. Their placement should be indicated in the right-hand margin of the manuscript. Preferably figures should be reduced by pho- tography to 5% inches (for single-column fig- ures, allowing for 50% reduction in printing), or to 12 inches (for double-column figures). The maximum height, for either width, is 14 inches. Photographs should be printed on glossy paper. Do not send original drawings to the Scien- tific Editor; if they, rather than the photo- graphic reductions, are needed by the printer, the Scientific Publications Staff will request them. Each table should start on a separate page. Consistency in headings and format is desirable. Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be numbered sequentially in arable numerals. To avoid confusion with powers, they should be placed to the left of numerals. Acknowledgments, if included, are placed at the end of the text. Literature is cited in the text as: Lynn and Reid (1968) or (Lynn and Reid 1968). All papers referred to in the text should be listed alphabetically by the senior author's surname under the heading "Literature Cited." Only the author's surname and initials are required in the literature cited. The accuracy of the lit- erature cited is the responsibility of the author. Abbreviations of names of periodicals and serials should conform to Biological Abstracts List of Serials with Title Abbreviations. (Chemical Ab- stracts also uses this system, which was devel- oped by the American Standards Association.) Common abbreviations and symbols, such as mm, m, g, ml, mg, °C (for Celsius), %, "/oo and so forth, should be used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbreviations. We prefer that measurements be given in metric units; other equivalent units may be given in parentheses. FORM OF THE MANUSCRIPT The original of the manuscript should be typed, double-spaced, on white bond paper. Please triple space above headings. We would rather receive good duplicated copies of manuscripts than carbon copies. The sequence of the material should be: TITLE PAGE ABSTRACT TEXT LITERATURE CITED APPENDIX TEXT FOOTNOTES TABLES (Each table should be numbered with an arable numeral and heading provided) LIST OF FIGURES (entire figure legends) FIGURES (Each figure should be numbered with an arable numeral ; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. Bruce B. Collette, Scientific Editor Fishery Bulletin National Marine Fisheries Service Systematics Laboratory National Museum of Natural History Washington, DC 20560 Fifty separates will be supplied to an author free of charge and 100 supplied to his organiza- tion. No covers will be supplied. Contents — continued SILLIMAN, RALPH P. Effect of crowding on stock and catch in Tilapia mossambica 685 WENNER, CHARLES A. The occurrence of elvers of Synaphobranchns affinin on the continental slope off North Carolina 687 PEARCY, WILLIAM G., DANIEL A. PANSHIN, and DONALD F. KEENE. Catches of albacore at different times of the day 691 ^'^?6-l9l^ AMERICA'S FIRST INDUSTRY ixG^O 696-295 ^^^"^ °^^o. ■^'■Ans o< '' Fishery Bulletin National Oceanic and Atmospheric Administration • National Marine Fisheries Service r \ Vol. 73, No. 4 '] October 1 975 \\ PARRISH, JAMES D. Marine trophic interactions by dynamic simulation of fish species 695 VREELAND, ROBERT R., ROY J. WAHLE, and ARTHUR H. ARP. Homing behavior and contribution to Columbia River fisheries of marked coho salmon released at two locations 717 MAYNARD, SHERWOOD D., FLETCHER V. RIGGS, and JOHN F. WALTERS. Mesopelagic micronekton in Hawaiian waters: Faunal composition, standing stock, and diel vertical migration 726 KRYGIER, EARL E., and ROBERT A. WASMER. Description and biology of a new species of pelagic penaeid shrimp, Bentheogennema hurkenroadi, from the northeastern Pacific 737-r YOSHIDA, HOWARD 0. The American Samoa longline fishery, 1966-71 747 De VLAMING, VICTOR L. Effects of photoperiod-temperature regimes and pinealectomy on body fat reserves in the golden shiner, Notemigonus crysoleucas 766 WIEBE, PETER H., STEVEN BOYD, and JAMES L. COX. Relationships between zooplankton displacement volume, wet weight, dry weight, and carbon 777 DAWLEY, EARL M., and WESLEY J. EBEL. Effects of various concentrations of dissolved atmospheric gas on juvenile chinook salmon and steelhead trout 787 GOPALAKRISHNAN, K. Biology and taxonomy of the genus Nematoscelis (Crus- tacea, Euphausiacea) 797 --f BRAY, RICHARD N., and ALFRED W. EBELING. Food, activity, and habitat of three "picker-type" microcarnivorous fishes in the kelp forests off Santa Barbara, California 815 PALM, WILLIAM J. Fishery regulation via optimal control theory 830 ACKMAN, R. G., C. A. EATON, and B. A. LINKE. Differentiation of freshwater characteristics of fatty acids in marine specimens of the Atlantic sturgeon, Acipenser oxyrhynchus 838 BAILEY, JACK E., BRUCE L. WING, and CHESTER R. MATTSON. Zooplankton abundance and feeding habits of fry of pink salmon, Oncorhynchus gorbuscha, and chum salmon, Oncorhynchus keta, in Traitors Cove, Alaska, with speculations on the carrying capacity of the area 846 KROUSE, JAY S., and JAMES C. THOMAS. Effects of trap selectivity and some population parameters on size composition of the American lobster, Homarus americanus, catch along the Maine coast 862- (Continued on back cover) Seattle, Washington U.S. DEPARTMENTOFCOMMERCE Rogers C. B. Morton, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Administrator NATIONALMARINE FISHERIES SERVICE Robert W. Schoning, Director Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 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. Bruce B. Collette Scientific Editor, Fishery Bulletin National Marine Fisheries Service Systematics Laboratory National Museum of Natural History Washington, DC 20560 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. William H. Bayliff Inter-American Tropical Tuna Commission Dr. Roger F. Cressey, Jr. U.S. National Museum Mr. John E. Fitch California Department of Fish and Game Dr. William W. Fox, Jr. National Marine Fisheries Service Dr. Marvin D. Grosslein National Marine Fisheries Service Dr. Edward D. Houde University of Miami Dr. Merton C. Ingham National Marine Fisheries Service Dr. Reuben Lasker National Marine Fisheries Service Dr. Jay C. Quast National Marine Fisheries Service Dr. Paul J. Struhsaker National Marine Fisheries Service Dr. Austin Williams National Marine Fisheries Service Kiyoshi G. Fukano, Managing Editor The Secretary of Commerce has determined that the publication of this periodical is necessary In the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget through May 31, 1977. Fishery Bulletin CONTENTS Vol. 73, No. 4 October 1975 PARRISH, JAMES D. Marine trophic interactions by dynamic simulation of fish species 695 VREELAND, ROBERT R., ROY J. WAHLE, and ARTHUR H. ARP. Homing behavior and contribution to Columbia River fisheries of marked coho salmon released at two locations 717 MAYNARD, SHERWOOD D., FLETCHER V. RIGGS, and JOHN F. WALTERS. Mesopelagic micronekton in Hawaiian waters: Faunal composition, standing stock, and diel vertical migration 726 KRYGIER, EARL E., and ROBERT A. WASMER. Description and biology of a new species of pelagic penaeid shrimp, Bentheogennema burkenroadi, from the northeastern Pacific 737 YOSHIDA, HOWARD 0. The American Samoa longline fishery, 1966-71 747 De VLAMING, VICTOR L. Effects of photoperiod-temperature regimes and pinealectomy on body fat reserves in the golden shiner, Notemigonus crysoleucas 766 WIEBE, PETER H., STEVEN BOYD, and JAMES L. COX. Relationships between zooplankton displacement volume, wet weight, dry weight, and carbon 777 DAWLEY, EARL M., and WESLEY J. EBEL. Effects of various concentrations of dissolved atmospheric gas on juvenile chinook salmon and steelhead trout 787 GOPALAKRISHNAN, K. Biology and taxonomy of the genus Nematoscelis (Crus- tacea, Euphausiacea) 797 BRAY, RICHARD N., and ALFRED W. EBELING. Food, activity, and habitat of three "picker-type" microcarnivorous fishes in the kelp forests off Santa Barbara, California 815 PALM, WILLIAM J. Fishery regulation via optimal control theory 830 ACKMAN, R. G., C. A. EATON, and B. A. LINKE. Differentiation of freshwater characteristics of fatty acids in marine specimens of the Atlantic sturgeon, Acipenser oxyrhynchus 838 BAILEY, JACK E.', BRUCE L. WING, and CHESTER R. MATTSON. Zooplankton abundance and feeding habits of fry of pink salmon, Oncor'hynchus gorbuscha, and chum salmon, Oncorhynchus keta, in Traitors Cove, Alaska, with speculations on the carrying capacity of the area 846 KROUSE, JAY S., and JAMES C. THOMAS. Effects of trap selectivity and some population parameters on size composition of the American lobster, Homarus americanus, catch along the Maine coast 862 (Continued on next page) Seattle, Washington For sale by the Superintendent of Documents, U.S. Government Printing Office, Contents— continued SILLIMAN, RALPH P. Experimental exploitation of competing fish populations . . . 872 CHEN, LO-CHAI, and ROBERT L. MARTINICH. Pheromonal stimulation and me- tabolite inhibition of ovulation in the zebrafish, Brachydanio rerio 889 OLLA, BORI L., ALLEN J. BEJDA, and A. DALE MARTIN. Activity, movements, and feeding behavior of the cunner, Tautogolabrus adspenms, and comparison of food habits with young tautog, Tautoga onitis, off Long Island, New York 895 BOTSFORD, LOUIS W., and DANIEL E. WICKHAM. Correlation of upwelling index and Dungeness crab catch 901 HOPKINS, THOMAS L., and RONALD C. BAIRD. Net feeding in mesopelagic fishes 908 Notes RUCKER, ROBERT R. Gas-bubble disease: Mortalities of coho salmon, Oncorhynchus kisutch, in water with constant total gas pressure and different oxygen-nitrogen ratios 915 WEBER, DOUGLAS D., and HERBERT H. SHIPPEN. Age-length-weight and dis- tribution of Alaska plaice, rock sole, and yellowfin sole collected from the southeastern Bering Sea in 1961 919 SLATICK, EMIL, DONN L. PARK, and WESLEY J. EBEL. Further studies regard- ing effects of transportation on survival and homing of Snake River chinook salmon and steelhead trout 925 PATTEN, BENJAMIN J. Comparative vulnerability of fry of Pacific salmon and steelhead trout to predation by torrent sculpin in stream aquaria 931 CRAMER, S. P., and J. D. McINTYRE. Heritable resistance to gas bubble disease in fall chinook salmon, Oncorhynchus tshawytscha 934 INDEX, VOLUME 73 939 Vol. 73, No. 3 was published on 28 August 1975. The National Marine Fisheries Service (NMFS) does not approve, rec- ommend or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION OF FISH SPECIES James D. Parrish' ABSTRACT A mathematical model was developed for performing dynamic simulations of groups of interacting animal species. The energy balance of the individual animal was modeled so that growth and reproduction respond to food consumption after metabolic expenses are met. Populations change in response to recruitment (based on parental spawning) and mortality from natural causes, predation, starvation, and (where applicable) human exploitation. The forms of the various component mathema- tical functions were derived from the available ecological sources. Functions and parameters are especially applicable to marine fish species. Trophic webs of any size or form can be constructed using this basic species model. Computer solution of the essentially continuous differential model gives a time history of trophic and population variables for all species in the web. Models of trophic webs of 2, 3, and 4 levels were constructed and exercised. These were used to examine effects of age class structure, reproductive time lag, and population regulation by starvation mortality and fecundity control. Competition between species and the effects of a top predator on competitors, with and without human exploitation, were studied. Thus far in the history of trophic ecology there has been little effort to bring together the important results of the diverse studies which provide the components of the total trophic system into a con- struct that will permit analyzing the effects of metabolism, food consumption, reproductive ef- fort, and the structure of the trophic web upon the weight, population, and biomass of the various species involved. Perhaps the most complete and useful approaches in the literature are those of Menshutkin and Kislyakov (1967, 1968), Menshutkin (1968), Menshutkin and Prikhodko (1968, 1969, 1970), Karpov et al. (1969), Krogius et al. (1969), Menshutkin and Umnov (1970), Lassiter and Hayne (1971). The present work is an attempt to create a complete model for fish in the natural environment and to employ it for the stated type of total trophic analysis. The mathematical "trophic anatomy" of the generalized species modeled contains certain functions which represent trophic interactions with other species. The trophic web consists of an arbitrary number of such interacting species, coupled in this way into any arbitrary design; e.g., with any number of trophic "levels" (or coupled across levels), any number of species at each level, any number of predator species on a single prey species, etc. The trophic properties of the 'Present address: Massachusetts Cooperative Fishery Unit, U.S. Fish and Wildlife Service, 204 Holdsworth Hall, University of Massachusetts, Amherst, MA 01002. Manuscript accepted February 1975. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. generalized, modular species are established by specifying a set of equations which define its various ecological functions, such as respiratory metabolism, feeding, natural mortality, and reproduction. The composite nature of the model species' trophic anatomy permits considerable structural flexibility in model development. A particular ecological function, such as feeding rate as a function of prey abundance, may be expressed differently in different simulation runs by changing a single component equation. The separate identity of each species is determined primarily by the numerical values of the parameters in its component functions, but the form of functions may be different in different species where the data dictate. The model approach used allows a number of different levels of approximation. In the simula- tions performed here, no differentiation is made between sexes in the populations. The sexes could easily be represented separately at the cost of more computing time and a larger data base. A common and convenient simplification that is used in most of the present simulations is construction of an entire species population of identical in- dividuals. Thus, the individual must be given a set of characteristics and parameter values that are in some sense representative of the entire life his- tory after recruitment. A population with separate age classes has also been created explicitly with the present model. 695 FISHERY BULLETIN; VOL. 73, NO. 4 The model was employed in a number of simulations for a variety of trophic webs. There were not sufficient data at hand for all the species of a real trophic web to permit simulation of such a web in this way. The parameters and initial values used in these simulations are, therefore, reason- able illustrative values for fish in the natural en- vironment, based mostly on the literature. Despite the scarcity of the real data that would be required to use the model effectively for quan- titative prediction of real systems at present, models of this sort have considerable immediate value. Representing and interrelating animal functions analytically enforces a discipline in thinking which tends to clarify perceptions of the trophic relations. Formulation of a system in mathematical functions makes clear the nature of the data required, so that effort in gathering data can be applied efficiently. Component functions for a single species can be collected from a variety of sources and fused. The trophic behavior of the resulting model species can be studied, at least qualitatively, to see if the model behaves as the animal appears to behave. If the species model appears to represent the animal reasonably well, and if a trophic web is constructed from such animals, some confidence may be placed in its predictions of the behavior of the real system— a system which may be much more difficult to evaluate independently of the model. THE MODEL The basic model used for each species in every trophic web was developed from an energy balance of the individual and a formulation of the population dynamics of the species. Table 1 con- tains a glossary of symbolic notation used in the model. (A) The Energy Balance The energy balance was written by equating assimilated food intake, kC, to the sum of the three physiological uses of the assimilated food: res- piratory metabolism, Q, reproductive material produced for spawning, S, and growth, G. kC = Q + S + G (1) (1971), and elsewhere. Mann (1965) has made one of the very few attempts to include the S' term quantitatively in the balance. Most workers (e.g., Winberg 1956:209; Mann 1967, 1969) find that to a very acceptable ecological approximation, most fish under most circumstances assimilate a fairly constant fraction, A-aO.8, of the food, C, consumed (C, the feeding rate, is commonly called the ration and will be so designated herein). Ten (1967) should be consulted for a minority opinion on the effective constancy of k. Kostitzin (1939:180) and Beverton and Holt (1957:113) also deal with the form of possible variation. All the above terms are time rates. In the present simulations, the time unit used is the year. Since G is dW/dt, the instantaneous value of body weight, W, can be found by integration of G. Each term in Equation (1) can be expressed as energy or as the equivalent weight of body tissue, wet (live) or dry. In these simulations, all terms for all species are expressed in wet weight of tissue, based on a standard conversion factor of 1 kcal/g wet weight (Winberg 1956; Mann 1969). Recent results (Davis 1968; Kausch 1968; Brett et al. 1969) on changes in water content of fish tissues at various nutritional states suggest that a dry weight basis may be noticeably more accurate where data are available. The use of different conversion factors for different species or condi- tions, when known, introduces no conceptual problems. That Equation (1) can be balanced using experimental values of A-, C, Q, and G determined simultaneously in the laboratory for a group of fish over a range of sizes, ambient temperatures, and nutritional states, has been demonstrated by Kausch (1968), using the carp, Ci/piinus carpio. The results of many investigations indicate that respiratory metabolism can be expressed approximately as a function of body weight, W, by the relation Q = aW"^, (2) Similar expressions are found in Winberg (1956:210, 1962), Ivlev (1961a), Warren and Davis (1967), Mann (1967, 1969), Davis and Warren where Y is some fractional power. For most fish species a value of Y = 0.8 appears to be sufficiently reliable for ecological pui'poses (Winberg 1956:149, 1962; Mann 1965, 1969; Paloheimo and Dickie 1966). Where a more accurate value of 7 is known for a particular species, the model will accept it readily. For the purpose of the present simulations, a level of a for a constant (or long-term average) temperature of 10°C is used. Based on a large 696 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION Table 1. -Glossary of symbolic notation used in the model. [Notation defined in referenced literature, when different from that of the present model, is not repeated here.] Symbol Definition Symbol Definition 8 8 b C max Cp = Cp I t-'CICUi E e F G G 9„ 31.92 ', i k M m Numerical parameter in the Beverton and Holt reproductive function Species biomass Biomass of the /th species Biomass of the food base Numerical parameter in the Beverton and Holt reproductive function Actual ration; actual rate of food consumption by an individual r^/laximum ration w/hen feeding to satiation Ration of a predator Consumption of the /th prey species by a predator Numerical predation parameter in the linear feeding function Instantaneous coefficient of natural increase of an "exponential grovjth" type food base Rate of change of population due to fishing mortality Rate of change of population due to natural mortality Rate of change of population due to predatory mortality Rate of change of population due to starvation mortality Total egg production rate of a population Base of natural logarithms Coefficient of instantaneous fishing mortality Actual growth rate of an individual Maximum growth rate; growth rate when C = C^^^^ Growth rate predicted by the von Bertalanffy growth function Fraction of the species population that is sexually mature Numerical parameters in the linear sexual maturity function for g^ Constant of integration for critical starvation mortality time Convenience combination of variables in Equation (A-2), Appendix Recruit body weight coefficent Ration assimilation coefficient; fraction of the ration assimilated Numerical parameter (theoretical maximum body length) in the von Bertalanffy growth function for length Coefficient of instantij--30us natural mortality Numerical parameter in the starvation mortality function Fish "1. N w, N N'n, Wo Wp n P Q Q R R. r S i u W Fish w, w, a ffstarv V Numerical predation parameter in the Holling feeding function Species population Population of a fish species Population of the /th species or age class Sexually mature species population Initial population Predator population Standard equilibrium value of species population Numerical parameter in the starvation mortality function Prey abundance Rate of food base biomass input to the system in a "constant input" model Rate of respiratory metabolism of an individual Rate of respiratory metabolism of an individual at maximum ration Rate of reproductive recruitment Standard equilibrium recruitment rate Fraction of the maximum ration actually consumed Fecundity: actual rate of production of reproductive material by an individual Maximum fecundity at current body weight Computed numerical parameter in the starvation mortality function Time Critical time to 100% starvation mortality Fecundity coefficient of an individual Coefficient of predator preference for the /th prey Body weight of an individual Body weight of an individual fish Body weight of an individual of the /th species or age class Initial body weight of an individual Body weight of a prey individual Standard equilibrium body weight of an individual Numerical parameter (theoretical maximum body weight) in the von Bertalanffy growth function Respiratory metabolism coefficient Respiratory metabolism coefficient at maximum ration Respiratory metabolism coefficient at zero ration Numerical parameter (weight exponent) in the respiratory metabolism function Numerical parameter in the von Bertalanffy growth function Numerical predation parameter in the Ivlev feeding function collection of data from the experimental litera- ture, a level of a appropriate to an average spon- taneous activity level was chosen. The instan- taneous value of a is allowed to vary in response to ration according to the equation a a starv- + («r a starv c Cr (3) where Cis the actual instantaneous ration, a^j^j^. is the value corresponding to minimum metabolic rate at complete starvation, and a^ax corresponds to the maximum metabolic rate when feeding to satiation at ration C^ax- Equation (3) is a linear expression that generally approximates the best results from the few applicable long-term fish feeding and growth experiments (Davis and Warren 1965, 1971; Paloheimo and Dickie 1965, 1966; Beamish and Dickie 1967; Warren and Davis 1967; Brett et al. 1969). Use of Equation (3) in Equation (2) gives Q for any size fish at any feed- ing level. Fecundity of fish must be dependent on size and, at least in some limiting sense, on nutrition. A number of workers have noted reduced fecundity in overcrowded, undernourished fish populations and have speculated on how this reduced fecundity might tend to regulate the population (Woodhead 1960; Nikolskii 1961, 1962; Scott 1962; Bagenal 1967; Mackay and Mann 1969). There is good 697 FISHERY BULLETIN: VOL. 73, NO. 4 evidence (Simpson 1951; Bagenal 1957, 1967; Beverton 1962; Pitt 1964; LeCren 1965; Bagenal and Braum 1971) that in most fish with adequate food supply above metabolic demands, fecundity is strongly dependent upon body weight. Regres- sions on weight usually fit better than regressions on length or age (Bagenal 1957, 1967; Nikolskii 1962). It seems reasonable to represent the rate of production of reproductive material, S, (or ac- cumulation of body stores for that purpose) as a simple function of weight. Although more general functions have been proposed (Bagenal and Braum 1971), apparently most regressions so far fitted using data from specimens have been quite close to the linear expression. S = uW, (4) where u is a constant. In the present simulations, u = 0.1 in all cases, based on average values for several species and both sexes (Bagenal 1957, 1967; LeCren 1958, 1962; Mann 1965; Norden 1967; Phillips 1969). The linear function is truncated near its lower end at a weight corresponding to sexual maturity. This is consistent with the general observation that the onset of sexual ma- turity in fish appears to be a function of size rather than age (Beverton and Holt 1959; LeCren 1965). Exceptions for individual species are noted in Bagenal (1957). Trophic factors regulate the animal's fecundity through their effect on body weight. Also, when food intake becomes sufficiently low, there must not be enough energy above metabolic demands for normal fecundity. The scanty field data available suggest that usually fish sacrifice growth for reproduction, so that as food intake decreases, fecundity stays at or near normal (with decreased growth) until the net energy above metabolic ex- penditures is less than the normal fecundity requirements, after which fecundity decreases (Mackay and Mann 1969). The model operates in this way. The ration, C, under any instantaneous set of conditions, is obtained from the maximum ration, Cn,ax . and the current abundance of the prey which constitutes the food supply. C^^^ is dependent on body weight, and its current value can be deter- mined from Equation (1) if the maximum growth rate, Gn,^^ , is known. Since Gma.x is a function of the current size of the individual, this function is required. Data from appropriate ad libitum feed- ing experiments with a particular species of interest could be fitted to the appropriate function to give continuous values of G^^^ . The von Ber- talanffy growth function is a convenient one to which growth data from a large number of fish species have been fitted (e.g., Beverton and Holt 1959; Ursin 1967). In its differential form it expresses growth GyB-'^iWoo'^m^^'-W), (5) where k and Woo are numerical fitting parameters (k corresponds to the ^k of Ursin 1967, and to 3 times the K of Beverton and Holt 1959). Woo corresponds to a theoretical maximum weight, asymptotically approached. Values of k and \\^ for the present simulations are taken for certain illustrative species from Beverton and Holt (1959) and Ursin (1967:2421-2423). Equation (5) is employed in the model with a constant coefficient of 4.0 as an arbitrary standard adjustment to represent the highest feeding conditions. This gives a relationship between values over the full feeding range (e.g., zero, maintenance, and maximum ration) consistent with those observed in long-term feeding and growth experiments. With the C^ax term thus expressed, the Q^^^ term is simply Equation (2) with a = a^^^ and using the current weight, W. The S^^^ term comes from Equation (4). Thus '^max k max (6) There is a considerable and developing body of theory, for which evidence continues to ac- cumulate, that where environmental conditions are fairly stable, a predator's ration may be expressed as a fraction, r, of its maximum ration, r being a simple function of the abundance of prey, P. This approach is taken as a useful long-term ecological approximation, in which short-term behavioral factors, factors affecting the acces- sibility of the prey, etc. are smoothed out. Several expressions for this relationship have been proposed. Three alternative expressions for simple preda- tion with no explicit competitive effect between predator individuals were used in the model in different simulation runs. These are: Linear: Ivlev: Holling: r = l-e-^''(Ivlev 1961b) P (7A) (7B) m. (Holling 1959) (7C) 698 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION where f ,i. I, w„,^. are numerical parameters. For prey species modeled as described here, expressing P in terms of prey numbers rather than biomass seems to have system stability advantages. It is through Equation (7) that this species interacts with the next lower species in the trophic chain (web). The instantaneous rate of production of reproductive material and growth at the current body weight and prey abundance can be deter- mined by use of the above terms in Equation (1). Using ^ = ¥r from Equation (7) in Equation (3) ^max gives the current value of a to be used in Equation (2) to give the current value of Q. When food sup- ply is adequate; i.e., when kC - Q>S^^^ , "fecun- dity" is and, from Equation (1), positive growth is dW dt = G = kC-Q-S, (8) When food supply is so low that 0 mortality due to preda- tion, -Dpj^gp, and starvation mortality, ^g^ARv (^^^ sign of the reproductive term is positive; all the other terms have negative signs). Equation (11) is used in essentially this form for the representative individual model. For the age class model, the last three terms appear for all age classes. Instead of including the first term, the appropriate number of recruits is simply introduced as a pulse into the youngest age class at the appropriate times in the simulation. The recruitment rate, R, is expressed as a func- tion of the rate of egg production, E, by the Beverton and Holt (1957:49) reproductive function R = a + b^ E (12) where a and h are numerical parameters. A simple relationship such as Equation (12) is appropriate for the present model where the response of a system of essentially adult populations to purely trophic variables is of interest. The egg production rate, E, is the cumulative spawn of the entire ma- ture population, A^^^; i.e., E = N^S. (13) For the age class model, this involves summing over all mature age classes and over the entire year. All real species have some reproductive time lag or "generation time." In all except the simplest animals, this lag is significant and can have im- portant influence on the dynamics of the popula- tion. Such lags of any desired length are in- troduced in the simulations by properly coding the programs so that the E produced in 1 yr is stored and used in Equation (12) to compute the R for the appropriate later year. Except for fishing mortality, natural mortality is the only kind expressed in most fishery models. The position taken is that all mortality not due to fishing is "natural" and may be measurable in an unexploited stock or by eliminating the fishing mortality from statistics on an exploited stock by some analytical technique. Thus defined, natural mortality is almost invariably represented in 699 FISHERY BULLETIN: VOL. 73, NO. 4 C = rC max i (Coupling to lov/tr trophic level) a = a. (P) tc.r\ Q = S , max ' S = 5 max — - G - kC - Q - Sj^ax at If 0 < kC - Q < Sn^axr S - kC - Q dW — = G = 0 dt If kC - Q < 0, S = 0 dW dt = G = kC - Q V E = NS = NuVJ R = 1 a + b E dt Dnat = -MN I °PRED "_EPliP W (coupling to higher trophic level) FISH = -FN ^STARV °n ^PRED ^ ^STARV + D FISH Figure 1. -Relationships between principal component equations describing a single species. fishery works by the simple decaying exponential function /)nat = - MN, (14) where M is a numerical parameter, the "coefficient of instantaneous natural mortality." The function has been applied very widely, whatever the predation rate may be (even where it is zero) and in situations where starvation probably does not occur. Lacking data to support another assump- tion, it seems reasonable to use Equation (14) tO' express the more limited category of natural mor- tality of the present model also. In the model, then, natural mortality is all mortality not due to predation or starvation (the trophic controls) or 700 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION fishing, and would include death due to disease, senility, accident, environmental stress, etc. Field data for this limited class of mortality are rather scarce. For the hypothetical species in these simulations, approximate conventional M values were taken from Beverton and Holt's (1959) ta- bles, and for species under predation, these were at times modified. In the age class models, different values of M are used for different age classes. Beverton and Holt (1959) discussed the variation of mortality with age more fully. The expression for mortality due to predation, -^PRED' comes directly from the ration of the predator species, Cp , modeled as described in the preceding discussion of energy balance. Thus, the rate of change of prey population due to predation, -^PRED' IS D PRED ^ _^pCpNp w (15) prey where Np is the number of predators, each with ration, Cp, W^^^yis the weight of a prey individual, and the summation is over all predator species which consume the particular prey. Equations (15) and (7) provide the coupling between each model species and the other species with which it interacts in the trophic web. Despite the scarcity of knowledge on starvation in fish, it would seem that a complete model for a system controlled by trophic variables should include some reasonable attempt at a formulation of this source of mortality. An expression was developed that can approximate the general form of the survival versus time curves from Ivlev's (1961b:266) starvation experiments with fish. This expression states that under pressure of starva- tion alone, the surviving number, A'^, of an initial population, Nq, after time, t, will be 1 A^ = NQ—{m + n - me-''). (16) The m and n are numerical parameters, and the parameter, .s, comes from the boundary condition at 100% mortality, after the critical time, t^, , to extinction has been found from the integrated form of the energy balance equation under star- vation conditions (see Appendix). The form of the function of Equation (16) is plotted in Figure 2. Use of Equation (1) in computing t,. and s provides an appropriate curve for any ration. The model uses the differential form of Equation (16), _l N=N(,"n mtn - me^' < L -l t^ioo z2 > --^^^ Li_> ^^^^ O^ ^\^ 3 ^\. 00 ^v ss x^ Cd \^ -LU \. ^ °° 50 \ OS ^" \ gi \ _i \ Z) \ Q. \ o \ CL \ 0 1 1 1 \ Q2 0.4 TIME STARVING, YRS. 0.6 Figure 2. -Survival curve of a population undergoing only star- vation mortality at zero ration. m ^STARV = -^oTT^e'', (17) as the fourth term in Equation (11). When exploitation by man is included in the system, the only modification is the addition of another term to Equation (11). In accordance with conventional fishery theory and the concept of chance encounter between fish and fishing gear, this term is exactly like Equation (14); i.e., fishing mortality, />fish> is D FISH = -FA^. (18) The numerical parameter, F, is an expression of the intensity of fishing effort and the vulnerability of the prey to the fishing gear. Equations (12), (14), (15), (17)-and (18) where appropriate-provide all the terms for determin- ing rate of change of population from Equation (11). (Figure 1 summarizes their relationships.) Numerical integration of Equation (ll)-less the first term for the age class model-gives "con- tinuous" values of population of the species over the entire time span of the simulation. Biomass of an entire species population at any instant is the product of instantaneous values of W and N (summation of a group of such products in age class models). Production over any desired period is obtained by integrating the incremental growth rates, G, and reproductive products, S (if desired), over that period. 701 FISHERY BULLETIN: VOL. 73, NO. 4 (C) The Trophic Web All the fish species in a trophic web can be modeled more or less as described above. Many invertebrates that serve as fish food can be modeled in much the same way, with some appropriate changes in individual component functions and by use of the proper parameter values (see Winberg 1962, and Mann 1969 for dis- cussion relative to invertebrates). Since this model uses feeding functions based on prey abundance, an operational limitation is imposed that the ul- timate resource base-the lowest item in the food chain— cannot be modeled fully in this way. In terms of total ecosystems, this is natural enough. Although the ultimate autotroph might be thought to "prey" upon inorganic nutrients, and models for plant growth as a function of nutrient abundance exist, the present model is obviously not appropriate for autotrophs. Thus, any trophic web modeled in this way must have at its base an arbitrarily defined species or group of species. The purpose of this first exercise with the model is to explore trophic interactions among fishes. Therefore, the cause of clarity seems best served by modeling all the species of interest as fishes. The level(s) below the lowest fish species-the food base for the fish community-is then given only the simplest representation. Two types of food base have been used in these simulations: 1) the "constant input", and 2) the "exponential growth." Properties of the constant input base are that biomass, B^ , enters the system at a constant rate, P^ , and is reduced only through predation by the higher level, fully modeled species. Thus, the rate of change of food base biomass is at (19) where the summation is over all predators with their individual rations, Cp , and populations, Np . Ecologically, this system might correspond to a fish community whose base prey enters the com- munity feeding area at a constant rate; e.g., as brought in by water circulation or by migration as prey individuals continuously reach a particular life stage. Because of its extreme simplicity, this type food base model is preferred for studying the trophic relationships of fishes higher in the web. Properties of the exponential growth food base are that biomass is produced at a rate directly proportional to the current standing crop of biomass and is reduced only through predation by the higher level species. Thus (IB I = c,B, ^CpNp, (20) where (\ is a numerical parameter corresponding to the "instantaneous coefl^cient of natural increase" of classical population growth theory. Again the summation is over all predator species preying on the food base. Without predation, B^ would of course increase exponentially and indefinitely. This makes stability of such a system precarious, a fact borne out by experimentation with the model. For these simulations, numerical values of P^ and c'l were selected arbitrarily to be compatible with the standard equilibrium state of the trophic webs constructed. SIMULATION TECHNIQUE Combination of the previously described func- tions produces the basic species model. A single such model species, with one of the food base models described above as prey, was exercised over a range of conditions and with some variety in certain component functions, in an effort to become familiar with some of the dynamic properties of the basic species model. Groups of such model species were then interconnected in various ways to explore the behavior of various trophic webs. Interactions between species occur through Equations (7) and (15). Where a predator feeds on more than one prey species, P in Equation (7) for that predator is the total abundance of all the )t species. For each of the u prey species, the predation mortality imposed by that predator is given by Equation (15) in which the Coupon the fth prey is ^Pj ~ ^p TOTAL „ 2v,A^, (21) where A^, is the current population of the iih species, and /', is a coeflRcient expressing predator preference and availability (vulnerability) of the prey. The two (or more) elements contained in v can be separately expressed by making r a product of separate coeflficients. By Equation (21), the predator tends to adjust the makeup of its diet 702 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION proportionately to the abundance of the various prey, but bias is allowed for known preferences or differences in the ease with which various prey can be taken. These features, together with the basic structure of the age class model, allow that predators and prey interacting with any species may be different for different age classes of the species or may change in their degree of impor- tance. All the model species created for these simula- tions are hypothetical. To avoid resorting to pure fantasy and to get some consistency among cer- tain species properties, each model fish species was based on a real fish species (see Table 2). Real species were selected which are sympatric, and in fact, each of the predator/prey relationships modeled has been reported in the literature in a nonquantitative way. A major simplification in Table 2.-Values of parameters and of basic variables at standard equilibrium state as used in simulations. [Values shown in parentheses are alternate values used in some simulations.] Namesake species Clupea Clupea Clupea Scomber Sard a Parameter sprattus sprattus sprattus scombrus sarda or standard equilibrium variable Model species Units Source E A B C D O'st.rv g0.2.yr-l.o See MODEL 1.0 1.0 0.7 1.0 1.0 "ma« g0.2.y|--|.O section (A) 7.0 7.0 4.9 7.0 7.0 U yr-^ See MODEL section (A) 0.1 0.1 0.1 0.1 0.1 y See MODEL section (A) 0.8 0.8 0.8 0.8 0.8 k See MODEL section (A) 0.8 0.8 0.8 0.8 0.8 K yr-i Ursin (1967): 2421-2423 1.75 1.75 1.75 1.2 0.539 Woo g Ursin (1967): 2421-2423 30.6 30.6 30.6 516.0 9,400.0 M yr-i Beverton and 1.0 1.076 1.076 0.9 0.9 Holt (1959) [0.1169] [0.1169] [0.0325] w/ g 10.0 15.16106 15.16106 150.0 600.0 N' 1.0 X 105 0.61032 X 0.61032 X 5,000 500 S 105 (1.22064 X 105) 105 (1.22064 X 105) R * 1.0 X 105 0.65680 X 0.65680 X 4,500 450 s 105 (1.31360 X 105) 105 (1.31360 X 105) a Computed 0.66667 1.09232 X 1.09232 X 0.13333 0.13333 based on R^ X 10-5 10-5 (0.54616 X 10-5) 10-5 (0.54616 X 10-5 X 10-3 xlO-2 b g-yr-1 Computed based on R^ 0.33333 0.39808 0.39808 6.66667 26.66667 c See 0.49584 X 0.60178 X 0.60178 X 0.37902 0.79657 X 10-' 10-' 10-' X 10-5 10-1 Ccl SIMULATION 0.45217 X 0.45217 X 3.03435 0.65706 X SI 10-' 10-' X 10-6 10-1 m TECHNIQUE 1.21158 X 1.21158 X 0.20750 0.10219 X mc section. 10' 10' X 10<> 105 m Selected to fit form of 1.0 1.0 1.0 1.0 1.0 n Ivlev (1961b: 266) 40.0 40.0 40.0 40.0 40.0 V See SIMULA- TION TECH- NIQUE sec- tion 1.0 1.0 9| See RESULTS section (A) -1.02147 -1.02147 -1.00000 -0.71428 92 g-' See RESULTS section (A) 0.13333 0.13333 0.013333 0.00286 F yr-i See RESULTS 0.05 0.05 0.05 section (D) to to to 0.90 0.90 0.20 •Values of variables at standard equilibrium state. 703 FISHERY BULLETIN: VOL, 73, NO. 4 the simulations is the extremely limited range of diet of the model species; their factual namesakes have rather catholic tastes. For these hypothetical species, a set of values for an arbitrary standard equilibrium condition was established as follows. Populations for all species were arbitrarily set at values that seemed reasonable relative to each other and in respect to the various body weights and reproductive rates. Using the various parameters selected, for each species, the r value corresponding to the standard equilibrium state was then computed. From this r and the equilibrium population of the prey, the predation parameters of Equation (7) were com- puted. Using this procedure for each fish species, working up the trophic chain, a complete set of equilibrium values for all species became avail- able. A compatible trophic web was thus created arbitrarily, having at least static stability; i.e., (IN /(It and dW/dt were zero for all species. Table 2 provides the values of parameters and of basic variables at standard equilibrium state for the model species used in these simulations. Where a consistent set of laboratory and field data on species in a real trophic web were available to be used in the model for predictive purposes, some of these procedures would be unnecessary. Like all simulations, those run with the model require that initial conditions be specified. Typically in these runs, the initial conditions were those of the standard equilibrium state with the exception of some single variable value which was displaced so as to perturb the system. For example, a simulation run started with all variables at equilibrium except B^ might be analogous to the natural occurrence of sudden catastrophic mor- tality in a prey species. Initial conditions are dis- cussed further under RESULTS. In each case, the simulation was allowed to run for an arbitrary length of time, or until automatically terminated when some variable reached a prescribed limiting value. Usually runs were continued until a stable state (the original standard equilibrium or other- wise) was approached, or until a distinct mono- tonic trend with a predictable outcome was de- tected. All the simulations were programmed using the IBM^ System/360 CSMP (Continuous System Modeling Program) (International Business ''Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Machines Corporation 1969, 1971) and run on an IBM 360/50 Data Processing System. RESULTS A limitation of the approach taken here, as with any simulation model for numerical solution, is that mathematically exact and general solutions are not obtainable. A full solution of the system represents a very complex multidimensional re- sponse surface. In the very simplest case of one modeled fish species and a food base species, there are three basic dependent variables whose in- tegrated values appear in the solution; viz., N^-^^^ , H^Fish , and B^.lm. "representative individual" model with n fish species and the food base, there are 2n + I basic dependent variables, and in a similar model with x explicit age classes per species, there are 2nx + 1 basic dependent varia- bles. A system with the complexity and nonlineari- ties of this type of model is capable of behaving quite differently in different regions of the state space. Since it is impossible to explore the entire response surface thoroughly, measures must be taken to limit simulation effort to regions of interest. Eventually a detailed and systematic exploration of regions of known interest using es- tablished optimization techniques (e.g., Box et al. 1953; Box 1954; Box and Hunter 1957) may be useful with the model. For the present, the scope of simulation effort has been limited by selecting parameter values that seem reasonable and compatible for each of a small group of rather common fish species and by building out from a system already investigated to a larger system of which the original is a subset. In a number of cases where moderate changes to values of parameters or even to the form of com- ponent functions have been made, system dynamics have been somewhat altered or the sys- tem has even moved toward a new stable state. Usually, however, in a system with any regula- tory capacity (stability) at all, the change has not been drastic. Rather large perturbations in initial values of the basic dependent trophic variables of such a system have not usually displaced the sys- tem to a distant stable region or resulted in breaking the trophic web (eliminating one or more species). This behavior of most of the systems simulated gives evidence that there is at least one region of some useful size in the total state space-i.e., the region in which the arbitrary 704 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION equilibrium state has been placed— in which the system models are fairly stable. For these reasons, it is believed that the basic form of the behavior of systems demonstrated in these simulations has some generality beyond the specific cases tested. However, in all cases, the results shown here are simply examples of interest from an infinite number of possibilities. The simulation technique used here is amenable to use in sensitivity analyses; i.e., for determina- tion by a systematic program of successive trials how sensitive the result is to the numerical value of a parameter or an initial variable value. Such analyses not only help define useful regions for particular results; they also give an indication of how accurately particular parameters must be measured in the field or laboratory, so that effort is applied where it is important to the system result. In describing the following results, the shorthand notation used to reference the trophic webs has the following form: 1 ,1st 2 2nd 1 3rd 0 4th 0 5th Trophic level The digit in each column indicates the number of species at that trophic level. Where two species appear at a common level, they compete for prey at the next lower trophic level and are preyed upon equally by the next higher trophic level. The lowest level is always occupied by the food base with biomass By. (A) Regulation of Body Weight and Population The basic species model seems to have a con- siderable capacity for self-regulation; i.e., it can return to an equilibrium state after sizable displacements of some of the variables in the sys- tem. The return usually involves a series of os- cillations above and below the equilibrium values, with the degree of damping depending on the exact structure and parameter values. One of the most common and interesting per- turbations involves displacement of prey abun- dance. Figure 3 illustrates a P 11000 trophic web with a representative individual model of a species A fish preying on an exponential growth type food base with an Ivlev feeding function. The system 5^ z 3 < cc 99 Z -J _ 3 2 O O LiJ ^Q _J OC =) < a. Q o b ^ CC a. ° 30 40 TIME, YEARS Figure 3. -Response of a single species to an initial perturbation in the abundance of its prey. N = predator population; W = predator body weight; B j = prey (food base) biomass. responded to an initial condition in which the fish population was at the standard equilibrium value and the prey abundance was initially about 71.4% of the standard equilibrium value. The system re- turned to the standard equilibrium state with damped oscillations. The purely population con- trols, natural mortality and reproduction (/>nat vs. R), were satisfied initially, but the system was unbalanced trophically because of the scarcity of prey. Regulation resulted from the response of body weight and resulting fecundity to food con- sumption, balanced by natural mortality respond- ing to the changing population level. Similar stable responses were demonstrated with the model for cases of initial perturbation due to high prey abundance, B^, high predator popula- tion, A^, and high predator body weight, W. At suflficiently low values of prey abundance and production, substantial starvation mortality can occur. This is particularly true when the prey abundance decreases suddenly, since the normal population response of the predator through reduced fecundity is delayed by the reproductive time lag. Figure 4 shows such mortality for a population for four age classes modeled explicitly. The abundance of each year class decreased with time until those which had become 4-yr-olds were decimated at about 1.6 yr after the start. The. 705 FISHERY BULLETIN: VOL. 73. NO. 4 Figure 4. -Effects of starvation mortality on the 4 age classes of a species population and on the total population. A new class of recruits entered one year after the start of the simulation. Arrows indicate when starvation mortality began for each year class. ( actual population of an age class; population of an age class in the absence of starvation mortality; N/Ng, where N = total species population, Ng = standard equilibrium value of total species population.) dashed lines indicate the course of natural mor- tality. As interesting as the fate of individual year classes is the substantial effect on the total population (shown by the upper broken line in Figure 4). Theoretically, starvation mortality should be capable of regulating the population. However, based on a considerable range of simulation runs, it appears that with the usual sets of reasonable parameter values for the species considered here, body weight and fecundity normally respond to produce regulation so that a starvation condition is not reached. With longlived, slowly growing species, starvation would tend to become a more important factor. In some of the cases simulated, where trophic conditions were sufficiently extreme to produce heavy starvation mortality, total ex- tinction occurred. Figure 4 represents such a case, in which the food base biomass, B^, was initially 20% of the standard equilibrium value and c^ was 10%. Extinction occurred during the initial 2-yr reproductive time lag before fecundity changes could be reflected in recruitment. Some simula- tions were run in which fecundity was made unresponsive to actual body weight so that the effects of starvation could be better observed. With the nutritional control on fecundity thus removed, starvation occurred for some systems and started with B^ at 50% of standard equilibrium and c■^ at 90%. With normal nutritional control on fecundity, these systems had survived. Experience with the model indicates that for the types of species used, where extreme trophic con- ditions exist, disruption of the system is more likely to result from excessive stunting of growth and resultant failure of spawning than from star- vation mortality. In age class models, the stunting can be observed directly in the failure of in- dividuals of a year class to grow normally while in the recruited population. Where food supply is ex- tremely low, actual weight loss by an individual can also be observed. In representative individual models, these separate effects are combined in the single continuous variable, W. A smaller-than- standard W represents a population that, on the whole, is undersize. If the population as a whole becomes sufficiently stunted, at some point, egg production will be reduced to zero. This corres- ponds to a population unable to reach sexual ma- turity. If the entire population fails to spawn for enough successive years, extinction of the local species population must result. It is difficult to see how a species could persist if it failed to spawn for a continuous period as long as its lifespan. In most of the results shown here, no effort has been made to impose a sexual maturity limit on fecundity; i.e.. Equations (4) and (13) with N^^^ = N determine the egg production, E, at any body weight. E = N,„uW Nu W. (22) For exploring the limits of stability of systems against perturbations, it seems useful to represent the attainment of sexual maturity in the model. A "knife-edge" representation— one in which fecun- dity has a substantial positive value or functional form above some age, length or weight, and zero below it— has been used for simplicity in much fishery work. This may be a justifiable approximation of nature for some species, especially those with short lives, fast growth, and infrequent spawning. However, a smoother and more realistic representation seems desirable. For some species, data exist on the fraction of all in- 706 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION dividuals mature as a function of age, length, or weight. In some cases (e.g., Bagenal 1957) the weight function appears reasonably linear. This means that in Equations (13) and (22) the mature spawning population, A^,„, can be expressed as and ^,n = 9,,/^, 9m ^ 9l+ 9-2^^ (23) (24) where g^^ and g^g ^re numerical parameters, for all values of IF between that which gives g^^ = 0 and that which gives ,9„, = 1.0. At lower and higher values of W,g,,, is 0 and 1.0 respectively. For species A and C, information from Bigelow and Schroeder (1953) permitted a rough fitting of this function. The length corresponding to the body weight at which all were mature agreed reasonably well with the ratio: length at maturi- ty/theoretical maximum length (Lqq) of Beverton and Holt (1959) for both species. The same sort of weight limits for the function were assumed for species D, in the absence of better data. The stan- dard equilibrium weight and the 100% sexual ma- turity weight were made to coincide in each species. This means that in this particular modified model, any reduction below standard equilibrium weight decreases the number of mature spawners. The effect of the linear sexual maturity of Equa- tion (24) on the total population egg production is shown in Figure 5. Figure 6 illustrates the response of a single PllOOO trophic web with sexual maturity of species A modeled in this way. For the first 4 yr of the simulation run, input production at the food base level was about 20% of the standard equilibrium value; subsequently it was always at the standard equilibrium value. The reduced food supply resulted in stunting the population so much that from year 6 through 9 there was no recruit- ment. The subsequent reduced total food con- sumption by the greatly reduced population tend- ed to bring the system into balance. If it survived, it would eventually return to standard equilibrium conditions. However, for this species with a life- span of 6 yr or less, an interruption of recruitment for 4 yr is very dangerous. This, combined with a minimum population of about 1.4% of the standard at one point in the simulation, suggests that the condition reached here was very near a critical one for survival of the local species population. This E en CO > < Q > 5 10 15 20 INDIVIDUAL BODY WEIGHT, gm. Figure 5.— Relationship between inaiviaual fecundity, S, and body weight and between total population egg production, E, and body weight. Sexual maturity is a linear function of body weight. (3a: LU 30 Q. era: o . <2 _ LU 170 - 150 - 100 30 40 50 TIME, YEARS 70 Figure 6.-Response of a single species, with sexual maturity a linear function of weight, to an initial 4-yr perturbation of low production by its prey. (Pj = 20% of standard equilibrium value for the first four years.) Reproduction ceased between years 6 and9. N = predator population; W = predator body weight; R = predator recruitment. 707 FISHERY BULLETIN: VOL. 73, NO. 4 approach seems to offer a means of predicting the Hmits of stability of trophic webs against pertur- bation. (B) Reproductive Time Lag A limited study of the effects of the length of reproductive time lag was made using a represen- tative individual model of the simplest trophic web, PI 1000 (a food base and the fish predator, species E). The food base was of the exponential growth type and the predator employed an Ivlev feeding function. Reproductive lags of 0, 2.50, and 6.25 yr were tried with a model that was otherwise basically the same. These three alternatives correspond respectively to the assumptions: 1) that offspring are mature when spawned, 2) that they take 2.50 yr to reach the "representative" stage, 3) that they take 6.25 yr to reach this stage. The second assumption is reasonable for species E. The system was initially perturbed by starting with species E at 20% above its standard equilibrium population. The results for the biomass of species E are shown in Figure 7. It is clear that with increasing reproductive lag, the UJ 3 $ CD o UJ t- cr o Q UJ a: X in O (/) CO < o CD 150- 100 Figure 7.-Effects of reproduc-tive time lag on the response of a single species to an initial perturbation in its population. The three cases illustrated have reproductive time lags of 0.00, 2.50, and 6.25 yr, respectively. regulation of the system about its standard equilibrium becomes weaker; i.e., the biomass of species E, and other variables (not shown), reach more extreme oscillatory amplitudes. Larger amplitudes always incur greater risk of disaster. For example, in the runs shown here, a different sexual maturity criterion was used-knife-edge maturity at 80% of the standard equilibrium body weight. In the 6.25-yr lag run, this weight was reached at about 59 yr into the simulation, and after the 6.25-yr lag, it so reduced recruitment and the species E population that the system became unstable. This instability, which did not occur in the other runs, was due to the long lag in recruit- ment response to change in fecundity with changing food availability. Except where otherwise stated, all the results presented here are for representative individual models having 2.50-yr lag and age class models having 2.00-yr lag. These are reasonable for the species involved. They are mutually consistent because in the age class model, reproductive products are summed over a full year and produce recruits 2.00 yr after the end of the year. Thus the average lag is about 2V2 yr for the age class model also. (C) Age Class Effects A number of simulations were run with an explicit 4-age class model. Some results involving starvation have been shown above. Other exercises investigated the capabilities of this more accurate type of population model to regulate in the normal manner. Figure 8 illustrates the response of a simple food base-fish predator PllOOO system with Ivlev feeding function to an initial perturbation of the fish predator population. The "mean total population" is a variable obtained by summing the 4 populations of all the age classes, ^^ N^, during each computational increment of the year and taking the arithmetic average of these values. "Population mean annual biomass" is obtained by similarly summing and averaging the biomass values, ^ ^i^i ■ These variables are shown in i = 1 Figure 8 as percents of their standard equilibrium 708 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION LiJ ZD _l " UJ < 3 > - 2 > 3 140 17 2 cr Z) CD 130 \ a: _l ' m ZD o 120 I N _j LU ;l 3 Q 1 10 ^,- s , ~. _--" cc 0 0 LiJ (Y 70 60 I 1 , I.I.I, 1 . 1 1 i ir Q- a. 10 20 30 40 50 60 70 TIME, YEARS FiGURE 8. -Effects of age class structure and condition of recruits on the response of a single species to an initial over- population. (Curve A - 4-age class model with recruitment at standard recruit weight; curve B - 4-age class model with Kj recruitment; curve C - representative individual model.) The three upper curves represent total species population, N; the three lower curves represent total species biomass, B. values. In curve A, the system was represented by an explicit 4-age class model in which recruitment occurred at "standard recruit weight" (standard equilibrium weight for recruitment age). In cur\'e B the model was identical except that recruitment occurred at K^ times "standard recruit weight." The coefficient, K^ , is the ratio of the weight of the class ending its first recruited year to its standard equilibrium weight. Its use makes recruitment weight more consistent with current conditions. In curve C the system was represented by the corresponding representative individual model. Because of the way weight at recruitment is expressed, curves B and C are most directly com- parable. It is clear that there are some real differences in dynamics among all three models. These simula- tions and others indicate that for some detailed studies of the dynamics of trophic systems, age class models can provide additional information not available through representative individual models. However, much of the information of basic interest is contained in the representative in- dividual solution. The final stable state is predicted accurately. Because of its lower damping, this model gives a conservative (maximum) estimate of the time required for the system to return to within any given range of this state, and this maximum is close enough to the actual time to be useful. For most variables and most perturbations, the maximum amplitudes of the age class model tend to be less than those of the representative individual model, so that the latter tends to predict an envelope of reasonable size within which the actual values will lie. These characteristics make the representative individual model especially useful for predicting stability. The results shown and others suggest that the representative individual model can be used to approximate the behavior of the much more difficult and expensive age class model sufficiently well to justify use of the simpler model for many purposes. However, the quality of the approxima- tion depends upon the characteristics of the par- ticular system to be simulated. Where species are included that display a large range of sizes and ecological differences among the age classes of the recruited population, the representative in- dividual approximation is likely to be less accept- able. (D) Competition and Predation in More Complex Webs A major purpose of the model developed here is to serve as a tool for study of more complex trophic systems. A few examples of particular interest are presented below. In Figure 9 the trophic chain is extended by one link in the simplest possible manner to make a PlllOO web. Species C preys on species A which preys on a constant input food base. Both interac- tions employ Holling feeding functions. The populations of all three trophic levels oscillated as the system returned from the initial perturbation of low food base biomass. Within about 3V2 to 4 cycles (»35 to 40 yr), all variables were within about 1% of standard equilibrium values again. The phase sequence of population and biomass rapidly became level 1, level 2, level 3 as would be expected. The population phase displacement was complicated by the 2V2-yr reproductive lags and the effect of predation on the species A population (as species A lost weight, species C ate more species A individuals to meet its energy demands). Such an extremely simple trophic web would be 709 FISHERY BULLETIN; VOL. 73, NO. 4 TIME, YEARS Figure 9.-Response of a simple 3-level trophic web (PUIOO) to an initial low prey abundance. ( food base; 2nd level fish species; — 3rd level fish species.) rare indeed in nature, but it is interesting for at least two reasons. The various species might be thought of as representing in some sense v^hole trophic levels of more extensive natural systems, each level consisting of a rather homogeneous group of species. If all the species at a single level have identical trophic parameters, this simple web in fact represents them exactly. This follows logically and has also been verified in simulation. Thus, the PlllOO web model provides a base line for comparison for later unequal competition runs in a 3-level system. It represents the behavior of any n equally competing species at a trophic level, each with a population of \/nth the total popula- tion. The PllOOO web model provides the same kind of base line for competition in 2-level sys- tems. Figure 10 illustrates the simplest web, PUllO, with four trophic levels: the food base and three fully modeled fish species. It consists of the PlllOO web with species D added as a top predator. Again, an initial perturbation of low food base biomass caused oscillation of trophic variables of all species. After about 8 to 9 cycles (»70 yr), all were within a very few percent of standard equilibrium values again. The same population and biomass phase sequence appears. Predation by species D caused quicker response of the species C popula- tion, reducing the phase displacement between species A and species C. The large phase displacement now occurred between species C and species D. The basic period of oscillation was also shortened from about 10 yr to about 8 yr. In all the above cases, maximum oscillation amplitudes of variables were less than the initial perturbation, and they rapidly became substantially less. Effects produced on the system by feeding competition between species were of particular interest in these studies. No attempt was made to formulate explicit (interference) competition. Implicit competition was studied by constructing models with two predator species utilizing a com- mon prey species. The PllOOO and PlllOO models above represent exactly equal competition at the second level in 2- and 3-level webs respectively. The abstraction of exactly equal competition is not likely nor very interesting ecologically. A simple type of unequal competition is modeled by replacing one of the two species A-type competi- tors with species B, which is identical except that it has the advantage that its a^^^.^. anda^iax are 70% of the species A values. Thus it has lower metabolic requirements and grows more for a given food intake. Figure 11 illustrates a simulation of this 10 20 30 40 TIME, YEARS 50 70 Figure lO.-Responseof a simple 4-level trophic web (PI 11 10) to an initial low prey abundance. ( food base; 2nd level fish species; 3rd level fish species; 4th level fish species.) 710 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION trt CO (T cr O o 5 - ^ S Q Q Nb UJc/1 CC UJ U- Lj Q-D to —i 00 — .<^ 1— 1— ___ LiJ ^ ii o tr O CD - si u. O Oo O LU LJ O m Q 5Q Z K ^(T < .< - Q -? 2 Wb s"§ z < < h- ~ i_-t- Lfi 1^ ^CO z W o OsS ' — ^ < 3 00 J^ -- 100 § Q Q- O a. 1 1 1 1 1 1 o CO I 0 20 30 TIME, 40 YEARS 50 60 70 Figure 11.— Course of competition between two fish species with unequal metabolic demands, competing for a common food base. Species B is favored, since ttg = 0.7 cCf^- N ^ = population of species A; N p = population of species B; W^ = body weight of species A; Wg = body weight of species B. 2-level P12000 web started without initial pertur- bation, with identical body weights and popula- tions of the two species. This situation is somewhat analogous to the simultaneous entry of the two competitors into an environment where the prey biomass and input are fairly close to the standard equilibrium values. The system moved away from the even start with oscillations which were firmly damped toward an apparent new steady state. In this state, the population and weight of the more capable competitor were increased relative to those of the less able con- tender. Their final relative positions might be characterized by the biomass ratio B^/B^ = 1.30. It is interesting to compare this prediction with that derived from the simpler graph theory analysis (Saila and Parrish 1972). This was ac- complished by using the variable values from the present model for Q, B, and M to calculate the parameters q, h, a, b, and m for the graph theory model. These parameter values were then used in Equation (18) of Saila and Parrish (1972) to com- pute the biomass ratio B2/B2, = 1.56 of the compe- titors. This ratio is directly comparable with the ratio B^/B^ = 1.30 from Figure 11. In view of the considerable differences in the two approaches, the agreement seems too good to be entirely for- tuitous. The above simulation represents simple unequal competition with the competitors' populations controlled by natural mortality and fecundity. Considerable theoretical and practical interest at- taches to the influence of predatory mortality on such a system. Questions arise concerning whether more competing species can coexist, or whether competitors can coexist on a more even basis, where they are utilized by a common predator than in an otherwise similar environment without such top predation. Paine (1966) dealt with these ques- tions by observation and field experiment and suggested that some intertidal systems seemed able to support more competing species when a top predator was present. Parrish and Saila (1970) explored a small number of cases by dynamic simulation of systems using Lotka-Volterra type interactions. Some competitive situations were found in which two unequally competing species persisted longer in more equal numbers when utilized by a top predator. Subsequently, May (1971) did a neighborhood stability analysis of the same systems and determined stability criteria in terms of competitive and predatory coefficients. Using coefficient values picked on this basis, Cramer and May (1972) used the Parrish and Saila model to demonstrate a case where an unstable two-species competition became stable when a common top predator was added to the system. Figure 12 illustrates the behavior of a system with species C added as a top predator on the P12000 web of Figure 11. After some oscillation, the system moved to a new stable state with species C reduced to a level such that the competi- tors could support the total mortality. The stable relative biomasses of the competitors still reflect the competitive advantage of species B, but the ratio Bg /B^ = 1.23 is less than in the comparable 2-level system; i.e., the competitors occur in more nearly equal numbers. The result obtained by us- ing the graph theory parameter values in Equa- tion (19) of Saila and Parrish (1972) is B^ /B^ = 1.39. When compared with the B2 /B^ = 1.56 for the P12000 web, this also represents a more even standing among the competitors. Table 3 sum- marizes the B^/B^ values obtained by dynamic simulation and by graph theory. The same trend toward more equal biomasses of two species competing in the q coefficient when a common predator was present was found (Saila and Parrish 1972) using an independent set of "rough coefficients" provided by Menshutkin (1969). These comparisons of biomass ratios are 711 FISHERY BULLETIN: VOL. 73, NO. 4 10 20 30 40 TIME, YEARS 50 60 70 Figure 12. -Course of competition between two fish species, A and B, with unequal metabolic demands, competing for a com- mon food base and utilized equally by a common predator, C. {a^ = 0.7 ttA)- Table 3.-Biomass ratios for competing species. Biomass of species with lower a Biomass of species with highera Trophic system Dynamic model Graph theory prediction, Bg/B^ prediction, B^/Bj PI 2000 web P12100 web 1.30 1.56 1.23 1.39 related to the concept of equitability diversity (Lloyd and Ghelardi 1964). Some effects of human exploitation on systems of this kind have been briefly examined. Human exploitation on any species in any trophic web is expressed by the addition of Equation (18) to Equation (11) for that species (see MODEL sec- tion). Exploitation has been applied to two iden- tical competitors in simple P12000 webs which were initially at standard equilibrium. It has produced the expected result of reducing both populations. Since the system is energy-controlled, there is always an accompanying increase in the competitors' body weights (which are always equal), and an increase in food base biomass, B^ . The total biomass of the competitor trophic level remains essentially constant. Differential exploi- tation of the two competitors affects the ratio of Figure 13.-Effect of differential exploitation on the biomass ratio, B2/B3, of two equally competing fish species: dynamic simulation prediction. The coefficient of instantaneous fishing mortality for species 3 is always F3 = 0.3. their numbers, and therefore also their biomass ratio. Figure 13 shows an example, using identical species A-type competitors that have arbitrarily been designated species 2 and species 3. This is the kind of curve produced by graph theory analysis for exploitation situations by Saila and Parrish (1972); e.g., their Figure 6, curve B. Parameter relationships are considerably different in the two papers. Natural mortality, M, in the present case is about 10 times its value in the Saila and Parrish paper. For another set of parameter values and a particular series of values of the exploitation coefficient, F, the stable i?2 ^^3 ratio was predicted by both the dynamic simulation and the linear graph theory technique, as shown in Figure 14. Exploitation of a comparable 3- level trophic web has also been simulated. A common predator, similar to species C except smaller, was added preying equally on two competitors almost iden- tical with species A. A stable state for this unexploited system was found. Exploitation was applied to the competitors at various /' values that had been used previously with the P12000 web. At sufficiently low values of F (in the range of Figure 14), in the new exploited steady state, the food base biomass, i?, , increased with exploitation of the competitors. The competitor with the lower F value increased in absolute population and biomass, while the more heavily exploited compe- titor decreased in both. Again, total biomass at the second trophic level remained essentially constant. In all cases, population and body weight of the predator decreased markedly when the competi- 712 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION Figure 14. -Effect of differential exploitation on the biomass ratio, B2/B3, of two equally competing fish species: predictions of dynamic simulation (DS) and graph theory (GT). The coefficient of instantaneous fishing mortality for species 3 is always F3 = 0.1. tors were exploited. The ratio B.^/B^ from the simulation at a given E,/F^ ratio was slightly less extreme than in the P12000 web. However, the difference was too slight to permit a meaningful check against graph theory calculations. The top predator in this system could also be exploited using low values of F. The results were qualitatively similar to the last case above. The ^2/53 ratio changed only very slightly in this par- ticular system to values intermediate between those of the last case above and those obtained with the P12000 web. At even the lowest F values for the competitors in Figure 14, if exploitation of the top predator was carried above about F = 0.2, the top predator was lost from the system. The same result occurred with F<0.2 on the top preda- tor if slightly higher lvalues than those in Figure 14 were applied to the competitors. The two lower trophic levels persisted stably. This vulnerability of the top predator represents another limit to the stability of the larger system. Although it has not been explored, it appears to have implications for possible effects of exploiting real multispecies fisheries. The particular combinations used in this brief investigation were far from optimum for explor- ing a large range of exploitation intensities in multilevel webs (The choices were made primarily for similarity to other cases studied previously). The low permissible levels of predation and exploitation that the 3-level system would tolerate, together with the high natural mortality, made for difficulty in comparing results with those of trophic systems previously examined; e.g., by graph theory. However, the considerable similarity of the predictions in Figure 14 and Ta- ble 3 by these very different approaches seems highly suggestive. AVAILABILITY OF MODEL AND COMPUTING DETAILS Written descriptions of various portions of the model and their sources in somewhat more detail are available from the author. The basic computer software package used (IBM 1969) and a more ad- vanced version (IBM 1971) are described in the manufacturer's literature with enough detail in the former case for ready use by the reader. The CSMP package is sufficiently user-oriented that no further interface program is required; the model is written directly into the CSMP structure using simplified FORTRAN-like statements. Program listings and card decks for sample trophic models are available from the author, together with tables of input values used and a glossary of code names of variables and parameters. ACKNOWLEDGMENTS Part of this work was performed while the author was the recipient of a National Institutes of Health Fellowship, 5F01GM48175-02, General Medical Sciences. This work was based on part of a dissertation submitted in partial fulfillment of the requirements for the Ph.D degree at the Univer- sity of Rhode Island (Graduate School of Oceanography). S. B. Saila provided general coun- sel and encouragement and reviewed earlier ver- sions. W. H. Krueger and N. Marshall also read earlier versions of the manuscript. R. Sternberg was helpful in the solution of a differential equa- tion. LITERATURE CITED Adelman, H. M., J. L. Bingham, and .J. L. Maatch. 1955. The effect of starvation upon brook trout of three sizes. Prog. Fish-Cult. 17:110-112. Bagenal, T. B. 1957. The breeding and fecundity of the long rough dab Hippoglussoides p/afe.s.so/rfe.s (Fabr.) and the associated cycle in condition. J. Mar. Biol. Assoc. U.K. 36:339-375. 1967. A short review of fish fecundity. In S. D. Gerking 713 FISHERY BULLETIN: VOL. 73, NO. 4 (editor). The biological basis of freshwater fish produc- tion, p. 89-111. Blackwell, Oxf. and Edinb. Bagenal, T. B., and E. Braum. 1971. Eggs and early life history. In W. E. Ricker (editor), Methods of assessment of fish production in fresh waters, p. 166-198. Blackwell, Oxf. and Edinb. Beamish, F. W. H., and L. M. Dickie. 1967. Metabolism and biological production in fish. In S. D. Gerking (editor), The biological basis of freshwater fish production, p. 215-242. Blackwell, Oxf. and Edinb. Beverton, R. J. H. 1962. Long-term dynamics of certain North Sea fish populations. In E. D. LeCren and M. W. Holdgate (edi- tors). The exploitation of natural animal populations, p. 242-259. Blackwell Sci. Publ. Ltd., Oxf. Beverton, R. 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. 1959. A review of the lifespans and mortality rates of fish in nature, and their relation to growth and other physiological characteristics. Ciba Found. Colloq. Age- ing 5:142-180. Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Box, G. E. P. 1954. The exploration and exploitation of response surfaces: Some general considerations and examples. Biometrics 10:16-60. Box, G. E. P., L. R. Connor, W. R. Cousins, 0. L. Davies, F. R. Himsworth, and G. p. Sillitto. 1953. The design and analysis of industrial experiments. 0. L. Davies (editor), Oliver and Boyd, Edinb. Box, G. E. P., AND J. S. Hunter. 1957. Multi-factor experimental designs for exploring re- sponse surfaces. Ann. Math. Stat. 28:195-241. Brett, J. R. 1962. Some considerations in the study of respiratory me- tabolism in fish, particularly salmon. J. Fish. Res. Board Can. 19:1025-1038. Brett, J. R., J. E. Shelbourn, anitC. T. Shoop. 1969. Growth rate and body composition of fingerling sockeye salmon, Oncorhyncus nerka, in relation to temperature and ration size. J. Fish. Res. Board Can. 26:2363-2394. Cramer, N. F., and R. M. May. 1972. Interspecific competition, predation and species diversity: a comment. J. Theor. Biol. 34:289-293. Davis, C. C. 1968. Quantitative feeding and weight changes in Poecilia reticulata. Trans. Am. Fish. Soc. 97:22-27. Davis, G. E., and C. E. Warren. 1965. Trophic relations of a sculpin in laboratory stream communities. J. Wildl. Manage. 29:846-871. 1971. Estimation of food consumption rates. In W. E. Ricker (editor). Methods for assessment of fish produc- tion in fresh waters, p. 204-225. Blackwell, Oxf. and Edinb. Dawes, B. 1930. Growth and maintenance in the plaice {Pleuronectes plafesm L.). Part I. J. Mar. Biol. Assoc. U.K. 17:103-174. HOLLING, C. S. 1959. Some characteristics of simple types of predation and parasitism. Can. Entomol. 91:385-398. International Business Machines Corporation. 1969. System/360 continuous system modeling program (360A - CX - 16X) user's manual H20-0367-3. 1971. Program product continuous system modeling program (CSMP III) and graphic feature (CSMP III graphic feature) general information manual. Program no. 5734-XS9. IVLEV, V.S. 1961a. [On the utilization of food by plankton-eating fishes.] Tr. Sevastopol'. Biol. Stn. 14:188-201. (Transl. 1963, Fish. Res. Board Can. Transl. Ser. 447.) 1961b. Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven, 302 p. Karpov, V. G., F. V. Krogius, I. M. Krokhin, and V. V. Menshutkin. 1969. [Model of the ichthyocoenosis in Dalnee Lake realized on an electronic computer.] Tr. Vses. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. (VNIRO) 67:76- 87. (Transl. 1970, Fish Res. Board Can. Transl. Ser. 1572.) Kausch, H. 1968. [The effect of spontaneous activity on the metabolic rate of young carp {Cyprinus carpio L.) when starving and fed.] Arch. Hydrobiol./Suppl. 33 (3/4):263-330. KOSTITZIN, V. A. 1939. Mathematical biology. George G. Harrap & Co. Ltd., Lond., 238 p. Krogius, F. V., E. M. Krokhin, and V. V. Menshutkin. 1969. [The pelagic fish community in Lake Dalnee. An experiment in cybernetic modelling.] Akad. Nauk SSSR, KOTINRO Inst. Evol. Fiziol. Biol. I. M. Sechenov AN SSSR. Leningrad, "Nauka," p. 1-86. (Transl. 1970, Fish Res. Board Can. Transl. Ser. 1486.) Lassiter, R. R., and D. W. Hayne. 1971. A finite difference model for simulation of dynamic processes in ecosystems. In B. C. Patten (editor). Sys- tems analysis and simulation in ecology, 1:367- 440. Academic Press, N.Y. Lawrence, W. M. 1940. The effect of temperature on the weight of fasting rainbow trout fingerlings. Trans. Am. Fish Soc. 70:291-296. LeCren, E. D. 1958. Observations on the growth of perch (Percafliiviafilifi L.) over twenty-two years with special reference to the effects of temperature and changes in population densi- ty. J. Anim. Ecol. 27:287-334. 1962. The efficiency of reproduction and recruitment in freshwater fish. In E. D. LeCren and M. W. Holdgate (editors). The exploitation of natural animal populations, p. 283-302. Blackwell, Oxford. 1965. Some factors regulating the size of populations of freshwater Fish. Mitt. Int. Ver. Theor. Angew. Limnol. 13:88-105. Lloyd, M., and R. J. Ghelardi. 1964. A table for calculating the 'equitability' component of species diversity. J. Anim. Ecol. 33:217-225. Mackay, I., AND K. H. Mann. 1969. Fecundity of two cyprinid fishes in the River Thames, Reading, England. J. Fish. Res. Board Can. 26:2795- 2805. Mann, K. H. 1965. Energy transformations by a population of fish in the river Thames. J. Anim. Ecol. 34:2.53-275. 1967. The cropping of the food supply. In S. D. Gerking 714 PARRISH: MARINE TROPHIC INTERACTIONS BY DYNAMIC SIMULATION (editor), The biological basis of freshwater fish produc- tion, p. 243-257. John Wiley & Sons, N.Y. 1969. The dynamics of aquatic ecosystems. Adv. Ecol. Res. 6:1-81. May, R.M. 1971. Stability in multispecies community models. Math. Biosci. 12:59-79. Menshutkin, v. V. 1968. [Fish populations competition for food analyzed.] Zool. Zh. 47:1597-1602. (Transl. Natl. Tech. Inf. Serv., Spring- field, Va., JPRS 47, 345.) 1969. [Graph theory applied to aquatic organism communi- ties.] Zh.Obshch.Biol. 1969(l):42-49. (Transl. Natl. Tech. Inf. Serv., Springfield, Va., JPRS 47,689.) Menshutkin, V. V., and Y. Y. Kislyakov. 1967. [Simulation of a population of a commercial fish as- suming a variable growth rate.] Zool. Zh. 46:805- 810. (Transl. Div. Fish. Res., U.S. Bur. Sport Fish., Wildl.). 1968. [Simulation of the influence of the food supply on the dynamics of a fish population.] Zool. Zh. 47:341- 346. (Transl. Div. Fish. Res., U.S. Bur. Sport Fish. Wildl.). Menshutkin, V. V., and T. I. Prikhodko. 1968. [Investigation of production process on models of the simplest populations of water animals.] Gidrobiol. Zh. 4:3-11. (Transl. Off . Foreign Fish., Natl. Mar. Fish. Serv.) 1969. Productive properties of stable populations with a prolonged period of reproduction Hydrobiol. J. 5(l):l-7. 1970. Simulation of the population of planktonic crus- taceans by electronic computer. Oceanol. Acad. Sci. USSR 10:261-265. Menshutkin, V. V., and A. A. Umnov. 1970. A mathematical model of a very simple aquatic ecosystem. Hydrobiol. J. 6(2):18-23. NiKOLSKlI, G. V. 1961. [Concerning the cause of Huctuations in the abun- dance of fishes.] Vopr. Ikhtiol. 1:659-665. (Transl. 1962, Fish. Res. Board Can. Transl. Ser. 389.) 1962. On some adaptations to the regulation of population density in fish species with different types of stock structure. In E. D. LeCren and M. W. Holdgate (editors), The exploitation of natural animal populations, p. 265- 282. Blackwell, Oxford. NORDEN, C.R. 1967. Age, growth and fecundity of the alewife, Alosa pseudoharengufi. (Wilson), in Lake Michigan. Trans. Am. Fish. Soc. 96:387-393. Paine, R.T. 1966. Food web complexity and species diversity. Am. Nat. 100:65-77. Paloheimo, J. E., AND L. M. Dickie. 1965. Food and growth of fishes. I. A growth curve derived from experimental data. J. Fish. Res. Board Can. 22:521-542. 1966. Food and growth of fishes. II. Effects of food and temperature on the relation between metabolism and body weight. J. Fish. Res. Board Can. 23:869-908. PaRRISH, J. D., AND S. B. Saila. 1970. Interspecific competition, predation and species diversity. J. Theor. Biol. 27:207-220. Phillips, A. M., Jr. 1954. Effect of starvation upon brook and rainbow trout. Prog. Fish-Cult. 16:124. 1969. Nutrition, digestion, and energy utilization. In W. S. Hoar and D. J. Randall (editors). Fish physiology, 1:391- 432. Academic Press, N.Y. Pitt, T. K. 1964. Fecundity of the American plaice, Hippogldnytoides platessoides (Fabr.) from Grand Bank and Newfoundland areas. J. Fish. Res. Board Can. 21:597-612. Saila, S. B., and J. D. Parrish. 1972. E.xploitation effects upon interspecific relationships in marine ecosystems. Fish. Bull., U.S. 70:383-393. Scott, D. P. 1962. Effect of food quantity on fecundity of rainbow trout, Salmo gairdneri . J. Fish Res. Board Can. 19:715-731. Simpson, A. C. 1951. The fecundity of the plaice. Fish. Invest. Minist. Agric. Fish Food (G.B.), Ser. II, 17(5):l-27. Ten, V.S. 1967. [Trophic interactions of simple predator-prey pairs among aquatic organisms.] Respubl. Mezhvedomstven- nyi Sbornik, Ser. "Biol. Morya" Akad. Nauk Ukrain. SSR "Naukova Dumka" Kiev, Pap. 2:16-43. (Fish. Res. Board Can. Transl. Ser. 972.) Ursin, E. 1967. A mathematical model of some aspects of fish growth, respiration and mortality. J. Fish. Res. Board Can. 24:2355-2453. Warren, C. E., and G. E. Davis. 1967. Laboratory studies on the feeding, bioenergetics and growth of fish. In S. D. Gerking (editor), The biological basis of freshwater fish production, p. 175-214. Blackwell, Oxf. and Edinb. WiNBERG, G.G. 1956. [Rate of metabolism and food requirements of fish.] Nauchnye Tr. belorussk. Gos. Univ., Minsk., 253 p. (Transl. 1960, Fish. Res. Board Can. Transl. Ser. 194.) 1962. [The energy principle in studying food associations and the productivity of ecological systems.] Zool. Zh. 41:1618-1630. (Fish. Res. Board Can. Transl. Ser. 433.) WOODHEAD, A. 0. 1960. Nutrition and reproductive capacity in fish. Proc. Nutr. Soc. 19:23-28. 715 FISHERY BULLETIN: VOL. 73, NO. 4 APPENDIX The parameter, s, in the starvation mortality Equations (16) and (17) is found as follows: During starvation, since kC 1 'M\ V» \ is Qi j \( \ 0 REG ON }| «» ♦ * ■ A 400- ♦ 3-26-72 500- Rnn- » 5-04-73 * 6-17-73 ° 7-30-71 - 10-02-72 * 12-09-73 Figure l.-Temperature-depth profiles taken by expendable bathythermographs in the sampling area. on the University of Hawaii RV Teritu, 29 Sep- tember to 3 October 1972 (Table 1). Sunrise oc- curred at about 0630 h, sunset at 1830. The moon was in its last quarter; on the last day of the cruise it rose at 0330 and set at about 1530. All hauls were oblique and sampled primarily during descent. Two types of tows were made: deep tows 0-ca. 1,200 m (Figure 2a) and shallow tows 0-ca. 400 m (Figure 2b). Two deep-tow series (four day tows, five night tows total) were separated by a 24-h shallow-tow series (six day tows, six night tows); three twilight tows (no. 180, 184, 197) were also made, but the data were not included in computations of mean standing stocks. The two types of tows were designed to complement each other and provide data on diel vertical migrations; the two deep series were in- tended to examine day-to-day catch variability. The sampling depths were based on the results of previous horizontal sampling which indicated that nearly all micronekton resided between 400 and 1,200 m during the day. 727 FISHERY BULLETIN: VOL. 73, NO. 4 Table l.-Tow data for TEUTHIS-18. Tow number Local time Date Begin End Number flowmeter revolutions 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 9/29 9/30 10/1 10/2 1703 2040 0235 0657 1221 1912 0015 0522 0716 0904 1051 1236 1429 1620 1758 1937 2117 2257 0036 0232 0421 0630 1134 1856 2340 2025 0213 0408 1203 1710 0002 0500 0657 0851 1040 1220 1408 1607 1751 1928 2110 2250 0029 0210 0409 0554 1117 1612 2330 0434 517,814 998,460 290,325 1,061,855 841,941 884,407 872,342 399,056 307,434 336,306 284,953 346,136 307,172 306,759 275,434 295,355 284,199 293,163 274,276 308,441 305,249 935,456 844,093 857,374 935,631 0000 0100 TIME 0200 0300 OUOO TIME 2130 2230 Max. depth (m) 1,350 1,500 365 1,160 1,250 1,320 1,350 390 400 400 400 500 405 410 400 400 395 440 415 420 400 1,200 1,240 1,310 1,210 0500 ° 400 Figure 2.-Time-depth records for typical tows, depth plotted every 5 min. a. Deep tow, no. 204. b. Shallow tow, no. 196. A stepped sampling strategy was adopted. The ship speed was 100 m/min (3.5 knots) for all sampling except during retrieval when the ship was nearly stopped. For the shallow tows, 100 m of cable were let out at 50 m/min every 5 min until 1,100 m of cable were out. This placed the trawl at about the 400-m depth. After 5 min at this depth, the trawl was retrieved as fast as possible. Total duration of shallow tows was about 1.5 h, that of the retrieval phase about 0.3 h. For deep tows, the first 1,100 m of cable were let out in the same manner as for the shallow tows. Beyond 1,100 m, 200 m of cable were let out at 50 m/min every 10 min until 3,400 m of cable were out. After 10 min at this depth (about 1,200 m), the trawl was retrieved as fast as possible. Total duration of deep tows was about 5 h; retrieval took about 1 h. The distribution of mean sampling times in each 100-m depth in- terval is shown in Table 2. By inspection, the variability seemed small enough to treat all tows within each type equally. Table 2.-Distribution of mean sampling time per tow (minutes) and standard deviation (SD) with depth. Shallow Deep Day (6 tows) Night (7 tows) Day (4 tows) Night (5 tows) Sample depth interval (m) Mean time/ tow (min) Mean time/ tow SD (min) SD Mean time/ tow (min) SD Mean time/ tow (min) SD 0-100 21.0 0.0 21.3 1.4 19.3 4.9 17.4 3.2 101-200 22.0 1.4 21.3 1.0 20.3 6.7 17.8 2.8 201-300 21.5 1.7 22.2 2.4 24.3 0.6 21.4 3.9 301-400 26.0 1.4 25.2 3.1 26.0 4.6 28.4 3.8 >400 3.5 0.5 2.7 4.6 0-400 94.7 2.7 92.9 2.3 90.0 7.9 89.0 4.5 401 501 601 701- 801- 901- 1,001- 1,101. 1,201- 1,301- 1,401- 500 600 700 800 900 1,000 1,100 1,200 1,300 1,400 1,500 24.7 3.1 25.0 2.1 24.7 11.6 18.8 4.4 23.0 1.7 29.6 4.4 22.3 2.8 21.8 2.7 39.7 7.1 34.0 5.6 23.7 7.8 20.6 5.6 17.7 8.6 14.8 2.4 19.7 14.4 13.4 3.5 10.7 9.7 11.0 2.0 0.0 10.6 7.2 0.0 7.4 16.6 Total 291.0 13.7 292.0 23.3 Sample Processing For this study, micronekton was defined as the pelagic marine animals longer than 1 cm caught by the 10- foot I KMT with the trawling methods described above. Each catch was preserved in buf- fered 7% seawater-Formalin. The animals were originally sorted to family or genus but sub- sequently were lumped into larger groups chosen to roughly discriminate between vertical migra- tors and non-migrators, as well as among trophic groups (Table 3). Group names are capitalized in the text. All micronekton except siphonophores were counted, and all groups were weighed (blot- ted wet weight ±0.01 g). Siphonophore biomass was included in the Cnidaria group, but the 728 MAYNARD ET AL.: HAWAIIAN MESOPELAGIC MICRONEKTON Table 3. -Group composition used to sort each catch. 1. Myctophidae. 2. Cyclothone (Gonostomatidae). 3. Other Gonostomatidae. 4. Sternoptychidae, 5. Other Stomiatoidei: Astronesthidae, Chauliodontidae, Idia- canthidae, Malacosteidae, Melanostomiatidae, Stomiatidae. 6. Anguilliformes; Cyemidae, Eurypharyngidae, leptocephali, Nemichthyidae, Serrivomeridae. 7. Miscellaneous fishes: Alepisauridae, Apogonidae, Bathylagi- dae, Bregmacerotidae, Brotulidae, Ceratioidei, Cetomimidae, Chiasmodontidae, Evermanellldae, Giganturidae, Macrouri- dae, Melamphaeidae, Neoscopelidae, Omosudidae, Opls- thoproctidae, Paralepididae, Scorpaenidae, Trachipteridae, Zeidae, Zoarcidae, larval neritic, unidentified larval mid- water. 8. Caridea: Opiophoridae, Pandalidae, Pasaphaeidae. 9. Penaeidae: Penaeidae, Sergestidae. 10. Euphausiacea: Bentheuphausidae, Euphausiidae. 11. Mysidacea: Eucopeidae, Lophogastridae. 12. Miscellaneous Crustacea: Amphipoda, Isopoda, Ostracoda. 13. Cephalopoda. 14. Tunicata: Pyrosomidae, Salpidae. 15. Cnidaria: Hydrozoa, Scyphozoa, SIphonophora. 16. Miscellaneous invertebrates: Annelida, Cfenophora, Hetero- poda, Pteropoda, Nemertea. 17. Zooplankton: Copepoda, larval Stomatopoda, other mero- plankton, organisms <, 1 cm, residue. Pooled classifications Total fishes = Groups 1-7. Total Crustacea = Groups 8-12. Other invertebrates = Groups 14-16. Total micronekton = Groups 1-16. number of siphonophores could not be determined from the assorted zooids (Pugh 1974), and the numerical standing stock of Cnidaria is thus slightly underestimated. Organisms larger than 10 g/individual were weighed separately, but their weights and abundance were included in group totals. Most animals with greatest linear dimen- sions of about 1 cm or less, such as the euphausiid Stylocheiron spp., were placed into the zooplank- ton group although some sergestids in this size range were included in Penaeidea. Standing stocks of zooplankton shrimps were calculated from subsamples (Folsom splitter) of one deep tow. Calculations The volume of water filtered by each tow was determined by multiplying the distance travelled (as determined by the flowmeter) by the area of the net mouth. Mouth areas from 7.08 to 8.19 m^ have been reported for the 10- foot I KMT (Brooks et al. 1974). We used 7.7 m^ for our trawl. Zooplankton biomass was calculated for a mouth area of 7.7 m^ the full IKMT mouth, and 0.785 m^ the area of the cod end mouth. The true zooplank- ton concentration probably lies somewhere between these values because the anterior por- tions of the trawl funnel some zooplankton into the cod end while others pass through the meshes (Banse and Semon 1963; Hopkins 1966; Friedl 1971). To calculate the number of organisms or biomass of each group per 100 m^of ocean surface, the catch was divided by the volume of water filtered; this quotient was multiplied by the maximum depth of the tow and the product was then multiplied by 100. This computation assumes all depths were sampled equally. RESULTS Standing Stock The standing stocks from deep-day and deep- night tows were not significantly different (Ntest, P<0.05) in either number of organisms or biomass for most groups, including total micronekton. Only the numbers of miscellaneous fishes and Mysidacea showed significant diel differences. Likewise there were no significant differences (P<0.05) between the two series of deep tows (tows 183-186 vs. 201-204). Consequently we treated all deep tows as replicates and pooled the data to compute mean micronekton standing stocks for the 0- to 1,200-m deep water column (Tables 4, 5). Shallow-day (tows 188-193) and shallow-night (tows 182, 195-200) data were obviously different and were treated separately in these tables. The percentage composition of the fauna by group is illustrated in Figure 3 for each of the three classes of tows. The mean standing stocks of total micronekton for the 0- to 1,200-m water column are about 900 organisms and 500 g wet weight/ 100 m^ of ocean surface (Tables 4, 5). Fishes comprised over one- half of both the total numbers and biomass; crus- taceans constituted about one-third of the numbers and one-fifth of the biomass, while the cephalopods contributed only one-hundredth of the numbers but one-tenth of the biomass (Tables 4, 5). Cijclothone were more than twice as numerous as any of the other 15 groups, totalling almost 35% of the individuals caught (Figure 3a). The distribution of biomass among the groups varied less than the distribution of the abundance. No group contributed more than the Myctophidae which comprised 13% of total biomass (Figure 3a). A comparison of group rank by biomass with rank by abundance indicates that most of the more 729 FISHERY BULLETIN: VOL. 73, NO. 4 Table 4.-Micronekton standing stock, mean number of or- ganisms per 100 m'^ ocean surface. Standard deviation in parentheses. Group 0-1,200 m Day 0-400 m Night 0-400 m Myctophidae 108.13 (45.71) 1.02 (1.15) 80.72 (22.65) Cyclothone 308.29 (103.34) 6.07 (8.84) 19.48 (39.01) Other Gonostomatidae 25.77 (3.70) 3.25 (3.84) 22.13 (8.56) Sternoptychidae 22.97 (5.98) 0.42 (0.78) 4.36 (1.47) Other Stomiatoidei 3.92 (2.04) 0.00 2.68 (1.70) Anguilliformes 11.43 (4.46) 2.64 (1.04) 2.96 (2.23) Misc. fishes 27.49 (7.64) 16.31 (5.68) 21.08 (5.83) Caridea 40.18 (5.93) 1.70 (2.21) 27.52 (11.18) Penaeidea 137.90 (34.99) 6.83 (3.74) 132.90 (21.69) Euphausiacea 97.79 (28.16) 6.79 (5.77) 69.01 (11.50) Mysidacea 9.89 (3.02) 0.00 8.73 (6.67) Misc. Crustacea 4.08 (6.72) 0.00 1.88 (1.54) Cephalopoda 8.44 (2.04) 3.60 (1.98) 5.95 (1.81) Tunicata 28.05 (14.10) 50.26 (24.73) 25.01 (10.97) Cnidaria 10.80 (9.41) 11.82 (2.49) 11.47 (8.81) Misc. invertebrates 52.93 (25.02) 55.55 (7.73) 42.04 (10.03) Total micronekton 898.07 (149.50) 166.27 (16.45) 477.94 (69.27) Total fishes Total Crustacea Cepholopoda Other invertebrates 508.00 (133.50) 289.84 (54.25) 8.44 (2.04) 91.78 (25.22) 29.71 (7.48) 153.41 (60.43) 15.32 (4.92) 240.04 (13.91) 3.60 (1.98) 5.95 (1.81) 117.63 (19.26) 78.52 (22.71) No. organisms caught No. tov\/s 12,037 9 1,576 6 5,136 7 Table 5.-Micronekton standing stock, mean biomass, grams wet weight per 100 m^ ocean surface. Standard deviation in parentheses. Group 0-1 ,200 m Day 0-400 m Night 0-400 m Myctophidae 65.71 (20.36) 0.19 (0.17) 69.85 (12.26) Cyclothone 45.91 (11.26) 0.46 (0.75) 1.20 (2.92) Other Gonostomatidae 14.92 (8.48) 0.41 (0.49) 29.24 (25.22) Sternoptychidae 25.05 (14.02) 0.39 (0.91) 7.45 (2.75) Other Stomiato'del 15.56 (12.53) 0.00 13.36 (14.73) Anguilliformes 48.18 (43.65) 2.40 (3.04) 1.53 (1.63) Misc. fishes 41.37 (40.16) 2.65 (0.93) 9.46 (6.35) Caridea 50.49 (21.07) 0.15 (0.28) 30.27 (13.11) Penaeidea 31.59 (10.61) 0.13 (0.10) 22.71 (3.88) Euphausiacea 18.52 (3.98) 0.86 (0.80) 12.28 (1.57) Mysidacea 8.80 (9.76) 0.00 4.42 (3.45) Misc. Crustacea 1.09 (1.92) 0.00 1.09 (0.78) Cephalopoda 48.71 (47.46) 2.02 (2.02) 13.84 (16.28) Tunicata 34.07 (42.52) 5.90 (3.50) 21.68 (15.00) Cnidaria 40.86 (46.86) 10.50 (5.50) 11.34 (12.87) Misc. invertebrates 3.38 (3.49) 6.61 (2.35) 1.39 (0.33) Total micronekton 494.20 (99.30) 32.68 (6.58) 251.11 (53.46) Total fishes Total Crustacea Cephalopoda Other invertebrates 256.70 110.49 48.71 78.31 (81.30) (36.15) (47.46) (53.50) 6.50 1.14 2.02 23.01 (2.99) (1.06) (2.02) (8.34) 132.09 70.77 13.84 34.41 (39.33) (12.30) (16.28) (21.93) Zooplankton' Zooplankton^ 48.12 471.95 (21.64) (212.26) 15.24 149.45 (4.07) (39.94) 49.20 482.57 (12.30) (120.91) 'Calculated assuming 7.7 m^ net mouth, full 10-foot IKMT mouth. 'Calculated assuming 0.785 m^ net mouth, cod end mouth area. abundant groups are comprised of small in- dividuals (Figure 3a). Biomass estimates for zooplankton in deep tows ranged from about 10 to 100% of the total micronekton, depending on which mouth area was used for calculation (Table 5). % NUMBERS 5 15 25 35 % BIOMASS CYCLQI 1 PENAF 1 MYCT. 1 EUPH. I M. INV. ZT CARID. TUNIC. M FISH. J O. GON. J STERN. J ANGUI. , CNID. MYSID. CEPH. (a) M. CRUS. O. STOM. 15 JMYCT. CARID. CEPH. ANGUI. CYCLO. M FISH. CNID. TUNIC. PENAE STERN. EUPH. O STOM. O. GON. MYSID. M INV. M. CRUS. % NUMBERS % BIOMASS 5 1 15 25 35 1 1 1 1 1 M INV. 1 TUNIC. 1 , 1 M. FISH. 1 [(b) CNID. PENAE. EUPH. CYCLO. CEPH. O. GON. ANGUI. CARID. MYCT. STERN. 15 J L_ 25 J L_ C N I PI 35 I . I . I M INV TUNIC. M. FISH. ANGU I. CEPH. EUPH. CYCLO. O. GON. STERN. MYCT. CARID. PENAE. % NUMBERS % BIOMASS 10 20 30 J J 1 ] J 1 PENAEJ MYCT. 1 EUPH M INV I CARID. f TUNIC. 1 O. GON. ZJ M. FISH. Zf CYCLO. T CNID. T MYSID. J CEPH. j STERN. ANGUI. (C) O. STOM. M. CRUS. 20 30 MYCT. I CARID. O GON. PENAE. TUN IC. CE PH. O. STOM. EUPH. CNID. M. FISH. STERN. MYSID. ANGUI. M. INV CYCLO. M. CRUS. Figure 3. -Mean faunal composition by group as percent of total micronekton standing stock, number of organisms and wet- weight biomass. a. Deep-tows, 0-1,200 m.b. Shallow-day tows, 0-400 m. c. Shallow-night tows, 0-400 m. ANGUI. = An- guilliforme.s, CARID. = Caridea, CEPH. = Cephalopoda, CNID. = Cnidaria, CYCLO. = Cyclothone, E\]Vli. = Euphausiacea, M. CRUS. = Miscellaneous Crustacea, M. FISH. = Miscellaneous fishes, M. INV. = Miscellaneous invertebrates, MYCT. = Myc- tophidae, MYSID. = Mysidacea, 0. GON. = Other Gonostoma- tidae, 0. STOM. = Other Stomiatoidei, PENAE. = Penaeidea, STERN. = Sternoptychidae, TUNIC. = Tunicata. The small euphausiids (<1 cm) which were sorted into the zooplankton group constituted 3.9 g wet weight and 831 individuals/ 100 m^ ocean surface 730 MAYNARD ET AL.: HAWAIIAN MESOPELAGIC MICRONEKTON for one deep tow (no. 183). This is about nine times the mean abundance of micronektonic euphausiids, but only one-fifth of the mean euphausiid biomass. To establish a rough index of the contribution of neritic and meroplanktonic animals to the pelagic catch for future comparison, the standing stock of stomatopod larvae (otherwise included in the zooplankton group (Table 5)) was compared to the total micronekton stock (Table 6). The stomatopod concentration is large relative to shallow-day micronekton catches but quite small with respect to shallow-night and especially to deep-tow catches. Table 6.-Mean standing stock of stomatopod larvae expressed as percent of total micronekton standing stock per 100 m- of ocean surface, number of organisms and wet-weight biomass. Item 0-1,200 m Day 0-400 m Night 0-400 m Number (%) Biomass (%) 4.3 0.6 58.3 12.9 11.4 2.9 Diel Vertical Migration Because there were no significant differences between standing stock estimates of deep-day and deep-night tows, we concluded that any trawl avoidance was not due to diel factors and that increased shallow-night catches were primarily the result of vertical migration. Thus, the amount of vertically migrating micronekton was the difference between the shallow-day stock and the shallow-night stock (Tables 4, 5, 7). The percent migrating was then computed by dividing the amount migrating by the amount of micronekton deeper than 400 m during the day (deep-tow standing stock minus shallow-day stock). Percent- ages larger than 100 are considered sampling ar- tifacts. Of the total micronekton which resided deeper than 400 m during the day, 43% of the individuals with 47% of the biomass migrated into the upper 400 m at night (Table 7). This dramatic diel change in the catch rates of shallow tows is illustrated in Figures 4 and 5. The most numerous migrators were crustaceans, but most of the biomass was fishes. Between day and night tows the difference in composition of the 0- to 400-m layer of fauna is quite pronounced (Figure 3b, c). The average weight per organism in each group for the 0- to 1,200-m deep water column was com- puted by dividing the mean group biomass in Ta- ble 5 by the mean number of organisms in the group shown in Table 4. The results are presented in Table 8 along with computations of the average biomass of individual migrators and non-migra- tors, based on the data in Table 7 and assuming that the non-migrator standing stock was the Table 7.— Mean standing stock of vertically migrating micronekton, grams wet-weight biomass and number of organisms per 100 m^of ocean surface between 0 and 1,200 m, groups ranked by biomass and abundance. Percent migrating represents the portion of the group residing deeper than 400 m during the day which migrated into the 0- to 400-m layer at night. Biomass Abun( No./ jance g/ % % Group 100 m2 migr. Group 100 m2 migr. Myctophidae 69.66 106 Penaeidea 126.07 96 Caridea 30.12 60 Myctophidae 79.70 74 Other Gonostomatidae 28.83 199 Euphausiacea 62.22 68 Penaeidea 22.58 72 Caridea 25.82 67 Tunicata 15.78 56 Other Gonostomatidae 18.88 84 Other Stomiatoidei 13.36 86 Cyclothone 13.41 4 Cephalopoda 11.82 25 Mysidacea 8.73 88 Euphausiacea 11.42 65 Misc. fishes 4.77 43 Sternoptychidae 7.06 29 Sternoptychidae 3.94 17 Misc. fishes 6.81 18 Other Stomiatoidei 2.68 68 Mysidacea 4.42 50 Cephalopoda 2.35 49 Misc. Crustacea 1.09 100 Misc. Crustacea 1.88 46 Cnidaria 0.84 3 Anguilliformes 0.32 4 Cyclothone 0.74 2 Cnidaria (') Misc. invertebrates (') Misc. invertebrates (') Anguilliformes (') Tunicata (') Total fishes 125.59 50 Total Crustacea 224.72 82 Total Crustacea 69.63 64 Total fishes 123.70 26 Cephalopoda 11.82 25 Cephalopoda 2.35 49 Other invertebrates 11.40 21 Other invertebrates (') Total micronekton 218.43 47 Total micronekton 311.67 43 iO-400-m day stock > 0-400-m night stock. 731 600 FISHERY BULLETIN: VOL. 73, NO. 4 J I I OTHER INVERTEBRPTES CEPHflLOPOOfl, — \ 1 r-^ TIME: 0600 0800 1000 1200 IHOO t t t t t HAUL: 187 188 189 190 191 -I 1 1 — 1600 1800 2000 t t t t 192 193 194 195 =F 1 1 — 2200 2400 0200 t t t t 196 197 198 199 0400 0600 t 200 Figure 4. -Shallow-tow (0-400 m) standing stock of micronekton abundance over 24 h. Time is plotted for midpoint of each tow. difference between the deep-tow stock and the migrator stock. In general, the average weight per migrating fish is greater than the weight of the non-migrators, while the opposite holds for the crustaceans and cephalopods, that is, the smaller- sized members of the groups migrate. The occurrence of organisms larger than 10 g (wet weight) per individual was highly variable both with respect to abundance and biomass (Ta- ble 9). None occurred in the groups of Cyclothone, Sternoptychidae, Penaeidea, Euphausiacea, mis- cellaneous Crustacea, or miscellaneous inver- tebrates. The only large organism in shallow-day tows was one tunicate. In deep tows, large animals were less than 1% of the number of total micronekton but 30% of the biomass. With respect to biomass, about one-quarter of the total fishes. one-eighth of the total Crustacea, four-fifths of the Cephalopoda and one-half of the other inver- tebrates were made up of individuals larger than 10 g each. In shallow-night tows these proportions are smaller for all groups except other inver- tebrates which is about the same. DISCUSSION Restrictions The interpretation of our data must be con- sidered with several restrictions. Larger, highly mobile micronekton, especially fishes and cephalopods, probably avoid or escape the trawl (Pearcy and Laurs 1966; M. R. Clarke 1969; T. A. Clarke 1973, 1974) resulting in underestimation of 732 MAYNARD ET AL.: HAWAIIAN MESOPELAGIC MICRONEKTON 300 -I 1 1 1 ' L 250 - u a. u. (C z 0- to 1,200-m stock. 20- to 400-m day stock >0- to 400-m night stock. ^Siphonophores were not enumerated. single-cruise data. In other parts of the world, investigators have reported seasonal biomass fluctuations of tw^o- to sevenfold for some micronekton groups (Legand 1969; Blackburn et al. 1970; Tranter 1973), and abundance fluctuations of some species up to fifteen- to fortyfold (Pearcy 1964; Legand et al. 1970). Previous work in Hawaiian waters has demonstrated temporal variability in primary productivity (Gordon 1971:1132), epipelagic zooplankton standing stock (Nakamura 1967; Shomura and Nakamura 1969), and such mesopelagic micronekton as myctophids (T. A. Clarke 1973) and stomiatoids (T. A. Clarke 1974). The limited data available suggest at least twofold temporal fluctuations in micronekton standing stocks might be expected in Hawaiian waters. At this time we cannot predict seasonal oscillations in standing stock. Thus, until better seasonal data are available, our reported standing stock and faunal composition values should only tentatively be considered characteristic for Hawaii. Standing Stock On the basis of fish distributions, Amesbury (1975) has definied the 400- to 1,200-m depth range off Hawaii as the mesopelagic zone. Very few animals have been taken deeper than 1,200 m in opening-closing tows. Thus, standing stock values determined from our deep tows are probably reliable estimates for the micronekton of the whole water column except near-bottom waters. In spite of the shortcomings, the data add con- siderably to our knowledge of micronekton especially because we sampled nearly the whole depth range of the fauna, used a fully lined net with small mesh, towed at a relatively high speed, sampled when the moon had a minimal effect on avoidance (cf. T. A. Clarke 1973), took several sample replicates, monitored sampling volume and depth, and determined standing stocks for all components of the catch. The general lack of diel- related avoidance agrees with the findings of T. A. Clarke (1973) and Atsatt and Seapy (1974) but is in contrast to the results of Pearcy and Laurs (1966). Table 9.-Mean standing stock of micronekton larger than 10 g (wet weight)/individual, by group, a. No. organisms/100 m- ocean surface, h. Grams biomass/100 m- ocean surface. Standard deviation in parentheses. a. Numoer. Day Ni ight b. Biomass, grams/100 m^. Day Night Group 0-1,; 200 m 0-400 m 0-400 m Group 0-1,200 m 0-400 m 0-400 m Myctophidae 0.08 (0.22) 0.00 0.10 (0.26) Myctophidae 1.08 (3.25) 0.00 1.06 (2.82) Other Gonostomatidae 0.15 (0.28) 0.00 0.38 (0.55) Other Gonostomatidae 2.99 (6.41) 0.00 13.24 (20.52) Other Stomiatoidei 0.31 (0.36) 0.00 0.18 (0.32) Other Stomiatoidei 9.38 (11.25) 0.00 7.77 (16.46) Anguilliformes 0.81 (0.83) 0.00 0.00 Anguilliformes 26.60 (36.06) 0.00 0.00 Misc. fishes 0,56 (0.88) 0.00 0.00 Misc. fishes 19.25 (37.84) 0.00 0.00 Caridea 0.53 (0.56) 0.00 0.09 (0.25) Caridea 10.93 (12.41) 0.00 1.14 (3.01) Mysidacea 0.22 (0.49) 0.00 0.00 Mysidacea 3.06 (6.71) 0.00 0.00 Cephalopoda 0.54 (0.47) 0.00 0.27 (0.51) Cephalopoda 40.03 (49.24) 0.00 8.15 (15.80) Tunicata 0.83 (1.57) 0.10 (0.23) 0.54 (0.45) Tunicata 26.28 (12.95) 1.39 (11.54) 15.29 (14.92) Cnidaria 0.52 (1.16) 0.00 0.18 (0.49) Cnidaria 14.58 (35.30) 0.00 2.69 (7.11) Total micronekton 4.62 (2.27) 0.10 (0.23) 1.76 (0.65) Total micronekton 153.98 (54.80) 1.39 (11.54) 49.34 (14.81) Total fishes 1.97 (1.21) 0.00 0.67 (0.41) Total fishes 59.31 (58.50) 0.00 22.08 (20.18) Total Crustacea 0.76 (2.76) 0.00 0.09 (0.25) Total Crustacea 13.77 (17.36) 0.00 1.14 (3.01) Cephalopoda 0.54 (0.47) 0.00 0.27 (0.51) Cephalopoda 40.03 (49.24) 0.00 8.15 (15.80) Other invertebrates 1.35 (1.74) 0.10 (0.23) 0.73 (0.77) Other invertebrates 40.86 (44.70) 1.39 (11.54) 17.98 (18.57) No. organisms 61 1 19 No. tows 9 6 7 734 MAYNARD ET AL.: HAWAIIAN MESOPELAGIC MICRONEKTON We searched the literature for data to compare the standing stock and composition of micronekton off Hawaii with that of other regions. Few studies provided data from more than one sampling period; each used different gear and techniques, and all are subject to most of the restrictions cited previously. We feel that meaningful regional comparisons are premature until data on the tem- poral variability of the fauna are available from samples covering the entire depth range of the fauna. Diel Vertical Migration About one-half of the total micronekton wet- weight biomass in our study area appeared to migrate from day-depths greater than 400 m to night-depths shallower than 400 m. During the day, about 90% of the mean total micronekton standing stock biomass lived deeper than 400 m (Tables 5, 7). Because most vertically migrating fishes have a lower water content than non- migrators (Childress and Nygaard 1973) the per- cent of total micronekton dry weight which migrates would be even higher. The same probably holds for cephalopods. If migrators have shorter life spans and higher metabolic rates than non- migrators (cf. Childress and Nygaard 1973; T. A. Clarke 1973; Meek and Childress 1973), then the percent of the annual micronekton production represented by migrating animals would be especially high. ACKNOWLEDGMENTS This study would have been impossible without the help of many friends, whom we are happy to acknowledge. We are especially grateful to Richard E. Young who has provided guidance and inspiration throughout the project. T. A. Clarke, T. K. Newbury, and V. L. Ridge read the manuscript and made valuable critiques; we appreciate their efforts. E. M. Kampa graciously supplied her un- published irradiance data for our use. S. S. Ames- bury helped us process the fishes. We also extend our mahalo to Red Scholtz and his crew of the late RV Teritii for their able seamanship. This project was supported in part by NSF grant GA-33659, funds from the University of Hawaii Department of Oceanography, and by the Hawaii Community Scholarship Program. LITERATURE CITED Amesbury, S. S. 1975. The vertical structure of the micronektonic fish com- munity off leeward Oahu. Ph.D. Thesis, Univ. Hawaii. Atsatt, L. H., and R. R. Seapy. 1974. An analysis of sampling variability in replicated mid- water trawls off southern California. J. Exp. Mar. Biol. Ecol. 14:261-273. Banse, K., and D. Semon. 1963. On the effective cross-section of the Isaacs- Kidd mid- water trawl. Univ. Wash. Dep. Oceanogr., Tech. Rep. 88, 9 p. Barkley, R. a. 1971. Island wakes and their effects. Pac. Sci. Congr. Proc. 12(1):145. Blackburn, M., R. M. Laurs, R. W. Owen, and B. Zeitschel. 1970. Seasonal and areal changes in standing stocks of phy- toplankton, zooplankton and micronekton in the eastern tropical Pacific. Mar. Biol. (Berl.) 7:14-31. Brooks, A. L., C. L. Brown, Jr., and P. H. Scully-Power. 1974. Net filtering efficiency of a 3-meter Isaacs-Kidd mid- water trawl. Fish Bull, U.S. 72:618-621. Childress, J. J., and M. H. Nygaard. 1973. The chemical composition of midwater fishes as a function of depth of occurrence off southern Califor- nia. Deep-Sea Res. 20:1093-1109. Clarke, G.L. 1971. Light conditions in the sea in relation to the diurnal vertical migrations of animals. In G. B. Farquhar (edi- tor). Proceedings of an international syposium on biological sound scattering in the ocean, p. 41-.50. Maury Cent. Ocean Sci., Wash., D.C. Clarke, M. R. 1969. Cephalopoda collected on the SOND cruise. J. Mar. Biol. Assoc. U.K. 49:961-976. Clarke, T. A. 1973. Some aspects of the ecology of laternfishes (Myc- tophidae) in the Pacific Ocean near Hawaii. Fish. Bull., U.S. 71:401-434. 1974. Some aspects of the ecology of stomiatoid fishes in the Pacific Ocean near Hawaii. Fish. Bull., U.S. 72:337-351. Devereaux, R. F., and R. C. Winsett. 1953. Isaacs-Kidd midwater trawl, final report. Scripps Inst. Oceanogr., Oceanogr. Equip. Rep. 1, 18 p. Doty, M. S., and M. Oguri. 1956. The island mass effect. J. Cons. 22:33-37. Friedl, W. a. 1971. The relative sampling performance of 6- and 10-foot Isaacs-Kidd midwater trawls. Fish. Bull., U.S. 69:427-432. GiLMARTIN, M., and N. REVELANTE. 1974. The 'island mass' effect on the phytoplankton and primary production of the Hawaiian Islands. J. Exp. Mar. Biol. Ecol. 16:181-204. Gordon, D. C, Jr. 1970. Chemical and biological observations at Station Gollum, an oceanic station near Hawaii, January 1969 to June 1970. Hawaii Inst. Geophys., Tech. Rep. HIG-70-22, 44 p. 1971. Distribution of particulate organic carbon and ni- trogen at an oceanic station in the central Pacific. Deep- Sea Res. 18:1127-1134. 735 FISHERY BULLETIN: VOL. 73, NO. 4 GUNDERSEN, K., C. W. MOUNTAIN, D. TaYLOR, R. OhYE, AND J. Shen. 1972. Some chemical and microbiological observations in the Pacific Ocean off the Hawaiian Islands. Limnol. Oceanogr. 17:524-531. Hopkins, T. L. 1966. A volumetric analysis of the catch of the Isaacs-Kidd midwater trawl and two types of plankton nets in the Antarctic. Aust. J. Mar. Freshwater Res. 17:147-154. King,J.E.,andT.S. HiDA. 1954. Variations in zooplankton abundance in Hawaiian waters, 1950-52. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 118, 66 p. 1957a. Zooplankton abundance in Hawaiian waters, 1953- 1954. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 221, 23 P- 1957b. Zooplankton abundance in the central Pacific. Part II. U.S. Fish Wildl. Serv., Fish Bull. 57:365-395. Legand, M. 1969. Seasonal variations in the Indian Ocean along 110°E. VI. Macroplankton and micronekton biomass. Aust. J. Mar. Freshwater Res. 20:85-103. Legand, M., P. Bourret, R. Grandperrin, and J. Rivaton. 1970. A preliminary study of some micronektonic fishes in the equatorial and tropical western Pacific. In W. S. Wooster (editor), Scientific exploration of the south Pacific, p. 226-235. Natl. Acad. Sci., Wash., D.C. McGary, J. W. 1955. Mid-Pacific oceanography, Part IV. Hawaiian offshore waters, December 1949-November 1951. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 152, 138 p. Meek, R. P., and J. J. Childress. 1973. Respiration and the effect of pressure in the mesopelagic fish Anoploganter coniuta (Beryciformes). Deep-Sea Res. 20:1111-1118. Nakamura, E. L. 1967. Abundance and distribution of zooplankton in Hawaiian waters, 1955-56. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 544, 37 p. Pearcy, W. G. 1964. Some distributional features of mesopelagic fishes off Oregon. J. Mar. Res. 22:83-102. Pearcy, W. G., and R. M. Laurs. 1966. Vertical migration and distribution of mesopelagic fishes off Oregon. Deep-Sea Res. 13:153-165. PUGH, P. R. 1974. The vertical distribution of the siphonophores collect- ed during the SOND cruise, 1965. J. Mar. Biol. Assoc. U.K. 54:25-90. Schuert, E. a. 1970. Turbulent diffusion in the intermediate waters of the north Pacific Ocean. J. Geophys. Res. 75:673-682. Sette, 0. E. 1955. Considerations of midocean fish production as related to oceanic circulatory systems. J. Mar. Res. 14:398-414. Shomura, R. S., and E. L. Nakamura. 1969. Variations in marine zooplankton from a single locality in Hawaiian waters. U.S. Fish Wildl. Serv., Fish. Bull. 68:87-100. Tranter, D. J. 1973. Seasonal studies of a pelagic ecosystem (meridian 110° E). In B. Zeitzschel (editor), The biology of the Indian Ocean, p. 487-520. Springer- Verlag, N.Y. I 1 736 DESCRIPTION AND BIOLOGY OF A NEW SPECIES OF PELAGIC PENAEID SHRIMP, BENTHEOGENNEMA BURKENROADI, FROM THE NORTHEASTERN PACIFIC Earl E. Krygier- and Robert A. Wasmer^ ABSTRACT The new species of pelagic penaeid shrimp lacks the richly plumose arthrobranch described for the genus and has a single pair of terminal spines on the telson. It is found mainly in transitional water of the North Pacific between 500 and 1,000 m by day and 150 and 1,000 m at night. Examination of testes and ovaries, and the structures of the petasma and thelycum, indicates a 4-5 mo spawning season and an equal male to female sex ratio. Generation time was estimated to be 2 yr. This paper describes the systematics and biology of a new species of pelagic penaeid shrimp of the genus Bentheogennema. Since 1961, studies of the fauna and ecology of the mesopelagic waters off the coast of Oregon have been conducted by members of the School of Oceanography, Oregon State University. Several unusual species of macrurous decapod Crustacea have been obtained. The discovery and identification of this new species of Bentheogennema was by Carl Forss, who entrusted his material to the authors. Subsequent sampling with mid-water trawls has provided de- tailed information on the distribution and biologyof this shrimp, as well as abundant material for taxonomic description. METHODS AND MATERIALS Material for the zoogeographic distribution was collected in Isaacs-Kidd Mid-water Trawls (IKMT) from the research vessels Yaquina, Endeavor, John R. Manning, and Hugh M. Smith in the northeastern Pacific, normally within 320 m of the surface (Wasmer 1972). Information on vertical distribution, reproductive biology, and growth of this species was obtained from samples taken on five cruises aboard RV Yaquina at a single sampling station 65 nautical miles (120 km) off the central Oregon coast (NH 65-lat. 44°35'N, long. 'Research supported by the Ofhce of Naval Research (Contract NOOO- 14-67- A-0369-0007 under project NR 083-102) and the Atomic Energy Commission (Contract AT[45-1] 2227, Task Agreement 12). Publication number RL0-2227-T12-51. -School of Oceanography, Oregon State University, Corvallis, OR 97331. 'School of Oceanography, Oregon State University, Corvallis, Oreg.; present address: Bass Memorial Academy, Lumberton, MS 39455. 125°25'W) in 1972-73. Samples at this station were taken both day and night, using an 8-foot IKMT with a five net opening-closing cod end section similar to the one described by Pearcy and Mesecar (1971). All samples were preserved at sea in 10% buf- fered Formalin.' The samples were later sorted, identified, sexed when possible, and measured. Carapace length (measured from the postorbital margin to the median posterior edge of the carapace) was used as an indication of size. All figures were drawn with the aid of a camera lucida. In males, sexual maturity was based on three characteristics: 1) petasmata joined; 2) well- developed accessory lobe on anterior surface of the petasma; 3) and dilated vas deferens with large terminal ampoule (indicative of developed sper- matophore) at the base of the fifth pereiopod. The combined characteristics of fully developed thelycum and the posterior lateral lobe of the ovary swollen with eggs at the base of the fifth pereiopod were used as signs of sexual maturity in females. Estimates of growth are presented from analysis of length-frequency data. Section Penaeidea Family Penaeidae Bate Subfamily Aristaeinae Alcock Series Benthesicymae Bouvier Bentheogennema hurkenroadi n. sp. ri/pes.-Holotype (USNM 150835), male, carapace length (c.l.) 18 mm, from Station lat. Manuscript accepted January 1975. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 737 FISHERY BULLETIN: VOL. 73, NO. 4 51°26'N and long. 138°28'W, Mid-water Trawl No. 857 (MT 857); Allotype (USNM 150836), female, 14.5 mm c.l., station Newport Hydrographic Line 65 nautical miles (120 km) offshore (NH 65), MT 2130; Paratypes: 1 male (USNM 150837), 15 mm c.l., from NH 265, MT604; 2 males (USNM 150838), 8 and 6.5 mm c.l., NH 65, MT 2088; 1 male (USNM 150839), 14 mm c.l., NH 65, MT 2131; 1 male (USNM 150840), 13.5 mm c.l., NH 65, MT 2130; 4 females (USNM 150841), 14.5, 14, 10, 18 mm c.l., NH 65, MT 2130; 4 males (USNM 150842), 12.5 13.0, 13.2, 17.5 mm c.l., NH 65, MT 2130; 1 male (USNM 150843), 9.0 mm C.I., NH 50, MT 570; 1 male (USNM 150844), 9.5 mm c.l., lat. 40°28', long. 133°46', MT 613; 3 females (USNM 150845), 13.3, 14.0, 15 mm c.l., NH 65, MT 2121 Net #5; 1 female (USNM 150846), 20.0 mm c.l., NH 65, MT 2133 Net #1; 2 females (USNM 150847), 7 and 10 mm c.l., NH 65, MT 2070 Net #5; 2 males, 1 female (BMNH 1975:10), 14.2, 16.4, 13.2 mm c.l, NH 65, MT 2175 Net #5; 1 female (BMNH 1975:10), 15.7 mm c.l., NH 65, MT 2178 Net #4; 1 female (BMNH 1975:10), 12.6 mm c.l., NH 65, MT 2302 Net #4; 1 male (BMNH 1975:10), 11.5 mm C.L, NH 65, MT 2301 Net #1. Other, nonparatype, material deposited at Los Angeles County Museum; Fisheries Research Board of Canada Biological Station, Nanaimo, British Columbia; and School of Oceanography, Oregon State University, Corvallis, Oreg. Z)iagwo.sis.— Benthesicymae with podobranch on second maxilliped to third pereiopod inclusive; first ma.xilliped with single rudimentary arthrobranch; only sixth abdominal somite with middorsal carina; telson distally truncate, usually with single pair of mobile terminal lateral spinules. Accessory lobe of petasma characterized by large upturned terminal hook. Plate of thelycum on sixth thoracic sternite triangular and elevated, projecting ventrally in strong ridge; plate on eighth thoracic sternite pentagonal with anteriormost angle concave and anterolateral margins bearing spines. Description. -Rostrum extending to level of eye tubercle, well elevated above middorsal carina of carapace (Figure 1). Margin between rostral tip (apex) and dorsal spine with usual setal fringe (although broken in type). Middorsal carina of carapace bearing minute tubercle posterior to Figure 1. -Mature female Bentheogennema burkenroadi n. sp. illustrating the (a) anterior and (b) posterolateral lobes of the ovary. 738 KRYGIER and WASMER: NEW NORTHEASTERN PACIFIC PELAGIC PENAEID SHRIMP dorsal spine; carina absent between well-defined cervical and postercervical sulci and on posterior- most portion of the carapace. Mid-lateral longi- tudinal carina consisting of strong antennal carina continuous with hepatic and branchial carinae. Branchiostegal spine small and branchiostegal carina distinct; hepatic sulci con- tinuous from branchiostegal spine towards lower margin of carapace; post-hepatic carina orientated dorsoventrally from longitudinal hepatic carina toward inferior margin of carapace. Antennal angle obtuse and infra-antennal angle acute (Figure 2a). Only sixth abdominal somite with middorsal carina; second through fifth abdominal somites with weak lateral ridges in approximately dor- soventral position extending from mid-lateral to ventrolateral edge of the pleuron. Fourth, fifth, and sixth abdominal somites with prominent, roughly transversal, lateral ridges which together form "half moon" area (Figure 1). Fourth and fifth abdominal somites bearing small mid-lateral tooth on posterior margins. Antennal flagellum (Figure 1) similar to Gen- nadas (Foxton 1969), having proximal and distal sections divided by short series of annuli forming kink in flagellum; proximal section rigid, bearing scattered short nonplumose setae; distal section bearing paired arched plumose setae with small plumose setae perpendicular to flagellum at irregular intervals between bases of some arched pairs. Second element of antennular peduncle, along dorsal midline, 0.7 ultimate element (Figure 2b). Antennal scale (Figure 2c) little less than 3 times as long as greatest width; distinct spine (outer margin of scale), slightly convex, terminal end free, not extending beyond narrow apex of blade. Mandible (Figure 2d) with two segmented palp; palp thickly covered with setae on medial and lateral margins, distal element not quite as long as widest portion of basal element. Endopod of first maxilla (Figure 2e) distally narrow, with tip rounded; proximal gnathobasic lacinia (endite of coxa) subequal in width to distal lacinia (endite of basis), both terminating in strong spines among setae fringe. Anterior lobe of proximal lacinia (endite of coxa) of second maxilla (Figure 2f) strongly constricted behind apex, not broader than posterior lobe of distal lacinia (endite of basis); anterior lobe of distal lacinia very broad; endopod distally long and narrow, with two (sometimes three) curved spines at base of apical portion. Endopod of first maxilliped (Figure 2g) reaching beyond endite of basis but falling short of exopod; endopod of four elements, third less than twice second; fourth extremely minute; first element bearing usual compliment of three curved spines on distomesial margin. Exopod bladelike, without constricted, segmented distal portion. Merus of second maxilliped (Figure 2h), including anterior prolongation, 1.9 times as long as wide; dactylus with single strong apical spine surrounded by medium and small spines back to proximal end of propodus; merus and carpus with numerous spines and setae; podobranch present. Third maxilliped (Figure 2i) reaching to, or beyond, middle of ul- timate joint of antennal peduncle; ischium nearly 3 times as long as greatest width; merus usually twice as long as greatest width; carpus slightly longer than propodus; dactylus with long slender terminal spine; podobranch present. Merus of first pereiopod (Figure 3a) 1.4 times length of carpus and 1.7 ischium; fingers slightly setose. In second pereiopod (Figure 3b), carpus 1.2 times length of propodus; merus 1.2 carpus and 1.5 propodus; chela with heavy tufts of bristles. Merus and carpus of third pereiopod (Figure 3c) of equal length, each twice ischium; fingers of chela similar to those of second pereiopod. Carpus and propodus of fourth pereiopod nearly equal, each approximating two-thirds of merus which is 2.4 times ischium. Propodus of fifth pereiopod subequal to carpus which is subequal to merus; ischium slightly more than one-third of merus. Outer scale of appendix masculina (Figure 3d) longer than inner; proximal half of lateral margin expanding slightly then tapering toward base. Inner scale broadly rounded distally; spines on distomesial margin (few to many) long and thin, spines on distal margin smaller, stronger, and of uniform length. Telson with single pair of mobile terminal- lateral spines (Figure 3e) fringed with setae on terminal and distal two-thirds of lateral margins (of the large number of specimens inspected, only two mature males had any indication of more than one pair of mobile spines (USNM 150839, 150840), each with two pair of mobile spines on terminal edge of telson). No mobile nonterminal-lateral spines present on telson. Lateral margins of lateral rami of uropods (Figure 3f) bearing spine at 0.78 total length. Mesial rami about 0.73 lateral rami. Each half of petasma (Figure 3g, h), distally divided into three lobes (external, median, and in- 739 FISHERY BULLETIN: VOL. 73, NO. 4 a Figure 2.-Benfheogennema bvrkpnroadi n. sp. (Holotype, male 18 mm carapace length) a, carapace; b, antennular peduncle; c, antennal scale; d, mandible; e, first maxilla; f, second maxilla; g, first maxilliped; h, second maxilliped; i, third maxilliped. Scale equals 1 mm. ternal (Balss 1927) which are equivalent to Burkenroad's (1936) distoventral, distolateral, and distomedian lobes). External lobe bipartite; lateral part elongate projection with minute terminal teethlike protuberances distally; mesial part curv- ing inward with apex directed toward median lobe. Median lobe broadly rounded; subdistally, accessory lobe on anterior face of petasma, characterized by large upturned terminal hook (Figure 3g) with free margin attaching to base of 740 KRYGIER and WASMER: NEW NORTHEASTERN PACIFIC PELAGIC PENAEID SHRIMP Figure Z.—Bentheogennema hurkenroadi n. sp. (Holotype, male 18 mm carapace length) a, first pereiopod; b, second pereiopod; c, third pereiopod;d,appendixmasculina;e, distal half of telson;f,uropod; g, anterior view of petasma (EL = external lobe, ML = median lobe, IL = internal lobe, AL = accessory lobe); h, posterior view of petasma; i, anterior view of petasma from young male (8.0 mm c.l.); j, anterior view of petasma from young male (9.0 mm c.l.); k, anterior view of petasma from young male (9.5 mm c.l.); 1, thelycum of female (17 mm c.l.), arrow pointing to right sperm receptacle. Scale equals 1 mm. 741 FISHERY BULLETIN: VOL. 73, NO. 4 median lobe, attachment area distinguishable to level of elongate projection of external lobe. In- ternal lobe undivided, bearing rigid hooks con- tinuous with row of cincinnuni, holding two halves of petasma together. The accessory lobe develops early in the juvenile stage (Figure 3i, j, k) and together with characteristic spination of telson and presence of podobranchs behind second max- illiped, young of this species were discernable to a size of 6 mm c.l. (the smallest size captured). Thelycum (Figure 31) with plate on eighth thoracic sternite pentagonal, with anteriormost angle concave, anterolateral margin bearing long spines (this plate exhibits greatest variation dur- ing growth, being more rectangular in young females, changing to the pentagonal shape of ma- turity but becoming almost bilobed in very large females). Plate of seventh thoracic sternite bear- ing three anterior directed projections; lateral pair, shortest, bearing short spines; center projec- tion exhibits varying amount of concavity after maturity such that distolateral margins may ap- pear as raised wings. Elevated plate on sixth thoracic sternite triangular and inverted "V" shaped, with apex pointing anteriorly; apex not reaching anterior limit of sternite. Sperm recep- tacles located toward lateral edges near bases of inverted V. Coloration at time of capture varying from deep red over entire body to medium red on cephalothorax and lighter on abdomen. Black pig- ment fleck on distolateral edge of ocular peduncle just below corneal region (Figure 2a). Other small flecks of purple pigmentation often observed on carpus and propodus of third maxilliped and first and second pereiopods, on carpus of third pereiopod, and on ventral surface of abdominal somites just anterior to lateral edge of base of each pleopod. /Jemarfcs. -Burkenroad (1936) proposed the genus Bentheogennema for those species of Gennadas Bate which possesses podobranchs on the second maxilliped to third pereiopods inclusive. Other generic characters he included were: arthrobranch of first maxilliped large and richly plumose; exopod of first maxilliped without a constricted, segmented distal portion; dorsal carina on sixth abdominal somite only; telson with truncated apex and more than a single pair of mobile lateral spinules. As is often the case, the addition of a new species changes the generic formula for that group. The new species is similar to Gennadas in the armature of the telson but more closely resembles Bentheogennema with podobranchs on the second maxilliped through third pereiopod. We agree with Kemp (1909) and Burkenroad (1936) that the presence of podobranchs, a primitive characteristic, is a more important generic trait than the number of pairs of spines on the telson. We found that the two species of Bentheogen- nema—B. borealis (Rathbun) and B. burkenroadi n. sp.— from the Oregon coast lack the large, richly plumose arthrobranch on the first maxilliped that Burkenroad (1936) included as a generic characteristic. Both have small rudimentary arthrobranchs similar to Gennadas. We assume that Burkenroad (1936) did not have samples of B. borealis but included this arthrobranch structure as a generic characteristic from samples of B. in- termedium (Bate) and B. pasithea (Man). Although Tirmizi (1959) stated that the endopod of the first maxilliped is five-segmented in Gen- nadas and apparently only four-segmented in Bentheogennema, we have found that Gennadas propinquus Rathbun off the Oregon coast has a four-segmented endopod. Hence these characters are not reliable to distinguish these two genera. Bentheogennema burkenroadi can be separated from B. borealis, B. intermedium, B. pasithea, and B. stephenseni (Burkenroad) by the armature of the telson, and the structures of the petasma and thelycum. The telson of B. burkenroadi typically possesses only a single pair of terminal-lateral spines, whereas the other members of this genus possess two or more pairs of lateral spines: B. borealis and B. stephenseni two pairs; B. pasithea three pairs; B. intermedium (as described by Tir- mizi 1959) four pairs. The number of spines present on the telson should not be held as an invariable characteristic, there is undoubtedly a small percentage of variation as exemplified by the two males of B. burkenroadi (USNM 150839, 150840) which possess two pairs of terminal spines. It is possible that one of the two specimens of Gennadas calmani (Kemp) (synonymy: B. borealis), which Kemp (1909) illustrated with two pairs of terminal spines is also an example of such variation. The petasma of this new species is unique and easily distinguishable from that of other members of the genus. The combined structures of the ac- cessory lobe with its mode of attachment, its large 742 KRYGIER and WASMER: NEW NORTHEASTERN PACIFIC PELAGIC PENAEID SHRIMP size, and its terminal hook (present in mature in- dividuals) and the shape of the bipartate external lobes make identification, even of the juvenile stages (Figure 3g-k), possible. The thelycum differs from that seen in other species by the pentagonal shape of the plate on the eighth thoracic sternite and the elevated triangular plate on the sixth sternite (Figure 31). We have named B. hurkenroadi after Martin D. Burkenroad, whose work on Crustacea, especially the Penaeidae, is well known. GEOGRAPHICAL AND VERTICAL DISTRIBUTION Shrimps were examined from mid-water trawl collections taken over much of the North Pacific (Figure 4). Bentheogennema hurkenroadi was found only in collections from the northeastern sectors (lat. 52-34°N and east of long. 142°W) (Wasmer 1972). Pearcy and Forss (1966, 1969) ob- served B. hurkenroadi off the coast of Oregon, as close as 28 km to the northern end of the coast and occurring >92 km off the central and south coast. Wasmer (1972) found the greatest concentration in the Transitional Water Mass (Figure 5), with a few individuals occurring in the Pacific Subarctic and eastern North Pacific Central Water Masses. It is assumed to be a transitional species, although as is the case for many shrimps, it is not totally confined to a single physicochemically definited water mass (Wasmer 1972). Since B. hurkenroadi is a deep mesopelagic species and most of the available geographical collections were from shallow depths, the known geographic range will undoubtedly be increased by more systematic deep trawls in the eastern Pacific. This species was captured in opening-closing nets from the surface to 1,000-m depth. It ap- parently demonstrates a diel vertical migration. The depth distribution is, with few exceptions, below 500 m during the day and below 100 m at night (Table 1). Neither day nor night distribu- tions are confined to a narrow depth stratum but are diffused in concentration over a broad range. The nocturnal migration into the upper waters appears to entail only a small segment of the population with the main concentration remaining at depth. Those migrating above 500 m included both sexes, though the immature shrimp (11 mm c.l.), mature females (>12 mm c.l.), and sexually immature male and female Bentheogennema hurkenroadi at a sampling location 65 nautical miles (120 km) off Newport, Oreg. Oat. 44°35'N-long. 125°30'W). Time of Depth (m) year Size group 0-50 50-100 100-150 150-200 200-300 300-400 400-500 500-600 600-700 700-800 800-900 900-1,000 DAY June 1972 Mature males Mature females Immatures 1 3 3 5 4 2 6 2 3 2 2 2 2 Sept. 1972 Mature males Mature females Immatures 1 2 2 6 8 2 4 2 1 1 4 1 Nov. 1972 Mature males Mature females Immatures 1 1 1 2 2 2 3 1 4 3 1 5 1 1 Mar. 1973 Mature males Mature females Immatures 1 1 '3 '1 '1 Total number 1 1 1 1 9 19 41 19 14 Total volume n 1,000 m3 103.5 149.9 103.2 94.7 474.7 497.9 310.9 259.6 132.6 182.3 183.3 162.4 Number/1,000 mJ (day) 0 0 0.0097 0.0106 0 0.0020 0.0032 0.0347 0.1433 0.2249 0.1037 0.0862 NIGHT June 1972 Mature males Mature females Immatures 1 1 3 1 2 3 2 1 1 Sept. 1972 Mature males 22 1 1 1 4 5 21 2 1 1 Mature females 1 2 5 2 2 3 1 4 Immatures 3 1 1 1 5 Nov. 1972 Mature males 1 5 2 7 3 5 Mature females 21 4 7 2 1 5 1 1 4 Immatures 1' 1 1 3 1 Mar. 1973 Mature males 3 3 1 1 1 1 1 Mature females 1 1 4 2 1 1 2 Immatures ? 1 1 Total number 3 2 8 17 30 19 12 20 19 3 15 Total volume 1 n 1,000 mJ 111.8 530.6 184.8 292.8 251.0 435.9 288.6 261.6 147.2 170.7 132.3 129.3 Number/1,000 m3 (night) 0.0268 0 0.0108 0.0273 0.0677 0.0688 0.0658 0.0459 0.1359 0.1113 0.0227 0.1160 'Twilight 1 h after sunset. 'Twilight 1 1 before sunrise. 744 KRYGIER and WASMER: NEW NORTHEASTERN PACIFIC PELAGIC PENAEID SHRIMP Male Female 30- 20 c 10 0 30 20 10 Night n I N=I25 ^ • T I T I 1 I I — I — I — I — I 1 — T Day N=I32 -I — I — I — r— 1 — r I ' — n — r-1— i 6 8 10 12 14 16 18 6 8 10 12 14 16 IB 20 Carapace Length (mm) Figure 6. -Day and night length-frequency distributions of male and female Bentheogennema burkenroadi n. sp. night with no progressive drop in concentration with increased depth to 1,000 m (Table 1). The slight upward movement of this species may be related to its morphology. Vinogradov (1968) considers reduced musculature and a thin integument, which we observed in Bentheo- gennema burkenroadi, to be a means of achieving buoyancy. Because of weak swimming muscula- ture, they may swim too slowly to keep pace with the upward movement of the stimulating isolume, resulting in broad day and night distributions (Donaldson 1973). REPRODUCTION Since penaeid shrimp do not brood their eggs, a description of the breeding cycle must rely on anatomical changes, especially in the development of the ovary and ova. The female reproductive system consists of a bilaterally symmetrical ovary and paired oviducts internally, and externally of a thelycum. Each half of the mature ovary has an anterior lobe angling from the cervical sulcus and almost reaching the base of the eye, and then folding back along itself (Figure la). The anterolateral lobe lies over the hepatopancreas extending approximately one-half the way down the body wall. The posterolateral lobe, of such a mature ovary, will have visible distinct ova, measuring up to 240-288 jum crossectional diameter, and will extend ventrally, making a pouchlike structure at the base of each of the fifth pereiopods (Figure lb). The posterior lobes ex- tend beneath the dorsal abdominal muscle bands. becoming swollen in the first abdominal segment and then extending on toward the end of the third segment. Females were considered to have reached maturity after attaining a size of 12 mm c.l. and males at a size of 11 mm c.l. The reproductive cycle, as judged from the sex- ual condition of the testis and ovary, appears to consist of a 4- to 6-mo spawning season and a 6- to 8-mo resting phase. Based on samples collected in 1972 and 1973, the carapace of females in June is fairly rigid, though the ovaries are not ripe. Some males, from external observation, appear to be ready to release sperm, though most display only partial swelling of the terminal ampoule and vas deferens or lack swelling at all. By fall, females exhibit developing ovaries (two females were in spawning condition), and the carapace is corre- spondingly rigid. Most males have full, ripe looking testes and dilated terminal ampoules. By the end of November, spawning is in evidence. Most females are mature with readily distinct ova; some mature females have evidently spawned as the thoracic cavity appears empty; the carapace is correspondingly nonrigid, due to the spent ovary which had crowded much of the other organs; others have developing ovaries distended by small diameter ova. All males at this time have ripe testes and dilated terminal ampoules. By February, 50% of all females exhibit signs of spawning activity; the rest have probably spawned because their thoracic cavities appear empty and the carapace nonrigid due to the flaccid ovary. Most males still exhibit ripe testes and enlarged terminal ampoules. The sex ratio for adult males to females (A^ = 440), when all tows are included, was: 1:1, 1:1.08, 1:1, 1:1.02, and 1.03:1 for the respective cruises. This approximate 1:1 sex ratio, if it applies to all ages, indicates that there is no selective mortality by sex for this species (Geise 1959). GROWTH If spawning occurs from November through February and young (6-7 mm c.l.) enter the population April through June (Figure 7), the in- tervening egg and larval stages must take 3 to 5 mo. Based on size frequency diagnosis, about 12 additional months are required to reach maturity (11-12 mm c.l.) and another 5 to 6 mo are required before spawning commences Thus the generation time is estimated to be about 2 yr. The largest 745 FISHERY BULLETIN; VOL. 73, NO. 4 Mole Female 6' '8 10' 'l2' 'l4' 'l6' 'l8 '20 Carapace Length (mm) Figure 7.— Length-frequency histx)granis of Bentheogennema burkenroadi n. sp., from the five cruises (1972-73) off the Oregon coast. shrimp captured were a 17-mm c.l. male and a 20-mm c.l. female. Since the mesh size of the net liner was small enough to retain the young (6-7 mm c.l.) and we assume equal chance of capture of young and adults, then adults apparently live more than a year or two after first spawning because the number of adults captured is greater than the number of immature. In fact, the 12- to 15-mm mode must consist of greater than one age class since by itself it exceeds the juveniles in number. This overlap of age classes at >12 mm c.l. indicates that growth slows after maturity is reached. ACKNOWLEDGMENTS We thank R. J. LeBrasseur (Fisheries Research Board of Canada, Nanaimo, B.C. [POG]) and E. C. Jones (Hawaii Area Fishery Research Center in Honolulu [POFI]) for providing us with shrimp specimens collected during their oceanographic exploration. We thank Robert Carney for his aid and instruction in the illustrations, and W. G. Pearcy for his critical review and financial sup- port. LITERATURE CITED Balss, H. 1927. Macrura der Deutschen Tief see-Expedition. 3. Na- tantia, Teil B. Wiss. Ergeb. dtsch. Tiefsee-Exped. "Val- divia" 23(6):247-275. BURKENROAD, M. D. 1936. The Aristaeinae, Solenocerinae and pelagic Penaeinae of the Bingham Oceanographic Collection. Materials for a revision of the oceanic Penaeidae. Bull. Bingham Oceanogr. Collect., Yale Univ. 5(2), 151 p. 1940. Preliminary descriptions of twenty-one new species of pelagic Penaeidea (Crustacea Decapoda) from the Danish Oceanographical Expeditions. Ann. Mag. Nat. HisL.Ser. 11,6:35-54. DONALDSON.H. A. 1973. The ecology of the genus Sergestes (Decapoda, Na- tantia in an area near Bermuda. Ph.D. Thesis, Univ. Rhode Island, Kingston, 234 p. FOXTON, P. 1969. The morphology of the antennal flagellum of certain of the Penaeidea (Decapoda, Natantia). Crustaceana 16:33-42. 1970a. The vertical distribution of pelagic decapods [Crus- tacea: Natantia] collected on the Sond Cruise 1965. I. The Caridea. J. Mar. Biol. Assoc. U.K. 50:939-960. 1970b. The vertical distribution of pelagic decapods [Crus- tacea: Natantia] collected on the Sond Cruise 1965. II. The Penaeidea and general discussion. J. Mar. Biol. As- soc. U.K. 50:961-1000. GlESE, A. C. 1959. Comparative physiology: Annual reproductive cycles of marine invertebrates. Ann. Rev. Physiol. 21:547-576. Kemp, S. 1909. The decapods of the genus Gennadas collected by H.M.S. 'Challenger.' Proc. Zool. Soc. Lond. 1909:718-730. Man, J. D. de. 1911. The Decapoda of the Siboga Expedition. Part I. Family Penaeidae. Siboga Exped. Monogr. 39a, 131 p. Pearcy, W. G., and C. A. Forss. 1966. Depth distribution of oceanic shrimps (Decapoda; Na- tantia) off Oregon. J. Fish. Res. Board Can. 23:1135-1143. 1969. The oceanic shrimp Sergestes similis off the Oregon coast. Limnol. Oceanogr. 14:755-765. Pearcy, W. G., and R. M. Laurs. 1966. Vertical migration and distribution of mesopelagic fishes off Oregon. Deep-Sea Res. 13:153-165. Pearcy, W. G., and R. S. Mesecar. 1971. Scattering layers and vertical distribution of oceanic animals off Oregon. In G. B. Farquhar (editor), Proc. Int. Symp. Biol. Sound Scattering Ocean, p. 381-394. Ocean Sci. Program, Mar. Cent. Ocean Sci., Dep. Navy, Wash., D.C. Rathbun, M. J.. 1902. Descriptions of new decapod crustaceans from the west coast of North America. Proc. U.S. Natl. Mus. 24:885-905. TiRMizi, N.M. 1959. Crustacea: Penaeidae, Part II. Series Benthesicymae. The John Murray Expedition 1933-1934 Sci.Rep. 10(7):319-383. Vinogradov, M. E. 1968. Vertical distribution of the oceanic zooplankton. A.N. SSSR Inst. Okeanol. (Translated from Russian, Israel Program Sci. Transl., 1970, 338 p., Natl. Sci. Found. TT 69-59015.) Wasmer, R. A. 1972. Zoogeography of pelagic shrimps (Natantia: Penaeidea and Caridea) in the North Pacific Ocean. Ph.D. Thesis, Oregon State Univ., Corvallis, 232 p. 746 THE AMERICAN SAMOA LONGLINE FISHERY, 1966-71 Howard 0. Yoshida' ABSTRACT Aspects of the longline fishery based at American Samoa covering the period from 1966 to 1971 are described. The fishery is discussed primarily as it relates to the albacore, Thunnus alalunga, and to a small extent the yellowfin tuna, T. albacares. The landings of albacore fluctuated between 17,722 and 28,310 metric tons from 1966 to 1971. Although no downward trend was evident in the relation between total landings and total effort, the relation between CPUE (catch per unit of effort) and effort showed a definite downward trend. Generally, fishing effort was confined to the north of lat. 20°S in the first and fourth quarters. In the second and third quarters large amounts of effort were also expended south of lat. 20°S. The length-frequency distribution of albacore showed that albacore sizes were stratified by latitude. North of lat. 20°S the size of albacore was rather uniform, in that only a single mode was evident in the length-frequency distributions. South of lat. 20°S two or more modes were evident. The Honolulu Laboratory of NMFS (National Marine Fisheries Service) has been involved in assessing and monitoring the fisheries resources and developing the high seas fishing industry of the territories and island possessions of the United States in the Pacific Ocean. Part of this work included an investigation of the longline fishery based in American Samoa, which resulted in a report describing the history of the fishery and the distribution, apparent abundance, and size com- position of albacore, Thunnus alalunga, landed from 1954 to 1965 (Otsu and Sumida 1968). The present report describes the status of the American Samoa longline fishery from 1966 to 1971; it is timely because the fishery has changed considerably since 1965, particularly with regard to the apparent abundance of albacore. Data published by Otsu and Sumida will also be used, especially where they are useful in illustrating certain continuing trends. The earlier report included a rationale for confining the study to the albacore, the principal species of tuna taken in the fishery. The data were reliable only with respect to albacore because the catches of the other species were often not sold in their entirety to the canneries and, therefore, were not included in the catch reports by the vessel operators. However, as will be discussed later, the vessel operators have expended a considerable amount of effort to catch yellowfin tuna, T. al- bacares, in recent years. It is believed that most of the yellowfin tuna are now included in the catch 'Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, Honolulu, HI 96812. Manuscript accepted February 1975. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. reports and this species has become an important factor in the American Samoa longline fishery, and can no longer be ignored. The tuna canneries, operated by Star-Kist Samoa, Inc. and the Van Camp Sea Food Com- pany, depend entirely upon deliveries made by foreign flag vessels and fishermen. One of the most notable changes in the fishery over the years has been in terms of vessel nationality. The fishery began in 1954 with seven Japanese vessels. Vessels from Korea entered the fishery in 1958, and from the Republic of China (Taiwan) in 1964. The Japanese continued to increase their participation until 1963, but thereafter began a gradual withdrawal. On the other hand, the vessels from Korea and Taiwan greatly increased in number until the fleet reached a peak in 1967. Due largely to the withdrawal of the Japanese, the fleet has decreased since 1967. During the last quarter of 1971 there were 209 vessels in the fleet, consisting of 4 from Japan, 90 from Korea, and 115 from Taiwan. SOURCES OF DATA The data in this report were obtained through the operation of a field station in American Samoa, established in 1963, and manned con- tinuously through December 1970 by personnel from NMFS, Honolulu. In January 1971 the field station was taken over by the Office of Marine Resources, Government of American Samoa. In the begining, the length, weight, and sex of 50 albacore, randomly chosen, were obtained from each trip landing. For various reasons, e.g.. 747 FISHERY BULLETIN: VOL. 73, NO. 4 changes in cannery operating procedures, changes in the sampling procedures were necessitated in subsequent years. The collection of sex and weight data was discontinued in early 1971. Catch and effort data have been provided voluntarily by the fishing vessel operators from about 75% of the fishing trips. The longliners on some occasions fish widely scattered areas on a single fishing voyage. Since there is no way to determine the origin of each fish in the catch and because the fish are sampled at random at the docks, it is probable that some samples included fish from several different loca- tions. This problem was minimized by summariz- ing the length data by large geographical areas. ANNUAL LANDINGS OF ALBACORE AND YELLOWFIN TUNA Except for dips in 1961 and 1964, the albacore landings increased steadily from 1954 to 1965 (Ot- su and Sumida 1968). In the 6 yr that followed, the landings fluctuated between 17,722 and 28,310 metric tons (Table 1). The gain in importance of yellowfin tuna can be seen in the increased landings in the later years (Table 1). After a period (1954 to 1964) of reported landings of less than 2,500 metric tons, the reported landings increased substantially and fluctuated between 4,514 and 8,567 metric tons from 1965 to 1971. APPARENT ABUNDANCE OF ALBACORE AND YELLOWFIN TUNA Otsu and Sumida (1968) used various indices of apparent abundance in discussing the status of the fishery for albacore during 1954 to 1965. These indices included catch per trip, catch per day, and catch per 100 hooks. However, because data on number of hooks fished per day were not available for the entire period of their study, they elected to use the fishing trip as the basic measure of effort in analyzing the apparent abundance of albacore. They also examined the relation between catch and effort, and CPUE (catch per unit of effort) and effort, in evaluating the effect of fishing on the stock. Ideally, in considering the mean annual CPUE as an index of apparent abundance, the fishery should affect all portions of the stock(s) under consideration equally throughout the years. Table L-Total annual albacore and yellowfin tuna landings in American Samoa, 1954-71. Landi ngs (metric tons) Year Albacore Yellowfin tuna 1954 338 597 1955 1,760 1,628 1956 3,680 2,113 1957 5,873 1,537 1958 9,869 2,458 1959 10,292 1,780 1960 10,852 1,134 1961 9,740 1,331 1962 13,326 1,406 1963 13,972 2,057 1964 10,652 2,452 1965 15,591 4,514 1966 25,278 6,531 1967 28,310 5,326 1968 17,722 7,337 1969 18,767 8,207 1970 23,875 7,689 1971 22,193 8,567 However, since the geographical limits of the fishery have been expanding each year, the situa- tion is almost certainly not ideal. In this section I will extend some of the analyses of Otsu and Sumida to determine if any changes have occurred in the apparent abundance of albacore from 1966 to 1971. Since there are now 9 yr of effort data in terms of the number of hooks fished, I will make greater use of the catch per 100 hooks to determine the apparent abundance of albacore. It is assumed that fishing efficiency is not influenced by the na- tionality of the vessels, for Skillman (1975) found no evidence to suggest that any gear modification or change in the nationality of the fleet has caused any change in the catchability coefficient of al- bacore in the Samoan fishery. Apparent Abundance of Albacore Otsu and Sumida (1968) analyzed the mean an- nual CPUE of albacore from 1954 to 1965 in terms of the catch per trip. They believed that the catch per trip was a satisfactory measure of apparent abundance because their analysis showed a close relationship between the monthly average catch per trip and the monthly average catch per day. However, there are some basic shortcomings in the catch per trip. One is that the catch per trip of any vessel is limited by its fish-holding capacity. Also, as indicated by Otsu and Sumida (1968), the catch per trip is influenced by changes in the number of days fished per trip and by changes in the size composition of the vessels in the fleet. Changes in the number of hooks fished per day can also affect 748 YOSHIDA: AMERICAN SAMOA LONGLINE FISHERY the catch per trip. The data do indeed show that the mean number of days fished per trip has increased during recent years (Table 2). The increase in the mean number of days fished per trip may also, in part, indicate change in the size of the vessels. That is, larger vessels, w^hich presumably have larger fish-holding capacities, probably fish more days to obtain a full load of fish. Table 2.-Total trip length, days spent fishing, and traveling time per trip by longline vessels based at American Samoa. Mean number of days Year Total trip length Fishing time per trip Traveling time per trip 1963 1964 1965 1966 1967 1968 1969 1970 1971 42.35 41.26 48.79 50.58 62.13 68.02 67.34 70.74 84.16 26.12 27.39 32.09 33.74 38.03 43.20 44.25 45.03 52.03 16.23 13.87 16.60 17.05 23.50 24.82 23.08 25.72 32.13 In their analysis of the apparent abundance of albacore from 1954 to 1965, Otsu and Sumida (1968) indicated that the mean catch per trip increased from 1954 through 1960 and then fell slightly and stabilized at a lower level in 1963-64. A continua- tion of this analysis (Figure 1) showed that the mean catch per trip continued to fluctuate around this lower level with no definite upward or down- ward trend. It is possible that the mean catch per trip is fluctuating near the mean fish-holding capacity of the vessels or near a level that is relat- ed to the profitability of the fishing trip. That is, a vessel may fish as many days as are needed to obtain a full load or until a catch that is at least profitable is obtained. The trip, as an index of ef- fort, does not take this factor into consideration, and, therefore, the catch per trip is not an accurate indicator of the apparent abundance. It would be useful then, to compare the catch per trip with the catch per day, which is not affected by as many variables as the catch per trip. The mean annual catch per trip from 1959 to 1971 fluc- tuated between 29.9 and 38.2 metric tons (Figure 2) and, as noted earlier, did not reveal any obvious trends. The mean annual catch per day during the same period fluctuated between 0.7 and 1.7 metric tons, and, contrary to the catch per trip, declined after 1962 suggesting that the longline vessels are fishing more days to compensate for the reduced catch per day. The mean total length of a fishing Z s s a in z o tn 3 O < 3 Z Z < "1 — \ — I — r MEAN CATCH /TRIP T — I — I — I — I — I — I — r 3 o 1954 1958 1962 1966 1970 Figure 1. -Total number of fishing trips, mean catch of albacore per trip, and annual albacore landings, 1954-71. z o < UJ s "1 1 r CATCH /TRIP CATCH/ DAY in z o I960 1962 1964 1966 1968 1970 Figure 2.-Comparison of the mean catch per day and mean catch per trip of albacore, 1959-71. trip, number of days spent fishing, and number of days spent traveling on each trip (Table 2) all showed an increasing trend from 1963 to 1971, which indicates that, in general, the vessels are traveling farther away from the home base to fish and are fishing more days per trip. 749 FISHERY BULLETIN: VOL. 73, NO. 4 The relation between the annual total landings of albacore and the total fishing effort (number of fishing trips) indicated that the annual landings increased with increasing fishing effort from 1954 to 1965. Based on this analysis, Otsu and Sumida (1968) concluded that the point of maximum yield of albacore had not been reached in the American Samoa-based fishery. The fishing effort continued to increase subsequent to 1965, and reached a peak of close to 800 fishing trips in 1967. The albacore landings also continued to increase with the increased effort. The relation between the annual landings and the effort, in total number of days fished, from 1959 to 1967 also showed that the landings increased with increasing fishing effort (Figure 3). The mean catch per day plotted against effort in total number of days fished, however, shows a decline in the CPUE from 1959 to 1971 with increasing effort (Figure 4). As noted earlier, our laboratory has been ob- taining, from the vessel operators, effort data in terms of number of hooks fished since 1963. Using these data, the mean monthly catch per 100 hooks of albacore from 1963 to 1971 was computed (Figure 5). A salient feature of Figure 5 is that the mean monthly CPUE fluctuated much more from 1966 to 1971 than they did from 1963 to 1965. It is not clear what caused this change in trends in the monthly CPUE after 1965. One possibility is that it is related to a geographical change in fishing effort. As will be discussed in more detail in another section, in the years after 1965 more fishing effort has been expended in latitudes south of lat.20°S where good CPUE's of albacore are obtained in the second and third quarters of the year. This fact could also account for the definite peak in abundance of albacore in June or July during 1966 to 1971. It is also evident, however, that there is a slightly declining trend in the CPUE from 1963 to 1971. A plot of the total annual catch of albacore against the estimated total annual effort in number of hooks fished from 1963 to 1971 is shown in Figure 6. During this period the estimated total effort ranged from about 13,165,000 to 51,092,000 hooks and the annual albacore catch from 10,652 to 28,310 metric tons. With some minor exceptions, there was a strong positive relation between the annual catch and effort for the 1963-71 period. Generally, the catches increased with increased fishiijg effort. Suda (1971) also found a similar relationship between albacore catch and fishing effort in the South Pacific from 1952 to 1968. 35 1 r 2 O a Z 20 3 O I o z o z < < Z Z < "T 1 ' ! 1 1 1 \ ] 1 r 18 20 22 24 26 28 30 32 EFFORT ( THOUSANDS OF DAYS FISHED ) Figure 3.— Relation between annual landings of albacore and effort. UJ Z X o z < UJ s 0 2 4 6 e 10 12 14 16 18 20 22 24 26 28 30 32 34 EFFORT (THOUSANDS OF DAYS FISHED) Figure 4.— Relation between mean catch of albacore per day and effort. In Figure 7 is plotted the CPUE in number of fish and in weight per 100 hooks fished against the estimated total annual number of hooks fished from 1963 to 1971. Both plots show a negative relation between CPUE and effort; the CPUE decreased with increased fishing effort. Thus, although the catch has been increasing with increasing effort, it appears that the fishery has had some effect on the stock size in that the CPUE has been declining with increasing effort. The analyses above reflect average conditions of 750 \ YOSHIDA: AMERICAN SAMOA LONGLINE FISHERY qI I I I I '' I I I I I I I I I I I I I I I I I I I I I I ]] I II Ill I I I I I I I I I I I I I I I ill 1963 1964 1965 1966 1967 1968 1969 1970 1971 Figure 5.-Monthly mean albacore CPUE and effort, 1963-71. z O 25 o a: o I 1 1 1 1 1 1 I I 1 ^1967 - - I966p-''^ / / 1970 >DI97I / 1968/ ^kIsbs - n / y/l965 - - ,J/ - 1 1 1 1 1 1 1 i 1 - 0 5 10 15 20 25 30 35 40 45 50 55 ESTIMATED TOTAL ANNUAL FISHING EFFORT (MILLIONS OF HOOKS) Figure 6.-Relatlon between total catch of albacore and total effort, 1963-71. the fishery taken as a whole. It is of interest to compare the conditions north and south of lat. 20°S, especially since the geographic boundaries of the fishery have been expanding over the years. That is, it would be useful to see how the CPUE in the older fishing grounds north of lat. 20°S com- pares with that in the more recently exploited grounds south of lat. 20°S. The annual mean CPUE was computed for the area north of lat. 20°S, which represents the older fishing ground, and the area south of lat. 20°S, which represents the newer grounds (Figure 8). This analysis must be viewed with some caution, however, since total effort expended south of lat. 20°S is less than that expended in the north. Southern waters are fished heavily only in certain quarters of the year and these are the quarters when good catch rates are obtained. The mean catch rates in the southern waters therefore, may be overly weighted by the catch rates obtained in these quarters. However, the mean annual CPUE is higher in the newer grounds than in the old. The mean annual CPUE's 751 FISHERY BULLETIN: VOL. 73, NO. 4 — .09 (/> z o >- 08 o o o I o o - - - 1963 ■ - - 1964(1,,^ 1966 - - 1965 / ■^1967 - - / - - 1970 - 1969 \ - N)I97I- - - - - 1 1 1 1 1 1 1 1 1 Z 8 I o o 1 1 1 19630 I964(i,„__^ 1 1 1965 1 1 1966 1 1 ^1967 - 1968 -~^ 1966 - 1965 "^~~^I967 - I96ec Xl969 ^^r^l970 - ''isri " - - 1 1 1 1 1 - 10 15 20 EFFORT (MILLIONS OF HOOKS) Figure 8.-Relation between CPUE of albacore and effort in areas south (A) of lat. 20°S and north (B) of lat. 20°S. should be repeated here, however, that despite the expansion into new fishing grounds the overall trend for the fishery has been a decline in CPUE (Figure 5). Furthermore, in the southern grounds to which the fishery has expanded, the albacore catches, as will be shown in a later section, are composed of a large proportion of smaller fish. These smaller albacore may not be the optimum size at which to harvest the stock. from 1963 to 1971 in the old grounds were all less than five albacore per 100 hooks, whereas in the new grounds they were all greater than five per 100 hooks during the same period. Another difference is that in the old grounds the mean CPUE has been declining with increasing effort. In the new grounds the CPUE increased with increasing effort from 1963 to 1966. In 1967, the CPUE declined slightly and from that year to 1971 the CPUE appears to have stabilized around six albacore per 100 hooks. These facts suggest that the albacore fishery has been able to maintain itself by continuously expanding geographically to take advantage of better CPUE in newer areas. It Apparent Abundance of Yellowfin Tuna The discussion of the apparent abundance of yellowfin tuna will be primarily in terms of how it relates to the apparent abundance of albacore in the Samoan longline fishery. For one thing, although albacore are selectively fished for by the fleet, yellowfin tuna (and small numbers of other species) are also caught by the longlines, and the CPUE for one species may affect the CPUE of another (Rothschild 1967). The CPUE for one species may affect that for another species because, in computing the CPUE, the total effort expended was applied to the catch of each species without regard to the competition of the species 752 YOSHIDA: AMERICAN SAMOA LONGLINE FISHERY for space on the gear. Furthermore, there is another compHcating variable: the fishermen ap- parently seek out yellowfin tuna when the catches of albacore are poor. They do this by modifying the longlines to fish shallower and by fishing closer to the equator where yellowfin tuna catches are known to be better. The fact that the price of yellowfin tuna increased from an average of $270 a ton in 1965 to $394 a ton in 1970 may also have been a factor. As noted earlier, the CPUE of albacore in rela- tion to increased fishing effort has been declining, especiallv in the years subsequent to 1967. During 1968 to 1971, tne yellowfin tuna CPUE was rela- tively good. It is apparent that as albacore fishing deteriorated, the vessels expended more effort to catch yellowfin tuna. The relation between al- bacore and yellowfin tuna CPUE in the Samoan longline fishery from 1963 to 1971 is shown in Figure 9. It appears that an inverse relation existed between albacore and yellowfin tuna CPUE: When yellowfin tuna CPUE was high, al- bacore CPUE was low, and vice versa. The correlation coefficient (r = -0.6636; df = 7), however, was not significant. < z 3 s o o X o o 1 1 , ' ' ' • 1968 1971 • aigeg •i965 - • 1964 • 1970 • 1966 - • 1963 • 1967 - 111 - 2.00 300 400 CATCH / 100 HOOKS - ALBACORE Figure 9.— Relation between albacore and yellowfin tuna CPUE. SPATIAL AND TEMPORAL CONSIDERATIONS Observations on the American Samoa fishery during the period from 1954 to 1965 indicated that the longline vessels shifted fishing grounds with season (Otsu and Sumida 1968). Seasonal and geographical differences in CPUE, however, were not readily apparent in the early years. In the following sections I will examine the spatial and temporal distribution of effort and CPUE in the fishery from 1966 through 1971. Effort In the early years of the fishery, the longliners confined their fishing largely to the vicinity of the Samoa Islands. Over the years, the vessels gradually extended their operations to more dis- tant waters, and by 1965 the fishing grounds reached from about long. 175°E to about long. 120°W between the equator and lat. 30°S (Otsu and Sumida 1968). From 1966 to 1971 there was a further extension of the fishing grounds; the ves- sels fished from off the east coast of Australia to long. 100°W and from about lat. 10°N to about lat. 40°S. The fishing effort, however, was not uniformly distributed throughout the geographical limits of the fishery. Rather, there were distinct seasonal patterns in the spatial dis- tribution of fishing effort. A composite geographic distribution of the fishing effort on a quarterly basis for 1966 to 1971 summarized by 2° squares between the equator and lat. 40°S and east of long. 150°E is shown in Figure 10. As composite charts they can reflect only "average" conditions. In the first and fourth quarters, the vessels generally fished to the north of lat. 20°S, and areas of concentrated fishing effort developed in about the same area each year. In the second and third quarters, the vessels expanded their operations to the south of lat. 20°S and, in addition to the usual area of concentrated effort to the north, an area of high fishing effort also developed to the south of lat. 20°S. However, there apparently has been a change from earlier years in the operations of the fleet because prior to 1966 the vessels fished in the north during both the first and second quarters (Otsu and Sumida 1968). These figures indicate that the vessels are moving south in the second quarter, earlier than previously. In any event, these seasonal changes in the concentration of fishing effort have been interpreted to reflect the movement of albacore in the South Pacific Ocean (Suzuki 1961; Otsu and Sumida 1968). Catch Per Unit of Effort The mean quarterly CPUE for 1966-71 plotted by 2° square areas is shown in Figure 11. In a 753 FISHERY BULLETIN: VOL. 73, NO. 4 = ° h °^ « §^ — 1 1 1 2 1 0 a> 1 O) u. ~ 0 o» 1 CD 0 i VI !2 01 - 0 a: Q < - Z < -> - I j 1 1 1 1 1 1 1M 1 1 a> = SI ?l 0!^ S n 1 1 V tn s Si ii to (M I 0 t »o ml Si «■■ <£ U3 n ~ 1 - O o CD UD S*^ rO| CM ~l - (O - o - 9M ml m' V CM (y u> (VJ (M CD U^ ? f =!i fOJ fO - m t evji 1 K (M m 9i 5 A - ^ h m CM (VJ e CD -^ en (O CMI fOI -j - CM ( 1 cj 5' to a> s m en '1 <7»i ci t CM ~ K 1 CM m <0 ~ in (Ci w' Sj 1 m * 1 •r> to at ?: CD' Si ?1 Oi t h- eo 5 ~l a io m 1 Wl...... to (7>' in. CM, ~! a> CMJ 1 ~ i g CM ^m o| sii SI CMI CM Si = - m' 1 j ff)' 5 v! M ^■i ? ^ 9' in to. 0 * m' 1 10 l =i 1 "1 CM m im lO, | Oi fO. CD 5 Si r- ^ sl mi Sj ? sP m - (M (M -| t a>{ i i Si Oi « $, ^. fOI r-l lO| -o! in| cu ~i * 1 R!i pO;-, '■ ■ (Ml' , " 5^ (Ml s lO! ? 81 031 CMI "•I fO| CM 8! 81. \ t£ ^M ^m (Ml 51 K 0 (Ml 1 cmI J 1 0! Si 9' € im tf> A O) h-l S -: oil o ic| ^i^V s m ^^ (M' h- "■j mi (0 CM 1 Ml 1 8 s; 5 S m §iM W to mi ^-, 5 -1 j (M| ID. 1 a> «■ 1 i P ?i< 5 (M. s: s 1 1 (01 m (M SI CNi (y[ tOi -1 ;s Si 5| m V ! e 1 Pi 3 ? M (T. CD 5 ■T 5' <0! (Ml ~i ■«■ S I;^ll ? m C h3 §p <7> Si SI 01 CJ K s m i § P (^ CM ?i <0' SI Si 1 1 1 1 1 i CJ (Ml 1 m 3 CM (M ^ ~ 5 s S!| 5' i ~l •n 1 ' » ^: s M B|| ^le V f-: iSi (M tf) 1 (M i 9 a>i <^ S CM CD' h-l c cm! 1 1 !. ^ i ftj OEM a S:. sP (7>| 3^ -1 *; 1 V Is^ (Ml ♦ 10 CM J *1 V «)) 1 1 2 ^ « f^ i 1 1 1 1 \ K in (M m g (M 1 1 t 1 i 0 « m tn (« tt, m 1 1 i A i i UJ ff> mI \ .— L_^ 1 1 ?* ^ 1 ^ V„ ! « -a to 3 O J3 3 XI 6 D C O B O 3 fi. be o o g o_ lU z -3 1 _l (T < i VI !2 ro Al ^ ^W ■ c CJ n h n tt CVJ (U * (0 eo = N CJ * tr> lO CM CJ CJ - OJ w s D CD * < r CJ Oi CJ 8 n c ri 5 lO (M (0 ? 8 (0 V V CM CM CM 1 S s CM CM - t * 9 i ^ ? P ffl CJ - (B to S ? in CJ 1 CM s CJ 0 Fl O S a> V - s ^P f S s ^ (£ tf i 2 s S o> to CM «o K * ^[™^ m r- R in CM ?1 m (J» « 0 F* 1 S R^ ? (0 8 CM <7i : S; s ^ O o M 1* 1 o I7> <£ - t~ 8 CJ ^'- ?i cj - w 5 P- - •o lo lO - s s s ^i 8 « tf> s P CJ V S (D h^ (D V R o CJ ^ o> OI CM lO If 1 W R * - K» - A in 3 o CD 1 o ■ K m 8 IS is 1 CJ ro ? ^ cm' (0 o> u CJ r^ ^^H — r^^^HrO OI - «■ 01 ^ ^ S^ in CJ c - o en in ? fO ro (O ? ^ £ -■■n K ^ 3 1 fO^H f CM c - » «■ u> « i V m to lO S o CJ m = S S - m CJ 0> O CM in ■ in s^ <7> 3; in O ... ^'xVV^ 8 (O Kl - CJ CJ s 5 p !S r ^ CM CJ •: ■) CJ 9M CM, S U> CJ OI a> m CJ (0 O / k. 5 in S l! "> o (O (O K CJ OI oo r m ,^' -^ - cu r- h ( = o» It ^ CO 8 ffl K) m lO in t: r CM fO s C ? in OD CJ tn ffl CJ - C o V CJ tn CJ I N r^ S s ff> lO O ^ V s s c 3 J CJ w * CJ r^ U) 0 D ^ CJ - o> c U u> OJ * eu CJ V ^^ ?^ ^^ J z LJ V_ "^ k. T3 ID 3 _C '■+J s o o l_ m c c« cn 3 O 0) 3 O) s 3 C 0) J3 OS o o B O 3 Cu cS Ul o O) >> C8 3 I Ed a: D O 755 FISHERY BULLETIN: VOL. 73, NO. 4 ° ° g s - fe' 1 IMBER OF HOOKS 49,999 3,000-299,999 )0,000 - 449,999 450,000 (£ lU OD UJ i- UJ g b rO 1 1 ll j b in b 1 s UJ b CM 1 z VI i£? rO Al 1 :- M ■ 1 fl- V lO - - lO lO * Oi CD CD CM CM CM OD - m in s 8 CM ID (D CM CM in to - ., CM s h- CM lO <0 g 5 in to cu CM CM U9 S cn CM tn CM CM = R u> to lO CD CM ? § s * 0) P sp^" OD CD CM at 1 K> CM lO - tf> Z (O fe o^ CM s £> V O tn (M 1 t£ — ^ M (D to lO 1 Ms CM CD lO ■■ to U> & s to fO, K 2? S m s «■ tn CM a m tn 00 CM rO lO ;: ■:■: ID' 1 ~i 1 CM Si sM m 2? g « tn CM m s CM p S^8: s^ (NJ, CD to CM 8 K h- s to 00 CM Si ' <7) j K) ■& \^ CM s CM V = to " O m - 5 i O 1 1 to OD CM gi 1 ^: CM CM - . CM CJl ~b S <7> m 1 1 »^ ^ rS CD 5 CD (D (7> o (0 8 s CD> » tf> m 1 - u lO 8 ?i 5 O = c 8 N Si id; 1 II in m C V ?1 ftl r^ g CD = tn OD (J). 8 r- ^: % ?i 1 to CD: s CM Ift tn CM — cu ^ s S|l C\J 0) ?^ ? 2 S' s W at to S ? m ■? ■■U) 8 .-^ o ^P 00 CM II OD 2 ID ? ^ ^ in 5 s X N CM i a s s CM CD S tn m CM s s^ IS fO (VJ (O CM lO o III ?ly CM g (D s fO g p o Si ? g CD ?h-^ K IS z CM * K s W lO CM 1 (X> «' 5^ ? s <£ - "- CM CM en 8 CD m K *' ' tf> CD; St , cm: a* s CM CM f" * «■ ? s s - CD CM f- ^^ tn in CM CO Si t r f"- tn CM CM 8 ? K Ot o CM e 0» N s CM CM V W c fO h- lO c IT) 1 1 _ ?^ L . /^ r '-' a> 3 _C '-M c o I CO s c« 1/3 3 O 3 be e 3 C O) 05 o 3 o. c« he o bo 3 I b: o 756 YOSHIDA: AMERICAN SAMOA LONGLINE FISHERY T3 1 1 O ^ — O in iri o VI cvi ro Al UJ OQ UJ 1- a. UJ i 1 1 o ^ M o w& ■I ; U) CM m CO ■ tf) r 1 o: c\j: m'. •T) -■:^ o Im OJ. •^ o PO'' in rO^^I mi b: CD: O; H-: oj; oj; ^^^^ <£> to ?^° ^ ff> ^0 to d to b '"Ki>:o^ m >"" s S^ in ^Htn ^Hro \ b cy ^■■.x- V (£ o a> b i^ pO; u> 'i N <7» ^^' (£> k i ^^CD^^I(T>^^ ' q- ll 1 gl 1 i e 6 b cn -^ •^ K> CJ :\, ^ S HI b b I^ pO ml S-: o: OD CM |. cn OJ b ^1 m m (T> CO b ^m »:"" PO; o: OJ CD'^^ 1 1 1 I i B B B m ? ^-■ o- IJ3 ^ i V •^i Kil 'W ^1 «■ 1 o BrO «^^HcM ^ro J3^^|u) 6 o ^' u> fOi ^ ■4 ■^1 to lO >$■- '^ ,^ i^ s- o^^ T ^^^ 0> U3 in 3rO^^ cn «n o o «i 1 lO 1 , Kl ^ ^ fO O CM o eg ^ P 1 O O o o o o CM ^ .X-L-^ -J I ^^ J vJ '■-J _ a; 3 o o OS D Oh o o XI e o s J3 ■5 o. CS u be o ho s :« B O « 0) CC I D 760 YOSHIDA: AMERICAN SAMOA LONGLINE FISHERY C c O I JO as Oh O o u s o 3 J5 O. ea I-. be o (N = 3,124) ■n rfH tTH" llln. 20' -29"S J- 1 (N=< J. 385) " 1 -rrfl rnJ T>>^ 30»-39»S _ r 30» -39* S JKn n r (N- *,8I5) T ■ TVJlrfVn (N = : 5,518) ' _ _1T ■TTTfTTfl Th TTTH^ ..J L K 60 70 80 90 FORK LENGTH (cm) 60 70 80 90 FORK LENGTH (cm) 100 120 Figure 13.— Length- frequency distributions of albacore arranged by sex and 10° bands of latitude, 1966-70. 763 FISHERY BULLETIN: VOL. 73, NO. 4 albacore in each of the latitudinal subdivisions. In the areas north of lat. 20°S, both male and female length-frequency distributions had a single well- defined mode. In the areas south of lat. 20°S the modes were less well defined. Koto and Hisada (1967) found a similar pattern in the length- frequency distribution of albacore in the South Pacific in 1961. Otsu and Sumida (1968) computed the mean length of albacore in the fishery during the period from 1962 to 1965. They divided the fishery into 5° bands of latitude and noted that the fish were smallest near the equator, largest at lat. 20° to 25°S and tended to be smaller again south of lat. 25°S. The results agree in general with those of Otsu and Sumida; however, as seen above, the length-frequency distributions indicate a more complex situation than do the mean sizes. North of lat. 20°S the mean sizes may be good indicators of the general size of albacore because the length- frequency distributions showed that the fish were composed of single uniform size groups. South of lat. 20°S, the catches were composed of several size groups of fish and the mean does not indicate the presence of different size groups of fish. For example, although Otsu and Sumida stated that albacore south of lat. 25°S tended to be smaller, my data show that large fish were also present in these latitudes. Although it is not readily apparent in the length-frequency distributions, there appears to be a declining trend in the mean size of albacore in the fishery. Otsu and Sumida (1968) noted that albacore taken in 1964 and 1965 were shorter on the average than those caught in 1963. My data show that the declining trend in the mean length of albacore has continued (Table 3). Table 3.-Mean lengths of albacore, sexes combined, 1963-71. Me an length Mean length Year (cm) Year (cm) 1963 95.0 1968 89.8 1964 91.6 1969 90.4 1965 91.3 1970 88.6 1966 91.7 1971 91.5 1967 90.8 The mechanisms that produced such a pattern in the distribution of sizes in the fishery are probably complex. The unique character of the length- frequency distributions among the four lati- tudinal bands must have resulted from some nonrandom distributional process. The larger numbers of smaller albacore in the more southern waters suggest that albacore are initially recruit- ed into the fishery in the area south of lat. 20°S. The high CPUE experienced in the second and third quarters south of lat. 20°S may be an in- dication of this. Also, it has been shown that juvenile albacore which originate from spawning that takes place north of lat. 20°S migrate south as they grow larger (Yoshida 1971). SUMMARY A comparison of various indices of apparent abundance of albacore indicated that catch per day and catch per 100 hooks were better indicators of apparent abundance than catch per trip. The mean catch per day and catch per 100 hooks of albacore have generally declined over the years, which suggests that the apparent abundance of albacore has declined in the American Samoa longline fishery. To compensate for the reduced CPUE, the longliners fished more days per trip and traveled farther from the home base seeking areas of good catch rates. The fishery apparently has had an effect on the albacore stock. Although the annual landings have continued to increase with increased fishing ef- fort, the CPUE has declined. That the apparent overall effect was not greater was due to the fact that the fishing grounds have expanded, especially into areas south of lat. 20°S where good catch rates were obtained. The mean annual CPUE plotted against fishing effort for selected, discrete areas north and south of lat. 20°S indicated that the fishery has not as yet had as great an effect in the south as it has to the north. There are at least two possible reasons for the better condition of the fishery to the south. First, the area south of lat. 20°S has not been exploited as long as the area to the north. Second, it was shown that the albacore are first recruited into the fishery in the latitudes south of lat. 20°S, which may account in part for the higher apparent abundance. There was some indication that there were temporal changes in apparent abundance of al- bacore south of lat. 20°S. Because of poor weather conditions or because the fishermen have learned through experience that catch rates are better during certain seasons, or a combination of these and other reasons, fishing effort expended south 764 YOSHIDA: AMERICAN SAMOA LONGLINE FISHERY of lat. 20°S fluctuates seasonally. Areas of con- centrated fishing effort were evident in the second and third quarters in the area south of lat. 20°S. Very little effort was expended in these waters in the first and fourth quarters. Good CPUE's were experienced in these areas of high fishing effort in the second and third quarters. Although there were some indications that the apparent abun- dance of albacore was low in the southern waters in the first and fourth quarters, more data are needed to show this conclusively. The apparent temporal changes in CPUE in the southern waters may be related to seasonal changes in recruitment of albacore into the fishery. The length-frequency distribution of al- bacore in waters south of lat. 20°S showed that several size groups of fish were represented in the catches, including groups of small fish not found north of lat. 20°S. The good CPUE in the second and third quarters may indicate periods of active recruitment. Generally, the albacore were stra- tified by size latitudinally; however, no such stra- tification was evident longitudinally. North of lat. 20°S the catches were composed of fish of a single size group. South of this latitude, as already in- dicated, the catches were composed of several size groups of fish. ACKNOWLEDGMENTS I thank Ray F. Sumida, who diligently processed or supervised the processing of the data for this report, and William H. Bayliff of the Inter- American Tropical Tuna Commission and Grant L. Beardsley of the Southeast Fisheries Center, NMFS, Miami, Fla., for reviewing this paper. LITERATURE CITED Griffiths, R. C. 1960. A study of measures of population density and of concentration of fishing effort in the fishery for yellowfin tuna, Neothxinnus macropterus, in the Eastern Tropical Pacific Ocean, from 1951 to 1956. Inter-Am. Trop. Tuna Comm. Bull. 4:41-136. HONMA, M., AND T. KaMIMURA. 1957. Studies on the albacore. V. The fishing condition and size of albacore taken in the South Pacific Ocean. [In Jap., Engl, summ.] Rep. Nankai Reg. Fish. Res. Lab. 6:84-90. Koto, T. 1966. Studies on the albacore. XL Distribution of albacore in the tuna longline fishing grounds of the South Pacific Ocean. [In Jap., Engl, summ.] Rep. Nankai Reg. Fish. Res. Lab. 23:43-53. Koto, T., and K. Hisada. 1967. Studies on the albacore. XIII. Size composition of South Pacific albacore caught by longline. [In Jap., Engl, summ.] Rep. Nankai Reg. Fish. Res. Lab. 25:37-47. Otsu, T., and R. F. Sumida. 1968. Distribution, apparent abundance, and size composi- tion of albacore {Thunnus alalunga) taken in the longline fishery based in American Samoa, 1954-65. U.S. Fish Wildl. Serv., Fish. Bull. 67:47-69. Rothschild, B. J. 1967. Competition for gear in a multiple-species fishery. J. Cons. 31:102-110. Skillman, R. A. 1975. An assessment of the South Pacific albacore, Thunnus alalunga, fishery, 1953-72. Mar. Fish. Rev. 37(3):9-17. Suda, a. 1971. Tuna fisheries and their resources in the IPFC area. [Jap. summ.] Far Seas Fish. Res. Lab. (Shimizu), S Ser., 5, 58 p. Suzuki, G. 1961. On tuna fishing at the Samoa base in 1956. [In Jap.] Tuna Fish. (Maguro Gyogyo) 76:13-18. YOSHIDA, H.O. 1971. Distribution, apparent abundance, and length com- position of juvenile albacore, Thunnus alalunga, in the South Pacific Ocean. Fish. Bull, U.S. 69:821-827. 765 EFFECTS OF PHOTOPERIOD-TEMPERATURE REGIMES AND PINEALECTOMY ON BODY FAT RESERVES IN THE GOLDEN SHINER, NOTEMIGONUS CRYSOLEUCAS Victor L. de Vlaming' ABSTRACT Various photoperiod-temperature regimes were examined for their effects on total fat content (excluding gonads) in Notemigonus crysoleucas; experiments were conducted during several different phases of the reproductive cycle. In Notemigonus, fattening normally occurs in fall and early winter concomitant with the early phases of gonadal development. Body fat stores are progressively depleted during the prespawning and spawning seasons. Warm temperature (25°C) normally favored body fat depletion in Notemigonus. Short photoperiod (9L/15D) accentuated the lipid depleting effects of warm temperatures. Low temperatures (12°-15°C) usually promoted lipid deposition. Short photoperiods complimented the lipid anabolic effects of low temperatures. Thus, a given photoperiod can have opposite effects on body fat levels depending on temperature. A long photoperiod, in combination with warm temperatures, is required for final gonadal maturation and results in a reduction of lipid reserves. Short photoperiod-warm temperature regimes have similar effects on fat levels, but bring about gonadal regression. Thus, the effects of photoperiod-temperature regimes on lipid metabolism are apparently not totally dependent on the effects of these environmental factors on reproduction. The effects of pinealectomy on lipid reserves varied depending on the phase of the natural reproductive cycle when the organ was removed, as well as, with the photoperiod-temperature regime under which the experimental animals were maintained. At 25°C and under a 15.5L/8.5D photoperiod, fat levels were frequently lower in pinealectomized than in sham animals. The opposite was usually true for fish exposed to a 9L/15D-25°C regime. Lipid reserves were normally greater in pinealectomized than in sham operated fish maintained on 15.5L/8.5D-12°C regime. Body fat composition was frequently less in pinealectomized than in sham operated animals exposed to a 9L/15D-12°C regime. Pinealectomy reverses the effects of photoperiod on lipid metabolism at a particular temperature. These results suggest that the pineal body is involved in regulating physiological functions and may serve as a photoreceptor and/or transducer of photoperiod information. In most temperate-latitude aquatic environments food availability varies seasonally and annual cycles of growth, reproduction, and fattening are normally observed in teleost fishes. Lipid reserves may be used to meet the energy demands of reproduction, and seasonal fattening cycles in teleosts may be related to sexual cycling (Lovern 1934; Luhmann 1953; Love 1957; Idler and Bitners 1960; Woodhead 1960; Nikolsky 1963; Wilkins 1967; Lasker 1970; de Vlaming 1971). Environmental factors such as photoperiod and temperature are used as cues to maintain annual reproductive cycles in fishes (for reviews see de Vlaming 1972, 1974), but little is known about environmental control of fattening. The pineal body of most fishes has sensory organ characteristics (e.g., Riideberg 1966, 1969; Omura and Oguri 1969; Owman and Riideberg 1970; 'Biology Department, Marquette University, Milwaukee, WI 53233. Bergmann 1971; Oksche et al. 1971). Histological examination of the pineal in various teleosts also reveals secretory gland characteristics (e.g., Takahashi 1969; Cheze and Lahaye 1969; Cheze 1970; Rizkalla 1970; Hafeez 1971). In mammals the pineal appears to function as an endocrine gland and the indolamine, melatonin, may be one of the hormones of this organ (cf. Reiter 1973). His- tochemical and biochemical data show that the teleost pineal has an active indolamine me- tabolism (Quay 1965; Hafeez and Quay 1969, 1970; Fenwick 1970; Owman and Rudeberg 1970). Very little, however, is known about the physiological role of the pineal in teleost fishes. De Vlaming, Sage, Charlton, and Tiegs (1974) showed that melatonin treatment results in body lipid depletion in Fundulus similii^ and Cyprinod/m variegatus acclimated to a long pho- toperiod. In F. similis acclimated to a short pho- toperiod during May, melatonin therapy also resulted in fat depletion, but body fat deposition Manuscript accepted March 197.5. FISHERY BULLETIN; VOL. 73, NO. 4, 1975. 766 de VLAMING: CONTROL OF FATTENING IN NOTEMIGONUS was observed in F. similis acclimated to a short photoperiod during July and treated with this in- dolamine (de Vlaming, Sage, Charlton, and Tiegs 1974). These investigators concluded that the pineal might be somehow involved in regulating body fat reserves in teleosts. The data of de Vlaming, Sage, Charlton, and Tiegs (1974) and others (Fenwick 1970; Urasaki 1972a, b, c) indicate that the influence of the pineal on physiological functions in teleosts may vary depending on season and photoperiod conditions. The objectives of the present investigation were to examine the effects of various photoperiod- temperature regimes on body lipid reserves in the cyprinid teleost Notemigonus crysoleucas and to determine if pinealectomy altered the response of this fish to the experimental regimes. The possible relationship between reproductive activity and body fat reserves was also examined. That is, the effects of photoperiod-temperature regimes and pinealectomy on reproductive activity were de- termined and compared to the data on fat me- tabolism. The effects of pinealectomy on reproductive activity and fattening were examined during different phases of the natural sexual cycle (at various times of the year) to de- termine if physiological responses vary seasonally. MATERIALS AND METHODS Samples of A'^. crysoleucas were collected in ponds around the area of Menomonee Falls, Wis. (lat. 43°10'N) at several different times during the year. The reproductive cycle consists of a spawn- ing season which extends from May through July. There is a postspawning season during August and September in which the gonads regress. From October through February there is a gonadal preparatory period, in which spermatogonia proliferate slowly and spermatocytes appear in the testes. Vitellogenesis is initiated during this period. March and April can be referred to as the prespawning period; during this time, final gonadal maturation occurs (i.e., spermatozoa fill the testes and ovaries are distended with mature oocytes). Several fish from each field sample were sacrificed, the gonads examined and body lipid levels determined at the time of collection; these fish served as a reference for the experiments that followed. In the following discussion the fish sacrificed at the time of collection will be identified as initial controls. Sham operated and pinealectomized fish were maintained under various photoperiod and con- stant temperature regimes (see Results) in 114- or 285-liter tanks supplied with aerated and filtered dechlorinated tap water. Temperatures selected for these experiments are within the range nor- mally experienced during the year in nature by this species. Illumination was a combination of incandescent and cool white fluorescent bulbs which gave a light intensity of 200 to 275 Ix at the surface of each tank. Fish were fed ad libitum on a commercial fish food (Tetra-Min)-; animals main- tained at warm temperatures were fed twice daily whereas fish at low temperatures were fed only once a day. All Notemigonus used in these studies weighed between 12 and 17 g. For pinealectomies, fish were anesthetized in buffered tricaine methane-sulfonate (1:4,000). Each fish was then wrapped in cheesecloth and submerged in water so that only the top of the skull was emergent. The section of skin covering the pineal area was cut and folded back to reveal the parietal bone. Using a diamond-edged wheel saw (diameter = 2.2 cm) attached to a dental drill, three sides of a rectangle (5x4 mm) were cut in the parietal bone. This bone flap was then lifted forward toward the animal's mouth to expose parts of the cerebrum and midbrain. The pineal could then be easily removed using a gentle suc- tion applied through a Pasteur pipette. After removal of the pineal, the parietal bone and the epithelial flaps were individually sealed into place with Eastman's 910 Adhesive. Sham operations consisted of raising the parietal bone flap without removing the pineal. Removal of the pineal can be completed within 2 min in this species. The effects of pinealectomy on reproductive function were assessed by gravimetric and his- tological techniques. Fish were sacrificed by severing the spinal cord. Body weight and gonadal weight were recorded immediately after sacrifice. Gravimetric data are expressed in terms of the gonosomatic index (GSI) (gonadal weight/body weight X 100) since gonadal size in this species depends on body weight. After weighing, gonads were fixed in Bouin's solution and embedded in paraplast for histological examination. The data obtained on the effects of pinealectomy on reproductive function in Notemigonus are the subject of another report (de Vlaming 1975). GSI -Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 767 FISHERY BULLETIN: VOL. 73, NO. 4 data are presented here without statistical com- parisons or extensive discussion so that com- parisons can be made with the results on body lipid resen'es. After removal of the gonads, the bodies of the fish were extracted to remove lipids. The procedure used to measure body fat content has been previously described (de Vlaming, Sage, Charlton, and Tiegs 1974; de Vlaming et al. in press). Basically this technique consists of ex- tracting in a methanol:chloroform:ether solution (1:1:1). Body fat content is expressed as a function of drj^ body weight. RESULTS Preparatory Season The effects of pinealectomy on body lipid resen'es were first examined during the gonadal preparatory period (January). The results of this experiment are summarized in Table 1. Body lipid reserves were significantly depleted in both male and female sham operated fish main- tained at 25°C (whether on a long or short pho- toperiod) compared to the initial January controls. Table l.-The effects of pinealectomy on body lipid reserves in Notemigonus maintained on various photoperiod-temperature regimes during the gonadal preparatory season (30-day treat- ment). Dry lipid Gonosomatic index' index2 Treatment n Sex X±SE X±SE Initial controls 5 M 248 ± 1 1 1.48 ±0.09 (January) 9 F 207 ± 10 3.03 ± 0.25 15.5L/8.5D photoperiod: 25°C: Sham operated 6 M 136 ±7 2.82 ± 0.07 8 F 164 ± 11 4.67 ± 0.42 Pinealectomized 5 M 119±5* 0.34 ±0.18 10 F 139 ±8* 2.88 ± 0.37 12°C: Sham operated 6 M 235 ± 10 1.95 ±0.16 8 F 211 ±7 3.42 ± 0.08 Pinealectomized 6 M 246 ±9 1.20 ±0.04 10 F 237 ± 5* 3.47 ±0.13 9L/15D photoperiod: 25°C: Sham operated 5 M 118±8 0.94 ± 0.09 8 F 96 ±6 1.78 ±0.22 Pinealectomized 6 M 106 ±5 0.89 ± 0.16 9 F 109 ± 11 1.64 ±0.21 12°C: Sham operated 6 M 263 ± 12 1.28 ±0.11 10 F 244 ±9 3.40 ±0.21 Pinealectomized 5 M 250 ± 15 1.19 ±0.18 8 F 221 ± 6* 4.12 ±0.35 'Dry lipid index = mg lipid, less gonadal lipids/g dry body wt. 'Gonosomatic index = wt of gonads (g)/g body wt x 100. 'Significantly (P<0.05) different than sham operated controls maintained under same photoperiod-temperature regime. In both sexes of sham operated fish exposed to the long photoperiod-low temperature regime body Hpid stores were maintained at the initial levels. Short photoperiod-low temperature treatment, however, caused a significant increase in fat stores in sham operated females, but not in sham oper- ated males. In sham operated control fish maintained on a long photoperiod, body lipid levels were sig- nificantly higher in the group at 12°C than in the group at 25°C. Fat levels were also significantly greater in sham operated Notemigonus main- tained at 12°C than at 25°C on a short photoperiod. These data suggest that low temperatures, regardless of daylength, either maintain or favor body lipid deposition in this species. At 12°C, a short photoperiod was more effective in maintaining or stimulating lipid deposition than a long photoperiod. Specificially, body fatness (in both sexes) was significantly greater in sham operated fish exposed to the short photoperiod-low temperature regime than in animals maintained on the long photoperiod-low temperature regime. Thus, short daylengths seem to compliment the effects of low temperatures on fat deposition. Body lipid levels were significantly lower in sham operated animals maintained on the short pho- toperiod-warm temperature regime than in fish exposed to the long photoperiod-warm tempera- ture regime. Body fat depletion in Notemigon iis at warm temperatures is therefore accentuated by short daylengths. In animals maintained on the 15.5L/8.5D-25°C regime, body lipid levels were significantly lower in the pinealectomized group than in sham operated group. Pinealectomy also retarded fat deposition in female fish exposed to the 9L/15D- 12°C regime. In contrast, body fat levels were significantly greater in the female pinealec- tomized fish than in the sham operated females maintained on the long photoperiod-low tempera- ture regime. Body lipid levels in pinealectomized fish did not differ significantly from lipid levels in control sham operated animals under any of the other experimental conditions. These data suggest that the effects of pinealectomy on lipid me- tabolism depend on photoperiod and temperature conditions. The effects of pinealectomy on reproductive function are discussed elsewhere (de Vlaming 1975). One should note, however, that pinealec- tomy retarded the stimulatory effects of the 15.5L/8.5D-25°C regime on gonadal maturation; 768 de VLAMING: CONTROL OF FATTENING IN NOTEMIGONUS under these same conditions body lipid levels were also significantly lower in pinealectomized than in sham operated fish. Ovarian GSI was significantly higher in the pinealectomized than in sham operated group exposed to the 9L/15D-12°C regime; however, body fat levels were sig- nificantly lower in pinealectomized than in sham operated fish under these conditions. These data imply that the effects of pinealectomy on lipid deposition do not necessarily depend on the effects of this organ on reproductive activity. Prespawning Season The effects of pinealectomy on body fat reserves were examined again during the pre- spawning season (Table 2). Regardless of photoperiod, low temperature treatment of both sexes of sham operated Notemigonus resulted in a significant increase in body fatness, compared to the initial March con- trols. Warm temperature treatment, however, caused a depletion of body lipids in sham operated fish compared to the initial controls; this fat depletion was observed in both photoperiod groups. Body fatness (in females) was significantly greater in sham operated control fish exposed to the 9L/15D-12°C regime than in animals main- tained on the 15.5L/8.5D-12°C regime. These data further confirm the suggestion that short daylengths compliment the effects of low temperatures on fattening. Body lipid levels (both sexes) were significantly lower in sham operated animals maintained on the short photoperiod- warm temperature regime than in sham operated fish exposed to the long photoperiod-warm temperature regime. The lipid depletion which oc- curs at warm temperatures is thus accentuated by short daylengths. Body fat levels (both sexes) were significantly lower in the pinealectomized group than in the sham operated group maintained on the 15.5L/8.5D-25°C regime. In animals maintained on a long photoperiod-low temperature regime, body lipid levels were significantly greater in the pmealectomized fish than in the shams. Pinealec- tomized females contained significantly more fat than sham operated females exposed to the short photoperiod-warm temperature regime. In con- trast, body lipid levels (both sexes) were sig- nificantly lower in pinealectomized than in sham operated fish maintained on the 9L/15D-12°C Table 2.-The effects of pinealectomy on body lipid reserves in Notemigonus maintained on various photoperiod-temperature regimes during the prespawning season (21-day treatment). Dry lipid Gonosomatic index' index2 Treatment n Sex X±SE X±SE Initial controls 5 M 209 d: 10 1.96 ±0.17 (March) 9 F 184 di 12 3.42 ± 0.09 15.5L/8.5D photoperiod: 25°C: Sham operated 6 M 140 ± 11 1.35 ±0.07 8 F 131 ±5 2.41 ±0.15 Pinealectomized 6 M 112± 10* 0.80 ± 0.09 7 F 107 ±8* 1.58 ±0.22 12°C: Sham operated 6 M 232 ±8 2.00 ±0.10 7 F 244 ±7 3.41 ±0.26 Pinealectomized 6 M 251 ±6* 1.79 ±0.14 7 F 264 ± 7* 3.18 ±0.33 9L/15D photoperiod: 25°C: Sham operated 8 M 129 ±5 2.02 ±0.15 6 F 114±6 1.74 ±0.11 Pinealectomized 6 M 143 ±9 1.62 ±0.04 7 F 132 ±8* 2.03 ±0.13 12°C: Sham operated 5 M 243 ±5 1.31 ±0.05 5 F 267 ±8 3.79 ± 0.28 Pinealectomized 6 M 187±11* 1.20 ±0.07 4 F 240 ± 6* 3.43 ± 0.27 'Dry lipid index = mg lipid, less gonadal lipids/g dry body wt. ^Gonosomatic index = wt of gonads (g)/g body wt x 100. 'Significantly (P<0.05) different than sham operated controls maintained under same photoperiod-temperature regime. regime. At a low temperature, pinealectomy in- terferes with the complimentary effects of short photoperiods on lipid deposition. At d warm temperature, however, pinealectomy reverses the accentuating effects of short photoperiods on body lipid depletion. Interestingly, body lipid levels in pinealec- tomized females maintained on the 9L/15D-12°C regime did not differ significantly from fat levels in sham operated females exposed to the 15.5L/8.5D-12°C regime. Fat levels were not sig- nificantly different in pinealectomized females on the 15.5L/8.5D-12°C regime and sham operated females on the 9L/15D-12°C regime. Similarly, body fat levels in pinealectomized fish (both sexes) exposed to the 9L/15D-25°C regime did not differ significantly from lipid levels in shams maintained on the 15.5L/8.5D-25°C regime. These data in- dicate that pinealectomy reverses the effects that photoperiod has on lipid levels at a given temperature. Pinealectomy reverses the stimulatory effects of a long photoperiod-warm temperature regime, causing gonadal regression (de Vlaming 1975); under these conditions body lipid levels in pinealectomized animals are lower than in shams. Short photoperiods in combination with warm 769 FISHERY BULLETIN: VOL. 73, NO. 4 temperatures induce gonadal involution in Notemigonus. Pinealectomy under these condi- tions prevents gonadal regression, stimulating gonadal development and spaw^ning. Body lipid reserves were significantly greater in pinealec- tomized fish than in shams maintained on the 9L/15D-25°C regime. Under the other pho- toperiod-temperature regimes, pinealectomy resulted in changes in body fat reserves without appreciably altering reproductive activity. Possi- bly then, changes in lipid metabolism due to pinealectomy may influence reproductive activity, but apparently the effects of pinealectomy on fat deposition are not totally dependent on changes in sexual activity. Spawning Season The effects of pinealectomy on body fat stores were again examined during the early spawning season (late April, Table 3). In sham operated fish maintained at 25°C (both photoperiod groups), the lipid index was sig- nificantly lower than that of the initial controls. Body lipid reserves were also significantly deplet- ed in sham operated female fish exposed to the 9L/15D-25°C regime, but not in sham operated Table 3.-The effects of pinealectomy on body lipid reserves in Notemigonus maintained on various photoperiod-temperature regimes during the early spawning season (21-day treatment). Dry lipid Gonosomatic index' index2 Treatment n Sex X±SE 7±SE Initial controls 6 M 91 ± 2 3.10 ±0.22 (Late April) 6 F 88 ±3 6.53 ± 0.97 15.5L/8.5D photoperiod: 25°C: Sham operated 7 M 56 ±4 2.79 ± 0.25 7 87 ±6 5.13 ±1.04 Pinealectomized 5 M 45 ±5 2.10 ±0.09 8 62 ±4* 4.08 ± 0.73 15°C: Sham operated 5 M 91 ±5 2.16 ±0.09 6 114±6 5.88 ± 0.89 Pinealectomized 5 M 132 ±8* 2.38 ±0.14 6 129 ±6* 7.81 ± 1.22 9L/15D photoperiod: 25°C: Sham operated 5 M 68 ±5 1.16±0.15 9 74 ±6 2.87 ± 0.52 Pinealectomized 6 M 110±6* 2.67 ±0.16 5 101 ±7* 3.50 ± 0.88 15°C: Sham operated 6 M 155 ± 10 3.12 ±0.17 5 138 ±8 6.43 ±0.66 Pinealectomized 5 M 124 ±9* 3.01 ±0.15 7 117±5* 5.89 ± 0.57 'Dry lipid index = mg lipid, less gonadal lipids/g dry body wt. ^Gonosomatic index = wt of gonads (g)/g body wt x 100. •Significantly (P<0.05) different than sham operated controls maintained under same photoperiod-temperature regime. females maintained on the 15.5L/8.5D-25°C regime compared to the initial controls. Low temperature treatment during the spawning season caused a significant increase in body lipid content of sham operated female fish regardless of photoperiod. The short photoperiod-low tempera- ture regime stimulated fat deposition in sham operated males; however, body lipid levels did not differ significantly in the initial controls and sham operated males maintained on the long pho- toperiod-low temperature regime. At 25°C, body lipid stores were slightly higher in sham operated female fish maintained on a long photoperiod than in females exposed to a short photoperiod; the reverse was true for males. At 15°C, sham operated fish maintained on a short photoperiod were significantly fatter than sham operated fish on the long photoperiod regime. Lipid reserves were significantly lower in female pinealectomized than in sham operated female fish exposed to the 15.5L/8.5D-25°C regime. In fish (both sexes) maintained on the long photoperiod-low temperature regime, body lipid stores were significantly greater in pinealec- tomized than in sham operated control fish. Body fat content was significantly greater in both sexes of pinealectomized fish than in shams exposed to the 9L/15D-25°C regime. In contrast, lipid content in pinealectomized fish was significantly lower than in sham operated animals on the 9L/15D- 15°C regime. Body lipid reserves were not significantly different in pinealectomized female fish main- tained on the 9L/15D-15°C regime compared to the females exposed to the 15.5L/8.5D-15°C regime. Fat composition of pinealectomized females exposed to the long photoperiod-low temperature regime did not differ significantly from fat composition in sham operated females which experienced the short photoperiod-low temperature regime. No significant difference was observed in body fatness in pinealectomized females on the 9L/15D-25°C regime and sham operated females on the 15.5L/8.5D-25°C regime. Similarly, body lipid stores were approximately the same in pinealectomized females maintained on the long photoperiod-warm temperature regime and in sham operated females exposed to the short photoperiod-warm temperature regime. Many of the fish maintained on the 15.5L/8.D- 25°C regime spawned; under these conditions pinealectomy resulted in the initiation of gonadal regression. Compared to sham operated animals 770 de VLAMING: CONTROL OF FATTENING IN NOTEMIGONUS exposed to the long photoperiod, body lipid levels in pinealectomized fish were significantly lower. Gonadal regression was also initiated in sham operated fish maintained on the 9L/15D-25°C regime; spawning was observed in pinealec- tomized animals on this regime. Body lipid levels were significantly greater in pinealectomized fish than in sham operated fish maintained on the short photoperiod-warm temperature regime. In animals maintained at a low temperature (both photoperiods), GSIs did not differ significantly in the pinealectomized and sham operated groups. Sham operated fish, however, were significantly fatter than the pinealectomized animals experiencing the 9L/15D-15°C regime. Further- more, under the 15.5L/8.5D-15°C regime, pinealectomized fish contained significantly more fat than the sham operated controls. DISCUSSION Both temperature and photoperiod have a dis- tinct effect on body lipid reserves in N. crysoleucas. During the prespawning and spawn- ing seasons, low temperature treatment (12°- 15°C) favors increases in body fat stores in both sexes of Noteniigonus, regardless of photoperiod conditions. Low temperature treatment of Noteniigonus during the gonadal preparatory season did not result in lipid deposition, but did maintain body fat composition at a level equivalent to that in the initial controls (sacrificed at the onset of the experiment). Animals collected during the preparatory period (January) were very fat. In fact, body fat stores in this species reach a peak in late December, January, and early February. The failure of laboratory low tempera- ture treatment to stimulate lipid deposition dur- ing the preparatory season could be due to the presence of sufficient fat reserves in the initial controls. Regardless of season or photoperiod, high temperature (25°C) acclimation favors body lipid depletion in males. During the preparatory and prespawning seasons, body fat depletion is also observed in females exposed to warm tempera- tures. Compared to initial controls (animals sacrificed at the beginning of the experiment), lipid levels in Noteniigonus maintained at warm temperatures during the early spawning season were not appreciably altered. The failure of warm temperature treatment to deplete lipid reserves in females during the early spawning season may be due to the relatively low levels of fat in the initial controls. Fish maintained at warm temperatures were fed twice daily whereas fish exposed to low temperatures were fed only once a day; warm temperature animals consumed four to five times more food than the low temperature fish. Since the fish in all experiments were fed ad libitum, the differences in body lipid levels should not be due entirely to higher metabolic rates at elevated temperatures. Lipid synthesis and deposition is also promoted at low temperatures in several other teleost species (Blazka 1958; Brown 1960; Dean and Goodnight 1964; Knipprath and Mead 1968; de Vlaming and Pardo 1975). The means by which temperature acts to control lipid metabolism is not fully understood, but Kinne (1960) reported that the efficiency of food conversion in Cyprinodon macularis is maximal at lower temperatures. Furthermore, enzyme systems (Hochachka 1969) and hormones (de Vlaming and Pardo 1975; Pardo and de Vlaming in press) involved in lipid me- tabolism in fishes appear to be temperature sensi- tive. At low temperatures, short photoperiods are more effective than long photoperiods at stimulating lipid deposition in Noteniigonus. In all of the experiments reported here, body lipid reserves were higher in female fish exposed to the short photoperiod-low temperature regimes than in females maintained on long photoperiod-low temperature regimes. With the exception of the experiment conducted during the prespawning season, similar results were obtained with males. Short photoperiods also compliment the lipid depleting effects of warm temperatures. In two of the three experiments summarized here, body fat reserves were significantly lower in fish (both sexes) exposed to the short photoperiod-warm temperature regime than in animals maintained on the long photoperiod-warm temperature regime. Thus, in Noteniigonus, the effects of pho- toperiod on lipid metabolism are temperature dependent. Specifically, in combination with a low temperature, short photoperiods favor body fat deposition, but at a high temperature, short pho- toperiods accelerate depletion of lipid reserves. The fact that short photoperiods have opposing effects on lipid metabolism depending on temperature is, at the present time, an enigma. Apparently, however, changing environmental temperatures can differentially sensitize Noteniigonus to daylength. Roberts (1964) showed that photoperiod changes can alter metabolic pat- 771 FISHERY BULLETIN: VOL. 73, NO. 4 terns in sunfish. In F. similis, short photoperiods promote fattening whereas long photoperiods result in lipid depletion (de Vlaming, Sage, Charl- ton, and Tiegs; de Vlaming et al. in press). Other than these studies, little is known concerning the potential role of daylength in controlling fatten- ing cycles in teleosts. The results presented here on temperature and photoperiod effects on fat storage are consistent with environmental data. Lipid levels are lowest in Notemigoniis collected during July, August and early September; environmental temperatures are high during this time and daylength is decreasing. From mid-September through December daylength and temperature continue to decrease. Fat stores increase progressively during this time. Beginning in mid or late February, lipid reserves are progressively depleted until late June or July. During this time, daylength and temperature are increasing. A progressive decrease in body fat reserves was observed in Notemigonus collected during the preparatory, prespawning and spawning seasons respectively. These data indicate that there may be a relationship between lipid stores and reproduction. Other investigators presented evidence of body fat depletion associated with increasing gonadal activity (Lovern 1934; Liih- mann 1953; Idler and Bitners 1960; Woodhead 1960; Wilkins 1967; Lasker 1970). A long photo- period, in combination with warm temperatures, is required for final gonadal maturation and spawn- ing in Notemigonus (de Vlaming 1975). These conditions result in depletion of fat stores in this species. The fat depleted from body storage sites could possibly be utilized for the energy demands of reproduction; in females, some of the body lipids may also be converted to yolk precursors and transported to the developing oocytes. So gonadal activation by long photoperiod-warm tempera- tures regimes may result in mobilization of body lipid reserves. Possibly, however, gonadal ma- turation may depend on the prior activation of lipid mobilization enzyme systems by long pho- toperiod-warm temperature regimes. The former hypothesis gains some support from observations of several investigators (e.g., Kobayashi 1953; Egami 1955; Oguro 1956) which indicate that sex steroids stimulate lipid synthesis in fishes. In vitro studies with Notemigonus liver preparations imply that estradiol-17^ stimulates synthesis and transport of lipid by this tissue (Shing and de Vlaming unpubl. data). My intention is not to imply that only sex steroids are involved in regulation of lipid metabolism. Indeed, other hor- mones such as insulin (de Vlaming and Pardo 1975) and prolactin (see below) have distinct effects on fat metabolism in Notemigonus. Low temperatures, regardless of photoperiod, maintain vitellogenesis and spermatocyte proliferation in Notemigonus, but will not stimulate final ovarian or testicular maturation (de Vlaming 1975). Low temperatures also main- tain or increase body lipid reserves in this species. These observations lend further support to the suggestion that fat stores are in some way related to reproductive activity. Short photoperiod-warm temperature regimes cause gonadal regression in Notemigonus (de Vlaming 1975). Body fat depletion also occurs under these conditions. Obviously this fat deple- tion is not associated with increased gametogenic activity. Body lipid reserves also decreased in fish maintained on long photoperiod-warm tempera- ture regimes. Therefore, depletion of body fats may be primarily associated with increased energy requirements at high temperatures. In Notemigonus, however, there is some indication that sex steroid secretion is stimulated or remains high in fish maintained on a short photoperiod- warm temperature regime. Specifically, fish ex- posed to these conditions either develop or main- tain nuptial coloration. If such is the case, sex steroids may be involved in mobilization and/or utilization of lipid reserves. In a majority of the experiments reported here, pinealectomy had a pronounced effect on body fat reserves in Notemigonus. The effects of pinealec- tomy on fat metabolism depend on the pho- toperiod-temperature regime to which the experimental animals are exposed. In all three experiments, body lipid levels were significantly lower in pinealectomized than in sham operated females exposed to the long photoperiod-warm temperature regime; similar results were obtained with males in one experiment. Body lipid content was significantly greater in pinealectomized than in sham operated females in two of the three experiments where fish were exposed to a short photoperiod-warm temperature regime; similar results were obtained in only one experiment with males. These data indicate that, in fish maintained at warm temperatures, the effects of pinealec- tomy depend on photoperiod conditions. During the prespawning and spawning seasons, pinealec- tomy reversed the effects of photoperiod on fish 772 de VLAMING: CONTROL OF FATTENING IN NOTEMIGONUS exposed to a warm temperature. For example, lipid levels were not significantly different in pinealectomized fish exposed to the short pho- toperiod-warm temperature regime and sham operated animals maintained on the long pho- toperiod-warm temperature regime; nor was fat "content significantly different in sham operated animals exposed to the short photoperiod-warm temperature condition and pinealectomized fish maintained on the long photoperiod-warm temperature regime (Tables 2, 3). In all three of the experiments summarized here, body fat reserves were significantly greater in pinealectomized than in sham operated females maintained on the long photoperiod-low tempera- ture regime; similar results were recorded in two of the experiments with males. Body fat composi- tion was significantly lower in pinealectomized than in sham operated females exposed to a short photoperiod-low temperature regime; similar differences were noted with males in two of the experiments. These data further confirm the suggestion that the effects of pinealectomy on Hpid metabolism in Notemigonus depend on pho- toperiod. In two of the experiments, fat content did not differ significantly in pinealectomized fish maintained on the long photoperiod-low tempera- ture regime and sham operated animals exposed to the short photoperiod-low temperature regime; nor were significant differences noted in fat levels when the reverse comparison was made (Tables 2, 3). The data obtained at both high and low temperatures indicate that the pineal in Notemigonus may have some role in receiving and/or integrating light information. Such a suggestion seems likely since the pineal may either facilitate or retard lipid deposition in this species. Several morphological and elec- trophysiological studies suggest that the pineal in some teleosts functions as a photoreceptor (cf. de Vlaming 1974). Light microscope studies on the pineal of Notemigonus also indicate a sensory function (Vodicnik and de Vlaming unpubl. data). If the pineal in Notemigonus is a photoreceptor involved by some means in measuring daylength, then removal of this organ from fish maintained under different photoregimes might be expected to have variable effects on lipid metabolism. Urasaki (1972a, b) has also shown that the effects of pinealectomy on reproductive function in Oryzias latipes vary with photoperiod conditions. The effects of the pineal on lipid metabolism in Notemigonus, however, do not depend entirely on light information. For example, in fish maintained on a long photoperiod during the prespawning and spawning seasons, pinealectomy accentuated lipid deposition at low temperatures and lipid depletion at a high temperature. Temperature may not, however, act on the pineal directly. High temperatures cause lipid catabolism and low temperatures favor lipid deposition. Temperature may act directly on lipid metabolism enzyme sys- tems or indirectly to stimulate hormone secretion from various endocrine glands. In Notemigonus, light information serves only to modify the effects of temperature on lipid metabolism. Thus, the pineal could function at all temperatures as a light receptor and/or integrator. Light information may be differentially interpreted (at different temperatures or at different times of the year) at some other level such as the hypothalamus and/or pituitary. Whether the pineal in Notemigonus exerts its effects on lipid metabolism via neural or hormonal pathways is not presently known. Most morphological studies on the teleost pineal have stressed the dual sensory and secretory ap- pearance of this organ (cf. de Vlaming 1974). A dual sensory-secretory function is also indicated by light microscope studies on the pineal of Notemigonus. Histochemical and biochemical data show that the teleost pineal has an active in- dolamine metabolism (Quay 1965; Hafeez and Quay 1969; Fenwick 1970; Owman and Riideberg 1970). Melatonin has inhibitory effects on reproductive function in various teleosts (Fenwick 1970; Urasaki 1972c; de Vlaming, Sage, and Charl- ton 1974). Melatonin treatment decreases lipid reserves in F. similis maintained on a long pho- toperiod-low temperature regime (de Vlaming, Sage, Charlton, and Tiegs 1974); if melatonin acts as the mediator of pineal action in Notemigonus, one might then expect pinealectomy to increase fat levels in fish exposed to a long photoperiod-low temperature regime. Pinealectomy did indeed have these results under this regime. During July, melatonin therapy of F. similis exposed to a short photoperiod-low temperature regime stimulated lipid deposition (de Vlaming, Sage, Charlton, and Tiegs 1974). If the mediator of pineal activity in Notemigonus is melatonin, one might predict that pinealectomy would decrease body lipid stores in animals exposed to a short photoperiod-low temperature regime. Such results were observed in the experiments reported here. Interestingly, 773 lighting conditions were reported to alter the secretory activity in the glandular appearing pineals of Gambusia affinis and Sijmphodus melops (Cheze and Lahaye 1969; Cheze 1970). Pos- sibly then, the pineal in some teleost species may function as a neuroendocrine transducer of pho- toperiod information. If the pineal is a neuroen- docrine organ, melatonin could conceivably be one of the hormones produced by this organ. Although the pineal of Notemigonus does seem to be in- volved in photoreception, available evidence does not allow one to conclusively state that this organ is neuroendocrine in nature or that pineal- produced melatonin functions as a chemical mes- senger. The pineal may modify fat stores in Notemigonus by influencing hypothalamic and/or pituitary function. Indeed, de Vlaming and Vodicnik (in press) showed that pinealectomy alters hypothalamic gonadotropin releasing ac- tivity and pituitary gonadotropin levels. Several investigators reported that prolactin has a pronounced affect on lipid metabolism in teleost fishes (Lee and Meier 1967; Meier 1969; Mehrle and Fleming 1970; Joseph and Meier 1971; Meier et al. 1971; de Vlaming and Sage 1972; Sage and de Vlaming 1973; de Vlaming et al. in press; Pardo and de Vlaming in press). Furthermore, melatonin treatment significantly reduces pituitary prolac- tin activity in F. similis (de Vlaming, Sage, Charlton, and Tiegs 1974). These authors suggest- ed that the effects of melatonin on lipid me- tabolism in this species may be due in part to the effects of this indolamine on pituitary prolactin release. Whether pinealectomy alters pituitary prolactin secretion is not presently known. Inves- tigations are presently in progress to examine this possibility. Prolactin does stimulate lipid deple- tion from in vitro liver preparations of Notemigonus incubated at high temperatures and promotes fat synthesis in liver preparations in- cubated at low temperatures (Pardo and de Vlam- ing in press). The effects of pinealectomy on lipid reserves in Notemigonus may result from changes in reproductive activity. This suggestion seems rather unlikely since pinealectomy frequently resulted in significant changes in body fat levels without appreciably altering gonadal activity. The data presented here favor the view that the pineal is a photoreceptor or integrates light infor- mation and plays an important role in regulating physiological processes in teleost fishes. FISHERY BULLETIN: VOL. 73, NO. 4 ACKNOWLEDGMENTS I am grateful to J. Flanagan, M. J. Vodicnik and R. J. Pardo for technical assistance. This work was supported by NSF Grant GB-41338. LITERATURE CITED Bergmann, G. 1971. Electron microscopic studies of the pineal organ in Pteraphyllum scalare Cuv. et Val. (Cichlidae, Teleos- tei). Z. Zellforsch. Mikrosk. Anat. 119:257-288. Blazka, p. 1958. The anaerobic metabolism of fish. Physiol. Zool. 31:117-128. Brown, W. D. 1960. Glucose metabolism in carp. J. Cell. Physiol. 55:81-85. Cheze, G. 1970. fetude morphologique, histologique, et experimentale de I'epiphyse de Symphodus melops (Poisson, Labridae). Bull. Soc. Zool. Fr. 94:697-704. Cheze, G., and J. Lahaye. 1969. Etude morphologique de la region epiphysaire de Gambusia affinis holbrooki G. Incidences histologiques de certains facteurs externes sur le toit diencephalique. Ann. Endocrinol. 30:45-53. Dean, J. M., and C. J. Goodnight. 1964. A comparative study of carbohydrate metabolism in fish as affected by temperature and exercise. Physiol. Zool. 37:280-299. de Vlaming, V. L. 1971. The effects of food deprivation and salinity changes on reproductive function in the estuarine gobiid fish, Gillichthys mirabilis. Biol. Bull. (Woods Hole) 141:458-471. 1972. Environmental control of teleost reproductive cycles: A brief review. J. Fish Biol. 4:131-140. 1974. Environmental and endocrine control of teleost reproduction. In C. B. Schreck, Control of sex in fishes, p. 13-83. Sea Grant and V.P.I. & S.U. Press, V.P.I.- S.G.-74-01. 1975. Effects of pinealectomy on gonadal activity in the cyprinid teleost, Notemigonus crysoleucas. Gen. Comp. Endocrinol. 26:36-49. DE Vlaming, V. L., and R. J. Pardo. 1975. In vitro effects of insulin on liver carbohydrate and lipid metabolism in the cyprinid teleost, Notemigonus crysoleucas. Comp. Biochem. Physiol. 51B:489-497. de Vlaming, V. L., and M. Sage. 1972. Diurnal variation in fattening response to prolactin treatment in two Cyprinodontid fishes, Cyprinodon variegatus and Fundulus similis. Contrib. Mar. Sci., Univ. Tex. 16:59-63. DE Vlaming, V. L., M. Sage, and C. B. Charlton. 1974. The effects of melatonin treatment on gonosomatic index in the teleost, Fundulus similis, and the tree frog, Hyla cinerea. Gen. Comp. Endrocrinol. 22:433-438. DE Vlaming, V. L., M. Sage, C. B. Charlton, and B. Tiegs. 1974. The effects of melatonin on lipid deposition in cyprinodontid fishes and on pituitary prolactin activity in Fundulus similis. J. Comp. Physiol. 94:309-319. 774 de VLAMING: CONTROL OF FATTENING IN NOTEMIGONUS DE Vlaming, V. L., M. Sage, and R. Tiegs. In press. A diurnal rhythm of pituitary prolactin activity with diurnal effects of mammalian and teleostean prolactin on total body lipid deposition and liver lipid metabolism in teleost fishes. J. Fish Biol. DE Vlaming, V. L., and M. J. Vodicnik. In press. Effects of pinealectomy on pituitary gonadotropin potency and hypothalamic gonadotropin releasing ac- tivity in the cyprinid teleost, Notemigonus crysoleucas. J. Fish Biol. Egami, N. 1955. Effect of estrogen and androgen on the weight and structure of the liver of the fish, Oryzias latipes. Annot. Zool. Jap. 28:79-85. Fenwick, J. C. 1970. Demonstration and effect of melatonin in fish. Gen. Comp. Endocrinol. 14:86-97. Hafeez, M. a. 1971. Light microscopic studies on the pineal organ in teleost fishes with special regard to its function. J. Morphol. 134:281-314. Hafeez, M. A., and W. B. Quay. 1969. Histochemical and experimental studies of 5- hydroxytryptamine in pineal organs of teleosts {Salmo gairdneri and Atherinapsis calif orniensis). Gen. Comp. Endocrinol. 13:211-217. 1970. Pineal acetylserotonin methyltransferase activity in the teleost fishes, Hesperoleucas symmetricus and Salmo gairdneri with evidence for lack of effect of constant light and darkness. Comp. Gen. Pharmacol. 1:257-262. HOCHACHKA, p. W. 1969. Intermediary metabolism in fishes, /n W.S. Hoar and D. J. Randall, (editors). Fish physiology, Vol. 1, p. 351- 389. Academic Press, N.Y. Idler, D. R., and I. Bitners. 1960. Biochemical studies on sockeye salmon during spawn- ing migration. IX. Fat, protein and water in the major internal organs and cholesterol in the liver and gonads of the standard fish. J. Fish. Res. Board Can. 17:113-122. Joseph, M. M., and A. H. Meier. 1971. Daily variations in the fattening response to prolactin in Fundulus grandis held on different photoperiods. J. Exp. Zool. 178:59-62. Kinne, 0. 1960. Growth, food intake, and food conversion in a euryplastic fish exposed to different temperatures and salinities. Physiol. Zool. 33:288-317. Knipprath, W. G., and J. F. Mead. 1968. The effect of the environmental temperature on the fatty acid composition and on the in vivo incorporation of l-"C-acetate in goldfish (Carassius auratus L.). Lipids 3:121-128. KOBAYASHl, H. 1953. Effects of estrone upon the structure, weight and fat content of the liver in the fish, Misgurnus an- guillicaudatus. Annot. Zool. Jap. 26:213-216. Lasker, R. 1970. Utilization of zooplankton energy by a Pacific sardine population in the California current. In J. H. Steele (editor), Marine food chains, p. 265-284. Univ. Calif. Press, Berkeley. Lee, R. W., and A. H. Meier. 1967. Diurnal variations of the fattening response to prolactin in the golden topminnow, Fundulus chryso- tus. J. Exp. Zool. 166:307-316. Love, R. M. 1957. The biochemical composition of fish. In M. E. Brown (editor), The physiology of fishes. Vol. I, Metabolism, p. 401-418. Academic Press, N.Y. LOVERN, J. A. 1934. Fat metabolism in fishes. IV. Mobilisation of depot fat in the salmon. Biochem. J. 28:1955-1960. LUHMANN, M. 1953. Uber die Fettspeicherung bei Ostseeheringen und ihre Beziehung zum Fortpflanzungszyklus. Kiel Meeres- forsch. 9:213-227. Mehrle, P. M., AND W. R. Fleming. 1970. The effect of early and midday prolactin injection on the lipid content of Fundulus kansae held on a constant photoperiod. Comp. Biochem. Physiol. 36:597-603. Meier, A. H. 1969. Diurnal variations of metabolic responses to prolactin in lower vertebrates. Gen. Comp. Endocrinol. Suppl. 2:55-62. Meier, A. H., T. N. Trobec, M. M. Joseph, and T. M. John. 1971. Temporal synergism of prolactin and adrenal steroids in the regulation of fat stores. Proc. Soc. Exp. Biol. Med. 137:408-415. NiKOLSKY, G. V. 1963. The ecology of fishes. Academic Press, N.Y. Oguro, C. 1956. Some observations on the effect of estrogen upon the liver of the three-spined stickleback, Gasterosteus aculeatus aculeatus L. Annot. Zool. Jap 29:19-22. Oksche, a. M., M. Ueck, and C. ROdeberg. 1971. Comparative ultrastructural studies of sensory and secretory elements in pineal organs. Mem. Soc. En- docrinol. 19:7-25. Omura, Y., and M. Oguri. 1969. Histological studies on the pineal organ of 15 species of teleosts. Bull. Jap. Soc. Sci. Fish. 35:991-1000. OWMAN, C, and C. RiJDEBERG. 1970. Light, fluorescence, and electron microscopic studies on the pineal organ of the pike, Esox lucius L., with special regard to 5-hydroxytryptamine. Z. Zellforsch. Mikrosk. Anat. 107:522-550. PaRDO, R. J., AND V. L. DE VLAMING. In press. In vivo and in vitro effects of prolactin on lipid metabolism in the cyprinid teleost, NotemigonUs crysoleucas. Copeia. QUAY.W.B. 1965. Retinal and pineal hydroxyindole-0-methyl transferase activity in vertebrates. Life Sci. 4:983- 991. Reiter, R. J. 1973. Comparative physiology: Pineal gland. Ann. Rev. Physiol. 35:305-328. RlZKALLA, W. 1970. The morphology of the pineal organ in the teleost, Clarias lazera C. V. Acta Biol. Acad. Sci. Hung. 21:25- 35. Roberts, J. L. 1964. Metabolic responses of fresh-water sunfish to seasonal photoperiods and temperatures. Helgolander wiss. Meeresunters. 9:459-473. 775 FISHERY BULLETIN: VOL. 73, NO. 4 RUDEBERG, C. 1966. Electronmicroscopic observations on the pineal organ of the teleosts Mugil auratus (Risso) and Uranoscopus scaber. Pubbi. Stn. Zool. Napoli 35:47-60. 1969. Structure of the parapineai organ of the adult rainbow trout, Salmo gairdneri Richardson. Z. Zellforsch. Mikrosk. Anat. 93:282-304. Sage, M., and V. L. de Vlaming. 1973. A diurnal rhythm in hormone effectiveness and in pituitary content of prolactin in Fundulus similis. Tex. Rep. Biol. Med. 31:101-102. Takahashi, H. 1969. Light and electron microscope studies on the pineal organ of the goldfish, Carassius auratus L. Bull. Fac. Fish., Hokkaido Univ. 20:143-157. Urasaki, H. 1972a. Effect of pinealectomy on gonadal development in the Japanese killifish (Medaka), Oryzias latipea. Annot. Zool. Jap. 45:10-15. 1972b. Role of the pineal gland in gonadal development in the fish, Orzyiax latipea. Annot. Zool. Jap. 45:152-158. 1972c. Effects of restricted photoperiod and melatonin ad- ministration on gonadal weight in Japanese killifish. J. Endocrinol. 55:619-620. WiLKINS, N. P. 1967. Starvation of the herring, Clupea harengus L.: Sur- vival and some gross biochemical changes. Comp. Biochem. Physiol. 23:503-518. WOODHEAD, A. D. 1960. Nutrition and reproductive capacity in fish. Proc. Nutr. Soc. 19:23-28. 776 RELATIONSHIPS BETWEEN ZOOPLANKTON DISPLACEMENT VOLUME, WET WEIGHT, DRY WEIGHT, AND CARBON^ Peter H. Wiebe,^ Steven Boyd,- and James L. Cox^ ABSTRACT Interconversion of various measures of zooplankton biomass have great utility in studies requiring nondestructive techniques, or for interpretation of past data. In establishing predictive relationships between such measures, the appropriate regression to use is the geometric mean estimate, which provides a regression line in which the regressions of .Y on Y and Y on X are identical. We have employed this type of analysis in determinations on samples from diverse sea areas in different seasons and have determined that statistically significant relationships exist between carbon, wet weight, displacement volume, and dry weight when a constant technique is used. The slope of the regression line for log transformed values for carbon vs. dry weight and wet weight vs. displacement volume was sufficiently close to unity to assume a straight percentage conversion between these values. Carbon was 31-33% of dry weight and wet weight was 72-73% of displacement volume, according to our techniques. Comparability of different techniques for a biomass measurement may be poor, especially in the case of displacement volume and wet weight measurements due to variations in the interstitial water content. Moreover, interstitial water content varies inversely with total biomass density, which accounts for the absence of a simple percentage relationship between wet weight or displacement volume and other measures of zooplankton biomass. Biomass is a classic and useful measure of the zooplankton standing crop. A number of tech- niques exist to measure it. Four commonly used techniques involve measurement of displacement volume (Yentsch and Hebard 1957; Frolander 1957; Sutcliffe 1957; Tranter 1960; Ahlstrom and Thrailkill 1963), wet weight (Nakai and Honjo 1962), dry vi^eight (Lovegrove 1966), and carbon (Curl 1962; Piatt et al. 1969). For most studies, especially those determining energy flow through food chains, carbon is the most fundamental of these gross measures. Many zooplankton collec- tions frequently serve several purposes and the destructive techniques required to determine car- bon or dry weight frequently cannot be employed. An alternative is to measure displacement volume or wet weight, and convert the data into either dry weight or carbon. These latter techniques, if done properly, are nondestructive since the organisms can still be identified when re-suspended in liquid. There is an obvious need for conversion factors that reliably define the relationship between the various biomass measures. This need also arises 'Contribution No. 3433 from the Woods Hole Oceanographic Institution, Woods Hole, MA 02543. This study was supported by NSF GA 29303, ONR N00014-66-CO-241, and NRO 83-004. ^Woods Hole Oceanographic Institution, Woods Hole, MA 02747. 'Southeastern Massachusetts University, North Dartmouth, MA 02747. when data based on different techniques are com- pared. Although conversion factors exist in the literature, they often are based on data from re- stricted sea areas. Further, in some cases, biomass determinations were made by techniques which are no longer recommended (see Lovegrove 1966). The objective of this paper is to more satisfactorily define the relationships between the biomass measures mentioned above. Using both data derived from samples collected from diverse oceanic areas over the past 6 yr and data selected from the literature, we have empirically deter- mined linear regression equations relating pairs of biomass measures. THEORETICAL CONSIDERATIONS Ideally, any two biomass measures, X and Y should be related by a constant of proportionality, a, such that Y=aXP, (1) Manuscript accepted March 1975. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. where fi = 1.0. A measurement error or bias which occurs as a constant fraction of the biomass results only in a change in the value of a. When natural variability or an error factor(s) in X or Y is disproportionate or inversely proportional to the amount of biomass, j8 cannot be assumed to equal 777 FISHERY BULLETIN: VOL. 73, NO. 4 1.0 and Y is not a simple percentage of X If /3 is a constant, then the log ^q of the two measures will be linearly related: Log io(iO = Logio(a) + /3Log io(X). (2) In the sections which follow, we will show that in most cases j8 9^ 1.0 and equation (2) is adequate for describing the linear relationship between log transformed measures of biomass. METHODS The station locations where samples were collected are shown in Figure 1 (symbol key given in Table 1). At many of these stations, more than one sample was collected. A single symbol may represent a number of collections as indicated in Table 1. Not shown are the stations of Gosnold 140, a cruise to the coastal upwelling region off Peru. Collections were made with 70-cm or 100-cm diameter ring nets, 70-cm diameter Bongo nets (McGowan and Brown 1966), or the 50x 50 cm net (Be et al. 1971), all equipped with a flowmeter (Table 1). In shallow regions, Buzzards Bay, Atlantis II 52 (continental Atlantic shelf), Gosnold 140 (Peru Current), Gosnold 166 (New York Bight), tows were made to near the bottom. In deeper waters, oblique tows were generally made to below 300 m. Ring net collections were generally split with a plankton splitter (McEwen et al. 1954). One-half was preserved in 10% buffered Formalin^ for displacement volume analysis and the other half was frozen in a chest freezer for wet weight, dry weight, and carbon analyses (re Bermuda Table 1: Menzel and Ryther (1961) do not state how this half was stored prior to analysis). A similar procedure was carried out for Bongo net collec- tions; one of the paired samples was preserved in Formalin while the other was frozen. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 100* 9Cf .,„„„„.. 80- 70- 60- Figure 1. -Location of zooplankton collection sites. For symbols, see Table 1. 50' 778 WIEBE ET AL.: RELATION OF VOLUME, WET AND DRY WEIGHTS, AND CARBON Table 1. -Number of observations and symbol designation for each cruise or area from which zooplankton samples were collected. The symbols are used in Figures 1 and 3-5. Number of observations Displacement Wet Dry Cruise or area Symbol Date(s) volume vi^eight weight Carbon Type of net (mesh) Buzzards Bay O Jan. -June 1972 15 16 16 16 70 cm (240//,m) diam. Slope A June-Aug. 1972 14 12 14 14 100 cm (333 /im) diam. Bermuda' ^ 1957-59 52 0 52 0 100 cm (366)U,m) diam. Gosnold 140 X May 1969 20 0 33 33 70 cm (240/U,m) diam. Gosnold 166 O June 1970 32 0 33 33 70 cm (240 /im) diam. Atlantis II 48 ^ Nov. 1968 0 0 20 20 70 cm (240/U.m) diam. Atlantis II 52 ^ Sept. 1969 0 0 37 37 70-cm (240/im) diam. Bongos Atlantis II 71 + Sept. 1972 27 43 43 42 100 cm (333 /im) diam. Chain 111 □ Feb. 1973 13 13 13 0 100 cm (333)U.m) 70-cm Bongos Knorr35 II X Nov. 1973 10 11 11 0 100 cm (333)Um) 70-cm Bongos Be North Atlantic O 3 229 229 229 0 50 X 50 cm (202|tim) Be South Atlantic + '' 176 192 193 0 50 X 50 cm (202;im) iData from Menzel and Ryther (1961). ^Omitted, bad data. 3See Be et al (1971) for geographical and seasonal coverage. "•Data from Be (footnote 5). Displacement volumes were measured by one of two techniques. The Mercury Immersion method of Yentsch and Hebard (1957) was used to deter- mine the values given by Menzel and Ryther (1961-Bermuda), by Be et al. (1971-North Atlantic) and Be (1973-unpubl. data for the South Atlantic).^ A modified version of the Mercury Immersion technique was used to measure displacement volumes on Gosnold 140, but further work has shown that the method has significant variable errors and is unreliable (Grice and Wiebe unpubl. data). We have not, therefore, used the Gosnold 140 displacement volume data. All other displacement volumes were measured by the method described by Ahlstrom ami Thrailkill (1963) after removal of all organisms larger than 5 cc. For split samples, organisms larger than 5 cc were removed prior to the split. On Gosnold 166, displacement volumes were run prior to sample preservation (see Vaccaro et al. 1972 for data) and again 2 yr after preservation. Contrary to the findings of Ahlstrom and Thrailkill (1963), shrinkage did not occur (Figure 2). These samples were, however, heavily dominated by copepods (Wiebe et al. 1973) which are least likely to under- go shrinking. Wet weight was measured by straining the plankton through a 333-/i,m plankton gauze, rins- ing with freshwater, and blotting the remaining mass on absorbant paper towels until water was no longer absorbed onto the towel. The biomass was then transferred to a pre-weighed glass jar with a stainless steel spatula. The jar was weighed on a Mettler balance to ± 2.5 mg and the wet weight of plankton determined by subtracting the jar weight from the total. Each jar was then dried to constant weight in an oven at 60°C. This 15-1 13- 0^11 Ct ^ 7- 3- 1- X A =.055 [-'•)'■ r^^^ -"^-■f- 9 '■V ^'v 23 X '■;'';'-■ Xv^v^v fev -llVf.l'f; ikp^ - ""'x ^ V^'x 'v- ■ -/ -,' -l' ';■.■''■ '-V^-V'-> :V# ^ - s'^'v ^ ^', .'.■ "'V~*','*V' ; .■ -f*.* ■,\v^'^^x,\- , v' '' \ ^ V ^ 'v' -•' '^V' ": : ••■■.• ;y- : . - , ■ -if^: j:;; ■, ,.V,-/^ :'y-.A^f4( . . . c.^'.'.^V 1 -4 ~£ ) .( ) .£ > A 1 .6 ^B€, A. W. H. 1973. Studies of zooplankton standing stock in the South Atlantic. Unpubl. final tech. rep. to Natl. Sci. Found., 14 p. DIFFERENCE BETWEEN 2 YEAR OLD DISPLACEMENT VOLUME 8 LIVE VOLUME (cc/m^j Figure 2.— Distribution of differences between displacement volumes measured 2 yr after preservation and the live displacement volume. The live displacement volumes ranged from 30 to 321 cc (0.57 to 2.53 cc/m^) and the 2-yr values ranged from 38 to 340 cc (0.58-2.40 cc/m'). 779 FISHERY BULLETIN: VOL. 73, NO. 4 frequently took 2 wk or longer owing to the large volumes of plankton collected. Dried samples were pulverized and an aliquot(s) of the powder used to determined carbon in either a Perkin-Elmer No. 240 or a Hewlett Packard No. 185 B carbon, hydrogen, nitrogen analyzer. A number of exceptions to this procedure are evident in Table 1. In some cases, wet weight was not measured; in others, carbon was not determined. All data presented below were standardized to biomass per cubic meter and then logarithmically transformed (base 10) before use in the regression analyses. RESULTS Several regression lines can be used to express the relationship between pairs of variables (Ricker 1973). The appropriate one is determined by the frequency distribution of the parent population as well as the nature of the error sources in the measurements (natural or measurement error). Since the biomass measures are all subject to na- tural variability and measurement error and since the observations presented cannot be assumed to be a random sample from a bivariate normal population, the "geometric mean (GM) estimate of the functional regression of Y on X" (Ricker 1973:412) is appropriate. As Ricker points out, this regression line minimizes the sum of the products of the vertical and horizontal distance of each point from the line. Thus, the GM regression lines of F on X and XonY are identical. Given the GM regression equation: where Y' = log (Y) and X' = log {X), one can determine both an X given Y or Y given X. Although Ricker's (1973) paper should be consulted for an in depth discussion of the assumptions and computations, we note that the slope, v, is given by: V = r V 2 nij'-Y)'' ^{x'.-xy (4) where b is the slope of the predictive regression of Y' on X' and r is the correlation coefficient. The Y'-axis intercept, u, is easily determined by: = Y'- vX'. (5) Y' = u + vX', (3) Plots of the values used in the GM regressions are given in Figures 3-5. The equations are listed in Table 2. All equations have slopes significantly different from zero (P<0.001). As indicated above, in the case where ^of Equation (1) {v of Equation (3)) is equal to 1.0, one biomass measure is a straight percentage of another. The only regres- sions with a V approaching 1.0 compare dry weight to carbon and displacement volume to wet weight. In these cases, predicted carbon varies from 31 to 33% of zooplankton dry weight and predicted wet weight varies from 72 to 73% of displacement volume. In all other regressions, a variable bias is present which causes v to deviate from 1.0. We believe that a large portion of the bias is caused by the interstitial water present in displacement volumes and wet weights. This bias is inversely proportional to the sample size; i.e., a small sample Table 2.— Functional (geometric mean) regression equations for pairs of biomass measures. Carbon: C; dry weight: DW; wet weight: WW; displacement volume: DV; Be et al. (1971) and Be (footnote 5) wet weight: BWW; Be et al. (1971) and Be (footnote 5) displacement volume: BDV; Be et al. (1971) and Be (footnote 5) dry weight: BDW; Piatt et al. (1969) dry weight: PDW; Piatt et al. (1969) carbon: PC. Logarithms to the base 10. Equation Regression equation a; Variance of slope r^ 1 Log (DV) = -1.429 + 0.808 Log(C) 87 0,0003187 0.96 2 Log (WW) = -1.537 + 0.822 Log(C) 70 0.0008303 0.92 3 Log(DW) = 0.508 + 0.977 Log(C) 193 0.0001438 0.97 4 Log(DV) = -1.828 + 0.848 Log(DW) 161 0.0001814 0.96 5 Log (WW) = -1.983 + 0.922 Log(DW) 93 0.0005800 0.94 6 Log(DV) = 0.670 + 0.950 Log(WW) 77 0.0013729 0.90 7 Log(BWW) = -1.897 + 0.835 Log(BDW) 420 0.0009725 0.63 8 Log(BDV) = -1.826 + 0.754 Log(BDW) 404 0.0011106 0.56 9 Log(BDV) = -0.219 + 0.848 Log(BWW) 403 0.0006079 0.75 10 Log(PDW) = 0.558 + 1.024 Log(PC) 45 0.0148049 0.39 11 Log(DV)i = 1.048 + 0.821 Log(DW) 110 0.0010510 0.83 12 Log(WW)i = 0.975 + 0.946 Log(DW) 94 0.0010173 0.90 13 Log(DV)i = 0.078 + 1.026 Log (WW) 75 0.0022271 0.85 'Note that biomass data used to determine equations 1-10 were standardized to per cubic meter while the data used to dete>mine equations 11-13 were not standardized. 780 WIEBE ET AL.: RELATION OF VOLUME, WET AND DRY WEIGHTS, AND CARBON I 0.01 0.001 ' 1 — H 1 I I I I III 1 1 I I 0.1 1- 10. CARBON (mg/m^) 100. i I I 10. 1.:: 0.1 0.01 1^ K I I I I I III H 1 — I I I I I II -I 1 I I I I III 0.1 1. 10. 100. CARBON (mg/m3) 1000. 1000. T 100. I 10. 1. :: 0.1 Figure 3.-Plots of data used in cal- culating geometric mean regression lines relating dry weight, wet weight, and displacement volume to carbon. For symbols, see Table 1. 0.01 H 1 — I I I I I II 1 1 — I I I I I II 1 1 — I I I I I II 1 1 — I I I I I II 0.1 1. 10. 100. 1000. CARBON (mg/m^) appears to have a larger percentage of interstitial water than a large sample. It is evident in our log transformed raw data as well as the data stan- dardized to biomass per cubic meter and then log transformed (see Table 4). The reason why the bias is not significantly influenced by the standardiza- tion to biomass per cubic meter results from the fact that the volume of water filtered in collecting most samples was quite similar, between 100 and 1,000 m^, while the biomass per cubic meter varied by as much as four orders of magnitude. As a result of the variable bias, it is not valid to assume a simple percentage relationship between the other pairs of biomass estimators. For example, dry weight is approximately 5% of displacement volume for low biomass per cubic meter and approximately 13% for high biomass per cubic meter. 781 FISHERY BULLETIN: VOL. 73, NO. 4 I 5:: 10. 1.:: 0.1 O01;; 0.001 ^ o H — I I I I I II -I — I I II III 0.1 10. 100. DRY WEIGHT (mg/m^) lOOO k. k. 1. T 0.1 :: O01 :: 0001 H — I I I I I II H 1 — I I I I I II 1. 10. DRY WEIGHT (mq/rrr>) 100. 10. T 0.01 0.1 1. WET WEIGHT (g/m^) 10. Figure 4. -Plots of data used in calculating geometric mean regression lines relating wet weight and displacement volume to dry weight and displacement volume to wet weight. For symbols, see Table 1. Confidence limits can he calculated for predicted values of Xor Y. Following Ricker (1973:411), the general form of the variance estimate for a single estimate of Y' given X' is: 'yx f '\2 -^ SSX' )■ (6) where S^v^ is the variance of observations from the regression line in the vertical direction, A'^ is the number of observation pairs in the regression, and Pj.' is the value of X' used to estimate Y'. In the reverse case where X' is being predicted, S^.y^, SSY', Py., and Y' are substituted for Sy/ ,SSX\Py, and X'. Because we have used GM regression equations rather than predictive regression equa- tions, the use of Expression (6) is not strictly legi- timate. However, Ricker (1973:413) finds the error involved is small and concludes that "... it is pos- sible to recommend using ordinary symmetrical confidence limits for the GM regression. They are a reasonable approximation to the true limits and will rarely lead to incorrect conclusions." The values required to use Expression (6) to cal- culate confidence limits for predicted A''s or Y's are given in Table 3. This variance and the ^95 value are used to construct confidence limits for the logarithms: 782 WIEBE ET AL.: RELATION OF VOLUME, WET AND DRY WEIGHTS, AND CARBON 10. T i I 5:! 0.01 0.001 0.01 01 1. 10. DRY WEIGHT (mg/m^) i 1 I I M III 100. 1^" t 0.1 S K 0.01 i 0.001 0.0001 0.01 o H — I 1 II Minn 1 — I I I I nil 1 — I I 0.1 1. 10. DRY WEIGHT (mg/m^) 100. 10.x i I I 0.01 O001 0.0001 0.001 0.01 0.1 WET WEIGHT (g/m^) Figure 5.-Plots of Be et al. (1971) and Be (footnote 5) data used in calculating geometric mean regression lines relating wet weight and displacement volume to dry weight, and displacement volume to wet weight. For symbols, see Table 1. Y' ± fgsVVarFor X' + t^^V^airX'. Antilogging provides multiplicative limits for the untransformed data. For example, suppose an es- timate of carbon is desired having measured a displacement volume (Y) of 0.1 cc/m^. Using Equation 1 in Table 2 and Expression (6) the following values result: Log (upper Upper Log (lower Lower Log (Y) Log (X) X(mg/m') limit) limit limit) limit -1.0 0.557 3.61 0.916 8.25 0.198 1.58 Thus, the antilogged estimate of carbon is 3.6% of the displacement volume with upper and lower 95% limits of 8.3% and 1.6%. Comparison of the regressions based upon the data of Be et al. (1971) and Be (footnote 5) with the 783 FISHERY BULLETIN: VOL. 73, NO. 4 Table 3.- Values required to calculate 95^ limits for values of X or 1' predicted from regression equations in Table 2. The ^95 value is based on the number of observations for each regression. For comparison abbreviations see Table 2 caption. son '95 Pi rediction of Y P rediction of X Compari X ssx S 2 y ssy S 2 DVvs. C 1.98 0.8310 69.3796 0.022095 -0.7573 47.1915 0.032483 WW vs. C 2.00 0.2076 22.0394 0.018282 -1.3663 16.1224 0.024992 DWvs. c 1.96 0.6456 117.7726 0.016987 1.1383 115.5700 0.017311 DVvs. DW 1.96 0.8369 106.4068 0.019299 -1.1181 79.6486 0.025782 WW vs. DW 1.98 0.6473 19.9736 0.011558 -1.3868 18.0252 0.012808 DVvs. WW 1.99 -1.3706 17.3509 0.023822 -1.2347 17.4390 0.023701 BDV vs. BDW 1.96 0.2408 85.9032 0.093843 -1.6446 63.8768 0.092518 BWWvs. BDW 1.96 0.2343 87.7505 0,085121 -1.7010 73.2612 0.077210 BDV vs. BWW 1.96 -1.6813 89.1699 0.054240 -1.6439 85.8228 0.056355 PDW vs, , PC 2.02 0.6992 1.9394 0.028706 1.5742 2.0334 0.027378 DVvs. DW 1.99 0.3918 25.6916 0.026995 1.3693 17.3094 0.040067 WW vs. DWi 1.99 0.1432 10.9063 0.011083 1.1101 9.7621 0.012383 WW vs. DVi 1.99 1.0998 7.1279 0.015876 1.2063 7.5091 0.015070 'Calculated values based on biomass data which was not standardized to per cubic meter. regressions based solely on our data reveals two notable features. First, the slopes of the regres- sions based on the same biomass estimators; i.e., displacement volume versus dry weight, wet weight versus dry weight, and displacement volume versus wet weight, are significantly different (P<0.05). This was tested by calculating approximately 95% limits for the difference in slopes using standard normal distribution theory: {i\r Vq^)±1 .96/Var v^, + V ar /'g^ . As was true in our cases, if A/' ± 95% limit does not cross 0, the slopes are significantly different. In all cases, slopes of the regressions derived from our data are closer to 1.0. The second feature is that there is a significant difference (P<0.005) in the variance of observa- tions from the regression lines. The Be et al. (1971) and Be (footnote 5) variance for displacement volume versus dry weight is 4.9 times larger than that calculated for our data; for wet weight versus dry weight, it is 7.4 times larger; for displacement volume versus wet weight it is 2.3 times larger. These differences are probably due in large part to the differences in methods used to determine displacement volume and wet weight. The mer- cury immersion method Be et al. (1971) and Be (footnote 5) used to measure displacement volume provides estimates substantially more variable than the technique used by us (Grice and Wiebe unpubl. data). The increased variability of their wet weights may have resulted from their use of a vacuum to remove some of the interstitial water. One implication of the lower slopes for the Be et al. (1971) and Be (footnote 5) data is that it appears the percentage of interstitial water in their samples may change more radically with increasing biomass than in our samples. This inference is drawn from the calculated values relating dry weight to wet weight and displacement volume in percent (Table 4). The alternate explanation is that as biomass per cubic meter increases, the percentage of wet weight or displacement volume that constitutes dry weight increases as a result of a decrease in intracellular water. It seems unlikely that this accounts for the differences between the two sets of data. Seasonal effects have been minimized by collection of samples at various times of the year and geographical effects should be similar since both studies covered wide geographical ranges. Table 4.-Regression ecjuation prediction of the percentage of displacement volume (DV) or wet weight (WW) that is dry weight (DW) for selected dry weight concentrations. DW Beet al. (1971) and B6 (footnote 5) % DV % WW This study DWI (g) This study (mg/m3) % DV % WW % DV % WW 0.1 1.0 10.0 100.0 3.80 5.39 6.70 7.89 11.80 11.53 20.80 16.87 5.10 6.95 9.48 12.94 8.95 10.95 11.27 12.65 0.1 1.0 10.0 100.0 5.9 9.0 13.5 20.5 9.4 10.6 12.0 13.6 'Calculated values based on biomass data which were not standardized to per cubic meter. 784 WIEBE ET AL.: RELATION OF VOLUME, WET AND DRY WEIGHTS, AND CARBON DISCUSSION Piatt et al. (1969), in a comparison of the seasonal changes in dry weight, carbon, and caloric values of zooplankton collected from St. Mar- garet's Bay, Nova Scotia, found a fivefold varia- tion in caloric content per unit dry weight. As a result they concluded ". . . that there is no single conversion factor that will serve to convert biomass of zooplankton, expressed as dry weight, to its energy equivalent." A similar conclusion was inferred for the conversion of dry weight to car- bon. They found, however, that the carbon content of zooplankton could be used to predict the energy equivalent. These results appear to contradict our finding that a statistically significant relationship does exist between pairs of the different measures of biomass including dry weight and carbon. The explanation for this discrepancy lies in the fact that the data of Piatt et al. represents a small segment of the extensive range of biomass per cubic meter which occurs in marine waters. This fact, coupled with high variation of the dry weight to carbon ratios, appeared to them to provide a nonsignificant relationship. We have used their data (as tabulated by Piatt and Irwin 1968, table 4) to examine the fit of their data to our regression line. After transformation to logarithms (base 10), a linear GM regression line was calculated for their 45 pairs of dry weight and carbon values. While the slope of this line was significantly different from zero (P<0.001), it was nonsig- nificantly different (P>0.05) from ours (Table 2). However, the intercept was substantially different. This is a reflection of the fact that their carbon values average 14% of dry weight, whereas in our data the average is 32%. The wet-combus- tion method (described by Strickland and Parsons 1965) which they used to determine carbon ap- parently provides lower estimates (an average here of 58% lower) than the high temperature combustion technique we used. Sharp (1973) found that persulfate oxidation yields an average 22% lower values than high temperature combustion when these methods are used to measure total or- ganic carbon in seawater. In terms of variability, the observations of Piatt et al. (1969) have a variance from the regression line significantly (P<0.01) larger than ours by a factor of 1.6. It is clear from the comparisons of biomass measures we have carried out, and from other un- published work performed at this laboratory, that the techniques used by various investigators in determining a particular biomass measure (such as displacement volume) provide substantially different answers which are not readily compara- ble. This is particularly true of displacement volume and wet weight and to a lessor degree, carbon. A similar conclusion was reached by Nakai and Honjo (1962). Only the procedure for measur- ing dry weight described by Lovegrove (1966) seems to have been widely adopted and values presented by various investigators using this technique seem to be intercomparable. With displacement volume and wet weight, the problem stems largely from the differing amounts of in- terstitial water adhering to the zooplankton at the time of measurement. We have found, as did Nakai and Honjo (1962), that for a given technique, the amount of interstitial water varies inversely with the amount of biomass being measured. The amount, however, varies from technique to tech- nique. Efforts to significantly reduce the amount of interstitial water present appear to create ad- ditional error. Rather than simply concentrating on the reduction of interstitial water, it is more important to establish a reproduceable procedure that generates values which can be directly related to a more absolute standard such as carhop as we have tried to do. The data on which Equations 1 to 6 and 11 to 13 in Table 2 are based were developed using methods which appeared to us to involve the least amount of technique-derived error and which required little complex instrumentation. The zooplankton biomass values used in this study encompass a significant part of the range of values an investigator is likely to encounter working in either coastal waters or the open ocean. Thus, the equations we have presented should be useful in a wide variety of situations providing the same techniques to measure biomass are employed. It is important to bear in mind, however, that situations do occur in which these equations may not apply. One example is where marine populations are dominated by salps, doliolids, jellyfish, or chaetognaths. The very high percentage of intracellular water in these or- ganisms may cause the relationships between displacement volume or wet weight and dry weight or carbon to deviate strongly from our predicted relationships. In such cases, which in our experience occur infrequently, we recommend that dry weight or carbon be measured directly. 785 FISHERY BULLETIN: VOL. 73, NO. 4 ACKNOWLEDGMENTS We express our appreciation to George Grice for assistance and discussion in the development of this project; to David Menzel for permission to use his unpublished data resulting from cruises All 52 and All 48; and to Allan Be for permission to use his biomass data. Woollcott Smith provided help- ful comments regarding statistical procedures. Loren Haury and Peter Ortner critically read the manuscript. LITERATURE CITED Ahlstrom, E. H., and J. R. Thrailkill. 1963. Plankton volume loss with time of preservation. Calif. Coop. Oceanic Fish. Invest. Rep. 9:57-73. Be, a. W. H., J. M. FoRNS, and 0. A. Roels. 1971. Plankton abundance in the North Atlantic Ocean. In J. D. Costlow, Jr. (editor), Fertility of the sea 1:17- 50. Gordon and Breach Sci. Publ., N.Y. Curl, H., Jr. 1962. Analyses of carbon in marine plankton organisms. J. Mar. Res. 20:181-188. Frolander, H. F. 1957. A plankton volume indicator. J. Cons. 22:278-283. LOVEGROVE, T. 1966. The determination of the dry weight of plankton and theeffectof various factors on the values obtained. InH. Barnes (editor), Some contemporary studies in marine science, p. 429-467. George Allen and Unwin Ltd., Lond. McEwEN, G. F., M. W. Johnson, and T. R. Folsom. 1954. A statistical analysis of the performance of the Fol- som plankton sample splitter, based upon test observa- tions. Arch. Meteorol. Geophys. Bioklimatol., Ser. A, 7:502-527. McGowan, J. A., and D. M. Brown. 1966. A new opening-closing paired zooplankton net. Univ. Calif., Scripps Inst. Oceanogr. Ref. 66-23, 56 p. Menzel, D. W., and J. H. Ryther. 1961. Zooplankton in the Sargasso Sea off Bermuda and its relation to organic production. J. Cons. 26:250-258. Nakai, Z., and K. Honjo. 1962. Comparative studies on measurements of the weight and the volume of plankton samples. A preliminary ac- count. Indo-Pac. Fish. Counc. Proc., 9th Sess., Sect. II, p. 9-16. Platt, T., V. M. Brawn, and B. Irwin. 1969. Caloric and carbon equivalents of zooplankton biomass. J. Fish. Res. Board Can. 26:2345-2349. Platt, T., and B. Irwin. 1968. Primary productivity measurements in St. Margaret's Bay. 1967. Fish. Res. Board Can., Tech. Rep. 77, 123 p. RiCKER, W. E. 1973. Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. Sharp, J. H. 1973. Total organic carbon in seawater— comparison of measurements using persulfate oxidation and high temperature combustion. Mar. Chem. 1:211-229. Strickland, J. D. H., and T. R. Parsons. 1965. A manual of seawater analysis. 2nd ed., revised. Bull. Fish. Res. Board Can. 125, 203 p. Sutcliffe,.W. H., Jr. 1957. An improved method for the determination of preserved plankton volumes. Limnol. Oceanogr. 2:295-296. Tranter, D.J. 1960. A method for determining zooplankton volumes. J. Cons. 25:272-278. Vaccaro, R. F., G. D. Grice, G. T. Rowe, and P. H. Wiebe. 1972. Acid-iron waste disposal and the summer distribution of standing crops in the New York Bight. Water Res. 6:231-256. Wiebe, P. H., G. D. Grice, and E. Hoagland. 1973. Acid-iron waste as a factor affecting the distribution and abundance of zooplankton in the New York Bight. II. Spatial variations in the field and implications for monitoring studies. Estuarine Coastal Mar. Sci. 1:51-64. Yentsch, C. S., and J. F. Hebard. 1957. A gauge for determining plankton volume by the mercury immersion method. J. Cons. 22:184-190. 786 EFFECTS OF VARIOUS CONCENTRATIONS OF DISSOLVED ATMOSPHERIC GAS ON JUVENILE CHINOOK SALMON AND STEELHEAD TROUT Earl M. Dawley and Wesley J. Ebel' ABSTRACT Bioassays in shallow tanks (25 cm deep) with dissolved nitrogen and argon gas concentrations ranging from 100 to 125% of saturation in water at 15°C were conducted to determine lethal and sublethal effects on juvenile chinook salmon, Oncorhynchus tshawyfscha, and steelhead trout, Salmogairdneri. Significant mortality of both species commenced at 115% saturation of nitrogen and argon (111% saturation of total dissolved atmospheric gas pressure). Over 50% mortality of both steelhead and chinook occurred in less than 1.5 days in water at 120 and 125% of saturation. Significant differences in swimming performance, growth, and blood chemistry were measured in groups of fish tested at sublethal exposures in various concentrations of dissolved gases. Sublethal stress for 35 days at 110% dissolved nitrogen (106% total atmospheric gas) decreased normal swimming ability of chinook. Growth of both steelhead and chinook was affected by sublethal exposures in water saturated with atmospheric nitrogen and argon at 105, 110, and 115%. Blood chemistry was affected at sublethal exposures in water at 115% saturation. Supersaturation of atmospheric gas (mainly ni- trogen) in waters of the Columbia and Snake rivers-caused by spillway discharges from dams— has been well documented as a serious problem to valuable stocks of Pacific salmon, On- corhynchus spp., and steelhead trout, Salmo gairdneri. Gas bubble disease resulting from this supersaturation causes both direct and indirect mortalities. Direct mortality results from air em- boli in the heart and gill filaments, destruction of vital organs, or characteristic red blood cell hemolysis (Marsh and Gorham 1905; Pauley and Nakatani 1967; Bouck et al. 1970-). Indirect mor- tality is a consequence of later invasion by disease organisms (Coutant and Genoway 1968') or of increased predation due to reduced performance capabilities of the fish as the result of sublethal exposure to supersaturation. The lowest level of nitrogen supersaturation at which juvenile salmon or steelhead trout can be exposed continually with no detrimental effects is 'Northwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA. 2725 Montlake Boulevard East, Seattle, WA 98112. 'Bouck, G. R., G. A. Chapman, P. W. Schneider, Jr., and D. G. Stevens. 1970. Observations on gas bubble disease in adult Columbia River sockeye salmon (Oncorhynchus nerka). Pac. Northwest Water Lab. [Fed. Water Qual. Adm., Corvallis, Oreg.], June 30, 1970. Unpubl. manuscr., 19 p. 'Coutant, C. C, and R. G. Genoway. 1968. Final report on an exploratory study of interaction of increased temperature and nitrogen supersaturation on mortality of adult salmonids to U.S. Bur. of Commercial Fisheries, Seattle, Washington. Battelle Mem. Inst. Pac. Northwest Lab. Richland, Wash., November 28, 1968, 28 p. Manuscript accepted February 1975. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. not known. Several investigators have recorded the lowest level observed during various experiments where mortalities occurred from gas bubble disease; however, very little attention has been given to determining the effect of sublethal exposure on physiological and behavioral perfor- mance. Harvey and Cooper (1962) indicated 108- 110% saturation produced gas bubble disease and subsequent mortalities in sockeye salmon alevins, 0. nerka; Rucker and Tuttle (1948) indicated a level somewhere between 110 and 115% as being the critical range for trout. Shirahata (1966) con- ducted the most comprehensive study to date on the effects of various levels of nitrogen gas on rainbow trout (rainbow trout is the nonanadromous form of 5. gairdneri, whereas the steelhead trout is the anadromous) from hatching to the swim-up stage, but such detail is lacking for other species of salmonids. In many experiments on gas bubble disease, either the water tempera- tures, nitrogen gas concentrations, or life stages of the test fish were omitted from record, thus making the results incomplete for critical applica- tions. Costs involved in alleviating the supersatura- tion problem in the Columbia and Snake rivers will be considerable. The extent of these costs will depend on the degree of protection required to afford a safe environment for the aquatic biota. It is imperative, therefore, that regulatory measures established to govern the level of saturation be 787 FISHERY BULLETIN: VOL. 73, NO. 4 based upon a thorough understanding of the ef- fects of dissolved gases on aquatic organisms. This paper describes the results of dissolved gas bioassays with juvenile steelhead trout and spring Chinook salmon, 0. tshatvytscha, conducted by the National Marine Fisheries Service during the spring of 1972. These experiments were designed to assess lethal and sublethal effects of supersat- uration of atmospheric gases on test fish at levpls found in the Columbia and Snake rivers during the spring freshet. Atmospheric nitrogen concentra- tions' were of major concern and test levels ranged from 100 to 125% of saturation. Special note is made of testing procedures and ramifications of the effects of these on the outcome of our tests. METHODS Bioassays were carried out in the laboratory in shallow tanks (25-cm water depth) to negate the effects of hydrostatic pressure compensation. These facilities were similar to those described by Ebel et al. (1971). Water flow into each test tank was maintained at 3 liters/min at a temperature of 15° ± 0.5°C. Test tanks were partitioned with perforated fiberglass plates to form four sections — in-flow area, test area A, test area B, and out- flow section (Figure 1). Supersaturated water was produced by meter- 'Atmospheric nitrogen-nitrogen gas (98.8% by vol) plus argon gas (1.2% by vol) hereafter referred to as nitrogen or N2 + ^'^■ ing 0.7 liter/min air into the suction side of a cen- trifugal pump which recirculated water through a 197-liter (52-gallon) closed receiver at a rate of about 190 liters/min (50gal/min). Water pressure throughout the system was at 1.4 kg/cm- (20 psi) except in a short section of pipe on the discharge side of the pump where it was increased to 3.2 kg/cm- (45 psi) by use of a valve for additional back pressure necessary to achieve the required supersaturation. Water remained in the recircula- tory system for about 10 min before passing to the test tanks. This arrangement supersaturated the water to about 145% of air saturation. Water was then piped to the test tanks where it passed over a series of perforated fiberglass plates into an inlet box with air bubbling through a bottom plate of porous polyethylene. The number of fiberglass plates and volume of air were adjusted to yield the various levels of saturation. An increase of air to water interface directly decreased the excess dis- solved gas content. Water samples for dissolved gas analyses were collected throughout the tests near the center of each test tank directly in front of the partition between A and B testing areas and in some tests at the center of each section of the tank. Frequency of analysis varied from once an hour to once a day depending on duration of test. Procedure for analysis of dissolved nitrogen was from Van Slyke and Neill (1924) using manometric blood gas apparatus; dissolved oxygen was analyzed using modified Winkler procedures Vinyl air supply tube -50 cm -50em- -124cm PLAN VIEW Supersoturoted constant-temperature water source Oiilesj air ^ supply j( .1 Mechonicol woter equilibrotor . -^ f with adjustoble fiberglass 'or ilo 24cm Perforated fiberglass plates Trar)slucent riberglass cover Figure L-Pian and cross-sectional views of test tank used for bioassay of dissolved gas. 35cm Polyethylene plote ^ Air bubbling porous to air ^g,^, equilibrotion CROSS SECTION 788 DAWLEY and EBEL: EFFECTS OF DISSOLVED GASES ON SALMONIDS (American Public Health Association et al. 1971). Gas concentrations at saturation (100%) were taken from Weiss (1970). To obtain the dissolved gas levels for the various tests, we adjusted the water equilibrators of each tank (screens plus air bubbling boxes) until the nitrogen concentration remained within ±2% of the desired value. The oxygen concentration was then measured and we found that the saturation value was 5 to 10% lower than that of Ng+Ar for each tank. This did not differ appreciably from prevailing oxygen saturations in the Columbia and Snake rivers which are usually 5 to 10% lower than dissolved nitrogen values (Beiningen and Ebel 1971; Ebel 1971). After introducing fish, however, we noted that the oxygen concentration dropped further (presumably because it was con- sumed), resulting in values from 8 to 28% of sat- uration below that of Ng + Ar, particularly in test area B and the outlet area of the tank. Due to large numbers of fish required for experiments on the survivors of these bioassays, and the complexity of changing the dissolved gas ratios of the water source, we did not alter the O2 concentrations in the tests but carefully documented the mid-tank gas concentrations. Data affected by this drop in oxygen partial pressure are discussed later in this report. One-year-old spring chinook salmon from Leavenworth National Fish Hatchery, Leaven- worth, Wash., and steelhead trout from the Washington Department of Game Hatchery at Aberdeen, Wash., were used in the tests. Test populations were acclimated to laboratory water at 15°C with normal dissolved gas concentrations for at least 2 wk before testing. Groups of 30 or 60 fish were placed simultaneously in control (100% atmospheric nitrogen saturation) and test tanks set at 105, 110, 115, 120, and 125% of Ng + Ar sat- uration and one to four replicates of each test were made, depending on test level. When 60 fish were being tested, 30 were in each of the two test sections A and B. Fish were randomized before introduction into individual test tanks. Mean sizes of the fish at completion of the tests are indicated in Table 1. Measurement of size at the beginning of the tests was omitted to avoid placing addi- tional stress on the test animals. Feeding of fish during the test period began 48 h after introduc- tion to test tanks; thereafter they were fed to satiation once each weekday. Lethal exposure times to 10 and 50% mortality (LEjo and LE50) were averaged for lots of test fish Held in tank sections A and B during the same time period, and the mid-tank gas concentrations were used for analysis with the exception of the steelhead groups stressed at 115% nitrogen; in these tests, exposure times and gas concentrations were measured separately for A and B sections of the tanks. In addition, lethal exposure times to 100% mortality (LEjoq) for chinook and steelhead at all levels of supersaturation were taken only from groups held in the A section of the tanks. Observations of behavior, progression of exter- nal signs of gas bubble disease, and mortality were recorded continuously for the first 6 h then every V2 h for 24 h and every 3, 6, or 12 h thereaf- ter-depending on test concentration-until ter- mination of the bioassay at 35 days. Observations of change in degree of external disease signs among test fish after a recovery period in normally saturated (100%) water also were made from selected groups. Sublethal effects of supersaturation were as- Table l.-Comparison of mean weigrhts and lengths of surviving test and control fish held in 15°C water w^ith N2 +Ar levels at 100 to 125% of saturation, February-April 1972. Test level (% of saturation Testing period' (mo/day) Duration {individual tests) Test fish Control fish ( Weight (g) 100% N^ + Ar) of Nj + Ar) Weight (g) Le ngth (mm) Length (mm) Spring ch inook salmon 105 2/8-3/14 35 days 13.6 115 15.5 119 110 3/7-4/11 35 days 17.5 125 17.9 126 115 2/8-3/14 35 days 13.6 115 15.5 119 120 2/8-3/3 i55h 16.2 120 18.0 122 125 2/8-3/1 <38h 16.8 117 16.8 118 Steelhead trout 105 4/3-5/8 35 days 18.8 130 22.8 135 110 4/10-5/15 35 days 20.0 130 22.0 132 115 4/13-5/13 £35 days 19.8 130 20.9 132 120 4/3-4/18 i53h 20.6 124 — — 'Replicates of tests at 115-125% levels were made at various time intervals throughout the indicated test period; others lasted the full indicated period. 789 FISHERY BULLETIN: VOL. 73, NO. 4 sessed by using measurements of maximal swim- ming performance, blood chemistry, and photic response. Measurements were made on groups of sur\'ivors from lethal exposure tests immediately after the LEjo and LE50 points were reached or following a 2-wk recovery period in 100% saturat- ed water. Swimming performance was measured by distance gained and time of swimming against a constant water current of 1.25 m/s within a U- shaped inclined trough (14 m long and 8 cm wide). Blood samples were analyzed on a Techni- con Sequential Multiple Analyzer (SMA 12/60).' Pooled serum samples were analyzed for Ca, Na, PO4 , K, CI, albumin, total protein, cholesterol, alkaline phosphatase, glucose, urea, uric acid, total bilirubin, lactic dehydrogenase and serum glu- tamic oxaloacetic-acid transaminase. Photic re- sponse was evaluated by electrophysiological monitoring of the optic tectum during retina stimulation with flickering light. A more detailed description of the methods used in the swimming performance and blood chemistry measurements appear in reports by Schiewe (1974) and by New- comb (1974),*^ respectively. RESULTS Relationships Among Mortality, Exposure Time, and Gas Concentration Mean exposure times at which 10, 50, and 100% mortality occurred at 120 and 125% Ng + Ar sat- uration indicate no substantial difference between susceptibility of juvenile chinook and steelhead trout (Table 2). However, at 115% N2 + Ar saturation, steelhead appeared to be more susceptible than chinook; i.e., steelhead reached the 50% mortality level within 35 days, whereas LE5Q was never reached in test groups of chinook. Mortalities of control fish for all tests (105-125%) ranged from 0 to 3.3% throughout the 35-day test periods. Because of the comparatively minor losses of controls, data from test groups are given as observed (not compensated for loss of controls). Mortalities observed in tests at 105 and 110% of nitrogen saturation were 5% or less for both species, and gas bubble disease was not the ap- parent cause of death. The onset of mortality attributable to gas supersaturation occurred at about 115% dissolved nitrogen among both steelhead and chinook. At about 120% nitrogen saturation the means of lethal exposure times to 50% mortality (LE50) were 26.9 and 33.3 h for chinook and steelhead, respectively. LE5q's for chinook and steelhead at 125% nitrogen saturation were 13.6 and 14.2 h, respectively, which are similar to those (11.3 and 14.0 h) observed in earlier tests by Ebel et al. (1971) at test concentrations of 125 to 130% N2+ Ar. Test fish stocks used previously were from different hatcheries and earlier brood years and were slightly larger (spring chinook-23 g and 135 mm, steelhead-54 g and 179 mm). Table 2. -Mean values of lethal exposure time for juvenile steelhead and chinook acclimated to 15°C and then subjected to various levels of gas saturation' from 100 to 125% in shallow tanks (25-cm depth). Percent eotiiration Percent mortality Exposu re time (h) (Nj + Ar) Steelhead Chinook 125 10 50 100 10.3 14.2 223.0 10.6 13.6 232.1 120 10 50 100 26.0 ■33.3 240.0 19.3 26.9 255.0 115 10 50 100 2258.0 2486.0 Not reached (7% mortality In 792 h) Not reached Not reached 110 105 100 Mortality of 5% or less recorded for either steelhead or chinook after 35 days at these concentrations. Gas bubble disease was not apparent cause of deaths. ^rade names referred to in this publication do not imply en- dorsement of commercial products by the National Marine Fisheries Service, NOAA. 'Newcomb, T. W. 1974. Changes in juvenile steelhead (Salmo gairdneri) blood chemistry foUovi^ing sublethal exposure to various levels of nitrogen supersaturation. Northwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, Wash. Unpubl. man user. 'Percentage saturation of nitrogen and argon was set as Indi- cated in the table {^ 2%). Oxygen concentrations ranged be- tween 87 and 98% saturation in tanks set at 100-110% nitrogen plus argon saturation; in tanks set a 115-125% nitrogen satura- tion, Oj levels ranged between 98 and 115%. 2Exposure times indicated for test replicates of section A only. Mortality in section B had not reached indicated level at termi- nation of test. Effect of Oxygen Concentrations on Time to Death Measurements The role of atmospheric gases other than ni- trogen (particularly oxygen) in causing gas bub- ble disease has been questioned by several inves- tigators. Arguments for and against the assump- tion that dissolved atmospheric nitrogen is the exclusive cause of gas bubble disease are prevalent throughout the literature (Marsh and Gorham 1905; Doudoroff 1957; Egusa 1959, 1969; Shirahata 790 DAWLEY and EBEL: EFFECTS OF DISSOLVED GASES ON SALMONIDS 1966; Bouck 1972^; Rucker 1972). Most of the comprehensive studies, however, have been analyzed in terms of nitrogen concentration, as- suming it to be the controlling influence upon the effects of gas bubble disease (even greater than indicated by the 80/20 ratio of the partial pres- sures Ng /O2 ). This assumption was based upon supposed biochemical decrease of the effective oxygen partial pressure within the fish. In comparing data from this experiment with that from past research we should acknowledge that our primary criterion during planning and set-up stages was dissolved nitrogen + argon concentration. At the outset of these experiments, oxygen levels were monitored primarily for documentation of overall water quality rather than for use in analysis of their effect upon the test organisms. However, upon examination of initial results derived from each of the tests carried out to lethal exposures, we found that the times for LEjq and LE50 were consistently less in test section A than in Section B. Analyses of in- dividual gas pressures in each of the two sections of the tanks were made to determine whether variations occurred among the component gases. We found that nitrogen concentrations were con- stant in both areas, but oxygen concentrations remained consistently lower (5-10%) in section B than in section A. The lower oxygen concentra- tions-thus lower (1-2%) total dissolved gas (TDG) saturations-appeared directly correlated with the lower mortality rates in section B of the test tanks. For example, when we examined mortality rates of individual groups of steelhead from A and B test sections at 115% Ng + Ar, we found: Ng + Ar saturation (in section B) of 116.0% and 88.2% of Og saturation (TDG at 110.0% of saturation) caused no mortality in 35 days for one replicate of 30 fish, whereas Ng + Ar saturation (in section A) of 116.0% and 98.8% O2 (TDG at 112.1%) caused 50% mortality in an average of 20 days for two replicates. Effect of Supersaturation Stress on Growth Exposure to sublethal concentrations (concen- trations at which no substantial mortality oc- 'Bouck, G. R. 1972. Effects of gas supersaturation on salmon in the Columbia River. West. Fish. Toxicol. Stn., Environ. Prot. Agency, Corvallis, Oreg. Paper presented at Ecol. Soc. Am. Symp. Aug. 1972, 29 p. ^ ^ curred within 35 days) of Na+Ar appeared to af- fect growth of both juvenile chinook and steelhead. Mean weights and lengths of test fishes after 35 days in dissolved nitrogen concentrations of 105, 110, and 115% of saturation (Figure 2) were in each instance less than those of controls. « 10 lesl control STEELHEAD m E 120 E no 115 105 PERCENT SATURATION Nj + A( Figure 2.-Comparison of mean weights (W) and lengths (L) for test and control groups of juvenile chinook salmon and steelhead trout after 35 days at saturation levels of 100% (control), 105%, 110%, and 115%. A statistical test of the hypothesis-that the slopes of the regression of mean weight of control fish groups and mean weight of test fish groups were equal -yielded a value of t = 4.938 (P<0.02 at 4 df ). The same statistical test of mean lengths of control vs. test groups yielded t = 1.36 (P<0.25 at 4 df). The lower t value calculated from length data is attributable to the duration of the test not being long enough to significantly overcome the varia- tion in lengths between individuals within groups. We attribute the difference between size of test and control lots to the effect of supersaturation on the normal growth of the test fish. After 30 days of testing at the 115% level, feed- ing response of the chinook fingerlings became lethargic. Many of the test fish had spinal flexures, exophthalmia, and large buccal cavity gas blisters and were unable or unwilling to move and accept food when made available. By contrast, control fish exhibited aggressive feeding behavior throughout the tests. Gross gas bubble disease signs and behavioral changes were less evident at 110% Ng -I- Ar and nonexistant at 105%. Testing for changes in the condition factor of juvenile fall chinook and steelhead during long- term (2-4 mo) holding in water saturated 100 to 127% Ng+Ar is currently underway. Results of these tests may provide further information on effects of gas supersaturation on growth rate. 791 FISHERY BULLETIN: VOL. 73, NO. 4 Figure 3. -Gas bubbles (arrow) in lateral line of juvenile chinook sal- mon. Progression of Gas Bubble Disease Observations on the progression of external signs of gas bubble disease in spring chinook ex- posed to various levels of supersaturation revealed that the first developments such as bubble forma- tion in the lateral line (Figure 3) appear within 2 h of exposure at 125%. Subcutaneous gas blisters between fin rays of at least one fin were present on each of the test animals before 11.5 h at 125%, and before 55 h at the 120% level. Several days' ex- posure were required before these signs occurred on fish tested at 115%. After 35 days, 56% of the fish at 110% had lateral line bubbles but only 4% had fin bubbles. Exophthalmia or "popeye," hyphema, cutaneous blisters of the head and buc- cal cavity, and spinal flexures were absent among fish tested at 120 and 125% but began appearing after 6 days on fish held at 115% and after 11 days on those held at 110% of nitrogen saturation. Ap- parently at the higher saturation levels, the fish died from cardiac occlusion or branchial artery occlusion (Figure 4) before development of these signs. By the end of 35 days, fish held at 115% Figure 4.-Gas emboli occluding gill filaments and branchial artery of chinook salmon held in 125% nitrogen saturation for 20 h. 792 DAWLEY and EBEL: EFFECTS OF DISSOLVED GASES ON SALMONIDS exhibited more than a 75% incidence of exophthalmia, 20% of the fish had spinal flexures, and 25% of the fish in section A became more or less immobile. After 35 days of exposure at 110% N2 +Ar, only 12% of the test fish exhibited signs other than the lateral line bubbles. No apparent signs of gas bubble disease were observed in fish tested at 105% nitrogen. Development of gas bubble disease signs in steelhead was similar to that of chinook-the signs occurred in the same sequence but the exposure time required to produce the signs was slightly less. Recovery From Gas Bubble Disease Observations on disappearance of gas bubble disease signs and delayed mortality following tests were made on groups of survivors of fish stressed to the LE50 level at 120, 125, and 130% of saturation. These survivors were placed in water at 100% gas saturation for up to 15 days. No delayed mortality could be attributed to prior ex- posure to supersaturation in either the chinook or the steelhead. The only significant mortality in any recovery group was a 10% loss of one replicate of steelhead subjected to 125% N2+ Ar until LEg^ , followed by a burst swimming performance test. Some mortality occurred after 102 h of recovery time, but the only observable disease sign was the presence of lateral line bubbles on one fish. Other mortalities during recovery were less than 3% of the fish held; no gas bubble disease signs were found. All external symptoms that were readily visible at the time the fish were removed from the recovery tanks had disappeared after 15 days in both species. Steelhead that had undergone 16, 24, and 35 days' exposure at 115% nitrogen saturation still showed gas bubbles after being held 3 days in normally (100%) saturated water. After 1.5 days' recovery, 64% exhibited lateral line bubbles or fin ray gas blisters and one fish (7%) retained unilateral exophthalmia; after 2 days' recovery, 88% of another group retained signs of lateral line bubbles and fin gas blisters; at 3 days, 54% of the third group retained like signs of gas bubble disease. After 15 days' recovery, no gas bubble disease signs were observed on groups of test fish examined. Effect of Supersaturation Stress on Survivors Burst swimming performance and blood chemistry were examined as potential indices of stress from sublethal exposures to supersaturated water. Swimming performance (Schiewe 1974) of chinook that survived from tests at 110-125% was significantly lower than that of control fish. Visual observations of behavior during swimming per- formance tests indicated genuine debilitation (inability to swim in some cases) which in turn resulted in lower swimming performance (i.e. less distance gained and less swimming time against a constant water current stimulus). No difference was apparent between performance of chinook salmon tested at 105% saturation and the control fish. Swimming performance of steelhead trout that survived tests at 105-125% was not significantly different from the performance of control fish. Performance of test and control lots of steelhead trout was highly variable. Fish stressed by ex- posure to supersaturation often responded in an irritated or stimulated fashion, which often resulted in a high measure of performance. Further tests with steelhead are needed to deter- mine whether swimming performance is a useful index of stress from supersaturation and, if so, whether test results in the laboratory apply to survival of fish in the river. Blood serum from groups of chinook and steelhead surviving supersaturation tests to LE^q and LE50 were analyzed (Newcomb see footnote 6) using a SMA 12/60. A 5% decrease in serum calcium was noted in chinook exposed to 115% ni- trogen plus argon when compared to those exposed to lower levels of supersaturation. Steelhead ex- posed to 115% nitrogen showed a 10 to 17% decrease in serum calcium and a decrease in serum albumin, total protein, serum chloride, cholesterol, and in alkaline phosphatase activity when com- pared to controls and those exposed to lower sat- urations. No significant changes in blood serum components were observed in samples taken from test groups exposed to levels of 105 and 110% of saturation when these were compared with con- trols. Measurements of photic response of salmonids failed to provide any consistent evidence of stress-related phenomena due to supersaturation so these tests were discontinued. 793 FISHERY BULLETIN: VOL. 73, NO. 4 DISCUSSION Data from these tests indicate that the critical level of supersaturation of nitrogen where juvenile spring chinook and steelhead began to show mortality was about 115% N2+Ar when Og saturations were about 95% (111% TDG). These data agree closely with the findings of Shirahata (1966), who indicated that the critical level for 2- mo-old rainbow trout was about 111.3%, N2+Ar and 99.7% 02(109% TDG). Although mortality from supersaturation did not occur until fish were exposed beyond 110% (± 2%) N2 + Ar, swimming performance measurements with juvenile chinook showed some effect from stress caused by exposure to supersat- uration at levels as low as 110% N2 + Ar (106% TDG). We believe that one can infer from the results of these tests, that something less than normal survival will result when juvenile chinook and steelhead are exposed for 35 days or longer at or above 110% N2+ Ar (106% TDG). Results of our testing program indicate that oxygen as well as nitrogen is responsible for caus- ing gas bubble disease, even when O2 concentra- tions are below saturation. The immediate conclusion drawn from this observation would be that total dissolved gas is the cause rather than any one or combination of component atmospheric gases. However, fish tolerance research by Egusa (1969) and by Rucker (1975) with various ratios of dissolved gas indicate that mortality from gas bubble disease is not necessarily in linear correla- tion with TDG. Egusa showed that oxygen sat- uration values of 400 to 500% were required to produce initial mortality of goldfish, Carassius auratus, and an eel Anguilla japonica when ni- trogen concentrations were near 100% (TDG 160- 180%). In earlier work with the same two species, however, Egusa (1959) recorded high mortality of goldfish with N2+Ar at 132% and O2 at 75% of saturation (TDG 123%), and of eel with N2+ Ar at 124%, O2 at 66% (TDG 112). Rucker found that mortality rate of juvenile salmon declines con- siderably if the ratio of oxygen to nitrogen is increased even though the same TDG pressure is maintained. It is apparent from our tests and those of Egusa and Rucker that the ratio of O2 and N2 must be considered as well as TDG when assessing possible effects from supersaturation. Additional information is needed to quantify the effects of various gas ratios (nitrogen to oxygen) on tolerance limits of fish in general. It is probable that most fish could tolerate higher total gas pressure if the major portion of the excess gas were oxygen. Dissolved gas measurements and resulting per- centage saturations for the Columbia and Snake rivers (Ebel 1969, 1971; Beiningen and Ebel 1971) have been based on surface or atmospheric pres- sure plus vapor pressure. Corrections for the hydrostatic pressure (or depth) at which a sample was taken were not made. Thus, the calculations of percentage saturation were made as though the samples were collected at the surface. This is con- venient when limnologists or oceanographers wish to compare values taken at various depths, but leads to confusion when attempting to assess how a given saturation measurement will affect a fish at depth. The depth that populations of fish travel must be considered when one attempts to determine the effects of an exposure to supersaturated levels of dissolved gases. Bubble formation in the circula- tory system or tissues of fish is directly dependent on the external hydrostatic pressure. For example, a fish traveling at a depth of only 1 m will be provided with enough hydrostatic pressure to compensate for a gas pressure in excess of 10% (110% saturation at surface pressures). A fish traveling at 3 m can compensate for 30%, or 130% saturation at surface pressures; a fish traveling at 10 m can compensate for an excess of 100% of saturation and so on. These tests were conducted in shallow tanks at essentially zero hydrostatic pressure with only a few centimeters depth com- pensation possible. The lethal exposure times we measured could only be applied directly to fish populations that could not compensate by sound- ing. Much more information is needed to deter- mine how a given gas level in a river affects the population inhabiting the river. Information regarding the behavior of fish is obviously essen- tial. We believe, however, that data from our tests support the 110% maximum allowable limit es- tablished by the Environmental Protection Agency primarily because significant mortalities did not occur until concentrations exceeded 110% TDG. Gas bubble disease signs either singly or in combination with one another did not correlate well with mortality. Those generated from stress conditions of 120% saturation and higher seemed to be nearly the same at LEjq as at LEjoo (gas blisters in the fins and lateral lines of most live and 794 DAWLEY and EBEL: EFFECTS OF DISSOLVED GASES ON SALMONIDS dead fishes). Signs that developed at low^er levels (110-115%) were obviously different from those appearing at the higher saturations; i.e., gas blisters in and around the eye, exophthalmia, cu- taneous gas blisters on the head and in the mouth, and spinal flexures. Neither set of signs (low-level or high-level types) correlate by percent of in- cidence or severity, with accumulative mortality. But they showed that one could determine with reasonable accuracy, whether fish observed in the river had been exposed to supersaturation for a long or short duration. Populations with signs of chronic exposure (exophthalmia, spinal flexures, etc.) could have been either 1.5 to 2.0 m deep in highly supersaturated water (130-135%) or near the water surface at near 115% saturation. SUMMARY AND CONCLUSIONS Bioassays in shallow tanks (25 cm) with dis- solved nitrogen and argon gas concentrations ranging from 100 to 125% of saturation were con- ducted to determine lethal and sublethal effects on juvenile chinook salmon and steelhead trout. Juvenile steelhead (130 mm fork length) reached the LE50 level within 35 days when exposed to 115% of nitrogen and argon saturation (112% TDG), whereas mortality of juvenile chinook (115 mm) did not exceed 7%. There appeared to be no substantial difference between susceptability of chinook and steelhead at 120 or 125% saturation N2 +Ar. No mortality related to supersaturation occurred in either juvenile chinook or steelhead trout exposed to 110 or 105% saturation Ng -^-Ar. Signs of gas bubble disease (such as bubbles in lateral line and exophthalmia) were evident on both species, however, after 35 days exposure to 110%. Time to death decreased in test tanks with higher oxygen concentrations (thus higher TDG) even though nitrogen and argon concentrations were identical, indicating that oxygen as well as nitrogen and argon concentrations must be con- sidered when time to death values are compared. The first notable sign of gas bubble disease was appearance of bubbles in the lateral line which appeared in some degree at all gas concentrations tested. Exophthalmia, dermal gas blisters of the buccal cavity and cephalic regions, and spinal flexures did not occur with short-term exposure (6 days) or at the higher levels (120 and 125%) but was prevalent after long exposure at both 115 and 110% saturation Ng + Ar. External gas bubble disease signs disappeared within 15 days when fish were placed in normally saturated water (100%). Fish stressed with supersaturation at sublethal levels for 35 days grew less than controls and the swimming performance of juvenile chinook ex- posed for sublethal periods to 110-125% nitrogen saturation was significantly lower than controls. Blood chemistry measurements indicated that significant differences occurred between blood samples taken from test and control chinook and steelhead after they were exposed to levels of 115% saturation. Serum calcium, for example, was 10- 17% lower in samples taken from test groups of steelhead. We concluded from these experiments that: 1. Significant mortality of both juvenile chinook and steelhead trout commences at about 115% sa- turation of nitrogen and argon (111% TDG). 2. Sublethal exposures to various concentra- tions of dissolved gas significantly affects swim- ming performance, growth and blood chemistry of chinook, and growth and blood chemistry of steelhead trout. 3. The first externally evident sign of gas bub- ble disease on juvenile chinook and steelhead trout exposed to supersaturation occurs as bubbles in pores of the lateral line. 4. Fish returned to normally (100%) saturated water appear to recover within 15 days from ex- posure to supersaturated water. LITERATURE CITED American Public Health Association, American Water Works Association, and Water Pollution Control Federation. 1971. Standard methods for the examination of water and wastewater. 13th edition. Am. Public Health Assoc, Wash., D.C., 874 p. Beiningen, K. T., and W. J. Ebel. 1971. Dissolved nitrogen, dissolved oxygen, and related water temperatures in the Columbia and lower Snake rivers, 1965-69. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Data Rep. 56, 60 p. on 2 microfiche. Doudoroff, R. 1956. Water quality requirements of fishes and effects of toxic substances. In M. E. Brown (editor), The physiology of fishes 2:403-430. Academic Press Inc., N.Y. Ebel, W. J. 1969. Supersaturation of nitrogen in the Columbia River and its effect on salmon and steelhead trout. U.S. Fish Wildl. Serv., Fish. Bull. 68:1-11. 1971. Dissolved nitrogen concentrations in the Columbia and Snake Rivers in 1970 and their effect on chinook salmon and steelhead trout. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-646, 7 p. Ebel, W. J., E. M. Dawley, and B. H. Monk. 1971. Thermal tolerance of juvenile Pacific salmon and 795 FISHERY BULLETIN: VOL. 73, NO. 4 steelhead trout in relation to supersaturation of nitrogen gas. Fish Bull., U.S. 69:833-843. Egusa, S. 1959. The gas disease of fish due to excess of nitrogen. [In Jap., Engl, summ.] J. Fac. Fish. Anim. Husb., Hiroshima Univ. 2:157-182. 1969. Yozon sanso kajo ni yoru sakana no gasu-byo ni tsuite (Gas disease of fish due to excess dissolved oxygen). Fish Pathol. 4:59-69. (Partial transl., Bur. Sport Fish Wildl., West. Fish. Dis. Lab., Seattle, Wash.) Harvey, H. H., and A. C. Cooper. 1962. Origin and treatment of a supersaturated river water. Int. Pac. Salmon Fish. Comm., Prog. Rep. 9, 19 p. Marsh, M. C, and F. P. Gorham. 1905. The gas disease in fishes. [U.S.] Bur. Fish., Rep. U.S. Comm. Fish. 1904:343-376. Pauley, G. B., and R. E. Nakatani. 1967. Histopathology of "gas-bubble" disease in salmon fingerlings. J. Fish. Res. Board Can. 24:867-871. Rucker, R. R. 1972. Gas-bubble disease of salmonids: a critical review. U.S. Fish Wildl. Serv., Bur. Sport Fish Wildl., Tech. Pap. 58, 11 p. 1975. Gas-bubble disease: Mortalities of coho salmon, On- corhynchus kittiutcli, in water with constant total gas pressure and different oxygen-nitrogen ratios. Fish. Bull., U.S. 73:915-918. Rucker, R. R., and E. M. Tuttle. 1948. Removal of excess nitrogen in a hatchery water supply. Prog. Fish-Cult. 10:88-90. SCHIEWE, M. H. 1974. Influence of dissolved atmospheric gas on swimming performance of juvenile chinook salmon. Trans. Am. Fish. Soc. 103:717-721. Shirahata, S. 1966. Experiments on nitrogen gas disease with rainbow trout fry. [In Jap., Engl, abstr.] Bull. Freshwater Fish Res. Lab. 15(2): 197-211. Van Slyke, D. D., and J. M. Neill. 1924. The determination of gases in blood and other solu- tions by vacuum extraction and manometric measuremint. I. J. Biol. Chem. 61:523-573. Weiss, R. F. 1970. The solubility of nitrogen, oxygen, and argon in water and seawater. Deep-Sea Res. Oceanogr. Abstr. 17:721-735. 796 BIOLOGY AND TAXONOMY OF THE GENUS NEMATOSCELIS (CRUSTACEA, EUPHAUSIACEA) K. GOPALAKRISHNAN' ABSTRACT The seven species of Nemafoscelis, N. difficilis, N. megalops, N. gracilis, N. microps, N. tenella, N. atlantica, and A^. lobata, are described in a comparative manner, and keys for their identifications are provided. The key to the larvae is based on structural differences in the carapace and rostrum of f urciUa stages, whereas the key to adults is mostly based on diagnostic features of the first thoracic leg (maxilliped) and a male secondary sexual structure, the petasma. Nematosceiis gracilis is represented by two distinct forms; they are considered ecophenotypes, since their patterns of geographical dis- tribution appear correlated with differences in environmental characteristics, particularly the dis- tribution of dissolved oxygen in the water column. Diagnostic features of these forms are pointed out. The antennule and carapace are sexually dimorphic in adults of all Nematosceiis. Abdominal pho- tophores in the males show species-specific patterns of enlargement. The genus Nematosceiis consists of seven species. It was described by G. 0. Sars (1883, 1885) as con- sisting of A'', megalops, N. tenella, N. microps, and A^. rostrata. Hansen synonymized A'', rostrata with A^. microps and added four species: A'^. gracilis and A^. atlantica in 1910, A^. difficilis in 1911, and A^. lobata in 1916. Nematosceiis lobata was not found by subsequent workers but the other species were discussed by Ruud (1936), Boden (1954), Boden et al. (1955) and Mauchline and Fisher (1969). Taxonomically Nematosceiis has been a difficult genus. Like other species of euphausiids, Nematosceiis species have been identified mainly on the basis of differences in the male copulatory organ, the pe- tasma. Since the petasma is an adult character, it has been difficult to identify immature specimens and mature females. Characters such as the shape of the eyes and structure of the second thoracic leg have been used to discriminate species. Einarsson (1942) showed structural differences in sper- mathecae (thelyca) of females, but such differences appear slight and are difficult to examine. Mauchline and Fisher (1969) pointed out difficulties encountered in the identification of species of this genus. In the present study an at- tempt is made to point out the diagnostic value of the first thoracic leg (maxilliped) in discriminating all species of Nematosceiis. This appendage 'Scripps Institution of Oceanography, La Jolla, CA 92037; present address: Hawaii Institute of Marine Biology, University of Hawaii, P.O. Box 1346, Kaneohe, HI 96744. Manuscript accepted October 1974. FISHERY BULLETIN; VOL. 73, NO. 4, 1975. usually remains attached to the animal caught by nets (as compared with the elongate second leg which is usually lost). It can be used as a diagnostic character in both sexes. The structural differences among the petasmae are also reexamined. The morphology of individual species will not be given separately, but species differences will be pointed out in a comparative manner. Since all species of this genus are sexually dimorphic, it is necessary to describe both sexes. Another aspect to which little attention has been paid is the significance of developmental features in determining phylogenetic associa- tions. Larvae of Nematosceiis are often difficult to separate as to species. The taxonomy of the larvae is yet to be worked out because the recognized adult characteristics are of no use in the larval identification. Gopalakrishnan (1973) summarized the available information on the sequential morphological development of an individual species. As Gordon (1955) and others have pointed out, larval characteristics may be more useful than those of adults in recognizing phylogenetic in- terrelationships of species. Adults show a greater degree of differentiation than the larvae, and their characters are more useful in the identifica- tion of the species than determining phylogeny. Moreover, part of the morphological variability observed in adults is sometimes ascribed to non- genetic modification that probably has no phylogenetic significance. Usually phylogenetic interrelationships are summarized in a classifica- tion. In this connection, larval characters are used 797 extensively in classifying insect groups, such as Diptera and Hymenoptera. It is the intention of this taxonomic study to provide as much informa- tion as possible so that one can examine the sys- tematic value of both larval and adult characters of Nematoscelis. MATERIAL AND METHODS The material used in this study consisted of 286 Isaacs- Kidd Mid-water Trawl (10 feet) collections and 1,950 plankton samples, including those collected during the International Indian Ocean Expedition (1960-65). These materials came from different geographical regions of the Atlantic, Indian, and Pacific oceans, mostly between lat. 40°N and 40°S. They are deposited at the Scripps Institution of Oceanography, La JoUa, Calif. All the trawl collections were not quantitative for es- timating species abundance. A paper on the dis- tribution and abundance of Nematoscelis based on plankton samples was already published (Gopalakrishnan 1974). Measurements were taken on 55 adult morphological characters for examin- ing relative degree of differences. Ten males and ten females of each species were selected for these measurements. Statistical significance was deter- mined on the basis of nonoverlapping confidence levels (95%) of the means. For making drawings, 10-20 individuals of each species were treated in heated 10% aqueous KOH to remove nonchitinous tissues. They were then stained in 1% aqueous Chlorazol Black E. Materials treated in KOH solu- tion could be kept in 60-80% glycerol without shrinkage. Drawings were made with the aid of a camera lucida fitted to monocular and binocular microscopes. The larval key was prepared on the basis of fur- cilia characters only. However, a few comments are made on characters of calyptopis and juvenile stages thought to have some diagnostic value. The adult key was prepared based on features of the first thoracic leg (maxilliped), eyes, antenna, and the carapace. Diagnostic features of the petasma are also included in the present key. Many characters are, therefore, used in the present key to facilitate its use on juveniles and adults of both sexes. Most of the commonly used adult characters are illustrated in Figure la. The terminology used here is the same that has been followed by most other workers. Larval terminology is defined and described in Gopalakrishnan (1973). FISHERY BULLETIN: VOL. 73, NO. 4 RESULTS Larval Development Between hatching and sexual maturity all species of Nematoscelis pass through four developmental phases: metanauplius, calyptopis, furcilia, and juvenile. A modified version of the nomenclature of larval development of euphausiids is given in Gopalakrishnan (1973). The metanauplius phase consists of one developmental stage, calyptopis of three (Cj, C2, and C^ and fur- cilia of three (F^, Fg, and F3). The strong differentiation of mouth parts and other thoracic legs shows similarities among larvae of all species of this genus. The development of larvae follows either of two pathways: A'', difficilis and A^". megalops follow one pathway and the other five species follow the other. During the third furcilia stage the second thoracic leg develops spines on both dactylus and propodus in N. difficilis and N. megalops, whereas in the rest of the species spines develop only on the dactylus. In all species of Nematoscelis this leg becomes the longest of all thoracic appendages. Other developmental differences between the two species groups are as follows: during the juvenile phase, the maxillules of A^. difficilis and A'', megalops develop pseudexopods from the posterior face of the coxa as the four-setose larval exopods disappear; but in the remaining species, at about the same stage, the larval exopod disap- pears without the development of a pseudexopod. The lacina externa (lobes of basis) of the maxilla is trilobed in larvae of all species, but becomes bilobed in adults of A'', difficilis and A^'. megalops and single lobed in adults of the remaining species. The differences in the sequential development of pleopods and telson spines (terminal) are sum- marized in Figure 2. These features are consistent and appear to be characteristic of each subgroup. The terminal spines of the telson in species of both subgroups show differences not only in their sequential reduction but also in their external morphology (Figure 3A). This structural difference can be seen even during calyptopis stages of all species of this genus. This is a diag- nostic feature and may have significance in un- derstanding the evolution of the larvae. The dorsal keel and rostrum of the carapace also appear to be of important diagnostic value for furcilia (Figure 3B) 798 GOPALAKRISHNAN: BIOLOGY AND TAXONOMY OF NEMATOSCELIS Figure l.-a. Nematoscelis atlantica, male (length = 13.2 mm; Indian Ocean, position: lat. 28°08'S, long. 66°09'E); b. N. microps, female Oength = 16.3 mm; Indian Ocean, position: lat. 10°06'S, long. 41°5rE). an, antennule; ca, carapace; car, carpus of second leg; da, dactylus of second leg; en, endopod of antenna; gi, gill; is, ischium of second leg; me. merus of second leg; mp. mandibular palp; pe, petasma; ph, photophore; pi, pleopod; pr, propodus of second leg; ro, rostrum; sc, scale of antenna; te, telson; thy, first thoracic leg; th2, second thoracic leg; ur, uropod. The furcilia phase of Nematoscelis was defined in Gopalakrishnan (1973) as follows: compound eyes no longer under carapace, but project outside; antenna retains larval natatory function; pleopod becomes functional, each appears first as non-se- tose rudiment which develops setae at following moult; furcilia, therefore, with different states of development of pleopods. The following key for identifying furcilia larvae to species is based mostly on diagnostic features of the carapace. There are three developmental stages in the fur- cilia phase. 799 FISHERY BULLETIN: VOL. 73, NO. 4 GENUS NEMATOSCELIS GO SARS N. LOBATA HANSEN. N GRACILIS HANSEN N MICROPS CO SARS N.ATLANTICA HANSEN N TENELLA GO SARS NDIFFICILIS N-MEGALOPS HANSEN GOSARS NUMBER OF TELSON SPINES (TERMINAL) SEQUENTIAL DIFFERENTIATION OF PLEOPODS FURCILIA FURCILIA 2 FURCILIA 3 NUMBER OF TELSON SPINES (TERMINAL) Figure 2.-Developmental sequence of differentiation of pleopods and terminal spines of the telson in species of Nematoscelis (adapted from Gopalakrishnan 1973). Key for Identifying Furcilia Larvae of Nematoscelis la. Each terminal spine of telson with a pair of conspicuous lateral subspines (Figure 3A,a); setules present only above these subspines. Fj stage with seven terminal spines and one pair of non-setose pleopods; Fg with five terminal spines, one pair of setose pleopods and three pairs of non-setose pleopods; F3 with three terminal spines, four pairs of setose pleopods and one pair of non-setose pleopods (Figure 2) 2 lb. Each terminal spine of telson without subspines, but with lateral setules along three- fourths of its length (Figure 3A,b). Fj stage with seven terminal spines and two pairs of non-setose pleopods; Fg stage also with seven terminal spines but two pairs of setose pleopods and three pairs of non-setose pleopods; F3 stage with five terminal spines and five pairs of setose pleopods (Figure 2) 5 2a. Rostrum rectangular with truncated anterior end (Figure 3B,b,c) 3 2b. Rostrum triangular with pointed anterior end (Figure 3B,a,d) 4 3a. Carapace keel very large and hump shaped; rostrum usually curved downwards (Figure 3B,b); length of sixth abdominal segment relatively short and its ventral margin largely convex A. tenella 3b. Carapace keel small, usually triangular shaped (Figure 3B,c), ventral margin ot sixth abdominal segment not convex A'', gracilis 4a. Rostrum broad and stough, keel large and platelike; carapace compressed anteroposteriorly (Figure 3B,a) N. microps 4b. Rostrum elongate and slender; keel less conspicuous (Figure 3B,d) N. atlanfica 5a. Rostrum broad, thick and triangular (Figure 3B,e) N. megahps 5b. Rostrum slender and narrow (Figure 3B,f ) N. difficilis Furcilia larvae of N. atlantica and N. lobata are complete the larval description of A^. lobata. This difficult to separate. More samples are necessary to species is known to occur only in the semi-isolated 800 GOPALAKRISHNAN: BIOLOGY AND TAXONOMY OF NEMATOSCELIS B Carapace lateral view Rostrum dorsal view 01 mm b c d f Figure Z.-A, terminal spines on lai^^al telson: a, Nematoscelis gracilis, N. microps, N. tenella, N. atlantica, and A'^. lobata; b, N. difficilis and N. megalops. B, carapace and rostrum of furcilia larvae of Nematoscelis: a, N. microps; b, N. tenella; c, N. gracilis; d, N. atlantica; e, N. m£galops;f, N. difficilis. Sulu and Celebes seas in the region of the Indo- Australian Archipelago (Gopalakrishnan 1974). The distribution of A^. atlantica appears not to extend into this area. Within each species, the body sizes of metanauplii ranged from 0.8 to 1.0 mm. However, body size differences among all species became apparent from the first calyptopis stage onwards (Table 1). It is at this stage that the larvae start feeding (Gopalakrishnan 1973). Evidently N. tenella larvae are the smallest, and N. megalops the largest. Length measurements of furcilia larvae also show size differences among Table l.-Body lengths (mm) of calyptopis stages of Nematoscelis. (Medians were based on 10 to .36 individuals of each species for each stage.) Stage 1 Stage 2 Stage 3 Species Median Range Median Range Median Range N. gracilis 1.4 1.3-1.4 2.1 2.0-2.2 2.6 2.5-2.7 N. microps 1.4 1.3-1.4 2.0 2.0-2.1 2.7 2.6-2.8 N. atlantica 1.5 1.4-1.5 2.2 2.1-2.3 2.8 2.6-2.9 N. tenella 1.2 1.2-1.3 1.8 1.7-1.9 2.5 2.4-2.7 N. megalops 1.6 1.5-1.7 2.4 2.1-2.5 3.2 2.9-3.3 N. difficilis' 1.4 1.3-1.5 2.0 1.9-2.1 2.8 2.2-2.8 'Data taken from Gopalattrishnan (1973). 801 FISHERY BULLETIN: VOL. 73, NO. 4 Table 2.-Body lengths (mm) of furcilia stages of six species of Nematoscelis. (Medians were based on 6 to 20 individuals of each species for each stage.) Stage 1 Stage 2 Stage 3 Species Median Range Median Range Median Range N. gracilis 2.8 2.6-3.0 3.0 2.9-3.4 3.3 3.2-3.5 N. microps 2.9 2.6-3.1 3.2 3.0-3.3 3.5 3.3-3.5 N. atlantica 3.4 3.2-3.5 3.9 3.9-4.1 4.3 4.2-4.5 N. tenella 2.7 2.6-2.9 3.2 3.1-3.3 3.5 3.5-3.6 N. wegalops 3.7 3.6-4.0 4.3 4.2-4.3 4.8 4.7-5.0 N. dilficilis* 3.1 2.7-3.2 3.5 3.4-3.7 3.9 3.6-4.2 'Data taken from Gopalakrishnan (1973). species (Table 2). Between the species pair N. megalops and A'', difficilis, there is a significant size difference during both the calyptopis and furcilia stages. Juveniles of all species of Nematoscelis are identified on the basis of their morphological similarities to adults, especially in such characters as the carapace, the rostrum, and the eye. During the juvenile stages of A'^. tenella, the carapace becomes elongate and narrow as in adults; the dorsal keel on the carapace elongates anteriorly and posteriorly; the broad and curved larval ros- trum becomes short and pointed; and the larval eye develops narrow upper and lower lobes (in the adult eye only the lower lobe remains narrow). These diagnostic features help to distinguish juveniles of N. tenella from similar stages of other species. The propodus of the first thoracic leg (max- illiped) is a useful character to identify juveniles, as it is in adults. Although the number of setae on the propodus of this leg is fewer in juveniles than in adults, it is possible to examine the differences in the "style" of setation among juveniles of Nematoscelis. For example, in adults of A^. gracilis there are two rows of setae on the propodus of the maxilliped (Figure 6c); in the juvenile stage this segment develops at least one seta from each position of these two rows. In A^. atlantica the same segment has only one row (marginal) of se- tae in adults, and in juveniles at least one seta is present at the position of this marginal row. This difference in the style of setation can be used to distinguish A'', gracilis and A^. atlantica juveniles. The propodus of the maxilliped of N. microps also has one row of marginal setae, but its inner mar- gin is convex; the carpus of this appendage is shorter than its propodus. The prominent dorsal keel on the carapace is a good diagnostic feature of A'', microps juveniles. Nematoscelis G. O. Sars— Generic Characters The shape of the rostrum variable in males and females; eyes large and bilobed; the peduncle of the first antenna slender in females and thicker in males. Dactylus of the first thoracic leg (max- illiped) triangular, flattened and furnished with comblike setae on its inner lateral margin. The second pair of legs greatly produced and with spines on the distal segment or on both the penul- timate and the distal segments. The endopod of the seventh leg biarticulate in the female, lacking in the male. Eighth leg a simple setose plate. All the four processes— proximal, terminal, lateral, and spine-shaped process— present in the petasma; the spine-shaped process always straight and the lateral without any hooks. Key for Identifying Adult Species of Nematoscelis The following key is adopted from Hansen (1910, 1912), Boden (1954), Boden et al. (1955), and Mauchline and Fisher (1969). It is modified to include additional information: la. Second pair of thoracic legs with long spines from both terminal segment (dactylus) and distal end of propodus. Third to sixth thoracic legs with three segments beyond knee. Maxillule with well-developed pseudexopod. Basis of maxilla bilobed. Ventrolateral spine on coxa of antenna greatly produced (Figure 4A,e,f ). Carapace with conspicuous cephalic ridge (Figure 5f ,g). Eyes large (Figure 4B,e,f ). Propodus of first thoracic leg with setae 802 GOPALAKRISHNAN: BIOLOGY AND TAXONOMY OF NEMATOSCELIS arranged in three rows (Figure 6e,f ). Proximal process of petasma with serrations in two rows (Figure 7a,b) 2 lb. Second pair of thoracic legs with long spines from terminal segment (dactylus) only. Third to fourth thoracic legs with only two segments beyond knee, and fifth and sixth with only one. Maxillule without a pseudexopod. Basis of maxilla with one lobe only. Ventrolateral spine on coxa of antenna small or greatly reduced (Figure 4A,a-d). Cephalic ridge on carapace absent or inconspicuous (Figure 5a-e). Eyes relatively small (Figure 4B,a-d). Propodus of first leg with setae arranged in one or two rows (Figure 6a-d). Proximal process of petasma without serrations, or when present in one row only (Figures 7c and 8) 3 2a. Proximal process of petasma reaching almost middle of serrated margin of terminal process (Figure 7a). Serrated part of terminal process slightly curved towards median lobe. Propodus of first leg usually with six setae in outer (dorsal) row and five in middle row A^. megalops 2b. Proximal process reaching much beyond middle of distal part of terminal process (Figure 7b). Distal end of terminal process greatly curved towards median lobe, reaching slightly over distal end of proximal process. Propodus of first leg usually with five setae in dorsal row and four in middle row N. difficilus 3a. Propodus of first thoracic leg with setae arranged in two separate rows (Figure 6c,d). Lateral process of petasma much longer than both terminal and spine-shaped processes (Figures 7c and 8a). Distal end of lateral process serrated 4 3b. Propodus of first thoracic leg dorsoventrally flattened and furnished with setae in one row only (Figure 6a,b). Lateral process of petasma much smaller than both terminal and spine-shaped processes (Figure 8b-d). No serrations on lateral process or any other processes of petasma 5 4a. Ventrolateral spine on coxa of antenna highly reduced to a hump (Figure 4A,d). Lower part of eye much smaller than upper part (Figure 4B,d). A long seta projecting from dorsal surface,of dactylus of first thoracic leg (Figure 6d). Distal end of lateral process always reaching beyond distal end of proximal process (Figure 8a) A^. tenella 4b. Ventrolateral spine on coxa of antenna not reduced to a hump (Figure 4A,a). Lower part of eye larger than or nearly equal to upper part (Figure 4B,a). No seta on dorsal surface of dactylus of first thoracic leg (Figure 6c). Distal end of lateral process not reaching beyond distal end of proximal process (exception: "old forms" from the Pacific Ocean to be discussed in a later section) (Figure 7c) N. gracilis 5a. Upper part of eye slightly narrower than lower part (Figures 4B,b and 9c). Propodus of first thoracic leg with less convex inner margin (Figure 6b and 9d). Keel on carapace less prominent and without conspicuous hump. Abdominal segments without any elevated dorsal keels. Shapes and relative lengths of processes of petasma as shown in Figure 8c,d. Lengths of propodus and carpus of first leg nearly equal 6 5b. Upper lobe of eye slightly wider than lower lobe; lateral evagination much deeper in upper than in lower lobe (Figure 4B,c). Propodus of first thoracic leg with highly convex inner margin (Figure 6a). Keel on carapace quite prominent and with conspicuous hump (Figures 5c and lb). Fourth and fifth abdominal segment characterized by less elevated dorsal keels. Carpus of first leg shorter than propodus. Shapes and relative lengths of processes of petasma as shown in Figure 8b A^. microps 6a. Proximal process thick; terminal process much shorter than both proximal and spine- shaped processes. Lateral process small. Median lobe greatly flattened and broad; its outer and inner margins broadly convex, inner margin forming an acute distal angle with outer margin (Figure 8c). Adult female without lateral denticle on carapace N. lobata 6b. Proximal process thin, a little shorter than terminal process; lateral process slightly curved toward median lobe, its distal end reaching to or almost to distal end of proximal process (Figure 8d). Adult female with lateral denticle on carapace A'^. atlantica 803 FISHERY BULLETIN: VOL. 73, NO. 4 B a """A '' J "'^*— _- .^ c ,,^_ e 1.0 m m B Figure \.—A , antenna (ventral view) and B, eye (lateral view) of Nematoscelis: a, N. gracilis; b, N. atlantica; c, N. microps; d, N. tenella; e, N. difficilis;/, N. megalops. en, endopod; sc, scale; sp, spine on proximal segment of protopod; ul, upper lobe; II, lower lobe. The present study shows that there are two dis- tinct forms in Nematoscelis (jracilis. They are distinguished as ecophenotypes and are referred to here as the "old form" and the "new form." The old form is identical in morphological characters with the typical form described by Hansen (1910) from the waters of the Indo-Australian Archipelago, and the new form is distinguished from the typical form on the basis of morphological differences on the proximal process of the petasma. There is also an apparent size difference between the two forms: Body length of the old form is significantly larger than that of the new form (cf. Figure 5a, b). The upper lobe of eye in the new form is slightly narrower than that in the old form. Both forms are distinguishable only as adults. The old form occurs mostly in the northern section of the tropical Indo-Pacific subregion and has maximum abundance in the oxygen-poor waters of the Arabian Sea, Bay of Bengal, and eastern tropical Pacific Ocean. The new form occurs in the region of the South Equa- torial Current. Along the equatorial zone, where the two forms overlap, an "intermediate" of the two forms, with regard to the length of the proximal process of petasma, is also encountered. Geographical distributions of these forms are described in Gopalakrishnan (1974). There are ap- parent morphological differences between the old forms of the Indian and Pacific oceans. The following key is prepared for identifying the forms of A^. gracilis: 804 GOPALAKRISHNAN: BIOLOGY AND TAXONOMY OF NEMATOSCELIS MALES FEMALES lOmm a-d LO mm e-g Figure 5.-Carapace of A^ematosce/ts: a, iV.gfracihs "old form" (length: male = 15.1 mm, female = 16.9mm);6, Ar.g'mct7is"newform" (length: male = 11.7 mm, female = 15.0mm);c,A^.ww'crops (length: male = 15.3 mm, female = 17.0mm);d,A'^.a(ton o .10 0.90 - 0.70 0.50 aa a ■ii* A AA A A • •• •ssf Nematoscelis gracilis A "new form" • "intermediate" o "old form"- Indian Ocean • "old form"- Pacific Ocean o oo l°° 8 o o ^o ° o 20 3.0 Carapace Length (mm) 40 Figure 10.— Nematoscelis gracilis forms: ratio of proximal process to median lobe of petasma plotted against carapace. propodus of first thoracic leg; number of marginal spine on lacina externa of maxillule; number of setae on second segment of mandibular palp. Sexual Dimorphism in Nematoscelis Sexual dimorphism in euphausiids was best documented by Hansen (1910, 1912). Einarsson (1942) and Nemoto (1966) provided further details. The most sexually dimorphic characters are: the lateral denticle, keel, and rostrum of carapace; antennule; eyes; sixth and seventh thoracic legs; first and second abdominal pleopods; preanal spine. The states of the lateral denticle, keel, and ros- trum in both sexes of Nematoscelis are illustrated in Figure 5. The carapace and rostrum are shorter in males than in females. The rostrum in the male is rarely variable; in the female it is always long and slender, except in A'^. gracilis and A^. tenella. The lateral denticle is absent in all species except in males of A'^. microps and both sexes of A^. atlantica. McLaughlin (1965) reported the oc- currence of this denticle in both subadult and adult stages of A^. difficilis caught from the northeast- ern Pacific Ocean. In the present study, the lateral denticle on the carapace was found only in immature specimens of this species, but not in adults. These individuals were collected from the North Pacific Drift and California Current areas. No sexual dimorphism was observed in the shape of eyes in Nematoscelis. The antennular peduncle in males has the two distal segments much thicker than in females; the second segment somewhat shorter and the third segment much shorter than in females (Figure llA,a,b). The lower flagellum of the antennule has the basal segment much thickened in males and furnished with tufts of sensory setae. In A^. gracilis males, the proximal part of this flagellum bends downward so as to accommodate the enlarged basal segment (Figure llA,a). Sexual dimorphism in abdominal photophores is characteristic of Nematoscelis. Einarsson (1942) pointed out a few examples of enlargements of abdominal photophores in this genus. James (1973) reported the existence of this feature in the North Atlantic species of N. tenella, N. atlantica, and N. microps. In the course of examining the material from all oceans, certain interspecific differences of photophore enlargement have become evident. In females of all species of Nematoscelis, the pho- tophores on each of the first four abdominal seg- ments are more or less alike in size and shape (Figure lb). However, in males, one or more of these photophores often show considerable en- largement. The patterns of this enlargement ap- pear to be consistent, species specific, and therefore of diagnostic value. Associated with photophore enlargement is the occurrence of paired chitinous saddle-shaped plates on the dorsal side of the abdominal segment anterior to that in which the photophore is enlarged (Figure IIB). Types of photophore enlargement in species of Nematoscelis are shown in this figure. In the Indian Ocean, A'^. gracilis males have the first abdominal photophore enlarged and lack chitinous plates on the dorsal side of the abdomen. Nematoscelis microps males have the second photophore enlarged and either with a dorsal hump (Figure llB,b) or paired chi- tinous plates on the dorsal side of the first ab- dominal segment. Taniguchi (1966) reported oc- currence of this hump on N. microps collected from the northeastern Indian Ocean. Humped males of this species are frequently found in the tropical 810 GOPALAKRISHNAN: BIOLOGY AND TAXONOMY OF NEMATOSCELIS A 1. 0mm A B Figure U.-A, dimorphic antennule of Nematoscelis gracilis: a, male; b, female. B, enlarged photophores of Nematoscelis males: a, N. gracilis forms; b, N. microps;c, first form of N. tenella and N. atlantica; d, second form of N. tenella and N. atlantica. ep, enlarged photophore; hu, dorsal hump; cs, chitinous "saddle." 811 FISHERY BULLETIN: VOL. 73, NO. 4 Indo-West Pacific subregion. Nematoscelis microps males with paired chitinous plates on the first abdominal segment occur mostly in the tropical regions of the Pacific, Atlantic, and Indian oceans. James (1973) found this form of N. microps occurring in the northeast Atlantic south of lat. 20°N. Adult males of N. microps without any photophore enlargement were also found in all oceans, but they occur mostly in the subtropical regions. Nematoscelis tenella and A^. atlantica males have two forms of photophore enlargement (Figure llB,c and llB,d): one form has the second and third photophores enlarged along with the presence of saddle-shaped paired plates on the dorsal side of both the first and second abdominal segments; the other form shows photophore enlargement on the fourth abdominal segment along with paired plates on the third abdominal segment. These two species occur together in the northern and southern subtropical provinces of all oceans. They also occur together in the tropical regions of the Atlantic Ocean. When the two species occur in the same geographical area, the males of both species do not show the same pattern of photophore enlargement. In such cases the pat- tern of photophore enlargement is as follows: Species Nematoscelis tenella Nematoscelis atlantica Subtropics Tropics First form * Second form * * Second form * * First form * *Saddle-shaped plates on first and second abdominal segments and photophore enlargements on second and third. '*Saddle-shaped plates on third abdominal segment and pho- tophore enlargement on fourth. Nematoscelis atlantica does not occur in the tropical regions of the Pacific and Indian oceans; A^. tenella males in this region have the second form of photophore enlargement. There is no clear evidence of photophore enlargement in N. difficilis and N. megalops. Even though a pattern of photophore en- largement appears to be characteristic of each species, not all mature males show this feature. When present, about 60-80% of the males in each sample had specific patterns of photophore en- largement as described above. Structure of the petasma of these forms did not differ from the typical forms. One is tempted to speculate on the evolutionary significance of this feature. The oc- currence of the dorsal chitinous plate and the enlargement of the photophore in adult males may have some joint functional importance, probably related to sexual behavior. Evidently, these ab- normal conditions are not random phenomena; patterns are of specific nature and have clear as- sociation with sexual maturity. The fact that N. tenella and A^^. atlantica do not have the same form of photophore enlargement when they occur together in a geographical area suggests the pos- sible role of these specific patterns in enhancing species recognition for mating. DISCUSSION Among the morphological characters of a species, feeding and reproductive structures af- ford specific features that have diagnostic value. The specificity of the feeding appendages reflects presumed niche specializations. The structural uniqueness in the reproductive system ensures reproductive isolation of the species upon which the biological species concept is founded (Mayr 1942). Therefore, a key based on these characters should be the best in distinguishing individual species. The selection of the maxillliped as a diag- nostic character has the advantage that the same key may be used for both sexes. However, it did not prove possible to make such a key for the larva. The species of Nematoscelis can be grouped into two subgroups, one with N. difficilis and N. megalops and the other with the rest of the five species. The present study has brought out both ecological and systematic evidences to support such a grouping. The sequential development of larval characteristics also suggests phylogenetic differences between the two subgroups. Hansen (1912), using the structure of adult maxillule, proposed a division of Nematoscelis into these two groups. A similar division was also made by Mauchline (1967) on the basis of structural differences in the adult maxilla. When closely related species are partially sym- patric, behavioral mechanisms might operate to insure reproductive isolation of the species (Mayr 1966). Usually the presence or absence of in- tergradation between sympatric populations serve as indicators of interbreeding or reproduc- tive isolation. Presumably, the absence of in- tergradation between species of Nematoscelis oc- cupying the same geographical area suggests reproductive isolation. However, the absence of reproductive isolation between the two forms of A^. gracilis is probably shown by the occurrence of sexually mature intermediate forms in the 812 GOPALAKRISHNAN: BIOLOGY AND TAXONOMY OF NEMATOSCELIS overlapping regions of their distribution. The fact that the observed diagnostic feature hes on the reproductive structure would suggest probable genetic separation. Nonetheless, until more is known about the ecology and behavior of these forms, I do not wish to formally describe them as species or subspecies and will consider them ecophenotypes. The pattern of geographical dis- tribution of these forms appears to be correlated with differences in environmental characteristics, particularly the distribution of dissolved oxygen in the water column (Gopalakrishnan 1974). Allopatric populations are inferred to have un- dergone reproductive isolation if they are morphologically distinct and do not show any overlap in their diagnostic features. Nematoscelis diffieiUfi. and A'^. megalops are allopatric, occupying the northern and southern transitional zones of the Pacific respectively. Nematoscelis megalops also occurs in the Atlantic and Indian oceans. On the basis of similarities in the structure of the petasma of A'^. difficUis and A^. megalops, Karedin (1971) questioned the validity of A^. difficilis. Brinton (1962) considered them a sibling species pair evolved as a result of complete geographical separation. Although closely related, there are certain morphological features that distinguish one species from the other. Both quantitative and qualitative features of the reproductive system, the petasma in this case, indicate significant differences. The petasmae of A^. megalops in the South Pacific, South and North Atlantic, and South Indian oceans show no apparent structural difference. These populations are probably in continuum facilitating gene exchange. (Com- munication between North and South Atlantic populations of N. megalops appears possible from the fact that only in this ocean does the characteristically subtropical species A", atlantica occur also in the tropics. This suggests that in the Atlantic the low-latitude boundaries of distribu- tion of subtropical species approach the equator, permitting at least occasional north-south com- munication). The validity of A", difficilis can be supported by the fact that this species and N. megalops do not overlap geographically in the Pacific Ocean. Both species live in comparable en- vironments (narrow mid-latitude zones) and therefore are probably exposed to similar selection pressures. In such a situation, even though geographical isolation would be complete, morphological differentiation might be slow. In the absence of gene exchange, the populations would be expected to have diverged genetically. The relative lengths and shapes of the median lobe, proximal, terminal, and lateral spines of pe- tasmae of the two species differ, supporting the validity of A^. difficilis. In Hansen's words (1916) these differences are "certainly so sharp, so im- portant, and so constant that they are sufficient for separating N. difficilis from A'^. megalops." The A'^. atlantica population in the North Pacific is spatially separated from its counterpart in the South Pacific Ocean. No morphological distinct- ness was evident in this population, although it may prove to be genetically separate from others. Nematoscelis lobata is endemic to the Sulu and Celebes seas in the Indo-Australian Archipelago (Gopalakrishnan 1974). According to Hansen's (1916) description, this species is very similar to N. microps. The present study indicates that in many morphological characters N. lobata is more related to A'^. atlantica than to A", microps. Nematoscelis lobata and A'', atlantica are allopatrically dis- tributed and have probably acquired characters which promote or guarantee their reproductive isolation. The observed differences in the structure of the first thoracic leg (maxilliped) of species of Nema- toscelis indicate presumed specialization in feed- ing habits. All species of this genus are recognized to be omnivores. From a comparison of the first thoracic legs, it appears that species would be ex- pected to show different types of feeding. Nema- toscelis microps, N. atlantica, and N. lobata, hav- ing marginal setae (one row) on their propodus of the first thoracic legs, may be better fitted for filtering a large proportion of phytoplankton in their food, whereas N. difficilis and A^. megalops, having three rows of setae on the propodus of the first leg, may select more animal food. In this re- spect A^. gracilis and A'^. tenella are intermediate. Existing information on the gut contents of species of Nematoscelis, e.g., Nemoto (1967) and Weigmann (1970), is inadequate to substantiate this. ACKNOWLEDGMENTS This work was supported in part by National Science Foundation Grant GA31783 and in part by the Marine Life Research Program, the Scripps Institution of Oceanogrpaphy's component of the California Cooperative Oceanic Fisheries Inves- tigations, a project sponsored by the Marine Research Committee of the State of California. I 813 FISHERY BULLETIN: VOL. 73, NO. 4 thank E. Brinton, M. M. Mullin, and P. Dayton for their comments on the manuscript. I also thank my wife, Vijaya, for her assistance in the preparation of figures. LITERATURE CITED BODEN, B. P. 1954. The euphasiid crustaceans of southern African wa- ters. Trans. R. See. S. Afr. 34:181-243. BoDEN, B. P., M. W. Johnson, and E. Brinton. 1955. The Euphausiacea (Crustacea) of the north Pacific. Bull. Scripps Inst. Oceanogr., Univ. Calif. 6:287-400. Brinton, E. 1962. The distribution of Pacific euphausiids. Bull. Scripps Inst. Oceanogr., Univ. Calif. 8:51-269. Einarsson, H. 1942. Notes on Euphausiacea I-III. On the systematic value of the spermatheca, on se.xual dimorphism in Nematos- celis and on the male in Bentheuphaiisia. Vidensk. Medd. Dan. Naturhist. Foren. Kbh. 106:263-286. Gopalakrishnan, K. 1973. Developmental and growth studies of the euphausiid Nematoscelis difficilis (Crustacea) based on rearing. Bull. Scripps Inst. Oceanogr., Univ. Calif. 20:1-87. 1974. Zoogeography of the genus Nematoscelis (Crustacea, Euphausiacea). Fish. Bull., U.S. 72:1039-1074. Gordon, I. 1955. Importance of larval characters in classification. Na- ture (Lond.) 176:911-912. Hansen, H. J. 1910. The Schizopoda of the Siboga expedition. Siboga Exped. 51(37):1-123. 1911. The genera and species of the order Euphausiacea, with account of remarkable variation. Bull. Inst. Oceanogr. (Monaco) 210:1-54. 1912. Reports on the scientific results of the expedition to the eastern tropical Pacific, in charge of Alexander Agassiz, by the U.S. Fish Commission steamer "Alba- tross," from October, 1904, to March, 1905, Lieut. Com- mander L. M. Garrett, U.S.N., commanding. 27. The Schizopoda. Mem. Mus. Comp. Zool. Harvard Coll. 35(4): 175-296. 1916. The Euphausiacean crustaceans of the "Albatross" Expedition to the Philippines. [Sci. Results Philippine Cruise Fish, steamer "Albatross," 1907-1910, No. 33.] Proc. U.S. Natl. Mus. 49:635-654. James, P. T. 1973. Distribution of dimorphic males of three species of Nematoscelis (Euphausiacea). Mar. Biol. (Berl.) 19:341-347. Karedin, E. p. 1971. On the similarity between Nematoscelis meqalops G. 0. Sars, 1885, and A^. difficilis Hansen, 1911 (Euphausiacea, Crustacea) and on the substantiation for the isolation of N. difficilis Hansen into a separate species. [Translated from Russian by Edith Roden.] Tikookean. Nauchno-issled. Inst. Rybn. Khoz. I. Okeanogr. (TINRO) 75:121-129. Mauchline, J. 1967. The feeding appendages of the Euphausiacea (Crus- tacea). J. Zool. (Lond.) 153:1-43. Mauchline, J., and L. R. Fisher. 1969. The biology of euphausiids. Adv. Mar. Biol. 7:1-454. Mayr, E. 1942. Systematics and the origin of species. Columbia Univ. Press,N.Y.,334p. 1966. Animal species and evolution. Belknap Press, Camb., Mass., 797 p. McLaughlin, P. A. 1965. A redescription of the euphasiid crustacean, Nema- toscelis difficilis Hansen, 1911. Crustaceana 9:41-44. Nemoto, T. 1966. Thysanoessa euphausiids, comparative morphology, allomorphosis and ecology. Sci. Rep. Whales Res. Inst., Tokyo 20:109-155. 1967. Feeding pattern of euphausiids and differentiations in their body characters. Inf. Bull. Planktology Jap., 61st Annu. No. 143-160. RuuD,J.T. 1936. Euphausiacea. Rep. Dan. Oceanogr. Exped. 1908-1910 Mediter. Adjacent Seas 2D6 (Biol.), 86 p. Sars, G. 0. 1883. Preliminary notices on the Schizopoda of H.M.S. Challenger Expedition. Forsch. Vidensk. Selsk. Krist. 7:1-43. 1885. Report on the Schizopoda collected by H.M.S. Challenger during the years 1873-76. Rep. Sci. Res. H.M.S. Challenger 13(37), 228 p. Taniguchi, a. 1966. Abnormal specimens of a euphausiid, Nematoscelis microps, collected in the Indian Ocean. Inf. Bull. Plank- tology Jap. 13:106-107. Weigmann, R. 1970. Zur Okologie und Ernahrungsbidogie der Euphausiaceen (Crustacea) im Arabishen Meer. ME- TEOR Forsch. Ergeb. 5:11-52. 814 FOOD, ACTIVITY, AND HABITAT OF THREE "PICKER-TYPE" MICROCARNIVOROUS FISHES IN THE KELP FORESTS OFF SANTA BARBARA, CALIFORNIA Richard N. Bray and Alfred W. Ebeling' ABSTRACT Diets, daily activity, and habitat preference were compared between the kelp perch, Brachyistius frenatus; the senorita, Oxyjulifi californica; and the white seaperch, Phanerodon furcatus, all of which co-occur in areas of reef and kelp off Santa Barbara, Calif. The kelp perch and senorita often clean ectoparasites off larger host fishes, whereas the white seaperch is a more generalized picker-type microcarnivore. The kelp perch and senorita, which co-occur in the kelp canopy, showed the least amount of total overlap in resource use, expressed as a combination of individual overlaps in food, activity, and habitat. The senorita had the narrowest breadth of diet but the widest breadth of habitat (within the kelp-bed areas). Senoritas and white seaperch ate mostly bryozoans encrusted on plants, whereas kelp perch ate mostly plankton and other tiny motile prey. As species, neither the kelp perch nor the senorita derives substantial amounts of food from cleaning, although some individual senoritas may. Unlike the two "cleaners," the white seaperch also ate substantial numbers of bottom prey. None of the species forage at night, when all are relatively inactive, and when the senorita actually buries itself in patches of rubble and sand on the reef. The two perches showed the greatest overlap in daytime activity, as measured both by bi-hourly counts of feeding bites in the field and of swimming movements in a laboratory tank. Fishes that exploit the same class of environmen- tal resources in similar ways may be thought of as forming a "guild" of species having similar ecological roles regardless of their taxonomic affinities (Root 1967). In and about the forests of giant kelp off southern California, fishes that can select relatively small prey from mid-water and from kelp or other surfaces form a foraging guild of "pickers" (cf. Hobson 1971). Hubbs and Hubbs (1954) stressed the fact that two common and taxonomically unrelated pickers have remarkably similar mouth structures and dentitions: the kelp perch, Brachyistius frenatus, which is in the primarily temperate family of surfperches Em- biotocidae, has evolved the same general type of pointed snout, tiny jaws, and protruding canines that characterize the senorita, Oxyjidis califor- nica, and most other members of the primarily tropical family of wrasses Labridae. Hobson (1971) noted that the habit of cleaning ectoparasites off larger fishes is widespread 'Marine Science Institute and Department of Biological Sciences, University of California, Santa Barbara, CA 93106. Manuscript accepted January 1974. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. among the picker-type fishes. Senoritas, kelp perch, and young of another embiotocid, the sharpnose seaperch, Phanerodon atripes, are the most consistent "cleaner fishes" of the kelp beds (Limbaugh 1961; Hobson 1971). Compared to some small tropical wrasses (Randall 1958), however, these species are less specialized as cleaners: their cleaning activities are sporadic and/or confined to certain individuals, and so their principal forage must be elsewhere (Hobson 1971). With this in mind, we compared the diets, daily activity patterns, and habitat preferences of the senorita and kelp perch, which are the principal cleaner fishes in the Santa Barbara area, with those of a more generalized picker, the white seaperch, P. furcatus. These three species have been studied off San Diego (Limbaugh 1955; Quast 1968a, b; Hobson 1971). Yet little has been published on their habits and distribution off Santa Barbara. Here the Channel Islands and the east-west oriented coastline protect kelp beds from swells, enabling giant kelp to anchor on low- relief and soft bottoms as well as to high-relief reefs. Also, species with centers of distribution located far to the north are more frequently en- countered (Quast 1968a; Ebeling et al. 1971). 815 FISHERY BULLETIN: VOL. 73, NO. 4 METHODS Kelp perch, white seaperch, and senoritas were observed in the field and laboratory. Over a 2-yr period, scuba divers spent more than a total of 125 h watching and collecting fish both day and night at depths ranging from 1 to 20 m. Study localities included areas of reef and kelp in the Santa Bar- bara Channel— off the Santa Barbara mainland and off Santa Cruz Island located across the Channel some 42 km to the south. Food During 27 scuba dives made between February, 1971 and March, 1973, we collected a total of 115 kelp perch (measuring 43-142 mm, averaging 103 mm standard length). 111 white seaperch (74-203, 139 mm), and 65 senoritas (110-227, 169 mm). Gut contents were found in and identified from 50 kelp perch, 55 white seaperch, and 53 senoritas. All fish were taken with a small, 15-prong pole spear. Later they were slit ventrally, fixed in 10% For- malin,- washed, and preserved in 45% isopropanol. During the analysis, fish were identified by num- bered tag only so as to minimize any bias that might result from forehand knowledge of time of collection, etc. For all three species, the simple, tubular gut, which is "stomachless" in the sense of Chao (1973), was excised, measured, and divided from front to back into three sections of equal length, here ar- bitrarily called the "fore-," "mid-," and "hindgut," respectively. Fullness of each section was scored subjectively from 1 (empty) to 5 (full and distend- ed). Scores were plotted against time of day. Because fish were sampled throughout the year, their times of collection were seasonally adjusted relative to times of sunrise and sunset as listed in solar tables for the particular dates. Adjusted time of collection, rounded to the nearest 2-h interval, was measured on the relative scale with sunrise arbitrarily set at 0600 h and sunset at 1800 h. Displayed under a dissecting scope, the contents of the foregut were sorted into broad taxonomic categories of food items, which were segregated in a small, partitioned tray. Then the percent volume of each item in the array was estimated by eye. Estimates were made quickly to the nearest per- cent, and their total over the array often exceeded -Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 100% per fish. Yet at the outset, independent es- timates of the same item did not vary substan- tially among successive trials, and series tended to regress toward a mean value (an observer's overestimation of volume on one trial was often countered by his underestimation of the same volume on the next). Item volumes were later standardized to 100%. In computing species means, fish with empty guts were not counted and all others were weighted equally, regardless of fish size or gut fullness (cf . Zaret and Rand 1971). The frequency of occurrence of a dietary item was expressed as the percent of fish with non- empty foreguts that contained the item. The rank order of item frequencies was highly correlated with that of volumes. Kendall's tau correlation coefficients for the kelp perch, white seaperch, and senorita were 0.51, 0.85, and 0.70, respectively {P4 mm) 3.3 18.8 — — 2.7 11.8 Nauplius larvae 0.2 2.1 — — — — Zoea larvae 4.3 18.8 — — 1.7 11.8 Fish eggs 0.2 2.1 — — — — From process of "cleaning" 2.9 — — — 5.4 — Parasitic copepods 1.4 22.9 — — 3.5 11.8 Gnathiid isopod larvae 1.4 2.1 — — 0.2 3.9 Fish scales 0.1 6.2 — — 1.7 7.8 Primarily substrate oriented: kelp, other algae, and bottom 45.0 — 81.8 — 87.1 — Free moving 36.2 — 18.1 — 9.7 — Mysids (kelp) 4.1 18.8 — — 0.4 7.8 Isopods 0.7 10.4 1.9 9.8 1.3 7.8 Gammarid amphipods 25.7 62.5 8.4 49.0 4.2 27.4 Caprellid amphipods 4.7 12.5 3.3 23.5 2.2 5.9 Decapod shrimps 1.0 8.3 3.2 5.8 0.4 5.9 Unident. crustaceans — — — — 1.2 5.9 Pycnogonids — — 1.3 5.8 — — Attached 9.0 — 63.7 — 77.4 — Bare kelp and other algae 1.5 12.5 0.3 1.9 5.6 15.7 Plant-encrusting bryozoans 7.3 10.4 63.4 63.4 65.5 80.4 Hydroids 0.2 2.1 — — 5.2 17.6 Serpulid worms — — — — 1.1 2.0 Primarily substrate oriented: bottom only — — 17.7 — 0.2 — Crushed shells and debris — — 12.6 27.4 — — Polychaete worms — — 3.7 5.8 0.2 2.0 Cumaceans — — 0.5 3.9 — Brittle stars — — 0.9 1.9 — — Small crustaceans that normally move freely on and about the kelp surfaces were almost equal to plankton in dietary importance. Gammarid amphipods, which may cluster just as abundantly about the kelp as in and about the tufted mat of plants and animals on the bottom, ranked second in overall abundance and frequency. Surprisingly unimportant were the so-called "kelp mysids," which are very abundant in the canopy and are commonly eaten by other fishes (Clarke 1971). 819 FISHERY BULLETIN: VOL. 73, NO. 4 Forage on attached organisms was less impor- tant. Cheilostomate bryozoans, principally Membra nipora ("plant-encrusting bryozoans"), ranked a distant third in overall volume. Memhranipora is the dominant bryozoan encrust- ing kelp, where it often covers large areas of the plant (Woollacott and North 1971), and most of the bryozoans in the gut contents were associated with bits of kelp blades. Kelp perch apparently ate no benthic prey. Cleaning activity was but a minor food source. Parasitic copepods, gnathiid isopod larvae, and fish scales were the only items likely to have been ingested in the process. The combined items never contributed more than 5% to the foregut food volume in a single fish. White Seaperch Virtually all prey of the white seaperch were substrate oriented, probably picked from off the kelp or bottom (Table 2). Plant-encrusting bryozoans predominated, and when present, averaged 85% of the foregut contents of individual fish. Moving prey, primarily amphipods and shrimps, were much less important. Many of the gammarid amphipods were quite small (<2 mm long); in one fish, e.g., all of 70 individuals did not fill the foregut. Only the white seaperch ingested appreciable amounts of bottom items. Crushed shells and sand particles, which often were cemented into tubes, ranked second in overall abundance and third in frequency. The remains of polychaete worms were found in but 3 of 14 sand-containing guts, which did, however, include substantial numbers of the gammarids that commonly inhabit such burrows in the tufted mat on the bottom. Relatively large amounts of loose sand in the mid- and hindguts indicated that these fish generally do not winnow non-food items in their mouth. Senorita Most of its prey was substrate oriented, proba- bly picked from off the kelp (Table 2). Like white seaperch, senoritas contained a predominance of plant-encrusting bryozoans, but unlike perch, had almost no bottom prey in the foregut. A third of all fish examined contained only the bryozoan Memhranipora encrusted on pieces of kelp, and bits of bare plant material were found frequently among the encrusted pieces; of a total of 18 categories of food items found in senorita guts, no other so dominated the contents of even a single fish. Hydroids, another item attached to plants and other substrates, ranked third in overall impor- tance. Moving prey, primarily amphipods, were less important. Unlike kelp perch, senoritas did not exploit plankton as a major source of food. Although some items occurred frequently, they contributed but little to the overall volume. Cleaning activity did not produce substantial forage, although it contributed relatively more to the diet of senoritas than to that of kelp perch. Of 10 adult senoritas, 142-184 mm long, that con- tained items likely to have been ingested during cleaning (parasitic copepods, gnathiid isopod lar- vae, and/or fish scales), the diets of seven were dominated by other food items. Ectoparasites and scales in guts of most senoritas were mixed with other food items, suggesting that the fish had both cleaned and foraged during the same day. However, guts of two of the remaining three fish contained nothing but parasitic copepods and scales. One specimen, collected at 1400 h, contained 465 fish scales, about 90% of the total contents, and 45 parasitic copepods. Both items were distributed more or less evenly throughout the length of the gut, indicating that this fish had cleaned during most of the day. Diel Forage All three species fed mostly, if not exclusively, during the day. Foreguts of kelp perch apparently were beginning to fill soon after dawn, were generally full by midmorning, and contained variable amounts of food through dusk (Figure 1). Of 54 day and 38 late-night (midnight-dawn) foreguts examined, 89% and 13%, respectively, contained food. Guts of white seaperch seemed to reach maximum fullness during midmorning and late afternoon. Of 64 day and 22 late-night foreguts examined, 88% and 4%, respectively, con- tained food. Fullness of mid- and hindguts of both species generally substantiated this daily cycle of feeding (Figure 2). Most foreguts were empty by midnight, when midguts still averaged at least half full and hindguts usually more. Then, by dawn most hindguts were empty, while foreguts were beginning to fill, a general pattern shown by fish whether collected during moonlit or dark nights. Senoritas seemed to feed actively through early afternoon, showing maximum gut fullness about 820 BRAY and EBELING; THREE "PICKER-TYPE" FISHES 5 4 3 2 t- (6) (17)^ (6)J» (5) (12) (8) (7) J9) 0600 1200 \J»^io)J.'S 1800 2400 0600 a> o o to 10 c 3 (;oiJ^U^ (7) (4K j;'- 'VI' {\oy (6) 0600 5r 4 3 (10) 1200 (3) 1800 2400 0600 \4I0J/ v{l4) (II) (7) -L. 0600 1200 1800 2400 0600 Time, h Figure 1. -Scored fullness (1, empty - 5, full) of foreguts of: top, kelp perch; middle, white seaperch; and bottom, seiiorita (which buries itself at night). Each point represents the mean value for (n) individuals collected over a 2-h interval. Time is measured relative to sunrise (0600 h) and sunset (1800 h-see text). midday. Of 65 diurnal foreguts examined, 78% contained food. At night the fish bury themselves in soft areas of bottom (see next section). However, six of seven guts of fish collected at dawn were completely empty. The duration of passage of food through the guts of the two perches was estimated from their diel feeding cycles. Assuming -that feeding stops at dusk when almost half the foreguts were full or nearly so, and that almost all hindguts have emp- tied by dawn when almost all were, the retention time is probably no more than 10-12 h. Activity Feeding Rate Field observations of feeding bites indicated Kelp perch White seoperch Senorita I u 0600 - 0800 -^^ 0800 - 1000 1000 - 1200 1200 - 1400 \//////^ r VyyyyyA- I V/////2^ --^m- I ^???^ T?????^ -^^ 1400 - 1600 I g^ 1 1600 - 1800 1800 - 2000 , ^^^ I -Y/y/yyA- 2000 - 2200 :^M 2200 - 2400 j'yyyy- ■H I K/yyyy^^ 2400 - 0200 0200 - 0400 0400 - 0600 fullness Figure 2.-Scored fullness of: foreguts (left open bars), midguts (middle hatched bars), and hindguts (right open bars) for the three fishes over 2-h intervals, beginning at dawn. Heights of the bars, scaled at the bottom of the figure, represent mean scores for the numbers of individuals indicated in Figure 1. that kelp perch fed frequently, at a maximum average rate of 20 bites/min around midday, decreasing to zero toward sunset (Figure 3). White seaperch and solitary senoritas fed much more slowly, at maximum rates of only 3.0 and 1.0 bites/min, respectively. Whereas both perches were seen feeding actively throughout the day, senoritas seemed to feed much less intensely after midafternoon. None of the particular individuals followed during the last two daytime intervals were seen to bite. During this time, however, a few other fish were observed picking away at bits of kelp. But this does not modify our general impression that, during the late afternoon hours, most serioritas feed much less actively than earlier in the day. 821 FISHERY BULLETIN: VOL. 73, NO. 4 o = bites • = grids 3 C to x> E 3 240 120 c E in (/) to o en E 3 0600 1200 1800 Time 2400 0600 Figure 3.-Feeding and swimming activity of; top, l Species S B B/S s b b/s Kelp perch White seaperch Sefiorita 6 5 6 6 6 6 4.19 4.59 5.53 0.70 0.76 0.92 H 6.0(2-6) 6.0 5.0(2-6) 3.52(NS) 2.47(1.16-4.63) 3.47(2.09-4.85) 4.41(1.95-5.82) 2.39(NS) 0.52(0.29-0.77) 0.59(0.35-0.81) 0.96(0.87-1.00) 11.23* 'Significant at P<0.005 824 BRAY and EBELING: THREE "PICKER-TYPE" FISHES Table 7.-Habitat breadths of the three fishes, measured relative to kelp-density, bottom-relief, and position-in-water-column classifications of the cinetransects (see text and Table 2). Sample size is the number of cinetransects from which the species was recorded; S is the number of habitat categories (maximum of 12) in which fish were photographed. See Table 4 for further explanation (note that the nature of habitat breadth precludes sample estimates). Species Sample size e B/S Kelp perch White seaperch Senorita 109 42 201 11 12 12 6.32 7.61 9.84 0.57 0.63 0.82 most in total resource use, with food the main contributor (Table 8). Yet their large food overlap was caused not by any overall similarity in dietary arrays, but by their sharing one predominating item, the plant-encrusting bryozoans. In fact, rank orders of their food items were not significantly correlated (tau = 0.20, P = 0.16). The kelp perch had the least amount of total overlap with others. Sharing the kelp-canopy area to a great extent, the kelp perch and senorita overlapped most in habitat even though the senorita had the broader overall spatial distribu- tion within the kelp bed. Also, the two species' small food overlap and different activity patterns tended to minimize their total overlap. Actually, rank orders of their food items were significantly correlated (tau = 0.50, P<0.001), because the two species shared similar proportions of a number of minor items. Yet overlap was small because they did not share the same predominating item. A low amount of overlap in total resource use was shown by the two perches. Although their activity pat- terns were similar, their diets and habitat preferences differed markedly. In diet, they shared neither a predominating food item nor an array of minor items, and rank orders of their items were uncorrelated (tau = 0.06). DISCUSSION The kelp perch and senorita-the two species most often involved in cleaning activity— co-occur to a great extent in the sunlit upper waters and eat few if any benthic prey. Kelp perch typically feed in loose aggregations of a few to over 30 in- dividuals. Constantly changing direction and depth, feeding individuals flit about to pick par- ticles from mid-water or, occasionally, from the surfaces of kelp and from other fishes. In calm, clear water, these aggregations often extend to the more open areas between kelp plants. In strong currents, however, the fishes gather in back of kelp columns where the water is quieter and where food swept off the surfaces of kelp may be consumed. Solitary senoritas occasionally nip at various substrates and large drifting particles, but they feed most intensely when in large schools. These schools move in and about kelp stands, momentarily dispersing for individuals to pick and tear at kelp fronds and encrustations, then re-as- sembling and moving on to another stand of kelp. The habit of kelp perch and senoritas of cleaning other mid-water fishes probably is incidental to the co-occurrence of the two species in the kelp canopy. Cleaning is not the principal occupation of either species (Limbaugh 1961, Hobson 1971). Their presence in the canopy better relates to their ability to pick small prey from off and from about the kelp blades. Seldom straying far from the heavy foliage where prey may become densely concentrated (Wing and Clendenning 1971), kelp perch also select plankton from incoming currents. Like other diurnal planktivores (Hobson 1974), the kelp perch has a relatively slender body, deeply forked caudal fin, and a slightly upturned mouth (Hobson 1971). Senoritas, which range more widely in the water column from canopy to near bottom, eat much less plankton. They favor at- tached food, primarily plant-encrusting bryozoans, either from the drift or torn from liv- ing plants. White seaperch, which usually range nearer the bottom and only clean occasionally (Hobson 1971), Table 8.— Overlap in resource use between members of all pairs of the three fishes. Activity overlap is the mean of two independent estimates: from feeding bites observed in the field, and from swimming movements observed in the laboratory. Habitat overlap is measured relative to kelp-density, bottom-relief, and depth classifications of the cinetransects (see Table 3). Total overlap is somewhere between the minimum and maximum estimates. Food F Activity Habitat H Total overlap Feeding bites T Swimming movements M Activity A = {T + M)/2 Overlap between: Minimum FAN Maximum (F+A+H)/Z Kelp perch and white seaperch Kelp perch and seiiorita While seaperch and senorita 0.25 0.21 0.92 0.92 0.85 0.72 0.95 0.79 0.78 0.94 0.82 0.74 0.63 0.79 0.74 0.15 0.14 0.50 0.16 0.61 0.80 825 FISHERY BULLETIN: VOL. 73, NO. 4 show more generalized behavior and are less specialized for picking than either kelp perch or senoritas. Like senoritas, white seaperch ate mostly plant-encrusting bryozoans, but their foraging behavior is quite different. White seaperch typically feed alone or in very small and loose aggregations. Feeding individuals often hover head down within 1 m of the substrate and, judging from their eye movements, search carefully for food. Even so, the substantial amounts of sand and other debris mingled with the more select items in white seaperch guts in- dicate that once the fish find their sedentary bot- tom prey, they engulf it in relatively large and indiscriminate mouthfuls. Underwater disturbances attract white seaperch and senoritas. For example, the two fishes commonly aggregate and feed where bat ray, Myliobatis californica, are stirring up the bottom with their wings. They are also quick to follow and assemble about actively working scuba divers. This seems to be an adaptation to forage opportunistically in the wake and disturbance left by others, a strategy which is commonly used by tropical wrasses (Hobson 1974). In contrast, the kelp perch appears to be much less aware of such disturbances and often seems oblivious of an ob- server at close range. Indirect evidence suggests that the plant material ingested with the bryozoans is not a primarily source of food for the fish. Only 10'^ of the ingested material was bare of bryozoans, in- dicating that white seaperch and senoritas select the heavily encrusted bits. Also, their relative gut lengths are less than expected for herbivores and many omnivores. Odum (1970) noted that the ratio of gut to fish length is usually less than unity in carnivores, one to three in omnivores, and greater than three in herbivores. Mean ratios from 74 white seaperch and 65 senoritas are only 0.82 + 0.028 (95% confidence interval) and 0.75 + 0.036, respectively. They do not differ significantly from the mean ratio of 0.76 ± 0.024 from 95 kelp perch, which ingest comparatively little plant material. Likewise, Chao (1973) found no evidence that the cunner, TautogolahrKs adspersu.s, a temperate labrid from off the Atlantic Coast, as- similates the algae it ingests. Small undigested amounts from the intestine of the cunner are usually associated with digested epiphytic animals, including bryozoans. Primarily a shellfish eater, the cunner also has a gut ratio that is less than unity. Individual diets of kelp perch and senoritas vary considerably from fish to fish, but this is not likely attributable to facultative cleaning. Diets of white seaperch, which do not commonly clean, were no more concordant than those of the other two species. Instead, opportunistic feeding in general may account for most of the variability. Kelp perch may switch from one patch of plankton to another, or feed on the kelp surface as the opportunity arises. Individuals were seen to dart back and forth between open areas and the kelp surface, selecting prey from either source. Although most sefioritas eat large amounts of bryozoans, many select small crustaceans, especially amphipods. Hobson (1971) noted that senoritas not only eat mid-water plankton, but are occasionally seen picking about on the bottom. We observed that they are usually among the first to arrive at un- derwater chumming stations where sea urchins are broken open. Yet cleaning contributes to the food breadth of kelp perch and senoritas by adding items that can be taken only by that process. And this points out a major problem in measuring food breadth by the "richness" or number-of-items measure, S. The categories of food items cannot be defined objec- tively, from the fish's point of view at least. For example, the total number of items recorded for the white seaperch would obviously increase if we further diversified the benthic categories (which are not exploited by the cleaners) by— say— distin- guishing gastropods from bivalves within the cate- gory of "crushed shells." Even though cleaning increases S, its total nutritional importance to the cleaner species may be negligible. Likewise, it is diflRcult to conclude whether or not cleaners have specialized diets. The total items eaten by either cleaner exceeded that eaten by the supposedly more generalized white seaperch, and the unsealed food breadth of the kelp perch was greatest of all three species. But the kelp perch and, to a lesser extent, the senorita are in fact limited to smaller items because they have smaller mouths. The 25 subjectively determined food cate- gories included some 15 "small items" (usually <3 mm in diameter) but only 10 "large" (usually >3 mm). Therefore, the diet of the kelp perch ap- peared to be relatively broad because it includes all of the small items, several of which are exclusively planktonic. On the other hand, the diet of the white seaperch, which rarely visits the canopy, appeared to be more narrow because it includes relatively few of these small prey. Yet having a 826 BRAY and EBELING: THREE "PICKER-TYPE" FISHES larger mouth, the white seaperch may eat not only small items, but also an array of items too large to be ingested by the other two. Other studies indicate that white seaperch forage opportunistically in a relatively broad range of kelp-bed and adjacent habitats. Although plant-encrusting bryozoans were by far their major food in the Santa Barbara areas of kelp and reef, they were of minor importance in fish collected off San Diego. Quast (1968b) reported that 18 fish from a kelp bed contained mostly small crustaceans, worms, and bivalves, while Hobson (1971) noted that 5 fish from shallow areas of surf grass contained small crustaceans, especially caprellid amphipods. DeMartini (1969) concluded that the white seaperch is almost "cosmopolitan" among habitats, including bays and artifacts far from the kelp beds. He observed that, unlike the kelp perch, it has uniformly broad and densely set pharyngeal teeth and commonly eats large, hard- shelled items like barnacles and clams. Although the kelp perch and senorita have superficially similar feeding mechanisms, they do not overlap broadly in their diets. Off Santa Bar- bara, in fact, food overlap is least between senorita and kelp perch and greatest between senorita and white seaperch, whose mouth structure and denti- tion are more generalized. These relations prevail because the kelp perch does not eat substantial amounts of the plant-encrusting bryozoans, the overwhelmingly predominate food item of the other two. Disregarding bryozoans, the remaining (minor) food array of the senorita actually resem- bles more closely that of the kelp perch than that of the white seaperch. Likewise, off San Diego, kelp perch favor copepods and gammarid amphipods (Quast 1968b), and senoritas favor bryozoans (Quast 1968b; Hobson 1971) but may eat a variety of small crustaceans associated with giant kelp as well (Limbaugh 1955). Because food overlap between the two cleaners is effectively small, they may co-occur with minimal mutual interference, even though their habitat overlap in the upper kelp bed is relatively broad. Also, their daytime activity patterns differ noticeably. Whereas kelp perch dart sporadically among the kelp blades and seem to feed almost continuously, senoritas move continuously about in open water as well as in dense kelp and seem to feed more sporadically. Also, solitary kelp perch continue their rapid picking about well past mid- afternoon after senoritas were observed to curtail their feeding activity. It would seem that the senorita and white seaperch are greater potential competitors because they overlap almost completely in both food and habitat within the kelp-bed area. But even so, it is doubtful that availability of their principal food, bryozoans, is a limiting factor in the Santa Barbara area, where encrustations are widespread over the kelp and other substrates. Furthermore, the frequency of occurrence of white seaperch within the kelp bed is quite low compared to that of the senorita. Even though fairly large aggregations are seen occasionally over the reef, the center of abundance of white seaperch may be in more peripheral areas where alternative prey are readily available. The senorita, which belongs to the large tropical family of wrasses, is more specialized in diel behavior than are the kelp perch and white seaperch. Whereas at night the perches simply slow down and become less responsive, the senori- ta buries itself in pockets of sand or gravel on the reef. Wrasses in general are strickly diurnal: they seek cover and become quiescent at night, as has been observed for tropical species (Hobson 1965, 1968, 1972, 1974; Stark and Davis 1966; Collette and Talbot 1972; Smith and Tyler 1972) and for other temperate species (Chao 1973; 011a et al. 1975). Various species hide in holes, bury them- selves, and/or protect themselves with mucus en- velopes (Hobson 1965, etc.). In the tropics, they are among the first fishes to take cover at dusk and the last to emerge at dawn, a practice that may minimize their vulnerability during the crepus- cular hours when predation is most intense (Hob- son 1968, 1972; Collette and Talbot 1972). In the kelp beds of the temperate zone, there may be relatively few nocturnal piscivores as compared with the tropics. Thus, the senorita may retain the burying habit of its family (which implies a complex genetic basis) simply because there are no pressures actively selecting against this trait (cf . Hobson 1972). Many tropical "pickers" have elongated snouts and small mouths with projecting teeth for select- ing and removing tiny prey from otherwise inac- cessible places (Alexander 1967; Hobson 1968, 1974). These are also adaptations for picking ec- toparasites from larger fishes, and indeed many of the small and sharp-nosed tropical-reef fishes are part-time or "facultative" cleaners (Hobson 1971, 1974; Losey 1972). Likewise, the tendency of kelp perch and senoritas to clean may vary among sit- uations or individuals (Hobson 1971) and may 827 FISHERY BULLETIN: VOL. 73, NO. 4 provide most fish w^ith only a minor dietary supplement. CONCLUSIONS In and about kelp beds off Santa Barbara, the kelp perch, the senorita, and to a lesser extent the white seaperch, belong to a foraging guild of picker-type microcarnivorous fishes. Throughout the year, the kelp perch and senorita, which com- monly pick ectoparasites from larger fish, spend most of the day in the sun-lit upper waters in and about the kelp canopy. Here they can discern and pick small prey from various surfaces and from the open water. A more generalized picker, the white seaperch, occurs a bit deeper in the water column and, unlike the two cleaner fishes, eats substantial amounts of benthic prey. Even though the two co-occurring cleaner fishes have superficially similar feeding mechanisms, they seem to minimize mutual interference in resource use by foraging in somewhat different ways. Thus their total overlap in resource use is relatively small because the kelp perch feeds ac- tively all day and does not eat substantial amounts of plant-encrusting bryozoans, the predominate staple of the senorita and white seaperch. Within the kelp-bed area, the senorita has the widest habitat breadth. It broadly overlaps the white seaperch's range below the canopy and near the bottom. Their sharing of food and habitat would seem to make these species the greater po- tential competitors. But even so, they may seldom actually interfere with one another because the white seaperch is not limited to the kelp bed and occurs there less frequently than the senorita. None of the three species forages at night, when all are relatively inactive and the senorita buries itself in soft substrates on the reef. Neither the kelp perch nor the senorita obtains substantial amounts of food from cleaning, although some individual senoritas may. Of the two species, the senorita is more specialized in its diel behavior and may be somewhat more nutri- tionally dependent on cleaning. ACKNOWLEDGMENTS We thank Stevan Arnold, Edmund Hobson, Michael Neushul, and two anonymous reviewers for critically reading the manuscript and offering many helpful suggestions. We thank Ralph'Lar- son, Fred Steinert, and James Cook for assisting the diving operations; Sharon Horn for drafting the illustrations; and Norm Lammer for his in- valuable technical assistance with equipment and boating operations. This study was supported by the following grants: NSF GA 38588 and Sea Grants GH 43 and GH 95; and USDC Sea Grants 2-35208-6 and 04-3-158-22, R-FA-14. LITERATURE CITED Alexander, R. M. 1967. Functional desig^n in fishes. Hutchinson and Co., Lond., 160 p. BORTONE, S. A. 1971. Studies on the biology of the sand perch, Diplcctrum forrniisum (Perciformes: serranidae). Fla. Dep. Nat. Resour. Tech. Serv. 65;l-27. Brown, D.W. 1974. Hydrography and midwater fishes of three contiguous oceanic areas off Santa Barbara, California. Los Ang. Cty. Mus. Contrib. Sci. 261:1-30. Chao, L. N. 1973. Digestive system and feeding habits of the cunner, Taufogolahruf: adspersus, a stomachless fish. Fish. Bull., U.S. 71:565-586. Clarke, W. D. 1971. Mysids of the southern kelp region. In W. J. North (editor). The biology of giant kelp beds (Macrocystis) in California, p. 369-380. Nova Hedwigia 32. 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. Bull. Los Ang. Cty. Mus. Nat. Hist. Sci. 14. DeMartini, E. E. 1969. A correlative study of the ecology and comparative feeding mechanism morphology of the Emhiotocidae (surf-fishes) as evidence of the family's adaptive radia- tion into available ecological niches. Wasmann J. Biol. 27:177-247. Ebeling, a. W., W. Werner, F. A. Dewitt, .Jr., and G. M. Cailliet. 1971. Santa Barbara oil spill: Short-term analysis of macroplankton and fish. U.S. Dep. Commer., Environ. Prot. Agency Off. Water Qual. Proj. 15080 EAL, 68 p. Hobson, E. S. 1965. Diurnal-nocturnal activity of some inshore fishes in the Gulf of California. Copeia 1965:291-302. 1968. Predatory behavior of some shore fishes in the Gulf of California. U.S. Fish Wildl. Serv., Res. Rep. 73, 92 p. 1971. Cleaning sybiosis among California inshore fishes. Fish. Bull., U.S. 69:491-523. 1972. Activity of Hawaiian reef fishes during the evening and morning transitions between daylight and dark- ness. Fish. Bull., U.S. 70:715-740. 1974. Feeding relationships of teleostean fishes on coral reefs in Kona, Hawaii. Fish. Bull., U.S. 72:915-1031. Horn, H.S. 1966. Measurement of "overlap" in comparative ecological studies. Am. Nat. 100:419-424. HUBBS, C. L., AND L. ('. HUBBS. 1954. Data on the life history, variation, ecology, and rela- tionships of the kelp perch, Braclu/isfiux frenatus, an 828 BRAY and EBELING: THREE "PICKER-TYPE" FISHES embiotocid fish of the Californias. Calif. Fish Game 40:183-198. Levins, R. 1968. Evolution in changing environments. Princeton Univ. Press, N.J., 120 p. LiMBAUGH, C. 1955. Fish life in the kelp beds and the effects of harvest- ing. Univ. Calif. Inst. Mar. Res., IMR Ref . 55-9, 158 p. 1961. Cleaning symbiosis. Sci. Am.205(2):42-49. LosEY, G. S., Jr. 1972. The ecological importance of cleaning symbiosis. Copeia 1972:820-833. MacArthur, R. H. 1972. Geographical ecology: Patterns in the distribution of species. Harper & Row, Publishers, Inc., N.Y., 269 p. Odum, W. E. 1970. Utilization of the direct grazing and plant detritus food chains by the striped mullet Mugil cephalui^. In J. H. Steele (editor), Marine food chains, p. 222-240. Univ. Calif. Press, Berkeley and Los Ang. Olla, B. L., a. J. Bejda, and A. D. Martin. 1975. Activity, movements, and feeding behavior of the cunner, Tautoqolahrus adsperxxi^, and comparison of food habits with the young tautog, Taiitoga onitia off Long Island, New York. Fish. Bull., U.S. 73:895-900. Pianka, E. R. 1974. Niche overlap and diffuse competition. Proc. Natl. Acad. Sci. U.S.A. 71:2141-2145. QUAST, J. C. 1968a. Fish fauna of the rocky inshore zone. In W. J. North and C. L. Hubbs (editors). Utilization of kelp-bed resources in Southern California, p. 35-79. Calif. Dep. Fish Game, Fish Bull. 139. 1968b. Observations on the food of the kelp-bed fishes. In W. J. North and C. L. Hubbs (editors). Utilization of kelp-bed resources in Southern California, p. 109- 142. Calif. Dep. Fish Game, Fish Bull. 139. Randall, J. E. 1958. A review of the labrid fish genus Labroides, with descriptions of two new species and notes on ecology. Pac. Sci. 12:327-347. Root, R. B. 1967. The niche e.xploitation pattern of the blue-gray gnatcatcher. Ecol. Mongr. 37:317-350. Simpson, E. H. 1949. Measurement of diversity. Nature (Lond.) 163:688. Smith, C. L., and J. C. Tyler. 1972. Space resource sharing in a coral reef fish community. In B. B. Collette and S. A. Earle (editors). Results of the Tektite program: Ecology of coral reef fishes, p. 125-170. Bull. Los Ang. Cty. Mus. Nat. Hist. Sci. 14. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. W. H. Freeman & Company, San Franc, 776 p. Starck, W. A., II, AND W. p. Davis. 1966. Night habits of fishes of Alligator Reef, Florida. Ichthyol. Aquarium J. 38:313-356. Tate, M. W., and R. C. Clelland. 1957. Nonparametric and shortcut statistics in the social, biological, and medical sciences. Interstate Printers and Publishers, Inc., Danville, 111., 171 p. Wing, B. L., and K. A. Clendenning. 1971. Kelp surfaces and associated invertebrates. In W. J. North (editor). The biology of giant kelp beds (Macrocystis) in California, p. 319-339. Nova Hedwigia 32. Woollacott, R. M., and W. J. North. 1971. Bryozoans of California and Northern Mexico kelp beds. //( W. J. North (editor). The biology of giant kelp beds (Macrocystis) in California, p. 455-479. Nova Hed- wigia 32. Zaret, T. M., and a. S. Rand. 1971. Competition in tropical stream fishes: Support for the competitive exclusion principle. Ecology 52:336-342. 829 FISHERY REGULATION VIA OPTIMAL CONTROL THEORY^ William J. Palm^ ABSTRACT This paper attempts to show how control theory can be used to formulate a regulatory scheme for fisheries. The regulatory mechanism considered is a limit imposed on fishing effort. It is shown that static optimization methods, such as maximum equilibrium yield analysis, need to be supplemented with dynamic methods, such as optimal control theory, which take into account the variable nature of a fishery. The dynamic analysis is used to show that the size of a limit on effort should be a feedback function of the variables in the state of the fishery. The concept of the Linear-Quadratic Optimal Control Problem is introduced as a method for devising such a feedback scheme for fishery regulation. A single- variable logistic model is used to introduce the basic concepts. A model with three variables is then analyzed to show how the techniques are easily extended to the general multivariable case. Details of the general method are given in an Appendix. The need for fishery regulation is apparent and will become even more important with the es- tablishment of resource management zones off our coasts. Regulatory mechanisms include catch quotas and limits on fishing effort (number of boats permitted entry into the fishery, number of hooks used, etc.). A mathematical model of the fishery, which includes biological and perhaps economic factors, is useful for determining the best regulatory scheme. Some of the more familiar examples of these models are given by Schaefer (1954, 1968), Beverton and Holt (1957), Ricker (1958), Larkin (1963, 1966), Pella and Tomlinson (1969) and Fox (1970). The above models are said to be dynamic because they utilize differential equa- tions to describe how the fishery changes with time. The inclusion of economic factors, multiple species, and other biological variables, such as size and age, results in multivariable models which are quite complex. Much of the analysis of fisheries is based on the concept of an equilibrium. Perhaps the best known is the maximum equilibrium yield analysis. However, equilibrium is an idealization and is never actually encountered in reality because con- tinually changing environmental infiuences act as disturbances which displace the system from its equilibrium condition. For unstable systems this is disastrous because equilibrium is never regained. 'Part of this work is a result of research sponsored by NOAA, Office of Sea Grant, Department of Commerce under Grant #9X-2()-6807C. Department of Mechanical P]ngineering and The Institute of Knvironmental Biology, University of Rhode Island, Kingston, HI 02881. and for stable systems with large time constants, the return to equilibrium might take so long as to negate the assumptions and usefulness of the equilibrium-based analysis. Thus "static" or equilibrium-based analysis should be supplement- ed with dynamic methods which take into account the variable nature of the fishery. A purpose of this paper is to show that the above considerations in- dicate that any regulatory scheme should contain "feedback"; that is, the size of any quota or limit should be a function of the state of the fishery. Also, the concept of the Linear-Quadratic Optimal Control Problem will be introduced as one way of devising such a feedback scheme for fishery regulation. The Linear-Quadratic Optimal Control Problem, which has been widely applied in engineering, is one method within the larger framework of op- timal control theory. Other optimal control methods have recently been applied to problems in fishery management which are unlike the problem treated here. Goh (1969, 1973) applied the so-called "singular" control method to the problem of maximizing yield with a single-species model. Saila (in press) describes Goh's results in more detail. Clark et al. (1973) analyze the problem of optimal reduction of effort in an overexploited fishery. They calculate the fishing mortality func- tion which maximizes the total present value of all profits and utilize a Beverton- Holt model for the fishery. Clark (1973) has presented a similar analysis for a logistic fishery model. The above three analyses lead to control functions which have been loosely described as a "bang-bang" control Manuscript accepted Januar\' 1974. FISHERY BULLKTIN: VOL. 73, NO. 4, 1975. 830 PALM: FISHP:RY REGULATION VIA CONTROL THEORY because the optimal values of the control variable lie at its boundaries. Thus the control variable switches between a lower value (usually zero) and an upper value which might be difficult to specify. There are advantages as well as limitations with the linear-quadratic approach as compared with the bang-bang control approach. With the linear- quadratic approach, quantities such as yield and present value of profits are not directly maximized to obtain the feedback control function, as is done with the bang-bang approach. Rather, the maximization is first done with static methods, and then a feedback control function is construct- ed to keep the system near the resulting equilibrium condition. To do this, the system equations are linearized about the equilibrium. If disturbances carry the system far from equilibrium, the linearization breaks down. However, this is generally not a serious limitation, since the feedback control function is designed to counteract disturbances and to keep the system near equilibrium. The method is not restricted to equilibrium analysis, and frequently the two approaches are combined by using bang-bang control methods, instead of static methods, to compute an optimal "open-loop" control function. Linearization of the model around the resulting trajectory enables the linear-quadratic method to be used to synthesize a closed-loop (feedback) con- trol function to keep the system on the optimal trajectory (Ho and Bryson 1969). A significant advantage of the linear-quadratic approach is that it allows the use of linear control theory, whose techniques are more highly developed and easier to apply than the nonlinear techniques required for bang-bang control analysis. Powerful methods of compensating for incomplete information, uncertainties in measurements, model parameters, and model structure are available for the linear-quadratic approach but are scarce for the bang-bang control approach. Also, solutions to bang-bang control problems are extremely difficult to obtain if the model contains more than two variables. First a single-variable model is used to illustrate the basic concepts. A model with three variables is then analyzed to show how the techniques are easily extended to the general multivariable case. The details of the general method are in the Ap- pendix. There it is also shown in more detail why static optimization methods, such as linear programming, and dynamic optimization methods, such as optimal control theory, should not be treated as competing methods but rather should be used together as part of the total approach to the problem because they are mutually complementary methods. This is mentioned because there is a tendency among economics- oriented analysts to use static methods, whereas analysts with control- theory backgrounds tend toward dynamic methods. It is assumed that the reader is familiar with the fundamentals of differential equations and ma- trix operations. A matrix will be denoted by brackets [ ]; a matrix transpose by [ ]^; and a column vector by a bar underneath, as x. SINGLE- VARIABLE MODEL The following model is the Schaefer or logistic model: dN ^=-aN-bN'~-qfN, (1) at where A^ is the biomass or number of catchable fish in the fishery, t is time, q is the catchability coefficient, and/is the fishing effort. The constant a is the intrinsic rate of natural increase of the population, and the constant h is related to the carrying capacity of the environment c by the relation: b — ale. The system's equilibrium (A' /eq) is found by setting the derivative in Equation (1) equal to zero: 0 = oA^eq - ^^'-eq - ^/'eq^eq- The equilibrium yield Y ^^ is: >^eq = '//eq^'eq = (« "^A^eq ) ^^eq • To find the maximum equilibrium yield, we differentiate Y ^ with respect to A'^^, and set this result equal to zero. eq Solving this for the population size and fishing effort corresponding to maximum equilibrium yield, we obtain: N^ = a/26. .' eq {a - />A\,, ) A^,„ a «) ' eq q N eq 2q. 831 FISHERY BULLETIN: VOL. 73, NO. 4 Note that a static optimization method, calculus, has been used to find an optimal equilibrium. To analyze the model's behavior in the vicinity of the equilibrium point. Equation (1) is linearized by expanding the right-hand side in a two-variable Taylor series about the point (A^eq-Zeq)' ^.nd keep- ing only the first-order terms (see Appendix). This gives: d£ dt («I'*(t" where: g = aN - bN^ x = N- N\ qfN " =f-fe^- eq (2) (3) The new variables .r and u are the deviations in population density and fishing effort from their equilibrium values. After evaluating the deriva- tives of g at the equilibrium point, we obtain: d£ dt aq 2h (4) If the fishing effort is kept constant at its equilibrium value, then n = Oand dx _ a It ~ "2'^' This system is stable for all positive values of a, which means that if disturbed from equilibrium, the population will eventually return to it. The solution is: x(0 = .r(ge" 2 "-'"', where x(f„) is the deviation at time f„ . Since the time constant for this system is 2/o, it will take that amount of time for the deviation to decay by 6S9r and for four time constants to decay by 98%. If the constant a is small, this time can be very large. Also, by keeping the fishing effort constant, we cannot take advantage of higher yields obtainable when x{to)>0, and risk overexploiting when x{t„)<0. We will show that by making the fishing effort a function of population level, we can change the system's time constant and also avoid the above difficulties. For example, the results of Schaefer (1954) give the following values for the Pacific halibut: a h 0.67 3.05 X 10-9 3.95 X 10-5 where A^ is in pounds, t in years, and./Mn number of skates. A standard skate of halibut gear consists of eight lines of 300 feet each in length, with shorter lines with hooks attached at 10-foot inter- vals (Carrothers 1941). The time constant is 3 yr, and thus 12 yr are required for a deviation in population to disappear, assuming no other dis- turbances act during that time. If we now specify, by means of a "performance index," that we wish to keep A'^ near Aeq while minimizing the variation in./' required to do so, we can design a regulatory procedure which will keep the fishery near the maximum equilibrium yield condition. The performance index J which specifies this desire is the so-called quadratic index: J = J (Q.r2 + Rn^ dt. The squared terms indicate that we make no dis- tinction between positive and negative deviations from equilibrium. The positive constants Q and R are the weighting factors which indicate the rela- tive importance placed on keeping N near Agq (.r near 0) versus keeping./' near ./'gq {n near 0). The infinite upper limit indicates that we are interest- ed in long-term as well as short-term effects of our fishing effort regulation. The problem of determining the function u, which minimizes the performance index, is solved by the application of optimal control theory. Since the system. Equation (4), is linear, and the index is quadratic, the problem formulated above is referred to as the Linear-Quadratic Optimal Con- trol Problem. The solution for the control function is (see Appendix): -Kx (5) K = 1 R 26 117 p where Pis the positive steady-state solution of the so-called Riccati equation: dP dt = -aP R W^' + Q 832 PALM: FISHERY REGULATION VIA CONTROL THEORY with initial condition P(0) is: 0. Tlie solution for P P = aR + a '^R'^ + qhi-RQ/h- q^a^2b" Thus P and K are functions of the weighting fac- tors R and Q, which must be specified. Note that this method yields three results: 1) that the optimal control function for u is a linear function of x (a linear feedback law); 2) the means to calculate the feedback gain K, once R and Q are specified; and 3) that A' is negative in this example (we assume that a, h, and q are positive). The third result indicates that the control law, Equation (5), opportunely calls for an increase in fishing effort when the population increases (.r > 0), and con- servatively calls for a decrease in effort when the population decreases (,r<0). In this simple single-variable case we can utilize the first result and avoid specifying R and Q by substituting wfrom Equation (5) into Equation (4). The result is: t \2b Zj 26 2 The time constant for this system is: T = a 2" 2b ^ (6) Using this approach it is possible to choose Kso as to give a desired value of the time constant. Alternately, K may be chosen by specifying the magnitude of the deviation in fishing effort we will allow in order to counteract an expected deviation in population level. Written in terms of magnitudes, Equation (5) becomes: ■r. where: .f,„ = maximum magnitude expected for x u„, — maximum magnitude specified for u. Once K has been determined,./' as a function of A^ can be found by substituting .r and u from Equa- tions (2) and (3) into Equation (5) to obtain: / = f - K{N ^eq). (7) To evaluate the effects of the above regulation scheme under various conditions, the above expression is substituted into Equation (1), which can then be solved by computer for A^ and / as functions of time. As an example with the previously mentioned results of Schaefer (1954) for the Pacific halibut, a maximum deviation in N of 5% from A^eq was pos- tulated, and a maximum deviation in /of 5% from ./'eq was specified. Thus: A^ eq a/26 = 1.098 x lO^ ./;„ = a/2q = 8.48 x 10^ eq x„, = O.OSA^eq u„, = 0.05/ eq- Using the second method for computing K, we obtain: K = - 0.05/-, eq 0.05A^ = -0.772 X 101 eq From Equation (6) the new time constant is found to be 1.5 yr, which is one-half the value for the case without feedback control. The fishing effort found from Equation (7) is: '■^ (»-fe) =-A^= 0.772 q X 10 -4 A^. (8) In view of the impossibility of continuously and instantly measuring population size and varying fishing effort, /as given by Equation (8) was in- terpreted as follows. It was assumed that a limit is imposed on fishing effort at the beginning of each year and held constant during that year, and its value / is calculated from Equation (8), with A^ being the average population over a yearly inter- val terminating three-tenths of a year before the imposition of the new limit. That is, three-tenths of a year is allowed for collecting and analyzing the population data used to calculate the next year's limit. With this discretized version of / computer simulation results show that the system time constant is 1.8 yr, which is reasonably close to the 1.5 yr predicted by the continuous model. Thus it is possible to use the analysis based on the con- tinuous model in the realistic situation involving data-collection limitations and limit-imposition constraints. THREE-VARIABLE EXAMPLE An advantage of the optimal control method is 833 FISHERY BULLETIN: VOL. 73, NO. 4 its ability to accommodate multivariable system models such as multispecies models; models describing economic as well as biological phenomena; and detailed population models in- corporating size, age, temperature, food supply, etc. Once a three-variable example is presented, generalization of the technique to models with more than three variables is straightforward. The following model of a single species population was developed by Timin and Collier (1971) and contains three state variables: A'', the population density; W, the mean biomass per organism; and E, the food density. The model is given in dimensionless form, and thus the values of the model variables are relative to reference values. The system's dynamics are described by the following equa- tions: ^ = {h-d)N - f dt dE dt — = a -qN-dE (9) (10) h = 3.8T^2_3 8^ + 0.95 e = 0.1 d = 0.19/(2IF-1) !l = 0.2 a= 3 (■ = 0.05 \r^E fi = w^-'^ ' 1 + O.IE Static optimization can be used to determine the maximum equilibrium yield condition. For./gq = 0.005, the equilibrium values are: A'^^eq = 0.16, fi'eq = 20.3, W^ = Wh^ = 0.8. Following the procedures outlined in the Appendix (Equations (A-2) through (A-4)), Equations (9), (10), and (11) were linearized around this equilibrium to obtain: 'd,' dt dxo dt f/.rg jjt _ = 0.03 0 -0.56 = -5.73 -0.12 -0.81 0.14 0.02 -2.02 •*"2 + -1 0 0 u (12) = gq - {W + c)b - fiW - dt N (11) where: x^ = N - N_ eq where: t = time measured in a dimensionless unit equal to the time required for the or- ganism to metabolize an amount of food equal to its own dry weight (usually between two and four weeks for commercial fish species) b,d= birth and death rates per individual /= fishing rate g = the ratio of the quantity (energy in- gested minus energy not assimilated, minus energy expended to catch, in- gest and assimilate the ingested food) to the amount of energy ingested q= food ingestion rate per individual c= coeflficient of energy loss associated with births /A = metabolic heat loss coefficient a = rate of food supply $= proportionality constant for the rate of food leaving the system through decay or flushing Wf, = mean organism biomass of harvested individuals. Functional forms and parameters given as typical by Timin and Collier are: 834 ■^•2 = E- -^eq ■^3 = W -w. U = ./'- J eq- eq The following performance index J describes our desire to keep the system near the desired equilibrium (Appendix, Equation (A-6)): CO ■J= J (Qll.*'? + Q22■vl+Q33•»■i + ^"')^^ (13) 0 Here the weighting matrix [Q] becomes: [Q] = Qn 0 0 0 Q'22 0 0 0 Q33 and the matrix [R] becomes a scalar R. A sub- stantial difference between the single-variable and multivariable cases is that in the latter case we can no longer easily determine the feedback gains by specifying the desired values of the time cons- tants. Instead, the gains are calculated by I PALM: FISHERY REGULATION VIA CONTROL THEORY specifying the components of the weighting ma- trices. A common procedure for doing this is to choose the components by the rule: Qn=- 1 Im R = 1 where .ri^ is the maximum desired magnitude of the deviation .ri of the population A'', and )(„, is the maximum desired magnitude of the deviation u of the fishing rate./'. The components Q22 and Q33 are chosen in a similar manner. Here we assume the maxima are specified to be: 5% deviation from A^^^ = 0.008 1% deviation from W^^^ = 0.008 50% deviation from/^q = 0.0025. Thus: Iw 3m = 1.6 X 104 = 0.1 Assuming that the variation .(•2 in the food density is not of direct interest, we set Q22 = 0- Since ./ from Equation (13) depends only on the relative magnitudes of the weighting factors, we can choose these factors to be: R = 1 Qii = Q33 = 0.1 Q22 = 0. For this three-variable model, the symmetric Ric- cati matrix [P] has nine elements, three of which are redundant. Computer solution of the six coupled differential equations resulting from Equation (A-9) and use of Equations (A-7) and (A-8) yield the following feedback control func- tion: u = -[K]x = 0.149.ri + 0.00027.r2 + 0.026.r3. (14) Use of Equation (14) and the definitions of x^, x^, x^, and u. gives the optimal fishing effort as a feed- back function of the system variables: /=/eq + 0.149 (iV-iV,q) + 0.00027 {E - E,^)+ 0.026 fPT - W^^). Substitution of the equilibrium values gives: /= -0.045 + 0.149iV + 0.00027£' + 0.026^^.(15) Substitution of u from Equation (14) into Equation (12) gives the set of linearized equations describing the behavior of the model under feed- back control. The matrix [A] to be used in Equa- tion (A-5) becomes [A] = -0.119 -0.00027 0.534 -5.73 -0.12 -0.81 0.14 0.02 -2.02 The roots of Equation (A-5) are: .'^■ = -2.09, -0.096 ± 0.17j, where j = \fA. The dominant time con- stant is the negative reciprocal of the least negative real part, and here is equal to 1/0.096 = 10.4 time units. In a similar way the dominant time constant for the system without feedback is 45.5 time units. Thus the feedback control given by Equation (15) reduces the effects of disturbances in one-fourth the time. These linearized results have been verified by simulation of the original nonlinear model. Other simulations are discussed bv Palm (1975). Before concluding this example, we note from Equation (15) that ./' is a function of all three variables. This is due to the coupling between the three equations. Also, although the choice of the weighting factors is somewhat arbitrary, this should not obscure the fact that the Linear- Quadratic Optimal Control Problem provides a systematic method for determining the feedback gain matrix [A^. A systematic approach is needed because the number of components of [K] becomes so large for multivariable problems that a trial- and-error approach is prohibitive. As long as [Q] and [R] are chosen to be positive-definite, the resulting [K\ will stabilize the system. Various choices of [Q] and [/?] merely affect the time con- stants and form of response (oscillatory vs. non- oscillatory return to equilibrium). This is the main advantage of this technique. With this model the effects of mesh size regula- tion can be studied by using Wh as an additional control variable. Also, the food supply rate a is another possible control variable if the model is used to analyze fish farming. The linear-quadratic control technique could be used in both cases. CONCLUSION In this introductory paper we have presented only the deterministic case of the Linear-Quadrat- 835 ic Optimal Control Problem. In order to set limits on fishing effort which are functions of system variables such as population density or mean or- ganism weight, it is necessary to measure these variables. Any measurement process is stochastic or noisy, and it is necessary to compensate for this in the design of a feedback regulation scheme. In many engineering applications this has been suc- cessfully accomplished by the use of the Kalman- Bucy filter (Athans 1971). In addition, it may be impossible even to measure some variables. This problem of incomplete information has been frequently solved by the use of the Observer Theory (Kwakernaak and Sivan 1972). Also there will be uncertainties in the deter- mination of the model constants. In fact the "con- stants" may not be constants at all, Init merely the representation of several effects lumped together. Thus there is also error in the model structure, since the model constants are actually variables dependent upon a variety of effects. For the Schaefer model these effects would be interspecies interactions, age structure, availability and vulnerability of the age groups, and physical en- vironment influences on the biological processes. Such difficulties are amenable to solution by add- ing a "noise" term to the model equations and by modifying the linear-quadratic techniques to ac- commodate these stochastic effects (Athans 1971). It should also be pointed out that compensation for modeling errors is one of the purposes of feedback control. The change in model parameters with time can be compensated for by regularly recomputing the feedback gains as more data becomes available. Finally, while no pretense is made of being able to predict exact time paths, the methods described in this paper should prove useful in pro\iding management guidelines. The effects of stochastic pr(K-esses and uncertainties can be handled in a manageable way by computer simulation, and l)rediction of the future course of the managed fishery, in an average sense, can be made with appropriate error bands placed on the predictions. ACKNOWLEDGMENT I thank Saul Saila for bringing (loh's results lo my attention, and the referees for their comments on the bang-bang control approach and the effects of uncertainties in model parameters. FISHERY BULLETIN: VOL. 73, NO. 4 LITERATURE CITED .\THANS, M. 1971. Thf rcile and u.se of the .stochastic linear-(|uadratic- gaussian iirohlem in control sy.stem de.sign. Inst. Electr. Electron. Eng., Trans. Autom. Control. AC-1(5:529-.5.51. Beverton. R. .]. H.. .AND S. J. Holt. 1957. On the dynamics of e.xploited fish populations. Fish Invest. Minist. Agric. Fish. Food (G. B.), Ser. II, 19, 533 p. Carrothers, W. A. 1941. The British Columbia fisheries. Univ. Toronto Press, Toronto, 136 p. Clark, C, G. Edwards, and M. Friedlander. 1973. Beverton-Holt model of a commercial fishery: Optimal dynamics. J. Fish. Res. Board Can. 30:1629-1640. Clark, C.W. 1973. The economics of overexploitation. Science (Wash., D.C.) 181:630-634. Fox, W. W., -Jr. 1970. An e.xponential surplus-yield model for optimizing exploited fish populations. Trans. Am. Fish. Soc. 99:80-88. GoH.B.S. 1969. Optima! control of the fish resource. Malayan Sci. 5:65-70. 1973. Optimal control of renewable resources and popula- tions. (Summary of paper presented at the 6th Hawaii International Conference on System Sciences) 5 p. Ho, Y. C, and a. E. Bryson. 1969. Applied optimal control: Optimization, estimation and control. Blaisdell, Waltham, Mass., 481 p. Kwakernaak, H., and R. Sivan. 1972. Linear optimal control systems. Wiley - Interscience, N.Y.,.575p. Larkin, p. a. 1963. Interspecific competition and exploitation. J. Fish. Res. Board Can. 20:647-678. 1966. Ex])loitation in a t\-pe of predator-prey rela- tionship. .J. Fish. Res. Board Can. 23:349-3.56. Palm, W.J. 1975. An application of control theory to population dynamics. In E. 0. Roxin. P. T. Liu, and R. L. Sternberg (editors), Differential games and control theory, p. 59-70. Marcel-Dekker, N.Y. PELLA, ,J. J., AND P. K. TOMLINSON. 1969. A generalized stocl; production model. Inter-Am. Trop. Tuna Comm,, Bull. 13:419-496. Ricker, W. E. 1958. Handbook of computations for biological statistics of fi.sh populations. Fish. Res. Board Can., Bull. 119, 300 p. Saila, S. B. In press. Some applications of ojitiinal control theory to fisheries management. Trans. Am. Fish. Soc. Schaefer, M. B. 19.54. Some aspects of the dynamics of ))()pulations impor- tant to the management of commercial marine fisheries. Inter-Am. Trop. Tuna Comm., Bull. l:2.5-,56. 1968. Methods of estimating effects of fishing on fish populations. Trans. Am. Fish. Soc. 97:231-241. TiMiN, M. E., and B. 1). Collier. 1971. A model incorporating energy utilization for the dynamics of single species poinilations. Theor. Popul. Biol. 2:237-251. 836 PALM: FISHERY REGULATION VIA CONTROL THEORY APPENDIX The Linear-Quadratic Optimal Control Problem and its solution are now outlined. For a thorough discussion, see Ho and Bryson (1969) or'Kwaker- naak and Sivan (1972). By use of state variable notation, any set of time-invariant ordinary differential equations can be put into the follow- ing form: df = g i!i,.t) (A-l) where y is a general /^-dimensional vector func- tion, y is the ^-dimensional state vector for the model, and/'is the ///-dimensional input or forcing function vector. If this system has an equilibrium (^eq'/eq- the followlng Set of algebrais equations must be satisfied: Q =^(//eq./eq)- The values of //p,, and f\,^ depend on the system's parameters, and static optimazation methods such as calculus or linear programming can be applied to find the optimal j/gq, /eqand system parameters according to some criterion. This was done in the first example to determine the condition of maximum equilibrium yield. Since any real system is subjected to varying conditions and distur- bances, it will be continually displaced from equilibrium. Thus for unstable systems or stable systems with large time constants, a static method of analysis is not suflicient. In such a case the next step is to apply a dynamic optimization method, such as the method presented here, to devise a control scheme which ensures that the system will return to equilibrium with a satisfactory time constant. Thus static and dynamic methods should not be viewed as alternative approaches to op- timization, but rather as mutually complementary methods. After the equilibrium is determined, Equation (A-l) is linearized by expanding the function g in a Taylor series in y and/, and keeping only the first-order terms. This gives the linearized model: where: ^=[A]x + [B]ii (A-2) [A] = [B] = (A-3) (A-4) in which the subscript eq indicates that the partial derivatives of _^ are evaluated at the equilibrium. The stability of the equilibrium can be determined from the roots of the determinant equation: s [/] - [A] = 0 (A-5) where [/] is the {)! x //) identity matrix. The equilibrium is stable if and only if all of the roots s have negative real parts. By finding the function n which minimizes the following quadratic performance index, £ and n are kept near zero and thus the system is kept near equilibrium. J ^ i (£'r[Q]x+ i±T[R]u)dt. (A-6) 0 The feedback control function which minimizes J has been shown to be: . = -[Al£. (A-7) The feedback "gain" matrix [K] is calculated from: [K] = [R]-'[Br[P] (A-8) where the Riccati matrix [P], an (n x n) symmet- ric matrix, is the steady-state solution of the Ric- cati matrix differential equation: dJF] dt = [Q]+[AnP] + [P][A] - [P][B][R]-nBV [P] (A-9) with the initial condition: [pm = [0]. The matrix [P] is usually found by numerically solving the Riccati equation until all the com- ponents of the solution [P] become constant. This will always occur. 837 DIFFERENTIATION OF FRESHWATER CHARACTERISTICS OF FATTY ACIDS IN MARINE SPECIMENS OF THE ATLANTIC STURGEON, ACIPENSER OXYRHYNCHUS R. G. AcKMAN, C. A. Eaton, and B. A. Linke' ABSTRACT Lipids and fatty acids of two marine-caught specimens of the Atlantic sturgeon, Acipenser oxyrhynchus, which spawns and also feeds in freshwater, were examined. Specific fat contents, re- spectively 47.2 and 25.0% in orange-colored dorsal tissue and 8.5% in both livers, were high but not unexpected for sturgeons generally. In each fish a very consistent basic fatty acid composition of depot fats showed that this fat in various parts of the body had a common function. Depot fat from the fatter fish had high iodine values (ca. 190) while in the leaner fish values were lower (ca. 135) and more typical of sturgeons generally. The fatty acid details of depot fats showed some characteristics of marine fats, such as the presence of the unusual wl and tr)4 polyunsaturated fatty acids, the low figure for linoleic acid and relatively high proportions of long chain polyunsaturated fatty acids, but were more typical of freshwater fats in the virtual absence of eicosenoic and docosenoic acids. Broadly speaking, the fatty acids of the Atlantic sturgeon seem to place it in a special class of fish with fats generally resembling freshwater fish fats in composition, despite its marine origin. The Atlantic sturgeon, Acipenser oxyrhynchus Mitchill, is widely distributed along the Atlantic coast of North America and is to be distinguished from A. sturio, the common sea sturgeon of Europe (Scott and Grossman 1973). The Atlantic sturgeon is an anadromous fish, spawning in freshwater,- whereas some other sturgeon species, such as the lake sturgeon, A. fulvescens, are re- stricted to freshwater. The most recent and de- tailed study of sturgeon lipids and fatty acids has been based on an A. sturio specimen, apparently of freshwater origin, as it showed a fatty acid pattern which is characteristic of lipids in fresh- water fish (Reichwald and Meizies 1973). The standard reference book on fatty acids states that sturgeon fats are "of the freshwater type" (Hilditch and Williams 1964) although this view was based on a single analysis of a specimen of A. sturio caught in the North Sea (Lovern 1932). We wish to report that a study of two saltwater A. oxyrhynchus shows that, during its marine period, the Atlantic sturgeon deposits fat with some composition details corresponding to marine fatty acid characteristics. However the fat definitely 'Environment Canada, Fisheries and Marine Service, Halifax Laboratory, P.O. Box 429, Halifax, Nova Scotia B3J 2R3 Can. ^Sturgeon were once so common that those blocked by the falls in the Hudson river were taken in large quantities and marketed in New York City as "Albany veal." They appear to be returning to their former habitat in large numbers (R. Severo, New York Times, 9 July 1975). lacks other details characteristic of fats of higher marine organisms and thus reinforces the published viewpoint based on the common sea sturgeon of the Northeast Atlantic. MATERIALS AND METHODS Samples Two A. oxyrhynchus were acquired from fish traps located in an area of the entrance to Halifax Harbour known as Eastern Passage. Both were male, that taken on 12 October 1968 (A) being 150 cm in length and that taken 30 August 1973 (B) being 155 cm. Fish A was frozen whole at -40°C until dissected in March 1969. Fish B was held overnight in an aquarium and dissected immedi- ately after sacrifice. In both cases sections were cut transversely through the middle of the fish. In fish A this was done while frozen and the cut section included liver which was recovered for study, while in fish B the liver was removed separately from the fish. Both fish showed a soft fatty orange layer between the dorsal skin and muscle, of one or more centimeters in thickness, but thinning down the flanks. Parts of this layer penetrated the muscle, especially between myotomes, and streaks of similar colored material appeared in the muscle. A section through fish B was observed to have a Manuscript accepted January 1975. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. 838 ACKMAN ET AL.: FATTY ACIDS IN ATLANTIC STURGEON white layer 2-3 mm in thickness between the epidermis and the soft orange layer. Lipid from the orange tissue of fish A was ex- tracted by blending with n-hexane in a Waring Blender.' All other lipids were extracted by the method of Bligh and Dyer (1959). For fish A the samples examined, and lipid recoveries, were liver, S.59c, orange tissue, 47.2'^; muscle freed of all visi- ble orange tissue and fat, 2.0%; and whole steak section, 7.2%. Lipid recoveries from fish B samples were liver, 8.5%; orange tissue, 25.0%; muscle freed of all visible orange tissue and fat, L2%; and sub- dermal white layer, L3%. Total lipids from the fish A samples were saponified and non-saponifiable materials removed. The recovered fatty acids were converted to methyl esters with boron trifluoride in methanol. White layer and orange tissue samples from fish B were treated similarly, but the muscle and liver lipids were fractionated on a column of divinylbenzene copolymer beads and eluted with benzene. The purity of the various fractions was monitored by thin-layer chroma- tography on silicic acid. The major fractions, the triglycerides and the polar lipids with the mobility of phospholipids, were transesterified with BFg- MeOH. Analyses of recovered methyl esters of fatty acids are given in Tables 1 and 2. Further details of these methods, including gas-liquid chromatography of methyl esters on open-tubular polyester columns, will be found elsewhere (Sipos and Ackman 1968; Ackman, Hooper, and Frair 1971; Ackman and Hooper 1973). RESULTS AND DISCUSSION The triglycerides in the two marine A. oxyrhynchus were distributed throughout the body in the dorsal fatty layer, in the form of muscle infiltration by this fatty layer, and also in the liver. This is most clearly made evident by comparing the iodine values of the different fats isolated from sturgeon B (Table 2). The white subdermal layer, the orange tissue fat, estimated from thin-layer chromatograms to be >95% triglycerides, and the triglycerides isolated from muscle and liver lipids, all have calculated iodine values in the range 132-139. The two phospholipid fractions have much higher iodine values, as ex- pected for this class of lipids (Ackman 1966). In fish A, the orange tissue fat (a clear oil, estimated to be >95% triglycerides from thin-layer chroma- tograms) had a calculated iodine value of 186 (Ta- ble 1). The liver lipid, also estimated to be mostly Table 1. -Fatty acid composition, in weiglit percent,' for lipids recovered from four tissues of Atlantic sturgeon A. Muscle Whole Orange (no visible steak Fatty Acid2 tissue fat) section Liver 14:0 5.13 2.80 3.61 2.82 ISO 15:0 0.35 0.21 0.27 0.21 Anteiso 15:0 0.16 0.11 0.14 0.10 15:0 0.78 0.53 0.90 0.78 ISO 16:0 0.23 0.12 0.19 0.14 16:0 16.21 17.05 16.34 17.38 ISO 17:0 0.29 0.21 0.31 0.26 Anteiso 17:0 0.15 0.10 0.12 0.09 17:0 0.48 0.39 0.40 0.33 3,7,1 1,15-TMHD 0.37 0.44 0.46 0.09 Iso 18:0 0.11 0.08 0.14 0.12 18:0 1.65 2.91 1.83 2.58 Total saturates 25.3 25.2 25.0 25.1 16:10)9 0.42 0.59 0.44 0.39 16:1(1)7 7.45 5.61 6.25 6.45 16:10)5 0.29 0.26 0.17 0.26 17:10)8 0.42 0.30 0.38 0.45 18:10)11+9 15.69 15.22 15.11 18.27 18:10)7 4.04 4.35 3.93 5.49 18:10)5 0.47 0.45 0.31 0.38 19:10)9 0.33 0.39 0.32 0.56 20:10)11 0.35 0.20 0.41 0.16 20:10)9 1.23 0.96 1.34 1.19 20:10)7 0.86 0.84 1.25 0.81 22:10)13 + 11 0.15 0.05 0.13 0.08 Total monoenes 31.9 29.4 30.2 34.7 16:20)6 0.26 0.12 0.26 0.04 16:20)4 1.42 1.03 1.17 0.78 18:20)6 0.85 0.77 0.85 0.84 18:20)4 0.28 0.31 0.29 0.57 20:20)6 0.24 0.15 0.35 0.26 16:30)4 1.61 0.93 1.47 0.43 18:3o)6 0.08 0.13 0.10 0.13 18:30)4 0.40 0.20 0.32 0.30 18:30)3 0.28 0.26 0.30 0.33 20:3o)6 0.18 0.16 0.28 0.21 20:3o)3 0.10 0.10 0.12 0.12 16:40)3 0.17 0.19 0.15 0.03 16:4o)1 1.40 0.90 1.25 0.15 18:40)3 2.78 1.62 2.25 1.28 18:40)1 0.83 0.54 0.64 0.67 20:40)6 1.37 1.63 1.53 1.72 20:40)3 1.64 1.41 1.64 1.97 22:40)6 0.13 0.19 0.20 0.26 20:50)3 18.37 19.15 19.40 14.33 21:50)3 or 2 0.99 0.84 0.92 0.94 22:50)6 0.38 0.74 0.54 0.42 22:50)3 2.21 3.10 2.82 3.40 22:60)3 6.85 10.88 7.91 10.94 Total polyenes 42.8 45.4 44.8 40.2 Calc. iodine value 186 199 194 182 ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 'NSA = no significant amount. Average percentages for some minor components were: Iso 14:0, 0.07%; 4,8,12-TMTD, 0.03%; 2,6,10,14-TMPD, 0.03%; 19:0, 0.09%; 20:0, 0.04%; 14:1tD7, 0.02%; 15:10)8, 0.3%; 19:1o)10, 0.07%; 19:1o)8, 0.06%; 20:1o)5, 0.01%; 22:10)9, 0.04%; 22:10)7, NSA; 22:10)9, NSA; 16:2o)7, 0.02%; 18:20)9, NSA; 20:2o)9, NSA; 16:30)3, NSA. 23,7,1 1,15-TMHD (phytanate) = 3,7,1 1,1 5-teframethylhexadeca- noic acid. 4,8,12-TMTD = 4,8,12-trimethyltrldecanoic acid. 2,6,10, 14-TMPD (pristanate) = 2,6,10,14-tetramethylpentadecanoic acid. 839 FISHERY BULLETIN; VOL. 73, NO. 4 Table 2.-Fatty acid composition, in weight percent,' for lipids recovered from four tissues of Atlantic sturgeon B. White Muscle (no visible fat) Liver ciihrlarmQl Orange tissue Fatty acid2 layer Triglyceride Polar lipid Triglyceride Polar lipid 14:0 3.75 3.81 3.50 0.75 2.42 1.53 ISO 15:0 0.37 0.35 0.31 0.21 0.20 0.14 Anteiso 15:0 0.20 0.26 0.22 0.11 0.09 0.04 15:0 0.71 0.66 0.52 0.43 0.62 0.72 ISO 16:0 0.18 0.20 0.14 ND' 0.16 0.11 16:0 15.46 15.24 14.75 22.42 18.32 25.72 7-MHD 0.33 0.25 0.22 0.11 0.19 0.20 2,6,10,14-TMPD 0.22 0.28 0.18 ND 0.16 NSA ISO 17:0 0.35 0.29 0.28 0.10 0.33 0.24 Anteiso 17:0 0.19 0.25 0.18 0.07 0.20 0.16 17:0 0.27 0.31 0.27 0.27 0.32 0.74 3,7,11. 15-TMHD 1.08 1.38 0.55 0.14 0.05 0.27 Iso 18:0 ? 0.14 0.20 0.20 0.08 0.23 0.35 18:0 2.17 1.98 3.14 8.71 2.67 11.77 19:0 0.15 0.13 0.12 0.33 0.07 0.43 20:0 0.18 0.19 0.14 0.09 0.08 0.31 Total saturates 25.9 26.1 25.1 33.9 26.3 43.0 16:1o»9 0.20 0.25 0.21 0.03 0.20 0.61 16:1a>7 6.22 6.08 5.47 1.01 4.50 1.13 17:1«8 0.75 0.70 0.63 0.31 0.52 0.09 18:1M9 25.94 25.67 25.75 11.81 31.16 9.03 18:1»7 4.00 3.91 5.94 3.94 4.46 4.44 18:1o»5 0.43 0.49 0.71 0.21 0.56 0.21 19:10)10 + 8 0.54 0.32 0.61 0.10 0.43 0.14 19:10)6 0.20 0.22 0.48 0.05 0.21 0.07 20:10)11 1.27 1.29 1.05 0.13 0.31 0.80 20:10)9 2.76 2.70 2.40 0.55 2.44 2.00 20:10)7 2.41 2.40 1.72 0.27 1.41 0.89 20:15 0.22 0.23 0.17 0.02 0.13 0.10 22:10).13+11 0.51 0.46 0.18 0.10 0.14 0.62 22:10)9 0.26 0.36 0.16 0.01 0.19 0.13 22:10)7 0.15 0.16 0.05 ND 0.10 0.09 Total monoenes 46.0 45.5 45.7 18.7 46.9 20.8 16:2016 0.15 0.21 0.16 ND 0.04 0.16 16:2a>4 0.95 0.85 0.84 0.12 0.52 0.24 18:2«6 0.84 0.84 0.77 0.22 0.68 0.52 18:2»4 ? 0.21 0.20 0.26 0.10 0.51 ND 20:2tt9 0.12 0.07 0.15 Trace 0.19 0.03 20:2(1)6 0.52 0.56 0.48 0.19 0.35 0.27 20:20)4 0.25 0.18 0.18 Trace 0.12 ND NMID [20:2] 0.44 0.48 0.24 0.03 0.10 0.03 NMID [22:2] 0.50 0.52 0.18 0.01 0.20 0.05 16:3(1)4 0.65 0.64 0.70 Trace 0.32 ND 18:3(06 0.26 0.29 0.19 Trace 0.07 Trace 18:30)4 0.74 0.68 0.52 ND 0.48 ND 18:30)3 0.48 0.41 0.38 0.11 0.24 0.18 20:30)6 0.26 0.19 0.12 0.05 0.14 0.06 20:30)4 0.16 0.30 0.15 ND 0.35 ND 20:31*3 0.18 0.27 0.13 0.05 0.31 0.08 16:4(d1 0.61 0.60 0.52 0.02 0.08 0.01 18:4(143 1.29 1.18 1.12 0.07 0.51 0.09 18:40)1 ? 0.38 0.37 0.47 0.03 0.33 0.01 20:40)6 1.40 1.28 1.37 4.78 0.84 6.90 20:40)3 1.42 1.41 1.18 0.26 1.24 0.37 22:40)6 0.17 0.51 0.20 0.20 0.29 0.32 20:50)3 0.62 10.21 10.19 18.61 6.95 7.35 21:50)2 or 3 0.64 0.66 0.67 0.06 0.62 0.05 22:5(»6 0.56 0.44 0.44 0.26 0.56 0.64 22:50)3 2.44 2.52 2.57 2.02 4.87 2.86 22:6«3 3.85 3.32 4.40 18.77 5.10 16.46 Total polyenes 28.1 28.4 29.2 47.4 26.8 36.2 Calc. iodine value 139 136 136 201 132 161 'NSA = no signficant amount. Average percentages for some minor components from subdermal layer, orange tissue, and triglycerides were: 12:0, 0.04%; 13:0, 0.01%; iso 14:0, 0.04%; 4,8,12-TMTD, 0.05%: anteiso 16:0, 0.02%; anteiso 18:0, 0.02%; iso 19:0, 0.05%; anteiso 19:0, 0.02%; 22:0, 0.01%; 14:1o»7, 0.02%; 14:1«»5, 0.01%; 16:1(i>11, NSA; 16:1o)5, 0.08%; 17:1o)6, 0.03%; 18:1o)13, NSA; 18:10)11, trace; 20:1o)15(?), 0.06%; 22:1(»5, NSA; 18:20)9, 0.05%; 22:2o>6, 0.11%; 16:3o)3, trace in TG; 20:3o)9, 0,03%; 16:40)3, 0.01%; 22:40)3, 0.06%. 27-MHD = 7-methylhexadecanoic acid from 7-methylhexadecenoic acid (Hooper et al. 1973), measured in hydrogenated esters. 2,6,10,14-TMPD (pristanate) = 2,6,10,14-tetramethylpentadecanoic acid. 3,7, 11, 15-TMHD (phytanate) = 3,7,11,15-tetramethylhexadecanoic acid, NMID = non-metfiylene-interrupted dienoic acids. 4,8,12- TMTD = 4,8,12-trimetfiyltridecanoic acid. 3ND = not determined. 840 ACKMAN ET AL.: FATTY ACIDS IN ATLANTIC STURGEON triglyceride by thin-layer chromatography, had a similar iodine value, as did the lipid from the steak section of high fat content. This high fat content appears to be normal as Fraser et al. (1961) reported 6.2% fat in a steak from this species. Evidently, the fatty acid compositions of triglycerides for fish A would give high iodine values similar to those for phospholipids, for example in the lean muscle extract which would include about half of each type of lipid. The actual iodine value of the triglyceride of sturgeon A is unusually high for marine fish triglycerides (Ack- man 1966) but a considerable range of iodine values appears possible for sturgeon depot fats. The "peritoneal cavity" depot fat of the A. ^turio examined by Lovern (1932) had an iodine value of 126.5, and that of the corresponding liver lipids was 125. A Pacific coast sturgeon (species un- known) had body and liver oils with respective iodine values of 90 and 95 (Bailey et al. 1952), and the iodine values of fats of three types of flesh from the freshwater A. sturio of Reichwald and Meizies (1973) were also low. Two yl. baeri kept in captivity for several years in the Freshwater Fisheries Research Laboratory in Tokyo were slightly different from each other in fatty acid compositions (Table 3) but the fats in each body sample, dorsal flesh and ventral flesh, were re- spectively quite similar in each fish although they differed in some details from the liver fatty acids (Shimma and Shimma 1968). These authors specifically note the absence of mesentary fat (cf. Lovern 1932), although they found the testes to be unexpectedly high in fat. Oil from American stur- geon of unspecified origin had an iodine value of 125.3 (Bull 1899) and the liver oil from A. mikadoi an iodine value of 157.7 (Tsujimoto 1926). Russian data shows Caspian and Atlantic sturgeon fats as having respective iodine values of 122 and 125 (Zaitsevetal. 1969). In comparative detail, the fatty acid analyses from sturgeon A show little differences between the fat from steak section, orange tissue, lean muscle, and liver (Table 1). However, the high proportions of 22:6 3 for either conversion or catabolism. The two Atlantic sturgeon do not appear to have as much fat in the liver as observed in other species or perhaps in animals from other habitats. Zaitsev et al. (1969) show 8-16^ oil in livers of other than Danube sturgeon, and 8-20% in the latter. The liver of an A. mikadoi taken at sea near Hokkaido yielded 52% oil by boiling (Tsujimoto 1926) and Shimma and Shimma (1968) report 38 and 51% lipid in livers of two A. baeri. Among the unusual fatty acids observed in the fat of fish B, special mention should be made of the NMID (nonmethylene-interrupted dienes) [20:2] and [22:2] (Ackman and Hooper 1973; Paradis and Ackman 1975). In vertebrate lipids these unusual fatty acids are apparently deposited in parallel with 20:1 and 22:1. The generally lower levels of the latter in the lipids of fish A resulted in the NMID [20:2] and [22:2] being barely detectable (<0.01%) and they are not included in Table 1. The food of sturgeons on the Nova Scotian shelf is probably basically bottom invertebrates (Scott and Grossman 1973). Many of these organisms are potential sources of these unusual acids (Ackman and Hooper 1973; Watanabe and Ackman 1974; Ackman et al. in press). These acids do not appear to occur significantly in the oils from pelagic fish and it can be assumed that their occurrence in the sturgeon is a food web effect rather than a peculiarity of the species. The absence of 16:lttll and 18:10.01% and identifications are speculative. The importance of marine algae (the primary source of phytol from which phytanic acid is derived) in the diet of sturgeons is not known (Scott and Grossman 1973). In freshwater, Atlan- tic sturgeon do eat algae (Leim and Scott 1966). The unusually high levels of 16:2(d4, 16:3<»4, and 16:4u)l, all of which are primarily algal in origin (Ackman et al. 1968), indicate that plants provide a significant proportion of dietary lipids. The ten- tatively identified higher homologues 18:222:6a)3 agrees with our data, as also reported in less comprehensive studies of European sturgeon (Mangold 1973; Meizies and Reichwald 1973). The depot fats of the marine sturgeon we have investigated had 22:6«3 at about half the level of 20:5 0)3, and we interpret the analysis of the marine A. sturio by Lovern (1932) to agree with our data. It appears that at some point in the animals' spawning migration the proportions of these two fatty acids could reverse. Interestingly enough, only one out of four fresh- water fish oils (from maria or Lota lota) examined earlier contained larger proportions of 22:6«*3 than of 20:5a)3 (Ackman 1967). The spawning period for Canadian Atlantic sturgeon is presumably in early summer. The two male fish examined showed no gonad development and, therefore, if mature presumably they had spawned and returned to the ocean. Oleic acid (218:1) at 44-49% and palmitic acid (16:0) at 21-23% are indicated by Reichwald and Meizies (1973) and other studies (see above) to be the major com- ponents of freshwater sturgeon fats. The depot fats of the two marine Atlantic sturgeon we have investigated differ in that both 18:1 totals are 20- 30% (magnitude inversely related to iodine value) and 16:0 is about 15%. The original marine A. sturio had about 36% 18:1 and 16-19% 16:0, or in other words, the fats examined by Lovern (1932) displayed a composition for these two fatty acids intermediate to two more recent studies. The very different iodine values for sturgeon A and B are accounted for mainly by the differences in percentages of 20:5«)3, 22:5w3, and 22:6»3 in lieu of 18:1 and other monoethylenic acids, as total sa- turated acids are in the same proportion of fat in both fish and the proportions of most minor unsa- turated acids are not important enough to matter. The monoethylenic fatty acids of fish B (total about 45%) probably are more normal as judged by the various low iodine values in the literature for sturgeon fat. The resolving power of open-tubular gas-liquid chromatograph for methyl esters of monoethylenic fatty acids extends our knowledge of the fatty acid biochemistry of the two Atlantic sturgeon in this study. Virtually no 18:lo)ll, which could only come from 20:l«i>ll, was observed. This indicates that the fish were depositing fat and not catabolizing it. On the other hand, the percentage of 20:120:la)7 appears to be important, suggesting a period of fatty acid biosynthesis rather than of deposition of ex- ogenous fatty acids. Both fish had about the same percentages of 16:1. Fish A had higher proportions of fat in the muscle than Fish B, and 18:lo)7 was about a quarter of 18:lo)9 in this fat. In the leaner fish B, 18:lo)7 was only about one-fifth of 18:la)9, suggesting less activity in de novo biosynthesis. It is possible that the diet of fish A was rich in the polyunsaturated fatty acids, the deposition of which was disturbing the typical species-fat com- position. Accordingly, fish A may have been more actively engaged in synthesizing monoethylenic fatty acids, via 16:0>16:la)7*18:lo)7, to achieve this composition. Seals, whose depot fat has a higher iodine value than that of most whales, may show the same monoethylenic fatty acid activity (Ack- man, Epstein, and Eaton 1971). Earlier work on oils from four freshwater fish showed l-37f 20:1 and about 0.3-0.4% 22:1 (Ackman 1967). Marine fish oils, in our experience, usually show 109c or more of 20:1 and 5% or more of 22:1. The absence of large proportions of 20:1 and 22:1 acids in the marine sturgeon depot fat, even in fish B with the less unsaturated fat, is the key reason for our placing the fat of the marine Atlantic sturgeon in a rather special class of marine fat, or more broadly, in the generally freshwater class of fish fats, as recorded by Hilditch and Williams (1964). ACKNOWLEDGMENTS J. Hingley assisted in the technical studies. P. J. Ke gave valuable advice on oriental species and W. J. Dyer on observations on the Atlantic sturgeon. Personal communications from Y. Shimma and I. Reichwald clarified anatomical problems. LITERATURE CITED Ackman, R.G. 1966. Empirical relationships between iodine value and polyunsaturated fatty acid content in marine oils and lipids. J. Am. Oil. Chem. Soc. 4.3:38.5-.388. 1967. Characteristics of the fatty acid composition and biochemistry of some fresh-water fish oils and lipids in comparison with marine oils and lipids. Comp. Biochem. Physiol. 22:907-922. Ackman, R. G., and C. A. Eaton. 1971. Investigation of the fatty acid composition of oils and lipids from the sand launce {Ammodytes americanus) from Nova Scotia waters. J. Fish. Res. Board Can. 28:601-606. \ckman, R. G., C. a. Eaton, J. C. Sipos, S. N. Hooper, and J. D. Castell. 1970. Lipids and fatty acids of two species of North Atlantic krill (Meganyctiphaneii norvegica and Thysanoessa iner- mis) and their role in the aquatic food web. J. Fish. Res. Board Can. 27:513-533. Ackman, R. G., S. Epstein, and C. A. Eaton. 1971. Differences in the fatty acid compositions of blubber fats from Northwestern Atlantic finwhales {Balaenoptera physalus) and harp seals (Pagophilus groenlandica). Comp. Biochem. Physiol. 40B:683-697. Ackman, R. G., S. Epstein, and M. Kelleher. In press. A comparison of lipids and fatty acids of the ocean quahaug {Artica islandica) from Nova Scotia and New Brunswick. J. Fish. Res. Board Can. Ackman, R. G., and S. N. Hooper. 1968. Examination of isoprenoid fatty acids as distin- guishing characteristics of specific marine oils with par- ticular reference to whale oils. Comp. Biochem. Physiol. 24:.549--565. 1973. Non-methylene-interrupted fatty acids in lipids of shallow-water marine invertebrates: A comparison of two molluscs (Littorina littorea and Lunatia triseriata) with the sand shrimp {Crangon septemspinosus). Comp. Biochem. Physiol. 46B:153-165. Ackman, R. G., S. N. Hooper, and W. Frair. 1971. Comparison of the fatty acid compositions of depot fats from fresh-water and marine turtles. Comp. Biochem. Physiol. 406:931-944. Ackman, R. G., J. D. Joseph, and A. Manzer. 1974. Tentative identification of an unusual naturally-oc- curring polyenoic fatty acid by calculations from precision open-tubular GLC and structural element re- tention data. Chromatographia 7:107-114. Ackman, R. G., C. S. Tocher, and J. McLachlan. 1968. Marine phytoplankter fatty acids. J. Fish. Res. Board Can. 25:160.3-1620. Albrecht, M. L., and B. Breitsprecher. 1969. Untersuchungen uber die chemische Zusammenset- zung von Fischnahrtieren und Fischfuttermitteln. Z. FischereiN.F. 17:143-163. Bailey, B. E., N. M. Carter, and L. A. Swain (editors). 1952. Marine oils with particular reference to those of Canada. Bull. Fish. Res. Board Can. 89, 413 p. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid extraction and purifica- tion. Can. J. Biochem. Physiol. 37:911-917. Bull, H. 1899. Ueber die Bestimmung stark ungesattiger Fett- sauren in den Thranen. Chemiker-Zeitung 23:1043-1044. Farkas, '1'. 1971. A possible explanation for the differences in the fatty acid composition of freshwater and marine fishes. Ann. Biol. Tihany 38:143-152. Farkas, T., and S. Herodek. 1967. Investigations of the fatty acid composition of fishes from Lake Balaton. Ann. Biol. Tihany 34:3-13. Eraser, D. I., A. Mannan, and W. J. Dyer. 1961. Proximate composition of Canadian Atlantic lish. III. Sectional differences in the flesh of a species of Chondronfei, one of Chimaerae, and of some mis- cellaneous teleosts. J. Fish. Res. Board Can. 18:893-905. 844 ACKMAN ET AL.: FATTY ACIDS IN ATLANTIC STURGEON HiLDITCH, T. P., AND P. N. WILLIAMS. 1964. The chemical constitution of natural fats. 4th ed. Wiley, N.Y., 745 p. Hooper, S. N., M. Paradis, and R. G. Ackman. 1973. Distribution of f raHs-6-hexadecenoic acid, 7- methyl-7-hexadecenoic acid and common fatty acids in lipids of the ocean sunfish Mola nwla. Lipids 8:509-516. Ikekawa, N., M. Matsui, T. Yoshida, and T. Watanabe. 1972. The composition of triglycerides and cholesteryl esters in some fish oils of salt, brackish and fresh water origins. Bull. Jap. See. Sci. Fish. 38:1267-1274. Jangaard, p. M. 1965. A rapid method for concentrating highly unsaturated fatty acid methyl esters in marine lipids as an aid to their identification by GLC. J. Am. Oil Chem. Soc. 42:845-847. Klenk, E. 1963. Uber die Bildung von C-^o-und €22" Polyensaureii ausA'^-^-^^-^^-Hexadecate-traensaure Dei der Ratte. Hoppe-Seyler's Z. Physiol. Chem. 331:50-55. Leim, a. H., and W. B. Scott. 1966. Fishes of the Atlantic Coast of Canada. Fish. Res. Board Can., Bull. 155, 485 p. Lovern, J. A. 1932. Fat metabolism in fishes. II. The peritoneal, pancrea- tic and liver fats of the sturgeon {Acipenser isturio). Biochem. J. 26:1985-1988. Mangold, H. K. 1973. "Unsichtbare" Fette und andere Lipide in Susswas- serfischen, Wissenschaftliche Veroffentlichungen der Deutschen Gesellschaft fiir Ernahrung, Band 24, -Un- sichtbare" Fette und Lipoide in Lebensmitteln. (Dr. Dietrich Steinkopff Verlag, Darmstadt 1973.) p. 32-38. Maxwell, J. R., R. G. Cox, G. Eglinton, C. T. Pillinger, R. G. Ackman, and S. N. Hooper. 1973. Stereochemical studies of acyclic isoprenoid com- pounds. II. The role of chlorophyll in the derivation of isoprenoid-type acids in a lacustrine sediment. Acta Geochim. Cosmochim. 37:297-313. Meizies, a., and I. Reichwald. 1973. Die Lipide im Fleisch und im Rogen frischer und geraucherter Fische. Z. Ernahrungswissenschaft 12:248-251. Morris, L.J. 1966. Separations of lipids by silver ion chromatography. J. Lipid Res. 7:717-732. Paradis, M., and R. G. Ackman. 1975. Occurrence and chemical structure of nonmethylene- interrupted dienoic fatty acids in the American oyster Crassostrea virginica. Lipids 10:1-8. Reichwald, I., and A. Meizies. 1973. Die Fettsauren der Lipide im Fleisch von Stisswas- serfischen und Seefischen. Z. Ernahrungswissenschaft 12:86-91. Scott, W. B., and E. J. Grossman. 1973. Freshwater fishes of Canada. Fish. Res. Board Can., Bull. 184, 966 p. Shimma, Y., and H. Shimma. 1968. Fatty acid composition of Acipenser baeri cultivated in Tokyo. Bull. Freshwater Fish. Res. Lab. (Tokyo) 18:179-184. Sipos, J. C, and R. G. Ackman. 1968. Jellyfish (Cijanea capillata) lipids: Fatty acid com- position. J. Fish. Res. Board Can. 25:1561-1569. Tsujimoto, M. 1926. Aquatic animal oils. J. Chem. Soc. Jap. Ind. Chem. Sect. 29:71-75. Watanabe, T., and R. G. Ackman. 1974. Lipids and fatty acids of the American (Crassostrea virginica) and European flat {Ostrea. edulis) oysters from a common habitat, and after one feeding with Dicrateria inornata or Isochrysis galbana. J. Fish. Kes. Board Can. 31:403-409. Wessels, J. P. H., AND A. A. Spark. 1973. The fatty acid composition of the lipids from two species of hake. J. Sci. Food Agric. 24:1359-1370. Zaitsev, v., L. Kizevetter, L. Lagunov, T. Makarova, L. Minder, and V. Podsevalov. 1969. //( Fish curing and processing. Mir Publishers, Mos- cow (translated by A. de Merindol), p. 68, 562. 845 ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY OF PINK SALMON, ONCORHYNCHUS GORBUSCHA, AND CHUM SALMON, ONCORHYNCHUS KETA, IN TRAITORS COVE, ALASKA, WITH SPECULATIONS ON THE CARRYING CAPACITY OF THE AREA Jack E. Bailey, Bruce L. Wing, and Chester R. Mattson' ABSTRACT Juvenile pink salmon, Oncorhynchus gorbuscha, and chum salmon, 0. keta, 28 to 56 mm long (fork length) from Traitors River in southeastern Alaska, fed little in freshwater but fed heavily in the estuar>', mainly on pelagic zooplankters. Fry did not feed on cloudy moonless nights. The rate of evacuation of pink salmon stomachs ranged from 6 h at 12.8°C to 16 h at 8.5°C. The abundance of zooplankton ranged from 9 to 154 organisms per liter and quantitatively did not change noticeably while fry were in the estuary. In 1964, 1965, and 1966, the estimated numbers of fry in Traitors Cove was 7, 1, and 4 million, respectively. An attempt was made to estimate the carrying capacity of Traitors Cove, using food consumption and evacuation rates in conjunction with estimates of standing crop of zooplankton. It was concluded that 50 to 100 million additional fry from hatcheries would probably exceed the carrying capacity of the estuary. With the rapidly growing demand for animal pro- tein and the emergence of new hatchery tech- niques for pink salmon, Oncorhynchus gorbuscha, and chum salmon, 0. keta (Bams 1972; Bailey and Heard 1973; Bailey and Taylor 1974), we believe that it is timely to speculate on the capacity of estuaries to support more fry. The Japanese, Rus- sians, and Canadians have a number of major pink and chum salmon hatcheries and spawning chan- nels in operation. Japanese hatcheries released over 800 million pink and chum salmon fry in 1973 (source: Japan Fishery Agency). Individual Rus- sian hatcheries are capable of releasing up to 120 million fry annually (Kanid'yev et al. 1970). The Qualicum River in British Columbia, Canada, now produces about 50 million chum salmon fry an- nually through a combination of flow control in the natural spawning areas and the operation of a spawning channel (Eraser 1972). The problem of evaluating the carrying capacity of estuaries for artificially produced fry is most pertinent. What, for example, would be the impact of 100 million fry on the available food in Traitors Cove? Recent technological advances in rearing salm- on in hatcheries and spawning channels now make it possible to release tens of millions of pink and chum salmon fry into individual estuaries, but 'Northwest Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 155, Auke Bav, AK 99821. lack of knowledge of the food requirements of these two species in nature makes even the immediate results of such releases uncertain. It is conceivable that a spawning channel or hatchery operation could produce such large numbers of fry that their migratory behavior might be altered, or growth and survival might be reduced because of severe competition for a limited food supply. Ivankov and Shershnev (1968) reported that young pink and chum salmon (50 to 80 mm) had fuller stomachs in years of "scarcity" of salmon than in years of "abundance" in the coastal zone of the southern Kuril Islands. The survival of fry to a large extent depends on their rate of growth and on their ability to escape from predators. Rapid growth requires suitable temperature, an abundance of food, and a rapid transition from endogenous nutrition, based on yolk reserve, to exogenous feeding on small aquat- ic organisms. In a study of size-selective preda- tion, Parker (1971) demonstrated that predation decreases with increase in size of the prey species. The study reported in this paper was under- taken in a southeastern Alaska estuary. Traitors Cove (Figure 1), in 1964-66 to gain further insight into the food requirements and feeding habits of pink and chum salmon fry. Questions asked were: How soon in life does feeding begin? How does the diet of the fry compare with the available food organisms? What are the food consumption rates Manuscript accepted January 1975. FISHERY BULLETIN: VOL. 73, NO. 4. 846 BAILEY ET AL.; ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY °"^°'- Ef-TR^^ INDtX MAP Figure 1. -Traitors Cove estuary, Revillagigedo Island, Alaska, 1963-65 (from McLain 1968), showing locations of plankton sampling stations. for fry in relation to water temperature? How many fry can the estuary support based on es- timates of abundance of food organisms and grazing rates? Traitors Cove is about 50 km nortli of Ketchikan, Alaska. Several tributaries used by pink and chum salmon enter Traitors Cove, the major one being Traitors River, which has about 55,000 m^ of spawning grounds. The dominant feature of the estuary is a narrow constriction with a sill, 1 or 2 m below mean low water, which divides the estuary into two basins. The inner bay is about 5.9 km long and 0.7 km wide and has a maximum depth of 46 m. The outer bay is about 6.5 km long and 1.3 km wide and has a maximum depth of 130 m. The tidal range of about 7 m and the constricted flow at the sill create exceptionally strong currents and a reversing tidal falls throughout the year. The tur- bulence and surface currents affect distribution and movement of fry for at least 0.5 km on both sides of the constriction. We measured surface temperatures of 5° to 13 °C in the estuary when fry were present. Some aspects of the oceanography of Traitors Cove have been described by McLain (1968). Pink and chum salmon fry from the tributary streams enter the estuary from mid-April to late June. Schools with thousands of fry are typically present until late June. METHODS To determine if juvenile salmon feed while still in Traitors River, we compared the contents of the entire digestive tracts of individuals excavated from redds with those trapped in nets while migrating downstream at night. All specimens were preserved whole in 10% Formalin solution.- The contents of the digestive tract were later removed in the laboratory and examined under a stereoscopic microscope. To determine the kinds and numbers of food organisms eaten in the es- tuary, we compared stomach contents of fry samples collected in the estuary in 1964, 1965, and -Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA." 847 FISHERY BULLETIN: VOL. 73, NO. 4 1966. Individual food items were measured to the nearest 0.01 mm of body length and diameter with an ocular micrometer to determine volume. Fry in the estuary were collected with a dip net from a skiff and by floating traps anchored near the shoreline. The dip nets and traps were effective in collecting fry less than 60 mm long, which are the subject of this report, but were not effective in collecting larger salmon. The larger fish were able to evade capture by sounding. Most of the fry examined for stomach contents were collected from the estuary during daylight (1100 to 1500). On three occasions, however, fry were collected during nights (0230) when the sky was overcast or moonless and incident light intensity was 0.0 footcandle near the water surface. No stomachs were collected during bright moonlight nights, which were rare. To estimate the volume of water grazed per day by fry, we measured velocities of water currents close to shore-oriented fry while observing their behavior in relation to the current and food items. Current velocities close to shore-oriented schools of fry were measured by two methods. One method was to record the time it took suspended particles in the water to drift 1 to 5 m along a floating anchored line graduated to 0.1 m. The second method was to measure the velocity by holding a current meter near a school of fish; the meter was attached to the end of a rod about 4 m long. The current meter dial was calibrated to read to the nearest 3 cm/s. Both methods required the ob- server to operate either from an anchored skiff or from shore. Polaroid glasses were used to reduce glare from the water surface and improve visibility. Feeding at night by fry was tested in two experiments in an aquarium with known densities of zooplankton. The aquarium consisted of a 7-mil plastic hag suspended in the estuary from a float and containing 76 liters of seawater. The fry were captured in the outer bay and held in a 1-m no. 10 mesh (158-ju,m openings) plankton net for 20 h to deny them food and to ensure that their stomachs were empty. The starved pink and chum salmon fry together with a known quantity of zooplank- ton were then placed in the aquarium and held under various light intensities. After tim.ed inter- vals in the aquarium, fry were removed, killed, and their stomach contents removed. In the first experiment, fry (length, 32 to 41 mm) in groups of five were placed in an aquarium that contained 240 zooplankters per liter of seawater. A cursory examination of the zooplankters revealed that they were predominantly copepods and bar- nacle nauplii. Each group of fish was allowed to feed 13 to 28 min before being removed and preserved in 10*? Formalin solution. The experiment was started in the evening before sunset and continued until the light meter read 0.0 footcandle. In the second experiment, 14 fry were placed in an aquarium containing about 260 zooplankters per liter; they were kept there for 4 h and 20 min at night before they were removed and preserved in Formalin. The light meter read 0.0 footcandle throughout the experiment. The time required for fry to evacuate their stomach contents was determined experimentally at 8.5°, 10.0°, and 12.8°C. The procedure was to capture 200 to 300 salmon fry in the estuary and place them in strained seawater in a floating cage of no. 10 mesh plankton net, which prevented entry of prey from the surrounding water. At the start of each test, 10 pink and 5 chum salmon were killed and preserved; at hourly intervals thereafter 5 fish of each species were killed and their stomach contents examined until all 10 fish of two successive samples contained no food in their stomachs. The pink salmon fry examined in these tests ranged from 32 to 57 mm in length and the chum salmon from 34 to 54 mm. Water tempera- tures were recorded by a thermograph to the nearest 1°C. The sensing probe of the ther- mograph was located 1 m below the water surface. The zooplankton in Traitors Cove was sampled only in 1965 and 1966 while fry were in the bay. A 5-inch Clarke-Bumpus sampler with a no. 10 mesh net was towed at a depth of about 0.5 m until about 50 liters of water (2 to 10 s) had been strained. Seventy-nine samples were collected-in 1965, 7 stations in the outer bay were each sampled in 1 day; and in 1966, 10 stations in the inner bay and 14 in the outer bay were each sampled on 3 different days (Figure 1). Only one sample was taken at each station. The zooplankton catch was preserved in 5% buffered Formalin solution. The plankton samples were subsampled by two methods for analysis. In the first method, each of the 79 samples was analyzed from 1-ml sub- samples (approximately 1/100 of the sample) placed in a Sedgewick-Raf ter chamber. The kinds, numbers, and size of the various plankters were determined. Volumes of the different plankters were computed from lengths and average diameters, assuming a cylindrical shape for each plankter. In the second method, all 7 of the 1965 848 BAILEY ET AL.: ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY samples and 66 of the 1966 samples were examined to determine numbers of plankters from larger subsamples (764 to '4 of the sample) taken with a Gushing subsampler (Gushing 1961). A comparison of the results of the two analyses indicated that the data from 1-ml subsamples overestimated the number of organisms by an average of 20"^^ (range 15 to 30%). Therefore, estimates of zooplankton densities using the first method were reduced by 20%. Plankton samples contained protozoans (mostly tintinnids) and phytoplankton, but these were not included in the estimate of standing stock of plankton because salmon fry consumed only the larger zooplankters. Rotifers and copepod nauplii were the smallest plankters included in the counts. FEEDING IN TRAITORS RIVER Although most of the pink and chum salmon excavated from redds in Traitors River contained items such as sand or detritus in their digestive tracts, only a few individuals contained food or- ganisms. Ghironomids (dipterans) were the most frequently observed food item. Seventy juvenile pink salmon (fork length, 33 to 41 mm) were collected from spawning gravels for analysis of contents; only three contained food— a chironomid pupa and some unidentifiable insect remains (Ta- ble 1). Seventy juvenile chum salmon (fork length, Table 1. -Frequency of occurrence of items in digestive tracts of 70 pink and 70 chum salmon juveniles excavated from redds in Traitors River in 1964-65. Pink sa Imon Chum salmon Item Number Percent Number Percent Arachnids 0 0 2 3 Ephemeropterans 0 0 2 3 Plecopterans 0 0 2 3 Dipterans 1 1 4 6 Insect remains 2 3 3 4 Plant detritus 15 21 16 23 Fine sand 33 47 44 63 Empty 22 31 12 17 Table 2. -Frequency of occurrence of items in digestive tracts of 40 pink salmon fry and 40 chum salmon fry trapped in nets while migrating down Traitors River, May 1964. Pink salmon Chum salmon Item Number Percent Number Percent Plecopterans 0 0 2 5 Dipterans 0 0 9 22 Insect remains 0 0 2 5 Detritus 1 2 1 2 Fine sand 8 20 22 55 Empty 31 78 19 48 33 to 41 mm) were collected from spawning gravels; nine contained food. Ghum salmon had eaten only chironomid larvae and pupae, plecopteran nymphs, ephemeropteran nymphs, and an arachnid (spider). One chum salmon (41 mm) contained the remains of 24 chironomid pupae, 2 chironomid larvae, 3 ephemeropteran nymphs, and 3 plecopteran nymphs. The other eight chum salmon that contained food were 37 to 38 mm long and had eaten only one to three items each. Although none of the 40 downstream-migrating pink salmon fry (length, 32 to 37 mm) contained food, 9 of the 40 chum salmon (length, 35 to 42 mm) contained substantial numbers of chironomid pupae and plecopteran nymphs (Table 2). The average for those that contained food was 6.7 food items (range 1 to 27 items). Fine sand (diameter, 0.05 to 0.90 mm) and plant detritus were common items in the digestive tracts of both the gravel-resident and the migrat- ing pink and chum salmon (Tables 1, 2). The sand and detritus were more common in fish taken from the redds than in those captured in the down- stream traps. FEEDING IN THE ESTUARY We studied four aspects of the feeding of pink and chum salmon fry in the estuary: 1) stomach contents, 2) feeding behavior in relation to water currents, 3) effect of daylight on feeding, and 4) time required for evacuation of stomach contents. Stomach Contents of Pink Salmon Fry In the springs of 1964, 1965, and 1966, a total of 140 pink salmon (length, 28 to 56 mm) were collected from the estuary during daylight, and 30 (length, 31 to 58 mm) were collected at night (Ta- ble 3). All of the stomachs from the fry collected in daylight contained food. Gopepods (calanoids and cyclopoids) occurred in 94% of the stomachs and constituted 77% of the total volume of stomach contents. Barnacle nauplii (cirripedes) and cladocerans (Podon sp. and Evadne sp.) each oc- curred in 56% of the stomachs and constituted 6% of the total volume. The remaining 11% of the food volume consisted of various other planktonic forms and occasional epibenthic organisms. Most of the food items were between 0.3 and 3.0 mm long. The smallest item in pink salmon stomachs was a disc-shaped diatom and the largest were fish 849 FISHERY BULLETIN: VOL. 73. NO. 4 Table 3.-Zooplankters and other organisms from stomachs of 140 pink salmon fry (length, 28 to 56 mm) collected in daylight and 30 (length, 31 to 58 mm) collected at night in Traitors Cove, 1964-66, and percentage relative importance by volume. Collected in daylight Collected at night Percentage stomachs containing item Mean items stomach per Percentage relative importance by volume' Percentage stomachs containing item Mean items stomach per Percentage relative importance by volume' Item Numbe r Percent Numbe r Percent Diatoms 32 3.3 3 + 26 0.4 2 + Rotifers 15 4.0 3 + 0 0.0 0 0 Bryozoans (cyphonautes) 2 0.0 0 + 0 0.0 0 0 Gastropods (veligers) 12 0.5 0 + 3 0.1 1 + Pelecypods (veligers) 26 0.9 1 + 6 0.1 1 + Polychaetes (larvae) 31 1.1 1 1 9 0.3 2 1 Arachnids 2 0.0 0 + 3 0.0 0 + Cladocerans 56 10.3 8 6 9 0.7 4 3 Copepods 94 70.7 52 77 76 10.7 67 85 Cirripedes (nauplii) 56 18.6 14 6 53 2.1 13 5 Cirripedes (cyprids) 25 1.9 1 2 9 0.4 2 3 Cirripedes (casts) 2 0.1 0 + 9 0.1 1 + Mysids 4 0.1 0 + 0 0.0 0 0 Cumaceans 3 0.1 0 + 0 0.0 0 0 Isopods 1 0.0 0 + 0 0.0 0 0 Amphipods 4 0.0 0 + 0 0.0 0 0 Euphausiids (larvae) 2 0.1 0 1 0 0.0 0 0 Decapods (zoeae) 9 0.3 0 1 0 0.0 0 0 Unidentified crustaceans (nauplii) 23 1.4 1 + 9 0.1 1 + Dipterans (larvae) 3 0.0 0 + 3 0.0 0 + Dipterans (pupae) 6 0.1 0 + 3 0.0 0 + Larvaceans 26 1.8 1 3 9 0.2 1 2 Eggs (Invertebrates) 49 20.8 15 3 23 0.8 5 1 Fish 4 0.0 0 + 0 0.0 0 0 '+ indicates less than 0.5% larvae (up to 8 mm long). Unidentifiable material occurred in only 11% of the stomachs and consti- tuted an insignificant fraction of the volume. The 30 pink salmon fry collected from the es- tuary at night all had food in their stomachs, but they probably had not feed recently. Many more food items were found in the stomachs of pink salmon fry collected in daytime than in those collected at night-an average of 136 items versus 16. Also, digestion had not progressed as far in the daytime fry-only 11% of their stomachs contained unidentifiable items, whereas 80% of the stomachs from nighttime fry contained unidentifiable items. On three moonlight nights, fry were seen dimpling the water surface while apparently feeding. Incident light intensity at the water sur- face at such times was 0.016 to 1.0 footcandle. Stomach Contents of Chum Salmon Fry In the springs of 1964, 1965, and 1966, a total of 124 chum salmon (length, 32 to .51 mm) were collected from the estuary during daylight and 20 (length, 35 to 43 mm) were collected at night (Table 4). All of the fry taken during daylight con- tained food. Copepods occurred in 73% of the stomachs and constituted 30% of the total food volume. Larvaceans occurred in 54% of the stomachs and constituted 34% of the total food volume. Dipteran (chironomid) pupae occurred in 51% of the stomachs and constituted 11% of the volume. The remaining 25% of the food volume was primarily other planktonic forms (including cladocerans and eggs) but also a few epibenthic animals. Unidentifiable material occurred in 20% of the chum salmon stomachs but constituted an insignificant fraction of the volume and was not included in the final comparisons. Food items eat- en by chum salmon fry were similar in size to those eaten by pink salmon, mostly from 0.3 to 3.0 mm long. The largest item was a larval fish 20 mm long. Chum salmon fry, however, tended to feed on larger (Table 5) and harder shelled items than pink salmon, as evidenced by the greater incidence of harpacticoid copepods, collembolans (intertidal springtails), cumaceans, and chironomids in the chum salmon (Tables 3, 4). The chum salmon fry could have picked up some of the so-called epibenthic or intertidal organisms in the form of neuston, or drift material. Many more food items were found in the stomachs of the chum salmon collected in daytime than in those collected at night-an average of 124 items versus 4. Only 20%of the stomachs collected in daytime contained unidentifiable items versus 70% at night. 850 BAILEY ET AL.: ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY Table 4.-Zooplankters and other organisms from stomachs of 124 chum salmon fry (length, 32 to 51 mm) collected in daylight and 20 (length, 35 to 43 mm) collected at night in Traitors Cove, 1964-66, and percentage relative importance by volume. Collected in daylight Collected at night Percentage Mean items per Percentage Percentage Mean tems per Percentage stomactis stomach relative stomachs stomach relative containing item importance by volume' containing item importance by volumei Item Number Percent Numbe r Percent Diatoms 15 0,4 1 + 4 0.2 5 + Rotifers 7 0.5 1 + 0 0.0 0 0 Gastropods (vellgers) 3 0.2 0 + 0 0.0 0 0 Pelecypods (veligers) 14 0.4 1 + 0 0.0 0 0 Polychaetes (larvae) 21 1.3 2 2 4 0.3 7 6 Arachnids 9 0.1 0 2 0 0.0 0 0 Cladocerans 58 12.9 18 8 0 0.0 0 0 Ostracods 1 0,0 0 + 0 0.0 0 0 Copepods 73 16.3 22 30 39 1.7 41 37 Cirripedes (nauplii) 34 2.3 3 1 29 0.5 12 4 Cirrepedes (cyprids) 20 0.6 1 1 14 0.1 2 1 Cirripedes (casts) 1 0.0 0 + 0 0.0 0 0 Cumaceans 6 0.1 0 1 0 0.0 0 0 Isopods 2 0.0 0 + 0 0.0 0 0 Amphipods 3 0.0 0 + 0 0.0 0 0 Euphausiids 1 0.0 0 + 0 0.0 0 0 Decapods (zoeae) 21 0.4 1 2 0 0.0 0 0 Unidentified crustaceans (nauplii) 10 0.8 1 + 4 0.0 0 + Collembolans 18 0.4 1 1 0 0.0 0 0 Dipterans (larvae) 10 0.2 0 + 9 0.1 2 2 Dipterans (pupae) 51 3.4 5 11 59 1.3 31 50 Dipterans (adults) 4 0.1 0 + 0 0.0 0 0 Unidentified insect remains 6 0.1 0 + 0 0.0 0 0 Larvaceans 54 18.2 25 34 0 0.0 0 0 Eggs (invertebrates) 19 14.0 9 4 4 0.0 0 + Fish 6 0.1 0 3 0 0.0 0 0 indicates less than 0.5% Feeding Behavior in Relation to Water Currents Our visual observations of individual chum and pink salmon fry in shore-oriented schools indicat- ed that their feeding varied with the speed of the water currents. At velocities of 0 to 10.7 cm/s, a fry would typically swim a darting course as much as three times its body length to capture a food item. At higher velocities, 10.8 to 19.8 cm/s, schools of fry sometimes held position relative to the shore or bottom while facing the current, and an individual would typically deviate up, down, or to the sides no more than one-third of its body length to capture oncoming food. At still higher velocities, 19.9 to 24.4 cm/s, fry in schools often held a constant position relative to shore or bottom but did not feed. Fry that appeared to be in visual contact with the shore or bottom avoided currents above 24.4 cm/s unless frightened. Effect of Daylight on Feeding The cessation of feeding at night by pink salm- on fry was confirmed by the two feeding experiments we conducted in the aquarium. In the first experiment, feeding rate was directly related to light intensity. During a 78-min period when light intensity ranged from 65 to 170 footcandles (three tests), the average consumption was 2.2 to 3.1 zooplankters per minute per fry (Figure 2). At light intensities of 2 footcandles or less, the average feeding rate was only 0.5 zooplankter per minute per fry (three tests). In the second experiment, performed entirely in darkness, little feeding took place. One fry had eaten 48 plankters (less than 0.2 plankter per minute), and the remaining 13 had eaten 13 plankters (0 to 0.001 plankter per minute). These observations agree with laboratory experiments of Hoar (1942) in which young salmon fed little during darkness. Z Q !i; o 5 u o.s - • • - • • - • ' ]• 1 1 1 al J • 1 1 1 1 1 1 1 1 0 0 016 0.032 0 06S 0 HO 0 260 O.SOO 16 11 65 130 260 FOOT CANDLES Figure 2. -Effect of darkness on feeding rate of pink salmon fry confined in an aquarium. Each dot represents a single test of feeding rate. 851 FISHERY BULLETIN: VOL. 73, NO. 4 Table 5.— Average size of zooplankters and other organisms collected by Clarke-Bumpus sampler and present in the stomachs of pink and chum salmon fry at Traitors Cove, 1965-66. Number Len gth (mm) Diamete r (mm) Average Item and place collected measured Average Range Average Range volume (mm3) Diatoms: Clarke-Bumpus sampler 24 0.09 0.06- 0.11 0.18 0.12-0.23 0.0023 Pink salmon 9 0.10 — 0.28 0.21-0.31 0.0062 Chum salmon 4 0.10 — 0.31 0.30-0.33 0.0075 Tintinnids: Clarke-Bumpus sampler 1 0.25 — 0.14 — 0.0038 Pink salmon — — — — — — Chum salmon — — — — — — Hydromedusans: Clarke-Bumpus sampler 2 0.18 — 0.31 — 0.0136 Pink salmon — — — — — — Chum salmon — — — — — — Rotifers: Clarke-Bumpus sampler 9 0.24 0.18- 0.31 0.16 0.11-0.20 0.0048 Pink salmon 9 0.32 0.30- 0.36 0.19 0.16-0.23 0.0091 Chum salmon 2 0.32 0.31- 0.33 — — 0.0091 Gastropods (veligers): Clarke-Bumpus sampler 1 0.20 — 0.14 — — Pink salmon 9 0.58 0.30- 0.75 0.38 0.20-0.49 0.0658 Chum salmon 1 0.68 — 0.44 — 0.1034 Pelecypods (veligers): Clarke-Bumpus sampler 4 0.28 0.19- 0.40 0.18 0.15-0.21 0.0071 Pink salmon 12 0.34 0.22- 0.42 0.31 0.22-0.38 0.0257 Chum salmon 6 0.35 0.28- 0.38 0.32 — 0.0281 Polychaetes (larvae): Clarke-Bumpus sampler 2 0.46 0.38- 0.55 0.17 0.13-0.21 0.0104 Pink salmon 11 0.94 0.61- 1.60 0.24 0.19-0.30 0.0425 Chum salmon 10 2.04 1.01- 4.00 0.29 0.29-0.47 0.1347 Arachnids: Clarke-Bumpus sampler — — — — — — Pink salmon 1 1.20 — 1.20 — 1.3572 Chum salmon 9 1.38 0.30- 1.80 1.38 0.18-3.50 2.0641 Crustaceans (nauplii): Clarke-Bumpus sampler 11 0.30 0.23- 0.43 0.14 0.10-0.19 0.0046 Pink salmon 7 0.37 0.28- 0.45 0.17 — 0.0084 Chum salmon 7 0.45 0.30- 0.50 0.21 — 0.0156 Cladocerans: Clarke-Bumpus sampler 11 0.50 0.31- 0.71 0.28 0.20-0.39 0.0308 Pink salmon 33 0.62 0.32- 0.91 0.33 0.17-0.49 0.0530 Chum salmon 31 0.60 0.20- 1.10 0.32 — 0.0482 Copepods: Clarke-Bumpus sampler 13 0.62 0.40- 1.23 0.19 0.12-0.32 0.0176 Pink salmon 98 1.00 0.26- 3.20 0.37 0.10-1.12 0.1075 Chum salmon 62 1.12 0.30- 3.20 0.41 0.11-1.19 0.1479 Cirripedes (nauplii): Clarke-Bumpus sampler 24 0.39 0.39- 0.61 0.22 0.16-0.42 0.0148 Pink salmon 28 0.47 0.28- 0.82 0.29 0.17-0.50 0.0310 Chum salmon 10 0.55 0.30- 1.20 0.34 0.18-0.56 0.0499 Cirripedes (cyprids): Clarke-Bumpus sampler 1 0.53 — 0.28 — 0.0326 Pink salmon 18 0.85 0.60- 1.00 0.37 0.26-0.46 0.0914 Chum salmon 14 0.84 0.62- 1.00 0.37 0.27-0.46 0.0903 Mysids: Clarke-Bumpus sampler — — — — — — Pink salmon 1 2.10 — 0.25 — 0.1031 Chum salmon — — — — — — Cumaceans: Clarke-Bumpus sampler — — — — — — Pink salmon 2 1.80 1.50- 2.10 0.46 — 0.2991 Chum salmon 3 2.11 1.52- 2.50 0.54 — 0.4832 Isopods: Clarke-Bumpus sampler — — — — — — Pink salmon 1 0.76 — 0,50 — 0.1492 Chum salmon 3 0.62 0.52- 0.80 0.31 0.26-0.40 0.0468 Amphipods: Clarke-Bumpus sampler — — — — — — Pink salmon 1 1.48 — 0.59 — 0.4046 Chum salmon 3 1.05 0.90- 1.25 0.42 — 0.1455 Euphausilds: Clarke-Bumpus sampler — — — — — — Pink salmon 3 2.67 2.50- 2.70 0.48 — 0.4832 Chum salmon 1 2.70 — 0.49 — 0.5092 852 BAILEY ET AL.: ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY Table 5.-Continued. Len gth (mm) Diameter (mm) Number measured Average Item and place collected Average Range Average Range volume (mm3) Decapods (zoeae): ClaFke-Bumpus sampler — — — — ^__ Pink salmon 10 1.79 0.54- 5.80 0.39 0.12-1.27 0.2138 Chum salmon 10 2.05 1.28- 3.04 0.45 0.28-0.67 0.3260 Collembolans: Clarke-Bumpus sampler — — — — .^ Pink salmon __ Chum salmon 14 1.57 0.67- 1.94 0.41 0.20-0.50 0.2073 Dipterans (larvae): Clarke-Bumpus sampler — — — — Pink salmon 3 1.43 1.20- 1.60 0.13 0.11-0.14 0.0190 Chum salmon 8 2.96 1.40- 4.00 0.27 0.13-0.37 0.1695 Dipterans (pupae): Clarke-Bumpus sampler — — — — Pink salmon 1 2.00 — 0.39 — 0.2389 Chum salmon 21 1.80 1.20- 2.90 0.43 0.27-0.70 0.2614 Dipterans (adults): Clarke-Bumpus sampler — — — ^_ Pink salmon — — — — Chum salmon 4 2.58 1.60- 3,50 0.39 0.24-0.52 0.3082 Unidentified insect remains: Clarke-Bumpus sampler — — — — -» Pink salmon — ^^ ^_ Chum salmon — 0.52 — — 0.0616 Larvaceans: Clarke-Bumpus sampler 1 0.20 0.14 0.0031 Pink salmon 9 0.69 0.50- 1.05 0.39 0.28-0.60 0.1498 Chum salmon 30 0.69 0.41- 1.30 0.39 0.23-0.74 0.1498 Polyzoans: Clarke-Bumpus sampler — — — — Pink salmon 1 0.60 — 0.19 — 0.0032 Chum salmon — Eggs (invertebrate): Clarke-Bumpus sampler 7 — — 0.32 0.20-0.40 0.0172 Pink salmon 11 — — 0.31 0.11-0.40 0.0156 Chum salmon 14 — — 0.34 0.10-0.42 0.0206 Eggs (vertebrate): Clarke-Bumpus sampler — — — Pink salmon 5 — — 0.88 0.80-0.98 0.3568 Chum salmon 1 — 0.85 0.3216 Fish: Clarke-Bumpus sampler — — — — —, Pink salmon 3 5.67 4.00- 8.00 0.14 — 0.0873 Chum salmon 3 16.70 15.00-20.00 0.42 — 2.3137 Stomach Evacuation The time required for satiated fry in the aquarium to evacuate food in their stomachs was inversely related to temperature. In tests at 12.8°C, the stomachs of two of five pink salmon were empty after 2 h without food. However, 6 h elapsed before successive samples of five fish con- tained no food, and 6 h was therefore accepted as the time required for pink salmon to evacuate their stomachs at a temperature of 12.8°C (Table 6). For chum salmon, the first empty stomach was observed after 1 h without food at 12.8°C. Only after 10 h did successive samples of five chum salm- on have empty stomachs. This longer evacuation time for chum salmon probably resulted from the larger and different kinds of organisms eaten. Using the same criterion for time of evacuation as described above for 12.8°C, pink salmon fry Table 6.— Time required for pink and chum salmon fry to evacuate food from their stomachs at various temperatures. Temperature (°C) Pink salmon (hours) Chum salmon (hours) 8.5 10.0 12.8 16 9 6 10 confined without food had empty stomachs after 9 h at 10°C and after 16 h at 8.5°C. We did not test chum salmon at the lower temperatures. ZOOPLANKTON ABUNDANCE AND DISTRIBUTION The abundance of zooplankton in the near-sur- face waters was determined from the samples we collected in the inner and outer bays of Traitors 853 FISHERY BULLETIN; VOL. 73, NO. 4 Cove in June 1965 and in April, May, and June 1966 when salmon fry were present. The lowest abun- dance in the inner bay, an average of 9 organisms per liter, occurred in April 1966, when the abun- dance was comparatively high in the outer bay, 51 per liter (Table 7). During the rest of the 1966 season, mean numbers ranged from 27 to 28 or- ganisms per liter in the inner bay and 24 to 40 in the outer bay. The highest numbers were observed in the outer bay in June 1965 after most of the fry had passed through the estuary. Zooplankters tended to be more abundant at the mouth of the bay, near the constriction, and at the head of the bay than at intervening points along the shoreline. Fifty- two categories of zooplankters were iden- tified from the Clarke-Bumpus samples, and seasonal qualitative and quantitative changes were evident in the composition of the zooplank- ton (Table 8). The peak abundance for polychaete larvae and cirrepede (barnacle) nauplii occurred in April, whereas the peak for other invertebrate larvae occurred in May. Rotifers, copepods (including nauplii), and barnacle nauplii were also very abundant in May. Cladocerans did not become abundant until June. Variation between years is indicated by the high abundance of ro- tifers in June 1965 (~120,000/mO and the much lower abundance of rotifers (-3,000/m') and possi- bly higher abundance of other forms in June 1966. The predominant zooplankters during the period of fry outmigration v/ere larvae of bar- nacles, polychaetes, and molluscs and nauplii and early copepodites of the copepods Acartia clausii, A. longiremis, and Oithona helgolandica. Over 98% of the zooplankters in the outer bay on 16 April 1966 were larvae, and as late as 7 June 1966 larvae constituted more than 65% of the zooplankton. In the inner bay on 18 April 1966 and 7 June 1966, the proportions of larvae in the zooplankton were 72 and 58%, respectively. Late copepodites and adults of calanoid and cyclopoid copepods were the next most abundant groups of zooplankters and con- tributed relatively more to the zooplankton as the season progressed. An abundance of larval forms was also characteristic of another southeastern Alaska estuary, Auke Bay (Wing and Reid 1972). Rotifers, although of minor importance in the diet of salmon fry, were often the most abundant zooplankters in the samples. Cladocerans and lar- vaceans were rare in April and May but by June constituted a significant portion of the zooplank- ton. Adults and juveniles of benthic invertebrates were rare in the plankton samples. Species com- Table 7. -Abundance of zooplankters determined from Clarke- Bumpus sampler with no. 10 mesh net (158)U,m) at Traitors Cove. Inner bay Outer bay Number of Organisms per liter Number of Organisms per liter Date samples Mean Range samples Mean Range 16 June 1965 7 154 8-563 16-18 Apr. 1966 10 9 1-28 14 51 6-180 16 May 1966 10 28 2-76 14 24 10-44 7 June 1966 10 27 2-62 14 40 4-95 position of zooplankters differed between the inner and outer bays of Traitors Cove (Table 8). The plankton samples contained zooplankton of the kinds and sizes eaten in great numbers by pink and chum salmon fry as well as smaller plankters, which were not important in the diet of fry. As a result, the average size of plankters in the net was slightly smaller than the average size of items eaten (Table 5). DISCUSSION Initiation of Feeding Neither pink nor chum juvenile salmon ate very much before leaving Traitors River, although chum salmon fed more than pink salmon. Some fry may have fed before they emerged from the redds. The size (41 mm) of the largest fry collected in the river suggests that at least a few individuals ac- tually grew as a result of exogenous feeding before they finally left the river. Mason (1974) collected chum salmon fry up to 70 mm long from Lymn Creek on Vancouver Island, British Colum- bia, where they moved into and out of high- salinity water and apparently fed in both media over a period of 1 to 4 wk or more. Immature stages of chironomids were most commonly eaten, but other bottom-dwelling aquatic organisms also occurred in stomachs of pink and chum salmon from Traitors River. Two workers (Disler 1953; Sparrow 1968) reported that zooplankton and bottom-dwelling aquatic or- ganisms occurred in the diet of chum salmon in freshwater. Although pink salmon apparently eat little or nothing while migrating seaward in short streams (Kazarnovskii 1962; Kobayashi 1968), as at Traitors River, they are more likely to feed while migrating long distances from large rivers (Levanidov and Levanidova 1957; McDonald 1960). Once they had left the stream, pink and chum salmon fry in Traitors Cove fed extensively on such zooplankters as calanoid copepods, lar- 854 BAILEY ET AL.: ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY Table 8. -Average species composition of zooplankton samples during salmon fry outmigration at Traitors Cove, June 1965' and April to June 1966. (See Table 4 for number of samples. Numbers of zooplankters per cubic meter rounded to nearest whole number. Percentages rounded to 0.1%; + indicates less than 0.5%.) Inner bay Outer bay 18 Apr. No. 1966 % 16 May 1966 No. % 7 June 1966^ 16 June 1965 16 Apr. No. 1966 % 16 May 1966 No. % 7 June 19662 Item No. % No. % No. % Hydromedusans: Bougainvillia sp. — — 1 + — — — — — — 2 + — — Obelia sp. 4 + 26 + — — 17 + — — 5 + — — Phialidium sp. — — 2 + — — — — — — 1 + — — Sarsia tubulosa (M. Sars) — — 1 + — — — — — — 1 + 45 + Unidentified 2 + — — — — 15 + — — — — — — Ctenophores: Pleurobrachia pileus (MiJIIer) — — 1 + — — — — — — — — — — Nemertines (pilidium) — — — — — — — — — — 3 + — — Rotifers 930 10.0 1,523 5.4 229 0.8 124,670 80.8 10 + 1,514 6.4 2,802 6.8 Bryozoans (cyphonautes) 10 + 30 + — — — — — — 108 + 89 + Molluscs; Littorina scutulata Gould (egg cases) — — 41 + 107 + 35 + 8 + 280 1.2 46 + Gastropoda (veligers) 221 2.4 345 1.2 84 + 53 + 102 + 127 0.5 — — Pelecypoda (veligers) 716 7.7 1,008 3.6 228 0.8 322 + 22 + 694 2.9 1,929 4.7 Polychaetes (larvae) 1,370 14.8 685 2.4 — — 62 + 891 1.7 547 2.3 205 0.5 Tardigrades 2 + — — — — — — — — — — — — Cladocerans: Evadne nordmanni Loven 2 + 4 + — — 419 + 4 + 8 + — — Podon leuckarti Sars — — 3 + — — 670 + — — 3 + — — Unidentified — — — — 1,096 3.8 — — — — — — 2,104 5.1 Ostracods — — 1 + — — — — — — — — — — Copepods (late copepodites and adults): Acartia clausii Giesbrecht 8 + 783 2.8 — — 3,087 2.0 2 + 47 + — — A. longiremis (Lilljeborg) 17 + 723 2.6 — — 336 + 110 + 148 0.6 — — Acartia spp. 79 0.9 4,571 16.2 — — 4,498 2.9 45 + 2,852 12.1 — — Calanus finmarchicus (Gunnerus) — — 399 1.4 — — — — — — 159 0.7 — — Centropages abdominalis Sato — — — — — — — — — — 9 + — — Metridia sp. 2 + 155 0.0 — — — — 2 + 111 + — — Pseudocalanus minutus (Kr^yer) 84 0.9 1,621 5.8 — — 95 + 143 + 862 3.6 — — Tortanus discaudatus (Thompson and Scott) 1 + 16 + — — — — — — 4 + — — Calanoids spp. 14 + — — — — 81 + — — 1 + — — Oithona helgolandica Claus 486 5.2 687 2.4 — — 981 0.6 469 0.9 514 2.2 — — Cyclopoids spp. 58 0.6 158 0.6 — — — — 7 + 134 0.6 — — Harpacticoids spp. 58 0.6 14 + — — 34 + 2 + 21 + — — Unidentified — — — — 10,422 36.5 — — — — — — 9,201 22.4 Copepods (nauplii) 1,980 21.4 10,237 36.3 — — 9,972 6.5 909 1.8 4,264 18.0 — — Cirripedes (nauplii) 2,334 25.2 4,041 14.3 10,261 35.9 6,994 4.5 48,067 93.9 9,714 41.1 13,706 33.4 Cirripedes (cyprids) — — 236 0.8 131 + 70 + — — 893 3.8 92 + Cumaceans: Cummella vulgaris Hart 1 + — — — — — — — — — — — — Amphipods: Corophiidae — — 1 + 50 + — — — — — — — — Euphausiids (calyptopis) — — — — — — — — — — 10 + — — Euphausiids (nauplii) 99 1.1 429 1.5 — — — — 206 + 168 0.7 — — Carids (zoeae) — — — — — — — — 2 + 1 + — — Brachyurans (zoeae) — — 16 + — — — — — — 7 + — — Pagurians (zoeae) — — — — — — — — 2 + — — — — Crustaceans (nauplii): Unidentified — — — — 5,955 20.8 — — 6 + 1 + 10,418 25.4 Chaetognaths: Sagitta elegans Verrill 4 + 2 + — — — — 9 + — — — — Echinoderms: Echinopleutel 4 + 32 + — — — — 1 + 19 + — — Bipinnaria — — — — — — — — — — 3 + — — Tunicates: Fritillaria borealis Lohmann 5 + 12 + — — — — 8 + — — — — Oikopleura sp. — — — — — — — — 11 + — — 92 + Tunicata (larvae) — — — — — — 107 + 2 + — — — — Tunicata (eggs) — — 20 + — — 12 + 33 + — — — — Unidentified invertebrate larvae 17 + 39 + — — 51 + 9 + 15 + — — Unidentified invertebrate eggs 764 8.2 309 1.1 — — 1,765 1.1 86 + 379 1.6 — — Fish larvae — — — — — — — — 1 + 1 + — — Fish eggs — — 5 + — — — — — — — — 362 0.9 Total 9,272 — 28,177 — 28,563 — 154,346 — 51,169 — 23,630 — 41,091 — 'No sampling was done in the inner bay in 1965. ^June 1986 samples were not available fortaxonomic breakdown. 855 FISHERY BULLETIN: VOL. 73, NO. 4 vaceans, barnacle nauplii, cladocerans, and other small crustaceans. Chum salmon fry tended to eat more larger hard-shelled organisms and epibenthic organisms than did pink salmon fry. The food of pink and chum salmon fry at Traitors Cove in general was similar to that reported at Uala and Anapka bays on the east side of the Kamchatka Peninsula (Andrievskaya 1968), the San Juan area of northern Washington (Annan 1958), the Strait of Georgia in southern British Columbia (Barraclough 1967; Robinson et al. 1968), Chatham Sound off the northern coast of British Columbia (Manzer 1969), and Moser Bay of southeastern Alaska (Chamberlain 1906). In con- trast, in Puget Sound epibenthic organisms (especially harpacticoid copepods) were more im- portant than pelagic zooplankters to pink and chum salmon fry (Gerke and Kaczynski 1972). Food Selection Salmon fry in Traitors Cove did not eat the same kinds and sizes of zooplankters in the same rela- tive numbers as they appeared in the samples of zooplankton, i.e., the fry fed selectively. Selective feeding in relation to sizes of prey and juvenile chum salmon has been reported by LeBrasseur (1969). The average size of the zooplankton eaten bv the fish was greater than the zooplankton collected by the Clarke-Bumpus sampler (Table 5). A coarser net such as a no. 6 mesh (233 /u,m) would probably have collected the zooplankters that were usually eaten by salmon fry and would not have collected so many of the small forms that are sel- dom eaten such as tintinnids, rotifers, and others. Selective feeding by pink and chum salmon fry was also demonstrated by the occurrence of cer- tain food items relatively more often in the stomachs of fry (Tables 3, 4) than in the plankton samples (Table 8). Relatively more cladocerans, decapod zoeae, and larvaceans were eaten by salm- on than appeared in the plankton samples. Another example of the marked disparity is the barnacle nauplii which were very abundant in most of the plankton samples (4 to 94% of the number of plankters) but constituted only 14% of the animals actually eaten by pink salmon and only 3% of the number of food items eaten by chum salmon. The high incidence of larvaceans in the stomach samples, especially in the chum salmon, may be the result of selective feeding on a scarce but very visible plankter. Larvaceans, in particular Oikopleura spp., form mucous feeding nets which may increase the visibility of the larvacean to the salmon fry. Once learning to capture Oikopleura, the fry may prefer that food item. Benthic and intertidal forms of mysids, cumaceans, isopods, amphipods, and insects were rare in the plankton samples and their presence in some of the stomachs shows that pink and chum salmon fry did on occasion feed in these ecological niches. This type of feeding behavior could not predominate at Traitors Cove because most of the shoreline is rocky and precipitous and offers little opportunity for benthic feeding. Grazing Rate The average number of zooplankters comsumed daily by a pink salmon fry in Traitors Cove was calculated from estimates of average stomach contents and evacuation rates. Stomachs of pink salmon collected from Traitors Cove estuary dur- ing daylight contained an average of 136 zooplankters. Stomach evacuation required 6 and 16 h at temperatures of 12.8°C and 8.5°C, respec- tively, although Brett and Higgs (1970) observed slower stomach evacuation rates at comparable temperatures in sockeye salmon fingerlings that had been fed a commercial pelleted food. The fry did not feed during darkness, which extended from about one-half hour after sunset to one-half hour before sunrise on cloudy or moonless nights. The duration of feeding at Traitors Cove when fry are present typically is about 16.5 h (range 15 to 18 h); the water temperature at 1 m ranges from 5° to 13°C. Thus, it appears that fry would consume a volume of food required to fill their stomachs once a day at cooler temperatures (8.5°C) and four times a day at warmer temperatures (12.8°C). The number of zooplankters consumed daily would, therefore, range between 136 and 544 per pink salmon fry for temperatures that are normal dur- ing the time fry are in Traitors Cove. By the same Hne of reasoning, chum salmon would consume about 120 to 480 food items per fry per day in Traitors Cove. Some insight into the availability of food for salmon fry at Traitors Cove was obtained by con- sidering the abundance of plankton in relation to the feeding habits of the fry. For example, fry 39 mm long that were holding a position relative to the shore while feeding in a current of 11 cm/s were in effect grazing a cylindrical mass of water 856 BAILEY ET AL.: ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY at the rate of about 3.5 liter/min. Even at the lowest observed abundance of 1 zooplankter per Hter (Table 5), each fry would theoretically en- counter about 3.5 zooplankters per minute, which is slightly greater than the estimated feeding rate of 3 zooplankters per minute in floating aquaria at 10°C. At this rate of feeding, a single fry could fill its stomach in about 39 to 155 min and could therefore easily ingest zooplankters faster than they could be evacuated. Abundance of zooplankton in the outer bay of Traitors Cove ranged from 4 to 563 organisms per liter (Table 7), and this was theoretically enough to satiate feeding fry as shown above. Furthermore, the abundance of zooplankton as estimated from Clarke-Bumpus samples did not decrease during the time that salmon fry were in the estuary. Therefore, we conclude that there was an abun- dant food supply in Traitors Cove for salmon fry. LeBrasseur et al. (1969), who conducted feeding experiments with wild juvenile pink and chum salmon in the Eraser River estuary in 1967 (an off-cycle year of pink salmon in the Eraser sys- tem), arrived at a similar conclusion for that area. Carrying Capacity of Traitors Cove Fry of pink and chum salmon emerged from the gravel of Traitors River at night, and most of them migrated to the estuary before dawn. Some of the fry, as evidenced by their size and the con- tents of their digestive tracts, lingered a few days in the stream where they fed on freshwater or- ganisms. The tendency to linger and feed in freshwater, most pronounced for chum salmon, has been described by Mason (1974). After the fry left Traitors River, they gathered in schools close to shore and began feeding and migrating ocean- ward. The time spent in the estuary is unknown but was probably from a few days to a few weeks. We estimated the abundance of pink and chum salmon fry in Traitors Cove by making counts each day along the shore from a moving skiff or by a mark-and-recapture technique. In 1965, the great- est number estimated from counts on any day was 7 million fry, but in 1966 the greatest estimate was under 1 million fry. The number of salmon fry in Traitors Cove in 1968 was estimated by mark and recapture to be 4 million (il.3 million, 95% confidence limits). The mark-and-recapture es- timate was made on a different annual fry migration than those covered by this study of feeding habits, but it strengthened our confidence in the visual estimates of fry abundance in 1965 and 1966. It did not appear that the Traitors Cove estuary was overgrazed by wild fry at the time of this study. In 1966, zooplankton abundance was always greater than 1.0 zooplankter per liter, which would allow maximum feeding rates by fry. During May and June 1966, when 1 million fry were present, the average abundance was about 29 zooplankters per liter. In June 1965, abundance was 154 zooplankters per liter after 7 million fry passed through the estuary. The number of fry that migrate through Trai- tors Cove each year is probably limited to less than 20 million by the productivity of the spawning grounds in Traitors River and Margaret Creek, the major salmon streams in the cove. We used stream survey data from Martin (1959) and applied a correction factor of 0.5 to correct for pools and stream bottoms of mud, sand, and bedrock to calculate 66,000 m^ of spawning grounds-55,000 m2in Traitors River and 11,000 m2 in Margaret Creek. These spawning grounds would yield about 7 million fry if they produced 100 fry per square meter or about 20 million fry if they produced 300 fry per square meter. Fry den- sities of 0.1 to 589 per square meter (average 250 fry per square meter) have been observed in Trai- tors River,''' but these densities were in sections of the stream consistently favored by spawning salm- on. Less favored areas were not sampled. The installation of a hatchery or spawning channel in a drainage system such as Traitors River could potentially result in a production of 100 million fry annually, or 5 to 100 times the es- timated production of wild fry. Available data are inadequate to determine the carrying capacity of Traitors Cove with certainty, but it is possible to make very speculative estimates based on the standing crop of zooplankton. Before presenting the estimates of carrying capacity, we wish to cite 10 necessary assumptions (required because we lack knowledge of the ecology of estuarine nursery areas) and some of the factors which may invalidate the estimates. 1. Zooplankton abundance was the same in Behm Canal as in Traitors Cove. Plankton samples ''Mattson, C. R., and J. E. Bailey. 1966. Chum and pink salmon studies at Traitors Cove, September 1963 to September 1964. On file, Auke Bay Laboratory. 'Mattson, C. R., and R. G. Rowland. 1963. Chum salmon studies at Traitors Cove Field Station June 1960 to March 1963. On file, Auke Bay Laboratory. 857 FISHERY BULLETIN: VOL. 73. NO. 4 were not collected outside Traitors Cove in Behm Canal. Several years later extensive plankton collections were made in open channels and several small adjacent bays of northern southeastern Alaska as a part of the 1972 MARMAP ■ investiga- tion. Within the May 1972 samples, average zooplankton abundance was nearly twice as great at 14 outside stations as at 4 stations within bays. Therefore, our estimates of carrying capacity based on influx of zooplankton from Behm Canal would be conservative. 2. Salmon fry were the only predators on zooplankton. We ignored the requirements of all other planktivorous animals of the area. The requirement of local planktivores other than salm- on are only qualitatively and poorly known. Herring were not seen in large numbers during the years of this food study. A school of herring entered the inner bay in 1967 while being fed on by a whale. We do not know how long these herring remained in Traitors Cove, but they were not conspicuous 2 wk after their entry. 3. Zooplankton concentrations were constant. We ignored the strong seasonality of reproduction and growth in the holoplankton and the fact that meroplankton may be present for only a limited time. We ignored the probability that some larval forms reach a life history stage where their behavior would make them unavailable to the salm- on fry. We ignored natural mortality of larval forms other than from predation by salmon fry. Some of these factors would increase zooplankton concentrations while others would decrease them. In the absence of information on reproduction, growth, mortality, and life histories, we assumed these factors would balance so that the zooplank- ton concentration would be constant. 4. Distribution of the zooplankton was uniform. Physical and biological factors controlling the patchiness of zooplankton in estuaries and near- shore environments are poorly understood and not easily modeled. 5. All zooplankton were equally available, equally desirable, and of equal quality as feed for salmon fry. We ignored the size selectivity and preference for calanoid copepods shown in our own data. It is highly probable that the species of zooplankton vary in quality as food. 6. Salmon fry had a constant feeding requirement of 544 zooplankters per day. This is Marine Resources Monitoring, Assessment, and Predic- tion-program sponsored by National Marine Fisheries Service on a nationwide scale. the highest of our estimates of pink salmon feed- ing rates and ignores variations in food requirements that would accompany variations in physical environment and physiological state. 7. No behavioral changes in either the salmon fry or zooplankton were induced by changes in densities, physical environment, or biological states. 8. The number of salmon fry was constant. 9. All the zooplankton would be utilized as food. If this actually occurred, no survivors would be left to produce new zooplankton crops or to replenish stocks of other resources that have planktonic lar- val stages such as herring, crabs, and shrimp. 10. Models of circulation in the estuary would be of the simplest type. We do not know the flushing rates in Traitors Cove or the potential of transport of zooplankton food to and from the bay by es- tuarine circulation. Some additional assumptions peculiar to each estimate are described with each estimate. Only the outer bay is considered because fry in Traitors Cove appeared to move quickly through the inner bay and then spend a longer time in the outer bay. Our first estimate of carrying capacity is based on standing stock of zooplankton in the top meter of water of the outer bay. Fry were in the outer bay in relatively high densities for about 30 days each year. The surface area of the outer bay is about 7.6x10^ m^, and the average density of zooplankters was estimated to be 24,000 per cubic meter or higher. The product of area and plankton density divided by 544 (the estimated maximal number of organisms consumed per day by pink salmon fry) results in a plankton stock equivalent to 335 X 10^ fry feeding days. This estimate divid- ed by 30 days expresses the food supply in fry months— 11 million fry could feed for 1 mo on the standing stock of food in the surface meter of outer bay. This establishes a lower limit for the carrying capacity of Traitors Cove because it ig- nores saltwater entrainment by outflowing fresh- water and the consequent addition of plankton from deeper water. For our second estimate, we calculated the quantity of zooplankton that would be brought into the outer bay from Behm Canal each day by a combination of two factors: circulation due to freshwater runoff from Traitors River and cir- culation due to tidal action. Records of the U.S. Geological Survey indicate that discharge from Traitors River generally averages about 8 mVs in 858 BAILEY ET AL,: ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY the spring when fry are migrating. Assuming an equal flow of seawater with plankton density of 24,000 organisms per cubic meter into Traitors Cove from Behm Canal and surface entrainment near the constriction, we calculated that 16.5 x 10^ organisms would be brought daily into Traitors Cove by freshwater-driven circulation. Dividing the number of organisms by 544 (the high estimate of organisms eaten by one fry daily) yields a con- servative estimate of 30 million fry that could be fed by an amount of food added daily by circula- tion. Although it is naive to assume that all of the plankton brought into Traitors Cove as a result of circulation would become available to the fry, the upwelling and thorough mixing that occur at the constriction between the two bays result in a con- tinual resupplying of zooplankton to the upper meter of depth where grazing apparently takes place. Field observations did indicate that the larg- est concentrations of fry were consistently found in eddies near the constriction, lending some credence to the theory that upwelling of deep water created a favorable supply of food in this area. The effects of tidal circulation and freshwater- runoff-driven circulation are often additive. Therefore, we calculated the influx of food or- ganisms by tidal exchange. We assumed that the surface waters were flushed completely by the outgoing tide; that complete mixing of incoming water with water present occurred on each tide; and that all zooplankton in the upper meter had been consumed before the waters were mixed. The influx of new food can then be estimated from the tidal prism as F = [T/V]xP where F is the net influx of new food organisms as zooplankters per cubic meter per tide; T is the volume of the tidal prism; V is the volume of the outer bay; and P is the density of zooplankton outside the bay. (We used P = 24x lOVm^ because we assumed that abundance was the same outside the bay as it was inside.) The resulting calculation assuming a mean tidal range of 4.11 m (McLain 1968) and a mean depth of 90 m gives for the net influx of organisms per tide: F = 4.11m /tide x 24 x lO^/m^ = 1.09 x 10^ 90 m zooplankters per cubic meter per tide. Only those in the upper meter are available, and since there are two tides per day, the calculated quantity of new food available to salmon fry per day is: Q = 2xFxrarea of bayjxl m = 2 tides/ day (1.09 x 10^ zooplankton per cubic meter per tide) x (7.6x lO^m^x Im) = 16.6 X 10^ zooplankters per day. This number will feed 30x10^ fry per day (16.6 x 10^ zooplankters -i- 544 zooplankters per day per fry). The estimate is high because mixing is not complete, as implied by the calculations. By adding fry that could be fed from the effects of freshwater runoff to fry that could be fed by tidal action, we get an upper estimate of carrying capacity of 60 million fry. The numbers of fry that could theoretically be fed by the two sources of zooplankton, i.e., stand- ing crop in the surface water and plankton in the net circulation, are not strictly additive. Although some plankton in deep seawater would be con- tinuously entrained upward to flow seaward on the surface, some would never reach the surface of the bay and a portion of the surface stock would be removed from the bay by outflow. Therefore, it would seem prudent to consider that populations numbering more than 30 million pink and chum salmon fry might cause reduced growth of fry (because of the competition for food). Also, such large populations might stimulate a more rapid migration of fry through the estuary to areas where food organisms were more abundant. On the basis of available spawning grounds, it seems unlikely that Traitors Cove has ever had to support more than 20 million pink and chum salm- on fry, although it is possible that 11 to 60 million fry could be supported in years when food abun- dance equaled or exceeded that observed in 1966. The release of 50 to 100 million additional hatchery fry into this estuary would probably exceed the carrying capacity of the area. Competition for food, especially if zooplankton production were lower than average, and increased potential in- fection by disease, parasitism, and predation could theoretically result in increased mortality, slower growth, or accelerated movement of fry out of the estuary. Further, a great increase in numbers of salmon fry in Traitors Cove could deplete plank- tonic food and planktonic larvae required to sup- port other fisheries. We have used Traitors Cove to discuss carrying capacity of estuaries only because 859 FISHERY BULLETIN: VOL. 73, NO. 4 observations on fry and food were available. We know of no plans for the operation of a hatchery in Traitors Cove. The discussion is merely intended to focus attention on an important factor to be considered in choosing sites and operating salmon fry hatcheries. LITERATURE CITED Andrievsk.aya, L. D. 1968. Feeding of Pacific salmon fry in the sea. [In Russ.] Izv. Tikhookean. Nauchno-issled. Inst. Rybn. Khoz. Okeanogr. 64:73-80. (Translated by Transl. Bur., Dep. Secretary of State of Canada, 1970, 16 p.) Annan, M. E. 1958. Notes on the food of the young of three species of Pacific salmon in the sea. Can. Fish Cult. 23:23-25. Bailey, J. E., and W. R. Heard. 1973. An improved incubator for salmonids and results of preliminary tests of its use. U.S. Dep. Commer., NOAA Tech. Memo. NMFS ABFL-1, 7 p. Bailey, J. E., and S. G. Taylor. 1974. Salmon fry production in a gravel incubator hatchery, Auke Creek, Alaska, 1971-72. U.S. Dep. Commer., NOAA Tech. Memo. NMFS ABFL-3, 13 p. Bams, R. A. 1972. A quantitative evaluation of survival to the adult stage and other characteristics of pink salmon (On- corhynchus gorbuscha) produced by a revised hatchery method which simulates optimal natural conditions. J. Fish. Res. Board Can. 29:1151-1167. Barraclough, W. E. 1967. Number, size and food of larval and juvenile fish caught with a two boat surface trawl in the Strait of Georgia April 25-29, 1966. Fish. Res. Board Can., Biol. Stn. Nanaimo, B.C., Manuscr. Rep. Ser. 922, 54 p. Brett, J. R., and D. A. Higgs. 1970. Effect of temperature on the rate of gastric digestion in fingerling sockeye salmon, Oncorhynchus nerka. J. Fish. Res. Board Can. 27:1767-1779. Chamberlain, F. M. 1906. Some observations on salmon and trout in Alaska. Rep. U.S. Comm. Fish., 1906 (Doc. 627), 112 p. Gushing, C. E., Jr. 1961. A plankton subsampler. Limnol. Oceanogr. 6:489-490. DiSLER, N.N. 1953. Ecological and morphological characteristics of the development of the Amur autumn chum salmon-Ow- corhynchus kefa (Walb.). |ln Russ. J Tr. Soveshch. Ikhtiol. Kom. Akad. NaukSSSR 1:354-362. (Translated by Israel Program Sci. Transl, 1961; In Pacific salmon, selected articles from Russian periodicals, p. 33-41; available U.S. Dep. Commer., Off. Tech. Serv., Wash., D.C., as OTS 60-51139.) Fraser, F. 1972. Evaluation of chum spawning channels (Abstr.). In J. E. Bailey (editor). Proceedings of the 1972 northeast Pacific pink salmon workshop, p. 55-64. Alaska Dep. Fish Game, Inf. Leafl. 161. Gerke, R. J., AND V. W. Kaczynski. 1972. Food of juvenile pink and chum salmon in Puget Sound, Washington. Wash. Dep. Fish., Tech. Rep. 10, 27p. Hoar, W.S. 1942. Diurnal variations in feeding activity of young salmon and trout. J. Fish. Res. Board Can. 6:90-101. IVANKOV, V. N., AND A. P. ShERSHNEV. 1968. (Biology of juvenile pink salmon and chum salmon in the sea.) [In Russ.] Rybn. Khoz. 44:16-17. Kanid'yev, a. N., G. M. Kostyunin, and S. A. Salmin. 1970. Hatchery propagation of the pink and chum salmons as a means of increasing the salmon stocks of Sakhalin. J. Ichthyol. 10:249-259. Kazarnovskii, M. Ya. 1962. Food of migrating fry of Oncorhynchus gorbuscha and Saivelinus nialma in the rivers of Sakhalin. [In Russ.] Rybn. Khoz. 38(6):24-25. (Abstr. translated from Russ. in Biol. Abstr. 41:(20992). KOBAYASHI, T. 1968. A note on the seaward migration of pink salmon fry. [In Jap., Engl, summ.] Sci. Rep. Hokkaido Salmon Hatchery 22:1-5. LeBrasseur, R. J. 1969. Growth of juvenile chum salmon {Oncorhynchus keta) under different feeding regimes. J. Fish. Res. Board Can. 26:1631-1645. LeBrasseur, R. M., W. E. Barraclough, 0. D. Kennedy, and T. R. Parsons. 1969. Production studies in the Strait of Georgia. Part III. Observations on the food of larval and juvenile fish in the Fraser River plume, February to May, 1967. J. E.xp. Mar. Biol. Ecol. 3:51-61. Levanidov, V. Y., and I. M. Levanidova. 1957. Food of downstream migrating young summer chum salmon and pink salmon in Amur tributaries. [In Russ.] Izv. Tikhookean. Nauchno-issled. Inst. Rybn. Khoz. Okeanogr. 45:3-16. (Translated by Israel Program Sci. Transl., 1961; In Pacific salmon, selected articles from Russian periodicals, p. 269-284; available U.S. Dep. Commer., Off. Tech. Serv., Wash., D.C.. as OTS 60-51139.) Manzer, J. I. 1969. Stomach contents of juvenile Pacific salmon in Chatham Sound and adjacent waters. J. Fish. Res. Board Can. 26:2219-2223. Martin, J. W. (editor) 1959. Stream catalog of eastern section of Ketchikan management district of southeastern Alaska. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 305 [615 p.]. Mason, J. C. 1974. Behavioral ecology of chum salmon fry (Oncorhynch us keta) in a small estuary. J. Fish. Res. Board Can. 31:83-92. McDonald, J. 1960. The behavior of Pacific salmon fry during their downstream migration to freshwater and saltwater nur- sery areas. J. Fish. Res. Board Can. 17:655-676. McLain, D. R. 1968. Oceanographic surveys of Traitors Cove, Revillagigedo Island, Alaska. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 576, 15 p. Parker, R. R. 1971. Size selective predation among juvenile salmonid fishes in a British Columbia inlet. J. Fish. Res. Board Can. 28:1503-1510. Robinson, D. G., W. E. Barraclough, and J. D. Fulton. 1968. Number, size composition, weight and food of larval 860 BAILEY ET AL.: ZOOPLANKTON ABUNDANCE AND FEEDING HABITS OF FRY and juvenile fish caught with a two-boat surface trawl in water of British Columbia. J. Fish. Res. Board Can. the Strait of Georgia May 1-4, 1967. Fish. Res. Board 25:599-602. Can., Pac. Oceanogr. Group, Nanaimo, B.C., Manuscr. Wing, B. L., AND G. M. Reid. Rep. Ser. 964, 105 p. 1972. Surface zooplankton from Auke Bay and vicinity, southeastern Alaska, August 1962 to January 1964. U.S. Sparrow, R. A. H. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Data Rep. 1968. A first report of chum salmon fry feeding in fresh 72, 765 p. on 12 microfiche. 861 EFFECTS OF TRAP SELECTIVITY AND SOME POPULATION PARAMETERS ON SIZE COMPOSITION OF THE AMERICAN LOBSTER, HOMARUS AMERICANUS, CATCH ALONG THE MAINE COAST- Jay S. Krouse and James C. Thomas^ ABSTRACT Information collected aboard commercial lobster boats along the Maine coast (1971-73) revealed, among other things, high numbers of sublegal lobsters (<81 mm carapace length) being handled by fishermen while sorting their catches. Throw-back ratios of illegal to legal lobsters (1.8 to 12.4:1) varied in association with lath spacing. Traps with spacings of 1% to 1% inches accounted for markedly fewer sublegals than those traps with 1'4- to lV2-inch spacings. Selectivity curves calculated for research traps with escape ports of V/2, 1%, and 1% inches and a trap escapement study demonstrate that a spacing of 1% inches is large enough to allow escapement of most sublegals yet small enough to retain legal lobsters. A regression of carapace length on carapace width shows that only an insignificant percentage of legal lobsters could physically squeeze through a P4-inch opening. Thus, results of this study led us to recommend that with a minimum legal length of 81 mm, traps should have 1%-inch escape vents. While riding aboard commercial lobster boats along the Maine coast (1971-73) to collect detailed catch and effort information, we frequently ob- served lobster fishermen sort and throw back from their traps excessive numbers of sublegal lobsters (<81 mm carapace length). When one considers that Maine lobstermen presently haul their traps more than 20 million times each year, the magni- tude of this sorting becomes apprent. Lobstermen not only lessen the efficiency of their fishing operations by needlessly handling sublegal lob- sters but they also inadvertently increase the lob- sters chances of becoming a cull (missing claw[s]) or a victim to predatory fish while descending to the ocean floor (D. G. Wilder, pers. commun.), which in either case represents an economic loss to the industry. A solution to this detrimental fishing practice became apparent to us after a cursory analysis of data from our earlier boat trips revealed an in- verse relationship between lath spacing and numbers of sublegal lobsters. Templeman (1939) and Wilder (1945, 1948, 1954) also reported the same relationship based on size composition of catches from traps of various lath spacings. 'This study was conducted in cooperation with the Department of Commerce, National Marine Fisheries Service, under Public Law 88-.309, as amended. Commercial Fisheries Research and Development Act, Project 3-153-R. -Maine Department of Marine Resources, West Boothbay Harbor, ME 0457.5. Although these Canadian scientists have long ad- vocated the use of wider latch spacings to allow sublegals to escape, presently only Newfoundland has a lath spacing regulation of 1% inches. Because of the management implications of this association between lath spacing and size com- position of the catch, we undertook this investiga- tion to quantitatively assess this situation with several independent approaches, namely: 1) selec- tivity curves; 2) trap escapement study; and 3) certain morphological dimensions of lobsters. Certain facets of this study were also valuable in corroborating some previously estimated popula- tion parameters such as natural mortality rates, sex ratios, and spawning stock structure and size. These analyses have become increasingly im- portant because we have recommended raising the legal minimum size from the present 81 mm (3'Vi6 inches) to 89 mm (3V2 inches) carapace length. Then this study not only has application for the present situation, but also provides pertinent in- formation for management of lobsters in the fu- ture. METHODS Samples from Commercial Gear From 1971 through 1973, we spent 21 days riding aboard nine different commercial lobster boats (three boats were sampled more than once) from Manuscript accepted December 1974. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. 862 KROUSE and THOMAS: EFFECTS OF TRAP SELECTIVITY ON LOBSTERS four coastal areas. These boats were selected on a nonrandom basis because not all vessel-captains could or would accommodate us, nor could we reallocate committed time from our ongoing sur- veys of the natural lobster population and the commercial catch. While aboard these vessels we recorded the following: 1) numbers of sublegal and legal lobsters for each trap haul; 2) carapace length and sex of lobsters from a systematic sample of the catch, along with the corresponding measurements of lath spacings in these traps; 3) time expended in actual fishing as well as fishing time for each trap (number of set-over-days); 4) whether the fisherman was hauling one trap at a time (singles) or two attached traps with one buoy (pairs), or three or more attached traps (trawls) with two buoys, one at each end of the string; and 5) amount and kind of bait used. Carapace lengths were measured in millimeters from the posterodorsal edge of the eye socket to the posterior margin of the carapace. In most cases, we attempted to measure all the lobsters in every nth trap (depended on whether traps were set as singles, pairs, or trawls); however, some- times with two samplers, we were able to measure and record all the lobsters in each of the total number of traps hauled for the day. Length compositions of the catches for each boat trip were used to calculate what we refer to as retention curves. These curves are simply an ac- cumulative percentage of the number of lobsters by 1-mm carapace increments that were retained in the systematic sample of the traps hauled, along with measurements of the lath spacings of these traps. Because lath spacings were not uniform for each of the traps hauled per boat, the term "modal spacing" was used to imply that at least a majority of the traps per boat had a spacing more frequently measured by us than any other. Samples from Research Gear Since 1968 we have recorded the carapace length, weight, sex, condition (hard or soft shell, lost appendages) of individual lobsters caught in our research traps. Our research gear consisted of: 1) modified wooden traps, with plastic escape vents of IV2, 1%, and 1% inches (Figure 1), and 2) 1 x 1 inch wire meshed traps especially designed to catch sublegal lobsters. The modified commercial gear was fished from July 1972 through 1973, while the wire traps were used since 1968. We also conducted a trap escapement study Figure 1.— Modified commercial lobster trap equipped with a plastic escape vent. whereby lobsters of known sizes were placed in wooden traps with vents of IV4, IV2, 1%, and 2 inches. Because the heads or entrances were sealed, any escapement should have been ac- complished between the laths. Through a 2-wk period traps were usually checked daily for es- capement. Comparison of Samples from Commercial and Research Gear Following the methodologies of Beverton and Holt (1957), Pope (1966), and Gulland (1969), we calculated selectivity curves which were based upon carapace lengths of lobsters retained in the commercial gear with modal lath spacings of IV4, 1%, and 1% inches. These data were proportioned with the same range of lengths retained in the 1x1 inch wire meshed research traps. Both sets of data, commercial and research, were weighted by trap-haul-set-over-days (THSOD). These com- parisons were from the same general area, but not with the same groups of traps nor necessarily during the exact same period of time. In addition, we used the cited methods to make selectivity determinations from the modified commercial traps that had specific lath spacings of IV2, 1%, and 1% inches. These spacings were proportioned with the data from the wire research traps (1x1 inch mesh). In this case, the modified commercial and wire research traps were fished 863 FISHERY BULLETIN: VOL. 73, NO. 4 simultaneously with the same spatial and tem- poral arrangements. Body Proportions of Lobsters To circumvent the spatial and temporal problems between commercial (boat trips) and research gear to a certain extent, we took body measurements of 217 lobsters, specifically carapace length, width, and height for sizes between 70 and 90 mm carapace length. These measurements should enable us to reach a more objective determination concerning the retention and escapement potential of various sized lobsters through different lath spacings. RESULTS AND DISCUSSION Samples from Commercial Gear For the 21 boat trips with commercial fishermen, we counted their entire catch of 12,071 lobsters of which there were 2,311 legal lobsters (Table 1). This catch resulted from 4,026 trap hauls for a catch of 0.57 legal lobsters per trap haul or 0.22 legal lobsters per THSOD. There are omissions in some of the data categories per boat trip because the sampling procedure evolved from successive trips aboard vessels; thus samplers learned by experience and observation what could or could not be ac- complished under different physical conditions in each vessel. Nevertheless, subtotals can be gleaned from the boat trips with the more complete information. For example, there were 156 berried and/or "V"-notched females from 18 of the 21 boat trips. Even though this is a subtotal, it is an alarmingly low number. Of course, such things as season of year, area fished, and availability of berried females could affect this number. Still, we continue to be concerned about the possibility of a precarious limit of an adequate spawning stock (Thomas 1973; Krouse 1973). The percent females is another estimate related to the reproduction potential of the exploited population of lobsters. For those lobsters that we measured and determined sex, 52.9% were females (sublegal and legal). This estimate is close to the 49.0 to 53.8% females that we estimated by year (1966-73) from the survey of the commercial (legal lobsters) and natural (mostly sublegal) population of lobsters. These estimates are in conflict with the expec- tation that there should be more males than females in the commercial catch because berried and/or V-notched females must be returned to the ocean by law. Again, this situation points to a low number of sexually mature females. To reach definite conclusions concerning the stock-progeny-recruitment relationships, we should follow the procedures of Beverton and Holt (1957) and Ricker (1958). This will be possible with continued support of this program and continued surveys on the commercial and natural popula- tions of lobsters. Length Frequencies We measured the carapace length of 3,595 lob- sters; the sex ratio (male:female) was 1:1.2. A his- togram of these length frequencies (Figure 2) portrays the same situation that we have demon- strated from the commercial and natural surveys. That is, there are relatively large numbers of lob- sters (2,937 or 81.7%) under the legal minimum size, while there are considerably fewer lobsters (658 or 18.3%) at and above the legal minimum size of 81 mm (3'^/ 16 inches) carapace length. In fact, 94.0% of the legal catch is constrained within a Va-inch size range immediately above the legal minimum size. These conditions confirm the high exploitation rate of 0.86 that can be calculated from Thomas (1973). Considering the modal lath spacings of traps used in each boat trip, there is a marked difference in the number of sublegal- and legal-sized lobsters. N.3595 CARAPACE LENGTH (MM) Figure 2. -Length-frequency distributions of lobsters caught in traps several lath spacings) used by commercial fishermen (1971-73). 864 KROUSE and THOMAS: EFFECTS OF TRAP SELECTIVITY ON LOBSTERS CO OS E a, •§. 3 o u t~ 1—* (0 « 0"D Z O L. (0 1 CO 0)(B ra ^ "O . 0) £ S- o = m CO Oo E TJ u a) o o o 0) TS c £ rt o J= n o eS O c coi: •o 73 0) o 0) cnJ^ O) .C £ c o c (0 CO T3 "a 0) m £ a.c C S I- I I I o in m cj ui CM to eg CNj eg CO CM CO CM (O I I CM CM 1- ■* w CO in CM CM CM CD CD in ■*- T— 1- ▼- CO CO CO T- 1— T— in ID CO CM T- .1- T- in o lo o lo CO o si CD CD CO CO S-' N CO lO CM CO o o CO s-' in c> o CM T-^ cm' CO N.' o o I CO s.' I C> p CO CO CM I I CM s« a> o) o oo n O O) o> s. m CO 1- CM CM »- •<»■ o in CM TT CD CO T- I- CO CO CM CO CM '- CM eg eg CO 0> Q O) Tf I?) t- CM o ■^ CD JO CO -^ CO r^ CO Si 1- »- T- ,- CM C> r^ CM s- 1- T- CO 00 O) CM CO CO CO m eg eg CM •r-' CO -r^ -r-" T-' a> CM 1^ CO CO CM CM »- ■ CD 1- r^ CD CM r^ CO O) CO o o o o o o o CO in TJ- in o CO ^' CO o> si eg CD CD ■<»■ CT m in m 1 m Ti- c3> Tf ^ CO m •» o CO o CO ■ 1- 7- 7- CM T- xr CO t 1- CD CO CM T- ■<»■ CM 1- CO O l-~ CO CM eg o CO CO m o o (? o o o o O '- O T- o cooo) Tj-min ^-oos r^ CM^^ O^O OCDCO •^ T- T-' r-^ CO 7- eg eg 1-' cm' r^-^LO rococo oco*^ r- CM TT CM 7- ^ CO in ID ooo c3C>o cJoci CCJCD l_^'~~. ,^^ °^'* ^ '^.^. ^. ■*<>? o n1 I £ s^ £ o> ^ Tt ^ T-' Tf T-' ^' T-^ eg' mm m m m^f mmm mmm com .s-ocom r^eocg t-ok cjom cmtt ^com^ m^s. ^cocm n ao a> T-cg icocM 7- Tt'-co CMcgm S-CO mCDS. CD III S.'J-CD ■fCMCD eg ■•- 1- 1- ■>- I I I m m CD CM C3> 00 CM CM 7- T- C\l O CO CO CD ■»■ O O O •* CO o I I m •>>• 1- •* o g T- o m o CM 2 o o o o S? 1 1 1 0) a. h- •r- y- 1- •^ T- •1- ^ 1- T- 1- T- 1- 1- ^ 1- CO Tl- CO •* ,- CO •^ CO o> s. co m m co CM CO CO CO CO C3> f^ eg eg o o 00 CM 00 eg CM 1- CM 00 m CO -f OT 1^ m CO ^ 00 ■<»■ CJ> CO •» c 0) CO <0 •0- 00 CM O) CO 00 ■* h^ o S- CD •» CO h~ Tf T- T- CD 0> •a- 00 ■■- CO o> •» o cj> CM eg CO ■>»■ •» O CO CM 1^ ■>«■ O) ■>!• CO eg m eg r^ TT CM CO ^ CO ^ CO CO CO eg CM eg CM ■o- ■■- CO m m o 0) <%3 cflC/DO<<^ Q. D 0) -^ a> o a> <^OT ™ CO Z W . . o ■ ■ • 00 <3> -o O 7- CM 0) CO .~f;r CM !^ ^^"^ coo-)<«a,<->-)«co ~« '.~^~.^^.'>^~. O'-eg' cosloo'o>'5c) T 2 O r^ r^ a •— o •o o c "E ^^E « o o (]> (D (a -IWQ 865 For example, the catches of boat trips numbered 3, 5, 7, 8, 20, and 21 (modal lath spacings of 1% or 1% inches) consisted of a low number of sublegal lob- sters compared to the high number of sublegals in catches of boat trips numbered 2, 4, 6, 9, 17, 18, and 19 (modal lath spacings of IV4 to IV2 inches) (Table 2). Effects of Throw Backs on the Fishery The throw-back ratios of illegal (sublegal plus berried and/or V-notched females) to legal lob- sters which ranged from 1.8 to 12.4:1, confirmed FISHERY BULLETIN: VOL. 73. NO. 4 our earlier observations that a considerable number of lobsters are being handled needlessly (Table 1). It is not difficult to envision sublegal, V-notched, or berried female lobsters spending a portion of their lives airborne. One of the impor- tant considerations in this situation is whether these lobsters suffer a higher natural mortality than those lobsters less than 51 mm (2 inches) carapace length which are seldom caught because of their possible secretive behavioral patterns and the selectivity of lobster traps. However, we might reach some tentative conclusions from the lengths of lobsters collected in the present study along Table 2. -Length-frequencies by 1-mm increments for lobsters collected in 21 commercial boat trips, September 1971 through Sep- tember 1973 (successive boat-trip numbers are identical to those in Table 1). Carapace Successive boat rips length (mm) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Totals i59 (') 2 — — — 1 (') — — — — — — 9 4 11 — 1 28 60 1 — 5 1 5 1 1 14 61 1 2 1 2 2 1 — 4 1 4 — — 18 62 1 2 1 — 3 1 — 2 10 7 2 1 — 30 63 2 1 1 — — 1 1 1 3 — 2 9 3 6 3 — 33 64 1 1 2 — 1 2 2 1 1 2 3 — 8 2 10 1 — 37 65 1 — 4 — — 3 2 3 1 — 3 2 10 6 9 1 — 45 66 1 2 — 6 — 1 10 2 — 2 1 1 1 10 3 12 3 — 55 67 — 4 — 4 1 — 5 3 3 4 1 1 1 23 6 18 1 — 75 68 1 1 4 — 5 — 1 2 7 5 2 3 4 2 15 10 16 5 — 83 69 6 1 3 — 3 2 1 12 2 3 6 3 2 3 14 9 10 6 — 86 70 1 — 7 — 4 1 — 8 11 13 9 1 4 2 22 7 14 5 3 112 71 2 2 2 1 8 1 2 8 13 8 10 — 2 — 21 8 13 8 8 117 72 5 1 6 1 9 1 2 12 10 7 3 1 1 3 20 11 15 11 9 128 73 2 7 3 7 4 4 21 16 6 6 2 5 8 20 6 14 15 10 156 74 7 2 2 7 18 10 6 9 11 21 10 1 1 1 33 9 18 20 16 202 75 3 7 8 11 20 9 6 14 29 14 14 1 1 6 29 12 22 23 11 240 76 5 11 7 11 14 10 7 14 29 18 15 3 3 5 34 13 10 20 22 251 77 2 13 8 13 11 14 8 23 23 15 13 2 2 — 35 10 14 21 31 258 78 2 15 6 12 23 10 12 18 31 20 15 2 2 7 11 14 15 23 29 267 79 6 13 11 14 16 19 18 13 31 31 20 1 1 1 27 11 14 17 29 293 80 2 17 5 14 19 7 14 19 33 32 15 3 1 3 30 7 10 18 40 289 81 1 12 7 5 8 9 3 10 12 7 11 — — 4 3 2 8 3 15 120 82 2 4 2 3 1 5 4 2 5 3 6 1 — 1 4 1 5 3 8 60 83 1 10 1 8 1 3 5 7 3 2 2 2 — 1 6 2 9 2 5 70 84 — 3 2 3 1 5 5 4 7 2 4 4 1 — 5 2 5 — 1 54 85 5 2 2 2 3 4 — 1 5 4 1 — — — 1 2 2 6 40 86 2 2 — 4 5 2 4 2 4 4 6 3 — 4 2 — 2 2 7 55 87 — 4 — 1 1 5 1 1 2 4 1 1 — — 4 1 4 1 6 37 88 — 3 — 4 2 4 5 5 8 4 4 — — 1 2 2 5 3 5 57 89 — 3 1 2 1 3 3 2 g 1 4 3 — 1 1 1 3 3 2 43 90 1 1 2 2 2 2 2 3 11 3 2 3 — 3 4 — 4 2 5 52 91 1 2 1 3 1 2 4 — 5 5 — — 1 — 1 1 5 2 4 38 92 — 2 — 1 — 1 2 1 5 3 2 — — 2 2 2 5 2 4 34 93 — 2 — 3 2 2 2 — 3 2 1 1 2 — 5 1 3 3 1 33 94 — 2 — — — — — 2 6 2 2 1 — — — — 1 — — 16 95 — 2 — 1 — — — 1 3 2 3 1 — — 1 1 1 — 1 17 96 97 — — — 1 1 — — — 1 1 2 2 — 5 1 2 1 1 z 1 __ _^ 13 7 98 99 z 1 — z z — z — z J. z z z 1 1 — 1 2 1 — *— 4 4 100 — — — — — — — 1 1 1 — — — — — 1 — — 5 101 — — — — — — 1 — — — — — — 2 — — — — — 3 102 — — — — — — — — — — — — — — — — 1 — 2 104 — — — — — — — — — — — — 1 — — — — — 2 105 — — 1 — — — 1 1 — — — — — — — — — — 3 106 — — — — 1 — — — — — — — — — — 1 107 1 — — — — — — — — — — — — — — — — 1 108 — — — — — 1 — — — — — — — — — 1 ^110 — — 1 2 — — — 1 1 — — — — — 1 — — — — 6 Totals 55 144 109 134 204 136 128 245 347 260 201 57 43 72 441 180 327 232 280 3.595 iNo len gths taken. 866 KROUSE and THOMAS: EFFECTS OF TRAP SELECTIVITY ON LOBSTERS with those from the commercial and natural sur- veys (Thomas 1973; Krouse 1973). The commercial survey shows that about 6% of the yearly catch are culls (one or both claws missing). Because most of the legal catch is recently recruited, this may in- dicate that frequent removal from traps of sublegal lobsters is, in part, responsible for this percentage of culls in the commercial catch. This could occur because the claws of those sublegals might grasp laths, knitted heads and parlor en- trances, hands of fishermen, and the like. When such lobsters are pulled from the traps by fisher- men, the claws are occasionally broken off. Another contributing factor might be that sublegal and legal lobsters sometimes extrude their claws through the lath spacings as the trap is hauled aboard the vessel. In this way, claws could be broken off. The design of the proposed "vent- ed" trap, discussed later, takes this situation into consideration. In order to evaluate more fully the possibility of a higher natural mortality due to handling, we used three independent approaches as follows: 1. Our observations aboard vessels show that the percentage of culls of sublegals is between 5 and 109c. This might indicate that natural mortality has not increased due to handling because of the similarity of the percentage of culls in the sublegal and legal size range of lobsters. Autotomy of the lobster could also confound the percentage of culls; however, we theorize that this particular percentage should not be different from the sublegal to legal sizes that we are studying. 2. Another insight on the effect of natural mor- tality would be the length frequencies of the sublegal lobsters caught by research gear in the sampling of the natural population (Figure 3), as well as the length frequencies from sampling aboard commercial vessels (Figure 2), although gear selectivity is a factor in this case. We should expect a higher mor- tality due to handling to show a significant decline in the number of sublegal lobsters as the size range increases by 1-mm increments from 70 (fully vulnerable size) to 81 mm (legal minimum size). Then the number of lobsters at, say, 80 mm should be less than those at 70 mm, not only due to the higher incidence of handling this larger size, but also because of the natural mortality that would occur without handling. These numbers at the specific sizes do not show this decline that could be attributed to a higher mortality due to increased handling (Figure 3). 3. As a supplement to the incidence of handling lobsters and the resultant natural mortality, we feel that our observations on the storage of lobsters in "pounds" (this procedure is described in Thomas 1973) might give infor- mation on the amount of natural mortality in the natural population and that mortality due to handling. The pound owners, stocking at the rate of one to two lobsters per square foot, tell us that a reduction of around 5% in numbers is normal for legal lobsters stocked to those reclaimed 3 to 5 mo later. Under these adverse conditions of crowding and handling in the pound as opposed to the situation in the natural environment, we infer that the annual natural mortality is low in the ocean (5 to 15%) and that handling has a minimal effect. The loss in lobster pounds is sometimes much higher than 5%, but in most of these situations the higher loss can be attributed to disease, adverse environmental conditions, and escapement. Despite these speculative premises concerning the negligible effects of handling on natural mor- tality, the fishermen should still eliminate this needless sorting of large numbers of sublegal lob- sters to reduce: 1) the time spend sorting sublegal from legal lobsters in traps, and 2) the eventual number of culls in the legal catch. Culls not only lessen the total poundage of the commercial catch but possibly the growth rate of culls may be slower than that of noncuUs; Stewart and Squires (1968) suggest that molting of unduly stressed lobsters may be inhibited. The section on selectivity will CARARkCE lENGTM (MM) Figure 3.-Length-frequency distributions of lobsters collected with wire traps (1x1 inch mesh) at Boothbay Harbor (1972-73). 867 FISHERY BULLETIN: VOL. 73, NO. 4 demonstrate the potential benefit of proper lath spacing. Retention Curves The accumulative length frequencies by 1-mrn increments from two selected boat trips reflect a characteristic sigmoid curve (Figure 4). Snedecor lOOr 70 78.0 80 95" CARAPACE LENGTH (MM) AT CAPTURE Figure 4. -Retention curves for two commercial boat trips (refer to boat trips 4 and 5 in Tables 1 and 2) in Boothhay Harbor (1972). Boat trips 4 and 5 had modal vent spacings of V/t and 1% inches, respectively. (1956) and others caution that although these curves are characteristic of normal distributions, other types of distributions could result in a sig- moid curve. For our purposes, we assume that these data represent normal distributions. These curves are interesting in themselves because the 50% accumulative point demonstrates the influence of different lath spacings on the size, composition of the catch. For instance, boat trips 4 (lV4-inch vent) and 5 (1%-inch vent) had 50% ac- cumulative points of 75.6 and 78.0 mm, respec- tively; thus demonstrating that traps with IVi-inch vents are more selective for smaller lobsters (85% of catch ^81 mm) than traps with 1%-inch vents (70% of catch <81 mm); while we are not advocat- ing using these curves in place of selectivity curves, however we do suggest that these "reten- tion curves" could be used as a quick, preliminary approximation of the influence of different lath spacings (gear selectivity) on the size composition of the catch in a trap fishery. Selectivity Curves Based upon the length composition of our catches with modified commercial traps equipped with escape ports (plastic vents) of IV2, lys, and 1% inches and the 1x1 inch wire meshed traps, we calculated selectivity curves in accordance with the methodology of Beverton and Holt (1957). These catch data by 1-mm increments were weighted by THSOD and then the resultant values for each of the three vents (lath spaces) were proportioned with those of the wire traps over the same range of carapace lengths. Traps with the same lath spacings had similar selectivity curves for the 1972 and 1973 catches while conspicuous differences are evident between the various size vents (Figure 5). In both years the lV2-inch vent was selective for the smaller sizes (50% retention ranged from 68.2 to 68.6 mm carapace length), the 1%-inch vent for the inter- mediate sizes (50% retention ranged from 71.4 to 73.5 mm carapace length), and the P/4-inch vent for larger sizes (50% retention ranged from 75.4 to 78.8 mm carapace length). Contrary to most selectivity studies the important consideration in this study is not the mean selection length (50%^ point at which half the lobsters escape and half are re- tained); but rather, the proximity of the curve to the minimum legal size (81 mm carapace length) and whether or not the 100% retention point occurs below or above the minimum legal size. According 868 KROUSE and THOMAS: EFFECTS OF TRAP SELECTIVITY ON LOBSTERS 60 70 80 CARAPACE LENGTH (MM) Figure 5.-Selectivity curves for each of the three vent sizes (1%, 1%, and 1% inches) of the modified commercial traps compared to wire (1x1 inch mesh) traps for 1972 and 1973. to the selectivity curves, all three vents would prohibit the escapement of legal-sized lobsters. However, the l^-inch curve falls closer to the 81- mm line, thus demonstrating that this size vent allows a greater percentage of sublegal lobsters to escape than the 1%- or lV4-inch vents. Effects of Vent Size on Trap Efficiency Catches with modified commercial traps reveal an inverse relationship between vent size and the ratios of sublegal to legal lobsters (Table 3). Overall catches with traps having a 1%-inch vent always consist of more legal than sublegal lobsters Table 3.— Ratios of sublegal to legal lobsters captured with wire (1x1 inch mesh) and modified commercial (IV2-, 1%-, and 1%-inch vents) lobster traps, 1972 through 1973. Actual numbers of sublegal and legal lobsters appear in parentheses. Vent size (inches) Year 1 X 1 (wire) 11/2 15/8 13/4 1972 1973 11.73:1 (962:82) 28.09:1 (927:33) 3.94:1 (71:18) 5.86:1 (164:28) 2.60:1 (78:30) 1.44:1 (133:92) 0.75:1 (21:28) 0.76:1 (104:136) while this size composition is reversed in the catches from traps with smaller vents. This further substantiates our contention that exces- sive handling of short lobsters in the lobster fishery can be minimized with the addition of a 1%-inch vent to all lobster traps. Throughout this study, traps with 1%- and 1%- inch vents not only retained fewer sublegal lob- sters but seemed to capture proportionally more legal-sized lobsters than did those traps with l'/2- inch vents. To assess this situation, we calculated separate catch-effort values (numbers of lobsters per THSOD) for legal-sized and all-sized lobsters combined for each of the three vent sizes (Figure 6). Indeed, our data indicate that traps with larger vents (1% and 1% inches) are more success- ful in retaining greater numbers of legal lobsters than traps with smaller vents. However, because of our limited field sampling, we cannot validly conclude that this disparity in efficiency between vents is conclusive evidence, but rather that our data strongly suggest this possibility. tEGALS & SUBLEGALS 2 3 SET-OVER- DAYS Figure 6. -Comparisons of the number of lobsters (legals, sublegals, and legals combined) per trap-haul-set-over-day for modified commercial traps with vents of l'/2, 1%, and P4 inches (1972-73). Body Proportions of Lobsters The escapement of lobsters of given sizes from traps with varying lath spacings depends upon certain morphological dimensions such as carapace length, width, and height. We contend that the width of the carapace is more important than the height because we observed in the laboratory that 869 FISHERY BULLETIN; VOL. 73, NO. 4 lobsters, attempting to escape through different lath spacings, would twist on their sides when en- countering a tight fit between laths. This, coupled with the fact that the width is always smaller than the height for any carapace length, led us to the opinion that the relationship between carapace length and width in association with lath spacing is the important consideration for gear selectivity studies. We calculated the regression of carapace length (X) on width (Y) for 217 lobsters (114 females and 103 males) by the method of least squares. The calculated equation was Y = -4.367 + 0.649 A' (Figure 7). Data for sexes were com- bined because analysis of covariance on the regression coefficients (Steel and Torrie 1960) showed that carapace length-width ratios of males and females did not differ significantly. According to this relationship, lobsters at the minimum legal size of 81 mm carapace length would be expected to have a mean carapace width of 48.2 mm + 0.18 with individual widths within the 95% prediction interval ranging from 45.6 to 50.9 mm. The magnitude of these measurements relative to a rV4-inch (44.5-mm) lath spacing sug- gests that only a very small percentage of legal- sized lobsters might escape through that size vent. We should mention that some compression of the shell, particularly if the lobster is newly molt- ed, is possible as a lobster struggles to get through the lath spacing of a trap. However, based upon our laboratory observations, we would not expect this compression to exceed 2 to 3 mm for soft- shelled lobsters (1 to 2 wk since ecdysis) and 1 mm for a hard-shelled lobster. Because this soft- shelled condition is of rather brief duration and 40 1-75" ! " M.n,m«m lagal i,i. 70 EfO 85 CARAPACE LENGTH (MM) 90 the frequency of trap hauls is greatest during the shedding season (a shorter period of time for es- capement), these situations should minimize es- capement. Trap Escapement Study Retention curves, based on the escapement of lobsters of known lengths from traps with lath spacings ranging from IV4 to 2 inches, graphically display the pronounced effect escape vent size has on lobster escapement (Figure 8). Retention of sublegal lobsters was high for the 1*4- and 1 '/2-inch traps while most sublegals were able to escape from traps with the 1%- and 2-inch vents. With the present minimum size of 81 mm (3-^/i6 inches), a 2-inch vent would be unsatisfactory as many legal lobsters could escape, whereas escapement of legals through a 1%-inch vent would be extremely minimal. Although the curve for the r'A-inch vent did show some escapement, we believe this es- capement is exaggerated by the methodology (plotting midpoints) employed in the derivation of this curve. This contention is further substantiat- ed by the fact that only one of seven lobsters with a carapace length of 82 mm escaped and there was no escapement for lobsters larger than 82 mm. 60 70 80 90 CARAPACE LENGTH (mm) Figure 7.-Carapace length-width relationship for lobsters with 95% confidence and prediction intervals. Figure 8. -Retention curves for lobsters placed in modified com- mercial traps with VA-, VA-, and 1%-, and 2-inch vent dimensions. 870 KROUSE and THOMAS: EFFECTS OF TRAP SELECTIVITY ON LOBSTERS RECOMMENDATIONS Based on the foregoing analysis of the effects different lath spacings have on the size composi- tion of lobster catches, we recommend that all lobster traps fished along the Maine coast have an escape vent of 1% inches. Of course, if the minimum legal size (81 mm) is increased then the vent size should be altered accordingly. We emphasize that it is not necessary for the entire trap to consist of the desired lath spacing; but rather, only one lath spacing either on the side or end (preferably near the bottom) of the parlor section of the trap. The remaining laths could be spaced at the fisherman's discretion. We believe an escape port (vent) fabricated from some type of durable material and manufac- tured to our specifications could be incorporated into any conventional lobster trap (Figure 1). Merits of this vent would be: 1) easy installation in both new and old traps without requiring drastic modification; 2) modest cost to the fishermen; and 3) retention of its original dimensions over time (unlike wooden laths which eventually wear, caus- ing a larger opening, thus permitting escapement of legal lobsters). If this recommendation of venting traps is adopted as a conservation measure, we would ex- pect reductions in: 1) the number of culls (which in turn would increase the weight of the total land- ings) and, if of consequence, the natural mortality; 2) time expended by lobstermen in sorting their catches; 3) perhaps the illicit trade of sublegal lobsters (shorts) which is considered by some dealers and fishermen to be of an alarming mag- nitude; and 4) if a real problem, the number of lobsters imprisoned in lost traps. ACKNOWLEDGMENTS We thank the fishermen who allowed us, often at inconvenience to themselves, to accompany them on their fishing trips. Cecil Pierce, a fisherman from Southport, Maine, designed the plastic vent. Gary Robinson, Louis Kazimer, Curt Crosby, Andrew Dolloff, and William Sheldon of the Maine Extension Service made many of these trips. Their comments and experience aided greatly in collect- ing meaningful data. We express gratitude to Gareth Coffin, formerly with the Northeast Fisheries Center, National Marine Fisheries Ser- vice, West Boothbay Harbor, Maine, for his pho- tographic work. LITERATURE CITED Beverton, R. 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. GULLAND, J. A. 1969. Manual of methods for fish stock assessment. Part 1. Fish population analysis. FAO (Food Agric. Organ. U.N.)Man.Fish.Sci.4, 154p. Krouse, J.S. 1973. Maturity, sex ratio, and size composition of the na- tural population of American lobster, Homarus americanus, along the Maine coast. Fish. Bull., U.S. 71:165-173. Pope, J. A. 1966. Manual of methods for fish stock assessment. Part III. Selectivity of fishing gear. FAO (Food Agric. Organ. U.N.) Fish. Tech. Pap. 41, 41 p. Richer, W.E. 1958. Handbook of computations for biological statistics of fish populations. Fish. Res. Board Can., Bull. 119, 300 p. Snedecor, G. W. 1956. Statistical methods applied to experiments in agricul- ture and biology. 5th ed. Iowa State Coll. Press, Ames, 534 p. Steel, R. G. D., and J. H. Torrie. 1960. Principles and procedures of statistics with special reference to the biological sciences. McGraw-Hill, N.Y., 481 p. Stewart, J. E., and H. J. Squires. 1968. Adverse conditions as inhibitors of ecdysis in the lob- ster Homarus americanus. J. Fish. Res. Board Can. 25:1763-1774. Templeman, W. 1939. Investigations into the life history of the lobster {Homarus americanus) on the west coast of New- foundland, 1938. Newfoundland Dep. Nat. Resour., Res. Bull. (Fish.) 7, 52 p. Thomas, J. C. 1973. An analysis of the commercial lobster {Homarus americanus) fishery along the coast of Maine, August 1966 through December 1970. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-667, 57 p. Wilder, D.G. 1945. Wider lath spaces protect short lobsters. Fish. Res. Board Can., Atl. Biol. Stn. Circ. G-4, 1 p. 1948. The protection of short lobsters in market lobster area. Fish. Res. Board Can., Atl. Biol. Stn. Circ. G-11, 1 p. 1954. The lobster fishery of the southern Gulf of St. Lawrence. Fish. Res. Board Can., Gen. Ser. Circ. 24, 16 p. 871 EXPERIMENTAL EXPLOITATION OF COMPETING FISH POPULATIONS Ralph P. Silliman' ABSTRACT Populations of the guppy, Poecilia reticulata, and the swordtail, Xiphophorus maculatus x X. helleri, were grown both independently and in competition under controlled conditions. Independent popula- tions were permitted to grow for about a year and then successively exploited at two different rates for each species. In the control (unfished) pair of competing populations, both species grew for about 30 weeks, followed by decline and extinction of the swordtail and fluctuations in the guppy. Similar initial growth in the test pair was followed by exploitation of both species at various combinations of rates. Measures of recruitment were available as weights of juveniles returned to adult tanks from separate nursery tanks. Data from fitted curves showed that guppy recruitment exceeded that of the swordtail under both independent and competing conditions. Depression of recruitment by competition was greater in the swordtail than in the guppy. A mathematical model for competing populations consisted of a pair of differential equations including elements of the Volterra competition formulae and the Fox exponential surplus-yield model. By using the exploitation rates applied in the experiments, and constants from the independent populations, the model was applied to biomass data from the control pair of competing populations. Successive trials resulted in a reasonably good fit, and competition coefl^cients from this were used to fit data from the exploited test pair. Yield isopleths calculated from the fitted model showed that maximal yields were obtained when exploitation for the swordtail was lower than for the guppy, suggesting lower productivity in the swordtail. The maximum sustainable yield represented about 20% of food placed in tanks, and indicated at least as great efficiency from competing populations as from independent ones. Results from the experiments clearly suggest that exploiting both members of a competing pair is preferable to exploiting either alone, provided fishing rates are adjusted in relation to the productivity of each species. Classical studies of fishery dynamics, such as those discussed in the works of Beverton and Holt (1957) and Ricker (1958), deal mostly with single popula- tions treated as if they existed independently. Fishery biologists have come to recognize, however, that in many situations the fish stock cannot be so treated (Larkin 1963; Murphy 1973). The exploited population of interest is inter- dependent with others (which may be either exploited or unexploited) through competitive or predator-prey relations. Any effect of exploitation on one stock may produce a reaction in another, resulting in readjustments in both populations, and invalidating the expected response to exploi- tation based on single-species dynamics. A familiar example of an apparent competitive situation is contained in the population histories of the Pacific sardine, Sardinops sagax, and the northern anchovy, Engraulis mordax, off the coast of California. The sardine suffered a cata- 'Northwest Fisheries Center, National Marine Fisheries Ser- vice, NOAA, Seattle, Wash.; present address: 4135 Baker NW, SeatUe, WA 98107. strophic decline in the mid-1940's, followed by an increase in the anchovy. An analog computer model of Silliman (1969a, b) demonstrated that at least part of the change in the anchovy population size could be simulated with data on the sardine population size and the differential equations of Volterra (1928). Murphy (1973) provided recent verification of the sardine-anchovy relation and suggested that similar relations may prevail in the Japanese and South African sardines. Laboratory experiments on the exploitation of self-sustaining fish populations have been report- ed fairly extensively (Silliman and Outsell 1958; Silliman 1968; Nagoshi et al. 1972). Experiments with competing populations have included such diverse organisms as yeast cells (Cause 1932); Protozoa (Cause 1934); Daphnia (Frank 1957); beetles (Park 1962); and warblers (MacArthur 1958). To the best of my knowledge, however, exploitation of competing laboratory fish popula- tions has not previously been reported. The purpose of the experiments reported below was to ascertain experimentally the reaction of two competing fish populations to exploitation. Manuscript accepted January 1975. FISHERY BULLETIN: VOL. 73, NO. 4, 1975. 872 SILLIMAN: EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS Within this general objective, it was desired to determine which combination of exploitation rates applied on the two species would provide the maximum sustainable yield. To approach this problem, test and control pairs of populations were established in aquariums and allowed to grow several months under controlled conditions. Various combinations of exploitation rates were then applied to the test pair to determine popula- tion interactions and total yields. A "base line" for evaluating the results was obtained by growing and exploiting each of the competing species independently. APPARATUS AND PROCEDURES Experimental Animals A lengthy fund of experience (Silliman 1948, 1968; Silliman and Outsell 1958) built up with the guppy, Poecilia reticulata, dictated this as one of the experimental fishes. For the other, it was desired to have a species that was similar to the guppy in size, reproduction, and feeding habit but readily distinguishable from it in a mixed popula- tion. A fish that met these requirements fairly well was the red swordtail hybrid, Xiphophorus maculatus x X. helleri, which will be referred to simply as "swordtail." It is somewhat larger than the guppy, but lived in the same sized aquarium. It is a live-bearer, like the guppy, and will readily eat the foods commonly fed guppies. Its brilliant red color permits easy distinction in the adults, and even the newborn young are pink or orange and may be distinguished from newborn guppies. In both species, adult males can be distinguished from adult females by external inspection. Dis- tinguishing male characters are the modified anal fin (gonopodium) and fin and body color in the guppy and the elongated lower caudal fin ("sword") in the swordtail. Aquarium Tanks Fish were grown in four conventional glass- walled aquariums, each with a water surface of 44 cm by 24 cm, a water depth of 19 cm, and a volume of 20 liters. Inside each tank was an air-stone and a fiber-charcoal filter. Tanks were placed in a row with their longer axes parallel, and lettered A, B, C, D from left to right. Tanks A and C were for juvenile and adult fish together. They each had a refuge in the left front corner for the escape and subsequent removal of recently born fish, or "fry." This refuge was formed with a fence consisting of 21 cm by 3 mm glass rods placed 1.5 mm apart, enclosing a right isosceles triangular space of 15 cm hypotenuse. Although guppies could survive and achieve population growth in the above-described tanks, preliminary experiments showed this not to be true for the swordtails. Tanks B and D were therefore provided as "nurseries" for the tem- porary relocation of newborn young from tanks A and C, respectively. The juveniles were placed back in tanks A and C when they had grown to such size that they would no longer pass through a sieve consisting of 3-mm plastic rods placed 2 mm apart, thus making them recruits to the fishable stock. Food and Feeding A diet previously developed for guppies (Silliman 1968) was fed to all fish (Table 1). The food Artemia nauplii, however, requires special mention. The original intention was to feed the fish in the nursery tanks one-half the amount fed those in the adult tanks. The weight of nauplii produced was mistakenly believed to be directly proportional to amount of Artemia eggs placed in the culture beakers and, therefore, one-half the amount of eggs placed in the beakers for the adult tanks was placed in those for the nursery tanks. Production tests (Table 2) based on duplicate hatchings produced under standard conditions, however, showed production not to be proportional to the amount of eggs. Amounts of eggs inserted were kept the same, nevertheless, on the chance that unhatched eggs were eaten (observed on one occasion). Artemia nauplii provided so small a proportion (1/100%) of the total diet that the lack of propor- tion noted above would have no significant effect on total food intake. The small amount of living food provided by the nauplii was regarded in the same sense as vitamins in human nutrition: as something required in small amounts for good health, but not furnishing a significant proportion of total food intake by weight. It is pertinent to note that the smallest number of nauplii (1,500) indicated by any of the tests (Table 2) would provide over four nauplii per fish for the largest 873 FISHERY BULLETIN: VOL. 73, NO. 4 Table l.-Food placed in tanks, grams. 3-wk Day of Adult tanks Nursery tanks Dates Frozen Artemia Frozen Artemia included periods week Dryi Artemia nauplil Total Dry' Artemia nauplil Total 25 Oct. 1965 0-71 Sun. 0.1 0.1 0.05 0.05 to Mon. 0.1 1.0 {') 0.05 0.5 (2) 0.55 6 Dec. 1969 Tues. 0.1 1.0 1?) 0.05 0.5 (2) 0.55 Wed. 0.1 1.0 (^) 0.05 0.5 (^) 0.55 Thurs. 0.1 1.0 (') 0.05 0.5 (2) 0.55 Fri. 0.1 1.0 (') 0.05 0.5 (^) 0.55 Sat. Total 0.1 — (^) 0.1 0.05 — P) 0.05 0.7 5.0 — 5.7 0.35 2.5 — ■ 2.85 7 Dec. 1969 71-124 Sun. 0.1 0.1 0.05 — 0.05 to f^on. 0.1 1.0 {}) 0.05 0.5 F) 0.55 1 Jan. 1973 Tues. 0.1 1.0 (') 0.05 0.5 (2) 0.55 Wed. 0.1 1.0 (2) 0.05 0.5 P) 0.55 Ttiurs. 0.1 1.0 (2) 0.05 0.5 P) 0.55 Fri. A.M. 0.1 1.0 (^) 0.05 0.5 (') 0.55 Fri. P.M.3 Total 0.1 — P) 0.1 0.05 — P) 0.05 0.7 5.0 — 5.7 0.35 2.5 2.85 'Tropical fish food. 2Hatch from 0.4 g (adult) or 0.2 g (nursery) of eggs (Table 2). Silliman and Outsell (1958) found tfiat the hatch from 0.4 g of eggs weighed 0.125 mg. Test hatches of nauplii were not proportional in weight to the amount of eggs used (Table 2), and since the total weight would be only 1/100% of the diet, no weight is indicated in the table. 3This was combined with the Friday A.M. feeding in 35 out of 161 wk and with the Sunday feeding once. Table 2.-Artemia production tests. All 48-h hatches at 24°C in 800 ml 3% salt water. Counts from 20 samples for each test. Samples were 0.3 ml, withdrawn by pipette from vigorously stirred cultures killed in 0.75% formaldehyde, and replaced. Mean no. Est. ■ 1,000s Source in samoles in CL ilture^ Test dates of eggs' eggs (g) Nauplil Eggs3 Nauplii Eggs3 1970: 1/28-30 A 0.2 2.10 7.50 5.6 20.0 1/28-30 A 0.4 1.45 18,05 3.9 48.1 3/23-25 A 0.2 1.85 6.95 4.9 18.5 3/23-25 A 0.4 2.05 16.30 5.5 43.5 4/13-15 B 0.2 1.75 10.35 4.7 27.6 1971: 5/ 3- 5 B 0.4 0.55 32.95 1.5 87.9 5/17-19 B 0.2 0.85 23.30 2.3 62.1 t/ 24-26 C 0.2 3.50 30.70 9.3 81.9 10/ 4- 6 C 0.2 3.65 33.40 9.7 89.1 11/29-12/1 C 0.2 2.75 34.30 7.3 91.5 1972: 5/15-17 C 0.2 6.85 27.80 18.3 74.1 11/13-15 C 0.2 2.75 36.10 7.3 96.3 lAII were commercial suppliers. 'Sample numbers times 800/0.3. ^Includes shells (from hatched eggs) and unhatched eggs. number of fish recorded in any tank (343 guppies in tank C during the last week of 3-wk period 65). Dry food was placed on the surface of the water and sank slowly if not eaten immediately (as oc- curred with large populations). Frozen food sank and was eaten as it thawed. Artemia nauplii were hatched in 800-ml glass beakers (Table 2). The en- tire water mass, including shells and unhatched eggs, was poured through a cloth filter which was rinsed with freshwater and then rinsed into the fish tanks. Cleaning and Treatment Detritus including uneaten food (none in large populations) was siphoned daily from the tanks onto a cloth filter and the siphoned water returned to the tanks. Once a week all the water was removed from the tanks and one-half the volume was replaced with tap water aged for 1 wk. At this time the tanks and their equipment were thoroughly cleaned, and the filter fiber and char- coal were replaced. Also, fish in the adult tanks were treated for 15 min in a 1:200 solution of a commercial aquarium disinfectant "Fungistop."- Water Characteristics Water temperature in tanks A and D (Tables 3, 4; Figure 1) was recorded daily (Saturday excluded during 3-wk periods 71-124). These end tanks were chosen to reveal any temperature gradient that might exist. Although there was a slight tendency for tank D to vary from tank A (Figure 1), the differences were mostly less than 1°C and are not believed to have significantly affected population growth. It will be shown in the section on oscilla- tory fluctuation that deviations of population size from the theoretical were not correlated with tank temperature. -Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 874 SILLIMAN: EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS Table 3. -Mean temperatures, °C, 3-wk periods, competing populations. Daily readings (21) to period 70, Sunday to Friday (18) from period' 72 on. Competing populations Tank sriod A D 0 224.2 224.6 1 24.5 24.6 2 24.9 24.7 3 24.6 24.5 4 24.7 24.5 5 24.6 24.6 6 24.5 24.6 7 24.7 24.8 8 24.5 24.8 9 24.4 24.6 10 24.4 24.7 11 24.6 24.8 12 25.6 25.6 13 25.9 25.9 14 26.4 26.0 15 25.6 25.7 16 25.5 25.4 17 25.0 25.1 18 25.0 24.8 19 25.6 25.1 20 26.4 25.1 21 26.2 24.9 22 26.0 25.5 23 25.9 24.8 24 25.2 24.5 25 25.6 24.6 26 26.0 25.0 27 25.7 25.2 28 26.4 25.8 29 26.2 25.9 30 26.3 25.8 31 25.9 26.0 32 24.5 25.1 33 23.8 24.5 34 23.1 23.8 35 23.7 23.6 36 23.4 23.8 37 23.7 23.8 Tank Period 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 23.6 23.8 23.9 23.5 23.3 24.1 24.5 24.1 24.9 24.9 24.3 24.8 24.1 24.1 23.7 23.7 23.5 23.4 23.7 23.9 23.6 23.5 23.2 24.3 24.5 24.9 24.7 23.8 23.9 24.4 21.5 24.0 24.0 '23.9 23.7 23.8 23.9 23.9 23.9 24.2 24.4 23.8 23.5 24.4 24.7 24.7 25.2 25.1 23.5 24.9 24.3 24.4 23.9 23.8 23.6 23.9 23.9 23.8 23.6 23.7 23.5 24.5 24.5 24.9 24.9 23.8 23.8 24.3 21.4 24.0 23.8 '23.8 23.4 23.5 23.5 23.5 'One Friday and two Saturday readings missing, period 71. ^Based on 1 wk only. Table 4. -Mean temperatures, °C, 3-wk periods, independent populations. Sunday to Friday readings (18). Tank Period Tank Period A D A D 79 24.6 24.1 102 24.5 24.3 80 23.5 23.2 103 24.7 24.2 81 25.0 24.5 104 24.5 23.6 82 24.7 24.3 105 24.4 23.6 83 24.4 24.2 106 24.3 23.5 84 24.5 24.5 107 24.1 23.8 85 24.3 24.3 108 24.2 23.8 86 23.9 23.7 109 24.3 23.9 87 24.2 23.8 110 24.4 23.9 88 24.2 23.6 111 24.3 23.8 89 24.6 23.9 112 24.5 23.9 90 24.6 23.9 113 24.6 23.9 91 24.6 24.2 114 24.6 23.9 92 24.5 23.8 115 24.1 23.5 93 24.5 23.9 116 25.8 25.2 94 24.6 23.7 117 25.6 24.9 95 24.2 24.0 118 25.4 24.7 96 24.3 23.9 119 24.9 24.1 97 24.2 23.6 120 24.5 23.6 98 24.4 24.2 121 24.7 23.6 99 25.6 25.4 122 24.1 22.9 100 25.7 25.4 123 24.5 23.3 101 24.7 24.5 124 '24.7 '24.2 BIdg. heat failure "1 I I I I 1 1 1 1 1 1 T r 10 20 30 40 50 60 70 3-week periods Aquarium: 27- independent popula ions 26- K 25- ^. A. A\ 24- ] fv j; \..,. ■■■■'"'■■''"'■■ f \ ^ 23- :/ 22- s s s 21- 80 90 100 no 120 3- week periods 'Last three readings missing. Figure 1. -Water temperatures (3-wk means). "S" indicates summer periods (approximately 20 June to 20 September). Heaters were placed in tanks during periods 1-32 (Figure 1) but caused excessive temperature fluctuation and one instance of mortality from a nonfunctioning thermostat. During periods 33-174 the tank water was at room temperature. This was thermostatically controlled except that no cooling was available in the summer. Summer tempera- tures were thus somewhat higher than during the balance of the year (Figure 1), but the change was the same for all tanks. Measurements of dissolved oxygen and carbon dioxide concentrations and pH were made at irregular intervals during the course of the experiment (Table 5). All Og readings were within or above the 3-5 ppm range considered satisfactory for warmwater fishes by Lewis (1963), and the COg and pH readings were within the range he con- sidered safe (CO2<30 ppm., pH 5.0-9.0). Light was provided from overhead fluorescent fixtures with standard tubes to period 11. After period 11, pink tinted tubes were used. Handling, Enumeration, Exploitation, and Weighing Areas behind the refuge fences described under "Aquarium Tanks" were inspected daily. If any newborn fish were found there or in other corners of the tanks, they were removed by netting or siphoning, counted, and placed in the nursery 875 Table 5.- Water condition on selected dates. 3-wk 0,, ppm. CO,, ppm. PH Tank Tank Tank Tank Tank Tank Dates period A D A D A D 1968: Aug. 9 16 48 48 7.0 6.0 7.0 ^^ 29 49 6.6 6.4 10 10 6.9 6.9 Sept 6 49 6.2 6.8 — — — — 13 49 6.2 6.4 — — — — 20 50 6.4 6.4 10 10 — — 27 50 6.4 6.4 — — — — Oct. 4 50 6.4 6.4 — — — — 11 51 6.0 6.4 — — — — 18 51 6.4 6.8 — — — — 25 51 6.2 6.4 — — — — Nov. 8 52 6.0 6.6 — — — — 15 52 6.4 7.2 — — — — 28 53 6.4 7.0 — — — — Dec. 6 53 6.4 6.8 — — — — 13 54 6.4 6.6 — — — — 26 54 6.4 6.8 — — — — 1969: Jan. 2 55 6.4 7.0 — — — — Sept. 17 67 6.6 6.6 10 10 8.0 8.0 Oct. 2 68 7.8 7.8 — — — — 9 68 7.4 8.2 — — — — 23 69 6.6 7.6 — — — — 30 69 6.2 7.6 — — — — Dec. 4 71 6.8 7.4 10 10 8.0 8.7 1971: Feb. 25-26 92 6.6 8.2 10 10 8.0 8.0 1972: Mar. 23 111 5.0 7.2 10 10 8.5 8.0 tanks. At the time of weekly cleaning, the water in the nursery tanks was poured through the sieve also described under Aquarium Tanks. Any fish remaining on the sieve were placed back in the adult tanks. Fish were counted and weighed weekly during periods 0-71. During periods 71-124, this was done only at the approximate brood intervals of the fish, which were 3 wk for the guppy and 4 wk for the swordtail. Fish were counted simply by netting them from one container to another. Counts were categorized into "immature" (those whose sex could not be determined from external inspection), male, and female. Fry and juveniles were counted when moved between adult and nursery tanks. Dead fish found in tanks were recorded as mor- talities. Exploitation was done at the time of counting. To apply an exploitation rate of 1//;, each »th fish was removed (n was always an integer). This was applied equally to juveniles and adults, but not at all to fish in the nursery tanks. Population and catch weights were also deter- mined at the time of counting. Fish were drained and placed in a previously weighed container of water. Total weight was measured and fish weight obtained by subtracting the tare. FISHERY BULLETIN: VOL. 73, NO. 4 COURSE OF POPULATIONS Independent Populations Although chronologically the competing populations preceded the independent popula- tions, the more logical order of presentation is to deal with the independent populations first. This arrangement will be followed in the remainder of the report. Separate populations of guppies and swordtails were started on 19 May 1970 (3-wk period No. 79). Each of these was permitted to grow for an initial period of about 1 yr (Figures 2, 3) before exploita- tion was begun. Even though complete equilibrium had not been reached, exploitation was started at 25% per brood interval (3 wk) for the guppy and 10% per interval (4 wk) for the swordtail. The lower rate for the swordtail was based on previous experience showing lower productivity for that species. Initial rates were maintained during weeks 289-334 for the guppy and 290-334 for the swordtail. Final rates were 33.3% for the guppy (weeks 337-373) and 16.7% for the swordtail (weeks 338-374). Responses were in accord with expectations for the guppy (Figure 2) and the early swordtail history, but there was an increase in both number and weight in the sword- tail in the last five brood intervals (Figure 3). This anomaly will be discussed under "Oscillatory Fluctuation." 320 ^80- 240- 200 X tn "- I SO- LI. O a, '20- OD 2 80- Z) -z 40- i5 30- _ 25- E ° 20 i 15- g lOH 5 0- WEEK 3^«eek period 80 240 110 Figure 2.-Course of independent guppy population. Numbers indicate target exploitation rates. 876 SILLIMAN: EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS 280- 240 1200 U- U.160 o £ 120 CD I 80 z 40 35 30- _25- E |20H El5- i£ Ui glO- 5 240 WEEK 3-weeh period 80 Figure 3. -Course of independent swordtail population. Numbers indicate target exploitation rates. Competing Populations The two mixed populations, each composed of both guppies and swordtails, were started 24 Oc- tober 1965 (week 0). Three-week period 0 started with week 1, since weights were not recorded in week 0. In the control (unexploited) pair of populations (Figure 4), both species grew for an initial period of about 30 wk. The swordtail population then began to decline and disappeared at week 129. This will be discussed in the section on competitive exclusion. Extinction of the swordtail was followed by a large oscillatory fluctuation in the guppy (about -39"^^ to +269f of the asymptotic level), which will be discussed in the section on such fluctuations. Initial growth in the test (exploited) pair was similar to that in the control pair (Figures 4, 5). Exploitation was started first on the swordtail 360- 320 280- ^ 240- ^ 20O Ll. o cr 160H UJ m 5 120- Z. 80- 40- 0- WEEK 0 3-week period 6 —1 — 10 — 1 — 20 30 40 — I — 50 I 60 70 Figure 4.— Course of competing populations, control pair. Solid line, guppy; broken line, swordtail. 877 FISHERY BULLETIN: VOL. 73. NO. 4 Week 3-week period 0 Figure 5.— Course of competing populations, test pair. Solid line, guppy; broken line, swordtail. Numbers indicate target exploitation rates. (week 30), under the mistaken impression that the greater biomass then achieved by the swordtail indicated a greater productive capability. Exploi- tation produced a rapid decline in the swordtail to a low population level (Figure 5). Cessation of swordtail exploitation at week 59 and initiation of guppy exploitation at week 62 did not lead to recovery of the swordtail, in spite of a drastic decline in guppy abundance (Figure 5). By week 74, it became apparent that the sword- tail would require a lengthy period for recovery, even if guppy abundance were further reduced. To accelerate the study of exploitation, the popula- tions v/ere reconstructed during weeks 74-85, us- ing fish from exploited populations that had been placed in a reserve tank. After reconstruction, the populations approximated fairly closely their number and weight at the time exploitation was started (compare week 85 with week 30 in Figure 5). Exploitation rates after week 85 were adjusted to keep both the guppy and swordtail at productive levels while trying as wide a range of pairs of exploitation rates as possible. RECRUITMENT RELATIONS Juvenile fish were counted both when removed from and returned to the adult tanks; it was thus possible to obtain a measure of recruitment. Numbers were converted to weights by use of factors (mean w^eights per fish) based on weighings of the juveniles: guppy, 0.0656 g based on 1,417 fish in 126 weighings spread over 199 wk; swordtail 0.0678 g, 337, 61, 196, respectively. At the beginning of the experiments, when few fish were in the nursery tanks, it was possible to distinguish individual groups of recruits by size, count them, and thereby estimate the "reproductive lag" from birth to recruitment. The lag was found to be approximately one brood interval (guppy, 3 wk; swordtail, 4 wk) for each species. In constructing the stock-recruitment relations, the recruitment for each brood interval was compared with the mean stock in the adult tanks during the preced- ing brood interval. During periods of exploitation, the catch was subtracted from the stock at the beginning of the interval. 878 SILLIMAN; EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS Because of the great variability in the recruit- ment data, group means were used for both species. The basis of the grouping was an interval of 5 g (0.0-4.9, 5.0-9.9, etc.) in the adult tank stock weight. Data used in calculating recruitment cur\'es were pairs consisting of the mean adult stock weight and mean recruit weight for each group. The stock-recruitment data for the guppy (Figure 6) could be fitted by a Ricker (1958) curve of the type: where Rj^+i is recruitmen|: during brood interval N + 1 and 5^ is mean adult stock during brood interval N, both in grams. Fitting of the curves shown (Figure 6) was by least squares to the rec- tilinear logarithmic form of the relation: log,^jv+i/ S^ = \og,a - hS^. Values of the constants are given in Table 6. Data for the swordtail (Figure 7) did not con- form well to the Ricker relation, as shown by the parabolic nature of the points for the competing stock, and were fitted better by a simple parabola: symbols as above. Curves shown (Figure 7) were fitted directly to the grouped data by least squares. Values of the constants are in Table 6. Comparison of results for the two species (Figures 6, 7; Table 6) reveals both similarities and differences. Despite the different types of Independent Table 6.-Stock-recruitment data comparing populations of the two species_with data from the relations: guppy, /?^ ^ j = aSj^fCxp (-65^); swordtail, R^ = aSj^j - bS Recruitment at value Nn nf Constants, of S^ for R^^,, inde- pairs of per brood pendent situation (g) Species and obser- vations' interval Per brood Per situation a b interval^ week Guppy: Independent 76 0.343 0.061 2.07 0.69 Competing 117 0.092 0.028 0.95 0.32 Swordtail: Independent 32 0.229 0.007 1.90 0.48 Competing 86 0.068 0.003 0.42 0.10 Total 311 'In fitting curves, data were grouped by 5-g intervals of S,^. 2Guppy, 3 \Nk; swordtail, 4 wk. recruitment curves, there is the suggestion in both species that maximum recruitment occurs at in- termediate rather than very high or very low adult stock levels, as has been observed for other species (Ricker 1958; Silliman 1969b). Also, recruitment for both species was depressed by competition. However, the depression was greater for the swordtail, as shown in the comparison of standard- ized recruitment (last column. Table 6). This is in keeping with the finding to be reported below that the guppy is more productive than the swordtail. This finding is also supported by the fact that guppy recruitment was greater than swordtail recruitment in both independent and competing situations (last column. Table 6). Some additional information on recruitment was obtained from a study of count discrepancies. Because all additions to and removals from the adult tanks were recorded, it was possible to cal- culate an "expected" count (the previous count Swordtail A, a C Independent Figure 6. -Stock-recruitment relation for the guppy. Data grouped by 5-g intervals of S^ . N is the number of the 3-wk brood interval. Figure 7.-Stock-recruitment relation for the swordtail. Data grouped by 5-g intervals of Sfj . N is the number of the 4-wk brood interval. 879 plus recruits and minus catch and deaths) to be compared with the actual counts made. Some of the observed discrepancies were no doubt due to unobserved deaths (dead fish eaten by others before seen) or to errors in counting. That some errors occurred is not surprising. Each expected- actual comparison involved as many as 17 separate counts. During each main count while exploiting the stocks the counter had to keep in mind the total number, the number caught, the state of maturity of each fish, and the sex of each mature fish. The distribution of discrepancies for selected (to provide representative data) periods (Table 7) Table 7.-Count discrepancies for selected periods: swordtail, April 1970 to June 1972; guppy, March 1970 to June 1972. Values represent "expected" number subtracted from the actual count. Guppy 1 1 1 2 1 1 1 2 1 3 3 2 3 1 1 1 3 1 2 31 Discrepancy Swordtail -20 — -12 — -11 — -10 3 -9 — -8 2 -7 2 -6 1 -5 3 -4 3 -3 3 -2 1 -1 2 0 1 + 1 1 +2 — +4 — +5 — + 6 — +7 — + 8 — +9 1 Total 23 shows that negative discrepancies (actual less than expected) exceed the positive for both swordtails and guppies. This no doubt arose from the unrecorded natural mortalities mentioned above. The two positive discrepancies for the swordtail probably represent counting errors. For the guppy, however, the fairly large proportion of positives exceeding three fish suggests that unrecorded recruitment occurred. Apparently some of the guppy "fry" escaped detection, even though a thorough search of the tanks was made. This phenomenon is in keeping with the observed greater hardiness of the guppy, and suggests that the superiority in recruitment for the guppy was even greater than indicated in the stock-recruit- ment relations reported above. FISHERY BULLETIN: VOL. 73. NO. 4 SIMULATION MODEL Mathematical Derivation Data of population weight reflect growth of in- dividual fish as well as recruitment and mortality, and all of the analyses below will be in terms of weight. Development of the formulae requires a fairly extensive list of symbols, which are defined below. P = Total population weight in grams. t — Time from start, in 3-wk periods. X = Fishing effort in arbitrary units. q = Catchability coefficient. F = Instantaneous rate of fishing mortality { = qX). m = Three-week rate of fishing mortality. G = Constant of the Gompertz growth curve. k = Constant of the Gompertz growth curve and of the Fox (1970) population model. . J — Empirical constants. cji Adding a term for the effect of fishing to the formulae of Volterra (1928) gives a pair of differential equations: rfPj/f/f =,/'(Pi)-/(P2)-./'(^i), dP^ldt = f{P,) -f{P^) -f{X^). (1) (2) In these equations, the first term of the right hand side is for population growth; the second, for competition; and the third, for the effect of fishing. The development is exactly parallel for the two equations, and only that for Equation (1) will be outlined below. For the growth term Volterra (1928) used/(Pj ) = JiPj-Z^jPf. This is the logistic growth curve, which requires symmetrical population growth. Growth for the guppy under fishing (equilibrium yield) was shown by Silliman and Gutsell (1958) and Silliman (1968) to be distinctly asymmetrical. The Gompertz (1825) curve, introduced as a population yield model by Fox (1970), is suitable for asymmetrical growth and will be shown in the section on determination of constants to be suit- able for initial population growth in both the guppy and the swordtail. This is expressed: P, = Poexp[G-Gexp(-A-OJ. (3) It can be shown by mathematical analysis of 880 SILLIMAN; EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS Equation (3) that the Hmit Pqo = -fo^xp (G), and by substituting this in Equation (3), differentiating and taking logs a growth term may be derived for Equation (1): /(Pi) = AA-,(log,Pi^-log/'i). (4) For the competition term, Volterra (1928) used ./'(P2 ) = f'l •^i-f*2 • Preliminary experimentation showed that this term was unsatisfactory for the guppy-swordtail experiments, since it was impos- sible to obtain even a reasonably good fit using it. I also experimented with /(f 2) = ^1 (-^1 + -^2) ^^ the theory that the sum of the populations, rather than their product, might be controlling, but it was equally unsatisfactory. The most suitable term proved to be simply: /(P2) = C,P2 (5) 40- Guppy - Populolion growth Asymptote 36- 32- •^^'"^ 28- / "^ E 24- / /• ^20- 0 m 16- / / /. 12- •/ / 8- y • 0- — > / 4 -i — r I 1 I T n This term agrees with the reasonable idea that the competitive effect on one population is propor- tional to the size of the other. For the fishing term I adopted from Fox (1970): 6 8 10 12 BROOD INTERVALS (3 weeks) 16 18 f{X,) = q,X,P, = F,P,. (6) Substitution of Equations (4), (5), and (6) in Equation (1) provides the model for the first population: Figure 8. -Initial growth for the independent population of the guppy, with fitted Gompertz curve. fitting the Fox (1970) model. The zero points plus two other exploitation rates (relatively stable periods considered to be equilibrium points: guppy, weeks 316-334 and 355-373, Figure 2; swordtail, 322-334 and 342-354, Figure 3) gave three fitting points for each species (Figures 10, C?Pj/C?f = PjA-i(l0g,Pioo-l0g,Pi) -fiPg-^lA- C'') ^^1 Swordtoil- Pop Growth By exactly parallel derivation the model for the second population is: dP^ldt = P,k 2(log,P2oo - log,^2) -C2P1- F,P,. (8) Thus the model for the competing populations represents a modification of the Fox (1970) ex- poential surplus-yield model, with the addition of a term for competition. Determination of Constants Growth data were obtained from the indepen- dent populations. Gompertz (1825) curves were fitted to the initial growth period for both species (Figures 8, 9), using the analog computer method of Silliman (1967). Asymptotic levels were 38.7 g for the guppy and 33.7 g for the swordtail. These values were used for the zero exploitation levels in Asymptote 0 2 4 6 8 10 12 14 BROOD INTERVALS (4 weeks) Figure 9.— Initial growth for the independent population of the swordtail, with fitted Gompertz curve. 881 FISHERY BULLETIN: VOL. 73, NO. 4 ;o.6- 1.4-1 Swordloil - Fo« model A ^ 1.0- o V ^~~-^^^^ 1 0.8- • ^^~~~~- '^^^,^__^^ • a — ^...^ 1 0.6- 0.2- 0- 1 1 1 1 1 1.2 1.6 2.0 EFFECTIVE EFFORT .6 .8 1.0 EFFECTIVE EFFORT 1.2 1.6 1.8 20 25 30 BIOMASS (groms) 45 Figure 10. -Fox (1970) model fitted to yield data for the guppy. Exploitation rates are 0.000, 0.257, and 0.326 per 3-wk period fleft-to-right in upper panel, reversed in lower panel). 12 16 20 BIOMASS (grams) Figure U.-Fox (1970) model fitted to yield data for the sword- tail. Exploitation rates are 0.000, 0,100, and 0.157 per 4-wk period (left-to-right in upper panel, reversed in lower panel). 11). Effective exploitation rates shown varied slightly from the "target" rates because of lack of infinite divisibility of the populations and because of errors. The fitted Fox models yielded values of A- of 0.260 per 3 wk and 0.321 per 4 wk for the guppy and swordtail, respectively. Comparable values for the Gompertz curves were 0.193 and 0.260. It was considered more appropriate to use the values from the Fox model because the analyses were based on that model. To place the swordtail on the same time scale as the guppy, the value of k was multiplied by %, or %(0.321) = 0.241. Data of catch and biomass for the competing populations (Table 8) were used to calculate exploitation rates. Again the effective rates varied from the target rates as explained in the preceding paragraph. Also, it was again necessary to adjust the effective rates for the swordtail to the same time scale as the guppy. This was done by the formula m = 1 - {1 - m.'y\ where ni' is the unadjusted rate. Finally, for use in the differential equations, the 3-wk rates must be converted to instantaneous rates. The formula is: F = -log^, (1 - m), from Ricker (1958). The use of instantaneous exploitation rates as employed herein assumes that P declines con- tinuously, whereas the experimental technique was to remove all the fish at the beginning of the brood interval. It can readily be shown, however, that the reduction in population resulting from the application of rn at the beginning of a period is exactly the same as the application of the equivalent F throughout the period, even if both are superimposed on a constant natural mortality. A summary of all the constants used in applying Formulae (7) and (8) to biomass data from the competing populations is given in Table 9. Where both unadjusted and adjusted data are shown, the latter were the ones used. Application of the Model Using standard analog computer techniques (Ashley 1963) values of guppy and swordtail 882 SILLIMAN: EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS Table 8.-Exploitation of competing populations. Rates are per brood interval: guppy, 3 wk; swordtail, 4 wk. Target Effective Biomass Catch Target Effective Biomass Catch Week rate rate (g) (g) Week rate rate (g) (g) Gi jppy 203 206 20.2 23.1 3.6 4.7 62 0.333 0.335 38.0 12.1 65 28.3 11.1 209 0.250 0.259 21.0 5.3 68 24.2 7.7 212 19.9 5.1 71 22.5 7.0 215 218 16.6 •M.1 4.5 3.6 86 0.333 0.344 14.8 5.2 221 12.4 3.5 89 10.4 3.7 224 10.6 2.8 92 7.4 2.4 227 8.3 1.9 95 5.9 2.0 98 3.9 1.3 Swordtail 101 0.100 0.113 3.0 0.4 30 0.333 0.346 22.3 8.0 104 2.8 0.4 34 13.5 3.9 107 2.7 0.5 38 10.2 4.0 110 2.3 0.1 42 6.7 2.6 113 2,5 0.1 46 4.3 1.5 116 2.6 0.2 50 2.8 0.8 119 2.6 0.1 54 2.2 0.8 122 3.2 0.3 58 1.5 0.4 125 3.5 0.1 116 0.100 0.112 19.0 2.6 128 4.0 0.7 120 18.2 2.4 131 4.2 0.5 124 19.1 1.3 134 4.6 0.4 128 20.2 2.5 137 5.3 0.4 132 19.2 2.3 140 5.4 0.3 136 20.1 1.9 143 6.2 0.6 146 6.2 0.4 140 0.250 0.254 21.2 5.9 149 6.3 0.7 144 18.8 4.9 152 6.8 0.7 148 16.9 4.5 155 5.4 0.4 152 15.0 3.0 158 161 6.2 6.8 0.4 0.8 156 0.100 0.112 15.2 1.7 164 6.8 0.5 176 0,100 0.088 26.2 1.2 167 7.1 0.8 180 26.6 1.2 170 6.2 0.6 184 26.5 1.5 173 6.1 0.7 188 22.0 2.2 '175 6.7 2.0 192 18.0 4.0 179 10.3 1.2 2197 14.4 0.2 182 11.0 1.0 200 12.1 1.0 185 12.2 1.1 204 11.5 1.3 188 14.5 1.5 208 12.1 1.6 191 15.5 3.4 212 12.8 1.7 194 16.2 1.8 216 9.3 1.0 197 200 0.200 0.223 17.5 16.3 6.4 2.5 220 224 228 8.6 7.5 5.6 0.8 0.6 0.5 iShould have been 176. 2Shoi lid have been 196. biomass were simultaneously generated using Formulae (7) and (8). A number of trials were made on data from the control populations (Table 10) to find the most suitable values of the compe- tition coefficients c, and Cn Values of Ci = 0.071 (guppy) and c, = 0.120 (swordtail) produced curves (Figure 12) which fitted reasonably well except for the oscillatory variations to be discussed below. Table 9.-Constants used in fitting simulation model to biomass data for competing populations. Swordtail Gu ppy ^2 = 0.321 (4 •wk) k = 0.260 (3-wk) k, = 0.241 (3 -wk) P, 00 = 38-2 9 ''200 = 32.7 g 3-wk m, F, 3-wk period Target Effective Adjusted Adjusted period Target Effective F2 20-23 0.333 0.335 0.408 9-19 0.333 0.346 0.273 0.319 28-32 0.333 0.344 0.422 38-45 0.100 0.112 0.085 0.088 33-64 0.100 0.113 0.120 46-50 0.250 0.254 0.197 0.219 65-68 0.200 0.223 0.252 51 only 0.100 0.112 0.085 0.088 69-75 0.250 0.259 0.300 58-75 0.100 0.088 0.067 0.069 883 FISHERY BULLETIN: VOL. 73, NO. 4 T.ABLE lO.-Biomass levels, 3-\vk means. control popu lations. Table ll.-Biomass levels, .3- wk means test popu ations. We ght (g) Period Weight (g) Period Weight (g) Period Weight (g) Period Guppy Swordtail Guppy Swordtail Guppy Swordtail Guppy Swordtail 0 1.6 8.2 38 21.0 0.3 0 2.0 9.6 38 2.6 18.7 1 2.1 13.2 39 21.4 0.3 1 2.0 '12.6 39 2.5 17.9 2 2.4 10.6 40 23.4 0.3 2 3.8 13.0 40 3.1 18.4 3 4.5 10.5 41 25.0 0.3 3 5.6 14.2 41 3.4 18.5 4 6.6 14.0 42 27.8 0.3 4 6.8 14.3 42 3.7 19.5 5 8.2 17.0 43 29.0 0.0 5 7.6 16.0 43 3.8 19.2 6 10.1 18.5 44 29.1 0.0 6 9.4 17.2 44 4.4 18.5 7 11.5 20.4 45 30.2 0.0 7 10.6 18.9 45 4.9 19.5 8 12.6 21.4 46 30.8 0.0 8 12.6 20.5 46 5.2 18.9 9 14.0 21.7 47 31.4 0.0 9 13.4 22.3 47 6.2 18.6 10 15.2 22.1 48 31.6 0.0 10 15.1 13.4 48 6.3 15.9 11 16.6 22.0 49 32.7 0.0 11 16.6 11.1 49 6.1 14.5 12 18.6 19.4 50 31.4 0.0 12 18.7 8.9 50 6.6 14.0 13 19.8 17.8 51 32.5 0.0 13 21.5 6.6 51 5.4 14.5 14 21.0 17.3 52 34.1 0.0 14 25.0 4.3 52 6.1 14.7 15 20.4 23.5 53 35.2 0.0 15 27.9 3.1 53 6.5 16.5 16 19.3 20.9 54 37.0 0.0 16 30.8 2.5 54 6.8 18.3 17 19.7 20.3 55 36.4 0.0 17 35.4 2.1 55 6.8 20.6 18 20.9 19.5 56 36.5 0.0 18 37.1 1.5 56 6.2 22.7 19 21.4 16.9 57 40.7 0.0 19 38.5 1.2 57 6.4 ■26.0 20 23.4 15.4 58 42.6 0.0 20 34.3 1.1 58 7.5 26.8 21 25.1 14.2 59 '47.3 0.0 21 25.6 1.2 59 9.5 26.4 22 27.3 13.4 60 44.1 0.0 22 21.6 1.2 60 10.9 26.2 23 29.1 12.8 61 43.9 0.0 23 19.7 1.1 61 12.0 24.8 24 30.8 11.1 62 42.4 0.0 24 (') P) 62 13.7 22.2 25 32.3 10.8 63 42.7 0.0 25 (') {') 63 14.6 18.3 26 32.6 9.4 64 45.5 0.0 26 {') {') 64 15.3 16.3 27 32.7 8.0 65 44.4 0.0 27 {') P) 65 16.9 13.1 28 32.2 7.2 66 34.4 0.0 28 12.7 22.0 66 15.6 11.6 29 32.7 5.7 67 29.7 0.0 29 9.2 21.4 67 18.5 10.9 30 32.0 4.0 68 32.5 0.0 30 6.7 22.4 68 20.1 11.6 31 30.1 3.0 69 35.8 0.0 31 5.2 22,9 69 19.3 11.7 32 27.1 2.0 70 36.7 0.0 32 3.5 22.0 70 17.3 11.0 33 24.8 1.5 71 234.1 0.0 33 2.9 20.9 71 316.3 38.8 34 23.1 1.0 72 230.9 0.0 34 2.7 20.3 72 314.2 39.0 35 23.3 0.9 73 228.0 0.0 35 2.6 19.9 73 312.4 38.3 36 22.5 1.0 74 225.5 0.0 36 2.3 20.1 74 310.5 37.4 37 21.9 0.6 75 1.222.8 0.0 37 2.5 20.1 75 38.7 36.1 'Based on two observations. 2Based partly on interpolation. 'Based on two observations. ^Reconstruction interval. 3Based partly on interpolation. These values were used in applying Equations (7) and (8) to data from the exploited test populations (Table 11). Curves for the test populations (Figure 13) followed the general trend of the hiomass levels, even though oscillatory deviations were great. 50-1 40- 30 < 20 s 10 • • • • •• - • ^ ••• • ^^^^ ^,**<'. .* • z"^- ..•* /^ • • , / • . /a * •••..• X / a*\«»a / / '^^\t. / * N* 7 / • ^'^ :7 /. Pl'">iAAAAA 0 10 20 30 40 50 60 70 3-WEEK PERIODS Figure 12. -Fitting of simulation model to control populations. Dots are for guppy, triangles for swordtail. Solid lines are fitted curves. 30 3-WEEK PERIODS Figure 13.-Fitting of simulation model for test populations. Dots are for guppy, triangles for swordtail. Solid lines are fitted curves. 884 SILLIMAN: EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS Oscillatory Fluctuations The substantial oscillatory deviations evident in comparisons of observed and simulated biomass levels (Figures 12, 13) suggest the need for special study. Deviations can be evaluated more readily if they are plotted along a straight baseline (Tables 12, 13; Figure 14). Viewed in this manner, devia- tions appear at least roughly regular with respect to time. Also, they tend to be similar for control and test populations of the guppy for comparable periods. They did not, therefore, result solely from perturbations due to exploitation, although in the test populations deviations were somewhat greater post-exploitation than pre-exploitation. Oscillations of the type described above seem to be basic to many populations. Walter (1973) points out that such a fluctuation occurred in Lake Michigan alewives. He developed delay-differen- tial equations which are compatible with "an os- cillatory sort of behavior centered about the equilibrium level." Although it would be of interest to apply such models to the guppy-sword- tail data, it is unlikely that they would provide substantially different results insofar as the basic relations between exploitation and yield are con- cerned. It is with such relations that I am primarily concerned in this paper. Since there were some fluctuations in water temperature (Figure 1) during the course of the experiments, it seemed possible that these might have caused all or part of the oscillatory population changes. To test this, a regression was erected for periods 33-75, when temperature fluctuations were fairly regular. The independent variable was water temperature, and the dependent variable was deviation of the control guppy population biomass from that predicted by the simulation model (Table 12, Figure 14). Results showed no significant correlation between biomass devia- tions and temperature (r = 0.087, P>0.1). It is worthwhile in this discussion of oscillatory Table 12. -Deviations of actual biomass from theoretical, on basis of fitted model, guppy. Table 13.— Deviations of actual biomass from theoretical, on basis of fitted model, swordtail. Deviation Control (g) Test Period Devi ation (g) Period Deviation (g) Period Devi ation (g) Period Control Test Control Test Control Test 0 -0.4 -1.0 38 -15.7 -2.7 0 2.2 0.1 38 0.3 -8.8 1 -0.4 -1.8 39 -15.5 -2.9 1 4.5 0.3 39 0.3 -8.0 2 -1.1 -1.5 40 -13.7 -2.6 2 -0.8 -1.8 40 0.3 -6.2 3 -0.4 -1.5 41 -12.2 -2.8 3 -3.3 -2.6 41 0.3 -5.0 4 -0.1 -2.3 42 -9.6 -3.0 4 -1.9 -4.1 42 0.3 -3.1 5 -0.5 -3.7 43 -8.4 -3.5 5 -0.6 -3.7 43 0.0 -2.6 6 -0.7 -4.1 44 -8.4 -3.6 6 -0.4 -3.3 44 0.0 -2.7 7 -1.5 -5.2 45 -7.3 -3.9 7 0.5 -2.1 45 0.0 -1.0 8 -2.7 -5.1 46 -6.8 -4.5 8 0.9 -0.8 46 0.0 -0.9 9 -3.2 -6.0 47 -6.2 -4.6 9 0.9 1.0 47 0.0 1.3 10 -3.8 -6.2 48 -6.0 -5.5 10 1.3 -2.4 48 0.0 0.7 11 -4.0 -6.6 49 -5.0 -6.9 11 1.2 -0.8 49 0.0 1.1 12 -3.5 -6.5 50 -6.3 -7.6 12 -1.2 0.1 50 0.0 2.0 13 -3.7 -5.7 51 -5.2 -9.9 13 -2.4 0.2 51 0.0 3.8 14 -3.8 -3.9 52 -3.6 -10.1 14 -2.4 0.3 52 0.0 4.0 15 -5.6 -2.6 53 -2.5 -10.4 15 4.5 2.1 53 0.0 4.7 16 -7.7 -1.0 54 -0.7 -10.6 16 2.6 2.5 54 0.0 5.6 17 -8.2 2.4 55 -1.3 -10.9 17 2.6 2.1 55 0.0 7.1 18 -7.7 3.1 56 -1.2 -11.7 18 2.7 1.5 56 0.0 8.5 19 -7.9 3.8 57 3.0 -11.6 19 0.7 1.2 57 0.0 11.2 20 -6.4 -1.0 58 4.9 -10.5 20 0.1 1.1 58 0.0 11.4 21 -5.2 0.0 59 9.6 -8.6 21 -0.4 1.2 59 0.0 11.2 22 -3.5 1.6 60 6.4 -7.3 22 -0.4 1.2 60 0.0 11.4 23 -2.1 3.4 61 6.2 -6.3 23 -0.2 1.1 61 0.0 10.3 24 -0.9 (') 62 4.7 -4.7 24 -1.1 (') 62 0.0 8.0 25 0.3 (') 63 5.0 -3.9 25 -0.6 (') 63 0.0 4.5 26 0.3 (') 64 7.8 -3.2 26 -1.2 (') 64 0.0 2.7 27 0.0 (') 65 6.7 -1.7 27 -1.7 (') 65 0.0 -0.2 28 -0.8 0.2 66 -3.4 -0.9 28 -1.6 5.8 66 0.0 -1.7 29 -0.7 -1.2 67 -8.1 3.2 29 -2.0 4.1 67 0.0 -2.4 30 -1.7 -1.9 68 -5.3 5.6 30 -2.5 3.6 68 0.0 -1.8 31 -4.1 -1.9 69 -2.0 5.6 31 -2.0 2.4 69 0.0 -2.0 32 -7.4 -2.3 70 -1.1 4.6 32 -1.1 0.2 70 0.0 -3.1 33 -10.2 -1.8 71 -3.7 4.6 33 1.5 -2.3 71 0.0 -5.7 34 -12.4 -2.0 72 -6.9 3.2 34 1.0 -4.2 72 0.0 -6.0 35 -12.6 -2,2 73 -9.8 2.0 35 0.9 -5.6 73 0.0 -7.1 36 -13.7 -2.7 74 -12.3 0.6 36 1.0 -6.3 74 0.0 -3.5 37 -14.6 -2.7 75 -15.0 -07 37 0.6 -6.9 75 0.0 -5.2 'Reconstruction interval. 'Reconstruction interval. 885 FISHERY BULLETIN: VOL. 73, NO. 4 z o (- > 10 20 30 40 50 3-WEEK PERIODS 60 70 80 Figure 14.— Deviations from simulation modeL Broken lines show lags for comparable portions of test and control popula- tions. fluctuations to consider the models fitted under independent and competing conditions. The former were based on equilibrium population con- ditions, whereas the latter recognized non- equilibrium conditions and used nonlinear differential equations capable of expressing con- tinuous variation in population and yield under stable-limit cyclic variation. Conclusions for management may be different under the second type of formulation, nevertheless I feel that the conclusions drawn below are of value. Finally, it is pertinent to discuss what seems likely to have been the start of an oscillatory fluc- tuation in the independent population of the swordtail. As mentioned under "Course of Populations," number and biomass increased dur- ing the final five brood intervals, contrary to what might be expected as a result of the 16.7% exploi- tation rate applied. The incipient oscillation may have been triggered by the low level of biomass reached just before it began, through overcom- pensation of the population. This level was lower than any that had been in effect since the initial growth of the population, and it may have moved the population toward the high recruitment rates .shown in Figure 7. CONCLUSIONS Extinction of the swordtail population in the control pair (Figures 4, 12) as mentioned under "Course of Populations," is compatible with the theory of competitive exclusion first advanced by Cause (1934). He stated that where two popula- tions are fully competing, one will have a slight advantage in growth or aggression and eventually displace the other. This occurrence illustrates one of the values of conducting population experiments over a suflficient period for natural phenomena to develop. The extinction of the swordtail could hardly have been anticipated dur- ing the first few months of the experiment, when growth of the swordtail actually outstripped that of the guppy. Cause's phenomenon of "mutual depression" (Cause and Witt 1935) also was exemplified in the experiments. Quantitative measures of this were provided by the coefficients of competition, r, and<:-2, determined (by succes- sive trials) for Formulae (7) and (8). These were 0.071 and 0.120 for the guppy and swordtail, re- spectively. These values show greater depression for the swordtail than for the guppy and, therefore, the superior competitive ability of the guppy. Growth advantage for the guppy was in- dicated by the values of k and /^ (Table 9), both of which were greater for the guppy. My greatest interest in these experiments was to discover what combination of exploitation rates would produce the greatest sustainable yield for the two populations. This problem can be approached by calculating equilibrium yields for pairs of population sizes P, and P.,- At equilibrium, the left hand sizes of Equations (7) and (8) are equal to zero; with the constants already deter- mined, F] and i^gcan be calculated for any pair of values P] and Pg- To obtain 3-wk yields, F^ and F., were converted back to »?jand m.^^ by the formula m = I - exp(-F). Then total 3-wk yields, com- parable to the yields actually obtained in the experiments (Table 8), represent the sum of /??jP, (guppy) and m^2 (swordtail). Yields are directly comparable for the guppy, but values in Table 8 must be multiplied by three fourths for the swordtail. I expressed the total yields {m jPj -I- m^2^ ^^ ^^e form of yield isopleths (Figure 15). Inspection of 886 SILLIMAN: EXPERIMENTAL EXPLOITATION OF FISH POPULATIONS .30- < 1.25- o 0.05) in 18 of 19 pairs. For the single exception, Cera- toscopeh(>< ivarniiiigi from tow 161, the mean size of fish from the catcher size was larger (P<0.05; .f (SL): 38 vs. 34 mm). However, since the distribu- tions had considerable overlap, this set was included in the study. Prey Abundance in Stomachs It was necessary to apply a square root trans- form ( /X-l-0.5 ) to the data on number of prey per stomach since the high frequency of empty 910 HOPKINS and BAIRD: NET FEEDING IN MESOPELAGIC FISHES Table 2.-ReIative abundance of principal (top 3) food items in fish stomachs and in plankton taken concurrently with fish, with 333-jLun mesh nets mounted in the mouth of the double trawl (see Figure 1): F = "fish-catcher" set of fish; P = cod end plankton net set. Numerical ab undance Numerical Numerical Top 3 prey of prey in stomachs (%) abundance Top 3 Items abundance items in in plankton in plankton in plankton Species Tow stomachs F P X nets (%) nets nets (%) Argyropelecus 142 Oncaea 45 42 43.5 12 Eucalanus 41 hemigymnus Conchoecinae 17 23 20.0 5 Oncaea 12 Eucalanus 12 9 10.5 41 Scolecithricidae 10 Benihosema 137 Oncaea 23 24 23.5 9 Conchoecinae 12 suborbitale Pleuromamma 21 17 19.0 9 Pleuromamma 9 Conctioecinae 11 10 10.5 12 Clausocalanus 9 Ceratoscopelus 137 Limacina 13 17 15.0 3 Conchoecinae 12 warmingi Conchoecinae 12 14 13.0 12 Pleuromamma 9 Siphonophores 12 11 11.5 2 Clausocalanus 9 Gonostoma 144 Stylocheiron 16 21 18.5 2 Sagitta 19 elongatum Pleuromamma 13 33 23.0 1 Conchoecinae 18 Conchoecinae 13 9 11.0 21 Oithona 14 Lampanyctus 141 Pleuromamma 23 25 24.0 4 Oithona 14 alatus Stylocheiron 16 13 14.5 4 Sagitta 8 Conchoecinae 8 6 7.0 8 Conchoecinae 8 Lepidophanes 137 Pleuromamma 25 31 28.0 9 Conchoecinae 12 guentheri Euphausia 14 6 10.0 <1 Pleuromamma 9 Conchoecinae 11 9 10.0 12 Clausocalanus 9 Valenciennellus 142 Oncaea 24 12 18.0 12 Eucalanus 41 tripunctulatus Pleuromamma 19 22 20.5 3 Oncaea 12 Euchaeta 10 ( 6) 8.0 1 Scolecithricidae 10 Eucalanus ( 3) 15 9.0 41 Valenciennellus 143 Oncaea 40 38 39.0 9 Pleuromamma 15 tripunctulatus Pleuromamma 23 25 24.0 15 Conchoecinae 14 Conchoecinae 6 8 7.0 14 Oithona 10 Valenciennellus 135 Eucalanus 19 16 17.5 20 Pleuromamma 23 tripunctulatus Pleuromamma 16 14 15.0 24 Eucalanus 20 Euchaeta 10 ( 9) 9.5 <1 Euphausia 7 Oncaea ( 3) 19 11.0 1 stomachs resulted in significant skewness in the distributions of many untransformed data sets. In three set comparisons there were significant differences (F- tests, 0.05 >P>0. 025) in variance {BentJwsema suborbitale, tow 137; C. ivanningi, tow 137; Lampanijctus alatus, tow 167). In these cases, tests comparing means of normal distribu- tions when population variances are unequal were applied as described by Johnson and Leone (1964:226). There were significant {t-tests, P<0.05) differences in 5 of 19 comparisons of number of prey items per stomach. Lampanyctus alatus in two collections contained more prey items per in- dividual in fish taken from the plankton net cod end (tow 141: 0.025 >P>0.01; tow 167: 0.05 >P >0.025). However, for Gonostoma elongatum in two sets (tows 144, 145: 0.05>P>0.025), and C. warm- ingi in one set (tow 161:P<0.005), individuals from the fish-catcher side averaged more prey per stomach. Because of possible diurnal feeding periodicity in mid-water fishes (Anderson 1967; Hoiton 1969; DeWitt and Cailliet 1972; Baird et al. 1975), fish entering the trawl at different periods in their feeding cycle may be satiated or have a different predisposition to feed in varying degrees. The five sets of fish showing significant differences in number of food items, howeVer, are not conspicuously grouped in any single time period (see Table 1) and no general relationship is apparent in our results between time of capture and relative abundance of prey in fish from either side of the trawl. Mean Prey Size In 8 of 19 data sets, mean prey size was smaller in cod end fish. The major size modes were coin- cidental in all 19 set comparisons as judged from visual inspection. A ^-test of the grand means (mean of 19 individual means for each cod end type), however, revealed no significant (P>0.05) difference in mean size of food item for fishes in either side of the trawl (variance of means homogeneous). Though the sensitivity of this test is weakened to some degree by comparing different species of fish collected at different times, a strong bias in prey size resulting from net feeding is not apparent. Prey size distributions for 14 paired sets were also compared using the contingency chi-square test. Significant (P<0.05) differences were found 911 FISHERY BULLETIN: VOL. 73, NO. 4 in only two pairs: Lepidophane^ guentheri, tow 167 (P<0.001) and Valenciennellus. fripunctulafus, tow 135 (0.005 >P>0.001). In the former, those in- dividuals from the cod end took more prey items in smaller size classes while in the latter the reverse occurred. We have no simple explanation for these results. It is difficult to attribute them, however, to net feeding since other diet characteristics tested showed no significant differences for these same sets. Additionally, paired samples of the same species from other collections revealed no sig- nificant differences. Prey Diversity In comparison to the coarse mesh fish-catcher, the unobstructed cod end net of the adjacent trawl contained a much greater variety of plankton and consequently a more diverse potential food source for net feeding. A comparison was made of diver- sity of food items in stomachs of fish from each side of the trawl using 12 (of 19) species-pair collections represented by sets of approximately equal numbers of individuals for each cod end type. Total diversity was scored for each set of fishes, yielding two diversity values for each species-pair collection. Diversity scores were then summed to give grand means for each cod end type. On the basis of a f-test on logjo transformed data, no significant (P>0.05) difference was in- dicated for the two cod end types though total diversity was considerably greater in fishes from the plankton net cod end in some sets (e.g., Ar- gyropelecus aculeatus, tow 144; L. alatus, tow 141). Fish Scales Anderson (1967), in his analysis of the diet of Bafh)jlagi(>i stilbiiis, frequently encountered fish scales in stomachs yet no other remains of fish of the size indicated by the scales. This, in addition to the absence of scales in intestines and the oc- currence of scales and copepods in the mouths of fish, he considered as evidence of net feeding. In the present study, fish in half the sets of samples (6 of 12) for which data are presented contained no fish scales. In four of the remaining six pairs, more scales were found in fish from the cod end where scales would be expected to accumulate during the course of a tow, but none of the differences were significant {f-test on \/A' + 0.5 transformed data; P >0.05). The occurrence of fish scales in stomachs does not necessarily stem from predation on smaller fish or from eating scales abraded from fish within the trawl. Fish scales appear to be common in the water column and thus available as separate forage items. In a series of paired 30-liter bottle casts made between 0 and 1,000 m in August 1972, in the eastern Gulf of Mexico where most of the fish examined were taken, scales occurred in collections (60 liters/sample) from 7 of 15 depths sampled at densities of 17-83 per m^. Scales ranged from 0.5 to 5 mm in diameter. No fish were taken in the sample bottles and the probability of con- tamination from other sources appears low. Taxonomic Composition of Stomach Contents and Plankton Table 2 presents the principal taxonomic com- ponents of prey found in nine sets of fish from both sides of the trawl. The principal diet item was the same in both sets of fish in six of nine collections, the same prey constituted the top three food items by number in seven of nine collections and the prey taxa were in the same rank order in five of nine collections. The principal three prey taxa in fish from either side of the trawl were within ±3% of the mean value for both sides from each tow in 25 of 29 food item comparisons and all values were within ±10% of the means. These results show that the taxonomic composition of at least the principal components of the diet was similar in fish from both sides of the trawl for all comparisons. Comparison of food items in stomachs of fish from the cod end, where net feeding is assumed mostly likely to occur, with plankton catches reveals little similarity in the top three taxonomic components. In none of the nine collections was the principal taxon the same in either the plankton net catch or in the stomachs of fish from the cod end. Of particular importance are tows 137, 141, 144, 152, and 161 which sampled relatively narrow depth zones and consequently were potentially less influenced by vertical stratification of plank- ton. Also, three species of fish collected in the same haul (tow 137) each contained a different principal food item, none of which matched the most abun- dant taxon in the cod end plankton catch. The major diet components for Bciifliosema siihorbi- fale, C. icarmhigi, and L. guenfheri from tow 137 were Oncaea, Limacina, and Pleuromamma; the most abundant plankton in the cod end net were ostracods (Conchoecinae). This particular haul' was 912 HOPKINS and BAIRD: NET FEEDING IN MESOPELAGIC FISHES a horizontal tow in which a discrete depth was maintained throughout. DISCUSSION While studies of the behavior of mesopelagic fishes in small mid-water trawls used for research are nonexistent, there is considerable information available on fish behavior in larger commercial trawls of many kinds (e.g., Ben-Tuvia and Dickson 1968). Generally, fish move in front of or away from the walls of the trawl until they are exhausted and are collected in the cod end. Mesopelagic fishes from trawl cod ends often show signs of abrasion (Harrisson 1967); consequently, the likelihood of active and extensive net feeding would appear low. Several authors (e.g., Collard 1970; Hopkins and Baird 1973) have suggested that trauma induced by stress conditions in the trawl environment would operate against active feeding behavior. Reflexive gulping or "pseudo" feeding behavior, however, resulting in the ingestion of significant amounts of prey from the plankton rich cod end is a potential mechanism whereby stomach contents could be biased by net feeding. Several studies have revealed diel periodicity in feeding in mid-water fishes which indicates that at certain times, at least, net feeding cannot be extensive (Holton 1969; DeWitt and Cailliet 1972; Baird et al. 1975). The possibility of fishes foraging in front of the cod end or fish-catcher can also be evaluated. At standard trawling speeds (3.7-4.6 km/h) the trawl moves at a rate of 1.0 to 1.3 m/s. For those epipelagic species which have been examined, foraging and cruising speeds range from about 1 to 4 body lengths per second and maximum burst speeds are on the order of 10 to 30 body lengths per second (e.g., Blaxter 1969; Baird et al. 1975). Assuming similar swimming capabilities for mid- water fishes and a fish size of 3 inches (76 mm), a conservative estimate of foraging speeds should be less than 0.3 m/s and burst rates of 0.8 to 2.3 m/s. All of the species examined here were less than 76 mm in length except G. elongatum which may have somewhat limited swimming capabili- ties (Marshall 1971). In view of the swimming speeds required, extensive foraging in front of the cod end appears remote. In addition, the data from L. yuentheri (tow 98), where prey of individuals gilled in the net were compared with those from the cod end, failed to reveal indications of net feeding. The present results support the contention that if net feeding does occur, it is not extensive in the relatively small fragile fishes typical of the oceanic mesopelagic enrivonment. Only the data on prey abundance in stomachs of L. alatus could be con- strued as statistical evidence of net feeding. In both of these collections, however, mean size of prey items and taxonomic composition of diet were very similar in both sets of fish, while the diet showed little agreement in terms of principal taxonomic components (Table 2) with plankton in the cod end, as might be expected from net feed- ing. In three collections (tows 144, 145, 167), G. elongatum (2 sets) and C. ivarmingi (1 set) from the "control" side contained more food items than fish from the cod end. Here again comparisons of mean prey size, taxonomic composition of diet, and major taxa in diet with that in cod end plankton samples failed to reveal evidence of net feeding. Furthermore, there were often substantial differences between principal taxonomic com- ponents of diet and plankton from cod end catches from the same haul. Mean prey size (with two ex- ceptions) and composition of principal taxa of diets were nearly identical for all sets of com- parison which further indicate the limited nature of net feeding in this study. The use of fish scales as a criterion for net feed- ing poses a number of difficult problems. Our hydrocasts reveal, for instance, that fish scales oc- cur naturally in the water column. Further, several studies of both marine and freshwater teleost fishes have shown that scales (probably also the covering mucous and epidermis) can serve as a major component of the diet, appear to be easily digested, and may possibly have considerable nu- tritive value (e.g., Roberts 1970, 1973; Carr and Adams 1972, 1973). Scales were relatively rare in stomachs examined here but did occur in fishes from both sides of the trawl. Since scales are present in the natural environment, may have nu- tritive value, and are possibly easily seen, cap- tured, and eaten, they could serve as a natural food source or provide appropriate stimuli to elicit ingestion. Until more evidence is obtained con- cerning the role of scales in the natural diets of fishes and their abundance in oceanic environ- ments, the presence of scales in the stomachs of mid-water fishes cannot be used with assurance as an indicator of net feeding. Because of the difficulty of replicating trawl conditions and obtaining sufficient material for analysis, the variability in distributions of mid- 913 FISHERY BULLETIN: VOL. 73, NO. 4 water fishes with respect to time and space, and possible variations in feeding cycles, the present collections are not ideal in all respects. The study did include representatives of most of the major groups of common mesopelagic fishes from a variety of depths and times, and the results may be expected to be broadly applicable to many mid- water environments. Considering the simul- taneous time-depth collections with the double trawl of both "control" fish and those with the opportunity to ingest food in the cod end, this study provides the first reasonably good test of net feeding in mesopelagic fishes. The relatively small differences in the mean number and taxonomic composition of prey items in most sets of stomachs are encouraging. The results presented here sug- gest that the published literature on the diets of mesopelagic fishes is not seriously biased by net feeding and that existing collections can be used for trophic investigations. ACKNOWLEDGMENTS We acknowledge with pleasure contributions to this work by R. C. Beckett of the Naval Research Laboratory and E. E. Gallaher, D. M. Milliken, J. K. Rolfes, and W. R. Weiss of the University of South Florida. Critical review was provided by D. F. Wilson, B. J. Zahuranec, and C. Woodhouse. Support was received from the State University System Institute of Oceanography (Florida), the Naval Research Laboratory and Environmental Protection Agency contract No. 5444 and NSF Grant DES- 75-03845. LITERATURE CITED Anderson, R. 1967. Feeding chronology in two deep-sea fishes off California. M.S. Thesis, Univ. South California, Los Ang., 22 p. Baird, R. C, T. L. Hopkins, and D. F. Wilson. 1975. Feeding chronology of Diaphux taaningi Norman in the Cariaco Trench. Copeia 1975:356-365. Ben-Tuvia, a., and W. Dickson (Editors). 1968. Proceedings of the conference on fish behaviour in relation to fishing techniques and tactics. FAO (Food Agric. Organ. U.N.) Fish. Rep. 62:1-47. Blaxter, J. H.S. 1969. Swimming speeds of fish. FAO (Food Agric. Organ. U.N.) Fish. Rep. 62:69-100. Carr, W. E. S., and C. a. Adams. 1972. Food habits of juvenile marine fishes: Evidence of the cleaning habit in the leatherjacket, Oligoplites saurus, and the spottail pinfish, Diplodus hulbrooki. Fish. Bull., U.S. 70:1111-1120. 1973. Food habits of juvenile marine fishes occupying seagrass beds in the estuarine zone near Crystal River, Florida. Trans. Am. Fish. Soc. 102:511-540. Collard, S. B. 1970. Forage of some eastern Pacific midwater fishes. Copeia 1970:348-354. DeWitt, F. a., Jr., and G. M. Cailliet. 1972. Feeding habits of two bristlemouth fishes, Cyrlofhone acclinidens and C. signata (Gonostomatidae). Copeia 1972:868-871. DeWitt, H. H., and T. L. Hopkins. In press. Aspects of the diet of the Antarctic herring Pleuragramma antarcticum Boulenger. //(. G. A. Llano (editor), Third Symposium on Antarctic Biology: Adap- tations within Antarctic Ecosystems. Harrisson, C. M. H. 1967. On methods for sampling mesopelagic fishes. Symp. Zool.Soc.Lond. 19:71-126. HOLTON, A. A. 1969. Feeding behavior of a vertically migrating lan- ternfish. Pac. Sci. 23:325-331. Hopkins, T. L., and R. C. Baird. 1973. Diet of the hatchetfish Sternaptyx diaphana. Mar. Biol. (Berl.) 21:34-46. Hopkins, T. L., R. C. Baird, and D. M. Milliken. 1973. A messenger-operated closing trawl. Limnol. Oceanogr. 18:488-490. Johnson, N. L., and F. C. Leone. 1964. Statistics and experimental design in engineering and the physical sciences. Vol. 1. John Wiley & Sons, N.Y., 523 p. Judkins, D. C, and a. Fleminger. 1972. Comparison of foregut contents of Sergestes similis obtained from net collections and albacore stomachs. Fish. Bull., U.S. 70:217-223. Marshall, N. B. 1971. Explorations in the life of fishes. Harvard Univ. Press, Camb., Mass., 204 p. Nesis, K. N. 1965. Distribution and feeding of young squids Gonatus fabricii (Licht.) in the Labrador Sea and the Norwegian Sea. Oceanology 5(1):102-108. Roberts, T. R. 1970. Scale-eating American characoid fishes, with special reference to Proholodiis heteroxtomuus. Proc. Calif. Acad. Sci. 38:383-390. 1973. The glandulocaudine characoid fishes of the Guayas Basin in Western Ecuador. Bull. Mus. Comp. Zool. 144:489-514. SOKAL, R. R„ and F. J. ROHLF. 1969. Biometry. The principles and practice of statistics in biological research. W. H. Freeman and Co., San Franc, 776 p. 914 NOTES GAS-BUBBLE DISEASE: MORTALITIES OF COHO SALMON, ONCORHYNCHUS KJSUTCH. IN WATER WITH CONSTANT TOTAL GAS PRESSURE AND DIFFERENT OXYGEN-NITROGEN RATIOS' A review of the literature regarding gas-bubble disease can be found in a recent publication by Rucker (1972); one by the National Academy of Science (Anonymous in press); and an unpublished report by Weitkamp and Katz (1973).- Most dis- cussions on gas-bubble disease have dealt with the inert gas, nitrogen-oxygen was given a secondary role. It is important to know the rela- tionship of nitrogen and oxygen when we are concerned with the total gas pressure in water. Where water becomes aerated at dams or falls, oxygen and nitrogen are usually about equally sat- urated, however, many of the samples analyzed from the Columbia River indicate that nitrogen is often about 79f higher than oxygen when expressed as a percentage. When oxygen is removed from water by metabolic and chemical action, or when oxygen is added to the water by photosynthesis, there is a definite change in the ratio of oxygen and the inert gases (mainly ni- trogen with some argon, etc.). This present study shows the effect of varying the oxygen and ni- trogen ratio in water on fingerling coho salmon, Oncorhiinchus. kisutch, while maintaining a con- stant total gas pressure. The primary purpose of these experiments was to determine differences in lethality of various gas ratios of oxygen and nitrogen at a constant total gas pressure of 119%. I also wished to deter- mine whether there was a difference in suscep- tibility between sizes and stocks of juvenile coho. Also to be examined was the effect of reducing the oxygen while holding the nitrogen constant. Methods Juvenile coho salmon averaging 6 cm in length, obtained from the Quilcene National Fish Hatchery, Quilcene, Wash., and the Northwest Fisheries Center of the National Marine Fisheries Service, NOAA (National Oceanic and Atmo- spheric Administration), Seattle, Wash., were used during all the tests concerning differences in lethality of O2 /N2ratios. During these tests water temperatures were 13.6°+ 0.1°C. Gas concentra- tions usually varied slightly from the desired ratios. The tank facility consisted of six troughs, two of which were used to hold experimental fish at normal saturation (100%) and two pairs of troughs used to test fish at different gas ratios. Control of gas concentrations and the test ap- paratus is described in a subsequent section. Dur- ing initial testing of the gas control system, I de- termined that a ratio of 114% O2 to 121% N2 could be achieved by merely allowing air to be sucked into the intake side of the recirculation pump. Since this gas ratio did not require injection of either oxygen or nitrogen, the resultant concen- tration (114% 02and 121% N2) was used as a quasi control for comparison with the other gas ratios. Several replicates were completed at this concen- tration. Water saturated at this ratio and concen- tration was also used to test for differences in size and stock and to provide base line data in deter- mining effect of reduced oxygen concentrations while maintaining a constant nitrogen level. In all the tests free carbon dioxide was near normal, or about 2 ppm. Oxygen is expressed as "O2" and the inert gases as "Ng." The number of days required to kill 25% of the fish at the different gas levels is expressed as the lethal exposure-LEgsand to kill 50%- LE50. Apparatus shown in Figure 1 was used to supply water with a definite oxygen and nitrogen con- tent. The tank (1) was divided so that two experiments could be carried on simultaneously with similar equipment. Water was circulated by a centrifugal pump (2) with a valve (3) on the effluent side to cause a controlled back pressure as read on a gauge (4). This created a vacuum on the inflow side (5) so that air could be introduced into the water with either oxygen or nitrogen (6) 'Research performed under contract with the U.S. Army Corps of Engineers. -D. E. Weitkamp and M. Katz. 1973. Resource and literature review of dissolved gas supersaturation in relation to the Columbia and Snake River fishery resources. Submitted to Northwest Utilities Cooperative, c/o Idaho Power Co., Boise, Idaho, Apr. 3, 197.3, by Seattle Marine Laboratories, Div. of Xelco Corp., Seattle, Wash., 55 p. (Typewritten.) 915 through a "Y" tube. Circulation of the water caused an increase in temperature which was maintained at approximately 13.6°C by means of a refrigeration system (7) and recorded on a ther- mograph (8). Water level in the tank was main- tained by float valves (9). Each trough was supplied with 1 liter per minute of water regulated with flow meters (10). The water used was from the municipal supply, was soft, and was passed through activated charcoal to remove the chlorine. A greater depth of water was needed for absorp- tion of the gases than was afforded by the tank (1), so two towers (11) were added to the system. The spout (12) at the top of the towers was to direct possible overflow water back into the system. Inside dimensions of troughs in the fish holding area (13) were 104.5 x 23.5 x 20 cm high. Water depth was maintained at 14 cm. Each trough could Figure 1. -Apparatus for subjecting fish to constant-temperature, flowing water with a definite oxygen and ni- trogen content. be separated into three compartments with screens— "A" was at the inflow end of the trough, "B" middle, and "C" outflow end. In a few cases a compartment was divided longitudinally so that two groups of fish could be subjected to almost identical conditions. Results Effect of Variation in Oj /N^ Ratios on Mortality Times to death (LE2.-, and LE50 ) of juvenile coho salmon at various concentrations of 0^ and N2 during constant total gas saturation of 119^^ ap- pear in Table 1 are shown graphically in Figure 2. All tests were run in duplicate with 50 fish per test, except one at 229% 02and 90% Ng which involve 50 fish but one test. With one exception (192% 02and 916 Table l.-Timetodeathof groups of juvenile coho salmon (about 6 cm long) in 13.6°C water with total gas pressure of 119% of saturation and different ratios of 02and N2. Time to death (days) RflS conCP'nt rntinn (% sati jration) N2 25% of Range all fish Average 30% 0I all fish O2 Range Average X Q 50 138 1.8- 1.9 1.9 3.2-4.0 3.7 75 131 1.8- 2.7 2.3 3.5-4.3 3.9 114 121 3.2- 4.1 3.8 6.3-7.3 6.9 0 159 109 3.2- 5.3 4.5 6.5-9.1 8.2 h- 173 105 33.5-35.3 34.4 — (') U) 192 100 — 232.0 — (') >- 229 90 — {') — {') < 'Not reached, 28% mortality in 39 days. 20ne replicate reached 24% mortality in 30 days; the other, 25% in 32 days. 3Not reached, test concluded at 33 days. "Not reached, 20% mortality in 35 days. XO2 %Nj 101 A 50 138 B 75 131 C 114 121 0 159 109 E 173 105 F 192 100 G 229 90 20 40 Figure 2.— Mortality pattern of 6-cm coho salmon reared at different O2/N2 levels at 13.6°C with a 119% total gas pressure. 100% N 2), all increases in ratioof02toN2resulted in increased tolerance to the total gas saturation. A marked increase in tolerance to total gas pres- sure occurred between concentrations of 159%/ 109% and 173%/ 105% saturation of 02andN2 (Figure 3). Effect of Size and Stock of Fish on Mortality A number of tests were carried out in the water containing 114% Ogand 121% N2 to determine ef- fect of size and stock of fish on susceptibility to gas supersaturation (Table 2). Two groups of 3.8-cm coho from the Northwest Fisheries Center which had just started feeding were initially tested. One group of 96 fish reached LEgsin 22.9 days and LEgg after 30 days. The other group of 50 fish reached LE25 in 10.9 days. No further losses occurred until the 27th day. Loss at 30 days was 34%. Averages of the two groups placed LE25 at 16.9 days. Average loss at 30 days was 32%. Two groups of 50, 4.6-cm fish from the Quilcene National Fish Hatchery were also tested. These tests produced LE25of 15.1 and 18.3 days. LE50 was 40n 30- 20- 10- •LE50 ALE25 0 %N2-^0 % 02-^0 138 50 131 75 121 114 1 1 1 109 105 100 159 173 192 Figure 3. -Relationship between Oo/N^levels and time to death of 6-cm coho salmon fingerlings at 13.^C and total gas concen- tration of 119%. Table 2.— Time to death of groups of juvenile coho salmon of different body length and stock composition in 13.6° + 0.1°C water with gas concentrations of 114% 02and 121% N2. Average body length Time to death (days) stock of fish 25% of all fish 50% of all fish 3.8 cm (Seattle) 16.9 Not reached in 30 days 4.6 cm (Quilcene) 16.7 27.4 10 cm (Seattle) 2.1 2.6 10 cm (Quilcene) 2.9 4.2 reached in 24.7 and 30 days. Averages of the above placed LE25 at 16.7 and LE50 at 27.4 days. Five groups of fish (8, 12, 16, 16, and 16 in number), and approximately 10 cm long, from the Northwest Fisheries Center were then tested. The average for all groups gave an LE25 of 2.1 days and an LEgQof 2.6 days. Three groups of 12 fish each approximately 10 cm long from the Quilcene National Fish Hatchery were similarly tested. Averages were 2.9 days for LE25and 4.2 days for LEg^. These results indicate that the larger fingerlings approximately the same age are definitely more subject to harm from excess air in the water than the smaller fish. These data agree with those of Meekin and Turner (1974) and Dawley etal.' 'E. Dawley, B. Monk, M. Schiewe, and F. Ossiander. 1974. Sal- monid bioassay of supersaturation of dissolved gas in water. Northwest Fish Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, Wash., unpubl. manuscr. 917 Although the data are limited, there appears to be little difference between susceptibility of the Montlake and Quilcene stocks. Effect of Reduced O^ Concentration on Mortality Fish held in compartments in a trough utilize oxygen so that the water in compartment C (out- flow end) would have less oxygen than in com- partment A (inflow end). Compartment B in the central part of the trough would have 0, levels somewhere between those in A and C. Nitrogen levels in these compartments, however, were the same. To demonstrate the effect of reduced oxj'gen in relation to gas-bubble disease, 48 coho of 8.5 cm fork length were randomly distributed into compartments A, B, and C. Two additional replicates of the C compartment tests were run using 32 coho (8.5 cm) in each trial. These are listed as Ci and C2 in Table 3. Table 3. -Time to death of groups of juvenile coho salmon (about 8.5 cm long) in 13.6°C water with 121% N2 and different concen- trations of 02- Gas concentration Trough compart- (% sati jration) Time to death (days) Total 25% of 50% of ment 0, pressure all fish all fish A 113 119 2.5 3.3 B 110 118 3.6 5.3 C 105 117 3.8 5.3 c, 105 117 4.2 6.6 C2 105 117 5.4 6.6 Inspection of these data indicates that when 121% N2 is maintained, oxygen plays a more sig- nificant role above 110% than below 110%. Some of the data obtained when the oxygen-ni- trogen ratio tests were done also illustrated the effect of reduced oxygen on the mortality rate. This was apparent in the e.xperiment using 173% Ogand 105% Ng. At 173% 0 2 there were losses of 26 and 30% in 39 days, whereas slightly larger fish at the lower end of the troughs subjected to 169% O2 had losses of only 7% in 39 days. Pathology Generally the fish died suddenly in the higher nitrogen concentrations. Never was tissue damage or any progressive pathology demonstrated. The fish always seemed to die from gas embolism, re- stricting the flow of blood through the gills. When the nitrogen was near normal and the oxygen high, the fish were moribund for many days before succumbing. These fish had blebs in the mouth which interfered with feeding and caused emaciation. Summary Coho salmon fingerlings were subjected to a to- tal gas pressure of 119% at 13.6°C with the O2/N2 varying from 50%/138% to 229%/90%. The small fish (3.8 to 6 cm) were the most resistant and the larger fish (8 to 10 cm) the least resistant to gas- bubble disease at the gas concentrations used. A drastic decrease in lethal effect of individual ra- tios of Ogto Naoccurred between 159% 02/109% N2 and 173% 02/105% N2at the same total gas pres- sure (119%). Acknow^ledgments Facilities for this work were furnished by the Western Fish Disease Laboratory, U.S. Bureau of Sport Fisheries and Wildlife, Department of the Interior, Seattle, Wash. The work was started while the author was an employee of this organization. The information in the section on pathology was furnished by William T. Yasutake, pathologist. Western Fish Disease Laboratory. Literature Cited Anonymous. In press. Total dissolved gases (supersaturation). In David Gates et al., Water quality criteria, p. 135-139. Natl. Acad.Sci. Meeking, T. K., and B. K. Turner. 1974. Tolerance of salmonid eggs, juveniles, and squawfish to supersaturated nitrogen. Wash. Dep. Fish., Tech. Rep. 12:78-126. RUCKER, R. R. 1972. Gas-huhhle disease of salmonids: a critical review. U.S. Fish Wildl. Serv., Bur. Sport Fish. Wildl. Tech. Pap. 58, lip. Robert R. Rucker Northwest Fisheries Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, WA 9^112 918 AGE-LENGTH- WEIGHT AND DISTRIBUTION OF ALASKA PLAICE, ROCK SOLE, AND YELLOWFIN SOLE COLLECTED FROM THE SOUTHEASTERN BERING SEA IN 1961 Japanese fishing companies explored the trawl fish resources of the eastern Bering Sea in 1929 and 1931. They began commercial fish meal production in 1933 and continued until 1937; a frozen fish operation was initiated in 1940 but was interrupt- ed by World War II (Bourgois 1951).' In 1954, Japan resumed trawling in the eastern Bering Sea, again producing fish meal and frozen fish. The Soviet Union began sending bottom trawl fleets to the eastern Bering Sea in 1959, and combined an- nual catches of flatfishes (excluding Pacific halibut, HippoglossKf^ stoiolepsif^) by Japan-USSR rose to a peak in 1961 when it exceeded 600,000 metric tons (Fadeev 1965). In the years following 1961, eastern Bering Sea flatfish catches by Japan decreased and in the period 1963-1970 have averaged less than 20% of the 456,890 metric tons caught in 1961 (International North Pacific Fisheries Commission 1973). Comparable Soviet data are not available. Prior to intensive exploitation of eastern Bering Sea resources, there were two groups of surveys in which samples of flatfish were taken to assess the age-length structure of the population. One such series was conducted by the U.S. Fish and Wildlife Service in 1947-49 (King 1949; Ellson et al. 1950; Wigutoff and Carlson 1950). The other surveys were made 10 yr later by the Soviet Union (sum- marized by Moiseev 1965). Age-length determina- tions from flatfish samples collected in 1949 were reported by Mosher (1954); the Soviet collections of 1957-60 were studied by Fadeev (1963), Mineva (1964), and Shubnikov and Lisovenko (1964). In July-August 1961, personnel of the Bureau of Commercial Fisheries (now National Marine Fisheries Service) conducted a trawl survey of the southeastern Bering Sea. This survey, although conducted principally to estimate the abundance of Alaska king crab, Paralithodes spp., provided an opportunity to sample several flatfish species. The purpose of the present report is to present biological information on the distribution, age. 'Bourgois reported that Alaska pollock was the principal species taken by these Japanese fisheries. However, Alverson et al. (1964) pointed out that the areas of the eastern Bering Sea fished by Japanese trawlers from 1933 to 1941 were the same locations as post-World War II flounder fisheries, so there is reason to doubt the complete accuracy of Bourgois' information as to species. length, and weight by sex for three commercially important species of Bering Sea flatfish: yellowfin sole, Limanda a^tpera (Pallas); rock sole, Lepidop- setfa bilineata (Ayres); and Alaska plaice, Pleuronectes quadritiiberculatus Pallas. Methods and Materials Sample Collection Otter trawl hauls of 1-h duration were made at 51 predesignated stations 20 nautical miles (37 km) apart (Figure 1). The trawling speed of the vessel was about 2.5 knots (4.6 km/h). The trawl was a 400-mesh, Eastern type, as described by Greenwood (1958). A 1.5-inch (3.8 cm) mesh liner was laced into the cod end to retain small specimens which might otherwise pass through the 3-inch (7.6 cm) meshes in that part of the trawl. At the completion of each haul, the catch was examined, and the weight of each major com- ponent was estimated. At five of the stations where one or more of the target species was abundant, samples of yellowfin sole, rock sole, and Alaska plaice was selected for length-weight-age determination. Specimens were measured to the nearest centimeter to obtain a representation of individuals throughout the available length range. Each fish to be retained was then frozen in- dividually in a plastic bag which was sealed to prevent shrinkage and weight loss through dehydration. At the laboratory, 3 mo after collection, the specimens were thawed, the total length (snout to longest rays of the tail fin) was measured, the weight recorded to the nearest gram, and sex de- termined from an examination of the gonads. BERING SE« =8> Figure 1.— Sampling station pattern and location of sample collections for 7 July-4 August 1961 survey. 919 Both otoliths were removed and placed in 95% ethyl alcohol .- Age Determination Procedure Studies by Hatanaka (1968) of yellowfin sole from the southeastern Bering Sea indicate that the translucent zone of an otolith is formed once a year during the winter months. In our readings on otoliths from all species, each translucent zone was considered an annular mark. The outermost edge of each otolith was also translucent except in the younger fish where there was evidence of some beginnings of opaque summer growth. Thus, the ages recorded are the number of translucent rings starting with the smallest observable and includ- ing the outermost. For example, a fish captured in July 1961 with 10 rings on its otolith was con- sidered to have been spawned and hatched in 1951. Fadeev (1965), through gonad examination, stated that Bering Sea yellowfin sole spawn in June- August, Alaska plaice in April-June, and rock sole in February-May. Shubnikov and Lisovenko (1964), who reported that rock sole in Bristol Bay spawn in March-June, are in general agreement with Fadeev. For reading, the otoliths were immersed in water in a petri dish with a black mat background and examined at 10 x under reflected light with a dissecting microscope. Both otoliths were con- sidered in age determination, but when a discrepancy occurred between the two otoliths, a decision was based on the eyed-side (right) otolith. In situations where the annular rings were not clear, the otoliths were ground on fine, water soaked, carborundum paper. In most samples grinding improved interpretation of annular rings, but the grinding of rock sole otoliths often exposed additional opaque and translucent zones to further confuse the readers. Consistency of Age Determinations Without reference to fish size, otoliths were in- terpreted by each author. A third, independent interpretation was made by an experienced otolith reader at the Northwest Fisheries Center. The observed ages, as agreed upon between the two authors, were compared with ages determined by the experienced reader. Initial agreement between authors and reader was 76% for yellowfin sole, 72% for Alaska plaice, and 85% for rock sole. Disagreements were not confined to a particular age class; only 3% differed by more than 1 yr, and these differences were equally negative and posi- tive. The similarity of results by authors and reader suggests that the method used produced consistent age-growth data. Otolith interpreta- tions not in agreement between the authors and the reader were reread and a joint decision was made on the most probable age of the fish. Results and Discussion Distribution Yellowfin sole is the most abundant flatfish taken in the eastern Bering Sea. Alaska plaice (33% by weight of yellowfin sole caught) was usually encountered together with yellowfin sole and share a similar distribution within the sampling area (Figure 2). Fadeev (1970) and Maeda et al. (1967) note that yellowfin sole con- centrate in the colder waters of Bristol Bay during the spring and summer months. In July of 1961, a tongue of cold water extended into the sampling area and the greatest concentrations of yellowfin sole and Alaska plaice were taken at bottom temperatures of 3°C or less. The distribution of rock sole within the sampling area (28% by weight of yellowfin sole catch) was spotty with the denser concentrations occurring toward the eastern edge of the area inhabited by yellowfin sole and Alaska plaice. The 1961 observations are compatible with the conten- tion of Shubnikov and Lisovenko (1964) that rock sole disperse during the summer into shallower water than they occupy in winter and spring. Age-Length Observations The age-length-weight composition for the three southeastern Bering Sea flatfish species sampled in 1961 are given in Tables 1-3.' It is difl^cult to compare the 1961 data with any earlier reports except in a generalized manner since in only one instance (Pruter and Alverson 1962) is there a determination of age by sex. The data presented in Tables 1-3, and also in studies of eas- -All otoliths from the 1961 collection are inpermanent storage at the Northwest Fisheries Center, Seattle, Wash. 'Individual age determinations and related lengths and weights, by sex, are available upon request from the Northwest Fisheries Center, National Marine Fisheries Service, NOAA, Seattle, Wash. 920 BERING SEt ^ Dam /^ ^ 'Release location ♦'Recovery site Figure 1. -Columbia and Snake rivers, showing release and recovery sites of migrating chinook salmon and steelhead trout. 925 Table 1.— Number of transported and non transported (control) juvenile chinook salmon and steelhead trout that were marked and released, 1969-70 (figures adjusted for tag loss)'. Release site (experimental group of fish) 1969 1970 Chinook Steelhead Chinook Steelhead Ice Harbor Dam (control) John Day Dam (transported) Bonneville Dam (transported) Total 24,217 14,782 13,529 52,528 25,313 20,430 45,743 8,624 10,159 10,173 28,956 18,347 20,935 31,282 70,564 'Initial tag loss was determined for the control releases by ex- amination of juveniles after recovery at Ice Harbor Dam, 1969- 70; tag loss for the test groups were determined by fish held at release sites after transport. migration. These included returns to the sport, commercial, and Indian fisheries in the Lower Columbia River; to Ice Harbor and Little Goose dams on the Lower Snake River; to Rapid River and Dworshak hatcheries in Idaho; and to the spawning grounds. Most of the tagged adults were captured at Ice Harbor Dam or Little Goose Dam. At Ice Harbor Dam about 809c of the run of adult fish ascend the south ladder enroute to the spawning grounds. At Little Goose Dam all fish must ascend the single ladder installed there. Adults were recovered at Ice Hai'bor Dam by a detector-separator device that intercepted tagged salmon and trout (Durkin et al. 1969). At Little Goose Dam, recoveries were made by an improved but similar detector ap- paratus. A major modification of the system included a Denil-type fishway instead of the pool- and-overfall ladder used at Ice Harbor Dam.- Improvements incorporated in the facility at Lit- tle Goose Dam increased detection efficiency markedly in 1970. Results Returns of Adult Spring and Summer Chinook to Ice Harbor and Little Goose Dams Numbers of returning adult salmon successfully detected, separated, and identified at the adult separator are listed in Table 2. It should be stressed that the observed return of adults represents only a fraction of the total return of marked fish to Ice Harbor and Little Goose dams. The observed tally is low for the following reasons: 1) approximately 20'^i of the adult run at Ice Har- bor Dam passed up the right bank (north) fishway which did not have a tag detection device; 2) at Little Goose Dam, the barrier gates at the en- trance to the automatic separator were open at night (2100-0500) allowing some adults to pass undetected; 3) some tag loss had occurred between tagging and recovery as adults; 4) the tag detec- tion system was less than W07c eflScient; 5) -Slatick E. 1974. Laboratory evaluation of a Denil-type steep- pass fishway with various entrance and e.xit conditions for pas- sage of adult salmonids and American shad. Unpubl. manuscr. Natl. Mar. Fish. Serv., NOAA, Pasco, Wash. Table 2.— Percentage of transported and nontransported (control) juvenile chinook salmon (released in 1969 and 1970) that were recaptured as adults at Ice Harbor and Little Goose dams, 1 April through 18 August 1971-73. Release site and Number Number (in parentheses) of recaptured Percentage reti jrn as adults experimental juveniles as group of fish released' ad lults Observed Estimated^ 1969: Ice Harbor Dam (control) 24,217 47 0.194 0.497 John Day Dam (transported) 14,782 19 0.129 0.356 Bonneville Dam (transported) 13,529 33 0.244 0.581 1970: Ice Harbor Dam (control) 8,624 17 0.197 0.323 John Day Dam (transported) 10,159 7 0.069 0.113 Bonneville Dam (transported) 10,173 29 0.286 0.467 'Adjusted for initial tag loss. 'Based on a comparison of the known recovery of fish with magnetized wire tags at Ice Harbor and Little Goose dams and the subsequent recovery of these and other marked fish at a hatchery upstream. Returning fish iden- tified at the dam were marked with dart tags and released to continue their migration upstream. Numbers of dart-tagged fish arriving at Rapid River Hatchery were compared with the recovery of other wire-tagged fish not previously detected and identified at Ice Harbor and Little Goose dams. 926 presumably some adults could have passed up- stream through the navigation locks at Ice Harbor and Little Goose dams. Throughout this section of the report, percent- age figures are given which indicate either an increase or decrease in survival of groups of juveniles transported downstream in comparison to control groups not transported but released near the collection point. Some of the increases are statistically significant, some are not; generally those that are significant are indicated. We present the data even though some of it is not statistically significant because it parallels earlier data reported by Ebel et al. (1973). The combined adult returns— of spring- and summer-run chinook salmon from juveniles transported from Ice Harbor Dam and, sub- sequently, released at Bonneville Dam— were greater than adult returns from control releases made at Ice Harbor Dam. The combined transpor- tation benefit (Table 2) for spring- and summer- run chinook salmon released in 1969 was 27%; in 1970, 47%. An analysis of comparative survival to adults for spring- and summer-run chinook salmon by year of transport are presented in Table 3. The transportation benefit indicated for juveniles released in 1969 was 27% for spring-run chinook salmon and 29% for summer-run chinook salmon. Benefits from the 1970 release were 40% for spring-run chinook salmon and 57% for summer- run chinook salmon. Combined spring and summer adult returns from the John Day release were 34% less in 1969 and 65% less in 1970 than returns from the con- trols. Although the lower adult returns from juvenile releases at John Day are unexplained at this time, it is possible that the cumulative stress from collection, handling, and hauling combined with the stress from having to pass two dams (The Dalles Dam and Bonneville Dam) may have been detrimental for fish released at this site. Returns of Adult Steelhead Trout to Ice Harbor and Little Goose Dams Table 4 lists the returns of adult steelhead trout (released as juveniles in 1969-70) that were suc- cessfully detected, separated, and identified at the automatic separator at Ice Harbor and Little Goose dams. We identified 148 adult steelhead trout from those released in 1969. Of these, 46 were from the control release and 102 from the John Day transport release, which give a transportation benefit of 174%-a significant (X^ = 34.370; df = 1) increase. Adult steelhead trout returns from the 1970 juvenile releases totaled 324 fish. Of these, 71 were from the control release, 75 from the John Day transport release, and 178 from the Bonneville transport release. The transportation benefit from the Bonneville release was 47% (X- = 7.315; df = 1); however, no benefit was derived from transport of juveniles to the John Day release site (adult re- Table 3. -Comparison between transported (released at Bonneville and John Day dams in 1969-70) and nontransported (control) groups of chinook salmon based on numbers of transported and nontransported juvenile fish recaptured as adults at Ice Harbor and Little Goose dams, 1971-73. Release site (of Juveniles) and seasonal race of salmon' No adu . of salmon Its at Ice Hi Goose recaptured as arbor and Little dams2 Transportation benefit or deficit (-) (Percent) Transported Nontransported 1969 Below Bonneville Dam: Spring Chinook salmon 38 30 27 Summer chinook salmon 22 17 29 Below John Day Dam: Spring chinook salmon 23 30 -23 Summer chinook salmon 8 17 -53 1970 Below Bonneville Dam: Spring chinook salmon 14 10 40 Summer chinook salmon 11 7 57 Below John Day Dam: Spring chinook salmon 4 10 -60 Summer chinook salmon 2 7 -71 'Seasonal races of chinook salmon in the Columbia River system are classified as spring, summer, or fall chinook depending on the time of year that the adults enter the river to spawn. ^Numbers recaptured adjusted in relation to numbers released (Table 1). 927 Table 4.-Percentage return and benefit or deficit (-) of transported to nontransported (control) juvenile steelhead trout (released in 1969-70) that were recaptured as adults at Ice Harbor and Little Goose dams, 1970-73. Release site and (in parentheses) experimental group offish Number of juveniles One Number recaptured as adults Two Three Percentage return as adults Percentage transported to control benefit or released' ocean ocean ocean Total Observed Estimated^ deficit (-) 1969: Ice Harbor Dam (control) John Day Dam (transported) 1970: Ice Harbor Dam (control) John Day Dam (transported) Bonneville Dam (transported) 25,313 43 3 0 46 0.182 0.792 20,430 76 25 1 102 0.499 1.600 18,347 12 58 1 71 0.387 0.729 20,935 8 66 1 75 0.358 0.610 31,282 14 162 2 178 0.569 0.924 174 -7 47 'Adjusted for initial tag loss. 2Based on comparison of the known recovery of fish with magnetized wire tags at Little Goose Dam and the subsequent recovery of these and other marked fish at a hatchery unstream from Little Goose Dam. Returning fish identified at the dam were marked with dart and law tags and released to continue their migration up- stream. Numbers of externally-tagged fish arriving at Dworshak Hatchery were compared with the recovery of other wire-tagged fish not previously detected and identified at Little Goose Dam. turns from this release were 7% less than returns from controls). Recovery of Marked Chinook Salmon in Commercial and Sport Fisheries Although only 43 adult chinook salmon (Table 5) were recovered in the commercial and sport fisheries from juvenile releases in 1969, returns indicate a definite benefit from transportation. The benefit of transported fish (John Day- Bonneville releases combined) was 19%. It was not possible to distinguish between re- turns of adults to the fishery from juvenile releases at Bonneville and John Day because of the loss of the identifying brands. Brands which would have enabled identification by release site were obliterated by gillnet abrasion. Transported and control groups of juveniles could be distinguished as adults by magnetic tags, but only two codes Table 5.-Comparison between transported and nontransported groups of chinook salmon based on numbers of transported and nontransported juvenile fish (released in 1969) that were cap- tured as adults by commercial and sport fisheries in the lower Columbia River, February through August 1971 and 1972. No. of salmon recaptured as adults Transported Nontransported Location of fisheries Upstream from Bonneville Dam (Indian fishery) 8 Downstream from Bonneville Dam 17 Total 25 4 14 18 were used— one for the controls and one for the transported fish (Bonneville and John Day com- bined). However, if the percentage of adult re- turns obtained at Ice Harbor and Little Goose dams— where brands of fish returning from releases at Bonneville and John Day were visible-is applied to the total returns of adults as obtained in the commercial fishery, the benefit from transporting juveniles becomes 59% for chinook salmon transported to Bonneville Dam. Adult recoveries in the lower river commercial and sport fisheries from juvenile chinook salmon released in 1970 were insufficient (seven trans- ported and eight control fish) for analysis of transport to control return ratios. Returns of Adult Chinook Salmon to Spawning Grounds Spawning ground surveys (Figure 2) and examination of tagged adult chinook salmon at Rapid River Hatchery near Riggins, Idaho, provided further information concerning benefits at their "home" destination from transport of juvenile spring- and summer-run chinook salmon. In 1971, 12 tagged adult fish (from the 1969 juvenile release) were recovered from the Rapid River Hatchery; an additional 15 were from sport fishermen and spawning ground surveys. Of the total, 15 adults were from the transported groups and 12 from the control group. By adjusting from the ratio of John Day to Bonneville adult returns, we estimated that 12 of the 15 transported fish 928 O Control 69 • Transport '69 D Control 70 ■ Transport'70 Figure 2.— Location of recoveries of tagged adult chinook salmon returning to spawning grounds from 1969-70 experiments. were from the group released at Bonneville Dam. The transport benefit for the groups of juveniles released at Bonneville becomes 78% when com- puted on the basis of the number of juveniles released per group. Too few tagged adult chinook (five Bonneville transports and one control) from the 1970 juvenile releases were collected in 1972 from all sources to make conclusions regarding the effect of trans- portation. Discussion Results from this study, which was a continua- tion of a study begun by Ebel et al. (1973), corroborated earlier findings, i.e., homing of adults after transportation downstream as juveniles was not seriously affected and survival was increased. Throughout this study, we found no evidence of straying among adults returning from the experimental releases. All comparisons between the adult returns from transported and control groups of juvenile chinook salmon and steelhead trout indicated that survival was definitely increased by transporting juvenile fish to a release site downstream from Bonneville Dam. We have been particularly concerned with how the percentage return from these experiments might compare with that of unhandled or undis- turbed juvenile migrants. Some insight into this matter is shown by a comparison between es- timated adult returns from juveniles marked and released as controls and returns of unhandled adult fish to Rapid River Hatchery in Idaho (Table 2); the data indicate that survival of chinook salm- on released in our 1968 experiment was greater than that indicated for salmon returning to the Rapid River Hatchery. Adult returns from con- trols released in 1969 were comparable to hatchery returns, but returns from those released in 1970 were lower than returns to the hatchery. It is assumed that some stress was placed on juveniles in the collection, handling, marking, and transport processes. These cumulative stresses were not outwardly apparent in the physical con- dition of the juvenile smolts at the time of handling, but differences in survival of returning adults indicated that condition of the fish at the time of marking must have varied among years. Our collection methods were changed in 1970 by addition of a fish pump; this added a pumping stress to our fish handling process. Although Park and Farr (1972) indicate no immediate mortality or observed stresses due to pumping from the 929 facility, it is possible there could have been sig- nificant delayed effects. The effect of pumping on juvenile chinook salmon and steelhead trout— when added to other cumulative stresses associated with handling in our transport process —is indicated by the lower percentage of adult returns from control releases of juveniles in 1970. Although smaller numbers of juvenile chinook were released in 1969-70 than in 1968 and a correspondingly small number of adults returned, we believe that the lower percentage of returning adults does indicate that stress factors due to handling were higher in 1970 than in 1969. The addition of two dams-Lower Monumental and Little Goose-placed in operation in 1969 and 1970, upstream from Ice Harbor Dam, must also be considered. Fish had to pass through Lower Monumental Reservoir and Dam in 1969 before being collected at Ice Harbor Dam. In 1970, they had to pass through both reservoirs and dams before being collected. Supersaturation of dis- solved nitrogen also became a problem between Little Goose and Ice Harbor dams at this time. Turbines from both Lower Monumental and Little Goose dams were not scheduled for installation until after the spring freshet and as a result large volumes of water had to be passed over the spill- ways, causing dissolved gas concentrations to be high; a large percentage of the fish arriving at Ice Harbor Dam e.xhibited obvious signs of gas bubble disease. If we use the percentage adult returns in rela- tion to juveniles released at the Rapid River Hatchery in Idaho as an indicator of the rate of return of naturally migrating chinook salmon and we compare our percentage return figures, we find our estimate of return of controls was 4.3%' in 1968-much higher than the 0.48% adult return recorded for Rapid River Hatchery.' The estimat- ed control return of 0.497% for the 1969 outmigra- tion is comparable to the 0.493% return to Rapid River Hatchery, but estimated returns from con- trols released in 1970 dropped to 0.323% whereas the return to the Rapid River Hatchery was 0.477%. Thus, the stresses placed on juvenile fish prior to collection, in addition to those involved in the handling process, conceivably were in- strumental in causing the lower return of adults from the 1969-70 e.xperiments. When we examine adult returns from juvenile control releases of steelhead trout, we find that the percentage return from control releases of steelhead trout in 1969-70 were much greater than for comparable juvenile releases of chinook salm- on. This indicates that the ability of steelhead trout to withstand the cumulative effects of stress is greater than that of chinook salmon. Using the adult return percentage of steelhead trout to the Dworshak Hatchery from juvenile migrants released at that site in 1970 as a base indicator of the adult return of naturally migrat- ing steelhead trout to Idaho streams, we find that our estimated adult returns from control releases to Ice Harbor and Little Goose dams of 0.792 and 0.729% (in 1969 and 1970, respectively) were somewhat greater than the 0.682%' return to Dworshak Hatchery. When our adult control re- turns are adjusted for the upriver sport catch on steelhead trout, our revised return (0.713% from the juvenile control releases in 1969) was com- parable to the 0.682% return to Dworshak Hatchery. The return from the 1970 (control) release of 0.598% was, however, less than the hatchery return of 0.682%. Based on the foregoing rationale, we believe that our control releases of juvenile chinook salm- on and steelhead trout in 1969 returned as adults at rates comparable to those of natural migrat- ing salmonids and that benefits on survival to adults indicated for our transported salmon and steelhead trout represent real increases. Studies to further define stress problems as- sociated with diversion, collection, and handling of naturally migrating juveniles are currently un- derway. To maximize the effectiveness of a collection and transportation system, stresses from all sources must be minimized. Conclusion The homing of adult fish, captured during their seaward migration as juveniles and transported downstream (from Ice Harbor Dam to Bonneville Dam), was not reduced by the transport operation. Although numbers of returning adults were small, comparisons of returns of transported fish versus control fish to Ice Harbor Dam, the spawning grounds, and hatcheries in Idaho indicated that they "homed" satisfactorily. No evidence of straying of transported fish was observed in our surveys. Ters. commun. Evan Parrish, Hatchery Manager, Rapid River Hatchery, Riggins, Idaho. 'Pers. commun. Einer Wold, Hatchery Pathologist, Dworshak Hatchery, Ahsahka, Idaho. 930 Adult returns indicate a definite benefit is achieved from transporting juvenile chinook salm- on and steelhead trout from a collector dam (Ice Harbor) to a release site below Bonneville Dam. Transport benefits were lower than reported from releases made in 1968, but a benefit of 27-47% was still indicated. No steelhead trout were released at Bonneville Dam in 1969, but a 47% benefit was realized from transportation of juveniles to that site in 1970. Data from returning adults indicate that in general the John Day release site was a poor one. In 1969, however, returns from juvenile steelhead trout releases there were 174% greater than con- trols. The reduced transport benefit for our John Day release can probably be best explained by the fact that juveniles must still pass over The Dalles and Bonneville dams before entering the ocean. These further stresses probably nullify any initial transport benefit. The rate of adult return from those juvenile fish transported in 1969 was better than the adult re- turns from those transported in 1970. Data sug- gest that stresses to juveniles encountered prior to collection at Ice Harbor and the changed handling procedures in 1970 were a factor. Literature Cited Bentley, W. W., and H. L. Raymond. 1969. Passage of juvenile fish through orifices in gatewells of turbine intakes at McNary Dam. Trans. Am. Fish. See. 98:723-727. DuRKiN, J. T., W. J. Ebel, and J. R. Smith. 1969. A device to detect magnetized wire tags in migrating adult coho salmon. J. Fish. Res. Board Can. 26:3083-3088. 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. Jefferts, K. B., p. K. Bergman, and H. F. Fiscus. 1963. A coded wire identification system for macro-or- ganisms. Nature (Lond.) 198:460-462. MiGHELL, J. L. 1969. Rapid cold-branding of salmon and trout with liquid nitrogen. J. Fish. Res. Board Can. 26:2765-2769. Park, D. L., and W. E. Farr. 1972. Collection of juvenile salmon and steelhead trout passing through orifices in gatewells of turbine intakes at Ice Harbor Dam. Trans. Am. Fish. Soc. 101:.381-384. Emil Slatick Donn L. Park VFesley J. Ebel Northiret^t Fisheries Center National Marine Fisheries Service, NOAA 2725 Montlake Bouleva rd East Seattle. WAD8112 COMPARATIVE VULNERABILITY OF FRY OF PACIFIC SALMON AND STEELHEAD TROUT TO PREDATION BY TORRENT SCULPIN IN STREAM AQUARIA Predation on fry of salmon and trout by sculpin, Cottuii spp., is intense in certain situations (Hunter 1959; Sheridan and Meehan 1962; Patten 1962, 1971a, 1972) or of little consequence in others (Ricker 1941; Patten 1971a, 1972). Variation in in- tensity may be related to such important causes as the environment or to specific differences of the predators or prey. In this paper I report the comparative ability of steelhead trout, Salmo gairdneri, and of five species of Pacific salmon, Oncorhynchus spp., to avoid predation by torrent sculpin, C. rhofheus, in a fixed environment-stream aquaria. The vulnerability of a species of salmon or steelhead trout, as determined from this study, is related to known information on the duration of residency and behavior of a species in streams. These results help in the assessment of natural causes of mor- tality that affect the productivity of salmon and steelhead trout. The study was conducted in stream aquaria adjacent to Cedar River near Ravensdale, Wash., in 1966. Facilities and Procedures The facilities consisted of two stream aquaria and eight holding aquaria that received water from the Cedar River (more fully described by Patten 1971b). Two stream aquaria used for tests of predation were 2.4 m long and 0.6 m wide and high; water depth ranged from 2 to 18 cm depend- ing on bottom contour. The eight holding aquaria used in the study (to incubate the eggs and main- tain the young fish before tests) were 34 cm wide by 41 cm long by 36 cm high; water depth was 18 cm. Water from the Cedar River was taken at a low dam and supplied by gravity flow to the head box and then to the stream aquaria. Each aquarium had a continuous flow. The water was usually clear, and temperatures recorded at 0800 ranged from 5° to 10°C during the course of the study. The experimental procedure exposed salmon or trout fry to predation by torrent sculpin under pseudo-natural but controlled conditions. Torrent sculpin were collected by electrofishing in Soos Creek, Wash.; the salmon and steelhead trout fry were reared from eggs to insure that they had no previous experience with predators. 931 Eggs from pink salmon, O. gorbuscha, chum salmon, 0. keta, sockeye salmon, 0. nerka, fall Chinook salmon, O. tsliairijfscha, coho salmon, O. kisutch, and from steelhead trout of Puget Sound stocks were placed in holding aquaria and covered with coarse gravel. The salmonid fry were subjected to predation tests as soon as the yolk sacs were absorbed. Since the time of emergence from the gravel by the fry of these six species varies, the tests extended from March to June, during which period water temperatures (Table 1) and day lengths differed. The salmon and trout fry were not fed but could be seen mouthing particles entering the holding aquaria. I assume grow^th of fry negligible and size differences to be fixed by the species and race used. Observations of viability and vigor of fry in the holding aquaria were made before, during, and after testing as a standard of comparison for test fish. Samples of the salmon and trout fry were measured in millimeters from snout to fork of tail (Table 1); their volumes were determined by displacement in a graduated tube. Sculpins were measured in millimeters from snout to end of tail (Table 1). Twenty sculpins were placed in holding aquaria without food the first day of the experiment. On the second day 10 fry of one species were placed in each stream aquarium and on the third day, 10 sculpins were quietly introduced at the down- stream end of each stream aquarium. On the fifth day the fry surviving after 48 h were counted; then both predators and prey were removed. These subjects were not used again. Two to seven replicate tests were made for each species of salmon or trout (Table 1). Comparative Survival of Salmon and Trout Against Predation The positions and activities of the salmon, trout, and sculpins in the stream aquaria are first described because these varied between species, affecting predator-prey interrelations. The following sections report on the viability and vigor of fry and on the survival rates of the species of salmon and trout. The positions and activities of a species of salm- on or trout during daylight tests varied. Fish in the stream aquaria maintained positions and ap- parently fed; Chinook and coho salmon displayed intraspecific aggression, indicating accommoda- tion to the enclosure. All species were observed in the deepest areas of the stream aquaria where they distributed themselves vertically 1 cm from the bottom to the water surface. Most of the steelhead trout fry and some pink and chum salm- on fry hid under rocks, but this behavior was seldom exhibited by the other salmon species ex- cept for short periods when they were frightened. Torrent sculpin typically spaced themselves through the deeper parts of the stream aquaria. They were distributed through its length with the greatest number at the upstream end. They were inactive and curled around large rocks or partially buried themselves in areas with soft bottoms. The concealment of the sculpins was so complete that I often had to search for as long as 20 min to remove all of them after an experiment. After the sculpins were placed in the stream aquaria, the salmon and trout fry, on recovering from the disturbance, modified their vertical dis- tribution. Salmon fry reacted to an active sculpin by moving away laterally and upward. In the presence of sculpins all salmon fry increased their distance from the bottom to about 5 cm. Steelhead trout fry, that usually hid under rocks when un- disturbed, moved off the bottom and maintained positions near the water surface when sculpins were present. Behavior of the steelhead trout fry was apparently more disturbed by sculpins than was that of the salmon fry. Sculpins rarely stalked the fry in bright daylight but waited immobile for them to come Table 1.— Survival of salmon and trout fry subjected to predation in 1966 by torrent sculpin, Cottus rhotheus. Water Test (prey) fish Me an length Number Number of survivors Date of tempi Length range of predator of Percentage Species testing (°C) Number (mm) (mm) tests Total Rs inge/test survival Chinook salmon 3-25 to 4-15 6.2 60 38-42 92.0 6 31 1-8 51.7 Chum salmon 5-16 to 5-23 8.7 60 35-38 88.7 6 3 1-2 5.0 Coho salmon 3-25 to 4-15 6.2 70 36-39 91.7 7 53 6-9 75.7 Pink salmon 3- 2 to 3-11 6.2 60 35-37 92.5 6 1 0-1 1.7 Sockeye salmon 4-15 to 4-19 6.1 20 41-43 97.5 2 4 1-3 20.0 Steelhead trout 5- 4 to 5-13 8.9 60 29-31 91.4 6 14 0-7 23.3 'Average temperature (°C) at 0800 for 2 days of test. 932 near. Then they made short, quick lunges at the prey. Little predation occurred during the day and I never ol)served a sculpin catching a salmon. Stocks of fry used for testing appeared normal, healthy, and vigorous. A reserve of fry of a species was maintained in holding aquaria during and after testing without mortality— in fact dead or inferior fry were never observed over 2 yr. Pretest salmonids held in stream aquaria often main- tained positions in the faster moving water. In- dividuals that may have been inferior as indicated by use of slow water shallows, by swimming at the downstream end, or by impingement on the outlet screen were never observed. Results showed variation between rates of predation on a prey species, size of prey species, and on temperature and length of daylight during testing. Predation by torrent sculpin was least on coho and chinook salmon, intermediate on sockeye salmon and steelhead trout, while practically complete on pink and chum salmon (Table 1). Chi- square analysis showed significant differences between all species except for 2 of the 10 com- binations tested: sockeye salmon-steelhead trout and pink-chum salmon. The number of survivors per test varied considerably for the chinook salm- on and steelhead trout. Steelhead trout were relatively deep bodied but shorter than salmon fry and among the salmon, chum and pink were thin bodied (Table 1 shows lengths; body volume de- terminations indicated chinook, coho, and sockeye salmon had as much as twice the displacement of the other species). Testing of chum salmon and of steelhead trout was a month or two later in the spring when temperatures were higher (Table 1) and duration of daylight was longer than for other species. Innate Predator Avoidance of Species Differences in rates of predation on the study species are not well explained by observed differences in behavior, size of prey, ambient con- ditions, or predator related effects but may be due to innate behavior after emergence of fry from the gravel. The only species with greatly divergent behavior in the stream aquaria was the steelhead trout. Remaining near the water surface during day effectively removes them from the influence of sculpin predators; however, they may settle to the substrate at night, a time when sculpins are more effective predators (Patten 1971b). The larger prey species, those having the lon- gest body lengths and being relatively deep bodied, were not always those with the higher survival. Chinook, coho, and sockeye salmon were the largest. Chinook and coho had the highest sur- vival but the sockeye salmon, the largest prey, had survival similar to the steelhead trout, the smallest prey. Chum and pink salmon were slim and as long as coho and longer than steelhead trout, but their survival was lowest of the species studied. If size of prey or satiation of predators from greater food volumes influenced rates of predation, these fac- tors were apparently less important than other effects on a species level. Length of day or temperature had no apparent effect on rate of predation. Sculpins are most predaceous on salmon at times of marginal light intensity (Patten 1971b), which might suggest they are more serious predators at times of shorter day lengths. Trends between intensity of sculpin predation on fry and temperatures observed dur- ing this and other studies have never been ob- served. The data show strong interspecific variations of the study species in vulnerability to predation by the torrent sculpin. I suspect a difference in innate behavior exists; some species are better able to evade predation. Furthermore, the early life his- tory and behavior of the study species may be linked to their predator avoidance abilities. Chum, pink, and sockeye salmon quickly migrate from a stream environment to the sea or a lake where they form schools (Mason 1974, has observed chum salmon forming loose aggregations in estuaries). Schooling may aid these species in avoiding predation (Shelbourn 1966). Chinook and coho salm- on and steelhead trout on the average form loose aggregations in streams during a period of growth before migrating to the sea. Forming loose aggregations would increase feeding opportuni- ties in streams. Density of predators may be high in this situation (Patten 1971a) and survival is at- tained by a well-developed avoidance response for chinook and coho salmon. Steelhead trout fry had a comparatively high mortality among stream resident species that may have been related in part to their behavior during tests, to their small size or an inferior predator avoidance response. Their survival, at least during the early fry stage, may be increased by unavailability through selection of a protective habitat. Hartman (1965) described the microhabi- tat of recently emerged steelhead trout and coho salmon in the Chilliwack River, British Columbia, 933 as shallows at stream edges or in close proximity to physical objects. Recently emerged steelhead trout fry, observed adjacent to my study area in the Cedar River in 1965-66, were rarely found along sandy shore areas but were commonly seen among rocks at depths of 1 to 5 cm -when dis- turbed they hid under the rocks. The use of ex- treme shallows by steelhead trout fry may in part be an innate response to predators since this type of habitat in streams is relatively barren of other fish. Literature Cited Hartman, G.F. 1965. The role of behavior in the ecology and interaction of underyearling coho salmon (Oncorhynchus kisufch) and steelhead trout {Salmo gairdneri). J. Fish. Res. Board Can. 22:1035-1081. Hunter, J. G. 1959. Survival and production of pink and chum salmon in a coastal stream. J. Fish. Res. Board Can. 16:835-886. Mason, J. C. 1974. Behavioral ecology of chum salmon fry (Oncorhynchus kefa) in a small estuary. J. Fish. Res. Board Can. 31:8:3-92. Patten, B. G. 1962. Cottid predation upon salmon fry in a Washington stream. Trans. Am. Fish. Soc. 91:427-429. 1971a. Predation by sculpins on fall chinook salmon, On- corhynchus ffshawytscha. fry of hatchery origin. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Spec. Sci. Rep. Fish. 621, 14 p. 1971b. Increased predation by the torrent sculpin, Cottus rhotheus. on coho salmon fry, Oncorhynchus kisufch, during moonlight nights. J. Fish. Res. Board Can. 28:1352-1354. 1972. Predation, particularly by sculpins, on salmon fry in fresh waters of Washington. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Data Rep. 71, 21 p. on 1 microfiche. RiCKER, W. E. 1941. The consumption of young sockeye salmon by predaceous fi.sh. J. Fish. Res. Board Can. 5:293-313. Shelboukn, .1. E. 1966. Influence of temperature, salinity, and photoperiod on the aggregations of chum salmon fry (Oncorhynchus keta). i. Fish. Res. Board Can. 23:293-304. Sheridan, W. L., and W. R. Meehan. 1962. Rehabilitation of Big Kitoi outlet stream, Afognak Island, .-Maska. Alaska Dep. Fish Game, Div. Biol. Res. Inf.Leafl.il. 13 p. Benjamin G. Patten Northtveat Fisheries Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, WA 9S112 HERITABLE RESISTANCE TO GAS BUBBLE DISEASE IN FALL CHINOOK SALMON, OSCORHYSCHUS TSHA W'YTSCHA' Construction of a series of dams on the Columbia River has resulted in air-supersaturation of the river during spring and early summer. Air-super- saturation is caused by the entrainment of air in water at depths as great as about 15 m in the plunge basins of the spillways below each dam. The level of air-supersaturation varies according to the amount of water-flow over the spillways (Ebel 1969). Supersaturation levels which are known to be fatal to salmonid fishes (Rucker and Hodgeboom 1953; Westgard 1964; Ebel 1969; and Blahm et al. 1975) are often sustained in the Columbia River from April through July, the period when many juvenile salmonids emigrate to the ocean. Salmonids vary greatly in their tolerance for supersaturation (Ebel 1969). If a portion of this variability is related to additive genetic factors, an increase in the average tolerance of salmon populations to air-supersaturation can be expected as a result of selection. The purpose of this study was to estimate the influence of genetic factors on resistance to gas bubble disease for fall chinook salmon, Oncorhynchus tshawytscha. Specifically, the objectives were: 1) To determine the heri- tability of resistance to death from gas bubble disease for a stock of Columbia River fall chinook salmon, and 2) to determine the inherent level of resistance to gas bubble disease for several fall chinook salmon stocks. Methods Estimation of Heritability Juvenile fall chinook salmon representing 80 families were reared at the Abernathy Salmon Cultural Development Center, near Longview, Wash. The families were produced by mating 20 males to 80 females, 4 females per male, in a nest- ed breeding experiment. One hundred fish from each family were marked by cold-branding (Everest and Edmundson 1967) when they were 4 mo old and their weights averaged 2 g. Each group of 100 fish received a unique mark. 'This work was carried out in cooperation with the U.S. Fish and Wildlife Ser\'ice, Oregon Fish Commission, Oregon Wildlife Commission, and Oregon State University. 934 Thirty marked fish from each family were put into each of three tanks (1.8 m in diameter and 0.3 m deep) at the Abernathy Center. These groups of 30 fish will be referred to as tank-families. The remaining 10 fish from each family were put into a similar tank as a control. The test tanks were supplied with 18.9 liter/min of water which was air-supersaturated to 130 +. 1.5%. Two variables, time to death for each fish after exposure to air-supersaturated water and the percentage survival for each family after 36 h of exposure, were examined. Water to be air-supersaturated was directed in- to a pump to create a line pressure of 1.4 kg/cm-. A controlled amount of air was injected into the line through an air stone inserted at a joint in the line. Aeration occurred under pressure in a 1.5-m ver- tical section of line. The test water then entered a pressurized tank where excess air was vented. Air-supersaturated water from the pressure tank was then jetted below the surface into the test tanks. Stock Comparisons Since differences between stocks of fish in their resistance to gas-bubble disease arise from both genetic and environmental factors, the fish used in the present experiments were reared in one loca- tion under controlled conditions to minimize differences related to environment. Differences in the groups of fish tested then were assumed to have a genetic basis. Locations on or near the Columbia River that are discussed in this report are the following approximate distances (kilometers) upstream from the Pacific Ocean: Abernathy Salmon Cul- tural Development Center, 72; Kalama Hatchery, 105; Bonneville Dam, 234; Little White Salmon Hatchery, 265; Little Goose Dam, 635. Experiment L-In the fall of 1972, eggs were taken from mature fall chinook at Little Goose Dam on the Snake River, and at Trask River Salm- on Hatchery. Smolts migrating downstream from Little Goose Dam must pass over seven dams and swim through water which may be air-super- saturated up to 130% (Beiningen and Ebel 1971). The Trask River enters the Pacific Ocean about 80 km south of the Columbia River and has never been known to contain lethal levels of air-super- saturated water. Eggs obtained at the Trask River Salmon Hatchery were taken from a large number of crosses, and eggs obtained at Little Goose Dam were from crosses between two males and two females. Fertilized eggs from each source were trans- ported to Oregon State University where they were incubated at 9.5°C. The fry were fed for 2 mo before being exposed to 127% air-supersaturated water. At the time of testing, the fish weighed between 1.3 and 1.5 g. Air-supersaturated water was produced by aerating water under a hydrostatic head of 3 m in a vertical column of 15.2-cm pipe. A regulated amount of air was injected into the lower portion of the column through four air stones. Water drawn from the bottom of the column was 123% air-supersaturated. This water was then heated to 13.5°C to attain the test level of 127 + 2% super- saturation. Experiment IL-In 1973, fall chinook eggs were obtained from Abernathy Salmon Cultural Development Center, Little White Salmon Hatchery, and Kalama Hatchery-all on the Columbia River-and from the coastal Trask River Salmon Hatchery. Eggs from fish at Columbia River hatcheries were taken on 2 October, and those from fish at Trask River Hatchery on 28 November. All eggs were taken to Oregon State University for incubation, rearing, and testing. Because of differences in ages, the experimen- tal groups had to be tested at different times. W^e held the test fish in a constant environment at equal densities during the rearing period. The fingerlings were reared at 13.5°C in a 5.2 m x 0.3 m X 0.3m Plexiglas^ tank which was divided into 16 sections. The sections were divided into four blocks of four tanks each. Fifty fish from each of the four stocks were put into one section in each of the four blocks establishing a randomized block design. Fish were reared for 50 days, at the end of which time they weighed from 1.0 to 1.7 g. Seven days before testing, each group of fish was marked with a group-specific cold brand. The fish were exposed to 127 +. 2% air-supersa- turated water at 11 °C in a 16.5-liter tank. Time to 50% mortality, proportion dead in 96 h, and proportion dead in 150 h were determined. An ap- paratus similar to that described above for tests at the Abernathy Center was used in this experiment. In all tests, we measured the total -Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 935 uncompensated hyperbaric dissolved gas pressure with a Weiss saturometer. Results Estimation of Heritability Initial mortalities occurred 8 h after the test fish were placed in 130''? air-supersaturated water. All fish were dead after 132 h. The grand means of time to death (hours) in the three tanks were 22.62, 24.66, and 25.04. Two fish died in the control tank. Because of counting errors at time of marking, individuals per tank- family ranged from 5 to 54. Few tank-families varied greatly from the ex- pected number per family (28.4) as was indicated by the harmonic mean (26.8). Because of the une(iual numbers of individuals per tank-family and the large number of observations, an un- weighted means analysis of variance was used (Kempthorne 1957). First, the unweighted means of each family in each tank were subjected to an analysis of variance (Table 1). Second, all obser- vations were used in a one-way analysis of variance to compute the within tank-families (error) sum of squares. Because the distribution of time to death for fish in each tank followed a pois- son distribution, a square root transformation of time to death was applied before the analysis of variance was carried out. The square root trans- formation is the most appropriate for poisson data (Bartlett 1936). Variance components were estimated as: '^.= {MS^^-MS,jpr)/f = 0.041 o/,,^ = {MSj^f^r-i^/ri,)MS^) = 0.024 1 ' , a ' Male-females by tanks {MFT) (mM) (M) = 158 0.128 7 1 2 Error (£)i N . . . -mit = 6,566 2.800 <. w where: m = number of males 1 = number of females per male t = number of tanks (replicates) N . . . = total number of individuals n„ = harmonic mean number of Individuals per tank family = 26.82 'Error mean square obtained in separate analysis (see Kempthorne 1957:459). 936 Table 2.— Hours to 50% mortality (ET^q), and percentages dead in 96 h (P96) and 150 h (P150) for juvenile chinook salmon exposed to air-supersaturated water in Experiment II. Each value represents an average of four replicates; ranges are shown in parentheses. Table 3.-Analysis of variance in time to 50% mortality for Trask and Columbia River chinook salmon (Experiment II). Stock ET, P96 P150 Abernathy Little White Salmon Kalama Trask 92.5 (70-116) 86.5 (82-94) 73.8 (64-79) 62.0 (48-75) 53 61 64 70 (43-64) (57-67) (60-70) (64-82) 68.5 (63-80) 77.0 (75-80) 86.0 (73-95) 86.8 (75-96) significant (Table 3). Variation between the three Columbia River stocks was not significant. Similar comparisons were made from data summarized after 96 and 150 h of exposure (Table 2). On the average, differences between Columbia and Trask stocks remained significant, but varia- tion between Columbia River stocks became sig- nificant only after 150 h of exposure {F = 5.01). This difference was between the Kalama stock and other lower Columbia River stocks, suggest- ing that a difference in resistance to gas bubble disease exists even between stocks separated by relatively short distances. The reason for the similarity of resistances between lower Columbia River stocks probably was their common origin: Abernathy brood stock were originally taken from Spring Creek and Willard hatcheries, both of which are located upstream from Bonneville Dam. The much greater difference in time to 50% mortality between fish taken from Little Goose Dam and fish from the Trask River (80.5 + 3.39 h) than between combined lower Columbia stocks and the Trask stock (22.25 ± 6.37 h) indicates that fall chinook salmon migrating as far as Little Goose Dam are more resistant to gas bubble disease than are lower Columbia River stocks. This conclusion could be made only by comparing results from Experiments I and II, and by assum- ing that the results were not biased by the small number of crosses made at Little Goose Dam. Discussion Differences between stocks indicated that selection for phenotypes with greatest resistances to gas bubble disease has occurred in the Columbia River. This conclusion was supported by the ob- servation that stocks with the longest histories of exposure to air-supersaturated water were most resistant to gas bubble disease. Because additive genetic variance contributing to the observed differences probably has been reduced, and reduced at an unknown rate, it is Source of variation Degrees of freedom Mean squares Total 15 Blocks 3 367.7 3.02 Between stocks 3 739.5 6.07* Columbia vs. Trask 1 1,485.2 12.19** Within Columbia 2 366.6 3.01 Error 9 121.8 *, **, statistical significance at the 0.05 and 0.01 levels, re- spectively. impossible to estimate accurately the selection in- tensities that must have occurred in the past to produce the differences in resistance observed between Trask and Columbia River stocks. The low heritability for resistance to gas bubble disease in fall chinook salmon indicates that no great increases in resistance can be expected even at relatively high selection intensities. The results further indicate that stocks trans- ferred from coastal streams to hatcheries within the Columbia River drainage may experience high levels of mortality from gas bubble disease. On the other hand, Columbia River stocks may provide a source of brood fish that are resistant to gas bubble disease for stocking in other waterways. Acknowledgments We thank Director Bobby Combs and personnel at the Abernathy Salmon Cultural Development Center and personnel at the various locations from which eggs were obtained. The manuscript benefitted from reviews by P. H. Eschmeyer, J. A. Lichatowich, and an anonymous reviewer. Literature Cited Bartlett, M.S. 19.36. The square root transformation in analysis of variance. J. R. Stat. Soc. (Suppl.) 3:68-78. Beiningen, K. T., and W. J. Ebel. 1971. Dissolved nitrogen, dissolved oxygen, and related water temperatures in the Columbia and lower Snake Rivers, 1965-69. U.S. Natl. Mar. Fish. Serv. Data Rep. 56, 60 p. Blahm, T. H., R. J. McConnell, and G. R. Snyder. 1975. Effect of gas supersaturated Columbia River water on the survival of juvenile chinook and coho salmon. NOAA Tech. Rep. NMFS SSRF-688, 22 p. DiCKERSON, G. E. 1959. Techniques for research in quantitative animal genetics. In Techniques and procedures in animal production research, p. 56-105. Am. Soc. Anim. Prod., Beltsville, Md. 937 Ebel, W. J. 1969. Supersaturation of nitrogen in the Columbia River and its effect on salmon and steelhead trout. U.S. Fish. Wild!. Ser\'., Fish. Bull. 68:1-11. Everest, F. H., and H. E. Edmundson. 1967. Cold branding for field use in making juvenile sal- monids. Prog. Fish-Cult. 29:175-176. Kempthorne, 0. 1957. An introduction to genetic statistics. Iowa State Univ. Press, Ames, 545 p. RUCKER, R. R., .\ND K. HODGEBOOM. 1953. Observations on gas-bubble disease of fish. Prog. Fish-Cult. 15:24-26. Westgard, R. L. 1964. Physical and biological aspects of gas-bubble disease in impounded adult chinook salmon at McNary spawning channel. Trans. Am. Fish. Soc. 93:306-309. S. P. Cramer Present address: Oregon Wildlife Commission Gold Beach, OR 97iU Oregon Cooperative Fishery Unit Department of Fisheries and Wildlife Oregon State University Carvallis, OR 97331 J. D.McIntyre 938 INDEX Fishery Bulletin Vol. 73, No. 1-4, 1975 Acids, fatty differentiation of freshwater characteristics of, in marine specimens of Atlantic sturgeon 838 Acipenser oxyrkynchus-see Sturgeon, Atlantic ACKMAN, R. G., C. A. EATON, and B. A. LINKE, "Differentiation of freshwater characteristics of fatty acids in marine specimens of the Atlantic sturgeon, Acipenser oxyrhynchus" 838 "Activity, movements, and feeding behavior of the cunner, Tautogolabrus adspersus, and comparison of food habits with young tautog, Tautoga onitis, off Long Island, New York," by Bori L. Oila, Allen J. Bejda, and A. Dale Martin "Acute toxicity of ammonia to several developmental stages of rainbow trout, Saimo gairdneri," by Stanley D. Rice and Robert M. Stokes "Additional evidence substantiating existence of northern subpopulation of northern anchovy, Engraulis nwrdax," by Michael F. Tillman "Additional studies of the fishes, macroinvertebrates, and hydrological conditions of upland canals in Tampa Bay, Florida," by William N. Lindall, Jr., William A. Fa- ble, Jr., and L. Alan Collins "Age and growth of Pacific hake, Merluccius productus," by Thomas A. Dark "Age-length-weight and distribution of Alaska plaice, rock sole, and yellowfin sole collected from the southeast- ern Bering Sea in 1961," by Douglas D. Weber and Her- bert H. Shippen AGNELLO, RICHARD J., and LAWRENCE P. DON- NELLEY, "The interaction of economic, biological, and legal forces in the Middle Atlantic oyster industry" Albacore-see Tuna, albacore Algae Amchitka Island, Alaska experimental studies of canopy interactions in a sea otter-dominated kelp community 230 Alosa sapidissima—see Shad, American Amchitka Island, Alaska experimental studies of algal canopy interactions in a sea otter-dominated kelp community at 230 "(The) American Samoa longline fishery, 1966-71," by Howard 0. Yoshida 747 Anchovy, northern additional evidence substantiating existence of northern subpopulation 212 field criteria for survival criteria for successful first-feeding larvae 460 895 207 212 81 336 919 256 effect of temperature on feeding laboratory-determined criteria for successful feed- ing shipboard experiments with first-feeding larvae . . . study area vertical distribution of larvae Anchovy, northern Adriatic concentration of mercury, copper, nickel, silver, cad- mium, and lead in ANDERSON, LEE G., "Optimum economic yield of an internationally utilized common property resource". . . . ANDERSON, WILLIAM W., JACK W. GEHRINGER, and FREDERICK H. BERRY, "The correlation between numbers of vertebrae and lateral-line scales in western Atlantic lizardfishes (Synodontidae)" Antilles Current velocity and transport northeast of Bahama Islands. . ARP, ARTHUR H.-see VREELAND et al. Atlantic Ocean, northwestern teleconnections between northeastern Pacific Ocean and Gulf of Mexico and Bahama Islands velocity and transport of Antilles Current northeast of BAILEY, JACK E., BRUCE L. WING, and CHESTER R. MATTSON, "Zooplankton abundance and feeding habits of fry of pink salmon, Oncorhynchus gorbuscha, and chum salmon, Oncorhynchus keta, in Traitors Cove, Alaska, with speculations on the carrying capacity of the area" BAIRD, RONALD C.-see HOPKINS and BAIRD Bairdiella effects of acclimation on temperature and salinity tolerance of larvae salinity tolerance upper thermal tolerance effects of temperature and salinity on fertilization, embryonic development and hatching acclimation of spawning fish to low salinity capture and maintenance of fish deformed larvae embryonic mortality fertilization incubation incubation time induced maturation and spawning maturity of spawning fish normal development response surfaces spermatozoan activity survival of starved larvae viable hatch 459 454 457 457 460 193 51 202 626 306 626 846 250 250 5 2 12 8 3 3 7 2 6 7 13 5 12 12 939 Bairdiella icistia-see Bairdiella Baja California phytopiankton in upvvelling waters off BAKER, L.D.-see REEVE and BAKER BARNETT, MICHAEL A.-see JOHNSON and BAR- NETT BEJDA, ALLEN J.-see OLLA et al. Bentheogennema burkenroadi northeastern Pacific Ocean description and biology of a new species of Bering Sea, southeastern age-length-weight and distribution, 1961 Alaska plaice rock sole yellowfin sole BERRIEN, PETER L., "A description of Atlantic mackerel. Scomber scombrus, eggs and early larvae" . . . BERRY, FREDERICK H.-see ANDERSON et al. "Biology and taxonomy of the genus Nematoscelis (Crustacea, Euphausiacea)," by K. Gopalakrishnan Bivalve larvae reevaluation of combined effects of temperature and salinity on survival and growth Crassost rea virginica Mercenaria mercenaria Muiinia lateralifs BLACKBURN, MAURICE, and FRANCIS WILLIAMS, "Distribution and ecology of skipjack tuna, Katsuwonus pelamis, in an offshore area of the eastern tropical Pacific Ocean" Bonitos systematics and morphology Allothunnua color pattern Cybiosarda Gymnosarda key to species of Sardini meristic characters morjjhometric characters OrcynopHis osteology Sarda scales soft anatomy systematics BOTSFORD, LOUIS W., and DANIEL E. WICKHAM, "Correlation of upwelling index and Dungeness crab catch" BOYD, STEVEN-see WIEBE et al. Brachydanio rerio-see Zebrafish Brachyistiusfrenatus-see Perch, kelp BRAY, RICHARD N., and ALFRED W. EBELING, "Food, activity, and habitat of three "picker-type" 940 38 737 919 919 919 186 797 87 382 614 521 594 612 520 524 524 591 533 597 523 524 590 901 microcarnivorous fishes in the kelp forests off Santa Barbara, California" Brevmrtia smithi-see Menhaden, yellowfin Brevoartia tyrannuK—see Menhaden, Atlantic CAIN, THOMAS D., "Reproduction and recruitment of the brackish water clam Ranyia cuneata in the James River, Virginia" Cancer magister—see Crab, Dungeness CARR, WILLIAM E. S., and JAMES T. GIESEL, "Im- pact of thermal eflluent from a steam-electric station on a marshland nursery area during the hot season" CASTRO, WALTER E.-see SANDIFER et al. "Catches of albacore at different times of the day," by William G. Pearcy, Daniel A. Panshin, and Donald F. Keene CHAO, LABBISH N.-see COLLETTE and CHAO Chauliodus filoani inverse correlation between meristic characters and food supply CHEN, LO-CHAI, and ROBERT L. MARTINICH, "Pheromonal stimulation and metabolite inhibition of ovulation in the zebrafish, Brachydanio rerio" CHITTENDEN, MARK E., JR., "Dynamics of American shad, Alosa sapidi.^sima, runs in the Delaware River" . . "Chlorinated hydrocarbons in sea-surface films and sub- surface waters at nearshore stations and in the North Central Pacific Gyre," by P.M. Williams and K. J. Robertson Clams reproduction and recruitment in Virginia distribution and recruitment environmental data histological study of reproductive cycle hydrographic data larval setting relationship of reproductive cycle to environmental data set collectors sex ratio study area CLARK, ROBERT C, JR., and JOHN S. FINLEY, "Up- take and loss of petroleum hydrocarbons by the mussel, Mytiliix edulis, in laboratory experiments" CLARKE. THOMAS A.-see HARTMANN and CLARKE Cod in m introduction in New England waters COLLETTE, BRUCE B., and LABBISH N. CHAO, "Systematics and morphology of the bonitos (Sarda) and their relatives (Scombridae, Sardini)" COLLIER, TRACY K.-see ROUBAL and COLLIER COLLINS, L. ALAN-see LINDALL et al. 815 412 67 691 290 889 487 445 427 414 415 423 421 425 414 420 413 508 215 516 Columbia River coho salmon homing behavior and contribution to "Comment. Introduction to Codimv in New England waters," by Victor L. Loosanoff "Comparative vulnerability of fry of Pacific salmon and steelhead trout to predation by torrent sculpin in stream aquaria," by Benjamin G. Patten Computer program analysis polymodal frequency distributions (ENORMSEP), FORTRAN IV " "(A) computer program for analysis of polymodal frequency distributions (ENORMSEP), FORTRAN IV," by Marian Y. Y. Yong and Robert A. Skillman "(The) concentration of mercury, copper, nickel, silver, cadmium, and lead in the northern Adriatic anchovy, Engraulia encraxicholus:, and sardine, Sardina pUchar- dus," by Malvern Gilmartin and Noelia Revelante Conger oceanicus— see Eel, conger Control theory, optimal fishery regulation via "(The) correlation between numbers of vertebrae and lateral-line scales in western Atlantic lizardfishes (Synodontidae)," by William W. Anderson, Jack W. Gehringer, and Frederick H. Berry "Correlation of upwelling index and Dungeness crab catch," by Louis W. Botsford and Daniel E. Wickhani . . Cottiii^ rhotheus—see Sculpin, torrent COX, JAMES L.-see WIEBE et al. Crab, Dungeness Pacific coast statistical relationship between upwelling intensity and annual catch CRAMER, S. P., and J. D. McINTYRE, "Heritable resis- tance to gas bubble disease in fall chinook salmon, Oh- curhynchus tshawytucha" Craf