^^^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 to 5, 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 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 14.5 48.5 24.4 14.8 13.4 43.5 35.8 38.8 1.9 63.1 43.2 7.4 5.7 33.7 52.6 60.1 16.8 48.7 15.3 2.7 28.9 81.6 87.5 41.6 18.4 69.9 49.4 59.9 52.5 62.3 87.8 81.3 50.8 30.3 77.2 76.7 82.0 67.4 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 II I 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 40 19.2 13.9 7.8 4.4 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 66.7 25.0 27.6 1.9 62.6 34.3 24.2 8.0 67.7 22.6 40.9 11.9 12.0 1.0 82.5 40.4 37.0 7.1 39.8 53.5 51.4 9.9 66.0 34.1 41.5 4.5 14.4 1.0 0.9 88.4 94.7 72.7 72.8 96.0 75.6 66.3 66.3 76.2 79.8 784 42.9 1.0 85.9 62.2 4.1 76.5 92.9 68.8 47.5 96.0 89.0 80.0 41.2 89.9 88.4 79.2 31.9 2.0 73.8 60.0 6.2 2.0 20.0 16.3 33.3 28.3 10.1 1.3 7.1 22.6 30.7 7.0 7.5 29.3 35.0 2.0 23.5 44.6 38.3 2.8 4.1 30.6 29.8 1.5 3.1 77.9 77.9 50.6 18.1 70.0 63.0 34.6 68.2 70.8 57.5 38.2 15.2 68.4 59.7 38.8 11.9 80.2 70.2 53.8 81.1 81.0 54.7 28.6 14.6 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 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 19.7 73.8 42.2 24.2 89.8 53.4 2.5 31.1 88.5 0.9 18.7 79.2 21.2 85.3 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 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 — — 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 = (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 5 10 15 10 ABREOJOS i ^24 *°''" 100 (m 45' 30' 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 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 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 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 m, 14.94°C at 50 m) than at Station 1 (14.53°C at 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 « ■6 1 II Q CD (D UJ CD to TO £ to S CO to SI -J Is H -J (0 to §1 '5 c c (0 CO s: 2 c CO Q. N Q. £ tn Qj to :::; ^- CO » .0 to io CO :3 CO C 1 = tr LU I 1- d. Q. CO 0) CO. a 11 "2 " to CO Q> ■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 0.09 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 0.03 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 17 316 b .017 .731 .177 .304 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 16 417 b .027 .457 .118 .793 1.101 2.50 38 a 4 60 19 528 17 628 b .009 .289 .093 1.545 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 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 2 27 0.14 0.14 0.06 0.01 0.005 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 6 0.23 0.16 0.06 0.20 0.04 0.04 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 43 0.03 0.01 0.01 0.03 0.01 0.015 Guinardia flaccida 69 109 66 24 17 2 50 0.29 0.46 0.28 0.10 0.07 0.008 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 47 364 0.61 0.46 0.71 0.22 0.12 0.05 0.37 Thalassiosira rotula 442 420 969 585 57 34 454 0.59 0.56 1.29 0.78 0.08 0.05 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 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 = 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^) = 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 _ (b) (0 a^y- =U^+ XaG^ - 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\ = 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 6 4 2 s° Z 4 UJ go Q 6 a 6 4 2 6 4 2 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 1,005 42 6 22 449 8,853 93.2 Anchoa hepsetus 208 1 1 210 2.3 Bairdiella chrysura 19 4 29 41 1 1 95 1.1 Chaetodipterus faber 9 1 1 16 5 1 1 34 05 Syngnathus scovelli^ 4 1 7 5 17 0.2 Lagodon rhomboides 2 6 3 3 14 0.1 Gobiosoma bosci 1 1 2 4 2 3 13 0.1 Pogonias cromis 1 1 2 8 1 13 0.1 Lucania parva^ 1 1 2 2 4 10 0.1 Diapterus plumierl^ 4 2 2 1 9 0.1 Eucinostomus argenteus 7 1 1 9 0.1 Menticirrhus americanus 1 5 2 1 9 0.1 Cynoscion arenarius 3 1 2 2 8 0.1 Orthopristis chrysoptera 4 1 5 0.1 Eucinostomus gula 4 4 0.1 Trinectes maculatus^- 4 4 0.1 Cynoscion nebulosus 3 3 0.0 Microgobius gulosus 1 1 1 3 0.0 Opisthonema oglinum 1 2 3 0.0 Sciaenops ocellata 1 1 1 3 0.0 Sphoeroides nephelus 2 1 3 0.0 Chilomycterus schoepfi^ 1 1 2 0.0 Epinephelus itajara^ 1 1 2 0.0 Gobiosoma robustum'' 1 1 2 0.0 Hippocampus erectus^ 1 1 2 00 Leiostomus xanthurus 2 2 00 Malaclemys terrapin 1 1 2 0.0 Monacanthus hispidus'' 1 1 2 0.0 Syngnathus louisianae'' 1 1 2 0.0 Archosargus probatocephalus 1 1 0.0 Arius felis 1 1 0.0 Elops saurus^ 1 1 0.0 Harengula pensacolae^ 1 1 0.0 Hippocampus zosterae^ 1 1 0.0 Lactophrys quadricornis^ 1 1 0.0 Invertebrates: Lolliguncula brevis 1 1 11 7 17 21 13 15 86 0.9 Callinectes sapidus 4 11 8 5 4 10 2 3 1 1 49 0.5 Penaeus duorarum 1 5 6 1 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 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% 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 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 • • • -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 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 6 7 4 GEORGES BANK 9 22 n3 2 I 3 2 10 15 C r 10 15 C u. o o z < X Q UJ z Q 3 Z 03 - < GULF OF MAINE I 9 II 3nii7n2 T" 10 15 20 C NOVA SCOTIAN SHELF 5 3 I 3 II 2 3 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 FISHERY BULLETIN: VOL. 73, NO. 1 GEORGES BANK 2 9147I 6 5 2 1 in i^ 10 15 20 C 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 I 3 I3["l0 6 1 I GEORGES BANK .5. h n 10 15 20 C 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 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 l O 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 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 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 12 4 3 6 7 4 2 -I I - 0- GULF OF MAINE "T 5 I 10 15 20 C 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 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 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 to 100% with a mean of 40%' ; and for pinfish it varied from 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 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 7 II natatory 3 1 7 III natatory 1 4 7 IV dominant natatory 5 5 variant natatory 5 7 variant natatory 5 6 variant natatory 5 4 V dominant juvenile 5 3 variant juvenile 5 5 variant juvenile 5 4 variant juvenile 5 2 VI juvenile 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 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 °-^ " 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 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 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 45.5 ± 4.6 45 100 nOO ± 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 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 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 ± 12.9 100 ±0 11.5 ± 10.2 8.4 ± 10.7 1100 ±0 145.5 ±16.1 25.8 ±11.9 7.0 ± 4.9 49.2 ±14.0 100 ±0 100 ±0 185.2 ± 33.4 2.1 ± 5.6 100 ±0 17.5 ±7.1 i100±0 177.4 ± 12.7 11.5 ± 4.0 1.2 ± 4.0 37.5 ±10.2 100 ±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 < 001 10 20 30 AGE IN DAYS 40 50 100 10 ct Q X < 01 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 (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%. 25%o 1 * - - 35%. \ • ^^^ X. ^ "-^ 34 : A • ^--.^ ^^^^ 32 — ^v ^^^.^^^ - — ^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^ 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 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 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 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 1 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 5 5 3 3 8 1 7 12 9 15 10 16 5 11 2 16 4 10 2 6 1 7 1 7 1 2 3 1 3 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 20 100 80 60 - 40 - 20 CATALINA U- GUADALUPE ( ^^ " \ / r^ 60 - \/ >) — -^ / v 40 - J \ " / 1 >* ' 20 - ' ■> / / \ .,.o — o... 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^"^^- ■'' 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 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 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. 61:434-48