..<' °?°^. ^ o c ><» '^'•^TES O* ^ Fishery BuJIetin National Oceanic and Atmospheric Administration • National Mat'ine Fisheries Service OD i P Vol. 76, No. 1 January 1978 COUCH, JOHN A. Diseases, parasites, and toxic responses of commercial penaeid shrimps of the Gulf of Mexico and south Atlantic coasts of North America 1 HENRY, KENNETH A. Estimating natural and fishing mortalities of chinook salm- on, Oncorhynchus tshawytscha, in the ocean, based on recoveries of marked fish 45 BIGFORD, THOMAS E. Effect of several diets on survival, development time, and growth of laboratory-reared spider crab, Libinia emarginata, larvae 59 FABLE, WILLIAM A., JR., THEODORE D. WILLIAMS, and C. R. ARNOLD. De- scription of reared eggs and young larvae of the spotted seatrout Cynoscion nebulosus 65 BULLARD, FERN A., and JEFF COLLINS. Physical and chemical changes of pink shrimp, Pandalus borealis, held in carbon dioxide modified refrigerated seawater compared with pink shrimp held on ice 73 MARKLE, DOUGLAS F. Taxonomy and distribution of Rouleina attrita and Rouleina maderensis (Pisces: Alepocephalidae) 79 GRABE, STEPHEN A. Food and feeding habits of juvenile Atlantic tomcod, Mi- crogadus tomcod, from Haverstraw Bay, Hudson River 89 BERRIEN, PETER L. Eggs and larvae of Scomber scombrus and Scomber japonicus in continental shelf waters between Massachusetts and Florida 95 COLIN, PATRICK L. Daily and summer-winter variation in mass spawning of striped parrotfish, Scarus croicensis 117 CALKINS, DONALD G. Feeding behavior and major prey species of the sea otter, Enhydra lutris, in Montague Strait, Prince William Sound, Alaska 125 HOBSON, EDMUND S., and JAMES R. CHESS. Trophic relationships among fishes and plankton in the lagoon at Enewetak Atoll, Marshall Islands 133 BROUSSEAU, DIANE J. Spawning cycle, fecundity, and recruitment in a popula- tion of soft-shell clam, Mya arenaria, from Cape Ann, Massachusetts 155 SMITH, W. G., J. D. SIBUNKA, and A. WELLS. Diel movements of larval yellowtail flounder, Limanda ferruginea, determined from discrete depth sampling 167 WAHLE, ROY J., and ROBERT R. VREELAND. Bioeconomic contribution of Co- lumbia River hatchery fall chinook salmon, 1961 through 1964 broods, to the Pacific salmon fisheries 179 KINNER, PETER, and DON MAURER. Polychaetous annelids of the Delaware Bay region 209 ROSS, STEPHEN T. Trophic ontogeny of the leopard searobin, Prionotus scitulus (Pisces: Triglidae) 225 J (Continued on back cover) Q Seattle, Washington U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Richard A. Frank, Administrator NATIONAL MARINE FISHERIES SERVICE Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Begiiming 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 jjapers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, nimiber 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription fix)m the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available tree in limited numbers to libraries, research institutions. State and Federal agencies, and in exchange for other scientific publications. EDITOR Dr. Bruce B. CoUette Scientific Editor, Fishery Bulletin National Marine Fisheries Service Systematics Laboratory National Museum of Natural History Washington, DC 20560 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. William H. Bayhff Inter-American Tropical Tuna Commission Dr. Roger F. Cressey, Jr. U.S. National Museum Mr. John E. Fitch California Department of Fish and Game Dr. William W. Fox, Jr. National Marine Fisheries Service Dr. Marvin D. Grosslein National Marine Fisheries Service Dr. Edward D. Houde University of Miami Dr. Merton C. Ingham National Marine Fisheries Service Dr. Reuben Lasker National Marine Fisheries Service Dr. Sally L. Richardson Oregon State University Dr. Paul J. Struhsaker National Marine Fisheries Service Dr. Austin Williams National Marine Fisheries Service Kiyoshi G. Fukano, Managing Editor The Fishery Bulletin is published quarterly by Scientific Publications Staff, National Marine Fisheries Service, NOAA, Room 450, 1107 NE 45th Street, Seattle, WA 98105. Controlled circulation postage paid at Tacoma, Wash. 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 tfirough 31 December 1978. Fishery Bulletin CONTENTS Vol. 76, No. 1 January 1978 COUCH, JOHN A. Diseases, parasites, and toxic responses of commercial penaeid shrimps of the Gulf of Mexico and south Atlantic coasts of North America 1 "** HENRY, KENNETH A. Estimating natural and fishing mortalities of chinook salm- on, Oncorhynchus tshawytscha, in the ocean, based on recoveries of marked fish 45 BIGFORD, THOMAS E. Effect of several diets on survival, development time, and growth of laboratory-reared spider crab, Libinia emarginata, larvae 59 -^ FABLE, WILLIAM A., JR., THEODORE D. WILLIAMS, and C. R. ARNOLD. De- scription of reared eggs and young larvae of the spotted seatrout Cynoscion nebulosus 65 BULLARD, FERN A., and JEFF COLLINS. Physical and chemical changes of pink shrimp, Pandalus borealis, held in carbon dioxide modified refrigerated seawater compared with pink shrimp held on ice 73 MARKLE, DOUGLAS F. Taxonomy and distribution of Rouleina attrita and Rouleina maderensis (Pisces: Alepocephalidae) 79 GRABE, STEPHEN A. Food and feeding habits of juvenile Atlantic tomcod, Mi- crogadus tomcod, from Haverstraw Bay, Hudson River 89 BERRIEN, PETER L. Eggs and larvae of Scomber scombrus and Scomber japonicus in continental shelf waters between Massachusetts and Florida 95 COLIN, PATRICK L. Daily and summer-winter variation in mass spawning of striped parrotfish, Scarus croicensis 117 CALKINS, DONALD G. Feeding behavior and major prey species of the sea otter, Enhydra lutris, in Montague Strait, Prince William Sound, Alaska 125 "tr HOBSON, EDMUND S., and JAMES R. CHESS. Trophic relationships among fishes and plankton in the lagoon at Enewetak Atoll, Marshall Islands 133 M' BROUSSEAU, DIANE J. Spawning cycle, fecundity, and recruitment in a popula- tion of soft-shell clam. My a arenaria, from Cape Ann, Massachusetts 155 SMITH, W. G., J. D. SIBUNKA, and A. WELLS. Diel movements of larval yellowtail flounder, Limanda ferruginea, determined from discrete depth sampling 167 WAHLE, ROY J., and ROBERT R. VREELAND. Bioeconomic contribution of Co- lumbia River hatchery fall chinook salmon, 1961 through 1964 broods, to the Pacific salmon fisheries 179 KINNER, PETER, and DON MAURER. Polychaetous annelids of the Delaware Bay region 209 ROSS, STEPHEN T. Trophic ontogeny of the leopard searobin, Prionotus scitulus (Pisces: Triglidae) 225 (Continued on next page) Seattle, Washington For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402 — Subscription price: $11.80 per year ($2.95 additional for foreign mailing). Cost per single issue — $2.95. Contents-continued HAYNES, EVAN. Description of larvae of the humpy shrimp, Pandalus goniurus, reared in situ in Kachemak Bay, Alaska 235 BEN-TUVIA, ADAM. Immigration of fishes through the Suez Canal 249 LOVE, MILTON S., and ALFRED W. EBELING. Food and habitat of three switch- feeding fishes in the kelp forests off Santa Barbara, California 257 COLLETTE, BRUCE B., JOSEPH L. RUSSO, and LUIS ALBERTO ZAVALA- CAMIN. Scomberomoris brasiliensis, a new species of Spanish mackerel from the western Atlantic 273 Notes ROGERS, CAROLYN A., DOUGLAS C. BIGGS, and RICHARD A. COOPER. Aggregation of the siphonophore A^anomta cara in the Gulf of Maine: observations from a submersible 281 "^ DAHLBERG, MICHAEL L. Computer program for analysis of the homogeneity and goodness of fit of frequency distributions, FORTRAN IV 285 GILMORE, R. GRANT, JOHN K. HOLT, ROBERT S. JONES, GEORGE R. KULCZYCKI, LOUIS G. MacDOWELL III, and WAYNE C. MAGLEY. Portable tripod drop net for estuarine fish studies 285 WELLINGTON, G. M., and SHANE ANDERSON. Surface feeding by a juvenile gray whale, Eschrichtius robustus 290 "^ SCHOLZ, ALLAN T., JON C. COOPER, ROSS M. HORRALL, and ARTHUR D. HASLER. Homing of morpholine-imprinted brown trout, Salmo trutta 293 ELDRIDGE, PETER J., FREDERICK H. BERRY, and M. CLINTON MILLER, III. Diurnal variations in catches of selected species of ichthyoneuston by the Boothbay neuston net off Charleston, South Carolina 295 Vol. 75, No. 4 was published on 30 December 1977. 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. DISEASES, PARASITES, AND TOXIC RESPONSES OF COMMERCIAL PENAEID SHRIMPS OF THE GULF OF MEXICO AND SOUTH ATLANTIC COASTS OF NORTH AMERICA^ John A. Couch* ABSTRACT A reference work and review of both infectious and noninfectious diseases of commercial penaeid shrimps of the Gulf and South Atlantic region of the United States is presented. Disease is second only to predation and periodic physical catastrophes in limiting numbers of penaeid shrimps in nature and second only to nutritional and reproductive requirements in limiting aquacultural successes with p>enaeid shrimps. Infectious agents causing disease in penaeid shrimps are a virus, bacteria, fungi, protozoa, hel- minthes, and nematodes. A well-described Baculovirus infects larval and adult shrimp and is as- sociated with mortality, particularly in larval shrimp. Bacteria of the genera Vibrio, Beneckea, and Leucothrix are associated with disease in penaeid shrimps, but bacterial roles in mortality are unclear. The same is largely true for fungi with members of the genera Lagenidium and Fusarium causing pathogenesis in cultured shrimp. Lagenidium causes severe destruction of larval shrimp tissues. Of the many protozoan groups represented in and on penaeid shrimps as tissue parasites and commensals, the MicrospKjrida of the genera Nosema, Thelohania, and Pleistophora are the most destructive. The ciliate protozoa Zoo^/iamratum sp., Lagenophrys sp., and Parauronema sp. may cause dysfunction in shrimp. An undescribed apostome ciliate is associated with black gill disease. A suctorian, Ephelota sp., is an ectocommensal of larval shrimp, attaching to the cuticle. The six species of gregarines reported cause little or no pathogenesis, and a single reported flagellate si)ecies role in shrimp health is uncertain. Flatworms found in penaeid shrimps are metacerceu"iae of a species o{ Microphallus in muscles and viscera, metacercariae of Opecoeloides fimbriatus in viscera, plerocercoid larvae of Prochristianella hispida in the hepatopancreas and hemocoel, and four other cestode developmental stages. Nematodes found are Thynnascaris sp., Spirocamallanus pereirai, Leptolaimus sp., and Croconema sp. Noninfectious disease agents in penaeid shrimps are chemical pollutants, heavy metals, and en- vironmental stresses. Organochlorine, organophosphate, and carbamate pesticides all have adverse eflfects in penaeids. Fractions of petroleum, particularly the naphthalenes, are very toxic to shrimp. Little other work has been done on the effects of petroleum on penaeid shrimps. Cadmium causes black gills in shrimp by killing gill cells. Mercury is accumulated by penaeids and may interfere with their osmoregulatory abilities. Many chemotheropeutic chemicals used routinely in treatment offish dis- eases are toxic to shrimp at certain determined concentrations. Spontaneous pathoses found are a benign tumor, muscle necrosis, and gas bubble disease. "Shell disease" is discussed from points of view of possible causes. A syndrome of "broken backs" is reported in jienaeid shrimps for the first time. An overview is presented for general needs in penaeid shrimp health research. Recent attempts to culture penaeid shrimps in large quantities have stimulated renewed interest in the pathobiology of crustacean species. Patho- gens and disease, in general, have been indicted as causes for many failures in maintaining various life-cycle stages of Crustacea. Therefore, consid- 'Contribution No. 283 from the Gulf Breeze Environmental Research Laboratory. ^U.S. Environmental Protection Agency, Environmental Re- search Laboratory, Gulf Breeze, PL 32561. Manuscript accepted May 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. erable amounts of new information and data on known and recently discovered diseases of penaeid shrimps have been published or reported in the last decade. This recent information, along with an older but equally valuable series of publica- tions, presents a substantial body of knowledge which describes and defines problems of disease encountered in the biology, management, and massive culture of penaeid shrimps. Major contributions to the study of shrimp dis- eases in North America have been made by sev- 1- eral individuals. Sprague (1954, 1970, footnote 3), Kruse (1959, 1966), Hutton et al. (1959), Iversen and Manning (1959), Hutton (1964), and Iversen and Van Meter (1964) were early explorers in penaeid shrimp infectious diseases. More recently the works of Overstreet (1973), Lightner (1974, 1975), Lightner and Fontaine (1973), Johnson (1974), Feigenbaum ( 1973, 1975), Couch ( 1974a, b, 1976) and Sindermann'' have contributed to the general fund of data. Overstreet's 1973 paper is particularly valuable because it gives prevalence data for many of the parasites of penaeid shrimps of the northern Gulf Many other authors of single, significant works on penaeid diseases will be cited in specific sections later in this paper. The scientific reports and reviews mentioned above, along with much unpublished experience, present a consensus which impresses me with the high significance of disease to the overall ecology and biology of penaeid shrimps. In its broadest sense, disease is probably second only to predation and periodic physical catastrophes (e.g., freshets, temperature fluctuations) as a continuous en- vironmental factor limiting numbers of penaeid shrimps in nature. In attempts at massive culture of penaeid shrimps, infectious disease may rank below only reproductive and nutritional require- ments as a limiting factor. Toxicants, in the form of pollutants, are threats to the well being of es- tuarine species, particularly in certain chronically polluted regions. Toxic responses in penaeid shrimps have been studied experimentally re- cently, and, therefore, some data are available on this subject. This paper is concerned with the present status of diseases, parasites, and toxic responses of four commercial species of penaeid shrimps from the Gulf and South Atlantic region of North America. These are the pink shrimp, Penaeus duorarum; the brown shrimp, P. aztecus; and the white shrimp, P. setiferus. Occasional reference will be made to parasites of P. braziliensis which occupies a marginal portion of the U.S. range of the three other species. The subjects will be treated in the following order: Infectious diseases and parasites; noninfectious diseases and toxic responses; and overview and future research. ^Sprague, V. 1950. Notes on three microsporidian parasites of Decapod Crustacea from Louisiana waters. Occas. Pap. Mar. Lab., La. State Univ. 5:1-8. ■•Sindermann, C. J. 1974. Diagnosis and control of mariculture diseases in the United States. Tech. Ser. Rep. No. 2, Natl. Mar. Fish. Serv., NOAA, Highlands, N.J., 306 p. FISHERY BULLETIN: VOL. 76, NO. 1 INFECTIOUS DISEASES AND PARASITES Viruses To date, only a single virus disease has been described for shrimps. Couch (1974a, b) and Couch et al. (1975) have described a rod-shaped virus (Figures 1-3) which has many characteristics of the baculoviruses (nuclear polyhedrosis viruses) previously described only from insects or mites. The virus has been named Baculovirus penaei (Couch 1974b). This virus commonly has been found to infect the hepatopancreas of juvenile and adult stages of pink and brown shrimp in nature. Laboratory- reared larval brown shrimp (protozoea and mysis stages) have been found with virus-infected mid- gut and hepatopancreas. Infected hepatopancreatic cells in pink shrimp display striking cytopathological changes when compared with normal, noninfected cells. Nuclear hypertrophy (Figure 3), chromatin diminution (Figure 3), nucleolar degeneration (Figure 3), and polyhedral inclusion body (PIB, Figure 2) produc- tion are characteristic of patent virus infections observable with bright field or phase contrast mi- croscopy. Electron microscopy (EM) reveals the rod- shaped virions (269 nm x 50 nm) in infected, hypertrophied nuclei prior to, during, and after the PIB is formed. Various stages of the virus replicative cycle are observable with EM of thin sections of moderately to heavily infected hepatopancreas. The ultimate cytopathological ef- fect of the virus is destruction of the host cell through rupture or lysis. This is accomplished usually by the growth of the PIB to a size too large for the host cell to accommodate (Figure 4), con- comitant with virus-induced nuclear hypertrophy and probable stressing of nuclear membranes. The PIB's produced during infections are pat- ently diagnostic for the baculovirus of penaeid shrimp (Figures 4, 5). To find a single characteris- tic PIB in tissue squashes of shrimp hepatopan- creas or midgut is to diagnose infection. Quantita- tion of patent infections (PIB's present) can be made on a relative basis by hemocytometer counts of PIB's in aliquots of fresh tissue. Degree of latent infections, however, may be estimated only with great difficulty through laborious EM examina- tions. Over 2,000 PIB's/mm^ of hepatopancreatic tissue are considered a heavy infection as deter- mined by hemocytometer counts. Heavy patent COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS ^ i * * ^»m 'M "T-^ ^;rw'..J A.Ji>M. ^ Figure l.—Bacuhvirus virions in nucleus of hepatopancreatic cell of pink shrimp; note rod form (arrows) and outer envelope surrounding nucleocapsid (electron micrograph), x 70,000. Figure 2. — Polyhedral inclusion body (PIB) in virus-infected nucleus; note characteristic triangular form, and rod-shaped virions in PIB (arrows); also note heterochromatin diminution and granular nucleoplasm. x22,260. FISHERY BULLETIN; VOL. 76, NO. 1 Figure 3. — Two hepatopancreatic cells withBacuZowrus-infected nuclei; note nuclear membrane proliferation (arrow) and nuclear hypertrophy, x 14,400. COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS FIGURE 4.— Phase contrast micrograph of fresh squash preparation of heavily, patent, virus-infected hepatopancreas from pink shrimp, note hypertrophied nuclei (arrows) and characteristic refringent PIB's. x 1,000. FIGURE 5.— Phase contrast micrograph of fresh squash showing PIB's (arrows) of varying sizes, some free of nuclei following nuclear rupture, x 1,000. FISHERY BULLETIN: VOL. 76, NO. 1 infections are obvious in fresh squash prepara- tions because PIB's fill every microscopic field. Prevalence of virus in feral pink shrimp from several locations on the northern gulf coast of Florida has varied among samples collected. There appears to be no seasonal intensification of prevalence that is statistically significant; how- ever, fall samples have been best for recovering heavy infections. To date, of 4,676 shrimp examined, 808 have been patently infected. In the laboratory, virus prevalence and intensity have increased repetitively in 20- to 30-day periods in different lots or samples of feral shrimp held under crowded, sublethally stressful conditions (Couch 1974b). This increase in prevalence associated with crowding provides indirect evidence for the infectious nature of the shrimp baculovirus. There is also increasing evidence, from our research, that exposures to low levels of certain chemicals, such as polychlorinated biphenyl (PCB), enhance spread of virus through captive populations (Couch and Courtney 1977). We have induced a 50% increase in prevalence in captive shrimp by exposing shrimp to sublethal levels of PCB's (Aroclor 1254).^ Transmission in nature probably is achieved via cannibalism of infected shrimp by noninfected shrimp. Laboratory transmission has been minimally successful when hatchery-reared or nonpatently infected juvenile or adult shrimp were fed heavily infected hepatopancreas. Only about 20% of fed shrimp show patent infections 20 to 30 days after initial feeding. Degree of infection in adult shrimp is not useful in predicting mortal- ity of shrimp. Recently the shrimp baculovirus was associated with massive mortality of larval and postlarval brown shrimp in a commercial aquaculture at- tempt. Brown shrimp, hatched and reared to pro- tozoal and mysid stages in laboratory tanks, suf- fered a mass mortality in a 48-h period (95% of several million larvae). Water quality was not found to be at fault and there were no toxicants known to be in the water. Upon careful histologi- cal examination of a sample of surviving and dead larvae, I discovered that 19.4% in - 139) had patent virus infections, mostly heavy, in midgut and hepatopancreatic cells (Table 1). Subsequent electron microscopical study confirmed that 60 to 90% of hepatopancreatic cell profiles in larvae had infections, many with prepatent stages of the virus. Present in higher prevalences in these dying shrimp were a flagellate protozoon and a ciliate protozoon. The relative roles of the three pathogens in the shrimp mortality will be dis- cussed in later sections of this paper (Tables 1, 2). Table l. — Relative prevalence of pathogens in 139 larval (late protozeal and mysid stages) brown shrimp, Penaeus aztecus,^ in April 1974. Condition Not infected Flagellate Ciliate Virus 'Whole mount slides with Protargol stain (Bodian-activated protein silver). Table 2. — Prevalence and concurrent infections of pathogens in 139 larval brown shrimp examined in April 1974. [Concurrent vs. single infections.] Number of Percent of larvae affected total examined 41 29.5 89 64.0 40 28.8 27 19.4 Number of Percent of Types of pathogens larvae affected total examined None 41 29.5 Flagellate only 38 27.3 Ciliate only 1 0.7 Virus only 8 5.8 Flagellate and ciliate 32 23.0 Flagellate and virus 12 8.6 Ciliate and virus 0 0.0 Flagellate, virus, and ciliate 7 5.0 ■^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA, or USEPA. Bacteria The role of bacteria in diseases of penaeid shrimps is presently being investigated seriously for the first time. A few scattered reports deal with bacteria as pathogens, contaminants, or ectocom- mensals in shrimps. Cook and Lofton (1973) reported isolation of three genera of bacteria, Beneckea, Vibrio, and Pseudomonas, from penaeid shrimp suffering from "shell disease," also known as black spot dis- ease. This disease (Figure 6) is characterized by brown to black spots on the external carapace or cuticle of shrimp and has been observed in brown, pink, and white shrimps. In advanced cases of the disease, considerable erosion and destruction of the cuticle occurs. This disease has been reported from many other decapod Crustacea (Rosen 1970). Chitinoclastic bacteria such as Beneckea sp. have been thought to be the causative agents of black spot disease, although attempts to experimentally produce the disease in shrimp by innoculating Beneckea have had uncertain results (see section on "shell disease" under Noninfectious Diseases). Mechanical injury to shrimp that results in breakage in the normal cuticle probably plays an COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Figure 6. — "Shell disease" in pink shrimp; note black spots of varying sizes (arrows); none of these have penetrated cuticle of shrimp at this stage. Figure 7. — a. Filaments of Leucothrix mucor (bacteria) in heavy infestation on gills of pink shrimp. x400. b. Wet mount preparation of L. mucor from heavy gill infestation; note granules in some filaments. x900. c. Single filaments of L. mucor showing attachment to end of gill filament; note few bacterial filaments in light infestation shown here. x900. FISHERY BULLETIN: VOL. 76, NO. 1 initiating role in the genesis of black spot disease (Cook and Lofton 1973). The effects of black spot disease on individual shrimp is apparently a breakdown of cuticular protection, thus permitting loss of hemolymph and invasion by internally destructive pathogens. Black spot disease in penaeids is fairly common, at least in early manifestation. However, the disease probably plays a minor role in mortalities of feral shrimp because shrimp probably tolerate the ini- tial lesions well. Vanderzant et al. (1970) isolated Vibrio para- hemolyticus from white shrimp from the Gulf of Mexico. This bacterium is one etiological agent for human gastroenteritis in Japan and possibly in the United States (Krantz et al. 1969). The pathogenicity of V. parahemolyticus for Crus- tacea, including shrimp, has not been conclusively established. One should remember that natural seawaters, particularly from inshore regions, may be considered "gram negative bacterial soups." Therefore, the presence of Vibrio sp. and other gram negative rods on marine organisms living in the "soup" should be expected. The role that Vibrio plays in the health of shrimps is uncertain. Ulitizur (1974) has pointed out that certain strains of Vibrio parahemolyticus isolated from sea water have very short generation times (12-14 min) at higher temperatures (39 °C). In subtropi- cal areas where temperatures might soar in hot seasons, particularly in ponds, the role of Vibrio sp. as pathogens of shrimp might be enhanced. Lightner (1975) discussed at length the suspect role of Vibrio spp. in penaeid shrimp health. Ectocommensal bacteria may play a significant role in the well being of penaeids, particularly those held in crowded volumes of water where rich organic substrate and optimum temperatures prevail. Pertinent among this group is the filamentous bacterium Leucothrix mucor (Oers- ted), a widespread epiphyte of marine animals and plants (Johnson et al. 1971). Leucothrix has been found in high numbers attached to the gill fila- ments of brown, white, and pink shrimp (Figure 7 a, b). The filaments are nonbranching, attached singly to the cuticle of the gills (Figure 7c), have a modal diameter of 2 fxin, and consist of septate chains of almost square-shaped bacteria. Each bacterium has several mesosomes along its cyto- plasmic membrane (Figure 8). A study was conducted with EM to determine the mode of attachment of^ Leucothrix to shrimp gill cuticle. Figure 9a, b shows cross sections of a Figure 8. — Electron micrograph of single filament of Leuco- thrix mucor showing nearly square cell profiles; note nucleoids (N) and mesosomes (M) (arrows) of bacterial cells plus septa separating each cell in filament x25,900. basal portion of a filament at its point of adhesion to gill cuticle. The bacterium does not possess a differentiated holdfast. There is no penetration of the epicuticle, and apparently the filament is se- cured to the gill epicuticle by an electron-opaque mucouslike substance. I presume that this sub- stance is secreted by the bacterium. Leucothrix grows best on penaeid shrimps when the shrimps are crowded and when there is a rich organic seawater medium. Salinities of 20-35%o and temperatures of 20°-25°C have been adequate for overgrowths of Leucothrix on gills of shrimp. Terminal gonidia were not searched for or ob- served in the fresh natural infestations on shrimp that I have studied with phase contrast, bright field, and electron microscopy. 8 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS The major adverse effect o^ Leucothrix infesta- tions on shrimp is probably interference with gas diffusion across gill cuticle, particularly in mas- sive infestations (Figure 7a, b). In experiments at my laboratory, I found that pink shrimp when exposed to various levels of an ethylene glycol- containing waste in bioassay systems had heavy growths of L. mucor on their gills, whereas nonex- posed, control shrimp had little or no growth on their gills. Mortality of the exposed shrimp was proportionate to the extent of growth of Leucothrix on their gills. Indications from EM studies are that the mucoid substance with which L. mucor at- tached to gills may cover gills (Figure 9a, b) in r%. . . ^ '. 0^ FIGURE 9. — a. Relationship of Leucothrix mucor filaments to gill cuticle of pink shrimp; note electron-opaque mucoid substance (etrrow) at point of attachment and adjacent to base of bacterium; no penetration of cuticle occurs, demonstrat- ing that the bacterium is not invaisive. X 14,400. b. Higher magnification ofL. mucor at point of attachment to gill; note distribution of electron-dense mucoid substance probably secreted by bacterium (arrow); shrimp cuticle is intact, x 28,500. 9b FISHERY BULLETIN: VOL. 76, NO. 1 heavy infestations. Massive amounts of this sub- stance overlying gill cuticle could block normal gas diffusion across gill surfaces. Fungi Our knowledge of fungal diseases of penaeid shrimps is in a state similar to that of our know- ledge of bacterial diseases. The only clear-cut case of a fungal pathogen affecting large numbers of penaeid shrimps in the United States was reported by Cook (1971) and by Lightner and Fontaine ( 1973). These authors described infections of white shrimp larvae by a phycomycete, Lagenidium sp., an estuarine fungus. The fungus infects the second protozoal stage of white shrimp, and disappears by the time the first mysis stage is reached. Figure 10a shows a heavily infected protozoea. According to Lightner and Fontaine (1973), the major pathogenic effect is almost complete tissue de- struction and replacement by invasive fungal mycelia (Figure 10a). Hyphae of the fungus are branched, septate, with thin walls, and range from 8.0 to 1 1 /u,m in diameter. Under bright field micro- scopy the hyphae were yellow-green and con- tained round oil droplets (Lightner and Fontaine 1973). The lifecycle of Lagenidium sp. in penaeid shrimps involves a sporulation phase. This begins when a hyphal extension penetrates the cuticle of the shrimp from within (Figure 10b). Following formation of a vesicle in the apical region of the extension, planonts (flagellated zoospores) are formed in the vesicle. The whole extension be- comes a discharge tube, releasing motile planonts (8.7-12 /xm) which presumably infect other shrimp. Lightner and Fontaine ( 1973) were able to infect larval brown shrimp (protozoea I) with planonts and hyphae on a large scale ( 2,000 larvae). Result- ing mortality in the experimentally infected shrimp was- 20%. Approximately 60 h were re- quired for infections to become patent. The role of this fungus in natural shrimp populations is not known. In aquaculture the fungus could be a definite limiting factor in the survival of shrimp larvae. Brown shrimp larvae in commercial hatcheries have been found to die of this disease (Cook 1971). The only other report of natural fungal infection in penaeid shrimps in the United States was that of Johnson.*' He briefly described a Fusarium species which infected the gills and antennal scales of Penaeus duorarum. Less than 5% of shrimp studied were infected and the spread of the fungal mycelium in the body of affected shrimp was slow. Solangi and Lightner (1976) have described the cellular inflammatory response ofPenaeus aztecus and P. setiferus to experimental infections of Fusarium sp. According to these authors, both species of shrimps showed "complete resistance to infection by the fungal spores when normal or wounded shrimp were held in seawater containing the spores or when spores were injected directly into the shrimp in low concentrations." Cellular "melanization" and encapsulation of the micro- and macroconidia occurred in gill tissues of penaeid shrimp. Only massive doses of 3.2 x 10^ spores injected into brown shrimp resulted in death of shrimp; this lethality was a result of mechanical blockage, by spores, of the blood sinuses of the shrimp's gills. Gills of affected shrimp sometimes were blackened. Protozoa More than any other phylum, the Protozoa as pathogens and parasites have had significant, known effects on shellfish populations. Represen- tatives of every class of Protozoa are found as sym- bionts, commensals, parasites, or pathogens in penaeid shrimps. Certain groups such as the Mi- crosporida have a long history as pathogens of not only penaeid shrimps, but arthropods in general. Only recently, however, species of such groups as the Ciliophora and the Sarcomastigophora have been indicted as serious pathogens of decapod Crustacea, including penaeid shrimps. Herein Protozoa associated with shrimps will be classified according to the scheme of the Honig- berg Committee in "A Revised Classification of the Phylum Protozoa" (Honigberg et al. 1964). Sprague and Couch (1971) published an annotated list of protozoan parasites, hyperparasites, and commensals of decapod Crustacea. This list in- cludes most of the known species of Protozoa as- sociated with penaeid shrimps. However, since its publication, several undescribed species have been found and will be included herein. Subphylum Sporozoa Leuckart 1879 This subphylum includes the gregarines and ^Johnson, S. K. 1974. Fusarium sp. in laboratory-held pink shrimp. Texas A&M Univ., Fish Disease Diagnostic Lab. Note FDDL-51, 1 p. 10 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Figure lO. — a. Lagemdium sp. hyphae throughout body of larval penaeid shrimp; note fungus has invaded antenna near eye (arrow). x300. b. Lagenidium sp. sporulation stage; note sporulation vesicle, filled with planonts, on end of hyphal extension (arrow) that has penetrated larval shrimp cuticle. x400. Figure ll. — Cephalohbus penaeus, gregarine trophonts attached to lappet of gastric mill from pink shrimp; note nucleus (arrow) and mode of attachment. xl50. Figure 12. — a.Nematopsis sp. trophonts in syzygy, from midgut of pink shrimp; arrows point to nuclei of trophonts. x900. b. Single, young trophont ofNematopsis sp. from gut of pink shrimp; note protomerite and septum separating it from rest of primite. X 1,000. 11 FISHERY BULLETIN: VOL. 76, NO. 1 coccidians. The only group considered here are the gregarines of penaeids. Gregarines, in general, are not highly pathogenic to their hosts. There- fore, information presented here is brief and the reader is referred to referenced works for details. Class Telosporea Schaudinn 1900 Subclass Gregarinia Defour 1828 Order Eugregarinida Leger 1900 Family Cephaloidophoridae Kamm 1922 Cephalolobus penaeus Kruse 1959 This species attaches to chitinous walls and terminal lappets of the stomach filter in Penaeus aztecus andP. duorarum (Figure 11). Usually the attached stage is a trophozoite consisting of a pri- mite with an anterior protomerite division that is modified into a holdfast organ. The single nucleus is in the center of the primite (Figure 11). The primite, including the protomerite, is from 100 to 200 /xm long. Often attached to the primite pos- teriorly will be 1 or 2 satellites (young tropho- zoites). Spores, sporozites, and cysts have not been observed. Overstreet (1973) reported this species in P. setiferus from Louisiana, extending its range from Florida as previously reported. I have ob- served this species in pink shrimp occasionally from Pensacola, Fla. This gregarine apparently has no harmful effect on the shrimp host. It may be possible that large numbers attached to the filter apparatus of the host could interfere with filtra- tion of particles bound for the hepatopancreatic ducts or passing through the stomach. Cephalolobus sp. Feigenbaum 1975 This form, reported from Penaeus brasiliensis, utilizes the stomach filter as position of attach- ment within host. Trophozoites consist of protom- erite and deutomerite separated by a septum. As in C. penaeus, the anterior end is modified into a holdfast organelle. This species has been reported in shrimp only from Biscayne Bay, Fla., and dif- fers from C. penaeus in that the trophozoites occur solitarily and are smaller (43-100 /xm long) than those of C. penaeus. Family Porosptjridae Labbe 1899 Nematopsis penaeus Sprague 1954 This species has been reported from brown, pink, and white shrimps. It is found in the intesti- nal tract. Figure 12a, b show specimens of 12 Nematopsis from the gut of a pink shrimp. These may be N. penaeus or N. duorari (see below). Works by Sprague (1954, see footnote 3), Sprague and Orr (1955), Kruse (1959, 1966), Button et al. (1959), and Hutton (1964) give information on hosts including the intermediate moUuscan hosts, for A^. penaeus. Overstreet (1973) discussed the prevalence and morphology of N. penaeus and pointed out that syzygy is multiple with up to seven trophozoites in line attached to one another reaching a length of over 0.5 mm. Characters for distinguishing A^. penaeus andN. duorari are size of gymnospore and number of different molluscan intermediate hosts. No pathogenesis is associated with this form. Nematopsis duorari Kruse 1966 This gregarine is restricted to the gut of pink shrimp. Kruse (1966) attempted to transmit it to brown and white shrimp, but could not. Figure 12a shows an immature association of a trophozoite of Nematopsis sp. in syzygy. Since two of the known Nematopsis species of penaeids appear identical in their trophozoite stages, no attempt will be made here to identify the specimens in Figure 12 to species. Nematopsis sp. Kruse 1966 Kruse (1966) described, but did not name, this species from concurrent infections ■w'lih.N. duorari in pink shrimp in Florida. This form had smaller gymnospores than did A^. duorari. Nematopsis brasiliensis Feigenbaum 1975 This is a recently described species oi Nematop- sis in a penaeid shrimp. Found in the intestine of Penaeus brasiliensis, this species consists of both individual trophozoites and syzygies of biassocia- tions (two trophs). It has been described from Bis- cayne Bay only. Hutton (1964) reported N. penaeus from P. brasiliensis. However, Feigen- baum (1973) believes that the species Hutton re- ported asN. penaeus may have been N. brasilien- sis. Subphylum Cnidospora Doflein 1901 Class Microsporea Corliss and Levine 1963 Order Microsporida Balbiani 1882 Microsporida are highly pathogenic to shrimps COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS and are probably one of the most destructive groups of pathogens to penaeid hosts. Rarely, however, have epizootics been recorded in which large numbers of penaeids have been lost to mi- crosporidan infections. Infection prevalences in samples of penaeids from nature and aquaculture rarely exceed 10'7c . Due to their highly pathogenic nature, however, emphasis is placed on the impor- tance of these protozoa to the health of penaeids. Table 3 summarizes salient characteristics of species of Microsporida discussed below. Kelley (1975) described histopathological changes in pink shrimp infected with Microsporida. Family Nosematidae Labbe 1899 Nosema nelsont Sprague 1950 This species is widespread, found in Penaeus duorarum, P. aztecus, and P. setiferus along the South Atlantic and Gulf coasts of the United States. The spores are found singly (one spore per sporont) in masses in infected tail muscle (Figure 13). As with certain other Microsporida, A^. nelsoni causes white discoloration of muscle or viscera giving infected shrimp a cotton or paper-white color (Figure 14). Fishermen call these shrimp "milk" or "cotton" shrimp. The spores of A^. nelsoni are 1.7 to 2.5 ixm long by 1.0 to 1.5 fj-m wide. Their polar filaments are 20 to 25 /Lim long. This parasite kills shrimp, and massive single infections with whole musculatures affected are found (Figure 15a, b). Thelohatiia penaei Sprague 1950 Members of this genus have eight spores in each sporocyst (Figure 16a, b). Found originally in the reproductive organs of Penaeus setiferus in Louisiana, this species has been reported from Mississippi, Texas, and Georgia. It infects muscle, gonads, and is seen grossly along the middorsal region of the abdomen and in appendages as white spots or clusters (Figures 17, 18). Spores are pyriform and occur in two size classes (2.0 to 5.0 /xm long and 5.0 to 8.2 /xm long). The polar fila- ment is unusual in that it has a thin distal half and a thick proximal half. Sprague (1970) reported that this is probably the microsporidan that Vio- sca (1943) observed in the reproductive organs of about 90*^ of P. setiferus along the Louisiana coast in 1919. This epizootic is one of the few reported in which penaeids have suffered en masse from a microsporidan. Viosca reported that the reproduc- tive organs of the white shrimp were destroyed by the parasite. Iversen and Kelly (1976) reported the first suc- cessful experimental transmission of a micro- sporidan {T. penaei) in shrimp. Postlarval pink shrimp fed T. penaei spores, conditioned by pas- sing through seatrout, showed tissue infections. Overstreet ( 1973) reported that pink and brown shrimps reared together in ponds showed only gill infections of T. penaei. Thelohania duorara Iversen and Manning 1959 This organism was first reported from Penaeus duorarum from the Dry Tortugas. A similar spe- cies has been reported from brown and white shrimps (Kruse 1959) in Florida. Overstreet (1973) reported that this species occurs in pink shrimp in the Mississippi Sound, and Iversen and Van Meter (1964) found it in P. brasiliensis in south Florida. Spores are 5.4 ixm x 3.6 /xm. This microsporidan parasitizes the muscle of shrimp causing white or "cotton" shrimp. The extent of impact it has on wild populations of penaeids is not understood. According to Sprague and Couch Table 3. — Characteristics of Microsporida in penaeid shrimps. Spores/sporont Spore size Species (averages) (Mm) Tissues Host(s) Locales Nosema nelsoni 1 2.0 X 1.2 Muscle P. aztecus Gulf coast Sprague 1 950 P duorarum P. setiferus Georgia coast Thedohania penaei 8 2.0 X 5.0 Gonads P. setiferus Gulf coast Sprague 1950 5.0 X 8.2 Muscle Georgia coast Tf^eolohania duoara 8 5.4 X 3.6 Muscle P. aztecus Gulf coast Iversen and Manning 1959 P duorarum P. setiferus Florida east coast Pleistophora sp. 16 to 40 + 2.6 X 2.1 Muscle P aztecus Gulf coast Baxter el al. 1970 Heart P setiferus Constransitch 1970 Gills P. duorarum Southeast Rorida Kruse (in Hepatopancreas sprague 1970) Iversen and Kelly 1976 13 FISHERY BULLETIN: VOL. 76, NO. 1 '# y / f » Figure 13. — Nosema nelsoni spores in fresh squash preparation of muscle from pink shrimp, x 1,500. Figure 14. — White or cotton appearance of organs and muscle of penaeid shrimp infected \N\ih. Nosema nelsoni, and Thelohania penaei; note opaque white appearance of gonads (arrow). Figure 15. — a. Abdominal musculature heavily infected with Nosema nelsoni; note long spore masses between and around every muscle bundle (arrows). xlOO. b. Higher magnification of spore masses of A^osema in histological section of muscle. x500. (1971), Thelohania hunterae (a nomen nudum) was probably T. duorara. Roth and Iversen (1971) reported attempts to transmit T. duorara to uninfected pink shrimp in the laboratory. They were unable to do this with their method of feeding heavily infected tissue. These authors did supply some clues as to the possible modes of transmission in nature. They observed that spores of T. duorara found between old cuticle and new cuticle at time of molting could infect shrimp that feed on cast cuticles. Therefore, 14 transmission could depend only on molting of the exoskeleton and not on death of the infected host. Iversen and Kelly ( 1976) have reported concur- rent infections of T. duorara and T. penaei in single specimens of pink shrimp. Pleistophora ( = Plhtophora) penaei Constransitch 1970 Members of this genus are characterized by sporocysts that contain 16 or more spores. Kruse COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Figure 16. — a. Thelohania penaei sporocysts and spores; note approximate size of sporocysts with eight spores each; dark bodies are trophozoites or early sporonts. x 1 ,000. b. Thelohania penaei sporocysts, higher magnification; note that each sporocyst contains about eight spores; dark body (arrow) is probably an early sporont or trophozoite of this species, x 1,500. (in Sprague 1970) first reported the genus Pleis- tophora in penaeid shrimps (Penaeus aztecus and P. setiferus from Louisiana). Constransitch (1970) named the species from Louisiana Pleistophora penaei. Tissues infected were tail muscle, cardiac muscle, hepatopancreas, and intestinal and stomach walls. Baxter et al. ( 1970) then reported a similar species from the same hosts from Texas. The Texas Pleistophora consisted of sporocysts that contained 40 or more spores. Recently, Iversen and Kelly (1976) reported a Pleistophora sp. from the pink shrimp for the first time. Therapeutic Measures for Microsporidosis Very little work has been done on attempting to control or treat microsporidan infection in reared shrimp. Quick removal of "cotton" or obviously infected shrimp from tanks or ponds should aid in preventing spread of infections. Overstreet (1975) has reported some success in treating blue crabs with the drug Buquinolate to prevent infection by Nosema michaelis, a common microsporidan in blue crabs. He fed the drug to crabs in food con- taminated with A^. michaelis spores. He also fed the drug in food without spores 48 h preceding or following spore feeding. Control crabs were fed spores, but no drug. Drug and spore-fed blue crabs had significantly fewer infections develop than did crabs fed spores only. Whether Buquinolate or other drugs would be helpful in preventing mi- crosporidosis in shrimp remains to be determined. Even if a drug is useful in therapy of a disease in 15 FISHERY BULLETIN: VOL. 76, NO. 1 Figure 17. — Whole shrimp showing dorsal areas of white that indicate microsporidan infection (arrows), in this case, Thelohania penaei. FIGURE 18.— White clusters of Thelohania penaei sporocysts in antennal scale of pink shrimp (arrow). 16 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS cultured shrimp, the problem remains for depura- tion of the drug from tissues prior to human con- sumption of the shrimp. Subphylum Ciliophora Doflein 1901' Ciliate Protozoa are very common associates of penaeids. As commensals, parasites, and patho- gens, they are among the Protozoa more often en- countered in or attached to penaeid shrimp. Their role, however, in the health of penaeids has not been conclusively demonstrated in most ciliate- penaeid relationships. Sprague and Couch (1971) presented a list of ciliates (and other Protozoa) found on or in decapod Crustacea. Since that re- port, several new finds of ciliates in penaeid shrimps have been made. Ciliates discussed herein will be presented in order of their frequency of occurrence in penaeid shrimps (common to rare). Class Ciliatea Percy 1852 Order Peritrichida Stein 1859 Suborder Sessilina Kahl 1933 Famih Vorticellidae Ehrenberg 1838 Genus Zootbanniin>n Bory 1826 Zoothamnitim sp. An heretofore undescribed species of peritri- chous ciliate, of the genus Zoothamnium, has been reported on penaeid shrimps along the coast of the southeastern United States Villella et al. (1970), Overstreet ( 1973), Johnson (1974), D. V. Lightner (pers. commun.), and I have found the colonial, stalked peritrich to be very common and fre- quently abundant on the gills of three commer- cially valuable species of penaeid shrimps. Stalked peritrichs of the genera Vorticella, Zoothamnium, Epistylis, Carchesium, Rhabdos- tyla, and Opisthostyla are found attached to many hard substrates in the marine environment. The vast majority of species in these genera have not been studied, described, and named. Therefore, with this background in mind, I propose to de- scribe, but not to formally name, the common species of Zoothamnium on gills and body of adults, juveniles, protozoea, and mysis ofPenaeus aztecus, P. setiferus, andP. duorarum. This species will be named after further study and comparison with other species in the genus Zoothamnium. ^Most ciliatologists and many protozoologists now consider the Ciliophora as a phylum, but herein the Honigberg et al. (1964) classification scheme is followed. Description. Vorticellid; colonial, rarely ob- served as individuals; 3 to 30 trophonts per colony (Figure 19); usually attached to the tips of gill filaments of hosts listed above; trophonts variable in form but usually resemble an inverted bell (45.2 )u.m X 33.9 ^(-m — means of measurement of 30 in- dividuals); with long, branching stalks (8.1 /Am in diameter); phase contrast and silver-stained (pro- targol) specimens show that myonemes in stalks are continuous and joined, and the diameter of myonemes averages 2.0 jum (Figure 20a, b). Silver-stained specimens (Figure 20c) also reveal adoral kineties consisting of a three-component polykinety (peniculus) and a haplokinety; telo- troch (Figure 21) produced by division of stalked trophont, slightly smaller than stalked trophont; lifecycle direct, that is, the telotroch may swim free of mother colony and attach to surface of gill or body of shrimp, secrete a stalk, and become progenitor of a colony; sexual reproductive cycle not observed for this species, but probably is a conjugative process as in other peritrichs having microconjugants and macroconjugants. I have ob- served only pairs and small colonies (3, 4 trophonts) oi Zoothamnium sp. attached to body surfaces of larval (mysis and protozoea) brown shrimp. Overstreet (1973) gave extensive data on the frequency of occurrence of Zoothamnium on penaeid shrimps. He found that an increase in density of hosts held in captivity was paralleled by an increase in density of peritrichs on gills. This is similar to what Couch (1971) observed for blue crabs infested -withLagenophrys callinectes Couch (1967), a gill peritrich. Overstreet ( 1973) also was able to correlate, positively in one test, increased mortality in shrimp with heavy infestations by Zoothamnium on their gills. However, he was not convinced that the correlation was valid. More extensive work on this relationship is needed. The mechanism of injury to penaeids infested with peritrichous ciliates would probably be oxy- gen starvation or asphyxiation due to blockage of gas exchange at the gill surface. The attachment stalk oi Zoothamnium sp. does not penetrate the cuticle of shrimp. Famih Lagenophryidae Kahl 1935 Genus Lagenophrys Stein 1852 Lagenophty liinatiis, Imamura 1940 A species of Lagenophrys was reported from the cuticle of Penaeus setiferus by Johnson (1974) and 17 FISHERY BULLETIN: VOL. 76, NO. 1 y 20b 20c Figure 19. — Colonies of Zoothamnium sp. attached to end of gill filaments in pink shrimp; this represents a light infestation; heavy infections would cover all filaments. x200. Figure 20. — a. Phase contrast photomicrograph of Zoothamnium colony showing stalk myonemes (M) that are continuous with one another, the major distinguishing characteristic of the genus; note inverted bell shape of contracted trophonts (T) and thickness of stalk sheath that surround myonemes (arrow). x500. b. Protargol treated specimens of Zoothamnium; note beltlike, horseshoe-shaped macronucleus and Protargol -positive myonemes of stalk (arrow). xl,200. c. Protargol-treated Zoothamnium; trophont ( in focus) shows peniculus (P) in infundibulum ( arrow); note pattern of kineties (K) making up peniculus. X 1,200. 18 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Figure 21.— Telotroch stage (arrow) of Zoothamnium produced from divi- sion of trophont (phase contrast); this is the dispersal stage for the species; the telotroch is motile and possesses a ventral girdle of cilia. Note the trophont at upper right with extended adoral ciliature (arrow). x500. by Lightner (1975) in Texas. From a photomicro- graph Icindly loaned to me by Johnson, I have tentatively identified this loricate peritrich as Lagenophrys lunatus. This species is commonly found on the cuticle of paleomonid shrimps along the east coast and gulf coast of the United States, but Johnson's report, if accurate, is the first for a penaeid. It is possible that the species of shrimp examined by Johnson was a grass shrimp, Paleomonetes sp. Species of Lagenophrys are usu- ally host specific, and though I have examined many penaeid shrimps, I have not observed Lagenophrys sp. on any. Couch (1971) gave a de- tailed discussion of the possible effects of Lagenophrys spp. on the cuticles and gills of de- capod Crustacea with particular reference to L. callinectes on the gills of the blue crab, Callinectes sapidus. Erosion of cuticle surface and interfer- ence with gas exchange at the gill surface in heavy infestations are possible effects of Lagenophrys. Order Apostomatida Chatton and Lwoff 1928 Family Foettingeriidae Chatton 1911 Genus Uncertain The encysted form (phoront) of an undescribed apostome ciliate has been observed on the gills of Penaeus duorarum (Figures 22, 23) in northwest Florida. The cysts are decumbent, ellipsoidal bodies that are 41 jxra wide by 60 /u,m long (range: 20.7-41.4 /Ltm by 27.6-60.0 /xm). The cyst wall is from 1 to 3 /xm thick and is semitransparent. Heavy infestations of this ciliate occur on gills of pink shrimp during periods of warm to moderately cool weather when shrimp are held under crowded conditions (Figure 22). The cysts are most often attached to the gills at the point of branching of the distal processes variously termed lamellae, filaments, or tertiary structures (Figure 22). The lifecycle of this ciliate has not been elucidated, and it cannot be assigned to a genus until silver- 19 FISHERY BULLETIN: VOL. 76, NO. 1 Figure 22. — Cysts (phoront) of apostome ci Hate (arrows) on gills of pink shrimp; this is a moderately heavy infestation, x 150. Figure 23. — Single cyst (phoront) of unidentified apostome attached near base of gill filament; note ellipsoid form, x 1,000. 20 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS staining studies and lifecycle studies are more complete. Reports by Chatton (1911), Chatton and Lwoff (1935), Debaisieux (1960), and Bradbury (1966, 1973) have demonstrated the common occurrence of apostomes on Crustacea that occupy ecological niches near that of the pink shrimp. The present species has not been found associated with mortal- ity in shrimp, although severe infestations may cover much gill surface and blackened areas of infested gills are found. Species of two known apostome genera, Synophyra and Terebrospira, cause considerable damage by penetrating the cuticle of their crustacean hosts (Chatton and Lwoff 1926; Bradbury 1974). R. M. Overstreet (pers. commun.) has found similar cysts on gills of brown and white shrimp, and Feigenbaum (1973) reported cysts similar to those described above on gills of Penaeus brasiliensis from Biscayne Bay. The cysts of apos- tomes could be confused with the loricae of species of Lagenophrys, Care should be taken to distin- guish them. Loricae of Lagenophrys spp. have apertures surrounded by liplike structures (Couch 1973). Order Scuticociliatida Small 1967 Genus Parauronema Thompson 1967 Parauronetua sp. An undescribed species of ciliate was observed in the hemocoel of protozoeal, mysid, and juvenile stages of living, moribund, and dead brown shrimp from a mass mortality which occurred at a com- mercial shrimp hatchery^ during April 1974. In a sample of 139 larvae examined, 28.8% were in- fected by the ciliate (Tables 1, 2). The ciliate is ovoid to pyriform in shape, ranging in length from 23.6 to 31.6 ;u.m, and in width from 9.2 to 12.2 ju.m (Figures 24, 25). It has a uniform body ciliature originating from longitudinal kineties (Figure 25) as revealed by Protargol silver staining. The ciliate was observed swimming about in hemolymph of infected shrimp larvae and juveniles. Often the affected shrimp were still alive and active, but several that were dead or quite moribund contained ciliates. John Corliss (University of Maryland) tentatively identified the ciliate as a species of Parauronema. More studies are required in order to name this ciliate. ^Mortality was that reported on preceding pages (under virus section). Several microorganisms were associated with this mor- tality. Apparently the ciliate causes mechanical injury in infected shrimp by replacing and dislodging tissues. I have been unable to determine from lim- ited observations whether or not the ciliate is his- tophagous. In some shrimp the ciliates were numerous enough to fill the entire hemocoel and abdomen. The fact that living shrimp larvae were infected by the ciliates strongly suggests that the ciliate probably contributes to pathogenesis and mortality and that it is an opportunistic invader following initial breaks in the host's defense mechanisms due, possibly, to the presence of other pathogenic microorganisms {the Baculovirus and a flagellate to be described next). Tables 1 and 2 show the relationship of prevalence of ciliate with virus and flagellate in a sample of young brown shrimp from a stock suffering mortality. Subclass Suctoria Haeckel 1866 Order Suctorida Claparede and Lachmann 1858 Family Ephelotidae Kent 1881 Genus Ephelota Wright 1858 Ephelota sp. Protozoeal and mysid stages of brown shrimp were found infested on a single occasion with an undescribed species of Ephelota. The larval shrimp were examined in March. Each larva had from one to seven individual Ep/ze/ota sp. attached to their cuticles usually on the pleural plates or on the telson. The suctorian possesses a characteris- tically striated attachment stalk and a trophont with both suctorial and prehensile tentacles. These Protozoa were not abundant enough to cause embarrassment to the larval shrimp. Subphylum Sarcomastigophora Honigberg and Balamuth 1963 Class Zoomastigophorea Calkins 1909 Order Kinetoplastida Honigberg 1963 Suborder Trypanosomatina Kent 1880 Family Trypanosomatidae Doflein 1901 Genus Leptomonas Kent 1880 Leptornonas sp. An undescribed species of flagellate was as- sociated with the mass mortality of brown shrimp larvae (see Baculovirus and Parauronema sec- tions) (Figure 26). This form is tentatively as- signed to the genus Leptomonas based on sub- sequently described characteristics. The flagellate was studied alive (bright field and phase contrast), fixed, and stained with Harris' hematoxylin and 21 FISHERY BULLETIN: VOL. 76, NO. 1 ■***^ 27 28 Figure 24. — Trophont o{ cihate, Parauronema sp., in hemocoel of browTi shrimp lai^a; note body form and longitudinal rows of kinetosomes on body surface (arrows) (Protargol). x 1,300. Figure 25. — Two trophs of Parauronema sp. in body of brown shrimp larva; in living shrimp these ciliates swim about in hemolymph. x900. Figure 26. — Cells of Leptomonas sp., a flagellate, from hemolymph space in appendage of larval brown shrimp; note flagellar base as revealed by Protargol stain (arrow); compact nucleus is also visible, x 1,000. Figure 27. — Head and anterior appendages of larval brown shrimp heavily infected with Leptomonas sp.; note antennae, antennules, and thoracic legs filled with flagellate (arrows). Figure 28. — Cystlike stages of Leptomonas in hemocoel of larval brown shrimp (Protargol). xl,000. 22 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Protargol silver protein. It is the first flagellate reported to be associated with shrimp mortalities. The flagellate occurred in the hemocoel, abdomen, and all appendages of protozoel and mysid stages of brown shrimp during April 1974 (Figure 27). The flagellate was found in 649f of larvae examined from the mortality; living, moribund, and dead larvae were infected (Tables 1, 2). The flagellates were variable in form ranging from 7.8 to 11.7 ^im with an average diameter of 9.4 yLtm. A compact nucleus (2 or 3 /um) containing a large endosome was situated medianly. The cytoplasm ranged from clear to opaque and often contained various inclusions. In life, the flagellate was slightly pyriform with a terminal, single flagellum (Figure 29). Specimens stained with protargol clearly demonstrated a flagellar base, parabasal body, or blepharoplast (karyomastig- ont) (Figures 26, 29). A possible cyst stage (7-9 /u.m) was observed in advanced or heavy infections in the hemocoel (Figures 28, 29f). Dividing stages, observed occasionally, contained nuclei undergo- ing division without loss of nuclear membranes (Figures 29e). The role, if any, that Leptomonas sp. plays in the mortality of shrimp larvae is unknown. Other than mechanical damage, there appears to be lit- tle evidence of a pathogenic mechanism for the flagellate. It is possible that the flagellate is a secondary invader of a weakened host, possibly from encysted forms which may exist in the hindgut of the host. Platyhelminthes Flatworms have been described as parasites of all commercial species of penaeid shrimps in the United States. These include digenetic trematodes Figure 29. — a. Leptomonas sp. drawn from life with flagellum. b, c, d. Forms of the flagellate (possibly amastigote stages) as they appear in Protargol-stained body (hemolymph) of brown shrimp, e. Cell division in flagellate showing karyokinesis and longitudinal cytoplasmic fission, f. Possible cyst stage Lep/omonas from hemocoel of larval shrimp. Note Protargol-positive kinetoplast near nucleus (arrow points to kinetoplast). (All figures x2,900.) 23 FISHERY BULLETIN: VOL. 76, NO, 1 and cestodes. The role of these worms as agents of disease in shrimps is uncertain. Most of the species reported, to date, appear to have little effect on individual shrimp infested, and probably little significant effect on populations of penaeids. How- ever, flatworms in penaeid shrimps are often con- spicuous and, thus, attract considerable attention. Penaeid shrimp usually play the role of inter- mediate host for most, if not all, flatworms they harbor; therefore, shrimps play a significant role in the ecology of parasites that may be transmitted through the food web to higher vertebrate hosts. Class Trematoda Rudolphi 1808 Subclass Digenea Carus 1863 Famly Microphallidae (Travassos 1920) Genus Microphallm Ward 1901 Microphalltis sp. Hutton et al. (1959) reported an undescribed species of microphallid trematode metacercariae from pink shrimp. They found that from two to three metacercarial cysts up to hundreds (from 1.2 to 1.5 mm in diameter) were encysted in muscle tissue surrounding internal organs, particularly the cephalothoracic and abdominal musculature. No effect on the shrimp host was reported. Overstreet (1973) also reported an unidentified microphallid metacercaria from abdominal mus- cles of white shrimp from Barataria Bay, La. The cysts were 93-95 ;u,m to 77-83 ixm, much smaller than those reported from pink shrimp from west Florida by Hutton et al. (1959). Family Opecoelidae Ozaki 1925 Genus Opecoeloides (Odhner 1928) Opecoeloidei finihriatus (Linton 1934) Sogandares-Bernal and Hutton 1959 Metacercariae of this trematode (Figure 30) en- cyst in hepatopancreas, other internal organs, and beneath the exoskeleton ofPenaeus duorarum, P. setiferus, and P. aztecus. This is a very common parasite of penaeids, occurring in up to 90% of some samples of pink shrimp taken during the summer from Apalachee Bay, Fla. No extreme pathogenesis in shrimp has been reported as- sociated with O. fimbriatus. The worm is approxi- mately 1.5 to 2.0 mm long when excysted and is quickly identified by its possession of an extremely pedunculate acetabulum (Figure 30). The sexu- ally mature worm (adult) is found mostly in fishes of the family Sciaenidae which feed on shrimps. The metacercaria is found in penaeids from the Gulf and Georgia coasts. Class Cestoidea Rudolphi 1809 Order Trypanorhyncha Diesing 1863 Family Eutetrarhynchidae Guiart 1927 Genus Prochristianella Dolfus 1946 Prochriitiatiella hispida (Linton 1890) Campbell and Carvajal 1975 Synonyms: Khynchohothrittm hispidum Linton 1890; P. penaei Kruse 1959 Plerocercoid larvae of this tapeworm are very common in Penaeus setiferus, P. duorarum, and P. aztecus. I have found up to 95% of large samples of P. duorarum from northwest Florida to harbor the cestode. This cestode is found mainly in the hepatopancreas of the host (Figure 31), and most often fails to elicit any strong pathologic response from the shrimp. Sparks and Fontaine (1973) and Feigenbaum and Carnuccio (1976) reported a strong host reponse to the plerocercoid when it encysted in hepatopancreas. I have not observed this in several hundred hosts examined, but host destruction of trypanorhynchan plerocerci may occur rarely in shrimp. Most evidence suggests a long and relatively tolerant relationship between shrimp and cestode. Often a single shrimp will have one to two dozen encysted larvae in its hepatopancreas. According to my measurements, the worm (Fig- ure 32a, b) has the following mean dimensions: length — 1.12 mm; bladder or blastocyst = 0.58 mm long by 0.37 mm wide; and scolex ( below both- ridia) =0.11 mm wide by 0.35 mm long. These measurements are close to those of Kruse's (1959) description. Though no lifecycle has been experi- mentally completed for a trypanorhynchan, the hosts for adult worms of this group are probably sharks and rays. From nature, cestodes of this order have been found in the spiral valves of elas- mobranchii (Kruse 1959). Aldrich (1965) and Ragan and Aldrich ( 1972) gave host-parasite data on this species. Parachristianella monomegacantha Kruse 1959 P. diniegacantha Kruse 1959 Kruse (1959) described two other trypanorhyn- chan plerocercoid larvae from Penaeus duorarum. These species were found in the hepatopancreas of shrimp from the northern gulf coast and are dis- tinct from one another "in hook arrangement and 24 COUCH; DISEASES AND PARASITES OF PENAEID SHRIMPS M 31 32b ^J> V* 33c 4," • r *' *C. >i. ♦ « • % 33b Figure 30. — Metacercaria oWpecoeloides fimbriatus, digenetic trematode, from hepatopancreas or hemocoel of adult pink shrimp. This species is quickly identified by its large, pedunculate acetabulum (arrow). x70. FIGURE 31. — Section of plerocercoid larva of Prochristianella hispida encysted in hepatopancreas of pink shrimp; note cyst wall and lack of host cellular response (Feulgen picro-methyl blue stain). x50. Figure 32. — a. Fresh wet mount of plerocercus of P. hispida; note scolex and blastocyst. x50. b. Scolex ofP. hispida; note tentacles (T) and bothria (B) (au-rows). x50. Figure 33. — a. Larvae of an unidentified cestode commonly found in hemocoel of penaeid shrimps; this figure shows a mass of larvae against the midgut lining (dark line). x25. b. Unidentified cestode larvae showing calcareous corpuscles and large sucker (arrows). x25. 25 FISHERY BULLETIN: VOL. 76, NO. 1 in the relative sizes of their bothridia, bulbs, and post-bulbosal regions." The genus differs from Prochristianella in the morphology of the blastocyst; species of the latter genus having a division between anterior and posterior portions, with large granules contained in the anterior division of the blastocyst. These worms apparently do not harm their hosts sig- nificantly. Pa rachriitia nella heterotnegaca nth lis Feigenbaum 1975 The most recent species to be described is from Penaeus brasiliensis from Biscayne Bay. Twenty percent of this shrimp were infected with fewer than 1.5 worms occurring in each infected shrimp. Corkern ( 1970) found an average of 2.3 specimens of P. dimegacantha per infected brown shrimp from Galveston Bay, Tex. Prevalence data from Corkern's work shows 239^ brown shrimp infected, a figure close to that of Feigenbaum's (1975) 20% for P. heteromegacanthus. Tentacle hook ar- rangements in P. heteromegacantha differed from those in P. monomegacantha and P. dimega- cantha. Family Renibulbidae Feigenbaum 1975 1 Genus Renihiilhiis Feigenbaum 1975 Renihulhus penaeus Feigenbaum 1975 To date, this species was found in 14.3% ofPen- aeus brasiliensis examined from Biscayne Bay. The short kidney-shaped bulbs in the scolex of this cestode set it apart from other trypanorhynchan cestodes in penaeid hosts. No organ site of infec- , tion was given by Feigenbaum (1975) for this worm, and no pathogenesis was indicated. Unknown Cestode Larva Hutton et al. (1959), Kruse (1959), Overstreet (1973), Feigenbaum (1975), and I have found a small pyriform cestode larval stage ( Figure 33a, b) commonly in the intestine of penaeid shrimps from the Gulf and Atlantic coasts of Florida. This worm also is found in large numbers in several tissues of infected shrimp, namely, the muscles and hemocoel. The worm possesses a large an- terior sucker and many refringent calcareous cor- puscles in its body, and is approximately 0.61 to 0.81 mm long by 0.12 to 0.22 mm wide. Large numbers of this worm may occlude the intestinal lumen or cause perforation of the intestinal wall. Several hundred larvae have been counted in a single shrimp. Hosts, to date, include Penaeus duorarum, P. aztecus, P. setiferus, and P. brasiliensis. Nematodes Phylum Aschelminthes Grobben 1910 Class Nematoda (Rudolphi 1809) Cobb 1919 Superfamily Ascaridoidea (Railliet and Henr> 1915) Genus Thynnascaris Dolfus 1933 Thynnascaris sp. Overstreet (1973) reported that the nematode larvae identified by Kruse (1959), Hutton et al. ( 1959), and Corkern ( 1970) as Contracaecum sp. in penaeid shrimps should be considered species of Thynnascaris. Norris and Overstreet (1976) have found that at least two species occur in penaeid shrimps in North America. Characteristics of this genus are short intestinal caecum and longer ven- tricular appendix combined with the position of the excretory pore near the nerve ring. Figures 34 and 35 are photomicrographs of Thynnascaris sp. recovered from hepatopancreas and cephalo- thorax of Penaeus duorarum near Pensacola. I have not found it commonly in shrimp from west Florida, but Overstreet (1973) reported that Donald Norris of his laboratory found up to 31% of white and brown shrimp from Mississippi Sound and adjacent waters infected during summer months. Thynnascaris sp. juveniles measure 1.02 to 2.40 mm long by 0.06 to 0.10 mm wide." Overstreet (1973) reported two specimens of Spirocamallanus pereirai Olsen 1952, in the intes- tine of Penaeus setiferus from near Biloxi, Miss. These were third stage larval nematodes which measured 1.00 mm long by 0.03 mm wide. Over- street suggested that the shrimp may serve as a paratenic host and that copepods may serve as a more common source or vector for this nematode which normally matures in fishes. Several species of free-living nematodes, com- monly found in shrimp habitat, have been re- ported as facultative commensals or inquilines of penaeids. Shrimp may take these worms in larval stages when they feed on detritus or bottom or- ganisms in nature or in artificial ponds. Speci- mens of Leptolaimus sp. and Croconema sp. have been found by Overstreet (1973) in brown and white shrimps from Mississippi. Other than phys- 26 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS X p'.i. . .^ ^*r ^«jt#«"^e»^ fi r"- / 34 Figure 34. — Thynnascaris sp. larvae in tissue squash from pink shrimp. Whole worm larva in view; note cellular arrangement at posterior of worm (arrow). x50. Figure 35. — Higher magnification of Thynnascaris sp.; note the intestinal caecum that turns anterior from the intestine (arrow). xlOO. 27 FISHERY BULLETIN: VOL 76, NO. 1 ical disruption of tissues, no mechanism of pathogenesis is apparent for nematodes in shrimp. NONINFECTIOUS DISEASES Toxic Responses In the last decade, because of interest in aquatic pollution, some research has been done on toxic responses of penaeid shrimps to a variety of chem- icals and heavy metals. Most of this w^ork has been done in pollution-oriented laboratories; however, few attempts have been made to apply results to interpretation of field conditions. Results obtained have been reported mostly as toxicity of specific chemical agents in terms of short-term lethality or longer-term mortality. Unfortunately, little indi- cative cellular or tissue changes caused by toxi- cants has been described for penaeid shrimps. I shall divide this section into categories of toxi- cants that have been tested or studied in penaeids. The following categories will be covered: or- ganochlorines, organophosphates, carbamates, oil or petroleum products, heavy metals, and chemo- therapeutic chemicals. Organochlorines Since World War II many kinds of pesticides and industrial chemicals containing or consisting of chlorinated hydrocarbons have been inadver- tently or intentionally released into the envi- ronment. Aquatic life is exposed to these com- pounds because the aquatic portion of the biosphere often behaves as a "sink" or receptacle for these compounds due to runoff or fallout. Some Table 4. — Comparative toxicity of pesticides to three estuarine taxa — most sensitive (1) to least sensitive (3).' Pesticide Penaeid shrimp Fish Oysters Chlordane DDT Dieldrin Endrin Heptachlor Toxaphene 2 2 2 3 2 3 1 3 3 2 3 2 3 Guthion Malathlon Parathion 2 2 2 3 3 3 Carbaryl Carbofuran 2 2 3 3 2,4-D (BEE) Atrazlne 3 1 2 2 3 3 Du-ter Difolatan 3 3 2 2 1 1 ' This table was prepared by Jack I. Lowe who graciously granted permission for Its use here The table has not been published previously of these compounds or their metabolites are re- fractory to breakdown, and thus tend to accumu- late in various compartments of the aquatic envi- ronment. Experimental shrimp have been found to accumulate certain chlorinated compounds in the laboratory and feral shrimp have possessed detectable levels when taken directly from con- taminated or apparently "clean" waters. Jack Lowe of the USEPA Laboratory, Gulf Breeze, has found, over several years of testing, that penaeid shrimps generally are far more sensitive to toxic effects of most insecticides than are fishes or mol- lusks (Table 4). The effects of some of the better known compounds will be reviewed here. DDT White shrimp, which died as a result of DDT exposure, accumulated up to 40.40 ppm DDT and DDE in hepatopancreas after 18 days exposure to 0.20 ppb in flowing seawater (Nimmo et al. 1970). Exposure to DDT concentrations greater than 0.10 ppb was lethal to pink shrimp in 28 days. A physiological effect of DDT exposure in pink and brown shrimps was loss of certain cations in the hepatopancreas (Nimmo and Blackman 1972). Sodium and potassium concentrations in shrimp exposed to 0.05 ppb DDT for 20 days were lower than in those not exposed. Magnesium, however, was not significantly lowered. The significance of reduced cations in the hepatopancreas of shrimp for the pathophysiological behavior of shrimp is not known. Blood protein levels also have been found to drop in shrimp exposed to DDT. There are no reports of histopathological changes in penaeids following exposure to DDT. In acute, high-concentration exposures, shrimp showed tremors, hyperkinetic behavior, and paralysis, classic signs of DDT poisoning in arthropods. After extended exposure to low concentrations of DDT, shrimp did not become paralyzed, but sank into lethargy, refused food, and then died. Dieldrin Pink shrimp were more sensitive to dieldrin than were grass shrimp in test exposures. How- ever, both species died when exposed to concentra- tions of dieldrin in the low parts-per-billion range. Pink shrimp had a 96-h LC^y of 0.9 ppb dieldrin (Parrish et al. 1973). No histopathological effects of dieldrin in penaeid shrimps have been re- ported. 28 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Mirex Juvenile pink and brown shrimps died after ex- posure to low concentrations of mirex. Twenty-five percent of a sample of pink shrimp died during 7 days exposure to 1.0 ppb mirex. However, all sur- vivors from this test died after 4 days in mirex-free seawater, demonstrating a delayed toxic effect of mirex (Lowe et al. 1971). I have examined both shrimp and blue crabs exposed to low concentrations of mirex for long periods (30 days or more) for histopathological ef- fects. No pathologic effects at the tissue level were found in the animals which I examined. Organs studied were muscle, hepatopancreas, and gonads. PCBs (Polychlorinated Biphenyls) These industrial chemicals have been at large in the aquatic environment for many years due to leakage from water and waste effluents, disposal of dielectric fluids, and other industrial sources (Broadhurst 1972). It is a well-established fact that certain fresh and marine bodies of water are contaminated with various compounds of PCB (Sodergren et al. 1972; Nimmo, Blackman, Wil- son, and Forester 1971; Nimmo, Wilson, Blackman, and Wilson 1971; Nimmo et al. 1975). As recently as 1970, Duke et al. reported PCB, Aroclor 1254, in water, sediments, and tissue of animals (including penaeid shrimps) from Escam- bia Bay, near Pensacola. At the U.S. Environmental Protection Agency Laboratory (Gulf Breeze, Fla.), much research has been done on the effects of PCB's on estuarine species with emphasis on pink and brown shrimps. These two penaeids were killed in 2-wk exposures to 0.9, 1.4, and 4.0 ppb Aroclor 1254 in flowing seawater. The minimum level causing mortality was 0.9 ppb. Penaeid shrimps appeared to suffer greatest mortality when exposed during premolt (just before molting) and during molt. Most ex- posed shrimp became lethargic, stopped feeding, and did not dig into the substrate (digging is a normal activity for penaeids). Subtle to dramatic chromatophore changes in the cuticle of exposed shrimp were more frequent and obvious than in control shrimp. On the light microscopical level, no lesions were consistently found that were indicative of PCB exposure in shrimp (Couch and Nimmo 1974a). However, several interesting cytopathic changes were noted in exposed shrimp studied with EM. Pink shrimp were exposed to 3 ppb Aroclor 1254 in flowing seawater for 30 to 52 days. During these exposures, up to 50'7f of the animals died. Living and dead shrimp were analyzed by gas chromatog- raphy and from 33 ppm to 40 ppm Aroclor 1254 was found in their hepatopancreatic tissues. Aro- clor uptake in hepatopancreas was linear with time (Couch and Nimmo 1974b). Hepatopancreas was fixed and processed for EM. Hepatopancreatic absorptive cells from exposed shrimp revealed the following departures from those of controls: 1 ) 30 to 50*^ of cells had increased or proliferated rough endoplasmic reticulum (Figure 36); 2) production of membrane whorls with enclosed lipid droplets (Figure 37); and 3) nuclear degeneration charac- terized by the occurrence of vesicles in the nu- cleoplasm (20-50 nm and 100-700 nm in diameter) (Figure 38a, b). The proliferation of smooth endoplasmic re- ticulum in hepatocytes of higher animals has been described as indicative of toxic responses to drugs or chemicals such as phenobarbitol, dilantin, diel- drin, and carbon tetrachloride. This proliferation has been related to detoxification of poisons and may, in shrimp, represent an attempt, on the part of hepatopancreatic cells, to metabolize PCB ab- sorbed from the lumen of hepatopancreatic ducts. If this is the case, cellular alterations at the ultra- structural level may be valuable as early indi- cators of sublethal effects of certain pollutants in penaeid shrimps. Another PCB, Aroclor 1016, has been more re- cently introduced for limited use in the United States. This compound has been tested for toxicity in brown shrimp. Aroclor 1016 was found to have nearly the same toxicity for penaeid shrimp as Aroclor 1254: 0.9 ppb Aroclor 1016 in flowing sea- water killed 87c of test shrimp in 96 h; 10 ppb Aroclor 1016 killed 43'7f of test shrimp in 96 h (Hansen, Parrish, and Forester 1974). It is apparent from research results now pub- lished that PCB's as pollutants pose a threat to penaeid shrimps which show a high level of sen- sitivity to these compounds. In this regard, Nimmo, Blackman, Wilson, and Forester (1971) and Nimmo, Wilson, Blackman, and Wilson (1971) demonstrated that pink shrimp could ab- sorb a PCB (Aroclor 1254) from sediments taken from a PCB-polluted estuary — Escambia Bay, Fla. Hansen, Schimmel, and Matthews (1974) found that some estuarine species could avoid waters contaminated with Aroclor 1254, but pink shrimp showed no avoidance reaction when given 29 r" / • -1 4 \. ^^ FISHERY BULLETIN: VOL. 76, NO. 1 %X' ^'^K* 36 X Figure 36. — Electron micrograph of profile of hepatopancreatic cell from pink shrimp exposed to 3 ppb Aroclor 1254 (PCB) for 52 days; note endoplasmic reticulum proliferation and beginning formation of cytoplasmic whorls (arrow). ^ 14,400. Figure 37. — Membrane whorls (myeloid bodies) surrounding lipid in hepatopancreatic cells of shrimp exposed to 3 ppb Aroclor 1254 (arrows). Control nonexposed shrimp did not produce profiles with these configurations, x 28,500, / '■«% ^'WS- choices of clean or PCB-contaminated water. These and other data suggest that PCB's, as pol- lutants, could have influence on relative survival and abundance of penaeid shrimps in natural wa- ters. Organophosphates and Carbamates Few organophosphate compounds have been tested in species of crustaceans. Howevei", those tested have shown approximately 1,000 times greater toxicity to shrimps than most other pes- ticides tested (Butler 1966), and penaeid shrimps have shown greater sensitivity than fishes or mol- lusks (Table 4). Baytex ( Bayer 29, 493 ) was very toxic to penaeid shrimp (Butler and Springer 1963) in the labora- tory. Naled (1,2 dibromo-2,2-dichloroethyl-di- methyl phosphate) had little effect in field tests on shrimp. Fast dilution and instability without per- sistence of compounds may be reasons for lack of mortality of shrimps in field tests of organophos- phates. In the laboratory, Dibrom is lethal to post- larval brown shrimp at 2.0 ppb, and at 5.5 ppb it is lethal to adult pink shrimp (5.5 ppb = LC^^ for 48 h exposure). Malathion, at 14 ppb, caused hyperactivity, paralysis, and death in penaeids, and parathion 30 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS jr^"? »'*•' 38. ^ / 38b Figure as. — a. Hepatopancreatic cell profiles revealing nuclei with small vesicles (20-50 nm) (white arrows) in nucleoplasm from shrimp exposed to 3 ppb Aroclor 1254; also note cytoplasmic degeneration (black arrow); compare with more normal cell in lower right comer, x 14,400. b. Hepatopancreatic cell profile showing nucleus with major large vesicles (100-700 nm) (arrows) in nucleoplasm; note also dense bodies in nuclear envelope from PCB-exposed shrimp, x 28,500. 31 FISHERY BULLETIN: VOL. 76, NO. 1 lethal concentration for 48 h in pink shrimp was 0.2 ppb (D. Coppage, pers. commun.). No his- topathogenesis has been reported for penaeids ex- posed to organophosphates. Conte and Parker ( 1975) found Malathion ae- rially applied to flooded marshes in Texas caused from 14 to 809c mortality in brown and white shrimps held in cages. They recommended that Malathion not be applied to flooded marshes that maintained shrimp. Both organophosphates and carbamates are po- tent acetycholinesterase ( AChe) inhibitors. Little evidence of early, presyndromic inhibition of AChe activity in the ventral nerve cord of pink shrimp was found, but inhibition as high as 75% was found in moribund shrimp exposed to Mala- thion (Coppage and Matthews 1974). Carbamate pesticides have not been tested much in regard to penaeid shrimps, but it is known that Sevin is lethal to other shrimps and crusta- ceans when applied to field sites in the marine environment (Haven et al. 1966). J. Lowe (pers. commun.) has found carbaryl (Sevin) to be quite toxic to penaeids (Table 4) in laboratory tests. Petroleum Very little information exists on the effects of petroleum or oil products on penaeid shrimps. This is surprising because many offshore oil producing areas are also penaeid shrimp producing regions. Anderson et al. ( 1974) and Cox^ reported results of studies on the toxicity of No. 2 fuel oil on the brown shrimp. The 24-h median tolerance limits of juvenile brown shrimp exposed to components of No. 2 fuel oil (naphthalenes, methylnaphthalenes, and dimethyl napthalenes) ranged from 0.77 to 2.51 ppm. The naphthalenes were the most toxic components of fuel oil. Refined oils. No. 2 fuel oil, and Venezuelan bunker C oil were more toxic to brown shrimp than was Louisiana crude oil. Cox reported that the higher content of toxic aromatics in the refined oils above accounted for their higher toxicity to penaeids. Yarbrough and Minchew'° reported several his- tological lesions in penaeids exposed to 2.0 ppm ^Cox, B. A. 1975. The toxicity of no. 2 fuel oil on the brown shrimp Penaeus aztecus. In Program of the first workshop on the pathology and toxicology of penaeid shrimps. U.S. EPA, Gulf Breeze, Fla., 12 p. '"Yarbrough, J. D., and D. Minchew. 1975. Histological changes in the shrimp related to chronic exposure to crude oil. In Program of the first workshop on the pathology and toxicology of penaeid shrimps. U.S. EPA, Gulf Breeze, Fla., 12 p. sonified crude oil. Nonspecific lesions were de- scribed in the cuticular chitin, the lining of the gastric mill, and the mouth region of shrimps. The proliferation of cells and necrosis in the basal por- tion of gill filaments was reported as a more specific lesion associated with exposure. These ef- fects should be examined carefully in relation to "shell" disease resulting from natural conditions. Heavy Metals Cadmium Unusually high levels of cadmium have been reported from certain estuarine areas in which penaeid shrimps commonly occur (i.e., Laguna Madre, Corpus Christi, Tex.). This metal is also a pollutant component from several industrial effluents that are emptied into aquatic systems. In experiments at Gulf Breeze, Nimmo et al. (1977) observed that in pink shrimp exposed to approximately 760 ppb cadmium (as CdCla) for 9 days or longer an unusual darkening of gills oc- curred which eventually led to complete blacken- ing of gills of a significant number of exposed shrimp. Control shrimp did not develop black gills. In other tests, it was found that the hC^^ of cad- mium in 30 days was 718 ppb, and during these tests many exposed shrimp developed the black gill syndrome prior to death (Figure 39). I have completed light and electron microscopic studies of gill tissues from exposed blackened gills and control gills of surviving pink shrimp which Nimmo supplied from his tests (Couch 1977). My findings indicate that the gross blackening of gills results from necrosis of subcuticular tissues (gill epithelial tissue) (Figure 40a, b). This necrosis stems from the death of cells in the distal gill filaments (smallest unit in gill of shrimp). Actual cell death occurs prior to gross blackening in tiny foci, followed by gradual involvement of the whole filament. Electron microscopy reveals polymor- phic black deposits in the cytoplasm of moribund or necrotic cells (early around mitochondria, later throughout). A complete loss of structural and, probably, functional integrity of the gill soft tissue (Figures 41, 42a, b) leads to organ necrosis. How- ever, the cuticle and epicuticle remain intact at the ultrastructural level and hold the moribund or necrotic soft tissue within their boundaries. Grossly, apparent melanization of injured gill filaments account for the blackening syndrome. However, EM (Figure 42a, b) does not present 32 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS 8 10 u 13 Figure 39. — Pink shrimp with black gill syndrome (above) associated with exposure to cadmium chloride. Control, nonexposed shrimp shown below (scale in inches). evidence for the presence of melanosomes, melanocytes, or melanophores. An alternative possibility, that cell death and necrosis lead to the deposition of metal sulfides or other black deposits in necrotic tissues in the living animal, could ac- count for the blackened gill syndrome. At any rate, the interesting concept of cell and tissue death preceding organismic death is represented in the pink shrimp's response to cadmium exposure. Death of cells (in the gills) concerned with os- moregulation and respiration would lead to dys- function and eventual death of shrimp. Bahner'i has studied the uptake of cadmium in "Bahner.L. H. 1975. Mobilization ofcadmium in the tissues of pink shrimp, Penaeus duorarum. In Program of the first work- shop on the pathology and toxicology of penaeid shrimps. U.S. EPA, Gulf Breeze, Fla., 8 p. the tissue of pink shrimp. He found that between 1 and 10 ppb Cd in water elicited uptake by hepatopancreas, gills, and exoskeleton. Below concentrations of 1 ppb Cd in water, there was no accumulation of the metal in shrimp tissue. Little is known concerning cadmium effects on feral shrimp in nature. Mercury Mercury as a metal has not been suspect in toxic effects on organisms. Mercuric salts and methy- lated mercury, however, are extremely toxic with both short-term and long-term chronic effects. Mercuric chloride is used in a variety of histologi- cal fixative fluids because of its protein-precipi- tating effects in tissues of invertebrates (Sparks 33 FISHERY BULLETIN: VOL 76, NO. 1 Figure 40. — a. Histological appearance of early black gill lesion; note that black- ening occurs first near tips of gill fila- ments; normal gill filament (arrow) is to right of blackened filaments. x580. b. Histological appearance of advanced black gill in cadmium-exposed pink shrimp; note complete necrosis of gill filaments, but clear line of separation from more normal tissue below. x580. 1972). Few studies have been reported concerning effects of mercury compounds on penaeid shrimps. Petrocelli et al.^^ studied the uptake and gross distribution of mercuric chloride in brown shrimp. '^Petrocelli, S. R., G. Roseijadi, J. W. Anderson, B . J. Presley, and R. Sims. 1975. Brown shrimp exposed to inorganic mercury in the field. In Program of the first workshop on the pathology and toxicology of penaeid shrimps. U.S. EPA, Gulf Breeze, Fla., 1 p. These authors also examined the effects of mer- curic chloride exposure on ability of brown shrimp to adjust to salinity changes. They found that after 2 h exposure to 0.5 ppb mercuric chloride in seawa- ter, residue level of mercury in shrimp was 285 ppb with only d% of the mercury in the meat (mus- cle) and Ql'/f in the shell. This suggested a surface adsorptive process for mercury in brown shrimp 34 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Tl .V *^' M :^ CM Figure 41. — Electron micrograph of normal gill cuticle (arrow) and underlying osmoregulatory and respiratory epithelium; note mitochondria (M), cell membranes (CM), hemolymph sinus (S), and cuticle (C). X 14,400. exposed for brief periods. These authors also re- ported that shrimp obtained from off Louisiana's Southwest Pass had natural levels of only 4.6 ppb mercury distributed as 64'7<^ in the muscle and 36*7^ in the cuticle. Brown shrimp are active regulators of blood chloride levels (ion regulators). Petrocelli et al. (see footnote 12) found that exposure of brown shrimp to mercury and to salinity changes re- sulted in interference with the shrimp's ability to adjust their internal ion levels to external salinity changes. Therefore, mercury could prove to be det- rimental to penaeid shrimps if it were present in form and amount enough to prevent their adjust- ment to freshets or high saline conditions that result from rapid changes in estuaries or tide- lands. Chemotherapeutic Chemicals Certain inorganic and organic chemicals have been tested for toxic effects in penaeid shrimps because they are used routinely as chemo- therapeutic agents in aquatic animal disease con- trol. 35 FISHERY BULLETIN: VOL. 76, NO. 1 4Z 42b, ^ ^K Figure 42. — a. Electron micrograph of comparable gill region (to Figure 40) in cadmium-exposed shrimp with black gill sjTidrome; note cell necrosis, black deposits around mitochondria (arrows); note loss of membrane integrity, x 14,400. b. Higher magnification of black cytoplasmic deposits in gill epithelial cells of cadmium-exposed shrimp; note polymorphic nature of deposits. x28,500. 36 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Johnson ( 1975, footnotes 13, 14) has determined toxic concentrations in penaeid shrimps for the following chemicals: Formalin, potassium per- manganate, potassium dichromate, copper sul- fate, acriflavine, malachite green, and methylene blue. His results are reported below. Essentially, Johnson found that for Formalin the 96-h LC^o at 28 °C for pink shrimp was 235 to 270 ppm in seawater. He reported that 25 ppm Formalin applications for killing of external pro- tozoa on penaeid shrimps would be safe for in- definite periods. Potassium permanganate LC50 at 96 h for pink shrimp was 6 ppm. At this concentration a precipi- tate was formed on the gills of shrimp and death may have resulted from asphyxiation. Potassium dichromate, which may be of some use as an antibacterial agent, was found to be nontoxic for shrimps below concentrations of 5 ppm for short term exposures. Copper sulfate has been of use as a herbicide and protozoan control agent in fisheries research. It was found that copper sulfate at low concentra- tions (0.5-1.0 ppm) was reasonably safe for penaeids. Acriflavine, an antibacterial agent, had a 96-h LCgp for pink shrimp of 1.0 ppm in seawater. This compound was probably not safe for shrimps at effective bacteriostatic concentrations. Malachite green, a parasiticide for freshwater fishes, has a toxic effect in shrimp associated with molting. Johnson (see footnote 14) reported that newly molted shrimps are much more sensitive to malachite green than intermolt shrimps. From 2.5 to 20 ppm of the compound in seawater resulted in death of all exposed newly molted shrimps. Adult, nonmolting, penaeid shrimps seemed to tolerate higher concentrations of malachite green (20 ppm). Johnson believed that malachite green holds promise as a fungistat for use in penaeid shrimp culture. Methylene blue should be usable below concen- trations of 1.0 ppm for prophylaxis of fungi and protozoa in penaeids. Quinaldine (product of Eastman Kodak Com- pany) was used by Johnson (see footnote 13) as an anesthetic for white shrimp. He found that shrimp become anesthetized when exposed to all concen- '^ Johnson, S. K. 1974. Use of Quinaldine with penaeid shrimp. Texas A&M Univ., Fish Disease Diagnostic Lab. Note FDDL-S4, 2 p. '■* Johnson, S. K. 1974. Toxicity of several management chemi- cals to penaeid shrimp. Texas A&M Univ., Fish Disease Diag- nostic Lab. Note FDDL-S13, 10 p. trations of quinaldine, but after 48 h, 10%, 20%, and 20% losses occurred respectively in 25-, 30-, and 35-ppm treatment groups. A 25-ppm concen- tration was set as the minimum effective anes- thetic level with white shrimps. This concentra- tion, however, results in death of some shrimp as indicated above. Johnson also reported that spon- taneous muscle necrosis occurred in abdominal musculature of some shrimp that became hyper- kinetic at concentrations of 25 ppm and above. SPONTANEOUS PATHOSES Under this heading are included diseases of penaeid shrimps for which etiologic agents are not known, or are uncertain. Tumors There have been no invasive neoplasms re- ported for decapod crustaceans. Tumorlike growths have been reported in lobsters (Herrick 1895, 1909; Prince 1897), in a crab (Fischer 1928), and in a paleomonid shrimp (Savant and Kewal- ramani 1964). To date, the only published report of a tumorlike growth in a penaeid shrimp is that of Sparks and Lightner (1973). They reported a papilliform, tumorlike growth on the right ventrolateral as- pect of the sixth abdominal segment of a specimen oi Penaeus aztecus. This shrimp had been taken from an experimental rearing pond at Palacios, Tex. The growth was tentatively diagnosed as a benign neoplasm, consisting of hypertrophied and normal tissue. Robin Overstreet (Gulf Coast Research Lab- oratory) recently presented me with two larval penaeid shrimp each of which had one small growth on an abdominal segment. Light micros- copy and EM revealed that these enlargements contained only striated muscle and sacroplasmic reticulum (Figure 43). There was no evidence that the growths were neoplastic or that parasites (in- cluding viruses) were involved. Overstreet is pres- ently completing a detailed study of this condition and is describing the growths as hamartomas, pos- sibly related to polluted water conditions from which the affected shrimp were collected. Spontaneous Muscle Necrosis Penaeid shrimps often respond to handling, temperature, and chemical stress by developing a 37 FISHERY BULLETIN: VOL. 76, NO. 1 4,' 0-»^ Figure 43. — Electron micrograph of striated muscle and sarcoplasmic reticulum from abnormal growth on abdominal appendage of penaeid shrimp, x 14,400. 38 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS Figure 44. — Spontaneous necrosis in pink shrimp exposed to low tem- peratures (10°C); muscle affected is in whitened area in tail; note uropod and tail degeneration associated with necrotic condition. Shrimp was alive at time photograph was taken. white or opaque abdominal musculature (Figure 44). Rigdon and Baxter ( 1970) first reported this disease as spontaneous muscle necrosis and de- scribed the histological condition as "degenerated foci of striated muscle" in brown shrimp. Shrimp with this condition are debilitated and usually die unless stress ceases and extent of necrosis is small and limited. Shrimp will recover in many cases, however, if stress ceases. The muscle fibers affected appear lysed microscopically, and their structural integrity is lost. This syndrome may be related to oxygen starvation of muscle tissue when the shrimp is pressed to its physiological tolerance limits for high or low temperatures or hyperkine- tic muscular activity. The white appearance of the shrimp abdomen caused by spontaneous muscle necrosis should not be confused with "cotton" shrimp which are infected by microsporidan para- sites (diffential diagnosis depends on finding spores of Microsporida in whitened tissue). Gas Bubble Disease Lightner et al. (1974) reported that juvenile brown shrimp developed a disease characterized by the presence of many small and large bubbles of gas in gill and other tissues. This condition was related to heated water in which the shrimp were held and from which excess gas was not allowed to escape. These authors pointed out the potential threat of gas bubble disease to shrimp held in culture situations utilizing heated water. The ex- tent of the threat of this disease in penaeid culture is unknown. This syndrome has not been reported in feral shrimp, but is a well-known disease in salmonid fishes that contact waters of varying temperatures and gaseous supersaturation. , "Shell Disease" and Black Gills Blackened, pitted, and eroded exoskeleton is not uncommon in many decapod crustaceans as previ- ously stated. These degenerative changes in cuti- cles of crabs, lobsters, and shrimps have been termed collectively "shell disease" (Rosen 1970). Lesions ranging from tiny, pinhead-size black holes in the cuticle to massive blackened, eroded area of the cuticle (Figure 6) are often observed in penaeid shrimps. Rosen (1970) reports that the disease is definitely contagious, but the identifica- tion of the infectious agents is not known for most species of decapods (see section on Bacteria, under Infectious Diseases). He believes that the necrotic pits in the cuticle act as "miniature niches" for several taxonomic groups of chitinoclastic mi- crobes (bacteria and fungi). The only successful demonstration that chitinoclastic bacteria caused the disease was that of Bright et al.^^. They iso- lated bacteria from lesions on Alaskan king crabs and introduced them into mechanical abrasions on healthy king crab and shell disease developed. "Shell disease" may have many different causes in different species of crustaceans. Couch (1977) and Lightner (pers. commun.) found that black- ening necrosis of gill tissues in pink shrimp (see Toxic Response Section — Cadmium), as well as blackened cuticular lesions occurred in shrimp exposed to cadmium, suggest that high concentra- tions of some heavy metals may cause a form of shell disease. i^Bright, D. B., F. E. Durham, and J. W. Knudsen. 1960. King crab investigations of Cook Inlet, Alaska. Unpubl. contract rep., Allen Hancock Found., Univ. South. Calif , Los Ang. to BCF Biol. Lab., Auke Bay, Alaska. Available Northwest and Alaska Fisheries Center Auke Bay Laboratory, Natl. Mar. Fish. Serv., NOAA, P.O. Box 155, Auke Bay, AK 99821. 39 FISHERY BULLETIN: VOL. 76, NO. 1 .^,,.M.^_»J^^^^^^^^^^ ^^ .: ■'■'■'■'■'■" -"''^'^^'*^^^^^^^---' ■ ■■■■■<* s F ; ^*^^ ^ *•* ^jmHPf Figure 45. — Black gill in feral shrimp not exposed to any known pollutant; grossly resembles cadmium-associated black gill syndrome. Black gills are often observed in shrimp taken from natural populations (Figure 45). Grossly, the black gills of feral shrimp and those of shrimp experimentally exposed to cadmium are indistin- guishable. The cause of black gills in feral pen- aeids is unknown, but I have found shrimp heavily infested w^ith apostome ciliate phoronts to have considerable areas of black gill. Therefore, black gill has been associated with heavy metal expo- sure, protozoan infestation, and with fungal infec- tion [Fusarium: Solangi and Lightner 1976), suggesting multiple causes. Probably, any injury that causes death of cells in gills of shrimp could cause some form of blackened gill due to necrotic tissues, and, perhaps, melanization. Broken-Back Syndrome Shrimp suffering from severe salinity, cold temperature, and handling stresses in combina- tion, display a characteristic dorsal separation of the pleural plates covering the third and fourth abdominal segments (Figure 46). This results in bulging of muscle through the separation. I have observed this in 100% of 1,800 captive pink shrimp dying from a sudden drop in salinity (15-18%o to 3%o) combined with cold water (8°C). The separa- tion of cuticular plates and bulging of muscle ap- parently results from uptake of water and severe flexures of the abdomen in shrimp attempting to escape unfavorable conditions. OVERVIEW AND FUTURE RESEARCH Some major problem areas in our knowledge of penaeid shrimp diseases become apparent in a re- view such as this. Although considerable parasitology has been done for penaeid shrimps, new protozoan and worm parasites, some pathogenic, continue to be found. Until recently no viruses were reported for shrimp; now at least one is known. Mycology and bacteriology have yet to contribute in major ways to our understanding of penaeid shrimp diseases and health. Relatively 40 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS i. Figure 46. — Pink shrimp from mortality related to salinity drop and cold-water temperatures; note dorsal region between third and fourth pleural plates where muscle is protruding. Middle shrimp was still alive when photo was taken; note beginning break in dorsal cuticle (arrow). Top and bottom shrimp died just prior to photograph. little is known of the toxic responses of penaeids to such environmentally abundant pollutants as oil, oil products, pesticides, heavy metals, industrial chemicals, and domestic sewage. The question of acquisition of resistance to infectious disease or toxicants in penaeid shrimps is unanswered. There is a pressing need to begin detailed studies of pathogenesis of disease and mechanisms of pathogenesis. With the knowledge that penaeid shrimps have cosmopolitan distribution comes the realization that the disease problems of so narrow an area as encompassed in the review merely hint at the vastness of the potential problems of shrimp dis- eases worldwide. This is not the case for many other decapod Crustacea which have relatively restricted ranges (i.e., Homarus americanus, Cal- linectes sapidus) and which do not assume the worldwide commercial value of penaeid shrimps. The old truisms concerning crowding of large numbers of penaeid shrimps in mariculture at- tempts and rapid spread of infectious diseases still apply as future problems to be studied. Along with this, continual need for better chemotherapeutic agents and an understanding of their effects on penaeid shrimps is apparent. Because penaeid shrimps are components in the human food chain (wherein man is the final con- 41 FISHERY BULLETIN: VOL. 76, NO. 1 sumer), a better knowledge of their accumulative, metabolic, and storage abilities of toxicants, par- ticularly carcinogenic chemicals, from the envi- ronment is needed to safeguard human health as well as shrimp health. Penaeid shrimps are known to be very sensitive to certain classes of chemical pollutants such as organochlorines and heavy metals (e.g., cadmium) and, therefore, should be utilized more in the future as indicator organisms in environmental quality studies. ACKNOWLEDGMENTS Appreciation is expressed to Lee Courtney who helped prepare the plates of figures and aided in collecting data and certain of the figures included. Steve Foss prepared some of the figures. Don Lightner furnished two micrographs, and John Corliss aided in identification of ciliates. Jack Lowe provided summarized data on toxicity of cer- tain compounds to penaeid shrimps. Robin Over- street discussed some of the taxonomic problems concerning helminths and brought to my atten- tion several recent important references. Dean Davenport and Chris Howell are thanked for con- tributions of larval shrimp for disease study. LITERATURE CITED ] ALDRICH, D. V. 1965. Observations on the ecology and life cycle of Pro- christianella penaei Kruse (Cestoda: Trypanorhyn- cha). J. Parasitol. 51:370-376. Anderson, J. W., J. M. Neff, B. a. Cox, h. e. tatem, and G. M. HIGHTOWER. 1974. The effects of oil on estuarine animals: Toxicity, uptake and depuration, respiration. In F. J. Vernberg and W. B. Vernberg (editors), Pollution and physiology of marine organisms, p. 285-310. Academic Press, N.Y. BAXTER, K. N., R. H. RIGDON, AND C. HANA. 1970. Pleistophora sp. (Microsporidia: Nosematidae): A new parasite of shrimp. J. Invertebr. Pathol. 16:289- 291. BRADBURY, P. C. 1966. The fine structure of the mature tomite ofHyalophy- sa chattoni. J. Protozool. 13:591-607. 1973. The fine structure of the cj^ostome of the apostomat- ous ciliate Hyalophysa chattoni. J. Protozool. 20:405- 414. 1974. The fine structure of the phoront of the apostomat- ous ciliate, Hyalophysa chattoni. J. Protozool. 21:112- 120. BROADHURST, M. G. 1972. Use and replaceability of PCB's. Environ. Health Perspect. 2:81-102. BUTLER, P. A. 1966. The problem of pesticides in estuaries. In A sym- posium on estuarine fisheries, p. 110-115. Am. Fish. See. Spec. Publ, 3. BUTLER, P. A., AND P. F. SPRINGER. 1963. Pesticides — a new factor in coastal environments. Trans. 28th North Am. Wildl. Nat. Resour. Conf., p. 378- 390. CAMPBELL, R. A., AND J. CARVAJAL. 1975. A revision of some trypanorhynchs from the western North Atlantic described by Edwin Linton. J. Parasitol. 61:1016-1022. CHATTON, E. 1911. Cilies parasites des cestes et des pyrosomes: Perika- ryon cestiocola n.g., n. sp., etConchophrys davidoffi, n.g., n. sp. Arch. Zool. Exp. Gen., 5e ser., 48(notes et rev.):8- 20. CHATTON, E., AND A. LWOFF. 1926. Les Synophrya, infusoires parasites internes des crabes. Leur evolution a la mue. Leur place parmi les Foettingeriidae. C. R. Acad. Sci. 183:1131-1134. 1935. Les cilies apostomes, morphologie, cytologie, ethologie, evolution, systematique. Arch. Zool. Exp. Gen. 77:1-453. CONSTRANSITCH, M. J. 1970. Description, pathology, and incidence of a new species oiPlistophora penaei n. sp. (Microsporidia, Nose- matidae), a parasite of commercial shrimp. M.S. Thesis, Northwestern State Univ., Natchitoches, La., 35 p. CONTE, F. S., AND J. C. PARKER. 1975. Effect of aerially-applied malathion on juvenile brown and white shrimp Penaeus aztecus and P. setiferus. Trans. Am. Fish. Soc. 104:793-799. COOK, D. W., AND S. R. LOFTON. 1973. Chitinoclastic bacteria associated with shell disease in Penaeus shrimp and blue crab (Callinectes sapidus). J. Wildl. Dis. 9:154-159. COOK, H. L. 1 97 1 . Fungi parasitic on shrimp. F AO Aquaculture BulL 3:13. COPPAGE, D. L., AND E. MATTHEWS. 1974. Short-term effects of organophosphate pesticides on cholinesterases of estuarine fishes and pink shrimp. Bull. Environ. Contam. Toxicol. 11:483-488. CORKERN, C. C, JR. 1970. Investigations of helminths from the hepatopan- creas of the brown shrimp, Penaeus aztecus Ives, from Galveston Bay, Texas. M.S. Thesis, Texas A&M Univ., College Station, 51 p. COUCH, J. A. 1971. Form, morphogenesis, and host-ciliate relationship of Lagenophrys callinectes (Ciliatea: Peritrichida). Ph.D. Thesis, Florida State Univ., Tallahassee, 153 p. 1973. Ultrastructural and protargol studies of Lageno- phrys callinectes (Ciliophora: Peritrichida). J. Protozool. 20:638-647. 1974a. Free and occluded virus, similar to Baculovirus, in hepatopancreas of pink shrimp. Nature (Lond.) 247:229-231. 1974b. An enzootic nuclear polyhedrosis virus of pink shrimp: Ultrastructure, prevalence, and enhance- ment. J. Invertebr. Pathol. 24:311-331. 1976. Attempts to increase Baculovirus prevalence in shrimp by chemical exposure. Prog. Exp. Tumor Res. 20:304-314. 1977. Ultrastructural study of lesions in gills of a marine 42 COUCH: DISEASES AND PARASITES OF PENAEID SHRIMPS shrimp exposed to Cadmium. J. Invertebr. Pathol. 29:267-288. COUCH, J. A., AND L. Courtney. 1977. Interaction of chemical pollutants and virus in a crustacean: a novel bioassay system. Ann. N.Y. Acad. Sci. 228:497-504. COUCH, J. A., AND D. NIMMO. 1974a. Ultrastructural studies of shrimp exposed to the pollutant chemical polychlorinated biphenyl (Aroclor 1254). Bull. Soc. Pharm. Ecol. Pathol. 2:17-20. 1974b. Detection of interactions between natural patho- gens and pollutants in aquatic animals. Proceedings: Gulf Coast Regional Symposium on Diseases of Aquatic Animals, p. 261-268. La. State Univ. Cent. Wetland Re- sour. Publ. No. LSU-SG-74-05. COUCH, J. A., M. D. SUMMERS, AND L. COURTNEY. 1975. Environmental significance of Baculovirus infec- tions in estuarine and marine shrimp. Ann. N.Y. Acad. Sci. 266:528-536. Debaisieux, p. 1960. Ciliates apostomes, parasites de Palaemon. Cel- lule 60:333-352. DUKE, T. W., J. I. LOWE, AND A. J. WILSON, JR. 1970. A polychlorinated biphenyl (Aroclor 1254) in the water, sediment, and biota of Escambia Bay, Flori- da. Bull. Environ. Contam. Toxicol. 5:171-180. FEIGENBAUM, D. L. 1973. Parasites of the commercial shrimp Penaeus van- namei Boone and Penaeus brasiliensis Latreille. M.S. Thesis, Univ. Miami, Coral Gables, 84 p. 1975. Parasites of the commercial shrimp Penaeus van- namei Boone and Penaeus brasiliensis Latreille. Bull. Mar. Sci. 25:491-514. FEIGENBAUM, D. L., AND J. CARNUCCIO. 1976. Comparison between the trypanorhynchid cestode infections of Penaeus duorarum and Penaeus brasiliensis in Biscayne Bay, Florida. J. Invertebr. Pathol. 28:127- 130. Hansen, d. J., P. R. Parrish, and J. Forester. 1974. Aroclor 1016: Toxicity to and uptake by estuarine animals. Environ. Res. 7:363-373. Hansen, d. j., s. C. Schimmel, and e. Matthews. 1974. Avoidance of Aroclor 1254 by shrimp and fishes. Bull. Environ. Contam. Toxicol. 12:253-256. HAVEN, D., M. CASTAGNA, P. CHANLEY, M. WASS, AND J. WHITCOMB. 1966. Effects of the treatment of an oyster bed with Poly- stream and Sevin. Chesapeake Sci. 7:179-188. HERRICK, F. H. 1895. The American lobster: A study of its habits and development. Bull. U.S. Fish Comm. 15:1-252. 1909. Natural history of the American lobster. Bull. [U.S.] Bur. Fish. 29:149-408. HONIGBERG, B. M., W. BALAMUTH, E. C. BOVEE, J. O. CORLISS, M. GOJDICS, R. P. HALL, R. R. KUDO, N. D. LEVINE, A. R. LOEBLICH, JR., J. WEISER, AND D. H. WENRICH. 1964. A revised classification of the phylum Protozoa. J. Protozool. 11:7-20. HUTTON, R. F. 1964. A second list of parasites from marine and coastal animals of Florida. Trans. Am. Microsc. Soc. 83:439-447. HUTTON, R. F., F. SOGANDARES-BERNAL, B. ELDRED, R. M. Ingle, and K. d. Woodburn. 1959. Investigations on the parasites and diseases of salt- water shrimps iPenaeidae) of sports and commercial im- portance to Florida. Fla. State Board Conserv., Tech. Ser. 26, 38 p. IVERSEN, E. S., and J. F. KELLY. 1976. Microsporidiosis successfully transmitted experi- mentally in pink shrimp. J. Invertebr. Pathol. 27:407- 408. IVERSEN, E. S., AND R. B. MANNING. 1959. A new microsporidan parasite from the pink shrimp (Penaeus duorarum). Trans. Am. Fish. Soc. 88:130-132. IVERSEN, E. S., AND N. N. VAN METER. 1964. A record of the microsporidian, Thelohania duorara, parasitizing the shrimp, Penaeus brasiliensis. Bull. Mar. Sci. Gulf Caribb. 14:549-553. JOHNSON, P. W., J. M. SIEBURTH, A. SASTRY, C. R. ARNOLD, AND M. S. DOTY. 1971. Leuchothrix mucor infestation of benthic Crustacea, fish eggs, and tropical algae. Limnol. Oceanogr. 16:962-969. JOHNSON, S. K. 1974. Ectocommensals and parasites of shrimp from Texas rearing ponds. Texas A&M Univ., Sea Grant Publ. TAMU-SG-74-207, 19 p. 1975. Handbook of shrimp diseases. Texas A&M Univ., Sea Grant Publ. TAMU-SG-75-603, 19 p. KELLEY, J. 1975. A description of the histological structure of normal and microsporidan-infected pink shrimp, Penaeus duo- rarum Burkenroad. M.S. Thesis, Univ. Miami, Fla. KRANTZ, G. E., R. R. COLWELL, and E. LOVELACE. 1969. Vibrio parahaemolyticus from the blue crab Cal- linectes sapidus in Chesapeake Bay. Science (Wash., D.C.) 164:1286-1287. KRUSE, D. N. 1959. Parasites of the commercial shrimps, Penaeus az- tecus Ives, P. duorarum Burkenroad and P. setiferus (Lin- naeus). Tulane Stud. Zool. 7:123-144. 1966. Life cycle studies ofNematopsis duorari n. sp., (Gre- garinia: Porosporidae). A parasite of the pink shrimp (Penaeus duorarum) and pelecypod molluscs. Ph.D. Thesis, Florida State Univ., Tallahassee, 174 p. LIGHTNER, D. V. 1974. Normal postmortem changes in the brown shrimp, Penaeus aztecus. Fish. Bull., U.S. 72:223-236. 1975. Some potentially serious disease problems in the culture of penaeid shrimp in North America. Proc. U.S. -Japan Natural Resources Program, Symposium of Aquaculture Disease. Spec. Publ. Fish. Agency, Japanese Government and Jap. Sea Reg. Fish. Res. Lab., Niigata. LIGHTNER, D. V., AND C. T. FONTAINE. 1973. A new fungus disease of the white shrimp Penaeus setiferus. J. Invertebr. Pathol. 22:94-99. LIGHTNER, D. v., B. R. SALSER, AND R. S. WHEELER. 1974. Gas-bubble disease in the brown shrimp [Penaeus aztecus). Aquaculture 4:81-84. LOWE, J. I., P. R. Parrish, A. J. Wilson, Jr., P. D. Wilson, AND T. W. DUKE. 1971. Effects of mirex on selected estuarine organisms. Trans. 36th North Am. Wildl. Nat. Resour. Conf., p. 171- 186. NiMMO, D. R., AND R. R. BLACKMAN. 1972. Effects of DDT on cations in the hepatopancreas of penaeid shrimp. Trans. Am. Fish. Soc. 101:547-549. 43 FISHERY BULLETIN: VOL. 76, NO. 1 NIMMO, D. R., R. R. BLACKMAN, a. J. WILSON, JR., AND J. FORESTER. 1971. Toxicity and distribution of Aroclor 1254 in the pink shrimp Penaeus duorarum. Mar. BioL (Berl.) 11:191- 197. NiMMO, D. R., D. J. HANSEN, J. A. COUCH, N. R. COOLEY, P. R. PARRISH, and J. I. LOWE. 1975. Toxicity of Aroclor 1254 and its physiological activ- ity in several estuarine organisms. Arch. Environ. Con- tarn. Toxicol. 3:22-39. NiMMO, D. R., D. V. LIGHTNER, AND L. H. BAHNER. 1977. Effects of cadmium on the shrimps Penaeus duorarum, Penaeus pugio, and Paleomonetes vulgris. In Physiological responses of marine biota to pollutants, p. 131. Academic Press, N.Y. NiMMO, D. R., A. J. WILSON, JR., AND R. R. BLACKMAN. 1970. Localization of DDT in body organs of pink and white shrimp. Bull. Environ. Contam. Toxicol. 5:333- 341. NIMMO, D. R., P. D. WILSON, R. R. BLACKMAN, AND A. J. WILSON, JR. 1971. Polychlorinated biphenyl absorbed from sediments by fiddler crabs and pink shrimp. Nature (Lond.) 231:50-52. NORRIS, D. E., AND R. M. OVERSTREET. 1976. The public health implications of larval Thynnas- caris nematodes from shellfish. J. Milk Food Technol. 39:47-54. OVERSTREET, R. M. 1973. Parasites of some penaeid shrimps with emphasis on reared hosts. Aquaculture 2:105-140. 1975. Buquinolate as a preventive drug to control micro- sporidosis in the blue crab. J. Invertebr. Pathol. 26:213-216. PARRISH, P. R., J. A. COUCH, J. FORESTER, J. M. PATRICK, JR., ANDG. H. COOK. 1973. Dieldrin: Effects on several estuarine organisms. Proc. 27th Annu. Conf Southeast. Assoc. Game Fish Comm., p. 427-434. PRINCE, E. E. 1897. Special report on the natural history of the lob- ster. 29th Annu. Rep., Dep. Mar. Fish. Can. Supl. 1, Parts I-IV, p. 1-36. RAGAN, J. G., AND D. V. ALDRICH. 1972. Infection of brown shrimp, Penaeus aztecus Ives by Prochristianella penaei KruselCestoda: Trypanorhyncha) in southeastern Louisiana bays. Trans. Am. Fish. Soc. 101:226-238. RIGDON, R. H., AND K. N. BAXTER. 1970. Spontaneous necroses in muscles of brown shrimp, Penaeus aztecus Ives. Trans. Am. Fish. Soc. 99:583-587. ROSEN, B. 1970. Shell disease of aquatic crustaceans. In S. F. Snieszko (editor), A symposium on diseases of fishes and shellfishes, p. 409-415. Am. Fish. Soc. Spec. Publ. 5. ROTH, J. N., AND E. S. IVERSEN. 1971. Attempts to transmit experimentally the microspor- idan Thelohania duorara, parasitizing the pink shrimp, Penaeus duorarum. Trans. Am. Fish. Soc. 100:369-370. SAVANT, K. B., AND H. G. KEWALRAMANI. 1964. On a new record of host species of isopod parasite, Bopyrus. Curr. Sci. 33:217. SODERGREN, A., BJ. SVENSSON, AND S. ULFSTRAND. 1972. DDT and PCB in south Swedish streams. Environ. Pollut. 3:25-36. SOLANGI, M. a., and D. v. LIGHTNER. 1976. Cellular inflammatory response of Penaeus aztecus and P. setiferus to the pathogenic fungus, Fusarium sp., isolated from the California brown shrimp, P. californien- sis. J. Invertebr. Pathol. 27:77-86. SPARKS, A. K. 1972. Invertebrate pathology: Noncommunicable dis- eases. Academic Press, N.Y., 387 p. SPARKS, A. K., AND C. T. FONTAINE. 1973. Host response in the white shrimp, Penaeus setifer- us, to infection by the larval trypanorhjTichid cestode, Prochristianella penaei. J. Invertebr. Pathol. 22:213- 219. Sparks, a. k., and d. v. lightner. 1973. A tumorlike papilliform growth in the brown shrimp {Penaeus aztecus). J. Invertebr. Pathol. 22:203-212. SPRAGUE, V. 1954. Protozoa. In P. S. GaltsofF (coordinator). Gulf of Mexico, its origin, waters, and marine life, p. 243- 256. U.S. Fish Wildl. Serv., Fish. Bull. 55. 1970. Some protozoan parasites and hyperparasites in marine decapod Crustacea. In S. F. Snieszko (editor), A symposium on diseases of fishes and shellfishes, p. 416- 430. Am. Fish. Soc. Spec. Publ. 5. Sprague, v., and J. Couch. 1971. An annotated list of protozoan parasites, hyperpara- sites, and commensals of decapod Crustacea. J. Pro- tozool. 18:526-537. SPRAGUE, v., AND P. E. ORR, JR. 1955. Nematopsis ostrearum and A^. prytherchi. (Eug- regarinina: Porosporidae) with special reference to the host-parasite relations. J. Parasitol. 41:89-104. ULITZUR, S. 1974. Vibrio parahaemolyticus and Vibrio alginolyticus: Short generation-time marine bacteria. Microb. Ecol. 1:127-135. Vanderzant, C, R. Nickelson, and J. C. Parker. 1970. Isolation oi Vibrio parahemolyticus from Gulf coast shrimp. J. Milk Food Technol. 33:161-162. VILLELLA, J. B., E. S. IVERSEN, AND C. J. SiNDERMANN. 1970. Comparison of the parasites of pond-reared and wild pink shrimp (Penaeus duorarum Burkenroad) in south Florida. Trans. Am. Fish. Soc. 99:789-794. VIOSCA, P., JR. 1943. A critical analysis of practices in the management of warm-water fish with a view to greater food produc- tion. Trans. Am. Fish. Soc. 73:274-283. 44 ESTIMATING NATURAL AND FISHING MORTALITIES OF CHINOOK SALMON, ONCORHYNCHUS TSHAWYTSCHA, IN THE OCEAN, BASED ON RECOVERIES OF MARKED FISH Kenneth A. Henry ' ABSTRACT In this paper I demonstrate the method of calculating estimates of fishing mortality (F) and natural mortality (M) occurring in the ocean for 1961 and 1962 brood Columbia River hatchery fall chinook salmon, Oncorhynchus tshawytscha, based on assumed values of the proportion of fish that mature annually {m) and on recoveries of marked fish. The advantages of this method over the method of assuming fixed natural mortality rates and back calculating estimates are discussed. It was possible to develop estimates of 1962 Spring Creek data up to the fourth year of life and to compare these estimates with values for the 1961 brood whereas no estimates had been possible with the back calculation method. Thus, estimates of Af j are higher for the 1962 brood; estimates of Mj are very similar for the two broods and the estimates of M, are slightly higher for the 1962 brood. A major difference between the two methods is that natural mortality was assumed to be constant for the back calculation method whereas estimates of natural mortality were obtained separately each year using assumed proportions maturing. Thus, for the 1962 brood general marked fish, an A/ = 0.60 was used in the back calculation method while estimates of Mj = 5.814, Afj = 0.510, M3 = 0.653, and M^ = 0.727 were obtained by assuming varying proportions maturing. A series of graphs are developed that permit a quick analysis of any combination of proportions offish maturing, fishing mortality, and natural mortality and which clearly depict the relationship between these various factors. Cleaver (1969) developed a method for estimating fishing mortalities and percentages of maturing fish for each age group of fall chinook salmon, Oncorhynchus tshawytscha,^ from the Columbia River using selected values of natural mortality. Cleaver's estimates were based on data obtained from a cooperative marking experiment by fishery agencies along the Pacific Coast. This experiment started in 1962 and was designed to measure the contribution of fall chinook salmon from Columbia River hatcheries to the various fisheries. Cleaver's analysis was specifically directed towards returns for the 1961 brood year. The procedure used catches and escapements, by age, along with selected natural mortality values to back calcu- late, from year 5 to year 2, annual estimates of fishing mortality and proportion of fish that ma- ture annually. Henry (1971) utilized Cleaver's method to ob- tain similar estimates for the 1962 brood releases of Columbia River hatchery fall chinook salmon. 'Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 2725 Montlake Boulevard East, Seat- tle, WA 98112. ^Seasonal races of chinook salmon in the Columbia River system aie classified as spring, summer, or fall depending on the time of year that the adults enter the river to spawn. Manuscript accepted June 1977 FISHERY BULLETIN: VOL. 76, NO 1, 1978. Lander and Henry (1973), in analyzing returns from marking experiments for Columbia River coho salmon, O. kisutch, pointed out two methods for estimating the various pertinent parameters mentioned above from salmon mark/recovery data: 1) assume selected values for M (natural mortality) and 2) assume selected values for m (proportion maturing). Although both methods gave identical esti- mates of the parameters, their concepts differ. In selecting a value for natural mortality, as was done by Cleaver (1969) and Henry (1971), one has to start at the end of the life cycle and work back- wards since the calculated parameters are sequen- tially dependent in that manner (Cleaver and Henry also assumed a constant M for all ages to simplify computations); by selecting values for the proportion of fish that mature annually, one be- gins at the younger age-groups and calculates the various parameters sequentially towards the end of the life cycle. This method more closely parallels the actual life history of the salmon. Furthermore, today's salmon management schemes are directed at preserving existing runs and their fisheries, i.e., changing diets, releasing fish at different times and at different sizes, transporting fish to avoid 45 FISHERY BULLETIN: VOL. 76, NO. 1 excessive mortalities (related to passage at dams and unfavorable environmental conditions caused by dams and reservoirs), or transporting fish to make a more direct input to a certain fishery. All of these efforts may affect the maturity, growth, fishing mortality, and the natural mortality for a particular stock offish. In this paper, I describe a method by which such changes can be accounted for in the estimating procedure as soon as they are determined. Thus, the present method reduces the need for assumptions regarding constancy of natural mortality in salmon stocks, and the re- sults may be more realistic, particularly if the maturity values selected are reasonable. In discussing their method of selecting values for the proportion offish that mature annually and then calculating the remaining parameters for coho salmon. Lander and Henry ( 1973) pointed out that the procedure also could be applied to chinook salmon, although they also noted that ". . . this gets to be very complicated to display graphically . . . .", since coho salmon have a much simpler life history than fall chinook salmon — m (proportion offish that mature annually), M (natural mortal- ity), and F (fishing mortality) need to be estimated for 1 yr only for each brood of coho salmon, but these parameters need to be estimated for three separate years for each brood of chinook salmon. Furthermore, the estimated values from this method are quite complicated to apply to chinook salmon. In fact for each m, (the subscript repre- sents the different years of life covered by the calculations) value selected, there is a series of possible Wj values, and for each of the possible m2 values there is again a series of possible mg values. Thus, if n separate calculations are made for each m,, and there are three of them, as for the chinook salmon, the total calculations potentially needed for a brood year would he n^ -\- n^"^ + n^^. METHOD OF ESTIMATING PARAMETERS In this paper I demonstrate the method of cal- culating estimates of fishing mortality (F) and natural mortality (M) based on assumed values of the proportion offish that mature annually (m ) for the 1961 and 1962 brood Columbia River fall chinook salmon. In particular, I compare data for the 1961 and 1962 broods of Spring Creek fish. To aid in understanding the various parameters I estimate, in Figure 1 1 have portrayed graphically certain features of the fall chinook salmon's life history, particularly the various parameters for the period from the release of the fish as smolts until final return to the Columbia River as adults — approximately 54 mo. Figure 1 shows that as a result of this series of events, I end up with eight items of observed data: 1) number of smolts released (A'^o^; 2) number maturing as 2-yr-olds (E-^); 3) number caught by the ocean troll and sport fisheries as 3-yr-olds (Cj); 4) number maturing and return- ing to the river as 3-yr-olds {E^Y, 5) number caught by the ocean troll and sport fisheries as 4-yr-olds (C2); 6) number maturing and return- ing to the river as 4-yr-olds (E^); 7) number caught by the ocean troll and sport fisheries as 5-yr-olds (Cg); and 8) number maturing and re- turning to the river as 5-yr-old fish (£^4). From these eight known values I want to estimate: 1) monthly fishing mortality rate on 3-, 4-, and 5-yr- old fish (Fj, F2, and F^, respectively) over the last 6-mo period of each year; 2) monthly natural €> D, t e-i8M, g-6Mj e-6 = 08 = 1-01-02-^3-^4-^5-^6-^7 (7) whereD = Di+D2+Ds+D4+D^+Dq+D-j = Total fish dying naturally. (8) 47 FISHERY BULLETIN; VOL. 76, NO. 1 ^1 = ^i/A^o h = Ci/A^o h = ^2/A^O h = C2/iVo h = ^3/A^O h = Cg/iVo h = i?4/A^o K = 1-01-^2- .03- -04- -h- -9e -0, The maximum likelihood estimators of the 0^ are: (9) (10) (11) (12) (13) (14) (15) (16) A maximum likelihood estimator of a function of the parameters 6, is obtained by replacing the parameter values by the corresponding maximum likelihood estimates, 6, (Graybill 1961). Beyond that, however, there exists no unique transforma- tion, or function, to obtain maximum likelihood estimates of mi, ma, mg, F^Fg, Fg, Mj, Mg, M^, and Af 4. Any given set of observed data can generate a variety of combinations of parameter estimates. Since no unique solution exists, the only practi- cal solution is to assume values for one of the unknown parameters and solve the equations for the remaining parameters. Thus Cleaver (1969) and Henry (1971) assumed values for M, (natural mortality) for hatchery chinook salmon and calcu- lated values for the remaining parameters. How- ever, they assumed M to be constant (Mj) throughout the life of the salmon to simplify com- putations. Lander and Henry (1973), on the other hand, assumed values for m (proportion of fish that mature annually) for coho salmon and then calculated the remaining parameters. Assuming fixed values for the proportion offish that mature annually (m,) permits a unique solu- tion to Equations ( l)-(8), combined with Equations (9)-(16), so that with: mi = nil (fixed) {di) = ,3 (20) and since g-^<^2+'W3) = ^-6F2-12A/3+6M3 ^ g-6(F2+i2M3H6M3 = ginfe4-infe2+6M3 (from Equation (18)) and (from Equation (19)) rin/j4-ln/e2+12M3"| F2 ~ L 6 J _ -(Infe4-Infe2+12M3) Fo+Mo " rin/e4-ln/e2+12M3l ~ -(ln/?4-ln/22+12M3)+6M3 -L § — -J^^3 ln/e4-ln/22+12M3 ln/24-ln/e2+12M3 ln/e4-ln/z2+12M3-6M3 ~ ln/e4-ln/22+6M3 then Equation (20) becomes §. ln/e4-ln/f2+12M3 jz ^3 rei8^i = /?2e-^^3 , u ^ u a^, (l-ei"''4-infc2+6M3) ^ f^ (text Equation (23)). (l-mi)(l-m2) ^ ln/e4-ln/22+6M3 / j v m Since e-^6^'i^i2M2) = /^^ and e-(«^2+i2iW3) = ^ 56 HENRY: NATURAL AND FISHING MORTALITIES OF CHINOOK SALMON then, Equation (14) can be written as ^7 ASM (l-mi)(l-m2)(l-m3) glSMi = g-6Ai2e-6(^'l+M2)g-6M3g-6(F2+M3)g-6M4g-6(F3+M4) k4 = ^ :i± g-6M4g-6F3-6M4 ^2 ^^g-(6F3.12M4) = ^^ The natural logarithm of k^ In/Jg = ln/j4-(6F3+lZ^4) (21) which can be solved for F3 as follows: -6F3 = ln/26-ln/i!4 + 12.\/4 -[In/e6-ln/?4+12M4] ^3 = Q (text Equation (26)). (22) Equation (13) can be written (l-mi)(l-m2)(l-m3)^ ''2^2 F3+M4 ^' ^ ^ ^^ (23) and since e^^^^^^^'> = 6-6^3-1 2M4+6M4 = g-(6F3+i2M4)+6M4 ^ ginfc6-infe4+6iW4 ^^^^^ Equation (21)) and (from Equation (22)) Infe6-Infe4+12M4 F3 6 -[lnfc6-ln^4+12M4] F3+M4 ln/26-ln/j4+12M4 ,. -(ln/e6-ln/e4+12M4)+6M4 6 ^^ ln/j6-ln/24+12A/4 ln/e6-ln/e4+12M4 ln/e6-ln/e4+12M4-6M4 ln/j6-ln/24+6M4 then Equation (23) becomes ^6 IBM, . fi/if •n/?6-ln/?4+12M4 , ^ , ^ ^a^. (l-..)(l-m,)(l-m3) ^'"" = V-'^ n^^M^S^TeiuT d-'"""*""*) = *5 (text Equation ,26,). 57 EFFECT OF SEVERAL DIETS ON SURVIVAL, DEVELOPMENT TIME, AND GROWTH OF LABORATORY-REARED SPIDER CRAB, LIBINIA EMARGINATA, LARVAE Thomas E. Bigfordi ABSTRACT Survival, development time, and growth were determined for larvae of the spider crab, Libinia emarginata, reared with nine diet combinations of algae, rotifers, copepods, ciliates, and Artemia. Percent survival was greater and development times shorter for diets of A. salina nauplii, either alone or in combination with other food sources. Zoeal survival was higher in diets of Artemia at 6 nauplii/ml than at 3 nauplii/ml. Megalopal survival was more variable, being highest in cultures with Artemia and the rotifer Brachionus plicatilis as food. No significant differences were noted in carapace mea- surements of larvae reared on the six diets which supported development beyond stage I zoea. The literature includes many descriptions of de- capod crustacean larval culture in the laboratory. Much of this work has been directed at deriving culture techniques and optimum levels of factors such as temperature and salinity. The "standard" diet has been newly hatched Artemia nauplii, a highly successful, convenient, but increasingly expensive food source. Research trends have been to seek substitute or supplemental diets for the brine shrimp. Foods investigated have included barnacle nauplii (i,awiriski and Pautsch 1969; Reed 1969), the rotifer Brachionus plicatilis (Brick 1974; Sulkin 1975; Sulkin and Epifanio 1975), various ciliates (Sulkin 1975), polychaete larvae (Roberts 1974; Sulkin 1975), and oyster larvae (Roberts 1974). This study was designed to evaluate possible diets, in addition to Artemia nauplii, which will support larval development of the spider crab, Libinia emarginata Leach. Normal larval de- velopment of this species consists of two zoeal stages and one megalopa (Johns and Lang 1977). Parameters used to estimate diet success were survival of larvae to each stage, time to each molt, and carapace size. Development of a satisfactory diet, in combina- tion with the short larval development time, could establish Libinia as a very suitable bioassay or- ganism. The culture methodology described is re- latively simple, further increasing the potential for continued laboratory study. MATERIALS AND METHODS Ovigerous female L. emarginata were collected by otter trawl in Narragansett Bay, during July and August 1976. Females were placed in contain- ers of aerated seawater and immediately trans- ported to the laboratory; storage in the laboratory was in a 1.2-m diameter ( 195-1 volume) Fiberglas^ tank provided with flow-through ambient temper- ature (approximately 20°C) seawater. As the eggs ripened, the females were transferred into tubs containing 8 1 of filtered seawater at 20° and 29- 31%o. After hatching occurred the female was re- moved and the water changed. Within several hours of hatching, the larvae were placed 5/dish in 8.75-cm diameter culture dishes containing 75 ml of filtered seawater. Temperature and salinity were maintained at 20°C and 29-3 l%o. This type of static system has been used commonly to rear other species of crabs (Brick 1974; Sulkin and Norman 1976; Sulkin et al. 1976). The density of 1 larva/15 ml was chosen to allow sufficient room for developing megalopae. Food organisms used included newly hatched San Francisco Bay Brand Artemia salina nauplii, the cihate Euplotes vannus, the copepod^urj'^em- ora affinis, the green flagellate algae Dunaliella viridis, and the rotifer Brachionus plicatilis (Table 1). These organisms are available at the Environ- mental Research Laboratory (Narragansett, R.I.) 'U.S. Environmental Protection Agency, Environmental Re- search Laboratory, South Ferry Road, Narragansett, R.I.; present address: The Center for Natural Areas, 1525 New Hampshire Avenue, NW, Washington, DC 20036. 2 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA or USEPA. Manuscript accepted May 1977. FISHERY BULLETIN: VOL. 76. NO. 1, 1978. 59 FISHERY BULLETIN: VOL. 76, NO. 1 TABLE 1. —Laboratory diets used in rearing Libinia larvae. emarginata Diet symbol Diet Size (Mm) Concen- tration (no./ml) Replicates and no. of larvae used A, Artemis salina 350-400 3 2/80 Ai A salina 350-400 6 1/40 D Dunaliella viridis 15-20 10^-103 1/40 BD Brachionus plicatilis D vindis 55-200 15-20 25 lO'-IO^ 2/80 BD/ABD Stage 1 B. plicatilis D. viridis 55-200 15-20 25 10'-102 2/80 Stage II - M A. salina B. plicatilis D viridis 350-400 55-200 15-20 3 15 10'-10^ EED Eurytemora affinis Euplotes vannus D. viridis 140-243 80-100 15-20 5 5 lO'-IO^ 1/40 ABO A. salina B plicatilis D. viridis 350-400 55-200 15-20 3 15 10'-102 2/80 ABDE A. salina B. plicatilis D. viridis Eurytemora affinis 350-400 55-200 15-20 140-243 3 15 10'-102 5 1/40 S Starved 0 2/80 in mass cultures. Each species is an active swim- mer, thereby satisfying the raptorial feeding re- quirements oiLibinia. As noted in Table 1, several of the diets were replicated with 40 larvae (8 dishes of 5) in each of two trials; the remaining diets were investigated only once. Different trials utilized zoeae from different hatches; all 40 larvae in each diet replicate were from the same hatch. Concentrations of food organisms listed in Table 1 remained constant and were not adjusted as mor- tality occurred. One exception was diet BD/ABD, where the food organism composition was altered after the molt into stage II to include Artemia nauplii. Food and culture water were changed daily. Culture dishes were scrubbed clean in freshwater twice weekly. Larvae were transferred by wide-bore pipette to minimize body damage. Molts were recorded when exuviae appeared in the dishes and were verified under a compound mi- croscope. The criteria for death was complete ab- sence of a heartbeat. Larvae and juvenile crabs were preserved in 10% buffered Formalin for carapace measure- ments. These measurements were determined with an ocular micrometer, with the carapace lengths and widths taken at maximum dimen- sions (Figure 1). Comparisons of development times and measurements were made by one-way analysis of variance, with significant differences Figure l. Body proportions oi Libinia emarginata measured and the lines of measurement used. (A) zoea, (B) megalopa, (C) juvenile, (SH) spine height, (CW) carapace width, (CL) carapace length. (P<0.05) between diets tested by a Scheffe pos- terior comparison test (Nie et al. 1970). RESULTS Survival Figure 2 shows the survival of spider crab larvae reared on each of the nine diets. Experiments con- tinued for 25 days, at which time all larvae had either metamorphosed into the first crab stage or died. Survival data after each stage are shown in Table 2. Only six of the nine diets permitted de- velopment to proceed beyond stage I; in three diets (EED, D, and S) all zoeae died in the first stage. Starved control larvae survived a maximum of 7 days, by which time mortality was 100% (Figure 2). After day 3, all larvae were moribund. Addition o{ Dunaliella viridis (D) did not en- hance either survival or molting. All stage I zoeae were motionless by day 4, but a heartbeat was observed up to day 10. No molts occurred. The dark red or orange chromatophores typically observed 60 BIGFORD: EFFECT OF DIETS ON SPIDER CRAB LARVAE 100 ~ starved (S) 50 \. \A Euplotes vannus, Dunaliella viridis 8* Eurytemora of finis ( E ED) D. viridis (D) ArtemJQ salina, Brochionus plicatilis, D. viridis, and Eurytenriora affinis (ABDE) ~ ° ~ B plicotilis a D. viridis / A. soling, B. plicatilis, and D viridis ( BD/ABD) \ VvA "■^■■- 1 » ^'A "v/o °-v 1 -A- 8 plicatilis a D. viridis (BD) « \ V"-.; ol \ \\\ ~ ° ~ A solino , B plicatilis a D viridis (ABD) \ • 'S \ *• )J ~ ■ ~ A_. sol ing (A|) • \ \ \ V- V- A - A. solino (Ao) \\ ^ \ 'Tv \ ■ — ■— BCia— a 8 10 AGE (days) Figure 2. Percent survival at each day for Libinia emarginata larvae reared on nine laboratory diets. Refer to Table 1 for concentrations and sizes of food organisms in each diet. Table 2. — Survival data and percentages to each stage of Libinia emarginata on the diets permitting larval development past stage I. A^i, Nn, andNn^ represent number of larvae surviv- ing to each stage; Af^ equals initial number. Molt i-m \-*'M l-^J Diet N\INo % A/ll/A/o % nmINq % A, 59/80 74 34/80 43 8/80 10 A2 33/40 83 30/40 75 3/40 8 ABD 58/80 73 40/80 50 10/80 13 BD/ABD 36/80 45 3/80 4 3/80 4 ABDE 32/40 80 18/40 47 BD 29/80 36 on the carapace were absent in nearly all larvae reared on diet D. Survival on diet EED (ciliate, copepod, and al- gae) was only slightly higher than the starved controls (Figure 2). No molts were observed. Mor- tality was 100% by day 8. A diet of Brachionus and Dunaliella (BD) al- lowed development into stage II. With this diet 36% (29/80) of the stage I zoeae molted into stage II, but all died by day 11. Food organisms offered during stage I in diet BD/ABD were identical to diet BD. Artemia nau- plii were added for all ensuing stages. Survival was 45% to stage II and 4% to both the megalopae and juvenile stages. Diet ABD, identical to diet BD/ABD after stage I, allowed 73% survival to stage II, 50% to the megalopae, and 13% to the first crab stage. Higher survival to stage II was achieved by diet ABDE, which included copepod subadults. On this diet, 80% of the zoeae molted successfully into stage II; 47% molted into megalopae. No larvae metamorphosed into the crab stage although sev- eral died during ecdysis. Two diets of newly hatched Artemia nauplii were tested. Diet Ai, with 3 nauplii/ml, yielded 74% survival to stage II, 43% to megalopae, and 10% to the first stage crab. A second diet, Ag (6 nauplii/ml), yielded higher survivals to stage II and megalopae, 83% and 75%, respectively, than any other diet. Survival to the juvenile stage was 61 Development Times Diets supplying A r^em JO nauplii in stage I re- sulted in highest survival to stage II and the shortest development times (Table 3). Of the four diets grouped in the first subset (Table 4) for the molt into stage II, diet ABDE was the best. Diets BD and BD/ABD, although identical in content during stage I, were significantly different. For the molt from zoeal stage II into megalopae diet ABDE again resulted in the shortest de- velopment time. Grouped with ABDE in homogeneous subset I was A2, with the latter diet sufficiently similar in molt time to diet Aj to also be included in subset II. As in the first molt, diet BD/ABD had the longest time to molt. Table 3. — Development times of Libinia emarginata larvae from hatching to each molt for each diet. Diets EED, D, and S did not allow development past stage I. Molt Diet I- l-^M I -►J A, ABD BD/ABD ABDE BD X SD Range X SD Range X SD Range X SD Range X SD Range X SD Range 4,66 060 4-7 4.42 050 4-5 462 0,64 4-6 6,56 1,36 5-9 4.25 0,44 4-5 572 1,28 4-8 1029 1,14 9-14 987 0.51 9-11 10.30 0.85 9-12 13.00 1.73 11-14 9.39 0.50 9-10 18.86 2.48 16-22 1867 3,79 16-23 19,00 2,21 16-24 21,67 306 19-25 FISHERY BULLETIN: VOL. 76. NO. 1 In the last molt, from megalopae to the first crab stage, all four diets tested were grouped as one subset. Of the four, Aj was ranked as the best in terms of development times and BD/ABD was the worst. Carapace Measurements Spine height, carapace length, and carapace width measurements were analyzed by a Scheffe posterior comparison test (Table 5). Zoeal stage II and juvenile crab measurements were not sig- nificantly (P = 0.05 level) different and were grouped into one homogeneous subset; carapace lengths of megalopae were similar in all diets. Only the carapace widths of megalopae proved statistically different, with two subsets describing the measurements of the larvae reared on differ- ent diets. Ranking within each subset provides an indica- tion of possible trends in size with respect to the diets tested. This trend is most evident in zoeal stage II; in both spine height and carapace length the ordering of diets was identical, with A2 superior and ABD second. In megalopae and Table 4. — Homogeneous subsets of diets tested on Libinia emarginata larvae as determined by analysis of variance and Scheffe posterior comparisons (P< 0.05) of development times. Shortest times are listed in subset I and longest in subset III. Subset Stage I -HI ll-^M M-^J ABDE. A2. ABD, A, BD BD/ABD ABDE, A 2 A2, A,, ABD BD/ABD A,, A2. ABD, BD/ABD Table 5. — Carapace measurements for stage II, megalopa, and juvenile Libinia emarginata reared on various diets. Mean values (in millimeters) of spine height (SH), carapace width (CW), and carapace length (CL) are given in ranked order within each homogeneous subset of similar values. Roman numerals followdng the diet symbol denote replicate number, if applicable. Diets not represented, e.g., A, in stage II, could not be analyzed because of insufficient data. Larval stage Parameter measured Subset Ranked order of means Zoea II Megalopa Juvenile CL BD/ABD- II ABDE ABD-II A2 0,859 0,865 0 936 0 970 SH BD;ABD-II ABDE ABD-II A2 0,311 0.316 0 338 0 360 ^~^-, CW A, -II A2 A,-l ABD-I 0,938 1.037 1,044 1 100 1 A2 A,-l ABD-I ABDE ABD-II 1,037 1,044 1,100 1.136 1.153 CL ABDE A,-ll A,-l ABD-II A2 ABD-I 1,232 1 258 1 260 1 265 1 289 1 380 CW BD/ABD-I A1-II A,-l ABD-II ABD-II A2 1,233 1.284 1,340 1.347 1 393 1.420 CL A2 ABD-II BD/ABD-I A, -II ABD-I A,-l 1.560 1.567 1.575 1.644 1.690 1.705 62 BIGFORD: EFFECT OF DIETS ON SPIDER CRAB LARVAE juveniles, diet ABD (replicates I and II) often re- sulted in the largest measurements. DISCUSSION Survival Based on survival, laboratory diets that in- cluded Artemia salina nauplii were better than diets consisting solely of rotifers, algae, ciliates, or copepod nauplii. However, when offered in combi- nation with brine shrimp nauplii, rotifers and copepods may provide some nutritional value to the larvae. Survival percentages to zoeal stage II were very high with diet ABDE: diet ABD pro- duced the best survival to the first stage juvenile. Johns and Lang (1977; unpubl. data), using an excess diet of A. salina and a compartmented box culture system, got 20% survival to first stage crab. The success oi Artemia nauplii as a laboratory diet is well documented (e.g.. Brick 1974; Sulkin et al. 1976). Studies by Brick ( 1974) also showed that survival ofScylla serrata to megalopae increased as the concentration of Artemia nauplii was in- creased. Results showed a 25% survival to megalopae at concentrations of 5 nauplii/ml and 44% at 16 nauplii/ml. Differences in survival on various diets is com- monly observed in laboratory studies. Diets that permit partial development, e.g., diet BD in this study, normally yield correspondingly lower sur- vival. This trend has also been observed in diet studies on larvae of the sand shrimp, Crangon septemspinosa (Bigford^). Division of molt times into three subsets during zoeal development infers thatL. emarginata may prefer certain food types or sizes at different stages. Diets including Artemta also consisted of the largest size food particles, with copepods, roti- fers, ciliates, and algae being smaller. This possi- ble discriminate particle selection was not ob- served in megalopae; all diets were consumed equally and development times were similar. All larvae surviving to first stage crabs were reared on Artemia, alone or in combination, during stage II and megalopae. The lack of development observed in diets D, EED, and S, plus the partial development in BD, is supported by the literature. Studies by Sulkin (1975) have shown that algae and ciliates do not satisfy the nutritional requirements of brachyuran zoeae. Broad (1957) concluded that various algal diets were similar to starved con- trols, with the addition of animal matter required for metamorphosis in grass shrimp, PaZaemone^es, larvae. Particle size and biochemical composition, among other factors, may limit development and survival. Conversely, rotifers have been found to enhance survival and development of several other decapod larvae, most notably the blue crab, Callinectes sapidus (Sulkin and Epifanio 1975). Food size appears to be the controlling factor in selection of the rotifer as food for early stage zoeae of the blue crab. Although ABDE was a successful diet in the zoeal stages, it did not sustain metamorphosis to the crab stage in this study. Perhaps at differing concentrations of Artemia and Eurytemora the diet would prove more successful for megalopae. Development Times The diets resulting in the shortest development times closely parallel those yielding the highest survival percentages. These diets all include Ar- temia nauplii (Tables 3, 4). For the molt from zoeal stage I to stage II the shortest development times were recorded for diets ABDE and A2, which also are the diets yield- ing maximum survival to stage II. These same diets continue to rate high in terms of survival and molt time for the second molt also. ^Bigford, T. E. 1975, The effects of diet on larval development of the early stages of the sand shrimp Crangon septemspinosa Say. Unpubl. manuscr. U.S. Environmental I^search Lab., Nar- ragansett, R.I. Carapace Measurements There does not appear to be a significant differ- ence in carapace size between the diets studied. Instead, the effects of diets were manifested in terms of development rate. Larvae apparently molt upon reaching a certain biomass, with the postmolt sizes similar in most cases. Carapace length measurements for second stage zoeae and megalopae (Table 5) for diets Aj and Ag compare favorably with the values reported by Johns and Lang (1977) in their description of the larvae reared on excess concentrations of Artemia. Their mean measurements of 0.94 mm and 1.21 mm, respectively, were only slightly below the values reported here. Differences in measuring 63 FISHERY BULLETIN: VOL. 76, NO. 1 techniques could account for the larger megalopa carapace lengths reported in this paper. CONCLUSIONS The results of this experiment suggest that a combined diet including at least 5 Artemia nauplii/ml would produce highest survival in the zoeae. Additional food organisms may be required by megalopae. Faster development times as- sociated with diet Ag, compared with Aj, em- phasize the importance of food concentration in addition to food type. Limited success of diet ABDE in the zoeal stages implies that Eurytemora affinis subadults may provide some nutritional substance to spider crab larvae. Replication of the copepod diet alone would be required to verify the potential oi Eurytemora. Each of the diets permitting development to proceed through metamorphosis resulted in a low percent survival. This could be partially explained by the static dish system used to culture the lar- vae. Flow-through designs would control water quality and perhaps microbial infestations. With an improved culture design, a satisfactory diet, and the short development time, L. emarginata could prove to be a very satisfactory bioassay or- ganism. The biochemical content o^ Artemia nauplii may account for their value in the diet of spider crab larvae. As determined by Sulkin ( 1975), A. salina contain 30 total lipid/unit dry weight, a value far superior to that o{ Brachionus plicatilis (9%). A diet of fertilized polychaete (Hydroides dianthus) eggs, containing 20% total lipid, also sustained complete development of Callinectes sapidus in Sulkin's experiments. The lipid content of Eurytemora was not determined. Each of the diets tested in this experiment re- sulted in a normal progression of larval develop- ment forL. emarginata (Johns and Lang 1977). No supernumerary stages or characters appeared ACKNOWLEDGMENTS I thank Allan D. Beck, Richard Brooks, Neal Goldberg, D. Michael Johns, William H. Lang, and Leslie Mills, all of the Environmental Research Laboratory at Narragansett, for assistance during the course of the experiment. The graph was drawn by Annette Doherty; photographs were completed by James Brennan. The manuscript was typed by Josephine DeVoU. LITERATURE CITED Brick, R. W. 1974. Effects of water quality, antibiotics, phytoplankton and food on survival and development of larvae ofScytla serrata (Crustacea: Portunidae). Aquaculture 3:231- 244. Broad, a. C. 1957. The relationship between diet and larval develop- ment of Palaemonetes. Biol. Bull. (Woods Hole) 112:162-170. JOHNS, D. M., AND W. H. LANG. 1977. Larval development of the spider crab, Libinia emarginata (Majidae). Fish. Bull., U.S. 75:831-841. L'awinski, L., and F. Pautsch. 1969. A successful trial to rear larvae of the crab Rhi- thropanopeus harrisi (Gould) subsp. tridentatus (Mait- land) under laboratory conditions. Zool. Pol. 19:495-504. NiE, N. H., D. H. Bent, and C. H. Hull. 1970. Statistical package for the social sciences. McGraw-Hill, Inc., N.Y., 343 p. Reed, P. H. 1969. Culture methods and effects of temperature and sa- linity on survival and growth of Dungeness crab (Cancer magister) larvae in the laboratory. J. Fish. Res. Board Can. 26:389-397. Roberts, M. H., Jr. 1974. Larval development of Pagurus longicarpus Say reared in the laboratory. V. Effect of diet on survival and molting. Biol. Bull. (Woods Hole) 146:67-77. Sulkin, S. D. 1975. The significance of diet in the growth and develop- ment of larvae of the blue crab, Callinectes sapidus Rathbun, under laboratory conditions. J. Exp. Mar. Biol. Ecol. 20:119-135. Sulkin, S. D., E. S. Branscomb, and r. e. Miller. 1976. Induced winter spawning and culture of larvae of the blue crab, Callinectes sapidus Rathbun. Aquaculture 8:103-113. Sulkin, S.D., AND C. E. Epifanio. 1975. Comparison of rotifers and other diets for rearing early larvae of the blue crab, Callinectes sapidus Rathbun. Estuarine coastal Mar. Sci. 3:109-113. Sulkin, S. D., and K. Norman. 1976. A comparison of two diets in the laboratory culture of the zoeal stages of the brachyuran crabs Rhi- thropanopeus harrisii and Neopanope sp. Helgol. wiss. Meeresunters 28:183-190. 64 DESCRIPTION OF REARED EGGS AND YOUNG LARVAE OF THE SPOTTED SEATROUT, CYNOSCION NEBULOSUS William A. Fable, Jr., Theodore D. Williams, and C. R. Arnold ' ABSTRACT Adult spotted seatrout, Cynoscion nebulosus, were induced to spawn in the laboratory by controlling temperature and photoperiod. Development of eggs and larvae, reared at 25°C, is described to 15 days after hatching. The pelagic, spherical eggs have a mean diameter of 0.77 mm, and usually contain one oil globule averaging 0.22 mm in diameter. Hatching occurs about 18 h after fertilization. Standard length at hatching is between 1.30 and 1.56 mm. Spotted seatrout average 4.4 mm standard length at notochord flexion. The larvae, which were fed the rotifer, Brachionis plicatilis, tmd nauplii oiArtemia sp., grew to about 4.5 mm standard length in 15 days. The spotted seatrout, Cynoscion nebulosus, is one of the most important fishes to both recreational and commercial fishermen in the Gulf of Mexico and southeastern United States. In the Gulf it ranks first in weight landed by sports anglers (Deuel 1973) and seventh by weight taken com- mercially (U.S. Department of Commerce 1975). Despite its value, the eggs and youngest larval stages have not been adequately described in pre- vious literature. Four early works (Welsh and Breder 1923; Hil- debrand and Schroeder 1928; Pearson 1929; Hil- debrand and Cable 1934) provided descriptions of spotted seatrout development. Welsh and Breder (1923) described juvenile C. nebulosus as small as 28 mm, collected from North Carolina and Chesapeake Bay waters. Hildebrand and Schroeder (1928) illustrated a spotted seatrout presumably 120 mm long, apparently from Chesapeake Bay. Spotted seatrout from Texas as small as 7.8 mm were described by Pearson (1929). The most complete description of young spotted seatrout was by Hildebrand and Cable ( 1934). The smallest seatrout described by them was 1.8 mm long and was taken off North Carolina. The only other illustrations of larval spotted seatrout were of 3.0 and 5.0 mm SL fish from south Florida by Jannke (1971). The first description of C. nebulosus eggs was by Miles ( 1950, 1951 ). He stated that eggs measured from 0.70 to 0.98 mm in diameter and contained one to four small oil globules. Later, Tabb (1966) stated that eggs were spherical and normally had one oil droplet, but sometimes two or three. In this paper, we provide detailed descriptions of eggs and young larvae of spotted seatrout, based on laboratory spawned and reared specimens. PROCEDURES Adult spotted seatrout were caught by hook and line at Port Aransas, Tex., in August 1973. Eleven fish (seven males and four females) were brought into the laboratory and maintained in a 30,000-1 seawater tank. The tank was constructed of fiber glass and measured 6 x 3 x 1.5 m. It contained seawater which was recirculated through a shell- and-gravel filter. The fish were fed shrimp and fish, both live and dead. Temperature and photoperiod in the laboratory were adjusted to simulate spring and, subsequently, summer conditions. Spawning began 1 mo after conditions were stabilized at 15 h of light, 9 h of dark, and 26°C. Details of the methods to induce spawning by spotted seatrout are described by Arnold et al. (in press). In a 1-yr period, the spotted seatrout have spawned during each month for a total of 82 times. On several occasions more than one female spawned. Eggs described in this paper were spawned by a single female on 8 September 1975. They were preserved hourly in S9c buffered Formalin^ from 'Southeast Fisheries Center Port Aransas Laboratory, Na- tional Marine Fisheries Service, NOAA, Port Aransas, TX 78373. Manuscript accepted March 1977. nSHERY BULLETIN: VOL. 76, NO. 1, 1978. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 65 FISHERY BULLETIN; VOL. 76, NO. 1 the time of spawning until hatching. Larvae de- scribed in this report were from eggs spawned on 7 October 1975 and were transferred to rearing aquaria. Samples of larvae were preserved daily in Wc Formalin for 15 days. Larvae were reared in 57-1 aquaria which were filled with algae and rotifer culture water 2 days prior to the introduction offish. Algal growth was enhanced by a constant light source above the aquaria. Seatrout were fed the rotifer, Brachionis plicatilis, daily at a rate of at least 20/ml of water for 4 days. On the fifth day rotifers and brine shrimp, Artemia sp., nauplii (3-5/ml) were both introduced. This combination was fed until larvae were 8 days old, then brine shrimp were used as the only food source. Temperatures in the aquaria were maintained at 24.0° to 26.0°C. Eggs and larva were measured using an ocular micrometer in a dissecting microscope. Measure- ments included total length, standard length, snout to anus length, snout length, head length, eye diameter, and body depth. Illustrations are of preserved specimens. In discussing seatrout eggs, the three stages described by Ahlstrom and Ball (1954) are used. EMBRYONIC DEVELOPMENT Spotted seatrout eggs are pelagic and spherical, the chorion is clear and unsculptured, and the yolk is homogeneous. The perivitelline space in live eggs is narrow, occupying approximately 4% of the egg diameter. One hundred live eggs and 100 Formalin-preserved eggs were measured at vari- ous stages of development. No differences in diameters of eggs or oil globules were noted at different stages of development. Diameters of both live and preserved eggs averaged 0.77 mm. The diameter of live eggs ranged from 0.73 to 0.82 mm while diameters of preserved eggs ranged from 0.70 to 0.85 mm. The eggs usually contain one yellow oil globule, but some eggs (2% ) have two or three globules. Oil globules in preserved eggs range from 0.18 to 0.26 mm in diameter, with a mean of 0.22 mm. Oil globules in live eggs range from 0.22 to 0.27 mm in diameter, with a mean of 0.23 mm. When more than one globule is present, sizes vary greatly. Early Stage Eggs Duration of the early stage is about 8 h. Eggs preserved in Formalin have yellowish oil globules. opaque cells, and, in this early stage, a shrunken and disorganized yolk (possibly due to poor pre- servation). Eggs float with the oil globule(s) on top and the developing cells on the bottom. Development proceeds as follows: IV2 h, 16- or 32-cell stage; 2 h, morula stage; 3 h, blastula stage; 4 h, gastrulation begins; 6 h, gastrula encircles two-thirds of the yolk and primitive streak is evi- dent; 8 h, blastopore closure. At the onset of gas- trulation, numerous small droplets form around the oil globule. By blastopore closure, optic vesi- cles are visible in most eggs and the notochord can be seen in some. Myomeres are not discernable and no pigmentation is present on the egg or em- bryo. 0.5 mm 1.0mm Figure 1 . — Spotted seatrout embryos: A) 15 h after fertilization; B) at hatching (SL 1.46 mm). 66 FABLE ET AL.; DESCRIPTION OF EGGS AND LARVAE OF SPOTTED SEATROUT Middle Stage Eggs Duration of the middle stage is about 4 h. At 9 h after fertilization, the notochord develops further and the forebrain begins to develop. Small mel- anophores are present for the first time around the optic vesicles and in no apparent pattern along the body. In the 10th hour, six to eight myomeres can be seen with difficulty on the posterior one-third of the embryo. By the 12th hour, the embryo extends over about one-half the circumference of the egg. Late Stage Eggs The tip of the tail of the embryo has separated from the yolk and the finfold is evident on both the posterior dorsal and ventral caudal regions at 13 h. Eighteen to 20 myomeres are present. Melano- phores, which are present over the entire body of the embryo, are concentrated around the dorsal surfaces of the eyes, on either side of the notochord, and along the base of the finfold. At 15 h (Figure lA), the tail of the embryo is well past the oil globule and has developed a marked curve. The finfold surrounds the posterior half of the embryo and 24 to 25 myomeres can be counted. Internal organs show some differentiation, while anteriorly the eyes are pronounced and the hind- brain is developing. One hour later, the embryo occupies three-fourths of the circumference of the egg. Twenty-five myomeres are apparent. Hatching occurs 16 to 20 h after fertilization, when incubation temperatures are approximately 25°C. In other experiments, hatching occurred in 15 h at 27 °C and in 21 h at 23 T. LARVAL DEVELOPMENT Hatching (Figure IB) Standard lengths of 20 newly hatched larvae ranged from 1.30 to 1.56 mm and averaged 1.46 mm. At hatching the oil globule is located at the posterior end of the yolk sac. Some scattered melanophores like those in the embryos are still found, but most are indistinct, especially those along the finfold. No pigment is visible in the yolk or on the oil globule. Sixteen Hours Posthatching (Figure 2A) At 16 h, larval standard lengths ranged from 1.89 to 2.10 mm and averaged 2.03 mm. The finfold is large and clear with no fin differentiation. The mouth is undeveloped, only a little yolk remains, and the oil globule is still in a posterior position. Otocysts are faintly visible within the otic cap- sule. Pectoral fin buds are evident for the first time. The alimentary canal is straight, terminat- ing at the anus in the anterior half of the body. Body pigments are in four vertical bands located above the abdomen, above the anus, and one-third and two-thirds of the distance from the anus to the tip of the notochord. Small melanophores are con- centrated in these bands, but many disappear with preservation. The most prominent of the bands is located one-third of the way from the anus to the notochord tip. Pigmentation in preserved speci- mens is most distinctive in the head region. Sev- eral small dendritic melanophores are located above and behind the eye. Two dendritic melano- phores are located on the dorsomedial surface of the head. Some slight black pigmentation is visi- ble above the abdomen where the first pigment band is located. Numerous granular melano- phores are also found on the finfold at the dorsal and ventral body margins at the notochord tip. Forty Hours Posthatching At 40 h, the larvae average 2.10 mm SL, the mouth is formed, and the yolk sac is almost com- pletely gone. The head has grown very deep, and the brain appears dorsally over the eyes. In pre- served fish the eye is totally black, and pectoral fins stand out from the sides. Internal organs are increasing in size and complexity, but the alimen- tary canal is still straight, although thicker than at hour 16. Pigmentation undergoes distinctive changes prior to 40 h of age. The four vertical bands which occur on the 16-h larva are absent, and only one wide, diffuse band is found just forward of the half-way point between the anus and the tip of the notochord. Melanophores are intensifying along the dorsal and ventral body margins within the band and anteriorly over the abdomen. The granu- lar melanophores on the finfold at the tip of the notochord are somewhat fewer in number. Dendri- tic melanophores are on the dorsal surface of the abdomen. Pigmentation on the lower jaw is heaviest at the angle and posteriorly. A few small melanophores are anterior to this and at the tip of the lower jaw. The pigment which remains least distinct and disappears after a short period in Formalin is that 67 FISHERY BULLETIN: VOL. 76, NO. 1 B 1.0mm around the eye and dorsal surface of the head. Concentrations of small amber chromatophores are found ventral and posterior to the eye, while several larger yellow chromatophores are found above the eye. Several amber chromatophores are also located medially on the dorsal surface of the head. 68 Sixty-Four Hours Posthatching (Figure 2B) Larvae at 64 h past hatching range from 2.06 to 2.15 mm SL and average 2.12 mm SL. The yolk is completely absorbed, the gut has become convo- luted, and the intestine is very thick. FABLE ET AL.: DESCRIPTION OF EGGS AND LARVAE OF SPOTTED SEATROUT 1.0 m m 1.0 mm Figure 2.— Spotted seatrout larva: A) 16 h posthatching (SL 2.03 mm); B) 64 h posthatching (SL 2.12 mm); C) 112 h posthatching (SL 2.12 mm); D) 232 h posthatching (SL 2.71 mm); E) 328 h posthatching (SL 4.21 mm). Eye pigmentation is complete and very reflec- tive. The diffuse band found in 40-h fish is still present but is indistinct. Basic pigment patterns and melanophore placement remain similar to 40-h fish except in the following cases. Pigment is increasing along the dorsal surface of the abdo- men, and anteriorly towards the eye. The melano- phores on the tip of the lower jaw are more distinct. Some pigment is also present on the ventral sur- face of the abdomen. Four and Five Days Posthatching (Figure 2C) In a typical spotted seatrout 112 h old, standard lengths vary from 2.04 to 2.15 mm and average 2.12 mm. The mouth is well-developed and the maxillary is prominent. Dendritic melanophores are found from the upper surface of the abdomen posteriorly to two- thirds of the length of the tail along the ventral midline. They radiate ventrally over the outer ab- dominal surface. Melanophores on the tail radiate dorsally from the ventral margin and ventrally from the dorsal margin. Large dark melanophores are present on the preserved larvae at this age but are somewhat variable. One is found immediately ahead of the anus (an important characteristic in sciaenid larvae), and two to three more occur an- teriorly below the abdomen. Another is located at the angle of the lower jaw. One or two are on the 69 FISHERY BULLETIN; VOL. 76, NO. 1 dorsal surface of the body above the abdomen. Melanophores on the finfold at the tail vary greatly; they are found both on the dorsal and ventral body margins in varying numbers. A single dendritic melanophore is present anterior to the eye, and two or three more are posterior to the eye. Six Through Eight Days Posthatching At this age, there is little difference in body form and structure from that in Figure 2c. Standard lengths at 160 h average 2.06 mm and range be- tween 1.80 and 2.23 mm. The preopercle can be seen on some larger specimens. Pigmentation has become more intense and is expanding. Principal changes in the dendritic pigments involve the ventral expansion of melanophores on the upper surface of the abdo- men, and the coalescence of tail pigmentation into dark stripes. Indistinct pigment occurs from the eye to the tip of the snout. Melanophores are still found anterior to the anus and have increased in number below the abdomen. A melanophore spot is still found on the tip of the lower jaw. Nine Through Eleven Days Posthatching (Figure 2D) During this 3-day period the larvae begin to grow appreciably in length. By 11 days, standard lengths average 2.92 mm and range from 2.37 to 3.48 mm. Six small teeth are present on the upper jaw and four on the lower jaw at this age. The preopercle is more evident and a small spine can be seen. Branchiostegal rays are present for the first time. The pectoral fin is still membraneous. Some larvae have a presumptive hypural plate below the notochord tip, but no notochord flexion is observed. Pigmentation undergoes only minor changes in this period. Principal body pigment gives the ap- pearance of a dark stripe from snout to tail. Melanophores are now evident on the lateral line giving the impression of a series of dashes. Tiny melanophores are present on the midlateral tail region and both ahead of and behind the eye within the pigment stripe. Twelve Through Fifteen Days Posthatching (Figure 2E) Standard lengths at 12 days average 3.35 mm, and increase to 4.59 mm at 15 days. The preoper- cular spine is prominent, and on the larger speci- mens second and third spines are visible below the first. By the 14th day (at a size of 4.4 mm SL) notochord flexion has occurred in all specimens. As many as 18 caudal rays are first seen at 13 days (4.0 mm SL), and by 15 days (4.4 mm SL), 25 dorsal rays and 10 anal rays are evident. Teeth are found on both jaws (10 on the upper and 6 on the lower). At this age, the pigmentation still gives the appearance of a stripe from the snout through the eye to the upper abdomen, and on the lateral line and ventral tail surface. Melanophores are still located at the tip and posterior to the angle of the lower jaw, on the tip of the upper jaw, and along the ventral margin of the abdomen. The spot an- terior to the anus is indistinct. Pigmentation around the eye is localized in an anterior and pos- terior position within the pigment stripe. The den- dritic melanophores on the upper abdominal sur- face are still large and distinct. Dendritic melanophores are heavily concentrated along the lateral line and also along the ventral margin of the tail. The dorsal tail margin has less pigmenta- tion. A single large dendritic melanophore is found on the base of the caudal fin. Other pigmen- tation is widely scattered over the entire tail. Seatrout preserved for long periods seem to lose the melanophore on the caudal fin but other body melanophores remain visible. GROWTH Larval spotted seatrout grew from about 1.5 mm SL at hatching to about 4.5 mm SL in 15 days. A. K. Taniguchi (pers. commun.) at the University of Miami has observed faster growth of larval spot- ted seatrout. He raised larvae at various tempera- tures and fed them copepods. At 2 wk of age, we noted cannibalism in our seatrout larvae even though ample food of appropriate size appeared to be present. Measurements were made of preserved larvae. The data were tabulated according to size and age (Table 1). Standard lengths of larvae were consis- tently 93 to 95% of the total length until flexion of the notochord occurred at 14 or 15 days; then the standard length decreased to 88'yfof total length. Preanal lengths at 1 day posthatching were 44% SL, 36% SL at 5 days, and 54% SL at 15 days. This indicated that the preliminary decrease in gut length appeared to be associated with yolk absorp- tion. After 5 days, the gut length steadily in- 70 FABLE ET AL.: DESCRIPTION OF EGGS AND LARVAE OF SPOTTED SEATROUT Table l. — Average age (hours) and measurements (millimeters) of preserved larval spotted seatrout of known size. Standard length Number of Snout to Snout Head Eye Body range Age specimens anus length length length diameter depth 1.70-1 89 118 4 0.79 0.09 0.43 0.21 0.50 1.90-2,09 92 32 085 0.10 0.44 0.22 0.53 2 10-2.29 122 52 0.91 012 0.49 023 0.54 230-249 216 6 1 22 0.17 0.69 0.28 065 250-269 248 3 1.38 0.21 083 0.30 0.75 2.70-289 244 4 1.41 0.20 082 031 0.77 290-3.09 253 7 1.47 0.21 0.88 0.32 0.80 3 10-329 274 8 1 60 0.26 1.01 0.34 0.90 3.30-349 290 7 1.72 0.29 1.06 035 0.91 3.50-369 304 3 1.78 0.27 1 08 0.35 0.96 3.70-3.89 316 4 1 95 0.35 1.20 040 1-01 3.90-4.09 323 5 2.06 0.34 1.29 0.42 1 07 4.10-4.29 323 5 2.20 0.40 1.37 0.38 1.13 4.30-4.49 344 3 2.46 041 1.50 0.45 1.24 4.50-4 69 — 0 — — — — — 4.70-4 89 352 1 2.56 0.43 1 63 0.48 1.35 490-5.09 342 5 268 0.45 1 68 0.48 1 36 5.10-5.29 — 0 — — — — — 5.30-5.49 352 1 3.05 0.47 1.84 0.52 1.48 creased relative to standard length. Snout length increased relative to standard length from 3'7( at 1 day to 97c at 15 days. Similarly, head length in- creased relatively from 19-207f SL to 347^ SL. Both these changes were due to rapid development of the mouth and head. Eye diameter and body depth varied only slightly during development. Eye diameter was between 9 and 119c SL at all stages, while body depth varied from 22 to 28'7f SL at all ages. ACKNOWLEDGEMENTS We thank Dinah Bowman for illustrating the eggs and larvae. Appreciation is also expressed to Jeff Messinger who assisted in many aspects of the study. We express our gratitude to Edward Houde and William Richards for reviewing drafts of this paper and for their informative critiques. LITERATURE CITED AHLSTROM, E. H., AND O. P. BALL. 1954. Description of eggs and larvae of jack mackerel ( Trachurus symmetricus) and distribution and abundance of larvae in 1950 and 1951. U.S. Fish Wildl. Serv., Fish. Bull. 56:209-245. ARNOLD, C. R., T. D. WILLIAMS, W. A. FABLE, JR., J. L. LASSWELL, AND W. H. BAILEY. In press. Methods and techniques for spawning and rear- ing spotted seatrout in the laboratory. Proc. 30th Annu. Conf Southeast. Assoc. Game Fish Comm. DEUEL, D. G. 1973. 1970 Salt-water angling survey. U.S. Dep. Com- mer., NOAA, NMFS Curr. Fish. Stat. 6200, 54 p. HILDEBRAND, S. F., AND L. E. CABLE. 1934. Reproduction and development of whitings or kingfishes, drums, spot, croaker, and weakfishes or sea trouts, family Sciaenidae, of the Atlantic coast of the United States. U.S. Bur. Fish., Bull. 48:41-117. HILDEBRAND, S. F., AND W. C. SCHROEDER. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43(1), 366 p. JANNKE, T. E. 1971. Abundance of young sciaenid fishes in Everglades National Park, Florida, in relation to season and other variables. Univ. Miami, Sea Grant Tech. Bull. 11, 128 p. MILES, D. W. 1950. The life histories of the spotted seatrout, Cynoscion nebulosus, and the redfish, Sciaenops ocellatus. Tex. Game Fish Comm. Mar. Lab. Annu. Rep. 1949-1950, 38 p. 1951. The life histories of the sea-trout, Cynoscion nebulo- sus, and the redfish, Sciaenops ocellatus: Sexual develop- ment. Tex. Game Fish Comm. Mar. Lab. Annu. Rep. 1950-1951, 11 p. PEARSON, J. C. 1929. Natural history and conservation of redfish and other commercial sciaenids on the Texas coast. Bull. U.S. Bur. Fish. 44:129-214. TABB, D. C. 1 966. The estuary as a habitat for spotted seatrout, Cynos- cion nebulosus. In R. F. Smith (chairman), A symposium on estuarine fisheries, p. 59-67. Am. Fish. Soc. Spec. Publ. 3. U.S. DEPARTMENT OF COMMERCE. 1975. Fishery statistics of the United States, 1972. NOAA, NMFS Stat. Dig. 66, 517 p. WELSH, W. W., AND C. M. BREDER, JR. 1923. Contributions to life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141-201. 71 PHYSICAL AND CHEMICAL CHANGES OF PINK SHRIMP, PANDALUS BOREALIS, HELD IN CARBON DIOXIDE MODIFIED REFRIGERATED SEA WATER COMPARED WITH PINK SHRIMP HELD IN ICE Fern A. Bullard and Jeff Collins* ABSTRACT Pink shrimp, Pandalus borealis, were held in carbon dioxide modified refrigerated seawater for 12.5 days and in ice for 11.5 days. Chemical tests for spoilage indicated that shrimp held in carbon dioxide modified refrigerated seawater were acceptable up to 9.5 days and those held in ice up to 6.5 days. Data on weight, yield, solids, carotenoids, protein, salt, and pH are given. When expressed on a constant basis (salt-free, 75% moisture), the yield of cooked product calculated from the gross weight of whole shrimp decreased rapidly during the first few days in either system. The yield of cooked meats from the carbon dioxide modified refrigerated seawater system decreased from 18.3% at 0.5 day to 15.3% at 4.5 days but varied in the ice system between 14.0 and 15.5% over the useful holding {jeriod of 6 days. The advantages and disadvantages of the refrig- erated seawater system (RSW) for holding fish and shellfish are well documented and were recently discussed by Barnett et al. (1971) and by Nelson and Barnett (1971). Based on bacteriological mea- surement and sensory evaluation, these authors showed that rockfish, Sebastodes flauidus, can be held in the RSW system modified by the addition of carbon dioxide (MRSW) for longer periods of time than in ice. The purpose of this study was to obtain detailed information on the physical and chemical changes that occur during time of holding of pink shrimp in the MRSW system compared with that of pink shrimp held in ice. EXPERIMENTAL Preparation of Pink Shrimp draining for 30 min resulted in nearly constant weight. The MRSW holding portion of the experiment was conducted as follows. Baskets of shrimp and loose shrimp were alternately placed in the MRSW tank containing a 3.5% brine at -1.7°C, previously treated with carbon dioxide to 3.92 pH. The final loading ratio of shrimp to brine was 1:1.4 (wt/wt). The ice holding portion of the experiment was conducted as follows. Samples of shrimp for ice holding were similarly rinsed, drained for 30 min, and adjusted to 2,100 g each before being placed in single layer cheese cloth "baskets" and covered with ice and 38.5 kg (85 lb) loose shrimp. Loose shrimp were mixed with ice to more closely simu- late boat holding conditions. Fresh ice was placed on the ice-held samples daily to insure a minimum 15-cm (6-in) cover over any given sample. Pink shrimp, Pandalus borealis, when received by the laboratory, had been held for 2 h at ambient temperature of -1.7°C (29°F) without ice aboard a commercial fishing vessel. Shrimp were separated from fish and after a brief rinse in cold freshwater were placed in fiber glass-coated hardware cloth baskets and rinsed again in cold tapwater for 2 min. The shrimp were then drained for 30 min and the weight of each sample was adjusted to contain 2,100 g (4.63 lb). It had been established that Holding Tank and Refrigeration Unit A 568-1 (150-gal) fiber glass holding tank was connected to a refrigeration unit by three 3.81-cm (IVa-in) flexible plastic hoses. The brine was circu- lated at 151 1/min (40 gal/min) through a shell and tube heat exchanger with the capacity to chill 454 kg (1,000 lb) of shrimp and brine from 10° to -1.7°C (50° to 29°F) in 3 h. Refrigeration was provided by a conventional Freon^ 12 condensing unit. The 'Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, P.O. Box 1638, Kodiak, AK 99615. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Manuscript accepted June 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. 73 FISHERY BULLETIN: VOL. 76, NO. 1 tubes in the heat exchanger were made of titanium to avoid corrosion. Carbon dioxide was metered into the suction side of the pump for maximum diffusion; at a rate of 0.2 1/min (0.2 ft^/h), thepH was lowered from 7 to 4 in 5 h. In this 14-day experiment, 8.8 kg (4 lb) of carbon dioxide were used. A chest-type home freezer was used as an insu- lated box to hold shrimp in ice. One day before its use the refrigeration was disconnected and the door raised to allow the ice to begin to melt in order to simulate conditions in a boat's hold. SAMPLING Sampling Procedure In this comparative holding experiment, a sam- ple was taken daily from each holding system and allowed to drain for 30 min before weighing (iced shrimp were first rinsed briefly in cold water to remove ice). Four subsamples were prepared from each sample — three to represent commercial practices and the fourth for laboratory analyses to determine chemical changes in the shrimp. The subsamples, stored at -34°C (-30°F) for later analyses, are as follows. Subsampling Procedure 1. Whole shrimp: The total weight of whole shrimp was determined at each period of hold- ing to simulate the weight of shrimp landed at the dock and to determine yield of products. Water uptake was determined by a solids analysis in a blended sample. 2. Hand-peeled, raw shrimp meats: This laboratory sample was used to determine basic chemical changes, mainly spoilage. 3. Hand-peeled, raw, washed shrimp meats: This sample simulated a machine peeled raw frozen product. Washing was required to ap- proximate the leaching action of commercial machine peelers. The hand-peeled meats were washed gently in cold water for 2 min, drained on hardware cloth for 10 min, then weighed, and frozen for later analyses. 4. Hand-peeled, washed, cooked shrimp meats: This sample simulated a cooked frozen product. A portion of the washed meats after frozen storage, as in 3 above, was thawed and cooked in boiling water for 2 min at a 12:1 ratio, drained 1 min, cooled 5 min, and blended for analyses. To prepare subsamples 2 and 3 for analyses, they were removed from the freezer left at room temperature for 2 h, stored in a refrigerator over- night to thaw, and then blended. Analytical Techniques After the frozen samples were thawed and blended, the following analyses were performed: total nitrogen, solids, total chloride (Horwitz 1975), total volatile base (TVB; Stansby et al. 1944), total volatile acid (TVA; Friedemann and Brook 1938), trimethylamine (TMA; Dyer 1945), and carotenoids (Kelley and Harmon 1972). Sodium and potassium were determined by using a Beckman Model B hydrogen-oxygen flame photometer on appropriate dilutions of a 20-g sample digested with nitric and perchloric acids. The pH of the brine was determined daily. RESULTS AND DISCUSSION Whole Shrimp The change in weight of whole shrimp held in these systems has commercial importance. The yield of product obtained in a processing plant is calculated from the weight of whole shrimp landed at the dock. The time of holding and the holding system affect the weight of landed shrimp (Table 1) and, therefore, the yield of the final product. Whole shrimp held in MRSW gained 5% in gross weight during the first 1.5 days and slowly gained an additional 2% during the next 7 days. A much Table l. — Change in gross weight and percentage solids with time of holding 2,100 g of whole pink shrimp in modified refrig- erated seawater (MRSW) and ice and pH of the brine. MRSW system Gross Ice system Holding Gross time weight Solids weight Solids (days) pH (g) (%) (g) (%) 0.5 6.85 2,173 22.1 2,215 20.7 1.5 6.50 2,198 18.8 2.333 18.6 2.5 6.40 2,191 19.0 2,333 17.6 3,5 640 2,165 18.8 2,365 18.7 4.5 6.10 2,214 18.6 2,330 17.5 5.5 6.30 2,214 18.0 2,323 17.2 6.5 6.40 2,212 18.3 2,355 16.8 7.5 6.35 2,226 18.0 2.315 16.9 8,5 6.30 2.250 18.3 2,331 17.1 9.5 6.50 2,200 17.8 2,254 16.0 10.5 6 70 2,177 17.6 2,263 16.3 11.5 6.67 2,245 17.7 2,239 14.5 12.5 6.57 2,221 17.7 74 BULLARD and COLLINS: PHYSICAL AND CHEMICAL CHANGES OF PINK SHRIMP higher gain in weight was observed in the ice-held shrimp. The ice-held whole shrimp gained 11% in the first 1.5 days, maintained this weight for 8.5 days, and decreased thereafter. These gross changes in weight were caused by changes in the water, solids, and salt content of whole shrimp with time of holding (Collins 1960, 1961). The pH of the brine was 3.92 at the beginning of the experiment and rose to 6.85 during the first 12 h but varied between 6.1 and 6.9 during the re- mainder of the experiment. The flow of carbon dioxide was regulated at approximately 0.2 1/min but was shut off occasionally to reduce excess loss of carbon dioxode to the environment and buildup of foam. Hand -Peeled, Raw, Pink Shrimp Meats Gross weights of hand-peeled, raw meats in- creased rapidly in both holding systems (Table 2). The salt and sodium content of the raw peeled meats from the MRSW system increased rapidly during the first 2 days to 2% and 0.85%, respec- tively, and remained at this level over the remain- der of the holding period. Potassium decreased during the holding period (MacLeod et al. 1960). In the ice system, however, the meats slowly lost salt, presumably due to the leaching effect of the ice melt. Based on chemical tests, the quality of shrimp held in the MRSW system was considered accept- able through 9.5 days. There was a slight increase in the total volatile acid value at 9.5 days and in the trimethylamine value at 10.5 days, suggesting that quality deteriorated slightly after 8.5 days. In the ice system, quality was acceptable up to 6.5 days, borderline at 7.5 days, and unacceptable thereafter. Because of the large excess of ice used in this holding experiment, the commercial limit for holding shrimp on ice would probably be less than 6.5 days. In this study, it appears that pink shrimp held in the MRSW system can be held for several days longer than in ice. Hand-Peeled, Raw, Washed, Pink Shrimp Meats The solids content (Table 3) for the hand-peeled, raw, washed meats when expressed as percentage composition, was nearly equal from the two hold- ing systems after the effects of salt were removed by subtraction. In both systems there was a rapid decrease in percentage solids (increased moisture) for the first 5 days, but the percentage solids re- mained about equal thereafter. Salt, sodium, and potassium followed the same trend as the raw, peeled meats but at a lower level due to the wash- ing. The data on gross weight and composition (Ta- ble 3) are not useful for direct comparison of recov- ery of meat between the two holding systems be- cause of differences in moisture and salt content. When recalculated on a constant basis (salt-free, 84% moisture), recoveries of raw, washed meats were much higher for the ice system than for the MRSW system (Figure 1). This observation was confirmed when recoveries of protein were com- pared for the two systems (Figure 1). The sharp drop in recovery at 6.5 days for the ice system suggested that soluble proteins were retained through the mild washing technique used in this experiment until spoilage became evident (the 6.5 day break point). In the MRSW system, however, the soluble proteins were leached gradually into the aqueous system. Table 2. — Change in weight and analytical values with time of holding 2,100 g of hand-peeled, raw, pink shrimp meats in modified refrigerated seawater (MRSW) and ice. MRSW system Ice system TVA TVA Holding Gross (meq TVB TMA Gross (meq TVB TMA time weight Solids NaCI Na K H+/ (mg N/ (mg N/ weight Solids NaCI Na K H + / (mg N,' (mg N/ (days) (9) (%) (%) (%) (%) 100 9) 100 g) 100 g) (g) (%) (%) (%) (%) 100 g) 100 g) 100 g) 0.5 759 19.6 1.5 0.66 0.16 0.04 10.2 0.1 743 19.0 0.6 0.26 0.27 0.10 11.0 0.0 1.5 797 19.0 1.9 0.78 0.10 0.05 48 0.3 787 18.0 0.5 0.26 0.25 0.10 40 0.3 25 790 183 20 0.83 0.09 006 2.8 0.3 815 17.1 0.5 0.25 0.21 008 4.5 0.2 3.5 803 18.0 2.1 0.85 008 0.16 7.2 0.2 819 16.9 06 0.26 0.21 0.06 8.8 0.2 4.5 807 17.7 2.1 085 008 0.28 7.0 0.4 827 16.5 0.6 0.25 0.21 0.21 10.8 0.2 5.5 793 17.5 2.1 0.84 0.09 0.26 7.0 0.6 822 16.6 0.6 0.26 0.20 015 10.9 0.3 6.5 812 17.5 2.2 0.85 0.09 0.39 6.6 0.5 830 15.7 0.5 0.24 0.18 0.32 12.4 0.3 7.5 814 17.6 2.2 0.84 0.08 0.31 7.2 0.5 837 15.7 0.5 0.23 0.18 0.46 12.8 1.1 85 822 17.5 2.2 0.90 0.09 0.21 7.0 0.8 853 15.9 0.5 0.24 018 051 18.5 3.2 9.5 812 17.3 22 090 0.09 0.43 6.8 0.8 855 15.2 04 0.20 0.15 0.58 18.9 5.1 10.5 805 17.1 2,2 0.90 009 0.40 7.3 1.1 850 15.2 0.5 0.21 0.16 0.87 26.2 11.4 11.5 836 16.9 2.2 091 0.09 050 9.1 1.1 848 13.0 0.2 0.11 0.07 0.50 15.8 6.5 12.5 827 16.3 2.1 0.85 0.08 0.46 7.5 1.1 75 nSHERY BULLETIN: VOL. 76, NO. 1 Table 3. — Change in weight and analytical values with time of holding 2,100 g of hand-peeled, raw, washed, pink shrimp meats in modified refrigerated seawater (MRSW) and ice. MRSW system Ice system Gross Gross Holding weight Solids Protein NaCI Na K weight Solids Protein NaCI Na K time (days) (g) (%) (%) (%) (%) (%) (g) (%) (%) (%) (%) (%) 0.5 749 17.8 15.4 1.2 0.48 0.14 761 16.4 15.2 0.5 0.21 0.20 1.5 827 17.6 13.7 1.7 0.65 0.08 820 159 14.7 0.5 0.21 0.19 2.5 816 166 13.6 1.8 0.69 007 844 15.3 14.1 0.4 0.21 0.17 3.5 783 165 13,5 1,8 0.70 0.07 862 15.1 13.8 0.5 022 0.16 4.5 799 16.1 13.1 1.9 0.71 0.07 859 14.9 13.8 0.5 0,22 0.16 5.5 809 15.8 12.8 1.9 0.69 0.06 859 15.0 13.9 0.5 0,25 0.17 6.5 812 16 1 12.9 1,9 0.73 0.07 846 14.3 13.1 0.4 0.21 0.14 7.5 801 16.4 13.3 1.9 0.69 0.07 856 14.3 13.3 0.4 0.20 0.14 8.5 828 16.2 13.0 1 9 0.70 0.07 866 14,4 13,2 0.4 0.20 0.14 9.5 794 16,1 13,1 20 0.71 0.07 864 139 126 0.4 0.17 0.12 10.5 793 159 11.1 2.0 0.72 0,07 850 14.0 12,8 0.4 017 0.13 11.5 807 15.7 12,6 2.0 0.71 0,07 825 12.2 11.2 0.2 009 0.05 12.5 796 16.0 12.8 2.1 0.75 0.07 < LU a UJ I < < 800 7 50- 700 650 600. 2 3 4 5 6 7 TIME OF HOLDING 8 9 days 10 11 12 Figure l. — Recovery of hand-peeled, raw, washed pink shrimp meats with time of holding 2,100 g of shrimp in modified refrig- erated seawater (MRSW) or ice, expressed on a salt-free, 84% moisture betsis and protein. Commercial shrimp peelers exert a strong mechanical and washing action on the shrimp, which leaches out soluble proteins. In part there- fore, the final yield would be a function of the gross weight of the landed whole shrimp and of the amount of soluble protein present, which would be influenced by the time and extent of action by bacteria or enzymes. Because the MRSW system reportedly extends holding time, ex-vessel shrimp — processed at an equal stage of quality (say, 4-day ice and 8-day MRSW) and at an equal water content — should give equal yields. In actual practice, however, machine peeling efficiency tends to be the controlling factor for yield. For example, if shrimp were to peel too easily on the machines, yields would be low because the meats would be rubbed off on the rollers. Yields would also be low if the shrimp were too difficult to peel because some unpeeled shrimp would be discarded at the inspection belt. Nelson and Barnett^ ob- tained a 19% raw meat yield from pink shrimp held in the MRSW system and processed through a Laitram (Model A) machine peeler. Hand-Peeled, Washed, Cooked, Pink Shrimp Meats The gross weights for hand-peeled, washed, cooked meats obtained from the 2,100 g of whole shrimp were considerably higher from the ice-held shrimp than from the MRSW-held shrimp (Table 4). Under commercial processing conditions, infill weights must be adjusted to compensate for the high moisture content which would otherwise cause low drained weights after retorting or freez- ing. Consequently, to equalize the variable water content between holding systems and samples, we calculated the weight on a constant basis (salt- free, 75% moisture) and found that the two holding systems gave nearly identical recoveries except for low recoveries during the first several days in ^Nelson, R. W., and H. J. Bamett. Improved shrimp quality by the use of RSW modified with CO2 gas. Unpubl. manuscr. Northwest and Alaska Fisheries Center Utilization Research Division, NMFS, NOAA, 2725 Montlake Boulevard East, Seat- tle, WA 98112. 76 BULLARD and COLLINS: PHYSICAL AND CHEMICAL CHANGES OF PINK SHRIMP Table 4. — Change in weight and analytical values with time of holding 2,100 g of hand-peeled, washed, cooked, pink shrimp meats in modified refrigerated seawater (MRSW) and ice. MRSW system ice system Holding Gross Solids Protein Carotenoid NaCI Na K Gross Solids Protein Carotenoid NaCI Na K time (days) weight (g) (%) (%) index (%) (%) (%) weight (g) (%) (%) index (%) (%) (%) 0.5 413 248 223 0 065 0.7 0.35 009 353 242 226 0047 0.3 0.16 0.14 1.5 389 256 226 0,077 049 006 395 233 21.5 0 062 03 0.17 0 14 2.5 352 27 6 24 1 0 089 049 005 390 226 205 0058 03 0.17 0 13 3.5 284 28 1 246 0 084 0.46 004 373 224 209 0067 03 0.17 0.12 4.5 309 286 246 0089 0.46 004 399 21 6 20 1 0064 03 0.16 0.11 5.5 286 296 258 0091 046 004 367 23.1 21 3 0 068 03 0.17 0 12 6.5 298 302 261 0 086 043 004 374 22 0 20.3 0068 03 016 0.13 7.5 280 29 9 269 0 092 042 004 383 21 9 200 0.074 03 015 0 10 8.5 289 29 9 256 0 098 043 004 398 220 202 0074 03 0 16 0.11 9.5 291 29.8 25.7 0 096 0.46 004 396 21.8 20.0 0.075 0.2 0 16 0.08 10.5 275 30 5 266 0 093 044 004 387 21 5 19.9 0.077 0.2 0.14 009 11.5 258 30.0 26.4 0.090 0.44 0.04 348 21.8 19.9 0.1 0.13 0.09 12.5 262 29.9 26.1 0.094 0.44 0.04 ice caused by poor peeling characteristics. These adjusted weights and the protein data (Figure 2) showed a rapid decrease from both holding sys- tems to 4.5 days, a leveling off to 10.5 days, and another decrease at 11.5 days. Under commercial fishing and processing condi- tions, payment for landed shrimp is based on weight, and weight depends upon time of holding and system used. In our equipment, ice-held shrimp gained more weight and gave a greater recovery of cooked meats than shrimp in the MRSW system. Based on the weight of whole shrimp (Table 1) and the weight of cooked meats (Table 4), therefore, MRSW-held shrimp gave much lower yields than ice-held shrimp, aver- aging 13.9 and 16.4%, respectively. This differ- ence in yield between systems would be reduced when the processor adjusts the weight of infill for a proper cut-out weight. Overall, the only difference in yield between systems is that caused by changes in water and salt content in the whole shrimp, i.e., landed weight. Under production conditions, MRSW has a slight advantage over ice because whole shrimp gain less in MRSW than in ice. It is believed that the laboratory data on the MRSW system would be representative of an MRSW hold- ing system on a boat, but icing techniques may vary considerably from laboratory to boat, and the results obtained in the laboratory may differ from those in commercial practice. Sodium chloride, sodium, and potassium fol- lowed the same general trends as the previous subsamples. The lower levels ( 1.1% NaCI, MRSW; 0.3% NaCI, Ice) were caused by cooking. The carotenoid index, previously used to indi- cate comparative quality between production var- iables (Collins and Kelley 1969), showed an in- crease with increase in time of holding shrimp in < UJ s :^ O O z UJ t— o 400 350 300 250 MRSW 90 80 70 60 3456789 10 11 TIME OF HOLDING, days 12 13 Figure 2. — Recovery of hand-peeled, cooked pink shrimp meats with time of holding 2,100 g of shrimp in modified refrigerated seawater (MRSW) or ice, expressed on a salt-free, 75% moisture basis and protein. both systems. The index, expressed on a dry basis, unexpectedly increased rather than decreased with holding time. We suggest that the peeling- washing technique used in this experiment was less severe than that used during commercial machine peeling and that the 26% loss of protein in cooked meats over the holding period caused a pseudoincrease in the carotenoid content. In 77 agreement with Nelson and Barnett (1971), the color of shrimp held in MRSW was much better than that for shrimp held in ice. LITERATURE CITED Barnett, H. J., R. W. Nelson. P. J. Hunter, S. Bauer, and H. Groninger. 1971. Studies on the use of carbon dioxide dissolved in refrigerated brine for the preservation of whole fish. Fish. Bull, U.S. 69:433-442. Collins. J. 1960. Processing and quality studies of shrimp held in refrigerated sea water and ice. Part 4 — Interchange of the components in the shrimp-refrigerated-sea- water sys- tem. Commer. Fish. Rev. 22(7):9-14. 1961. Processing and quality studies of shrimp held in refrigerated sea water and ice. Part 5 — Interchange of components in a shrimp-ice system. Commer. Fish. Rev. 23(7):l-3. Collins, J., and C. Kelley. 1969. Alaska pink shrimp, Pandalus borealis: Effects of heat treatment on color and machine peelability. U.S. Fish Wildl. Serv., Fish. Ind. Res. 5:181-189. FISHERY BULLETIN: VOL. 76, NO. 1 DYER, W. J. 1945. Amines in fish muscle. I. Colorimetric determina- tion of trimethylamine as the picrate salt. J. Fish Res. Board Can. 6:351-358. Friedemann, T. E., and T. Brook. 1938. The identification and quantitative determination of volatile alcohols and acids. J. Biol. Chem. 123:161- 184. HORWITZ, W. (editor). 1975. Official methods of analysis of the Association of Official Analytical Chemists. 12th ed. Assoc. Off. Anal. Chem., 1094 p. Kelley, C. E., and A. W. Harmon. 1972. Method of determining carotenoid content of Alaska pink shrimp and representative values for several shrimp products. Fish. Bull., U.S. 70:111-113. MACLEOD, R. A., R. E. E. JONAS, AND J. R. McBRIDE. 1960. Sodium ion, potassium ion, and weight changes in fish held in refrigerated sea water and other solutions. J. Agric. Food Chem. 8:132-136. Nelson, R. W., and H. J. Barnett. 1971. Fish preservation in refrigerated sea water modified with carbon dioxide. Proc. 13th Int. Congr. Refrig. 3:57-64. Stansby, M. E., r. W. Harrison, J. Dassow, and M. Sater. 1944. Determining volatile bases in fish. Ind. Eng. Chem., Anal. Educ. 16:593-596. 78 TAXONOMY AND DISTRIBUTION OF ROULEINA ATTRITA AND ROULEINA MADERENSIS (PISCES: ALEPOCEPHALIDAE)i Douglas F. Markle^ ABSTRACT Three Atlantic species o{ Xenodermichthys and Rouleina are recognized: X. copei, R. attrita, and R. maderensis. Bathytroctes mollis and B. aequatoris are considered junior synonyms of R. attrita. Anomalopterus megalops is considered incerta sedis. Diagnostic characters fori?, attrita are: no photophores, convoluted testes, 43-48 lateral line scales, 43-46 preural vertebrae, papillae on body near lateral line, and maturation at a size around 250-300 mm standard length. Diagnostic characters for/?, maderensis are: photophores present, lobate testes, 50-56 lateral line scales, 47-50 preural vertebrae, papillae usually peripheral to photophores on fins and fin bases, and maturation at a size around 200-250 mm standard length. The two species are sharply segregated by depth: 91% of alii?, maderensis were from bottom trawls made between 595 and 1,200 m while 88% ofaUR. attrita were from bottom trawls fished between 1,400 and 2,100 m. The Alepocephalidae are moderate to large deep- sea salmoniform fishes, most commonly encoun- tered below 1,000 m. In terms of biomass and species diversity, the family is one of the most important in the deep sea. Recent exploratory trawling has discovered commercial concentra- tions of alepocephalids west of the British Isles (Anonymous 1974) and in the northwestern At- lantic (Savvatimskii 1969). Off northwestern Af- rica, Golovan (1974) found about 20 species of alepocephalids and labeled the zone below about 1,000 m as "the kingdom of fishes of the family Alepocephalidae." As might be expected in a di- verse group of deep-sea fishes, there are still many problems with identification and nomenclature. One group of naked alepocephalids, those with approximately equal and opposite dorsal and anal fins, has been the subject of numerous descriptions and much confusion. Roule (1915) recognized two genera, Rouleina {=Aleposomus of Roule) and Xenodermichthys, the latter distinguished by a greater number (more than 25) of dorsal and anal fin rays. The two known species of Xenodermichthys, X. nodulosus and X. copei, have caused few taxonomic problems and are easily diagnosed. Both have photophores arranged approximately 'Contribution No. 825 from the Virginia Institute of Marine Science. ^Virginia Institute of Marine Science, Gloucester Point, Va.; present address: Huntsman Marine Laboratory, St. Andrews, N.B. EOG 2X0, Canada. Manuscript accepted April 1977. FISHERY BULLETIN; VOL. 76, NO. 1, 1978. in quincunx on the body and fin bases, two pyloric caeca, and no lateral line scales in adults. Xeno- dermichthys copei has 27-31 dorsal and 26-30 anal fin rays, 46-50 vertebrae, and an unrestricted gill opening; X. nodulosus has 32-33 dorsal and anal fin rays, 50 vertebrae, and a dorsally restricted gill opening which begins at the upper base of the pectoral (Markle 1976). The nomenclature of the Atlantic species, X. copei, has been confused be- cause the oldest of the three available names, Aleposomus copei Gill 1884, was originally de- scribed as: "an Alepocephalid, with the body as well as heads caleless (sic), which I shall describe as Aleposomus copei.'' Grey (1959) and Krefft (1973) have considered A. copei Gill 1884 a nomen nudum, but Gill's ( 1884) sentence clearly refers to an alepocephalid with a naked head and body, and in 1884 that was a sufficient amount of informa- tion to clearly distinguish it from all known alepo- cephalids, with the possible exception of X. nodulosus. In any case the inadequate statement satisfies Articles 11 and 12 of the International Code of Zoological Nomenclature and the name has been used frequently since 1884. Gill's holotype (USNM 33551) was subsequently de- scribed and figured by Goode and Bean (1895). The taxonomy of Rouleina is more confused, in part because there are 15 nominal species, many based upon damaged or poorly preserved speci- mens. All known species of Rouleina can be dis- tinguished from Xenodermichthys by having less than 25 anal fin rays, more than two pyloric caeca, 79 FISHERY BULLETIN: VOL. 76. NO. 1 and modified ringlike lateral line scales in the adults. Photophores are present or absent: their loss appears secondary. For example, in R. fune- bris the size and arrangement of photophores are identical to Xenodermichthys: in R. maderensis the photophores are smaller; in R. harperi only dark spots remain; and in R. attrita there are no photophores. The purpose of this paper is to dis- cuss the taxonomy and distribution of the two known Atlantic species, R. attrita and R. maderensis. METHODS Standard taxonomic measurements and counts were made (Hubbs and Lagler 1958) with the fol- lowing clarifications and additions. Caudal ver- tebrae were distinguished from precaudal verte- brae by the presence of a haemal arch and spine in the former. On radiographs there is a sharp de- marcation, characterized by a reduction in the length of the pleural rib on the last precaudal vertebra and/or the apparent intersection of the last pleural rib with the first haemal spine. The last caudal vertebra counted is that which articu- lates with the parahypural, even if fused to a ural centrum. The one or more ural centra are variable in alepocephalids and were not counted. The high water content and postpreservation shrinkage plus the damage inflicted on most alepocephalids during capture, causes a notice- able amount of variation in most measurements of a species or even in repeated measurements of an individual. The precision of alepocephalid morphometries is therefore relatively low. In addi- tion, most alepocephalid morphometries exhibit definite allometry (Parr 1949, 1956, 1960). Before the allometry of morphometries will be useful in identifying larvae and small juveniles, more smal- ler and less damaged specimens than are pre- sently available will be needed. MATERIAL The following type-material of Rouleina was examined from the U.S. National Museum of Natural History, Washington, D.C. (USNM); Museum National d'Histoire Naturelle, Paris (MNHN); Zoological Museum, University of Copenhagen (ZMUC); Zoological Museum, Berlin (ZMB); and Museu Municipal do Funchal, Madeira (MMF): Bathytroctes attrita, MNHN 85-166 and 85-169; B. mollis, MNHN B-2219; B. aequatoris, USNM 44085; B. harperi, USNM 92333; B. welshi, USNM 92332; Xenodermichthys funebris, USNM 99b3A,Anomalopterus megalops, USNM 170957; Aleposomus nudus, ZMB 17426; A. lividus, ZMB 22398; R. danae, ZMUC P1778; andR. maderensis, MMF 50, 2395, and 2396. Additional material was examined from the British Museum (Natural History), London (BMNH); University Museum, Tokyo (UMT); In- stitute of Oceanographic Sciences, Wormley, En- gland (lOS); Museum of Comparative Zoology, Harvard (MCZ); Field Museum of Natural His- tory, Chicago (FMNH); Rosenstiel School of Ma- rine and Atmospheric Sciences, Miami (UMML); Institut fiir Seefischerei, Hamburg (ISH); and Virginia Institute Marine Science, Gloucester Point (VIMS). These collections included four specimens of R. guentheri cataloged as BMNH 1898.7.13.19 and UMT 5785, 5785', and 20983; one specimen of R. danae, USNM 215490; 69 specimens of R. attrita, USNM 215479-215489 and 44085; ISH 123/73, 124/73, 950/73, 141/74, 163/74, 511/74, 512/74, 835/74, 844/74, 212/75, 234/75, and one uncatalogued; VIMS 3539, 3540, 3542, and 3543; FMNH 65711; UMML 22353; MCZ 40609; and lOS Discovery 8512#1; and 35 specimens of R. maderensis, USNM 215471- 215478; ISH 130/75; VIMS 3541; MCZ 39349; BMNH 1945.7.20.5; lOS Discovery 7431, 7432, and 7436; and ZMUC Dana 1183^ RESULTS The species oi Rouleina separate conveniently into two groups. The first group, which lacks photophores or their remnants, contains i?. attrita and/?, danae. Rouleina danae differs from/?, at- trita by its reduced maxillary dentition and much larger orbit (43.5% of head length (HL) vs. 24-29% HL at about 100 mm standard length (SD). The second gi'oup, which has photophores, contains/?. maderensis and several Indo-Pacific species which differ from it in having fewer anal fin rays (16-19 vs. 20-22). Although the two North Atlantic species,/?, at- trita and /?. maderensis, are easily distinguished with undamaged material, most specimens are damaged and the two species are very similar in gross morphology. The following key summarizes characters which have been found useful to sepa- rate these species. 80 MARKLE: TAXONOMY AND DISTRIBUTION OF ROULEINA Key to North Atlantic Species of Rouleina la. No photophores: testes ribbonlike with many convolutions in mature speci- mens but folds always connected, never with separate lobes (Figure 1); lateral line with 43-48 modified ringlike scales, undetectable in specimens less than 155 mm SL; preural vertebrae 19-22 (pre- caudal) + 22-26 (caudal) = 43-46 (total); papillae on body especially near lateral line, along bases of vertical fins, and along all fin rays; mature around 250- 300 mm SL R. attrita (Vaillant 1888) lb. Flat superficial photophores present, commonly abraded; testes discrete, separate lobes even when immature (Figure 1); lateral line with 50-56 mod- ified ringlike scales, undetectable at 131 mm SL; preural vertebrae 20-22 (precaudal) + 26-28 (caudal) = 47-50 (total); papillae restricted to fins and fin bases, usually peripheral to photo- phores which are more numerous below lateral line; mature around 200-250 mm SL R. maderensis Maul 1948 Rouleina attrita (Vaillant 1888) Figure 2 A Bathytroctes attritus Vaillant 1888:158, fig. 2 (holotype, MNHN 85-166 only; lat. 37°35'N, long. 29°26'W, 1,442 m; paratype, MNHN 85- 169, is Bellocia koefoedi). Bathytroctes mollis Koehler 1896:517, pi. 26, fig. 2 (holotype, MNHN B-2219, Bay of Biscay, 1,700 m). Bathytroctes aequatoris Goode and Bean 1896:44, fig. 50 (holotype, USNM 44085, lat. 01°03'N, long. 80°15'W, 1,355 m). Nomenclature Quero (1974) suggested that R. attrita be treat- ed as a nomen dubium since Vaillant (1888:158), using a 55-mm shred of skin from the caudal peduncle, had estimated 40-50 scale rows on the body and since Vaillant's dorsal and anal fin ray counts are wrong for Rouleina. The source of the problem is the nature of the skin of Rouleina and the fact that the remaining type-material repre- sents two different genera (Vaillant originally listed four specimens, but two could not be located V-.. ...-^^T- f^ ■'■■ mm Jm-A' V^.^:^A* •^.--^ B Figure l. — a. Rouleina maderensis, USNM 215476, about 275 mm SL, testes, showing completely separated lobes (arrow). B. Rouleina attrita, USNM 215483, 369 mm SL, testes, showing convolutions without the formation of separate lobes (arrow). 81 FISHERY BULLETIN. VOL. 76, NO. 1 Figure 2. — A. Rouleina attrita, redrawn from Koefoed (1927, plate 3, fig. 5). B. Rouleina maderensis, redrawn from Maul (1948, fig. 1), with photophore distribution based upon USNM 215478, 131 mm SL. in MNHN). Fortunately, Vaillant clearly indi- cated that the description of each species is based on a unique individual chosen from the collection (Bauchot et al. 1971). On the bottom of page 159, following a list of measurements of a 250-mm specimen, Vaillant ( 1888) made the notation "No. 85-166, Coll. Mus.," a clear designation of a holo- type. This specimen is now in very poor condition but a piece of skin clearly shows the typical ring- like lateral line scales (Figure 3) and indications of fluid-filled dermal compartments typical of^ Rou- leina. The latter could be mistaken for scale poc- PORE Figure 3. — Rouleina attrita, schematic of lateral line scale and subsequent pore from the midbody region. SCALE LATERAL LINE CANAL 82 MARKLE: TAXONOMY AND DISTRIBUTION OF ROULEINA kets and are very similar to the dermal compart- ments in Xenodermichthys as illustrated and described by Best and Bone (1976). Vaillant ( 1888, pi. 12, fig. 2) illustrated otoliths and gave a vertebral count (Vaillant 1888,159) "II y a 20 vertebres dorsales et 25 caudales." A radio- graph of the contents of the jar containing MNHN 85-166 showed that the otoliths were intact and there were 20 + 24 vertebrae. It is likely therefore that both observations came from the missing paratypes. A comparison of the illustrated otoliths with recently collected material o{ Alepocephalus agassizii, Xenodermichthys copei, Bathytroctes microlepis, Narcetes stomias. and Ron leina attrita shows they were undoubtedly taken from a Rouleina. Haedrich and Polloni (1974) found un- stated "significant differences" between their Rouleina otoliths and Vaillant's, but their descrip- tion and my examination of their specimens (ISH 950/73) shows them to heR. attrita. Therefore, the vertebral counts, lateral line scales, Vaillant's es- timate of number of (lateral line) scales, and oto- liths indicate that the holotype and probably the missing paratypes agree with recently collected material of R. attrita. The remaining paratype, MNHN 85-169 (lat. 15°48'N, long. 20°23' W, 3,655 m), is a specimen of Bellocia koefoedi Parr 1951. This identification is based on examination of the type series of B. koefoedi in the Zoological Museum, Bergen, and the presence of the following diagnostic characters in MNHN 85-169: palatine teeth present, gill rak- ers 4-1-14 on first arch, body scaled, dorsal in- serted in advance of anal, and a radiograph shows 22 + 18 = 40 vertebrae, 1 1 anal fin rays, and about 16 dorsal fin rays. The radiograph also shows oto- liths in the skull and a standard length of no more than 220 mm (Quero 1974 stated about 230 mm). The length, intact otoliths, and vertebral count indicate that Vaillant (1888) was not basing his description of /?. attrita on MNHN 85-169. How- ever, since its condition is somewhat better than the holotype, Vaillant's reference to scale rows and a minimum of 1 1 anal fin rays may have been based on comparison with this specimen. Description Accurate descriptions and illustrations can be found in Goode and Bean (1895, as B. aequatoris), Koehler (1896, as B. mollis), Koefoed (1927, as Talismania mollis). Grey (1959), Haedrich and Polloni ( 1974), and Pakhorukov ( 1976). Important diagnostic meristic characters are in Table 1. In addition, the present material showed the follow- ing meristic variation (number of specimens in parentheses): Pj6-7 (26), P26-7 without a splint bone (27), gill rakers on first arch [7-8] -I- 1 + [15-20] = [23-28] (23), branchiostegal rays 6 (5), and pyloric caeca 7-11 ( 16). Teeth are present only on the dentary, premaxillary, maxillary, third and fourth infrapharyngobranchials, fourth epi- branchial, and fifth ceratobranchial. Twenty-six specimens of/?, attrita, 57.1-378 mm SL, showed much morphometric variation and no noticeable differences with 19 R. maderensis, 86.7-323 mm SL. In both species smaller speci- mens have relatively shorter caudal peduncles. In addition, smaller specimens of R. attrita (<155 mm SL) lacked lateral line scales and the papillae on the body were relatively longer and more noticeable than in larger specimens. In one well-preserved large specimen, 347 mm SL (USNM 215481), the branchiostegal mem- branes, gill cavity, orbit, and bases of fins are bluish. The rest of the body is covered by thin black skin, under which is a network of longitudinally aligned, fluid-filled, oblong dermal compartments (Best and Bone 1976). The lateral line, which ex- tends onto the caudal fin, is a tube supported by Table 1. — Selected counts of Rouleina attrita and R. maderensis (superscript prefix indicates type material of: A — Bathytroctes attritus. B— 7?. maderensis, C — B. aequatoris, and D — B. mollis). Lateral line pores Species 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 R attrita R maderensis 114 2 1 2 1 1 1 Bi 2 Bi Precaudal vertebrae 19 20 21 22 22 23 Caudal vertebrae 24 25 26 27 28 R attrita R. maderensis 4 A,C.D22 7 2 2 Bi7 17 Dorsal fin rays 18 19 20 21 1 22 Dl A,Ci3 18 15 4 Bl9 Bi5 3 Anal fin rays 19 20 21 22 R. attrita R maderensis 5 Dio Cio 1 3 Bg B5 D7 10 Cg 1 5 Be 4 Total vertebrae 44 45 46 47 48 49 50 Dl A,Ci3 15 7 Be B22 8 1 83 FISHERY BULLETIN: VOL. 76, NO. 1 modified ringlike scales with pores usually situated midway between and not touching the scales (Figure 3). The skin along the dorsal mid- line, above the supracarinalis muscle, is typically split open, exposing dense fat deposits and mucus. Ventrally, the skin overlying the lower hypaxial muscles is also split open. In addition, the area ventral to the heart, between the cleithra, con- tains a mucus-filled network of connective tissue. Testes are thin ribbonlike structures in imma- ture males and become thick and convoluted in mature specimens. The convolutions, however, never become separate lobes (Figure 1). The ovaries, back to about the level of the pelvics, are completely enclosed by ovarian tunic medially and the body wall laterally. Posteriad the lateral ovar- ian surface is exposed. The ovary contains few eggs up to 3.2 mm in diameter. Rouleina maderensis Maul 1948 Figure 2B Rouleina maderensis Maul 1948:7, fig. 1 (holotype, MMF 2398, Madeira, 600-1,600 m depth range for type series). As a supplement to Maul's (1948) description. Table 1 summarizes important diagnostic meristic characters. In addition, the present material showed the following meristic variation (number of specimens in parentheses): Pj5-7 (13), P25-6 without a splint bone (13), gill rakers on first arch [6-8] + 1 + [15-21] = [22-30] (8), branchiostegal rays 6 (12), and pyloric caeca 10-11 (7). Dentition similar to R. attrita. Lateral line scales were absent in the two specimens <131 mm SL but were present in a 177-mm SL specimen. Photophores were present on the smallest specimen, 86.7 mm SL. Generally, photophores are more difficult to find in larger specimens. Black papillae are distributed along the base of the caudal, on primary caudal rays, dorsal and anal rays, on the supratemporal, and from the interorbital area to the snout. An irregularly ar- ranged row of papillae lies between the lateral line and dorsal profile. Small flat photophores are mostly located below the lateral line; a paratype (MMF 50) has nine photophores along the anal fin, two on the base of the lower caudal and one or two on the upper caudal base; body photophores are arranged approximately in quincunx. The super- ficial layer of black skin covers longitudinally aligned, fluid-filled, oblong, dermal compartments and is frequently split along the midline as in /?. attrita. The modified ringlike lateral line scales have a relatively broad and long posterior tab. Lateral line pores are usually at the end of the scale tab of the preceding lateral line scale, ap- proximately midway between scales but touching the anterior scale. Testes, even when immature, are always lobed (Figure 1). The ovary is similar to that in R. at- trita. Eggs are large, up to 3.7 mm. Incerta sedis Anomalopterus megalops Beebe 1933 An examination of Beebe's damaged and con- torted holotype (USNM 170957), now about 25 mm SL, indicates that it might be a Rouleina. The dorsal and anal origins appear approximately op- posite in contrast to Beebe's ( 1933) statement that the anal origin was under the middle of the dorsal. The "numerous small tubercles" which Beebe found abundant on the head and less so on the body are no longer visible. Beebe's (1933) description, the best source for deciphering the identity of the specimen, agrees with Rouleina, especially R. maderensis. However, the seven branchiostegal rays and anal fin extending well posteriad of the end of the dorsal fin are characters which are un- known in the available North Atlantic Rouleina. Identification of this specimen should be post- poned until more larval and juvenile material are available. ECOLOGY Direct sighting of two R. attrita <1 m from the bottom at 1,800 m off Virginia was made during DSRV Aluin dive 575, 4 June 1975. The moderate-sized individuals had a more rounded head than the more commonly sighted alepoceph- alid, Alepocephalus agassizii. The dorsal and ven- tral profiles of the snout and lower jaw regions are approximately equal arcs in /?. attrita (Figure 2A), while in A. agassizii the ventral profile of the lower jaw is straighter. The skin of/?, attrita also appears smoother since it is mostly scaleless, but both are about equally black in situ. An unexpected observation was that the two/?. attrita had shredded sheets of mucus hanging from their jaws and body. The two individuals drifted 84 MARKLE: TAXONOMY AND DISTRIBUTION OF ROULEINA motionless by the observation port, one head down, the other more or less on its side. Alepo- cephalus agassizii was observed in similar motion- less positions and were seen to move when disturbed, so that the motionless positions are probably not a sign of death. The observation of mucus is, as yet, uncorroborated by others. How- ever, Koehler (1896:518) described the fresh con- dition of the holotype of B. mollis as being flaccid as a holothurian and retrieved from the trawl in a thick mucus. The split skin along the dorsal and ventral midline commonly observed in preserved specimens of Rouleina may be related to fat and mucus concentrations in these regions of the body. The function of these concentrations and the mucus sheets is unknown. All of the R. attrita and most of the i?. maderen- sis were from bottom trawls, but two of the smaller R. maderensis, 86.7 and 177 mm SL, were from nonclosing midwater trawls. It is possible that the rather amorphous and almost degenerate photo- phores (based on microsections from a 236-mm SL specimen) of demersal adult R. maderensis repre- sent organs which are functional only in meso- pelagic juveniles. DISTRIBUTION Both species are known from the southeastern Pacific and North Atlantic, while i?. attrita is also known from the South Atlantic and southwestern Indian Ocean (Figure 4). The two species have been caught in the same net once in the western Atlantic and once in the southeast Pacific. Al- though the geographic distributions are similar, R. attrita andi?. maderensis segregate sharply by depth. Thirty of 33 specimens (91%) of/?, mad- erensis were from bottom trawls fished between 595 and 1,200 m. In contrast, 66 of 75 specimens (88% ) ofi?. attrita were from bottom trawls fished between 1,400 and 2,100 m. Off the east coast of the United States, the most consistent physical characteristic between 1,200 and 1,400 m is the 4°C isotherm (VIMS unpubl. data, Churgin and Halminski 1974a). However, in the Gulf of Mexico (Churgin and Halminski 1974b) and eastern North Atlantic (Lenz 1975), the 4°C isotherm is considerably deeper. A charac- teristic feature of the demersal ichthyofauna on the continental slope off Virginia is a sharp in- crease in mean weight of individual fish around 1,500 m (Markle 1976; C. A. Wenner and J. A. Musick pers. commun.). Consistent with this phenomenon is the observation of generally larger body size in the deeper dwelling R. attrita com- pared with its shoaler dwelling congener, R. maderensis. Although this suggests a possible bio- logical factor in their distribution, a lack of ap- propriate ecological data for most of the available collections precludes such a statement. Without comprehensive ecological information for all col- lections, the mechanism of bathymetric segrega- tion in the two Atlantic species of Rouleina re- mains unknown. ACKNOWLEDGMENTS I am grateful to the following individuals and institutions for loan of material: D. M. Cohen, National Marine Fisheries Service, Systematics Laboratory; R. H. Gibbs, Jr., S. H. Weitzman, and S. Karnella, USNM; M. L. Bauchot, MNHN; A. Wheeler, BMNH; G. Krefft, ISH; R. K. Johnson, FMNH; C. R. Robins, UMML; K. Liem and R. Schoknecht, MCZ; N. R. Merrett, lOS; T. Abe, UMT; J. Nielsen and E. Bertelsen, ZMUC; G. E. Maul, MMF; J. A. Musick, VIMS; and C. Karrer, ZMB. Travel expenses were partly defrayed by the 1976 Raney Award from the American Society of Ichthyologists and Herpetologists and a Grant- in-Aid of Research from Sigma Xi. This work was supported in part by NSF grant No. GA-37561 and OCE 73-06539, J. A. Musick principal inves- tigator. LITERATURE CITED ANONYMOUS. 1974. Deep water trawl fish tested for food market. Fish. News Int. 13(l):47-49. Bauchot, M.-L., T. Iwamoto, p. Geistdoerfer, and M. Rannou. 1971. Etude critique des resultats des expeditions scien- tifiques du "Travailleur" et du "Talisman". Nouvel exa- men des Macrouridae (Teleosteens Gadiformes). Bull. Mus. Natl. Hist. Nat., Paris, 3e Ser., Zool. 14:653-666. BEEBE, W. 1933. Deep-sea isospondylous fishes, two new genera and four new species. Zoologica (N.Y.) 13:159-167. BEST, A. C. G., AND Q. Bone. 1976. On the integument and photophores of the alepo- cephalid fishes Xenodermichthys and Photostylus. J. Mar. Biol. Assoc. U.K. 56:227-236. Churgin, J., and S. J. Halminski. 1974a. Temperature, salinity, oxygen, and phosphate in waters off United States. Vol. 1. Western North Atlan- tic. U.S. Dep. Commer., Natl. Ocean. Atmos. Admin., Environ. Data Serv., Key Oceanogr. Rec. Doc. 2, 166 p. 1974b. Temperature, salinity, oxygen, and phosphate in waters oflF United States. Vol. 2. Gulf of Mexico. U.S. 85 FISHERY BULLETIN; VOL 76. NO. 1 Figure 4.— a. Rouleina attrita, geo- graphic distribution of collections examined plus recent capture in South Atlantic of Pakhorukov (1976). B. Rou- leina maderensis , geographic distribution of collections examined. Larger sjTnbols indicate multiple captures. 86 MARKLE: TAXONOMY AND DISTRIBUTION OF ROVLEINA Dep. Commer., Natl. Ocean. Atmos. Admin., Environ. Data Serv., Key Oceanogr. Rec. Doc. 2. 117 p. GILL, T. N. 1884. Three new families of fishes added to the deep-sea fauna in a year. Am. Nat. 18:433. GOLOVAN, G. G. 1974. Preliminary data on the composition and distribu- tion of the bathyal ichthyofauna (in the Cap Blanc Area). Oceanology. Akad. Nauk. USSR 14:288-290. GOODE, G. B., AND T. H. Bean. 1895. Oceanic ichthyology. U.S. Natl. Mus., Spec. Bull 2, 553 p. Grey, M. 1959. Deep sea fishes from the Gulf of Mexico with the description of a new species, Squalogadus intermedius (Macrouroididae). Fieldiana Zool. 39:323-346. Haedrich, R. L., and p. T. Polloni. 1974. Rarely seen fishes captured in Hudson Submarine Canyon. J. Fish. Res. Board Can. 31:231-234. HUBBS, C. L., AND K. F. LAGLER. 1958. Fishes of the Great Lakes region. Revised ed. Cranbrook Inst. Sci. Bull. 26, 213 p. KOEFOED, E. 1927. Fishes from the sea-bottom. Rep. Sci. Res. "Michael Sars" North Atl. Deep-Sea Exped. 1910 4: 1- 148. KOEHLER, R. 1896. Resultats scientifiques de la campagne du "Caudan" dans le Golfe de Gascogne. Poissons. Ann. Univ. Lyon. 26:475-526. KREFFT, G. 1973. Alepocephalidae. In J. C. Hureau and T. Monod (editors). Check-list of the fishes of the north-eastern Atlantic and of the Mediterranean, p. 86-93. Unesco, Paris. Lenz, W. 1975. Untersuchungen zur inneren hydrographischen struktur des suedlichen and mittleren Atlantiks (0- 2000m Tiefe) mit zoogeographischen Anmerkun- gen. Ber. Dtsch. wiss. Komm. Meeresforsch. 24:1-22. MARKLE, D. F. 1976. Preliminary studies on the systematics of deepsea Alepocephaloidea (Pisces:Salmoniformes). Ph.D. The- sis, Coll. William and Mary, Williamsburg, Va., 225 p. Maul, G. E. 1948. Monografia dos peixes do Museu Municipal do Funchal. Ordem Isospondyli. Bol. Mus. Munic. Funchal 3(5):5-41. PAKHORUKOV, N. p. 1976. (Preliminary list of the bathyal bottom fishes of the Rio Grande Rise.) [In Russ.J In (Biology and distribu- tion of the deep-sea fishes), p. 318-331. Tr. Inst. Okeanol. P. P. Shirshova 101. Parr, A. E. 1949. An approximate formula for stating taxonomically significant proportions of fishes with reference to growth changes. Copeia 1949:47-55. 1951. Preliminary revision of the Alepocephalidae, with the introduction of a new family, Searsidae. Am. Mus. Novit. 1531, 21 p. 1956. On the original variates of taxonomy and their re- gressions upon size in fishes. Bull. Am. Mus. Nat. Hist. 110:369-397. 1960. The fishes of the family Searsidae. Dana Rep. Carlsberg Found. 51, 109 p. QUERO, J.-C. 1974. Rouleina mollis (Koehler, 1896) poissons, Clupei- formes, Alepocephalides en remplacement de Rouleina attrita ( Vaillant, 1888) nomen dubium. Rev. Trav. Inst. Peches Marit. 38:437-438. ROULE, L. 1915. Consideration sur les genres Xenodermichthys Giinth. et Aleposomus Gill dans la famille des Alepo- cephalides. Bull. Mus. Natl. Hist. Nat. Paris (for 1914):42-46. Savvatimskii, p. I. 1969. The grenadier of the North Atlantic. Tr. Polyam. Nauchno-Issled. Proektivnogo Inst. Morsk. Rhybn. Khoz. Okeanogr., p. 3-72. (Transl. 1974, Fish. Res. Board Can., Transl. Ser. 2879, St. Johns, Newfoundland). Vaillant, L. 1888. Poissons. /n Expeditions scientifiques du"Travail- leur" et du "Talisman". Paris, 406 p. 87 FOOD AND FEEDING HABITS OF JUVENILE ATLANTIC TOMCOD, MICROGADUS TOMCOD, FROM HAVERSTRAW BAY, HUDSON RIVER Stephen A. Grabe* ABSTRACT Juvenile Atlantic tomcod from Haverstraw Bay (Hudson River, N.Y.) were found to have a May-June diet of copepods and a July-December diet of amphipods, Neomysis americana, and isopods. This dietary shift occurred when mean length reached 90 mm during July. Growth paralleled feeding intensity: elevated during June, October, and November, and depressed July through September; feeding intensity decreased prior to spawning (December). Feeding and growth were inhibited at temperatures >24°C and dissolved oxygen <7mg/l. The Atlantic tomcod, Microgadus tomcod Wal- baum, is an inshore marine fish whose range ex- tends from southern Labrador (Bigelow and Schroeder 1953) south to Virginia (Massman 1957); freshwater populations are localized in Quebec and Newfoundland (Scott and Grossman 1973). The Hudson River may represent the southern extent of the tomcod's breeding range since it has not been reported from the Delaware River estuary (de Sylva et al.^) and its status in New Jersey waters is uncertain (Miller 1972; Heintzelman^). In the Hudson River tomcod were formerly considered to be a seasonal, migratory species (Curran and Ries 1937; Clark and Smith"*); more recent work, however, suggests that tomcod remain in the estuary for their entire life cycle (Lawler et al.^). Tomcod spawn as young-of-the-year and year- lings (Lawler et al.^) with egg deposition typically occurring during December and January ( Bigelow and Schroeder 1953; Booth 1967). First year growth, while initially rapid, slows in midsummer (Howe 1971) and resumes in early fall (Lawler et 'Lawler, Matusky and Skelly Engineers, Pearl River, N.Y.; present address: 95 Ash Street, Piermont, NY 10968. Me Sylva, D. P., F. A. Kalber, and C. N. Schuster, Jr. 1962. Fishes and ecological conditions in the shore zone of the Dela- ware River estuary, with notes on other species collected in deeper water. Del. Board Fish Game Comm., 164 p. ^Heintzelman, D. S. (editor). 1971. Rare or endangered fish and wildlife of New Jersey. N.J. State Mus. Sci. Notes 4, 23 p. *Clark, J. R., and S. E. Smith. 1969. Migratory fish of the Hudson River. /n G. P. Howells and G. J. Lauer (editors), Hudson River ecology, p. 293-319. N.Y. State Dep. Environ. Conserv. ^Lawler, Matusky and Skelly Engineers. 1975. 1974 Hudson River aquatic ecology studies. Bowline Point and Lovett Generating Stations. Prepared for Orange and Rockland Util- ities, Inc. ^Lawler, Matusky and Skelly Engineers. 1976. Environmen- tal impact assessment-water quaUty analysis: Hudson River. National Comm. on Water Quality. NTIS PB-251099. al. see footnote 5; Texas Instruments^; Dew and Hechts). Young-of-the-year Hudson River tomcod undergo a dietary shift, from calanoid copepods to Gammarus spp. amphipods, as they increase in size (Texas Instruments see footnote 7). My objec- tives were to define the diet and feeding intensity of juvenile tomcod within the vicinity of Haver- straw Bay, Hudson River, N.Y. MATERIALS AND METHODS Stomach contents of 577 juvenile tomcod were analyzed as part of the postoperational biological monitoring program for a fossil fuel steam electric generating station located at Hudson River mile- point 37.5. The study area (Figure 1 ) encompassed Hudson River milepoints 37.5-41.5, as measured from the Manhattan Battery. Tomcod were collected once monthly June- December 1973 and 1974 by a 9.1-m otter trawl with a 64-mm mesh cod end liner, towed against the tide at 1.5-2.0 m/s. Collections of plankton and juvenile fishes were made twice monthly June- August 1974 with a 1-m diameter plankton net of 571-^tm mesh mounted in an epibenthic sled and towed against the tide at 0.9-1.2 m/s. Tomcod from May and December 1975 trawl collections were also analyzed to provide a larger data base for these months. Manuscript accepted June 1977. FISHERY BULLETIN: VOL. 76, NO. 1. 1978. 'Texas Instruments, Inc. 1975. Hudson River ecological study in the area of Indian Point: 1974 annual report (draft). Prep, for Consolidated Edison Co. of N.Y. , Inc. *Dew, C. B., and J. H. Hecht. 1976. Ecology and population dynamics of Atlantic tomcod (Microgadus tomcod) in the Hudson River estuary. In Hudson River ecology. Hudson River Environ. Soc., Inc. 89 FISHERY BULLETIN: VOL. 76, NO. 1 .44 PCCKSKILL STONY POINT Figure l. — Sampling stations, depths, and collection methods for Atlantic tomcod, Haverstraw Bay 1973-75. Numbers along river indicate mile points above the Manhattan Battery. Station 1: 6.7 m; trawl, epibenthic sled. Station 2: 12.2 m; epibenthic sled. Station 3: 7.6 m; trawl. Station 4: 3.0 m; epibenthic sled. Station 5: 18.3 m; trawl, epibenthic sled. Station 6: 16.8 m; epibenthic sled. Station 7: 13.7 m; trawl, epibenthic sled. Bottom temperature (Figure 2), dissolved oxy- gen, and salinity (Table 1) were measured at sta- tion 2 (depth 12.2 m). Fish were preserved in 5% (epibenthic sled col- lections) or 10% (trawl collections) buffered For- malin.^ Total length of each fish was measured to the nearest millimeter. Fish >50 mm were weighed to the nearest 0.1 g; fish <50 mm were weighed to the nearest 0.01 g. Stomachs were re- moved and transferred to a 70% solution of eth- anol prior to analysis. One everted fish stomach, indicative of regurgitation, was excluded. Food organisms were identified, counted, and the entire contents, excluding obvious nonfood items (e.g., pebbles), of 401 stomachs were dried to a constant weight at 103°C. Only postlarval juveniles were studied; the dis- tinction between larval and juvenile tomcod was the completed differentiation of the fins (Balon 1975). Lower limits of adult fin ray counts were taken from Bigelow and Schroeder (1953). Appli- cation of this criterion showed that a total body length of 25 mm represented the lower size limit of juveniles. During 1973, young-of-the-year were distinguished from yearlings by examination of length-frequency histograms of larger sample sizes of tomcod (Lawleretal. see footnote 6; Lawler et al.^°). Fish collected during November and De- cember 1973, 148 and 160 mm, respectively, were considered to represent upper size limits of young-of-the-year. All fish collected during 1974 and December 1975 were condsidered young-of- the-year. Stomach content data were pooled by month and quantitative results for each taxon calculated as percent occurrence, percent composition, and im- portance (Windell 1971): Importance = / (% composition) (% occurrence). Percent relative importance was calculated by summing importance values at the lowest taxonomic level and dividing individual impor- tance values by that sum. A modified similarity index (Windell 1971) was then calculated to com- pare monthly changes in percent relative impor- tance of various food items, at the lowest compara- ble taxonomic level. Consecutive months were compared by selecting the lesser of two relative importance values for each food item and then summing them. This sum is the index of similarity and it may range from 0 to 100%. An index of fullness ilf) (Nikolsky 1963; Windell 1971), indicative of feeding intensity, was calcu- lated for each fish: // stomach content biomass (g* x 10^ weight (g) of fish ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. '"Lawler, Matusky and Skelly Engineers. 1974. 1973 Hudson River aquatic ecology studies: Bowline Point and Lovett Gen- erating Stations. Prepared for Orange and Rockland Utilities, Inc. 90 urxrtDIj. r \J\JLJ c\i^LJ r iJiJL/ii'^ \^r i irvui i o ^-^i- rv i L.i-vi'^ 1 1*^ & v^nivyv/xv Figure 2. — Bottom water tempera- tures at Station 2, Haverstraw Bay (mile point 37.5). • 26 •1973 i-^.-rv^^: 1974 24 1 r»"7P 1 9 ID ^^ . / ' \ V o22 /-''' •''■ 0 ^20 ^ \ Ul /' \~~'~ 5l8 / / ' '\ 1 1- / — < 16 / . oc / \i UJ . . Y ^ Q. 14 / V" 2 \ti\2 cr 10 / I ■ \^ LlJ • \ ►- R I / \ < ° 1 1 A ^ 6 1 \\ 4 1 \\ 2 V \ M M J J MONTHS N Table l. — Mean monthly bottom dissolved oxygen and salinity measurements at Station 2, Haverstraw Bay (mile point 37.5) 1973-75. Dissolved oxygen (mg/l) Salinity (%«) Month 1973 1974 1975 1973 1974 1975 Jan. (') (') 9.0 (') <0.03 2,65 Feb. C) (') 12.5 (') <0.03 292 Mar. (') (') 12.8 (') <0.03 0,97 Apr. (') 9.6 12.1 (') <0.03 1,16 May 7.8 9.3 9.5 C) <0.03 0,93 June 7.1 7.8 6.7 <0.03 1.44 1,48 July 5.9 7.7 5.8 1.32 3.07 3,34 Aug. 5.8 5.5 5.1 D 4.14 4,48 Sept. 6.9 9.1 7.8 2.98 1.72 1,39 Oct. 8.6 8.5 7.8 5.72 2.02 0,87 Nov. 9.2 10.2 8.1 3.51 0.85 0,03 Dec. 12.4 12.2 11.7 ■t).03 0.04 0.30 'Data not available. RESULTS average number of food items per stomach (Table 4), with feeding greatest during May, June, Oc- tober, and November, and lowest during July- September. Feeding also decreased during De- cember. Growth of the 1974 year class paralleled seasonal alterations in If (Figure 3). The trends of the above parameters suggested that seasonally fluctuating environmental vari- ables (e.g., temperature and dissolved oxygen) might be affecting feeding intensity and, there- fore, growth. Statistical tests to discriminate the Ranking dominant food items by importance (Table 2) revealed two distinct dietary regimes: a May-June diet of copepods and a July-December diet of amphipods, mysids, and isopods. The simi- larity index for consecutive months emphasized this shift by a markedly low value (39%) for June-July compared with a range of 54-80% for other months. Pooling June and July fish by 10-mm length intervals indicated that copepod importance de- creased and that of amphipods increased as mean length increased. At 90 mm, transition to an amphipod-dominated diet was complete (Table 3). A seasonal feeding cycle was distinguished by trends in If, percentage of empty stomachs, and Figure 3. — Index of fullness (If) and growth of juvenile Atlantic tomcod, Haverstraw Bay, June-December 1974. 91 Table 2. — Monthly summary of five most important food items of juvenile Atlantic tomcod from Haverstraw Bay, 1973-75. Month Sample size Taxon Percent Percent occur- compo- rence sition May 38 June 210 July Aug. Sept. Oct. Nov. Dec. 69 58 43 43 42 74 Index Copepoda 100.0 99.2 99.6 Eurytemora afftnis Ectocyclops sp. Halicyclops sp. Gammarus daiberi 10.5 0.6 25 Monoculodes edwardsi 2.6 0.1 0.5 Ostracoda 2.6 0.1 0.5 Copepoda 54.8 82.4 67.2 £. affin(s Cyclopoida Harpacticoida Unidentified nauplii G daiben 64.8 6.9 21.2 M edwardsi 37.6 2.7 10.0 Bosmina sp 224 3.0 82 Neomysis americana 19.5 0.9 43 G. daiberi 63.8 38.9 498 N. americana 30.4 19.8 245 M. edwardsi 31.9 18.4 24.2 Cyathura polita 23.2 4.7 10.5 Scolecolepides viridis 11.6 3.0 5.9 M- edwardsi 37.9 45.8 41.7 G daiberi 41.4 18.2 27.4 N. americana 25.9 15.0 19.7 Edotea triloba 22.4 5.2 108 C polita 12.1 3.1 6.1 G. daiben 72.1 53.1 61.9 M. edwardsi 34.9 28.6 31.6 N. americana 20.9 5.4 10.6 C. polita 14.0 2.1 5.4 Chaoborus punctipennis 11.6 1.8 46 G. daiben 93.0 70.9 81.2 M. edwardsi 34.9 20.2 26.5 C polita 25.6 2.2 7.6 Rhithropanopeus harrisii 14.0 1.5 4.6 Corophium lacustre 23 06 1.2 G daiberi 73,8 868 80.0 Crangon septemspinosa 40.5 7.1 16.9 N. americana 11-9 3.3 6.3 R. harrisii 16.7 1.4 48 M edwardsi 4.8 0.3 1.2 G daiben 95.9 68.9 81.3 Copepoda 9.4 24.9 15.3 M. edwardsi 16.2 2.3 6.1 Chironomidae larvae 18,9 1.4 5.1 Cyathura polita 162 0.7 3.3 Table 3. — Importance values of copepods, amphipods, and Neomysis americana in stomachs of June and July juvenile At- lantic tomcod pooled by 10-mm size intervals. Size interval Sample Neomysis (mm) size Copepods Amphipods americana 40-49 3 36.3 47.9 0.0 50-59 48 65.9 29.8 3.5 60-69 65 74.9 27.3 5.7 70-79 80 59.7 29.8 6.7 80-89 40 39.8 38.9 4.5 90-99 38 0.0 83.7 16.8 >100 5 8.8 75.5 0.0 FISHERY BULLETIN: VOL. 76, NO. 1 DISCUSSION Howe (1971) characterized tomcod as opportun- istic feeders; the data presented here qualify that hypothesis. Smaller tomcod, present during May and June, preyed upon copepods (Table 2) which have been the most abundant zooplankters col- lected by 76- and 150-/Lim mesh nets in this reach of the Hudson River (Lawleretal. see footnotes 5, 10; Lawler et al.i\ Lauer et al.^^). When total length reached 80-90 mm (June- July), food preference shifted to larger prey, e.g., amphipods (Table 3). Such a shift has been documented in a variety of species (Nikolsky 1963; Stickney et al. 1974; Werner 1974; Stickney 1976), including the re- lated species Gadus morhua (Kohler and Fitz- gerald 1969). This shift did not appear to be a response to changes in prey density, since abun- dance of copepods increased while that of amphi- pods decreased during June-August 1973-75 (Lawler et al. see footnotes 5, 10, 11). Copepods were a supplementary prey during December, occurring as frequently as the larger decapods Crangon septemspinosa (5.4%) and Rhithropanopeus harrisii (4.1%) which were rela- tively important during November (Table 2). Selection of smaller prey with the concomitant decrease of larger prey may be a response to the constriction of the alimentary canal by maturing gonads noted by Schaner and Sherman ( 1960). In Hudson River tomcod, gonadal biomass prior to spawning averages between 15 (males) and >30% (females) of the body weight minus the gonad weight. In contrast, female gonads in Hudson River Morone americana (Lawler et al. see foot- note 10) average about 8%, Alosa sapidissima about 22% (calculated from Lehman 1953), Tri- nectes maculatus less than 6% (calculated from Koski 1974), while those of Tautogolabrus adsper- sus from Long Island Sound averaged about 7% (Dew 1976) of the body weight minus the gonad weight. A decrease in prey (C. septemspinosa) avail- ability was not considered a factor in this change. In the Haverstraw Bay area, C septemspinosa effects of temperature from those of dissolved oxy- gen were not applied since the two parameters were highly correlated (r = -0.96). If was, how- ever, lowest when water temperatures were >24°C and dissolved oxygen (DO) <7 mg/1 and increased at temperatures <19°C and DO >7 mg/1 (Table 5). "Lawler, Matusky and Skelly Engineers. 1976. 1975 Hudson River aquatic ecology studies: Bowline Point and Lovett Generating Stations. Prepared for Orange and Rockland Utilities, Inc. i^Lauer, G. J., W. T. Waller, D. W. Bath, W. Meeks, R. Heffner, T. Ginn, L. Zubarik, P. Bibko, and P. C. Storm. 1974. Entrain- ment studies on Hudson River organisms. In L. D. Jensen (editor). Proceedings of the second entrainment and intake screening workshop, Feb. 5-9, 1973, p. 37-88. Johns Hopkins Univ., Baltimore, Md. 92 Table 4. — Mean length, weight, index of fullness, number of food items per stomach, and percent frequency of empty stomachs for juvenile Atlantic tomcod from Haverstraw Bay 1973-75. Number' Total length (mm) Weight (g) Mean SD Index of fullness Number of food items per stomach Frequency of empty stomachs (%) Month Mean SD Mean SD Mean SD May^ 36/38 28.9 3.2 0.3 0.1 21 809 14.630 29.3 14 1 0.0 June^ 100,210 688 11,0 35 18 17 224 9.645 68.4 178 5 0.0 July3 68 69 86 8 no 69 2,4 7,272 6,214 72 74 5.8 Aug/* 39 58 865 102 63 22 5387 5.333 5.0 6.2 10.3 Sept^ 30 43 909 99 74 28 7820 7453 9.2 94 2.3 Oct:> 40,43 986 122 98 3.7 25 317 41 485 18.7 169 2.3 Nov^ 42 42 1392 142 330 118 24403 22657 15.2 170 2.4 Dec" 46,74 143.8 12.9 352 12.2 12.902 7550 55.4 67.7 27 'Number of stomachs analyzed for index of fullness/total number of stomachs. ^Two dates. 1975 only ^1973 and 1974 "1973-75: no index of fullness for 1973 fish. Table 5. — Index of fullness of 1974 juvenile Atlantic tomcod, bottom water temperatures, and dissolved oxygen measure- ments, Haverstraw Bay. Sample Index of fullness Temp Dissol ved oxygen Date size Mean SD (•=C) mg/l % saturation 4 June 17 20.195 6226 175 8,2 85 11 June 44 16.839 10,040 203 8.4 91 26 June 25 17,977 12 066 21.7 7.2 82 29 June 14 13 482 5499 (') C) (') 10 July 24 7 062 5872 248 7.1 85 16 July 7 8 694 6 593 248 7.0 83 23 July 9 9798 9 264 248 6.8 81 8 Aug 12 7 895 5 986 25.9 6.9 84 13 Aug 18 3 288 3667 255 5.6 68 22 Aug 9 6241 6 087 267 5.4 79 10 Sept. 13 6.261 4610 234 6.8 79 26 Sept. 14 9.394 9 583 19.4 6.7 72 2 Oct 4 10 194 3634 18,9 7.6 81 8 Oct 13 22336 10 859 179 7.8 82 23 Oct 11 22,065 11,226 142 98 94 5, 8 Nov 14 20,695 18794 14,6 9.4 91 13 Nov. 13 27,898 22372 122 10.2 94 3 Dec. 22 12.370 8107 5,6 11.6 92 'Data not available. were relatively abundant in trawl collections Au- gust through November 1973 and 1974 (Lawler, Matusky and Skelly Engineers unpubl. data). Haefner (1976) found that greatest abundance of C. septemspinosa in channel areas of the York River and lower Chesapeake Bay occurred when water temperatures were 5°-10°C and was a result of migration from littoral areas to deeper, more saline areas; such a temperature regime occurs in Haverstraw Bay between mid-November and mid- December (Figure 1). Feeding intensity and growth followed similar seasonal patterns. Rapid growth and relatively intense feeding occurred during May, June, Oc- tober, and November (Table 4; Figure 3); growth and feeding were depressed during July- September. Prey density was not considered limit- ing during summer months since Neomysis ameri- cana was generally abundant. Also, resumption of feeding and growth occurred during October when macrozooplankton standing crop was lower than previous months (Lawler et al. see footnotes 5, 10, 11; Lauer et al. see footnote 12). Seasonally fluc- tuating abiotic factors, then, may be affecting growth and feeding. Food consumption in other species of gadids has been observed (Tyler 1970) or postulated (Sikora et al. 1972) to be inhibited at temperatures >20°C. Tomcod are considered to have a low thermal optimum (Huntsman and Sparks 1924; Bigelow and Schroeder 1953; Howe 1971). Retardation of growth during summer months when water tem- peratures exceed 24°C has been observed in the Hudson River (Lawler et al. see footnote 5; Texas Instruments see footnote 7; Dew and Hecht see footnote 8) and Weweantic River, Mass. (Howe 1971), populations. Growth of juveniles from the Woods Hole area during 1962 (maximum surface water temperature = 21.1°C) did not appear to cease during midsummer (Lux and Nichy 1971); however, only 22 young-of-the-year fish were caught between June and August. Concomitant with elevated water temperature is decreased dissolved oxygen. In separate reviews of dissolved oxygen requirements, Doudoroff and Shumway ( 1970) noted that feeding and growth responses to low DO levels have been variable, while Davis (1975) suggested that inhibition oc- curred at 509c of air saturation. Warren et al. (1973) found that growth and feeding of Onco- rhynchus kisutch and O. tshawytscha were inhib- ited when saturation was <100%, but that only a 10% decrease in production would occur at 70% saturation. Thatcher (1975; cited in McKim et al. 1976) found that O. kisutch acclimated at 15°C did not reduce food consumption or growth when DO was >5 mg/l (49% saturation). Tomcod feeding, measured hy If, was minimal at DO <7 mg/l during 1974; July-September percent saturation ranged from 68 to 85% (Table 5). In light 93 FISHERY BULLETIN: VOL. 76. NO. 1 of the above studies on salmonids, it seems unlikely that DO levels encountered in Haverstraw Bay are the primary variable affecting feeding and growth. The summer temperature regime of the Hudson River, then, appears to be near maximum for this species and may be capable of inhibiting feeding and retarding growth. ACKNOWLEDGMENTS Support for this study came from Orange and Rockland Utilities, Inc. I am indebted to my wife, Vincentia, for her encouragement and assistance throughout this investigation. I am also grateful to J. Berkun, R. Alevras, M. Baslow, T. C. Cosper, C. B. Dew, B. Lippincott, J. Matousek, S. Weiss, M. Weinstein, and R. Wyman for their criticisms and suggestions. LITERATURE CITED BALON, E. K. 1975. Terminology of intervals in fish development. J. Fish. Res. Board Can. 32:1663-1670. BIGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. BOOTH, R. A. 1967. A description of the larval stages of the tomcod, Microgadus tomcod, with comments on its spawning ecol- ogy. Ph.D. Thesis, Univ. Connecticut, Storrs, 43 p. CURRAN, H. W., AND D. T. RiES. 1937. Fisheries investigations in the lower Hudson River. In A biological survey of the Lower Hudson Watershed, p. 124-145. N.Y. State Conserv. Dep., Suppl. 26th Annu. Rep. DAVIS, J. C. 1975. Minimal dissolved oxygen requirements of aquatic life with emphasis on Canadian species: A review. J. Fish. Res. Board Can. 32:2295-2332. Dew, C. B. 1976. A contribution to the life history of the cunner, Tautogolabrus adspersus, in Fishers Island Sound, Con- necticut. Chesapeake Sci. 17:101-113. DOUDOROFF, P., AND D. L. SHUMWAY. 1970. Dissolved oxygen requirements of freshwater fishes. FAO Fish. Tech. Pap. 86, 291 p. HAEFNER, P. A., JR. 1976. Seasonal distribution and abundance of sand shrimp Crangon septemspinosa in the York River-Chesapeake Bay Estuary. Chesapeake Sci. 17:131-134. HOWE, A. B. 1971, Biological investigations of Atlantic tomcod, Micro- gadus tomcod Walbaum, in the Weweantic River estuary, Massachusetts, 1967. M.S. Thesis, Univ. Mas- sachusetts, Amherst, 82 p. HUNTSMAN, A. G., AND M. I. SPARKS. 1924. Limiting factors for marine animals. 3. Relative re- sistance to high temperatures. Contrib. Can. Biol., New Ser. 2:97-114. 94 KOHLER, A.C., AND D, N. FITZGERALD. 1969. Comparisons of food of cod and haddock in the Gulf of St. Lawrence and on the Nova Scotia Banks. J. Fish. Res. Board Can. 26:1273-1287. KOSKI, R. T. 1974. Life history and ecology of the hogchoker, Trinectes maculatus, in its northern range. Ph.D. Thesis, Univ. Connecticut, Storrs, 111 p. LEHMAN, B. A. 1953. Fecundity of Hudson River shad. U.S. Fish Wildl. Serv., Res. Rep. 33, 8 p. « Lux, F. E., AND F. E. NICHY. 1971. Numbers and lengths, by season, of fishes caught with an otter trawl near Woods Hole, Massachusetts, Sep- i tember 1961 to December 1962. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Spec. Sci. Rep. Fish. 622, 15 p. MASSMAN, W. H. 1957. New and recent records for fishes in Chesapeake Bay. Copeia 1957:156-157. MCKIM, J. M., R. L. ANDERSON, D. A. BENOIT, R. L. SPEHAR, AND G. N. Stokes. 1976. Effects of pollution on freshwater fish. J. Water Pollut. Control Fed. 48:1544-1620. Miller, R. R. 1972. Threatened freshwater fishes of the United States. Trans. Am. Fish. Soc. 101:239-252. NIKOLSKY, G. V. 1963. The ecology of fishes. Academic Press, N.Y. , 352 p. Schaner, E., and K. Sherman. 1960. Observations on the fecundity of the tomcod, Micro- gadus tomcod (Walbaum). Copeia 1960:347-348. Scott, W. B., and E. J. Grossman. 1973. Freshwater fishes of Canada. Fish. Res. Board Can., Bull. 184, 966 p. SiKORA, W. B., R. W. Heard, and M. D. Dahlberg. 1972. The occurrence and food habits of two species of hake, Urophycis regius and U. floridanus in Georgia es- tuaries. Trans. Am. Fish. Soc. 101:513-525. STICKNEY, R. R. 1976. Food habits of Georgia estuarine fishes IL Sym- phurus plagiusa (Pleuronectiformes: Cyno- glossidae). Trans. Am. Fish. Soc. 105:202-207. STICKNEY, R. R., G. L. Taylor, and R. w. Heard m. 1974. Food habits of Georgia estuarine fishes. L Four sp)ecies of flounders (Pleuronectiformes: Bothidae). Fish. Bull., U.S. 72:515-523. THATCHER, T. O. 1975. Some effects of dissolved oxygen concentration on feeding, growth and bioenergetics of juvenile coho salm- on. Diss. Abstr. 35, 5763-B TYLER, A. V. 1970. Rates of gastric emptying in young cod. J. Fish. Res. Board Can. 27:1177-1189. Warren, C. E., P. Doudoroff, and D. L. Shumway. 1973. Development of dissolved oxygen criteria for fresh- water fish. Environ. Prot. Agency, EPA-R3-73-019, 121 p. WERNER, E. E. 1974. The fish size, prey size, handling time relation in several sunfishes and some implications. J. Fish. Res. Board Can. 31:1531-1536. WINDELL, J. T. 1971. Food analysis and rate of digestion. In W. E. Ricker (editor), Methods for assessment offish production in fresh waters, 2d ed., p. 215-226. IBP (Int. Biol. Pro- gramme) Handb. 3. EGGS AND LARVAE OF SCOMBER SCOMBRUS AND SCOMBER JAPONICUS IN CONTINENTAL SHELF WATERS BETWEEN MASSACHUSETTS AND FLORIDA Peter L. Berrien' ABSTRACT Larval Scomber scombrus and Scomber japonicus from the western North Atlantic Ocean are com- pared. At 4 to 11 mm S. japonicus are deeper bodied, and at 3 to 15 mm have greater preanus lengths than S. scombrus of comparable sizes. Scomber scombrus larvae are more heavily pigmented than S. japonicus, particularly on the dorsal trunk surface and at the cleithral sympysis. In continental shelf waters between Martha's Vineyard, Mass., and Palm Beach, Fla., 1966-68, S. scombrus eggs occurred north of Cape Hatteras, N.C., mostly in the shoreward half of shelf waters, during spring and summer. Surface temperatures associated with egg occurrences varied from 6.3° to 16.9°C. Scomber japonicus eggs were taken south of Cape Hatteras, in the outer half of shelf waters, during winter and spring cruises. Surface temperatures associated with egg occurrences ranged from 20.4° to 25.4°C. Larval S. scombrus occurred north of Cape Hatteras during spring and summer with concurrent surface temperatures ranging from 12.3°to20.7°C. With the exception of three specimens, S. japonicus larvae occurred south of Cape Hatteras and were taken where the surface temperature rsmged from 16.0°to29.4°C. Despite an abundance of publications describing the young stages of Atlantic mackerel, Scomber scombrus Linnaeus, and their occurrences in the western North Atlantic (Dannevig 1919; Sette 1943; Bigelow and Schroeder 1953; Berrien 1975), very little information exists on young of the con- generic chub mackerel. Scomber japonicus Hout- tuyn, from the same area. There are no descrip- tions of S. Japonicus eggs, larvae, or juveniles from the western North Atlantic, although there are excellent descriptions of specimens from the Pacific Ocean (Fry 1936a; Orton 1953; Uchida et al. 1958; Kramer 1960; Watanabe 1970) and some brief descriptions of this species from European waters (Ehrenbaum 1924; Padoa 1956). Ehren- baum (1924), Padoa (1956), and Dekhnik (1959) compared larvae of the two species. Reports of young S. japonicus in the western North Atlantic are limited to those by Anderson and Gehringer (1958), Dooley (1972), Fahay (1975), and de Sylva.2 Although adults of S. japonicus are known to range from the Gulf of St. Lawrence (Leim and Scott 1966) to Bermuda and the Gulf of Mexico •Northeast Fisheries Center Sandy Hook Laboratory, Na- tional Marine Fisheries Service, NOAA, Highlands, NJ 07732. ^de Sylva, D. P. 1970. Ecology and distribution of postlar- val fishes of southern Biscayne Bay, Florida. Prog. Rep. to Div. Water Qual. Res., Water Qual. Off., U.S. Environ. Prot. Agency Contract FWQA 18050 Div. Rosenstiel School Mar. Atmos. Sci., Univ. Miami, 198 p. (Unpubl. manuscr.) (Briggs 1958) in the western Atlantic, they occur irregularly along the U.S. east coast. In various years they have been abundant, uncommon, or absent (Hildebrand and Schroeder 1928; Bigelow and Schroeder 1953). This species apparently in- habits warmer waters than does S. scombrus (Bigelow and Schroeder 1953; Matsui 1967). The purposes of this paper are: 1) to present descriptive, comparative information on two species of Scomber larvae, in order to facilitate their identification; and 2) to compare the spawn- ing areas of the two species as indicated by occur- rences of Scomber young taken between Mas- sachusetts and Florida. Specimens utilized in this study were taken primarily during ichthyoplankton survey cruises by the RV Dolphin in continental shelf waters from December 1965 to February 1968 between Martha's Vineyard, Mass., and Palm Beach, Fla. Some larvae in the descriptive section were taken on other cruises during April 1971 and June 1972, within the same area. PROCEDURES Sampling Eight plankton sampling cruises were con- ducted between December 1965 and December Manuscript accepted June 1977. FISHERY BULLETIN: VOL. 76, No. 1, 1978. 95 FISHERY BULLETIN: VOL. 76, NO. 1 1966 aboard the RV Dolphin in continental shelf waters, between Martha's Vineyard and Cape Lookout, N.C. Four cruises were made between May 1967 and February 1968 between New River Inlet, N.C, and Palm Beach (Figure 1). Gulf V samplers, with 0.4-m mouth and 0.52-mm mesh openings, were used for plankton tows. The tows were 0.5 h long at a speed of 9.3 km/h (5 knots) in a step-oblique pattern. Normally the nets were low- ered in six 3-m depth increments and towed for 5 min at each depth. One Gulf V net (net 1) sampled from 0 to 15 m, and a second net sampled from 18 to 33 m. While setting and retrieving net 2, contami- nation above 15 m was inevitable, since the nets were not equipped with closing devices. Plankton samples were preserved in 5% Formalin^ buffered with borax. Sampling time, whether day or night, was essentially random, in that there was no prearranged time schedule. At each station we measured surface water temperature, made a bathythermograph cast to a maximum depth of 275 m, and measured salinity with an in situ in- duction salinometer at 5-m intervals down to in- clude the plankton sampling depth. Additional de- tails on the sampling scheme and gear used, as well as temperatures, salinities, zooplankton vol- umes, and midwater trawl catches, were sum- marized by Clark et al. (1969, 1970). Identification of Eggs and Larvae Scomber scombrus eggs were identifiable using criteria summarized by Berrien (1975). Briefly, distinguishing features of this species' eggs are: they are spherical and have a diameter of about 1.0 to 1.3 mm; they have a single yellowish oil globule about 0.3 mm in diameter; and after blas- topore closure, melanophores occur on the head, trunk, and oil globule. Pigment is absent from the yolk except just prior to hatching when one melanophore occurs near each side of the embryo, immediately posterior to the head. Despite a lack of information on S. japonicus 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Figure l. — Ichthyoplankton survey area; transects designated by single letters were sampled eight times, December 1965 to December 1966; those with two letters were sampled four times. May 1967 to February 1968. Stations starting at 1 on the inshore end of each transect were numbered consecutively, progressing ocean ward. nOO *<" A MARTHA'S VINEYARD i MONT AUK POINT ■,■■••- C FIRE ISLAND ■ • 'D BARNEGAT INLET ■ ■ • • • E GREAT EGG INLET ;v S . . <-■ F CAPE HENLOPEN . . .: G ASSATEAGUE ISLAND . ■■ H PARRAMORE ISLAND . J CAPE HENRY . K CURRITUCK BEACH . L OREGON INLET . M CAPE HATTERAS ■ • . N OCRACOKE INLET . P CAPE LOOKOUT ., AA NEWRIVK INLET CAPE FEAR • CC MYRTLE BEACH . DD GEORGETOWN \ 7(f zi \ ■ EE CHARLESTON FF SAVANNAH GG BRUNSWICK • HH JACKSONVILLE .-,JJ MATANSAS INLET ..KK PONCE DE LEON INLET . LL CAPE CANAVERAL MM VERO BEACH . NN ST. LUCIE INLET PP PALM BEACH \ A, 3(f 2^ 80° In 96 eggs from the Atlantic Ocean, they have been well described from the Pacific Ocean, and are similar to eggs of S. scombrus in size and appearance (Fry 1936a; Kramer 1960; Watanabe 1970). The most obvious difference between eggs of the two species is the amount of pigment found on the yolk surface during the late or third stage in development. Scomber japonicus develops several melanophores on the yolk while S. scombrus has, at most, a pair of melanophores, as described above. Due to simi- larity of early-stage eggs of the two Scomber species, their identification must depend upon other information, such as spawning area, and the proximity of older identifiable stages. In separating Scomber spp. larvae from larvae of other fishes I found the descriptions and illus- trations by Bigelow and Schroeder ( 1953), Kramer ( 1960), and Watanabe (1970) to be especially help- ful. Scomber spp. larvae are characterized by the following: 1) they have 31 myomeres and lack preopercular spines, unlike other scombrid larvae in the western North Atlantic which have more myomeres and possess strong spines; 2) melanophores are present above the forebrain, midbrain, and gut, and along the postanus ventral edge of the trunk; 3) prominent recurved teeth form in larvae by about 4 mm, and are present well into the juvenile stage although somewhat em- bedded and obscured at sizes above about 15 mm; and 4) a large portion of Scomber spp. larvae be- tween about 7 and 15 mm have noticeably subter- minal mouths. Other larval fishes found along the U.S. east coast which grossly resemble one or the other of the two Scomber spp. include Sebastes marinus, Pomatomus saltatrix, Centropristis striata, and Stenotomus chrysops. Despite pigmentation similarities myomere counts alone will separate Scomber larvae (with 31 myomeres) from P. sal- tatrix (with 26) and C. striata and Stenotomus chrysops (each with 24 myomeres). Sebastes marinus can have the same number of myomeres (with 30 to 32) as Scomber and is pigmented in most of the same body areas as both species of Scomber. However, at lengths less than about 9 mm Sebastes marinus lack teeth and have dorsal and ventral trunk melanophores which are close enough together to appear as dorsal and ventral lines of pigment. Comparably sized Scomber lar- vae have prominent teeth and discrete dorsal and ventral trunk melanophores. Also Sebastes marinus larvae are more slender and have shorter snout-to-anus lengths than Scomber larvae. The presence of temporal and preopercular spines on Sebastes marinus and their absence on Scomber larvae separate the two species at lengths >9 mm, before fin-ray counts are distinguishable. Treatment of Specimens and Data Measurements, as defined by Kramer (1960), made in this study include: standard length (SL = anterior tip of snout to tip of notochord, or to posterior edge of the hypurals after notochord flexure); preanus length (PAL = anterior tip of snout to the most posterior edge of the anus); and body depth (BD = the vertical distance from the dorsal surface of the body directly above the dorsal point of the cleithrum to the ventral point of the cleithrum). Length measurements in this paper are standard lengths, unless otherwise stated. Osteological characters in developing Scomber larvae were investigated by examination of bone- stained specimens (Hollister's method in Clothier 1950) and radiographs. All Scomber eggs in samples containing <400 eggs were identified and tabulated. In larger sam- ples, the numbers of S. scombrus eggs were esti- mated from a random subsample of 200. To test the validity of this procedure S. scombrus eggs were identified from seven aliquots of 200 eggs from one sample. No significant differences were found be- tween aliquots (chi-square = 5.415, P = 0.5). Lengths for length-frequency diagrams were measured to the nearest 0.1 mm in fish <15 mm and to the nearest 0.5 mm in those >15 mm. Mea- surements were taken of all specimens from sam- ples of 100 or fewer fish and of 50 to 75 randomly selected specimens from larger samples. The numbers of Scomber spp. eggs and larvae taken during survey cruises are presented on charts. For these charts the catches from net 1 (0-15 m) and net 2 (18-33 m) were combined at stations where both were towed. Before these numbers were plotted some were adjusted in an attempt to standardize the catches. Because net 2 spent an estimated 3 min of the V^-h. tow being set and retrieved through the upper 15 m, the catch by net 2 was reduced by 10% of the net 1 catch to correct for contamination. In cases where there was insufficient water depth to allow lowering the plankton net for the standard of six 3-m depth increments, the towing scheme was altered. Dur- ing these tows we sampled for 15 min at each of two levels, or for 10 min at each of three levels. The resulting catch was reduced to one-third when two 97 FISHERY BULLETIN: VOL. 76, NO. 1 levels were sampled or to one-half when three levels were sampled. Fahay (1974) explained this procedure in more detail. COMPARISON OF TWO SPECIES OF SCOMBER LARVAE Scomber larvae occurred in samples from our northernmost transect, off Martha's Vineyard to our southernmost transect off Palm Beach. The larvae were of two types, the distinction between the two being more obvious in larvae smaller than 15 mm. One type, collected north of Cape Hat- teras, predominantly over the inshore and central portions of the continental shelf, during May, June, and August 1966, was tentatively identified as Atlantic mackerel, S. scombrus. A second type collected south of Cape Hatteras was tentatively identified as chub mackerel, S.japonicus. It occur- red predominantly in samples taken near the offshore edge of the continental shelf, during May and July 1967 and January and February 1968. The identities of the two types were confirmed by examination of some meristic characters of the large larvae and juveniles. Because of the similarity and possible confusion of these two species, the following descriptions and comparisons were compiled to facilitate future identifications. Three study areas were considered in larval development: meristic characters, mor- phology, and pigmentation. Meristic Characters Of the 12 characters listed by Matsui (1967, table 5) as distinguishing between the species of Scomber, four were found to be useful in identify- ing young stages dealt with here. These were: 1) first-dorsal-fin spine counts; 2) counts of pre- caudal and caudal vertebrae; 3) counts of first- dorsal-fin ptergiophores and the arrangements in relation to neural spines; and 4) the relative pos- ition of the first haemal spine and the first anal pterygiophore. Scomber japonicus has 9 or 10 first-dorsal-fin spines and S. scombrus has 1 1 to 14 (Matsui 1967). Examination of Formalin-preserved specimens under a dissecting microscope revealed that counts of 9 or 10 were attained by a length of 18.5 mm in S.japonicus and counts of 11 to 15 by 21.0 mm in S. scombrus. However, bone-stained specimens of both species had higher counts and earlier formation of spines than indicated in the above. Apparently some of the minute, posterior spines in the first dorsal fin, observed in bone- stained specimens, were obscured in nonstained specimens by surrounding muscle and epithelial tissue and by their position in the longitudinal groove. I observed a complement of 10 or 1 1 spines in S . japonicus as small as 11.9 mm long and 12 to 17 spines in S. scombrus 18.2 mm and greater (Table 1). Counts of vertebrae were made to help identify the two species of Scomber larvae. Scomber japonicus is reported to have 14 precaudal and 17 caudal vertebrae and S. scombrus to have 13 pre- caudal and 18 caudal vertebrae (Matsui 1967). The first caudal vertebra is the most anterior ver- tebra which has an elongate pointed haemal spine and lacks ribs. In Scomber larvae the haemal spine on the first caudal vertebra is noticeably longer than the haemal arch on the last precaudal vertebra. Also, rib articulation surfaces on haemal arches of posterior precaudal vertebrae are dis- tinctly flattened or truncated, rather than pointed as are haemal spines on caudal vertebrae. In my work counts of precaudal vertebrae were distin- guishable in bone-stained S.japonicus as small as 7.6 mm (indeterminate at 6.7 mm) and on radio- graphs by 9.3 mm. Precaudal counts characteristic of S. scombrus were observable in bone-stained larvae at 8.6 mm (indeterminate at 7.6 mm) and on radiographs by 11.2 mm (Table 1). A few of the S. scombrus specimens had precaudal and caudal vertebral counts different from those reported by Matsui (1967). Six of the 136 S. scombrus speci- mens bone-stained or X-rayed large enough for determination had 12 precaudal and 19 caudal vertebrae. In two other specimens the 28th and 29th vertebrae were fused together, as evinced by a total count of 30 and by the presence of two neural and two haemal spines on the 28th ver- tebra. One additional larva was observed with partial fusion of the same two vertebrae. The numbers of first-dorsal-fin pterygiophores separate the two species of Scomber. Matsui (1967) reported S. japonicus has 12 to 15 first- dorsal-fin pterygiophores and S. scombrus has 21 to 28. Full complements of pterygiophores, 13 or 14 in S. japonicus and 22 to 25 in S. scombrus, were found in bone-stained S. japonicus as small as 20.2 mm and on radiographs by 33.3 mm; they were found in bone-stained S. scombrus at 32.0 mm and on radiographs at 38.8 mm (Table 1). Because anterior pterygiophores ossify before posterior ones and because there is a difference 98 BEKKIEM: KUUS AINU LAKVAfc Ut !HJUMtft.ti Table l. — Some meristic characters in Scomber japonicus and S. scombrus young as determined in bone-stained (and two X-rayed) specimens. Dj refers to the first dorsal fin; pterygiophore counts were made between successive neural spines, starting in the second interneural space. ( — = count was indeterminate. X = X-rayed specimen. * = pterygiophore series completed. M = mutilated, spine(s) lost in handling.) SL (mm) Scomber japonicus Vertebrae D, spines D, pterygiophores Scomber scombrus SL (mm) Vertebrae D, spines D, pterygiophores 6.7 7.6 7.7 84 85 9.0 9 1 10.2 10.5 11.7 11-9 12.4 13.8 14.0 165 17.7 20.2 22.1 24.7 26.3 28 6X 333X 14 + — — — 14 + 17 1 — 14 + — 6 — 14 + 17 7 — 14 + 17 4 — 14 + 17 6 — 14 + 17 8 — 14 + 17 9 — 14 + 17 8 11121 14 + 17 11 1121 14 + 17 11 11121 14 + 17 11 112111 14 + 17 10 11121 14 + 17 10 11121111 14 + 17 10 1112111 14 + 17 11 1121111112r 14 + 17 10 111211120211 14 + 17 11 11121111 14 + 17 11 11121111112" 14 + 17 10 111211 14 + 17 11 11121111121" 7.6 8.6 9.3 13 + • 13 10.5 13 + 18 — — 10.7 13 + 18 — — 11.4 13 + — — — 11.6 13 + 18 — — 12.3 13 + 18 2 13.4 13 + 18 5 — 14.8 13 + 18 6 — 16.0 13 + 18 10 — 18.2 13 + 18 17 1122 19.8 12 + 19 15 1123 22 1 13 + 18 16 112221 24.3 13 + 18 13 11222 26.0 13 + 18 14 1122221 28.2 13 + 18 14 11222222 299 12 + 19 15 1123221 32.0 13 + 18 12M 1132212212221 34.2 13 + 18 13 112222223222" 36.6 13 + 18 13 1122322133221 38.6 13 -h 18 13 112322122222* between the two species in counts of pterygiophores in anterior, successive interneural spaces, the two species can be separated well be- fore the total complement is attained. A count of six pterygiophores in the 2d through 6th inter- neural spaces, characteristic of S. japonicus, was observed in bone-stained larvae as small as 11.7 mm and on radiographs at 20.2 mm; a count of six or seven pterygiophores in the 2d through 50th interneural spaces, characteristic of S. scombrus, was observed in bone-stained larvae as small as 18.2 mm and on radiographs at 20.1 mm. In S. japonicus the first anal pterygiophore is anterior to the first haemal spine while in S. scombrus the first anal pterygiophore is posterior to the first haemal spine (Matsui 1967). This was observable in bone-stained iS. japonicus at 11.7 mm and in S. scombrus at 32.0 mm, and on radio- graphs at 17.0 mm in S. japonicus and 32.0 mm in S. scombrus. Body Proportions Larvae of the two species differ noticeably in several body proportions. Scomber japonicus is deeper bodied and has a greater preanus length than S. scombrus. Measurements of body depth (BD) and preanus length (PAL) were converted to percentages of standard length (SL) and the re- sults were graphed (Figure 2). Although the sep- aration of the two species by these characters is not total, more than two-thirds of the larvae are separable by BD measurements at lengths of 4 to 11 mm and by PAL measurements at 3 to 15 mm long. Of these two characters the PAL difference is more useful, as it is present over a greater size range. Other morphological differences between the two species have been reported by previous work- ers. These contrasts were not considered strong enough in the larvae from this study to warrant elaboration. Padoa (1956) noted a larger eye, shorter lower jaw, and shorter snout relative to eye diameter in S. japonicus than in S. scombrus. Dekhnik (1959) presented a brief and generalized comparison of larvae of the two species. She re- ported S. japonicus larvae are more advanced than S. scombrus of the same length. Thus S. japonicus are smaller than S. scombrus at hatch- ing, at yolk and oil globule absorption, and at the initial formation of caudal fin rays. These differ- ences were not as striking in my specimens. In our survey both species apparently hatched at about 3 mm long, and yolk and oil globules were absorbed in both by a length of 4 mm. Caudal ray develop- ment varied between species; in S. japonicus the 99 FISHERY BULLETIN: VOL. 76. NO. 1 • — ■—-• B o --B>- • — ■- o-e- o o-E>o o -S- o n iH3- o D o--g-o o- r^.,,»j -■ • a -a I o— -Q— -o Q o o o z 3 z • — ■- o-B- o-Q-o •-*--■• ©e- o g. ••—■-• I >- Q • O -a o • O i.P o B- — o o s o • O Q- — -w..::.:-.^-J^ o 0 ■•■ G i-.■^■.■.^•.. M CO C 0) v bo •T3 C to u v o C8 >H -tf u is -O 4) C *J in =«, g en a> d) -a 1 J -O o E E S 'S b TS r 2 c h- 3 3 o ^ „ z to ^ o-- ja^ Sn 1 b' (T 6 +1 < >-> a m 4) z c -S < 4) (- CO &l 4) U! El C8 ... — m o u CD <3 4) s e o , O 4) ^ t "S g to 'O 4) U 01 a. to o » .(J o Vh *" o to C o t a o 1-1 a. >> "S PQ o 1 irt 1 (N Cd Bi P O E 1 s ' s Hi9N3T ayVQNVlS JO 39ViN3Dy3d T 100 BERRIEN: EGGS AND LARVAE OF SCOMBER rays were forming at a length of 5 mm and in S. scombrus at 7 mm. Pigmentation Differences in pigmentation were found be- tween larvae of the two species. Pigmentation over the gut and midbrain and on the caudal region is not described in detail because it does not differ between the two species. A series of 210 S. japonicus specimens ranging from 2.8 to 49.0 mm long and 187 S. scombrus, 2.6 to 21.6 mm long, were used in the pigmentation comparison. Figure 3 illustrates the development of pigmentation and various body features. Forebrain Scomber scombrus larvae usually acquire melanophores on the forebrain at smaller sizes than S. japonicus. They were present on S. scom- brus as small as 3.7 mm and were present on all larvae larger than 5.5 mm. The smallest S. japonicus with such pigment was 5.2 mm, and not until 8.7 mm was attained did all larvae have this pigment. Forebrain pigment should not be con- fused with that on the midbrain which larvae of both species possess at all sizes. Hindbrain Pigmentation on the hindbrain begins as a single melanophore then increases to three to five melanophores on the posterior and middle portion of the hindbrain. This pigmentation is increas- ingly obscured by overlying tissue after about 5 mm. All S. scombrus larvae examined had this pigment, but S. japonicus <3.5 mm did not. Snout Pigmentation on the snout refers to melanophores on, or within, epidermal tissue, not subsurface as on the forebrain. Melanophores ap- pear first near the tip of the snout. Scomber scom- brus generally develop snout pigmentation at smaller sizes than S. japonicus. The smallest S. scombrus with such pigmentation was 4.3 mm long and it was present in all that were 6.3 mm and greater. It was first observed in S. japonicus at 5.2 mm and was present in all specimens 10.5 mm and longer. Cleithral Symphysis Pigmentation at the symphysis of the cleithra, and on the isthmus immediately anterior to the symphysis, was lacking in all specimens of S. japonicus. However, in S. scombrus prominent melanophores were noted at this location in larvae as small as 3.7 mm and occurred in all larvae >8.0 mm (Figures 3, 4). Melanophores occurred on the isthmus of S. scombrus in: 13% of those 4.0 to 4.9 mm long; 41% of those 5.0 to 5.9 mm; 67% of those 6.0 to 6.9 mm; 95% of those 7.0 to 7.9 mm; and in all specimens 8.1 mm and longer. In larvae <8 mm the presence of melanophores at the cleithral symphysis indicates S. scombrus; however, the absence of this pigment at this size does not indi- cate either of the two species. At sizes >8 mm the presence of this pigment indicates S. scombrus and its absence indicates S. japonicus. Lower Jaw Melanophores on the lower jaw first appear at the mandibular symphysis, then spread laterally and posteriorly. Scomber scombrus acquire this pigment at a smaller size than S. japonicus. The smallest larval S. scombrus observed with lower jaw pigmentation was 4.6 mm long and it occurred in all specimens 6.2 mm and greater. The smallest S. japonicus with such pigment was 8.3 mm long and it occurred in all larvae of this species 11.7 mm and greater. Ventrum of Gut In his paper on the development of S. japonicus, Kramer (1960) referred to two or three charac- teristic, minute melanophores on the ventral sur- face of the gut, found after yolk absorption. During my study pigment in this location was observed in both S. japonicus and S. scombrus. The percent occurrence of melanophores on the ventrum of the gut in S. japonicus <12 mm long varied from 70% to 92% for each 1-mm size group, with an average of 88% occurrence. The occurrence for the same sizes of iS. scombrus varied from 10% to 41%, with an average of 28%. Dorsum of Trunk There are substantial differences between the two species in pigmentation on the dorsum of the 101 FISHERY BULLETIN: VOL. 76, NO. 1 G 2.9 K 117 L 15.1 Figure 3. — Scomber japonkus, A to F; S. scombrus, G to L; lengths (SL) are given in millimeters. trunk, posterior to the nape, particularly at lengths less than about 8 mm (Figures 3, 4). AUS. scombrus specimens examined, 2.6 mm and larger, possessed dorsal melanophores. At lengths less than about 5 mm this pigmentation consists of a single median series of dendritic melanophores, initially 3 to 6 in number, increasing to 4 to 13, located between myomeres 13 and 28. In larvae greater than about 5 mm the median series be- comes double, one row on each side of the develop- ing dorsal fin base, and increases in number of melanophores and extent so that by a length of 9.5 102 BERRIEN: EGGS AND LARVAE OF SCOMBER HATCH 4 t. 8 10 12 FORE BRAIN HIND BRAIN CLEITHRAL SYMPHYSIS LOWER JAW GUT VENTRUM TRUNK DORSUM FLANK MIDLATERAL TRUNK VENTRUM No S. joponicus pigmenfed at cleithfot symphyjii. STANDARD LENGTH (MMi Figure 4. — Acquisition of pigmentation of larval Scomber scombrus and S. japonicus. Dashed lines indicate some speci- mens have pigmentation; solid lines indicate all specimens have pigmentation. The upper of each pair of lines refers to S. scom- brus, the lower to S. japonicus. mm the dorsal edge of the trunk is pigmented from nape to caudal fin. With further growth melanophores form on the flanks, and spread downward from the dorsal row; this happens first in the abdominal area, then posteriorly. Scomber Japonicus larvae develop this pigmen- tation at larger sizes than S. scombrus. Only one S. japonicus (4.1 mm) <5.2 mm long possessed dorsal melanophores. Subsequent percent occur- rences of S. japonicus larvae possessing this pig- mentation were: 24'7f at 5.0 to 5.9 mm, 597c at 6.0 to 6.9 mm, and lOO'/f at 7.0 mm and greater. The largest S. japonicus lacking dorsal melanophores was 6.9 mm long. As in S. scombrus this pigmen- tation develops from a single median series into a double row and increases to extend from the nape to the caudal fin by a length of about 11.0 mm. Thus at sizes smaller than about 11 or 12 mm there is a difference in dorsal pigmentation be- tween the two species. While S. scombrus possess dorsal pigmentation many S. japonicus either lack melanophores in this location or have consid- erably less than comparably sized S. scombrus. This conclusion is in general agreement with ear- lier published statements. Padoa (1956) men- tioned that postanal pigmentation of S. japonicus is less intense than that of S. scombrus, but he did not specify whether he was referring to dorsal or ventral postanal pigment. Dekhnik (1959) re- ported that, between yolk absorption and a length of 6.18 mm TL, larval S. japonicus lack melanophores on the dorsal edge of the trunk while larval S. scombrus have melanophores in this area. Fry (1936a, figure 12G) illustrated a late yolk- sac stage S. japonicus with a small dorsal patch of melanophores near the 23d myomere, but did not comment in the text on the occurrence of this pig- mentation. Uchida et al. (1958) and Kramer (1960) referred to a similar dorsal patch of melanophores in some of their late yolk-sac stage S. japonicus. Watanabe (1970) did not illustrate such dorsal pigment in his paper on this species. None of the S. japonicus larvae in my study had this dorsal patch; however, I identified only two larvae <3.0 mm long. Flank A longitudinal row of melanophores develops along the midline of the lateral trunk surface in Scomber larvae. This row begins forming in S. japonicus at 8.3 to 9.6 mm long and in S. scombrus at 9.6 to 11.1 mm long. The pigment in this row, first observable as a few distinct melanophores in the postanal region, increases to form a line flanked by scattered melanophores. These scat- tered melanophores tend to occur along the myosepta; this tendency is more pronounced in S. scombrus than in S. japonicus. Postanus Ventral Pigmentation Both species possess postanus ventral pigmen- tation, at all sizes examined. This pigmentation occurs in the smallest larvae as a median row of 15 to 20 melanophores. This series occurs first near the dermal surface and becomes internally situated along the median ventral septum as the anal fin develops. A second, double series of melanophores forms on the dermal surface, on either side of the developing anal fin base. This second series appears first at lengths of 7.0 to 7.9 mm in both species and increases in number of melanophores, so that by a length of about 15 mm there is a line of melanophores along either side of the anal fin, continuous with a median group of melanophores between the anal and caudal fin. The initial median series of melanophores gradually becomes obscured by overlying tissue 103 FISHERY BULLETIN: VOL. 76, NO. 1 and pigmentation, so that by a length of 15 mm only one to four melanophores of that series are still visible, and these only under favorable light- ing conditions. Summary of Contrasting Characters The precaudal and caudal vertebral counts, 14 + 17 in S.Japonicus and 13 + 18 (or 12 + 19) inS. scombrus, are distinguishable in S.Japonicus as small as 7.6 mm and in S. scombrus at 8.6 mm. First dorsal fins, with 10 or 11 spines in S. japonicus and 12 to 17 spines in S. scombrus, at- tain their full complement by 13.0 and 17.0 mm in the two species, respectively. In S. japonicus a total complement of 13 or 14 first-dorsal-fin pterygiophores is attained by 20.2 mm while in S. scombrus a total complement of 22 to 25 is at- tained by 32.0 mm. Because anterior pterygio- phores ossify before posterior ones, and the counts differ between the two species, counts of pterygio- phores in the second through fifth or sixth inter- neural spaces serve to identify S.japon/cus by 11.7 mm and in S. scombrus by 18.2 mm (Table 1). The relative position of the first anal pterygiophore and the first haemal spine is first observable in S. japonicus at 11.7 mm and in S. scombrus at 32.0 mm. In S.Japonicus the first anal pterygiophore is anterior to the first haemal spine while in S. scombrus it is posterior. Scomber japonicus larvae are deeper bodied at 4 to 1 1 mm and have greater preanus lengths at 3 to 15 mm than comparably sized S. scombrus larvae. Scomber scombrus larvae are more heavily pigmented and acquire pigmentation earlier than S. japonicus at lengths less than about 15 mm (Figure 4). Of the two species S. scombrus is ear- lier in developing melanophores on the snout and lower jaw. Some specimens of both species possess a few minute melanophores on the ventrum of the abdomen, but their occurrence is more frequent in S.Japonicus larvae <4.2 mm. At given sizes up to 12 mm, where additional dorsal trunk pigmenta- tion is developing in both species, the melanophores are more numerous and larger in S. scombrus than in S.Japonicus. At lengths greater than about 12 mm this character is equally de- veloped in both species. Melanophores are not found at the symphysis of the cleithra in any S. japonicus larvae, but are present in S. scombrus larvae as small as 3.7 mm, then in increasing frequency of occurrence so that all S. scombrus larvae >8 mm possess this pigmentation DISTRIBUTIONS OF EGGS AND LARVAE Scomber scombrus. Egg Distributions During the May cruise, S. scombrus eggs were taken from Martha's Vineyard to below the mouth of Chesapeake Bay and were concentrated from Fire Island, N.Y., to Cape Henry, Va. (Figure 5). Spawning apparently extended northward in the inshore portion of shelf water in an area whose northeastern boundary roughly paralleled the surface isotherms. The egg distribution extended out to at least the edge of the continental shelf off Maryland to North Carolina on transects F, G, J, and K. By the time of the June cruise, spawning of S. scombrus had shifted to the northeast. Eggs were taken only on the three northernmost transects, the majority occurring in the inner half of shelf waters (Figure 6). Scomber scombrus. Larva Distributions During May, S. scombrus larvae were caught between Chespeake Bay and Oregon Inlet, N.C., across the breadth of the continental shelf and south of the area where eggs were taken during this cruise (Figure 7). These larvae were small, ranging from 2.5 to 8. 1 mm long with a mode of 3.0 to 3.9 mm. During the June cruise we took S. scombrus young over a greater area than in May. Larvae occurred from the offing of Martha's Vineyard, which was probably not the northern limit of their distribution, south to the offing of Currituck Beach, N.C. (Figure 8). The distribution of larvae overlapped that of eggs on the three northernmost transects and extended across the entire breadth of the continental shelf between Martha's Vine- yard and New Jersey. The largest numbers oc- curred off Montauk Point, N.Y. Most larvae taken in June were north of the area of larva occurrence in May. A marked increase in lengths of young, progres- sing from north to south, is shown in length- frequency data for this cruise (Figure 9). This in- crease may be due to earlier spawning or higher temperatures to the south which may enable the larvae to grow faster. The inordinately large increase in lengths be- tween transects D and E and decrease in lengths south of transect E may have been caused by the 104 BERRIEN; EGGS AND LARVAE OF SCOMBER ATLANTIC MACKEREL EGGS/STATION time sequence of sampling. We sampled transect E as much as 4 days after transects G, H, and K, and 8 or 9 days after transect D. If we had progressed southward over the whole cruise, the young taken on transect E probably would have been smaller by 8 or 9 mm and intermediate between the lengths of those found on transects D and G, as- suming Sette's (1943) calculated growth rate of about 1.0 mm/day in 20- to 30-mm S. scombrus is correct. During August we took S. scombrus larvae only on the two northernmost transects, off Martha's Vineyard and Montauk Point between about 10 and 90 km offshore. Relatively few larvae were caught, 76 in all. They were small, ranging from 2.6 to 7.7 mm with a mode of 3.0 to 3.9 mm long. Because 1) no S. scombrus eggs were taken on the August cruise and 2) larvae occurred only near the northeastern extreme of sampling at a time when the adults are knowm to be migrating toward the north and east, it follows that these larvae may have resulted from the last spawning within our survey area for 1966. In fact, they may have been spawned northeast of the survey area, for Bumpus and Lauzier (1965) report a southwesterly drift in continental shelf waters off Rhode Island and Long Island, N.Y., in August. Scomber scombrus. Catch Characteristics Statistical tests were run on catch characteris- tics, in order to summarize the data. These tests included: 1 ) comparison of catch sizes by net 1 (0 to 15 m) versus those by net 2 ( 18 to 33 m) for eggs; 2) the same comparison for larvae; 3) comparison of larva lengths taken by net 1 versus net 2 during day; 4) the same comparison during night; and 5) comparison of larva lengths taken during day ver- sus those taken during night. Because the samples were collected by open nets, net 2 catches were corrected for contamination. Results of tests 1 and 2 showed significant dif- ferences in the catch between nets 1 and 2. Net 1 caught 2.3 times as many eggs (chi- square = 1,533.956, P<0.005, with 19 df) and 6.1 times as many larvae (chi-square = 1,360.618, P<0.005, with 26 df) as net 2. The larger catch in the 0- to 15-m (net 1) tow is probably related to the occurrence of most eggs and larvae of iS. scombrus Figure 5. — Distribution of Scomber scombms eggs and selected surface isotherms (°C) during May 1966. 105 FISHERY BULLETIN: VOL 76, NO. 1 Figure 6. — Distribution of Scomber scom- brus eggs and selected surface isotherms (°C) during June 1966. ATLANTIC MACKEREL EGGS/STATION CRUISE D-66-7 JUNE 17-29, 1966 no° a \ X \ \^ w «e" above the thermocline as reported by Sette ( 1943). During Sette's study the thermocline occurred be- tween 17 and 19 m. During this survey, at stations where S. scombrus eggs or larvae were caught, the thermocline was situated so that the surface mixed layer was sampled by net 1 and was rarely deep enough for the surface layer to be sampled by net 2. I tested the two hypotheses that the mean lengths (SL) were equal in catches from net 1 and net 2 during both day and night tows, and found in both cases that the mean lengths were not sig- nificantly different between the paired catches. In another analysis I tested for differences in mean lengths between day and night tows. In this case the pairs tested were adjacent stations either on the same or adjacent transects. The result of the test was not significant, i.e., there was no sig- nificant difference between the means. I used analysis of variance in these tests for differences in mean lengths between the two nets and be- tween light regimes because this procedure segre- gates the known differences in lengths observed over the geographical distribution. 106 BERRIEN: EGGS AND LARVAE OF SCOMBER Figure 7. — Distribution of Scomber scom- brus larvae during May 1966. T" \ X \ KILOMETERS 70 '0 ^^=^^:^^ Scomber scombrus. Relationship of Temperature to Egg and Larva Occurrences Temperature dependence of spawning is suggested by the parallel relationship of the sur- face isotherms and the northeastward edge of the egg abundance contours in May (Figure 5). This temperature dependence is also implied by the June cruise results, i.e., while shelf waters warmed, with consequent northward and east- ward displacement of surface isotherms, the dis- tribution of eggs moved accordingly (Figure 6). While the northern extent of the egg distribution was defined only during the May cruise, the south- ern extent was defined during both the May and June cruises, falling within the 16.0°- to 16.9°C- temperature interval despite the northerly dis- placement of temperatures between the two cruises. Along wdth even higher water tempera- tures prevailing during the August cruise, spawn- ing had ceased entirely within the survey area by that time. Sette (1943) related his egg catches to surface temperature and reported a weighted mean of 10.9°C for all eggs taken in 1932, with 98% occur- ring at 9.0°C to 13.5°C. During the May cruise of our survey, similar surface temperatures were as- sociated with the eggs. The weighted mean surface temperatures for all eggs taken during May was 11.0°C, with 97% at 8.7°to 13.8°C and the temper- ature associated with all eggs in May ranged from 6.3° to 16.9°C. 107 FISHERY BULLETIN: VOL. 76, NO. 1 .^ \jf !> ,., iiiilh. ";>^ l/:4iiliiliiilli Mi iiHllliiiliiilnliiliiiliyilP' :::::i::ll:::::::l!llil::v'' . ■•■ Figure 13. — Distribution of Scomber japonicus larvae during May 1967. 113 during this survey than in other studies on this species in the western North Pacific Ocean by Uchida et al. (1958), Dekhnik (1959), and Watanabe ( 1970) and in the eastern North Pacific Ocean by Fry (1936b). Although there was some variation between these studies, all reported sur- face temperatures within the range of 15° to 21°C associated with spawning or with the majority of eggs caught. Scomber scombrus population estimates of 18 and 17 million spawners, based on our May and June 1966 cruises, respectively, were reported by Berrien and Anderson. "* As discussed by the au- thors, these point estimates, calculated from egg catches, probably understated the true population size due to cruise timing and the area sampled. Apparently the May cruise occurred prior to peak spawning intensity resulting in many spawners being unaccounted for in the point estimate. Dur- ing June, although the egg density was greater than in May, only a portion of the egg population was surveyed; therefore, the population was in- completely sampled. Other plankton survey efforts within the Mid- Atlantic Bight have resulted in higher and proba- bly more accurate, S. scombrus spawning popula- tion estimates. Sette (1943) reported a season- long, Mid-Atlantic Bight spawning population of 320 million spawners in 1932. Berrien and Ander- son (see footnote 4) reported a point estimate of 392 million spawners within the New York Bight during May 1975. ACKNOWLEGMENTS I thank L.A. Walford for his review of an early version of this paper; the editors at Sandy Hook Laboratory for their review; Alyce Wells for prep- aration of the graphs and charts; W.J. Richards and T. Potthoff for their critical review of the de- scriptive section; the technicians at Sandy Hook Laboratory for sorting specimens; the boat crew, technicians, and project biologists for their assis- tance in obtaining the samples aboard the RV Dolphin. ■•Berrien, P. L., and E. D. Anderson. 1976. Scomber scom- brus spawning stock estimates in ICN AF Subarea 5 and Statisti- cal Area 6, based on egg catches during 1966, 1975 and 1976. ICNAF (Int. Comm. Northwest Atl. Fish.) Res. Doc. 76/ XII/140, 10 p. FISHERY BULLETIN: VOL. 76, NO. 1 LITERATURE CITED ANDERSON, W. W., AND J. W. GEHRINGER. 1958. Physical oceanographic, biological, and chemical data — South Atlantic coast of the United States. MA^ Theodore N. Gill Cruise 5. U.S. Fish. Wildl. Serv., Spec. Sci. Rep. Fish 248, 220 p. BERRIEN, P. L. 1975. A description of Atlantic mackerel, Scomber scom- brus, eggs and early larvae. Fish. Bull, U.S. 73:186-192. BIGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish. Wildl. Serv., Fish. Bull. 53, 577 p. Briggs, J. C. 1958. A list of Florida fishes and their distribution. Bull. Fla. State Mus., Biol. Sci. 2:223-318. BUMPUS, D. F., AND L. M. LAUZIER. 1965. Surface circulation on the continental shelf off east- em North America between Newfoundland and Flori- da. Am. Geogr. Soc, Ser. Atlas Mar. Environ., Folio 7, 4 p., 8 pi. Clark, J,, w. G. Smith, A. W. Kendall, JR., and M. P. Fahay. 1969. Studies of estuarine dependence of Atlantic coastal fishes. Data Report I: Northern section, Cape Cod to Cape Lookout. R.V. Dolphin cruises 1965-66: Zooplankton vol- umes, midwater trawl collections, temperatures and salinities. U.S. Fish Wildl. Serv., Tech. Pap. 28, 132 p. 1970. Studies on estuarine dep)endence on Atlantic coastal fishes. Data Repwrt II: Southern section. New River Inlet, N.C. to Palm Beach, Fla. R.V. Dolphin cruises 1967-68: Zooplankton volumes, surface-meter net collections, temperatures, and salinities. U.S. Fish Wildl. Serv., Tech. Pap. 59, 97 p. Clothier, C. R. 1950. A key to some southern California fishes based on vertebral characters. Calif. Dep. Fish Game, Fish Bull. 79, 83 p. Dannevig, a. 1919. Biology of Atlantic waters of Canada. Canadian fish-eggs and larvae. In Canadian Fisheries Expedition, 1914-15, p. 1-74. Dep. Nav. Serv., Ottawa. Dekhnik, T. V, 1959. Reproduction and development oi Pneumatophorus japonicus (Houttuyn) off the coast of southern Sakha- lin. [In Russ.] Akad Nauk SSSR, Zool. Inst., Issled Dal'nevost. Morei SSSR 6:97-108 (Engl, transl. by M. Slesser, U.S. Nav. Oceanogr. Office, Transl. 307, 15 p., 1967). DOOLEY, J. K. 1972. Fishes associated with the pelagic sargassum com- plex, with a discussion of the sargassum communi- ty. Contrib. Mar. Sci. 16:1-32. EHRENBAUM, E. 1924. A. 11. Scombriformes. In Report on the Danish oceanographical expeditions 1908-10 to the Mediterra- nean and adjacent seas. Vol. 2 (8-9), Biology, 1-42 p. H^st and S^n, Copenh. FAHAY, M. p. 1974. Occurrence of silver hake, Merluccius bilinearis, eggs and larvae along the middle Atlantic continental shelf during 1966. Fish. Bull, U.S. 72:813-834. 114 BERRIEN: EGGS AND LARVAE OF SCOMBER 1975. An annotated list of larval and juvenile fishes cap- tured with surface- towed meter net in the South Atlantic Bight during four RV Dolphin cruises between May 1967 and February 1968. U.S. Dep. Commer., NOAA Tech. Rep. NMFS, SSRF 685, 39 p. FRY, D. H., Jr. 1936a. A description of the eggs and larvae of the Pacific mackerel [Pneumatophorus diego). Calif. Fish. Game 22:28-29. 1936b. A preliminary summary of the life history of the Pacific mackerel iPneumatophorus diego). Calif. Fish Game 22:30-39. HILDEBRAND, S. F., AND W. C. SCHROEDER. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish 43(1), 366 p. (Doc. 1024.) KRAMER, D. 1960. Development of eggs and larvae of Pacific mackerel and distribution and abundance of larvae 1952-56. U.S. Fish Wildl. Serv., Fish. Bull 60:393-438. Leim, a. H., and W. B. Scott. 1966. Fishes of the Atlantic coast of Canada. Fish. Res. Board Can., Bull. 155, 485 p. Matsui, T. 1967. Review of the mackerel genera Scomber and Ras- trelliger with description of a new species of Rastrelliger . Copeia 1967:71-83. Orton, G. L. 1953. Development and migration of pigment cells in some teleost fishes. J. Morphol. 93:69-99. Padoa, E. 1956. Divisione: Scombriformes. Famiglia 1: Scom- bridae. In Fauna e flora del Golfo di Napoli, Monografia 38. Uova, larve e stadi giovanili di Teleostei, p. 471-478. (Engl, transl. by J. P. Wise and G. M. Ranallo. Transl. 12, U.S. Bur. Commer. Fish., Trop. Atl. Biol. Lab., Miami, Fla.) Sette, O. E. 1943. Biology of the Atlantic mackerel (Scomber scom- brus) of North America. Part I: Early life history, includ- ing growth, drift and mortality of the egg and larval popu- lations. U.S. Fish Wildl. Serv., Fish. Bull. 50:149-237. UCHIDA, K., S. IMAI, S. MITO, S. FUFITA, M. UENO, Y. Shofima, T. Senta, M. Tahuku, and Y. Dotu. 1958. Studies on the eggs, larvae and juveniles of Japanese fishes. Series I. [In Jap. ] Kyushu Univ., Fac. Agric, Fish Dep., 2d Lab. Fish. Biol., Fukuoka, Jap. 89 p. (Engl, transl. by W. G. van Campen, 300 p.) Watanabe, T. 1970. Morphology and ecology of early stages of life in Japanese common mackerel. Scomber japonicus Hout- tuyn, with special reference to fluctuation of popula- tion. Bull. Tokai Reg. Fish. Res. Lab. 62, 283 p. 115 DAILY AND SUMMER- WINTER VARIATION IN MASS SPAWNING OF THE STRIPED PARROTFISH, SCARUS CROICENSIS Patrick L. Colin ^ ABSTRACT The "striped" phase of the striped parrotfish, Scarus croicensis, engaged in mass spawning during afternoon periods on a deep (24 m) coral pinnacle off Discovery Bay, Jamaica. During morning periods the fish occurred in a large foraging group on shallow reefs and moved to the spawning site in etirly afternoon. The occurrence of spawning rushes per day in June was about six times that during January. Chromis cyanea and Clepticus parrai fed on freshly released eggs of S. croicensis. Mass spawning by S. croicensis was similar to that of Sparisoma rubripinne. The striped parrotfish, Scarus croicensis Bloch (Figure 1), is the smallest (reaching 25 cm SL, standard length) but the most common member of this genus in the tropical western North Atlantic (Randall 1968; Bohlke and Chaplin 1968). Like other scarids, S. croicensis is a benthic herbivore grazing on algal-covered rock and coral surfaces and is seldom found at depths below 30 m. The species possesses dimorphic color phases, termed the "striped" phase (male and female) and the "terminal" phase (male only), believed derived from striped phase females by protogynous sex reversal (Ogden and Buckman 1973). Aspects of the general biology of this fish have been reported on by several authors. Ogden and Buckman (1973) followed movements of tagged individuals in Panama and found daily migrations between feeding and sleeping areas. Feeding was largely carried out in foraging groups of up to 500 individuals with a characteristic set of associate, but less numerous species. Buckman and Ogden (1973) described territoriality by striped phase females and terminal phase males. Barlow (1975) discussed the sociobiology of S. croicensis in com- parison with three other species of parrotfishes and described their feeding pattern, group sizes, density, and color variation. He also added some notes on spawning behavior of S. croicensis. Randall (1963) reported both mass spawning by the striped phase of S. croicensis and pair spawn- ing by terminal phase males and striped phase females. Randall and Randall (1963) described pair spawning at St. John, V. I., during February, 'Depsirtment of Marine Sciences, University of Puerto Rico, Mayaguez, PR 00708. Manuscript accepted May 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. Figure l. — Striped parrotfish, Scarus croicensis, with "con- trast" color pattern approximately 100 mm standard length at Discovery Bay, Jamaica. March, April, June, and August, and with their limited observations they felt that pair spawning accounted for most of the reproduction of the species. Buckman and Ogden (1973) commonly observed pair spawning at depths of 9-13 m, but also as shallow as 3 m, in Panama. Munro et al. (1973) found females of the striped parrotfish in ripe condition from March to May near Jamaica. In August 1971 a large spawning group of striped parrotfish was encountered on a deep coral platform (24 m) offshore from the Discovery Bay Marine Laboratory on the north coast of Jamaica. This species is by far the most common parrotfish along this coast, which is heavily fished using An- tillean fish pots. This spawning group consisted of several hundred individuals. Its reproductive ac- tivity was sufficiently regular and observable that investigation of diel patterning of spawning seemed feasible. Widely scattered observations from 1971 to 1975 indicated the continued pre- sence of this group. During January and June 117 FISHERY BULLETIN: VOL 76, NO. 1 1975 systematic observations of spawning be- havior were conducted. MATERIALS AND METHODS For purposes of determining diel variation of spawning activity, the daylight period (from sun- rise to sunset) was divided into 16 equal periods. As occasional checks during the morning indi- cated that the spawning population was not pre- sent at the spavwiing site and was not spawning elsewhere, only the latter eight periods of the day were included in this study. Since day length var- ied considerably between January and June ob- servations, the length of each period also varied by the same factor. The change of day length during each of the two series of observations was only a few minutes. During the winter observations (12-28 Janu- ary), the day length was 11 h 10 min with 42 min for each period. During summer ( 19-29 June), the day length was 13 h 10 min with 50 min for each period, an increase of 17% in day length. Water temperature at the study site varied between 26° and 29°C seasonally. The number of spawning rushes, the upward dash by groups of parrotfish culminating in the release of eggs and sperm, occurring during 15 min within the observation period was counted by an observer (wearing scuba equipment). This time was chosen as the minimum for measurements of spawning rush frequency due to the somewhat irregular occurrence of the rushes on a minute by minute basis. In the latter portion of the study, data were recorded minute by minute for the full 15-min period. The observers were tethered near the spawning site by lines attached to the bottom which caused them to float nearly motionless at 21 m depth, approximately 3 m above the substrate. This allowed observations to be made from a con- sistent location, minimized movement needed to stay in position, and decreased the depth of the observers slightly to allow more bottom time for observations with no or short decompression at the end of the dive. The presence of the observers did not seem to interrupt or affect the spawning be- havior as the population did not move away or cease spawning after the observers' arrival. Color motion pictures (16 mm) were made of spawning and feeding behavior of S. croicensis, including some at two times normal film speed for slow motion analysis of movement. Films were analyzed on a frame by frame basis. GENERAL BEHAVIOR A general profile of the area near the spawning site is presented in Figure 2. Sand channels run between fingers of reef directed seaward which gradually slope from a shallow reef crest to a zone dominated by the branched coral 'Acropora cer- vicornis at 10-13 m depth. At the seaward edge of this A. cervicornis zone, the reef slopes steeply to a sandy bottom at 24-25 m depth. Beyond this point the sand bottom either slopes rapidly downward to the near vertical dropoff or has an outer reef rising above it, often resembling a rounded pinnacle and somewhat trapping the sediment behind it. The pinnacle of "Dancing Lady Reef" was the location of the spawning observed in this study. On the 0 10 20 30 10 .>0 0 100 200 300 ACROPORA CKRVKORMS ZONK BOTTOM OF SAND (HANNKLS Figure 2. — Bottom profile of "Dancing Lady Reef" offshore from the Discovery Bay Marine Laboratory. Vertical exaggeration is2x. 118 COLIN: VARIATION IN MASS SPAWNING OF STRIPED PARROTFISH outer face of this pinnacle, the reef drops away steeply and at a depth of 50-70 m becomes nearly vertical in profile. In Jamaica S. croicensis occurred in foraging groups similar to those described by Ogden and Buckman ( 1973) in Panama. In the vicinity of the spawning site only one sizeable foraging group occurred. Although no tagging experiments were carried out, this group almost surely constituted the major portion of the spawning population studied. During morning hours this group ranged as much as 300 m inshore from the spawning area onto the shallow reefs to depths of as little as 7 m. They also ranged only about 100 m in either direc- tion parallel to shore along the reef. These foraging groups consisted of several hundred S. croicensis (the exact number being impossible to determine in most cases) plus a few other fishes. In one instance at least 410 individu- als of S. croicensis were visible in photos taken of the entire group. Only a few terminal phase males were seen in these groups. The group swam about 1 m above the substrate in the A. cervicornis zone and descended en masse at intervals to feed. Algae were scraped from rock surfaces of the reef, par- ticularly from the dead lower portions of the branches of A. cervicornis. Mixed foraging groups consisting largely of S. croicensis have been reported by Buckman and Ogden (1973) and Itzkowitz (1974). In the former two species of acanthurids (Acanthiirus chirurgus dLXidi A . coeruleus); aham\et,Hypoplectruspuella; a goatfish; and a few other parrotfishes were typi- cally found associated with the foraging groups. Similar composition of associated species was ob- served in the present study. Only A. coeruleus among the surgeonfishes occurred with the forag- ing group. However, A. chirurgus is relatively rare in the study area. A different species of ham- let, H. indigo, also occurred with the foraging group rather than H. puella. Among fishes ob- served occasionally joining foraging groups and not mentioned by Buckman and Ogden ( 1973 ) was Halichoeres maculipinna. The functionality of such schooling behavior has been commented on before. Various Indo-Pacific surgeonfishes form schooling groups which be- have much like the foraging groups of S. croicensis (Jones 1968; Randall 1970; Barlow 1974). Randall (1970), Barlow (1974), and Vine (1974) believed this foraging herd was a method for the sur- geonfishes to swamp the defenses of territorial food competitors, in the former instance an acan- FlGURE 3. — Striped phase individuals of Scarus croicensis at Discovery Bay, Jamaica, in the "contrast" color form (A) and "gray" form (B). Standard length is approximately 100 mm. thurid and in the latter a pomacentrid. This also seems to be the case in the present study. When the foraging group entered the territory of Eupo- macentrus planifrons, attacks were quickly di- rected at a few members causing an escape reac- tion in the few individuals near the center of attack. The group was largely undisturbed by the actions of the damselfish. Two color forms of striped phase S. croicensis were seen in both foraging and spawning groups. The first had two broad dark stripes separated by thinner pale stripes, the dorsal surface dark and the snout yellowish. This form is termed the "con- trast" (Figures 1,3). The second color form, termed the "gray" form does not have the sharp contrast between dark and pale stripes (Figure 3). The stripes are apparent on the head, but posteriorly they become much less distinct. The scales near the caudal peduncle, even in the center of the dark stripe, are pale-edged and resemble a checker- board pattern. In foraging groups one-fourth to one-half of the indivuals had the gray color pat- tern and the remainder were of the contrast pat- tern. No functional role could be assigned to these color forms. The possibility does exist that they represent male and female, but this could not be established. MASS SPAWNING BEHAVIOR Spawning occurred on the deep coral pinnacle (Figure 2) of Dancing Lady Reef at 24 m depth. This pinnacle is the feature with the greatest re- 119 FISHERY BULLETIN: VOL. 76, NO. 1 lief for a distance of several hundred meters along the outer face of the reef. Transects were swum along the sloping face for 200-300 m each direction from the study area while spawning was under- way at that site and no other spawning aggrega- tions were encountered. In one instance a group of several S. croicensis were observed spawning on the seaward face of a shallow reef immediately west of Dancing Lady Reef at 18 m depth. The spawning population did not arrive en masse at the spawning site, but rather appeared in small groups over a lengthy period of time. Whether the foraging group breaks up on the shal- low reef before the individuals move to the spawn- ing site is not known. The behavior of the striped parrotfish after arrival at the spawning area con- sists of swimming in small groups around the area within a few meters of the bottom ("milling") and bouts of feeding (from the substrate). The size of eight individuals speared from the spawning aggregations varied between 80 and 100 mm SL, relatively small for mature specimens. These are deposited in the University of Puerto Rico fish collection (UPR 3452). This sample is biased for small individuals since these were most easily approached and the mean size of specimens in the aggregation was certainly near or over 100 mm SL. The numbers engaged in milling and the speed and frequency of turns gradually increased. Often groups of 20 or more individuals broke away from the main group and swam as a school farther above the substrate than the milling individuals (Figure 4A, B). The separated group swam in- creasingly rapidly making abrupt lateral turns ("weaving"). The entire group or a portion of it rushed upward extremely rapidly a distance of several meters (Figure 4C, D) releasing eggs and sperm at the peak of the "rush." They returned to the substrate nearly as rapidly (Figure 4E). Be- cause of the large numbers of individuals present in the spawning aggregation, several separate weaving groups could be present and rush at near the same time. Rushes by some weaving groups began at the level of at least 3 m above the sub- strate as they were level with the observers' line of sight. From analysis of motion pictures of spawning behavior the number of fish engaged in a rush varied between 5 and 30 with the mean number about 15 individuals. Generally only about one- half of the group engaged in weaving actually participated in the rush and often a few individu- als starting the upward rush were left behind. The entire upward rush and return to the level of the weaving group took <1 s. Of seven rushes which were filmed in their entirety the time for the up- ward movement varied between 0.21 and 0.40 s and for the return 0.20 and 0.40 s. One rush with return occupied only 0.45 s total. Assuming a dis- tance of 3 m was covered during the upward rush (probably a conservative estimate), the average speed from leaving the weaving group until turn- ing at the point where the gametes are released was around 40 km/h. The sexual composition of the rushing groups has not been determined. Randall and Randall (1963) believed that the spawning groups of Sparisoma rubripinne were predominantly males and that a single female participated in the spawning rush with 3 to 12 males. A single terminal phase male Scarus croicensis was present at the spawning site. This fish vigor- ously defended a territory near the outer edge of the coral pinnacle and patrolled the area in the "bob-swim" manner with the caudal fin upturned as described by Barlow ( 1975). No attempts at pair spawning with striped phase females by this fish were observed. The only other parrotfish observed on the deep coral pinnacle was Sparisoma viride with only a few present. SPAWNING FREQUENCY The frequency of rushes during the daily periods for both January and June is presented in Figure 5. The summer spawning begins earlier in the day, continues later, and has a higher frequency of rushes than during the winter. It is impossible to determine the number of eggs released per rush and whether differences exist between summer and winter. No data are available concerning the number of fish participating in rushes during winter, but observations suggest this was also lower. Considering an equal number of eggs are expel- led on each rush, it appears that the production of eggs by this population of Scarus croicensis is about six times greater during a summer day than winter on the basis of the area beneath the curves derived from Figure 5. It is likely that S. croicen- sis, at least in the Caribbean, spawns year round, but the warm months are the most important period of egg production. During the summer the occurrence of spawning rushes might be referred to as epidemic. When the 120 COLIN: VARIATION IN MASS SPAWNING OF STRIPED PARROTFISH D Figure 4. — Spawning sequence of an aggregation o{ Scarus croicensis at Discovery Bay, Jamaica. A. A "weaving" group above a larger "milling" aggregation. B. The weaving group becomes tighter and makes more rapid turns. A few fish are joining the group as it moves toward a spawning "rush." C. A small group carries out a spawning rush (upper right) while a second, larger group engages in weaving behavior (left side). Part of the main aggregation is visible at the bottom of the photography. D. Rushing (center) and weaving (left) groups of S. croicensis. E. Return of group from a rush (upper left) to the aggreagation engaged in milling. 121 FISHERY BULLETIN: VOL. 76, NO. 1 I ! L <> 7 h 8 I w IN I KK ri-:HM»i)N I'KKIuM^ Figure 5. — Daily variation of spawning rushes during June (summer) and January (winter) by a population of striped parrotfish at Discovery Bay, Jamaica. The beginning of period 1 represents midpoint of the day and the end of period 8 sunset. Figures represent mean of two observations (periods 1-3) and four observations (periods 4-8). data are analyzed on a minute by minute basis, over 90*^ of the spawning rushes observed occur- red during only 33% of the 1-min periods. Since the group engaged in a spawning rush is consider- ably smaller than the total population at the spawning area, it is possible for several groups to carry out a spawning rush separately, but nearly simultaneously. The occurrence of the first rush by a group seems to trigger other groups to spawn. A flurry of rushes lasted a period of 1-4 min and in one case reached a frequency of 35 rushes in a 1-min period. This number may be underesti- mated due to the difficulty in observing and count- ing such rapid events. The period between groups of rushes was spent in milling about close to the substrate and feeding on exposed rock surface of the reef. The time between episodes of epidemic rushing varied during the day in summer periods. During early periods when some spawning occurred (period 3 and to a lesser extent period 4) often 5-7 min would elapse without any rushes occurring. In one case there was 9 min between rushes. Later in the day, at times of peak spawning (periods 5-7), these nonspawning periods were reduced to 1,2, and occasionally 3 min. PREDATION Mackerel (either cero, Scomberomorus regalis, or king mackerel, S. caualla) twice attempted to prey on Scarus croicensis at the top of the spawn- ing rush, once apparently successfully. These at- tacks interrupted the spawning behavior of the entire group. In one case only 1 rush occurred in the 10 min following the attack even though 67 rushes had occurred in the previous 15 min. On a third occasion, a lizardfish, Synodus sp., rushed upward from the substrate in an unsuccessful at- tempt to prey on Scarus croicensis and thus inter- rupted spawning for a short period. Chromis cynaeus and Clepticus parrai were ob- served to feed actively on the freshly released eggs of S. croicensis. Within 5-10 s after completion of the spawning rush, numerous Chromis cyaneus converged on the area of egg release, followed shortly by a lesser number of Clepticus parrai, and while remaining in a tightly bunched group ap- parently picked individual eggs from the water. It was estimated that as many as 200 Chromis cyaneus and 20-30 Clepticus parrai composed one group picking eggs released in a single spawning rush. The group remained tightly bunched and fed 122 COLIN: VARIATION IN MASS SPAWNING OF STRIPED PARROTFISH for about 1 min, moved slowly with the current (and presumably with the eggs), and dispersed quickly returning as individuals to a position closer to the substrate. Whether dispersion of the released eggs, depletion of the eggs by feeding, or some other factor caused cessation of the feeding by Chromis cyaneus and Clepticus parrai is not known. A few hundred predators, each ingesting at least one egg every few seconds for periods of nearly 1 min, could eliminate a significant portion of the eggs released in any given spawning rush. These groups of egg predators form after only a small percentage of spawning rushes. During the "epidemic" rushes of summer periods, there are too many eggs released at several locations for these predators to significantly deplete the number released. During winter periods when rushes were few, there did not seem to be sufficient gamete release for the egg predators to wait for rushes to occur and consequently no predation on eggs was observed during these periods. The pre- dation on newly released eggs of S. croicensis is obviously an intentional activity of the predators, not a chance occurrence, but probably serves only as a "bonus" for these fishes which normally spend lengthy portions of the day feeding on particulate zooplankton in the water column (Davis and Birdsong 1973). DISCUSSION The mass spawning behavior of S. croicensis is similar to that described for Sparisoma rubripinne by Randall and Randall ( 1963). The movement of the population to the deep-reef area in the early afternoon, its behavior before and during rushes, the epidemic rushes, and other behavior is nearly identical. This similarity in mass spawning be- tween genera lines in parrotfishes is interesting. It would be most informative to know the num- bers needed before both foraging aggregations and striped phase spawning aggregations occur. Small groups of 15-20 Scarus croicensis have been seen moving together between bouts of feeding, but seem easily deterred by damselfishes defending territories. At least on the north coast of Jamaica, mass spawning probably contributes most of the eggs produced by S. croicensis. Pair spawning was never observed in the vicinity of Discovery Bay although terminal phase males were present but never abundant. The summer season is certainly the most active reproductive period. The occurrence of mass spawning by parrot- fishes at specific locations on the reef is a relatively long-term phenomena. In the present case nearly 4 yr have elapsed since the initial encounter with the spawning group and the location of spawning has not varied. More interestingly, the spawning location of Sparisoma rubripinne at Reef Bay, St. John, investigated by Randall and Randall (1963), was visited in March 1977. Following the direc- tions provided by those authors, a group of approx- imately 200 S. rubripinne were found engaged in spawning during the late afternoon. The presence of a spawning aggregation in what is be- lieved the identical location on the reef after 17 yr in similar numbers to that previously reported indicates a stability and importance of spawning locations not previously documented. The occur- rence of spawning by S. rubripinne on 3-4 March extends the period reported by Randall and Ran- dall ( 1963) and supports their belief in year round spawning. Also the water temperature of 25.8°C was slightly lower than that previously reported. The reasons for the abundance of Scarus croicensis compared with some other scarids (such as Sparisoma rubripinne) are difficult to deter- mine. Randall (1967) reported three species of fishes {Mycteroperca interstitialis, M. venenosa, and Caranx ruber) which definitely preyed on Scarus croicensis; however, individals of Scarus (not identifiable to species) were found in guts of several other predatory fishes. Ogden and Buckman (1973) added Epinephelus striatus and Scomberomorus regalis as predators of Scarus croicensis. Due to overfishing, few large predatory fishes are found on the outer reef at Discovery Bay. Indeed, few of the larger species of Scarus and Sparisoma occur there for the same reason. This may be an important factor allowing relatively high numbers of Scarus croicensis to occur there and schooling behavior to be effective in over- whelming the defenses of territorial herbivores. Alevizon and Brooks (1975), in examining two coral-reef fish assemblages (Islas Las Aves, Venez. and Key Largo, Fla. ), found S. croicensis to be only a minor component of one (Florida) and of no con- sequence at the other (Venezuela). Possibly they sampled areas where S. croicensis was not abun- dant. In other areas S. croicensis may be absent, even though the environment seems typical of that in which it normally occurs. At Isla Desecheo, a small ( 1 km^) island 20 km west of Puerto Rico in the Mona Channel, extensive diving operations failed to reveal the presence of S. croicensis even 123 FISHERY BULLETIN; VOL. 76, NO. 1 though we have specifically searched for it. Other scarids occur there, and there seems no simple reason for the nonoccurrence of S. croicensis at this island. ACKNOWLEDGMENTS Deborah W. Arneson assisted in all the field observations and commented on the manuscript. The staff of the Discovery Bay Marine Laboratory, particularly Eileen Graham, made this project possible. John C. Ogden and Ileana Clavijo are thanked for commenting on the manuscript. Evangelina Hernandez prepared Figures 2 and 5. Observations at Isla Desecheo were carried out from the RV Corallina. Those and the observa- tions at St. John were made possible by a grant from the Oceanography Section, National Science Foundation (NSF Grant OCE76-02352), Patrick L. Colin, Principal Investigator. LITERATURE CITED ALEVIZON, W. S., AND M. G. BROOKS. ^ ~^ 1975. The comparative structure of two western Atlantic reef-fish assemblages. Bull. Mar. Sci. 25:482-490. Barlow, G. W. 1974. Extraspecific imposition of social grouping among surgeonfishes (Pisces: Acanthuridae). J. Zool. (Lond.) 174:333-340. 1975. On the sociobiology of four Puerto Rican parrotfishes (Scaridae). Mar. Biol. (Berl.) 33:281-293. BOHLKE, J. E., AND C. C. G. CHAPLIN. 1968. Fishes of the Bahamas and adjacent tropical waters. Livingston Press, Wynnewood, Pa., 771 p. BUCKMAN, N. S., AND J. C. OGDEN. 1973. Territorial behavior of the striped parrotfish Scarus croicensis Bloch (Scaridae). Ecology 54:1377-1382. DAVIS, W. P., AND R. S. BIRDSONG. 1973. Coral reef fishes which forage in the water col- umn. Helgol. wiss. Meeresunters. 24:292-306. ITZKOWITZ, M. 1974. A behavioural reconnaissance of some Jamaican reef fishes. J. Linn. See. Zool. 55:87-118. JONES, R. S. 1968. Ecological relationships in Hawaiian and Johnston Island Acanthuridae (Surgeonfishes). Micronesica 4:309-361. MUNRO, J. L., V. C. Gaut, R. Thompson, and P. H. Reeson. 1973. The spawning seasons of Caribbean reef fishes. J. Fish Biol. 5:69-84. Ogden, J. C, and N. S. Buckman. 1973. Movements, foraging groups, and diurnal migra- tions of the striped parrotfish Scarus croicensis Bloch (Scaridae). Ecology 54:589-596. RANDALL, J. E. 1963. Notes on the systematics of parrotfishes (Scaridae), with emphasis on sexual dichromatism. Copeia 1963:225-237. 1967. Food habits of reef fishes of the West Indies. Stud. Trop, Oceanogr. (Miami) 5:655-847. 1968. Caribbean reef fishes. T.F.H. Publ., Neptune, N.J., 318 p. 1970. Easter Island: an ichthyological expedition. Oceans 3:48-59. Randall, J. E., and H. A. Randall. 1963 . The spawning and early development of the Atlantic parrotfish, SP<^''isoma rubripinne, with notes on other scarid and labrid fishes. Zoologica (N.Y.) 48:49-60. VINE, P. J. 1 974 . Effects of algal grazing and aggressive behaviour of the fishes Pomacentrus lividus and Acanthrus sohal on coral-reef ecology. Mar. Biol. (Berl.) 24:131-136. 124 FEEDING BEHAVIOR AND MAJOR PREY SPECIES OF THE SEA OTTER, ENHYDRA LUTRIS, IN MONTAGUE STRAIT, PRINCE WILLIAM SOUND, ALASKA Donald G. Calkins' ABSTRACT Food habits and feeding behavior of sea otters were studied in Prince William Sound, Alaska, from May through August 1971. Otters fed primarily on clams, crabs, and sea stars: Saxidomus gigantea, Telmessus cheiragonus, and Evasterias troschelii, respectively, were the most important prey species identified in the major groups. Mean times for feeding dives were 67 s for females (mean water depth = 9.6 m) and 59 s for males (mean water depth = 11.9 m). Clams were dug from the bottom and opened with the aid of stones. Sea urchins and fishes were not identified as dietary components. The sea otter, Enhydra lutris, hunted to near ex- tinction by 191 1 in Alaska, is steadily reoccupying its former range. Several areas are being repopu- lated naturally (Kenyon 1969), while others have been restocked with otters translocated from Am- chitka Island in the Aleutians or from south cen- tral Alaska (Burris and McKnight 1973). In some areas of the Aleutian Islands, sea otters have be- come so abundant that an experimental harvest has been conducted by the Alaska Department of Fish and Game. Populations in Prince William Sound have become large enough to permit cap- ture of a small number of animals for restocking areas of former abundance. Large gaps still exist in our knowledge of the biology and life history of the sea otter. Past studies have dealt primarily with populations along the California coast and off Amchitka Is- land. No intensive study of sea otters in Prince William Sound has been completed, and the only available information from that area concerns re- stocking activities and population counts (Pitcher and Vania^). The lack of information on the biol- ogy of the sea otter in Prince William Sound and the impending development of oil reserves along the Alaska coast motivated this study. STUDY AREA This investigation took place in Montague Strait, Prince William Sound, Alaska (Figure 1). ^Alaska Department of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502, ^Pitcher, K. W., and J. S. Vania. 1973. Distribution and abun- dance of sea otters, sea lions and harbor seals in Prince William One week was spent in the field in September 1970. In May 1971, a camp was established at the northwestern end of Montague Island (lat. 60°15'54"N, long. 147°12'18"W). Observations were made from May through August 1971. The study area included the northwestern end of Mon- tague Island, from Stockdale Harbor to a logging camp 19 km southwest. Green Island, Little Green Island, and the adjacent waters were also included (see Figure 1). The area was selected as a location where sea otter populations have always existed. Although the population is still expanding, there has always been some sea otters in this area (Karl Schneider, Alaska Department of Fish and Game, pers. com- mun.). The area is characterized by a rugged coastline with rocky shores. Two sand beaches occur in the area, one south of Port Chalmers and one on the south side of Green Island. Several streams empty into the Sound from Montague Island: mud flats and small estuaries are common. The mud flats support stands of eel grass, Zostera sp., and pro- vide habitats for populations of clams — Macoma spp., Saxidomus gigantea, and Protothaca staminea. Approximately 55 km of coastline was included in the study area. Kenyon (1969:57) stated that "generally sea otters favor waters adjacent to rocky coasts near points of land" and that "coasts adjacent to extensive areas of underwater reefs are particularly attractive." Using these criteria, Manuscript accepted June 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. Sound. Unpubl. manuscr., 18 p. Available Alaska Department of Fish and Game, Anchorage, Alaska. 125 y PRINCE V^ «• i 5^ W I L L I A GULF OF ALASKA Figure l. — Montague Strait sea otter study area located in Prince William Sound, Alaska. at least 50 km of the coast within the study area seemed suitable for sea otters. The animals did not frequent the areas with sandy beaches or shallow estuaries. Feeding habits were studied at three main loca- tions at Montague Island: a small lagoon (Ook- shilk Lagoon, see de Laguna 1956) on the south side of Stockdale Harbor, the area outside Ook- shilk Lagoon to the north and west, and Port Chalmers south of Stockdale Harbor. Ookshilk Lagoon had water depths from 5 to 7 m and rock and mud beaches grading to subtidal sand which supported stands of eel grass, Zostera sp., and rockweed, Fiicus sp. The area outside Ookshilk Lagoon was characterized by water depths of 5 to 16 m, rock beaches and sand with reef shoals sub- tidally, and Fucus sp. beach and subtidal flora. Port Chalmers had water depths of 14 to 26 m with rock beaches and subtidal sand with reefs and shoals. Beach and subtidal flora in the Port Chal- mers area consisted of Fucus sp. and kelp, Nereocystis lutkeana. FISHERY BULLETIN: VOL. 76, NO. 1 METHODS All observations on feeding habits were made from advantageous locations on land. Spotting telescopes with magnification of 15 to 60 x were used to identify food organisms. Observation dis- tances ranged from 20 to 500 m. The dimensions of the organisms were estimated relative to the ot- ters paws, which were estimated to average 4 cm wide. Dimensions of octopuses were estimated across the tips of the tentacles, relative to the otter's body, and all sizes are reported in this man- ner. No identification of organisms was attempted beyond 100 m, but it was often possible to classify food items by categories such as clam, crab, sea star, etc., up to 500 m away. Dive and surface feeding times for a total of 14 feeding periods were measured with stopwatches. Timing of feeding periods began when other activities ceased and the otter dived for food and ended when the last bit of food was eaten and some other activity began. Prey species were collected at low tide, and taken to the University of Alaska for identifica- tion. Clams were collected on a gravel beach in Ookshilk Lagoon where otters fed. Work was confined to 1 h before until 1 h after low tide ( —0.86 m). Ten transects were dug 25 m apart with each transect running from the extreme high-tide mark to the water's edge. Sample holes of approximately 0.25 m^ were dug at 5-m intervals along each transect. Sample holes were dug to a depth of 25 cm. In areas where extensive observations were made, water depths were measured using a weighted line graduated at 25-cm intervals. RESULTS Types of Organisms Eaten All food organisms were bottom-dwelling in- vertebrates from three major groups of organisms: molluscs, crustaceans, and echinoderms. The per- centage occurrence of prey organisms in the diet is shown in Table 1. Five species of clams are found in this area (Table 1), and all were eaten. Empty shells and observations of feeding otters suggest that Saxidomus gigantea is the clam most com- monly eaten by otters. Several species were present in the area but never observed to be eaten by otters (Table 1). Each had been previously identified as food of sea otters (Barabash-Nikiforov 1947; Kenyon 1969). 126 CALKINS: FEEDING BEHAVIOR OF ENHYDRA LUTRIS Table l.— Bottom-dwelling invertebrates of Montague Strait, Alaska, and the percent of occurrence in the diet of sea otters. No.of Percent of times occurrence Food organism consumed In diet Arthropoda; Crustacea; Telemessus cheiragonus 43 7 Mollusca: Gastropoda: . Nucella( = Thais) lamellosa 0 0 Pelecypoda; Saxidomus gigantea^ Protothaca staminea^ Mya truncala'' 481 81 Macoma inquinata'' Macoma incongrua^ Mytilus edulis. musseP 2 0.3 Pododesmus macroschisma 0 0 Clinocardium nuttalli ' 0 0 Cephalopoda; Octopus sp 4 0.6 Echinodermata: Asteroidea; Evasterias troschelii 5 0.8 Echlnoidea; Strongylocentrolus drobachiensis 0 0 Holothuroidea 2 0.3 Unidentified 60 10 Total 597 100 'Each of these pelecypods was identified as a dietary item one or more times, but the relative frequency of use was not determined. ''Observations were made on two different occasions of otters feeding on mussels. The small mussels averaged around 2 to 3 cm each. This plus the fact that the observation distance was up to 100 m made it impossible to get an exact count. Shells of the snail Nucella ( = Thais) lamellosa; cockle, Clinocardium nuttallii; and the rock oyster or jingle, Pododesmus macroschisma, were abun- dant in the study area. Tests of sea urchins were rare. Octopuses consumed by otters ranged from 30 cm to 1 m across the tips of the tentacles. Crabs (Telmessus cheiragonus) eaten ranged from 5 to 15 cm across the carapace. The clams consumed (Mya truncata, Macoma inquinata, and M. incongrua) were approximately 2 to 3 cm long, with Pro- tothaca staminea andS. gigantea ranging from 2 to 10 cm long. Mussels (Mytilus edulis ) were 2 to 3 cm long. Sea cucumbers measured 15 cm long and sea stars (Evasterias troschelii) 20 to 30 cm across the rays. From the 30 stations occupied along the inter- tidal transects, a total of four clams (two Macoma spp., oneS. gigantea, and one P. staminea) and 56 mussels were collected. Feeding Behavior Otters usually rose vertically so that the shoul- ders were above the water surface before diving (also see Limbaugh 1961). In water depths <4 or 5 m otters usually sank to shoulder level before roll- ing forward into a dive. In deeper water they ordi- narily dove from the highest position of emergence, presumably to provide greater down- ward thrust. During the beginning of a dive, the forelimbs were kept close to the body. One otter often dove backward from a supine floating posi- tion by kicking its hind flippers and arching its back. The duration of feeding dives (average 66 s; Table 2) was approximately the same as that ob- served for sea otters in California (60-90 s; Lim- baugh 1961). Otters in Montague Strait ate crabs as described by Fisher ( 1939) for California otters and by Ken- yon ( 1969) for Aleutian otters. Otters removed the legs with one paw while clasping the crab to the chest with the other paw. Kenyon (1969:116) re- ports that "in the Aleutians the carapace was not among the stomach contents," whereas Fisher (1939:28) noted for California otters "when the legs are finished, the body is eaten." While holding otters in captivity prior to translocation from the Montague Strait area during 1965 and 1966, the animals were fed commercially available crabs (Cancer magister) (Ed Klinkhart, Alaska Depart- ment of Fish and Game, pers. commun.). The ot- ters consistently ate the chelipeds first and then the walking legs. Next the carapace was removed and the body eaten. Finally the carapace was gen- erally licked prior to discarding. Unconfined sea otters occasionally bit the carapace but usually discarded it after finishing the legs. Two crabs were often taken during one dive. Otters dug out clams with their forepaws while maintaining a head downward position (see Lim- baugh 1961 for similar shallow-water feeding be- havior of California otters). Holes or craters from 15 to 45 cm across and up to 50 cm deep, made by Table 2.— Results of 673 timed feeding dives of sea otters in Montague Strait, Alaska, listed by depth. No. of dives Mean divine Approx. water Sex observed time (s) depth (m) 1 F 20 3 4 2 M 80 47 4.8 F 60 49 3 M 3 108 10.6 F 14 83 4 M 14 83 13.3 F 406 73 5 M 6 118 13.3 6 F 26 83 16.3 7 U 44 69 17.6 Total F 526 67 '9.6 Total M 147 59 '11.9 Total both sexes 673 66 '.11.9 'Average depths for combined observations. 127 FISHERY BULLETIN: VOL. 76, NO. 1 the otters in this process, were abundant in inter- tidal and subtidal areas with gravel or sand bot- toms. A male otter was observed feeding on clams about 3 to 5 cm long; 38 clams were consumed in 35 min (1.08/min). A female and a large pup, ob- served at the same location, fed on clams of the same size range as those eaten by the male. Only the female successfully brought up clams al- though the pup dove with her. Together, they con- sumed 56 clams in 65 min (0.86/min). Both adults brought up as many as three clams per dive. Generally, clams 3 to 5 cm long were eaten in- tact including the shell. The otter pushed each clam into its mouth, crushed the shell, and swal- lowed the entire clam immediately. Larger clams (5 to 10 cm long) were cracked with the cheek teeth, usually breaking one valve in half (see Mil- ler et al. 1975). This has also been observed in Monterey Bay (H. Feder pers. commun.). Valves were then forced open by a rotating motion or were pulled apart with the paws, and the soft parts scooped or bitten out with the incisor teeth. Large males were occassionally able to crack clams >10 cm with their cheek teeth and pull the valves apart with their paws. However, they typi- cally opened larger clams by pounding them against each other or against a rock until the shell was fractured and the valves forced open. The size of the rocks ranged from 7 to 15 cm long but there was no preference for shape. Otters often used stones as tools for opening hard shelled invertebrates such as clams (Fisher 1939; Limbaugh 1961; Hall and Schaller 1964). With the stone lying on the otter's chest, the clam was struck against it with several quick, hard blows until the shell or the hinge was broken. Otters were typically nonselective when striking the clam against a rock; however, one otter consis- tently struck the hinge area which usually sepa- rated after three or four blows. A rock was not used more than once. Each rock was always discarded immediately by allowing it to slip off the chest. Otters obtained mussels by pulling up holdfasts of Laminaria sp. to which the bivalves were at- tached. The animal then floated with the algal frond across the body and picked individual mus- sels off with its forepaws and ate them whole. Otters never consumed algal material. Octopuses were eaten completely. One female consumed an octopus (60 cm across the tips of the tentacles) in slightly more than 6 min. The otter held the body of the octopus in its paws and bit into an arm or the body while pulling away with its head and pushing away with its paws. This left a piece of octopus in the mouth, which was pushed in while the remainder was held in the otter's axilla or against the chest. This procedure was repeated until the entire octopus was eaten. Pieces dropped during the feeding process were retrieved. Sea stars were not a preferred food. According to Kenyon( 1969: 119), "the otter usually tears off and eats one or two arms of a sea star . . . and discards the remainder." Otters in Montague Strait fed in a similar manner. Kenyon (1969) reported several species of sea stars are eaten by otters in the Aleu- tians. Only one sea-star species (Evasterias tros- chelii) was taken by otters in the present study, although others were available (Dermasterias im- bricata and Pycnopodia helianthoides). Sea cucumbers were rarely eaten and were also apparently of minor importance to Aleutian otters (Kenyon 1969). Sea cucumbers were torn open, a portion of the viscera and part of the body wall eaten, and the remainder discarded. Feeding periods ranged from 25 to 147 min, av- eraging 84.5 min. Elapsed times for eating at the surface during the 14 feeding periods ranged from 17 s for a clam to 6 min for an octopus, with a mean value of 38 s for all foods (see Table 2 for diving times and Table 3 for average consumption times of each food item). Table 3. — Range and mean of feeding times for individual food items measured in seconds for sea otters in Montague Strait, Alaska. No.of Surface feeding time Food item observations Range IVIean Clam 81 17-64 38.6 Crab 2 30-39 34.5 Sea star 4 25-41 30 Octopus 1 — 380 Unidentified 5 17-53 34 No food brought up 52 10-54 24.5 DISCUSSION The sea otter is an opportunistic feeder throughout its range. It generally feeds on bot- tom-dwelling invertebrates, but may select fishes if the invertebrate supply is depleted (Kenyon 1969 in Table 4). Mollusks were the most impor- tant food of otters in California and Montague Strait, echinoderms are apparently most impor- tant in the Commander Islands, and fishes most important in the Aleutians (Table 5). Crustaceans were second in importance at Pico Creek, Calif., 128 CALKINS: FEEDING BEHAVIOR OF ENHYDRA I.VTRIS Table 4. — Qualitative comparison of food of sea otters in Montague Strait, Alaska. Major food items consumed I oration and Method of reference analysis Molluscs Crustaceans Echlnoderms Fishes Amchitka Island. Stomach and Chiton Crabs Green sea urchin, Globe fish. Aleutian Islands. fecal analyses Cryptochiton Cancer sp Strongylocentrotus Cyclopterichthys Alaska (Kenyon and direct stellerf Placelron drobachiensis, glaber. 1969) observation Snails Bucanum sp. Argobuccinium oregonensis Mussels wosnessenski Red Irish lord. Hemilepidotus hemilepidotus, Musculus vernicosa Volsella volsella Octopus Octopus sp. Rock oyster Pododesmus > macroschisma Pico Creek, Calif. Direct Red abalone. Rock crab, (Ebert 1968) observation Haliotis rufescens. Gaper clam. Tresus nuttalli. Cancer antennanus. Monterey Bay, Calif. Direct Red abalone. Sea urchin (Limbaugfi 1961) observation Haliotis rufescens. Purple hinged scallop. Hinites gigantea. California mussel, Mytilus californianus. Strongylocentrotus franciscanus Point Lobos. Calif. Direct Mussel Crab Purple urchin. (Hall and Schaller observation Mytilus Cancer sp. Strongylocentrotus 1964) californianus Red abalone, Haliotis rufescens, purpuratus. Commander Islands. Direct Clam Crab Sea urchin USSR (Barabash- observat.on Mya truncata Telmessus Strongylocentrotus Hexigrammidae Nikikforov 1947) and fecal analysis Mussel Mytilus edulis cheiragonus drobachiensis Montague Strait. Pirect Clams Crab Sea star Alaska (this observation Saxidomus Telmessus Evasterias study) gigantea Protothaca staminea Mussel Mytilus edulis cheiragonus troschelii and Montague Strait with molluscs second in the Aleutians and echinoderms second at Point Lobos, Calif. Sea urchins seem to be a relatively minor part of the diet in Montague Strait. No living sea urchins were found in the intertidal zone and only an occa- sional test was found. Kenyon (1969:111) indi- cated that "the bones of those sea otters utilizing sea urchins . . . are stained purple by the bio- chrome polyhydroxynaphthoquinone ( Scott in Fox 1953)."' Of the six different sets of skeletal remains found on the beaches of Montague Strait during this study, none showed this diagnostic purple stain. Schneider (Alaska Department of Fish and Game, pers. commun.) reports that of the several skulls he obtained from Prince William Sound, none show purple pigmentation. Fishes are an important food source in the Aleu- tians when invertebrates become depleted. Ken- yon (1969:110) reported that "At Amchitka it appears that the otters fall into two groups — those eating mostly fish and those eating mostly in- vertebrates." Otters were not observed eating fishes in Montague Strait and fishes are probably not important here. During the latter part of this study pink salmon, Oncorhynchus gorbuscha, and chum salmon, Oncorhynchus keta, became abun- dant. Vania^ found that otters captured in Mon- tague Strait and held for translocation refused to eat chum and pink salmon for a period of 24 h. ^Vania, J. 1967. Sea otter. /« Marine mammal investigations. Alaska Dep. Fish Game, Vol. 7 Annual Project Segment Rep., Fed. Aid Wildl. Restoration, Proj. W-14-R-1 and -2, work plan G, p. 6-13. 129 FISHERY BULLETIN: VOL. 76, NO. 1 Table 5. Frequency of occurrence of major food items in the diet of sea otters in Montague Strait, Alaska, compared with other locations. Organisms from the Commander Islands study are shown according to relative abundance as indicated by plus signs, increasing plus signs indicate increasing abundance. Location and Amchitka Island. Pico Creek, Calif. Point Lobos, Calif. Commander Islands, USSR Montague Strait, reference Aleutian Islands. (Ebert 1968) (Hall and Schaller (Barabash-Nikiforov Alaska (this Alaska (Kenyon 1969) 1964) 1947) study) Method of Stomach and fecal Direct observation Direct observation Direct observation and Direct observation analysis analyses and direct observation fecal i analysis Food Item Percent Percent Percent Abundance Percent Mollusks. Clams 2.5 + + 81 Mussels 0.8 40 + + 03 Snails Chiton 0.4 0.8 Octopods 0.4 0,6 Abalone 63.4 9.9 Rock scallop 2.1 Total 37 69.2 51,1 81,9 Crustaceans Crabs Present 25.9 145 + + 7 Spiny lobster 0,6 Total 25.9 15 1 7 Echinoderms: Sea urchins Present 328 + + + Sea stars Present 0.6 0,9 Sea cucumbers Present 03 Total 33,4 1 2 Fishes 50 0,4 + + Others 13 49 9,9 Grand total 100 100 100 100 Prior to this study, little use of rocks as tools for opening clams had been observed in Alaska. Ken- yon (1969) did not observe this phenomenon in the wild, but saw a captive Alaskan otter pound a clam against the side of its cement pool. Schneider (pers. commun.) observed otters using rocks near Amchitka, but considers this behavior uncommon. Kenyon (1969) compared rock-pounding behavior in the sea otter to the use of gravity by gulls (Larus sp.) and ravens iCorvus corax). He also suggested that tool-using behavior is derived from "chest pounding, frustration behavior" (Kenyon 1969). Otters will often pound clams on their chest when the clams are particularly difficult to open (also see Hall and Schaller 1964). Limbaugh (1961) noted that otters used the same rocks on successive dives in California. This was not observed in Montague Strait. Although Kenyon (1969:123) felt that "clams which are buried are not dug from the bottom" and that only those exposed to view or with exposed parts are taken by the otters, otters in Montague Strait frequently and successfully dug clams. Saxidomus gigantea and Protothaca staminea are found at depths of 8 to 45 cm along the North Pacific coast (Fitch 1953: Quale and Bourne 1972: Paul and Feder'*). Miller et al. (1975) presented "Paul, A. J,, and H. M. Feder. 1976. Clam, mussel and oyster resources of Alaska. Univ. Alaska I.M.S. Rep, 76-4, Sea Grant Rep, 76-6, 41 p. evidence which suggests California otters have dug pismo clams, although no direct observations have been made. When otters dig in soft sediments characteristic of clam habitats, they undoubtedly locate clams by touch due to obscured vision and, in fact, Kenyon (1969) has shown that otters can locate food by tactile sense alone. One blind captive otter located food successfully and another normal individual used only forepaws in the selection of a preferred food (Mytilus edidis ) from a bucket of turbid water that also contained small crabs iPachygrapsus), and pebbles of various sizes. It is apparent that sea otters are able to subsist on a wide variety of bottom-dwelling inverte- brates and some fishes. Although they seem to have local preferences, they tend to exploit what- ever is available. As otter populations increase they can effect drastic changes in bottom com- munities. ACKNOWLEDGMENTS This study was made possible by funds provided under Federal Aid in Wildlife Restoration, Alaska, Project W-17-3, administered through the Alaska Cooperative Wildlife Research Unit. I am extremely grateful to Howard M. Feder, Univer- sity of Alaska, for his advice and assistance and tireless critical review of the manuscript; I thank 130 CALKINS: FEEDING BEHAVIOR OF ESHVnRA LUTRIS Peter Lent and Francis H. Fay, University of Alaska, for their advice: Karl Schneider, Alaska Department of Fish and Game, for his advice and critical review of the manuscript; Paul Marhenke III, College, Alaska, for his assistance in the field; Matt Dick and George Mueller, Aquatic Collection Center, University of Alaska Museum, for iden- tification of food organisms; and Janet Viale, An- chorage, Alaska, for her encouraging assistance and patient, devoted help throughout the study. LITERATURE CITED BARABASH-NIKIFOROV, I. I. 1947. The sea otter (Enhydra lutris L.) — Biology and economic problems of breeding. In The sea otter, p. 1-174. (Translated by Isr. Program Sci. Transl., 1962, 227 p., as OTS 61-31057.) BURRIS, O. E., AND D. E. MCKNIGHT. 1973. Game transplants in Alaska. Alaska Dep. Fish Game, Wildl. Tech. Bull. 4, 57 p. DE LAGUNA, F. 1956. Chugach prehistory. The archaeology of Prince Wil- liam Sound, Alaska. Univ. Wash. Press, Seattle, 289 p. EBERT, E. E. 1968. A food habits study of the southern sea otter, En- hydra lutris nereis. Calif Fish Game 54:33-42. Fisher, E. M. 1939. Habits of the southern sea otter. J. Mammal. 20:21-36. Fitch, J. E. 1953. Common marine bivalves of California. Calif Fish Game, Fish Bull. 90, 102 p. Fox, D. L. 1953. Animal biochromes and structural colours. Camb. Univ. Press, Lond. 379 p. Hall, K. R. L., and G. B. Schaller. 1964. Tool using behavior of the California sea otter. J. Mammal. 45:287-298. KENYON, K. W. 1969. The sea otter in the eastern Pacific Ocean. North Am. Fauna 68, 352 p. LIMBAUGH, C. 1961. Observations on the California sea otter. J. Mam- mal. 42:271-273. MILLER, D. J., J. E. HARWICK, AND W. A. DAHLSTROM. 1975. Pismo clams and sea otters. Calif Fish Game Mar. Resour. Tech. Rep. 31, 49 p. Quale, D. B., and N. Bourne. 1972. The clam fisheries of British Columbia. Fish. Res. Board Can., Bull. 179, 70 p. 131 TROPHIC RELATIONSHIPS AMONG FISHES AND PLANKTON IN THE LAGOON AT ENEWETAK ATOLL, MARSHALL ISLANDS Edmund S. Hobson and James R. Chess' ABSTRACT Trophic relationships among fishes and zooplankters in the nearshore lagoon at Enewetak differ sharply between day and night, and are strongly influenced by current patterns. Adults of most diurnal planktivorous fishes are numerous in certain places where tidal currents are strong, but few where such currents are consistently weak. Thus, the sea bass, Mirolabrichthys pascalus; the snapper Pterocaesio tile; and the damselfishes (Chromis agilis, C. caerulea, C. lepidolepis, C. margaritifer , and Pomacentrus coelestus) are numerous in strong-current areas near interisland passes, but relatively few or absent in weak-current areas close in the lee of islands or interisland reefs. The former areas are rich, the latter poor, in the major prey of these fishes — copepods, larvaceans, and fish eggs. On the other hsind, the zooplankton-poor waters close in the lee of islands and interisland reefs are rich in debris from the reefs, and fishes that can subsist on these materials are abundant. Dascyllus reticulatus is numerous here, although less so than where currents are strong, and takes algal fragments as an important, if secondary, part of its diet; Pomacentrus vaiuli, equally abundemt in both strong- emd weak-current areas, feeds largely on algal fragments, as does P. pavo, which is more numerous here than where currents are strong. In contrast, the major nocturnal planktivores are concentrated where currents are weak, but relatively sparse where these currents are strong. Included are: the soldier fishes Myripristis pralinus eindM. violaceus, and the cardinalfishes Apogo« ^nicj/is (youngalsofeedby day), A. novaeguinae, and A. savayensis. They are strictly carnivores that prey mostly on larger zooplankters — including large calanoids, mysids, isopods, gammarids, postlarval carideans, and brachyuran megalops — absent (ex- cept for the mysids) in the nearshore water column by day. These prey organisms generally find conditions unfavorable where strong currents flow. Most of them are sheltered on or near specific nearshore substrata during the day and enter the water colunm only at night; but others are in deeper water offshore by day and move inshore at night after rising toward the surface. Limited evidence indicates that planktivorous juvenile and larval fishes, as well as the tiny plankters on which they feed, follow patterns different from those followed by larger individuals. Many nearshore fishes find most of their food among the plankton. Clearly, the water column is a rich feeding ground. Nevertheless, fishes that would take plankters face problems perhaps not immediately apparent. Consider, for example, the feeding-related morphologies of planktivorous fishes, which obviously are products of strong selection pressures. Fishes that take plankters by day are characterized by modifications of head and jaws, including dentition, that permit even rela- tively large individuals to effectively consume tiny organisms in midwater, whereas fishes that take plankters at night tend to be large-mouthed species with specialized means to detect, and cap- ture, the larger organisms that are in the near- shore water column only after dark. Awareness of 'Southwest Fisheries Center Tiburon Laboratory, National Marine Fisheries Service, NOAA, 3150 Paradise Drive, Tiburon, CA 94920. Manuscript accepted May 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. these facts evolved from studies in tropical seas (Hobson 1965, 1968, 1972, 1974; Starck and Davis 1966; Davis and Birdsong 1973) and was em- phasized in more detailed study in warm temper- ate waters of southern California (Hobson and Chess 1976). Additional study has shown that many fishes which take plankters by day accen- tuate fusiform bodies and deeply incised caudal fins — features that promote rapid swimming, and which, significantly, are undeveloped among their nocturnal counterparts. Increased speed, it was suggested (Hobson 1974, 1975; Hobson and Chess 1976), has given diurnal planktivores that swim in the water column quicker access to shelter in response to severe pressures from piscivorous predators; that these speed-inducing features are comparatively undeveloped among the nocturnal species, the suggestion continued, reflects a sharp- ly reduced threat from piscivorous predators in the water column after dark. 133 FISHERY BULLETIN: VOL. 76, NO. 1 The present paper considers these aspects of the interactions among the plankton and adult plank- tivorous fishes as expressed in the lagoon of a coral atoll. It is based on a study over 21 days at Enewetak, Marshall Islands, during April 1976. STUDY AREA Enewetak Atoll (lat. 1 1=26 ' N, long. 162°22 ' E) is a ring of shallow coral reefs and low islands en- circling a lagoon about 37 km north to south and 56 km east to west. It sits amid the westward flowing North Equatorial Current and was buf- feted throughout our visit (as during most of the year) by trade winds from the east. So with surface waters generally moving to the west, it was not surprising that tidal currents in passes between the open ocean and the lagoon on the windward side of the atoll were strong on the flood, but weak on the ebb. Furthermore, water over the windward interisland reefs, driven by the incessant trade winds and seas breaking over the outer reef, flowed in just one direction — into the lagoon. Pre- sumably the situation was reversed on the lee- ward side of the atoll, as described for Bikini and Rongelap, two other Marshallese atolls (von Arx 1948). From most islands, and interisland reefs, a nar- row shelf of sand and isolated patch reefs extend several hundred meters into the lagoon. At the outer edge of this shelf, where the water in most places is about 20 m deep, the sea floor drops sharply to about 50 m, which is the approximate water depth over much of the lagoon. Our study centered on the lagoon's nearshore shelf along the eastern (windward) side of the atoll, where the waters are sheltered from the trade winds and prevailing seas. Initially, we made observations from Aoman Island in the north, to Enewetak Is- land in the south — a distance of about 32 km. Underwater visibility ranged from about 5 to over 30 m, and so at all times was suitable for observing activity. From these observations we gained a general impression of how the planktivorous fishes were distributed, as well as something of their activities. It was soon apparent that the distribution of the planktivorous fishes was strongly influenced by nearshore current patterns. This knowledge per- mitted us to select fruitful locations for more in- tensive work, including sampling the plankton and gut contents of planktivorous fishes. Because time was short, we limited intensive study to two 134 sites that represented opposing extremes in pre- vailing current velocities, weak and strong — a variance that proved to identify certain major influences on fish-plankton interactions. Currents were weak or nonexistent at our site in 7 m of water among coral heads on level sand about 100 m from Walt Island, close in the lee of the interisland reef (Figure 1, site A). These weak currents were most evident when water covered the reef, and always flowed from the reef We made observations here at all hours of day and night during both spring and neap tides, and our collec- tions sampled the full range of currents encoun- tered, from no perceptible water movement to a velocity of 9 cm/s. '^Bogen Is. ,^, Japton i Is '"\^^5:> LAGOON I62°20' EAST CHANNEL Parry Is. PACIFIC OCEAN Figure l.— The study area, Enewetak Atoll, Marshall Islands. Strong tidal currents fed by water entering the lagoon through East Channel periodically swept through our site in 13 m of water among coral heads on gently sloping sand about 600 m wind- ward of Bogen Island (Figure l,siteB). During our sampling here, currents ranged from 15 to 90, x = 51, cm/s, always on flood tide. Observations (but no sampling) were also made at this station at slack water and during ebb tide when there was little perceptible current. Although there was scant evidence of an ebb current at the collection HOBSON and CHESS; TROPHIC RELATIONSHIPS AMONG FISHES site, a slow outflow from the lagoon was evident in East Channel itself. Even though strong currents at this site were limited essentially to flooding spring tides, their impact was clearly visible on the substrate at all times. Most notable, the sand, which swirled about in the stronger currents, was piled high in the lee of the patch reefs. METHODS Plankton The methods used to collect plankton differed between the two primary sites owing to the con- trast in prevailing current velocities. Never- theless, all collections employed the same 0.333-mm mesh net and produced comparable as- sessments of the plankton at the two places, par- ticularly between day and night. Collecting Where Currents Were Weak When sampling at the Walt Island site, we pushed the net through the water around one patch reef (Figure 2), a circuit that always took 5 min. The procedure was similar to that used at Santa Catalina Island, Calif. (Hobson and Chess 1976). When swimming with the net by day, we could watch organisms in its path, and this gave us insight into which of them might be evading the net. Mysids, for example, could do so, and often did. But these organisms reacted to us less than expected, perhaps because the meter net's opening was large, and its approach was slow and quiet. Certainly our collections would have sampled these large mobile forms less effectively if the net had been preceded by the harness and tow line used when operating from a boat. Three series of collections were made during midday (between 1000 and 1400 h), and three series were made at night ( 1 h after last evening light, at midnight, and 1 h before first morning light). We spaced the noctunal collections over the night because earlier work had suggested that certain organisms are in the water column only briefly during specific periods of the night, a phenomenon we did not find among the diurnal plankters (Hobson and Chess 1976). Of the three collections in each series, one was made within 1 m of the bottom, one midway between bottom and surface, and one with the net breaking the surface. At night, ambient light in this clear water over white sand permitted us to collect without diving lights. Our stay at Enewetak spanned the period from full to new moon, so that we sampled both spring and neap tides, but generally there was no moonlight during the collections owing to cloud cover or time of night. Net speed was 28 cm/s, as calculated from read- ings of a current meter calibrated by the speed at which the smallest fragments of algae visible to us drifted along a measured course. We decided it was necessary to determine net speed only once at this station, because all collections were made by the same two swimmers who each time exerted about the same effort, and covered the same distance. FIGURE 2.— Collecting plankton at Walt Island, Enewetak Atoll, site of weak currents. The square frame per- mitted more accurate assessments at the surface and close to the sea floor. 135 FISHERY BULLETIN: VOL. 76, NO. 1 Collecting Where Currents Were Strong All collections at Bogen Island were made at the height of flood tide, when currents often were too strong to swim with the net, so here we worked from a boat anchored fore and aft above the study site. The net was secured to a line that passed from the boat, through a block anchored on the reef below, and returned to the boat. It was positioned at the three collecting depths — bottom, mid- depths, and surface — by pulling the line one way or the other through the block (Figure 3). The collections were extended to 15 min (compared with 5 min at the other station) to reduce error introduced by organisms taken during the few seconds it took to raise and lower the net. In pre- senting these data, however, we make the values equivalent to a 5-min collection. These collections depended on the current (which was measured with every collection) to carry plankton into the net, and the weakest current sampled, 15 cm/s, was judged close to the minimum necessary. Two series of collections were made during the day — at midday and in midafternoon — and two series were made at night — 1 h after last evening light and at midnight. There are problems in comparing data collected by these different methods at the two stations, but we had the advantage of sampling precisely defined positions — a critical requirement when relating the plankton to food habits of specific fishes. The volume of water filtered by this stationary net varied with the different current velocities. which strongly influenced the numbers of plankters taken. Nevertheless, these numbers ac- curately reflect the relative numbers of plankters available to fishes feeding in these currents. On the other hand, differences in volumes of water sampled must be considered when comparing es- timates of the plankton in the water column from one time or place to another. Therefore, plankton volumes from the strong-current site are pre- sented two ways: volumes actually sampled and volumes adjusted for current differences. In ad- justing for current, the volumes in all collections were made equivalent to those taken in a net mov- ing at the same relative speed that we pushed the net at the Walt Island site — 28 cm/s. These ad- justments also permit rough comparisons with data from California (Hobson and Chess 1976), where plankton were collected in the same way and by the same swimmers. Fishes A total of 154 fish specimens of 16 species were speared immediately after the plankton collec- tions. Species names are those used by Schultz et al. (1953, 1960), except where more recent taxo- nomic study has indicated change. The specimens were preserved in 10% Forma- lin^ immediately after collection. Later, food items in the gut were identified and their positions in the gut noted. The following data were recorded for ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Figure 3.— Collecting plankton at Bogen Island, Enewetak Atoll, site of strong currents. 136 HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES items ofeach food type: 1) number, 2) size range, 3) state of digestion (subjectively assessed on a scale of 5, from fresh to well-digested), and 4) an esti- mate of their representation among the gut con- tents as percent of the total volume. RESULTS Our widespread observations along the sandy shelf which rims the lagoon established that the planktivorous fishes were centered about the iso- lated patch reefs. At least a few planktivores for- aged in the water column above virtually every reef, but more of them were above some reefs than others and there were clear patterns to their dis- tributions. For example, during the day there tended to be more planktivores above reefs at the outer edge of the shelf than above similar reefs at comparable depths, and shallower, shoreward on the shelf. But diurnal planktivores were most numerous where strong tidal currents flowed through passes from the open sea, and least numerous where reefs or islands blocked the flow of water into the lagoon. On the other hand, the reverse was true of the nocturnal planktivores. Because the distributions and activities of these fishes proved to be closely related to current pat- terns, we judged that the contributing influences are best isolated by concentrating on the more extreme current situations. This was true even though in most places over the range of our obser- vations currents were variably moderate, and prevailing conditions intermediate between the two extremes. Where Currents Were Weak General Observations There is relatively little water movement near the lee shores of the islands and close behind the interisland reefs that block entry of water into the lagoon from the open sea. In some of these loca- tions there is enough circulation to permit rich coral growth and underwater visibility that ex- ceeds 15 m, but in other places the circulation is more limited, and living corals exist as small heads or encrustations on otherwise dead reefs, while underwater visibility often is <5 m. The lagoon floor in these regions generally is of rela- tively undisturbed, fine-grained sand. (A sample of sediment from the Walt Island site proved to be 75% foraminiferans, with a density of 1.32 g/ml. Grain size in over 80% of this sample was < 1 mm.) PLANKTON.— Usually we made no effort to detect the smaller plankters during our general diurnal observations, even though many of the copepods and others were visible with close inspec- tion. Dense swarms of mysids, however, were out- standing features of the daytime scene in many places where currents were weak, especially above sand close to the patch reefs. With increasing dis- tance from the bottom, their swarms were smaller and less numerous, though swarm-members al- ways were closely spaced. Juveniles predominated at the lower levels, adults were more numerous above. The swarms dispersed at night, when both adults and juveniles scattered near the bottom and at middepth, but only adults were near the sur- face. Although mysids were the only plankters routinely noted during the day, others were prom- inent after dark. Most conspicuous were large calanoid copepods — larger than any copepods pre- sent in daylight — that for a few hours after last evening light swarmed around us in dense num- bers whenever we turned on our diving lights. Highly motile epitokous nereids, as well as an opheliid, Polyophthalmus sp., were numerous polychaetes, with other forms including hyperid and gammarid amphipods, stomatopod larvae, reptantian zoea, and brachyuran megalops. None of these forms were seen in daylight. FISHES. — Adult diurnal planktivorous fishes were relatively few in these surroundings com- pared with their numbers elsewhere. Neverthe- less, this seemed a favored habitat for at least one species, Pomacentrus pavo, which was widespread in groups of four to six individuals 2 to 5 cm above low coral-rock outcroppings in the sand, usually in the vicinity of patch reefs. Pomacentrus vaiuli, another abundant species, was present only as solitary individuals that rarely moved more than a few centimeters from the larger patch reefs, yet most of its food was small organisms swimming or drifting free in the water immediately adjacent to the substrate. Dascyllus reticulatus was numerous by day in feeding aggregations up into the mid- waters, usually above large heads of branching corals, while at the same time Amblyglyphidodon curacao, which usually fed in groups of <10, often ranged up to the water's surface. Of the diurnal planktivores considered here that ranged into the 137 FISHERY BULLETIN: VOL. 76, NO. 1 water column, D. reticulatus and A. curacao were the only deep-bodied forms. Other diurnal plank- tivores were more sparse. The more prominent of these were species ofChromis that usually stayed within 2 m of the reef. Chromis caerulea,^ mostly juveniles, generally hovered in small aggrega- tions above heads of the coral Pocillopora, but C. agilis and C. margaritifer more often were solitary or in groups of just a few. At night all of these fishes were under reef shelter, and we saw no evi- dence of them feeding at that time. Despite the relative paucity of adult diurnal planktivores in this habitat, planktivorous juve- niles and larvae of at least several fish species frequently were numerous and fed by day. An out- standing example was the juveniles of Apogon 3 At the distances that most of our observations were made, we were unable to consistently distinguish Chromis caerulea from the very similar C. atripectoralis , and so referred all observa- tions to the former. Significantly, however, the behavior attri- buted to this species is consistent with that in all individuals observed. gracilis, well under 50 mm long, which hovered in large, umbrella-shaped aggregations above coral heads in open sand (Figure 4). Dense schools of larval fishes, 7 to 10 mm long, (often taken on first glance as mysid swarms) were sometimes promi- nent, but so close to the reefs that our net sampled only an occasional outlier. Although adult diurnal planktivores were com- paratively sparse in this habitat, their nocturnal counterparts tended to be especially numerous. During daylight, dense, inactive concentrations of Myripristis spp. abounded at openings of reef cre- vices. Prominent as these concentrations were, they represented only a small part of the tremend- ous numbers of their species packed into the reef interstices. We became fully aware of the immen- sity of these populations when, about 30 min after sunset, they abruptly streamed into the open and entered the water column. Shortly after emerging, most individuals of one species, M. murdjan, ap- parently moved elsewhere, because though they were numerous initially, relatively few were seen during the night, and their numbers did not in- FIGURE 4. — Juvenile cardinal fish, Apogon gracilis, approximately 25 to 30 mm long, feeding on plankton by day where currents are weak. 138 HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES crease again until just before dawn. In contrast, large numbers of M. praliniis and M. violaceus remained concentrated in the waters above the nearshore patch reefs throughout the night. Also prominent in daylight were Apogon spp., which concentrated close to reef cover. These in- cluded adults of A. gracilis, which schooled quietly at the bases of the same coral heads above which juveniles of the species (see above) actively fed; nevertheless, the true numbers of apogonids were fully appreciated only after nightfall, when many large species unseen during the day emerged from reef shelters. The most prominent of the larger apogonids entering the water column was A. savayensis, although some of the smaller species, notably A. ^ract/ts and A. nouaeguinae, were more numerous. Larger apogonids were solitary at night, but smaller ones often were loosely aggre- gated, including A. gracilis, of which the adults Table i. -Composition of plankton at Walt Island, Enewetak Atoll, site of weak currents. Day(r 1 =9) Nigtit (n = 9) Materials Mean vol (ml) Mean % of total vol Mean Mean % of vol (ml) total vol Zooplankters Algae fragments Crustacean molts Totals 3.4 3.6 0.5 7.7 38.3 51.8 9.9 100.0 10.7 79.0 3.1 21.0 0.0 0.0 13.8 100.0 joined the juveniles in the water column after dark. Samples From Walt Island PLANKTON. — Major materials (zooplankters, algae fragements, and crustacean molts) taken in the plankton net during day and night at the Walt Island site of weak currents are listed in Table 1. Zooplankters, grouped by major taxonomic categories and with data pooled from the three sampled depths (surface, middepth, and near bot- tom), are listed in Table 2. Additional data for calanoid copepods are presented in Table 3 to sup- port certain points developed in the Discussion. Table 3. — Size distribution of calanoid copepods, day and night, at Walt Island, Enewetak Atoll, site of weak currents. Midday Night Size (n = 9) 1 h after last hgfit (n = 3) Percent Mean no. Midnight and later (n = 6) (mm) Percent Mean no Percent Mean no. >3-5 >2-3 >1-2 <1 0 0 0 0 48 ='10.9 52 11.8 0 0 43 57 0 0 "40.1 53.2 24 31 31 14 M38.7 ^180. 5 M79.0 81.3 ' Including Euchaeta manna. Pleurommama xiphias. and Undinula vulgaris. ^Including Candacia sp.. E. marina. Neocalanus sp., Pleurommama xiphias. and U. vulgaris. ^Including Acartia sp.. Metndia sp., Pleurommama sp., and Scolothricella sp. "Including Acartia sp., Candacia sp., E. marina, and U. vulgaris. ^including Acartia sp , Candacia sp., and Euchaeta sp Table 2.- -Occurrence, number, and size of zooplankters collected day and night at Walt Island, Enewetak Atoll, site of weak currents. Day (n = 9) Night (n = 9) Plankton categories Size Percent Mean Size Percent Mean present (mm) occurrence number (mm) occurrence number Foraminiferans' 0.4-1.0 100 36.7 0.4-2.0 100 337.0 Siphonophores 4.0-6.0 11 0.4 4.0-8.0 38 2.6 Polychaetes — 0 0.0 3.0-25.0 33 28.3 Mollusk larvae 0.3-1.0 78 21.0 0.5-2.0 89 55.2 Pteropods — 0 0.0 2.0-5.0 33 2.0 Squid — 0 0.0 3.0-12.0 22 0.3 Ostracods 0.5-1.0 67 5.4 0.6-2.0 100 264 Calanoid copepods 0.5-2.0 89 22.7 0.5-5.0 100 579.5 Cyclopoid copepods 0.5-1.5 56 8.0 0.5-2.0 100 39.0 Harpacticoid copepods 0.5-1 0 22 0.3 0.5-2.0 89 9.3 Stomatopod larvae — 0 0.0 18.0-26.0 11 1.0 Mysids 2.0-8.0 89 21,398.7 1.0-8.0 100 33,031.8 Cumaceans — 0 0.0 1.0-1.5 56 12.4 Isopods — — 0.0 1.0-12.0 67 5.3 Hyperiid amphipods 0.6-2.0 33 5.0 1.0-4.0 100 17.8 Gammarid amphipods — 0 0.0 1.0-5.0 100 23.2 Caridean larvae 2.0-3.0 89 6.0 1.0-12.0 100 504.2 Caridean adults and juveniles — 0 0.0 4.0-15.0 100 20.0 Reptantian zoea 05-2.0 78 20.0 0.5-4.0 100 629.8 Brachyuran megalops 20-3.0 22 0.2 2.0-8.0 100 60.3 Chaetognaths 40-10.0 44 1.0 3.0-12.0 100 92.4 Larvaceans — 0 0.0 2.0 11 0.4 Apendicularian larvae 2.0 11 0.1 2.0 11 0.4 Fish eggs 1.0-2.0 100 40.0 0.5-3.0 100 273.6 Fish larvae 2.0-13-0 44 11.3 2.0-25.0 100 51.2 'Most of them planktonic stage of Tretomphalus. ^All appeared to be Mysinae sp. Mysids constituted 52.8% of the volume of daytime collections. ^Included Mysinae sp. and Sinella sp., the latter unseen in daylight. Mysids constituted 44.8% of the volume of nighttime collections. 139 FISHERY BULLETIN: VOL 76, NO 1 GUT CONTENTS OF THE PLANKTIVOR- OUS FISHES.— The gut contents of diurnal fishes collected at the same time, and in the same loca- tion, as the daytime plankton collections are listed in Table 4, and those from the nocturnal species, which were collected between midnight and first morning light on nights when the plankton were sampled, are listed in Table 5. Where Currents Were Strong General Observations Currents were periodically strong near the passes from the open sea, and here, where patch reefs and other hard substrata typically are co- vered with living corals, underwater visibility consistently exceeded 20 m.^ The lagoon floor in these areas generally is coarse, well-sorted sand ( a sample of the sediment at the Bogen Island site proved to be about 60% fragments of calcareous algae, Halimeda spp., with a density of 1.25 g/ml; grain size in over 80% of this sample was greater than 1 mm). PLANKTON. — Plankters were noted infre- quently during casual diurnal observations where currents were strong. Nevertheless, the mysids so prominent where currents were weak occurred here only in small, inconspicuous swarms that concentrated close in the lee of patch reefs when currents were running. The larger zooplankters, frequently so prominent after dark in weak- current areas, were not noted here in any abun- dance, although nocturnal observations underwa- ter in this habitat were limited. ■•Our concept of strong-current locations does not include those breaks in the interisland reefs where the lagoonward flow of water crossing the reef concentrated and spilled into the lagoon at sometimes exceptionally high velocities. These currents were localized and relatively shallow. Planktivorous fishes present were essentially those of nearby weak-current locations in the lee of these reefs, and although no collections were made here we would not have expected such currents to be rich in zooplankters, for reasons developed in the Discussion. FISHES. — During the day planktivorous fishes were especially numerous in these surroundings. Many diurnal species were concentrated here, the more prominent being: the serranid Mirolab- richthus pascalus, the lutjanid Pterocaesio tile, and the damselfishes Chromis agilis, C. caerulea, C. lepidolepis , C. margaritifer, Pomacentrus coe- lestus, and Dascyllus reticulatus. Pomacentrus Table 4.— Food habits of diurnal planktivorous fishes from Walt Island, Enewetak Atoll, site of weak currents. The value outside the parentheses is the rank of the item as food of that fish species ( based on incidence and volume in diet); of the two values in parentheses, the first is the percent offish of that species containing the item, the second is the mean percent of the total diet of that fish species represented by the item. 1 Apogon gracilis (juveniles) n = 10, 17-37, x = 27 mm SL 2. Pomacentrus pavo n =_5: 46-65, x = 57,2 mm SL 3. P. vaiuli n = 6. 40-51, x = 50 mm_SL 4 Dascyllus reticulata n = 5; 50-74, x = 63 7 mm SL Categories present Mean no.' 1 5. Amblyglyphidodon curacao n_= 5; 67-82, x = 74.2 mm SL 6. Chromis agilis n = 2, 50-54, x = 52 mm SL 7. C. caerulea n = 5: 44-73, x =58.6 mm SL 8. C. margaritifer n = 3; 43-50, x = 46 mm SL 3 4 5 6 7 Plankton: Foraminiferans 36.7 — — — — — * — — Siphonophores 0.4 — — — — — — — — Mollusk larvae 21.0 — — — — — — — — Ostracods 54 — — — — — . — — — Calanoids and cyclopoids 230.7 1(100:90) 2(100:16.2) 4(17:2.5) 1(100:62.7) 2(100:23) 1(100:75) 1(100:85.8) 2(100:41.7) Harpactlcoids 0.3 4(20:0.4) 6(17:0.3) — — — — — Mysids 1,398.7 — 4(17:0.8) — 5(33:1.0) — 4(50:6.0) 4(20:1.4) — Hyperids 5.0 — — — — — — — — Candean larvae 6.0 — 5(17:0.5) — 4(67:2.3) — — 5(20:1) — Reptantian zoea 200 — — — — — — — — Brachyuran megalops 0.2 — — — — — — — — Chaetognaths 1.0 3(40:3.6) — — — — — — — Larvaceans O — — — — — — 2(20:8) 4(33:3.3) Apendlcularian larvae 0.1 — — 6(17:0.5) — — — — — Fish eggs 40.0 2(40:60) 3(67:5.0) 2(50:5.3) 3(67:7.3) 3(40:5.0) 2(100:8.5) 3(60:1.8) 3(67:8.3) Fish larvae 11.3 — — — — — — — — Algal fragments — 1(100:77.2) 1(100:85) 2(67:15) 1(100:65) 3(50:10.5) — 1(100:46.7) Crustacean fragments - ' '■ — — — (33:17) — — — — Gurry — — — (33:10.0) (60:7.0) — (20:2.0) — Benthic: Cephalaspidean mollusks — — 5(17:1.7) — — — — — Compound ascidlans — — 3(33:5.0) — — — — — 'Numbers of plankters (from Table 2) provided only for rough measure of relative abundance. ^Calanoids and cyclopoids not separated in gut contents: both occurred in all fish species but calanoids predominated. ^Larvaceans not present in plankton collections but in two fish guts. 140 HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES Table 5. — Food habits of nocturnal planktivorous fishes from Walt Island, Enewetak Atoll, site of weak currents. See Table 4 legend for explanation of listed values. 1 . Myripristis pralinus n = 10:82-124, X = 100 mm SL 3. Apogon savayensis n = 9; 50-71 , X = 60.7 mm SL 2. M violaceus n = 1 1 ; 120-168. X =149 mm SL 4. A. novaeguinae n = 10: 20-42. X = 32.1 mm SL Plankton categories present tviean no.' 1 2 3 4 Foramlniferans 337.0 Siphonophores 2.6 — — — — Polychaetes^ 28.3 6(20:5.3) 1(91:45.4) 5(11:5.6) 7(10:5.5) Mollusk larvae 55.2 — — Pteropods 2.0 — — — — Squid 0.3 — 10(9:0.5) — — Ostracods 26.4 11(10:0.2) — — — Calanoids 579.5 31(100:37.0) =■5(36:1.4) 36( 11:2.2) 1(70:38.3) Cyclopoids 39.0 — — — — Harpacticoids 9.3 — — — — Stomatopod larvae 1.0 — 4(18:4) 7(11:1.7) — Mysids 3,031.8 2(80:17.5) 3(56:9.7) 2(67:26.1) 5(30:4) Cumaceans 12.4 — — — — Isopods 5.3 7(20:1.5) — 8(11:1.1) — Tanaids n 10(10:0.5) — — 9(10:1) Hyperids 17.8 — — — — Gammarids 23.2 8(20:1) 11(9:0.2) — 10(10:0.3) Caridean larvae 504.2 — — 9(11:0.6) 2(70:18.2) Caridean adults and juveniles 20.0 5(50:7.0) 8(9:3.7) 4(44:7.8) 3(30:16.0) Reptantian zoea 6298 — — — 11(10:0.2) Brachyuran megalops 60.3 3(50:17.3) 2(82:23.1) 1(100:28.3) 6(20:3) Chaetognaths 92.4 9(10:1) — — — Larvaceans 0.4 — — — — Apendiculanan larvae 0.4 — — — — Fish eggs 273.6 — — — 8(10:2) Fish larvae 51 2 4(30:8.2) 6(18:27) 3(56:20,6) 4(30:11.5) Fish adults and juveniles 0.3 — 7(9:4.7) — — Insects n — 9(27:0.9) — — Algal fragments — — — — Crustacean fragments — (9:2.3) (33:6.0) — Unidentified fragments — (18:1.4) — — 'Numbers-of plankters (from Table 2) provided only for rough measure of relative abundance. ^Most polychaetes in guts of fishes were nereid epitokes. ■'Predominant calanoids in the three larger fish species were Pleurommama xiphias and Euchaeta marina, which were relatively large (3 to 5 mm). ^Tanaids and insects were not present in plankton collections but were in several fish guts. Both are known from plankton collections elsewhere (e.g.. Hobson and Chess 1973, 1976). vaiuli was numerous, but perhaps no more so than where currents were weak (see above), and here too it confined itself to the immediate proximity of the reef The nature of the substrate can be important. Chromis caerulea and Dascyllus reticulatus, for example, swam in tight well-defined aggregations above specific growths of branching coral — particularly large heads oi Pocillopora spp. (Fig- ure 5A). Pomacentrus coelestus generally sta- tioned itself low in the water column above out- croppings of coral rock and rubble, its relation to the substrate much like that of the similarly hued, but deeper-bodied, P. pavo. Chromis agilis, C. lepidolepis, and C. margaritifer generally swam in small widespread groups over patch reefs. Com- pared with their congener C caerulea, they showed less affinity to specific substrata or loca- tions on the reef. Thus C. caerulea invariably re- sponded to a human intruder by sheltering among the branches of a large coral head directly below its feeding station (Figure 5B), whereas C. agilis, C. lepidolepis, and C margaritifer frequently re- sponded to the same stimulus by moving away, and taking shelter in a variety of places only when the stimulus was intensified. In places where many of these diurnal plankti- vores were concentrated, a relation was evident between their morphologies and the distances they swam from the reef: those with feeding sta- tions farther from the reef tended more toward cylindrical bodies and deeply incised caudal fins (Figure 6). This generalization proved valid de- spite exceptions among such deep-bodied forms as Dascyllus reticulatus (Figure 7) and Amblygly- phidodon curacao, in which the effect of their deeper bodies is even further enhanced by longer fin spines. Thus, for example, 7 D. reticulatus, 47 141 FISHERY BULLETIN: VOL. 76, NO. 1 Figure 5. — A. Chromis caerulea, and a few Dascyllus reticulatus (lower left), feeding on plankton above a head of Pocillopora at the Bogen Island site. The largest fish are about 70 mm SL; the coral head is about 1.5 m in diameter. B. Upon being threatened, the fish shown in 5A dive to shelter in the interstices of the coral head. 142 HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES "TT" J.**'^ E m m 31 Q. ,« --'«^S= Figure 6. — Planktivorous fishes where currents are strong. Major species in each of the zones identified in the photo by roman numerals are illustrated in the appropriate column below the photo (placement based on observations made at the scene). I. Pomacentrus vaiuli; II. a, Chromis agilis, b, C. margaritifer; III. a, C. caerulea, b, C. lepidolepis; IV. Mirolabrichthys pascalus; V. Pterocaesio tile. to 60 mm SL, x = 55.9, had longest dorsal fin spines that were 20.3 to 23.4%, x = 21.0%, of their stan- dard length, whereas these values for 13 individu- als of Chromis spp. (4 C. agilis, 4 C. caerulea, and 5 C. lepidolepis ), 52 to 70 mm SL, x = 59.4, were 12.3 to 16.1%, x= 15.3%. The significance of these data becomes clear when possible selective values of both fusiform and deep-bodied morphologies in planktivorous fishes are treated in the Discussion. Although most diurnal planktivorous fishes fa- vored conditions associated with current, the strongest currents observed at this site, approxi- mately 1 m/s, clearly exceeded optimum veloc- ities. When such currents flowed, most of the smal- ler planktivores were close to the reef, many of them concentrated in the lee, and their feeding rates had noticeably declined. In comparison to the great numbers of adult diurnal planktivores in these surroundings, the nocturnal planktivores were sparse. Although ob- servations underwater in this habitat at night were limited, only a relatively few individuals of 143 FISHERY BULLETIN: VOL. 76, NO. 1 Table 6. — Composition of plankton in 6 day and 6 night collec- tions at Bogen Island, Enewetak Atoll, site of strong currents.* Items Zooplankters Algae fragments Totals Day collections: Mean vol (ml): Collected 2.8 5.7 8.5 Adjusted Mean % of total vol 1.2 323 2.7 677 3.9 100.0 Night collections: Mean vol (ml); Collected 7.3 1.9 9.2 Adjusted Mean % of total vol 39 78.8 1.0 21.2 4.9 100.0 'Currents during diurnal collections_32 to 90 cm/s, x = 57; currents during nocturnal collections: 15 to 83 cm/s, x = 45. Figure 7. — Dascyllus reticulatus illustrates the tendency to- ward a deep body in certain diurnal planktivores that is in contrast to the tendency toward a more cylindrical body in many others. Myripristis spp. and Apogon spp. were seen. Furthermore, during extensive daytime observa- tions here we failed to note the dense concentra- tions of these and other nocturnal fishes in diurnal shelters that were widespread and obvious where currents were weak. Samples From Bogen Island PLANKTON.— The major materials taken in the net at the Bogen Island site of strong tidal currents were zooplankters and algae fragments (Table 6). To facilitate comparisons with collec- tions from the weak-current site, all volumes are standardized to a 5-min collection. The table lists 144 volumes of plankters actually collected, as well as volumes adjusted to the standard relative net speed of 28 cm/s (the net speed at the weak-current site). The zooplankters collected at Bogen Island, grouped by major taxonomic categories and with data pooled from the three collection depths (sur- face, middepths, and near bottom), are listed in Table 7. For the reasons given above concerning volumes, the table lists numbers of plankters ac- tually collected and numbers adjusted to the stan- dard relative net speed. Additional data on calanoid copepods (Table 8) are presented to sup- port certain points developed in the Discussion. Possibly zooplankters attempting to hold sta- tion above precise points on the sea floor would be sampled less effectively by the stationary net dur- ing the slower currents sampled at Bogen Island than by the moving net used at Walt Island. We discount this possibility as a significant source of error, however, because we did not see such or- ganisms during our underwater observations of the operation, or when examining collections that sampled a wide range of current velocities. GUT CONTENTS OF THE DIURNAL PLANKTIVOROUS FISHES.— The gut contents of diurnal fishes collected at the same time, and in the same location, as the daytime plankton collec- tions are listed in Table 9. Only a relatively few nocturnal planktivores (all of them Myripristis spp. and Apogon spp.) were seen during the limi- ted observations in this habitat after dark, and none were sampled. DISCUSSION We were unable to intensively sample more than two stations in the limited time available to us at Enewetak. Nevertheless, data collected at these two sites under a variety of conditions, HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES Table 7. — Occurrence, number ( actual and adjusted for current velocity), and size of zooplankters collected day and night at Bogen Island, Enewetak Atoll, site of strong currents. Day (n = 6) Night (n = 6) Size Percent Mean no Mean mo. Size Percent Mean no. Mean no. Plankton categones present (mm) occurrence (collected) (adjusted) (mm) occurrence (collected) (adjusted) Foraminiferans' 0.3-1.0 100 563 27.7 0,3-2 100 558 7 346 4 Siphonophores 4,0 17 0.6 0.3 4-8 50 5.7 3,5 Polychaetes 3.0 17 0.6 0.3 3-20 100 6.9 4,3 Mollusk larvae 0.5-2.0 100 11.1 5.4 0.5-2 100 31.1 19.3 Pteropods 0.5-6.0 100 6.3 3.1 2-12 83 33.7 20.9 Squid — 0 0.0 0.0 3-4 50 1.8 11 Cladocerons 0.7-1.0 33 0.9 0.4 — 0 0 0 Ostracods 0.5-2.0 100 14.1 6.9 0,5-2 83 107,6 66.7 Calanoid copepods 0.5-40 100 1,726.7 846.4 05-5,0 100 7,751.1 4,820 1 Cyclopoid copepods 0.5-2.0 100 8400 411.7 0.5-2,0 100 303.6 188.6 Harpaticoid copepods 0.5-2.0 100 23.0 11.3 0.8-2 67 6.2 3.8 Mysids 2.0 83 6.2 3.0 0,5-7 67 16.2 10.0 Stomatopod larvae — 0 0 0 20-25 17 0.2 0.1 Cumaceans — 0 0 0 2 17 0.1 <0.1 Tanaids 2.0 17 0.6 0.3 — 0 0 0 Isopods — 0 0.0 0.0 1-3 100 57.3 35.5 Hyperid amphipods 0.4-2.0 50 32.2 15.8 0.5-6 100 76.0 47.1 Gammarid amphipods 1.0 17 0.6 0.3 3-4 100 38.4 23.8 Euphausid larvae 0.5-7.0 50 5.0 2.5 0.8-1 33 9.1 5.6 Caridean larvae 1.0-4.0 100 81.2 39.8 1-10 too 386.1 239.4 Carldean adults and juveniles 2,0-6,0 33 2.8 1.4 5-15 83 13.1 8.1 Reptantian zoea 05-2.0 100 252.8 123.9 0.5-4 100 509.7 316.0 Brachyuran megalops 1.5 17 0.6 0.3 1-6 . 100 115,4 71,6 Ophiuroid larvae 2.0 17 0.5 0.3 — 0 0 0 Chaetognaths 2.0-15.0 100 75.3 36.9 3-55 100 440,0 272,8 Larvaceans 1.0-3 0 100 25.3 12.4 2-4 100 87,4 54,2 Salps — 0 0 0 (?) 17 1,1 0.7 Fish eggs 0.5-2.0 100 732.2 358.8 1-2 100 3.785.6 2,347.1 Fish larvae 2.0-6.0 50 6.6 3.2 23-90 100 46.6 28.9 ' Most of them planktonic stage of Tretomphalus ^A 90-mm leptocephalus iarva, TABLE 8. — Size distribution of calanoid copepods, day and night, at Bogen Island, Enewetak Atoll, site of strong currents. Size Midday (/ 1 = 6) Night {n = 6) (mm) Percent Mean no.' Percent Mean no.' >3-4 0 0 5 2246.7 >2-3 3 ^26.2 25 "1,203 6 >1-2 54 5453.7 60 ^2,888,6 <1 43 366 5 10 481 4 'Numbers from collections in varying currents adjusted for equivalence to collections from the Walt Island site. ^Including Euchaeta marina. ^Including Candacia sp and E marina "Including Candacia sp,, E. marina. Neocalanus sp, and Undinula vulgaris. 5 Including Acartia sp, and Euchaeta sp, 'Including Acartia sp,. E. marina, and Metridia sp. supplemented by widespread observations else- where, permit a synthesis that we hope stimulates needed additional study. The following discussion pertains to adults of the planktivorous fishes and to plankters collected by our 0.333-mm mesh meter net. All food items found in the fish guts occurred in these plankton collections, so the com- bined assemblage can be considered a trophic unit. The situation described from these data, however, may not apply to smaller individuals. Limited data, including that from Apogon gracilis, the only planktivore studied as an early juvenile, suggest that the smaller plankters which passed through our net, and their predators among juvenile and larval fishes, may follow significantly different patterns (see Miscellaneous Considera- tions below). Diurnal Relationships Probably diurnal planktivores concentrated where strong tidal currents fiowed into the lagoon through the passes because these waters were rich in zooplankters, particularly calanoid copepods (Table 7). We presume that at least many of these were oceanic zooplankters carried to within reach of inflowing tidal currents on the eastern side of the atoll by the westward flowing North Equato- rial Current — a phenomenon amplified by the trade winds. In addition, some of the materials carried from the lagoon on the preceding ebb tide probably return. Although this outflow is minimal on the windward side of the atoll, at least during the trade- wind season (see von Arx 1948), it prob- ably contains significant amounts of certain kinds of organisms. Gerber and Marshall (1974) noted that the waters of the Enewetak lagoon are much richer in zooplankton than the surrounding ocean. Describing the same condition at Bikini, Johnson (1949) stated: "Much of the oceanic plankton 145 FISHERY BULLETIN: VOL. 76, NO. 1 > T3 C a> bc 3 ca H a> CO c Cd c C8 C be o DQ E o 0] 0) J3 to 3 o o > c CO e 3 -5 ca 1 05 w J < ? i fc_];0 h.- 11 t^ E Ix "5. " -S "S ^•a « II Q.S 3 " ro c *- « 0)ci> = P m u 3 I lit (b r^ CO oi C/3 E c?^ EcT^E qIx<9 II r-: «^ II St- rain c o 1^ o „ «■- o W II 3 J c o 0) o 0) _ S 'J' c ra a. f^ CO CO ^ »— ^ C31 l-~ d O CO CO 0 CD CM incoococococnoq c3)T-cocbin(bc\jc3i >- .- .- CO I I I I I I I I I I I I 6 I I I II I d M CO o CD CD CD C\l CD rr CJ> ..- CO I s I I i II I I I I I I I I CO O CM I I O T- o C in C II III I 1^ CD - ° I I CM CD ISM II III CD O o CO CM ocT I CD O o o _ _ r^ CD I § I o o T- CM CO ,^ CM Mil SS M II § I I o M I I I r^ CM in in CM -r^ CO CO CT> -"^f O O CD CM CO ■- O CM CD O CM CD CO I • ■ I o CM ^ r-. >- CD CO CO 6 CD ■^ CO CO n CO I I I I CD O d CO I I I I iri d I o I I I I - g ^ i2 -D 2 ,„S.5?E§ca>c C7)l=Q.ir"EcE ©i5_~i2c3)aicn>, (n in It C7) .-^ ■*—(—•«— ~1 ■c= -"^ CJ '^ 3 ^ w m 5 2" :: ~ U- Ll LL < O c E . o O 0) CU.C ^D O C O > ^£- p o 3 o - . JD O O CO „ o co^ 2 — o *- o c -■D ■= o "ra £ raS D CO a ZO< 146 HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES swept into the lagoon thrives there and becomes concentrated so that the average concentration per cubic meter of the eleven most common animal groups is about four times higher than outside." In addition, by the time the incoming current passed our Bogen Island station it presumably had picked up lagoon materials upstream, so its contents probably were of diverse origin. Of course, currents in themselves enhance the planktivorous habit because planktivores holding station above a reef receive more plankters in cur- rents than in equally rich waters without cur- rents. Most of these fishes, however, take shelter by the time a current reaches 1 m/s, so that opti- mal velocities are somewhat below this. As the current increases, the advantage of receiving more plankters is progressively outweighed by the difficulty of holding station (as was pointed out for Chromis punctipinnis in California by Hobson and Chess 1976). The relatively few adult diurnal planktivores that foraged where currents were weak probably owed their low numbers to the lack there during the day of calanoids and other zooplankters suit- able as prey (Table 2). The many zooplankters that tidal currents carried to planktivores elsewhere were unavailable to fishes here, and those taken as prey or otherwise lost were not quickly replaced. Although the volume of zooplankters collected at the weak-current site by day (Table 1) actually exceeded the volume at the strong-current site (Table 6), it consisted largely of swarming mysids (Table 2) which are local residents seemingly un- available as prey to diurnal planktivores (possibly for reasons discussed below under Miscellaneous Considerations). The strong-current site was in fact much richer in copepods, caridean larvae, lar- vaceans, and fish eggs — the major prey of the diurnal planktivores (compare Tables 2 and 7). Locations in the lee of reefs, however, can be rich in drifting debris from these reefs (Gerber and Marshall 1974). This situation existed at the Walt Island site, where Pomacentrus pavo and P. vaiuli, the most numerous diurnal planktivores there, subsisted largely on algal fragments. Further- more, the only other diurnal planktivores numer- ous in weak-current areas, Amblyglyphidodon curacao and Dascyllus reticulatus, demonstrated a capacity to utilize algae even though both species are largely carnivorous. Gerber and Marshall (1974), too, found that D. reticulatus fed on algal fragments when zooplankters were sparse. Obvi- ously, the capacity to utilize algae as food is highly adaptive for planktivorous fishes that would live where drift from a reef is rich in algal fragments, though relatively poor in zooplankters (Table 1). Despite the adaptiveness of herbivory to plank- tivores under these circumstances, most of the fishes studied by us were strictly, or predomi- nantly, carnivores. Drifting algal fragments were plentiful in nearly all nearshore habitats, but where zooplankters were also numerous the algae were insignificant in the diets of most plankti- vores. To be sure, certain species capitalized on drifting algae even where zooplankters were numerous. For example, P. vaiuli, which we fre- quently observed plucking items from the water column, was herbivorous and numerous at the zooplankton-rich Bogen Island site, just as it was at the zooplankton-poor Walt Island site. And P. coelestus, which may have replaced P. pavo where currents were strong, fed heavily on algal frag- ments where zooplankters were readily accessible. Yet the pattern is clear — zooplankters were fa- vored by most. Generally Chromis spp. have been reported as strictly carnivores even where other planktivorous pomacentrids fed substantially on drifting algae (e.g., in the Marshall Islands by Hiatt and Strasburg 1960; in the West Indies by Randall 1967; and in Hawaii by Hobson 1974). Nevertheless, species of Chromis display some capacity to accept algal fragments, as we found in C. margaritifer and Gerber and Marshall (1974) found in C caerulea. Thus, where waters are rich in reef debris but poor in zooplankters, we should expect to find Chromis spp. in relatively low num- bers, just as we did at Walt Island. On the other hand, Mirolabrichthys pascalus (a serranid) and Pterocaesio tile (a lutjanid) are members of strictly carnivorous families, a fact that probably limits them to places adequately supplied with zoo- plankters. This view finds support from Gerber and Marshall (1974), who reported that M. pas- cales (as M. tuka) and P. tile fed entirely on zoo- plankters. They noted the same for A. curacao, C. agilis, and C. lepidolepis but did not indicate where any of these fishes had been collected, nor whether anything but zooplankters had been available to them. This may be important because one of their major stations was in East Channel, where their plankton collections were without al- gae, and though they found A. curacao strictly carnivorous, we found that it fed heavily on algal fragments where zooplankters were in short sup- ply (Table 4). Gerber and Marshall also noted that P. vaiuli fed mainly on algal fragments while 147 coocurring pomacentrids concentrated on zoo- plankters, but concluded from this that the species is a benthic grazer. Nocturnal Relationships Nocturnal planktivores probably concentrated where currents were weak because their prey — including polychaetes, large calanoids, mysids, isopods, gammarids, postlarval carideans, and brachyuran megalops — were most numerous there (Table 2). With the probable exception of at least most of the calanoids (see below), most of these zooplankters were local residents that rose into the water column at night after spending the daytime sheltered on or near the sea floor. This pattern has been adequately documented among these groups of organisms from both Atlantic and Pacific Oceans (Emery 1968; Williams and Bynum 1972; Alldredge and King 1977), and its impor- tance in shaping the activities of nocturnal planktivorous fishes has been stressed (Hobson 1968, 1972, 1974; Hobson and Chess 1976). Food- habit studies have shown that these groups in- clude the major prey of apogonids, holocentrids, and other tropical nocturnal planktivores (Atlan- tic Ocean: Randall 1967; Indian Ocean: Vivien 1973, 1975; and Pacific Ocean: Hobson 1974). Only a relatively few nocturnal planktivorous fishes occurred where currents were strong, prob- ably because prey suitable to them were relatively scarce there (Table 7). Many of the organisms on which these fishes feed most likely find conditions in places with strong currents adverse. For exam- ple, those nocturnal zooplankters that return each morning to shelter in specific habitats would likely be transported to foreign surroundings should they encounter strong currents while in the water column. The mysids, which include some of the strongest swimmers, probably cannot hold sta- tion in currents much over 15 cm/s (based on the maximum swimming speeds of several species: Steven 1961; Clutter 1969) and currents at the Bogen Island station regularly exceeded this six- fold. Organisms that need to spend only a few hours in the water column each night might time their emergence to avoid currents, as pointed out by Alldredge and King (1977), but probably even these would find it advantageous to live without this complex timing problem. Furthermore, many of these nocturnal forms rest in sediments by day (Hobson and Chess 1976; Alldredge and King 1977) and might find the coarse, unstable sand 148 FISHERY BULLETIN: VOL. 76, NO. 1 characteristic of strong-current areas unfavor- able. Only part of the increased numbers of zoo- plankters at night were suitable prey of the noc- turnal planktivores. These were individuals more than about 2 mm long, which predominated among the nocturnal visitors at the weak-current site but which were a much smaller segment of the zooplankters that appeared after dark at the strong-current site. Among calanoids, for exam- ple, only individuals longer than 2 mm (mostly Euchaeta marina, Pleurommama xiphias, and Undinula vulgaris) were important prey of such larger nocturnal planktivores as Myripristis spp., and while these larger calanoids were never seen or collected by us at the weak-current site during the day, they were more numerous than the small- er ones at that station after dark (Table 3). On the other hand, most of the dramatic increase in calanoids at the strong-current site involved only slightly larger individuals of essentially the same species that were there by day, including Acartia sp., Candacia sp., and E. marina (Table 8), and these were largely unexploited by nocturnal planktivores. At 3 mm or less, the majority may be too small to be taken by the relatively large mouths of most of the nocturnal fishes considered here (see Hobson and Chess 1976), although they were important prey of some of the smaller species, such as Apogon nouaeguinae. The daytime location of the many calanoids which appear above the reefs at night remains in question. Our nearshore plankton collections in southern California (Hobson and Chess 1976) showed far less increase in calanoids after dark, and we concluded they were in the nearshore water column day and night. But the dramatic increase in calanoids nearshore after dark at Enewetak suggests a different situation. We rec- ognize one or a combination of two possibilities: 1) that some calanoids reside under shelter on the sea floor by day, and join the plankton at night, or 2) that some calanoids reside elsewhere by day, and migrate, or are transported, to the nearshore waters only after dark. There is evidence for both possibilities. The large calanoids that swarmed around our lights shortly after last evening light (but not taken in our collections) could not have traveled far. Alldredge and King (1977) reported calanoids emerging at night from nearshore benthic substrata on the Great Barrier Reef in numbers that could readily account for the in- crease in calanoids we observed after dark at HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES Enewetak; but there may be a problem with All- dredge and King's sampling technique. Their sam- ples were taken with Plexiglas traps that rested on the bottom and collected zooplankters that rose into the water column at night; however, there were gaps between the rigid lower edges of these traps and irregularities on the sea floor. Conceiva- bly, as Alldredge and King themselves recognized, the samples could have included swimming or- ganisms from the base of the surrounding water column that entered the traps through these gaps. These collections need to be repeated with this possibility for error eliminated. While it would be surprising if the numbers of calanoids they col- lected had actually entered the traps through these gaps, we are concerned that the only calanoid identified in their samples, Acartia spp. (listed as cyclopoids), are of a genus known to include species that are exceedingly numerous in the water column during both day and night (e.g., Emery 1968; Hobson and Chess 1976). We would expect organisms that live in the substrate by day to have morphological features reflecting this habit that distinguish them from holoplanktonic relatives at the generic level or higher. So al- though there may have been nearshore residents among the calanoids whose numbers sharply in- creased after dark at Enewetak, we believe that at least most of them, especially the larger ones, ap- peared following regular movements from deeper water. The calanoids that visited the nearshore waters after dark seemed to be part of a nocturnal move shoreward made by many zooplankters, including chaetognaths and larval fishes. Because each of our primary collecting sites probably received noc- turnal visitors from different sources, the two are discussed separately. Walt Island Perhaps some of the nocturnal plankters that visited the weak-current site were carried from the open sea by the turbulent flow of water that crossed the interisland reef at higher tides, but this would have been a hazardous transit for most zooplankters, and we doubt that significant num- bers came this way. If many had come by this incidental route, at least some would still have been there al daybreak — probably somewhat dis- oriented in these foreign surroundings. But they were always gone by early morning twilight, suggesting they followed a well-established pat- tern with consistent and predictable arrivals and departures. Probably most of the nocturnal plankters that visited Walt Island came from the deeper waters of the lagoon, moving over the lagoon's shallow periphery as part of a regular nocturnal rise into the surface waters. The general rise of zoo- plankters at night in lagoons of the Marshall Islands has been documented (at Bikini by Johnson 1 949; and at Majuro by Hobson and Chess 1973). It has also been noted that by day the mid- lagoon is much richer in zooplankters than is the shallow periphery (Gerber and Marshall 1974), but a shoreward movement among zooplankters at night would reduce this difference between the two regions. Probably it is widespread that zoo- plankters rising from the depths at night spread out over shallow water near shore. At Kona, Hawaii, where great depths lie adjacent to a coast- al shelf (see Hobson 1974), one of us (E. Hobson) often observed myctophids (lanternfishes), and other deep-water forms, in <5 m of water close to shore after dark (unpubl. obs.). Swimming to the Walt Island site from the deeper water of the lagoon would usually entail moving against the drift from the reef. Although comparatively weak, this current would neverthe- less obstruct small or weak-swimming forms. The nocturnal shoreward movement of zooplankters at this location, then, would favor the larger, stronger swimming components of the plankton — forms like chaetognaths, larval fishes, and the larger calanoids. Likely for this reason most of the calanoids among the increased num- bers of zooplankters at Walt Island were >2 mm (Table 3), whereas at Bogen Island, where zoo- plankters were carried by currents, most of a much greater number were 1 to 2 mm long (Table 8). Distinction between the two locations is important because it is the larger zooplankters that were important prey of the nocturnal planktivores. Of course, the upcurrent swim from deeper water would take even the most mobile zooplankters some time. Thus, it is significant that larger calanoids were absent in the plankton collections made at Walt Island 1 h after last evening light, but were numerous in the collections made here at midnight and later (Table 3). Bogen Island We presume that most of the zooplankton col- lected in the flooding tidal currents at Bogen Is- 149 FISHERY BULLETIN: VOL. 76, NO. 1 land had been carried in through East Channel from outside the lagoon — ^just as during the day. The greatly increased numbers at night probably followed a general rise of zooplankton toward the surface waters in the open sea. Some of these zoo- plankters were larger than any that were present by day, but such forms represented a lesser propor- tion of the nocturnal plankton here than they did at the weak-current site. Presumably the collec- tions also included lagoon organisms from up- stream, but we would expect these to be relatively few because the entrance to East Channel is only about 1.2 km away (Figure 1). Although the in- coming tidal currents probably carried materials that had been transported from the lagoon on ear- lier ebb tides, we would not expect many of the larger mobile organisms to be among them. Most large mobile forms, it would seem, could avoid being transported from the lagoon by the com- paratively weak outgoing currents. But certainly the incoming tide could be returning substantial numbers of passive drifters, like fish eggs and algal fragments, in addition to forms like the smaller calanoids. In any event, we can under- stand the relative scarcity in the flooding tidal currents of the relatively large nearshore resi- dents (e.g., polychaetes, mysids, and postlarval carideans) that are so important in the diets of nocturnal planktivores. Probably at least some zooplankters from the deeper waters of the lagoon visited the Bogen Is- land site at night during periods between flooding tides, but we made no collections at these times. Nevertheless, it would seem that the impact of such forms on the area would be limited, consider- ing how long it takes them to travel without ben- efit of transport by current, and the fact that a flooding tide sweeps through here during much of most nights. Miscellaneous Topics The Nocturnal Increase in Fish Eggs Planktonic fish eggs represent a special case. Unlike most other zooplankters, which are mobile forms that strongly influence their own distribu- tions, fish eggs are passive drifters that are quickly carried from where they are released if there is any current. Presumably their relative numbers in the water column closely follow the incidence of their release by fishes on the reefs below, and certainly the circumstances of this re- 150 lease have been strongly influenced by the threat from predators that abound over the nearshore reefs. Planktonic fish eggs were a major food of diurnal planktivores (Tables 4, 9) but, despite an almost sevenfold increase in numbers at night (Tables 2, 7), they were insignificant in the diets of nocturnal planktivores (Table 5). Clearly these largely transparent eggs are relatively safe from predatory fishes after dark, probably because they are then invisible. Thus, it would be highly adap- tive for reef fishes to release planktonic eggs late in the day, or early in the night, when the eggs have maximum time for dispersing in the dark, relatively free of threat from planktivorous reef fishes. Possible Influences of Water Depth and Size Among the promising topics we lacked time to pursue during our short stay at Enewetak were ways that water depths, and the sizes of interact- ing fishes and zooplankters, may influence trophic relationships. We believe that the difference in water depth between our primary collecting sites (7 vs. 13 m) did not significantly influence our findings, espe- cially as the deeper station was well away from the deep part of the lagoon (Figure 1) — farther, in fact, than the shallower station. It was apparent to us, nevertheless, that water depth in the lagoon can, directly or indirectly, influence fish-zooplankton interactions. Obviously both fishes and zoo- plankters are physically limited in extreme shal- lows, especially in turbulent waters above shallow reefs. But probably the major depth-related influence stems from the general tendency of la- goon zooplankters to seek deeper water during the day (e.g., Johnson 1949; Hobson and Chess 1973) — a tendency that apparently increases with size. We suggest above that many of the larger zooplankters active above the nearshore shelf at night were in the deeper lagoon waters by day, when the water column of the nearshore shelf was largely without such forms. Perhaps the concen- trations of planktivores along the outer edge of the nearshore shelf during the day were in contact with the fringe of these deep zooplankton popula- tions. This leads to a possible influence related to size. Very small zooplankters (those passing through the mesh of our net, and so unrepresented in the collections), and their predators among juvenile and larval fishes, may follow patterns sig- HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES nificantly different from patterns followed by the larger forms studied here. The zooplankters de- scending into the depths by day tend to be the larger individuals, so we wonder where the very small ones are located. In sharp contrast to the relatively few adult planktivores active in weak- current areas of the nearshore shelf by day, large numbers of juvenile and larval fishes (Figure 4) clearly found planktonic food abundant. It may be that very small zooplankters, unsampled by our net and too small to be taken by most adult plank- tivores, remain numerous in shallow weak- current areas during the day. Mysids as Prey During the Day It is striking that when mysids swarm in dense numbers near many reefs during the day they are relatively unimportant as prey of the major planktivorous fishes. They seem to escape the in- terest not only of diurnal planktivores, but also of the many nocturnal planktivores (e.g., Myripristis spp. ) that hover within easy reach close among the coral. To be sure, a number of the fishes we studied took some of these mysids by day. Chromis caeru- lea, C. agilis, Dascyllus reticulatus, and Poma- centrus pavo included mysids as minor compo- nents of their diet at the weak-current site. Furthermore, Hiatt and Strasburg( 1960) reported that C. atripectoralis preyed significantly on mysids. But considering the preponderance of mysids in the water column at so many places during the day, these fishes took only token num- bers. Probably the relatively large size of the mysids is important in this context. The evolution of feed- ing morphologies in diurnal planktivores appears to have been determined by strong selective pres- sures to take tiny prey (Davis and Birdsong 1973; Hobson and Chess 1976). Significantly, most of the zooplankters taken by these fishes (e.g., copepods, larvaceans, and fish eggs) were <2 mm long, and the size range of mysids that swarmed around these reefs in daylight was 2 to 8 mm (Tables 2, 7). In reporting a similar situation in the tropical Atlantic Ocean, Emery (1968) speculated that planktivorous pomacentrids fail to prey on swarm- ing mysids because normally these fishes feed on smaller prey. The failure of Myripristis spp. and other large- mouthed nocturnal planktivores to exploit this diurnal resource cannot be attributed to the size of the mysids, however, because these fishes find the same mysids major prey at night. Apparently the nocturnal fishes simply do not react to these read- ily accessible mysids as prey during daylight. In warm-temperature waters of southern California the large juvenile olive rockfish, Sebastes serra- noides, feeds primarily on zooplankters after dark, but during the day sometimes preys on mysids that are within reach of the rockfish where it hov- ers in relatively inactive diurnal schools (Hobson and Chess 1976). However, predominantly noc- turnal habits seem to be characteristic of the olive rockfish only during its large juvenile stage — both before and after this stage it feeds mainly by day (Hobson and Chess 1976). Therefore, even at that time of its life when the olive rockfish feeds primarily at night, we should not expect it to be as strongly nocturnal as Myripristis spp. and the other more specialized nocturnal forms that ig- nore mysids by day at Enewetak. Possibly swarming mysids are protected from predators by the nature of their aggregations. Emery ( 1968) noted that mysid swarms respond to predators just as fish schools do. The analogy can be expanded. Like these nocturnal mysids, many nocturnal fishes congregate in dense numbers above the reef during the day, and at this time they too are relatively undisturbed by the many predators at large in the same area (Hobson 1965, 1968). It is widely believed that fishes are less vulnerable to predators when they aggregate (e.g., Bowen 1931; Springer 1957; Brock and Riffen- burgh 1960; Manteifel and Radakov 1961; Wil- liams 1964). Of the many theories that would ex- plain this circumstance, we favor the existence of a confusion effect, as advocated by Allen ( 1920) and others. This theory suggests that visually orient- ing predators which select individual prey have trouble singling out a target among the many al- ternatives they confront in an aggregation. That mysids achieve some safety from predators by ag- gregating is further supported by the experiments of Welty (1934), who found that goldfish, Caras- sius auratus, consumed fewer daphnia when these prey were concentrated. (These comments apply as well to the relative lack of diurnal predation on larval fishes, which, in their dense schools close to the reef, resembled swarming mysids.) Planktivore Morphology and Their Distance From the Reef It was suggested earlier (Hobson 1974) that in 151 their tendencies toward more fusiform bodies and deeply incised caudal fins, diurnal planktivores have acquired added speed that is adaptive in quickening their return to reef shelter when threatened. Expanding this suggestion, these fea- tures are more developed in planktivores that swim farther from the reef because threats from predators probably increase in more exposed loca- tions. Although morphology that permits faster swimming would also enhance holding station in a current, we believe the major selection pressures shaping these features in planktivores have come from predators. Despite the obvious adaptiveness of fusiform bodies and of deeply incised caudal fins in many planktivores, the morphologies of certain other highly successful diurnal planktivores have taken the opposite course. For example, among the fishes we studied, Dascyllus reticulatus (Figure 7) and Amblyglyphidodon curacao are among the deepest bodied of pomacentrids, and yet they range farther into the water column than the species ofChromis or Pomacentrus. Similarly, the many planktivor- ous chaetodontids in Hawaii (e.g., species of Chaetodon and Hemitaurichthys), all deep-bodied forms with truncate caudal fins, are highly suc- cessful planktivores that range widely in the water column (Hobson 1974). We suggest that whereas fusiform bodies in- crease the chance of eluding predators, deep bodies increase the chance of discouraging predators. The basis of this second suggestion is the fact that piscivores live with the danger of choking on spiny-rayed prey lodged in their pharynx or esophagus. Over the years we have seen many predators in this predicament — often fatally. Pis- civores generally swallow their prey head-first, frequently after manipulation to ensure proper orientation. Reasons for not swallowing a spiny- rayed fish tail-first are obvious. Assuming, then, that a prey fish is swallowed head-first, the danger of it becoming lodged in the pharynx or esophagus increases with its depth or width. Thus, predators equipped to take prey from among the variety of planktivores in the water column (where those at a given level tend to be about the same length) would find greater risk ingesting deeper bodied forms, especially those with prominent fin spines. Of course, this advantage of a deep body and prom- inent spines in thwarting predators extends beyond planktivores; the entire family Chaeto- dontidae, for example, would benefit (Hobson and Chave 1972). 152 FISHERY BULLETIN: VOL. 76, NO. 1 ACKNOWLEDGMENTS We thank Stephen V. Smith, Director, and Laboratory Mangers Philip and Janet Lamberson, of the Mid Pacific Marine Laboratory at Enewetak Atoll, for making facilities available to us. The laboratory is supported by the Division of Biomed- ical and Environmental Research of the U.S. Energy Research and Development Administra- tion and is operated as an extension of the Hawaii Institute of Marine Biology, University of Hawaii. For constructive criticism of the manuscript we thank Carl L. Hubbs and Richard H. Rosenblatt, Scripps Institution of Oceanogrphy; William M. Hamner, Australian Institute of Marine Science; Robert E. Johannes, Hawaii Institute of Marine Biology; and William Lenarz, Tiburon Labora- tory. John E. Randall, Bernice P. Bishop Museum, Honolulu, identified Mirolabrichthys pascalus; Kenneth Raymond, Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, drew Figure 1 and the fishes in Figure 6; and Alice Jellett, Tiburon Laboratory, typed the manuscript. LITERATURE CITED ALLDREDGE, A. L., AND J. M. KING. 1977. Distribution, abundance, and substrate perferences of demersal reef zooplankton at Lizard Island Lagoon, Great Barrier Reef Mar. Biol. (Berl.) 41:317-333. Allen, W. E. 1920. Behavior of loon and sardines. Ecology 1:309-310. Bowen, E. S. 1931. The role of the sense organs in aggregations of Ameiurus melas. Ecol. Monogr. 1:1-35. BROCK, V. E., AND R. H. RIFFENBURGH. 1960. Fish schooling: A possible factor in reducing preda- tion. J. Cons. 25:307-317. Clutter, R. L 1969. The microdistribution and social behavior of some pelagic mysid shrimps. J. Exp. Mar. Biol. Ecol. 3:125- 155. Davis, W. P., and R. S. Birdsong. 1973. Coral reef fishes which forage in the water col- umn. Helgol. wiss. Meersunters. 24:292-306. Emery, a. R. 1968. Preliminary observations on coral reef plank- ton. Limnol. Oceanogr. 13:293-303. Gerber, R. p., and N. Marshall. 1974. Ingestion of detritus by the lagoon pelagic commun- ity at Eniwetok Atoll. Limnol. Oceanogr. 19:815-824. HIATT, R. W., AND D. W. STRASBURG. 1960. Ecological relationships of the fish fauna on coral reefs of the Marshall Islands. Ecol. Monogr. 30:65-127. HOBSON, E. S. 1965. Diurnal-nocturnal activity of some inshore fishes in the Gulf of California. Copeia 1965:291-302. 1968. Predatory behavior of some inshore fishes in the HOBSON and CHESS: TROPHIC RELATIONSHIPS AMONG FISHES Gulf of California. U.S. Fish Wildl. Serv., Res. Rep. 73, 92 p. 1972. Activity of Hawaiian reef fishes during the evening and morning transitions between daylight and darkness. Fish. Bull., U.S. 70:715-740. 1974. Feeding relationships of teleostesin fishes on coral reefs in Kona, Hawaii. Fish. Bull., U.S. 72:915-1031. 1975. Feeding patterns among tropical reef fishes. Am. Sci. 63:382-392. HOBSON, E. S., AND E. H. CHAVE. 1972. Hawaiian reef animals. Univ. Press Hawaii, Hon- olulu, 135 p. HOBSON, E. S., AND J. R. CHESS. 1973. Feeding oriented movements of the atherinid fish Pranesus pinguis at Majuro Atoll, Marshall Islands. Fish. Bull., U.S. 71:777-786. 1976. Trophic interactions among fishes and zooplankters neeir shore at Santa Catalina Island, California. Fish. Bull., U.S. 74:567-598. JOHNSON, M. W. 1949. Zooplankton as an index of water exchange between Bikini Lagoon and the open sea. Trans. Am. Geophys. Union 30:238-244. Manteifel, B. p., and D. V. RADAKOV. 1961. The adaptive significance of schooling behaviour in fishes. Russ. Rev. Biol. 50:338-345 (Engl, transl. from Russ.). Randall, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. (Miami) 5:665-847. SCHULTZ, L. P., W. M. Chapman, E. A. Lachner, and L. P. Woods. I960. Fishes ofthe Marshall and Marianas islands. Bull. U.S. Natl. Mus. 202(2), 438 p. SCHULTZ, L. p., E. S. HERALD, E. A. LACHNER, A. D. WELAN- der, and l. p. Woods. 1953. Fishes ofthe Marshall and Marianas islands. Bull. U.S. Natl. Mus. 202(1), 685 p. Springer, S. 1957. Some observations on the behavior of schools of fishes in the Gulf of Mexico and adjacent waters. Ecology 38:166-171. STARCK, W. a., II, AND W. P. DAVIS. 1966. Night habits of fishes of Alligator Reef, Flori- da. Ichthyol. Aquarium J. 38:313-356. Steven, D. M. 1961. Shoaling behaviour in a mysid. Nature (Lend.) 192:280-281. VIVIEN, M. L. 1973. Contribution a I'etude de I'ethologie alimentaire de I'ichtyofaune du platier interne des recifs corralliens de Tulear (Madagascar). Tethys, Suppl. 5:221-308. 1975. Place of apogonid fish in the food webs of a Malag£isy coral reef. Micronesica 11:185-198. VON ARX, W. S. 1948. The circulation systems of Bikini and Rongelap la- goons. Trans. Am. Geophys. Union 29:861-870. WELTY, J. C. 1934. Experiments in group behavior of fishes. Physiol. Zool. 7:85-128. WILLIAMS, A. B., AND K. H. BYNUM. 1972. A ten-year study of meroplankton in North Carolina estuaries: Amphipods. Chesapeake Sci. 13:175-192. WILLIAMS, G. C. 1964. Measurement of consociation among fishes and comments on the evolution of schooling. Mich. St. Univ. Mus., Biol. Ser. 2:349-384. 153 SPAWNING CYCLE, FECUNDITY, AND RECRUITMENT IN A POPULATION OF SOFT-SHELL CLAM, MYA ARENARIA, FROM CAPE ANN, MASSACHUSETTS Diane J. Brousseau* ABSTRACT A population ofMya arenaria in the Annisquam River system, Gloucester, Mass. , was studied for 3 yr to determine spawning frequency, fecundity, and recruitment rates under natural conditions. This population was observed to spawn twice each year, in March-April and June-July. Temperature appeared to be a more critical factor in the timing of gonad maturation than in triggering the release of gametes. Female body sizes and oocyte production were positively correlated (1973, r = 0.95; 1974, r = 0.90). Regression lines were compared by analysis of covariance. Slopes of the lines did not differ significantly between years or between spawning cycles within years (P 3^0.05). Elevations of the lines differed significantly from one another (P«0.05) indicating annual and seasonal variability in fecun- dity. Sex ratios of Af. arenaria 25-95 mm shell length did not differ significantly from 1:1 over the 3-yr study period. In smaller individuals, male and female gonads were indistinguishable. No evidence of hermaphroditism or protandry was observed. Recruitment rates of juveniles fluctuated widely between spawning cycles as well as between years. Although the literature contains widely scattered references to the reproductive cycle of Mya arenaria in New England, there is no combined account of egg production ( = fecundity), spawn- ing, and recruitment of this species under natural conditions. Inferences about the time and fre- quency of spawning by M. arenartia have been made from observations on larvae in the plankton (Stevenson 1907; Stafford 1912; Sullivan 1948; Landers 1954; Pfitzenmeyer 1962); from first ap- pearances of newly settled juveniles (Belding 1930; Warwick and Price 1975); and from the presence of ripe gametes in the gonads (Battle 1932; Coe and Turner 1938; Shaw 1962; Stickney 1963; Ropes and Stickney 1965; Munch-Peterson 1973; Porter 1974). Observations on larvae and recently metamorphosed clams, however, are use- ful only as indirect measures of the frequency and duration of spawning, since larval abundance and juvenile recruitment are controlled by factors other than spawning alone. Conversely, evidence concerning gonad maturation and gamete release obtained by means of histological methods defines the spawning period without contributing to knowledge about recruitment. 'Department of Biology, Fairfield University, Fairfield, CT 06430. Manuscript accepted July 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. Most shallow-water marine animals reproduce in a cyclic manner, the time of spawTiing ulti- mately depending on environmental factors (Or- ton 1920; Giese 1959; Kinne 1963). As with most other commonly studied bivalves, the timing of spawning by M. arenaria has been linked to water temperatures (Nelson 1928; Belding 1930; Battle 1932). Nevertheless, it remains unclear whether gametogenesis, spawning, or both occur at a specific temperature or in a specific temperature range in M. arenaria. Reliable information on fecundity of M. arenaria is also unavailable. Laboratory methods for stripping eggs or inducing spawning in oysters and hard-shell clams (Brooks 1880; Churchill 1920; Galtsoff 1930; Belding 1930; Davis and Chanley 1956; Loosanoff and Davis 1963) are gen- erally unsuccessful with M. arenaria. Conse- quently, the only information on egg production by M. arenaria is an unsupported statement by Belding (1930) that a 2.5-in clam (63 mm) pro- duces about 3 million eggs per breeding season. In an effort to clarify the breeding habits of M. arenaria, this study was designed to determine 1) the reproductive cycle in a natural population, 2) the temperature at which gametogenesis and spawning begin in this locale, and 3) the total numbers of eggs produced by individuals of differ- ent sizes. 155 FISHERY BULLETIN: VOL. 76, NO. 1 MATERIALS AND METHODS The Annisquam River is a natural waterway approximately 3 mi long connecting Ipswich Bay on the north side of Cape Ann peninsula with Gloucester Harbor on the south (Figure 1). The 70°40' 42°40' Figure l. — Map showing locations of the Jones River study site (A) and the University of Massachusetts Marine Station, Hodgkins Cove (B). river consists of a dredged channel with extensive tidal mud flats or shallow water on both sides. The mean tidal amplitude at Gloucester Harbor is 3 m. The Annisquam River receives limited freshwater drainage, resulting in salinities of 28-33. 5%o. Water movement is largely dependent on the tides. Average monthly surface water tempera- tures (1 m depth) for the years 1973 and 1974 obtained from the University of Massachusetts Marine Station at Hodgkins Cove (Figure 1) indi- cate that monthly temperature fluctuations are great (Figure 2). Temperature data for 1975 were not available. The site for this study was located on a mudflat along the west bank of the Jones River, a small tributary opening at the northern end of the An- nisquam River (Figure 1). Historically, this area has been the site of a productive shellfish bed and is known to sustain numerous clams of differing age classes (Mass. Dep. Resour., Div. Mar. Fish, pers. commun.). The study began in February 1973 and was completed in October 1975. Clams were collected from the middle of the intertidal zone ( +1 m tidal level) once a month from October 1973 through February 1974 and October 1974 through October 1975, and twice a month from March through Sep- tember 1973 and April through August 1974. No samples were taken in September 1974 or in May, June, July, and September 1975. Sample sizes var- ied greatly. Samples collected during the spring and summer months consisted of 30 to 127 clams, 21-90 mm shell length. Those collected during the winter months consisted of 15 to 30 clams each in a similar size range. Large numbers of clams were 20- ' "T 1 1 1 1 1 1 r ^ •" M A M J J A S FIGURE 2.— Sea-surface (1 m depth) temperatures for Hodgkins Cove, Gloucester, Mass. Monthly means for 1973 (••) and 1974 (oo) are plotted. The dashed lines represent 8-yr average maxima, means, and minima for the period 1963-71, based on temperatures for the Portland Lightship (Chase 1965-1973) corrected for Hodgkins Cove. MONTHS 156 BROUSSEAU: MYA REPRODUCTION AND RECRUITMENT collected during the spawning season in order to insure sufficient numbers of "ripe" females for fe- cundity studies. A total of 2,480 clams were examined of which 11% were immature, leaving 2,206 mature clams that were used in the analysis of the reproductive cycle. The samples were returned to the laboratory where they were kept at 0°C for not more than 3 days before being dissected. Each clam was num- bered and its maximum length ( ±0.1 mm) deter- mined. The visceral mass (gonad, liver, and gastrointestinal tract) was taken out and fixed in 10% buffered Formalin^ (Humason 1967). The displacement volume of each visceral mass was taken to determine its size. The amount of gonadal tissue present was determined after sectioning by the planimetry method described below. The fixed mass was then dehydrated in alcohol, embedded in paraffin, sectioned at 8 fxm, and stained in Harris' hematoxylin and eosin. Each clam was classified with respect to gonad development and the number in each developmental stage was recorded for both sexes. Previous studies on the gonadal cycles of Mya arenaria have divided the developmental se- quence into five morphologically distinct phases: inactive, active ripe, spawning, and spent (Ropes and Stickney 1965; Porter 1974). Since semantic problems arise with this usage, several terms are redefined for use here! The term "indifferent" is preferred to "inactive" to describe low levels of oogenic and spermiogenic activity. As pointed out by Keck et al. ( 1975) in work on hard clam gonadal cycles, the term "inactive" is biologically in- accurate since it implies a "static condition where absolutely no morphological or biochemical activ- ity is proceeding." The term "developing" is used when describing the onset of gametogenesis since it can be argued that ripe and partially spawned gonads are active in the sense that gametogenic activity continues at a reduced level. Developing, ripe, and partially spawned stages are collectively termed "active," whereas spent and indifferent stages are termed "inactive." This distinction aids in defining peaks of spawning within the annual cycle. Recognition of the five phases of gonadal condi- tion was based on the same characteristics as those used by other investigators (Ropes and Stickney 1965; Porter 1974). ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. The number of oocytes present in each female gonad was determined in the following manner. Using an ocular grid, triplicate counts were made of the number of oocytes present per 0.49 mm^ of gonad for each female reported in a ripe condition. This area was then multiplied by the mean oocyte diameter (0.65 mm) in order to determine oocyte densities on a cubic basis. An estimate of the total number of oocytes in the gonad could then be cal- culated on the basis of gonad size. Analysis of variance confirmed that the number of oocytes per unit volume was constant throughout the ripened gonad (P^0.05). Mean oocyte diameter was determined for a rep- resentative sample of ripe females, selected at random from each of the reported spawning periods. Twenty oocytes per clam were measured using an ocular micrometer. Only those oocytes which were spherical in shape and ready for re- lease were selected for measurement. The relationship between the size of the ripe female gonads and the volume of the total visceral mass was determined as follows. Entire viscera from 17 ripe females (53-76 mm shell length) were sectioned at 12 /u,m. Next, 18 sections from each individual were chosen at random, mounted on a Plexiglas base and fitted into a 35-mm slide projec- tor and the projected tissue outlines were traced. A planimeter was used to estimate the percentage of gonad tissue present. A correction factor repre- senting the proportion of gonad in the total vis- ceral mass was used in estimating the total number of oocytes per individual (0.763 ±0.21, 95% C.I.). Photographs of representative stages of the female reproductive cycle were taken with a light microscope at 160 x and 100 x magnification using a 35-mm camera. High contrast, Panatomic X ASA32 film was used. Densities of juvenile M. arenaria were tab- ulated from the monthly samplings of the tidal flat during October 1973, from May to November 1974, and in November 1975. At each sampling period, 12 random samples (0.11 m^, 20 cm deep) were taken along a 90-m transect from mean low water shoreward to the marsh scarp. Samples were wet seived in the field (2-mm mesh) and the size-frequency distribution of the clams was de- termined. Cohorts in the population were isolated by the probability paper method (Harding 1949; Cassie 1954). 157 u a: 1973-1974 MONTHS FISHERY BULLETIN: VOL. 76, NO. 1 RESULTS Reproductive Cycle Reproductively active individuals were encoun- tered throughout the 3-yr study period, the largest numbers occurred in April and July of 1973, March and early July of 1974, and mid-March of 1975 (Figure 3). Due to the limited sampling undertaken during the summer of 1975, the sum- mer spawning peak cannot be determined with certainty. In February 1974, gametogenesis had begun in both sexes (Figure 4). Ripe and partially spawned clams were observed in mid-March. By late April, Figure 3. — Proportions o^Mya arenaria population with active or inactive gonads during 1973-74, 1974-75, and 1975-76. Cross-hatched portions of each bar represent inactive gonads (indifferent, no gamteogenesis, or spent); solid portions repre- sent active gonads (developing, ripe gametes, or partially spent). Observations on males and females were combined. CI] SPAWNING nSPENT 0 FIGURE 4 158 F ■ M ' A ■ m' J ' J ' A SO N D I J F M A M J J ' A ' S ' 0 ' N ' D I J —Proportions of male and female Mya arenaria with gonads in each developmental phase during 1973-74 and 1974-75. BROUSSEAU: MYA REPRODUCTION AND RECRUITMENT about 75^ had completely spawned and returned to the indifferent condition. Gametogenesis usu- ally resumed after spawning, and by early June about one-quarter of the clams were again ripe and partially spawned. The presence of cytolyzed unspawned gametes in the summer samples sug- gested that the same individuals had also been ripe earlier in the year. Thus the observed spawn- ing pattern was due to repeated spawning by the same individuals rather than asynchronous spawning of individuals within the population. A similar spawning pattern was observed in 1973, except that gametogenesis did not begin until April (Figure 4) and the summer spawning peak occurred in July rather than late June-early July. The data for both years indicate a more or less consistent recovery period between repro- ductive cycles. The data for the 1975 season indi- cate that spring spawning occurred in March as it did in 1974, but the summer sampling intervals were too irregular to describe details of the sum- mer spawning. Nevertheless, occurrence of a summer spawning is confirmed by the gonad con- dition of the clams in the August sample. Photomicrographs of representative female stages in the spring and summer peaks of the annual cycle are shown in Figure 5. The pattern of development in the clams during the spring cycle differs from that of the later summer one. In the female, the spring cycle is characterized by rapid gametogenesis, resulting in smaller oocyte size and fewer numbers of oocytes produced per unit of gonad tissue (Table 1), so different density values were used for calculations of fecundity (gonad vol- ume X density) in different seasons. A significant seasonal difference in the diameter of ripe oocytes of female M. arenaria was detected using one-way analysis of variance (P^0.05). Similarly, male clams appear to undergo rapid maturation and produce fewer gametes than during the summer spawning. Fully ripe males were not encountered in any of the spring samples, however, spent males were numerous, indicating that spawning had taken place. Spring spawning may be a facultative event, characterized by rapid maturation and the subsequent utilization of the abundant food sup- ply that is available during the major phyto- plankton "bloom" that occurs nearshore during this period. Temperature is an important factor influencing the gonadal cycle in a variety of marine bivalves (Loosanoff 1937a, b; Landers 1954; Giese 1959; Carriker 1961; Ansell et al. 1964; Galtsoff 1964; Calabrese 1970). If temperature is indeed a factor in the onset of reproduction in M. arenaria as previously believed (Nelson 1928; Belding 1930), short-term temperature patterns in winter and early spring should correlate with the annual tim- ing of gametogenesis (Figure 4). In fact, tempera- tures during January-March 1974 averaged about 2° higher than during the same period of the pre- vious year (Figure 2) and gametogenesis began a month earlier than in 1973. The actual role of temperature in the timing of gamete release remains unclear. Spring spawning peaks occurred at surface water temperatures of 4°-6°C and summer spawnings at 15°-18°C. Al- though the interstitial water of exposed tidal flats warms up considerably during midday spring lows (Johnson 1965), it is unlikely that interstitial temperatures would be high enough to account for these differences. If these is a critical minimum temperature for spawning it is at or above 4°-6°C. No maximum limit can be discerned from these data. The role of rapid temperature change in triggering spawning as suggested by other au- thors (Battle 1932; Stickney 1963) has not been assessed here. Sex Ratios and Fecundity The reproductive potential of a population de- pends, in large part, on the number of fertile females and the number of young produced per female. The proportion of females in all size- classes in three large samples from the Jones River in 1973 (n = 1,266), 1974 (n = 859), and 1975 (n = 150) did not differ signiflcantly from one-half. In size-classes <25 mm, male and female gonads were indistinguishable. No evidence of hermaphroditism or protandry was observed. The number oocytes produced was found to in- crease exponentially with increasing female body size. The regression equations for oocyte numbers (O) versus female shell length (S) are: Spring 1973: log^^O = Summer 1973: logj^O^ Spring 1973: log^oO^ 1.45 + 3.29 1og,.S -1.29 + 3.28 1ogj^S -0.90 + 2.91 \og^^S Summer 1974: log,oO= -1.42 + 3.32 log^^S Comparison of the regression lines by analysis of covariance indicated that the lines were parallel (P 3=0.05) but the elevations of the lines were sig- nificantly different (P^0.05). Total oocyte produc- tion during 1973 was greater than during 1974. 159 FISHERY BULLETIN: VOL. 76, NO. 1 -\^ • r • s &i*«" fi- ^ii^A :•- € ^ U^ ^' . 9 \? % ^ 4 ^ 4te^ ^. <• ♦ «f "si r^ • ..^. 160 BROUSSEAU: MYA REPRODUCTION AND RECRUITMENT f • « ». #. #► '# # ••- ■ • CD 05 3 sg >< 5 O > « a) be a. C M a >> Q ft- CO Tl"' •-1 i-H < « -T « o o m CO — • s <^ > > Q o. 2*2 s'^ >> cs .. . ^ Oi o e' '-' O I ^CJ = ^ « o o c o. 01 J3 CO o CO (D — 43 _aj CO cs a, „ fS f2 CO — OJ T3 T3 .— I r- CO >> § t> CS ^2S O x lO CO cj - a. J2 >< CB ^ o Eb Tt ^ O (N -^ ft< -I in 0) E > o O in O in O I I E 0) > o z I I I I E E o u O r- ro 4 mnrvi. * « * « . ^ ^ - T ^ '■■"^* ^A.-rr) . . . . •>«»♦< m l»• r?: ::::■:: : ^ —-^^ •m o ^^^"^"^'^—-'-^^ X o o ■:::::: ^ 0> E o ♦- o. r in CM in in cvi in CM in ed C ^ H ^ -*^ IS * -§^ -s c .c CO c ^ I o § CO a> CO o 6 c o ■« IZ to 05 « T3 ^ CO a> "-H 6 ^ 01 Ol o ^ >. o . o CO ^ S; S ^ j: fe CO O M c " T ^- t; « -O >. Q <0 't? J- a "" II |1 o 2 .2 ^ -a cB &-"§ -j: III tfc CO m l2 c pa bB c •E o. en °f DUDU9JD 'lAj ^0 'ON to W w .2 as ? D — O « " 163 FISHERY BULLETIN: VOL. 76, NO. 1 and 1974 approximated 8-yr monthly means so the bimodal pattern was not an atypical response to above average temperatures. Nevertheless, the temperature patterns of this locale are probably influenced substantially by local topography. More than 60% of the total area of the Annisquam River system is <6 m deep, 30% being intertidal (Jerome et al. 1968). Early spring warming and fall cooling trends would be expected here due to these nearshore influences. Lastly, it appears that the bimodal spawning pattern emerging here may be typical of some populations of M. arenaria found as far north as Plum Island Sound. A spring set of juvenile clams occurs annually on intertidal flats in Ipswich, Mass., (Richard Sheppard pers. commun.) and large numbers of 2- to 4-mm clams appeared in the May and June samples of Smith et al. (1955). Such evidence indicates that a semian- nual pattern may be more prevalent in northern Massachusetts than once believed. Orton (1920) first noted that some animals in temperate regions spawn when the temperature exceeds a critical level characteristic of the species, while for others the rate of change is im- portant. Nelson (1928) reported 10°-12°C as the critical spawning temperature for M. arenaria; Belding (1930) reported the exceptionally high figure of 22°C. The data for Gloucester indicate that spawning can occur with equal likelihood at either of the supposedly critical temperatures pro- vided that the gonad is ripe. The significant tem- perature appears to be that at which maturation of the gonad occurs. Similar significance of matura- tion temperature had been reported for the oyster, Crassostrea virginica, by Loosanoff and Davis (1950). Gonadal oocyte counts provide an accurate mea- sure of fecundity in M. arenaria since all oocytes are stored in the gonad prior to spawning and nearly total evacuation takes place at spawning. The fecundity values for M. arenaria indicate that the largest females produce the largest number of oocytes. This increase is undoubtedly due to in- creased gonad size made possible by increased shell volume. Average oocyte production by a 60-mm clam during a single breeding season (two spawning periods) is about 120,000; lifetime pro- duction would be in the order of 1.5 x 10^ oocytes. Although fecundity of M. arenaria is large, as is typical of species with planktonic larvae (Thorson 1950), these estimates are considerably lower than early unsubstantiated ones for this species (Belding 1930), as well as those reported for other 164 marine bivalves such as Crassostrea virginica and the hard-shell clam, Mercenaria mercenaria (Galtsoff 1930; Davis and Chanley 1956). High fecundity, however, is offset by high mor- tality during pelagic life, metamorphosis, and early settlement. It appears that sources of mor- tality such as predation, disease, and bottom character are more critical factors in explaining fluctuations in recruitment than variability in fecundity rates or spawning frequency. The spawning cycles in which the greatest number of oocytes were released did not correlate with periods of highest recruitment. In terms of spat densities, spring recruitment in both years studied was higher than summer recruitment. Success of some year classes and failure of others indicate that fluctuations in clam populations are largely natural occurrences and may result from things other than fluctuations in the number of oocytes or the number of juveniles or byssus-stage young. Spawning times and fecundities of individual females are critical factors in determining first, what constitutes a satisfactory breeding stock and secondly, how to protect it. Numerous studies have been conducted on methods of improving soft-shell clam fisheries (Belding 1930; Turner 1949, 1950; Smith et al. 1955; Smith^). Regulatory efforts have ranged from predator control to establishment of legal size limits for clams, closed seasons, and re- stocking of barren flats. All this work has pro- ceeded in the near absence of basic information of the reproduction and population dynamics of the clam. The dwindling yields of clams on the New England coasts indicate the ineffectiveness of present regulatory procedures and the need for revised management practices. In Massachusetts, any clam over 2 in long (51 mm) may be harvested. In effect this practice maximizes the removal of the reproductively most valuable individuals in the population. Murphy ( 1968), using genetic models, has shown that adult longevity and iteroparity ( = repeated reproduc- tion) are important adaptations for population stability in species like M. arenaria which exist under conditions of uncertain preadult survival and relatively stable adult survival (Brousseau ^Smith, O. R. 1952. The results of experimental soft clam farming in Plum Island Sound, Massachusetts. Third annual conference on clam research, U.S. Fish and Wildl. Service, clam investigations, Boothbay Harbor, Maine, p. 46-48. Unpubl. rep. BROUSSEAU: MYA REPRODUCTION AND RECRUITMENT 1976). Consequently, long-term stability of the re- source is endangered by present harvesting prac- tices which reduce the normal 10-12 yr lifespan of M. arenaria to 2 yr. Revision of existing regula- tions to include protection of sufficient breeding stock may be an effective way of insuring the long-term stability of the resource and minimizing the harmful effects of human predation. ACKNOWLEDGMENTS I thank D. Fairbairn, D. C. Edwards, and C. J. Berg for critically reviewing this manuscript and providing useful comments and suggestions; Robert Knowles for technical assistance in the field; and the staff of the University of Mas- sachusetts Marine Station who provided me with laboratory space and logistic support. LITERATURE CITED Ansell, a. D., K. F. Lander, J. Coughlan, and F. a. Loos- more. 1964. Studies on the hard-shell clam, Venus mercenaria, in British waters. I. Growth and reproduction in natural and experimental colonies. J. Appl. Ecol. 1:63-82. Battle, H. 1. 1932. Rhythmic sexual maturity and spawning of certain bivalve mollusks. Contrib. Can. Biol. Fish., New Ser. 7:255-276. Belding, D. L. 1930. The soft-shelled clam fishery of Massachusetts. Commonw. Mass. Dep. Conserv., Div. Fish Game, Mar. Fish. Ser. 1, 65 p. Brooks, W. k. 1880. The development of the oyster. Contrib. Chesapeake Zool. Lab. Johns Hopkins Univ. N. rV:l-115. BROUSSEAU, D. J. 1976. Life history parameters of Mya arenaria (Pel- ecypoda: Mollusca) and the population consequences. Ph.D. Thesis, Univ. Massachusetts, Amherst, 151 p. Calabrese, a. 1970. Reproductive cycle of the coot clam, Mulinia lat- eralis (Say), in Long Island Sound. Veliger 12:265-269. CARRIKER, M. R. 1961. Interrelation of functional morphology, behavior, and autecology in early stages of the bivalve Mercenaria mercenaria. J. Elisha Mitchell Sci. Soc. 77:168-241. Cassie, R. M. 1954. Some uses of probability paper in the analysis of size frequency distributions. Aust. J. Mar. Freshwater Res. 5:513-522. Chase, j. 1965-73. Oceanographic observations along the east coast of the United States (1963-1971). U.S. Coast Guard OceEinogr. Unit, Wash., D.C. Churchill, E. P., Jr. 1920. The oyster and the oyster industry of the Atlantic and Gulf coasts. Rep. U.S. Comm. Fish. 1919, Append. 8, 51 p. (Doc. 890.) COE, W. R., and H. J. TURNER, JR. 1938. Development of the gonads and gametes in the soft- shell clam (Mya arenaria). J. Morphol. 62:91-111. Davis, H. C, and P. E. Chanley. 1956. Spawning and egg production of oysters and clams. Biol. Bull. (Woods Hole) 110:117-128. Galtsoff, p. S. 1930. The fecundity of the oyster. Science (Wash., D.C.) 72:97-98. 1964. The American oyster, Crassostrea virginica Gme- lin. U.S. Fish. Wildl. Serv., Fish. Bull. 64, 480 p. GlESE, A. C. 1959. Comparative physiology: annual reproductive cy- cles of marine invertebrates. Annu. Rev. Physiol. 21:547-576. Harding, J. P. 1949. The use of probability paper for the graphical analysis of polymodal frequency distributions. J. Mar. Biol. Assoc. U.K. 28:141-153. HUGHES, R. N. 1970. Population dynamics of the bivalve Scrobicularia plana (Da Costa) on an intertidal mud-flat in North Wales. J. Anim. Ecol. 39:333-356. HUMASON, G. L. 1967. Animal tissue techniques. 2ded. W. H. Freeman Co., San Franc, 569 p. Jerome, W. C, A. P. Chesmore, and C. O. Anderson. 1968. A study of the marine resources of the Annisquam River-Gloucester Harbor coastal system. Mass. Dep. Nat. Resour., Div. Mar. Fish., Monogr. Ser. 8, 62 p. Johnson, R. G. 1965. Temperature variation in the infaunal environment of a sand flat. Limnol. Oceanogr. 10:114-120. Keck, R. T., D. Maurer, and H. Lind. 1975. A comparative study of the hard clam gonad de- velopmental cycle. Biol. Bull. (Woods Hole) 148:243- 258. Kellogg, J. L. 1905. rV. Conditions giveming existence and growth of the soft shell clam {Mya arenaria). Rep. U.S. Fish Comm. 29:195-224. KlNNE, O. 1963. The effects of temperature and salinity on marine £md brackish water animals. I. Temperature. Oceanogr. Mar. Biol., Annu. Rev. 1:301-340. Landers, W. S. 1954. Seasonal abundance of clam larvae in Rhode Island waters, 1950-52. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 117, 29 p. Loosanoff, v. L. 1937a. Seasonal gonadal changes of adult clams, Venus mercenaria. Biol. Bull. (Woods Hole) 72:406-416. 1937b. Development of the primary gonad and sexual phases in Venus mercenaria Linnaeus. Biol. Bull. (Woods Hole) 72:389-405. Loosanoff, V. L., and H. C. Davis. 1950. Spawning of oysters at low temperatures. Science (Wash., D.C.) 111:521-522. 1963. Rearing of bivalve molluscs. Adv. Mar. Biol. 1:1- 136. Mead, a. D., and E. W. Barnes. 1903. Observations on the soft-shell clam, Mya arenaria (4th paper). R.I. Comm. Inland Fish., 22d. Annu. Rep., p. 29-46. 165 FISHERY BULLETIN: VOL. 76, NO. 1 MUNCH-PETERSEN, S. 1973. An investigation of a population of the soft clam i.Mya arenaria L.) in a Danish estuary. Medd. Dan. Fisk. Havunders., New Ser. 7:47-73. Murphy, G. I. 1968. Pattern in life history and the environment. Am. Nat. 102:391-403. NELSON, T. C. 1928. On the distribution of critical temperatures for spawning and for ciliary activity in bivalve molluscs. Science (Wash., D.C.) 67:220-221. ORTON, J. H. 1920. Sea-temperature, breeding and distribution in marine animals. J. Mar. Biol. Assoc. U.K. 12:339-366. PFITZENMEYER, H. T. 1962. Periods of spawning and setting of the soft-shelled clam, Mya arenaria, at Solomons, Maryland. Chesapeake Sci. 3:114-120. Porter, R. G. 1974, Reproductive cycle of the soft-shell clam, Mya arenaria, at Skagit Bay, Washington. Fish. Bull,, U,S, 72:648-656. Ropes, J. W., and A. P. Stickney. 1965. Reproductive cycle of Mya arenaria in New Eng- land. Biol. Bull. (Woods Hole) 128:315-327. SHAW, W. N. 1962. Seasonal gonadal changes in female soft-shell clams, Mya arenaria, in the Tred Avon River, Mary- land. Proc. Natl. Shellfish. Assoc. 53:121-132. Smith, 0. R., J. P. Baptist, and E. Chin. 1955. Experimental farming of the soft-shell clam, Mya arenaria, in Massachusetts, 1949-1953. Commer. Fish. Rev. 17(6):1-16. Stafford, J. 1912. On the recognition of bivalve larvae in plankton collections. Contrib. Can. Biol. 1906-1910:221-242. Stevenson, J. R. 1907. Report upon observations and experiments on mol- luscs in Essex County during 1906. In Rep. Comm. Fish. Game 1906, Commonw, Mass., Public Doc. 25:68-96. Stickney, A. P. 1963. Histology of the reproductive system of the soft-shell clam (Mya arenaria). Biol. Bull. (Woods Hole) 125:344- 351. 1964. Salinity, temperature, and food requirements of soft-shell clam larvae in laboratory culture. Ecology 45:283-291, Sullivan, C. M. 1948. Bivalve larvae of Malpeque Bay, P.E.I. Fish. Res. Board Can., Bull. 77, 58 p. Thorson, G. 1950. Reproductive and larval ecology of marine bottom invertebrates. Biol. Rev. (Camb.) 25:1-45. Turner, H. J., Jr. 1949. Report on investigations of methods of improving . the shellfish resources of Massachusetts. Commonw. Mass. Dep. Conserv., Div. Mar. Fish., 22 p. Also Woods Hole Oceanogr. Inst., Contrib. No. 510. 1950. Investigations on the soft-shell clam, Mya arenaria. In Third report on investigations of methods of improving shellfish resources of Massachusetts, p. 5-15. Commonw. Mass. Dep. Conserv., Div. Mar. Fish. Also Woods Hole Oceanogr. Inst., Contrib. No. 564. Warwick, R. M., and R. Price. 1975. Macrofauna production in an estuarine mud- flat. J. Mar. Biol. Assoc. U.K. 55:1-18. 166 DIEL MOVEMENTS OF LARVAL YELLOWTAIL FLOUNDER, LIMANDA FERRUGINEA, DETERMINED FROM DISCRETE DEPTH SAMPLING W. G. Smith, J. D. Sibunka, and A. Wells ^ ABSTRACT A 72-h study to investigate diel movements of yellowtail flounder larvae indicated that they exhibited pronounced vertical movements that were repetitious from day to day. Collections at 3-h intervals with 20-cm bongo nets revealed that larvae were near the surface at night, and mostly at a depth of 20 m during the day. Ascent and descent occurred largely at sunset and sunrise, respectively. Thermal gradients at 10 to 20 m and 30 to 40 m had no apparent influence on the vertical movements. Amplitude of the movements increased with the size of larvae. Recently hatched larvae remained near the shallow thermal gradient. Intermediate sized larvae migrated from middepths during the day to surface and near-surface at night. Large larvae moved throughout the water column. The incidence of feeding was low but a daily feeding pattern was evident. Most larvae with gut contents were collected from 1900 to 0100 h on the first day; from 1600 to 2200 h on the second day; and from 1600 to 0100 h on the third day. The near-absence of gut contents in larvae caught during morning daylight hours suggests that the onset of feeding is triggered by something other than, or in addition to, light. Wind driven circulation near the surface was thought to transport larvae at night, when they moved towards the surface. Subsurface circulation was sluggish and ineffective as a transporting mechanism. Diel migrations by larval fishes play an important but largely unexplored role in dispersion during planktonic development. We became cognizant of the need to investigate this role after our initial ichthyoplankton survey, a series of cruises in the Middle Atlantic Bight to determine when and where coastal fishes spawn and to trace the disper- sal of planktonic eggs and larvae (Clark et al. 1969). Despite a full schedule of field work, the survey was only partially successful. We learned where and when many fishes spawn and recog- nized seasonal shifts in spawning areas (see Smith 1973; Fahay 1974; Kendall and Reintjes 1975; Smith et al. 1975), but we were unsuccessful in tracking disperson away from the spawning grounds. After realizing the shortcomings of the survey, we began to speculate on the significance of diel migrations and how they might interact with cir- culation to affect dispersion, especially where the water column is thermally stratified and surface and subsurface currents differ in velocity. We theorized that a study of the diel movements offish 'Northeast Fisheries Center Sandy Hook Laboratory, Na- tional Marine Fisheries Service, NOAA, Highlands, NJ 07732. Manuscript accepted June 1977. FISHERY BULLETIN: VOL. 76. NO. 1, 1978. larvae, when related to our survey data and to known circulation patterns, might provide us with better information on larval transport than we could obtain from continued surveys. In June 1972 we conducted a 72-h study of the diel movements of larval yellowtail flounder, Limanda ferruginea (Storer), an important species in the New England trawl fishery, and the most abundant flatfish lar- vae collected during our survey of the Middle At- lantic Bight. Our primary objectives were to de- termine whether the young flatfish undergo diel migrations, whether the migrations are repeti- tious in time and extent, and how they interact with circulation to affect dispersion. Yellowtail flounder range from the Gulf of St. Lawrence to Chesapeake Bay. Their center of abundance lies between the western Gulf of Maine and southern New England (Bigelow and Schroeder 1953). They spawn from March to Sep- tember in the Middle Atlantic Bight. Spawning progresses from south to north as the season ad- vances. The peak of the season in the bight occurs in early May with heaviest spawning off New York and northern New Jersey. Based on the catch of larvae <4 mm. Smith et al. (1975) determined that most spawning takes place between 4° and 9°C. 167 FISHERY BULLETIN: VOL. 76, NO. 1 METHODS We selected the general area for the 72-h study from results of the 1965-66 survey (Smith et al. 1975). The specific site, 98 km south of Montauk Point, N. Y., was selected by making trial plankton tows until we found the patch of larvae ( Figure 1 ). To stay within the patch, we deployed a free- drifting parachute drogue similar to that de- scribed by Volkmann et al. ( 1956). The parachute was attached 18 m below the staff buoy on our drogue. We sampled at 3-h intervals, from 1000 h on 15 June to 0700 h (EDT) on 18 June 1972. Tempera- ture and salinity observations preceded each tow during the first 2 days. We continued to take tem- peratures at 3-h intervals on the third day but recorded salinity data at 6-h intervals. When we started sampling, the summer solstice was only 6 days hence and a day was divided into 15 h of daylight and 9 h of darkness. Sunrise and sunset were at about 0530 h and 2030 h, respectively. By sampling at 3-h intervals, we made five tows dur- ing daylight and three tows at night during each day. Plankton samples were taken with an array of four 20-cm bongos fitted with 0.505-mm mesh nets. Each tow lasted 15 min. Towing speed was 5 kn (3 m/s). We chose the 20-cm bongo over the larger 61-cm bongo to keep both plankton volumes and numbers offish larvae at levels that would not exceed our laboratory capabilities. Catch com- parison tests between the 20- and 61-cm nets re- vealed no significant differences in the catch of larvae (Bjdrke et al. 1974; Posgay et al.^). Readings obtained from digital flow meters were used to calculate the amount of water sam- pled by one side of each bongo. With the exception of the surface-sampling net, the bongos were at- tached to the towing wire to sample near depths where temperature changes were greatest. They sampled at 8 m, which was just above the shal- lower of the two temperature gradients; at 20 m, below the shallow gradient; and at 48 m, which was below the deep thermal gradient and about 17 to 20 m above bottom. We preserved the contents from only the metered side of each bongo. Bathy- kymographs (BKG) were attached above two of the three subsurface nets to monitor sampling depth profiles. The sequence of attachment changed with each tow. Resultant BKG traces in- dicated that the average towing depth of each sub- surface net was ±2 m of the intended sampling depth. The bongos did not have opening-closing devices. We tried to minimize contamination dur- ing setting and retrieval by snapping the three subsurface nets onto the towing wire and lowering them into the water while the ship maintained just enough way to stay on course. Immediately after affixing the 8-m net, vessel speed was in- creased. The surface net was snapped in place and lowered into the water as the ship approached ^Posgay, J. A., R. R. Marak, and R. C. Hennemuth. 1968. Development and tests of new zooplankton samplers. Int. Comm. Northwest Atl. Fish., Res. Doc. 68/85, 7 p. Figure l.— Site of 72-h study of diel movements of yellowtail flounder lar- vae. Insert shows theoretical track of drogue and its position at 3-h sampling intervals. i kr' ... -s" © J ^.. c ^*^<" \ \, ST«t 13 / Q^ / Vt start ^T 0AY3 I 0 / '^^V— ^ / •n SIAKT ^ DAY 3 •(ftw / f , 12 11 10 09 oa TfVt WEST 168 SMITH ET AL: DIEL MOVEMENTS OF LARVAL FLOUNDER towing speed. At the end of each tow the nets were retrieved as we slowed to a stop. All yellowtail flounder larvae from each sample of <100 fish were counted and measured to the nearest 0.1 mm SL. If the count exceeded 100, a subsample of about 25% was randomly selected and measured. Then the number of larvae in each size increment was adjusted so that the sum cor- responded with the total sample size. Despite our efforts to minimize sampling contamination while setting and retrieving the nets, subsurface nets sampled more water than the surface net. To com- pensate for contamination, we standardized the volume of water filtered by each net by using the mean amount of water filtered by the surface net (88.8 m^) as the standard. We then adjusted the catch of each net to correspond with the adjusted amount of water filtered. These changes accounted for average reductions in the catch of < 1.0% in the surface net, 3.4% in the 8-m net, 4.4% in the 20-m net, and 13.4% in the 48-m net or a net reduction of 4.7% of the total catch. We inspected digestive tracts of young flounders for indications of a feeding pattern, i.e., presence or absence of gut contents, that might be related to vertical movements. We were able to make these observations simply by using a microscope and incident lighting. After grouping the adjusted larval catches into four size categories, «4.0, 4.1 to 8.0, 8.1 to 10.0, and >10.0 mm, we examined the data for homo- geneity of sampling variance by comparing within station catches by depth. Daylight tows were con- sidered replicates, as were night tows. Standard deviations were proportional to the means in the raw data, indicating that sampling variance was not homogeneous. The variance was stabilized by transforming the data to log J (J (x + 1). We used the UCLA BMD computer program 02 V, a multifactor ANOVA program (Dixon 1973), to test for differ- ences in mean catches by day, depth, time (day vs. night), and size of larvae (Table 1). To meet a program prerequisite, we balanced the number of day and night tows used in the analysis by ran- domly selecting three of the five tows for each daytime period. RESULTS Light conditions and sea state varied during the 3-day study in response to changing weather. The sky was cloudy when we began sampling on 15 June. Seas were moderate, stirred by 2 days of Table l. — Analysis of VEiriance of data collected during study of diel movements of yellowtail flounder larvae. Variables include days, time (day vs. night), capture depth, and length of larvae, grouped into size categories of €4.0, 4.1 to 8.0, 8.1 to 10.0, and >10.0mm. Data were transformed to log, gix + 1 ) and pertain to 3 day tows and 3 night tows taken during each day of the 3-day study. Source of variation df S.S. I^.S. F 1 (days) 2 0.21 0 11 0.93 2 (day-night) 1 23 83 2383 208 40" 3 (depth) 3 16.81 5 60 48 99*- 4 (size of larvae) 3 7349 24.50 21453" 1.2 2 0.23 0.11 1 01 1.3 6 1.79 030 2.61* 1. 4 6 1.45 0.24 2.11 2,3 3 4363 14.54 127 18" 2.4 3 3.94 1.31 11.47" 3,4 9 14.86 1.65 14.44" 1,2, 3 6 2.20 0.37 3.21" 1,2,4 6 0.45 0.07 0.66 1,3,4 18 237 0.13 1.15 2, 3,4 9 13.13 1 46 1276" 1. 2, 3, 4 18 2.80 0 15 1.36 Within replicates 192 21 95 Oil Total 287 223.14 •P«0.05. "P«0 01 brisk south to southwesterly winds of 15 to 20 kn (7-10 m/s). On the 16th the sky cleared but south- erly winds persisted. The 17th was cloudy with intermittent periods of light rain until evening when dense fog set in. We completed field work in heavy rain on the morning of the 18th. There was little or no measurable wind during the last 24 h of sampling. Water temperature in the Middle Atlantic Bight increases rapidly in the spring and the water column becomes thermally stratified during the summer (Norcross and Harrison 1967). At the time and site of our study, the surface temperature averaged 15.0°C, the bottom 5.7 °C. A thermal gradient of about 5°C, the predecessor of the more strongly defined summer thermocline, occurred at depths between 10 and 20 m. A second, weaker gradient existed between 30 and 40 m. Salinity increased from 31.3 %o at the surface to 32.8%onear the bottom. The most pronounced change in salin- ity occurred at about the same depths as the shal- low thermal gradient (Figure 2). Drift of the drogue was erratic and sluggish throughout the 72-h study. In 3 days it crossed its previous path 16 times, travelled a net distance of only 5.4 km in a southwesterly direction, and was never more than 7.2 km from the starting point. Net direction of drift was into the wind and the drogue travelled the greatest distance on the third day, when there was little or no wind. Because the drogue's direction of drift changed at approxi- mately 6-h intervals, we concluded that tidal 169 FISHERY BULLETIN: VOL. 76, NO. 1 Figure 2. — Vertical distribution of yellow- tail flounder larvae at 3-h intervals, based on percent contribution of adjusted catches in each of four nets. Mean temperature and salinity profiles during each day of the 3-day study are shown at right. STATION A TIME 1000 DAY 1 1600 1900 2200 0100 OCCURRENCE OF LARVAE ( % 1 TEMPERATURE ( °C ) 5 10 15 20 STATION J TIME 1000 DAY 2 K. L " M N O 1300 1600 1900 2200 0100 OCCUBIENCE OF L A fl V A E ( '■■-■ ■ 0400 0700 31.0 31.5 32.0 32.5 33,0 SALINITY %o TEMPEBAIURE I °C DAY 3 STATION R TIME 1000 OCCURRENCE OF LARVAE (% SALINITY %, V 0400 Z 0700 TEMPERATURE 1 °C 1 0 5 10 15 20 31.0 31.5 32.0 32.5 33.0 SALINITY %. 170 SMITH ET AL: DIEL MOVEMENTS OF LARVAL FLOUNDER circulation was largely responsible for its move- ments (see Figure 1 insert). The analysis of variance indicated that we stayed within the same patch of larvae throughout the study. Daily mean differences in both the number and size of larvae were not significant. There was, however, a highly significant differ- ence between means of day and night catches, and between catches at the four depths sampled. We attributed these differences to diel movements and the resultant shift in the distribution of most larvae toward the surface, where two nets fished, at night. The diel movements were repetitious in time and extent. There was no significant differ- ence in means of catches within daylight and night tows, or in their depth distribution at a given time during each day (Table 1). Larvae were most abundant in the 20-m net during daylight tows on the first day. None were caught by the surface net and the combined catch of the nets at 8 and 48 m contributed <15'7c of the daytime catch. The distribution of larvae changed significantly after dark. By 2200 h the catch in the surface net was greater than the combined catch of the other three nets and more than double that of any other net. When combined, the surface and 8-m catches accounted for nearly 17% of the 2200-h catch. At 0100 h larvae remained most abundant at the surface and, although the surface catch was less than at 2200 h, again the upper two nets accounted for >709c of the catch (Figure 2). At 0400 h, the last nighttime tow, most larvae were caught at 8 m (Table 2). The vertical movements of larvae throughout the second day were similar to those on the first day. Most larvae were taken at 20 m on each of the five daylight tows. Except for a single specimen in the 1600-h tow, none were caught at the surface during daylight. By 2200 h the distribution again changed significantly. Like the first night, the sur- face catch was greater than the total catch of the other three nets. The combined catch at the sur- face and 8 m made up 88% of the 2200-h catch. Unlike the first night, larvae were less abundant at the surface than at 8 m at 0100 h but the upper two nets again contributed >80'^ of the catch (Figure 2). At 0400 h larvae reoccurred in greatest numbers at the surface. This increase in the sur- face catch at 0400 h did not occur on the previous day (Table 2). Results of tows on the last day were much like those on the first 2 days. Larvae were most abun- dant at 20 m on all five daylight tows. Only one larva was taken at the surface, that at 1900 h. By 2000 h the distribution of larvae shifted towards the surface. The young flounder repeated their behavior of the previous day by descending at 0100 h. Most were at 8 m and, for the first time, the 20-m net caught more larvae than the surface net on a night tow. Despite the somewhat deeper distribu- tion, the combined catch of the surface and 8-m nets contributed nearly 80% of the 0100-h catch (Figure 2). The distribution of larvae at 0400 h was much like that at 0100 h. It differed from the other two 0400-h tows in that the contribution of the surface net was greatly reduced, and that of the 8-m net greatly increased (Table 2). The amplitude of diel movements increased with size of larvae but, within each of the four size groups, the movements were similar each day (Figure 3). The vertical movements of larvae ^4.0 mm were relatively insignificant compared with those of larger larvae. During daylight hours the recently hatched larvae were at an average depth of about 24 m, at night 20 m, a difference of only 4 m. Larvae 4.1 to 8.0 mm long were more active. They moved vertically from an average depth of 24 m during the day to about 9 m at night. The trend continued with larvae 8.1 to 10.0 mm long. During the day were at an average depth of 29 m. At night they ascended to an average depth of 5 m. Larvae > 10.0 mm exhibited the most pronounced vertical movements. During the day they were at an average depth of 41 m, at night 7 m. By not having a net near bottom, we failed to sample the entire depth range of larvae. However, it appears that our nets encompassed the depth distribution for nearly all larvae <10.0 mm. Only 5% of those < 10.0 mm were caught in the 48-m net and we assume that their numbers continued to decline below that depth. On the other hand, the daytime distribution of larvae >10.0 mm may have been deeper than our results indicate. Al- most half (46%) of the daytime catch of larvae >10.0 mm was caught in the deep net. None were caught at depths <20 m, and most (77%) of the large larvae caught at 20 m during the day were collected at 0700 h, probably during their morning descent. The incidence (percent) of larvae with visible gut contents was as high as 40% at one station but only 6% of the larvae caught during the 3-day study contained visible gut contents. The overall incidence was low, but our results indicate that most feeding occurred at about the same time on all 3 days. We found the highest incidence from 171 FISHERY BULLETIN: VOL 76, NO. 1 Table 2.— Adjusted catch of yellowtail flounder larvae by size group, depth, and time. Results are presented by day, beginning with the initial daylight tow, although we began sampling at 1000 h (Station A) and finished at 0700 h (Station Z). Net depth (m) Day 1 Day 2 Hour Size group (mm) 1 Total Size group (mm] 1 Total of Stn. tow <4 4-8 8-10 >10 No. % No.;m=' Stn. <4 4-8 8-10 :>10 No. % No./m3 Day tows: H 0700 Surt 0 0 0 0 0 0 0 Q 0 0 0 0 0 0 0 8 0 21 1 0 22 5 0.2 0 125 76 0 201 42 23 20 26 350 15 0 391 84 4.4 0 153 95 15 263 54 3.0 48 1 38 7 4 50 11 0.6 0 14 2 3 19 4 0.2 Total 27 409 23 4 463 100 0 292 173 18 483 100 A 1000 Surf 0 0 0 0 0 0 0 J 0 0 0 0 0 0 0 8 2 6 0 0 8 1 0,1 0 48 12 0 60 11 0.7 20 84 447 8 0 539 76 6.1 25 315 22 1 363 67 4.1 48 0 85 68 12 165 23 1.9 2 87 26 7 122 22 1.4 Total 86 538 76 12 712 100 27 450 60 8 545 100 B 1300 Surt 0 0 0 0 0 0 0 K 0 0 0 0 0 0 0 8 0 12 1 0 13 3 0.1 0 15 8 0 23 4 03 20 25 411 16 0 452 82 5.1 15 325 74 4 418 82 4.7 48 6 51 20 7 84 15 0.9 5 49 11 5 70 14 0.8 Total 31 474 37 7 549 100 20 389 93 9 511 100 C 1600 Surf 0 0 0 0 0 0 0 L 0 1 0 0 1 <1 <0.1 8 0 12 2 0 14 4 02 1 16 4 0 21 6 0.2 20 15 285 24 1 325 85 3.7 29 173 31 3 236 67 2.7 48 0 33 6 3 42 11 0.5 13 68 11 4 96 27 1.1 Total 15 330 32 4 381 100 43 258 46 7 354 100 D 1900 Surf 0 0 0 0 0 0 0 M 0 0 0 0 0 0 0 8 0 13 0 0 13 1 0.1 1 62 19 0 82 13 0.9 20 22 864 18 0 904 96 10.2 24 444 60 2 530 83 60 48 0 14 9 6 29 3 03 0 14 6 8 28 4 0.3 Total 22 891 27 6 946 100 25 520 85 10 640 100 All day tows: Surf 0 0 0 0 0 0 0 0 1 0 0 1 <1 <0.1 8 2 64 4 0 70 2 0.2 2 266 119 0 387 15 0.9 20 172 2.357 81 1 2.611 86 5.9 93 1.410 282 25 1,810 72 4.1 48 7 221 110 32 370 12 08 20 232 56 27 335 13 0.8 Total 181 2,642 195 33 3,051 100 115 1.909 457 52 2,533 100 Night tows: E 2200 Surt 10 922 342 30 1,304 54 14.7 N 4 795 215 24 1,038 53 11.7 8 13 421 93 18 545 23 6.1 29 504 111 36 680 35 7.7 20 61 411 53 15 540 22 6.1 16 169 17 3 205 11 2.3 48 1 21 2 0 24 1 0.3 1 23 2 1 27 1 03 Total 85 1.775 490 63 2,413 100 50 1,491 345 64 1,950 100 F 0100 Surt 4 329 126 20 479 43 5.4 0 0 374 146 24 544 30 6.1 8 10 268 29 2 309 27 3.5 0 886 62 18 966 54 109 20 49 230 11 1 291 26 3.3 29 216 10 5 260 15 2.9 48 3 34 3 0 40 4 0.5 1 22 0 0 23 1 0.3 Total 66 861 169 23 1,119 100 30 1,498 218 47 1,793 100 G 0400 Surt 4 274 33 4 315 25 35 P 5 551 126 0 682 47 7.7 8 11 505 59 7 582 47 6.6 12 338 102 18 470 33 5.3 20 46 244 7 5 302 25 3.4 12 227 0 0 239 17 2.7 48 7 27 0 0 34 3 0.4 5 35 2 2 44 3 05 Total 68 1.050 99 16 1,233 100 34 1,151 230 20 1,435 100 All night tows: Surf 18 1.525 501 54 2,098 44 7.9 9 1,720 487 48 2,264 44 8.5 8 34 1.194 181 27 1,436 30 5.4 41 1,728 275 72 2,116 41 79 20 156 885 71 21 1,133 24 4.3 57 612 27 8 704 13 26 48 11 82 5 0 98 2 0.4 7 80 4 3 94 2 0.4 Total 219 3,686 758 102 4,765 100 114 4,140 793 131 5,178 100 All tows: Surt 18 1.525 501 54 2,098 27 3.0 9 1,721 487 48 2,265 29 3.2 8 36 1.258 185 27 1,506 19 2.1 43 1,994 394 72 2,503 32 3.5 20 328 3.242 152 22 3.744 48 5.3 150 2,022 309 33 2.514 33 3.5 48 18 303 115 32 468 6 0.7 27 312 60 30 429 6 0.6 Total 400 6,328 953 135 7,816 100 229 6.049 1,250 183 7,711 100 1900 to 0100 h on the first day; from 1600 to 2200 h on the second day; and from 1600 to 0100 h on the third day. The evening ascent toward the surface occurred during the time of peak feeding, but the incidence of feeding remained highest in larvae caught at 20 m before, during, and after the even- ing ascent (Figure 4). We concluded that essential prey organisms occur throughout the water col- 172 umn and that diel movements and feeding are not directly related. DISCUSSION When Sette ( 1943) studied the early life history of Atlantic mackerel, Scomber scombrus, in the Middle Atlantic Bight in 1929, he made four tows, SMITH ET AL: DIEL MOVEMENTS OF LARVAL FLOUNDER Table 2.— Continued. Net depth (m) Day 3 3-day total Hour of Size group (mm) 1 Total Size group (mm) Total No,/m3 Stn. tow <4 4-8 8-10 •10 No. % No./m' ■ 4 4-8 8-10 ■10 No % (avg.) Day tows: Z 0700 Surf 0 0 0 0 0 0 0 0 0 0 0 0 0 0 e 1 13 5 0 19 3 02 1 159 82 0 242 15 09 20 11 271 218 71 571 93 64 37 774 328 86 1.225 79 46 48 0 17 2 3 22 4 0,2 1 69 11 10 91 6 0.3 Total 12 301 225 74 612 100 39 1,002 421 96 1,558 100 R 1000 Surt 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 17 32 0 49 13 0,6 2 71 44 0 117 7 04 20 6 151 123 15 295 79 33 115 913 153 16 1,197 74 4.5 48 0 21 5 2 28 8 03 2 193 99 21 315 19 1,2 Total 6 189 160 17 372 100 119 1,177 296 37 1,629 100 S 1300 Surf 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 2 0 2 1 <0,1 0 27 11 0 38 3 0.1 20 47 150 9 0 206 71 23 87 886 99 4 1,076 80 4.0 48 3 42 24 12 81 28 09 14 142 55 24 235 17 0,9 Total 50 192 35 12 289 100 101 1,055 165 28 1,349 100 T 1600 Sun 0 0 0 0 0 0 0 0 1 0 0 1 <1 <0.1 8 1 5 0 0 6 3 0,1 2 33 6 0 41 4 0.2 20 25 106 3 0 134 53 1.5 69 564 58 4 695 71 2.6 48 1 66 35 10 112 44 1,3 14 167 52 17 250 25 0.9 Total 27 177 38 10 252 100 85 765 116 21 987 100 U 1900 Surf 0 1 0 0 1 <1 <0.1 0 1 0 0 1 <1 <0.1 8 0 30 7 0 37 6 0-4 1 105 26 0 132 6 0.5 20 14 467 46 0 527 89 59 60 1,775 124 2 1,961 90 74 48 0 11 6 10 27 5 0-3 0 39 21 24 84 4 03 Total 14 509 59 10 592 100 61 1,920 171 26 2,178 100 All day tows: Surf 0 1 0 0 1 <1 <01 0 2 0 0 2 <1 <01 8 2 65 46 0 113 5 03 6 395 169 0 570 7 0.4 20 103 1.145 399 86 1,733 82 3.9 368 4,912 762 112 6,154 80 4,6 48 4 157 72 37 270 13 0-6 31 610 238 96 975 13 07 Total 109 1.368 517 123 2,117 100 405 5,919 1,169 208 7,701 100 Night tows: W 2200 Surf 4 574 235 28 841 53 9.5 18 2,291 792 82 3.183 54 119 8 8 374 206 28 616 39 6.9 50 1,299 410 82 1,841 31 6.9 20 12 68 6 0 86 6 1.0 89 648 76 18 831 14 3.1 48 0 23 3 2 28 2 03 2 67 7 3 79 1 03 Total 24 1,039 450 58 1,571 100 159 4,305 1,285 185 5,934 100 X 0100 Sun 5 197 43 0 245 10 28 9 900 315 44 1,268 23 48 8 0 1.488 274 16 1.778 70 20.0 10 2,642 365 36 3.053 56 11.5 20 32 412 3 0 447 17 50 110 858 24 6 998 18 3.7 48 7 57 0 1 65 3 07 11 113 3 1 128 3 0.5 Total 44 2,154 320 17 2,535 100 140 4,513 707 87 5.447 100 Y 0400 Surt 0 55 36 4 95 6 1.1 9 880 195 8 1.092 26 4 1 8 4 851 209 31 1.095 71 12,3 27 1,694 370 56 2.147 51 8.1 20 24 257 10 0 291 19 3-3 82 728 17 5 832 20 3,1 48 11 38 3 1 53 4 0-6 23 100 5 3 131 3 05 Total 39 1.201 258 36 1,534 100 141 3,402 587 72 4.202 100 All night tows: Surf 9 826 314 32 1,181 21 44 36 4,071 1,302 134 5.543 36 69 8 12 2.713 689 75 3,489 62 13-1 87 5,635 1,145 174 7,041 45 8.8 20 68 737 19 0 824 14 3.1 281 2,234 117 29 2.661 17 3.3 48 18 118 6 4 146 3 0,5 36 280 15 7 338 2 0.4 Total 107 4,394 1,028 111 5,640 100 440 12,220 2.579 344 15,583 100 All tows: Surf 9 827 314 32 1,182 15 17 36 4,073 1.302 134 5,545 24 2.6 8 14 2,778 735 75 3.602 46 5.1 93 6.030 1.314 174 7,611 33 3,6 20 171 1.882 418 86 2,557 33 36 649 7.146 879 141 8,815 38 4,1 48 22 275 78 41 416 6 0.6 67 890 253 103 1,313 5 0,6 Total 216 5,762 1.545 234 7.757 100 845 18.139 3.748 552 23.284 100 morning, noon, evening, and midnight, off Fire Island to investigate the vertical distribution of eggs and larvae. Royce et al. (1959) included a cursory presentation of data on yellowtail flounder larvae from Sette's series of discrete depth tows. Although Sette's nets were towed slower (1 kn vs. 5 kn) and the flounder larvae were smaller ix = 3.9 mm vs. X = 6.7 mm) than ours, the results of the two studies are similar in several aspects. For example, Royce et al. (1959) reported larvae at the surface at night, but not during daylight; the night catch was double the daytime catch; and the catch dropped off sharply in their deep net at night. Their larvae were most concentrated at a depth of 10 m on all four tows. Although this appears to differ from our results, we have shown that larvae <4 mm do not participate in the diel migrations but remain within a limited depth stratum. Thus 173 FISHERY BULLETIN: VOL. 76. NO. 1 10 20 33 a. LU Q 10 20 30 0 10 20 30 40 0 10 20 30 40 OOP il300 1600 LARVAE < 4.0 mm DAY 1 (N - 400) DAY 2 (N = 229) DAY 3 (N = 216) A .^-^ a- LARVAE 4.1 to 8.0 mm DAY 1 (N =6328) DAY 2 (N =6049) DAY 3 (N = 5762 A __^y_ A TIME 1 900 2200 0100 0400 0700 LARVAE 8.1 ^o 10,0 mm DAY 1 (N =953) DAY 2 (N = 1250) DAY 3 (N = 1545) LARVAE >10.0 mm DAY 1 (N = 135) DAY 2 (N = 183) DAY 3 (N = 234) 10 20 30 40 ALL LARVAE DAY 1 (N = 7816) DAY 2 (N = 7711) DAY 3 (N = 7757) A.... '■■■A- Figure 3.— Mean depth of occurrence of yellowtail flounder larvae, grouped by size. Samples were taken at 3-h intervals for 3 days. Water depth ranged from 63 to 68 m. 174 SMITH ET AL: DIEL MOVEMENTS OF LARVAL FLOUNDER 100 Figure 4. — Percent of yellowtail flounder larvae by depth and time (up- per graph); and percent of larvae with visible gut contents by depth and time (lower graph). Figure represents aver- aged results from 3-day study. the small larvae exhibited similar behavior in both studies. Although their larvae were concen- trated at shallower depths than ours, in both cases the temperature was about 10°C where larvae <4 mm were most abundant. See Sette (1943) for temperature profile pertaining to data presented by Royce et al. (1959). A ^test on our adjusted catch data from the 15 daylight tows and 9 night tows indicated that the catch at night was significantly greater than the daytime catch. Some of this difference might re- sult from avoidance during daylight but, based on our fast towing speed, which would curtail avoid- ance, and results of gear performance tests by Bjdrke et al. ( 1974) and Posgay et al. (see footnote 2), which showed the 20-cm bongo to be an effec- tive sampler, we concluded that the greater catch at night was largely attributable to a change in the vertical distribution of larvae and our sampl- ing depths. Comparisons of day might catch ratios of daily catches and catches at 20 m support our conclusion. Whereas the daymight catch ratios of the adjusted catch (larvae per cubic meter) were 1:1.56, 1:2.04, and 1:2.66 on days 1 through 3, respecitvely, the reverse was true at 20 m, where the ratios were 2.30:1, 2.57:1, and 2.10:1. Night catches were greater than day catches because most larvae migrated towards the surface at night, where two nets fished. The resultant con- centration of larvae in a confined depth stratum, and the "extra" net fishing within the stratum where larvae were concentrated, accounted for the significantly greater catch with less sampling ef- fort at night. After descending during the early morning hours, larvae were largely subjected to capture at 20 m, where the daytime catch was more than twice as great as the catch at night. If avoidance were the principal factor in the day:night differences, we would expect larger catches at all depths at night. Both Bridger ( 1958 ) and Wood ( 197 1 ) found that the daytime distribution of herring, Clupea ha- rengus, larvae depended on light conditions. Their larvae were nearer the surface on cloudy days than on sunny days. Although weather conditions changed from partly cloudy to sunny, followed by fog and rain, and sea conditions changed from moderate to calm as winds diminished, yellowtail flounder larvae showed little variation in their diel movements during our 3-day study. We caught only two larvae at the surface during day- light hours. On all 3 days larvae began to ascend after 1900 h and were at the surface in greatest numbers at 2200 h. During the early morning hours of darkness their numbers decreased at the surface but the young fish did not disappear from the surface until sometime between 0400 and 0700 h. Judging from our results and those of Royce et 175 FISHERY BULLETIN: VOL. 76. NO. 1 al. (1959), we presume both the daily ascent and descent occurred near sunset and sunrise, respec- tively. Ahlstrom (1959) studied the vertical move- ments of larvae of several fishes off the coast of California. He found no evidence that larvae moved through the thermocline. His collections showed that they migrated vertically but the movements were usually restricted to the upper mixed layer. In contrast, neither the salinity gra- dient at 10 to 20 m nor the temperature gradients beginning at 10 and 30 m had a noticeable effect on the vertical movements of yellowtail flounder lar- vae in our study. Our collections indicate that the small flounder that migrated between middepths and the surface routinely tolerated salinity differ- ences of 1 . 5%o and temperature changes of 5°C , and those that moved throughout the water column withstood changes of about 10°C. Such rapid changes in temperature seem deleterious but our survey collections indicated that larvae of most flatfishes spawned in the Middle Atlantic Bight are physiologically adapted to wide ranges in temperature. For example, in 1966, when yellow- tail flounder spawned mostly at bottom tempera- tures between 4° and 9°C, we caught their larvae where the surface temperature was 5°C in April and 23°C in August (Smith et al. 1975). The amplitude of the vertical migrations by yel- lowtail flounder larvae increased in proportion to their size. Similar behavior was reported for larval haddock, Melanogrammus aeglefinus (Miller etal. 1963), and larval Clupea harengus (Seliverstov 1974). Recently hatched yellowtail flounder re- mained most abundant beneath the shallow ther- mal gradient, whereas late-stage larvae exhibited extensive vertical migrations that included most or all of the water column. Larvae > 10 mm proba- bly spend some time on the bottom. Bigelow and Schroeder (1953) reported that young yellovd:ail flounder descend to the bottom when 14 mm long. Royce et al. (1959) concluded that they seek bot- tom when 12 to 19 mm long. Judging from this information and the advanced stage of develop- ment of some larvae we caught near the surface after dark, we concluded that the change from a pelagic to a demersal life is not abrupt. Larvae making the transition to a demersal life continue to migrate towards the surface at night. This noc- turnal behavior might reflect a gradual dietary change from planktonic to benthic organisms. Al- though we are unsure of how long they continue the vertical migrations, the 20.7-mm SL specimen 176 collected during our survey (see Smith et al. 1975) might represent the maximum size at which they ascend toward the surface. In his review of the "critical period" concept, May (1974) pointed out that field studies of larval feeding have produced highly variable results. He cited several investigations that found the feeding incidence of clupeoid larvae very low, others that found it very high, and discussed theories that have been advanced to explain this variability. They include rapid digestion; nutrition from dis- solved organics; low food requirements; daily feed- ing patterns; defecation upon capture and preser- vation; escapement by healthy, feeding larvae; and food availability. Our data on yellowtail flounder larvae support at least two of these theories, namely, a daily feeding pattern and rapid digestion. Both the highest and lowest inci- dence of feeding occurred at predictable times on all 3 days and, with the exception of five speci- mens, the guts of all larvae appeared to be empty within hours after the period of maximum feeding. Several studies report that fish larvae feed most actively at high light intensities, but others differ. For example, Kjelson et al. ( 1975) found the diges- tive tract of young Atlantic menhaden, Brevoortia tyrannus; pinfish, Lagodon rhomboides; and spot, Leiostomus xanthurus, fullest at midday. Ruda- kova (1971) estimated that an average of 25% of the Atlantic herring, Clupea harengus harengus, larvae that he caught fed during the day, only 3.2% at night. Feeding studies by Blaxter (1965), Schumann (1965), and Braum (1967) support the above studies. On the other hand, Marak (1974) reported that young redfish, Sebastes marinus, fed during day or night and Blaxter ( 1969) found that larval sole, Solea solea, feed at night. Shelbourne (1953) reported that all postlarval plaice, Pleuronectes platessa, that he collected between 1400 and 2000 h had food in their guts. The per- cent of feeding larvae declined to between 70 and 80% in his samples collected from 2000 to 0200 h, then dropped sharply until daylight when it again increased to 100% for a short time. Our results resemble Shelbourne's (1953), ex- cept that we caught fewer feeding larvae and we did not find an indication of feeding at sunrise. The near absence of feeding larvae during daylight morning hours suggests that something other than, or in addition to, light triggers feeding by yellowtail flounder larvae. It appeared to us that feeding intensity increased during afternoon and evening hours. Larvae that had food in their guts SMITH ET AL: DIEL MOVEMENTS OF LARVAL FLOUNDER at 2200 and 0100 h might have fed after dark or they might have stopped feeding after sunset. Further study is needed to determine whether yel- lowtail flounder larvae feed at night. After analyzing 10 yr of drifter releases, Bum- pus (1973) reported that surface currents in the Middle Atlantic Bight occasionally reach speeds of 15 mi/day (27 km/day), but they are usually less than 10 mi/day ( 18 km/day). He estimated bottom drift at 0.5±0.2 mi/day (0.9±0.4 km/day) and speculated that circulation near bottom was so random and sluggish that it was unrealistic to derive drift rates of bottom water from his data, except from nearshore releases, which stranded within a reasonable time frame. Howe (1962) con- cluded that coastal circulation between Cape Cod and New York was largely attributable to short- term wind effects and that waters inside the 90-m isobath were comparatively stagnant during the first half of the year. The sluggish performance of our drogue supports Howe's results and indicates that the velocity of middepth drift at the time and location of our study was similar to Bumpus' de- scription of bottom circulation. Returns from drift bottle releases indicate that surface water generally moves westward off Long Island then southward along the Middle Atlantic States (Bumpus and Lauzier 1965). However, both Norcross and Stanley (1967) and Bumpus (1969) found evidence of surface current reversals in the Middle Atlantic Bight during the summer, and Doebler (1966) showed that the direction of sur- face water transport off Delaware responded rapidly to changes in wind direction. On the basis of these reports, we assume that the brisk south to southwest wind during the first 48 h of our study propelled surface water towards southern New England. Although yellowtail flounder larvae were not at the surface during the day, 44% of our night catches were taken at the surface during the first two nights. During this time wind probably influenced their horizontal displacement. By pas- sing the 15 h of daylight at subsurface depths, it appears from the net drift of our drogue that the larvae were transported in the opposite direction to that at night. Assuming that our drogue's erratic and sluggish drift is representative of middepth circulation off Long Island in the spring, when spawning by yel- lowtail flounder peaks, and that effects of spring and summer winds on circulation are usually lim- ited to a few days at a time, we conclude that wind driven currents in the study area do not play a major role in dispersing the larvae. Our conclusion is supported by Royce et al. (1959). Similarities in patterns of distribution between eggs and larvae led them to conclude that larvae were demersal before much horizontal drift occurred. It seems worth noting here that the smallest larvae, those least able to swim with directed movements, did not ascend to the surface at night. They remained below the shallow thermal gradient, where they were unaffected by wind-driven circulation. Whether or not our interpretation of the effects of currents on the distribution of yellowtail floun- der larvae is correct, it is clear to us that research- ers must investigate the diel movements of larvae they are studying before hypothesizing on how circulation affects the distribution and survival of young fishes. It is common practice to overlook or ignore larval behavior and relate the transport of larvae from both day and night collections by ob- liquely towed nets to surface circulation. In many cases, this oversight produces an exaggerated es- timate of the distance larvae are transported and, perhaps, an erroneous estimate of the direction of transport. 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. BIGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S.Fish Wildl. Serv., Fish. Bull. 53, 577 p. BJ0RKE, H., O. DRAGESUND, AND 0. ULLTANG. 1974. Efficency test on four high-speed plankton samplers. In J, H. S. Blaxter (editor). The early life history offish, p. 183-200. Springer- Verlag, N.Y. BLAXTER, J. H. S. 1965. The feeding of herring larvae and their ecology in relation to feeding. Calif Coop. Oceanic Fish Invest. Rep. 10:79-88. 1969. Visual thresholds and spectral sensitivity of flatfish larvae. J. Exp. Biol. 51:221-230. BRAUM, E. 1967. The survival of fish larvae with reference to their feeding behavior and the food supply. In S. D. Gerking (editor). The biological basis of freshwater fish production, p. 113-131. Blackwell Sci. Publ., Oxf. BRIDGER, J. P. 1958. On efficiency tests made with a modified Gulf III high-speed tow net. J. Cons. 23:357-365. Bumpus, D. F. 1969. Reversals in the surface drift in the Middle Atlantic Bight area. Deep-Sea Res. 16(Suppl.):17-23. 1973. A description of the circulation on the continental shelf of the east coast of the United States. Prog. Oceanogr. 6:111-157. 177 FISHERY BULLETIN: VOL. 76, NO. 1 BUMPUS, D. F., AND L. M. LAUZIER. 1965. Surface circulation on the continental shelf off east- em North America between Newfoundland and Florida. Am. Geogr. Soc, Ser. Atlas Mar. Environ. Folio. 7, 4 p. Clark, J., W. G. Smith, A. W. Kendall, and M. P. Fahay. 1969. Studies of estuarine dependence of Atlantic coastal fishes. Data report 1; Northern section. Cape Cod to Cape Lookout. R. V. Dolphin cruises 1965-66: Zooplankton vol- umes, midwater trawl collections, temperatures and salinities. U.S. Bur. Sport Fish. Wildl., Tech. Pap. 28, 132 p. Dixon, W. J. 1973. BMD02V analysis of variance for factorial design. In BMD biomedical computer programs, 3d ed., p. 607- 621. Univ. Calif Press, Berkeley. DOEBLER, H. J. 1966. A study of shallow water wind drift currents at two stations off the east coast of the United States. U.S. Navy Underwater Sound Lab., USL Rep. 755, 78 p. Fahay, M. P. 1974. Occurrence of silver hake, Merluccius bilinearis, eggs and larvae along the middle Atlantic continental shelf during 1966. Fish. Bull., U.S. 72:813-834. Howe, M. R. 1962. Some direct measurements of the non-tidal drift on the continental shelf between Cape Cod and Cape Hatter- as. Deep-Sea Res. 9:445-455. Kendall, a. w.. Jr., and j. w. Reintjes. 1975. Geographic and hydrographic distribution of Atlan- tic menhaden eggs and larvae along the middle Atlantic coast from RVZ)o/p/!(>! cruises 1965-66. Fish. Bull., U.S. 73:317-335. Kjelson, M. a., D. S. Peters, G. W. Thayer, and G. N. Johnson. 1975. The general feeding ecology of postlarval fishes in the Newport River estuary. Fish. Bull, U.S. 73:137-144. Marak, R. R. 1974. Food and feeding of larval redfish in the Gulf of Maine. In J. H. S. Blaxter (editor). The early life history offish, p. 267-275. Springer-Verlag, N.Y. May, R. c. 1974. Larval mortality in marine fishes and the critical period concept. In J. H. S. Blaxter (editor), The early life history offish, p. 3-19. Springer-Verlag, N.Y. Miller, D., J, B. Colton, Jr., and R. R. Marak. 1963. A study of the vertical distribution of larval had- dock. J. Cons. 28:37-49. NORCROSS, J. J., AND W. HARRISON. 1967. Part L Introduction, /n W.Harrison, J.J. Norcross, N. A. Pore, and E. M. Stanley. Circulation of shelf waters of the Chesapeake Bight, surface and bottom drift of con- tinental shelf waters between Cape Henlopen, Delaware, and Cape Hatteras, North Carolina, June 1963- December 1964, p. 3-9. Environ. Sci. Serv. Admin. Prof Pap. 3. NORCROSS, J. J., AND E. M. STANLEY. 1967. Part IL Inferred surface and bottom drift, June 1963 through October 1964. In W. Harrison, J. J. Norcross, N. A. Pore, and E. M. Stanley. Circulation of shelf waters of the Chesapeake Bight, surface and bottom drift of conti- nental shelf waters between Cape Henlopen, Delaware, and Cape Hatteras, North Carolina, June 1963- December 1964, p. 11-42. Environ. Sci. Serv. Admin. Prof Pap. 3. ROYCE, W. F,, R. J. BULLER, AND E. D. PREMETZ. 1959. Decline of the yellowtail flounder iLimanda fer- ruginea) off New England. U.S. Fish Wildl. Serv., Fish. Bull. 59:169-267. RUDAKOVA, V. A. 1971. On feeding of young larvae of the Atlanto-Scandian herring (Clupea harengus harengus L.) in the Norwegian Sea. Rapp. P.-V. Reun, Cons. Int. Explor. Mer 160:114- 120. SCHUMANN, G. 0. 1965. Some aspects of behavior in clupeid larvae. Calif Coop. Oceanic Fish. Invest. Rep. 10:71-78. Seliverstov, a. S. 1974. Vertical migrations of larvae of the Atlanto- Scandian herring iClupea harengus ). In J. H. S. Blaxter (editor), The early life history of fish, p. 253-262. Springer-Verlag, N.Y. SETTE, O. E. 1943. Biology of the Atlantic mackerel (Scomber scom- brus) of North America. Part I. -Early life history, includ- ing growth, drift and mortality of the egg and larval popu- lations. U.S. Fish Wildl. Serv., Fish. Bull. 50:149-237. SHELBOURNE, J. E. 1953. The feeding habits of plaice post-larvae in the Southern Bight. J. Mar. Biol. Assoc. U.K. 32:149-159. Smith, W. G. 1973. The distribution of summer flounder, Paralichthys dentatus, eggs and larvae on the continental shelf be- tween Cape Cod and Cape Lookout, 1965-66. Fish. Bull., U.S. 71:527-548. Smith, W. G., J. D. Sibunka, and A. Wells. 1975. Seasonal distribution of larval flatfishes { Pleuronec- tiformes) on the continental shelf between Cape Cod, Massachusetts, and Cape Lookout, North Carolina, 1965-66. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-691, 68 p. VOLKMANN, G., J. KNAUSS, AND A. VINE. 1956. The use of parachute drogues in the measurement of subsurface ocean currents. Trans. Am. Geophys. Union. 37:573-577. WOOD, R. J. 1971. Some observations on the vertical distributions of herring larvae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 160:60-64. 178 BIOECONOMIC CONTRIBUTION OF COLUMBIA RIVER HATCHERY FALL CHINOOK SALMON, 1961 THROUGH 1964 BROODS, TO THE PACIFIC SALMON FISHERIES Roy J. Wahle and Robert R. Vreeland> ABSTRACT This experiment was designed to estimate the contribution to sport and commercial fisheries of the 1961 through 1964 broods of fall chinook salmon, Oncor/jvnc/zi/s^s/iau'.y^sc/ja, from 13 rearing facilities on the Columbia River. These facilities reared 909c of the Columbia River hatchery fall chinook salmon during the four brood years. Marks common to all facilities were applied to 21.3 million of the 213 million 1961-64 brood fish released. Special marks were applied to 9.6 million fish at 11 of the study hatcheries. Sampling for the marks took place from 1963 through 1969. During the 7 yr of sampling, 65,620 chinook salmon with common and 22,090 fish with special marks were estimated to have been caught in marine commercial and sport fisheries from Pelican, Alaska, to Avila Beach, Calif., and Columbia River fisheries. The potential contribution for the four broods from the 13 study facilities, after adjustment for the effects of marking, was 1,433,300 fish. The value of the contribution was estimated at $12,027,000. Costs applicable to rearing were $2,859,700, yielding an average benefit to cost ratio of 4.2 to 1. Benefit to cost ratios at the 11 special mark hatcheries ranged from 0.3 to 1 to 17.1 to 1. The Columbia River Development Program (sub- sequently referred to as "Program"), initiated in 1949, was created to counteract the severe loss of salmon, Oncorhynchus spp., and steelhead trout, Salmo gairdneri, resulting from the expansion of water-use projects in the Columbia River system. The Program is a cooperative effort of fish man- agement agencies of the States of Oregon, Wash- ington, and Idaho and the Federal Government and is administered by the Columbia Fisheries Program Office, National Marine Fisheries Ser- vice, NOAA, Portland, Oreg. The Program's role has included two major functions: 1 ) the protection and improvement of stream environment which has included improvement of natural habitat, such as clearing obstructions from nearly 2,000 mi of tributary streams, building 87 fishways past natural barriers, and installation of 570 screens in diversion ditches and canals; and 2 ) the production offish in hatcheries which has been accomplished by the construction or modernization of 21 salmon and steelhead hatcheries on the lower Columbia River and tributaries. A supplementary function of the Program is funding operational improve- ment studies to complement the hatchery system. Major achievements have been: 1) improved marking techniques through development of the implanted coded wire fish tag (Bergman et al. 1968); 2) increased natural production through rehabilitation of chinook salmon runs in the Clearwater River system in Idaho and the Wil- lamette River system in Oregon; 3) determination of the physiological factors controlling downstream salmonid smolt migration through understanding the development of osmotic and ionic regulation in coho salmon (Conte et al. 1966), chinook salmon (Wagner et al. 1969), and steelhead trout (Conte and Wagner 1965), thus improving hatchery release timing; 4) reduced natural competition and predation through the development of Squaxin,^ a selective toxin to squawfish (MacPhee and Ruelle 1969); and 5) im- proved fish diets through development of the Ore- gon Moist Pellet (Hublou 1963). There are two major reasons for concentrating on hatchery produced salmon and steelhead trout: their life histories allow successful hatchery prop- agation and these species are historically and economically important to the United States. Over the past three decades Pacific salmon have ranked first or second in landed value of commercial 'Environmental and Technical Services Division, National Marine Fisheries Service, NOAA, 811 NE Oregon Street, P.O. Box 4332, Portland, OR 97208. Manuscript accepted April 1977. FISHERY BULLETIN: VOL. 76, NO. 1. 1978. ^References to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 179 FISHERY BULLETIN: VOL. 76, NO. 1 finfishes to U.S. fishermen. The net economic value of salmon sport fishing in the United States was $77.7 million in 1970 (Wahle et al. 1974). Initially, Program hatcheries were constructed to emphasize rearing of fall chinook salmon rather than coho and spring chinook salmon and steelhead trout because of a serious decline of this run in the early 1950's (Van Hyning 1973). Releases of migrant size fall chinook salmon have ranged from 10 million fish from 6 hatcheries in 1949 to 94 million fish from 17 hatcheries in 1973. Prior to the study reported by Worlund et al. (1969), little was known about the contribution of these releases to the commercial and sport fisheries. Some marking experiments had demon- strated that hatchery releases contribute to fisheries, but because such experiments were lim- ited and designed for other purposes, the contribu- tion had not been estimated. Although reports were written for each of the four broods of fall chinook salmon (Worlund et al. 1969; Rose and Arp^; Arp et al.^; Wahle et al.^), brood years were not compared and individual hatchery contributions, values, and benefits were not evaluated or compared. No new studies of this scale on the Columbia River have been initiated to supersede the 1962 through 1969 data. In addi- tion, the contributions, values, and benefits in the individual brood year reports are not comparable with those presented for Columbia River hatchery coho salmon (Wahle et al. 1974). Therefore, we compiled this report to supplement, summarize, and, in some cases, replace previously reported Columbia River hatchery fall chinook salmon con- tribution and value data. The marking study discussed in this paper, in- itiated in 1962 by the Columbia Fisheries Pro- gram Office, was designed to estimate the con- tribution of Columbia River hatchery-reared fall chinook salmon to the fisheries. The effort was brought about by the Bureau of the Budget (now ^Joe H. Rose, and Arthur H. Arp. 1970. Contribution of Co- lumbia River hatcheries to harvest of 1962 brood fall chinook salmon (Oncorhynchus tshawytscha). Unpubl. manuscr., 27 p. U.S. Fish Wildl. Serv., Bur. Commer. Fish., Columbia Fish. Program Off., Portland, Oreg. ••Arthur H. Arp. Joe H. Rose, and Steven K. Olhausen. 1970. Contribution of Columbia River hatcheries to harvest of 1963 brood fall chinook salmon {Oncorhynchus tshawytscha). Unpubl. manuscr., 33 p. Natl. Mar. Fish. Serv., Columbia Fish. Program Off., Portland, Oreg., Econ. Feasibility Rep. 5Roy J. Wahle, Arthur H. Arp, and Steven K. Olhausen. 1972. Contribution of Columbia River hatcheries to harvest of 1964 brood fall chinook salmon (Oncorhynchus tshawytscha). Unpubl. manuscr., 31 p. Natl. Mar. Fish. Serv., Columbia Fish. Program Off., Portland, Oreg., Econ. Feasibility Rep. 180 the Office of Management and Budget) which had declared a moratorium on hatchery construction until there was proof that further expansion would be economically justified. The experiment was confined to 12 hatcheries and 1 rearing pond that during the marking phase of the study propagated nearly 90% of all fall chinook salmon artificially reared in the Colum- bia River system. Locations of the participating and nonparticipating hatcheries rearing fall chinook salmon during the study period are shown in Figure 1. The marking of four brood years, 1961 through 1964, began in 1962 and data collection was completed in 1969. This report contains: 1) the experimental de- sign; 2) a description of the field operations; 3) estimation of 10 individual hatchery contribu- tions, values to fisheries, benefit to cost ratios for study facilities, and comparisons between hatch- eries; 4) the contributions, values, and benefit to cost ratios for each brood year marked for all par- ticipating hatcheries combined, with a compari- son of brood years; and 5) the contribution and value to the Pacific Coast fisheries of fall chinook salmon from all Columbia River hatcheries. EXPERIMENTAL DESIGN The experimental procedures for this study were the same for the four brood years. The design of the study is described by Worlund et al. (1969), and will be reviewed here. In general, 10% of the fall chinook salmon production from the par- ticipating hatcheries was marked by clipping fins and maxillary bones. The commercial and sport fisheries along the Pacific Coast were sampled for these marks. Individual and collective hatchery contributions can be estimated from: 1) proportion offish marked, 2) number of marks actually recov- ered, 3) fractions of the total catches sampled for marks by time and area in each fishery, and 4) information on any bias associated with applica- tion or detection of marks. The execution of this entire study required the cooperation of personnel from the following agencies: the Alaska Depart- ment of Fish and Game, the Fisheries Research Board of Canada (now the Department of Envi- ronment), the Washington Department of Fisheries, the Fish Commission of Oregon and the Oregon Game Commission (now the Oregon De- partment of Fish and Wildlife), the California De- partment of Fish and Game, the Bureau of Com- mercial Fisheries (now the National Marine WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON Fisheries Service), and the U.S. Fish and Wildlife Service, Bureau of Sport Fisheries and Wildlife. Allocation of Marks The experiment was limited to 13 rearing facilities on the Columbia River. The hatchery locations ranged from Big Creek Hatchery, the lowermost station, 40 km (25 mi) above the Co- lumbia River mouth, to Klickitat Hatchery, the uppermost station, 290 km (180 mi) above the Columbia River mouth (Figure 1). Approximately 10% of the production at each of the 13 facilities was marked with a common mark 64 KILOMETERS 40 MILES PARTICIPATING 1 - GRAYS RIVER 2 —BIG CREEK 3-ELOKOMIN 4 —LOWER KALAMA 5 -KALAMA FALLS 6 - WASH0U6AL 7 -BONNEVILLE 8 - CASCADE 9— OXBOW 10 -LITTLE WHITE 11 -SPRING CREEK 12 —BIG WHITE REARING PONDS 13 —KLICKITAT NONPARTICIPA TING 14 — KLASKANINE 15 - 16 - 17 - 18 - 19 - ABERNATHY TOUTLE LEWIS RIVER SPEELYAI SANDY 20 - EAGLE CREEK Figure l. — Locations of participating and nonparticipating Columbia River hatcheries rearing fall chinook salmon, 1961-64 broods. 181 FISHERY BULLETIN: VOL. 76. NO. 1 (Table 1). This mark consisted of clipping the adipose fin (Ad) and a right or left maxillary (RM or LM). The maxillary clip was alternated from one brood year to the next. In addition, a portion (as discussed later) of the production at 11 of the study hatcheries was marked with special marks. A portion of four broods at Spring Creek National Fish Hatchery and Kalama River hatcheries (in this study, Kalama Falls and Lower Kalama Hatcheries were treated as one facility) were marked with the following special mark: adipose, a ventral, and a maxillary clip. Spring Creek was assigned the adipose, left ventral (LV), and left or right maxillary clip. The maxillary clip was alter- nated among brood years. The 1961 brood was marked Ad-LV-RM, the 1962 brood was marked Ad-LV-LM, and so on. Kalama River hatcheries were assigned the adipose, right ventral (RV), and left or right maxillary clip. Again, the maxillary clip was alternated among brood years. Combina- tions of a single ventral and maxillary were alter- nated among eight other hatcheries: Elokomin, OxBow, Grays River, Cascade, Klickitat, Big Creek, Bonneville, and Little White Salmon. Two different hatcheries were marked with this com- bination for each brood year. Sources of Variation and Error Two major sources of variation in contributions to fisheries are differences among brood years and differences among hatcheries. To evaluate the dif- ferences among broods, four broods were marked. The variations among hatcheries were evaluated by special marking at four hatcheries for each brood year. One possible source of error in estimating con- tributions is the combination of differential rela- tive survival and differential maturation time for marked and unmarked fish. If the difference in marked and unmarked ratios at release and re- turn were due primarily to delayed maturation caused by marking, then marked fish may have been subjected to more intense fishing pressure due to a longer time in the ocean. This could mean the ratio of marked to unmarked fish in the fisheries would be greater than the ratio at release from the hatcheries. If this were true, the potential contributions would be overestimated in this re- port. However, since we are making the best esti- mate of contribution and benefit for the hatch- eries, we are assuming all differences in marked to unmarked ratios at release and return are due to 182 Table l. — fieleases of marked fall chinook salmon from Colum- bia River study hatcheries, 1961-64 broods. Percent Number production Brood Hatchery Mark' marked marked 1961 All hatcheries Ad-RM 5,446,439 10 15 Spring Creek Ad-LV-RM 1,133.019 10-37 Kalama Ad-RV-RM 475,964 9.70 Elokomin LV-RM 480,533 30.51 OxBow RV-RM 450,446 9.90 1962 All hatcheries Ad-LM 5,249,079 10.00 Spring Creek Ad-LV-LM 866,892 10.31 Kalama Ad-RV-LM 437,669 9.52 Grays River LV-LM 241.494 17.76 Cascade RV-LM 541,158 12.83 1963 All hatchenes Ad-RM 5,986.464 9.96 Spring Creek Ad-LV-RM 751.243 10.06 Kalama Ad-RV-RM 456,158 9.34 Klickitat LV-RM 521,610 18.06 Big Creek RV-RM 579,967 29.21 1964 All hatcheries Ad-LM 4,638,237 992 Spring Creek Ad-LV-LM 600,953 9.17 Kalama Ad-RV-LM 319,412 9,14 Bonneville LV-LM 957,110 9.68 Little White Salmon RV-LM 797,345 953 'Ad: Adipose; LV: Left ventral: RV: Right ventral; LM: Lett maxillary; RM: Right maxillary. differential survival between marked and un- marked fish. This point is discussed in detail under assumption 4. Straying of wild fish into the hatcheries, thus diluting the marked to unmarked ratios at return, is another source of variation and/or error. This dilution would reduce the relative survival rates for marked fish. To minimize this effect of varia- tion and/or error, average relative survival figures for common and special marked fish were calcu- lated and used in the contribution computations. Estimating Procedures A formal account of the estimating procedures is presented in the report by Worlund et al. (1969). Simple numerical examples will be used to explain the procedure in this report. Estimating the poten- tial contributions and values of hatchery fall chinook salmon required four steps. First, the number of marked and unmarked hatchery re- leases had to be estimated. Second, the estimated catch of marked fish was calculated. Third, the total contribution of hatchery fish was estimated. Fourth, dollar values were applied to the contribu- tion estimates. Hatchery Releases The numbers of marked and unmarked fish in hatchery releases were estimated by sampling the hatchery population with a 10-part sampler (see Marking and Release Procedures). This device WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON was precalibrated from a number of trials with known numbers of fish to find the average per- centage retained by a single closed pocket. The following example illustrates the fish enumera- tion procedure for a pond of fall chinook salmon. Suppose a precalibrated pocket is found to remove a 10.1'7f sample. Also, suppose after passing all the fish in a pond through the sampler, the number offish retained by the closed pocket is found to be 20,200. The total number of fish in that pond is then estimated as 20,200 0.101 = 200,000. Sup- pose further that of the 20,200 fish retained in the pocket, 2,020 fish are found to be marked. Then 2,020/20,200 = 10% of the estimated 200,000 fish in the pond, or 20,000 are estimated to be marked and 180,000 unmarked. The total release, num- bers marked ( common and special ) and unmarked, were estimated for a hatchery by summing data from all ponds. Catch of Marked Fish To estimate the catch of marked fish in a given area and fishery, the following values were needed by time period: total catch; number of fish examined for marks; number of marked fish by species, mark type, and age; and the proportion of each age-group in the total catch. The sampling seasons were stratified into relatively small time units (usually 2-wk periodsK The estimated catches of a particular mark were summed over the entire fishing season for a given area and fishery. For example, during the period from 26 June through 9 July 1966 in the Ilwaco sport fishery, 1,193 chinook salmon from a total catch of 5,664 were examined for marks, for a 21.1% sam- ple. Samplers found one Ad-LM marked 1964- brood (2-yr-old) fall chinook salmon during this period. Then the estimated catch of 1964-brood Ad-LM marked fall chinook salmon during this period was 1/0.2106 = 5. Catches of 1964-brood Ad-LM marked chinook salmon for the Ilwaco sport fishery in 1966 were summed for 13 time periods. This resulted in an estimated catch of 196 Ad-LM marked fish. This procedure was carried out for each port sampled and each mark found. Catch data for each time-location stratum were provided by manage- ment agencies. Commercial catches were esti- mated from total landing weights and average fish size data or from total numbers of salmon landed and species composition estimates. Sport catches were estimated from measures of total effort and catch-per-unit-efTort or from salmon punch cards and independent sampling. All catch and sampl- ing information was transferred to computer cards and estimates were calculated by computer. Un- published reports of catch and mark data were produced for 1963 through 1969 by the Seattle Biological Laboratory, Bureau of Commercial Fisheries (now the Northwest and Alaska Fisheries Center, National Marine Fisheries Ser- vice, NOAA). Contribution of Hatcher) Fish Maxillary regeneration occurred during the ocean lives of some of the common and special marked chinook salmon, resulting in partial marks (see Assumptions). For example, a 1961- brood Kalama Ad-RV-RM mark could have regen- erated to an Ad-RV mark, or a 1962-brood Ad-LM common mark could have regenerated to an Ad- only mark. Partial marks were a result of this regeneration and/or an occurrence of naturally marked fish. If partial marks due to regeneration were not claimed as part of the marked hatchery fish total, the hatchery contribution would be un- derestimated considerably. Therefore, we examined the ocean catches of chinook salmon with partial marks to determine the number that could be claimed as hatchery fish. A comparison of maxillary regeneration rates of marked fish held at Bowman Bay (Worlund et al. 1969) and the occurrence of Ad-SV (adipose-single ventral ) and Ad-only partial marks in the fisheries (Table 2), led us to believe Ad-LV, Ad-RV, and Ad-only marks occurred because of maxillary re- TabLE 2. — Percent partial mark occurrence in the ocean and Columbia River fisheries and in hatchery returns, 1961-64 broods. Brood Partial marks' Region Ad-SV2 Ad SV Ocean fisheries 1961 15.8 14.6 74.9 1962 18.8 23.5 72.7 1963 8.0 9.1 36.4 1964 12.8 15.2 39.7 Columbia River fisheries 1961 10.3 7.8 51.0 1962 17.4 5.0 57.4 1963 95 6.0 7.0 1964 80 7.2 283 Hatchery returns 1961 109 16.1 27.7 1962 19.8 220 20.0 1963 8.3 8.6 2.0 1964 11.2 172 12.5 ' Figures are ratios, averaged for all years by brood, of estimated numbers of partial marks to estimated sum of partial marks and corresponding complete marks expressed in percent. ^SV signifies single ventral." Marks of same general type are combined. 183 FISHERY BULLETIN: VOL. 76. NO. 1 generation. This belief is also supported by the absence of Ad-LV and Ad-RV marks in the 1965- brood catches of chinook salmon (Bureau of Com- mercial Fisheries^' ''• ^; Fish Commission of Ore- gon^). The Ad-V marks were not assigned to the 1965-brood fish. Thus, we have claimed all Ad-RV, Ad-LV, and Ad-only marked chinook salmon as hatchery fish. However, the percentage occurrence of SV marks in the fisheries was much higher than 1 ) the maxillary regeneration rate, 2) the occurrence of Ad-SV marks in the fisheries, and 3) the occur- rence of SV marks in hatchery returns. Thus, we concluded SV marks occurred because of maxil- lary regeneration and natural marks. Two steps were required to determine the number of SV marked fish we would claim as part of the hatchery production. First, we assumed the maxillary regeneration rate for all special marked hatcheries was the same. The partial mark per- centages for Kalama River and Spring Creek com- bined were calculated for each fishery, year, and brood. For example, in the 1964 Washington commercial fisheries the estimated catch of 1961- brood Ad-LV-RM and Ad-RV-RM full marked fish was 1,001 and Ad-LV and Ad-RV partial marked fish was 232. The partial mark percentage for this year, fishery, and brood was then 232/1,001 = 23%. Second, full mark recoveries from other special mark hatcheries (Elokomin, OxBow, Grays River, Cascade, Klickitat, Big Creek, Bonneville, and Little White) for the corresponding brood, year of recovery, and fishery were multiplied by the Kalama-Spring Creek percentages. For example, the estimated full mark recoveries of Elokomin and OxBow 1961-brood chinook salmon in the 1964 Washington commercial fisheries were 48 and 58 fish respectively. The SV marked fish claimed as part of Elokomin and OxBow hatch- *Bureau of Commercial Fisheries. 1969. Data report: Colum- bia River fall chinook salmon hatchery contribution study; 1967 sampling season. Unpubl. manuscr., 519 p. U.S. Fish Wildl. Serv., Bur. Commer. Fish., Seattle Biol. Lab. 'Bureau of Commercial Fisheries. 1970. Data report: Colum- bia River fall chinook salmon hatchery contribution study: 1968 sampling season. Unpubl, manuscr., 437 p. U.S. Fish Wildl. Serv., Bur. Commer. Fish., Seattle Biol. Lab. ^National Marine Fisheries Service. 1971. Data Report: Co- lumbia River fall chinook salmon hatchery contribution study: 1969 sampling season. Unpubl. manuscr., 283 p. Natl. Mar. Fish. Serv., Seattle Biol. Lab. "Fish Commission of Oregon. 1972. 1970 fin-mark sampling and recovery report for salmon and steelhead from various Pacific coast fisheries. Unpubl. manuscr., 102 p. Fish Comm. Oreg., Biom. Sect., Clackamas. eries' production were then 48 x 0.23 = 11 and 58 X 0.23 = 13 respectively. In cases where the calcu- lated claimed partial marks were greater than the partial marks actually recovered, all partial marked fish were claimed. No SV marked fish were claimed for the southeastern Alaska or California fisheries because few Columbia River hatchery special marked fish were captured in these fisheries. The claimed partial marked fish estimates by year and fishery were summed for each special mark hatchery. The sums are the number of par- tial marked fish we claimed as part of the special mark hatcheries' catch (Table 3). Loss of maxillaries due to hooking occurred dur- ing the ocean lives of the marked fall chinook salmon (author's pers. obs. ), resulting in the possi- ble misidentification of marks. In some cases a marked chinook salmon was assigned to a certain brood year from scale analysis, but the fish had the wrong maxillary mark for that brood. For exam- ple, 1961-brood Ad-LM marked chinook salmon, 1962-brood Ad-LV-RM marked fish, 1963-brood LV-RM marked chinook salmon, and so on (see Table 1 for correct marks for each brood) were reported to have been caught in the fisheries. In some cases, double maxillary marks (1961-brood Ad-RM-LM, 1963-brood Ad-LV-RM-LM, etc.) were reported to have been caught. Duplication of marks or use of marks with the opposite maxillary for the same brood year were prevented by the Pacific Marine Fisheries Com- TABLE 3.— Estimated catches of 1961- to 1964-brood fall chinook salmon from Columbia River study hatcheries with full marks, misidentified marks, partial marks, and partial marks claimed as study hatchery fish by brood and hatchery. Misiden- Partial Total Full tified Partial marks estimated Brood Hatchery marks marks' marks claimed marks 1961 All study 18,906 621 2,710 2,710 22,237 Spring Creek 3,553 115 732 732 4,400 Kalama 1,955 34 186 186 2,175 Elokomin 174 18 533 43 235 OxBow 266 19 594 51 336 1962 All study 6,008 512 1,366 1,366 7,886 Spring Creek 769 26 172 172 967 Kalama 498 48 113 113 659 Grays River 177 8 373 30 215 Cascade 140 21 418 30 191 1963 All study 19,856 489 1,838 1,838 22,183 Spring Creek 2,210 48 149 149 2,407 Kalama 1,053 60 144 144 1,257 Klickitat 1,048 702 396 108 1,858 Big Creek 772 71 479 71 914 1964 All study 11,085 489 1,740 1,740 13,314 Spring Creek 3,798 99 509 509 4,406 Kalama 849 54 102 102 1.005 Bonneville 649 43 210 70 762 Little White 274 6 392 23 303 ' Double maxillary clips or the opposite maxillary for a particular brood year. 184 WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON mission. We are assuming aging was correct (see Assumptions). Therefore, we have assumed marked fall chinook salmon with a double maxil- lary or the wrong maxillary for a particular brood were misidentified. Thus we claimed these marked fish as part of the Columbia River hatch- ery marked fall chinook salmon catch (Table 3). Therefore, estimated catches of Columbia River hatchery marked fall chinook salmon (Tables 3, 8-14) include full, misidentified, and claimed par- tial marked fish. Before estimating the contribution of hatchery fall chinook salmon if no marking had taken place (hereafter referred to as potential contribution), the survivals of common marked fish had to be calculated. Three methods were used to estimate the common mark relative survival and a median relative survival was calculated from the three answers. METHOD 1.— All 13 study facilities were com- bined and four sums — marked releases, un- marked releases, marked returns, and unmarked returns — were obtained for each brood year. The marked to unmarked ratio at return was then divided by the marked to unmarked ratio at re- lease. The formula is: Marked returns Unmarked returns Marked releases Unmarked releases = Relative survival. METHOD 2.— If wild fish strayed into the study hatcheries, diluting the marked to unmarked ratios at return, method 1 would underestimate relative survival. Thus to allow for straying, in method 2 we have calculated relative survivals using releases and returns from four selected hatcheries. Cascade, OxBow, Little White Salm- on, and Spring Creek, on streams with no natural runs of fall chinook salmon. Relative sur- vivals were estimated for each brood in the same manner as described in method 1. METHOD 3.— Even for the four selected hatch- eries, straying of wild fish into hatcheries is a possibility, resulting in an underestimated rela- tive survival. To account for this possibility, a method was devised to estimate the number of wild fish straying into the four selected hatcheries. This was done in four steps. First, since the selected hatcheries are between Bonneville and The Dalles Dams, an estimate of the maximum number of fall chinook salmon spawning between the dams was obtained by subtracting both the Indian and sport fall chinook salmon catches be- tween Bonneville and The Dalles Dams as well as The Dalles Dam fall chinook salmon count from the Bonneville Dam fall chinook salmon count. Second, the maximum number offish spawning at sites other than the selected hatcheries was ob- tained by subtracting the four hatcheries returns from the total spawners between the dams. Third, the age of fish spawning at sites other than the selected hatcheries was approximated by applying age data from Columbia River gillnet fall chinook salmon catches. Fourth, straying factors (from observed straying offish marked at Spring Creek Hatchery) were applied by brood and age to the wild spawners to obtain the estimate of wild fish straying into the selected hatcheries. These estimates are maximum since we cannot account for mortalities, uncounted fish passing through navigation locks, double counting offish that fall back over dam spillways and again ascend the fish ladders, or fish straying from the four hatcheries. Also, we assumed wild fish had the same straying pattern as the hatchery fish in this study, i.e., they strayed to sites near their area of origin. Once the brood estimate of the number of wild fish entering the hatcheries was obtained, it was subtracted from the appropriate unmarked re- turns. The resulting unmarked hatchery return quantity for each brood was then used in the for- mula described in method 1 to calculate the third estimated common mark relative survival. Examples of the calculations used to obtain the three values for the common mark relative survi- vals are presented by Worlund et al. (1969). The median common mark relative survivals for the 1961-64 broods of Columbia River study hatchery fall chinook salmon are: Common mark 3 rood relative survival 1961 0.608 1962 0.477 1963 0.372 1964 0.448 Special mark relative survivals also had to be calculated to estimate contributions of special marked hatcheries. Calculating special mark rel- ative survivals for each hatchery was impossible because seven hatcheries (Elokomin, OxBow, 185 FISHERY BULLETIN VOL. 76, NO. 1 Grays River, Cascade, Klickitat, Bonneville, and Little White) had too few special mark returns to obtain reliable estimates of marked to unmarked ratios at return. Thus returns to only three hatch- eries (Spring Creek, Kalama, and Big Creek), hav- ing sufficient special mark returns, were used to calculate average special mark relative survivals for each brood. However, if special marked fish from the other seven hatcheries had lower relative survivals than the average, the contributions of these hatcheries would be underestimated using this method. Relative survivals of special marks to common marks were first calculated using the formula: Special mark return/Common mark return Special mark release/Common mark release " The relative survivals are: Table 4. — Mark percentages at release for common and special marked fall chinook salmon by brood year and hatchery. Spring Kalama Big Brood Creek River Creek 1961 0.526 0.800 — 1962 0.617 0.472 1963 0.535 0.498 0.797 1964 0.535 0.731 Percent of brood marked Mark type and hatchery 1961 1962 1963 1964 Common marks' All hatcheries 10.7 10.4 10.4 10.5 Special marks^ Spring Creek^ 7.8 7.3 7,6 7.0 Kalama River 97 9.5 9.3 9.1 Elokomin 30.5 — — — OxBow 9.9 — — — Grays River — 17.8 — — Cascade — 12.8 — — Klickitat — — 18.1 — Big Creek — — 29.2 — Bonneville — — — 9.7 Little White Salmon — — — 9.5 'Special marks not included. ^Common marks included with unmarked releases. ^Includes Big White Salmon pond releases. The potential contributions of the hatchery fall chinook salmon were calculated by dividing the estimated catch of marks by the marked fish rela- tive survival times the mark proportion at release. The formula for calculating the potential con- tributions of Spring Creek, Kalama River, and other special mark hatcheries is: Estimated catch of spec, marks (Spec, mark relative survival)! Spec, mark proper, at rel.) • From these values we concluded that special marked fish survived between 50 and 80% as well as common marked fish. Multiplying the common mark relative survivals by 50 and 80% for each brood year yielded the following average special mark relative survivals: Brood 1961 1962 1963 1964 Survival 0.395 0.310 0.242 0.291 The next step was to determine the mark pro- portions at release for common and special marks for each brood year. Special marks were excluded from the calculation of the common mark propor- tions. This was done for two reasons: special marked fish had a lower relative survival than the common or unmarked fish, and the special marks could be identified in the fisheries and related back to specific hatcheries. The common marked fish had to be treated as unmarked fish in calculating the special mark proportions at release because common mark catches could not be related to spe- cific hatcheries. These mark porportions at release are presented in Table 4. The potential contribution of all study facilities was calculated with the formula: Estimated catch of common marks (Common mark relative survival )( Common mark proper, at rel. ) + Potential catch of spec, marks. The potential catch of special marks is an esti- mate of the special marks that would have been caught if marking had not caused differential mortality. The formula used to calculate this po- tential catch is: Estimated catch of special marks Special mark relative survival Value of Hatchery Contribution With estimates of the potential contribution of Columbia River hatchery fall chinook salmon, the potential value of the catches could be calculated from average weight and unit price data. The av- erage weights for the commercially caught fish were obtained from common marked fish. Total weights of hatchery fish caught in the commercial fisheries are underestimated with this method be- 186 WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON cause marked fish are smaller than unmarked fish (Cleaver 1969). Weights for the ocean troll fish- eries are dressed weights and those for Columbia River net fisheries are round weights. Ex-vessel market prices have been used to represent esti- mated net values for commercially caught fish. The ex-vessel prices were obtained from Washington Department of Fisheries records for the appropriate years and age of fish. (D. Ward, Washington Department of Fisheries, pers. com- mun.) Washington troll prices were used for other commercial fisheries on the Pacific Coast. The net value for salmon and steelhead sport fishing is estimated to be $20/day of fishing. This value results from reconciling the existing re- search that is closely related to estimated net economic values of Columbia River sport caught salmon. The maximum potential benefits from sport fishing at a single market price is predicted at $20/fishing day (Brown et al.^"). The salmon catch per angler trip data were obtained from Washington, Oregon, and California publications (Campbell and Locke 1964, 1965, 1966, 1967, 1968, 1969; Nye and Ward undated a, b; Green- hood and Mackett 1967; Haw et al. 1967; Heimann and Frey 1968a, b; Heimann and Carlisle 1970; Pinkas 1970). An estimate of 1.09 salmon/angler trip was obtained by averaging data for the three States over the appropriate years. The $20/angler trip was divided by 1.09 salmon/angler trip to yield a value of $18.35/salmon. This value was used in the ocean sport and Columbia River sport fisheries for all broods and years of capture. Assumptions Six assumptions are required in our method for estimating contributions of hatchery fall chinook salmon to the fisheries. Three basic assumptions are: 1) a marked fish is identifiable as a marked fish throughout life, 2) all fish detected and re- ported with the kind of mark applied at the hatch- eries are hatchery fish, and 3) chinook salmon are correctly aged from scale examinations and in- formation on size offish and date of capture. Two assumptions as to the behavior of marked and unmarked hatchery fish are: 4) marked and un- marked hatchery fish have the same survival '"William G. Brown, Ashok K. Singh, and Jack A. Richards. 1972. Influence of improved estimating techniques on predicted net economic values for salmon and steelhead. Unpubl. man- user., 26 p. Oreg. State Univ., Agric. Exp. Stn., Corvallis. rates and maturity schedules, and 5) marked and unmarked hatchery fish have the same ocean dis- tribution and are equally vulnerable to the fisher- ies. Finally, because part of all hatchery releases bear the same mark, we assume: 6) common marks were applied to the same proportion of each hatch- ery's production in a given year. The appropriateness of the estimating proce- dures is dependent on the validity of these as- sumptions. Assumption 1 was tested by holding marked fish in saltwater ponds for periodic examination of the condition of the mark. There was no regeneration of the adipose fin. However, regeneration of ventral fins and maxillary bones did occur. In most cases, the ventral fin regener- ated to <25% of its original size. Greater regener- ation was identifiable by deformation of the fin rays. The high occurrence of maxillary regeneration (7-12%) for the 1961- and 1962-brood chinook salmon resulted in the removal of more of the maxillary bone in the 1963- and 1964-brood fish. This change in marking procedure resulted in a smaller percentage offish with regenerated maxil- laries (1-3%). Since single and double fin marks were associated with maxillary clips, even when maxil- laries completely regenerated, the fish were iden- tifiable as marked fish. Thus we believe assump- tion 1 to be true. The validity of assumption 2, the absence of natural marks on hatchery and wild fish, was tested in two ways: First, over 30 million hatchery fingerlings were examined during marking for naturally missing adipose and ventral fins. Only 156 missing adipose and 201 missing ventral fins (none together) were observed indicating the in- significance of naturally occurring marks on these fish. Second, the occurrence of natural marks out- side the hatchery system was checked by examin- ing 1965-brood chinook salmon catches for study marks. The allocation of study marks to any 1965 brood on the Pacific Coast was to have been pre- vented. Unfortunately, the attempt to prevent the application of study marks to this brood was not completely successful. However, no adipose- ventral-maxillary combinations were applied and none were found in the fisheries. Any occurrence of natural marks like those claimed as hatchery marks has been accounted for under Estimating Procedure. Therefore, we believe assumption 2 has been satisfied. Assumption 3 was evaluated by testing scale 187 FISHERY BULLETIN: VOL. 76, NO. 1 readers with chinook salmon scales of known age. Scales from 400 marked fish of known age were submitted to six readers: two from the Fish Com- mission of Oregon and one each from the Fisheries Research Board of Canada, Washington Department of Fisheries, Oregon Game Commis- sion, and Bureau of Commercial Fisheries. Length offish and date of capture were available for each scale. The six scale readers correctly aged 83*% of the 400 test scales ( Worlund et al. 1969). Thus, we believe that assumption 3 is reasonably well satisfied. The equality of marked and unmarked survival rates and maturity schedules, assumption 4, needs some additional study. A lowering of the marked to unmarked ratio at the hatchery from the time of release to the time of return indicated possible problems with this assumption. There are several possible reasons for this change in marked to un- marked ratio. They are: 1) errors in estimating the number of marked and/or unmarked hatchery fish at the time of release; 2) a difference in distribu- tion or timing of marked and unmarked fish, re- sulting in the marked fish being exposed to a more intense fishery; 3 ) a selectivity of some fisheries for marked fish; 4) a greater amount of straying for marked fish than unmarked fish; 5) a difference in maturation schedule for marked and unmarked fish; 6) differential survival between marked and unmarked fish because of marking; and 7) mis- takes in aging unmarked hatchery returns. It is unlikely the difference in the marked to unmarked ratios at the time of release and return could have been caused entirely by mistakes in estimating the ratio at release. The differences were too great, considering the randomness of the estimating procedures and the number of hatch- eries involved. There is no way to determine nor reason to believe differences in distribution, tim- ing, or straying between marked and unmarked fish caused the differences in the ratios at release and return. Nor is there any way to determine or reason to believe any fishery was selective for marked fish. Thus we rejected these as possible reasons for the change in marked to unmarked ratios between the time of release and return. There is some indication a difference in time of maturing did occur between marked and un- marked fish (Cleaver 1969). Examination of the marked to unmarked ratios at the hatcheries by year of return shows a trend of increasing ratios. This indicates the marked fish did not mature as soon as the unmarked fish. The marked fish ap- 188 peared to stay in the ocean longer and thus were subject to a higher natural and fishing mortality. It is also possible clipping fins and maxillary bones caused mortality after the fish were released from the hatchery. The unmarked fish would obvi- ously not be subjected to this mortality. Mistakes in aging of unmarked hatchery re- turns could easily have occurred because of the poor condition of the fish when entering the hatch- ery. The scales had been partially resorbed, mak- ing them difficult to read. Since the same marks were used in alternate brood years, the mark and size of the fish would aid the aging procedure for the marked fish. This would result in more accu- rate aging of marked than unmarked fish. How- ever, the errors in aging unmarked fish could have been self cancelling. Possible errors in aging seemed to be a very minor reason for the differ- ences in the marked to unmarked ratios. Thus the two most probable reasons for the change of marked to unmarked ratios from the time of release to return were differences in mat- uration schedule and differential survival of marked and unmarked fish. These two problems probably acted in combination. Since we have no way of separating the effects of delayed maturity and differential survival and since we are making the best estimate of hatchery contribution, we are assuming the change in marked to unmarked ratio was due only to differences in survival of marked and unmarked fish. Correction factors were applied to adjust for the differential survival. The validity of assumption 5, equal ocean dis- tribution and vulnerability to the fisheries for marked and unmarked fish, is supported by ocean tagging studies showing similar ocean distribu- tion for marked and unmarked hatchery fish (Cleaver 1969). Common marks were applied to 10 or 11% of the production at the 13 study facilities for the four brood years, 1961 through 1964 (Table 4). The percentages ranged from 9 to 1 1 among the hatch- eries for each brood. With these ranges we feel assumption 6, application of common marks to the same proportion of each hatchery's production, is satisfied. FIELD OPERATIONS Marking and Release Procedures Artificial propagation procedures were similar at all 13 study facilities. Adult fall chinook salmon WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON returned to these facilities and were spawned dur- ing September and October. Fry reached free swimming stage in February or March and were then placed in ponds. They were reared 90 to 120 days in the hatchery and released at an average length of 6 to 8 cm (2-3 in). Since there was consid- erable variation in time and size of release be- tween hatcheries and brood years, we have in- cluded Table 5 to complete the release procedure record. After the hatchery fall chinook salmon spent 1 to 6 yr in the ocean, where they were available to sport and commercial fisheries from southeastern Alaska to central California, they matured and returned to the Columbia River. The marking phase of this study extended from June 1962 through June 1965. Approximately 10*7^ of the 1961-64 broods were marked. A "10- part sampler," a modified sampling tool (Worlund et al. 1969), was used to obtain the sample offish for marking. The sampler consisted of a cylindri- cal liner containing a circular metal frame. The frame was divided into 10 equal pie-shaped sec- tions with a zipper-bottomed net pocket hung from each section. To obtain a sample for marking, the zippers on one or more pockets were closed, the frame and liner were placed in a water-filled tub, and 18 kg (40 lb) offish were placed into the liner. The closed pocket, or pockets, retained the desired sample when the liner and frame were lifted. The fish remaining in the tub were placed into another pond. This procedure was followed until all chinook salmon in each pond were processed. In the case of the special mark hatcheries, two or more pockets were closed. One pocket retained the fish for common marking and the other pockets retained those for special marking. The intention was to apply special marks to between 500,000 and 1.0 million chinook salmon at each of the special mark hatcheries. We felt this number would pro- vide a statistically sound number of special mark recoveries for each hatchery. The hatchery man- ager's estimate of the number of fall chinook salm- on on hand at the time of sampling was used to determine how many pockets to close at each hatchery to obtain the desired sample for special marking. These estimates were sometimes inac- curate, resulting in a smaller or larger sample than had been desired. Fish to be marked were anesthetized with MS-222 (tricaine methanesulfonate). The fins and maxillary bones were clipped with bent-nosed scissors. Marked fish were held in hatchery troughs until they recovered from the anesthetic, then returned to the group of unmarked fish from which they came. Mark quality control was main- tained by sampling 100 marked fish per marker at irregular periods each day and grading them ac- cording to quality of mark. Each year over 100,000 marked fish were sampled and graded. This grad- ing indicated a high mark quality was attain- ed. The entire production of fall chinook salmon at the study hatcheries was sampled with the 10-part sampler prior to release to estimate the marked and unmarked releases. The "107c" samples were set aside and resampled to obtain a "1%" sample which was sorted into marked and unmarked groups, counted, and weighed. The counts and the estimate of the proportion removed by the particu- lar sampler were used to estimate the numbers of marked and unmarked fish released. Over 213 million 1961-64-brood fall chinook salmon were released from the study hatcheries. Of these, 21.3 million were given the common mark and 9.6 million were given a special mark (Table 1). Table 5. — Size and date of release of 1961-64 broods of fall chinook salmon from Columbia River hatcheries participating in the fall chinook salmon study by hatchery and brood. 1961 brood 1962 brood 1 963 brood 1964 brood Hatchery Size' Date Size' Date Size' Date Size' Date Grays River 169 5 24 62 141 5/31/63 114 6/ 1 '64 108 5/26/65 Elokomin 202 5/24 62 206 5/20/63 181 5/9/64 134 5/12/65 Kalama Falls 356-202 6; 1-7/3 1/62 226 6/4/63 198 615/64 177 6/20/65 Washougal 187-107 5-6,62 180 5/22/63 153 5.25/64 139 5/2/65 Little White 180 6,2262 227- 83 6/5-8/15/63 200 6/18/64 177 6/65 Spring Creek 289-173 4 9-5,11 62 282-149 4,8-6/13/63 273-206 4/1 2-5/1 2'64 250-142 4/11-5/4/65 Big White 182 511,62 190 6/17/63 181 5/12 64 85 6/29/65 Klickitat 166 4,23;62 164 4/20/63 148 4/29/64 132 5/5/65 OxBow 217 5/10/62 195 5/14/63 189 5/6/64 170 6/19/65 Cascade 318 5/20/62 192 6/24/63 215 6/12/64 146 6/'29/65 Bonneville 312 6/6,62 152 6/19/63 136 6/26/64 154 6/24/65 Big Creek 174 52 62 137 5,7/63 102 513/64 91 6/2/65 Lower Kalama 261 6/2/62 199 5/18/63 139 518/64 169 5 18/65 'Fish per pound. 189 FISHERY BULLETIN: VOL 76, NO 1 Mark Recovery The sampling phase of this study began in 1963 and was completed in 1969. Table 6 shows the marks and the age of the marked fish in the fisheries during these years. Sampling and catch estimation procedures are explained under Catch of Marked Fish. Sampling for these fish occurred in the major ocean sport and commercial fisheries from southeastern Alaska to central California, the Columbia River fisheries (Figure 2, Table 7), at parent hatcheries, and certain natural spawn- ing grounds (Worlund et al. 1969; Rose and Arp see footnote 3; Arp et al. see footnote 4; Wahle et al. see footnote 5). During the first sampling year, 1963, only Washington and Oregon ocean fisheries, Columbia River fisheries, and hatchery returns were examined for marks. In 1964, the sampling was expanded to include most chinook salmon fisheries from Avila Beach, Calif, to Peli- can, Alaska. The Puget Sound sport fishery was not sampled in 1964. The British Columbia purse seine fishery was not sampled in 1966. The sam- pling of the southeastern Alaska gillnet fishery Table 6. — Ages of marked Columbia River fall chinook salmon in catches and escapements by brood (1961-64) and sampling years ( 1963-69). Mark' Hatchery Year of sampling Brood 1963 1964 1965 1966 1967 1968 1969 Ad-RM 12 hatcheries Years old 5 1961 2 3 . 4 Ad-LV-RM Spring Creek 2 3 4 5 Ad-RV-RM Kalama 2 3 4 5 RV-RM OxBow 2 3 4 5 LV-RM Elokomin 2 3 4 5 1962 Ad-LM 12 hatcheries 2 3 4 5 Ad-LV-LM Spring Creek 2 3 4 5 Ad-RV-LM Kalama 2 3 4 5 RV-LM Cascade 2 3 4 5 LV-LM Grays River 2 3 4 5 1963 Ad-RM 12 htacheries 2 3 4 5 Ad-LV-RM Spring Creek 2 3 4 5 Ad-RV-RM Kalama 2 3 4 5 LV-RM Klickitat 2 3 4 5 RV-RM Big Creek 2 3 4 5 1964 Ad-LM 12 hatcheries 2 3 4 5 Ad-LV-LM Spring Creek 2 3 4 5 Ad-RV-LM Kalama 2 3 4 5 RV-LM Little White Salmon 2 3 4 5 LV-LM Bonneville 2 3 4 5 No, of marks in catches and escapements 5 10 15 20 15 10 5 'Ad adipose: LV left ventral: RV right ventral, LM: left maxillary: and RM right maxillary Table 7. — Areas where catches were examined for marked fall chinook salmon of Columbia River origin by port or zone of landing and type of fishery. Area sampled Sport fishery Rod and reel Commercial fisheries Troll Gill net Dip net Purse seine Southeast Alaska British Columbia Washington ocean Puget Sound Oregon ocean California ocean Columbia River Sekiu Neah Bay La Push Westport llwaco Zones 6-12 Warrenton Depoe Bay Newport Florence Reedsport Coos Bay Gold Beach Brookings Crescent City Eureka Fort Bragg San Francisco Monterey Zones 1-5 Zones 1, 3-15, 18, 22 Alaska area Zones 29, 40-43, Area C, Seattle Neah Bay La Push Westport llwaco Astoria, Tillamook Nestucca, Depoe Bay Newport Florence Reedsport Coos Bay Port Orford Brookings Crescent City Eureka Fort Bragg San Francisco Monterey Zones 1, 6, 8, 11, 15. 18, 19 Zones 29, 40, 41-43 Juan de Fuca Strait Grays Harbor Willapa Bay Zones 40-43 Zones 1-6 Klickitat River 190 WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON Figure 2. — Ports and zones sampled for marked fall chinook salmon of Columbia River origin. 191 FISHERY BULLETIN: VOL. 76, NO. 1 was discontinued after 1966 and the Alaska troll fishery sampling stopped after 1967. Over the 7 yr of sampling, 3.3 million chinook salmon were examined for marks and 208,000 were sampled for age. This was an average sampling percentage of 20 and l^r for marks and age, respectively. The yearly mark sampling rate ranged from 14 to 28% of the catch and the age sampling ranged from 1 to 4%. Enumeration of Returns Returns to all study facilities were counted and examined for marks. Age, length, and sex data were also collected from 25 to 50 unmarked chinook salmon/wk at each hatchery. Returns to five other Columbia River hatcheries ( Abernathy, Speelyai, Toutle, Klaskanine, and Sandy) were also examined for marks. Total hatchery returns for the 1961-64 broods of fall chinook salmon were 155,783, of which 8,527 were marked. Hatchery and adjacent fall chinook salmon spawning streams were surveyed to estimate natural spawning of hatchery fish. The Klickitat, Big White Salmon, Little White Salmon, Wind, Washougal, Kalama, Lewis, Elokomin, and Grays Rivers and Plympton and Big Creeks were sur- veyed in 1964, 1965, and 1966. The surveys were designed to estimate the total spawning popula- tion and to gather mark, age, and length data. During the 3 yr, 62,400 chinook salmon were examined of which 1,600 were marked. The stream surveys were discontinued after 1966 be- cause of a funding reduction. INDIVIDUAL HATCHERY MARK CATCH AND POTENTIAL CONTRIBUTION, 1961-64 BROODS In this study 12 hatcheries and one rearing pond were marked with a common mark for four brood years. All but two of these facilities (Big White Salmon Pond and Washougal Hatchery) had a por- tion of at least one brood year's production marked with a special mark. A portion of all four brood years' production at Spring Creek and the two Kalama River hatcheries were marked with spe- cial marks. This special marking was done to give an indication of the migration patterns and con- tributions to the fisheries for each individual hatchery in the study. The estimated catches and potential contributions will now be presented for each hatchery with special marks. 192 Spring Creek National Fish Hatchery, 1961-64 Broods Spring Creek Hatchery was allocated the Ad, LV, combination mark for the four brood years. The RM mark was used in combination with the Ad-LV mark for the 1961 and 1963 broods and the LM mark was used with the 1962 and 1964 broods. Approximately 10% of Spring Creek's production was marked for each brood year. The number of fish given special marks ranged from 1.1 million for the 1961 brood to 600,000 for the 1964 brood. Spring Creek special marked chinook salmon were available to the ocean and Columbia River fisheries from 1963 through 1969. During this 7-yr period, we estimated 12,180 special marked fish were recovered in the fisheries (Table 8 ). Over 65% of the fish were captured in their third year of life, with nearly 27% taken as 4-yr-olds. Ocean re- coveries occurred primarily from the Columbia River mouth north to the west coast of Vancouver Island. Fisheries in the marine areas took 74% of the fish, with 26% being caught in the Columbia River commercial fisheries (Figure 3). The potential contribution of Spring Creek chinook salmon (had no marking taken place) was estimated at 401,700 fish for the four broods com- bined. The average Spring Creek contribution to the fisheries for the four broods combined was 12 Southeastern Alaska COMMERCIAL 0 % Bntisti Columbia COMMERC lAL 1 25% Washington SPORT 1 20% COMMERCIAL 1 24 % Oregon SPORT ] 3% COMMERCIAL ] 2% California SPORT 0 % COMMERCIAL 0 % Columbia River SPORT 0 % GILLNET 1 20% INDIAN 3 6% J 20 40 60 SO 10 PERCENTAGE OF CATCH Figure 3.— Percentage of catch of 1961- to 1964-brood fall chinook salmon from Spring Creek National Fish Hatchery taken by area and fishery, 1963-69. WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON Table 8. — Estimated catches of special marked fall chinook salmon from Spring Creek National Fish Hatchery and potential contributions by fishery type and brood (1961-64), 1963-69. Estimated catch of marked fish by year Catch Potential contribution Brood year and fishery type 1963 1964 1965 1966 1967 1968 1969 (in thousands) 1961: Marine sport 156 488 129 0 — — — 773 18.9 Marine commercial 4 2.031 269 5 — — — 2.309 56,4 Columbia River sport 0 14 16 0 — — — 30 0.7 Columbia River gillnet 11 388 633 17 — — — 1.049 25.6 Columbia River Indian' 11 147 81 0 — — — 239 58 Total 182 3.068 1.128 22 — — — 4.400 107 4 1962: Marine sport — 34 142 28 0 — — 204 64 Marine commercial — 0 234 135 14 — — 383 12,0 Columbia River sport — 0 0 0 0 — — 0 0,0 Columbia River gillnet — 10 242 88 0 — — 340 106 Columbia River Indian' — 0 40 0 0 — — 40 1,3 Total — 44 658 251 14 — — 967 303 1963: Marine sport — — 120 368 133 0 — 621 255 Manne commercial — — 23 966 282 9 — 1.280 52,6 Columbia River sport — — 0 0 0 0 — 0 0.0 Columbia River gillnet — — 15 151 203 7 — 376 15.4 Columbia River Indian' — — 14 13 95 8 — 130 5,3 Total — — 172 1.498 713 24 — 2,407 988 1964: Marine sport — — — 378 685 87 10 1.160 435 Marine commercial — — — 7 1,634 582 16 2,239 839 Columbia River sport — — — 0 0 0 0 0 0.0 Columbia River gillnet — — — 15 260 351 19 645 242 Columbia River Indian' — — — 0 201 156 5 362 13,6 Total — — — 400 2.780 1.176 50 4,406 165,2 'Setnet and dip net fisheries. fish per 1,000 released and 2.3 fish per pound of fish released. Kalama River Hatcheries, 1961-64 Broods The production at Kalama Falls Salmon Hatch- ery and Lower Kalama Salmon Hatchery was combined for this study. Common and special marks were applied to the production at both facilities. The Ad, RV, and M special mark was allocated to the Kalama facilities. The RM clip was used with the 1961 and 1963 broods, and the LM mark was used with 1962 and 1964 broods. For all brood years, approximately 109^ of both hatch- eries' fall chinook salmon production was marked with a special mark. We estimated 5,096 chinook salmon with spe- cial marks from Kalama River hatcheries were captured in the ocean and Columbia River fisheries from 1963 through 1969 (Table 9). Gen- erally for the four brood years, over half of the Kalama fish were caught in their fourth and fifth years of life. However, the age distribution did vary by brood year. The 1961 and 1964 broods were over 60% 4- and 5-yr-old fish while these two age-groups contributed less than 50% to the 1962- and 1963-brood catches. The Kalama chinook salmon contributed to the Alaska fisheries primarily as 4-yr-olds; and the larger the Cana- dian catch, the larger the Alaskan catch. In 1968 the Canadian catch of Kalama fish was large and no sampling took place in the Alaska fisheries. Thus a significant contribution to Alaska of 1964-brood Kalama fall chinook salmon in 1968 could have been missed. The potential contribution of Kalama River hatcheries fall chinook salmon totaled 172,400 fish for the four brood years (Table 9). The con- tributions ranged from a low of 22,300 fish for the 1962 brood to a high of 56,800 fish for the 1961 brood. The average contribution for all four broods combined was 43,100. This is an average potential contribution to Pacific coast fisheries of 9.6 fish for each 1,000 smolts released and 2.0 fish caught for every pound of fish released. Kalama chinook salmon contributed primarily to British Columbia, Washington, and Columbia River gillnet fisheries (Figure 4). The largest con- tribution was to British Columbia followed by Washington, Columbia River, Oregon, and Alaska, in that order. 193 FISHERY BULLETIN: VOL. 76, NO. 1 Table 9.— Estimated catch of special marked fall chinook salmon from Kalama River hatcheries and potentia 1 contributions by fishery type and brood (1961-64), 1963-69. Estimated catch of marked fish by year Catch Potential contribution Brood year and fishery type 1963 1964 1965 1966 1967 1968 1969 (in thousands) 1961: Marine sport 23 78 103 9 — — — 213 5.6 Marine commercial 0 618 683 106 — — — 1,407 36.7 Columbia River sport 0 0 0 0 — — — 0 0.0 Columbia River gillnet 0 38 402 111 — — — 551 14.4 Columbia River Indian' 0 0 4 0 — — — 4 0.1 Total 23 734 1.192 226 — — — 2.175 56.8 1962: Marine sport — 0 84 11 8 — — 103 3.5 Marine commercial — 0 240 194 23 — — 457 15.5 Columbia River sport — 0 0 16 0 — — 16 0.5 Columbia River gillnet — 6 21 46 10 — — 83 2.8 Columbia River Indian' — 0 0 0 0 — — 0 0.0 Total — 6 345 267 41 — — 659 22.3 1963: Marine sport — — 140 167 66 12 — 385 17.0 Marine commercial — — 0 366 320 53 — 739 32.7 Columbia River sport — — 0 0 0 0 — 0 0.0 Columbia River gillnet — — 7 32 44 50 — 133 5.9 Columbia River Indian' — — 0 0 0 0 — 0 0.0 Total' — — 147 565 430 115 — 1.257 55.6 1964: Marine sport — — — 38 61 40 0 139 5.2 Marine commercial — — — 0 132 533 69 734 27.6 Columbia River sport — — — 0 0 17 0 17 0.6 Columbia River gillnet — — — 3 0 41 68 112 4.2 Columbia River Indian' — — — 0 3 0 0 3 0.1 Total — — — 41 196 631 137 1,005 37,7 'Setnet and dip net fisheries. Southeostern Alosko COMMERCIAL British Columbio COMMERCIAL Woshinqton SPORT COMMERCIAL Oregon SPORT COMME RCIAL California SPORT COMMERCIAL Col umbio River SPORT GILLNET INDIAN 1 % 50% 15% 1 13% 1% 1% 0% 0% 1% 17%, 0% PERCENTAGE OF CATCH FIGURE 4.— Percentage of catch of 1961- to 1964-brood fall Chinook salmon from Kalama River hatcheries taken by area and fishery, 1963-69. Percentages do not add to 100% due to rounding. Elokomin and OxBow Hatcheries, 1961 Brood A portion of the 1961-brood fall chinook salmon productions at Elokomin and OxBow Hatcheries were given special marks. At Elokomin Hatchery, 480,500 or 30% of the production was LV-RM clip- ped. Approximately 450,400 or 10% of OxBow's fall chinook salmon production was marked with a RV-RM clip. These fish contributed to the fisheries from 1963 through 1966. During the 4 yr, 235 Elokomin and 336 OxBow fish with special marks \yere estimated to have been caught (Table 10). Chinook salmon from both hatcheries were taken primarily as 3-yr-olds. A larger portion of Elokomin fish than OxBow fish were taken as 4-yT-olds, and a larger portion of OxBow than Elokomin fish were taken as 2- and 5-yr-olds. Potential contributions were estimated at 2,000 and 8,500 fish for Elokomin and OxBow respectively. The catch per 1,000 fish released at Elokomin Hatchery was 1.3 fish and at OxBow 1.9 fish. The catches per pound of fish released at Elokomin and OxBow Hatcheries were 0.2 and 0.4 fish respectively. About one-half of the catch from the two hatch- 194 WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON Table lO.— Estimated catch of 1961-brood special marked fall chinook salmon and potential contribution from Elokomin and OxBow Hatcheries by fishery type, 1963-66. Estimated catch of marked fish by year Total catch Potential rontrihiitjon Hatchery and fishery type 1963 1964 1965 1966 (in thousands) Elokomin Hatchery Marine sport Marine commercial Columbia River sport Columbia River gillnet Columbia River Indian' 0 0 0 0 0 25 109 0 6 2 31 23 0 30 0 0 9 0 0 0 56 141 0 36 2 05 1,2 0.0 0.3 0.0 Total 0 142 84 9 235 2.0 OxBow Hatchery Marine sport Manne commercial Columbia River sport Columbia River gillnet Columbia River Indian' 18 0 0 0 3 78 107 0 27 0 6 41 17 3 0 16 6 0 14 0 118 154 17 44 3 3.0 39 0.4 11 0 1 Total 21 212 67 36 336 8.5 'Setnet and dip net fisheries. eries occurred in the Washington fisheries. (Fig- ure 5). Nearly 30% of the Elokomin catch was taken in the Washington commercial fisheries. Washington sport fishermen took over one-fourth of the OxBow catch. Fish from Elokomin appear to have a more northerly distribution than those from OxBow. Southeostern Alosko COMMERC lAL British Columbia COMMERCIAL 1 7|0/. Washington SPORT 1 COMMERCIAL Oreqon SPORT ZZi 6% COMMERCIAL 1 7"/. California SPORT 0% COMMERCIAL 3 3% Columbia River SPORT 0% GILLNET 1 15% INDIAN 1 1%. Elokomin hatchery Southeostern Aioska COM ME RC lA L British Columbia COMMERCIAL Washington SPORT COMMERCIAL Oregon SPORT COMMERCIAL Colifornio SPORT COMMERCIAL Columbio River SPORT GILLNET INDIAN PERCENTAGE OXBOW HATCHERY 60 OF CATCH ZZI 27% H 2 4% PERCENTAGE OF CATCH FIGURE 5. — Percentage of 1961-brood fall chinook salmon from Elokomin and OxBow Hatcheries taken by area and fishery, 1963-66. Grays River and Cascade Hatcheries, 1962 Brood The 1962-brood fall chinook salmon at Grays River Hatchery were given a LV-LM special clip and the Cascade fish were RV-LM clipped. Special marks were applied to approximately 18% or 241,500 Grays River and 13% or 541,200 Cascade Hatchery fish. These fish contributed to the fisheries from 1964 through 1967. Approximately equal numbers of Grays River and Cascade fall chinook salmon with special marks were estimated to have been taken in the fisheries (Table 11). Fish from both hatcheries were caught almost exclusively as 3- and 4-yr- olds. Few were taken as 2's and 5's. The potential contributions of Grays River and Cascade were 3,900 and 4,800 fish, respectively. For each 1,000 chinook salmon released at Grays River Hatchery, 2.9 were caught in the fisheries and 0.4 fish were caught per pound of fish released. The contribu- tion from Cascade Hatchery was 1.1 chinook salmon per 1,000 released and 0.2 per pound offish released. The catch distributions of Grays River and Cas- cade Hatcheries were very different (Figure 6); for example, a much greater portion of Cascade's than Grays River's fish were taken in the British Co- lumbia fishery. Most of Grays River's fish (65%) but only 24% of Cascade's fish were taken in the Washington sport fishery. Klickitat and Big Creek Hatcheries, 1963 Brood A LV-RM special mark was applied to 18% or 521,600 1963-brood fall chinook salmon at Klick- 195 FISHERY BULLETIN; VOL. 76, NO. 1 Table ll. — Estimated catch of 1962-brood special marked fall chinook salmon and potential contribution from Grays River and Cascade hatcheries by fishery type, 1964-67. Estimated catch of marked fish by year Total catch Potential contribution Hatchery and fishery type 1964 1965 1966 1967 (in thousands) Grays River Hatchery; Marine sport Marine commercial Columbia River sport Columbia River gillnet Columbia River Indian' 0 3 0 0 0 89 29 0 0 0 50 35 0 5 0 0 4 0 0 0 139 71 0 5 0 25 13 0.0 0.1 0.0 Total 3 118 90 4 215 3.9 Cascade Hatchery: Marine sport Marine commercial Columbia River sport Columbia River commercial Columbia River Indian' 0 0 0 3 0 19 66 0 6 0 28 38 0 24 4 0 3 0 0 0 47 107 0 33 4 1,2 2.7 0.0 08 0.1 Total 3 91 94 3 191 4.8 ^Setnet and dip net fisheries Southeastern Alaska COMMERCIAL British Columbia COMM ERG iflL Washington SPORT COM ME RClflL Oregon SPORT COMMERCIAL Colifornia SPORT COM MERCl AL Columbto River SPORT GILLNET INDIAN Southeastern Alasko COMMERCIAL British Columbia COMME RClAL Woshington SPORT COMME BCiAl Oregon SPORT COMMERCIAL California SPORT COM M ERCIAL Columbia River SPORT GILLNE T INDIAN GRAYS RIVER HATCHERY 0% 3 5% 0% : 2% 0% _!_ PERCENTAGE OF CASCADE HATCHERY 0% 1 2% 3 17% 2 % __l PERCENTAGE OF CATCH Figure 6.— Percentage of catch or 1962-brood fall chinook salmon from Grays River and Cascade Hatcheries taken by area and fishery, 1964-67. Percentages do not add to 100% due to rounding. itat Hatchery. At Big Creek Hatchery nearly 30% or 580,000 1963-brood chinook salmon were given RV-RM special clips. These fish contributed to the fisheries from 1965 through 1968. The estimated catches of chinook salmon with special marks from Klickitat and Big Creek Hatcheries were 1,858 and 914 fish, respectively 196 (Table 12). The Klickitat fish were caught primar- ily as 3- and 4-yr-olds, except in the ocean sport fishery where 2-yr-olds were predominant. In the marine commercial and Columbia River fisheries, the predominant age class was 3-yr-olds. Nearly 60% of Big Creek's special marked fish were caught in their third year of life, and about one- third were taken as 4-yr-olds. Klickitat and Big Creek Hatcheries' potential contributions to the fisheries were 42,500 and 12,900 fish, respectively. From Klickitat the con- tribution was 14.7 fish per 1,000 released and 2.2 fish for each pound of fish released. The contribu- tion per 1,000 chinook salmon released at Big Creek was 6.5 fish and 0.7 fish for each pound of fish released. Distribution of both facilities' catches can be compared by examination of Figure 7. Thirty-nine percent of Klickitat's fish were taken in the British Columbia commercial fisheries compared with 16% for Big Creek, suggesting a more north- erly distribution for Klickitat fish. Although Big Creek fish pass through only a small portion of the Columbia River commercial fishery, the portion taken in this fishery is larger (19%) than the Klickitat portion (10%). Over half of Big Creek's estimated catch was taken in the Washington marine fisheries. Bonneville and Little White Salmon Hatcheries, 1964 Brood About 10% (957,100) of the 1964-brood Bon- neville Hatchery fall chinook salmon were marked with a LV-LM clip. The RV-LM mark was applied to about 10% (797,300) of the Little White WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON Table 12. — Estimated catch of 1963-brood special marked fall chinook salmon and potential contribution from Klickitat and Big Creek hatcheries by fishery type, 1965-68. Estimated catch of marked fish by year Total Potential Hatchery and fishery type 1965 1966 1967 1968 catch (in thousands) Klickitat Hatchery: Marine sport 161 146 81 0 388 8.9 Marine commercial 3 633 617 32 1.285 29.4 Columbia River sport 0 0 0 0 0 0.0 Columbia River commercial 0 72 45 0 117 2.7 Columbia River Indian' 0 47 21 0 68 1.5 Total 164 898 764 32 1,858 425 Big Creek Hatchery: Marine sport 70 209 73 0 352 5.0 Marine commercial 0 240 144 7 391 5.5 Columbia River sport 0 0 0 0 0 0.0 Columbia River commercial 0 93 78 0 171 2.4 Columbia River Indian' 0 0 0 0 0 0.0 Total 70 542 295 7 914 12.9 'Setnet and dip net fisheries. Southeastern Alosko COMMERC tAL British Columbio COMMERCIAL Woshington SPORT COMMERCIAL Oregon SPORT COMMERCIAL California SPORT COMMERC lAL Columbia River SPORT GILLNET INDIAN Southeostern Alosko COMMERC tAL British Columbia COMME RC I AL Washington SPORT COMM ERCIAL Oregon SPORT COMMERCIAL Colifornio SPORT COMMERCIAL Columbio River SPORT GILLNET INDIAN KLICKITAT HATCHERY :j 39% n 14% 0% : 4% n 4% 40 60 PERCENTAGE OF CATCH 100% PERCENTAGE OF CATCH BIG CREEK HATCHERY ^ 0% 1 ' 1% H 3% ZJ 5% 0% 1 1% 0% 1 1=1% 0% °1 1 1 1 1 1 1 1 1 1 Figure 7. — Percentage of catch of 1963-brood fall chinook salm- on from Klickitat and Big Creek Hatcheries taken by area and fishery, 1965-68. Percentages do not add to 100% due to round- ing. Salmon National Fish Hatchery 1964-brood fish. Both groups contributed to the fisheries from 1966 through 1969. The estimated catches of special marked fish from Bonneville and Little White Salmon Hatch- eries were 762 and 303 fish respectively. Sig- nificant numbers of Bonneville special mark chinook salmon were caught in the ocean fisheries as 2-, 3-, and 4-yr-olds, while the Little White fish contributed as 3's and 4's (Table 13). The largest numbers of both hatcheries' fish were taken in the ocean commercial fisheries. The potential contributions for Bonneville and Little White were 27,100 and 11,000 fish, respec- tively. Bonneville produced 2.7 fish per 1,000 or 0.4 fish per pound of fish released. Little White produced 1.3 fish per 1,000 or 0.2 fish per pound of fish released. The distribution of the Bonneville Hatchery catch was more southerly than that of Little White Salmon Hatchery (Figure 8). Nearly 50% of the catch from both facilities occurred in the Washington fisheries. The British Columbia fisheries took most of the remaining Little White catch (41%). Common Mark Catch and Potential Contribution All Study Facilities Combined, 1961-64 Broods An Ad-M common mark was applied to a portion of the 1961-64-brood fall chinook salmon produc- tion at all 13 study facilities. The RM was clipped from the 1961- and 1963-brood fish, and the LM was clipped from the 1962- and 1964-brood chinook salmon. Common marks were applied to 21,320,000 (approximately 10%) of the 213,014,000 fall chinook salmon released over the four brood years from the 13 study facilities. We estimated 65,620 common marked fish were caught from 1963 through 1969 (Table 14). On the average over the four broods 76% of the common marked fish were taken in the ocean, with 56% caught in the ocean commercial fisheries. In the 197 Table 13. — Estimated catch of 1964-brood special marked fall chinook contribution from Bonneville and Little White Salmon National Fish type, 1966-69. FISHERY BULLETIN: VOL. 76, NO. 1 salmon and potential hatcheries by fishery Estimated catch of marked flsti by year Hatchery and fishery type 1966 1967 1968 1969 Bonneville Hatchery: Marine sport 99 70 95 f^^arlne commercial 62 230 172 Columbia River sport 0 0 0 Columbia River commercial 0 0 17 Columbia River Indian' 4 0 0 Total 165 300 284 Little White Salmon Hatchery: t^arlne sport 0 40 37 Marine commercial 4 84 125 Columbia River sport 0 0 0 Columbia River commerical 0 5 8 Columbia River Indian' 0 0 0 Total 4 129 170 Total catch Potential contribution (in thousands) 0 264 9.4 8 472 16.8 0 0 0.0 5 22 0.8 0 4 0.1 13 762 27.1 0 77 28 0 213 7,7 0 0 0,0 0 13 0.5 0 0 0,0 0 303 11,0 'Setnet and dip net fisheries. Southeastern Aloska COMMERCIAL British Columbia COMMERCIAL 1 Washington SPORT 1 1 10"/. Oregon SPORT 1 fir. COMMERCIAL 1 io"/„ California SPORT 0% COMMERCIAL Z) "% Columbia River SPORT GILLNET INDIAN Southeastern AtasKo COMMERCIAL British Columbia COMM ERC I AL Washington SPORT COMMERCIAL Oregon SPORT COMMERCIAL Colifornio SPORT COMMERCIAL Columbio River SPORT GILLNET INDIAN BONNEVILLE HATCHERY Zl 29% 0% n 3% % \ 10 PERCENTAGE 60 OF CATCH LITTLE WHITE SALMON HATCHERY H 11% 1 25% 2 2 % 0% P 3% 0% n 1 0% -L. 10 60 PERCENTAGE OF CATCH FIGURE 8.— Percentage of catch of 1964-brood fall chinook salm- on from Bonneville and Little White Salmon Hatcheries taken by area and fishery, 1966-69. ocean fisheries, the 3-yr-old exceeded the 4-yr-old catch. However, in the river the 4-yr-old catch was larger than the 3-yr-old. The Columbia River fall chinook salmon sport fishery was small and few marked fish were observed. The potential contribution for the four broods combined was 1,433,300 fall chinook salmon. The 198 contribution ranged from a low of 165,200 fish for the 1962 brood to a high of 602,200 for the 1963 brood. The contribution figures in Table 14 include fish with common and special marks as well as unmarked fish from the 13 study facilities. The average catch to release ratio was 6.7 fish per 1,000 released, with ratios of 6.7, 3.1, 10.0, and 6.5 for the 1961-64 broods respectively. The average catch per pound released was 1.2 fish with ratios by brood of 1.4, 0.6, 1.7, and 0.9 fish per pound released. The catch was distributed primarily among the British Columbia commercial (34%), the Washington marine sport and commercial (38%), and the Columbia River gillnet (19%) fisheries (Figure 9). CATCH TO ESCAPEMENT AND SURVIVAL Returns to Columbia River hatcheries, both study and nonstudy, and to streams adjacent to these hatcheries were examined for marked chinook salmon (see Enumeration of Returns). Mark return data were used to estimate catch to escapement ratios and total survival percentages for each special mark hatchery and all study hatcheries combined (Table 15). Only marked catches and escapements were used to develop the estimates to eliminate possible inflation of es- capement values due to unmarked wild fish in hatchery returns. Survival estimates were calcu- lated by dividing the potential marked catches and escapements by the marked releases. Poten- tial marked catches and escapements are those that would be expected if marking did not cause post release mortalities. Potential marks were es- WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON Table 14.— Estimated cat ches of common market fall chi nooksa Imon and potei itial contrib ution from all Columbia River study hatcheries by fishery type and brood. 1963-69. Potential Estimated catch of common marked fish by yea Total Catch contribution' BrcxxJ year and fishery type 1963 1964 1965 1966 1967 1968 1969 (in thousands) 1961: Marine sport 576 2,091 613 82 — — — 3,362 548 Marine commercial 3 8,778 3,034 366 — — — 12.181 198 1 Columbia River sport 0 21 0 0 — — — 21 0-4 Columbia River gillnet 98 1,651 3,407 197 — — — 5,353 868 Columbia River Indian^ 50 852 411 7 — — — 1,320 21 0 Total 727 13,393 7,465 652 — — — 22.237 361 1 1962 Marine sport — 204 773 166 27 — — 1,170 25 1 Marine commercial — 79 2,981 1,490 108 — — 4,658 970 Columbia River sport — 12 8 0 0 — — 20 0.5 Columbia River gillnet — 31 879 680 21 — — 1.611 33.9 Columbia River Indian^ — 11 392 21 3 — — 427 8.7 Total — 337 5,033 2,357 159 — — 7,886 165.2 1963: Marine sport — — 1,304 3,140 594 56 — 5,094 139.4 Marine commercial — — 71 9.016 3.317 284 — 12,688 344 5 Columbia River sport — — 0 0 0 0 — 0 00 Columbia River gillnet — — 88 1,194 2,168 315 — 3,765 101 0 Columbia River Indian^ — — 38 103 453 42 — 636 173 Total — — 1,501 13,453 6,532 697 — 22,183 602 2 1964: Marine sport — — — 797 1,966 466 4 3.233 74,2 Marine commercial — — — 53 4,757 2,492 108 7,410 169.8 Columbia River sport — — — 0 0 0 0 0 0.1 Columbia River gillnet — — — 27 692 1,034 188 1,941 43.9 Columbia River Indian^ — — — 1 405 307 17 730 16.8 Total — — — 878 7,820 4,299 317 13,314 304.8 'Special marks included. ^Setnet and dip net fisheries. Southeastern Alaska COMMERCIAL 3 2% British Columbia COMMERCIAL Washington SPORT 18 8% COMMERCIAL 19 3% Oregon SPORT 17% COMMERCIAL n 2 9% California SPORT 3% COMMERC lAL 0.4% Columbia River SPORT 0 1% GILLNET 18 5% INDIAN 4 5 /<, 33 7% 20 40 60 PERCENTAGE OF CATCH Figure 9.— Percentage of catch of 1961- to 1964-brood fall Chinook salmon from 13 Columbia River study facilities taken by area and fishery, 1963-69. Percentages do not add to 100% due to rounding. timated by dividing the mark recoveries by the appropriate special or common marked to un- marked relative survivals (see Contribution of Hatchery Fish). Catch to escapement and survival estimates are of limited value for several reasons. First, adjacent tributary streams were surveyed during only three of the seven return years of the study (1964- 66). Survey data are unavailable for at least one return year for each brood. Thus, all catch to es- capement ratios are probably overestimated and survivals underestimated. Second, only a portion of the fish returning to the streams could be examined for marks. Total mark recoveries had to be estimated from the survey samples. Third, in some cases fish were delayed in entering adult holding facilities and may have strayed to other areas. Thus, some marked hatchery fish may not have been counted. Fourth, use of average relative survivals limited the accuracy of potential mark catches and returns and thus the total survival percentages. Relative survivals for individual hatcheries could have differed greatly from the averages. Catch to escapement ratios and total survivals are needed to develop values for fisheries compen- sation and enhancement projects related to water use projects on the Columbia River system. Thus, 199 FISHERY BULLETIN; VOL. 76, NO. 1 TabLK 15. — Marked catches and escapements, catch to escapement ratios, and total survivals for fish from each special mark hatchery and all study facilities combined, 1961-64 broods. Catch Potential Marked Marked to marked catch Marked Total Hatchery Brood catch escapement escapement and escapement' releases survival Spring Creek 1961 4.400 613 7.2 12,691 1.133.019 0.011 1962 967 92 105 3.416 866.892 0.004 1963 2.407 374 64 11,492 751,243 0.015 1964 4.406 228 19,3 15,924 600,953 0026 Kalama River 1961 2.175 238 9,1 6,109 475.964 0013 1962 659 38 17.3 2.248 437,669 0005 1963 1.257 106 119 5,632 456.158 0012 1964 1.005 41 245 3.595 319.412 0 011 Elokomin 1961 235 33 7.1 678 480,533 0001 OxBow 1961 336 99 3.4 1.101 450,446 0002 Grays River 1962 215 5 43.0 710 241.494 0003 Cascade 1962 191 6 31.9 635 541.158 0001 Klickitat 1963 1.858 129 14,4 8.210 521.610 0016 Big Creek 1963 914 380 24 5.347 579.967 0 009 Bonneville 1964 762 27 28 2 1 2,711 957.110 0 003 Little White Salmon 1964 303 37 82 1 1,168 797.345 0001 All study facilities^ 1961 22.237 3.399 6.5 1 42,164 5,446.439 0 008 1962 7,886 675 11.7 1 17,948 5.249.079 0003 1963 22.183 2.737 8 1 1 66.989 5,986,464 0,011 1964 13.314 856 156 1 31,629 4,638,237 0 007 'Assuming no mortality due to marking ^Includes common marks only. despite the limitations, we have included the val- ues in this report. Catch to escapement ratios for special mark hatcheries (Table 15) ranged from 2.4 to 1 (Big Creek, 1963 brood) to 43 to 1 (Grays River, 1962 brood). Average catch to escapements for Spring Creek and Kalama River hatcheries were 9.3 to 1 and 12.0 to 1 respectively. The catch to escape- ment ratios for all hatcheries combined, common marks only, show much less yearly variation than those for the special mark hatcheries. The average catch to escapement, all hatcheries and broods combined, was 8.6 to 1. Only common marks were combined for all hatcheries because these marks show only the variations among broods, not those among marks. Total survivals ranged from 0.1% (Elokomin, 1961 brood; Cascade, 1962 brood; Little White Salmon, 1964 brood) to 2.6% (Spring Creek, 1964 brood). Average survivals for Spring Creek and Kalama River hatcheries were 1.3 and 1.0% re- spectively. For all hatcheries combined, the aver- age survival was 0.7% . Examination of Table 15 does not reveal any relationship between catch to escapement ratios and survivals. For example, at Spring Creek the 1964 brood had the highest catch to escapement ratio and percent survival. At Kalama River hatcheries, the 1 964 brood had the highest catch to escapement ratio and the second highest survival value. The 1961 brood had the lowest catch to escapement and highest survival. For all study facilities, the 1964 brood had the highest catch to 200 escapement ratio, and the 1963 brood had the highest total survival. The 1964 and 1961 broods had nearly equal survivals, but markedly differ- ent catch to escapements. The major reason for high 1964 brood catch to escapement ratios is the absence of adjacent stream surveys during three of the four return years for this brood. ECONOMIC EVALUATION A major purpose of this paper is to develop bene- fit to cost ratios for each of the special mark hatch- eries and for each brood of the combined study facilities. To develop these ratios, the cost of rear- ing the four broods of chinook salmon and their potential value to the fisheries had to be esti- mated. The development of benefit to cost ratios is explained in detail by Worlund et al. ( 1969) and Wahle et al. ( 1974), but certain modifications will be discussed here briefly. The values and benefit to cost ratios are higher in this report than those reported in our previous reports for five reasons: 1 ) the interest rate applied to capital costs is lower in this report (Wahle et al. 1974), 2) the sport value used is higher (see Value of Hatchery Contribution), 3) a lower marked fish relative survival figure was used for the 1961- brood (see Contribution of Hatchery Fish), 4) mis- identified and partial marks were included in this report (see Contribution of Hatchery Fish), and 5) the potential catch contribution figures were used in this report rather than estimated catches (see Contribution of Hatchery Fish). WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON Cost Accounting and Value Estimation Costs in Table 16 include capital and operation and maintenance costs applicable to the rearing of fall chinook salmon at each study facility. Capital costs for each facility were amortized over a 30-yr period froml940 to 1970 and divided among the species reared at the facilities. Capital costs applied to fall chinook salmon at all study facilities combined were $193,867, $169,616, $193,102, and $186,437 for the 1961-64 broods respectively. Operation and maintenance costs were divided into two categories at each facility: fish food and drugs, and other operational costs. Operational costs other than food and drugs include costs for labor, personal services, travel, transportation of items, communication services, equipment, supplies and materials, and administration. Total operational and maintenance costs for the 1961-64 broods were $554,171, $489,947, $534,146, and $538,418 respectively. Estimation of values is described under Value of Hatchery Contribution. Basically, the weights of commercial catches in each fishery were multi- plied by the appropriate ex-vessel prices. The numbers of sport caught fish in all fisheries were multiplied by $18.35. Valuation of the Potential Contributions The value of the potential contribution to the fisheries of fall chinook salmon from Spring Creek National Fish Hatchery and Big White Salmon Pond were combined (Table 16). This was done because Spring Creek Hatchery personnel oper- ated the Big White Pond, and Spring Creek fall chinook salmon stock was reared in the pond. Thus available Spring Creek operation and mainte- nance, and capital costs include the Big White facility. Values of Big White contributions were estimated using the ratio: Table 16. — Values of the potential contributions, costs of rear- ing, and benefit (B) to cost (C) ratios for fish from each special mark hatchery and all study facilities combined, 1961-64 broods.' Spring Creek value Big White value Spring Creek releases Big White releases' For example, the 1961-brood Spring Creek value was $797,300. Releases were 10,925,933 and 3,545,865 1961-brood chinook salmon for Spring Creek and Big White respectively (Worlund et al. 1969). Thus, the Big White Salmon Pond value was estimated at $258,700. Values for the other broods were calculated in the same manner. Hatchery Brood Value ($) Cost ($) B/C ratio Spring Creek^ 1961 1 ,056,000 99,900 10,5/1 1962 373,900 84,800 44/1 1963 1,131,400 99,200 11,4/1 1964 1,917,300 112,000 17.1/1 Kalama River 1961 481,900 100,700 4,8/1 1962 199.800 104,700 1.9/1 1963 582.000 97,600 6.0/1 1964 392.700 110,700 3.5/1 Elokomin 1961 16,900 53,400 0.3/1 OxBow 1961 93,100 42 100 2,2/1 Grays River 1962 56,100 38,800 1.4/1 Cascade 1962 44,800 57,800 0.8/1 Klickitat 1963 373,200 32,800 11.4/1 Big Creek 1963 141,400 33.700 4.2/1 Bonneville 1964 279,300 81,000 3.4/1 Little White Salmon 1964 108,200 99,400 1.1/1 All study facilities 1961 2,738,800 748,000 3.7/1 1962 1,306,100 659,600 2.0/1 1963 5,224,100 727,200 7.2/1 1964 2,758,000 724,900 3.8/1 'Values and costs rounded to the nearest $100 ^Includes Big White Salmon Pond values and costs Combined Spring Creek and Big White values ranged from $373,900 (1962 brood) to $1,917,300 (1964 brood). The average value was $1,119,600. The costs averaged approximately $100,000 per brood. Benefit to cost ratios ranged from 4.4 to 1 to 17.1 to 1 and averaged 11.2 to 1. The 1961 brood had the largest contribution to the fisheries, yet the 1963 and 1964 broods had higher values. The reason for this is the increase in prices paid for troll caught fish from 1963 to 1969. Values for the Kalama River hatcheries ranged from $199,800 (1962 brood) to $582,000 (1963 brood). The 1963 brood value was larger than the 1961 brood despite a smaller contribution for the 1963 brood. Again this was due to higher prices paid for troll chinook salmon in the later years of the study and also a larger 1963 than 1961 brood contribution to Washington and Oregon ocean sport fisheries. The average benefit over the four broods was $414,100. The average cost of rearing was $103,400 per brood. Benefit to cost ratios var- ied from 1.9 to 1 to 6.0 to 1 and averaged 4.0 to 1. The value of Elokomin Hatchery's potential con- tribution was $16,900 for the 1961 brood and the cost of rearing was $53,400. The benefit to cost ratio was then 0.3 to 1. OxBow's 1961 brood value was $93,100 and costs were $42,100 for a benefit to cost ratio of 2.2 to 1. The ratio was much higher for OxBow because OxBow chinook salmon contri- buted more heavily to ocean sport fisheries than Elokomin fish. Contributions of 1962-brood Grays River and Cascade Hatchery fish were valued at $56,100 and 201 FISHERY BULLETIN: VOL. 76, NO. 1 $44,800 respectively. The Grays River value is higher because of a larger contribution to the ocean sport fishery. The costs of rearing were $38,800 at Grays River and $57,800 at Cascade. The benefit to cost ratios were 1.4 to 1 and 0.8 to 1 for Grays River and Cascade respectively. Klickitat and Big Creek Hatcheries' potential contributions of 1963-brood chinook salmon were valued at $373,200 and $141,400 respectively. The costs of rearing were $32,800 and $33,700 for the two hatcheries respectively. Benefit to cost ratios were 11.4 to 1 for Klickitat Hatchery and 4.2 to 1 for Big Creek Hatchery. The values of the 1964 brood potential contribu- tions were estimated at $279,300 for Bonneville Hatchery and $108,200 for Little White Salmon National Fish Hatchery. Rearing costs were $81,000 and $99,400 for the respective facilities. The benefit to cost ratios were 3.4 to 1 and 1.1 to 1 for Bonneville and Little White respectively. Values of potential contributions for all study facilities combined ranged from $1,306,100 for the 1962 brood to $5,224,100 for the 1963 brood and averaged $3,006,800. Costs ranged from $659,600 to $748,000 for the 1962 and 1961 broods respec- tively. The average rearing costs were $714,900 per brood. Benefit to cost ratios ranged from 2.0 to 1 (1962 brood) to 7.2 to 1 (1963 brood) and aver- aged 4.2 to 1. During the later years of the study, fall chinook salmon carcasses from study hatcheries were sold to commercial processors or donated to various institutions and groups. The value of these carcas- ses was determined from the average price paid by commercial processors. The estimated value was $31,467 for the 1963 brood (Arp et al. see footnote 4) and $42,000 for the 1964 brood (Wahle et al. see footnote 5). Thus the total value of 1963- and 1964-brood study hatchery fall chinook salmon was $5,255,600 and $2,800,000 respectively. DISCUSSION Brood Year Comparison The 1963-brood Columbia River hatchery fall chinook salmon had the best potential contribu- tion and value to the Pacific coast fisheries (Tables 16, 17). The 1963 brood had a potential contribu- tion of 602,900 fish or 10 fall chinook salmon caught for every 1,000 releases and 1.7 fish per pound released. The 1963 brood contribution and catch to release ratios were followed in order by the 1961, 1964, and 1962 broods. The benefit to cost ratios followed a similar pattern, with the best ratio (7.2 to 1) for the 1963 brood followed by the 1964, 1961, and 1962 broods. The 1964 brood had a lower potential contribution than the 1961 brood, but a higher benefit to cost ratio because of higher prices paid for salmon when the 1964 brood was in the fisheries. Also total rearing costs for the 1964 brood were lower than the 1961 brood because fewer fish were raised. The ocean distribution of the fall chinook salm- on for all hatcheries combined was similar for all Table 17. — Potential contributions, numbers of smolts released, pounds of smelts released, contribution in fish caught per 1,000 released, and contribution per pound released for each special mark hatchery and all study facilities combined, 1961-64 broods. Contribution Releases (in thousands) Contr ibution Per 1,000 Per pound Hatchery Brood (in thousands) Number Pounds released released Spring Creek 1961 107.4 10,9259 48.0 9.8 22 1962 30.3 8,408.3 48.9 36 0.6 1963 988 7,467.6 34.7 13.2 2.8 1964 165.2 6,554.5 42.4 252 3.9 Kalama River 1961 56.8 4,9068 16.8 11 6 3.4 1962 22.3 4,599.3 21.0 48 11 1963 55.6 4,883.9 26.8 11.4 2 1 1964 37.7 3,4966 21.0 10.8 1.8 Elokomin 1961 20 1,575.0 8.1 1.3 0.2 OxBow 1961 8.5 4,5500 21 0 19 0.4 Grays River 1962 3.9 1,359 8 96 29 0.4 Cascade 1962 4.8 4,217.9 21.9 1.1 02 Klickitat 1963 42.5 2,888.2 19.5 14.7 2.2 Big Creek 1963 129 1,985 8 194 6,5 0.7 Bonneville 1964 271 9,887.6 62.1 2.7 0.4 Little White Salmon 1964 110 8,365.6 47.3 1.3 02 All study facilities 1961 361.1 536532 2509 6.7 1.4 1962 165-2 52,470-0 2785 3.1 0.6 1963 602.2 60,1121 350.7 10.0 1.7 1964 304.8 46,778.6 322 2 6,5 0.9 202 WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON four brood years (Table 18). Washington marine fisheries took the largest catch of Columbia River study hatchery fall chinook salmon followed by British Columbia, Columbia River, and Oregon fisheries. The combined Washington commercial and sport marine catches from the 1961-63 broods were equal to or greater than the British Colum- bia commercial catch and were between 33 and 39% of the catch of Columbia River study hatchery fall chinook salmon. For the 1964 brood the Washington catch was over IVz times as large as the British Columbia catch and approached one- half of the total 1964-brood study hatchery fall chinook salmon catch. The British Columbia commercial catch ranged from 27 to 39% of the study hatchery fall chinook salmon catch. The combined Columbia River sport and commercial catch by brood ranged from 20 to 30% of the study hatchery catch. The Oregon ocean portion of the catch ranged from 1 to 9%. The California portion was 1% or less. Less than 0.5% of Columbia River study hatchery fish were taken in the Alaska fisheries, but these fisheries were incompletely sampled. Kalama River and Spring Creek hatcheries, the only hatcheries with special marks all four brood years, did not follow the combined hatchery pat- tern. For the Kalama River hatcheries the 1961 brood had the largest contribution and best catch to release ratio, followed in order by the 1963, 1964, and 1962 broods (Table 17). The benefit to cost ratios, however, did not follow this pattern primarily because of higher prices paid for salmon in the later years of the study. The 1963 brood had the best benefit to cost ratio, followed by the 1961, 1964, and 1962 broods respectively (Table 16). Distribution of the Kalama fish was more northerly than the combined distribution for all study hatcheries (Table 18). About 1% of the Kalama fish were caught in the Alaska fisheries during the years when these fisheries were sam- pled. The British Columbia portion of the Kalama contribution ranged from 42 to 60%. The Washington marine fisheries took from 23 to 43% of the Kalama fall chinook salmon. When the Washington catch was at its highest (1963 brood), the British Columbia catch was at its lowest. The Columbia River sport and commercial catches of Kalama fish ranged from 11 to 26%. In general, the larger the percentage taken by the British Columbia and Washington fisheries, the smaller the percentage of Kalama fish taken by the Co- lumbia River fisheries. The Oregon ocean fisheries took 1 to 3% of the Kalama chinook salmon and the California fisheries took very few Kalama fish. The brood year comparison of Spring Creek con- tribution also differed from the comparison of all hatcheries combined. The 1964 brood showed the best potential contribution followed by the 1961, 1963, and 1962 broods (Table 17). The catch to release and benefit to cost ratios were best for the 1964 brood followed by the 1963, 1961, and 1962 broods (Table 16). The ocean distribution of the Spring Creek Table 18. — Percentage of catch of Columbia River study hatchery fall chinook salmon taken by each fishery, 1961-64 broods.^ Fishery British Columbia Hatchery Brood Alaska Columbia Washington Oregon California River Spring Creek 1961 0 23 43 4 1 30 1962 0 18 41 2 0 39 1963 0 34 41 3 {') 21 1964 0 24 45 7 (') 23 Kalama River 1961 2 48 23 1 e) 26 1962 1 58 24 2 0 15 1963 2 42 43 3 0 11 1964 0 60 23 3 1 13 Elokomin 1961 0 21 47 13 3 16 OxBow 1961 0 13 51 15 2 19 Grays River 1962 0 12 74 5 / 2 Cascade 1962 0 43 36 2 1 19 Klickitat 1963 0 39 32 15 4 10 Big Creek 1963 0 16 57 8 1 19 Bonneville 1964 0 29 46 17 4 4 Little White 1964 0 41 47 3 5 4 All study facilities 1961 (') 33 33 3 e) 30 1962 {') 39 33 1 1 26 1963 {') 36 39 5 {') 20 1964 (') 27 44 9 {') 20 'Percentages may not add to 100 due to rounding 2 Less than 0.5%. 203 FISHERY BULLETIN: VOL. 76, NO. 1 Hatchery fall chinook salmon was more southerly than those of the Kalama or combined study hatcheries (Table 18). The British Columbia catch ranged from 18 to 34% of the total Spring Creek contribution. The Washington marine fisheries took 41 to 45%. The catch of Spring Creek fish in the Oregon ocean fisheries ranged from 2 to 7%. The maximum California take of these fish was just over 0.5%. The Columbia River catch of Spring Creek fish (21 to 39% ) was higher than the percent catch of Kalama or all hatcheries com- bined. This is to be expected since the Spring Creek chinook salmon are exposed to more river fisheries because of the upriver location of the hatchery. Hatchery Comparison A hatchery comparison is made difficult by the great differences in contribution between brood years. Thus these comparisons are not a reflection of the value of any particular hatchery as a fall chinook salmon station nor are they a criticism of rearing techniques at any of the hatcheries. In general, the best catch to release and benefit to cost ratios occurred for the 1963-brood special marked hatchery fish (Tables 16, 17). The poorest ratios generally occurred for the 1962-brood spe- cial mark hatchery chinook salmon. This follows the pattern of the common marked fish. The 1964-brood Spring Creek fall chinook salmon had the best catch to release and benefit to cost ratios of 10 special mark hatcheries. The Cascade Hatch- ery 1962-brood chinook salmon had the poorest catch to release ratio, and the 1961-brood Eloko- min Hatchery fish had the poorest benefit to cost ratio. The general distribution of fall chinook salmon from special mark hatcheries was similar in that a majority of the fall chinook salmon were caught north of the Columbia River mouth in the Washington and British Columbia ocean fisheries (Table 18). However, the percent catch of each hatchery's fish varied greatly within each fishery. The percent catch ranged from 12% (Grays River 1962-brood falls) to 60% (Kalama 1964 brood) in the British Columbia fisheries. Percent catch by Washington ocean fisheries ranged from 23% for 1961- and 1964-brood Kalama River fish to 74% for 1962-brood Grays River chinook salmon. Washington fisheries took the largest portion of the catch for all but Kalama, Cascade, and Klick- itat hatcheries. The British Columbia exceeded the Washington catch for these facilities except for the 1963-brood Kalama fish where the Washing- ton catch was slightly higher. As the percentage of the catch taken by the British Columbia fisheries increased, the percentage taken by other fisheries (particularly Washington) naturally decreased. Percent catches in the Oregon fisheries ranged from 1 to n% for 1961-brood Kalama and 1964- brood Bonneville fish respectively. In the Califor- nia fisheries, percentages ranged from 0% for Spring Creek and Kalama fish to 7% for Grays River fish. Columbia River catch portions ranged from 2 to 39% for the Grays River and Spring Creek 1962-brood fish respectively. COLUMBIA RIVER HATCHERY CONTRIBUTION TO PACIFIC COAST FISHERIES This report covers the contributions of 13 fall chinook salmon study facilities on the Columbia River for brood years 1961 through 1964. These broods were also released from other hatcheries on the Columbia system. From 1962 through 1965, seven nonparticipating facilities released fall chinook salmon during one or more years (Table 19). Experimental releases made from three facilities were not included. A total of 26 million 1961-64-brood fall chinook salmon migrants were released from nonstudy hatcheries. We have as- sumed nonstudy hatchery releases had the same distribution and contribution as the study facility average. In this way, we have incorporated the catches of nonstudy hatchery fall chinook salmon into those from study hatcheries to estimate the total contribution and value of Columbia River 1961-64-brood hatchery fall chinook salmon. From 1963 through 1969 the estimated total catch in the fisheries sampled of the 1961-64-brood chinook salmon, wild and hatchery, was 9,894,200 (Table 20). Marine sport and commercial catches include three races of chinook salmon, i.e., spring, summer, and fall. Columbia River catches include Table 19.— Releases of 1961- to 1964-brood migrant fall chinook salmon from Columbia River nonstudy hatcheries. Hatche,-y 1961 brood 1 962 brood 1963 brood 1 964 brood Abernathy 1,077,519 1,806,164 836,375 719,228 Lewis River 477.462 0 275,965 0 Speelyai 456.550 0 0 0 Toutle 992,559 3.075,052 2,580,198 5,730,659 Klaskanine 568,032 137,132 252,216 191,636 Sandy 231,999 1 44,848 969,154 1,000,418 Eagle Creek 0 2,435,531 1,427,326 1,054,720 Total 3.804,121 7,598.727 6,341,234 8,696,661 204 WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON TABLE 20. — Percent contribution of Columbia River hatchery fall chinook salmon in the Pacific coast fisheries sampled for marks, 1961-64 broods. Estimated catch of Estimated total Fishery hatchery fall catch of Percent hatchery Region type Chinook salmon' Chinook salmon^ contribution Marine fisheries: Southeastern Alaska Commercial 26 754.3 0.3 British Columbia Commercial 496 1 4.048.4 12.3 Washington Sport 2763 897.4 30.8 Commercial 2860 576.5 49.6 Oregon Sport 246 97.6 25.2 Commercial 430 3026 14.2 California Sport 0.4 248 1 0.2 Commercial 56 2.171 0 0.3 Freshwater fishenes; Columbia River Sport 09 27.4 3.3 Glllnet 2736 658 3 41.6 Indian^ 58.5 112.6 52.0 Total All fisheries 1,467.6 9,8942 14.8 'Includes study and nonstudy Columbia River hatcheries which reared 1961- to 1964-brood fall chinook salmon. ^Marine catches include all races of chinook salmon; Columbia River catches include only fall chinook salmon 'Setnet and dip net fisheries. only fall chinook salmon. We estimated 1,467,600 fish or 14.8'7c were Columbia River hatchery fall chinook salmon. The proportions of fall chinook salmon in each of the fisheries sampled that were of Columbia River hatchery origin are presented in Figure 10. The percentages are averages ob- tained by summing the 1961-64-brood fall chinook salmon catches from Columbia River hatcheries and dividing by the total 1961-64-brood chinook salmon catches in the Pacific coast fisheries sam- pled for marks (Table 20). The importance of Columbia River hatchery fall chinook salmon to the Pacific coast fisheries is readily evident in Figure 10. Columbia River hatchery fall chinook salmon compose nearly one-half of the Washington commercial and nearly one-third of the Washington marine sport chinook salmon catches. The Oregon ocean sport catch of chinook salmon is one-fourth Columbia River hatchery fall chinook salmon. The low sampling percentage ( averaging < 5% ) may be the reason for the apparent lack of hatchery contribu- tion to the Columbia River sport fall chinook salmon fishery. The contributions to the fisheries from the seven Columbia River nonstudy hatcheries were 24,100, 22,700, 61,800, and 53,500 fall chinook salmon for the 1961-64 broods respectively. Values of the con- tributions were calculated using the ratio: Study hatchery value Nonstudy hatchery value Study hatchery contribution Nonstudy hatchery contribution' The values calculated for the nonstudy hatchery chinook salmon were $182,900, $179,100, $538,700, and $484,600 for the four broods respec- tively. The total values for all 1961-64-brood Co- lumbia River hatcheries by brood were $2,921,700, $1,485,200, $5,794,300, and $3,284,600 respectively. Southeastern Alaska COMMERCIAL British Columbia POMMERCI AL Washington SPORT COMMERCIAL Oregon SPORT COMMERCIAL California SPORT COMMERC lAL Co umbia River SPORT 6ILLNET INDIAN 0 3% 0 3% I 3 3% 5 8.4% 4 8.0%, Y////\ Fall Chinook contribution 20 40 60 80 PERCENTAGE OF CATCH 100 from Columbia River tiatctiery Chinook contribution trom other sources Figure lO. — Percentage contribution of 1961- to 1964-brood Columbia River hatchery fall chinook salmon to the total chinook salmon catch in each Pacific coast fishery, 1963-69. Marine fisheries include all races of chinook salmon; Columbia River fisheries include only fall chinook salmon. SUMMARY In 1962 a marking experiment was initiated to determine the bioeconomic contribution of Co- lumbia River hatchery fall chinook salmon. From 205 FISHERY BULLETIN: VOL. 76, NO 1 1962 through 1965, 30.9 million 1961-64-brood fall Chinook salmon were marked at 12 Columbia River hatcheries and one rearing pond. Four brood years were marked to examine the differences be- tween broods. A mark common to all 13 facilities was used for each brood. Common marks were applied to 21.3 million fish. To examine the differ- ences between hatcheries, four hatcheries were assigned special marks for each brood. Two hatch- eries, Kalama River (in this study Kalama Falls and Lower Kalama Hatcheries were treated as one facility) and Spring Creek, had special marks for all four brood years. Special marks were applied to 9.6 million fish. Sampling for these marked chinook salmon took place from 1963 through 1969. Major marine sport and commercial fisheries from southeastern Alaska to central California and Columbia River fisheries were sampled for marks, and scale sam- ples were taken for age determination. Mark sam- pling ranged from 14 to 28^^ of the catch, and age sampling ranged from 1 to 47c by year. During the 7 yr of sampling, 3.3 million chinook salmon were sampled for marks and 208,000 were sampled for age. Returns to the 13 study facilities, adjacent streams, and nonstudy hatcheries rearing fall chinook salmon were sampled for marked 1961- 64-brood fish. Hatchery returns of these broods numbered 155,800 fish, of which 8,500 were marked. The stream sampling was conducted from 1964 through 1966 with 62,400 chinook salmon examined and 1,600 marked fish found. Hatchery contribution estimation is dependent on the validity of six assumptions. Where practi- cal, these assumptions were tested with additional studies and data collections. Assumption 1 (that the marks were permanent) was tested by holding marked fish in saltwater ponds for periodic examination. Some regeneration did occur but, since double and triple marks were applied, the marked fish remained identifiable throughout their life. Assumption 2 (that fish detected and reported with the kinds of marks applied at the hatcheries are hatchery fish) was tested by exam- ining hatchery fingerlings and 1965-brood chinook salmon catches for study marks. Over 30 million hatchery fingerlings were examined, and only 201 missing ventral and 156 missing adipose fins (none together) were found. The attempt to keep 1965-brood chinook salmon from being marked with study marks was unsuccessful. How- ever, ocean and Columbia River catches of study marks were adjusted for those marks that ap- peared to have a natural origin. Assumption 3 (fish were correctly aged from scales) was exam- ined by having six scale readers from State, Pro- vincial, and Federal agencies read 400 scales from fish of known age. The readers correctly aged 83% of the scales. Hatchery returns showed survival adjustments had to be made for assumption 4 (equality of survival and maturity schedules for marked and unmarked fish). Assumption 5 (the equality of ocean distribution and catch vulnera- bility of marked and unmarked fish) is supported by ocean tagging studies showing similar dis- tributions for marked and unmarked hatchery fish. A 10-part sampler was used to select fish for marking thus insuring the validity of assumption 6 (the marking of equal proportions of each hatch- ery's production). Estimated catches of special marked fish from the 10 special mark facilities ranged from 191 (Cascade, 1962 brood) to 4,406 (Spring Creek, 1964 brood). During the 7 yr of sampling, 65,620 common marked fish were estimated to have been caught: 22,237, 1961 brood; 7,886, 1962 brood; 22,183, 1963 brood; and 13,314, 1964 brood. Columbia River hatchery fish were captured in marine fisheries from Alaska to California. Marine catches were primarily in British Colum- bia and Washington fisheries. Fall chinook salm- on from the Kalama River hatcheries had a more northerly distribution than those from other spe- cial mark hatcheries. Kalama fish had the highest percentage catches of any special marked hatch- ery chinook salmon in Alaska and British Colum- bia fisheries. The average common marked fish catch distributions in percent of the total chinook salmon catch for the 1961-64 broods combined were: 0.2, Alaska commercial fisheries; 33.7, British Columbia commercial fisheries; 38.1, Washington marine fisheries; 4.6, Oregon ocean fisheries; 0.4, California ocean fisheries; and 23.1, Columbia River fisheries. The potential contribution of Spring Creek 1961-64-brood fall chinook salmon ranged from 30,300 (1962 brood) to 165,200 (1964 brood) with an average of 100,500 fish per brood. The average catch to release ratio was 12 fish per 1,000 fish released from Spring Creek. The Kalama hatch- eries potential contribution ranged from 22,300 (1962 brood) to 56,800 (1961 brood) and averaged 43,100 fish per brood. The average catch to release ratio for the two Kalama facilities was 9.6 fish for each 1,000 released. Potential contributions at the 206 WAHLE and VREELAND: BIOECONOMIC CONTRIBUTION OF FALL CHINOOK SALMON eight other special mark hatcheries (OxBow, Elokomin, Grays River, Cascade, Klickitat, Big Creek, Bonneville, and Little White Salmon) ranged from 2,000 fish (Elokomin, 1961 brood) to 42,500 (Klickitat, 1963 brood). The range of catch per 1,000 fish released was from 1.1 (Cascade, 1962 brood) to 14.7 (Klickitat, 1963 brood). The potential contribution for all study facilities com- bined ranged from 165,200 ( 1962 brood) to 602,200 (1963 brood). The average contribution was 358,500 fall chinook salmon per brood. The aver- age catch per 1,000 smolts released was 6.7 fish. Catch to escapement ratios ranged from 2.4 to 1 (Big Creek, 1963 brood) to 43.0 to 1 (Grays River. 1962 brood). Total survivals ranged from 0.1% (Elokomin, 1961 brood; Cascade, 1962 brood; Lit- tle White Salmon, 1964 brood) to 2.6% (Spring Creek, 1964 brood). Spring Creek Hatchery's av- erage catch to escapement ratio was 9.3 to 1 and the average survival was 1.3% . The average catch to escapement and survival values for the Kalama River hatcheries were 12.0 to 1 and 1.0%. For all facilities and the four broods combined, the aver- age survival was 0.7% and the average catch to escapement was 8.6 to 1. Spring Creek Hatchery and Big White Pond values were combined because Spring Creek per- sonnel operated the Big White facility making costs inseparable. The average cost of rearing each brood at the two facilities was approximately $100,000. The average value of the potential con- tribution was $1,119,600. The average benefit to cost ratio was 11.2 to 1. The average cost of rearing the 1961-64 broods of chinook salmon at the two Kalama hatcheries was $103,400. The average benefit from their production was $414,100, yield- ing a benefit to cost ratio of 4.0 to 1. For the other eight special mark hatcheries, costs ranged from $32,800 (Klickitat, 1963 brood) to $99,400 (Little White, 1964 brood), benefits from $16,900 (Elo- komin, 1961 brood) to $373,200 (Klickitat, 1963 brood), and benefit to cost ratios from 0.3 to 1 (Elokomin, 1961 brood) to 11.4 to 1 (Klickitat, 1963 brood). The average cost of rearing the four broods, all study facilities combined, was $714,900. The average benefit was $3,006,800, for an average benefit to cost ratio of 4.2 to 1. Fall chinook salmon releases from seven nonstudy Columbia River hatcheries totaled 26 million fish for the 1961-64 broods. If we assume these fish had a catch distribution and contribu- tion like the 13 study facilities, then the estimated total catch of fall chinook salmon from all Colum- bia River hatcheries is 1,467,600 fish. The 1961- to 1964-brood fall chinook salmon caught in marine fisheries sampled from Alaska to California and Columbia River fisheries was 14.8% of the total chinook salmon catch. The portions of the total chinook salmon catch by region originating from fall chinook salmon raised at Columbia River hatcheries were: Alaska, 0.3%; British Columbia, 12.3%; Washington, 38.2%; Oregon, 16.9%; California, 0.2%; and Columbia River, 41.7%. The 1961-64-brood Columbia River hatchery (study and nonstudy) contributions were valued at $2,921,700, $1,485,200, $5,794,300, and $3,284,600 by brood respectively. ACKNOWLEDGMENTS This study was planned and implemented with the assistance of several agencies and many indi- viduals. The Canadian Government financed and conducted a mark sampling program in the British Columbia fisheries. Alaska, Washington, Oregon, and California State fishery agencies pro- vided research and management personnel and necessary catch data. We especially thank the fol- lowing individuals: Donald D. Worlund, National Marine Fisheries Service, for developing the de- sign of this study and serving as the primary mathematical consultant; Jack A. Richards, Na- tional Marine Fisheries Service, for developing the justification for the sport and commercial economic evaluation; Robert C. Lewis, Bonneville Power Administration, for improving the method of amortizing hatchery construction costs; and Harold Godfrey, Canadian Fisheries and Marine Service; Gary Finger, Alaska Department of Fish and Game; Richard E. Noble, Emanual A. LeMier, Samuel G. Wright, and Harry Senn, Washington Department of Fisheries; Fred E. Locke, formerly Oregon Game Commission; Ernest A. Jeffries, Earl F. Pulford, Thomas B. McKee, and Roy E. Sams, Oregon Department of Fish and Wildlife; Paul T. Jensen, L. B. Boydstun, and William H. Sholes, California Department of Fish and Game; and Harlan E. Johnson and Warner G. Taylor, U.S. Fish and Wildlife Service, for their help in the design, supervision, and data collection portions of this study. We also thank Arthur H. Arp, Dean A. Eggert, Steven K. Olhausen, William D. Parente, Joe H. Rose, and Paul D. Zimmer for data organi- zation and previous reports which have led to this report. Helpful editorial comments were contri- buted by Roger Pearson, Frederick C. Cleaver, 207 FISHERY BULLETIN: VOL. 76, NO. 1 Richard T. Pressey, John I. Hodges, Kenneth Henry, and Jack Richards, National Marine Fisheries Service; E. W. Lesh, CaHfornia Depart- ment of Fish and Game; Earl F. Pulford and Ken- neth A. Johnson, Oregon Department of Fish and Wildlife; Samuel G. Wright, Washington Depart- ment of Fisheries; and Glenn H. Petry, Washing- ton State University. Special thanks go to Kath- leen M. LaBarge and Vivian Dignan who typed the text and tables for this publication. LITERATURE CITED Bergman, P. K., K. B. Jefferts, H. F. Fiscus, and R. C. Hager. 1968. A preliminary evaluation of an implanted, coded wire fish tag. Wash. Dep. Fish., Fish. Res. Pap. 3:63-84. Campbell, C. J., and F. E. Locke (editors). 1964. 1964 annual report. Oreg. State Game Comm., Fish. Div., Portland, 315 p. 1965. 1965 annual report. Fish. Div., Portland, 133 p. 1966. 1966 annual report. Fish. Div., Portland, 137 p. 1967. 1967 annual report. Fish. Div., Portland, 156 p. 1968. 1968 annual report. Fish. Div., Portland, 154 p. 1969. 1969 annual report. Fish. Div., Portland, 149 p. Cleaver, F. C. 1969. Effects of ocean fishing on 1961-brood fall chinook salmon from Columbia River hatcheries. Res. Rep. Fish Comm. Oreg. l(l):l-76. Conte, F. p., and H. H. Wagner. 1965. Development of osmotic and ionic regulation in juvenile steelhead trout Salmo gairdneri. Comp. Bio- chem. Physiol. 14:603-620. CoNTE, F. P., H. H. Wagner, J. Fessler, and C. Gnose. 1966. Development of osmotic and ionic regulation in juvenile coho salmon Oncorhynehus kisutch. Comp. Biochem. Physiol. 18:1-15. Oreg. State Game Comm., Oreg. State Game Comm., Oreg. State Game Comm., Oreg. State Game Comm., Oreg. State Game Comm., Greenhood, E. C, and D. J. Mackett. 1967. The California marine fish catch for 1965. Calif. Dep. Fish Game, Fish Bull. 135:1-42. Haw, F., H. O. Wendler, and G. Deschamps. 1967. Development of Washington state salmon sport fish- ery through 1964. Wash. Dep. Fish., Res. Bull. 7, 192 p. Heimann, R. F. G., and J. G. Carlisle, Jr. 1970. The California marine fish catch for 1968 and histor- ical review 1916-68. Calif Dep. Fish Game, Fish Bull. 149, 70 p. Heimann, R. F. G., and H. W. Frey. 1968a. The California marine fish catch for 1966. Calif. Dep. Fish Game, Fish Bull. 138, 76 p. 1968b. The California marine fish catch for 1967. Calif. Dep. Fish Game, Fish Bull. 144, 47 p. HUBLOU, W. F. 1963. Oregon pellets. Prog. Fish-Cult. 25:175-180. MacPhee, C„ and R. RUELLE. 1969. A chemical selectively lethal to squawfish (Ptycho- cheilus oregonensis and P. umpquae). Trans. Am. Fish. Soc. 98:676-684. Nye, g. d., and w. d. ward. .Undated a. Washington salmon sport catch report from punch card returns in 1968. Wash. Dep. Fish., Olympia, 71 p. Undated b. Washington salmon sport catch report from punch card returns in 1969. Wash. Dep. Fish., Olympia, 60 p. PINKAS, L. 1970. The California marine fish catch for 1969. Calif. Dep. Fish Game, Fish Bull. 153, 47 p. Van Hyning, J. M. 1973. Factors affecting the abundance of fall chinook salmon in the Columbia River. Res. Rep. Fish Comm. Oreg. 4(1): 1-87. Wagner, H. H., F. p. conte, and J. L. Fessler. 1969. Development of osmotic and ionic regulation in two races of chinook salmon Oncorhynehus tshawytscha. Comp. Biochem. Physiol. 29:325-341. WAHLE, R. J., R. R. VREELAND, AND R. H. LANDER. 1974. Bioeconomic contribution of Columbia River hatch- ery coho salmon, 1965 and 1966 broods, to the Pacific salmon fisheries. Fish. Bull., U.S. 72:139-169. WORLUND, D. D., R. J. WAHLE, AND P. D. ZIMMER. 1969. Contribution of Columbia River hatcheries to har- vest of fall chinook salmon (Oncorhynehus tshawytsc- ha). U.S. Fish Wildl. Serv., Fish. Bull. 67:361-391. 208 POLYCHAETOUS ANNELIDS OF THE DELAWARE BAY REGION Peter Kinner^ and Don Maurer^ ABSTRACT Between 1967 and the present, 1,303 quantitative and 887 qualitative samples were taken from 10 different areas in the Delaware Bay region. Four major areas were examined: Delaware Bay proper, two smaller bays, the coastal areas, and offshore on the midcontinental shelf A total of 125 species of polychaetous annelids representing 34 families and 88 genera were identified. The greatest number of species (95) was collected at the offshore stations, which also had the highest genus to species ratio (1:1.6). Delaware Bay samples contained 83 species and the coastal areas 74 species. The smallest number of species was collected in the small bays (33). The dominant species on the midcontinental shelf were: Goniadella gracilis, Lumbrinerides acuta, Spiophanes bombyx, Exogone hebes, and E. verugera. In Delaware Bay, Heteromastus fUiformis, Nephtys picta, and Glycera dibranchiata were collected most regularly. The polychaete fauna of three epifaunal assemblages (mussel bed, serpulid "reef," and oyster beds) were also examined. Increasing numbers of Nephtys picta, Glycera dibran- chiata, and Heteromastus fUiformis were associated with sediments containing increasing amounts of silt-clay in Delaware Bay. Lumbrinerides acuta and Goniadella gracilis were associated with poorly sorted coarse sediments ( >1 mm) on the continental shelf. A zoogeographic analysis revealed this area to be the southern limit of the range for 11 species and the northern limit for 3 species. The Delaware fauna was more closely related to the northern fauna than to the southern fauna. A summary is given for some recent taxonomic changes in species present in the coastal waters of the eastern United States. This account was prepared to review the composi- tion, distribution, and general ecology of polychaetous annelids in the Delaware Bay re- gion. The most comprehensive treatment of polychaetes from the northeast Atlantic off the United States was presented by Pettibone ( 1963a). She reported 183 species from 29 families; cited records of depth, sediment preference, and repro- ductive condition; and collated and reviewed the works of Webster, Benedict, Verrill, Treadwell, Moore, Hartman, and others. Since then, she has published research on paraonids, spionids, sigalionids, pilargids, and nereids from the north- east Atlantic (Pettibone 1962, 1963b, 1965, 1966, 1970a, b, 1971). Hobson (1971) has added some additional records to the polychaetes of New En- gland. Deepwater polychaetes from the western Atlantic Ocean, including New England, were de- scribed by Hartman (1965) and Hartman and Fauchald ( 1971). Gosner (1971) prepared a key for invertebrates from Cape Hatteras to the Bay of Fundy and listed 213 species of polychaetes. Pratt^ 'College of Marine Studies, University of Delaware, Lewes, Del.; present address: Normandeau Associates, Inc., 15 Picker- ing Avenue, Portsmouth, NH 03801. ^College of Marine Studies, University of Delaware, Lewes, DE 19958. ^Pratt, S. D. 1973. Benthic fauna./n S. B. Saila (editor). Coast- al and offshore environmental inventory, Cape Hatteras to Nan- tucket Shoals, 5:1-70. Univ. R.I., Mar. Publ. Ser. 2. reviewed the literature on polychaetes from Nan- tucket to Cape Hatteras. In nearshore waters off North Carolina, Hartman (1945) reported 104 species of polychaetes and presented information on tube building, reproductive maturity, and faunal as- sociations. Wells and Gray (1964) listed 110 species from the Cape Hatteras area and mainly emphasized the zoogeographic affinities of the polychaetes. Day et al. (1971) analyzed distribu- tional patterns of the benthic fauna across the continental shelf off Beaufort, N.C., from the shore to 200 m in depth. Later, Day ( 1973) reported 229 species of polychaetes from the shelf study and prepared a guide for the species known from North Carolina. More recently, Gardiner ( 1975) provided a key to 163 species of errant polychaetes from intertidal and shallow subtidal zones of North Carolina. Wass (1972) compiled a valuable list of the benthic fauna of Chesapeake Bay, including polychaetes, with annotated records of ecological data. The earliest work on polychaetes in the Dela- ware Bay area was conducted by Leidy ( 1855) and Webster ( 1880, 1886). Polychaetes associated with oyster beds in Delaware were discussed by Maurer and Watling (1973). Wells (1970) and Curtis (1975) described reefs of "sand coral" (Sabellaria vulgaris) from the shores of Delaware. Manuscript accepted June 1977. nSHERY BULLETIN: VOL. 76, NO. 1, 1978. 209 FISHERY BULLETIN: VOL. 76, NO. 1 METHODS Since 1967, a large number of quantitative (1,303) and qualitative (887) samples of benthic invertebrates have been collected throughout the Delaware Bay region. The major collecting areas are designated with letters and presented in Fig- ure 1. Since the objectives of the various surveys differed, the sampling pattern and season, number of samples, frequency of sampling, collecting gear and sieve type, environmental data, and type of analysis also varied (Table 1). A local reference collection was established and verified with the polychaete collection in the U.S. National Museum. ENVIRONMENTAL SETTING The general environmental setting is discussed as four major areas: Delaware Bay proper, small bays, coastal areas, and offshore. Polychaetes ATLANTIC OCEAN 74°W Figure l. — Polychaete sampling in the Delaware Bay region. The sampling areas are: A. baywide, B. bay mouth, C. midbay , D. oyster beds, E, intertidal, F. small bays, G. Bethany Beach, H. Hen and Chickens Shoal, I. off Delaware Bay mouth, and J. midshelf site. were collected from a variety of habitats which have been designated as follows: Delaware Bay area (Figure 1) Delaware Bay proper Baywide (A) Bay mouth (B) Midbay (C) Sandy shoals (Brown Shoal, Lower Middle Shoal, Old Bare Shoal) Muddy sand bottom Epifaunal-infaunal assemblages (blue mussel assemblage; calcareous serpulid assemblage) Oyster beds (Delaware Bay; Broadkill, Mis- pillion, Murderkill, St. Jones, and Leipsic Rivers) (D) Intertidal — Cape Henlopen (E) Small Bays Rehoboth and Indian River Bays (F) Coastal areas Bethany Beach (G) Hen and Chickens Shoal (H) Off mouth of Delaware Bay (I) Offshore Midshelf site (J) The letters in parentheses refer to letters used to designate areas in Figure 1. Delaware Bay Proper The morphology, geology, and sediment dis- tribution of Delaware Bay (Figure 1, A) was de- scribed by Shuster,4 Kraft,^ Weil ( 1975), and Wat- ling and Maurer.^ Salinity values were 5-8%o at the northern limit of sampling and 30-3 l%o near the bay mouth, with the major part of the area being polyhaline (18-30%o) (Table 1). Sediment at the bay mouth was generally medium sand ( l-2(/)), with the coarsest material in the middle of the bay (Figure 2). Sand farther up the bay became finer (2-3. 5c^), with medium sand in the center channel. Sediments along both sides of the estuary were fine, wdth as much as 90% silt-clay in some sam- ples. In sediments from the northernmost tran- ■•Shuster, C. N. 1959. A biological evaluation of the Delaware River Estuary. Univ. Del. Mar. Lab., Inf Ser., Publ. 3, 77 p. ^Kraft, J. C. 1971. A guide to the geology of Delaware's coastal environments. Univ. Del., Coll. Mar. Stud. Publ. No. 2GL039, 220 p. *Watling, L., and D. Maurer (editors). 1976. Ecological studies on benthic and planktonic assemblages in lower Delaware Bay. NSF/RANN. Univ. Del., Coll. Mar. Stud. Publ., 630 p. 210 KINNER and MAURER: POLYCHAETOUS ANNELIDS OF DELAWARE BAY Table l. — Summary of collecting and environmental data for Delaware Bay area polychaetous annelids (areas shown in Figure 1). Area Sampling pattern. number of samples, frequency of sampling Collecting gear and processing Salinity Depth (m) Substrate Source Baywide (A) Transects; 207 samples; summer 1972, 1973 0 1-m^ Petersen grab. 1 O-mm mesh seive 5.0-31.0 1,0-50.0 Bay mouth ^-2(b: midbay 2-3. 5d>; Delaware side coarse sand; fine sediment along both shores Watling and Maurer (see text footnote 5) Bay mouth (B) Random spacing; 277 samples; Dec 1971, Mar. 1972, June 1972 0 l-m^ Petersen grab, 1 O-mm mesh sieve 23.0-29.0 1.0-30.0 100% silt-clay, medium to coarse sand in northwest Maurer et al. (see text footnote 6) Midbay (C) Selected stations; 170 samples; l^ay, Aug., Nov 1974. Feb , May 1975; 60 samples. August 1975 0 1-m^ Petersen grab. 1.0-mm mesh sieve; dredges 2 0, 1 0, 0 5. 0.25 mesh sieves 21 0-297 30-350 Well sorted shoal sands, mud (30% silt-clay, calcareous serpulid reef, bimodal sediment with silt and coarse sand) Watling and Maurer (see text footnote 5) Oyster beds in rivers, bay (D) Random spacing; ^ 800 samples from 1967 to 1971 Oyster dredge. 1 gal, sample. 0.25-mm mesh sieve 20-330 200-285 0 5-6 0 2 5-8 0 Hard shell bottom intercalated with mud and muddy shell bottom Maurer and Watling (1973) Intertidal (E) Transects. 200 samples, monthly from 1970 to 1972 25 X 25 cm core. 1 O-mm mesh sieve 26.0-31 0 Sediment ranged from coarse sand (>025% gravel Maurer et al. (1976) Watling et al. (1974) sects, medium sands (1.5-3.00) were restricted to the ship channel, grading rapidly into finer sedi- ments {7.04>) away from the channel. At the interidal site (E) just inside Cape Henlo- pen (Figure 1), salinity ranged from 26.0 to 31.0%o, but it became higher in trapped shallow ponds during the summer. Sediment consisted of a fine sand ( <2.0(t>) at the northwest end of the flat and a coarse sand ( >0) at the ocean end. Environmen- tal data, including sediment distribution, surface and bottom temperature, salinity, and dissolved oxygen, are discussed more extensively by Maurer et al. (1971), Maurer et al.,'' Kinner et al. (1974), and Watling and Maurer (see footnote 6). Small Bays Delaware has several small bays (F), which have received considerable attention in recent years (Logan and Maurer 1975; Watling 1975; Brenum 1976; Maurer in press; Jones et al.^). In Rehoboth Bay, salinity varied seasonally from 20.1 to 30.8%o and the average silt-clay in the sediment was 40.3%. Salinities in Indian River Bay ranged from 27.7 to 31.9%oat the mouth of the bay and 7.5 to 19.3%o near the Indian River. Sedi- ment was similar to that in Rehoboth Bay, except that the bay mouth contained coarse sand and shell fragments. Coastal Areas In coastal waters, collections were concentrated at three sites (Figure 1; G, H, I). The annual mean range of sahnity was 28.5-30. 7%o and 27.2-29.8%o at Bethany Beach (G) and Hen and Chickens Shoal (H), respectively. Sediment at the two sites can be characterized as medium sand. Occasional depressions and holes trapped finer grained sedi- ment. The deeper areas of Hen and Chickens Shoal 'Maurer, D., R. Biggs, W. Leathem, P. Kinner, W. Treasure, M. Otley, L. Watling, and V. Klemas. 1974. Effect of spoil dis- posal on benthic communities near the mouth of Delaware Bay. Univ. Del., Coll. Mar, Stud. Publ, 200 p. *Jones, R. D., L. D. Jensen, and R. W. Koss. 1974. Environmen- tal responses to thermal discharges from the Indian River sta- tion, Indian River, Delaware. Rep. 12, Cooling Water Studies for Electric Power Research Institute, Res. Proj. (RP-49). 211 FISHERY BULLETIN: VOL. 76, NO. 1 FIGURE 2.— Mean grain size for surface sediment in Delaware Bay. Dots represent the baywide (A) sampling stations. 212 KINNER and MAURER: POLYCHAETOUS ANNELIDS OF DELAWARE BAY also contained large rocks, small boulders, and mussel beds. A detailed account of these areas can be found in Maurer et al.^ Although Area I was about 20 km off the Delaware Bay mouth, the western portion of this area appeared to be influenced by the hydrography of Delaware Bay. Salinity ranged from 28.2 to 32.5%o and the sedi- ment varied from silty-sand to gravelly sand. However, a few sediment samples contained black mud (30-33'^ silt-clay and 2.63-3.64% organiccon- tent) (Watling et al. 1974). Offshore The oceanic or offshore area, termed midshelf site (Figure 1, J), has been the subject of several studies. An extensive review of the hydrography and geology was presented by Bumpus et al.^" and Milliman.ii Salinity was 31-40.0%o and the sedi- ment was dominated by clean sand with some peb- bles and dead shells at the collecting site. Ridge and swale microtopography influences sediment composition. Crests of the ridges contained clean sand and swales or troughs consisted of shell and flocculent material (Maurer et al. 1976). Results and Discussion A total of 125 species of poly chaetes, represent- ing 34 families and 88 genera, were identified from all the sampling areas. Eighty-three species and 25 families were collected within Delaware Bay proper (Table 2, columns A-E). The Delaware Bay samples usually showed less than 10 species and 250 individuals/m^. However, the most species (95) were collected in the offshore samples. The number of individuals per sample was much higher at stations in the midshelf area. This was also the only collection where the polychaetes dominated the fauna. Infaunal samples in both the bay and in the nearshore areas are otherwise dominated by members of the Mollusca (Maurer et al. see footnote 7; Watling et al. 1974). 'Maurer, D., J. Tinsman, W. Leathern, and P. Kinner. 1974. Baseline study of Sussex County, Delaware ocean outfalls. Rep. Sussex County Engineer, Sussex County Delaware. Univ. Del., Coll. Mar. Stud., 287 p. '"Bumpus, D. F., R. E. Lynde, and D. M. Shaw. 1973. Physical oceanography. In S. B. Saila (editor). Coastal and offshore en- vironmental inventory Cape Hatteras to Nantucket Shoals, 72 p. Univ. R.I., Mar, Publ. Ser. 2. "Milliman, J. D. 1974. Marine geolog>-. /n S. B. Saila (editor). Coastal and offshore environmental inventory Cape Hatteras to Nantucket Shoals. Univ. R.I., Mar. Publ. Ser 3. Delaware Bay Intertidal — Cape Henlopen (E) Eight core samples (25 cm diameter x 25 cm height) were taken each month for 25 mo on Cape Henlopen near the mouth of the bay, from 1970 to 1972 (Figure 1). The study area was on the bay side of the spit on a tidal flat with swash bars. Eighteen species of polychaetes were collected in the sampling area (Table 2, column E). The number of species decreased gradually from fine to coarse sand (Maurer unpubl. data). Large tube- building polychaetes (Diopatra cuprea) and bur- rowing infaunal species (Lumbrineris tenuis and Scoloplos fragilis) occurred in highest densities in the fine sand, whereas spionids and nephtyids were better represented where sediment grain size increased towards the ocean. Two species (S. fragilis and Spio setosa) were particularly abun- dant from the low to the high tide line. Scoloplos fragilis was most common just above the reducing layer in the sediment. Pista palmata was collected only in the sand flat area. Baywide (A) The polychaete fauna in the upper bay (5-15%o) was dominated by the deposit feeders Heteromas- tus fUiformis and Scolecolepides viridis (Table 2, column A). Glycera dibranchiata was also present at a number of stations. Sediments in this area ranged from M 4.2 to 7.9<^ (median grain size), vdth generally poor sorting (o- = 2.4-3.90). At all stations the numbers of individuals were very small, with four individuals being the most re- corded at one time. This paucity of individuals was also evident in other groups of the benthic fauna. Farther down the bay where salinities were 15- 25%o, there was an increase in number of species and individuals. Thirty-two species were col- lected, including all the six species recorded in the area of 5-15%o (Table 1). The sediment showed a much wider range of particle size (M 1.0-7.0200m), 3 = Chesapeake Bay, 4 = North Carolina; * = species at the southern extension of their range; *' = species at the northern extension of their range.] Polychaete species Ampharetidae: Ampharele arclica Malmgren Asabellides oculata (Webster) Hypaniola florida (Hartman) Melinna maculata Webster Amphictenidae (= Pectinariidae): Cistena gouldii (Verrlll) Aphrodltidae: Aphrodita hastata Moore Arabellidae: Arabella iricolor (Montagu) Driloneris longa Webster D magna Webster and Benedict Capitellidae: Capilella capitata (Fabricius) Heteromastus fililormls (Claparede) Mediomastus ambiseta (Hartman) Chaetopteridae: Spiochaetopterus oculatus Webster Cirratulidae; Caulleriella spp, Chaetozone setosa Malmgren Chaetozone spp. Cirralulus grandis Verrill Cirriformia filigera (Delle Ctiiaje) Tharyx acutus Webster and Benedict Tharyx sp, Dorvilleidae: Protodorvillea gaspeensis Pettibone Schistomeringos caeca (Webster and Benedict) S. rudolphi (Delle Ctiiaje) Eunicidae: Marphysa belli (Audouin and Milne-Edwards) M. sanguinea (Montagu) Flabelligeridae: Pherusa affinis (Leidy) Glyceridae; Glycera americana Leidy G. capitata Oersted G. dibranchiata Ehlers Goniadidae: Glycinde solitaria (Webster) Goniadella gracilis (Verrill) Hesionidae: Gyptis vittata Webster and Benedict Microphthalmus schzelkowii Mecznikow Podarke obscura Verrill Lumbrinendae; Lumbnnendes acuta (Verrill) Lumbnnens coccinea (Renier) L- fragilis (OF Muller) L. impatiens (Clapar6de) L latrielli (Audouin and Milne-Edwards) L. tenuis Verrill Magelonidae: Magelona sp A Magelona sp B (near riojai) Maldanidae Asychis elongata (Verrill) Clymenella mucosa (Andrews) Clymenella spp. C. torquata (Leidy) C. zonalis (Verrill) Clymenura borealis (Arwidsson) Praxillella sp. Neptityidae: Aglaophamus circinata Verrill Nephtys bucera Etilers N. incisa Malmgren N. picta Ehlers 10 20 80 20 1.770 40 10 150 20 40 20 10 30 20 150 490 220 900 1,500 560 10 10 10 180 50 10 10 10 60 40 30 30 30 30 10 30 :30 180 60 40 10 30 50 100 80 20 50 40 20 270 40 10 100 10 10 10 10 30 10 243 1,273 1,659 29 129 43 42 10 30 40 60 40 150 129 114 229 110 50 10 190 10 10 100 50 14 10 10 25 X 10 10 25 X 10 10 25 X 458 10 40 X 114 10 20 25 X X' 10 130 150 10 10 275 X 20 325 20 20 25 X 80 960 180 X 60 20 525 25 X' 40 50 X 60 100 50 X 10 50 X X 25 475 200 X 250 X 25 X 150 825 X 80 20 30 50 X 100 75 X 170 40 110 250 X 20 160 10 25 X X 157 10 70 X X X X 200 270 50 125 X X 43 100 40 40 175 X X X 257 80 10 X X X 10 110 3,950 X X 29 10 10 X x' X X 72 x X X 20 10 20 925 25 X X X X 10 10 150 75 150 425 X X X X X X X ,530 10 30 40 10 110 100 25 X x X X X 214 KINNER and MAURER: POLYCHAETOUS ANNELIDS OF DELAWARE BAY Table 2.— (Continued). Polychaete species A B c D E F G H 1 J 1 2 3 4 Nereidae: Nereis grayi Pettibone 11 125 X X X N succinea (Frey and Leuckart) 130 30 450 X 63 672 20 900 25 X X X Onuphldae; Diopatra cuprea (Bosc) 10 10 X 29 10 X X X Onuphis opalina (Verrill) 10 10 X Opheliidae: Ophelia bicornis Savigny 10 180 30 40 25 X X 0 denticulata Vernll 20 10 25 X X Ophelina cylindncaudala Hansen 25 X X Travisia carnea Vernll 30 10 380 10 75 X X Orbiniidae; Orbinia ornata (Vernll) 20 10 25 X X X 0 swam Pettibone 25 X- Scoloplos armiger (OF. Muller) 30 20 20 50 X X S, fragilis (Vernll) 60 30 130 X 3.024 858 10 25 X X X S, robustus (Verrill) 40 40 10 X X X Oweniidae: Myriowenia sp. A 25 Owenia lusiformis Delle Chiaje X 14 X X X Paraonidae: Aricidea cathennae Laubier 40 90 240 180 60 350 X X X A. suecica Eliason 125 X > ( X A wassi Pettibone 250 X X Cirrophorus branchiatus Ehlers 25 > ( X Paradoneis lyra (Southern) 10 130 10 125 X > ! Phyllodocidae: Eteone flava (Fabricius) 25 X- £ heteropoda Harlman 10 30 X 10 25 X X X E. lactea Clapar6de 10 X 558 30 25 X X X E. longa (Fabricius) 60 60 X £ trilineata (Webster and Benedict) 25 X' Eulalia bilineata (Johnston) 50 X Eumida sanguinea (Oersted) 40 1.240 X 143 X X X Paranaitis kosteriensis (Malmgren) 10 10 X P speciosa (Webster) 20 14 50 X X X Phyllodoce arenae Webster 20 60 14 30 50 75 X X X P maculata Linnaeus 50 10 10 25 X* P. mucosa Oersted ■ 20 25 X X X Pisionldae: Pisione rernota (Southern) 10 80 X Polynoidae: Harmothoe extenuata (Grube) 240 10 790 X 30 380 20 25 X X Lepidametria commensalis Webster 10 72 X X X Lepidonotus squamalus (Linnaeus) 20 270 10 50 10 X X L sublevis Vernll 10 30 270 X X X X X Sabellarlidae: Sabellaria vulgaris Verrill 70 150 2,310 X 57 720 120 X X X Sabellidae; Chone spp. 400 Euchone spp 100 Potamilla neglecta Sars 50 X X P. reniformis (Leuckart) 50 X X Sabella microphthalma Verrill X 14 25 X X X Scalibregmidae: Scalibregma inflatum Rathke 150 X ) < X Serpulldae: Hydroides dianthus (Verrill) 1,930 40 8,160 X 21 43 10 40 150 X X X Sigalionidae: Pholoe minuta (Fabricius) 25 X X Sigalion arenicola Verrill 40 10 50 75 X X Sthenelais limicola (Ehlers) 10 10 20 10 50 X X X S. boa (Johnston) 60 10 10 75 X X X Spionidae: Dispio uncinata Hartman 10 10 20 10 X X Parapionspio pinnata (Ehlers) 10 10 10 X X X Polydora caulleryi Mesnil 10 10 50 X P concharum Vernll 10 75 X* P ligni Webster 1,050 10 330 X 2,131 80 X X X P sociahs (Schmarda) 10 440 10 200 10 25 X X P webslen Hartman 20 X X X X Pnonospio cnstata Foster 25 X" P steenstrupi Malmgren 100 X X X Scolecolepides vihdis (Verrill) 40 20 X X X Scolelepis squamata (OF. Muller) 10 10 20 21 40 30 25 X X Spio setosa Verrill 10 5,450 42 150 40 40 50 X X X Spiophanes bombyx (Clapar6de) 70 110 320 70 160 2,550 X X X Streblospio benedicti Webster 160 10 590 X 86 10 120 X X X 215 FISHERY BULLETIN: VOL. 76, NO. 1 Table 2.— (Continued). 20 Polychaete species Syllidae; Brania clavata (Clapar^de) Exogone dispar Webster £ hebes (Webster and Benedict) E verugera (Claparede) Parapionosyllis longicirrata (Webster and Benedict) 10 Proceraea cornuta (Agassiz) 220 Sphaerosyll(S erinaceus Claparede S hyslrix Claparede Streptosyllis arenae Webster and Benedict S varlans Webster and Benedict Syllis cornuta Rattike S gracilis Grube Syltides sp Terebellidae: Amphitrite ornata (Leidy) Pista cristata (OF Mijller) P. palmata (Verrill) Poly cirrus eximius (Leidy) 190 20 30 50 40 1,140 40 10 40 20 40 830 25 850 1,425 1,125 175 50 125 175 75 50 100 100 estuary contained larger densities of carnivores and omnivores. One station on the most southerly transect in this salinity range had the coarest sed- iment found to this point (M 1.00) and the most diverse fauna. Eleven species were present repre- senting both sedentary (e.g., H. fUiformis, Streh- lospio benedicti, and Asabellides oculata) and er- rant types (e.g., Glycera dibranchiata, G. americana, Eteone heteropoda, and E. longa). Since all species mentioned occurred at both higher and lower salinities, species richness may be a response to the sediment type. Fifty-one species were collected in the estuary in salinities >25%o. The six species found in the upper bay all occurred here. Nineteen species col- lected in the midbay area were also found in the high-salinity samples. Twenty-six additional species found in the lower estuary were not found in salinities <25%o. They were equally divided between sedentary and errant types. The seden- tary deposit-feeding species are mainly sand- dweller types, such as Paradoneis lyra, Scolelepis squamata, and Spio setosa, while the errant species consisted of phyllodocids, nephtyids, and polynoids. Delaware Bay Temporal Studies To examine more closely the temporal changes in assemblages in different Delaware Bay sedi- ments, a program of quarterly sampling was un- dertaken in Area C ( Figure 1 ) . Three sandy shoals, three muddy sand bottoms, a polymodal sediment, and a calcareous serpulid assemblage were the selected sites ( Watling and Maurer see footnote 6). At all of the stations the salinity was >25%o. In addition to the quarterly samples, 20 replicate 216 grabs and 20 replicate dredge hauls were taken at a station representing each substrate to obtain a more accurate count of species abundances. SANDY SHOALS.— Two of the shoal stations were located in the middle of the bay on Brown Shoal and Lower Middle Shoal. Sediments were medium-well sorted (M 1.9-2. 9c/), cr = 0.30(f)) sand constantly subjected to strong tidal currents. The fauna was restricted to a few species of polychaetes throughout the year: Nephtys picta, N. bucera, Magelona sp. 2 (near riojai), and Spiophanes bom- byx. The species were always present in densities of <10 individuals/0.1 m^. The third shoal station on Old Bare Shoal was slightly different in faunal and sedimentary characteristics. The sediment was finer ( M 2.8-2.9(f), a = 0.3025%o, as well as in the nearshore and offshore communities. Pholoe minuta and Sigalion arenicola were present only in the coastal and offshore stations. Sigalion arenicola occurred in 12% of the November offshore samples, with many individuals being juveniles. None of these scale worms were ever found in large numbers in any sample. The in- crease in sigalionids offshore was not matched by the other major scale worm family, the Polynoidae. Polynoids were extremely numerous in the bay, particularly in the epifaunal-infaunal communities. In the offshore marine areas, only Harmothoe extenuata was present. The absence of collections from hard substrate offshore may affect the average numbers of offshore polynoids. The Sigalionidae typically are burrowing forms and may find the fine sandy substrate more suitable than do the polynoids. Maldanids were important in the three seasonal offshore sampling periods. Most of the individuals collected were juveniles, and thus difficult to iden- tify. Most adult specimens were Clymenella zonalis, C. torquata, and C. mucosa. 219 FISHERY BULLETIN: VOL. 76, NO. 1 ANIMAL-SEDIMENT RELATIONSHIPS To describe some of the sediment associations of the dominant species of Delaware Bay, correla- tions were made with median grain size and per- centage of silt-clay using Spearman's p (a = 0.05). Nephtys picta was collected in sediments with an unweighted mean grain size of 2.1c/) and in 1-10% silt-clay (x = 4.7%). Increasing abundance of A^. picta was associated with increasing amounts of silt-clay within the range in which it occurred. Glycera dibranchiata was found in sediment with up to 50% silt-clay (3c = 13.3%), and a mean size ranging from 0.8 to 6.6(^ (x = 2.7). There was a positive association between numbers of individu- als and increasing silt-clay content. No other as- sociations were significant. Two of the dominant species were found primar- ily in muddier sands. Heteromastus filiformis has been described as a member of soft sediment com- munities in Delaware Bay (Kinner et al. 1974) as well as elsewhere (Dean and Haskin 1964). The species inhabited a wide range of sediments, M 0.08-6.5(^ (x = 3.7), and was positively correlated with increasing silt-clay and increasing median and mean grain size. Streblospio benedicti occur- red in sediments with a wide range of silt-clay (2.5-59.0%). The distribution of the species was not correlated with median grain size, silt-clay, or mean grain size. Streblospio showed an even greater affinity for the areas along the Delaware and New Jersey shoreline than did H. filiformis. Correlations were also made between measures of sediment and five of the dominant polychaetes of the offshore assemblages. Lumbrinerides acuta (0.76-2.40(/)) and Goniadella gracilis (0.76-2.49(/)) were negatively associated (a = 0.05) with in- creases in median and positively correlated with an increase in the percentage of sediment >1 mm in diameter. Both species showed correlations of high density with more poorly sorted sediments. Nichols (1970) has postulated that although sort- ing is not well understood biologically, positive correlations with well sorted sediments may indi- cate niche specificity, while poor sorting suits a wider variety of needs. The larger sediment sizes probably facilitate burrowing. Aricidea catherinae (0.34-2.64(/)) was negatively associated with an increase in the size of the me- dian (I). This deposit-feeding species builds a flexi- ble mucous tube and is far less mobile than L. acuta and G. gracilis. Sediments containing parti- cles >1 mm may be difficult for this fragile species. 220 Aglaophamus circinata was not significantly as- sociated with any sediment parameters. However, it was found in a range of sediment (0.34-2.64(^) similar to that of the other species. Sediments which contained the greatest densities of Spiophanes bombyx were generally well sorted (o- = 0.21-0.570) with between 25% and 50% of the sediment >04>. There was a negative associa- tion (a = 0.05) between S. bombyx and sediment >1 mm. This species was also negatively as- sociated with an increase in the standard devia- tion of (/) indicating its preference for a well-sorted sediment. GENUS-SPECIES RELATIONSHIPS A comparison was made of the genus to species ratios for each of the estuarine coastal and offshore areas to obtain information on diversity and speciation. The midshelf station had the highest ratio of 1.0:1.6 with the Serpulidae and Mytilus assemblages second (1.0:1.4). Coastal areas were next with Hen and Chickens Shoal and Bethany Beach 1.3 and off the bay mouth 1.2. The areas within Delaware Bay and the small bays were as follows: baywide (1.3), intertidal (1.3), bay mouth (1.0), oyster beds (1.2), and small bays (1.1). The epifaunal-infaunal speciation ratio does not reflect the stability of the habitat, but rather the greater number of niches due to a mixed sub- stratum. Winter reductions in species diversity in the Mytilus assemblage due to storms and mussel mortality emphasize the fragile nature of the en- vironmental stability. TAXONOMIC NOTES Revisions and synonymies that appear in polychaete taxonomic literature are often not in- cluded in ecological publications for a long time. Based on suggestions from Marian Pettibone, we have included a section describing some of the systematic changes that affect the east coast of the United States. We formally acknowledge her for providing us with much of the information in- cluded in this section. Ampharetidae Hypaniola florida (Hartman) In a recent paper Pettibone (1977) has pre- sented the synonomy and distribution of the es- tuarine species, Hypaniola florida (Hartman). The KINNER and MAURER: POLYCHAETOUS ANNELIDS OF DELAWARE BAY species was reported as Amphicteis gunneri floridus by Hartman in 1951 from Florida and as Hypaniola grayi by Pettibone (1953) from Mas- sachusetts and by Kinner et al. ( 1974) from Dela- ware Bay. Wass (1972) listed the species as Lysipiddes grayi from Chesapeake Bay and Zottoli (1974) used the name Amphicteis floridus from New Hampshire. Pettibone stated that this species is distributed in estuaries from Maine to Florida and the Gulf of Mexico. Amphictenidae (= Pectinariidae) Lucus and Holthuis (1975) showed that the type-species of the well known generic name Pec- tinaria Lamarck was confused and had to be re- placed by Cistena Leach. Since the genus Pec- tinaria is no longer valid, the widely used family name Amphictenidae is now preferred to Pec- tinariidae. The single east coast representative should now be referred to as Cistena gouldii (Ver- rill) new combination. Capitellidae Mediomastus ambiseta (Hartman) was a do- minant species in the mussel and serpulid as- semblages. Hartman (1947) described the species as Capitata ambiseta from intertidal flats in California. Hartman-Schroder (1962) later synonymized Capitata with Mediomastus, and Hobson (1971) reported it for the east coast of the United States. The species has been reported as a dominant species in Newport Bay, Calif., and Baja California by Reish (1959, 1963) and in Florida by Dauer and Simon (1975, 1976a, b). Mediomastus californiensis has been reported from North Carolina (Day 1973), but it differs from M. ambiseta in a number of characteristics. Mediomastus californiensis lacks a caudal process, and spinous setae in posterior segments that are represented in M. ambiseta, and has a different positioning of the distal teeth of the hooked setae. Dorvilleidae According to a recent revision of the genera of the family Dorvilleidae by Jumars (1974), the new generic name Schistomeringos replaces Stauro- nereis as used by Pettibone (1963a) and Wass ( 1972) and Doruillea as used by Day ( 1973) for the species Schistomeringos caeca and S. rudolphi. Protodorvillea gaspeensis, described originally by Pettibone ( 1961) from the Gulf of St. Lawrence, was reported from Massachusetts by Hobson (1971) and now from the midcontinental shelf off Delaware. Magelonidae Two species of Magelonidae have been recorded from Delaware and designated as Magelona sp. A and Magelona sp. B. Meredith Jones is currently revising this group and he informs us that Magelona sp. B is near M. riojai (Jones 1963). Maldanidae In a revision of three species of Maldanidae from the east coast of the United States, Mangum (1962) included three species under Clymenella: C. torquata (Leidy), C. zonalis (Verrill), and C mucosa (Andrews). Day (1973) maintained the genus Axiothella for C. mucosa; however, Man- gum has pointed out that this separation, based on the position of segmental collars, is not warranted because of the presence of collars scattered throughout the family. Clymenella zonalis was re- ported by Day (1973) as Macroclymene zonalis. The genus, Macroclymene, was originally erected as a subgenus by Verrill for a specimen which had a much larger number of segments than his type. The subgenus was raised to generic status by Hartman ( 195 1 ) for a fragment found in the Gulf of Mexico. Mangum pointed to the great variation in segmental number even within populations and thus rejected Maroclymene. It has also been our experience that numbers of segments vary. We have found that juveniles particularly do not fit the characteristic segmental numbers, and as a re- sult, have used Clymenella spp. and Praxillella sp. Light (1974), in a comparison of Maldanidae specimens from San Francisco Bay and the east coast of the United States, followed Ardwidsson and referred Verrill's species Maldane elongata to Asychis ( including the synonymy). The species has been reported from Chesapeake Bay by Wass (1972) as Maldanopsis and from North Carolina by Hartman (1945) and Day (1973) as Bran- chioasychis americana Hartman. Orbiniidae In a study involving various growth stages of Scoloplos armiger, Curtis (1970) has shown S. acutus to be a juvenile form of S. armiger. The 221 FISHERY BULLETIN: VOL. 76. NO. 1 characters which were used to separate these two species were the specialized thoracic hooks and the abdominal papillae. Curtis documented the ap- pearance of first hooks, then papillae, with the increasing size of the animals. He also observed various intermediate stages with the population. Paraonidae In a revision of the family Paraonidae by Strel- zov(1973), Mcintosh's species of Sco/eco/epzWes (?) jeffreysii was shown to be an indeterminable Aricidea sp. The records of A . jeffreysii from New England (Pettibone 1963a) and from the Chesa- peake Bay (Wass 1972) were referred to A. catherinae Laubier by Strelzov (1973:91). The rec- ord by Day ( 1973) of A. cerruti (not Laubier) from North Carolina should also be referred to A. catherinae. Strelzov (1973:108) also has referred Cirrophorus lyriformis (Annekova) to C. bran- chiatus Ehlers. The species collected in our mid- shelf collection thus was referred to C. bran- chiatus. Both species were recorded by Day ( 1973) from North Carolina. These specimens probably require further examination. Sabellidae Banse (1970, 1972) revised the generic descrip- tions of both Chone spp. and Euchone spp. em- phasizing the branchial crown, setae, and anterior abdominal segments (Euchone). There were many specimens of Euchone spp. and Chone spp. on the continental shelf off Dela- ware. We experienced difficulty in distinguishing the species because many of our specimens were juvenile forms. Our specimens of Euchone spp. appear to have more variability than those re- ported by Banse. In addition, many specimens were damaged or lacked branchial crowns so the number of radioles and the palmate membrane could not be observed. The specimens of Euchone compared most favorably with E. incolor and E. elegans, and the specimens of Chone spp. were most like C. duneri. ZOOGEOGRAPHY Some 125 species of polychaetes (and 8 other species identified only to genus) were collected in the Delaware Bay area. Based on the literature, 116 species have been collected in areas off New England (Table 2, column 1). Sixty-seven species were cited from Chesapeake Bay (Wass 1972; Table 2 , column 3 ). The number of species common to the Chesapeake and Delaware Bay areas is lower than expected, considering their proximity. This was mainly because many of the offshore species encountered in our work were not included in Wass's list. However, work in progress on the mid-Atlantic shelf is expected to change this (D. Boesch, pers. commun.). Ninety-one of the species were common to North Carolina (Hartman 1945; Day 1973; Gardiner 1975; Table 2, column 4). Examination of the local species revealed that for 11 of them, this was the southern extent of their range; i.e., they were reported for New En- gland, but not from Chesapeake Bay or North Carolina (Table 2). Only three species were found to be at the northern limit of their range in the Delaware Bay area, having been found in Chesapeake Bay or North Carolina, but not New England. It appears that the polychaete fauna from the Delaware Bay area is more closely re- lated to the northern than the southern fauna. Two of the species with their northern range in this area, Prionospio cristata and Clymenella mu- cosa., were offshore species. The probability of lar- vae being carried north into the area by the Gulf Stream is great, as Lear and Pesch^^ have shown the intrusion of this water from offshore during the winter and summer months. Data from Hartman (1965) and Hartman and Fauchald (1971) showed that 14 species, which were collected in depths >200 m, were also found in our samples (Table 2, column 2). Seven of these species were recorded only in our offshore samples (J). The remaining seven species, Brania clavata, Paradoneis lyra, Lumbrineris fragilis, Ampharete arctica, Heteromastus filiformis, Chaetozone setosa, and Glycera americana, were also found in the estuary. It was interesting to note that of these seven species, H. filiformis and C. setosa belong to particularly difficult families taxonomically. In our work, H. filiformis was found in salinities as low as 5%o. The species was reported in depths of >1,000 m by Hartman (1965) and Hartman and Fauchald ( 1971). Lumbrineris fragilis, L. latrielli, Aricidea suecica, Prionospio steenstrupi, and Exogone dispar are other species given wide dis- tributions in the literature (M. Pettibone pers. commun.). The distribution of species over such a i^Lear, D. W., and G. G. Pesch. 1975. Effects of ocean disposal activities on the mid-continental shelf environment off Dela- ware and Maryland. EPA Reg. Ill Rep., 78 p. 222 KINNER and MAURER: POLYCHAETOUS ANNELIDS OF DELAWARE BAY wide salinity and depth range appears to be highly doubtful and emphasizes the need for more defini- tive taxonomic work in some of the errant, and in particular, the sedentary polychaete families. ACKNOWLEDGMENTS We thank Wayne Leathern for his help with some of the polychaete identifications, Jeff Tinsman for his invaluable assistance in the field, and Tom White for his help in sample collection. April Morris was extremely helpful in the prep- aration of this paper. This manuscript benefited greatly from numerous suggestions and improve- ments by Roland Wigley, Kristian Fauchald, David Dean, Daniel Dauer, and Meredith Jones. To Marian Pettibone we owe a special thanks for her helpful criticism and her large contribution to the taxonomic portions of this paper. Her knowl- edge and resources greatly improved the content of the paper. LITERATURE CITED Banse, K. 1970. The small species ofEuchone Malmgren (Sabellidae, Polychaeta). Proc. Biol, Soc. Wash. 83:387-408. 1972. Redescription of some species ofChone Kroyer and Euchone Malmgren, and three new species (Sabellidae, Polychaeta). Fish. Bull., U.S. 70:459-495. BOSENCE, D. W. J. 1973. Recent serpulid reefs, Connemara, Eire. Nature (Lond.) 242:40-41. BRENUM, G. 1976. A comparative study of benthic communities of dredged lagoons, tidal creeks, and areas of the open bay in Little Assawoman, Indian River, and Rehoboth Bays, Delaware. M.S. Thesis, Univ. Delaware, Lewes, 174 p. CURTIS, L. A. 1975. Distribution of Sabellaria vulgaris Verrill (Poly- chaeta: Sabellariidae) on a sandflat in Delaware Bay. Chesapeake Sci. 16:14-19. CURTIS, M. A. 1970. Synonymy of the polychaete Scoloplos acutus with S. armiger. J. Fish. Res. Board Can. 26:3279-3282. Dauer, D. M., and J. L. Simon. 1975. Lateral or along-shore distribution of the pwlychae- tous annelids of an intertidal, sandy habitat. Mar. Biol. (Berl.) 31:363-370. 1976a. Repopulation of the polychaete fauna of an interti- dal habitat following natural defaunation: Species equilibrium. Oecologia 22:99-117. 1976b. Habitat expansion among polychaetous annelids repopulating a defaunated marine habitat. Mar. Biol. (Berl.) 37:169-177. DAY, J. H, 1973. New Polychaeta from Beaufort, with a key to all species recorded from North Carolina. U.S. Dep. Com- mer., NOAA Tech. Rep. NMFS CIRC-375, 140 p. Day, J. H., J. G. FIELD, AND M. P. Montgomery. 1971. The use of numerical methods to determine the dis- tribution of benthic fauna across the continental shelf of North Carolina. J. Anim. Ecol. 40:93-125. DEAN, D., AND H. H. HASKIN. 1964. Benthic repopulation of the Raritan River estuary following pollution abatement. Limnol. Oceanogr. 9:551-563. Gardiner, S. L. 1975. Errant polychaete annelids from North Caroli- na. J. Elisha Mitchell Sci. Soc. 91:77-220. Garwood, E. J. 1931. The Tuedian beds of northern Cumberland and Rox- burghshire east of the Liddel Water. Q. J. Geol. Soc. Lond. 87:97-143. GOSNER, K. L. 1971. Guide to identification of marine and estuarine in- vertebrates, Cape Hatteras to the Bay of Fundy. Wiley- Interscience, N.Y., 693 p. HARTMAN, O. 1945. The marine annelids of North Carolina. Duke Univ. Mar. Stn. Bull. 2, 51 p. 1947. Polychaetous annelids. Pt. VIL Capitellidae. Al- lan Hancock Pac. Exped. 10:391-481. 1951. The littoral marine annelids of the Gulf of Mexico. Publ. Inst. Mar. Sci. 2:7-124. 1965. Deep-water benthic polychaetous annelids off New England to Bermuda and other North Atlantic ar- eas. Allan Hancock Found. Publ., Occas. Pap. 28, 378 p. Hartman, O., and K. Fauchald. 1971. Deep-water benthic polychaetous annelids off New England to Bermuda and other North Atlantic areas. Part II. Allan Hancock Monogr. Mar. Biol. 6, 327 p. HARTMANN-SCHRCDDER, G. 1962. Zweiter Beitrag zur Polychaetenfauna von Peru. Kiel. Meeresforsch. 18:109-147. HOBSON, K. D. 1971. Polychaeta new to New England, with additions to the description of Aberranta enigmatica Hartman. Proc. Biol. Soc. Wash. 84:245-252. Jones, m. l. 1963. Four new species of Magelona (Annelida, Poly- chaeta) and a redescription of Magelona longicornis Johnson. Am. Mus. Novit. 2164, 31 p. JUMARS, P. A. 1974. A generic revision of the Dorvilleidae (Polychaeta), with six new species from the deep North Pacific. Zool. J. Linn. Soc. 54:101-135. KINNER, P., D. MAURER, AND W. LEATHEM. 1974. Benthic invertebrates in Delaware Bay: Animal- sediment associations of the dominant species. Int. Rev. Ges. Hydrobiol. 59:685-701. LEIDY, J. 1855. Contributions towards a knowledge of the marine invertebrate fauna of the coasts of Rhode Island and New Jersey. J. Acad. Nat, Sci. Phila, 3:135-152, Light, W, j. 1974. Occurrence of the Atlantic maldanid Asychis elon- gata (Annelida, Polychaeta) in San Francisco Bay, with comments on its synonymy. Proc. Biol. Soc. Wash. 87:175-183. Logan, D, T., and D, Maurer. 1975. Diversity of marine invertebrates in a thermal effluent. J. Water Pollut. Control Fed. 47:515-523. 223 FISHERY BULLETIN: VOL. 76, NO. 1 LUCUS, J. A. W., AND L. B. HOLTHUIS. 1975. On the identity and nomenclature of "Pectinaria belgica (Pallas, 1766)" (Polychaeta, Amphic- tenidae). Zool. Meded. 49:85-90. MANGUM, C. p. 1962. Studies on speciation in maldanidpolychaetes of the North American Atlantic coast. I. A taxonomic revision of three species of the subfamily Euclymeninae. Postilla 65, 12 p. Maurer, D. In press. Estuarine benthic invertebrates of Indian River and Rehoboth Bays, Delaware. Int. Rev. Ges. Hydrobiol. Maurer, D., P. Kinner, W. Leathem, and L. Watling. 1976. Benthic faunal assemblages off the Delmarva Peninsula. Estuarine Coastal Mar. Sci. 4:163-177. Maurer, D., and L. Watling. 1973. Studies on the oyster community in Delaware: The effects of the estuarine environment on the associated fauna. Int. Rev. Ges. Hydrobiol. 58:161-201. Maurer, D., L. Watling, and R. Keck. 1971. The Delaware oyster industry: A reality? Trans. Am. Fish. Soc. 100:100-111. NICHOLS, F. H. 1970. Benthic polychaete assemblages and their relation- ship to the sediment in Port Madison, Washington. Mar. Biol. (Berl.) 6:48-57. PETTIBONE, M. H. 1953. A new species of polychaete worm of the family Am- pharetidae from Massachusetts. J. Wash. Acad. Sci. 43:384-386. 1961. New species of polychaete worms from the Atlantic Ocean, with a revision of the Dorvilleidae. Proc. Biol. Soc. Wash. 74:167-185. 1962. New species of polychaete worms (Spionidae: Spiophanes ) from the east and west coast of North Ameri- ca. Proc. Biol. Soc. Wash. 75:77-87. 1963a. Marine polychaete worms of the New England re- gion. Part I. Families Aphroditidae through Troch- ochaetidae. Bull. U.S. Natl. Mus. 227:1-356. 1963b. Revision of some genera of polychaete worms of the family Spionidae, including the description of a new species of Scolelepis. Proc. Biol. Soc. Wash. 76:89- 103. 1965. Two new species of Aricidea (Polychaeta, Paraonidael from Virginia and Florida, and redescription of Aricidea fragilis Webster. Proc. Biol. Soc. Wash. 78:127-139. 1966. Revision of the Pilargidae (Annelida: Polychaeta), including descriptions of new species, and redescription of the pelagic Podarmus ploa Chamberlin (Poly- noidae). Proc. U.S. Natl. Mus. 3525:155-207. 1970a. Two new genera of Sigalionidae (Polychae- ta). Proc. Biol. Soc. Wash. 83:365-386. 1970b. Revision of some species referred to Leanira Kin- berg (Polychaeta:Sigalionidae). Smithson. Contrib. Zool. 53, 25 p. 1971. Revision of some species referred to Leptonereis, Nicon, and Laeonereis (Polychaeta: Nereididae). Smithson. Contrib. Zool. 104, 53 p. 1977. The synonomy and distribution of the estuarine Hypaniola florida (Hartman) from the east coast of the United States (Polychaeta:Ampharetidae). Proc. Biol. Soc. Wash. 90:205-208. REISH, D. J. 1959. Part III. Benthic polychaetous annelids. In J. L. Barnard and D. J. Reish (editors). Ecology of Amphipoda and Polychaeta of Newport Bay, CaUfomia, p. 70-106. Allan Hancock Found. Publ., Occas. Pap. 21. 1963. A quantitative study of the benthic polychaetous annelids of Bahia de San Quintin, Baja California. Pac. Nat. 3:399-436. Sanders, H. L. 1958. Benthic studies in Buzzards Bay. I. Animal- sediment relationships. Limnol. Oceanogr. 3:245-258. Steimle, F. W., Jr., and R. B. Stone. 1973. Abundance and distribution of inshore benthic fauna off southwestern Long Island, N.Y. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-673, 50 p. Strelzov, V. E. 1973. Polychaete worms of the Family Paraonidae Cer- ruti, 1909 (Polychaeta, Sedentaria). [In Russ.] Akad. Nauk SSSR Murmanskii Morsk Biol. Inst., 170 p. WASS, M. L. (editor). 1972. A checklist of the biota of lower Chesapeake Bay. Va. Inst. Mar. Sci., Spec. Sci. Rep. 65, 290 p. Watling, L. 1975. Analysis of structural variations in a shallow es- tuarine deposit-feeding community. J. Exp. Mar. Biol. Ecol. 19:275-313. Watling, L., W. Leathem, P. Kinner, C. Wethe, and D. Maurer. 1974. Evaluation of sludge dumping off Delaware Bay. Mar. Pollut. Bull. 5:39-42. Webster, H. E. 1880. The Annelids Chaetopoda of New Jersey. Annu. Rep. N.Y. State Mus. Nat. Hist. 32:101-128. 1886. The Annelids Chaetopoda of New Jersey. Annu, Rep. N.Y. State Mus. Nat. Hist. 39:128-159. WEIL, C. 1975. Dynamics of sedimentary fades in Delaware Bay. Ph.D. Thesis, University of Delaware, Lewes, 300 p. WELLS, H. W. 1970. Sabellaria reef masses in Delaware Bay. Chesapeake Sci. 11:258-260. WELLS, H. W., AND I. E. GRAY. 1964. Polychaetous annelids of the Cape Hatteras area. J. Elisha Mitchell Sci. Soc. 80:70-78. ZOTTOLI, R. A. 1974. Reproduction and larval development of the am- pharetid polychaete Amphicteis floridus. Trans. Am. Microsc. Soc. 93:78-89. ZOTTOLI, R. A., AND M. R. CARRIKER. 1974. External release of protease by stationary burrow- dwelling polychaetes. J. Mar. Res. 32:331-342. 224 TROPHIC ONTOGENY OF THE LEOPARD SEAROBIN, PRIONOTUS SCITULUS (PISCES: TRIGLIDAE) Stephen T. Ross' ABSTRACT Ontogenetic feeding changes of the leopard searobin, Prionotus scitulus, from Tampa Bay, Fla., showed a shift from planktonic and epifaunal prey in small fish to infaunal prey in larger fish. Smaller fish utilized larval crustaceans, natantians, brachyurans, cumaceans, copepods, and gammarid am- phipods while larger fish showed increasing reliance on the lancelet, Branchiostoma floridae. Biomass and linear dimensions of prey increased exponentially with fish size for larger fish, but were relatively constant for small fish. Relative prey biomass was lowest for intermediate-sized P. scitulus (65-95 mm) and increased for both large and small predators so that small individuals were most similar to very large fish in terms of relative prey size. The switch to larger prey was preceded by rapid increases in mouth size and intestinal length, and was followed by attainment of minimum reproductive size and greater body weight per unit length. Spatial and trophic partitioning appear quite efficient in reducing potential intraspecific competi- tion. Our present understanding of energy resource partitioning among metazoans is based primarily on food analyses. However, the study of trophic relationships among fishes is frequently compli- cated by indeterminate gro^vth and the cooccur- rence of several size classes of a species at a single locality. A significant degree of prey variability of fishes may be due to size related changes. For instance, Darnell (1958) and Carr and Adams (1973) de- monstrated changes in food habits with increasing size for numerous juvenile marine fishes, and Northcote (1954), Ivlev (1961), Keast and Webb (1966), Wong and Ward (1972), and others have shown a close relationship between morphology (in particular mouth size and shape) and prey kind or size. Such results indicate that inter- and in- traspecific partitioning of energy resources in fish biofacies vary with fish size. This study examines ontogenetic changes in trophic biology of the leopard searobin, Prionotus scitulus Jordan and Gilbert, a common nearshore benthic fish in the eastern Gulf of Mexico. Mor- phological and developmental attributes of jaw size, intestinal length, growth, reproduction, and distribution are evaluated in relationship to trophic changes and to intraspecific resource par- titioning. MATERIALS AND METHODS I collected P. scitulus from three locations in Tampa Bay, Fla. (Figure 1). Numbers offish col- lected and inclusive dates for each station were Station 1, 489 specimens, July 1972-July 1973 Station 2, 838 specimens, August 1972-July 1973 Station 3, 690 specimens, April 1972-July 1973. I examined stomachs from 650 specimens of P. scitulus from Station 3, collected monthly from April 1972 to May 1973. I also identified stomach contents offish from August 1972 collections from GULF of MEXICO 28' 50 40 30 83 • 50 40 30' 20 'Department of Biology, University of Southern Mississippi, Southern Station Box 18, Hattiesburg, MS 39401. Manuscript accepted July 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. Figure l. — Collection localities of Prionotus scitulus in Tampa Bay, Fla., 1972-73. 225 FISHERY BULLETIN: VOL 76. NO. 1 Station 2(N ^ 22) and November and July collec- tions from Station 1 [N = 122). A total of 469 stomachs (72%) from all stations contained food items. April and May collections at Station 3 were made during the day; all other collections were from 1 to 5 h after sunset which was near the end of the greatest diel feeding activity. Ross (1977) demonstrated that searobins from the West Florida Shelf, including P. scitulus, had their greatest feeding activity during the day, but re- tained full stomachs through midnight. Collection depths averaged 5, 5, and 7 m, respec- tively, for Stations 1-3. Sampling gear was a 3.6-m otter trawl with 2.5-cm stretched mesh and a 0.5-cm cod end liner. Upon capture I injected all specimens intraperitoneally with \Q'7( Formalin. ^ Fish were fixed for 2 wk in 10'7( Formalin and then washed and transferred to 40% isopropanol for storage. I sorted prey by taxa from each 10-mm size class offish and measured a random sample (/?ss25) of each prey kind to the nearest 0.1 mm along the axis of greatest dimension. The level of prey iden- tification used in comparisons of size groups was the lowest taxon which was regularly identifiable for each prey kind. Since polychaetes were gener- ally fragmented, they were not measured. Mean number of prey per fish was based only on fish which contained food items. I used a volume displacement technique to mea- sure food items >0.05 cm^ and a squash technique, modified from Hellawell and Abel (1971), to mea- sure volume of food items <0.05 cm^ (Ross 1974). To establish minimum sample sizes for description of the ration I used the criterion t, obtained by plotting cumulative trophic diversity {H ), ) against cumulative stomachs examined ik). Actual num- bers of stomachs ik) varied between samples but had a lower limit of 17. The value of ^ was greater when specimens varied more in date or location of capture. Trophic diversity was determined by the Brillouin information function {H) according to Pielou (1966) and Hurtubia (1973). A horizontal asymptote, beginning at t, indicated a sufficient sample size so that examination of stomachs in excess of t would not yield an increase in trophic diversity. To compare trophic differences of size groups of P. scitulus I used an unweighted pair group, arithmetic average (UPGMA) cluster analysis ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 226 (Sneath and Sokal 1973) based on a Czechanowski similarity matrix (Bray and Curtis 1957). All linear regressions were based on the Berkson case of a model I regression (Sokal and Rohlf 1969). All fish lengths reported are standard length (SL), measured to the nearest 0.1 mm. Mouth width was measured externally between the pos- terior maxillary processes with the mouth fully closed. Internal mouth width was not routinely measured because of difficulty in working with preserved specimens. However, there was no dif- ference between external mouth width with the mouth fully closed and internal mouth width with the mouth fully opened on 36 specimens favorable to such a comparison (two-tailed paired t = 1.88; P>0.05). Measurement of mouth length followed Hubbs and Lagler (1958). To measure intestinal length I cut the hindgut distally at the anus and freed the intestine from the investing mesentery. Length (to the nearest millimeter) was measured from the stomach with the intestine fully extended, but not stretched. Wet weights of P. scitulus were taken to the nearest 0. 1 g after removing stomach contents and blotting the specimens with absorbent paper. Ovaries and testes were removed, blotted, and weighed to the nearest 0.001 g. To compare levels of gonadal activity I used a gonadosomatic index (GSI = (gonad weight/somatic weight) x 100). RESULTS Food Habits The dominant prey of P. scitulus based on per- cent occurrence and percent volume was the lance- let, Branchiostoma floridae, which composed 61% of the food volume and occurred in 60% of the fish examined (Table 1). Numerically, cumaceans were dominant, making up 40% of the total number of prey. On the basis of percent number, volume, and occurrence, the ration of P. scitulus was composed primarily of lancelets, polychaetes, natantians, brachyurans, gammarid amphipods, cumaceans, pelecypods, copepods, and larval crus- taceans. Ninety percent of the number of prey items and volume of prey items were accounted for by 6 and 7, respectively, of the 22 major food categories. I examined seasonal feeding patterns of P. sci- tulus from Station 3, using fish 100 mm or larger to eliminate effects of fish size. Branchiostoma floridae occurred in over 50% of the fish in 8 of the ROSS: TROPHIC ONTOGENY OK LEOPARD SEAROBIN Table l. — Food items utilized by Prionotus scitulus collected between April 1972 and May 1973 at three stations in Tampa Bay, Fla. Based on 469 specimens containing food items. Percent Number Volume Percent Number Volu me Prey category occurrence no. % cm^ % Prey category occurrence no. % cm^ 7o Teleostei; Penaeidae 8.4 99 0.50 2.42 3.68 Sciaenidae 0.2 1 0.01 006 0.09 Lucifer faxoni 1.3 13 0.07 0.02 0.03 Prionotus scitulus 0.4 2 001 0.15 0.23 Unidentified sfirimp 17.4 111 060 1.18 1.80 All fishes 2.3 11 006 081 1 24 Natantian larvae 0.4 2 0.01 (') Amphioxi: All stirimps 42.4 304 1.70 5.12 7.79 Branchiostoma flohdae 595 1,171 650 39.77 60.49 Ampfiipoda: Hemichordata: Gammaridea 56.1 4.615 25.6 2.44 371 Enteropneusta 0.6 3 0.01 002 0,03 Caprellidea 3.9 67 0.04 0.02 0.03 Echinodermata: Isopoda 13.1 87 0.50 0.19 0.29 Ophlurae 0.4 2 0.01 006 0.09 Cumacea 38.3 7,287 40.40 1.45 221 Brachyura: Mysidacea 7.7 119 0.70 0.14 0.21 Portunidae 3.0 25 001 0.40 0.61 Leptostraca: Xanthidae 3.6 42 0.20 1.27 1 93 Nebalia 0.8 4 0.02 0.08 012 Grapsidae 0.9 35 0.20 0.58 0.88 Cope pod a 12 8 2,882 16.00 0.12 0.18 Pinnotheridae 7.4 127 070 0.75 1.14 Ostracoda 88 83 050 0.07 0.11 Oxyrhyncha 0.2 1 001 0.01 002 Unidentified Crustacea 5.1 363 2.00 0.03 0.05 Unidentified crabs 13.9 138 0.80 1,56 237 Acarina: Bractiyuran megalops 8.4 140 080 Oil 0.17 Hydracarina 0.4 2 0.01 (') All crabs 44.6 507 2 82 4.70 7.13 Pycnogonida 0.2 1 0.01 (') Anomura: Annelida: Euceramus praelongus 94 71 040 0.58 088 Polycfiaeta 36.2 {') 9.11 13 86 Pagurldae 0,4 2 0.01 0,12 0.18 fvlullusca: Natantia: Pelecypoda 22.1 356 2.00 0.78 1,19 Leptochela serratorbita 4.9 45 0.20 0,79 1.21 Gastropoda 4.9 52 0.30 0.13 0,20 Palaemonldae 2.4 11 0.06 0.05 0.08 Bractiiopoda 0.2 1 0.01 0.01 0,02 Alpfieidae 0.6 11 006 0.21 0.32 Cnidaria 0.9 6 0,03 0.02 0.03 Processidae 1.7 8 004 0.43 0.65 Totals 17,992 65.75 Hippolytidae 0.4 4 0.02 0.02 003 'Only a trace amount of food present, ^An accurate count of individuals was not possible. 13 mo examined, dropping between 30 and 40^^ in September, January, and May. Number, volume, and percent occurrence for B. floridae all showed major peaks in utilization between June and Au- gust, and October and December 1972. Polychaetes were irregular in percent occurrence, but the data suggest a peak in spring and summer, while natantians and brachyurans showed in- creases in percent occurrence in the spring and fall. Amphipods, cumaceans, mysids, and pelecypods showed strong spring peaks in impor- tance. Nine size groups of P. scitulus (21-40, 41-60, 61-80, 81-90, 91-100, 101-110, 111-120, 121-130, 13 1-140) reached stabilized horizontal asymptotes of cumulative trophic diversity versus cumulative stomachs examined. The analyses of size changes in feeding are based on these groups. The percent occurrence of lancelets and polychaetes increased with increasing fish size, while gammarid amphipods decreased (Table 2). Brachyurans, cumaceans, copepods, larval crusta- ceans, pelecypods, and ostracods increased in per- cent occurrence for searobins up to 80-100 mm. Table 2. — Percentage of prey occurrence for size groups (millimeters standard length) of Prionotus scitulus from Tampa Bay, Fla., 1972-73. Only prey categories with an overall occurrence of 1% or greater were included. Prey category 21-40 41-60 61-80 81-90 91-100 101-110 111-120 121-130 131-140 iJTeleostei 0 5.3 4.6 5.9 0 3.7 2.9 0 3.7 Branchiostoma floridae 4.0 21,1 22.7 29.4 630 69.1 64.5 70.6 70.3 Brachyura 80 201 45.5 64.7 25.9 37.0 30.4 29.4 22.2 Natantia 20.0 263 136 17.6 11.1 29.6 34.1 32.4 48.2 Anomura 4,0 5,3 0 5.9 11.1 11.1 10.9 9.8 7.4 Gammaridea 880 895 54.5 52.9 48.1 51.9 59.4 48.0 48.2 Caprellidea 8.0 0 4.6 0 0 2.5 8.7 2.0 3.7 Isopoda 0 15 8 9.1 5.9 18.5 8.6 17.4 8.8 25.9 Cumacea 360 78.9 72.7 52.9 74.1 51.9 25.4 21.6 25.9 Mysidacea 80 10.3 0 0 25.9 17.3 5.8 1.0 7.4 Copepoda 8.0 26.3 63.6 41.2 37.0 9.9 3.6 4.9 7.4 Ostracoda 16.0 20.1 22.7 41.2 14.8 7.4 3.6 2.9 7.4 Crustacean larvae 40 21.1 54.6 23.5 3.7 2.5 0 0 0 Polychaeta 0 5.3 9.1 23.5 25.9 30.9 47.1 44.1 62.9 Pelecypoda 4.0 0 45.5 29.4 25.9 21.0 21.7 23.5 22.2 Gastropoda 0 0 18.2 11.8 7.4 2.5 5.8 2.0 7.4 No. of fish examined 25 19 22 25 27 81 138 102 27 227 FISHERY BULLETIN: VOL 76, NO. 1 and then decreased for larger fish. Other prey categories either did not show regular trends or remained relatively constant in occurrence be- tween size classes. The percent number of prey showed similar trends with increasing fish size. Crustacean larvae, copepods, gammarid am- phipods, and cumaceans were of greater impor- tance to small fish, while larger fish ( 100-140 mm) utilized more lancelets and pelecypods. The volumetric importance o{ Branchiostoma to the 41- to 60-mm size group resulted from one fish capturing a single large lancelet. Volumetrically, the diet of P. scitulus 80 mm and larger was domi- nated by lancelets and polychaetes, while cuma- ceans, copepods, and natantians (especially larval forms) were of greater importance to small fish (Figure 2). Brachyurans showed a more uniform pattern of distribution among size groups. uu- Misc 90- 80- 70- Natontio 60- 50- 40- Cumacea 30- Gammaridea 20- 10- Mysidacea Crust Larwoe Misc Teleoslei Bronchiosfoma Brochyuro Notcntia Cumacea Copepoda Gammoridea Misc Teleostei Branchiostoma Brochyuro Notontia Cumacea Copepodo" Gammoridea Teleostei Bronchiostomo Brochyuro Natontia Cumoceo Copepoda Gammoridea Misc Branchiostomo Brochyuro Notontio Cumoceo Gommandeo" Misc Polychoeta Teieostei BronchiosToma Brochyuro Notontia Cumoceo Gommorideo Misc Polychoeta Branchiostomo Brochyuro Notontia Gommorideo Misc Bivolvio Polychoeta Bronchiostofna Brochyuro Notontia GommoriBeo^ Misc Bivolvio Polychoeta Bronchiostomo Brochyuro Natontio 21-40 41-60 61-80 BI-90 91-100 lOI-IIO 111-120 121-130 131-140 (25) (19) (22) (25) (27) (81) (138) (102) (27) Fish Length (mm) Figure 2. — Changes in the percent volume of major prey categories for size classes ofPrionotus scitulus, Tampa Bay, Fla., 1972-73. Percent Similarity 0 10 20 30 40 50 60 70 80 90 100 r Fish Length (mm) 131 -140 121- 130 101- ■ no III- 120 91- 100 81- 90 41- -60 61 -80 21 -40 Trophic relationships among size groups were summarized by cluster analysis based on the per- cent occurrence of prey (Figure 3). Fish smaller than 81-90 mm and larger than 91-100 mm formed two major divisions, linking at 779^ similarity. The lower similarity between size classes of small- er searobins compared with larger size classes is indicative of the more rapid changes in trophic ontogeny occurring between small individuals. Figure 3. — Cluster analysis (UPGMA; unweighted pair group, arithmetic average) of prey similarity between size classes of Prionotus scitulus, Tampa Bay, Fla., 1972-73. Similarity was determined from percent occurrence of prey categories. 228 ROSS: TROPHIC ONTOGENY OF LEOPARD SEAROBIN The total amount of food ingested, as shown by the mean volume of stomach contents, increased rapidly with increasing fish size; log transformed values of total prey volume varied linearly with fish size over most size classes (Figure 4). The total number of prey per fish also increased rapidly with increasing fish size up to the 60- to 80-mm size class, but then declined markedly for larger size groups (Figure 4). The decline in number of prey ingested occurred somewhat prior to a detectable increase in mean prey size (cf. Figure 5). Searobins smaller than the 90- to 100-mm size group showed an asymptotic relationship of fish length and linear prey size, while prey sizes increased rapidly over the larger size groups. Since linear prey mea- surements may be misleading, I also examined the average volume (cubic centimeters) of prey items eaten by size classes of P. scitulus. Prey volume was calculated from the total sorted food volume from each 10- or 20-mm size class, divided by the total prey number for each size class. Mean prey volume did not increase over small size classes of searobins, but at 90-100 mm it initiated a rapid increase (Figure 6). Consequently, the rapid rise in total stomach volume of the leopard searobin occurred initally through the capture of increas- ing numbers of small prey, followed (after 90-100 mm) by the capture of fewer, but progressively larger, prey. Relative prey biomas's (mean prey volume/mean wet weight) was initially very high but then de- 500 . (E 300 : ^200 I 100 . u 50 a 30 z 20 < ^ 10 1 102. Meon Stomach Volume „, nn'O' 81 ;j"_.4' 27 '^ 27/*' 22,,-l-f 19 ,' 50 30 20 10 . 05 03 02 01 005 003 002 .001 UJ Z _) o > X o < o 1 — I I < I I I I I — I I I 25 35 45 55 65 75 85 95 105 115 125 135 FISH LENGTH (mm SL) FIGURE 4. — The relationship of mean volume of stomach con- tents (cubic centimeters) and mean prey number (logarithmic scales) to fish length for Prionotus scitulus, Tampa Bay, Fla., 1972-73. The vertical lines indicate 1 SE on either side of the mean, sample sizes are shown above the upper graph. creased with fish size to the 61- to 70-mm size class, followed by an increase for fish larger than the 90- to 100-mm size class (Figure 6). Increases in prey size with increasing predator size might occur through shifts in the utilization of progressively larger prey kinds, or through the 100 60 - 50 - 40 - 30 i25 N > UJ a: Q. z < UJ 20 - 15 - 10 - 5 ■ i¥ il 25 35 — I— 45 — r— 55 — T— 65 75 — r- 85 — I— 95 105 I I 125 135 Figure 5. — Mean prey length versus standard length groups ofPrionotus sci- tulus from Tampa Bay, Fla. Vertical lines are ranges; cross-bars and open rectangles arex ± 2 SE. FISH LENGTH (mm SL) 229 2 OO 8 « 6 ^ 4? 2^ 0 > 8 ui 6 J ? ■* FISH LENGTH (mm SL) Figure 6. — Mean prey volume (cubic centimeters) and relative prey size (x prey volume/x wet fish weight) for size classes of Prionotus scitulus, Tampa Bay, Fla., 1972-73. selection of larger sized individuals within a single prey kind. Only one prey item, B. floridae, exhibited a broad enough size range to meaning- fully test for differences between fish sizes. The mean size of lancelets, however, did increase with increasing fish size (P<0.001) (Figure 7), but the rate of increase was quite low compared with the overall increase in mean prey size (cf. Figure 5). 55- 50 - 'e 45- E -- 40 H B 35-1 0) 30 - >» 25 - o tL 20 - 15- 10 - 5 - 319 150 337 II 69 - {} ^^ ^^ 154 bB 59 /At 1 1 1 1 1 1 — 85 95 105 115 125 135 145 Fish Length (mm) FISHERY BULLETIN: VOL. 76. NO. 1 Morphology and Growth Trophic changes showed a critical size interval between approximately 60 and 100 mm, within which the mean prey number decreased, and after which the mean prey volume, length, and relative volume increased. These trophic changes suggest- ed the presence of certain morphological or de- velopmental correlates, of which I examined jaw size, intestinal length, and growth. Ontogenetic changes in mouth size were ex- pressed by relative jaw width and relative jaw length. Juvenile leopard searobins showed propor- tionately greater mouth widths and lengths com- pared with adults, but plots of both relative jaw length and relative jaw width versus SL showed considerably lower slopes by approximately 75 mm (Figure 8). Proportionate mouth length con- tinued to decrease with increasing fish size for fish >75 mm; however, proportionate mouth width remained constant for fish >75 mm. Mouth size thus increased rapidly with increasing fish size for early juvenile P. scitulus, but by 75 mm the rela- tionship between mouth size and fish length was essentially fixed. Intestinal length increased rapidly between the 45- and 65-mm size classes. Fish <50 mm had mean intestinal lengths of 70% SL, while fish >60 mm had mean intestinal lengths of 102% SL. Log transformed length-weight values of leopard searobins showed an increase in the slope of the regression line between approximately 55 and 75 mm (Figure 9). The fish were divided into two size groups, <75 mm and >75 mm, and sepa- 140 . 135 130 I i 125 J UJ _l ^ 120 . CO "^ 115 . H z UJ o 110 (E UJ ^ 105 100 ^6--, o o Mouth Width /SL i i Mouth Length/SL ^s:_ T- -r I I 1 I I 1 I 30 40 50 60 70 80 90 100 110 120 130 140 FISH LENGTH (mm SL) Figure 7. — The relationship between lengths of the dominant prey, Branchiostoma floridae, and its predator, Prionotus sci- tulus. See Figure 5 for explanation of symbols. 230 Figure 8. — Relative mouth width and relative mouth length versus fish length for Prionotus scitulus, Tampa Bay, Fla. Each data point is based on the mean of 20 individuals. ROSS TROPHIC ONTOGKNY OF l,KOI'ARD SEAROBIN I 10 o o 4. 3 3. X « 2. 1 . 0. -I . 2 Log W = -10.878 » 2.949 Log SL 0 6 75mm 65mml 55mm *^ i Log W= -9.083-2.451 Log SL N=77 SE=0.328 -P ■ I I I I < I I I I 3.0 3.2 3.4 36 3.8 4.0 42 4 4 46 4.8 5.0 5.2 LOGg FISH LENGTH (mm SL) Figure 9. — Length- weight relationships for size classes of Prionotus scitulus, Tampa Bay, Fla. rate length-weight regressions were calculated. The 75-mm size was chosen because of its associa- tion with changes in relative mouth size. The growth data showed that P. scitulus >75 mm were gaining weight much more rapidly than smaller fish, even after allowing for the expected exponen- tial rate of increase by using the log transforma- tion. Reproduction and Distribution Mean female GSFs remained below 0.4 for P. scitulus between 20 and 90 mm and these fish did not contain mature ova. Leopard searobins 100 mm and larger had mean GSI values between 3 and 6 and were sexually mature. Ross (1974, in press) showed that mean female GSI values dur- ing spring to summer spawning were 5 to 10. The GSI values reported here are lower because the fish were combined from all months of the study to avoid possible bias due to differences of spawning times of different size groups. Male searobins showed the same size-related pattern. Spatial separation between immature and ma- ture P. scitulus was quite pronounced (Figure 10). Juvenile searobins consistently had high relative abundances at Station 1 in Old Tampa Bay (5 m deep), while mature fish had high relative abun- dances near the mouth of Tampa Bay at Station 3 (7 m deep). Overlap between immature and ma- ture searobins was greatest during summer and fall 1972 at Station 2 ( 5 m). The percent occurrence of juvenile fish in combined collections was high- est between March and May (60-75%) and lowest between June and November (25-47%). Annual mean salinities varied significantly be- o z < o z Zi CD < > < UJ (T. O 100 80 60 40 20 0 Juveniles •:■>:■:■ .^^xN 58 .^93 June - Aug. 1972 1 Sepl.- Nov. J59 $a:^90 581 m ■•■'■'■'• 19! ..M248 giM24l Dec- Feb. 1973 m v/.v M m 65 March- May 310 a::^:M95 June- July 16 -1M168 $;^^39 I 2 3 TAMPA BAY STATIONS FIGURE 10.— Spatial distribution of juvenile ( <100 mm SL), and adult (>100 mm SL) Prionotus scitulus at three stations in Tampa Bay, Fla., 1972-73. Percent relative abundance is based on each sample site and date; numbers indicate sample sizes. Adults are indicated by hatching; juveniles by cross-hatching. tween stations (P<0.05); respective means for Stations 1-3 were 25.7, 28.1, and 33.2%o. Con- sequently, small leopard searobins were occupy- ing somewhat less saline water. Annual mean temperatures did not vary between stations (range = 13°-32°C). 231 FISHERY BULLETIN: VOL 76, NO. 1 DISCUSSION Ontogenetic changes in prey utilization by P. scitulus showed an early dependence on plank- tonic or epifaunal prey such as crustacean larvae, copepods, mysids, cumaceans, and gammarid am- phipods. Larger P. scitulus (>90 mm) ate more infaunal organisms such as lancelets and polychaetes. Separation by prey kind was greatest at 90 mm which corresponded to the transition size between immature and mature fishes. The greatest percent occurrence of juvenile fish (March-May) coincided with periods of higher utilization of brachyurans, natantians, cumaceans, amphipods, mysids, pelecypods, and polychaetes by adult fish, although lancelets re- mained the dominant prey. Consequently, size dif- ferences in food habits were not biased by seasonal unavailability of certain prey to adults or juve- niles. Also, Ross (1974) demonstrated that changes in food habits with increasing fish size were generally consistent between stations. Other studies on food habits of P. scitulus have indicated that small crustaceans and polychaetes were important prey (Reid 1954; Springer and Woodburn 1960; Ross 1977, in press). Ross (1977, in press) found that P. scitulus from offshore of Tampa Bay utilized principally brachyurans, polychaetes, cumaceans, gammarid amphipods, natantians, and lancelets. Total food consumption showed an accelerating rate of increase with fish length, but initially this occurred through a rapid rise in the number of prey consumed, rather than through an increase in prey size. Prey size did not increase with in- creasing fish size for searobins <90 mm. Although numerous studies have demonstrated positive cor- relations between prey and predator sizes (e.g., Northcote 1954; Hartman 1958; Wong and Ward 1972; Hespenheide 1973), Schoener (1969, 1971) predicted that prey size would decrease with de- creasing predator sizes to a lower horizontal asymptote. Essentially, the energy gained from progressively smaller prey gradually approaches the energy expended in obtaining and digesting prey. Data on prey size-predator size relationships supporting this prediction were reviewed by Schoener (1971), but did not include fishes as examples. Prey size (both length and volume) was posi- tively correlated with fish size for searobins 90 mm and larger. The increase in mean prey size relative to predator size occurred primarily through a 232 progressive shift to different, larger prey taxa, and only secondarily by size selection within a single prey taxon. The transition from numerous small prey to fewer large prey was preceded by rapid growth of jaw size relative to body size and by an increase in intestinal length. Since intestinal absorption may be increased through the development of folds and an increase in length or both (Siankowa 1966), the relative increase in intestinal length of P. scitulus is perhaps a response to increased energy demands of larger fish or to their utilization of larger prey items. Growth in fishes may occur as a series of stanzas which are entered by ecological and physiological size thresholds (Parker and Larkin 1959). Growth stanzas may be recognized by changes in weight- length relationships (Ricker 1975). The shift from small to large prey in P. scitulus was accompanied by a change in the weight-length relationship in- dicating the presence of two growth stanzas. Growth efficiency, measured as weight gained per ration weight per unit time, varies extensively with prey kind (Paloheimo and Dickie 1966). For instance, growth efficiency of trout increased as the ration progressed from hatchery mash to gammarid amphipods to minnows. The two growth stanzas in P. scitulus may thus reflect an increase in the proportion of food energy available for growth as small crustaceans are replaced by larger lancelets and polychaetes in the diet. Relative prey size showed a parabolic relation- ship with fish size. Consequently, small P. scitulus were, in effect, predators of large prey. Prey size distributions have been shown to follow a lognor- mal relationship in various communities (Whit- taker 1952; Schoener and Janzen 1968; Griffiths 1975), so juvenile leopard searobins were utilizing an apparently abundant energy source. However, since mean prey size did not increase with increas- ing fish size for searobins <90 mm, with growth, searobins tended toward being "small" predators due to the continued use of the same-sized prey items. Although prey availability was not moni- tored, P. scitulus between 20 and 90 mm were likely operating as number maximizers (cf. Griffiths 1975). Griffiths presented evidence that juvenile stages of several kinds of vertebrates pass through such a stage during which prey items are utilized in close proportion to their actual occur- rence. Searobins >90 mm showed an increase in rela- tive prey size, thus tending again towards being ROSS: TROPHIC ONTOGENY OF LEOPARD SEAROBIN predators of large prey. The data suggest a switch in feeding strategy to an energy maximizer (cf. Griffiths 1975) in which predators feed in such a manner as to maximize their energy intake. In P. scitulus this is perhaps accomplished by a switch in feeding behavior after achieving a critical size threshold requisite for capturing partially buried infaunal prey. The shift to utilization of large prey occurs slightly before the onset of reproduction. In- creased energy demands, or a decrease in foraging time, brought about by gonadal development and breeding activity or both, might be critical factors in selecting for the change in the feeding strategy of P. scitulus. Mature and immature P. scitulus were effec- tively segregated along both spatial and trophic dimensions in Tampa Bay. Spatial segregation might occur through the ability of juvenile searob- ins to occupy shallower water or to withstand lower salinity, a characteristic of many juvenile marine fishes (Gunter 1961). Trophic overlap in prey kind between immature and mature size groups was closely comparable with trophic over- lap between adult individuals of different species of searobins on the West Florida Shelf (Ross 1977). Consequently, P. scitulus in Tampa Bay were ef- fectively reducing the potential for intraspecific competition. ACKNOWLEDGMENTS This study is based, in part, on a segment of my doctoral dissertation submitted to the University of South Florida. I thank my major professor, J. C. Briggs, and committee members, D. G. Burch, B. C. Cowell, R. W. McDiarmid, and A. J. Meyer- riecks for their help. I thank B. C. Cowell and S. A. Bortone for help- ful comments on an early draft of this paper. The University of Southern Mississippi Ecology Forum contributed many helpful comments to a later draft. I am especially grateful to my wife Yvonne for her help during all phases of this study. The program for cluster analysis was writ- ten by J. G. Field, who is gratefully acknowledged. LITERATURE CITED Bray, J. R., and J. T. Curtis. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27:325-349. Carr, W. E. S., and C. a. Adams. 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 communi- ty. Publ. Inst. Mar. Sci. Univ. Tex. 5:353-416. Griffiths, D. 1975. Prey availability and the food of predators. Ecol- ogy 56:1209-1214. Gunter, G. 1961. Salinity and size in marine fishes. Copeia 1961:234-235. Hartman, G. F. 1958. Mouth size and food size in young rainbow trout, Salmo gairdneri. Copeia 1958:233-234. HELLAWELL, J. M., AND R. ABEL. 1971. A rapid volumetric method for the analysis of the food of fishes. J. Fish Biol. 3:29-37. Hespenheide, H. a. 1973. Ecological inferences from morphological data. Annu. Rev. Ecol. Syst. 4:213-229. HUBBS, C. L., AND K. F. LAGLER. 1958. Fishes of the Great Lakes region. Revised ed. Cranbrook Inst. Sci. Bull. 26, 213 p. HURTUBIA, J. 1973. Trophic diversity measurement in sympatric pred- atory species. Ecology 54:885-890. IVLEV, V. S. 1961. Experimental ecology of the feeding of fishes. (Translated from Russ.) Yale Univ. Press, New Haven, Conn., 302 p. keast, a., and d. Webb. 1967. Mouth and body form relative to feeding ecology in the fish fauna of a small lake, Lake Opinicon, Ontario. J. Fish. Res. Board Can. 23:1845-1874. NORTHCOTE, T. G. 1954. Observations on the comparative ecology of two species of fish, Cottus asper and Cottus rhotheus, in British Columbia. Copeia 1954:25-28. PALOHEIMO, J. E., AND L. M. DICKIE. 1966. Food and growth of fishes. III. Relations among food, body size, and growth efficiency. J. Fish. Res. Board Can. 23:1209-1248. Parker, R. R., and P. A. Larkin. 1959. A concept of growth in fishes. J. Fish. Res. Board Can. 16:721-745. PlELOU, E. C. 1966. The measurement of diversity in different types of biological collections. J. Theor. Biol. 13:131-144. REID, G. K., Jr. 1954. An ecological study of the Gulf of Mexico fishes, in the vicinity of Cedar Key, Florida. Bull. Mar. Sci. Gulf Caribb. 4:1-94. RICKER, W. E. 1975. Computation and interpretation of biological statis- tics offish papulations. Fish. Res. Board Can., Bull. 191, 382 p. Ross, S. T. 1974. Resource partitioning in searobins (Pisces: Trig- lidae) on the west Florida shelf Ph.D. Thesis, Univ. South Florida, Tampa, 205 p. 1977. Patterns of resource partitioning in searobins (Pisces: Triglidae). Copeia 1977:561-571. In press. Searobins (Pisces: Triglidae). Mem. Hourglass Cruises. 233 FISHERY BULLETIN: VOL 76, NO. 1 SCHOENER, T. W. 1969. Models of optimal size for solitary predators. Am. Nat. 103:277-313. 1971. Theory of feeding strategies. Annu. Rev. Ecol. Syst. 2:369-404. SCHOENER, T. W., AND D. H. JANZEN. 1968. Notes on environmental determinants of tropical versus temperate insect size patterns. Am. Nat. 102:207-224. SlANKOWA, L. 1966. The surface area of the intestinal mucosa in bream - Abramis brama (L). Stud. Soc. Sci. Torun., Sect. E (Zool.) 8:1-54. SNEATH, P. H. A., AND R. R. SOKAL. 1973. Numerical taxonomy, the principles and practice of numerical classification. W. H. Freeman and Co., San Franc, 573 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. Springer, V. G., and K. D. Woodburn. I960. An ecological study of the fishes of the Tampa Bay area. Fla. State Board Conserv. Mar. Lab., Prof. Pap. Ser. 1, 104 p. WHITTAKER, R. H. 1952. A study of summer foliage insect communities in the Great Smoky Mountains. Ecol. Monogr. 22:1-44. Wong, B., and F. J. Ward. 1972. Size selection oiDaphnia pulicaria by yellow perch iPerca flavescens) fry in West Blue Lake, Manitoba. J. Fish. Res. Board Can. 29:1761-1764. 234 DESCRIPTION OF LARVAE OF THE HUMPY SHRIMP, PANDALUS GONIURUS, REARED IN SITU IN KACHEMAK BAY, ALASKA Evan Haynes^ ABSTRACT Except for Stage I, identification of larval stages of Pandalus goniurus has not been verified by rearing the larvae from known parentage. Larvae of P. goniurus were reared in situ in Kachemak Bay, Alaska, from the first zoea (Stage I) through the first juvenile stage (Stage VII). Each of the seven stages is described and illustrated. The descriptions are compared with descriptions of larval stages of P. goniurus given by other authors. Studies on the early life history of pandalid shrimp in Alaskan waters were begun in 1972 by the National Marine Fisheries Service with the initial objective of describing pandalid shrimp larvae reared in the laboratory from known par- entage. I have reported on larvae of coonstripe shrimp, Pandalus hypsinotus Brandt, reared in the laboratory (Haynes 1976). In the present re- port I describe and illustrate larvae of humpy shrimp, P. goniurus Stimpson, reared in situ in Kachemak Bay, Alaska. A third report will de- scribe larvae of pink shrimp, P. borealis Krdyer, and compare the larvae of P. borealis with larvae of other local pandalid species, including P. goniurus. MATERIALS AND METHODS The laboratory technique used successfully for rearing larvae of P. hypsinotus (Haynes 1976) proved unsuitable for rearing P. goniurus beyond Stage II. Beginning with Stage III, molting fre- quency and number of larval stages of P. goniurus reared in the laboratory were inconsistent, mor- tality was high, and the larvae of a given stage were not always morphologically identical. Rear- ing P. goniurus in situ reduced mortalities and yielded larvae essentially identical morphologi- cally within each stage. Larvae were reared in situ from the first zoea (Stage I) through the megalopa and first juvenile (Stages VI and VII) in the following manner. Stage 'Northwest and Alaska Fisheries Center Auke Bay Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 155, Auke Bay, AK 99821. I zoeae of known parentage were obtained using the laboratory technique described by Haynes ( 1976). The Stage I zoeae were then transported to sea and placed in 500-ml flasks containing seawa- ter of about 35%o salinity and 4°C obtained from about 6 m depth with a plastic hand pump and hose. Subsurface seawater was used to avoid the lower salinity (about 28%o) of surface waters de- rived from local runoff which, as I had found dur- ing previous rearing studies, adversely affects larval development by resulting in delayed molt- ing and variable numbers of stages. One larva was placed in each flask. The mouths of the flasks were then covered with nylon screening of #0 mesh (0.571 mm); the flasks were placed in holding con- tainers and suspended upright at 15-20 m depth in water about 40 m deep. The #0 mesh size allowed plankton to collect in the flasks for food but pre- vented the larvae from escaping. Each flask was numbered and a record kept of the molting history of each larva in each flask. Flasks were checked at least every other day for cast skins and refilled with fresh subsurface seawater. When a larva molted, the cast skin was removed from the flask with a large-bore pipette and preserved in 5% formaldehyde for subsequent examination ashore. Identification of larval sequence and stage was verified using larvae obtained from plankton with a net of #0 mesh towed near the bottom at about 2 kn in water 60 m deep. The plankton sample was immediately placed in a glass receptacle contain- ing several liters of subsurface seawater. Stage I zoeae of P. goniurus were removed from the glass receptacle using a large-bore pipette, placed in 500-ml flasks, one zoea to a flask, and reared to Manuscript accepted Julv 1977. FISHERY BULLETIN: VOL. 76, NO. 1, 1978. 235- FISHERY BULLETIN: VOL. 76, NO. 1 postlarvae in the same manner as the Stage I zoeae obtained in the laboratory. To verify the validity of the sequence of the larval stages obtained from flasks, larvae of each stage were obtained from plankton and reared in flasks in the same manner for one molt. They were then removed from the flasks along with their cast skins, preserved, and replaced with a larva of the same stage. Thus, a Stage II zoea that had molted to Stage III in a flask was replaced with a Stage III zoea from plankton, the Stage III zoea being re- placed in like manner when it had molted to Stage 236 HAYNES: PANDALUS GONIURUS LARVAE IV. This procedure was done for each stage in- cluding the megalopa ( Stage VI). In addition to the larvae and cast skins obtained from rearing in flasks, molting sequence and stage were verified by monitoring the sequence of larval stages from local collections, obtained at least weekly in areas where larvae were abundant, and by examining larvae caught while in the process of molting. Only those morphological characteristics useful for readily identifying each stage are given. Figure l. — Stage I zoea of Pandalus goniurus: A, whole animal; B, antennule; C, antenna; D, mandibles (right and left); E, maxillule; F, maxilla; G, first maxilliped; H, second maxilliped; I, third maxilliped; J, second pereopod; K, telson. 237 FISHERY BULLETIN: VOL 76, NO. 1 Stages I and II are described in greatest detail because these stages are the most difficult to iden- tify. Terminology, methods of measuring, techniques of illustration, and nomenclature of gills and appendages follow Haynes ( 1976). Com- parison of larvae from plankton with cast skins from flasks was facilitated by first clearing the larvae in 10*?^ KOH. For each pair of appendages the left member is figured except for the mandi- bles, which are drawn in pairs and figured from the right side. For clarity, setules on setae are usually omitted but spinulose setae are shown. STAGE I ZOEA Total length of Stage I (Figure lA) 4.0 mm (range 3.7-4.2 mm; 10 specimens). Live specimens translucent with isolated areas of color: mouth- parts orange with a bright yellow chromatophore at base; internal thoracic organs greenish, espe- cially heart area; base of maxillipeds greenish orange; distinct yellow chromatophore at anus. Rostrum slender, spiniform, without teeth, about one-third length of carapace, and projects horizon- tally or slightly downward. Carapace with small, somewhat angular dorsal prominence at base of rostrum and a smaller rounded prominence near posterior edge. These two prominences occur in all zoeal stages. Pterygostomian spines present but usually hidden by sessile eyes. Three to four mi- nute spinules along ventral margin of carapace immediately posterior to pterygostomian spine (spinules not shown in Figure lA). These spinules usually occur in all zoeal stages but may vary in number from two to five not only between stages but among individuals within a given stage. ANTENNULE (Figure IB).— First antenna, or antennule, consists of a simple unsegmented tubu- lar basal portion with a heavily plumose seta ter- minally and a distal conical projection bearing four aesthetascs: one long, one short, and two of intermediate length. ANTENNA (Figure IC).— Consists of inner flagellum (endopodite) and outer antennal scale (exopodite). Flagellum unsegmented, slightly shorter than scale, styliform, and tipped by a spinulose spine. Antennal scale distally divided into five joints (the proximal joint incomplete) and fringed with nine heavily plumose setae. Two simple setae occur on outer margin, one terminal and adjacent to plumose setae and the other near 238 base of terminal segments. A small plumose seta usually occurs proximally near lateral margin in all zoeal stages. Protopodite bears spinous seta at base of flagellum but no spine at base of scale. MANDIBLES (Figure ID).— Without palps in all zoeal stages. Incisor process of left mandible bears four teeth in contrast to triserrate incisor process of right mandible. Left mandible bears a movable premolar denticle (lacinia mobilis) whereas right mandible bears two immobile premolar denticles. Truncated molar process of left mandible bears a subterminal tooth that occurs throughout all zoeal stages. MAXILLULE (Figure IE).— First maxilla, or maxillule, bears coxal and basial endites and an endopodite. Proximal lobe (coxopodite) bears stout seta near base, and seven spinulose spines termi- nally. Median lobe (basipodite) bears five stout spinulose spines on terminal margin, two of them especially thick with projecting teeth, and a large setose seta proximally. Endopodite originates from lateral margin of basipodite and bears three terminal and two subterminal setae; two of the setae are especially spinulose. MAXILLA (Figure IF). — Bears platelike exopo- dite ( scaphognathite) with four long plumose setae along distal and outer margins, and one slightly longer and thicker seta at proximal end. Endopo- dite gives indication of four partly fused segments and bears nine large plumose setae. Basipodite bilobed; each lobe bears six setae. Bilobed coxopo- dite bears 15 setae, 4 on distal lobe and 11 on proximal lobe. Four setae, one on each lobe of basipodite and coxopodite, bear a row of little spines along entire length. FIRST MAXILLIPED (Figure IG).— Most heavily setose of natatory appendages. Protopodite not segmented; bears 17-20 setae, several of them especially spinulose. Endopodite distinctly four- segmented; setation formula 4, 2, 1, 3. Exopodite a long slender ramus segmented at base; has two terminal and two lateral natatory setae. Epipodite a single lobe. SECOND MAXILLIPED (Figure IH).— Protopo- dite not segmented; bears nine sparsely plumose setae. Endopodite distinctly four-segmented; seta- tion formula 6, 2, 1, 3. Exopodite with two termi- nal, six lateral natatory setae. No epipodite. HAYNES: PANDALUS GONIURUS LARVAE THIRD MAXILLIPED (Figure II).— Protopodite bears four setae. Endopodite distinctly five-seg- mented; nearly as long as exopodite; setation for- mula 5, 2, 1, 0, 2. Exopodite with 2 terminal, 10 lateral natatory setae. No epipodite. PEREOPODS.— Poorly developed, directed under body somewhat anteriorly. First three pairs biramous (second pereopod shown in Figure IJ), last two pairs uniramous and slightly smaller than pairs 1-3. PLEOPODS.— Not evident. TELSON (Figure IK).— Not segmented from sixth abdominal somite; slightly emarginate dis- tally; bears seven pairs of densely plumose setae. Fourth pair of setae longest, length about one-half width of telson. Minute spinules at base of each seta except possibly last pair. Larger spinules along terminal margin between bases of four inner pairs and on setae themselves. Enclosed uropods visible. No anal spine. STAGE II ZOEA Total length of Stage II (Figure 2 A) 4.9 mm (range 4.5-5.3 mm; 10 specimens). Chromatophore color and pattern essentially identical to Stage I except chromatophores larger and color more pro- nounced, especially in mouth parts. From this stage on, zoeae become increasingly more orange and color pattern is not useful as an aid to specific identification. Rostrum still without teeth but not curved downward as strongly as in Stage I. Carapace has prominent supraorbital spine; an- tennal and pterygostomian spines clearly visible. These spines persist throughout all zoeal stages. Epipodite still not bilobed; pleurobranchiae not yet present. ANTENNULE (Figure 2B).— Three-segmented; bears on terminal margin a large outer and a smaller inner flagellum. Inner flagellum not seg- mented, conical, and bears one long spine termi- nally. Outer flagellum bears two groups of aes- thetascs, one group terminally consisting of seven aesthetascs, two of them larger than remaining five, and a second group of two aesthetascs on inner margin. A small budlike projection (not shown in Figure 2B) originates at base of the two flagella and bears three simple setae. Joint of proximal segment faint and may not be complete; bears about five dorsally projecting small plumose setae. Second segment has one lateral plumose seta and about five dorsally projecting plumose setae ringing terminal margin. Third segment has five lateral plumose setae. ANTENNA (Figure 2C).— Flagellum unseg- mented, still shorter than scale, styliform, and tipped by a short spine. Antennal scale fringed with 19 long, thin, plumose setae along terminal and inner margins; small seta on outer margin near base of terminal segments; has four joints distally but only the three most distal joints are complete. Protopodite bears minute spine at base of scale in addition to spine at base of flagellum. MANDIBLES (Figure 2D).— More massive than in Stage I. Both mandibles bear additional denti- cles and molar processes more developed. Curved lip of truncated end of molar process of right man- dible more developed. MAXILLULE. — Unchanged from Stage I except basipodite now bears two additional spinulose spines. MAXILLA. — Shape similar to Stage I except exopodite slightly longer proximally and now bears nine marginal plumose setae in addition to plumose seta at proximal end. No change in number of setae on basipodite or coxopodite. MAXILLIPEDS.— Essentially identical to Stage I but bear additional setae as follows. First maxil- liped bears 17-20 setae on protopodite; exopodite bears 6 natatory setae rather than 4 as in Stage I; no change in epipodite. Second maxilliped bears 7 setae on protopodite; exopodite bears 10 lateral natatory setae in addition to the 2 terminal setae; endopodite five-segmented, setation formula 5, 2, 1, 1, 3. Third maxilliped bears 2 setae on protopo- dite; exopodite bears 10 lateral natatory setae in addition to the 2 terminal setae; segments of en- dopodite may or may not bear an additional seta or 2, setation formula usually 5, 4, 0, 1, 2. FIRST PEREOPOD (Figure 2E).— Endopodite functionally developed; five-segmented and ter- minating in a simple conical dactylopodite; seta- tion formula 4, 2, 1, 0, 0. Protopodite bears no setae. Exopodite, longest among pereopods, has 2 terminal and 10 lateral natatory setae. 239 FISHERY BULLETIN: VOL. 76, NO. 1 0.25 mm SECOND PEREOPOD (Figure 2F).— Similar to first pereopod except endopodite shorter, setation formula 3, 2, 0, 0, 1. Protopodite bears no setae. Exopodite with two terminal and six lateral natatory setae. THIRD PEREOPOD (Figure 2G).— Endopodite five-segmented; one-fourth to one-third longer than exopodite. Dactylopodite slightly longer than in first two pereopods; bears two setae terminally. Propodite bears two setae; remaining segments without setae. Exopodite noticeably shorter than exopodites of first two pereopods; bears six lateral 240 natatory setae in addition to two terminal nata- tory setae. FOURTH AND FIFTH PEREOPODS.— Unseg- mented except at base; without exopodite or setae; directed under body somewhat anteriorly as in Stage I (Figure 2A). PLEOPODS (Figure 2A).— Present as minute buds. TELSON (Figure 2H).— Similar in shape to Stage I but distinctly segmented from sixth abdominal HAYNES: P AND ALUS GONIURUS LARVAE Figure 2. — Stage n zoea of Pandalus goniurus: A, whole animal; B, antennule; C, antenna; D, mandibles (right and left); E, first pereopod; F, second pereopod; G, third pereopod; H, telson. 241 FISHERY BULLETIN: VOL 76, NO. 1 somite; bears eight pairs of densely plumose setae. Uropods still enclosed. Anal spine present but mi- nute. STAGE III ZOEA Total length of Stage III 6.2 mm (range 6.0-6.6 mm; 10 specimens). Rostrum (Figure 3A) projects horizontally but curves slightly downward at tip; bears the beginning of a tooth at base. Epipodite of first maxilliped minutely bilobed; pleurobran- chiae present as minute buds. ANTENNULE (Figure 3B).— Inner flagellum un- segmented; about one-half to two-thirds length of outer flagellum. Outer flagellum unsegmented; bears three long and three shorter aesthetascs terminally and one group of two aesthetascs prox- imally. Each segment bears additionally one or two long plumose setae. Large spine projects vent- rally from proximal segment. ANTENNA (Figure 3C).— Flagellum three-seg- mented; about two-thirds length of scale and tipped by remnant terminal spine. Antennal scale slightly narrower than in Stage II and fringed with 21 plumose setae; two complete joints at tip. Spine on protopodite at base of scale somewhat larger than in Stage II. FIRST PEREOPOD (Figure 3D).— Has begun to acquire adult shape, particularly in widened prop- odite and carpopodite segments. Exopodite bears 12 natatory setae in addition to terminal pair. SECOND PEREOPOD (Figure 3E).— Endopodite bears a few additional setae and dactylopodite slightly more conical than in Stage II. Propodite not yet projected anteriorly. Exopodite of second pereopod bears 9-10 natatory setae in addition to terminal pair. THIRD PEREOPOD.— Essentially identical to third pereopod of Stage II except each segment of endopodite bears an additional seta or two. FOURTH (Figure 3F) AND FIFTH PERE- OPODS. — Have begun to acquire adult shape, especially in lengthened dactylopodite and slightly widened propodite. PLEOPODS (Figure 3G).— Bilobed, unseg- mented, and without setae. 242 TELSON (Figure 3H).— Uropods free. Endopodite undeveloped; about one-third length of exopodite and bearing two simple setae terminally. Anal spine clearly visible. STAGE IV ZOEA Total length of Stage IV 7.7 mm (range 6.8-8.3 mm; 10 specimens). Rostrum (Figure 4A) bears two teeth dorsally, no teeth ventrally; tip not yet bifid. Epipodite of first maxilliped fully bilobed; pleurobranchiae small but readily visible, project anteriorly. Epipodite on second maxilliped pre- sent as a small bud. No mastigobranchiae. ANTENNULE.— Shaped as in adult. Neither inner nor outer flagellum segmented. Outer flagel- lum bears an additional group of three aesthetascs proximally. ANTENNA (Figure 4B).— Flagellum six-seg- mented; longer than scale but does not extend past terminal setae of scale. Antennal scale without joints at tip. Other than increase in size, changes in antennal scale from Stage IV onward are neg- ligible. FIRST PEREOPOD.— Essentially no change from Stage III except exopodite may have an additional pair of natatory setae. SECOND PEREOPOD (Figure 4C).— Distal joint of propodite projects slightly anteriorly. Exopodite has 10-12 natatory setae in addition to terminal pair. THIRD PEREOPOD.— Shaped as in adult; exopo- dite with five pairs of natatory setae in addition to terminal pair. FOURTH AND FIFTH PEREOPODS.— Shaped as in adult. PLEOPODS (Figure 4D).— Still unsegmented; length of second pleopod about one-third height of second abdominal segment. Neither setae nor ap- pendix internae present. TELSON (Figure 4E).— Endopodite of uropod nearly as long as exopodite and fringed with about 20 setae. Lateral margins of telson nearly parallel but slightly wider posteriorly and bear two spines each. Terminal margin still slightly emarginate; HAYNES: PANDALUS GONIURUS LARVAE Figure 3. — Stage III zoea o{ Pandalus goniurus: A, carapace; B, antennule; C, antenna; D, first pereopod; E, second pereopyod; F, fourth pereopod; G, second abdominal segment and pleopod; H, telson. 243 FISHERY BULLETIN: VOL. 76, NO. 1 B 0. 5 mm 0. 5 mm 0. 5 mm Figure 4. — stage IV zoea of Pandalus goniurus: A, rostrum; B, antenna; C, second pereopod; D, second abdominal segment and pleopod; E, telson. 244 HAYNES: PANDALUS GONIURUS LARVAE bears six pairs of spines, the outermost (sixth) pair usually without spinules. STAGE V ZOEA Total length of Stage V 10.3 mm (range 8.2-11.3 mm; 10 specimens). Rostrum (Figure 5A) with five or six dorsal teeth and no ventral teeth; tip smooth but may bear small hump indicating future loca- tion of bifid tooth. Epipodite of second maxilliped lobed. Mastigobranchiae occur as minute buds on third maxilliped and pereopods 1-3. ANTENNULE.— Inner fiagellum usually four- segmented; still bears terminal spine. Outer fiagellum three-segmented; bears four groups of three aesthetascs each in addition to terminal aes- thetascs. ANTENNA (Figure 5B).— Fiagellum about 1.7 times length of scale; 11-12 segments. FIRST PEREOPOD (Figure 5C).— Propodite pro- jects anteriorly but not as much as in second pereopod; projection bears small spine terminally. Neither dactylopodite nor propodite projection bear subterminal spines. SECOND PEREOPOD (Figure 5D).— Chela well formed. Dactylopodite bears two spines subtermi- nally, and propodite projection one spine subter- minally. Carpopodite not segmented. PLEOPODS (Figure 5E).— Segmented; length about two-thirds height of second abdominal seg- ment. Flagella tipped with several simple setae, except first pair of pleopods bears setae only on outer fiagellum. TELSON (Figure 5F). — Uropods similar in shape to adult; telson margins somewhat parallel, bear two spines each. Terminal margin straight or only slightly emarginated, bears six pairs of spines. No evidence of transverse hinge of exopodite of uropod. STAGES VI AND VII (MEGALOPA AND FIRST JUVENILE) Total length of Stage VI (megalopa) 13.8 mm (11.1-15.8 mm; 6 specimens). Carapace without supraorbital spine. Rostrum (Figure 6A) shaped as in adult; posterior dorsoventral width not as pronounced nor ventral teeth as fully developed as in Stage VII; bears eight or nine teeth dorsally in addition to distinct bifid tip and four or five teeth ventrally. Usually one or two setae occur between several of the posterior dorsal teeth. Exopodites on third maxilliped and pereopods reduced. Mas- tigobranchiae larger but still not evident on fourth pereopod. Pleurobranchiae clearly lobulated. Inner fiagellum of antennule five- or six-seg- mented and outer fiagellum four-segmented. Inner fiagellum lacks terminal spine. Outer fiagel- lum bears subterminally six groups of three aes- thetascs each; terminal segment lengthened, without aesthetascs. Mouthparts shaped as in adult; mandibular palp present, two-segmented, without setae. Chelae of first and second pereopods shaped as in adult; carpal joints of left and right second pereopods 20 to 25 and 7 to 9, respectively. Meropodite of left second pereopod three- segmented. Pleopodal setae extend along entire lateral margins of both fiagella; tips of appendix internae bear several distinct cincinnuli. Telson (Figure 6B) shows, for first time, shape and spina- tion similar to adult; lateral margins narrow pos- teriorly but widen slightly at terminal margin. Typically three pairs of spines on lateral margins of telson but often a spine, rarely two, lacking. Terminal margin of telson rounded but not as much as in Stage VII; bears three pairs of stout spines. Transverse hinge of uropod exopodite com- plete. Total length of Stage VII (first juvenile) 14.9 mm (range 13.7-15.8 mm; 3 specimens). Rostrum (Figure 7 A) typically adult; posterior dorsoventral width slightly greater than in Stage VI; ventral teeth fully formed and one or two setae between most, if not all, teeth including bifid tip. No exopo- dite on third maxilliped or pereopods. Mastigo- branchia evident on fourth pereopod. Flagella of antennules lengthened as in adult; outer fiagel- lum nine-segmented, bears nine groups of three aesthetascs each; inner fiagellum six-segmented. Mandibular palp three-segmented, with spinous setae. Carpal joints of left and right second pereopods 29 and 11, respectively. Meropodite of left second pereopod 1 1-segmented. Telson (Figure 7B) adult in shape, typically bears four pairs of lateral spines although often lacks a single lateral spine. 245 FISHERY BULLETIN: VOL. 76, NO. 1 0. 5 mm Figure 5. — Stage V zoea of Pandalus goniurus: A, rostrum; B, antenna; C, first pereopod (terminal segments only); D, second pereopod (terminal segments only); E, second abdominal segment and pleopod; F, telson. 246 HAYNES: PANDALUS GONWRVS LARVAE 1 . 0 mm 0. 5 mm Figure 6. — Stage VI (megalopa) of Pandalus goniurus: A, ros- trum; B, telson. 1 . 0 mm 0. 5 mm Figure 7. — Stage VII (first juvenile) oi Pandalus goniurus: A, rostrum; B, telson. COMPARISON OF LARVAL STAGES WITH DESCRIPTIONS BY OTHER AUTHORS Ivanov (1965) described and illustrated the first stage zoeae of P. goniurus that he reared in the laboratory from known parentage. His descrip- tions agree in all aspects with mine except for the third maxillipeds: Ivanov's zoeae had 9 natatory setae on the exopodite compared with 12 natatory setae in my zoeae. The only other description of P. goniurus larvae known to me is that of Makarov (1967) who con- structed a series of zoeal stages from plankton of the western Kamchatka coast based on Ivanov's description of Stage I. Makarov's descriptions of each stage are brief and include primarily de- velopment of the rostrum, antennal flagellum, dactylopodite of the second pereopod, pleopods, and telson. Makarov's zoeae are essentially iden- tical to mine through Stage V but Makarov's Stages VI and VII possess mostly zoeal charac- teristics, rather than postzoeal as mine do. For instance, in Stage VI the rostrum of Makarov's specimens is not bifid and does not bear ventral teeth, and the telson still bears six pairs of spines terminally. In my Stage VI specimens, the ros- trum is bifid, bears five or six distinct ventral teeth, and the telson bears only four pairs of spines terminally. In Stage VII, the rostrum of Makarov's specimens is bifid but bears only three or four poorly developed teeth ventrally and the telson still bears six pairs of spines terminally. In my Stage VII specimens both the rostrum and telson are essentially fully developed as in the adult. Apparently P. goniurus from the western Kam- chatka coast has at least two more zoeal stages than P. goniurus from Kachemak Bay. The morphological differences between larval Stages VI and VII of P. goniurus from the western Kamchatka coast and from Kachemak Bay, Alaska, may reflect variation in number of molts in response to environmental conditions. Variabil- ity in number of molts required to reach a specific point in development in the Crustacea is well known. In a review of the literature, Costlow (1965) showed that variability in number of molts occurs in the Cirripedia, Euphausiacea, Natantia, Reptantia, Anomura, and Brachyura regardless of whether the larvae are reared in the laboratory or 247 from the natural environment. Regarding the Pandalidae, Pike and Williamson (1964) have shown variability in number of molts required to reach the megalopa stage in the plankton for Pan- dalina brevirostris (Rathke) and Pandalus pro- pinquus G. O. Sars, and that larvae of Dichelopan- dalus bonnieri (Caullery) and Pandalus montagui Leach reared in the laboratory have more larval stages than specimens from plankton. Berkeley (1931) mentioned the possibility of variation in number of molts in larvae of Pandalus danae Stimpson. Kurata ( 1964) speculated that larvae of P. borealis Kr0yer in Japanese waters may have six or seven stages. I have observed that both P. borealis and P. goniurus reared in the laboratory are capable of prolonging their normal interval between zoeal moltings (about 10-15 days) to as much as 5 wk, and that P. borealis may have as many as 1 1 zoeal stages before reaching the megalopa stage. Al- though the causes of molt retardation and mor- phological variation in pandalid larvae have not been established, the potential for variability exists not only in P. goniurus but in other pan- dalids as well. Variability in larval development from different geographical areas, therefore, is to be expected. FISHERY BULLETIN; VOL. 76, NO. 1 LITERATURE CITED Berkeley, A. A. 1931. The post-embryonic development of the common pandalids of British Columbia. Contrib. Can. Biol. 6(6):79-163. COSTLOW, J. D., Jr. 1965. Variability in larval stages of the blue crab, Cal- linectes sapidus. Biol. Bull. (Woods Hole) 128:58-66. HAYNES, E. 1976. Description of zoeae of coonstripe shrimp, Pandalus hypsinotus, reared in the laboratory. Fish. Bull., U.S. 74:323-342. IVANOV, B. G. 1965. (A description of the first larvae of the far-eastern shrimp {Pandalus goniurus).) [In Russ., Engl, summ.] Zool. Zh. 44:1255-1257. (Translated by U.S. Dep. Com- mer., NOAA, Natl. Mar. Fish. Serv., Div. Foreign Fish.) Kurata, H. 1964. Larvae of decapod Crustacea of Hokkaido. 3. Pan- dalidae. Bull. Hokkaido Reg. Fish. Res. Lab. 28:23-34. (Transl., Fish. Res. Board Can., 1966, Transl. 693.) MAKAROV, R. R. 1967. Larvae of the shrimps and crabs of the west Kam- chatkan shelf and their distribution. Natl. Lending Libr. Sci. Technol., Boston Spa, Yorkshire, 199 p. Pike, R. B., and D. L Williamson. 1964. The larvae of some species of Pandalidae (Decapo- da). Crustaceana 6:265-284. 248 IMMIGRATION OF FISHES THROUGH THE SUEZ CANAL^ Adam Ben-Tuvia^ ABSTRACT The number of Red Sea fishes found in the eastern Mediterranean amounts to 36 species. Twelve immigrants, namely: Spratelloides delicatulus , Herklotsichthys punctatus, Tylosurus choram, Sebas- tapistes nuchalis, Epinephelus tauvina, Autisthes puta, Pelates quadrilineatus, Silago sihama, Rhon- sicus stridens, Crenidens crenidens, Rastrelligerkanagurta.Scomberonwrus commerson, were found in the last 12 yr. The southward migration, from the Mediterranean to the Red Sea is almost negligible. Only Liza aurata, Dicentrarchus punctatus . and perhaps Carcharhinus plumbeus can be regarded as Mediterranean immigrants. In studying the immigration of fishes through the Suez Canal, three zooecological areas must be taken into consideration: 1) the northern Red Sea; 2) the eastern Mediterranean; and 3) the Suez Canal itself in which many marine animals from the two neighboring areas have found a perma- nent habitat (Steinitz 1968). The prevailing hydrographic conditions differ in these three areas, although the salinities and summer temperatures are to some extent similar (Morcos 1967, 1970; El-Saby 1968; Oren 1970; Oren and Hornung 1972). Temperature and salin- ity are the main abiotic factors influencing the distribution of organisms over large zoogeo- graphical areas. Often they also have a decisive influence on the ecological distribution of species in various biotopes of an area. The process of immigration is highly selective. Common species of the home seas are not necessar- ily successful immigrants in a new region. Similar effects have been shown to occur in many forms of colonization (MacArthur and Wilson 1967). The adaptation of a species to a new area requires adjustment of its reproductive processes, espe- cially with regard to the correct timing of spawn- ing in order to ensure suitable physical and ecolog- ical conditions for the development and survival of the young stages. It is evident that the direction of immigration is mainly from the Red Sea into the Mediterranean (Figure 1). The possible causes of such one way immigration have been discussed elsewhere ( Aron 'This paper was read at the 17th International Zoological Congress in Monte Carlo, 25-30 September 1972; some changes were introduced to include more recent information on immi- grants. ^Department of Zoology, Hebrew University of Jerusalem, Jerusalem, Israel. Manuscript accepted June 1977 FISHERY BULLETIN: VOL. 76, NO. 1, 1978. and Smith 1971; Ben-Tuvia 1971a, 1973; Por 1971a, b). Thirty-six Red Sea or cosmopolitan species can be regarded as Suez Canal immigrants. Twelve of them were found within the last 12 yr. Evidently, immigration is a continuous process, and over time the probability of suitable species of fishes entering the Suez Canal and colonizing the new region increases. Time also plays an essential role in the biological processes of adaptation of the species to the modified conditions of life. More resistant species, endowed with greater plasticity of genetic characters, can form local "races" within a few generations by natural selection in the new environment (Kosswig 1974). But first they need a firm foothold on the other side of the Canal, geo- graphically close to the parental stock and in places where conditions are not drastically dif- ferent from their normal habitat. Recently I had an opportunity to collect samples from the Gulf of Suez (Ben-Tuvia and Grofit 1973), Suez Canal (Steinitz and Ben-Tuvia 1972), and Bardawil Lagoon (Ben-Tuvia 1975a) which re- vealed interesting data on the distribution of im- migrants. Many of the species which have success- fully colonized the eastern Mediterranean, such as Saurida undosquamis , Leiognathus klunzingeri, Upeneus moluccensis, and U. asymmetricus, and which are abundant there, are also dominant species on the trawling grounds of the Gulf of Suez. High percentage of Red Sea fishes found in the hypersaline Bardawil Lagoon on the northern coast of Sinai indicates that it may serve as a stepping stone in the immigration of Red Sea fishes into the Mediterranean, especially if we re- gard it as a part of the system of lakes and lagoons of the Isthmus of Suez (Por 1971a). Among 55 249 FISHERY BULLETIN: VOL. 76, NO. 1 w "3 01 C J5 3 ■' CO CS 3 V CO CO ^ § I be'*' "5 k « c t: 3 ■as CO 3 O c 3^ CO s I a. -2 CO c^ r). ^ST :^ O « ST C ^ T3 C CO m ^^ CO 3 <<-< o CO m ca (N o o CO cd ■^ cn c 01 J5 CO 300 mm SL, standard length), which were consequently underrepresented in the collec- tions. Thus the samples probably reflect the usual size distribution of fish between ca. 100 and 300 mm SL over the reef (Table 1). In this way, 324 specimens were collected between 0900 and 1500 h during all seasons from March 1971 to June 1972. Of these, 80% had food in their stomachs. We made considerable effort not to bias stom- ach-content composition. Underwater chumming or disturbing the bottom were never used as ways to attract fish near the collector. Spearing was begun only after it was ascertained that no sport fishing involving chumming with live bait (usu- ally northern anchovy, Engraulis mordax) occur- red within visual range of the collecting site. An initial practice of securing individual fish in plas- tic bags or locking their mouths with paper clips was soon discontinued when no individual was seen to regurgitate food. All specimens were placed immediately in an ice chest aboard the div- ing skiff. In the laboratory, they were measured (nearest millimeter SL), slit open, and their intes- tines detached and measured (millimeters SL). Other trophic structures (jaw length, gill rakers on first arch, and greatest width between gill rak- ers) were measured on a few typical specimens of about 225 mm SL. Specimens were then fixed in 10% Formalin^ and preserved in 50% isopropanol. To investigate the effect of habitat on the olive rockfish's diet, one of us (Love) collected an addi- tional 110 individuals from One-Mile Reef, an open, rocky reef located 1.6 km offshore of Santa Barbara Harbor, about 20 km east of Naples Reef Of these, 72 (65.5% ) had stomachs containing food (Table 1). Too deep and turbid to support kelp, this reef is made up of a strip of rocky bottom at about 27 m depth, with 1.5-5.0 m high rock piles scat- tered along its length. From January to October, fish were caught by angling with artificial lures and by gill net. No sport fishing or chumming were seen to occur during collecting. Fish were pre- served and processed as before. Gut fullness was estimated before stomach con- tents were sorted and identified. Degrees of full- ness of stomach and of the first half of the intestine were scored from 1.0 (empty) to 5.0 (full). Stomach contents were sorted taxonomically into 26 food items (Table 2). The volume of each item was mea- sured by liquid displacement. The "nekton" cate- gory of items (prey type) included all nonlarval fish and squid prey. The substrate-oriented prey type included all prey (except fish) that live on or about reef and plant surfaces. Such prey are either motile like shrimps, amphipods, and small crabs, or attached like hydroids, bryozoans, and the algae itself. Plant material was identified as either kelp (Macrocystis) or other algae, mostly low lying browns and reds. In computing percent volumes and frequencies of occurrence of prey per ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table l. — Number, size, and food containment of specimens examined of the three species of kelp-bed fishes (blue rock- fish, kelp bass, and olive rockfishi from Naples Reef or One-Mile Reef (olive rockfish only) off Santa Barbara, Calif See also Figure 1. Total Specimens with food in their stomachs by size groups Locality specimens with with 50-150 mm SL 151-300 mm SL 301-400 mm SL and species examined food food No. Range N/ledlan No. Range Median No. Range IVIedian Naples Reef: Blue rockfish 122 97 79.5 30 78-149 118 5 67 1 50-262 193.0 — — — Kelp bass 102 86 843 — — — 67 167-296 209.0 19 304-400 328.0 Olive rockfish 100 86 86,0 13 82-150 122.5 73 151-274 196.0 One-lvlile Reef: Olive rockfish 110 72 65.5 — — — 72 158-290 222.0 — — — 258 LOVE and EBELING: FOOD AND HABITAT OF THREE FISHES Table 2. — Percent total volume and frequency of occurrence of 26 food items in stomachs with food of the three species of kelp-bed fishes in the 151- to 300-mm size group (Table 1, Figure 1) from Naples Reef or One-Mile Reef (olive rockfish only) off Santa Barbara, Calif Food items are listed by general characteristics and presumed major dajftime source. A tr indicates unmeasurable trace; a dash indicates none. Primarily planktonic (Sum =) Small crustaceans (0.5-5 mm long): Ostracods Cladocerans Zoea larvae Cope pods Megalops larvae Large crustaceans ("^10 mm): Euphausiids Pleuroncodes Small-medium sized, transparent (1-10 mm): Eggs Chaetognattis Tunicafes (small salps. larvaceans) Large, transparent (^^15 mm): Siptionophores. medusae, etc. Fish larvae (5-15 mm) Primarily nektonic (20-80 mm) (Sum = Fish Squid Ectoparasites of other fish: Parasitic copepods Primarily substrate oriented (Sum =) Free moving animals: Crabs Shrimps fwlysids Isopods Gammaridean amphipods Caprellid amphipods Hyperiid amphipods « Polychaete worms ' Hydroids Kelp, etc.: Kelp (including encrustfng bryozoans) Other algae (including encrusting bryozoans) Total volume of food consumed (ml) Total number of specimens examined (56.7) 0.6 0.2 0.6 0.6 1.9 51,5 0.7 0.6 (15.7) 7.4 8.3 (27.5) 0.3 0.2 1.6 0.1 0.1 0.3 13.1 10.5 1.3 171.2 20.9 22.4 11.9 1.5 1.5 10.4 40.3 4.5 10.4 13.4 3.0 6.0 3.0 17.9 1.5 1.5 1.5 16.4 25.4 9.0 67 (12,6) 0.3 1.5 0.1 2.5 7.8 0.4 (55,3) 51.0 4.3 (32.3) 0.8 0.7 0.5 0.2 2.2 7.5 0.4 8.7 8.8 2.5 141.3 6.0 7.5 3.0 1.5 16.4 1.5 46.3 6.0 3.0 1.5 9.0 1.5 13.4 13.4 7.5 16.4 16.4 14.9 67 (10.5) 0.3 0.3 2.6 0.1 5.4 2.9 15.1 15.1 24.7 2.7 6.8 1.8 16.4 (85,0) 84.2 54.8 0.8 4.1 tr 2.7 (4.5) 1.4 0.8 8.2 0.8 6.8 tr 1.4 20.5 Naples Reef One-Mile Reef Blue rockfish % vol. % freq. Kelp bass Olive rockfish % vol. % freq. Olive rockfish Food Item % vol. % freq. % vol. % freq. (41.8) tr 2.9 0.4 8.6 6.5 35.7 15,8 34,3 7.0 47,0 1.2 2.9 4.4 4.9 0.1 2.9 1.0 5.7 5.4 16.7 :55.2) 51.0 28.0 4.2 5.7 0.4 12.9 (2.6) 1.4 85.8 0.1 102.9 14.3 tr 1.4 tr 1.4 0.1 2.9 1.0 14.3 1.4 73 72 species (Table 2), fish with empty guts and of sizes outside the middle range of 151-300 mm SL (Ta- ble 1, Figure 1) were excluded. To test for communal switch feeding and dietary consistency, we examined variation among indi- viduals. We counted fish that contained mostly one food item or prey type and that 1) were of one species collected on the same day, 2) were of all three species collected on the same day (Table 3), and 3) were of all species collected at any time (Table 4). To examine seasonal variation in diet, stomach contents of each species were pooled by seasonal periods that correspond roughly to different oceanographic regimes off Santa Barbara. Brown (1974) concluded that in the Santa Barbara Chan- nel, cooling of surface water typically proceeds from December to July, first by surface mixing and small-scale upwelling associated with storms from December to April, then by large-scale upwelling from May through July. This precedes gradual surface warming from late June to December, with strongest thermal stratification and clearest water from August to December. Therefore, we delimited seasonal periods as: 1) December- February, a period of winter storms and the be- ginning of vertical mixing and surface cooling ( in- itial breeding season of many species); 2) March- May, a period of most intense upwelling of deep cold water (high surface productivity, zooplankton blooms, appearance of young-of-the-year fish, etc.); 3) June-August, a period of decreasing up- welling and the beginning of thermal stratifica- tion and surface warming (a transitional period); 259 FISHERY BULLETIN: VOL. 76, NO. 1 SIZE GROUP (STANDARD LENGTH) 80-150 151-200 201-300 30l-400mm (-) BLUE RKF. (30) (40) (27) P »3% ||7% 70% 70% N 15% 33% 38% 15% 8 m^^ 4e K 40% 26^ (-) KELP BASS OLIVE RKF. (NAPLES) OLIVE RKF. (1-MILE) P N S K (13) p 1 77'^ N ; 8% S 54% (-) p N S 83% 17% 17% 76% 1 33% 33% 222 P<0.005 05>P>0.025 M P^oos Figure l. — Percentage frequency of prey types (bars and num- bers) in stomachs offish in all size groups of the three species of kelp-bed fishes from Naples Reef (all three species) or One-Mile Reef (olive rockfish only) off Santa Barbara, Calif Prey types are designated: P, plankton; N, nekton; S, substrate-oriented prey; and K, kelp and other algae (with encrusting bryozoans), and are represented by any constituent food item under the appropriate prey-type heading in Table 2. Numbers in parenthesis are num- bers of fish stomachs examined. Hatching shows significantly different frequencies at the indicated probabilities determined by chi-square tests (see text). and 4) September-November, a period of warm, clear surface water with little vertical mixing. The 26 food items were ranked for each season by vol- ume, using data from all size groups of fishes to maximize sample size (Table 5). Seasonal varia- tion in diet was also tested by frequencies of oc- currence of subsets of items comprising major food categories, using data from the 151- to 300-mm SL size group only (Figure 2). Habitat Spatial distributions of the three species were determined from underwater movies taken for another project. Observations were made from 2.5-min Super-8-mm underwater movie strips in color (cinetransects) filmed by scuba divers swimming courses started at random either under the kelp canopy or just over the bottom at study sites near Santa Barbara and across the Santa Barbara Channel along Santa Cruz Island (Bray and Ebeling 1975; Ebeling, R. Larson, and W. Alevizon in prep.). An initial set of cinetransects was filmed in 1970 over a variety of habitats and areas at both localities. Then, during the fall sea- sons of 1971-74, transects were filmed over per- manent study sites at Naples Reef and at Santa Cruz Island west of Prisoner's Harbor. Fish were counted by species as the films were projected in the laboratory. Environmental characteristics were measured or scored either on station or dur- ing projection. Breadth and Overlap Breadth and overlap of resource use were com- puted from values of p, , the proportion of item i used by each species, either at Naples Reef (food and space) or off Santa Cruz Island (space only). For food, p^ is the proportionate volume of any of the 26 different food items included in the species total (S); for space it is the proportionate abun- dance of the species in any of the 297 cinetransects taken over Naples Reef or 331 cinetransects taken along Santa Cruz Island. Resource breadth, B = s l/Xp^, can be thought of as the theoretical number i= 1 of equally used food items (or spaces covered by cinetransects) yielding a value of B equal to the observed. For example, if all items are in equal proportions, B equals S, the total items in the spectrum (see Bray and Ebeling 1975). A Hill's (1973) ratio was used to estimate the degree of concentration of each species among cinetransects (the unevenness of distribution offish numbers): HR = exp(H')/5, where H' is the Shannon- s Weaver measure of diversity , -2p, Inp,. Since i/' 1=1 is more sensitive to changes in the small to medium values of proportionate abundances than is B, their ratio is a sample-size independent mea- sure of concentration of observations (Peet 1974). Overlap between two species, / = 1.0 - [0.5 s (X\pij - P'L I)], where p J is the proportion of item i 1=1 used by species 7 and s is the species total of food items eaten (or cinetransects in which recorded), is scaled from zero (complete discordance of item use) to 1.0 (all items used in equal proportions) (e.g., Whittaker 1960; Cody 1974; Ebeling and Bray 1976). 260 LOVE and EBELING: FOOD AND HABITAT OF THREE FISHES RESULTS Morphology, Size Groups, Gut Fullness Of the three species, the blue rockfish appeared best adapted to eat a diverse array of small prey. It has a shorter jaw (ca. 15% of SL) than the olive rockfish and kelp bass (ca. 17%). It has about the same number of gill rakers on the first arch as the others (34-37); but has significantly smaller inter- raker widths {X = 1.24 ±0.088 mm, 95% con- fidence limits, n = 10) than the others pooled ix = 1.80 ±0.076, n = 20). Blue rockfish have a sig- nificantly longer intestine (ratio, intestinal length/SL of x = 1.41 ±0.147, n = 15) than either kelp bass ix = 1.11 ±0.105, n = 18) or olive rockfish (x = 0.807 ±0.098, n = 19). Tests justified comparing diets offish within the 151- to 300-mm SL size range, which included 82% of all food-containing individuals (Table 1). Within this range, only the median length of olive rockfish from One-Mile Reef differed significantly from the others (Kruskal-Wallace ranks location test, P<0.05 including the One-Mile sample, P>0.1 excluding it). Also (Figure 1), diets as ex- pressed by frequencies of occurrence of prey types were not significantly heterogeneous between subgroups: largest chi-square value determined in tests of the resulting 14 contingency tables of di- mension two (presence or absence) by two (sub- groups within this size range) = 2.31 (P>0.1). However, tests showed less justification for in- creasing sample size by adding individuals from outside the 151- to 300-mm size range (Figure 1). Diets were often significantly heterogeneous be- tween subgroups when either smaller (blue rock- fish, olive rockfish) or larger (kelp bass) sizes were included: 5 of 11 chi-square values determined in tests of the resulting 11 contingency tables of di- mension two (presence or absence) by three (sub- groups both within and without the 151- to 300-mm range) were significant at P~ 0.05 or less. Scored stomach fullness in 151- to 300-mm Naples Reef fish was about the same for all three species: x = 2.72-2.75, an equivalent of about 46% full. Intestinal fullness averaged somewhat great- er: X = 2.76 (olive rockfish) to 3.00 (others). Blue rockfish and olive rockfish in the smaller size categories had fuller stomachs: x = 3.81-3.10, re- spectively. Olive rockfish from One-Mile Reef had less food in their stomachs ix = 2.15) but as much food as the others in their intestines {x = 3.05). Intestinal contents usually resembled stomach contents. Food Diets Blue rockfish ate mostly swimming, drifting, or attached organisms in midwater under and about the kelp canopy (Table 2, Figure 1). Tunicates, hydroids, kelp, fish, and smaller planktonic prey formed most of the fish's diet throughout the year. Recognizable fish prey included juveniles of pipefish, Syngnathus; blue rockfish; and C-O soles, Pleuronichthys coenosus; and adults of northern anchovy. Fish larvae made up but a small part of the blue rockfish's diet. Pelagic tunicates — the thaliaceans (salps) Salpa and Doliolum and the larvacean Oikopleura — constituted the largest volume of food consumed. Among the relatively large numbers of small plankters eaten, copepods ranked very low in vol- ume, but relatively high in frequency of occur- rence. Hydroids (especially Sertularia) ranked high in volume consumed. The blue rockfish were probably not merely ingesting hydroids to obtain the caprellid amphipods that live there (Gotshall et al. 1965), because caprellids were found along with hydroids in only 2 of 20 stomachs. Some 73% of the fish that contained kelp and other algae also contained detached hydroids and encrusting bryo- zoans (Membranipora). So most plant material may have once borne epiphytic prey now detached. And like tunicate tunics, algae per se was appar- ently passed undigested, so fish probably eat plants for the attached animals (Quast 1968d; Bray and Ebeling 1975). Kelp bass foraged primarily in midwater, but occasionally ate bottom organisms (Table 2, Fig- ure 1). They ate mostly fish, which ranked first in both total volume and frequency of occurrence. Recognizable fish prey included juveniles of rockfishes, pipefish, kelp greenling, Hexagram- mos decagrammus, topsmelt, Atherinops affinis, anchovy, and jack mackerel, Trachurus symmet- ricus, and adults of anchovy and agonids. Kelp bass ate no fish larvae and relatively less plankton than did the other species. Thaliacean tunicates (Salpa) contributed the largest volume of plankton consumed; copepods and other small crustaceans occurred at moderate frequency and in fairly large numbers in a few individuals. Bass ate relatively more substrate-oriented prey, with 261 FISHERY BULLETIN: VOL. 76, NO. 1 hydroids (especially Sertularia), caprellid am- phipods, and kelp ranking highest among such items. Most caprellid amphipods were found in stomachs containing substantial amounts of hy- droids and bottom algae, indicating that fish may ingest such turf for the contained animals. About a third of all pieces of kelp bore attached bryozoans (Membranipora) or hydroids. Whether speared from Naples Reef or angled from One-Mile Reef, olive rockfish ate relatively more fish than did the others (Table 2, Figure 1). Recognizable fish prey in Naples Reef individuals included juveniles of blacksmith, Chromis punctipinnis, anchovy, pipefish, blue rockfish, other olive rockfish, and adults of topsmelt and anchovy. One-Mile Reef fish had eaten adult an- chovies and a young pipefish. Fish larvae made up a relatively large part of the diets of olive rockfish from both localities. One-Mile Reef fish ate more kinds and greater numbers of small zooplankton. Individuals of all sizes ingested and retained such tiny prey as ostracods, cladocerans, and small copepods (e.g., Coryceus emarginata). During the winter, copepods and zoea larvae actually out- ranked fish prey in volumes consumed. Many polychaetes, which occurred commonly in fish from either area, were of the small nereid variety found in the kelp canopy (Quast 1968c) and swarming in the midwater plankton at night (Hobson and Chess 1976). Only olive rockfish con- tained parasitic copepods among their stomach contents. Although these copepods were identified as Caligus, an obligatory ectoparasite, olive rockfish were not observed to clean (i.e., pick such prey from off other host fishes). Individual Variation On any given day, individuals of the same species tended to select the same food item. Within particular collections of 2-9 individuals, 67% of a cumulative total of 96 blue rockfish, 60% of 72 kelp bass, and 60.5% of 86 olive rockfish had the same item dominating their stomach contents. Occasionally, individuals of all three species selected items from the same major prey category, although not necessarily the same item (Table 3). Plankton dominated the stomach contents of most individuals sampled together in a February and in an April collection, while nekton and substrate- oriented prey were favored by those in three May and in one October collections. Yet fish in two November and two January collections showed 262 little communality of diet. And even when they tended to select items from the same prey type, as in the February, April, May, and October collec- tions, they often selected different items. For example, most blue rockfish collected on 22 Feb- ruary 1972 had mostly salps or chaetognaths in their stomachs; kelp bass contained either salps or copepods; and olive rockfish contained larval fish. On the other hand, all blue rockfish and most kelp bass in the 21 January 1972 collection had eaten a single planktonic item, namely salps. Fish usually selected the same prey type during a particular feeding bout (Table 4). For all species pooled, 76% of the individuals contained more than 95% by volume of items in a single major prey category (prey type), and 39% contained but a single item (20% with relatively small items, 19% with large items). Combinations of prey types var- ied among the three species: usually plankton and substrate-oriented prey for kelp bass, and plankton and nekton for olive rockfish (Table 4). Of all fish containing kelp, etc. (Figure 1), about 40% also contained relatively large amounts of substrate-oriented prey, about 15% each also con- tained relatively large amounts of plankton or nekton, and the remainder contained kelp only. About 83% of 81 specimens with recognizable prey in both stomach and intestine had the same prey type dominating the contents of both. Seasonal Variation Considering all 26 food items, diets were weakly, though usually significantly concordant among seasons (Table 5). Fish ate relatively greater volumes of plankton during winter-spring periods, and more nekton or substrate-oriented prey during summer-fall. Showing the greatest seasonal variation (least concordance), the blue rockfish's diet included 93% plankton (by volume) in the winter, 75% in the spring, and less than 8% in summer-fall. Tunicates ranked high from De- cember to August, while kelp (with encrusting animals), hydroids, and, later, fish, ranked high from March to November. Similarly, olive rockfish from One-Mile Reef contained 80%, 25%, and < 10% plankton (by volume) in the first three sea- sonal periods, respectively. Small crustaceans ranked high from December to August, while fish and polychaetes ranked high from March to November. Individuals of both species ate larval fish during late winter and spring when such prey are most abundant. Seasonal trends for the others LOVE and EBELING: FOOD AND HAtMTAT OF THREE FISHES 5 cd C B o ^ a. ed Z ea ^ el 9 3 2 to ^ a> e o z ^ a> m O 7j O ^ r^ m ^ o ^— 5 it: CO C\J m C»i CD O *— >> 2 iC o m CM 05 O 5 ^ in m i O ^— a. < ^ C\J CQ Cvj CJ> o ^— ri LL ^ CM m CO cn O T— —3 i«: CNJ m C31 O ^— C to -3 ^ CM ffl 6 8 LL S i? 0) to to £ , — to 2 ■5 (A E o 5 >. — "- to (0 9; p ° E N ao o ro 5 S. "-L. c - a)tDa;tDclog>'SQ.£ 0 5 CJ 1- Q- -I 03 ^ p to C^ 5 i: 0)9 2 ;? 15 c tr ^ Jc -h; ^ u) to to C — .^' ^ j^ -^ ID — - c — to (O ^ ^ :>, ■ ■ Q. to CI filii|l i QI O ■D -o m SZ j= JZ TD £ P 9 !0 " >. O I to 3 _ o to c o OJ N o g n — (jiVi Q. C ~ 0) • O n to 263 FISHERY BULLETIN: VOL. 76, NO. 1 Table 4. — Numbers of the three species of kelp-bed fishes in the 151- to 300-mm size group (Table 1, Figure 1) from Naples Reef or One-Mile Reef (olive rockfish only) that contained more than 95% (by volume) of items composing prey types (plankton, P; nekton, N; or substrate-oriented prey, SOP) listed in Table 2. Species P N SOP P -^ N SOP + P SOP + N Naples Reef Blue rockfish Kelp bass Olive rockfish 24 11 20 3 23 28 18 23 5 1 1 7 13 3 3 1 6 4 One-Mile Reef Olive rockfish 43 15 — 5 3 — Totals 98 69 46 14 22 11 % of total (279) food- containing specimens 35.1 24.7 16.5 5.0 7.9 3.9 were less clear. Kelp bass ate tunicates from De- cember to May, but fish, kelp, and hydroids were important prey for much of the year. Olive rockfish from Naples Reef ate mostly fish throughout the year. To test for seasonal differences in diet, the fre- quencies of prey types were subjected to chi-square tests of homogeneity calculated from contingency tables of dimension two (presence or absence) by four (seasonal periods). Plankton frequencies were significantly heterogeneous, with highest values during winter-spring periods (Figure 2). As fewer kelp bass and olive rockfish ate plankton during the year, more ate nekton, primarily small fish. More blue rockfish and kelp bass ate more algae (with encrusting animals) later in the year. Species showed greater overlap in diet during periods when their stomachs were fuller of prey (Table 6). For all species, both stomach fullness and food overlap were greater during summer-fall than during winter-spring (Table 6). Fullness may relate to greater exploitation of nekton during summer-fall (Figure 2). For all species, stomachs Table 5. — Seasonal variation in diets of the three species of kelp-bed fishes in all size groups (Figure 1) from Naples Reef or One-Mile Reef (olive rockfish only) off Santa Barbara, Calif. The first five ranking food items with their percent volimie are listed in order for each time period. Sample size is the number of diets (fish) pooled per period; W is Kendall's "Vf" rank concordance (Tate and Clelland 1957) among seeisons for (n) total items. December-February March-May June-August September-November Species Item % Item % Item % Item % W Naples Reef: Blue rockfish Sample size 26 Sample size 24 Sample size 18 Sample size 29 20 total Tunicates 84.2 Tunicates 70.2 Kelp' 434 Hydroids 55.2 items Chaetognaths 8.1 Kelp' 13.6 Fish 232 Kelp' 31.8 0.37 Kelp' 5.8 Hydroids 93 Hydroids 18.4 Fish 11.3 Copepods 0.7 Copepods 2.1 Tunicates 3.6 Gammarid amphipods 1.5 Gammarid amphipods 0.5 Siphonophores, etc 1.9 Fish larvae 3.6 Megalops lan/ae 0.07 Kelp bass Sample size 25 Sample size 29 Sample size 17 Sample size 15 19 total Kelp' 32.9 Fish 57.6 Fish 47.4 Fish 63.4 items Tunicates 272 Tunicates 18.1 Squid 257 Hydroids 17.4 0.41* Squid 14.2 Caprellid amphipods 8.5 Kelp' 22.1 Kelp' 16.1 Eggs 8.0 Hydroids 63 Caprellid amphipods 1.6 Crabs 1.5 Fish 7.3 Kelp' 4.1 Shrimps 1.4 Tunicates 1.1 Olive rockfish Sample size 8 Sample size 39 Sample size 11 Sample size 28 14 total Fish 938 Fish 82.4 Fish 86.1 Fish 84.5 items Fish larvae 2.5 Fish larvae 4.9 Tunicates 48 Tunicates 5.4 0.51" Polychaete worms 1.9 Megalops larvae 4.0 Fish larvae 4.4 Polychaete worms 4.1 "Crustacean pieces" 0.9 Tunicates 3.5 Isopods 2 9 Mysids 2.5 Shrimps 0.5 Squid 1.0 Polychaete worms 09 Copepods 0.9 One-Mile Reef: Olive rockfish Sample size 17 Sample size 40 Sample size 10 Sample size 5 19 total Copepods 34.5 Fish 39.0 Fish 88.7 Fish 93.2 items Zoea larvae 17.6 Fish larvae 19.9 Megalops larvae 6.9 Fish larvae 3.3 0.44" Fish 149 Polychaete worms 17.9 Zoea larvae 3.6 Mysids 3.0 Pleuroncodes 14.7 Squid 5.6 Mysids 0.7 Copepods 0.1 Tunicates 13.9 Tunicates 5.1 Parasitic copepods 0.1 Zoea larvae 0.1 ' Including encrusfing bryozoans 'Significant at P = 0.05 "Significant at ; DS0.025. Table 6. — Seasonal variation in stomach fullness and interspecific dietary overlap in the three species of kelp-bed fishes in all size groups (Figure 1) from Naples Reef off Santa Barbara, Calif Stomach fullness is mean score, from 1.0 (empty) to 5.0 (full and distended). Food overlap with species in next row down is defined in the text. Stomach fullness Food overlap Species Dec.-Feb. Mar-May June- Aug. Sept. -Nov Dec.-Feb. Mar-May June- Aug. Sept.-Nov. Blue rockfish Kelp bass Olive rockfish Blue rockfish Unweighted mean 2.59 2.22 2.34 2.38 2.94 2.65 2.93 2.84 3.75 3.12 3.01 3.29 3.26 3.08 2.76 3.03 0.36 0.13 008 0.19 0 28 0.64 0.08 0.33 049 0.48 0.32 0.43 0.44 066 0.12 0.41 264 LOVE and EBELING: FOOD AND HABITAT OF THREE FISHES SEASON MAR-MAY JUN-AU8 (13) (8) BLUE RKF. KELP BASS OLIVE RKF. (NAPLES) P S N §21% S K Jll% |jl8% (7) (35) |l7% 50% 83% 33% 79% 36% 0.025 >P >QOI 0.05) P > 0.025 Figure 2. — Seasonal variation in percentage frequency of prey tjfpes (bars and numbers) in stomachs of fish in the 151- to 300-mm SL size group (Table 1) of the three species of kelp-bed fishes from Naples Reef (all three species) or One-Mile Reef (olive rockfish only) off Santa Barbara, Calif. Prey types (P-K) are designated in Figure 1; seasonal periods are explained in the text; and numbers in parenthesis aire numbers of fish stomachs examined. Hatching shows significant seasonal differences at the indicated probabilities determined by chi-square tests (see text). containing mostly nekton averaged fuller (weigh- ted means pooled among seasons = 3.12-3.48) than stomachs containing mostly other prey (2.14- 2.67). Table 7. — Numbers of the three species of kelp-bed fishes (excluding small juveniles) observed in movie strips (cinetran- secta) taken at Naples Reef or Santa Cruz Island study sites off Santa Barbara, Calif Cinetransects are classified as taken either in and about the kelp canopy or reef bottom (see text). Naples Reef Santa Cruz Island Cinetransect samples: CInetransect samples: canopy = 129, bottom = 168 canopy = 146, bottom = 185 Species Total No. in % in fish canopy canopy observed samples samples Total No. in % in fisfi canopy canopy observed samples samples Habitat Blue rockfish 3,305 2.953 89.3 919 636 69.2 Kelp bass 861 324 37.6 1,065 318 29.9 Olive rockfish 140 119 85.0 922 843 91.4 ing clear-water days over Naples Reef. Blue rock- fish often mingle with blacksmith, a specialized daytime planktivore with small mouth and com- pressed body. Blacksmith are quicker and more maneuverable than blue rockfish, which pick plankton more slowly and seem to have more difficulty repositioning themselves after feeding lunges. Small numbers of kelp bass and olive rockfish occasionally join the plankton pickers and feed at even lower rates. Although all plankton pickers may cooccur in the same field of view, they usually segregate by species. Larger individuals are usually lower in the water column. But even big kelp bass occasionally pick small particles from near the surface. All three species were more numerous over greater bottom depths (to about 12 m), where the reef-fish community is generally richer and more abundant (Table 8). Kelp bass and olive rockfish tended toward zones of greater underwater visibil- ity and kelp density, with kelp bass often prefer- ring the outer margin of the kelp bed. Both rockfishes occurred in greater numbers over high-relief rocky bottoms. Olive rockfish (juveniles and subadults) were more numerous higher in the water column. The three species occurred throughout the water column. However, most rockfish (juveniles and subadults) were recorded in canopy cinetran- sects (Table 7), and younger blue rockfish (reddish phase) usually clustered near the bottom close to shelter. In contrast, kelp bass were more abun- dant in bottom transects (Table 7). Relatively more blue rockfish and kelp bass were recorded in canopy transects over Naples Reef, where bottom and canopy tend to merge along the reef crest. One of us (Ebeling) has observed small- to medium-sized fish (ca. 100-250 mm SL) feeding together between middepth and kelp canopy dur- Table 8. — Correlations between numbers of the three species of kelp-bed fishes and environmental variables observed in an ini- tial set of 175 movie strips (cinetransects) taken over a variety of locations and subtidal habitats along ca. 24-km stretches of coastline at the mainland and Santa Cruz Island off Santa Bar- bara, Cfdif. Numbers are Kendall's tau coefficients of rank corre- lation, significant at P«0.05. Blue Kelp Olive Environmental variable rockfish bass rockfish Bottom depth 0.26 0.23 0.15 Height in water column (score) — — 0.32 Underwater visibility — 0.18 0.14 Bottom relief (score) 0.19 — 0.10 Kelp density (score) — 0.18 0.17 Toward outer margin of kelp (score) — 0.13 — Total fish numbers 0.19 0.24 0.23 Total fish species 0.40 0.31 0.20 265 FISHERY BULLETIN; VOL. 76, NO. 1 Resource Breadth and Overlap Olive rockfish from Naples Reef had the smal- lest food breadth, less than half as large as breadths of the others (Table 9). The Naples Reef fish, which occurred at relatively low density (Ta- bles 7, 10), ate mostly fish. Blue rockfish and kelp bass, whose diets were much more varied (Table 9), supplemented their fare with plankton and substrate-oriented prey. Olive rockfish from One-Mile Reef extended their diet with plankton. The kelp bass was the most widespread species both at Naples Reef and at Santa Cruz Island (Table 10). Kelp bass tended to aggregate more at Naples Reef, as indicated by a larger Hill's ratio and smaller spatial breadth. Blue rockfish were also more clumped at Naples Reef. Olive rockfish, which were relatively rare at Naples, were more evenly distributed there. In diet, the kelp bass overlapped the two rockfishes more broadly than either rockfish over- lapped the other (Table 11). The kelp bass and Table 9. — Food breadths of the three species of kelp-bed fishes in the 151- to 300-inm size group (Table 1, Figure 1) from Naples Reef or One-Mile Reef (olive rockfish only) off Santa Barbara, Calif The text defines the breadth measure B, which is based on proportionate item volumes. Sample size is the number of fish examined that had food in their stomachs; S is the number of food items eaten; and maximum % volume is of the dominant item (Table 2). Sample Maximum Dominant Species size S e % volume item Naples Reef: Blue rockfish 67 20 3.07 51.5 Tunicates Kelp bass 67 18 3.44 51.0 Fish Olive rockfish 73 14 1.40 84.2 Fish One-Mile Reef; Olive rockfish 72 19 3.32 51.0 Fish Table 10. — Spatial breadths of the three species of kelp-bed fishes from Naples Reef or Santa Cruz Island study sites off Santa Barbara, Calif The text defines the breadth measure B, which is based on proportionate abundances of the species in 297 Naples Reef or 331 Santa Cruz Island movie strips (cinetran- sects). Sample size is the total fish counted (cf. Table 7); S is the number of cinetransects in which the species was observed; and HR is a measure of concentration ( larger values indicate that more individuals are concentrated in fewer of the S cine- transects— see text). Sample Species size S B HR Blue rockfish; Naples Reef 3,305 185 42.8 1.66 Santa Cruz Island 919 151 51.0 1.61 Kelp bass: Naples Reef 861 218 65.4 1.78 Santa Cruz Island 1,065 217 90.2 1.52 Olive rockfish: Naples Reef 140 46 326 1 21 Santa Cruz Island 922 144 36.0 1.69 olive rockfish overlapped most in diet and over- lapped least in space both at Naples Reef and at Santa Cruz Island. The concordance of food and spatial breadths (Tables 9, 10) indicates that the arithmetic mean of food and spatial overlaps may be a realistic measure of total overlap in resource use (Cody 1974; Pianka 1974; Bray and Ebeling 1975). This is because concordance in breadths suggests that diet and spatial distribution may not vary inde- pendently; i.e., certain areas may be best for gathering one prey type, while other areas may be best for another. Total overlap does not vary markedly among the three species pairs because food and spatial overlaps are nearly complemen- tary (Table 11). Even so, total overlap between rockfishes is clearly less than that of either rockfish with the kelp bass. Table ll. — Overlap in food and space between members of all pairs of the three species of kelp-bed fishes from Naples Reef or Santa Cruz Island (spatial overlap only) study sites off Santa Barbara, Calif Thus food overlap, determined from dietary item volumes, and total overlap pertain only to the fish from Naples Reef Spatial overlap, determined from cinetransect fish counts, is measured separately for Naples and Santa Cruz Island fish. Paired species Blue rockfish x Kelp bass Blue rockfish « Olive rockfish Kelp bass x Olive rockfish DISCUSSION We first examine possible sources of sampling bias and how they were minimized. Then we argue that within the size range of individuals studied, the three species are indeed able to switch from one prey type to another, and that this ability is not a universal trait of fishes in general. We dis- cuss the circumstances under which the three species may change their diets and why their diets may vary from one place to another. Finally, we discuss coexistence of the three species from an evolutionary viewpoint. Sampling Bias Sport fishing activities may bias samples. Fish collected from partyboats often contain anchovies used as chum (Quast 1968d), and the mere pre- Naples Reef Santa Cruz Spatial (Sn) Total Food (F) Spatial (Sn) overlap (F+Sn/2) 043 0 .22 0.26 0.32 0.17 0.24 0.19 0.20 0.60 0.08 0.16 0.34 266 LOVE and EBELING: FOOD AND HABITAT OF THREE FISHES sence of regular sport fishing in particular areas may condition or disturb the fish fauna (Quast 1968b, c). Quast inferred that kelp bass move quickly from bare sites into more heavily foliaged, favored habitats as previous inhabitants are re- moved by fishing. In the present study, however, the influence of sport fishing was minimal because large partyboats visited Naples Reef infrequently from 1970 to 1973 (due to the erratic state of the Santa Barbara sport fishery then), and we made special effort to avoid the few skiff fishermen. Nonetheless, our samples may be biased in other ways. Quast ( 1968b) listed such sport-diving activities as shellfish gathering, which disturbs the bottom, and spearfishing among factors that condition fish behavior. Although we designed our sampling regime to minimize most hazards, we admit that spearing may induce wariness, espe- cially in kelp bass. Hence, our method of spearing fishes as they were encountered may have selected certain individuals by virtue of their size or condi- tion. Perhaps even more importantly, angling olive rockfish from One-Mile Reef, even with unbaited lures, may have selected hungrier or weakened individuals with empty stomachs. Randall (1967) noted that fish angled in tropical areas often have empty stomachs and some regurgitate their meal during the fight. Our One-Mile Reef specimens did in fact average less stomach fullness than did Naples Reef fish. But since they averaged greater intestinal fullness, they probably had been feed- ing normally. Our sampling may reflect some temporal bias. We collected most fish near midday when feeding may slacken. In the tropics, larger generalized carnivores feed mainly at dawn and dusk (e.g., Hobson 1974) or even at night if there is sufficient light (Randall and Brock 1960). In a study of kelp-bed fishes off Santa Catalina Island (ca. 160 km south of Santa Barbara) Hobson and Chess (1976) inferred that juvenile olive rockfish in the 65- to 157-mm SL range feed mostly at night. Quast (1968c) found that only 10-50*7^ of specimens of the three species collected during the day off San Diego contained food. In the present study, how- ever, most specimens contained substantial amounts of food in their stomachs, which were often more packed than their intestines. And indi- viduals were often seen feeding during the day but seldom at night, when they usually sit quietly on the bottom or hide in holes (Ebeling and Bray 1976). Similarly off central California, blue rockfish, at least, are typically active during the day (Gotshall et al. 1965; Miller and Geibel 1973). Evidence for Switch Feeding Are the three species indeed switch-feeding predators? They are certainly equipped to switch from large to small prey. All have large mouths for engulfing big items, yet have protrusible jaws and well-developed gill rakers for selecting and keep- ing small ones. In general, switch feeders show relatively weak preference for alternative prey and readily take the more abundant or otherwise more available kind (Murdoch et al. 1975). Switching mech- anisms may involve avoiding a previous prey or selecting a new one ( perhaps by acquiring a search image), spending more time in the area occupied by the new prey, or improving capture technique as the new prey becomes more abundant ( Murdoch et al. 1975). Any of these mechanisms should make individual fish specialize. We could not com- pare diets with prey density, which we did not measure. Indirectly, then, we wanted to see if a relatively large proportion of fish contain mostly one of an array of alternative kinds of prey. This seems to be the case. A fish usually con- tained mostly one and not a combination of prey types. Moreover, its stomach and intestinal con- tents usually matched, implying that it had fed on the particular prey type for a few hours (Windell 1971). Also, the percentage offish (167c) containing a single dominant food item is relatively large. It exceeds the estimated percentage (55%) for picker-type microcarnivores — small-bodied fishes with pointed, specialized mouths — which also in- habit the midwaters of the kelp bed (original data from Bray and Ebeling 1975). And it greatly ex- ceeds the small percentage (139^) for demersal microcarnivores — somewhat larger fishes (Em- biotocidae) with small mouths and fleshy lips — which usually inhabit the waters just above the reef surface (Ebeling and D. Laur in prep.). With food breadths exceeding 4.0, demersal microcar- nivores eat a diverse array of prey, but all of the substrate-oriented type, and seldom one item at a time. Fryer (1959) concluded that in Lake Nyasa (Malawi), Africa, switch feeding is easy for more generalized predatory fishes, but is difficult or im- possible for many of the more specialized species. If switching is a simple functional response (in the sense of Solomon 1949) to more of a particular 267 FISHERY BULLETIN: VOL. 76, NO. 1 prey type, fish may, e.g., switch to plankton when it is particularly dense. This implies that all switch feeders may eat mostly plankton on certain occasions and eat alternative prey on others. There did seem to be a tendency for species to eat mostly plankton during winter-spring when plankton volumes are characteristically large in this area (Smith 1971, 1974) or when other food may be relatively scarce. Yet a fish may spend more energy ingesting many plankters or tiny substrate-oriented prey than a few large prey. Quast (1968c:92) found it ". . . difficult to under- stand how the effort required to pick caprellids from kelp fronds may be rewarding to a fish as large as 200 mm SL." Reasons for Switching In the simple proximate sense, a fish should switch from a dwindling or less accessible type of prey to an increasing or more accessible type (e.g., Murdoch et al. 1975). Yet the factors that ulti- mately condition fish behavior and control food availability may be many and complex. Quast (1968b) listed predators, hunger, breeding condi- tion, water turbidity, temperature, and neighbor- ing species or conspecific individuals as such factors. Lowe-McConnell (1975) reviewed con- siderable evidence that generalized predators in tropical freshwaters eat different prey as their environment changes with time, as they occupy different geographic areas and habitats, or simply as they become able to choose among equally abundant food items in a plentiful array. There- fore, we discuss dietary variation with 1) season, 2) geographic areas and faunal mix, 3) habitat, and 4) the presence of large predators. Unlike wide-ranging, migratory fishes, the three species are limited to the food in their im- mediate environment. Tagging studies show that even adults have small home ranges. Off central ^California, juvenile blue rockfish move less than 90 m from their place of settlement unless dis- turbed by severe winter storms; adults either re- main as kelp-bed residents or migrate to deeper water and disperse more widely (Miller and Geibel 1973). Similarly, some 80% of thousands of adult kelp bass tagged off southern California were re- covered at or near the release site (Limbaugh 1955; Collyer and Young 1953; Young 1963), and but a small percentage had ventured as far as 8 km (Young 1963). Displaced individuals of Sebastes flavidus, a sibling of the olive rockfish, show re- 268 markable homing capabilities (Carlson and Haight 1972). Feeding habits of kelp-bed residents vary sea- sonally. All three species eat relatively more plankton on emptier stomachs during the cool- water seasons. Similarly, blue rockfish off central California feed less during winter and more dur- ing summer (Gotshall et al. 1965). Unlike Santa Barbara fish, however, their feeding increases during the spring upwelling season when they grow rapidly eating abundant plankton, and de- creases during the fall when they grow more slowly eating relatively more substrate-oriented prey and nekton (Miller and Geibel 1973). Like Santa Barbara fish, kelp bass off San Diego feed less during winter, when they are difficult to catch (Limbaugh 1955; Quast 1968c). Quast (1968c) concluded that feeding peaks during fall and late spring may relate to reproductive cycles. Yet in the present study, olive rockfish, which were mostly prereproductive, show the same seasonal feeding cycle as the others. Perhaps here, the sea- sonal cycle of switching among prey types simply reflects greater availability of larger or more eas- ily accessible prey when fish are most active dur- ing warmwater seasons. Seasonal variation in food overlap corroborates this. Overlap is greatest when stomachs are fullest during summer-fall, and least when stomachs are least full during winter. Zaret and Rand (1971) found that food overlap among sympatric Central American stream fishes was greatest during the food-rich wet season and least during the im- poverished dry season when intraspecific competi- tion was presumably greatest. Also, Lowe- McConnell ( 1975) summarized evidence that diets of species in large African lakes overlap most when food is abundant. Yet we have no direct evidence that smaller overlaps reflect greater competition, because we do not know when, if ever, food is limiting. Feeding habits vary geographically. Blue rock- fish seem to differ markedly in diet, distribution, and behavior between Santa Barbara and San Diego. Quast ( 1968d) noted that the few blue rock- fish sampled from a relatively sparse, marginally distributed population off San Diego (ca. 300 km southeast of Santa Barbara) had eaten little. This prompted him to suggest (1968d:132), "The blue rockfish may be poorly adapted to the envi- ronment of this region and the schools may com- prise expatriate populations." Off Santa Barbara, a denser population contains a larger size range of LOVE and EBELING: FOOD AND HABITAT OF THREE FISHES better-fed individuals. Similarly, near Monterey (ca. 300 km north of Santa Barbara), kelp beds abound with all growth stages (Miller and Geibel 1973) eating mostly plankton, but including less attached prey and more nekton as adults (Gotshall et al. 1965). Kelp bass also show differences. Compared with Santa Barbara fish, relatively more medium-sized bass from off San Diego contained clupeiform fishes (mainly anchovies, reflecting the bias due to sampling from partyboats) and motile substrate- oriented prey, such as crabs, shrimps, and am- phipods; but fewer contained plankton, algae, nonclupeiform fishes, and hydroids (Quast 1968c). Other, more cursory results (Limbaugh 1955; Young 1963) agree basically with Quast's. How- ever, Turner et al. (1969), who examined kelp bass speared from about oil platforms and other arti- ficial reefs off southern California, found, as we did, large numbers of pelagic tunicates in some individuals. These researchers saw bass eating chains of salps floating near the reefs. Bass would first bite out and ingest the viscera of large salps, then consume the tunics of the gutted prey; they swallowed small salps whole. Quast (1968c) con- cluded that larger kelp bass eat larger and more motile prey, especially fish, and ingest more kelp. Although we observed a similar trend, we have no evidence that, as Quast suggested, large bass mis- take kelp fragmented by boat propellers for fish prey. These feeding differences in kelp bass cannot be explained by distributional differences. Like San Diego fish (Limbaugh 1955; Quast 1968b, c), all sizes of Santa Barbara fish are frequently encoun- tered from surface to bottom, and prefer areas of dense kelp at the outer margins of the bed. Quast (1968b) concluded, however, that kelp bass also occupy reefs having little or no kelp. There is less information on geographic varia- tion in feeding habits of olive rockfish. South of Santa Barbara, olive rockfish and kelp bass repor- tedly cooccur and even intermingle (Quast 1968d; Turner et al. 1969), eat similar foods (Quast 1968d), and so may compete for the same cover and food (Feder et al. 1974). Off Santa Barbara, how- ever, the two may minimize interference by hav- ing a relatively small overlap in spatial distribu- tion. Considering the two species' superficial similarities in body form and color pattern, Lim- baugh (1955) suggested that olive rockfish may ecologically replace kelp bass north of Santa Bar- bara, where kelp bass dwindle in numbers (Quast 1968a; Miller and Geibel 1973). Geographic variation in a fish's feeding habits may reflect its environmental tolerances, range limits, and numbers of competitors, as well as its food supplies. Blue rockfish are more abundant off central California, kelp bass are more abundant off southern California, and olive rockfish occur abundantly in both regions but, unlike the others, are mostly restricted to Californian coastal waters (Limbaugh 1955; Quast 1968a, d; Miller and Geibel 1973). Because the Santa Barbara Channel is near the northern limit of the San Diegan fauna (Hubbs 1960; Quast 1968a), it harbors more cen- tral Californian cool-water species (Ebeling et al. 1971; Ebeling, R. Larson, and W. Alevizon in prep.). Hence all three species abound in Santa Barbara kelp forests, and here, for example, the olive rockfish may be better at capturing nekton, thus reducing supplies for the other two. Off San Diego, on the other hand, both rockfishes may occur more sporadically (Quast 1968d) and com- pete less intensely with the more numerous kelp bass. Generally reduced planktivory off San Diego may either reflect lower average plankton densi- ties there (Smith 1971, 1974), or greater abun- dances of larger, more preferred prey. Within the Santa Barbara area, habitat differ- ences may affect prey availability and the species' feeding habits. Like most areas of reef and kelp (Feder et al. 1974; Miller and Geibel 1973), Naples Reef may provide more refuges for larger prey. So here, as suggested generally both from experi- ments (e.g., Ivlev 1961) and theoretical models (e.g., Schoener 1971; Estabrook and Dunham 1976), predators may concentrate on fewer categories of larger, preferred prey in a greater overall abundance of food. One-Mile Reef, on the other hand, appears less intrinsically productive because it is deeper than Naples Reef and supports no giant kelp. So here larger prey may occur less predictably and olive rockfish must switch to plankton, including the tiniest of items, more fre- quently. Santa Cruz Island reefs are even more complex and productive than Naples Reef (Alevi- zon 1975; D. Laur pers. commun.). Thus Santa Cruz supports larger aggregations of olive rock- fish, which tend more to segregate from equally large aggregations of blue rockfish. Finally, food and space need not be the primary factors that limit the sizes of the swdtch-feeder populations. Severe storms, disease, and predators may eliminate certain numbers of individuals. Menge and Sutherland (1976) reviewed evidence 269 FISHERY BULLETIN: VOL 76, NO. 1 that for complex communities in stable environ- ments, predators may crop prey populations below their environmental carrying capacity. Hence, only top predators must partition resources to avoid competitive exclusion. Thus if adult switch feeders are heavily exploited by sharks, marine mammals, man, etc., or young are decimated by smaller predators, the three species may have lit- tle, if any, competitive effect on one another. Evolutionary Viewpoint Ultimately, the tendency to choose different prey may be an evolutionary response to coexis- tence with a close relative. The two rockfishes, which cooccur throughout much of their ranges (Phillips 1957; Quast 1968a), may have coevolved their divergent food habits. Most species of rockfish are spiny types that sit on the bottom and/or live in deep water (Phillips 1957). How- ever, the blue and olive rockfishes are members of a derived group of related species that have smoother, more streamlined bodies and inhabit the entire water column. Extending its distribu- tion from bottom to surface, the common ancestor of this species group could eat plankton and sur- face nekton as well as benthic prey. Such an ances- tor would have the ability to hunt in open water and exploit all three prey types by evolving a more streamlined morphotype. Then, during the pro- cess of speciation within the group, the blue and olive rockfishes may have themselves diverged in food habits as might be expected of two cooccuring congeners (e.g., Mayr 1963; MacArthur 1972). Thus even if their numbers are not limited by predators or other disturbances, the three super- ficially similar species may coexist by partitioning resources. As a more distantly related serranid, the kelp bass broadly shares the food spectrum with both scorpaeniform rockfishes: the plank- ton-eating and browsing blue rockfish and the fish-eating olive rockfish. Yet the kelp bass and olive rockfish have the greater dietary overlap and so tend to stay out of each others' way where both are common off Santa Barbara. And if conditions warrant it, kelp bass and olive rockfish can switch to plankton and other tiny prey although they are apparently less well adapted than blue rockfish to do so. ACKNOWLEDGMENTS We thank Richard Bray, Mark Hixon, Ralph 270 Larson, Robert Warner, and two anonymous re- viewers for critically reading the manuscript and offering a large number of helpful suggestions. Norm Lammer provided invaluable technical help with equipment and boating operations. Mary Ankeny typed several tables. This work is a result of research sponsored by NOAA, Office of Sea Grant, U.S. Department of Commerce, under grants no. 2-35208-6, 04-3-158-22 (Project R-FA- 14), and 04-6-158-44021 (R/NP-II); and by NSF Grant GA 38588 and Sea Grants GH 43 and 95. Supplementary funding was provided by a U.C.S.B. Faculty Research Committee grant (No. 369) for Computer Center user services, and by the Marine Science Institute, through the courtesy of Director Henry Offen, for interim project support. LITERATURE CITED ALEVIZON, W. S. 1975. Comparative feeding ecology of a kelp-bed em- biotocid {Enibiotoca lateralis). Copeia 1975:608-615. Bray, r. n., and A. w. ebeling. 1975. Food, activity, and habitat of three "picker-type" microcarnivorous fishes in the kelp forests off Santa Bar- bara, California. Fish. Bull., U.S. 73:815-829. BROWN, D. W. 1974. Hydrography and midwater fishes of three contigu- ous oceanic areas off Santa Barbara, California. Los Ang. Cty. Mus. Contrib. Sci. 261:1-30. Carlson, H. R., and R. E. Haight. 1 972 . Evidence for a home site and homing of adult yellow- tail rockfish, Sebastes flavidus. J. Fish. Res. Board Can. 29:1011-1014. CODY, M. L. 1974. Competition and the structure of bird communities. Princeton Univ. Press, Princeton, N.J., 318 p. COLLYER, R. D., and P. H. YOUNG. 1953. Progress report on a study of the kelp bass, Paralab- rax clathratus. Calif Fish Game 39:191-208. Ebeling, A. W., and R. N. Bray. 1976. Day versus night activity of reef fishes in a kelp forest off Santa Barbara, California. Fish. Bull., U.S. 74:703-717. Ebeling, a. W., W, Werner, F. A. DeWitt, Jr., and G. M. Cailliet. 1971. Santa Barbara oil spill: Short-term analysis of mac- roplankton and fish. U.S. Environ. Prot. Agency Water Pollut. Control. Res. Ser, 15080EAL02/71, 68 p. Estabrook, G. F., and a. E. Dunham. 1976. Optimal diet as a function of absolute abundance, relative abundance, and relative value of available prey. Am. Nat. 110:401-413. Feder, H. M., C. H. Turner, and C. Limbaugh. 1974. Observations on fishes associated with kelp beds in southern California. Calif Dep. Fish Game, Fish Bull. 160, 144 p. Fryer, G. 1959. The trophic interrelationships and ecology of some littoral communities of Lake Nyasa with especial refer- ence to the fishes, and a discussion of the evolution of a LOVE and EBELING: FOOD AND HABITAT OF THREE FISHES group of rock-frequenting Cichlidae. Proc. Zool. Soc. Lend. 132:153-281. GOTSHALL, D. W., J. G. Smith, and A. Holbert. 1965. Food of the blue rockfish Sebastodes mys- tinus. Calif Fish Game 51:147-162. HILL, M. O. 1973. Diversity and evenness: A unifying notation and its consequences. Ecology 54:427-432. HOBSON, E. S. 1974. Feeding relationships of teleostean fishes on coral reefs in Kona, Hawaii. Fish. Bull., U.S. 72:915-1031. HOBSON, E. S., AND J. R. CHESS. 1976. Trophic interactions among fishes and zooplankters near shore at Santa Catalina Island, California. Fish. Bull., U.S. 74:567-598. HUBBS, C. L. 1960. The marine vertebrates of the outer coast. Syst. Zool. 9:134-147. IVLEV, V. S. 1961. Experimental ecology of the feeding of fishes. [Translated from Russ.] Yale Univ. Press, New Haven, Conn., 302 p. LIMBAUGH, C. 1955. Fish life in the kelp beds and the effects of harvest- ing. Univ. Calif Inst. Mar. Res., IMR Ref 55-9, 158 p. LOWE-MCCONNELL, R. H. 1975. Fish communities in tropical freshwaters. Their dis- tribution, ecology and evolution. Longman Group Ltd., Lond., 337 p. MacArthur, R. H. 1972. Geographical ecology: patterns in the distribution of species. Harper and Row, N.Y., 269 p. Mayr, E. 1963. Animal species and evolution. Harvard Univ. Press, Camb., Mass., 797 p. Menge, B. a., AND J. P. Sutherland. 1976. Species diversity gradients: synthesis of the roles of predation, competition, and temporal heterogeneity. Am. Nat. 110:351-369. Miller, D. J., and J. J. Geibel. 1973. Summary of blue rockfish and lingcod life histories; a reef ecology study; and giant kelp, Macrocystis pyrifera, experiments in Monterey Bay, California. Calif Dep. Fish Game, Fish Bull. 158, 137 p. Murdoch, W. W., S. Avery, and M. E. B. Smyth. 1975. Switching in predatory fish. Ecology 56:1094- 1105. PEET, R. K. 1974. The measurement of species diversity. Annu. Rev. Ecol. Syst. 5:285-307. Phillips, J. B. 1957. A review of the rockfishes of California (family Scor- paenidae). Calif Dep. Fish Game, Fish Bull. 104, 158 p. PIANKA, E. R. 1974. Niche overlap and diffuse competition. Proc. Natl. Acad. Sci. 71:2141-2145. QUAST, J. C. 1968a. Fish fauna of the rocky inshore zone. In W. J. North and C. L. Hubbs (editorsi. Utilization of kelp-bed resources in southern California, p. 35-55. Calif Dep. Fish Game, Fish Bull. 139. 1968b. Estimates of the populations and the standing crop of fishes. In W. J. North and C. L. Hubbs (editors). Utili- zation of kelp-bed resources in southern California, p. 57-79. Calif Dep. Fish Game, Fish Bull. 139. 1968c. Observations on the food and biology of the kelp bass, Paralabrax clathratus with notes on its sportfishery at San Diego, California. In W. J. North and C. L. Hubbs (editors). Utilization of kelp-bed resources in southern California, p. 81-108. Calif Dep. Fish Game, Fish Bull. 139. 1968d. Observations on the food ofthe kelp-bed fishes. In W. J. North and C. L. Hubbs (editors). Utilization of kelp- bed resources in southern California, p. 109-142. Calif. Dep. Fish Game, Fish Bull. 139. Randall, J. E. 1967 . Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. (Miami) 5:665-847. Randall, J. E., and V. E. Brock. I960. Observations on the ecology of epinepheline and lut- janid fishes of the Society Islands, with emphasis on food habits. Trans. Am. Fish. Soc. 89:9-16. SCHOENER, T. W. 1971. Theor>' of feeding strategies. Annu. Rev. Ecol. Syst. 2:369-404. SMITH, P. E. 1971. Distributional atlas of zooplankton volumes in the California Current region, 1951 through 1966. Calif. Coop. Oceanic Fish. Invest., Atlas 13, 16 p., 144 charts. 1974. Distribution of zooplankton volumes in the Califor- nia Current region, 1969. Calif Coop. Oceanic Fish. In- vest., Atlas 20:15-17, charts 118-125. Solomon, m. E. 1949. The natural control of animal populations. J. Anim. Ecol. 18:1-35. Tate, M. W., and R. C. Clelland. 1957. Nonparametric and shortcut statistics in the social, biological, and medical sciences. Interstate Printers and Publishers, Inc., Danville, 111., 171 p. Turner, C. H., E. E. Ebert, and R. R. Given. 1969. Man-made reef ecology. Calif Dep. Fish Game, Fish Bull. 146, 221 p. WHITTAKER, R. H. 1960. Vegetation ofthe Siskiyou Mountains, Oregon and California. Ecol. Monogr. 30:279-338. WINDELL, J. T. 1971. Food analysis and rate of digestion. /wW.E.Ricker (editor). Methods for assessment offish production in fresh waters, 2d ed., p. 215-226. IBP (Int. Biol. Programme) Handb. 3. YOUNG, P. H. 1963. The kelp bass {Paralabrax clathratus) and its fishery, 1947-1958. Calif Dep. Fish Game, Fish Bull. 122, 67 p. Zaret, t. M., AND A. S. Rand. 1971. Competition in tropical stream fishes: Support for the competitive exclusion principle. Ecology 52:336- 342. 271 SCOMBEROMORUS BRASILIENSIS, A NEW SPECIES OF SPANISH MACKEREL FROM THE WESTERN ATLANTIC Bruce B. Collette,^ Joseph L. Russo,^' ^ and Luis Alberto Zavala-Camin^ ABSTRACT Scomberomorus brasiliensis is most closely related to S. sierra of the eastern tropical Pacific and more distantly related to S. maculatus £tnd S. regalis of the western Atlantic and to S. tritor of the eastern Atlantic. It differs from all four of these species in having a shorter ptelvic fin (3.6-5.9% fork length, x 4.53 compared with 4.4-7.1% in the other four species, means 5.07-5.71). Scomberomorus brasiliensis differs sharply from S. maculatus with which it has previously been confused in having fewer vertebrae (47-49 compared with 50-53). Scomberomorus brasiliensis is a more southern species than S. maculatus , occurring along the Atltmtic coasts of Central and South America from Belize to Rio Grande do Sul, Brazil, while S. maculatus is confined to the Gulf of Mexico and the Atlantic coast of the United States. RESUMO Scomberomorus brasiliensis e uma especie estreitamente relacionada com S. sierra, do Pacifico Orien- tal Tropical, tendo tambem relagao com S. maculatus e S. regalis, do Atlantico Ocidental e S. tritor, do Atlantico Oriental. Diferencia-se dessas quatro especies por ter a nadadeira ventral de menor tamanho (3,6-5,9% do comprimento zoologico, x 4, 53, comparado a 4,4-7, 1% nas outras quatro especies, que tem medias de 5,07 a 5,71). Scomberomorus brasiliensis difere claramente de S. maculatus, com a qual foi confundida anteriormente, por apresentar menor numero de vertebras (47-49 comparado a 50-53). Scomberomorus brasiliensis ocorre na costa Atlantica da America Central e America do Sul, desde Belize (Honduras britanica) ate o Rio Grande do Sul (Brasil), enquanto S. maculatus esta confinada ao Grolfo de Mexico e a cop.ta Atlantica dos Estados Unidos. While revising the tribe Scomberomorini (Collette and Russo in prep.), two apparently undescribed species of Scomberomorus were discovered, one from Australia and New Guinea and the other from the Atlantic coasts of Central and South America. Because completion of the revision will be delayed and because Atlantic Spanish macker- els are of recreational and commerical fishing con- cern, we describe the Atlantic species herein. Naming of this species adds one to the currently recognized (Rivas 1951; Mago Leccia 1958) three species of western Atlantic Scomberomorus — the king mackerel, S. cavalla (Cuvier), Spanish mackerel, S. maculatus (Mitchill); and cero, S. regalis (Block). 'Systematics Laboratory, National Marine Fisheries Service, NOAA, National Museum of Natural History, Washington, DC 20560. ^Department of Biological Sciences, George Washington Uni- versity, Washington, DC 20007. ^Divisao de Pesca Maritima, Institute de Pesca, Coor- denadoria de Pesquisa de Recursos Naturals, Santos, Brasil. METHODS AND MATERIALS The methods of counting, measuring, and dis- secting are those used by Gibbs and Collette ( 1967) in revising Thunnus and by Collette and Chao { 1975) in revising the Sardini. Extensive anatomi- cal data on the undescribed species will be pre- sented in a future revision of the Scomberomorini by Collette and Russo. Only relevant diagnostic characters plus standard descriptive meristic and morphometric data are presented here. Statistical tests were performed on the IBM 370-148 compu- ter'* at the George Washington University using computer programs written for the revision of the genus Scomberomorus following the statistical methods presented by Zar (1974). Material examined is in the following collec- tions: ANSP (Academy of Natural Sciences, Philadelphia); BMNH (British Museum, Natural History, London); CAS (California Academy of Manuscript accepted July 1977. FISHERY BULLETIN: VOL. 76. NO. 1, '•Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 273 1978. FISHERY BULLETIN: VOL. 76, NO 1 Sciences, San Francisco); FMNH (Field Museum of Natural History, Chicago); LACM (Los Angeles County Museum of Natural History); MCZ (Museum of Comparative Zoology, Harvard Uni- versity); MNHN (Museum National d'Histoire Naturelle, Paris); MPIP (Museu de Pesca do In- stituto de Pesca, Santos); MZUSP (Museu de Zoologia da Universidade de Sao Paulo); NHMV (Naturhistorisches Museum, Vienna); NMC (Na- tional Museum of Canada, Ottawa); RMNH (Rijksmuseum van Natuurlijke Historie, Leiden); ROM (Royal Ontario Museum, Toronto); SIO (Scripps Institution of Oceanography, La Jolla); SU (Stanford University, specimens now at CAS); UDONECI (Universidad de Oriente, Nueva Es- parta, Centro de Investigaciones, Venezuela); UF (Florida State Museum, University of Florida, Gainesville); UMML (Rosenstiel School of Atmos- pheric and Marine Science, University of Miami); USNM (National Museum of Natural History, Washington, D.C.); ZMA (Zoological Museum, Amsterdam); ZMH (Zoologisches Institut und Zoologisches Museum, Hamburg); ZMK (Zoologi- cal Museum, Copenhagen). SERRA SPANISH MACKEREL S comber omor us brasiliensis n.sp. Diagnosis. — A spotted species of Spanish mack- erel without a dip in the lateral line, without scales covering the pectoral fins, with a moderate number of vertebrae (47-49, usually 48) and gill rakers (12-16, usually 13-15), and with a short pelvic fin (3.6-5.9'7f FL). Scomberomorus brasiliensis is most closely re- lated to S. sierra Jordan and Starks of the eastern tropical Pacific and more distantly to S. maculatus (Mitchill) of the western Atlantic and to S. tritor (Cuvier) of the eastern Atlantic. It differs from S. maculatus in having fewer vertebrae (47-49 com- pared with 50-53, Table 1). Morphometrically, S. brasiliensis differs from its four closest relatives in having a much shorter pelvic fin (Figure 1): 3.56-5.86, X = 4.53% FL com- pared with S. sierra (4.71-6.37, x = 5.51), S. maculatus (4.59-5.76, x = 5.71), S. tritor (4.97- 7.14, X = 5.07), and S. regalis (4.41-6.33, 3c = 5.54). The linear regression ofpelvic fin length on fork length was tested by analysis of covariance. The slopes of the regression lines of all five species were not significantly different at the 0.01 level of significance (Table 2). The elevations of the five regression lines were significantly dif- ferent at the 0.001 level (P< 1.0 X 10 ■'). The Stu- dent Newman-Keules multiple range test indi- cates that the elevations of the regression lines for S. sierra, S. maculatus, S. tritor, and S. regalis were not significantly different (P>0.2); however, S. brasiliensis was found to be different from the other four species (P<0.001). Data from S. maculatus, S. sierra, S. tritor, and S. regalis were resubmitted to analysis of covariance after re- moval of S. brasiliensis. This reduced the calcu- lated F from 54.72 to 6.79 showing that most of the variance was caused by inclusion of S. brasiliensis with four other more or less homogeneous species. Table 2. — Regression equations of pelvic fin length on fork length for five species of Scomberomorus. Coefficient of Species N Y intercept Slope determination (r^) S. tritor 30 2.137 0.053 0.965 S brasiliensis 49 -0.013 0.045 0.963 S sierra 50 1,510 0,051 0-957 S. maculatus 32 1.029 0.054 0.960 S. regalis 37 1.179 0.051 0.933 Description. — Lateral line without a prominent dip in the region of the second dorsal fin (present only in S. cavalla among American and Atlantic Table l. — Numbers of precaudal, caudal, and total vertebrae in five species of Scomberomorus. Species S tritor (E Atlantic) S. brasiliensis (W, Atlantic): Central and northern South America Brazil S, sierra (E, Pacific): Mexico Central and South America S, maculatus (W. Atlantic): Eastern United States Gulf of Mexico S. regalis (Caribbean) Precaudal Caudal Total 18 19 20 21 22 N 26 27 28 29 30 31 32 N 45 46 47 48 49 50 51 52 53 N 6 30 24 2 59 2 14 16 36 18.8 26 20 1 62 200 18 17 23 4 27 16 2 18 5 47 2 54 20,0 20.1 21.1 21.1 19.9 2 30 4 4 21 6 53 3 12 3 13 5 36 1 13 9 10 36 27 1 6 29 1 25 62 18 17 28 18 54 27.8 280 28,0 279 30 7 30.3 28.1 3 21 8 50 2 14 2 14 4 42 5 20 3 28 51,9 19 7 1 18 51.4 9 55 48.1 36 45.9 26 48.0 62 47.9 18 48.0 17 47.9 274 COLLETTE ET AL: SCOMBEROMOROUS HRASILIENSIS NEW SPECIES 40 H X I— O z > 600 675 FORK LENGTH (mm) Figure l. — Regression of pelvic fin length on fork length in five species of Scomberomorus . The regression line for S. brasiliensis is significantly different from those for S. maculatus, S. sierra, S. tritor, and S. regalis. The regression lines for these four species do not differ significantly from each other so the same symbol is used for plotting specimens of the four species. species). Pectoral fin without scales (covered with scales in S. regalis). Vertebrae (19-21) + (27- 29) = 47-49 usually 20 + 28 = 48. Gill rakers (1- 3) + 1 + (9-12) = 11-16, usually 13-15 (Table 3) more than in S. cavalla (7-12) and many fewer than in S. concolor (21-27). Pectoral fin rays 21-24, usually 22 or 23, usually 21 in S. tritor, S. sierra, and S. maculatus , 21 or 22 in S. regalis (Table 4). Dorsal spines 17-19, usually 17 or 18; second dor- sal fin rays 17-19; dorsal finlets 8-10; anal fin rays 18-20; anal finlets 8-10, usually 8 or 9. Mor- phometric data are summarized in Table 5. Intes- tine with three limbs and two folds (Figure 2). Sides with several rows of round yellowish- bronze (in life) spots similar to S. maculatus and S. Table 4. — Number of pectoral fin rays in five species of Scomberomorus . Pectoral fin rays Species 20 21 22 23 24 N X S. tritor (E. Atlantic) 1 19 9 29 21.3 S. brasilierisis (W, Atlantic): Central and northern South America 6 13 11 1 31 22.2 Brazil 2 24 14 40 22.3 S sierra (E. Pacific): Mexico 7 16 7 30 21,0 Central and South America 4 18 7 1 1 31 21.3 S. maculatus (W. Atlantic): Eastern United States 2 13 5 20 21.2 Gulf of Mexico 13 1 14 21.1 S. regalis (Caribbean) 15 17 4 36 21.7 sierra but without any lines or streaks such as are present in S. regalis. The number of yellowish- TABLE 3. — Numbers of upper, lower, and total gill rakers on the first gill arch in five species of Scomberomorus . Upper Lower Total Species 1 2 3 4 N X 9 10 11 12 13 14 N X 11 12 13 14 15 16 17 18 N X S. tritor (E. Atlantic) 25 6 31 2.2 7 18 3 3 31 11.1 6 14 8 3 31 13.3 S, brasiliensis (W. Atlantic): Central and northern South America 14 7 21 2.3 1 2 10 8 21 112 1 2 6 9 3 21 13.5 Brazil 1 68 35 104 23 11 47 37 8 103 11.4 7 42 29 21 4 103 13.7 S. sierra (E. Pacific): Mexico 9 20 4 33 2.8 11 15 6 1 33 11.9 4 10 10 8 1 33 14.8 Central and South America 2 31 1 34 3.0 1 5 13 13 2 34 127 1 5 14 12 2 34 15.3 S. maculatus (W. Atlantic): Eastern United States 13 6 1 20 2.4 2 8 7 2 1 20 10.6 2 5 7 4 1 1 20 13.0 Gulf of Mexico 1 13 14 1.9 2 8 3 1 14 11.2 2 9 2 1 14 13.1 S. regalis (Canbbean) 8 26 4 38 2.9 1 4 12 17 4 38 12.5 1 2 2 16 11 5 1 38 15.4 275 FISHERY BULLETIN: VOL. 76, NO. 1 Table 5. — Summary of morphometric data of Scomberomorus brasiliensis expressed as percent fork length. Character N X Min Max SD Snout-anal distance 51 53.80 51.21 69.19 2.67 Snout-second dorsal distance 51 51.13 48.31 67 21 264 Snout-first dorsal distance 51 24.09 19.65 33.69 2 10 Snout-pelvic distance 51 25.15 21.75 35 86 221 Snout- pectoral distance 51 21.93 1908 31.71 2 12 Pectoral-pelvic distance 50 10.83 8.75 15.95 1.12 Head length 51 21 20 12 12 3090 2.34 Maximum depth 51 1986 16.40 26.31 1.57 Maximum width 49 7.99 5.42 11 38 1.09 Pectoral length 51 12 29 966 1432 1.00 Pelvic length 49 4.53 356 5.86 0.41 Pelvic insert-vent 49 27.51 23.87 34 86 1.65 Pelvic tip-vent 46 2283 19.72 29.82 290 Base first dorsal 50 2651 23.16 36.04 1.82 Height second dorsal 48 11 58 9.19 13.94 1.11 Base second dorsal 51 11 86 10.05 15.32 097 Height anal 48 11 30 8.27 14.86 1 18 Base anal 50 11 20 974 14.23 094 Snout (fleshy) 51 8 18 6.88 11.98 087 Snout (bony) 51 7.31 6.16 10.18 0.70 Maxillary length 50 1223 8.16 1883 1 50 Post orbital distance 51 948 8.43 12.70 0.69 Orbital (fleshy) 51 3.73 2.70 5.74 0.62 Orbital (bony) 51 5.27 4.15 7.66 0.71 Interorbital distance 51 566 4.78 10.65 0.83 Second dorsal-base caudal peduncle 50 49.34 42.75 59.37 3.29 SPLEEN GALL BLADDER GONAD LIVER CAECAL MASS INTESTINE 9- .ANUS Figure 2. — Ventral view of viscera in Scomberomorus brasiliensis, 556 mm FL, Belem, Brazil, dissected 17 June 1975. bronze spots on the sides of the body increases with the size of the fish, young specimens (200 mm) have about 30 spots; adults more, 45 spots (422 mm), 47 (455), 46 (470), 45 (516), and 58 (530). The spots are arranged in 3 or 4 rows (sometimes in 2 rows). The rows are not very well defined but it is possible to recognize them. The spots in S. maculatus are not arranged in rows. The first dorsal fin is black in the anterior half (first 7 membranes) and the posterior half is white with the upper edge black. Pectoral fin dusky; pelvic and anal fins white. In young specimens (192-240 mm) (collected from estuarine waters) the caudal and pectoral fins are yellow (in the pectoral fin, yellow over the dusky color) and the whole body and the anal fin are slightly yellow. Range. — Atlantic coasts of Central and South America from Belize at least as far south as Lagoa Tramandai, Rio Grande do Sul, Brazil (Figure 3). Not known to overlap with S. maculatus which occurs in the Gulf of Mexico and along the Atlantic coast of the United States. Replaced in the West Indies by S. regalis. Material examined. — Inasmuch as there is abun- dant material from Brazil, and because further study might show some differentiation within the range of S. brasiliensis, the type-material is re- stricted to the specimens examined from Brazil. Holotype. —USNM 217550 (502 mm FL); Belem market; 22 May 1975; B. B. Collette 1642. Paratypes. — 103 specimens (110-630 mm FL) from 54 Brazilian collections. USNM 217551-57 (7, 509-588); Belem market; May 1975; B. B. Collette 1639, 1642, and 1645. MCZ 17131 (1, 220); Para (= Belem). NHMV uncat. (2, 410-538); Para; Brasil Exped.; Steindachner. NHMV uncat. (1, 325); Maranhao; Brasil Exped.; 1903. USNM 188424 (3, 153-281); Oregon II stn. 4250, 2°23'S, 40°31'W; 12 Mar. 1973. CAS-SU 52981 (1, 483) Ceara, Fortaleza, Mucuripe. CAS-SU 52989 (1 359); Ceara, Fortaleza. CAS-SU 52988 (1, 300) Ceara, Fortaleza. CAS-SU 52987 (1, 220); Ceara Fortaleza. MZUSP 13263-4 (2, 375-405); Ceara; May 1976. MPIP 0001-2 (2, 354-380); Ceara; May 1976. MZUSP 13262 (1, 385); axial skeleton; Rio Grande do Norte; Feb. 1976. CAS-SU 52971 (1, 340); Pernambuco, Recife. CAS-SU 52973 (1, 236); Pernambuco, Recife. MCZ 48894 (2, 392-412); Re- cife market; Equalant Exped.; Chain; R. H. Bakus; 276 COLLETTE ET AL: SCOMBEROMOROUS BRASILIENSIS NEW SPECIES 95" »• 85- 80" 75- 70- 65* W 55" 50* 45' W 35' 30" raiiia^i„m.,i„m |iiM„i„inM,i;,intniMiMrnM (gm,jn>ni.,i„m.,i„m„i„M m m,,iMUninUM.,i„hiMJ„MT!iMi„titii,,i,,mjMmj,,m,,inMMi,,mMLMrf^^ MiiiMi„mMiMm mMiMiMtiiiMiMUniMinm„i..fcu„i..m„i,;t^ O S. brasiliensis • S. maculatus * S. sierra ^2j^>.cz>- \ Bloef.elcJs* ,^-' --^'^-. "mr An^fso^^ \. r»' I|iillllijn|li|illll|li|ii^a25' I'T'lillM'T'lM'T'imy Hu!ii''l"l Figure 3. — Distribution of Scomberomorus brasiliensis (stars in circles) and adjacent populations of S. maculatus (dots) and S. sierra (stars). The ranges of S. maculatus and S. sierra extend farther north and that of S. brasiliensis farther south. 3 Mar. 1963. NHMV uncat. (1, 378); Pernambuco; Brasil Exped.; Steindachner; 1904. CAS-SU 52983 (1, 403); Bahia, Salvador, Sobura. MZUSP 13265-7 (3, 278-319); Bahia; Dec. 1976. MPIP0003 (1, 292); Bahia; Dec. 1976. MPIP 0004 (1, 407); axial skeleton; Bahia; Jan. 1977. MZUSP 13628-9 (2, 283-354); axial skeleton; Jan. 1977. CAS-SU 52972 (1, 196); Espirito Santo, Vitoria. MZUSP 13270 (1, 483); Espirito Santo; Dec. 1976. MZUSP 13271-2 (2, 424-462); axial skeleton; Espirito Santo; Dec. 1976. MPIP 0005 ( 1, 477); axial skele- ton; Espirito Santo; Jan. 1977. MCZ 877 (3, 270- 300); Rio de Janeiro. MCZ 17261 (1, 252); Rio de Janeiro. MCZ 17236 (8, 234-307); Rio de Janeiro. MCZ 23802 (1, 630); Rio de Janeiro. BMNH 1896.6.29.9 (1, 480); Rio de Janeiro; Capt. Milner. BMNH 1923.7.30.305 (1, 395); Rio de Janeiro; Ternetz. ZMH 4029 (1, 282); Rio de Janeiro; 1885. NHMV 1874.1.532a (2, 253-284); Rio de Janeiro; Steindachner. NHMV 76740 (3, 255-286); Rio de Janeiro; 1857-59. MZUSP 13273-6 (4, 251-420); axial skeleton; Rio de Janeiro; Jan. 1976. MPIP 0006 (1, 435); Rio de Janeiro; May 1976. MPIP 0007-8 (2, 372-374); Rio de Janeiro; June 1976. MZUSP 13277 (1, 394); Rio de Janeiro; June 1976. CAS-SU 52985 (1, 490); Sao Paulo, Santos. CAS-SU 52984 (1, 383); Sao Paulo, Santos. MZUSP 878 (1, 110); Sao Paulo; Miranda Ribeiro; 1913. MZUSP 13279-80 (2, 304-365); axial skele- tons; Sao Paulo; Dec. 1976. MZUSP 13281-8 (8, 187-203) axial skeletons; Sao Paulo, Cananeia; Feb. 1977. MZUSP 13289 (1, 240); Sao Paulo, Cananeia; Feb. 1977. MPIP 0009-12 (4, 196-201); axial skeletons; Sao Paulo, Cananeia; Feb. 1977. MZUSP 13278 (1, 353); Sao Paulo; July 1977. MZUSP 13290-1 (2, 340-450); axial skeletons; 277 FISHERY BULLETIN: VOL. 76, NO. 1 Santa Catarina; Aug. 1976. MPIP 0013 (1, 425); Santa Catarina; Aug. 1976. MZUSP 1329-30 (2, 405-600); Santa Catarina; Dec. 1976. MPIP 0014 ( 1, 405); Santa Catarina; Dec. 1976. MZUSP 13294 (1, 372); axial skelton; Santa Catarina; Jan. 1977. CAS-SU 52986 (1, 416); Rio Grande do Sul. MZUSP 13295-6 (2, 240-245); Rio Grande do Sul. Lagoa Tramandai; May 1977; MCZ 17158 (4, 136- 216); Brazil. Other material. — 28 specimens (111-520 mm FL) from 15 collections arranged here by country from north to south. BELIZE: BMNH 1864.1.26.304-5 (2, 217-230); Salvin. HONDURAS: UF-TABL 67-106 (1, 243); 15°21'N, 83°34'W; 10 Apr 1967. Costa Rica: 3(172-194) from 2 collections LACM 30727-13 (2, 191-194); Canuita Bay; W Bussing and party. LACM 30726-3 (1, 172) Canuita Bay; W. Bussing and party. PANAMA 4(114-225) from 2 collections. ANSP 86721 (1, 225); Balboa; 5th G. Vanderbilt Exped. ; 1 1- 14 Apr. 1941. ANSP 45270 (3, 114-182); Colon market; D. E. Hanover; June 1945. COLOMBIA: USNM 217433 (1, 326); Choco cruise 6908, stn. 127, 9°22.1'N, 75°36.4'W; 6 Sept. 1969. VENEZUELA: 9(89-520) from 3 collections. ZMA 1 14.581 ( 1, 520); Puerto Cabello; 10 Aug. 1905. USNM 121802 (2, 296-330); Maracaibo market; L. P. Schultz; 15 May 1942. UDONECI 1071 (6, 89-198); Peder- nales; 3 July 1974. TRINIDAD: 8(260-311) from 5 collections. BMNH 1931.12.5.173 (1, 260); Gulf of Paria; Totten, Rodney. ANSP 94311 (2, 278-311); Brighton Pier; L. Wehekind; 10 May 1930. ANSP 94325 (2, 280-287); Brighton Pier No. 2; L. Wehekind; 7 May 1930. ANSP 94329 ( 2, 268-289); Brighton Pier No. 2; L. Wehekind; 17 May 1930. UF-TABL uncat. (1, 233); M/V Calamar cruise 67-B, stn. 260; 13 Nov. 1967. SURINAM: RMNH 24764(1, 111). DISCUSSION Although it is a common fish, Scomberomorus brasiliensis has not been formally described be- cause adults closely resemble S. maculatus in their spotted pattern. The juveniles are similar to S. regalis in having low vertebral counts (47-49) and have probably been confounded with that species (which is actually uncommon in the range of S. brasiliensis off the coasts of Central and South America). A fairly extensive literature pertains to S. brasiliensis (as S. maculatus) dating back to 278 Ribeiro ( 1915). Particularly important are a series of 30 papers on various biological and fisheries aspects of S. brasiliensis from Laboratorio de Ciencias do Mar da Universidade Federal do Ceara at Fortaleza, Brazil. Bastos (1966) sum- marized morphometric and meristic data for 90 specimens ( 163-553 mm FL). His gill raker counts (usually 2 + 1 + 11 = 14 or 3 + 1 + 11 =15) agree closely with ours (Table 3). His vertebral counts (26 specimens wdth 46 and 55 specimens with 47) are 1 less than ours (Table 1) because he presumably did not include the hypural plate in his counts as we did. Menezes ( 1972) also counted gill rakers and found no differences between counts for 225 males and 275 females; the most typical count was 3 -I- 1 + 11 = 15. The digestive tract was studied both grossly and histologically by Mota Alves ( 1969). The histology of the pyloric caeca of S. brasiliensis was found similar to that found in S. cavalla by Mota Alves and Tome ( 1970). The pyloric caeca were found to contain the same enzymes as the intestine in both species — lipase, maltase, and trypsin but the pyloric caeca in S. brasiliensis also contained pep- sin which was restricted to the stomach in S. cavalla. The food of S. brasiliensis in the State of Ceara was studied around the year by Menezes (1970). Fish composed the major part of the diet; penaeid shrimps and loliginid cephalopods also were im- portant. The most important fishes were, in order: Opisthonema oglinum, Engraulidae, Chloroscom- brus chrysurus, Hemiramphus sp., and Haemulon spp. The diet of S. maculatus in southeastern Florida is similar to this according to Klima (1959), consisting mostly of clupeids (especially Harengula pensacolae) plus Penaeus, engraulids, and other fishes. Mota Alves and Tome (1968a) reported on the sexual development of S. brasiliensis and recog- nized five developmental stages in the ovary. They also (1968b) described the sperm. Gesteira (1972) found that females first become sexually mature at about 460 mm FL at an age of III or IV. She presented equations for calculating fecundity based on length, age, and weight. Klima (1959) found that the smallest mature female S. maculatus from southeastern Florida was 250 mm FL and that both sexes matured at age I or II. Length-frequency data for S. brasiliensis (and S. cavalla) were collected and published annually, starting vdth the data for 1962 and continuing through 1969 by Costa and Paiva and then for COLLETTE ET AL: SCOMBEROMOROUS BRASILIENSIS NEW SPECIES 1971-73 by Costa and Almeida ( 1974). For 32,514 specimens of S. brasiliensis measured from 1962 through 1973, the size range was 267-1,250 mm FL. Of 16,170 specimens measured between 1962 and 1968, 9 were longer than 950 mm FL: 6 (951- 1,000); 1 (1,001-1,050); 1 (1,051-1,100); and 1 (1,201-1,250). More than 60% each year from 1962 to 1968 were in the size range 401-650 mm FL. Scomberomorus maculatus is a much smaller species; the largest of 1,279 specimens examined by Klima ( 1959) from southeastern Florida was 700 mm FL and most were between 300 and 500 mm. The length-weight relationship was deter- mined by Nomura and Costa ( 1968) for Brazilian S. brasiliensis: for males log W = -2.2051 + 2.973 log L, and for females log W = -2.154 + 3.035 log L. For 1971-73, the age composition of the catch was II to X, concentrated at III to VI, and mostly III or IV (Costa and Almeida 1974). Color pattern, possession of nasal denticles, lat- eral line curvature, and other characters suggest that S. brasiliensis, S. sierra, S. maculatus, and S. tritor are closely related. Scomberomorus regalis may also belong to this group of species and S. concolor Lockington of the eastern tropical Pacific is even more distantly related. Scomberomorus cavalla belongs to another species group, contain- ing S. commerson (Lacepede). The center of origin of Scomberomorus appears to be in the Indo-West Pacific as is the case with many other groups of fishes. It appears likely that an ancestor of S. tritor crossed the Atlantic and populated the tropical western Atlantic and eastern Pacific. When the Panamanian isthmus emerged, this population was divided into two, which subsequently dif- ferentiated into S. sierra (eastern Pacific) and S. brasiliensis. Scomberomorus maculatus is pre- sumably derived from the S. sierra-brasiliensis stock and developed a higher number of vertebrae along with its movements into more temperate waters along the U.S. east coast. COMPARATIVE MATERIAL EXAMINED Scomberomorus maculatus. East coast of United States: 24 specimens (163-712 mm FL) from 13 collections from Cape Cod, Mass.; New York; Cape Hatteras, N.C.; Charleston, S.C.; and Brunswick, Ga., at MCZ, NHMV, USNM, ZMH, and ZMK. Gulf of Mexico: 29 specimens (176-439 mm FL) from 16 collections from Captiva Key and St. Andrew Bay, Fla; Mobile, Ala.; Biloxi, Miss.; Atchafalaya Bay, La.; Aransas Bay, Tex.; Vera Cruz, Mexico; and Progreso, Yucatan, Mexico at BMNH, MCZ, NHMV, USNM, and ZMK. Scomberomorus sierra . SIO 62-338 ( 1 , 594 ) , La Jolla, Calif. Mexico: 42 specimens (183-685 mm FL) from 22 collections from Baja California, Guaymas, Mazatlan, and Sonora at BMNH, CAS, LACM, MCZ, NHMV, SIO, and USNM including the lectotype SU 1720 (332 mm) from Mazat- lan. Costa Rica: 6 specimens (237-515 mm FL) from 4 collections from Golfo Dulce and Golfo Nicoya at LACM. Panama: 15 specimens (226- 605 mm FL) from 8 collections at FMNH, MCZ, NHMV, SU, USNM, and ZMH. Colombia: 8 specimens (202-260 mm FL) from Buenaventura at USNM. Peru: 4 specimens ( 135-460 mm FL) from 3 collections at LACM and NHMV. Galapagos: 4 specimens (422-621 mm FL) from 3 collections at ANSP, CAS, and NMC. Scomberomorus tritor. Mediterranean: 2 specimens (365-475 mm FL) from Nice at NHMV and Florence. Gulf of Guinea: 36 specimens (69- 600 mm FL) from 25 collections from the Canary Islands, Senegal, Sierra Leone, Liberia, Cote d'lvoire, Ghana, Nigeria, and Angola at BMNH, CAS, MCZ, MNHN, NHMV, USNM, and ZMA including the holotype MNHN A. 6871 from Goree, Dakar. Scomberomorus regalis. Caribbean: 40 speci- mens (77-525 mm FL) from 27 collections from Florida, Bahamas, Cuba, Haiti, Jamaica, Puerto Rico, Virgin Islands, Lesser Antilles, and Bar- bados at BMNH, MCZ, NHMV, ROM, USNM, ZMA, ZMH, and ZMK. ACKNOWLEDGMENTS Material was examined through the courtesy of M.L. Bauchot (MNHN), J. E. and E. B. Bohlke (ANSP), M. Boeseman (RMNH), F. Cervigon M. (UDONECI), A. R. Emery (ROM), W. N. Esch- meyer (CAS), C. Gilbert (UF), K. Hartel (MCZ), R. K. Johnson (FMNH), P. Kahsbauer (NHMV), R. Lavenberg (LACM), K. F. Liem (MCZ), D. McAl- lister (NMC), N. A. Menezes (MZUSP), J. Nielsen (ZMK), H. Nijssen (ZMA), C. R. Robins (UMML), R. Rosenblatt (SIO), P. Sonoda (CAS), P. J. P. Whitehead (BMNH), and H. Wilkins (ZMH). Wil- liam J. Richards (Southeast Fisheries Center, Na- tional Marine Fisheries Service, NOAA, Miami, Fla.) has kindly made a series of vertebral counts of Atlantic species of Scomberomorus available to us. George Clipper X-rayed most of the material and read the radiographs. The figures were com- 279 FISHERY BULLETIN: VOL. 76, NO. 1 pleted from our sketches by Keiko Hiratsuku Moore. Drafts of the manuscripts were read by Mark E. Chittenden, Daniel M. Cohen, Eugene L. Nakamura, William J. Richards, and Naercio A. Menezes. LITERATURE CITED Bastos, J. R. 1966. Sobre a biometria da serra, Scomberomorus maculatus (Mitchill), da costa do Estado do Ceara. Arq. Estac. Biol. Mar Univ. Fed. Ceara 6:113-117. COLLETTE, B. B., AND L. N. CHAO. 1975. Systematics and morphology of the bonitos (Sarda) and their relatives (Scombridae, Sardini). Fish. Bull., U.S. 73:516-625. COSTA, R. S. da, and H. T. DE ALMEIDA. 1974. Notas sobre a pesca da cavala e da serra no Ceara - Dados de 1971 a 1973. Arq. Cienc. Mar 14:115-122. Gesteira, T. C. V. 1972. Sobre a reprodufao e fecundidade da serra, Scom- beromorus maculatus (Mitchill), no Estado do Cea- ra. Arq. Cienc. Mar 12:117-122. GIBBS, R. H., Jr., and B. B. COLLETTE. 1 967 . Comparative anatomy and systematics of the tunas , genus Thunnus. U.S. Fish Wildl. Serv., Fish. Bull. 66:65-130. KLIMA, E. F. 1959. Aspects of the biology and fishery for Spanish mack- erel, Scomberomorus maculatus (Mitchill), of southern Florida. Fla. State Board Conserv., Tech. Ser. 27, 39 p. Mago Leccia, F. 1958. The comparative osteology of the scombroid fishes of the genus Scomberomorus from Florida. Bull. Mar. Sci. GulfCaribb. 8:299-341. MENEZES, M. F. DE. 1970. Alimentagao da serra, Scomberomorus maculatus (Mitchill), em aguas costeiras do Estado do Ceara. Arq. Cienc. Mar 10:171-176. 1972. Niimero de rastros da serra, Scomberomorus maculatus (Mitchill), das aguas costeiras do Estado do Ceara. Arq. Cienc. Mar 12:86-88. MOTA ALVES, M. I. 1969. Sobre o trato digestivo da serra, Scomberomorus maculatus (Mitchill). Arq. Cienc. Mar 9:167-171. MOTA ALVES, M. I., AND G. DE SOUSA TOME. 1968a. Observagoes sobre o desenvolvimento maturativo das gonadas da serra, Scomberomorus maculatus (Mitch- ill, 1815). Arq. Estac. Biol. Mar Univ. Fed. Ceara 8:25- 30. 1968b. Algumas observagoes sobre o semen da serra, Scomberomorus mnculatus (Mitchill). Arq. Estac. Biol. Mar Univ. Fed. Ceara 8:139-140. 1970. On the pyloric caeca in fishes of the genus Scom- beromorus Lacep'ede. Arq. Cienc. Mar 10:181-184. Nomura, H., and R. S. da Costa. 1968. Length-weight relationship of two sp>ecies of Scom- bridae fishes from Northeastern Brazil. Arq. Estac. Biol. Mar Univ. Fed. Ceara 8:95-99. RIBEIRO, A. DE MIRANDA. 1915. Fauna Brasiliense - Peixes. V Eleuterobranchios, Aspirophorus (Physoclisti). Arch. Mus. Nac. Rio de J., 679 p. RIVAS, L. R. 1951. A preliminary review of the western North Atlantic fishes of the family Scombridae. Bull. Mar. Sci. Gulf Caribb. 1:209-230. Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall, Inc., Eng- lewood Cliffs, N.J., 620 p. 280 NOTES AGGREGATION OF THE SIPHONOPHORE NANOMIA CARA IN THE GULF OF MAINE: OBSERVATIONS FROM A SUBMERSIBLE Large concentrations of a physonect siphono- phore, Nanomia cara Agassiz 1865, were present in the Gulf of Maine during fall and winter of 1975. These gelatinous, colonial coelenterates were sufficiently abundant that they clogged trawl ne' s and occasioned considerable losses of time and money to commercial fishermen at several New England ports (Rogers in press). During October and November 1975, scuba divers on the Helgo- land habitat in 30-m shoals off Rockport, Mass. (Figure 1), noted concentrations of N. cara reach- ing 1 colony/m^ throughout the water column (R. A. Cooper and H. W. Pratt unpubl. data). Off Rockport again in late March 1976, divers esti- mated densities of 1 to 2 colonies of A^. cara/m^ in CASHES LEDGE 7-o,o6 ,8r» B-s O^-. 00-30 SUBMERSIBLE DIVES LOCATIONS- JUN. 1976 ■ NANOMIA CARA OBSERVED O N CARA NOT OBSERVED 40 M 80 M Figure l. — Distribution of siphonophores at dive sites of the submersible Nekton Gamma and the position of the Helgoland habitat (insert). water only 9 m deep (H. W. Pratt pers. commun.). In April and again in early May 1976, a series of 100-m to surface oblique plankton tows was taken in the Gulf of Maine along a transect from the Wilkinson Basin to Cape Ann, Mass., by AZfta^ross IV, a fisheries research ship of the Northeast Fisheries Center. In these deeper water areas, as well, high densities of N. cara apparently per- sisted through the winter months and were present at each station occupied, although the aggregations were somewhat less numerous and colonies appeared smaller than those encountered during fall 1975 (Rogers in press). The difficulties and limitations inherent in using plankton nets to sample quantitatively populations of siphonophores and other fragile gelatinous zooplankton have been reviewed by Hamner et al. (1975), who suggested in situ scuba observations as an alternative method for study- ing gelatinous taxa. In the present study we used the two-man research submersible Nekton Gamma to estimate the size and density of the iV. cara aggregations and to evaluate some of the biotic and abiotic factors which might influence their distribution below depths easily accessible to scuba divers. In June 1976 we made six dives along a transect from Provincetown, Mass., to Cape Ann (Table 1, Figure 1). Dives were of 90 to 160 min duration during which we surveyed the water column from surface to bottom. Other ob- servers made 25 additional shorter dives to look for siphonophores at adjacent stations. Observa- tions were narrated and recorded on tape through- out each dive. The submersible pilot relayed in- formation on temperature and depth and this was combined with comments on siphonophore col- onies such as size, density, swimming speed, as- sociated species, and other factors of interest. Photographs were taken with a 35-mm camera and a video tape camera with a sound track was also used to record and verify visual observations and estimates. After each dive information was transcribed from the tapes and videotapes were reviewed and discussed by the observers. Observations Gulf of Maine surface temperatures in mid-June 1976 ranged from 12.5°C in the Wilkinson Basin 281 Table l. — Station locations of Nekton gama dives to observe depth distribution and density of the siphonophore Nanomia cara, 15-28 June 1976. Dive Position Station depth (m) Bottom temp. (°C) Depth (m) where siphonophores were obsen/ed Estimated density (no./m^) of siphonophores station Lat. N Long. W 1 4r56.4' 70°19.7' 38 7.0 2 41°56.4' 70°18.6' 33 6.6 3' 42°128' 69°54 2' 207 7.5 67,101-205 1-2 4 42°05.6' 70°12.0' 37 11.1 5 42°04.7' 70t)6.1' 24 8.0 6' 42°11.4' 70°20 1' 34 5.5 7 42°11.4' 70°21.9' 46 5.5 8' 42°12.6' 70-03, 5' 128 6.0 88-126 0.1 9 42°15.r 70°07,0' 122 6.9 10 42°18.8' 70°11.r 55 6.0 11 42°38.0' 70°27.6' 107 5.5 12 42°380' 70°28.5' 85 5.0 13 42°38.0' 70°27.6' 85 60 14' 42°28.0' 69°52.6' 201 122-128 1 15' 42°36.6' 69°58.4' 180 6.8 146-177 2-4 16 42°385' 69°58.0' 136 91-120 1-2 17 42°38 2' 70°00,5' 183 7.0 120-181 7-8 18 42°39.1' 70°00.0' 168 5.5 140-166 19 42°39.1' 69°58.9' 192 7.0 82-183 0.1 20 43°01.9' 70°05.4' 107 5.5 21 43''00.5' 70°06.5' 146 5.5 107-122 0.1 22 43°01.0' 70°04.5' 53 6.5. 23 43°45.2' 69°00.0' 58 24 43°32.0' 69°35.5' 152 6.8 149-150 1 25 43°38.6' 69°38.4' 84 6.8 26 43°43.1' 69°41.2' 51 27 43°44.1' 69°40.5- 76 9-10 28' 42°55.0' 69°00.0' 72 7.3 45-70 0.05 29 42°54.4' 69°00.0' 98 7.0 76-85 91-98 <0.1 1.1 30 42°07.r 69°50.9' 124 6.0 31 42°07.3' 69°51.1' 122 7.1 'Dives conducted by authors. (Figure 1, Station 3) to 17°C off Cape Ann (Station 15). Bottom temperatures at stations deeper than 100 m ranged from 5.5° to 7.5°C (Table 1). In gen- eral, the thermocline shoaled from about 75 m off Cape Ann to about 30 m in the Wilkinson Basin (Stations 3, 8, 14-16); the estimated zone of twilight visibility extended to about 135 m. Lat- eral visibility on most dives exceeded 5 m both above the twilight zone and below it where the lights on the submersible were used. Large numbers of A^. cara were observed at all dive stations deeper than 125 m; they were also present, though less dense, at two shallower sta- tions, 28 (72 m) and 29 (98 m) (Table 1). During daylight, siphonophores were observed only below the thermocline. No dives were made at night so it was not possible to predict if transthermocline movement occurs during expected diurnal migra- tions. They appeared to be distributed in patches both horizontally and vertically. We estimated that patch diameters ranged from 5 to 30 m. At depths where N. cara was locally abundant, col- onies could be seen out of every viewport (Figure 2). The densest concentrations often occurred be- tween 3 and 45 m above the bottom where we estimated that their densities ranged between 1 and 7 colonies/m^. At Station 29 siphonophores occurred in two distinct layers: sparsely distri- buted from 76 to 85 m where concentrations were usually <0.1 colony /m^, and more densely aggre- gated above the bottom where concentrations were about 1 colony/m^. We found no correlation between colony density and substrate type. Colonies ranged in size from 0.2 to 3.7 m when suspended in fishing posture with the stem and tentacles relaxed. In this configuration the dis- tance between adjacent stem groups ranged from 10 to 15 mm. The largest colonies had over 200 salmon-colored feeding polyps (gastrozooids) and 30 to 40 swimming bells (nectophores). Unless swimming, most colonies oriented with the apical gas-filled float (pneumatophore) and nectophores upward. The rest of the flexible stem, which ap- peared neutrally buoyant, hung in three- dimensional series of loops and arcs. In high density localizations of A^. cara, colonies of several different sizes were often present. In areas of the aggregation peripheral to the highest densities of siphonophores, however, colonies were generally small, i.e., 20 to 40 cm long. Smaller colonies were also found higher in the water col- umn than the larger ones, or occurred singly. All 282 N Figure 2. — The siphonophore Nanomia cara photographed from a viewport of the submersible: p, pneumatophore; n, nec- tophore; g, gastrozooid. An excellent schematic drawing of this species can be found in Mackie (1964). colonies of A^. cara were extremely fragile and isolated pieces of stem were not uncommon. When siphonophores came into contact with the submer- sible, their tentacles frequently adhered while stem and nectophores fragmented and floated away. Most colonies were negatively phototactic and contracted their stem and tentacles as they drifted into the radius of the submersible's lights. Con- traction usually initiated an escape swimming re- sponse. The siphonophores could move rapidly away at any orientation, and some were observed swimming with pneumatophore and nectosome pointed directly downward. We estimated that es- cape speeds exceeded 20 to 30 cm/s. The most numerous invertebrates among or ad- jacent to the densest localizations of A'^. cara were the euphausiids, Meganyctiphanes norvegica and Thysanoessa inermis; mysids, principally Neomysis americanus; and hyperiid amphipods, principally Parathemisto gaudichaudii and Hyperia galba. We observed one siphonophore which had recently ingested an euphausiid; the others had no prey of this size in their feeding polyps. Calanoid copepods, among them the large species Calanus finmarchicus and Euchaeta nor- vegica, were also locally abundant among the aggregations of Nanomia cara. Plankton samples taken in June 1976 showed that these calanoids were rich in lipids, as a heavy slick of oil droplets formed after they were preserved in 4% Formalin.^ The copepods were apparently being eaten by N. cara as fragments of siphonophores removed from the same plankton samples were distended by lipids droplets inside feeding polyps and palpons, where lipids would concentrate during digestion of prey. Discussion The density of siphonophore colonies in the Gulf of Maine was considerably greater than Barham's (1963) estimate of the abundance (300 colonies/ 1,000 m^) of a congeneric species (N. bijuga) in the San Diego Trough. Barham concluded that the gas-filled floats of A'^. bijuga were of adequate di- mensions to act as strong sound scatterers and that at these densities this siphonophore could contribute significantly to scattering layer forma- tion. Thepneumatophores of A^. bijuga arvdN. cara are similar in dimension, and aggregations of N. cara should be equally effective sound scatterers. In fact, in fall and winter 1975-76, fishermen in the Gulf of Maine reported near-bottom, dense layers of sound-reflecting organisms in areas where trawl nets were being clogged with A^. cara (F. E. Lux pers. commun.). The cause of the aggregation of N. cara in the Gulf of Maine has not been determined. It is con- ceivable that widespread reproduction of N. cara occurred in fall and winter 1975-76 and that local patterns of circulation aided in concentrating and maintaining the aggregation and prey items. It is clear, however, that localization of siphonophores like A'^. cara at densities exceeding 1 colony/m^ will interfere with commercial fishing efforts by clog- ging the meshes of nets trawled for shrimp, silver hake, and redfish. Aggregations of siphonophores •Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 283 may produce serious indirect effects as well. Biggs (1976) has shown that siphonophores like N. cara can eat prey ranging in size from zooplankton nauplii to small fish. As Fraser (1962) and Zelik- man (1969) have proposed for aggregations of other gelatinous carnivores capable of eating zoo- plankton and larval fish, areas or seasons in which siphonophores are locally abundant could conceiv- ably suffer dramatic reductions of ichthyoplank- ton. Lough^ cited indirect evidence of possible heavy predation by siphonophores upon Atlantic herring larvae based on changes in population densities and distributions of the two species dur- ing winter 1975-76 in the Nantucket Shoals- Georges Bank area. Since the Gulf of Maine his- torically has been an important commercial fishing ground, future research on interaction be- tween siphonophores and ichthyoplankton could lead to a better understanding of the regional food chain and the factors which influence year class success of ichthyoplankton. Summary Aggregations of the physonect siphonophore Nanomia cara were observed at several dive sites in the Gulf of Maine from Nekton Gama. This siphonophore occurs throughout the Gulf of Maine although the vertical and horizontal distribution is patchy. Densities as high as 1 to 7 colonies/m^ were observed. Colony length ranged in size from 0.2 to 3.7 m and most aggregations included sev- eral different sizes. Nanomia cara was negatively photoactic and initiated escape swimming re- sponse at speeds which exceeded 20 to 30 cm/s. All siphonophores were observed below the thermo- cline and generally occurred only where water depth was >128 m. Euphausiids, mysids, and hyperiid amphipods were observed among populations of siphono- phores, but we observed only one colony which had eaten prey of this size. In seasons and areas of maximum abundance, siphonophores could con- ceivably influence the success of a year class of ichthyoplankton by heavy predation as well as cause losses of time and money to commercial fishermen by clogging trawl gear. Acknowledgments We are indebted to H. Wes Pratt of the North- east Fisheries Center Narragansett Laboratory, National Marine Fisheries Service, NO A A for his many observations and to Fred Lux of the North- east Fisheries Center Woods Hole Laboratory for his helpful information. Lianne Armstrong pre- pared the illustrations. Literature Cited AGASSIZ, A. 1865. North American Acalephae. Illus. Cat. Mus. Comp. Zool. Harv. 2, 234 p. BARHAM, E. G. 1963. Siphonophores and the deep scattering layer. Sci- ence (Wash., D.C.) 140:826-828. BIGGS, D. C. 1976. Nutritional ecology of Agalma okeni and other siphonophores from the epipelagic western North Atlan- tic Ocean. Ph.D. Thesis M.I.T.-W.H.O.I. Jt. Program Biol. Oceanogr., 141 p. Fraser, J. H. 1962. The role of ctenophores and salps in zooplankton production emd standing crop. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 153:121-123. HAMNER, W. M., L. P. MADIN, A. L. ALLDREDGE, R. W. GiLMER, AND P. P. HAMNER. 1975. Underwater observations of gelatinous zooplankton: Sampling problems, feeding biology, and be- havior. Limnol. Oceanogr. 20:907-917. MACKIE, G. O. 1964. Analysis of locomotion in a siphonophore col- ony. Proc. Roy. Soc. Lond., Ser. B., 159:366-391. ROGERS, C. A. In press. Impact of autumn-winter swarming of a siphonophore ("lipo") on fishing in coastal waters of New England. In J. R. Goulet, Jr. and E. D. Haynes (editors), Status of environment — 1975. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. ZELIKMAN, E. A. 1969. Structural features of mass aggregations of jellyfish. [In Russ.] Okeanologiya 9. (Engl, transl. in Oceanology 9:558-564). ^x Carolyn A. Rogers Northeast Fisheries Center Narragansett Laboratory National Marine Fisheries Service, NOAA R.R. 7 A, Box 522A, Narragansett, RI 02882 Douglas C. Biggs Marine Sciences Research Center State University of New York Stony Brook, NY 11794 *Lough, R. G. 1976. The distribution and abundance, growth and mortality of Georges Bank-Nantucket Shoals herring larvae during the 1975-76 winter period. Int. Comm. Northwest Atl. Fish. Res. Doc. 76Aa/123, 30 p. Richard a. Cooper Northeast Fisheries Center Woods Hole Laboratory National Marine Fisheries Service, NOAA Woods Hole, MA 02543 284 COMPUTER PROGRAM FOR ANALYSIS OF THE HOMOGENEITY AND GOODNESS OF FIT OF FREQUENCY DISTRIBUTIONS, FORTRAN IV Routinely, in the study of the dynamics of a fish population, one of the initial steps is the examina- tion of length measurements, viz, the frequency distribution of lengths, average length at age, and differential length distribution by gender. Often, length measurements are the only information available from which to estimate the age structure of the population. Standard statistical techniques such as chi-square tests are often used to analyze length-frequency distributions before pooling data, e.g., to estimate the age structure of the population (Yong and Skillman 1975). I have developed a computer program which forms frequency distributions from length mea- surements and then calculates a chi-square statis- tic which is used to test the homogeneity of the frequencies for the purpose of pooling. Theoretical frequencies from a normal distribution based upon the sample mean and variance of each length- frequency distribution are used in calculating chi-square tests of goodness of fit (Li 1959). The program does not partition the chi-square test of homogeneity but does pool adjacent class frequen- cies when expected frequencies are small in the case of the test of goodness of fit. Observed adja- cent class frequencies are pooled if their expected frequencies are too small and then the test of goodness of fit is calculated. The usual caution against using small samples and expected fre- quencies less than five in chi-square tests of good- ness of fit should be followed (Sokal and Rohlf 1969). Data required are either individual length mea- surements in millimeters (from 1 to 1,000 mm) or pairs of length class midpoint and frequency for each of up to five length-frequency distributions per data set; maximum frequency must be less than 1 million. Program storage could be in- creased to accommodate more than five length- frequency distributions, depending on the capac- ity of the computer being used. Class interval width must be specified; lengths are then tallied by up to 100 classes which are identified by mid- point on the output. Multiple data sets are pro- cessed sequentially without limit. Output includes listings of arithmetic mean, variance, standard deviation, standard error of the mean, total sample size, and chi-square statis- tic of goodness of fit for individual groups and for the pooled frequency distribution. The chi-square value for the test of homogeneity is printed with its degrees of freedom; appropriate tables should be consulted for critical values used in testing hypotheses. The goodness of fit test for the pooled data would not apply to the situation where the distribution is clearly multinomial. Histograms of all frequency distributions are produced as full- page printer charts, scaled if necessary to 50 units by up to 100 class intervals. The pooled frequen- cies and class midpoints are punched on cards to facilitate additional analyses. The program was developed on an IBM 360/65 OS System' and required 56,811 bytes of storage. A copy of the FORTRAN IV source program list- ing, example input and output, and an instruction manual are available from the author. Literature Cited LI, J. C. R. 1959. Introduction to statistical inference. Edward Bros., Ann Arbor, Mich., 553 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. YoNG, M. Y. Y., AND R. A. Skillman 1975. A computer program for analysis of polymodal fre- quency distributions (ENORMSEP), FORTRAN rV. Fish. Bull., U.S. 73:681. MICHAEL L. DAHLBERG Northwest and Alaska Fisheries Center Auke Bay Laboratory National Marine Fisheries Service, NOAA P.O. Box 155, Auke Bay, AK 99821 'Reference to trade name does not imply endorsement by the National Marine Fisheries Service, NOAA. PORTABLE TRIPOD DROP NET FOR ESTUARINE FISH STUDIES' Since the introduction of a portable drop net sys- tem by Jones et al. (1963) several designs have been utilized for freshwater and estuarine fish studies (Moseley and Copeland 1969; Kjelson and Johnson 1973; Kushlan 1974; Adams 1976). The value of these sampling systems in estimating the density and biomass of certain fish species has been well documented by these authors (Table 1). 1 Contribution No. 83 from the Harbor Branch Foundation, Inc. 285 Table l. — Basic drop net design characteristics of previous studies and the current net system. tvlethod of Fixed or tVlesh Sample sample Dominant species in Autfior portable size (mm) area (m^) collection the sample Hellier 1958, 1962 fixed 9.5 2529 1,011,7 seme Anchoa. Mugil Lagodon Hoese and Jones 1 963 fixed 190 118 seme Lagodon. Gobiosoma. Mugil Jones et al 1963: portable. 19.0 1004 pursed net Brevoortia. Mugil. Cynoscion Jones 1965 helicopter fVloseley and Copeland portable. 100 16 pursed net Brevoortia. Mugil. Cynoscion 1969 float Kjelson and Johnson 1973 portable, float 60 16 pursed net Anchoa. Lagodon, Eucinostomus Kjelson et al, 1975 fixed 3.0 4 pursed net Lagodon. Leiostomus. Anchoa Adams 1 976 portable, float 3.2 9 pursed net Anchoa. Lagodon. Orthopristis Current design portable 3.2 10 seme Gobiosioma. Lagodon. Eucino- stomus. Anchoa A drop net design was needed which would not significantly disturb the water surface and yet take an adequate sample. Some previous portable drop net designs sampled a larger area, but with greater water surface contact (Moseley and Cope- land 1969; Kjelson and Johnson 1973). This new gear design allows less water surface disturbance (i.e., noise and shading) than previous drop nets and yet is capable of sampling 10 m^ without com- promising portability. The sample area is rigidly controlled and all fishes are collected from the sample area. The design criteria and success of this drop net system is comparable with, and in some cases surpasses, previous drop net designs in the literature with regard to sample area control and the capture of certain small demersal fish species. This study was conducted to compare this new drop net system with a larger haul seine sys- tem sampling 1,160 m^ used concurrently for shal- low water estuarine fish studies. The duration of this study was from April to December 1976. Drop Net Description and Operation The drop net apparatus consists of two primary sections: the collapsible aluminum tripod with the trigger mechanism and the drop net (Figure 1). The 5.2-m tripod legs are held together by aluminum hinges at the upper end and three 4.0-mm flexible steel support cables attached to the legs below the upper hinges. Two sheaves are mounted to the upper ends of two of the tripod legs, one to carry the winch line (i.e., upper frame har- ness line) to hoist the net and the other to carry the drop frame harness line that is released as the net is triggered. After the sample site is straddled by the tripod, the drop net (3.16 x 3.16 m) is deployed using a pontoon boat. The boat is floated under the open tripod legs to prevent disturbing the bottom within the sample area. To lift the net, the drop frame harness plate and the upper frame harness plate are coupled together with a steel set pin (Figure la). The net is then lifted from the boat deck using the winch. After the net is in the set position, the drop frame harness line is set on the trip lever via a set ring (Figure lb), and the pon- toon boat is pushed out from under the net. The trip lever is held down with a notched trigger pin attached to the remote trigger line. The remote trigger line has a fluorescent floating jar attached to the distal end 20 to 30 m from the net apparatus. Once the net is set at the correct height, the steel set pin is pulled, and the drop frame plate and harness are free to fall when the trigger mechanism is tripped. Within 15 min three people can deploy a single net set to drop. The trigger mechanism and drop frame are re- leased with one pull of the remote trigger line. Once the net has fallen, the drop frame harness is undipped from its harness plate and a drop net seine, made of tubular aluminum and 3.2-mm mesh netting, is used to seine the enclosure (Fig- ure Ic). The seine fits closely against the inside walls of the drop net, and it is pulled by three people, two on either handle and one pulling a line attached to the bottom, center of the seine. The seine frame is kept firmly on the bottom and a standard five hauls are made to collect the sample. For night operations, an amber flashing light is attached to one tripod leg. Once the net has drop- ped, a lantern can be hung from the flexible steel support cable. Although night operations may take longer, V2 h is generally taken from the drop to complete sample removal. To store and disassemble the drop net the pon- toon boat is brought under the raised net. The net and frame are lowered onto the deck. The harness 286 UFHL SET RING TRIP LEVER FRAME Qmm.lOOmm 3.l€m par sId* Figure l. — Drop-net apparatus with insets of (a) harness plates, (b) trip lever mechanism, and (c) seine. UFHP = upper frame harness plate; UFH = upper frame harness; DFHP = drop frame harness plate; DFH = drop frame harness; DFHL = drop frame harness line; UFHL = upper frame harness line; UF = upper frame; DF = drop frame; SSP = steel set pin; FSSC = flexible steel support cable. clips to the upper frame harness and drop frame harnesses are released from their respective plates. The tripod (weight 56.3 kg) can now be collapsed and stowed with the drop net (weight 52.7 kg) on the pontoon boat. Disassembly of the drop net apparatus generally takes 10 min. Not counting the arbitrary waiting period between set and drop, the described procedure takes approxi- mately 1 h. The drop net was released 1 h after it was set once a month beginning in April 1976. These sam- ples were taken in a shallow seagreass bed (i.e., Thalassia, Halodule, and Syringodium ). This drop net design is limited to depths <1.2 m. A seine haul was made within an hour of each drop net sample in a seagrass bed approximately 75 m from the drop net site. A 62 x 1.8 m bag seine (3.2-mm mesh) was pulled with one end anchored on shore and the seaward end stretched perpendicular to shore. A 15.2 x 1.8 m barrier net (3.2-mm mesh) was set 30.5 m down the beach and parallel to the 62-m seine. The seaward end of the large seine was pulled by hand to the seaward end of the barrier net and then to shore covering approximately 1,160 m^/haul. The entire seine haul is made within 10 min. All specimens taken using both drop net and seine were identified, counted, measured, and weighed (wet weight). The percent occurrence was calculated based on the number of samples in which a species occurred out of the total number of samples taken. A comparison was then made between fish samples taken by both gear types (Table 2). Results and Discussion The drop net captured fewer individuals and species than the seine and mostly small demersal and semidemersal forms (Table 2). However, the total fish density and biomass values from drop net samples surpassed seine sample values. April to December drop net samples gave fish density val- ues from 1.8 to 19.3 fish/m^ (x = 9.0) and biomass values from 1.3 to 29.4 g/m^ (x - 15.0). In seine samples fish density ranged from 0.09 to 2.14 287 Table 2. — Partial species comparison, numerical catch, fish densities (no./m^), and percent occurrence in samples for simultaneous seine and drop net collec- tions (nine samples each). This is a partial species list, 17 of 61 species taken with the seine and 12 of 29 species taken with the drop net. Seine (10,440 m2) Drop net (90 m2) Type and species No. No./m^ Occurrence No. No./m^ Occurrence Schooling planktivores; Anchoa mitchilli 97,981 938 1 00 452 558 033 A hepsetus 539 .05 .78 0 — — A. nasuta 656 .06 .67 1 .01 .11 A. cubana 248 .02 .44 1 .01 .11 Harengula jaguana 2,725 .26 .67 0 — — Opisthonema oglinum 521 .05 .33 0 — — Sardinella anchovia 3 .00 .11 0 — — Semldemersal predators: Bairdiella chrysura 1,102 .11 1.00 14 16 22 Cynoscion nebulosus 22 .00 .44 2 02 .22 Diapterus auratus 944 .09 1.00 0 — — Euctnostomus sp. 1,404 13 1 00 43 .48 .67 Lagodon rhomboides 1.225 .12 1.00 191 2.12 1.00 Lutjanus griseus 23 .00 .89 1 .01 .11 Orthopnstis chrysoptera 326 .03 .56 25 .28 .33 Demersal species: Achirus lineatus 0 — — 3 .03 .22 Bathygobius soporator 6 .00 .22 0 — — Gobiosoma robustum 632 06 .44 336 4.15 .89 Gobionellus boleosoma 0 — — 6 .07 .44 Microgobius gulosus 8 .00 .33 18 ,22 .67 fish/m^ ix = 0.53) and biomass from 1.3 to 4.0 g/m^ (x = 2.0). The high fish density and biomass values of drop net methods versus lower values using seine methods has been demonstrated in previous studies (Kjelson and Johnson 1973; Kjelson et al. 1975). Schooling, nektonic species (e.g., anchovies and herring) and adults of larger species (>150 mm SL) were seldom taken in the drop net yet proved common in seine samples (Table 2). The drop net bias toward nonschooling fishes or those that do not have a clumped distribution has been documented by Kjelson and Johnson (1973) and Kjelson et al. (1975). However, the drop net de- signs of Hellier (1958, 1962), Hoese and Jones (1963), Jones et al. (1963), Jones (1965), and Moseley and Copeland (1969) captured large numbers of schooling fishes (e.g., Breuoortia and Anchoa; Table 1). These schooling fishes, because of their irregular occurrence (Table 2), occasion- ally presented a problem with subsequent sample analysis (Jones 1965). Small gobies (e.g., Gobio- soma robustum and Microgobius gulosus) were common in our drop net samples and were only occasionally seen in our seine samples. Most of those fishes captured by the drop net were grass flat residents and resident juveniles of adult popu- lations living elsewhere. The seine not only cap- tured grass flat residents and juvenile fish but adults and juveniles of migratory schooling forms and large top predators ( ^250 mm SL). When catch records of our drop net system are compared with those of others many sample similarities and differences are seen. Hellier's data demonstrates that drop nets with a smaller mesh size will capture a greater fish biomass when the sample area is kept constant (Hellier 1958). The current drop net design incorporates a 3.2-mm mesh (Table 1). This enables the capture of nearly all small fishes ( < 150 mm SL) present. Very small species (e.g., Gobiosoma robustum, 13-30 mm TL) were not commonly captured using other drop net methods, except in the samples taken by Hoese and Jones (1963) (Table 1). Gobiosoma robustum is a common seagreass bed resident from Corpus Christi, Tex., to the Indian River lagoon in eastern Florida (Hoese 1966; Springer and McErlean 1961); therefore, it would not be expected in the samples of Kjelson and Johnson (1973), Kjelson et al. (1975), and Adams (1976). Demersal flatfishes (e.g., Paralichthys, Etropus, Citharichthys , Sym- phurus, and Achirus) were captured in drop nets used by Jones et al. ( 1963), Mosely and Copeland (1969), Kjelson and Johnson (1973), Adams (1976), and our design. Juvenile commercial and sport fishes ( 15-50 mm SL) caught by the current drop net design were Cynoscion nebulosus, Lut- janus griseus, L. analis, L. synargris, Albula vul- pes, Archosargus probatocephalus, and Haemulon parrai. Lagodon rhomboides was also taken in large numbers ( 15-145 mm SL), showing densities 288 well over seine estimates. Other authors also found L. romboides to be common in their drop net samples (Table 1). The current drop net system is the only design to use a rigid frame seine and a solid aluminum drop frame in conjunction with 3.2-mm mesh netting. This probably accounts for the goby and flatfish captures and also accurately delineates the sam- ple area. It is possible that the sample area may change due to wind or current effects on falling pursing nets (Table 1; Jones et al. 1963; Kjelson et al. 1975). Disadvantages with the aluminum drop frame are its bulk, limited maneuverability, and operations limited to a level bottom. A collapsible frame or one which can be disassembled may eliminate the maneuverability problem. Moseley and Copeland (1969) indicated that noise and shadows may have affected their samples. We tried to eliminate the shadow effect and noise with as little water surface contact as possible using a tripod which suspended the net over the water with an open center. It may be possible to have vibrations in the tripod apparatus transmitted through the submerged portion of the tripod legs; however, this possibility and its effect is not known. Portable float and portable helicopter drop nets (Table 1) could drop in deeper water (depths of 2.5-4.6 m) than our system (1.2 m). Most other drop net designs require two people to operate. The helicopter drop net requires six while our design requires three. A smaller version of this tripod design would require fewer operators. It takes 60 min to set up, drop, retrieve the sample, and dis- mantle our drop net without the arbitrary 1 h waiting period. Kjelson and Johnson (1973) and Kjelson et al. (1975) were the only authors to pub- lish operational times and these were 25 min and 15 to 20 min respectively. The 10-m^ sample area in the current design is a compromise between maneuverability and sample size. The small sample precludes adequate capture of mobile fishes >150 mm SL. Fishes with a clumped distribution or that form schools will also occur in these drop net samples less frequently than if other gear were used (e.g., seines and trawls). However, to obtain an accurate fish den- sity and biomass estimate in nursery areas or of fish populations in which the adult size is small (e.g., gobioids) the current design has produced adequate samples. Literature Cited Adams, S. M. 1976. The ecology of eelgrass, Zostera marina (L.) fish communities. I. Structural analysis. J. Exp. Mar. Biol. Ecol. 22:269-291. HELLIER, T. R., Jr. 1958. The drop-net quadrat, a new population sampling device. Publ. Inst. Mar. Sci. Univ. Tex. 5:165-168. 1962. Fish production and biomass studies in relation to photosynthesis in the Laguna Madre of Texas. Publ. Inst. Mar. Sci. Univ. Tex. 8:1-22. HOESE, H. D. 1966. Habitat segregation in aquaria between two sym- patric species of Gobiosoma. Publ. Inst. Mar. Sci. Univ. Tex. 11:7-11. HOESE, H. D., AND R. S. JONES. 1963. Seasonality of larger animals in a Texas turtle grass community. Publ. Inst. Mar. Sci. Univ. Tex. 9:37-46. JONES, R. S. 1965. Fish stocks from a helicopter-borne purse net sampl- ing of Corpus Christi Bay, Texas 1962-1963. Publ. Inst. Mar. Sci. Univ. Tex. 10:68-75. Jones, r. S., W. B. Ogletree, J. H. Thompson, and W. Flen- NIKEN. 1963. Helicopter borne purse net for population sampling of shallow marine bays. Publ. Inst. Mar. Sci. Univ. Tex. 9:1-6. Kjelson, M. A., and G. N. Johnson. 1973. Description and evaluation of a portable drop-net for sampling nekton populations. Southeast Assoc. Game Fish. Comm., Proc. 27th Annu. Conf., p. 653-662. Kjelson, m. a., W. R. Turner, and G. N. Johnson. 1975. Description of a stationary drop-net for estimating nekton abundance in shallow waters. Trans. Am. Fish. Soc. 104:46-49. KUSHLAN, J. A. 1974. Quantitative sampling of fish populations in shal- low, freshwater environments. Trans. Am. Fish. Soc. 103:348-352. Moseley, F. N., and B. T. Copeland. 1969. A portable drop-net for representative sampling of nekton. Contrib. Mar. Sci. Univ. Tex. 14:37-45. Springer, V. G., and a. J. McErlean. 1961. Spawning seasons and growth of the code goby, Gobiosoma robustum (Pisces: Gobiidae), in the Tampa Bay area. Tulane Stud. Zool. 9:77-85. R. Grant Gilmore John K. Holt Robert S. Jones George R. Kulczycki Louis G. MacDowell III Wayne C. Magley Harbor Branch Foundation, Inc. RFD l.Box 196 Fort Pierce, FL 33450 289 SURFACE FEEDING BY A JUVENILE GRAY WHALE, ESCHRICHTWS ROBUSTUS Recently Ray and Schevill (1974) summarized in- formation on the feeding habits and feeding be- havior of Eschrichtius rohustus. The gray whale is primarily a bottom feeder whose diet consists mainly of six species of benthic gammaridean am- phipods taken in the Bering and Chukchi Seas during the summer months (Zimushko and Lenskaya 1970; Rice and Wolman 1971). It is gen- erally assumed that gray whales fast during mi- gration and while at the breeding grounds along the Mexican coast. Several reports, however, suggest the possibility that feeding may occur oc- casionally outside of the Arctic region and may include a wide array of different food items, e.g., smelt, anchovylike fish; planktonic crusta- ceans— Euphausia and Pleuroncodes (Howell and Huey 1930; Matthews 1932; Gilmore 1961; Bal- comb in Ray and Schevill 1974). In addition to these, reports of bits of woods, stones, tube worms, shell, etc., including kelp fragments have been reported in stomach contents of gray whales (Tom- ilin 1957). However, most of these items are prob- ably attributable to incidental swallowing. Herein we report observations made on a juvenile gray whale,* ca. 6-m long, exhibiting un- usual surface feeding behavior in a kelp, Macro- cystis angustifolia, bed near Refugio Beach State Park, 38 km west of Santa Barbara, Calif. Be- tween 1 and 9 April 1976, four visits were made to the area and a total of 8 h were spent detailing the observed behavior. Throughout the study period the whale's activities were confined to the exten- sive kelp bed situated between Refugio Beach State Park and Arroyo Hondo — a distance of 3.2 km. This feeding activity was restricted to the kelp canopy and occurred in shallow water (<5-10 m depth) and 50 to 200 m offshore. We last saw the whale on 9 April 1976. Apparently it left the area shortly thereafter as subsequent searches were made on 16 and 18 April 1976. Description of Feeding Behavior When first sighted, the whale's head was pro- truding a meter or more above the surface of the ' On a number of occasions the whale laid nearly horizontal on the surface of the water only a meter from our boat ( 7-m Boston Whaler), thus we were able to make a reasonably accurate esti- mate of it's overall length. Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. water in the center of a dense kelp bed (Figure lA). Shortly after surfacing snout first, its mouth opened and a large volume of water and kelp flowed into the oral cavity (Figure IB). Next the jaws closed (Figure IC) and in the process a small squirt of "excess" water issued from the most an- terolateral margins of the mouth. Within mo- ments entrapped water was forced out of the mouth across the baleen plates through the lips in a strong flush directed posterolaterally (Figure ID). This sequence was repeated several times before the whale submerged. Prior to submerging, the head was raised at an angle approximately 60° normal to the surface of the water. The body then slid backwards through the kelp canopy with its jaws slightly agape releasing the kelp present in its mouth. Resurfacing generally occurred a short distance away. There was little deviation from this pattern during the entire observation period. Visits were made at all hours of daylight during which the intensity of the feeding behavior ap- peared consistent. During a typical 27-min period when the whale was exhibiting feeding behavior, we noted that it emerged in the kelp, fed, submerged, and then reemerged a total of 18 times. A single feeding- submergence interval averaged 90 s, of which 56 s were spent feeding and 34 s submerged. Fre- quency of breaths during this period were recorded for 11.5 min. The average time from inhalation to exhalation was 48 s; the maximum was 70 s and the minimum 20 s. The act of breathing (i.e., ex- haling, then inhaling) at the surface averaged 2 s. These data clearly demonstrate that the whale was quite active in its behavior. At first impression the whale appeared to be "biting and eating" the kelp, but on closer inspec- tion the fronds and stipes of the kelp incurred little if any damage. While there is no direct evidence available from stomach analyses, we suggest the whale's activities among the kelp were di- rected to procuring quantities of the small kelp mysid crustacean, Acanthomysis sculpta. Sam- pling of the mysid fauna was accomplished using a 50-gal plastic trash can which was lowered into the water at a horizontal angle from the boat in such a fashion that the surface water down to 30 cm fiowed freely into the container. The mysids were subsequently filtered out, counted and vol- ume determinations made. A total of four repli- cates provided a conservative estimate of 5 to 10 mysids/1 at the canopy surface. The size range for individual mysids in our sample was 6 to 12 mm, 290 Figure l . — Time sequence photographs showing the observed feeding behavior: A, the gray whale first emerging in the kelp canopy; B, jaws extended open allowing surface water to enter mouth; C, mouth closed entrapping water and kelp fronds; D, water expelled through baleen in posterolateral direction. which falls well within the size range of the gam- maridean amphipods reportedly composing 95% of the whale's diet in Arctic seas (Rice and Wolman 1971). In addition to these observations, we noted that during feeding, water was expelled predominately through the right side of the mouth. Kasuya and Rice (1970) found that of 34 whales examined, 31 showed disproportionate wear of the baleen on the right side. Analysis from movie footage (8 mm) taken by us shows that of 31 consecutive expul- sions, water passed exclusively from the right side 20 times — in the remaining cases it was passed equally or nearly equally from both sides. At no time, however, was the water expelled on the left side exclusively. It is not clear what causes the wear on the baleen plates; perhaps it is unequal mechanical rubbing action of the tongue pushing water through the plates. Possibly related to this are observations made by Ray and Schevill (1974) on the captive juvenile gray whale, Gigi. At first this whale was hand fed by her trainers on the left side exclusively. Later, after hand feeding was discontinued and feeding became voluntary, food continued to be ingested solely on the left side. Interpretations and Conclusions of Observations Several aspects concerning the physical charac- teristics of our whale are worthy of comment. The mean length at birth (January) for a normal gray whale is reported to be ca. 4.9 m and by the time of weaning (August), the animal can be expected to reach a total length of 8.5 m (Rice and Wolman 291 1971). The size of our whale (ca. 6 m) would indi- cate a juvenile at the nursing stage. However, during our observations no large whale was noted in the vicinity which could have been interpreted as a parent. Thus we suggest that this animal may be a yearling runt. Further evidence in support of this notion is the fact that the epizoic barnacles iCryptolepas rhachianecti) were of a large class ( >2.5 cm), too large to be considered 4 to 5 mo of age, which would be the approximate age of the whale were it born in the most recent calving season. Also, since all barnacles were of only one distinct size class we further suggest that the whale we observed had not been south to the breeding grounds this year (1975-76). Rice and Wolman (1971) stated that northbound whales have two distinct size classes of barnacles, one adult and one juvenile (2-3 and 0.3-0.5 cm in diameter, respectively). We can only speculate on the events which may have occurred prior to our observations (e.g., abandonment or loss of the mother during the northbound journey in the previous year and con- sequent exploitation of an alternative food source, i.e., kelp mysids by a preweaned juvenile whale). However, we have been able to ascertain by com- parative photographic analysis of barnacle scar patterns (Figure 2) that this whale was present in the San Diego area (approximately 320 km south of Santa Barbara) from early January to early February 1976 (P. Zovanyi and H. Hall pers. commun.) — ^just over 4 mo prior to our encounter in April. In conclusion, this report would seem to indicate that gray whales can display plasticity in their feeding behavior. While conclusive evidence of feeding is lacking (i.e., gut content analysis), this appears to be the most logical explanation ac- counting for this unusual behavior. Acknowledgments We thank the following persons for critically read- ing the manuscript: E. Hochberg, G. V. Morejohn, G. C. Ray, D. Rice, W. Schevill, and C. Woodhouse. We are grateful to C. Engle for identifying the mysid and D. Kittle for bringing the whale to our attention. H. Hall and P. Zovanyi were helpful in allowing us to compare photographs of the same whale seen in the San Diego Area. Also, we thank H. Offen and the Marine Science Institute for sup- port in the research. Literature Cited GILMORE, R. M. 1961. The story of the gray whale. 2d ed. Privately pub- lished, San Diego, 17 p. Graves, W. 1976. The imperiled giants. Natl. Geogr. Mag. 150:722- 751. Howell, A. B., and L. M. Huey. 1930. Food of the gray and other whales. J. Mammal. 11:321-322. Kasuya, T., and D. w. Rice. 1970. Notes on baleen plates and on arrangement of parasitic barnacles of gray whale. Sci. Rep. Whales Res. Inst. 22:39-43. Matthews, L. H. 1932. Lobster-krill, anomuran Crustacea that are the food of whales. Discovery Rep. 5:467-484. Ray, G. C., and W, E. Schevill. 1974. Feeding of a captive gray whale, Eschrichtius robus- tus. In W. E. Evans (editor), The California gray whale, p. 31-38. Mar. Fish. Rev. 36(4). Rice, d. W., and a. a. Wolman. 1971. The life history and ecology of the gray whale (Es- chrichtius robustus). Am. Soc. Mammal., Spec. Publ. 3, 142 p. Tomilin, a. G. 1957. Mammals of the U.S.S.R. and adjacent countries. Vol. 9. Cetacea. Akad. Nauk SSSR, Moscow, 756 p. (Translated by Isr. Program Sci. Transl. Jerusalem, 1967, 717 p.) J. B Figure 2. — Line drawings of barnacle scar patterns on a gray whale: A, after Figure 1 A, seven barnacle scars on the gray whale seen in Santa Barbara in April 1976; B, drawn from photograph taken by H. Hall (Graves 1976) of a gray whale seen in San Diego in January 1976. The same seven barnacle scars are evident. 292 ZIMUSHKO, V. v., AND S. A. LENSKAYA. 1970. Feeding of the gray whale {Eschrichtius gibbosus Erx.) at foraging grounds. Ekologiya Akad. Nauk SSSR l(3);26-35. (Engl, transl., Consultants Bureau, Plenum Publ. Corp., 1971. Ekologiya 1(3):205-212.) g. m. wellington Shane Anderson Marine Science Institute and Department of Biological Sciences University of California Santa Barbara. CA 93106 HOMING OF MORPHOLINE-IMPRINTED BROWN TROUT, SALAIO TRUTTA Homing for the purpose of spawning is well documented for lake-run brown trout, Salmo trutta (Stuart 1957; Niemuth 1967), but the mechanism by which they find their natal trib- utary is not understood. Our own recent studies on related species — coho salmon, Oncorhynchus kisutch, and rainbow trout, Salmo gairdneri — suggest that they become imprinted to the odor of their natal tributary when they begin their downstream migration and later use this informa- tion for homing (Hasler and Wisby 1951; Scholz et al. 1973, 1975, 1976; Cooper and Scholz 1976; Cooper et al. 1976). In these experiments 18-mo-old hatchery-raised fish were exposed to a synthetic chemical, morpholine, for 40 days and then stocked in Lake Michigan. During the spawning migration the fish homed to a simulated home stream which was scented with morpholine. Since the life cycle of migratory brown trout is similar to that of coho salmon and rainbow trout, we con- ducted the present study to determine if odor im- printing could be extended to brown trout. The methods used in this study were similar to proce- dures reported by Cooper and Scholz ( 1976) since both experiments were conducted concurrently. Methods In 1972, hatchery-raised, 18-mo-old brown trout fingerlings were transported to South Milwaukee, Wis. (Figure 1). The fish were marked with fin clips, divided into three groups of 300 each, and held in separate tanks at the South Milwaukee Water Filtration Plant. Lake Michigan water was supplied to all three tanks from an intake crib Figure l. — Research area, South Milwaukee, Wis. (after Cooper et al. 1976). The solid triangle indicates the location of the hatchery where the fish were reared. Inset (A) shows detail of: 1) the water intake for the tanks at the South Milwaukee Water Filtration Plant, 2) the Oak Creek stocking site, and 3) the Milwaukee Harbor stocking site. located 1.5 km offshore. Morpholine (C4HgN0) was metered into one tank for 34 days in May and June. This period was selected because it is the time when brown trout would normally begin their downstream migration (Stuart 1957; Niemuth 1967). A concentration of 5 x 10"'^ mg/1 morpholine was maintained in the tank through- out the exposure period. The morpholine-exposed group and one unex- posed control group were then stocked in Lake Michigan at Milwaukee Harbor, 13 km north of Oak Creek (Figure 1). The second control group was released at the mouth of Oak Creek. During the spawning migration in fall 1972 and 1973, morpholine was metered into Oak Creek at the same concentration used for imprinting. The stream was surveyed for marked fish by gillnet- ting, electrofishing, and creel-census methods (summarized in Table 1). Fish were unable to move past a dam situated 1.5 km from the mouth. Surveys began before the spawning migration started and continued until no fish were left in the river. The results are recorded in Table 2. 293 Table l.— Summary of efFort spent in monitoring Oak Creek during the spawning migrations of brown trout in fall 1972 and 1973. Creel-census surveys were conducted three to five times each day and electrofishing surveys were made once or twice each week. A total of 51 marked brown trout were caught by anglers; 17, by electrofishing; and 2, in gill nets. Fall Creel census Electrofishing Gill net 1972 1973 274 451 Number of trips 11 24 62 54 Table 2. — Recoveries of brown trout at Oak Creek in fall 1972 and 1973 from those released in spring 1972. Morpholine- exposed and control fish were released 13 km north of Oak Creek and a second control group was released at the mouth of Oak Creek. Fin clip: RP, right pectoral; LP, left pectoral; LM, left maxillary. Experimental group Fin clip Number released Number recovered Percent of fishi 1972 1973 Total stocked Morpholine Control Oak Creek RP LM LP 300 300 300 23 1 3 30 2 11 53 3 14 177 1.0 4,7 Results A total of 53 morpholine fish (17.7% of the total number originally stocked) were captured as com- pared with 3 control trout (1.0%) released at Mil- waukee Harbor and 14 control trout (4.7%) re- leased at Oak Creek. Thus, the data show that morpholine-exposed brown trout returned to the scented stream in larger numbers than either con- trol group. Both control and morpholine fish ex- perienced uniform stocking procedures after the initial treatment. If the selection of the morpho- line-scented stream were attributed to a cue learned after the treatment, we would have ex- pected to capture as many control fish as morpholine-treated fish in the scented stream. The fact that this was not the case implies that the cue was morpholine. Therefore we conclude that morpholine-exposed brown trout used morpholine as a cue for homing. To locate the scented stream morpholine fish were able to search a distance of at least 13 km. This experiment should be repeated because of the low numbers offish stocked but the results are of interest because of the high percent- age of morpholine-exposed fish captured in the scented stream. Discussion Scotland. In one case brown trout were marked in one branch of a forked stream which flowed into the reservoir. After the fish had migrated to the reservoir, all of the water from the home fork was diverted into a new channel. The original channel was also maintained with water from the second fork. During the spawning migration, adult trout homed to the new channel in preference to the channel by which they had entered the reservoir. In the second instance Stuart reported that, when a different stream broke its banks, the stream bed below the break dried up and the en- tire flow of water was diverted into a marsh through which it percolated into the reservoir. During the spawning migration, brown trout con- gregated off the marsh where the percolating water entered the reservoir and not off the dry stream mouth. Both of Stuart's observations clearly indicate that the fish homed to water originating from the home tributary, rather than to a specific home location and are, thus, consistent with our conclu- sion that it is a characteristic of the home water, specifically odor, which provides brown trout with homing cues. Acknowledgments We thank Sy Drezweicki, Rod Smith, Terry Chapp, and Peter Johnsen for their help in the field. We acknowledge Dale Madison for advice in all aspects of this study; and Ron Poff, Russ Daly, Paul Schultz, and Jim Holzer of the Wisconsin Department of Natural Resources for technical and logistical support. The assistance of Ed Muel- ler and his advanced biology high-school students at South Milwaukee High School, and John Skorupski and Don Geiger at the South Mil- waukee Water Filtration Plant is also ap- preciated. Supported by grants to A. D. Hasler (NSF Grant GB 343, University of Wisconsin Sea Grant Program, Department of Commerce, NOAA 2-35209) and the Wisconsin Department of Natural Resources. Literature Cited COOPER, J. C, AND A. T. SCHOLZ. 1976. Homing of artificially imprinted steelhead (rain- In view of our findings it is of interest to consider two unpublished observations made by Stuart^ on homing of brown trout at Dunalastair Reservoir in 'Pers. commun. T. Stuart to A. D. Hasler, 7 March 1958. Letter No. Pu. 9 from Freshwater Fisheries Laboratory, Faskally, Pit- lochry, Perthshire, Scotland. 294 bow) trout, Salmo gairdneri. J. Fish. Res. Board Can. 33:826-829. Cooper, J. C, A. T. Scholz, R. M. Horrall, a. D. Hasler, AND D. M. Madison. 1976. Experimental confirmation of the olfactory hypo- thesis with homing, artificially imprinted coho salmon (Oncorhynchus kisutch). J. Fish. Res. Board Can. 33:703-710. Hasler, A. D., and W. J. Wisby. 1951. Discrimination of stream odors by fishes and rela- tion to parent stream behavior. Am. Nat. 85:223-238. NIEMUTH, W. 1967. A study of migratory lake-run trout in the Brule River, Wisconsin: brown trout. Wis. Dep. Nat. Resour. Fish. Manage. Rep. 12, 80 p. Scholz, A. T., J. C. Cooper, D. M. Madison, R. M. horrall, A. D. Hasler, A. E. Dizon, and R. J. Poff. 1973. Olfactory imprinting in coho salmon: behavioral and electrophysiological evidence. Proc. 16th Conf. Great Lakes Res., p. 143-153. Scholz, a. T., R. M. horrall, J. C. Cooper, and a. D. Hasler. 1976. Imprinting to chemical cues: the betsis for home stream selection in salmon. Science (Wash., D.C.) 192:1247-1249. Scholz, a. T., R. M. Horrall, J. C. Cooper, a. D. Hasler, D. M. Madison, R. J. Poff, and R. I. Daly. 1975. Artificial imprinting of salmon and trout in Lake Michigan. Wis. Dep. Nat. Resour. Fish. Manage. Rep. 80, 46 p. Stuart, T. a. 1957. The migrations and homing behavior of brown trout {Salmo truttaL.). Freshwater Salmon Fish. Res. 18, 27 p. of lobster larvae. Personnel participating in Cooperative Investigations of the Caribbean and Adjacent Regions (CICAR) activities have pre- pared a "Plan for Sampling the Early Develop- ment Stages of Pelagic Fish during CICAR Opera- tions" which describes the use of the neuston net (FAO^). The Boothbay neuston net, initially adopted as the standard for the Marine Resources Monitoring, Assessment and Prediction Program (MARMAP), consists of a pipe frame 2 m wide by 1 m deep with an 8.5-m long net."* Because little was known concerning the sampling performance of this gear, an experiment was designed to test the operating characteristics of two types of frame (galvanized pipe and aluminum pipe) and two lengths of net (4.9 m and 8.5 m with ratios of mouth to open mesh aperture areas of 1 : 6 and 1 : 1 1 , respectively). The nets were of 0.947-mm Nitex^ mesh. The results of the experiment defining the operating characteristics of the two types of frame and two lengths of net were described by Eldridge et al. (1977). The present report will describe mainly diurnal variations in catches of ichthyo- neuston during the latter experiment, which was conducted during 9-15 July 1973 utilizing the RV Dolphin. Allan T. Scholz Laboratory of Limnology University of Wisconsin Madison. WI 53706 JoN C. Cooper Laboratory of Limnology, University of Wisconsin Present address: Texas Instruments, Inc. Buchanan, NY 1 05 II Ross M. Horrall Arthur D. Hasler Laboratory of Limnology University of Wisconsin Madison, WI 53706 Materials and Methods The neuston net was towed from a boom extend- ing 3 m from the starboard side of the RV Dolphin, and the ship was ordered in an arc of radius 1 n.mi. or less to starboard to keep the net mouth out of the ship's wake. The net was towed so that one-half the height (0.5 m) was in the water. Towing speeds of 1, 2, and 3 m/s were employed with a total of 48 tows being conducted. Twenty- four daylight tows were made between 1107 and 1627 EST and 24 night tows between 2206 and 0432 EST. After setting (which took an average of 29 s), nets were towed 10 min and then retrieved DIURNAL VARIATIONS IN CATCHES OF SELECTED SPECIES OF ICHTHYONEUSTON BY THE BOOTHBAY NEUSTON NET OFF CHARLESTON, SOUTH CAROLINA^' ^ The Boothbay neuston net is becoming a standard gear for collection of ichthyoneuston. Sherman and Lewis (1967) reported using this gear for collection 'Contribution No. 74 from the South Carolina Marine Re- sources Center. This work is the result of research sponsored by the MARMAP Program, U.S. Department of Commerce, Na- tional Oceanic and Atmospheric Administration, National Marine Fisheries Service under Contract No. 4-35137. MAR- MAP Contribution No. 117. ^Contribution No. 451 from the Southeast Fisheries Center, National Marine Fisheries Service, NOAA, Miami, Fla. 3FA0-UNDP Fisheries Program, Mexico City. 1970. A plan for sampling the eggs and larvae of the fishes of Mexican waters. Unpubl. manuscr. ■•MARMAP is now using a 0.5 x 1 m neuston net. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 295 (average time of retrieval was 32 s). After each tow, the catch was drained on a 0.85-mm mesh sieve and preserved in 5% buffered Formalin. Sorting and identification of ichthyoplankton oc- curred at the Marine Resources Research Institute (MRRI). Fork lengths were measured in forked tail species; total lengths in all others. Relative volume of water strained was determined by the formula: Relative volume strained = (Speed)(To- tal tow time)(Average fraction of net in water). The reader should consult Eldridge et al. ( 1977) for further details concerning material and methods as well as the experimental design. Results The 4.9-m net was superior to the 8.5-m net in both ease of handling and minimizing damage to specimens after capture. There was no significant difference in catching ability of the two nets al- though the 4.9-m net actually caught more speci- mens during the experiment (Eldridge et al. 1977). The galvanized pipe frame was superior to the aluminum. A total of 10,621 specimens of ichthyoneuston were collected. The 20 most abundant taxa made up 85.6% (9,088) of the total number of specimens. The remaining 92 taxa composed 14.4% (1,533) of the total (see Table 1 for most numerous taxa collected). Analyses of variance and covariance tests re- vealed that total tow duration, speed, and relative volume strained did not vary significantly be- tween day and night tows (Eldridge et al. 1977). Thus, catches between diurnal periods did not ap- pear biased by the conduct of the experiment. Table l. — Numbers of individuals of selected ichthyoneuston collected in neuston experiment (+ = significantly more abundant for day or night, or no significant difference in catch between day and night at 5% level of significance). Taxon Total Number in Number in number caught night catches day catches Day Night Both Range total length (mm) Number of tows present Auxis sp 3,576 Exocoetldae 1 ,245 Scombridae 907 Gerreidae 513 Tetraodontldae 409 Mullldae 348 Mugil curema 230 Priacanthldae 223 Coryphaena hippurus 217 Caranx crysos 191 Gobiidae 180 Angullllformes 143 Carangldae 128 Psenes maculatus 125 Hemlramphidae 124 Decapterus punctatus 118 Monacanthus setifer 1 03 Scorpaenidae 102 Holocentridae 93 Caranx sp 91 Synodontidae 88 Euthynnus alletteratus 67 Monacanthus hispidus 64 Opislhonema oglinum 62 Istiophorus platypterus 59 Decapterus sp 54 Coryphaena equisetis 54 Aluterus sp 49 Trachinotus falcatus 48 Balistldae 46 Pomacentrldae 46 Labridae 44 Scomberomorus cavalla 41 Serranldae 40 Cynoglossidae 39 Kyphosus sp. 39 Selar crumenophthalmus 38 Bothus sp 34 Canthigaster sp, 33 Monacanthus sp 33 Dactylopterus volitans 30 Serioa sp 26 Seriola rivoliana 25 Caranx hippos 23 Syngnathidae 22 Apogonldae 22 Rachycentron canadum 1 9 573 3 700 545 906 1 229 284 15 394 7 341 77 153 222 1 188 29 67 124 179 1 142 1 125 3 125 0 97 27 46 72 11 92 92 10 93 0 87 4 88 0 67 0 3 61 55 7 26 33 53 1 50 4 6 43 31 17 21 25 29 17 43 1 39 2 40 0 39 0 15 24 18 20 33 1 31 2 3 30 3 27 4 22 0 25 21 2 19 3 22 0 19 0 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + v+ + + 2-16 26 4-83 45 4-15 8 5-14 43 4-14 29 5-21 29 6-18 40 3-30 25 9-62 34 7-37 42 5-14 22 6-84 24 3-5 21 6-49 20 6-57 31 24-47 32 9-35 30 3-11 29 3-17 21 4-32 24 5-32 18 3-10 18 14-58 22 5-15 20 3-18 26 7-11 18 8-18 21 1-105 17 7-18 19 3-12 23 5-20 28 5-18 18 5-10 18 3-16 15 5-16 16 7-21 18 6-69 15 3-22 16 3-17 14 8-30 11 8-30 11 5-18 14 14-43 9 6-32 14 7-69 15 4-10 12 6-13 11 296 Partial correlation analyses indicated that catches of flyingfishes, Exocoetidae, and silver driftfish, Psenes maculatus, were positively corre- lated with speed. Catches of the planehead filefish, Monacanthus hispidus, pygmy filefish, M. setifer, and dolphin, Coryphaena hippurus, increased with concentrations of sargassum weed which cor- responds to earlier observations by Dooley (1972). Catches of Exocoetidae were negatively correlated with manatee grass (Eldridge et al. 1977). Chi-square analyses indicated that catches of 41 taxa were significantly affected by changes in diurnal period (Table 1). Catches of 29 were great- er at night, whereas collections of 12 were greater during daylight hours. There was no evidence to suggest that catches varied significantly between diurnal periods for six species groups. Data in Table 1 indicate that specimens o^Auxis sp., Scombridae, Priacanthidae, Gobiidae, Anguil- liformes, Carangidae, Psenes maculatus, Holocen- tridae, Caranx sp., Synodontidae, Euthynnus al- letteratus, Decapterus sp., Coryphaena equisetis, Labridae, Scomberomorus cavalla, Serranidae, Cynoglossidae, Bothus sp., Canthigaster sp., Apogonidae, and Rachycentron canadum could be considered "faculative neuston" (Hempel and Weikert 1972). Specimens of Gerreidae, 7s- tiophorus platypterus, Balistidae, Pomacentridae, Kyphosus sp., and Selar crumenophthalmus ap- pear to be "euneuston" as defined by Hempel and Weikert (1972). Similarly, Mugil curema, Caranx crysos, and Decapterus punctatus appear to be "pseudoneuston." Mugil cephalus was identified as an euneustonic species by Hempel and Weikert (1972); whereas M. curema in our samples appeared to be pseudoneustonic. The difference may be real be- cause different species are involved or simply a sampling artifact. Similarly, although young stages of Exocoetidae were reported as rarely en- countered and as concentrating at the surface dur- ing daytime by Hempel and Weikert (1972), Exocoetidae were the second most abundant taxa in our samples and were taken mostly at night. The reason for this is unknown, but may be due to differences in location, species sampled, or random error associated with sampling of ichthyoneuston. Tetraodontidae, puffers, were taken most often during the day and were positively correlated with density of manatee grass. Results of the neuston gear experiment indi- cated that 1) the 4.9-m net is the superior net for routine surveys, and 2) choice of sampling hours should take into account variation in catches as- sociated with changes in light conditions. Acknowledgments We thank members of the crew and scientific party of RV Dolphin, who performed the field work, especially Bruce Stender, Bill Leland, and Oleg Pashuk. Thanks are also due to Howard Powles, Paul Sandifer, and Edwin B. Joseph, who reviewed the manuscript and to Patricia Dupree and Lexa Ford who typed the manuscript. Literature Cited DOOLEY, J. K. 1972. Fishes associated with the pelagic sargassum com- plex, with a discussion of the sargassum communi- ty. Contrib. Mar. Sci. 16:1-32. Eldridge, p. J., F. H. Berry, and M. C. Miller, III. 1977. Test results of the Boothbay neuston net related to net length, diurnal period, and other variables. S.C. Mar. Resour. Cent. Tech. Rep. 18, 22 p. Hempel, G., and H. Weikert. 1972. The neuston of the subtropical and boreal North- eastern Atlantic Ocean. A review. Mar. Biol. (Berl.) 13:70-88. Sherman, K., and R. D. Lewis. 1967. Seasonal occurrence of larval lobsters in coastal waters of central Maine. Proc. Natl. Shellfish. Assoc. 57:27-30. Peter J. Eldridge Marine Resources Research Institute South Carolina Wildlife and Marine Resources Department P.O. Box 12559, Charleston, SC 29412 Frederick H. Berry Southeast Fisheries Center National Marine Fisheries Service, NOAA 75 Virginia Beach Drive, Miami, FL 33149 M. Clinton Miller, III Department of Biometry Medical University of South Carolina Charleston, SC 29401 297 INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the following instructions. These are not absolute requirements, of course, but desiderata. CONTENT OF MANUSCRIPT The title page should give only the title of the paper, the author's name, his affiliation, and mailing address, including Zip code. The abstract should not exceed one double- spaced page. In the text, Fishery Bulletin style, for the most part, follows that of the U.S. Government Printing Office Style Manual. Fish names follow the style of the American Fisheries Society Special Publi- cation No. 6, A List of Common and Scientific Names of Fishes from the United States and Canada, Third Edition, 1970. 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 photog- raphy to 5% inches (for single-column figures, allowing for 50% reduction in printing), or to 12 inches (for double-column figures). The maximum height, for either width, is 14 inches. Photo- graphs should be printed on glossy paper. Do not send original drawings to the Scientific Editor; if they, rather than the photographic re- ductions, 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 re- ferred 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 literature cited is the re- sponsibility of the author. Abbreviations of names of periodicals and serials should conform to Bio- logical Abstracts List of Serials with Title Abbrevi- ations. (Chemical Abstracts also uses this system, which was developed by the American Standards Association.) Common abbreviations and symbols, such as mm, m, g, ml, mg, °C (for Celsius), %, %o and so forth, should be used. Abbreviate units of mea- sure 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 car- bon copies. The sequence of the material should be: TITLE PAGE ABSTRACT TEXT UTERATURE 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 arabic numeral; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. Jay C. Quast, Scientific Editor Fishery Bulletin Northwest and Alaska Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA P.O. Box 155, Auke Bay, AK 99821 Fifty separates will be supplied to an author free of charge and 100 suppUed to his organiza- tion. No covers will be supplied. Contents-continued HAYNES, EVAN. Description of larvae of the humpy shrimp, Pandalus goniurus, reared in situ in Kachemak Bay, Alaska 235 BEN-TUVIA, ADAM. Immigration of fishes through the Suez Canal 249 LOVE, MILTON S., and ALFRED W. EBELING. Food and habitat of three switch- feeding fishes in the kelp forests off Santa Barbara, California 257 COLLETTE, BRUCE B., JOSEPH L. RUSSO, and LUIS ALBERTO ZAVALA- CAMIN. Scomberomoris brasiliensis, a new species of Spanish mackerel from the western Atlantic 273 Notes ROGERS, CAROLYN A., DOUGLAS C. BIGGS, and RICHARD A. COOPER. Aggregation of the siphonophore Na^omia cara in the Gulf of Maine: observations from a submersible 281 DAHLBERG, MICHAEL L. Computer program for analysis of the homogeneity and goodness of fit of frequency distributions, FORTRAN IV 285 GILMORE, R. GRANT, JOHN K. HOLT, ROBERT S. JONES, GEORGE R. KULCZYCKI, LOUIS G. MacDOWELL III, and WAYNE C. MAGLEY. Portable tripod drop net for estuarine fish studies 285 WELLINGTON, G. M., and SHANE ANDERSON. Surface feeding by a juvenile gray whale, Eschrichtius robustus 290 SCHOLZ, ALLAN T., JON C. COOPER, ROSS M. HORRALL, and ARTHUR D. HASLER. Homing of morpholine-imprinted brown trout, Salmo trutta '. 293 ELDRIDGE, PETER J., FREDERICK H. BERRY, and M. CLINTON MILLER, III. Diurnal variations in catches of selected species of ichthyoneuston by the Boothbay neuston net off Charleston, South Carolina 295 ■50 mm experimentally selected final mean temperatures of 29.0°C at 34%o salinity to 19.5°C at 0%o salinity. In the field, fish ^50 mm remained seaward of the tide line in water of lower and more uniform temperature and higher and more uniform salinity than those recorded for mullet <50 mm. Mullet <50 mm occur seasonally when there are a maximum number of low tides sO.O mm and a minimum number of high tides &0.6 m. This allows the mullet increased time to feed undisturbed in areas where there are no predators and intraspecific and possible interspecific competitors for food and space. By the time fish reach 50 mm standard length, the tidal situation changes, allowing predators and competitors access to the shallow areas during low tide. When in the presence of predators, the schooling habit increases chances of survival for individual mullet. The marine environment includes the highly complex estuarine and intertidal habitats, which undergo continuous fluctuation. Organisms dwell- ing within these areas must be able to tolerate or escape from the consequences of extreme tempera- ture and salinity oscillations brought about by tidal and meteorological changes. Coral and rocky intertidal tidepools and the estuarine environ- ment serve as nursery and feeding grounds for the young of many species of fishes (Randall 1961; Norris 1963; Lauff 1967; Carr and Giesel 1975). The purpose of this study was twofold: to deter- mine whether young striped mullet, Mugil cephalus Linnaeus, select specific environmental conditions, particularly with respect to tempera- ture and salinity, as found in intertidal estuarine environments in Hawaii, and to explore the possi- ble causal mechanisms that might lead to the selection of such conditions. An experimental ver- tical thermal gradient (use of such gradients was 'Oceanic Institute Contribution No. 118. From part of a thesis submitted in partial fulfillment of requirements for the degree of Doctor of Philosophy, University of California at Santa Cruz. ^Center for Coa.stal Marine Studies, University of California at Santa Cruz, Santa Cruz, Calif; present address: P.O. Box 23720, L'Enfant Plaza Station, Washington, DC 20024. reviewed by Mantelman 1958; Ivlev and Lei- zerovich 1960; Fry 1964) in a tank was used to study the relationship between salinity and temp- erature and the distribution of schools of young striped mullet, and field observations were made of the distribution, feeding, and predator-prey be- havior of schooled mullet. METHODS AND MATERIALS Field Sites and Capture of Fish Young striped mullet were observed and col- lected in estuarine intertidal habitats at a number of locations on the island of Oahu ( Figure Din the Hawaiian Archipelago during 1972 and 1973. All experimental fish were captured with hand or beach seines near stream mouths or springs and on tidal mud flats in Maunalua Bay on the south- east side of the island. Schools were usually caught in the morning at low tide and transported <16 km to the Oceanic Institute, Makapuu, Oahu. Observations were made along Wailupe Beach, Wailupe Stream, and Kuapa Pond Streams (Ha- waii Kai Development drainage culverts) in Mau- nalua Bay, and along Kahana River and a silted Hawaiian fishpond in Kahana Bay on the east side Manuscript accepted July 1977. FISHERY BULLETIN: VOL 76. NO. 2. 1978, 299 FISHERY BULLETIN: VOL. 76, NO. 2 21-40 N / i ^ / ! J Kahora Boy S Kooeohe Boy Koneohe Mouho ^-^ ^^T fi.fport \ y — ' "^ ~— ^V:^ u 1 1 N^MOkopuu V ^Mojnoluo ^-^ Boy ° 0 AHU 1|0 km Figure l. — The island of Oahu, Hawaii, showing the major study areas and the locations at which environmental data were collected. of Oahu. Observations were made primarily dur- ing daylight hours, but a few night observations were made as well at Kahana River and Wailupe Beach. Field observations were recorded as I fol- lowed at a distance schools of mullet as they swam about estuarine intertidal regions and estuarine streams. Information about the distribution of mullet was also collected by using seines. Be- havioral and distributional records were kept and the temperature and salinity of the water through which the fish passed were measured with a tele- thermometer and compensated salinity refrac- tometer, respectively. Young Mugil cephalus were distinguished and differentiated from the young of a second sympat- ric species of mullet, Chelon engeli, an introduced species (Randall and Kanayama 1972), by differ- ences in body pigmentation pattern and opercle coloration. In addition, young C. engeli ^bO mm standard length (SL) occurred in the intertidal estuarine regions predominantly during the summer and fall, whereas, striped mullet pre- dominated in the winter and spring months. Only observations of mullet that were unquestionably identified as striped mullet were used in this re- port. I Experimental Methods Experiments were carried out at the Oceanic Institute during 1972 and 1973 (Table 1). An ex- perimental thermal vertical gradient was estab- lished in a 566-1 cylindrical Plexiglas^ tank, 91 cm high and 89 cm in diameter, inside a lighttight enclosure (Figure 2). Epoxy-coated copper coils spiraled around the inside of the tank, having entered through the surface above and through the side at the bottom. These separate sets of coils exited at midtank through the side. Pumps con- ^ Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Figure 2. — Experimental apparatus (diagramatic). Light excluding sides Etnd covering have been removed from frame. 300 MAJOR: ASPECTS OF ECOLOGY OF STRIPED MULLET Table 1. — Summary of experimental conditions and fish statistics. Experimental salinity (%«) (actual test salinity) 34(32) 32 34 34(33) 33 34(36) 34 32 32 15(16) (14) (15) 15(14) 15 Month and year of experiments Min-max tank temp CC) Length to nearest mm 17-19 20-29 30-39 40-49 50-59 60-69 70-79 80-130 Mean (range) Number fish in size range (mm SL) Total Experimental number group code fish (see Figure 3) Feb 1972 Mar 1972 Mar, 1972 May 1972 May 1972 July 1972 July 1972 Mar 1973 Mar 1973 Jan, 1972 Mar 1973 Mar, 1973 Aug, 1972 Mar. 1973 Feb 1972 Mar 1972 Mar. 1972 Mar 1973 May 1972 Jan. 1973 Mar. 1973 16.8-360 15.1-38.6 170-40.1 17.8-40.8 17.5-41.0 15.8-39.2 15 0-39.2 150-378 17.0-40.0 169-40.0 13.0-39 0 132-39.9 15.5-39.8 14.1-39.8 15.9-36.9 164-389 160-40.1 148-39.0 17.0-38.5 18.8-38.2 162-386 36 39 3 78 13 39 27 79 26 36 47 35 144 11 29 40 7 3 23 26 10 2 16 10 2 29 14 43 19 2 17 19 5 7 12 12 24 27 34 (19-28) (21-38) (29-43) 28 42 51 (19-43) (37-52) (46-59) 7 10 9 2 11 22 20 46 68 75 80 109 (37-59) (55-83) (56-98) (48-103) (82-129) 26 55 82 27 26 28 (48-129) (24-31) (19-29) (20-39) 18 10 26 88 98 (19-39) (69-114) (75-125) 28 26 22 27 26 (22-34) (19-26) (17-34) (21-31) 35 16 24 43 98 102 (17-34) (35-51) (76-118) (87-130) 122 ; 69 106 106 J 92 (69-125) 33, 51 99 (76-130) 52 A1, 2, 3 B1, 2, 3 CI, 2, 3 D1, 2, 3 El, 2. 3 F1. 2, 3 G1, 2. 3 HI, 2, 3 nected to reservoirs circulated v/ater, chilled as it passed through a set of coils in an ice bath, before entering the bottom of the tank, and water heated by a braided glass heating tape wrapped around a section of the coil, before it passed into the tank at the surface. The reservoirs allowed air bubbles to escape and the addition of water to the coils. Vinyl-coated thermistor probes (and leads), ex- tending through plastic tubes to various levels in the tank, ran above the top of the tank and out to a telethermometer recorder. Fine mesh plastic window screen was attached to the inside circumference of the coils to keep fish in the central area of the tank. The volume of water in the tank available to the fish was approx- imately 486 1, 88 cm in diameter and 80 cm deep. Observations were made through narrow eye- width slits cut at various levels in each side of the enclosure. A light-excluding cover surrounded the observer during observations. The viewing slits were closed when not in use. Water samples were taken from the surface, middepth, and bottom before and after each exper- iment for oxygen, pH, and salinity analysis. Oxy- gen measurements could not be made con- tinuously during the experiments, but oxygen measurements before and after each experiment did not change noticeably. In addition, respiratory movements of mouth and opercles, which might have been indicative of oxygen deficiencies, in the mullet did not change with increased heating. The above measurements were made primarily to en- sure that mullet were not orienting to factors other than temperature. Illumination was provided by two 15-W incan- descent light bulbs fixed in reflectors 84 cm above the surface of the water. Due to the position and low wattage of the light sources, a light gradient was established in the tank. The behavioral and distributional responses of the fish in a continu- ously changing thermal environment indicated that orientation to temperature and not light gra- dients occurred. The same observations were used to control for any orientation to pressure gra- dients, which inevitably existed in the 80-cm deep tank. One of three experimental salinity conditions, freshwater (range 0-2%o), 15%o (range 14-16%o), and 34%o (range 30-36%o), was established prior to placing the fish in the tank. Freshwater was 301 FISHERY BULLETIN: VOL. 76, NO 2 thoroughly mixed with seawater of 36%o salinity until a desired salinity was obtained in the tank. Frequently, schools of fish in the field were not caught in water of identical salinity to that used experimentally. Before placing these fish in the experimental tank, freshwater slowly ran into the container of water in which the fish were trans- ported to the laboratory. When the desired lower salinity was reached by dilution and overflow of the water in the container (this required about 60 min to accomplish), the fish were transferred by dip net to the experimental tank. The experimen- tal tank had a water temperature within 1° or 2°C of the water in which the fish were caught and transported or in which dilution occurred. Within 30-60 min after being placed in the tank, schools and individual mullet swam throughout the tank at relatively uniform speeds. Single and grouped individuals "grazed" along the sides and bottom of the screen in the tank. The behavior of the mullet at this time appeared to be similar to the behavior of undisturbed fish observed in the field. The water in the tank was cooled and heated simultaneously after the fish demonstrated what appeared to be "normal" schooling and grazing behavior. At half-hour intervals the temperature at various levels was recorded and the behavior and depth and temperature distribution of the mullet were noted. The observations required about 1 or 2 min to complete. Observations con- tinued until the vertical distribution of the mullet did not change with respect to specific water tem- peratures during two or three consecutive observa- tions. This occurred between 4 and 8 h after com- mencing the experiments, when the maximum water temperature in the tank was between 36.0° and 40.8°C, and the minimum temperature was between 13.0° and 19. 0°C. Upon termination of the observations, the heating and chilling equipment was turned off, water samples were collected, and one of the sides of the enclosure around the tank was slowly removed. The lights were dimmed slowly and turned off by a rheostat. The fish were left in the tank overnight, exposed to natural twilight conditions in the evening and morning, the room having a number of large windows. Dur- ing the evening and overnight periods the tank temperature gradually became more uniform. Well before a second series of observations was to be made during the next day, the side of the tank's enclosure was replaced, the lights having been turned on during the twilight period by a 302 rheostat. The tank was also oxygenated for ap- proximately 30 min, followed by at least an addi- tional 30-min period during which the tank was not disturbed. The second day of observations served to check on the experimental procedures and results obtained the first day, and to achieve higher tank temperatures than reached during the first day. Statistical comparisons between first and second day activity were not significant (P&0.05), indicating that the fish did not change their depth or temperature distributions. RESULTS Experimental Prior to the onset of heating and chilling, fish generally <50 mm SL (Table 1), with one excep- tion, initially concentrated near the surface or in the upper half (40-80 cm) of the tank. Fish in the 30- to 50-mm SL size ranges in 0%o salinity were initially concentrated in the lower half ( 0-40 cm) of the tank. Fish generally ^50 mm SL initially con- centrated near or on the bottom, or in the lower half of the tank. However, fish of all sizes continu- ously moved throughout the 80-cm deep tank, in- dicating that the light and pressure gradients and the depth of water were not limiting. When left for hours in the tank at constant temperature condi- tions at each test salinity, the mullet exhibited the same behavior, distribution patterns, and move- ments as those observed in mullet before heating and chilling were initiated during the experi- ments. These observations served as a control for the distributional patterns of mullet observed under test conditions. Experimental results were grouped and ana- lyzed according to test salinities (0, 15, and 34%o) and mullet size ranges (20-30, 30-50, and 2^50 mm SL) (Table 1). There was overlap and occasional fish outside of specific size ranges in some experi- ments due to the size composition of the individual mullet caught in schools in the field and sub- sequently put into the experimental tank. The experimental time period was divided for analysis into five observation intervals, each consisting of three observations (i.e., 1 complete hour of obser- vation with observation being made each half hour). Only data for observation intervals 1,3, and 5 are presented in Figure 3. Each graph consists of open histograms for the temperature and depth distributions for fish of specified size ranges, test salinity, and observation interval. Observations of MAJOR: ASPECTS OF ECOLOGY OF STRIPED MULLET depth distribution were made to within 5-cm in- tervals; the means given in the graphs being the mean of a given interval (e.g., 47.5 cm for 45-50 cm). Statistical comparisons for depth or tempera- ture distributions were significant (analysis of variance, Ps^O.OOl) in all but a few cases. The following comparisons were made, the exceptions to P ^0.001 values being noted in parentheses: 1) between observation intervals for given fish size ranges and test salinity (no exceptions), 2) be- tween salinities for given fish size ranges and ob- servation interval (depth distribution for fish ^50 mm SL in observation interval 5 and salinities 15 and 34%o (P = 0.27); depth distribution for fish 20-30 mm SL in observation interval 1 and salinities 0 and 15%o (P <0.003); temperature dis- tribution for fish 2=50 mm SL in observation inter- val 1 and salinities 0 and 15%o (P <0.002)), and 3) between size ranges for a given test salinity and observation interval (depth distribution for a sa- linity of 34%o in observation interval 5 and fish size ranges 30-50 and ^50 mm SL (P <0.004); temper- ature distribution for salinity 34%o in observation interval 5 and fish size ranges 20-30 and 30-50 mm SL (P = 0.46)). Changes in depth distribution offish are readily discernable in the histograms in Figure 3. At each test salinity, with the exception of 30-50 mm SL fish at 0%o salinity, mullet «50 mm SL moved downwards in the tank to a mean depth of 47.5 or 52.5 cm by the last observation interval. Fish ^50 mm SL moved up to mean depths of 32.5 to 42.5 cm. Fish 30-50 mm SL at 0%o moved from an initial distribution in the bottom half of the tank to a mean of 52.5 cm during the remainder of the ex- periments (observation intervals). As test salinities decreased so did the final depth distribu- tion for given fish size ranges, except for 30- to 50-mm SL fish. Changes in fish depth distribution were directly related to tank temperature, since water tempera- ture decreased with depth. However, temperature values changed rapidly between depths 60 and 20-30 cm and were relatively isothermal and cold between 20-30 cm and the bottom of the tank, and isothermal and hot above 60 cm. As a result only small differences in final depth distribution cor- responded with relatively large differences in final temperature distribution. The mean selected temperature tended to in- crease between observation intervals 1 and 5 at each test salinity for given fish size ranges. The exception to this tendency was fish ^50 mm SL at 0%o salinity (20.9°-19.5°C). Fish «50 mm SL (20-30 and 30-50 mm) tended to select higher final observation interval 5) mean temperatures (30.0°-32.3°C) than did fish ^50 mm SL (20.0°- 19.5°C) at each test salinity. For all fish size ranges the final mean selected temperature tend- ed to decrease as the test salinity decreased. This decrease was greatest for fish >50 mm SL (29.0°- 19.5°C). The depth and temperature distribution results taken together indicate that temperature selec- tion was the more important, depth distribution being secondarily related. Other gradients such as light, pressure, and oxygen, if present, did not appear to influence the distribution of the mullet at least to the extent that temperature did. Fish ^50 mm SL appeared to have a predilection to- wards the surface whereas fish ^50 mm SL ap- peared to be predisposed towards the bottom. This was evident before the initiation of heating and chilling in each experiment and during constant temperature control experiments. As the experi- ments progressed it also appeared as if mullet ^50 mm SL, in most instances, were 'Torced" down- wards by rising temperatures. Similarly, mullet ^50 mm SL were "forced" upwards by decreasing water temperature, and then downwards by rising temperature, such that their final temperature and depth distributions were somewhat lower than those for fish «50 mm SL in similar condi- tions. Just how important the actual temperature and depth distribution values selected by mullet are is unknown. What does appear to be important is the relative difference between distributions for fish generally <50 mm SL as compared with those for fish generally >50 mm SL at each salinity, and the relative changes which occurred between salini- ties for each fish length interval. The predisposi- tion offish <50 mm SL towards the surface (and higher temperatures) may be adaptive in the field. Warm water rises and the warmest (hottest) water is usually on the surface. By moving in the surface layer fish <50 mm SL may be able to orient to- wards the shallowest water inshore, which at low tide should also be the warmest. The predisposi- tion offish >50 mm SL towards the bottom of the tank (and cooler temperatures) may similarly be adaptive in the field. In this case movement away from areas subjected to tide pool formation may be important for survival. 303 FISHERY BULLETIN: VOL. 76, NO. 2 g I 34°'oo Salinity 1^ ^ ) 30 50 mm SL ) n - 414 XD = 52 5 cm XT - 28.2 °C. B2 n = 414 XD = 57 5 cm XT = 32.9°C 83 n = 76 XD = 47 5 cm XT = 32.2 °C -1 I p I ' r-i t--i I FT _lJ L-J I I I I 1 I II 1 1 Q j34°'ooSalinftyli ) > 50 mm SL < XD = 32.5cm XT = 24.6° C n = 599 f-\ —• — '— >UJ_ '""" - , -" ~- - ' '■■ -«. s-, C2 XD = 47.5cm n = 581 XT = 28.5° C _____r 1 r- ' ""'-Tl... - C3 XD = 42.5cm - n = 362 XT = 29.0° C ; 1— 1 1 1 1 1 ._ 1 1>u^,.,_ - o I— CD DC h- CO Q > o 3 O LU DC 40 p I 15°''oo Salinity l.-| XD = 57.5 cm 1 20 -30 mm SL • XT = 26.5° C n= 636 20- r'--— T^ 0 i^-tt: .. D2 n = 546 ^^_^ 20 0 XD = 57.5cm ^-^ J--, 1 xT = 30.4°c__| L.._; 'T_jT:->-^ 03 40 n = 573 XD = 47.5 cm 20 XT = 30.0° C 1 1 n .J — 1 60- 40 20 0 20 0 40 20 0 p ( 0°''oo Salinity l.i (20 30 mm SL ' n= 771 XD = 57.5 cm XT = 25.2° C F2 n=829 XD = 62.5 cm XT - 30.0° C F3 n = 925 XD = 47. 5 cm XT = 30.8° C 0 10 20 30 4^ I 31 0/ 30 50 mm SL 156 °° Salinity ii 1 <;i I G2 n = 99 XD - 52.5 cm XT = 29.4° C XD = 17 5 cm XT = 28.0° C j:: G3 n = 99 XD = 52.5 cm XT = 30.3°C JL H .tO°'oo Salinity I,., XD = 22.5 cm I > 50 mm SL ( XT = 20.9° C - n = 99 XD = 27.5 cm XT = 21.3° C 40 50 60j 70 80 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 XD = 32.5 cm XT = 19.5° C EXPERIMENTAL TANK DEPTH (cm) 15 20 25 30 35 40 15 20 25 30 35 4015 20 25 30 35 40 EXPERIMENTAL TANK TEMPERATURE (°c) 304 MAJOR: ASPECTS OF ECOLOGY OF STRIPED MULLET Figure 3. — Experimental temperature (dashed line) and depth (solid line) distributions for indicated mullet size ranges in the 80-cm deep tank at the indicated salinity (34, 15, or 0%o). No experiments were conducted with fish 30-50 mm SL at 15%o salinity. Each of the eight salinity-fish size range "conditions" are subdivided into three observation intervals: 1) observation interval 1, corresponding to the first three experimental obser- vations (1-3); 2) observation interval 3 (7th-9th observations); and 3) interval 5 (13th- 15th observations). Also presented are mean depth (XD) and temperature ( XT) data. Sample size (n) is based on the pooled data for the number of times each fish was observed within the observation interval for all experiments at a given salinity and for a given fish size range. Experiments varied in duration and number of observations made. Thus, the sample size fluctuates. See Table 1 for the actual (maximum) number of fish and experiments for each salinity-fish size range condition. Field Mullet Distribution The initial appearance of the 17- to 35-mm SL mullet prejuveniles, a distinct silvery, counter- shaded pelagic stage (Hubbs 1958), in the es- tuarine intertidal regions varied between 1972 and 1973. In 1972 they were observed and col- lected along the tide line (the most shoreward edge of falling or rising water, which is contiguous with deeper offshore water) in sand and mud flat tide pools and around freshwater streams and springs at the end of January. In 1973 they did not appear in these areas until the end of February. Pre- juveniles were particularly abundant in areas with the finest silt, mud, and/or sand particles near the outlets of freshwater rivers, streams, or springs. Fish 3=50 mm SL could be seen year around in the intertidal areas. The main body of this study concentrated on those fish that entered the inter- tidal in 1972. Also observed were juveniles of the 1971 year class, fish ^80 mm SL, in the intertidal in early 1972, and prejuveniles and juveniles of the 1973 year class, «50 mm SL, in early 1973. The disappearance of prejuveniles-juveniles from the low tide intertidal swash zone (Hedgpeth 1957) and tide line regions appeared to be com- pleted by the end of June each year, although occasional schools were seen as late as August. However, these schools were composed offish usu- ally about 40 mm SL or larger, that moved with the tide line. Prejuvenile mullet undergo metamorphosis to juveniles after entering the estuarine intertidal region. The most evident change is the loss of the pelagic silvery coloration with a general darken- ing of all surfaces, especially the dorsal side. How- ever, a general countershaded pattern remains. Other less obvious changes include: the elongation and convolution of the intestine, development of adipose eyelids, transformation of the third anal element from a soft ray to a spine, and changes in the morphology of lips and teeth. Metamorphosis is thought to be completed at about 50 mm SL (Jacot 1920). During metamorphosis, diet and feeding habits change. I found copepods in the stomach contents of some prejuveniles in Hawaii in the estuarine intertidal regions. Other prejuveniles and juveniles had plant and animal material as well as mud or silt in their stomachs. I found that sedi- ments constituted the bulk of the diet of juvenile mullet in some localities in Hawaii. In the es- tuarine intertidal region around the island of Oahu, the sizes of these injested particles ranged from 0.02 to 0.60 mm in diameter. Odum (1968) showed that fine particulate materials are a source of adsorbed organic matter and microor- ganisms, and are important in the diet of Mugil cephalus along parts of the east coast of the United States. The change in diet from copepods to plant material and mud or silt presumably occurs con- currently with changes in intestinal length, lips, and teeth. After metamorphosis is completed, the juvenile fish move into somewhat deeper interti- dal water. Prejuveniles and juveniles of all sizes formed schools ranging in size from tens to hundreds of individuals. Prejuveniles and juveniles <50 mm SL were always observed in the shallowest, warm- est water near shore wherever they occurred (Ta- ble 2). At low tide they were located along the tide line, the shallowest water along estuarine streams, and trapped in shallow mud flat and oc- casionally sand tide pools (the swash zone). In- tense continuous feeding activity was usually ob- served. The substrate in the areas in which the small mullet occurred was covered usually by the finest inorganic sediment (sand and silt 0.02-0.60 mm in diameter). Salinity and temperature values changed daily depending on tide level (water depth), wind, bot- tom type (particle size, color, etc.), insolation, and the location of springs and streams. Often a spatial as well as temporal kaleidoscope of temperature and salinity values was recorded, especially along Wailupe Beach. On 15 March 1972 the last hour (1200-1300 h local time) of a natural fish kill was observed in 305 FISHERY BULLETIN: VOL. 76. NO, 2 Table 2. — Summary of field data collected during observations of preju venile and juveniles Mugil cephalus in Hawaii during 1972-73. Salinity Water Size Temperature range depth School Locality Habitat (mm SL) Month Tide range (°C) (%o) (cm) size Remarks Wailupe Inshore of <50 Feb, to Low 19.8-37.2 2-29.5 0.6-15 10's-IOOs Mullet trapped in tide pools. Beach. tide line in May (22.0-26.9 in No predators observed. Maunalua swash zone. center of No mullet ^50 mm SL Bay' tide pools, freshwater springs springs) observed or collected. Coral rock/ »50 Feb. to Low 23.0-35.1 2-35 2.5-30 lO's-IOOs Seaward of tide line. rubble tide May; (22.0-269 (2-10 in Occasionally attacked pools with Dec, in springs) springs) by predatory fish. open connec- tions/channels to deeper water Mud/sand <50 Feb. to High 26.1-30 1 10-35 5-30 IO's-100's In wave wash (tide line); flats, sandy June attacked by predators. beach, coral 5=50 Feb to High 26.1-30.1 10-35 2.5-90 10's-IOOs Feeding, Attacked by rock/rubble June; Dec. predators. Wailupe Tidal stream 20- May Low/ 278 — 1 8-200 lOs-IOOs In shallowest water. Feeding Stream. 200 high during low tide. Attacked by Maunalua predators at high tide. Bay' Kaupa Cement <50 Feb. to Low 21 9-36 2 0-32 0.6-13 lO's Tideline. Feeding in Pond. drainage July shallowest water. Maunalua culverts Bay' Mud bottomed channel a 50 Feb to July Low 21 6-30.1 15-32 13-30 + 10s Feeding on surface. Unable to see below surface. <50 Feb. to July High 28.1-340 15-32 0.6-13 10s Along tide line — shallow- est water. a 50 Feb to July High 28 1-340 15-32 30 + 10's Feeding. Kahana Silted areas. <50 May Low — — 0.6-5 10's Feeding No predators Bay' fishpond observed Fishpond 30- May Low 24.0-26.5 2-15 0,6-15 + 10's Feeding. No predators channels 200 observed River; <50 May Low 241-296 2-30 0,6-5 10's Feeding in tide line/shallowest mud/sand water. spits (bars) S50 May Low 24.1-296 2-31 2,5-7,5 3-5 Feeding. <50 May High 20.0-25.0 — 2,5-15 10's Feeding in tide line/shallowest water amongst mangrove roots. Kahana River; <50 May Low/ — — 2.5-15 — In shallowest water Individuals Bay2 mangrove vegetation high spread out motionless on bottom. Wailupe Mudflats; <50 Feb to Low/ — — 2,5-15 — In shallowest water Spread out Beach. coral rocks/ July high or in compact groups Maunalua rubble s50 Feb. to Low/ — — 7.5 + — motionless on bottom Bay2 Aug high In shallowest water. Individuals spread out motionless on bottom 'Daylight observations of mullet. ^ Night observations of mullet. one of the cement culverts in Kuapa Pond (Hawaii Kai), Maunalua Bay. The tide was low and just starting to turn and flood. Prior to this day heavy rains washed down large amounts of silt, rocks, and debris to a depth of about 0.2-0.5 m. A narrow channel was cut through the mud by the trickling stream, and isolated "tide pools" or pockets of water were common. An estimated several thou- sand prejuvenile and juvenile mullet 18-80 mm SL were found dead along the bottom of the culvert. Although it appeared that the fish had a free exit at higher tide levels, via the shallow ( 1 cm deep at low tide) channel to the cooler tidal region (26°- 30°C), the fish were presumably trapped physi- 306 cally by the debris and/or by a very rapid increase in water temperature. Dead mullet were found in the pockets with water temperatures of 39.5°- 42.5°C. The only survivors observed were mullet 20-35 mm SL slowly swimming in small pockets at temperatures as high as 39.0°-41.1°C. Salinity measurements were not made, but would be ex- pected to be low. Juvenile mullet 3=50 mm SL were not observed or collected in tide pools at low tide. At low tide these larger fish occurred beyond the tide line in tide pools with open connections to deeper water or along sills and sand/mud flats which sloped into deeper water. These areas were characterized by MAJOR ASPECTS OF ECOLOGY OF STRIPED MULLET higher but more uniform salinities and lower but more uniform temperatures than areas in which fish <50 mm SL were generally found. At high tide, water temperature and salinity values were nearly uniform throughout a given location. Mullet <50 mm SL were concentrated along the tide line on the beach or along the sides of rivers in the shallowest water available. Schools were dense and often composed of more individu- als than found in schools at low tide in the same area. Little feeding occurred; evasion of predators was seen more commonly. At high tide, schools of mullet with individuals >50 mm SL moved into the former tide line and tide pool areas, and were observed feeding in the areas from which the smaller mullet retreated. The results of the experiments and field obser- vations demonstrate that dynamic differences occur in the behavior and distribution of schools of mullet composed of individuals generally <50 mm SL as compared with schools with juveniles of greater length. In the field young mullet <50 mm SL prefer and select areas characterized by mini- mal water depth, tide pool formation, relatively high fluctuating temperatures, and relatively low fluctuating salinities. Juveniles &50 mm SL, on the other hand, seek somewhat deeper water and tide pools with lower more uniform temperatures and higher more uniform salinities. Experimentally, a wide range of temperature values ( 13°-40.8°C) was available to the fish. How- ever, mullet <50 mm SL tended to concentrate in water of higher temperature near the surface of the tank and fish >50 mm SL tended to occur deeper in the tank at lower temperatures for each test salinity. Although the experimental tank al- lowed fish a means of behaviorally escaping lethal or near lethal conditions (as presumably did deeper tide pools for mullet >50 mm SL in the field), entrapment in shallow intertidal tide pools in the field did not. Fish <50 mm SL appeared to actively seek such near lethal conditions in the field (and experimentally), and as observed in a Kuapa Pond stream culvert, occasionally perished as a result. Predators Most predators observed interacting with mul- let during this study were solitary stalking or stationary "sit-and-wait" species (Table 3). At- tacks by predators upon schools of mullet with individuals <50 mm SL were almost nonexistant at low tide. Where deeper water was immediately contigu- ous with a shallow water shelf (e.g., along Kahana River), predators (e.g., barracuda) in the deeper water were observed orienting towards and paral- leling the movements of schools of the small mul- let feeding in the shallower water. When mullet strayed off the shelf into deeper water, they were attacked. Predation upon mullet &50 mm SL dur- ing low tide was occasionally observed as schools moved and fed in the deeper intertidal region. At high tide, schools of these larger mullet con- tinued to be attacked by predators, as they were at low tide. Similarly, during high tides and ebb and flood periods, predators attacked mullet schools with individuals <50 mm SL. At night, along Wailupe Beach and the Kahana River (Table 2), mullet schools broke up and indi- viduals spread out and remained relatively mo- tionless near the bottom. The fish slowly moved Table 3. — Predatory fish observed interacting with schools of mullet. Location Species Standard length (mm) Water depth (cm) Tide Remarks Wailupe Beach Lizardfish Saunda gracilis 60-175 5-23 Low/' high Needlefish 100-300 30-75 High Tylosurus crocodilus Great barracuda. 50-225 30-75 Low high Sphyraena barracuda Wailupe Stream Great barracuda 40-600 30-90 Low/high Hawaii Kai Great barracuda 30-250 15-75 High (Kuapa Pond) culverts Kahana Bay Great barracuda 125-500 15-60 Low/high River Eleotris sandwichensis 78 7 5-30 Low In tide pools with open connections and channels to deeper water Attacked and chased juveniles of all sizes. A possible predator, moved in with flood tide Observed following schools of mullet near surface. Moved inshore with flood tide; attacked and chased individuals of all sizes in schools Followed feeding individuals in schools. Swam slowly along the shoreline at high tide tvlullet <50 mm SL usually in shallowest water 1.8-75 cm deep. Barracuda in deeper water followed or paralleled movements of mullet in shallow water and attacked when the mullet strayed into deeper water Moved in with flood tide Water turbid: caught in seines with juvenile mullet No mullet found in stomach contents At low tide mullet • 50 mm SL in shallowest water 15 cm deep, barracuda followed (paralleled) schools cf mullet In shallower water (see above). Single unsuccessful attack on school of mullet of individuals about 40 mm SL. Caught after attack. 307 FISHERY BULLETIN: VOL. 76, NO. 2 (drifted) with the tides. The break up of schools may have been a result of reduced predation and/or lowered visual sensitivity thresholds (Munz and MacFarland 1973). When a school was attacked, it usually split into two or more segments and passed around behind the predator to reform a single school again. When a predator was successful in separating an indi- vidual from a school, a chase occurred, the results of which were seldom observed. Of the approxi- mately 50 lizardfish stomach contents analyzed, one contained a juvenile mullet. None of the 10 barracuda stomach contents analyzed contained juvenile mullet. Potential invertebrate predators were abun- dant in the various habitats where mullet occur- red. However, only individuals of a single crab species, Thalamita crenata, were observed stalk- ing and extending their chelipeds toward passing mullet. In one instance an individual crab did cap- ture a juvenile mullet, but only after it had been wounded by and escaped from a barracuda. DISCUSSION Mugil cephalus is a worldwide (lat. 42°N-42°S, Thomson 1966) inhabitant of the estuarine inter- tidal as well as freshwater and coastal marine environments (Broadhead 1953, 1955; Hendricks 1961; Thomson 1963, 1966; Johnson and McClen- don 1970). In Hawaii, selective pressures appear to have favored prejuvenile and juvenile mullet that are able to survive in the shallowest, warmest estuarine intertidal waters, waters that are characterized by temporal and spatial hetero- geneity with respect to temperature, salinity, and depth. Before discussing the adaptations evolved by striped mullet making possible survival in es- tuarine intertidal regions, a discussion of the en- vironmental variables important to young mullet in Hawaii might be in order. The monthly occurrence of mullet <50 mm SL observed in 1972 and 1973 in Hawaii is presented with data for 12 consecutive years (1962-73) of recorded (skycover, rainfall, seawater tempera- ture) and predicted (tidal) data in Figure 4. These appear to be the most important environmental factors that bear directly upon the lives of mullet in the estuarine intertidal region. Indirectly, the length of daylight (time from sunrise to sunset) may also be important; it is shortest (about 10.9 h) about 22 December each year, and longest (13.3 h) about 21 June each year. Visual observations and collections of mullet < 50 mm SL indicate that these mullet occur in the Hawaiian intertidal estuarine regions during the months when there are a maximum number of low tides ^0.0 m (mean tide level at Honolulu is 0.2 m (0.8 ft)). Perhaps of greater importance is the occurrence of mullet when there is a minimum number of high tides ^0.6 m (2.0 ft). In Maunalua Bay, tide pools begin to form when the tide level is approximately 0.06 m (0.2 ft). The number of tides that would result in tide pool formation at noon ( 1000-1400 h local time) begins to decrease during the time of the year mullet are undergoing metamorphosis in the intertidal estuarine region, but is still maximal when prejuveniles first enter the inshore areas. Thermal and salinity stresses should be maximal during the noon time period. It is not known whether this tidal-estuarine interti- dal situation is unique to Hawaii or of more wide spread occurrence. Also unknown is whether the peak occurrence of young mullet during such tidal relationships is fortuitous, or whether selection pressures have resulted in a shift of the peak oc- currence from either earlier or later in the winter-spring season to its present "position" in April. The extent to which stress occurs in the interti- dal estuarine region may be ameliorated by low ambient (oceanic) seawater temperatures and maximum cloud cover. During the time mullet <50 mm SL are found in the intertidal estuarine region in Hawaii, seawater temperatures are min- imal and increasing, and average cloud cover is seasonally maximal. During the late winter- spring, the lowest seasonal seawater tempera- tures occur in the tropical-temperate Northern Hemisphere, and increase until maximum levels are reached about September. The average maximum amount of cloud cover, which occurs in Hawaii during April (when young Figure 4.— The relative abundance of mullet <50 mm SL in the Hawaiian estuarine intertidal tide pools compared with environ- mental data collected (seawater temperature, rainfall, and sky cover) and predicted (tides). The monthly means (cormected by horizontal lines), ranges (vertical lines), and standard deviations (vertical boxes) were derived from data for the 12-yr period, 1962-73. Mullet abundance data were from field observations and collections in 1972-73. Sky cover data were derived from monthly average values. Sky cover and rainfall data were taken from Climatological Data, Hawaii, U.S. Weather Bureau, NOAA; tidal information from Tide Tables, West Coast of North and South America including the Hawaiian Islands, National Ocean Survey, NOAA; seawater temperature data were col- lected by the National Marine Fisheries Service, Honolulu. 308 MAJOR: ASPECTS OF ECOLOGY OF STRIPED MULLET JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 309 FISHERY BULLETIN: VOL. 76, NO 2 mullet are most abundant), reduces insolation and thus reduces the potential for the attainment of lethal conditions in tide pools. These relationships are particularly important since young mullet congregate in areas of springs and freshwater run-off. However, cloud cover during February through May varies more than during any other period of the year and points to the fact that the environment in which mullet <50 mm SL occur fluctuates tremendously within a season and from year to year. Seasonal rainfall is maximal during winter- spring in Hawaii. Rainfall in the mountains (Kaneohe Mauka Station) exceeds that in nearby shore regions (Hawaii Institute of Marine Biology Station). Fluctuations in rainfall from month to month and year to year are great and appear to be unpredictable, particularly during the season when mullet <50 mm SL are abundant in the intertidal estuarine region. Run-off is maximal during this season due to the heavier mountain rainfall. This run-off could contribute to poten- tially lethal or near lethal conditions in the inter- tidal region by reducing its salinity, particularly during periods when low tides occur during noon. Intertidal spring water temperatures, on the other hand, were recorded to be as much as 10° cooler than surrounding water of higher salinity (Table 2). This cooler water may serve to reduce overall intertidal estuarine temperatures, at least during nontide-pool forming tide levels, but at the same time it increases the thermal and salinity heter- ogeneity of the environment. Returning to a discussion of adaptations of mul- let, experimental studies indicate that tempera- ture acclimation is important in the ability of striped mullet ( at least larger juveniles ) to survive higher temperatures (Heath 1967; Sylvester 1974, 1975; Sylvester et al. 1974). Heath, although not providing fish length or salinity data at which tests were made, reported critical thermal max- ima (CTM) of 42.4°-43.rC for mullet in the north- ern Gulf of California. Sylvester (1974, 1975) and Sylvester et al. (1974) demonstrated increased CTM (29.0°-41.6°C) with increasing acclimation temperatures at a salinity of 32%o for juvenile striped mullet, 70-125 mm SL, in Hawaii. At a lower acclimation temperature, and a salinity of 0%o, the CTM's were reduced. In general, CTM's were lower at lower salinities. Sylvester (1974) also found that juveniles adjusted or acclimated faster at higher temperatures, and that their thermal resistance to lethal temperatures de- 310 creased slightly when they were exposed to fluc- tuating low, rather than constant, temperatures. Sylvester (1975) also demonstrated the exis- tence of increased CTM at noon and lower CTM in the morning and afternoon for fish 78-122 mm SL. There is good evidence that underlying bio- chemical changes are responsible for acclimation to changing thermal regimes (reviews in Hochachka and Somero 1971, 1973; Haschemeyer 1973; Somero 1975). Hochachka and Clayton- Hochachka (1973) provided some evidence that this may also be the case for striped mullet about 120 mm long in Hawaii. Whether the ability of mullet to tolerate in- creasingly higher temperatures, seasonally and daily, is a result of interacting endogenous factors (biological rhythms) as is indicated in other ani- mals (Sweeney and Hastings 1960; Wilkins 1965), and/or exogenous factors (direct exposure to in- creasing temperatures) is not known. Sylvester's (1974, 1975) studies were conducted between August and January and those of Heath ( 1967) in March and September. As the discussion above and Figure 4 indicate, ambient seawater temperatures are highest during August to Oc- tober and lowest during February to April in the tropical-temperate Northern Hemisphere. Al- though the seasons varied during which the exper- iments of Sylvester and Heath were conducted, the CTM data obtained were similar for the experi- mentally acclimated fish. Doudoroff (1957) and Allen and Strawn ( 1971) reported that relatively brief exposure to high nonlethal temperatures usually increased heat resistance in a number of species of fish. This increased resistance was not readily lost when fishes were subsequently ex- posed to low temperatures. This also appears to be true for striped mullet as Sylvester's ( 1974) study indicates. Thus, striped mullet appear to have an ability to modify their thermal tolerance in direct response to prevailing environmental conditions (exogenous factors). The ability to increase their heat resistance even after a brief exposure to high nonlethal temperature would be especially advan- tageous to mullet in the estuarine intertidal, at least in Hawaii. If exogenous factors are solely responsible for the ability of striped mullet to survive high tem- peratures, it is difficult to explain the differences in the distribution of striped mullet presented in this report. Prejuveniles enter the intertidal estuarine re- gions from the far more environmentally uniform MAJOR: ASPECTS OF ECOLOGY OF STRIPED MULLET oceanic waters, and appear to be "preadapted" to the near lethal conditions inshore. This may indi- cate the existence of an ontogenetic biological rhythm (cued by slight monthly changes in photo- period or water temperature?) in these mullet, which biochemically and physiologically pre- adapts these fish for life in the intertidal estuarine regions, while they are still in oceanic waters. Seasonal and daily acclimation to existing ther- mal and salinity regimes may then occur after the prejuveniles arrive inshore. Factors affecting the distribution of striped mul- let 3^50 mm SL are more complex. In Hawaii, field acclimated fish behaviorally select or prefer water temperatures well below their experimental CTM. The natural fish kill observed in March also indi- cates that the larger juveniles (and most pre- juveniles) were not able to survive exposure to high temperatures (at least for relatively long periods of time). Behavioral selection of tempera- ture regimes well below CTM has also been ob- served in the estuarine goby, Gillichthys mirabilis (de Vlaming 1971). Mullet &50 mm SL in Hawaii moved seaward to pools with open connections to deeper water with ebbing tides during the spring. During the non- tide-pool forming low tides in the late summer, the larger juveniles may be acclimated to tolerate higher ambient seawater temperatures in Hawaii as they are in the northern Gulf of California. In Hawaii, however, intertidal waters reach their maximum temperatures during the period from late winter to early summer when shallow tide pools are formed, although ambient oceanic temp- eratures are higher during late summer. Heath (1967) indicated that inshore water temperatures in the northern Gulf of California are highest dur- ing the late summer-early fall period. It is not known whether CTM's are similar for field accli- mated striped mullet ^50 mm SL in the spring in Hawaii and in the late summer in the northern Gulf of California, or whether mullet from both locations have CTM's parallelling seasonal changes in ambient (oceanic) seawater tempera- ture. Selected temperatures are lower for young mul- let ^50 mm SL compared with smaller fish at each salinity, and much lower at decreasingly lower salinities (Figure 3). Presumably, physiological changes mediated hormonally/biochemically occur during metamorphosis, resulting in a pref- erence for reduced temperatures by the larger juveniles. This decrease in temperature "toler- ance" with metamorphosis (age) is somewhat con- trary to the discussion thus far. It may only be a behavioral trait not directly correlated with CTM (i.e., CTM may actually be increasing). Behavioral selection of lower temperatures appears to be adaptive in the field during the period between April and August. In addition to having widely varying salinity values, many of the tide pools formed during this period may have been too shal- low for larger juveniles to feed and swim. Many of the pools were shallower than the body depth of the older juveniles. Thus, those individuals that remained seaward of the tide line as they com- pleted metamorphosis may have reduced or es- caped the possibility of entrapment and exposure to lethal conditions in tide pools and shallow wa- ter; conditions observed once during this study. These relationships may indicate the existence of an endogenous rhythm involved in the move- ment of (behavioral selection by) juvenile mullet towards deeper, relatively cooler, more saline water during or after metamorphosis. This rhythm may be acting in opposition to the pre- sumably increased acclimation to higher ambient (oceanic) seawater temperatures. The change in behavior with metamorphosis may be a result of endogenous rhythms perhaps coupled with exogenous factors, such as the slight monthly changes occurring in seawater temperature and/or photoperiod, or it may be due directly to these exogenous factors. The reproductive cycle of striped mullet appears to be coupled with both these environmental variables (Kuo et al. 1974; Kuo and Nash 1975), so presumably younger indi- viduals could use these same cues as well. It is difficult to separate cause from effect, but the shal- lowness and volume limitations of tide pools may be critical. Thus, selection may have favored those metamorphosed individuals with reduced physio- logical tolerance to high fluctuating temperatures and low fluctuating salinities as found in the es- tuarine intertidal (i.e., those individuals that be- haviorally moved away from such conditions). If selection favored those metamorphosed (metamorphosing) individuals that moved into deeper intertidal waters, what selection pressures may have favored individuals able to survive the kaleidoscopic conditions of the estuarine interti- dal tide pools? Experimental and field evidence demonstrate the importance of refugia for species from their competitors and/or predators (Cause 1934; Crom- bie 1946; Connell 1961; Paine 1969). Connell 311 FISHERY BULLETIN: VOL. 76, NO. 2 proposed that intertidal species are limited in their upper distributional range by physiological (and presumably biochemical) adaptive abilities to environmental stress. At the lower end of the range, organisms are limited by biotic factors such as competition and predation. Field observations indicated that there was essentially no predation, including that by birds, of mullet <50 mm SL when they occupied the intertidal estuarine tide line and swash zone areas at low tide (Tables 2, 3). At high tide, and during ebb and flood, mullet <50 mm SL were exposed to predators, but the poten- tial for being attacked and caught was reduced by occupying the shallowest tide line waters and by the schooling habit. The absence of predatory fishes in the shallow intertidal estuarine regions at low tide may be related to 1) a subminimal depth or area of water in which to maneuver, and 2) possibly, although data are lacking, an inability of predators to adjust to rapidly fluctuating thermal and salinity re- gimes. Predators escaped entrapment by remain- ing seaward of the tide line during ebbing tides just as juvenile mullet s^50 mm SL did. In addi- tion, potential invertebrate predators were absent from the shallow intertidal areas presumably es- caping seaward and/or, as often observed or caught, burrowing to a level below the surface of the mud/sand substrate. Juvenile mullet ^50 mm SL as well as adult mullet appear to be competitors with individuals <50 mm SL for food resources in the intertidal estuarine region. At high tides the larger fish moved into and fed in the areas used by the younger fish during low tides. With the incoming tide, the younger fish moved shoreward with the tide line. In addition, other species of fishes moved in with the flood tides. It is not known whether these fishes utilized the same food resources as the young mullet. Space also may be at a premium in the shallow intertidal estuarine regions, particularly in tide pools. As discussed previously, the volume of water in the tide pools as well as the depth of pool water may be critical. Formation of large schools is characteristic of larger juvenile mullet as it is for the species as a whole and inter- and intra- specific competition for space may occur. Other species of fishes, as well as the larger juvenile mullet, were not observed in the shallowest water during low tide. The ability of small mullet to occupy the shallowest, warmest water may also occur in the northern Gulf of California (Heath 312 1967), where they (no length data) are one of two species penetrating the farthest up seawater drainages along the margin of the desert. Mullet <50 mm SL formed loose schools with individuals constantly feeding during low tides, particularly in tide pools. At high tides, or in more exposed environments, feeding often ceased and tight dense schools were formed. This was evident when predators were nearby, approaching, or at- tacking. When exposed to predators at high tides and changing tides, the schooling habit confers an increased advantage to the mullet in terms of sur- vival (Major 1977, in press). Most of the attacks by predators on schools that I observed, failed. The formation of schools appears to be yet another adaptive feature in the behavioral repertory of mullet. The schooling habit increases the ability of individual mullet to survive as prejuvenile and juveniles in the intertidal estuarine region and presumably as prejuveniles in oceanic waters. Prejuvenile and small juvenile mullet have pre- sumably evolved the necessary biochemical and physiological adaptations to exist successfully in the fluctuating, often near lethal, intertidal es- tuarine environment in the spring months. In Hawaii, this has allowed them to use this interti- dal refugium to escape their predators and com- petitors for food and space, making possible undis- turbed feeding activity. Kinne's (1960) work with Cyprinodon tnacularius and Norris's (1963) study of Girella nigricans suggest that high tempera- tures increase the rate of food uptake and diges- tion. Food conversion (to growth) efficiencies are highest at lower temperatures, however. De Silva and Perera (1976) experimentally determined that young mullet grow more rapidly at 20%o sa- linity than at salinities of 30, 10, or l%o. This was comparable to Kinne's (1960) work with C. macu- larius. The widely fluctuating environmental variables in the estuarine intertidal in Hawaii may provide the necessary conditions for rapid grovvi;h in mul- let. This would allow metamorphosis to be com- pleted in all mullet by the time the tidal situation becomes less favorable, as predators gain access to the small mullet and intra- and interspecific com- petitors gain access to their feeding areas as well. The formation of large schools during all stages of life appears to be important in reducing predation and possibly also in competing with other species for food and space in the estuarine intertidal re- gion. MAJOR: ASPECTS OF ECOLOGY OF STRIPED MULLET It is interesting to note that the environmental conditions and the behavior of prejuveniles and juvenile striped mullet in Hawaii appear to be very similar in many instances to those of various species of western U.S. desert pupfish, Cyprinodon (Barlow 1958, 1961; Kinne 1960; Lowe and Heath 1969; Brown and Feldmeth 1971; Deacon and Minckley 1974), and to the African cichlid, Tilapia grahami (Coe 1966). Daily and seasonal changes in water temperature and possibly salinity (ionic) regimes in the low tide, tide pools in Hawaii and desert springs and pools appear very similar. The pupfish and T. grahami young live in the shal- lowest, hottest water often at near lethal tempera- tures. Adults, generally, did not occur in these areas at the same time as the young. Feeding appeared to be continuous and was directed at the substrate, as it was in mullet <50 mm SL. The shape, size, and length of the pupfish and T. grahami also appear to be very similar to those of the small mullet. The changes occurring in the behavior and dis- tribution of mullet prejuveniles transforming to juveniles is also very similar to the changes occur- ring in prejuvenile and juveniles opaleye, Girella, nigricans, in the intertidal areas along southern California and Baja California, Mexico (Norris 1963). The ability of certain life history stages of these diverse species of fishes to tolerate fluctuating conditions and/or near lethal thermal and ionic (salinity) regimes in shallow water possibly indi- cates convergence of adaptations to similar envi- ronments. The physiological adaptations may be mediated biochemically (hormonally) and may be a result of the interaction of both endogenous and exogenous factors or cues. The evolutionary driv- ing or selection forces operating appear to include predation and at least intraspecific competition for food and space. ACKNOWLEDGMENTS I thank G. W. Barlow, A. R. Moldenke, W. H. Neill, Jr., K. S. Norris, J. S. Pearse, M. Silver, and anonymous reviewers for editorial advice. Craig Emberson, Kenneth Dormer, Diane Henderson, Ching-Ming Kuo, George Lauder, William Mad- den, Elaine Major, and Robert Shallenberger pro- vided assistance during parts of the study. The figures were prepared by Doris Heinsohn (UCSC), and the Audio- Visual Graphics staff at Simon Eraser University, Canada. The experiments were conducted under NSF Grant No. GA 28195 to K. S. Norris and, subsequently, C. E. Nash, past and present Research Director, respectively. Oceanic Institute (Oceanic Foundation), Makapuu, Hawaii. LITERATURE CITED ALLEN, K. O., AND K. STRAWN. 1971. Rate of acclimation of juvenile catfish, Ictalurus punctatus, to high temperatures. Trans. Am. Fish. Soc. 100:665-671. Barlow, G. W. 1958. Daily movements of desert pupfish, Cyprinodon macularius, in shore pools of the Salton Sea, California. Ecology 39:580-587. 1961. Social behavior of the desert pupfish, Cyprinodon macularius, in the field and in the aquarium. Am. Midi. Nat. 65:339-359. BROADHEAD, G. C. 1953. Investigations of the black mullet, Mugil cephalus L., in northwest Florida. Fla. State Board Conserv., Tech. Ser. 7:1-34. 1955. Growth of the black mullet {Mugil cephalus L.) in west and northwest Florida. M.S. Thesis, Univ. Miami, 83 p. BROWN, J. H., AND C. R. FELDMETH. 1971. Evolution in constant and fluctuating environ- ments: thermal tolerances of desert pupfish (Cyprino- don). Evolution 25:390-398. CARR, W. E. S., AND J. T. GlESEL. 1975. Impact of thermal effluent from a steam-electric sta- tion on a marshland nursery area during the hot sea- son. Fish. Bull., U.S. 73:67-80. COE, M. J. 1966. The biology of Tilapia grahami Boulenger in Lake Magadi, Kenya. Acta Trop. 23:146-177. CONNELL, J. H. 1961. The influence of interspecific competition and other factors on the distribution of the barnacle Chthamalus stellatus. Ecology 42:710-723. Crombie, a. C. 1946. Further experiments on insect competition. Proc. R. Soc. Lond. Biol. Sci. 133:76-109. DEACON, J. E., AND W. L. MINCKLEY. 1974. Desert fishes. In G. W. Brown, Jr. (editor), Desert biology, Vol. 2, p. 385-488. Academic Press Inc., N.Y. DeSILVA, S. S., AND P. A. B. PERERA. 1976. Studies on the young grey mullet, Mugil cephalus L. I. Effects of salinity on food intake, growth and food con- version. Aquaculture 7:327-338. DE VLAMING, V. L. 1971. Thermal selection behaviour in the estuarine goby, Gillichthys mirabilis Cooper. J. Fish. Biol. 3:277-286. DOUDOROFF, P. 1957. Water quality requirements of fishes and effects of toxic substances. In M. E. Brovm (editor), The physiol- ogy of fishes. Vol. 2, p. 403-430. Academic Press Inc., N.Y. FRY, F. E. J. 1964. Animals in aquatic environments: fishes. In C. G. Wilbur (executive editor). Handbook of physiology. D. B. Dill (section editor), Adaptations to the environment, p. 715-728. Am. Physiol. Soc. 313 FISHERY BULLETIN: VOL 76, NO. 2 Cause, G. F. 1934. The struggle for existence. Williams and Wilkins, Bait., 163 p. Haschemeyer, a. E. V. 1973. Control of protein synthesis in the acclimation offish to environmental temperature changes. In W. Chavin (editor). Responses of fish to environmental changes, p. 3-33. Charles C. Thomas, Publ., Springfield, 111. HEATH, W. G. 1967. Ecological significance of temperature tolerance in Gulf of California shore fishes. J. Ariz. Acad. Sci. 4:172- 178. Hedgpeth, J. W. 1957. Sandy beaches. In J. W. Hedgpeth (editor). Treatise on marine ecology and paleoecology. Vol. I. Ecol- ogy, p. 587-608. Geol. Soc. Am. Mem. 67. Hendricks, L. J. 1961. The striped mullet, Mugil cephalus Linnaeus. In B. Walker (editor). The ecology of the Salton Sea, Califor- nia, in relation to the sportfishery, p. 95-103. Calif Dep. Fish Game. Fish Bull. 113. HOCHACHKA, P. W., AND B. CLAYTON-HOCHACHKA. 1973. Glucose-6-phosphate dehydrogenase and thermal acclimation in the mullet fish. Mar. Biol. (Berl.) 18:251-259. HOCHACHKA, P. W., AND G. N. SOMERO. 1971. Biochemical adaptations to the environment. In W. S. Hoar and D. J. Randall (editors). Fish physiology, Vol. 6, p. 99-156. Academic Press Inc., N.Y. 1973. Strategies of biochemical adaptation. W.B.Saun- ders Co., Phila., 358 p. HUBBS, C. L. 1958. Dikellorhynchus and Kanazawaichthys: nominal fish genera interpreted as based on prejuveniles of AfaZa- canthus and Antennarius respectively. Copeia 1958:282-285. IVLEV, V. S., AND H. A. LEIZEROVICH. 1960. An ecological analysisof the distribution of animals under the conditions of a temperature gradient. Tr. Murm. Morsk. Biol. Inst. 5:1-36. (Transl. from Russ.: Fish. Lab., Lowestoft, Suffolk, Engl., E. 712.) Jacot, a. P. 1920. Age, growth and scale characters of the mullets, Mugil cephalus and Mugil curema. Trans. Am. Microsc. Soc. 39:199-229. Johnson, D. W., and E. L. McClendon. 1970. Differential distribution of the striped mullet, Mugil cephalus Linnaeus. Calif. Fish Game 56:138-139. KINNE, O. 1960. Growth, food intake, and food conversion in a eury- plastic fish exposed to different temperatures and salini- ties. Physiol. Zool. 33:288-317. Kuo, C. M., AND C. E. Nash. 1975. Recent progress on the control of ovarian develop- ment and induced spawning of the grey mullet (Mugil cephalus L.). Aquaculture 5:19-29. Kuo, C. M., C. E. Nash, and Z. H. Shehadeh. 1974. The effects of temperature and photoperiod on ova- rian development in captive grey mullet (Mugil cephalus L.). Aquaculture 3:25-43. LAUFF, G. H. (editor). 1967. Estuaries. Am. Assoc. Adv. Sci., Publ. 83, Wash., D.C., 757 p. LOWE, C. H., AND w. G. Heath. 1969. Behavioral and physiological responses to tempera- 314 ture in the desert pupfish Cyprinodon macularius. Physiol. Zool. 42:53-59. MAJOR, P. F. 1977. Predator-prey interactions in schooling fishes dur- ing periods of twilight: A study of the silverside Pranesus insularum in Hawaii. Fish. Bull., U.S. 75:415-426. In press. Predator-prey interactions in two schooling fishes, Caranx ignobilis and Stolephorus purpure- us. Anim. Behav. Mantleman, I. I. 1958. Distribution of the young of certain species offish in temperature gradients. [In Russ.] Izvest. Vses. Nauchno- Issled. Inst. Ozemovoi Rechnovo Rybnovo Khoz. 47:1- 63. (Transl. in Fish. Res. Board Can. Transl. Ser. 257.) MUNZ, F. W., AND W. W. McFARLAND. 1973. The significance of spectral position in the rhodop- sins of tropical marine fishes. Vision Res. 13:1829-1874. NORRIS, K. S. 1963. The functions of temperature in the ecology of the percoid. fish Girella nigricans (Ayers). Ecol. Monogr. 33:23-62. ODUM, W. E. 1968. The ecological significance of fine particle selection by the striped mullet Mugil cephalus. Limnol. Oceanogr. 13:92-98. PAINE, R. T. 1969. The Pisaster-Tegula interaction: prey patches, pred- ator food preference, and intertidal community struc- ture. Ecology 50:950-961. Randall, J. E. 1961. A contribution to the biology of the convict sur- geonfish of the Hawaiian Islands, Acanthurus triostegus sandvicensis. Pac. Sci. 15:215-272. Randall, J. E., and R. K. Kanayama. 1972. Hawaiian fish immigrants. Sea Front. 18:144-153. SOMERO, G. N, 1975. Enzymic mechanisms of eurythermality in desert and estuarine fishes: genetics and kinetics, /n N. F. Had- ley (editor). Environmental physiology of desert organ- isms, p. 168-187. Halsted Press, Dowden, Hutchinson and Ross, Inc., Stroudsburg, Pa. Sweeney, B. M., and J. W. Hastings. I960. Effects oftemperature upon diurnal rhythms. Cold Spring Harbor Symp. Quant. Biol. 25:87-104. Sylvester, J. R. 1974. Thermal response of juvenile Hawaiian mullet Mugil cephalus (L.) to acclimation time and fluctuating low temperatures. J. Fish. Biol. 6:791-796. 1975. Critical thermal maxima of three species of Ha- waiian estuarine fish: a comparative study. J. Fish. Biol. 7:257-262. Sylvester, J. R., C. E. Nash, and C. E. Emberson. 1974. Preliminary study oftemperature tolerance in juve- nile Hawaiian mullet (MugiZcep/ia/us). Prog. Fish-Cult. 36:99-100. Thomson, j. M. 1963. Synopsis of biological data on the grey mullet Mugil cephalus Linnaeus 1758. C.S.I.R.O. Aust. Fish. Oceanogr. Fish. Synop. 1, 66 p. 1966. The grey mullets. Oceanogr. Mar. Biol. Annu. Rev. 4:301-335. WILKINS, M. B. 1965. The influence of temperature and temperature changes on biological clocks. In J. Aschoff (editor), Cir- cadian clocks, p. 146-163. North Holland Publ. Co., Amst. AGING OF GULF MENHADEN, BREVOORTIA PATRONUS William R. Nicholson and William E. Schaaf^ ABSTRACT Length-frequency distributions, returns of tagged juveniles, and scale annuli indicate that over 97% of Gulf menhaden, Brevoortia patronus, caught in the purse seine fishery are ages 1 and 2. Few fish survive to age 3. About 50% of the fish examined for the years 1971-73 could be aged by scale annuli. Those with no scale annuli or with indistinct or false annuli could be assigned to age 1 or younger or age 2 or older on the basis of length. When large numbers offish are routinely sampled for age and size distributions, they must be aged by some technique that consumes relatively little time. One common method is to count scale annuli, another is to group the fish by length frequencies. Counting otolith rings usually is impractical be- cause of the large amount of time and effort it takes to collect, prepare, and observe the otoliths. As pointed out by Struhsaker and Uchiyama ( 1976), "Attempts to age tropical fishes by conven- tional methods have generally been thwarted by the absence of well-defined annuli in calcarious structures and protracted spawning periods which make length-frequency mode progression analysis difficult." In temperate regions where the winter water temperature may not fall low enough to cause a cessation of fish growth for an extended period, aging fish by counting scale annuli may also be difficult. Gulf menhaden, Brevoortia patronus, range along the coasts of the United States and Mexico from Florida to Yucatan. They spawn offshore in the Gulf of Mexico from about October to April (Suttkus 1956; Turner 1969). The eggs hatch in about 48 h and the larvae are transported by on- shore currents to estuaries, where they metamor- phose into adult form (Fore and Baxter 1972). In late summer the juveniles, ranging from about 45 to 120 mm fork length (FL), congregate in the lower estuaries before moving to offshore waters (Kroger and Pristas 1975). Gulf menhaden, the basic resource for a large meal and oil industry, are caught exclusively in a purse seine fishery extending from western 'Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Center, NOAA, P.O. Box 570, Beaufort, NC 28516. Florida to eastern Texas. Processing plants, now operating only in Mississippi and Louisiana, for- merly operated in Florida and Texas also. During routine sampling of the catch during the fishing season, usually lasting from late April to October, scales have been removed from, and weights and fork lengths recorded for, about 13,000 fish annu- ally since 1964. Aging these fish by conventional methods has been a problem. Although some fish had well- defined rings that appeared to be annuli, others had no rings, or rings that were unclear or oddly spaced. Length-frequency distributions indicated two major age-groups with overlapping lengths, and a third group that appeared in late summer. Since neither length frequencies nor scale rings alone were satisfactory for aging all fish, ages sub- sequently were based on a combination of factors: appearance of scales, number and spacing of visi- ble rings, and length of fish at the time it was caught. This method of aging could be criticized as being too subjective. But until returns offish tag- ged at a known age, such as juveniles, were avail- able there were no distributions of known ages to which distributions of estimated ages based on scale rings and lengths could be compared. A study to resolve the problem was not begun until returns of juveniles tagged in late summer and early fall 1970-73 became available. In 1975 we began a study of fish collected from 1971 to 1973. We choose those years because returns of tagged juveniles of the 1970-73 year classes were available, and we limited our material to 3 yr to keep it manageable. Age-0 fish were defined as young-of-the-year that would have no scale ring, age 1 as those in their second year that should have one ring, and age 2 as those in their third year that should have two rings. Manuscript accepted August 1977. FISHERY BULLETIN: VOL. 76. NO. 2, 1978. 315 FISHERY BULLETIN; VOL. 76, NO. 2 Our primary objective was to determine if rings visible on some scales w^ere true annuli. If they were, our second objective was to determine if fish that had scales with no visible rings or with a profusion of unclear rings could be aged on the basis of length; if the rings were not annuli, our second objective was to explore other methods of aging Gulf menhaden. COLLECTION OF DATA Samples of the catch are taken daily by field personnel stationed at four ports, each comprising two or three plants grouped in close proximity. After weighing and measuring a fish, samplers remove a cluster of scales from just above the lat- eral line below the dorsal fin and deposit them in a 0.1% phenol solution. Later, six scales from each fish are cleaned and mounted between two glass slides. Each fish and its scales are identified by port, collection, and scale number. From 1964 to 1971, two samples of 20 fish each were taken daily at each port. Since 1972, three samples of 10 fish each have been taken. Juveniles ranging from about 80 to 120 mm FL were marked in late summer and early autumn with numbered internal ferro magnetic tags simi- lar to, but smaller than, those used for adults (Pristas and Willis 1973). Tags were recovered on magnets in various parts of the processing plants (Parker 1973), although all tags passing through a plant were not retained. To estimate the numbers of field tags not re- tained on magnets, batches of 100 fish marked with test tags were periodically planted in the catches. The percentage subsequently recovered was an estimate of the efficiency of the magents to recover field tags. The number of annual tests at each plant varied from 2 to 15 (200-1,500 tags). Annual recovery rates varied from 2 to 60% , but over one-half were between 15 and 40%. Some test tags were not recovered until 1, 2, or even 3 yr after they had been placed in catches (Table 1). Delayed recoveries were caused by 1) tags lodging in various parts of a plant before later being dislodged, 2) tags remaining in fish scrap stored for long periods before being ground, 3) tags remaining in various scrap storage areas before being mixed with new scrap. The number varied by plant and year and amounted to about 1% or less of the number of test tags applied, although in 1972 it was 6% at plant 58 and 5% at plant 57. Table l. — Number of test tags applied and number recovered at Gulf menhaden processing plants, 1971-73. Year tags applied No. of tags No. recovered Plant no. 1971 1972 1973 1974 54 1971 1972 900 1,300 28 0 703 (') (') (') 55 1971 1972 1973 900 1,400 700 100 4 598 6 13 255 0 0 7 56 1971 1972 1973 900 1,300 600 448 0 617 1 2 199 1 0 2 57 1971 1972 1973 200 1,300 1,200 40 1 301 1 59 322 0 3 7 58 1971 1972 1973 300 1,500 1,200 43 5 425 0 93 396 0 0 3 62 1971 1972 1973 1,200 1,300 1,100 470 1 654 0 1 658 0 0 2 63 1971 1972 1973 1,000 1,000 1,200 203 12 379 0 8 217 0 0 13 64 1971 1972 1973 930 1,200 1,100 395 8 314 0 4 241 0 0 6 65 1971 800 399 (') {') {') 68 1971 1972 1973 1,100 900 400 436 2 311 0 12 214 0 0 0 69 1971 1972 1973 1.000 1.100 1,300 163 1 265 0 3 273 0 0 1 71 1971 1972 1973 1,100 1.100 1.200 205 0 317 1 0 186 0 0 0 'Plant did not operate after 1972. ^Plant did not operate after 1971 OBSERVATION OF SCALES Scales were viewed on a scale projector at 48 x magnification. If no rings were evident, or if no one ring could be considered as an annulus, no mea- surements were made. If rings were evident, the distances from the focus to each ring and to the scale edge of the projected image were measured. Each ring had to be discernable on three or more scales or it was not measured. Each fish was as- signed an age corresponding to the number of rings on the scales except when the only ring visi- ble was in the area of the scale usually occupied by the second ring. Then the fish was called age 2 rather than age 1. The decision to assign a ring the number one or two position was based on the dis- tance of the ring from the scale focus. VALIDITY OF RINGS AS YEAR MARKS To determine if observed rings were true annuli, we examined three different sets of data: length- 316 NICHOLSON and SCHAAF; AGING OF GULF MENHADEN frequency distributions of sampled fish, returns of tagged juvenile menhaden, and spacing of rings on the scales. Length-Frequency Distributions From the general shape and the number of modes of a length-frequency distribution curve, it is often possible to infer the number of age-groups represented. Length-frequency curves of Gulf menhaden sampled during 1964-73 fishing sea- sons had two distinct modes. Since distributions in all years were similar, we have shown only those for 1967-70 (Figure 1). A prominent mode, usually evident in May at around 135 to 150 mm, shifted progressively to the right during the season and by September varied from about 155 to 170 mm. A smaller mode at about 170 to 180 mm in May tended to shift farther right during the season and disappear by midsummer, so that the curve be- came unimodal and greatly skewed. This small mode apparent in May appeared to be a continua- tion of the mode that was prominant during the preceding September. From the general shape of the length-frequency curves there appears to be only two dominant age-groups in the fishery. The younger and more numerous tends to dominate the fishery as the u Z o >- z UJ o 1970 N = 118 N = 1018 N = 2523 N = 828 N = 456 N = 2152 N = 23 100 120 140 160 180 200 220 100 120 FORK LENGTH IN MM N = 36 160 180 200 220 FIGURE 1.— Length-frequency distributions of Gulf menhaden in percent, by month, 1967-70. 317 FISHERY BULLETIN: VOL. 76, NO. 2 season advances, so that by the end of the season the catch is composed almost entirely of this age- group. The considerable variation in the relative numbers of the two major age-groups from year to year is probably a reflection of the differences in the relative abundance of each year class. Small numbers of a third age-group of still younger fish may enter the fishery in some years in August or September. Juvenile Tagging An advantage of tagging juveniles is that the age of each fish is known when it is recaptured. Unfortunately, recovery of some tags a year or more after they had entered a plant caused some fish to appear older at the time of recapture than they actually were. Although the number was rel- atively small, it had an important bearing on in- ferences pertaining to longevity and the propor- tion of older fish in the catch. By the end of the 1974 fishing season 1,137 field tags had been recovered (Table 2). Of these 1,069 (94.0%) were recovered in the 2 yr following tag- ging, 62 (5.5%) in the third year, and 6 (0.5%) in the fourth year. Only tags applied in 1970 had an opportunity to be recovered in the fourth year, and only tags applied in 1970 and 1971 had an oppor- tunity to be recovered in the third year. The tendency for test tags to remain in plants for one or more years leads us to believe that all field tags recovered in the fourth year and most of those recovered in the third year were holdovers from previous years. All of these tags were recovered at plants that had the highest percentage of test tag holdovers. We conclude, therefore, that at least 97% of Gulf menhaden caught in the purse seine fishery are age 1 or age 2, and that very few, probably less than 1% — live to age 3. If ring marks are valid annuli, they also should indicate that the catch is composed primarily of age-1 and age-2 fish. Incidence and Spacing of Scale Rings The scale length-fish length relation for Gulf menhaden was linear. Correlation coefficients for 1971, 1972, and 1973, based on fish ranging from 95 to 225 mm FL, was 0.790, 0.765, and 0.768, respectively; for log transformed data the coef- ficients decreased to 0.729, 0.695, and 0.681. Sam- ple sizes were 4,674, 4,457, and 4,902, respec- tively. The regression equation for the 3 yr Table 2. — Numbers of field tags of juvenile Gulf menhaden recovered from 10,458 fish tagged in 1970, 15,511 in 1971, 15,262 in 1972, by plant. Year No. recovered Plant no. tagged 1971 1972 1973 1974 54 1970 2 30 (') (') 1971 5 (') (') 55 1970 16 3 3 1 1971 14 31 5 1972 76 35 56 1970 29 30 3 2 1971 10 33 7 1972 54 30 57 1970 5 5 4 0 1971 4 14 2 1972 16 3 58 1970 2 19 18 1 1971 2 24 6 1972 6 12 62 1970 0 3 1 1 1971 18 13 2 1972 12 17 63 1970 3 2 0 0 1971 49 2 1 1972 28 4 64 1970 19 0 0 0 1971 95 1 1 1972 54 1 65 1970 7 {') n (') 68 1970 3 7 3 0 1971 45 16 2 1972 36 17 69 1970 1 7 2 1 1971 5 7 1 1972 9 6 71 1970 4 2 0 0 1971 43 1 1 1972 21 1 ' Plant did not operate after 1972. 2 Plant did not operate after 1971. combined was: Scale length = 5.392 + 0.865 Fish length. Not all fish had scales with clearly discernable rings. The percentage with no rings varied from 45.0 to 55.8. Of the fish considered as age 2, <2% had a ring in the number one position only and 25.1% in 1971, 39.1% in 1972, and 30.6% in 1973 had a ring in the second position only. A relatively small number offish had scales with three clearly discernable rings. The length frequencies from May to September of fish with one or two rings clearly indicated two distinct ages. In Tables 3-5 length groups below the first containing fish with both one and two rings have been lumped into one group. For one- ring fish the mode is the group, other than the lumped one, containing the most fish. The only exception is for May 1971 when the mode for fish with one ring was 155 mm. It is clear that the modes and means of one- and two-ring fish in- creased and that the distributions shifted toward larger sizes as the season progressed. 318 NICHOLSON and SCHAAF: AGING OF GULF MENHADEN Table 3. — Length frequencies and mean lengths of Gulf Menhaden, by month and number of scale rings (0-3), 1971. Fork May June July August September length (mm) 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 • 165 50 150 560 747 587 503 663 330 260 48 165-169 1 14 31 99 3 36 80 1 74 61 55 17 170-174 4 14 6 45 74 16 32 95 1 49 69 1 29 22 175-179 14 6 9 67 42 56 41 69 27 28 76 2 19 18 4 180-184 11 22 84 18 111 53 43 100 44 61 6 27 19 10 185-189 12 26 78 2 134 56 14 108 1 36 32 32 15 5 15 190-194 4 29 53 83 3 42 144 5 44 15 48 2 22 6 17 195-199 3 10 1 30 49 7 29 68 2 28 1 52 3 10 2 9 1 200-204 3 4 2 18 14 3 14 35 17 16 35 5 11 4 1 205-209 1 1 6 1 2 11 7 4 6 4 1 4 2 210-214 1 2 1 4 8 5 3 5 2 1 3 215-219 h 1 2 1 2 1 2 220-224 1 1 Total number 104 184 106 4 969 982 473 20 892 804 495 42 992 645 185 21 452 137 64 9 Mean length — 148 187 202 — 154 186 202 — 160 189 203 — 165 194 204 — 169 191 209 Table 4. — Length frequencies and mean lengths of Gulf menhaden, by month and number of scale rings (0-3), 1972. Fork May June July August September length (mm) 0 1 2 3 0 1 2 : 3 0 1 2 3 0 1 2 3 0 1 2 3 <155 288 353 264 338 198 102 97 39 41 6 155-159 64 79 3 158 134 204 107 136 51 14 7 160-164 109 94 5 170 114 1 339 169 280 129 47 11 165-169 109 75 12 192 88 3 236 126 347 134 1 33 19 1 70- 1 74 79 47 52 120 99 24 178 78 9 180 83 3 43 10 175-179 53 5 71 86 48 42 136 64 39 118 54 7 57 14 180-184 44 1 72 45 7 59 97 41 52 102 33 24 44 7 1 185-189 20 55 1 36 72 70 6 102 78 8 74 28 190-194 16 27 2 17 66 2 37 1 89 2 79 3 105 25 195-199 7 13 7 15 31 3 42 79 3 65 77 17 200-204 8 4 4 11 12 4 19 40 3 30 46 3 26 205-209 2 2 4 3 4 16 20 4 19 10 4 10 210-214 1 9 2 9 13 10 7 9 5 3 7 215-219 3 4 5 1 3 7 1 3 1 220-224 1 1 2 1 3 2 1 8 1 225-229 1 1 1 1 1 1 Total number 799 654 314 18 1,131 828 315 27 1.593 694 443 26 1.550 534 354 22 394 Mean length — 151 181 198 — 157 186 208 — 164 191 209 — 166 193 215 — 76 17 170 194 Table 5. — Length frequencies and mean lengths of Gulf menhaden by month and scale rings (0-3), 1973. Fork May June July August September length (mm) 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 <155 162 55 265 109 181 41 171 10 49 2 155-159 80 28 1 217 105 187 103 26 22 1 160-164 94 17 13 151 117 229 190 77 71 2 1 165-169 69 16 27 93 80 2 210 169 127 164 11 12 170-174 44 16 45 96 79 14 204 207 2 138 295 16 26 175-179 40 4 44 90 39 70 150 138 24 237 100 3 34 26 3 180-184 21 42 68 11 169 95 89 73 143 71 32 31 28 7 185-189 6 30 28 5 196 82 43 114 75 35 80 19 16 12 190-194 7 26 1 16 1 119 34 19 117 50 12 88 15 3 28 195-199 4 12 15 71 3 15 1 81 14 1 64 17 2 16 200-204 1 12 1 10 46 6 10 49 4 14 42 6 8 1 16 2 205-209 1 4 2 2 28 2 7 26 3 8 28 5 5 1 210-214 3 4 4 8 13 8 5 12 5 5 8 6 9 215-219 3 2 1 5 1 6 3 2 2 8 2 2 220-224 1 5 1 2 3 4 1 4 3 2 225-229 2 1 2 1 2 1 Total numtier 532 136 260 11 1.062 546 729 31 1.412 1.000 506 20 1,092 781 348 26 223 117 87 7 Mean length — 158 182 210 — 162 188 211 — 169 192 213 — 172 194 214 — 178 193 211 319 FISHERY BULLETIN: VOL. 76, NO. 2 Table 6. — Mean distance from last ring to scale edge expressed as the percent of the distance from focus to scale edge for Gulf menhaden. 1971 1972 1973 Period Age 1 Age 2 Age 3 Age 1 Age 2 Age 3 Age 1 Age 2 Age 3 1-15 May 36.7 99 5,0 24.6 11,6 — 29,6 13.3 6.7 16-31 May 37.0 10.6 — 35.8 12,5 5,2 326 15.2 7.3 1-15 June 40.1 12,4 7.7 39.1 13.8 5,7 34 8 17.4 88 16-30 June 42,1 126 7.3 39.9 15.6 5,8 380 18.9 8.7 1-15 July 432 13.0 7.4 42.1 15.7 6,7 40.4 20.7 8.6 16-31 July 43.5 12,7 7.4 41.6 15,0 6,7 41.1 19.6 8.8 1-15 Aug. 45.6 14.2 8.6 42.9 15,8 6-9 42 1 20.6 10.4 16-31 Aug. 47,5 13,4 89 44.9 14 1 7,5 43,7 19.6 11.0 1-15 Sept. 44,6 10,4 — 42.3 14,4 — 36,8 25.6 — 16-30 Sept. 452 138 87 47,3 15,0 — 40,3 21 8 87 1-15 Oct. — 16,0 — — — — 423 20.0 84 Fish with three rings were not clearly differen- tiated as a distinct age-group. Because of small numbers neither modes nor general shapes of the distributions could be clearly determined. Means tended to increase slightly as the season pro- gressed. There was considerable overlap of the lower end of the length range with the upper end of the length range of two-ring fish. If the rings we observed were true annuli, the distance from the last ring to the scale edge should have decreased as the number of rings increased, and should have increased throughout the season for fish having the same number of rings. Both of these trends were apparent (Table 6). Mean lengths back-calculated to the age of an- nulus formation from fish with one, two, or three rings were similar, and the mean lengths at the time of first ring formation calculated from two- ring fish were slightly smaller than those calcu- lated from one-ring fish, as would be expected (Table 7). This tendency for mean lengths back- calculated from successively older ages of the same year class to become progressively smaller is commonly known as Lee's phenomenon and may be caused by a variety of factors. Mean lengths calculated from three-ring fish, however, were slightly larger, rather than smaller, than mean lengths calculated from either one- or two-ring fish. Frequency distributions of lengths back-calcu- lated to the time of the first, second, and third ring Table 7.— Mean lengths (millimeters) of Gulf menhaden at the time of each ring formation calculated from one-, two-, and three-ring fish, 1968-72 year classes. First ring Second ring 2-ring 3-ring Third ring Year 1-ring 2-ring 3-ring 3-ring class fish fish fish fish fish fish 1968 — — 88.4 — 157.4 186,4 1969 — 95.2 96.6 164,5 1661 195 1 1970 90.3 85.1 92.9 165-0 166,3 1953 1971 95.0 86.3 — 158 9 — 1972 100 3 — — — formation were well separated from each other, with only a small overlap between one- and two- ring and two- and three-ring fish (Figure 2). Those at the time of the first ring formation were similar in shape, whether calculated from one-ring, two- ring, or three-ring fish. Those calculated from two-ring fish were shifted slightly farther to the left than those calculated from one-ring fish, as Year Class 1972 1970 >- u z o 1970 1970 1969 1969 100 120 140 160 180 FORK LENGTH IN MM 200 220 Figure 2. — Length-frequency distributions at time of first, sec- ond, and third ring formation, back-calculated from one-ring, two- ring, and three- ring Gulf menhaden, 1968-72 year classes. 320 NICHOLSON and SCHAAF: AGING OF GULF MENHADEN would be expected. Those at the time of the second ring formation, whether calculated from fish with only a ring in the second position or from fish with both rings, were nearly identical and were shifted slightly to the right of distributions calculated from fish with three rings. CONCLUSIONS On scales of most Gulf menhaden with one or two rings, the rings appear to be true annuli. A relatively large number of fish that do not form a ring at the end of the first year form a ring at the end of the second year. A very small number that form a ring at the end of the first year do not form a ring at the end of the second year. It is possible, therefore, to separate age-2 from age-1 fish by the number of rings, or the location of the ring if only one is visible. For fish having scales with more than two rings, or with two rings that are oddly spaced, it is difficult to differentiate between true and false annuli, or to determine to what year a particular ring should be assigned. On scales of some fish that could be age 3 on the basis of length, only two rings are visible, in what appears to be either the first and second or second and third positions. On some scales that have three well-defined rings, the spac- ing appears too unusual to be true annuli. For those fish that are called age 3, the lengths overlap those of age-2 fish, the mean lengths and ranges progress very little during the season, and the mean increments from the last annulus to the scale edge show little increase. We concluded that it is impossible to separate age-3 from age-2 fish with a high degree of certainty on the basis of the number or the location of scale rings. From late August until October a small number offish ranging from about 115 to 135 mm appear. We believe most of these fish, which have no scale rings, are age 0, but we cannot be certain because many of the fish in this size range of age-1 fish also have no scale rings. The small number of tags recovered after 2 yr from fish tagged as juveniles, or age 0, the scarcity in the catch of fish larger than those with two rings, and the small numbers of fish with more than two rings, indicate that few Gulf menhaden live to be older than age 2. Since both age-0 and age-3 fish compose <2'Jc of the catch, and since each age-group is either impossible or difficult to identify, we believe it is practical to recognize only two age-groups of Gulf menhaden: those age 1 or younger and those age 2 or older. If only two ages are recognized, fish with no annuli can be aged by length. For each month, those below a certain fork length can be called age 1 or under, those above a certain length age 2 or older. Those in between cannot be individually aged, but the number in each length class can be apportioned to each age-group on the basis of the percentage offish in each length class with one or two annuli. For example, in June 197 1 (Table 3) all fish <165 mm may be called age 1, all >185 mm age2. Ofthe31unagedfishl65-169mm,30(97%) are age 1 and 1 (3%) is age 2; of the 45 between 170 and 174 mm, 37 (82%) are age 1 and 8 (18%) are age 2. A question that may arise concerns the accuracy of previous aging methods. To shed some light on this question, we compared the percentages offish at each age for methods 1 and 2. In method 1, fish ages had been based on a combination of factors: the general appearance of the scales, the number and location of rings, the fish length, and the time of year the fish was caught. In method 2, fish that had been aged by the number and location of scale rings, and fish that could not be aged, were grouped in 5-mm size classes. For each size class the number of unaged fish were apportioned to each age-group by the same percentage as fish that had been aged. We retained the age-3 group for comparative purposes, but could not differentiate between age-0 and age-1 fish. The percentages at each age were remarkably similar, the differences between methods varying from only 0.1 to 2.7%. We concluded, therefore, that age compositions based on the previous method of aging are reliable and that valid infer- ences pertaining to population dynamics of Gulf menhaden can be based on them. LITERATURE CITED FORE, P. L., AND K. N. BAXTER. 1972. Diel fluctuations in the catch of larval Gulf men- haden, Brevoortia patronus, at Galveston entrance, Tex- as. Trans. Am. Fish. Soc. 101:729-732. KROGER, R. L., AND P. J. PRIST AS. 1975. Movements of tagged juvenile menhaden (Brevoor- tia patronus I in the Gulf of Mexico. Tex. J. Sci., 26:473- 477. Parker, R. O., Jr. 1973. Menhaden tagging and recovery: Part II — Recovery of internal ferromagnetic tags used to mark menhaden, genus Brevoortia. Mar. Fish. Rev. 35(5-6):36-39. 321 FISHERY BULLETIN: VOL. 76, NO. 2 PRISTAS, p. J., AND T. D. WILLIS. SUTTKUS, R. D. 1973. Menhaden tagging and recovery: Part I — Field 1956. Early life history of the largescale menhaden, Bre- methods for tagging menhaden, genus Brevoortia. Mar. voortia patronus, in Louisiana. Trans. 21st North Am. Fish. Rev. 35(5-6):31-35. Wildl. Conf., p. 390-407. STRUHSAKER, p., and J. H. UCHIYAMA. TURNER, W. R. 1976. Ageandgrowthofthenehu, Sto/ep/iorus purpureas 1969. Life history of menhaden in the eastern Gulf of (Ksces: Engraulidae), from the Hawaiian Islands as indi- Mexico. Trans. Am. Fish. Soc. 98:216-224. cated by daily growth increments of sagittae. Fish. Bull., U.S. 74:9-17. 322 VARIABILITY IN ZOOPLANKTON BIOMASS DISTRIBUTION IN THE NORTHERN SARGASSO SEA: THE CONTRIBUTION OF GULF STREAM COLD CORE RINGS ^ Peter B. Ortner, Peter H. Wiebe, Loren Haury, and Steven Boyd^ t ABSTRACT The scale and frequency of physical variability resulting from incursion of Gulf Stream cold core rings into the northern Sargasso Sea makes this fauna! province more heterogeneous than previously recognized. At any one time such rings may cover between 6 and IS'^ of the surface area of the northern Sargasso Sea. They are more productive than the surrounding Sargasso Sea and have a zooplankton biomass intermediate between the Sargasso Sea and the slope water. Cold core rings may augment by 3 to 7% the primary productivity and by 8 to 169c the zooplankton standing crop of the northern Sargasso Sea. Compared with either the surrounding Sargasso Sea or their parent slope water, an unusually large percentage of the 0-800 m biomass in rings is found at depths greater than 200 m. This distribution may be related to hydrographic and biological changes associated with ring decay. Because of their higher productivity, differences in vertical biomass structure, and the possibility that ring food chain efficiency is lower than that of the Sargasso Sea, rings may provide a disproportionately large fraction of the total supply of organic matter to the northern Sargasso deep Sea. A number of papers have characterized the zoo- plankton biomass of the northern Sargasso Sea (Menzel and Ryther 1961; Grice and Hart 1962; Be etal. 1971; Deevey 1971; Deevey and Brooks 1971; and others). Because of the variety of methods employed in both sampling and processing, the results of these studies are not readily compara- ble. In general, previous authors have portrayed the Sargasso Sea as a remarkably homogeneous faunal province. The scale and frequency of re- gional variability resulting from incursions of cold core rings into the Sargasso Sea have not been generally appreciated. Cold core rings are meso- scale hydrological features 150 to 300 km in diameter and up to several thousand meters in depth. They form when southerly directed Gulf Stream meanders become so accentuated as to separate from the Stream and move south, enclos- ing a core of cold and relatively fresh slope water within a remnant of the Gulf Stream (Parker 1971; Fuglister 1972; Richardson 1976). It is likely that in the northern Sargasso Sea, at any one time, there are 10 to 15 such rings (Lai and Richardson 'Contribution No. 3939 from the Woods Hole Oceanographic Institution. This paper is part of a thesis submitted by P. B. Ortner for a Ph.D. degree from the Woods Hole Oceanographic Institution. ^Woods Hole Oceanographic Institution, Woods Hole, MA 02543. Manuscript accepted September 1977. FISHERY BULLETIN; VOL. 76, NO. 2, 1978. 1977). Estimating the surface area of the northern Sargasso Sea as 32.9 x 10^ km2 (Jahn 1976), cold core rings may cover between 6 and 13*^ of this surface. (Throughout the ensuing sections, unless otherwise indicated, the terms ring, slope water, and Sargasso Sea denote hydrographic, not geo- graphic, entities.) An overview of the phytoplankton, zooplankton, and midwater fish populations inhabiting cold core rings has been given by Wiebe, Hulburt, Car- penter, Jahn, Knapp, Boyd, Ortner, and Cox (1976). The results of that study indicated that mean zooplankton biomass in the upper 750-800 m of rings between 3 and 10 to 12 mo of age was consistently higher than that in the surrounding Sargasso Sea. In these preliminary data the frac- tion of biomass below 250-300 m in depth was particularly large while the near surface was more similar to the Sargasso Sea. We have now taken vertically stratified hauls in the same ring 3 mo apart. The data from these hauls confirm our ini- tial interpretation. The objective of this paper is twofold. First, we describe the zooplankton biomass distributions characteristic of the northern Sargasso Sea, of a cold core ring, and to a lesser extent of the slope water — the source of ring water. Second, we will attempt to relate the patterns observed to sys- tematic variations in phytoplankton standing 323 FISHERY BULLETIN: VOL. 76, NO. 2 crop, primary productivity, and w^ater tempera- ture, and to explore the significance of ring biomass distribution. METHODS The major portion of the data to be presented in this paper w^as collected on RV Chain cruise 125 (August 1975) and on RV Knorr cruise 53 (November 1975). The ring sampled (designated Ring-D by the Naval Oceanographic Office), was formed in February 1975. It was, therefore, about 6 mo old when first sampled and 9 mo old when sampled again in November. In November the slope water was hydrographically complex. It is likely that some of our intended slope water tows (MOC 39 and MOC 40) may have been taken in the vicinity of a warm core ring (Saunders 1971). The upper 200 m of the water column at that station was warmer and more saline than is typical for the slope water. In addition, infrared satellite photo- graphs clearly show the presence of this warm ring during the period of sampling. Other slope water stations may have been influenced by the passage of a warm core ring. In analyzing the data, MOC 39 and 40 are considered separately and desig- nated warm core ring tows. Data corroborating specific points or conclusions have been drawn from collections made on KV Atlantis cruise 71, RV Chain cruise 1 11, and RVif/iorr cruises 35 and 38 (Table 1). Collections in Gulf Stream cold core rings, the northern Sargasso Sea, and slope water were made with three types of sampling gear: on the early cruises 1-m diameter ring nets or modified opening/closing 70-cm diameter bongo nets (McGowan and Brown 1966), on the two most re- cent cruises a multiple opening/closing net and environmental sensing system — MOCNESS (Wiebe, Burt, Boyd, and Morton 1976)— with a mouth area of 1 m x 1.4 m (effective area is 1 m^). All nets were constructed from 0.333-mm Nitex^ gauze; depth recorders and flow meters were used on all tows. The 1-m nets were hauled obliquely, ideally to a depth of 800 m. On some cruises a second haul was taken to a depth of 300 m. Bongo nets were towed obliquely within the depth intervals 0-250, 250- 500, and 500-750 m. With occasional exceptions, the MOCNESS sampled both from 800 m to the surface in 100-m intervals, and from 200 m to the surface in 25-m intervals. Sampling with 1-m and bongo nets was almost always done at night, while at most MOCNESS stations samples were taken both day and night. The types of tows taken on the five cruises are given in Table 1. All samples were preserved in 5-10% Formalin buffered to pH 8.0 with sodium tetraborate. In the vicinity of all plankton hauls, hydrographic casts were made yielding nearly concomitant vertical profiles of temperature, salinity, oxygen, chlorophyll, nu- trients, primary productivity, and phytoplankton species (see Wiebe, Hulburt, Carpenter, Jahn, Knapp, Boyd, Ortner, and Cox 1976 for methods). Zooplankton biomass was measured by the method of Ahlstrom and Thrailkill (1963) after removal of all organisms greater than 5 cm^. Dis- placement volumes were measured 5 to 9 wk after a cruise. No attempt has been made in this paper to partition the biomass according to taxa. The species composition of those samples already examined appears similar to that reported for the region by Grice and Hart (1962), Deevey (1971), and Deevey and Brooks (1971). ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table l. — Summary of slope water, ring, and Sargasso Sea zooplankton sample stations. Age of ring (mo) Number of samples (stations) Cruise Date Ring Sargasso Sea Slope water Type of net Atlantis inv 21 Sept.-14 Oct. 1972 10-12 8(4) 28(15) 4(2) 1 m Chain 111' 7 Feb.- 18 1973 3.5 6(2) 5(2) 2(2) 1 m and bongos Knorr 35' 23 N0V.-3 Dec. 1973 3.0 8(4) 1(1) 5(3) 1 m and bongos Knorr 38' 12 Feb. -3 Apr 1974 10-12 1 m Ctiain 1252 4 Aug. -17 1975 6.0 48(1) 32(2) 48(2) MOCNESS Knorr 53^ 17 N0V.-I Dec. 1975 9.0 48(1) 40(2) 64(3) MOCNESS 'Positions of stations Illustrated in Wiebe, Hulburt, Carpenter, Jatin, Knapp, Boyd, Ortner, and Cox (1976) ^Sargasso Sea; MOC 1,2 (35°37', 68°3r), MOC 3,4 (35°22', 68°17'), MOC 12 (34°ir, 7r40'), MOC 13,14,15 (34°10', 7r34'); ring: MOC 5,11 (34°29' 69°56'), MOC 6,7 (34°34', 69°52'), MOC 8 (34°31', 69°49'), MOC 10 (34°33'. 69°53'); slope water: MOC 16,17 (38°02', 69°59'), MOC 18,19 (38°05'.70°02'), MOC 20,21 (39°05', 70°12').AII positions are north latitude and west lonaitude, ^Sargasso Sea: MOC 23,24,25 (-32"44', 71°10'), MOC 26 (32°52', 7r08'), MOC 34 (34"12', 70°30'); ring: MOC 27, 28, 29 (33°49', 7r54'), MOC 31 (35°50', 7r48'), MOC 32 (33°56', 71 °54'), MOC 33 (34°03', 71 °56'): warm core nng: MOC 39,40 (40°04', 68°05'): slope water: MOC 35,36 (38°51 ', 67°47'), MOC 37,38 (38°55', 67°46'), MOC 41,42 (39°59', 69°00). All positions are nortti latitude and west longitude. 324 ORTNER ET AL.: SARGASSO SEA ZOOPLANKTON BIOMASS DISTRIBUTION RESULTS Regional Biomass The biomass collections obtained in the same ring 3 mo apart in August and November 1975 corroborate differences already noted in 0-800 m zooplankton biomass between the Sargasso Sea, cold core rings, and slope water (Figures 1, 2). In both months the slope water 0-800 m biomass was larger than either the Sargasso Sea or ring 0-800 m biomass {Mann- Whitney U-iest, P<0.01). In August, the contrast between slope water and the two other regions was particularly marked with average concentration in the upper 800 m approx- imately 10-12 times larger than in the Sargasso Sea and ring (Table 2, top;. At several depths in the August slope water stations the zooplankton biomass was dominated by Salpa aspera.^ Differ- ences rn abundance of this salp accounted for a large part of the variation between slope water stations. The high water content of these animals undoubtedly caused our estimate ofTnomass by displacement volume to be considerably higher than had we measured dry weight or organic car- bon content. It is clear, nonetheless, that the standing stock of zooplankton was exceedingly large. Relative to the Sargasso Sea, 0-800 m ring biomass was on the average 1.36 times^larger in August and 1.33 times larger in November (Table ■•The significant biomass contribution of S. aspera is discussed in Wiebe, P. H., L. P. Madin, G. R. Harbison, L. R. Haury, and L. M. Philbin. 1977. Diel vertical migration by Salpa aspera and its potential for large-scale particulate organic matter transport to the deep-sea. Submitted to Mar. Biol. Table 2. — Comparison of slope water, ring, and Sargasso Sea zooplankton biomass (cm^/m^) based on weighted averages of day and night samples collected at 0-800, 200-800, and 0-200 m depth intervals. Number of tows used to make average given in paren- thesis. August 1975 November 1975 Region 0-800 200-800 0-200 0-800 200-800 0-200 Sargasso Sea Ring fringe Cold core ring Slope water Warm core ring 16.6(4) 8.8(4) 8.1(8) 22.5(3) 256.2(4) 15.6(3) 6.4(6) 95.6(4) 121.8(6) 10.2(2) 91(1) 13.6(4) 33.8(4) 5.6(2) 4.3(1) 11.4(4) 23.0(4) 6.0(4) 48(1) 3.4(5) 12.8(6) 27.9(2) 17.2(2) 10.8(2) Month Ratio ring/Sargasso Sea 0-800 200-800 0-200 August November 1 36 1.33 1.77 2.03 0.79 0.56 2, bottom). These differences are consistent with data from previous cruises; using paired regional biomass averages observed on all cruises to date (Table 3) we see that the mean zooplankton biomass in rings and in the Sargasso Sea is sig- nificantly different (Sign test, P<0.05). In November zooplankton standing crop was consistently lower than in August with the most pronounced change occurring in the slope water. Although S. aspera was still present at the slope water stations, it no longer dominated the bio- mass. Comparing August and November 0-800 m biomass averages, segregating day and night samples for each region, we see that the overall seasonal decline is statistically significant (Sign test, P<0.05, computed using cubic centimeters per square meter in Figures 1 and 2). Indeed only one November 0-800 m biomass value was as large as the smallest 0-800 m biomass in the same re- gion in August. Table 3. — Average zooplankton biomass in slope water, ring, and Sargasso Sea — dry weight (mg/m*) in colurmi =750 m deep. Numbers of stations per area and range of biomass values (after colon) given in parenthesis. This table is an expanded version of table 3 in Wiebe, Hulburt, Carpenter, Jahn, Knapp, Boyd, Ortner, and Cox ( 1976). Note that the biomass units in the original table were incorrectly presented as milligrams per square meter. Ring age Ring ^ Slope -r Cruise Slope water Ring Sargasso Sea (mo) Sargasso Sargasso Atlantis II 71 7.43 (1) 2.24 (4:2.06-2.60) 1.68 (4:1,34-2.05) 10-12 1 33 4.42 Chain 111 4.21 (1) 4.95 ('2:2.21-7.68) 2.70 (23:1.99-3.96) 3.5 1.83 1.56 Knorr 35 6.06 (M:3.60-7.74) 3.67 ("3:3.15-4.33) 2.47 (1) 3.0 1.49 2.45 Knorr 38^ 5.74 (4:2.2-9.14) 3.10 (6:256-4.98) 2.34 (4:1.76-3.21) 10-12 1.32 2.45 Chain 125^ 9.10 (^2:6.85-1 1.35) 226 (3:1.73-2.57) 1.49 (4:0.90-1.97) 6 1.52 6.11 Knorr 53^ 3.43 (4:2.77-4.03) 1.09 (4:0.72-1.35) 0.83 (2:0.80-0.86) 9 1.31 4.13 Mean 7.3 1.47 3.52 'Each of the two values is Integrated for the water column based on three stratified bongo net tows. ^Two values are from oblique meter net tows: one is an Integrated value based on three stratified bongo net tows. ^Two values are from oblique meter net tows; one is an integrated value based on three bongo net tows; one is an oblique bongo net tow. "One value is an integrated value based on three stratified bongo net tows; two values based on two stratified bongo net tows. ^Cubic centimeters per cubic meter converted to milligramsper cubic meter using equation 4 table 2 in Wiebe etal. (1975). Note that this conversion affects the regional biomas ratios (see Figure 2) because the relationship between displacement volume and dry weight is not linear. ^Salp-rich tows. MOC 39 and MOC 40, excluded. 325 FISHERY BULLETIN: VOL. 76, NO. 2 T 1 — I — I 1 1 r o o o o o o o o f^ ^ a> 03 < UJ CD O CO CO < cr < o o. o I o O O- o o o c o o o o C3 cr UJ cr o o O o < UJ (r t- co 3 c:) o o o o o o o o si I o Q. LU Q ^" I < |_ CO < I- <1 o.-^ u aJ — 4) > T3 C ca a. X G CB a; > CO j= -s Q 2 ^ > 0) (C o o U5 2 < Ji . a. O 6 n O il S (U 2 -^ :s - C O o c J«! .2 « o S £ CO e s CO i; bO £ CO L< s s * a; M la *" 3 •E o- _ to -2 i en .5 I c ^- S w oi O fS^JlJW Mid JO 326 ORTNER ET AL.: SARGASSO SEA ZOOPLANKTON BIOMASS DISTRIBUTION DC LJ UJ CL O _J CO o cr Li_ o CO o CO CO < cc < CO e^ o z: o- o cr m o ^ III 1 m- cr o * ! o 2 ^ ^H CJ> o- " 2 cr o ■ T o i * 1^* H < o 2 c ^J ^ ) o o o o o o o o CJ t to a> cr LlI I- I LU a. o _i CO ^-*^^^^^^^ «\i C ) 0 T 7 0 t 1 0 0 If) 0 m 0 •■ CVJ iD -| 70 0- H Qi 0 . 5 L <^* o- ^^^S ^^^ ^^ s 00 1 iiif'--' C5 ^ Q . «r 0 i"^ 5 ^ 3 0 0 0 0 0 0 0 0 0 0 !SS=SSS&SSSS <_) o o •5 T) o o — T— o o — I 1 1 — 8 8 o o o o o -r- O o o o 3§ V 00 si !; ^ a c JC « U V CO > OJ o • ^ Co 4) C 2 $ > 0) w aj in j3 o ? It -I > « o -^ ?^ s 2 CO f-i o § .2 8 S C I) J= 5 -^ U C O.'O 3. ca g get - e a> « § > c CS X be cs CO ^ Cfi 0) o « a §■£ tn ■ — ■ QJ O M D. ca a. o (ShJ313l^) HldJO 9i E gz 327 FISHERY BULLETIN: VOL. 76, NO. 2 Average Vertical Structure Comparing Ring-D biomass partitioned accord- ing to depth, the upper 200 m in the ring con- tained, on the average, less biomass during both sampling periods than did the Sargasso Sea (Table 2, top). This was true both day and night during the August and November cruises. In contrast, ring biomass between 200 and 800 m was higher both day and night (Figures 1, 2). The range of 200-800 m biomass values in the ring and in the Sargasso Sea does not even overlap. The combina- tion of lower average surface biomass and higher average subsurface biomass in the ring is highly significant (Sign test, P<0.01, computed using sums of 0-200 m and 200-800 m cm3/l,000 m^ de- rived from Figures 1 and 2). The regional weighted averages of percent 0-800 m biomass present in the upper 200 m in August were Sl'/f , 347c , and 2T7c in the Sargasso Sea, slope water, and ring, respectively. In November these averages were 457c, 327c, and 257^ (Table 4). Although very dif- ferent sampling systems and tow strategies were employed, data from Atlantis II cruise 71 corrobo- rate the direction of difference of these observa- tions in that the percentages of 0-800 m biomass found at night in the upper 300 m were 64% and 52% for the Sargasso Sea and ring, respectively (Table 5). In addition, the 300-800 m biomass was 1.73 times larger in this latter ring than in the surrounding Sargasso Sea. Diel Migration Complicating these general observations and contributing to sample variability are day/night differences in biomass distributions (Table 4). In Table 5.- Area -Ring and Sargasso Sea zooplankton biomass- Atlantis II cruise 71 (mg/m^). 0-300 m 0-800 m 0-300 0-800 100 ■Ring Sargasso Sea 954 1,648 963 2,080 930 1,704 828 1,728 858 1,344 798 1,072 765 1,640 921 1,304 52% 64% all day/night sample pairs the fraction of 0-800 m biomass present in the 0-200 m interval is larger in the night sample (Sign test, P <0.01). This re- sults from either diel migration or day/night dif- ferences in avoidance within the comparatively well-illuminated surface layers. Avoidance does not appear to be an important factor because at some stations the day 0-800 m biomass exceeds the night 0-800 m biomass. This is true in all Sargasso Sea 0-800 m sample pairs and at one slope water station (Figures 1, 2). Furthermore, some species of zooplankton taxa already enumerated, e.g., euphausiids and pteropods, exhibit strong diel migration patterns in all three areas. Since we believe diel migration to be the ap- propriate explanation, the data further suggest that while essentially the same percentage of 0-800 m biomass was migrating into the surface layers of the Sargasso Sea (24-30% during both sampling periods), there was a reduced percentage migrating in the ring in November ( 2 1% in August versus 9% in November — Table 4). Although a smaller proportion of the biomass may have been migrating in the ring relative to the Sargasso Sea, there was a significantly greater (Mann- Whitney t/-test, P<0.05) day/night biomass ratio in the Table 4. — Percent of 0-800 m slope water, ring, and Sargasso Sea zooplankton biomass in the upper 200 m (800 m tows only). D = Day; N = Night. August : 1975 November 1975 Percentages of Percentages of Region individual tows D N N-D (D + N)/2 individual tows D N N-D (D + N)/2 Sargasso Sea Di = 32 N, = 57 D2 = 41 N2 = 69 39 63 24 51 z p II II 30 60 30 45 Ring fringe D, = 46 46 Cold core ring D, = 16 N, = 32 N2 = 42 16 37 21 27 N, = 29 Di = 21 D2 = 19 20 29 9 25 Slope water 'D, = 3 'N, = 93 D2 = 7 N2 = 61 7 61 54 34 Di = 32 Ni = 55 D2 = 13 N2 = 27 23 41 18 32 Warm core ring Di = 30 N, = 46 30 46 16 38 'On this tow series, MOC 18 and 19, salps were extremely dominant. These tows are excluded from averages. 328 ORTNER ET AL.: SARGASSO SEA ZOOPLANKTON BIOMASS DISTRIBUTION upper 200 m in the ring (Table 6). This apparent contradiction results from the fact, already noted, that the percentage of 0-800 m biomass present in the upper 0-200 m was very much greater in the Sargasso Sea. Day/night ratio of biomass in the upper water column is often used to measure in- tensity of diel migration; clearly the meaning of this ratio is highly dependent upon average verti- cal biomass distribution. Slope water day/night sample pairs may be in- terpreted as documenting diel migration, but the data are extremely variable both within and be- tween cruises (Table 4). There may have been a less well-developed migration in the fall, but the generality of this is questionable. Table 6. — Day night differences in slope water, ring, and Sar- gasso Sea zooplankton biomass in the upper 200 m. Ratio night day Region Augi ist 1975 November 1975 Sargasso Sea '1.78 '1.84 21.37 20.99 '1.37 21.86 Cold core ring '2.01 23 10 '2 48 21 86 Slope water '406 2.368 33 2.31482 '2 10 2227 21 70 'Based on 0-200 m tows 2Based on 0-800 m tows 3Ratio affected by extreme salp dominance. Shallow Biomass Structure In the 0-200 m biomass profiles, an intermediate biomass peak occurred between 50 and 100 m depth at nearly every station in August 1975 (Fig- ure 1: MOC 1, 3, 6, 7, 10, 16, 17). At all butone of the Sargasso Sea and ring stations this inter- mediate peak is the highest observed value in the 0-200 m tows. At slope water stations of the same cruise this intermediate peak is the second highest observed value. If we rank each interval in a profile in order of zooplankton abundance, we can test the significance of this observation. For in- stance, the individual summer tows in the ring and the Sargasso Sea exhibit significant concor- dance as to which depth intervals have the larger and which the smaller zooplankton biomass (Friedman 2-way analysis of variance on ranks, P<0.005). Given this result, the best estimate of the differences between intervals is the order of their summed ranks (i.e., 50-75 m>75-100 m>100-125 m>25-50 m>0-25 m>150-175 m>125-150 m>175-200 m). Applying a procedure for testing differences between individual depth intervals (Nemenyi 1963), we see that concor- dance results from the fact that the 50-75 m biomass is significantly greater than the biomass in the intervals 125-150, 150-175, and 175-200 m, and the 75-100 m biomass is greater than the 175-200 m biomass (P<0.05). An intermediate peak is not a notable feature of any of the 0-200 m profiles taken on the fall cruise with the exception of the Sargasso Sea sample pair (Figure 2: MOC 23, 26). DISCUSSION Wiebe, Hulburt, Carpenter, Jahn, Knapp, Boyd, Ortner, and Cox ( 1976) have discussed the forma- tion and decay of an idealized cold core ring. Ini- tially conditions inside a ring core are identical to those in the slope water just northward of the Gulf Stream at the time of ring formation. Through time the ring decays; the isotherms deepen, the water becomes more saline, the O2 minimum deepens, and the constituent flora and fauna either die off or become diluted by populations from the surrounding Sargasso Sea. Because zoo- plankton populations are generally suited to the environmental conditions they encounter within their normal range, this decay process may be viewed as the gradual imposition of a complex environmental stress upon an entire community. Wiebe, Hulbert, Carpenter, Jahn, Knapp, Boyd, Ortner, and Cox ( 1976) have documented some of the intermediate stages in this idealized process. In fact, this process can be aborted when a ring is reabsorbed by the Gulf Stream (Fuglister 1972; Richardson et al. 1977). All biological and physical properties are not equally conservative so their decay rates would not be the same. Regional Contribution of Cold Core Rings PRIMARY PRODUCTIVITY.— It is well known that slope water is more productive than the Sargasso Sea. Ryther (1963) estimated that slope water is about twice as productive on an annual basis ( 120 g C/m^ per yr versus 60 g C/m^ per yr). Although our own data are scanty, rings on the average are intermediate between slope water and the Sargasso Sea (Table 7). A few simplifying assumptions permit budgetary com- putations to be made regarding the overall effect of rings on the carbon budget of the northern Sar- gasso Sea. Let us assume an average ring life of 1 yr and a linear rate of decay of productivity (i.e., that annual ring production is the arithmetic mean of annual Sargasso Sea and slope water pro- duction). Allowing 6 to 13% as the areal contribu- 329 FISHERY BULLETIN: VOL. 76, NO 2 Table 7. — Summary of slope water, ring, and Sargasso Sea primary productivity (mg C/m^ per day), phytoplankton carbon' (mg/m^), and chlorophyll a (mg/m^) measurements. March 1974 August 1975 Phytoplankton carbon November 1975 Phytoplankton Region Productivity Chlorophyll Productivity Chlorophyll Productivity Chlorophyll carbon Sargasso Sea 2285 46.4 ^207 100 133 334 86.5 2522 12,0 21 Cold core ring 440 1 333.1 73.0 83 106 17.3 45 4832 155-5 1865 103 282 Slope water 1.025.5 70.4 175 50.5 1,302.2 368 4 270 280 824.0 376,2 363.7 39 287 'Based on counts of cells larger than 4-5 txm ^The high value in average mg C m^ per day observed at this station is a consequence of one unusually high surface value. tion of rings to the northern Sargasso Sea as explained earlier, and Ryther's estimate of a twofold difference in annual production, the net annual production of the geographic northern Sargasso Sea is then 3 to 7% higher than if it contained no rings (i.e., 6 x 1.5 = 9, 9 + 94 = 103 and 13 X 1.5 = 20, 20 + 87 = 107). Our assumption of linear decay is most certainly an oversimplifica- tion. In November 1975, the ring water column, like the slope water, began its winter overturn before the surrounding Sargasso Sea. Mixing eroded the seasonal thermocline that had been observed in Ring-D in August 1975. The decay we have assumed was reversed, and ring productivity was enhanced (Table 7). ZOOPLANKTON STANDING CROP.— Simi- lar calculations can be made regarding the rela- tive contribution of rings to the mean zooplankton biomass of the geographic northern Sargasso Sea. Neglecting one station which had anomalously high values due to extreme salp dominance, the average of slope water biomass values is 3.5 times the observed Sargasso Sea biomass (Table 3). Given this ratio and the same linear [i.e., (3.5 + 1) -i- 2 = 2.25] and areal assumptions made earlier, rings may augment the zooplankton standing crop of the geographic northern Sargasso Sea by 8 to 16% (i.e., 6 X 2.25 = 14, 14 + 94 = 108 and 13 x 2.25 = 29, 29 + 87-1 16). Our ratio of slope water to Sargasso Sea biomass may be compared with that of Grice and Hart (1962), who reported the slope water standing crop as three to four times that of the Sargasso Sea. They also excluded ex- tremely salp-rich samples in making this com- parison. Our assumption of 2.25 as an annual mean ring/Sargasso biomass ratio (i.e., linear de- cay) may be an overestimate considering the aver- age biomass ratio obtained on all cruises to date and the average ring age sampled (Table 3). On the other hand some rings do last longer than a year and the lowest ring:Sargasso Sea ratios that we have observed are approximately 1.3 (i.e., >1.0). We have noted a highly significant decline in 0-800 m biomass from August to November in slope water, ring, and Sargasso Sea both in data presented here and in data more recently collected.'^ This observation is consistent with those of Grice and Hart ( 1962) with respect to the slope water. They noted, however, no such decline in the Sargasso Sea. Neither is there a summer- to-fall decline in the Sargasso Sea data of Deevey (1971). The Sargasso Sea and slope water data of Fish (1954) exhibit irregular fluctuations in biomass throughout the summer and fall. Moore ( 1949) presented some Sargasso Sea data indicat- ing a progressive decline of biomass from a spring maximum to a fall minimum. Their data substan- tiate that interseasonal fluctuations in the Sar- gasso Sea are less marked than in the slope water. Vertical Structure We have pointed out that, compared with either the Sargasso Sea or slope water, an unusually small percentage of 0-800 m biomass is present in the upper 200 m of a ring. We found a relatively large fraction of the 0-800 m zooplankton biomass above 200 m in the northern Sargasso Sea. The netting employed by Leavitt ( 1935, 1938) was rel- atively coarse (1.0 mm) so it is difficult to compare our results with his. Nonetheless, at his two Sar- gasso Sea stations (2462, 2463) the percentages of ^Some of this data is presented in figure 4 of Richardson, P, L,, J, Schmitz, and P. H, Wiebe, 1977, Gulf Stream ring experi- ment, Polymode News 25:3. Unpubl. manuscr. 330 ORTNER ET AL.: SARGASSO SEA ZOOPLANKTON BIOMASS DISTRIBUTION 0-800 m biomass present in the upper 200 m were 42 and 49Vf which corresponds closely to our val- ues (Table 3). Both of our results are virtually identical with those obtained by Menzel and Ryther ( 1961). From their table 1 we can calculate the percentages of 0-500 m biomass and 0-1,000 m biomass present above 200 m. Averaging the re- sults, we find 449c of the 750-m biomass was pres- ent during the day above 200 m. lashnov (1961) presented data for the Sargasso Sea in which 90% of the 0-1,000 m plankton was present above 200 m, but he used a relatively fine mesh net (0.180 mm). Unfortunately, Deevey and Brooks (1971) characterized 500-m depth intervals with horizon- tal tows at the midpoint of each interval to 2,000 m, while Grice and Hart (1962) sampled only the upper 100-200 m. Several authors suggest that a vertical biomass structure similar to our slope water and Sargasso Sea observations is to be expected in temperate or subtropical oceanic environments relatively free of advective inputs. Vinogradov (1968: figure 47 and stations 3206 and 3829 in table 18) gave examples of oceanic regions with such a distribu- tion. Zenkevich and Birstein (1956) agreed that zooplankton biomass in the North Pacific rather steadily decreases from the surface downwards, although the most marked reduction they discuss might be below our lowest standard sampling depth. The one very deep tow series we obtained in a ring, however, gave no indication of such a re- duction (Figure 2, MOC 31). Zooplankton biomass profiles obtained by Murano et al. (1976) in the northwest Pacific above the Sagami Trough exhibit the expected decrease with depth. Reanalyzed in our manner, the data of Marlowe and Miller ( 1975) for Station P in the North Pacific support the above generali- zation; the percentage of their 0-500 m biomass found at night in the upper 200 m was 579^ . If one extrapolates their 500-m values as approximately applicable to the 500-800 m interval — a conserva- tive approach for this argument — the resulting percentage becomes 49*7^ (N). This is not unlike our average slope water percentage of 5 1% ( N ) and quite distinct from the average ring percentage of 33% (N) (Table 4). Station P is very different from Ring-D in respect to its vertical biomass distribu- tion. In slope water, the intermediate biomass peak in the upper 200 m approximately coincides with the depth of a nitrite maximum of the type discuss- ed by Vaccaro and Ryther ( 1960). Our results and those of Marlowe and Miller (1975) appear to dif- fer: they felt that the shallow nitrite peak of Sta- tion P was avoided by zooplankton. Since the levels of nitrite we have observed at the maximum are only slightly lower than those reported by Marlowe and Miller (0.2-0.5 fxg A-N-NOg/l versus 0.64 /Ltg A-N-NO2/I), our findings cast doubt on their speculation that nitrite toxicity might have been involved in the maintenance of the biomass minima they observed. Explanations for Ring Biomass Structure Given the relatively high zooplankton biomass of the slope water, it is clear why cold core rings have a higher average zooplankton biomass than the Sargasso Sea. Further, their higher average primary productivity appears responsible for this differential persisting 10-12 mo after ring forma- tion. Our data suggest the decline in ring biomass takes place rather slowly; the oldest rings sampled (10-12 mo) had ring/Sargasso biomass ratios only 20% smaller than the same ratios in the newest rings sampled (3.0 and 3.5 mo, Table 3). Although physically and chemically intermediate between slope water and Sargasso Sea, rings appear to be unique in their vertical distribution of biomass. We offer two logically distinct explanations for the small fraction of the 0-800 m biomass found within the upper 200 m of a ring. They are not mutually exclusive and the relative importance of these explanations is species dependent. The sim- pler argument stresses the importance of a physi- cal factor — temperature. If a slope water animal were physiologically restricted to a particular temperature range, its habitat would descend as the ring decayed and isotherms sank. To the ex- tent that the zooplankton population in the slope water exhibited this behavior, ring biomass dis- tributions would deepen. This could apply only to a species which in its home range — the slope wa- ter— remains beneath the seasonal thermocline (i.e., moderately deep-living and exhibiting li- mited diel migration). Such a species would most likely have to be either carnivorous or omnivor- ous. Wiebe and Boyd ( 1978) have documented such a phenomenon for the slope water euphausid species, Nematoscelis megalops. A more complex explanation stresses the impor- tance of a biological factor — food resources. The kinds of changes that accompany ring decay must have a substantial effect upon zooplankton- phytoplankton interactions. Using unpublished 331 FISHERY BULLETIN: VOL. 76, NO 2 data obtained in August 1975 from 5 nine bottle hydrocasts, the number of phytoplankton cells per liter averaged 10,000 in the slope water, 2,500 in the ring, and 2,000 in the Sargasso Sea. Cells smaller than 4-5 /xm were not enumerated and were, therefore, excluded from these computa- tions. Values were integrated from 0 to 200 m — a conservative procedure tending to reduce slope water versus ring or Sargasso Sea differences. The species composition of the ring, while distinct, was more like that of the Sargasso Sea than that of the slope water. Again, considering the 0-200 m depth interval, the number of different phytoplankton species an animal would have encountered in a liter of water would, on the average, have been 6.0 (slope water), 9.6 (ring), and 10.4 (Sargasso Sea). Converting the mean cell volume of each species to carbon (Strathmann 1967) and multiplying by the number of individuals present, yielded values of average phytoplankton carbon of 1,400, 200, and 140 ng C/1. Thus, to acquire the same ration of food, a herbivore would have had to filter more than five times more water in the ring than in the slope water, and even more in the Sargasso Sea. In addition, the evenness of species' carbon equiva- lence was 0.46, 0.75, and 0.76. That is, the total carbon per liter was more evenly distributed among different species in the ring and the Sar- gasso Sea than in the slope water. (Evenness equals HIH„^^^ (Pielou 1966) where H is the Shannon-Weaver diversity index computed upon species carbon equivalence rather than abun- dance and //max ^ ^ogj, S where S = number of species.) This last result implies that a herbivore capable of selecting by carbon content (i.e., parti- cle size) would have found it less advantageous to concentrate on a particular species in the Sargasso Sea and the ring than in the slope water. These properties of the phytoplankton popula- tion, i.e., species composition, carbon concentra- tion, cell concentration, and cell carbon distri- bution, have profound effects on a filter-feeding herbivore's harvesting ability. We believe that early in ring evolution herbivorous slope water species are deleteriously affected and, therefore, may be replaced by Sargasso Sea forms more quickly than deeper living carnivorous or om- nivorous slope water species. If we are correct, ring biomass distribution may deepen in part because a ring's 0-200 m biomass declines more rapidly than does its 200-800 m biomass. Identification of some of the taxa in August 1975 samples, although limited, support the argument that in Ring-D epizooplanktonic herbivores were replaced before epizooplanktonic carnivores or omnivores. The species list of Ring-D thecosoma- tous pteropods, a largely herbivorous group, was quite similar to that of the surrounding Sargasso Sea.^ Grice and Hart (1962) found that chaetog- naths, a purely carnivorous group, were consider- ably more abundant in the Sargasso Sea than they are in slope water. In 6 nine-net fine-mesh tow series (12.5 cm diameter, Clarke-Bumpus nets with 67 ^tm mesh) taken in August, chaetognaths were five to ten times more abundant in the sur- rounding Sargasso Sea than they were in Ring-D. Other epizooplanktonic carnivores, e.g., Stylo- cheiron suhmii and S. abbreviatum, which are routinely found in the Sargasso Sea were not found in Ring-D August MOCNESS tows. Organic Flux to Deep Sea Rings may contribute a disproportionate frac- tion of the utilizable organic material available to the northern Sargasso deep sea. We feel this is likely both because of their generally higher pro- ductivity and because of their unique zooplankton biomass distribution and the factors that have re- sulted in that distribution. Ring zooplankton biomass below 200 m, in that it exceeds Sargasso Sea biomass and ultimately declines to a similar level, contributes to this augmentation. Differen- tial seasonal mixing processes could also increase downward particulate flux. For example, in November 1975 we observed that winter mixing had proceeded further in Ring-D than in the sur- rounding Sargasso Sea water column. Herbivor- ous ring zooplankton (i.e., Sargasso forms) may have been unable to fully capitalize upon the sud- den opportunity afforded by the increased primary production that accompanied the mixing (Table 7). If so, a larger fraction of this enhanced phyto- plankton production would sink into the aphotic depths. Physical evidence obtained on two cruises undertaken to study rings during the summer has suggested to us that the seasonal thermocline may often be less stable in rings than in the Sargasso Sea. Finally, there is a possibility of enhanced con- tribution of organic matter into the deep sea due to a lower overall trophic efficiency within the upper 200 m of rings (and slope water). If we divide ^John Wormuth, unpubl. data; cited with permission. 332 ORTNER ET AL.: SARGASSO SEA ZOOPLANKTON BIOMASS DISTRIBUTION average 0-200 m zooplankton biomass (milligrams of carbon per square meter calculated using equa- tion 4, table 2 in Wiebe et al. 1975) by 0-200 m phytoplankton carbon (milligrams of carbon per square meter from Table 5), excluding salp-rich MOC 18 and 19, we obtain the following ratios: Aug. 1975 Nov. 1975 Sargasso Slope Sea Ring water 253 138 84 332 131 28 Ratios in the ring are low, as are those in the slope water. Lower ratios suggest to us lower overall trophic efficiency within the upper 200 m. Al- though biased in that many cells are quite small, particularly in the Sargasso Sea, phytoplankton carbon of cells >5 ixm is probably a reasonable estimate of the food available at the time of sam- pling to many of the herbivorous animals caught by our 0.333-mm mesh nets. The direction of dif- ference noted above conforms with ideas expressed by Menzel and Ryther ( 1961 ), Heinrich ( 1962), and others who argued that especially close phytoplankton-zooplankton coupling may charac- terize oceanic tropical-subtropical waters. The biomass data presented here illustrate the fact that geographic demarcation of oceanic faunal provinces is not sufficient. Hydrographic as well as faunal mapping is essential in explaining that portion of station-to-staion variability associated with mesoscale hydrographic variability resulting from phenomena like Gulf Stream cold core rings. ACKNOWLEDGMENTS We express our grateful appreciation to Alfred Morton for his assistance at sea; to Michael Stro- man who measured many of the displacement vol- umes; and to our typist Jane Peterson, for her remarkable tolerance and good humor. James Cox and Margo Haygood critically read the manu- script. This study was supported by ONR NOOO14-66-C-0240, NOOO14-24-C-0262 NR 083-004, NSF DES 74-02783A1, the Woods Hole Oceanographic Institution Graduate Education Program, and the Tai Ping Foundation. LITERATURE CITED AHLSTROM, E. H., AND J. R. THR.'MLKILL. 1963. Plankton volume loss with time of preserva- tion. Calif. Coop. Oceanic Fish. Invest. Rep. 9:57-73. Be, a. W. H., J. M. FORNS, AND O. A. ROELS. 1971. Plankton abundance in the North Atlantic Ocean. In J. D. Costlow, Jr. (editor), Fertility of the sea, Vol. 1, p. 17-50. Gordon and Breach Sci. Publ., N.Y. Deevey, G. B. 1971. The annual cycle in quantity and composition of the zooplankton of the Sargasso Sea off Bermuda. I. The upper 500 m. Limnol. Oceanogr. 16:219-240. Deevey, g. b., and a. L. Brooks. 1971. The annual cycle in quantity and composition of the zooplankton of the Sargasso Sea off Bermuda. II. The surface to 2000 m. Limnol. Oceanogr. 16:927-943. Fish, C. J. 1954. Preliminary observations on the biology of boreo- arctic and subtropical oceanic zooplankton populations. In Symposium on marine and fresh-water plankton in the Indo-Pacific, Bangkok, p. 3-9. Diocesan Press, Madras. FUGLISTER, F. C. 1972. Cyclonic rings formed by the Gulf Stream 1965- 66. In A. L. Gordon (editor). Studies in physical oceanog- raphy. A tribute to George Wiist on his 80th birthday, Vol. 1, p. 137-168. Gordon and Breach, N.Y. Grice, G. D., and a. D. Hart. 1962. The abundance, seasonal occurrence, and distribu- tion of the epizooplankton between New York and Ber- muda. Ecol. Monogr. 32:287-309. Heinrich, a. k. 1962. The life histories of plankton animals and seasonal cycles of plankton communities in the oceans. J. Cons. 27:15-24. lASHNOV, V. A. 1961. Vertikal'noe respredelenie massy zooplanktona tropicheskoi oblasti Atlanticheskogo Okeana (Vertical distribution of the mass of zooplankton throughout the tropical zone of the Atlantic). Dokl. Akad. Nauk. SSSR 136:705-708. JAHN, a. E. 1976. On the midwater fish faunas of Gulf Stream rings, with respect to habitat differences between Slope water and Northern Sargasso Sea. Ph.D. Thesis, Woods Hole Oceanogr. Inst., Woods Hole, 172 p. Lai, D. Y., and p. L. Richardson. 1977. Distribution and movement of Gulf Stream rings. J. Phys. Oceanogr. 7:670-683. LEAVITT, B. B. 1935. A quantitative study of the vertical distribution of the larger zooplankton in deep water. Biol. Bull. (Woods Hole) 68:115-130. 1938. The quantitative vertical distribution of macrozoo- plankton in the Atlantic Ocean basin. Biol. Bull. (Woods Hole) 74:376-394. Marlowe, C. J., and C. B. Miller. 1975. Patterns of vertical distribution and migration of zooplankton at Ocean Station "P". Limnol. Oceanogr. 20:824-844. MCGOWAN, J. A., AND D. M. BROWN. 1966. A new opening-closing paired zooplankton net. Scripps Inst. Oceanogr. Ref 66-23:1-56. MENZEL, D. W., AND J. H. RYTHER. 1961 . Zooplankton in the Sargasso Sea off Bermuda and its relation to organic production. J. Cons. 26:250-258. MOORE, H. B. 1949. The zooplankton of the upper waters of the Bermuda area of the North Atlantic. Bull. Bingham Oceanogr. Collect., Yale Univ. 12(2), 97 p. 333 FISHERY BULLETIN: VOL. 76, NO. 2 MURANO M., R. MARUMO, T. NEMOTO, AND Y. AlZAWA. 1976. Vertical distribution of biomass of plankton and micronekton in the Kuroshio water off Central Ja- pan. Bull. Plankton. Soc. Jap. 23:1-12. Nemenyi, p. B. 1963. Distribution-free multiple comparisons. Ph.D. Thesis, Princeton Univ., Princeton, 127 p. Parker, C. E. 1971. Gulf Stream rings in the Sargasso Sea. Deep-Sea Res. 18:981-993. PlELOU, E. C. 1966. The measurement of diversity in different types of biological collections. J. Theor. Biol. 13:131-144. Richardson, P. 1976. Gulf Stream rings. Oceanus 19(3):65-68. RICHARDSON, P. L., R. E. CHENEY, AND L. A. MANTINI. 1977. Tracking a Gulf Stream ring with a free drifting surface buoy. J. Phys. Oceanogr. 7:580-590. RYTHER, J. H. 1963. Geographic variations in productivity. In M. N. Hill (editor). The sea. Ideas and observations on progress in the study of the seas, Vol. 2, p. 347-380. Interscience Publ., Lond. > Saunders, P. M. 1971. Anticyclonic eddies formed from shoreward mean- ders of the Gulf Stream. Deep-Sea Res. 18:1207-1219. STRATHMANN, R. R. 1967. Estimating the organic carbon content of phyto- plankton from cell volume or plasma volume. Limnol. Oceanogr. 12:411-418. Vaccaro, R. F., and J. H. RYTHER. 1960. Marine phytoplankton and the distribution of ni- trite in the sea. J. Cons. 25:260-271. Vinogradov, M. E. 1968. Vertical distribution of the oceanic zooplankton. [In Russ.] Isdatel'stvo Nauka, Moskva, 319 p. (Isr. Prog. Sci. Transl., 1970, 339 p.) Wiebe, p. H., and S. H. Boyd. 1978. Limits of Nematoscelis megalops in the Northwest- em Atlantic in relation to Gulf Stream cold core rings. Part I. Horizontal and vertical distributions. J. Mar. Res. 36:119-142. Wiebe, P. H., S. Boyd, and J. L. Cox. 1975. Relationships between zooplankton displacement volume, wet weight, dry weight, and carbon. Fish. Bull., U.S. 73:777-786. Wiebe, P. H., K. H. Burt, S. H. Boyd, and A. W. Morton. 1976. A multiple opening/closing net and environmental sensing system for sampling zooplankton. J. Mar. Res. 34:313-326. Wiebe, P. H., E. M. Hulburt, E. J. Carpenter, A. E. Jahn, G. P. Knapp, III, S. H. Boyd, P. B. Ortner, and J. L. Cox. 1976. Gulf Stream cold core rings: large-scale interaction sites for open ocean plankton communities. Deep-Sea Res. 23:695-710. Zenkevich, L. a., and J. A. Birstein. 1956. Studies of the deep water fauna and related prob- lems. Deep-Sea Res. 4:54-64. 334 RELATIVE CONTRIBUTION OF HUDSON, CHESAPEAKE, AND ROANOKE STRIPED BASS, MORONE SAXATILIS, STOCKS TO THE ATLANTIC COAST FISHERY Thomas J. Berggren' and Joel T. Lieberman^ ABSTRACT Morphological characters were used in discriminant analysis to quantitatively estimate the relative contribution of striped bass, Morone saxatilis, stocks from various estuaries to the striped bass fishery along the Atlantic coast. Representative samples of the spawning stocks of the Hudson River, Chesapeake Bay system, and Roanoke River were collected and counts and measurements were taken on each specimen. Discriminant functions based on five morphological characters correctly classified approximately TS^c of the specimens. The effectiveness of three types of estimates based on these functions in accurately estimating stock proportions was investigated in a simulation study. Results of the simulation study indicated which type of estimate was least biased. A sampling design using geographical and temporal strata was then employed to sample the Atlantic coastal fishery from Cape Hatteras, N.C., to Maine. Observations for the morphological characters were taken on collected fish and the resulting data entered into discriminant functions obtained from spawning-stock collections. The specimens were classified by area of origin and the three types of estimates of relative contribution of the Hudson, Chesapeake, and Roanoke stocks were obtained. Results indicated that the Chesapeake stock was the major contributor to the Atlantic coastal striped bass fishery and the Hudson and Roanoke stocks were minor contributors. The striped bass, Morone saxatilis, is an important sport and commercial fish in the estuaries and coastal waters of the Atlantic seaboard from Maine to North Carolina (Koo 1970). Recruitment to the striped bass fishery is from various stocks of striped bass spawned and developed in rivers and estuaries along the Atlantic coast. Recapture loca- tions of tagged striped bass indicate that individu- als from all spawning areas north of Cape Hat- teras, N.C., utilize much of the Atlantic coast north of their respective spawning areas during a northward migration in the spring and a south- ward migration in the fall (Merriman 1941; Raney et al. 1954; Alperin 1966; Schaefer 1968; Flor- ence^; Texas Instruments^). The major spawning areas which potentially contribute individuals to the fisheries operating during the northward and southward migrations are the tributaries of 'Texas Instruments Inc., Buchanan, N.Y.; present address: Biometrics Unit, 337 Warren Hall, Cornell University, Ithaca, NY 14853. ^Texas Instruments Inc., P.O. Box 237, Buchanan, NY 1051 1. ^Florence, B. 1974. Tag returns from 1375 large striped bass tagged in two Maryland spawning rivers. Outdoor Message. Organized Sportsmen of Mass. Oct. 1974. ■•Texas Instruments Inc. 1976. Report on relative contribution of Hudson River striped bass to the Atlantic coastal fishery. Prepared for Consolidated Edison Company of New York, Inc., 101 p. Chesapeake Bay and the Roanoke and Hudson Rivers. Although tagging data have not led to quantita- tive estimates of relative contribution, they have led to conflicting ideas as to which major stock of striped bass predominates in the fishery: the Hud- son stock or the Chesapeake stock. Most published works have generally concluded that the striped bass stock from the Chesapeake Bay system is the major contributor to the fisheries north of Chesapeake Bay (Merriman 1941; Vladykov and Wallace 1952; Alperin 1966; Schaefer 1968; Porter and Saila^; Raney^). However, Clark'' and Goodyear^ concluded that the striped bass stock Manuscript accepted August 1977. FISHERY BULLETIN: VOL. 76, NO. 2. 1978. ^Porter, J., and S. B. Saila. 1969. Final report for the coopera- tive striped bass migration study. U.S. Fish. Wildl. Serv. Con- tract no. 14-16-005, 33 p. "Raney, E. C. 1972. The striped bass, Morone saxatilis, of the Atlantic coast of the United States with particular reference to the population found in the Hudson River. Testimony before USAEC Safety and Licensing Board for Indian Point, Unit no. 2. Docket no. 50-247, Oct. 30, 105 p. 'Clark, J. 1972. Effects of Indian Point Units 1 and 2 on the Hudson River aquatic life. Testimony before USAEC Safety and Licensing Board for Indian Point, Unit no. 2. Docket no. 50-247, Oct. 30, 63 p. ^Goodyear, C. P. 1974. Origin of the striped bass of the middle Atlantic coast. Testimony presented to the Committee on Mer- chant Marine and Fisheries of the U.S. House of Representa- tives. Feb. 19, 40 p. 335 FISHERY BULLETIN: VOL. 76, NO. 2 from the Hudson River is the major contributor to the coastal fishery from New Jersey to Mas- sachusetts because the number of striped bass tagged in Chesapeake Bay and recaptured outside the Bay was too low to indicate a large contribu- tion of Chesapeake stock to that fishery. Because of the controversy of which stock pre- dominates, we conducted a study to obtain quan- titative estimates of relative percentage of the major stocks in the coastal fishery. A previous study (Grove et al. 1976) demonstrated the feasi- bility of using discriminant analysis on mor- phological characters (counts and morphometric ratios) to distinguish among Hudson, Chesapeake, and Roanoke spawning stocks of striped bass. That study showed that adequate segregation of spawn- ing stocks within the Chesapeake Bay system was not possible. Quantitative estimates of stock com- position based on morphological characters and discriminant analysis have been obtained for sockeye salmon (Fukuhara et al. 1962; Anas and Murai 1969), pink salmon (Amos et al. 1963), and Atlantic herring (Messieh 1975). The present study establishes discriminant functions based on collections of spawning-stock specimens to classify striped bass collected in the Atlantic coastal fishery from southern Maine to Cape Hatteras. The percentage of specimens collected that were classified into each stock was used to estimate the relative contribution of that stock to the fishery. METHODS AND MATERIALS Collection of Spawning-Stock Specimens During the spawning season of 1975, mature striped bass were collected from the natal rivers of major stocks along the Atlantic coast. These fish were assumed to have originated from the rivers (i.e., that striped bass, like salmon and other anadromous fishes, home to their natal stream to spawn). This assumption was supported by tag- ging studies in which striped bass tagged on spawTiing grounds were recaptured on the same spawning grounds in successive years (Mansueti 1961; Nichols and Miller 1967). Collections were composed of 232 mature striped bass from the Chesapeake Bay tributaries (70 from the Rap- pahannock River, 53 from the Potomac River, 52 from the Choptank River, and 57 from the Elk River and Chesapeake and Delaware Canal), 168 from the Hudson River, and 99 from the Roanoke River. Only 19 sexually ripe striped bass were collected from the Delaware River above the en- trance to the Chesapeake and Delaware Canal, which confirms findings by Chittenden ( 1971) that spawning in the Delaware River is not substan- tial. Therefore specimens from the Delaware River were omitted from subsequent analyses. Collections were made primarily during April in the Chesapeake Bay tributaries, Delaware and Roanoke Rivers, and during May in the Hudson River. Most specimens were obtained fresh from commercial fishermen using pound nets, haul seines, and gill nets. Some were netted by study personnel. To assure an adequate representation of the sexes and multiple year classes in spawning-stock collections, sampling was designed to obtain nearly equal numbers of male and female striped bass and a minimum of 10 individuals in each of the following length categories: ^399, 400-549, 550-699, 700-849, and &850 mm. Discriminant functions based on male and female specimens from multiple year classes are needed to analyze an oceanic population which consists of a different sex ratio and broader age structure than that of the spawning stocks. Processing of Spawning-Stock Specimens Scale samples, counts, measurements, sex, and state of maturity were obtained from each speci- men while in fresh condition. Scale samples from above the lateral line between the first and second dorsal fins were pressed on acetate cards. Ages were determined by the scale annulus method (Mansueti 1961). Measurements from the focus to the first and second annuli were made on mag- nified scale images. The following counts and measurements were taken: number of lateral line scales, left pectoral rays, right pectoral rays, sec- ond dorsal rays, anal rays, upper-arm gill rakers, fork length, snout length, head length, and inter- nostril width. Methods used were those discussed by Hubbs and Lagler (1958) and Grove et al. (1976). Counts, measurements, and age determinations were replicated by a second observer and a set of tolerances was established to reduce observation error. When differences between replicated obser- vations exceeded tolerances, the observations were retaken. Means of the replicated counts and means of ratios of the replicated measurements were used in subsequent analysis. 336 BERGGREN and LIEBERMAN: RELATIVE CONTRIBUTION OF STRIPED BASS Analysis of Spawning-Stock Specimens Choice of morphological characters for segrega- tion of Hudson, Chesapeake, and Roanoke spawn- ing stocks followed three stages of statistical analysis: correlation analysis between each character and fork length (FL), analysis of the effects of sex and age on each character, and dis- criminant analysis. Analysis involved only speci- mens with observations on all counts and mor- phometric and scale-annulus measurements. Since spawning stocks do not include immature specimens which occur in the coastal waters, we chose only those characters that were independent (i.e., not highly correlated) offish size and could therefore be used to segregate specimens from the entire stock. Characters were considered to be in- dependent of length when variations (r^) attribut- able to length in any stock were ssO.lO. Characters not independent of length were used in further analysis when the distribution of character values had small overlap among spawning stocks since such characters help identify stock origin. Multivariate statistical tests were made to de- termine the effect of sex and age on the characters used to determine the discriminant functions, since one assumption of discriminant analysis was that each stock was homogeneous. Differences in character values among ages for males or females and between sexes within each stock were tested with a procedure that combined tests of equality of means and equality of covariance matrices (An- derson 1958). Assuming equal covariance ma- trices, rejection of the null hypotheses of equal distributions indicated that one or more of the character means differed among ages or between sexes. Multivariate discriminant analysis was used to gain maximum separation among stocks. Linear and quadratic discriminant functions (Anderson 1958; Kendall and Stuart 1968) for each spawning stock were determined from character values ob- tained from collections of that stock. A stepwise procedure on the linear function was used to indi- cate the subset of characters which best separated the stocks. The quadratic function based on this subset was formed if the assumption of a common covariance matrix among spawning stocks needed for the linear function was not met. The assump- tion in discriminant analysis that characters had a multivariate normal distribution was investi- gated with histograms. Ability of the discriminant functions to separate stocks and accurately estimate stock proportions was assessed using functions based on total spawning-stock collections and functions obtained from a cross-validation procedure (Mosteller and Tukey 1968). In this procedure collections were randomly divided in half and discriminant func- tions were determined from one-half and applied to each half. Percentages of correct classification and estimates of stock proportion were obtained for each subset and compared with those from the total sample. Comparisons were also made be- tween estimated and known spawning-stock per- centages. Although these estimates of stock percentages may accurately approximate true percentages in spawning-stock collections, they may deviate sub- stantially from stock percentages in oceanic col- lections. Fukuhara et al. (1962) stated that the bias in these estimates increased as stock percen- tages became more disproportionate. Since stock percentages in oceanic collections may be more disproportionate than stock percentages in spawning-stock collections (i.e., 347^ Hudson, 469c Chesapeake, and 20*^ Roanoke stocks), less biased estimates of stock percentages may be needed. Adjusting Estimates of Stock Percentages Two procedures were developed to obtain esti- mates of stock percentages that were less biased than the as-classified (i.e., classifications obtained directly from discriminant functions) estimates. The first procedure adjusted estimates using a technique described by Worlund and Fredin (1962) which generalized to the three population case methodology developed in Fukuhara et al. (1962). This procedure used percentages of speci- mens from each spawning stock that were mis- classified into other stocks to correct as-classified estimates for bias due to misclassifications. When adjusted estimates were negative, as-classified es- timates were modified by methodology developed by Schuermann and Curry. ^ The second procedure iteratively reclassified specimens based on updated prior probabilities that specimens originated from each of the spawn- ing stocks. The first stage of the procedure is the same as the as-classfied procedure; therefore as- ^Schuermann, A. C, and G. L. Curry. 1973. Notes on paramet- ric programming. Unpubl. manuscr. Dep. Ind. Eng. Texas A&M Univ., College Station. 337 FISHERY BULLETIN: VOL. 76, NO. 2 classified estimates of stock contribution are ob- tained at the end of this stage. However, these estimates are then used in the second stage as prior probabilities that specimens come from the three stocks. For example, the as-classified esti- mate of Hudson stock contribution obtained at the end of the first stage was used at the beginning of the second stage as our best guess of the proportion of specimens in the sample that originate from the Hudson. These prior probabilities are then used to weight the decision to classify each specimen into one of the stocks. Similarly, the proportion of specimens classified into each stock in the second stage were used as priors in the third stage. The procedure was carried out for nine stages. The effectiveness of adjusted and iterative esti- mates in reducing bias in the as-classified esti- mate due to misclassification was investigated in a simulation study. Discriminant functions from the cross-validation study were used to classify a subset of specimens from the independent half of the spawning-stock collections, and each of the three types of estimates of relative percentage were obtained and compared with the known stock percentage. For percentages of Hudson stock rang- ing from 0 to 907f , the difference between each estimate of Hudson percentage and the known percentage of Hudson specimens in the subsample was obtained as a measure of bias in the estimate. Collection, Processing, and Analysis of Atlantic and Hudson River Specimens Assessment of the relative contribution of vari- ous stocks of striped bass to the Atlantic coastal fishery required a stratified sampling design that provided samples from the entire coastal fishery and considered the migratory nature of striped bass; therefore a geographically and temporally stratified sampling design was used. The geo- graphical stratification consisted of 10 strata from southern Maine to Cape Hatteras, with 2 to 4 substrata within each stratum to compensate for variations in stock composition within the stratum ( Figure 1 ). The Rhode Island stratum was not subdivided because of its small size. Tempo- rally, the year was divided into six 2-mo periods to obtain estimates of stock composition by stratum throughout the year. Collections of striped bass from the coastal fishery were obtained primarily from sport and commercial fishermen; however, in areas where adequate sport and commercial fisheries did not exist, study personnel used haul seines and gill nets to collect specimens. Collections were limited to striped bass caught during the same day (i.e., within 24 h) to assure freshness. In many in- stances the entire catch was used, but due to the size of some catches, a random sample propor- tional to the number of small ( <550 mm), medium (550-850), and large (>850) striped bass caught was obtained. Oceanic and overwintering specimens were pro- cessed in the same manner as spawning-stock specimens. Two replicates of 10 counts and mea- surements were taken from each specimen, and scale samples were obtained for subsequent age and growth rate determinations in the laboratory. A total of 2,737 oceanic specimens with a complete set of meristic, morphometric, and scale charac- ters were processed (Table 1). Additionally, 79 striped bass overwintering in Croton Bay on the Table l . — Number of striped bass with complete character sets' collected by spatial stratum and period from Atlantic coastal fishery in 1975. Spatial stratum 8 9 10 Total Locality Legal/ subiegai^ S Maine-N Mass S. Mass. Rhode Is. E. Long Is. Sound W Long Is Sound E Long Is S. Shore W Long Is S Shore N.J, Del.-Md -N Va. S Va.-N C. Jan. -Feb. Mar.-Apr. May-June July-Aug. Sept -Oct. Nov. -Dec. Total Legal 82 58 74 214 Legal 91 . 90 82 263 Legal 60 43 56 159 Legal 96 140 99 335 Sublegal 5 1 6 Legal 1 38 14 15 89 157 Sublegal 2 42 85 10 139 Legal 1 89 102 86 106 384 Sublegal 8 17 19 44 Legal 30 58 93 120 124 425 Sublegal 4 11 15 Legal 34 113 28 73 117 365 Legal 71 3 6 100 180 Legal 27 24 51 28 180 672 531 755 571 2,737 'Measurements and counts taken on all variables used in the character set ^Sublegal-sized striped bass (< 406.5 mm FL) from New York waters (strata 4 to 7) were analyzed separately. 338 BERGGREN and LIEBERMAN: RELATIVE CONTRIBUTION OF STRIPED BASS ATLANTIC OCEAN ST CROIX R>vEft( Figure l.— Collection regions for the Atlantic coastal fishery for striped bass showing geo- graphical stratification and substratification; collection sites for spawning-stock speci- mens indicated by dots on source rivers. CEAN 10-3 ICAPE HATTERAS WILMINGTON, SC CAPE FEAR 339 FISHERY BULLETIN: VOL. 76, NO. 2 Hudson River from 6 December 1974 through 20 March 1975 were processed. Three estimates of stock contribution, i.e., "as- classified," "adjusted," and "iterative" estimates, were calculated for collections of legal-sized, sublegal-sized, and overwintering striped bass by geographical and temporal strata. Sublegal-sized <406.5 mm or 16 in FL) and overwintering striped bass collected in New York waters were not con- sidered to be a part of the coastal fishery and were analyzed separately. In each stratum, the percentage of striped bass allocated to a stock pro- vided an estimate of that stock's relative contribu- tion. Mean 1975 estimates of stock contribution of legal-sized striped bass were calculated by averag- ing strata estimates within periods then averag- ing across the six periods. Relative contribution estimates by age were also obtained. The influence of the Hudson stock in coastal strata adjacent to the Hudson River was investi- gated by comparing the relative contribution of Hudson, Chesapeake, and Roanoke stocks within "inner" and "outer" zones designed by the U.S. Nuclear Regulatory Commission.'*^ The inner zone encompassed western Long Island Sound (stratum 5), the New York Bight (stratum 7), and northern New Jersey (stratum 8-1), whereas the outer zone encompassed the remaining waters from Cape May, N.J., to Maine (strata 1 to 4, 6, 8-2, 8-3). Estimates of relative contribution for inner and outer zones were calculated for each period by summing the number of Hudson-, Chesapeake-, and Roanoke-classified fish within appropriate strata. Mean estimates of contribution within each zone were calculated for the year by averag- ing across temporal strata. RESULTS AND DISCUSSION Establishment of Discriminant Functions Five characters were established as the charac- ter set best able to discriminate among Hudson, Chesapeake, and Roanoke stocks. They are, in order of importance (as established by stepwise linear discriminant analysis): 1) the ratio of snout length/internostril width, 2) the scale ratio of first to second annulus/focus to first annulus measure, '"U.S. Nuclear Regulatory Commission. 1975. Final environ- mental statement related to operation of Indian Point Nuclear Generating Plant Unit no. 3 Consolidated Edison Company of New York, Inc. Office of Nuclear Reactor Regulation. Docket no. 50-286, Vol. 1:V-166-V-178. 3) a character index (Raney and deSylva 1953), 4) the upper-arm gill raker count (which includes rudimentary rakers), and 5) the lateral line scale count. The character index, i.e., the sum of left and right pectoral, second dorsal, and anal fin rays, was used since Grove et al. (1976) demonstrated that individual fin ray characters did not add sig- nificant discriminatory ability. The five characters satisfied the criterion for independence with fish length in each stock with only one exception. The snout length/internostril width ratio for the Roanoke stock has a coefficient of determination of nearly 0.20 but was retained because its distribution had the least overlap among spawning stocks of all characters, thus making it a potentially good discriminator. Results of the test of homogeneity indicated that only the Hudson stock was homogeneous among ages and between males and females. Significant differences (a = 0.05) were found among ages and between sexes in the Chesapeake spawning stock and among ages in the Roanoke spawning stock. Differences found in the Chesapeake spawning stock may have resulted from pooling collections from its four major tributaries. Quadratic functions (Table 2) were used to dis- criminate among stocks as a result of the investi- gation of underlying assumptions of discriminant analysis. Significant differences (a = 0.05) were found among covariance matrices of Hudson, Chesapeake, and Roanoke spawning stocks which suggested that quadratic functions would better discriminate among these stocks than linear func- tions. Histograms suggested that no radical de- parture of multivariate normality was evident, although normality of individual characters does not assure multivariate normality of the character set. Therefore multivariate normality of the character sets was assumed. Percentage of spawning-stock specimens cor- rectly classified by the quadratic functions and estimated stock percentages resulting from the use of these functions closely agreed with results obtained by the cross-validation procedure (Table 3). For the total set of collections, 76. 8^^ of Hudson specimens, 67.T7( of Chesapeake specimens, and 85. 9^^ of Roanoke specimens were correctly clas- sified, resulting in an overall correct classification of 74.4%. This was similar to overall percentages of 73.2 and 77.1 obtained for the cross-validation subsets. Estimated relative percentages for each stock varied <3 percentage points among the total set and cross-validated subsets, whereas varia- 340 BERGGREN and LIEBERMAN: RELATIVE CONTRIBUTION OF STRIPED BASS Table 2. — Quadratic discriminant functions' based on Hudson, Chesapeake, and Roanoke spawning-stock specimens of striped bass and used to classify spawning-stock, oceanic, and overwintering specimens.^ Hudson; ''hud =^ 1.489.070559 - + 0 090968 UZ + + 24 321052 X r Chesapeake ''CHES = - 1.368 946420 + 0.321075 C/Z - + 25 294896 X + Roanoke: '^ROAN = - 1.650 902863 + 0.228873 UZ + 25 512087 X - (0.077516 U^ i 0.256954 W^ > 0,047441 WX + 0.023246 WY + 7 985031 Y > 381 695141 Z - (0.089560 U^* 0 242459 W? 0.092151 WX 0,000861 WY 7.014936 Y » 323 469441 Z 1 171065X2 ^ 0 164200 WZ 2 536320 Y^ 0.457365 Xy 123 907000 Z2 2 799760 XZ - 0.019058 UW 2.861250 YZ) + <- 0.015160 UX - 0.007057 UY 8.776221 U + 28,127772 W 1 122690X2 + 2 155850/2 + 117 554000 Z^- 0.007099 UW + 0.005302 UX 2 363980 WZ + 0 381082 XY + 3.623860 XZ - 1 590090 VZ) + 1 1 316822 U - (0.107062 U2 0.293615 WX 22.351388 Y + 0.316254 W2 t + 0 129292 WY - 469 422957 Z 2 063540X2 1 009790 WZ 0.842590 Y^ + 139.577500 Z2 0 106776 Xy - 0 606466 XZ + 0 062826 UW + 0.015703 UX 4 416000 YZ) + 10 320202 U ^ 0.015500 uy 21 749040 W 0 043640 UY 27 000888 W 'Except for an additive constant ( -2.5 In 2tt) common to each function. 2F = discriminant score. U = lateral line scale count, W = character index, X = upper-arm gill raker count, Y measurement ratio, and Z = snout length mternostril width ratio. first to second annulus/focus to first annulus Table 3. — Comparison of correct-classification percentages and estimated and known stock percentages among the total set of spawning-stock specimens of striped bass and cross-validation subsets. Correctly Spawning classified stock (%) Random set' Known stock (%) Estimated stock (%) Correctly classified (%) Independent set2 Known stock (%) Estimatec stock (%) Correctly classified (%) T^tal^t^ Known stock (%) 'Randomly sampled half of total spawmng-stock collections used to determine quadratic functions for cross-validation. ^Remaining half of spawmng-stock specimens classified by quadratic functions based on the random set. ^AII specimens from spawmng-stock collections classified by quadratic functions based on the total set. Estimated stock (%) Hudson 81 0 337 36.5 72.6 33.6 35.2 76.8 337 36.9 Chesapeake 69 8 466 40.2 68 1 464 424 67.7 465 40.3 Roanoke 878 19.7 233 86.0 200 224 859 198 22.9 Overall 77 1 732 74.4 tions between estimated and known stock percentages within sets was as much as 9 percent- age points. The quadratic functions thus provided slightly biased estimates of stock percentages when applied to collections composed of 34^^ Hud- son, 46*^ Chesapeake, and 20'7( Roanoke stocks. Best Estimator of Relative Contribution The best estimate of the percentage of Hudson River specimens in subsamples from the simula- tion studies was the estimate from the third itera- tion of the reclassification procedure (Table 4). On the average, this iterative estimate was less biased than estimates from other iterations, the as-classified estimated (i.e. , estimate from the first iteration), and the adjusted estimate for most per- centages of Hudson stock considered. In addition, the variance of the bias of the iterative estimate was often less than that of the other estimates. For percentages of Hudson stock 'e 50*^ , the iterative and adjusted estimates closely agreed and the bias in each estimate was small ( 'SS percentage points). The iterative estimate will, therefore, be used to estimate Hudson stock contribution in oceanic collections, and the adjusted estimate will be used to substantiate estimates of Hudson con- TABLE 4. — Mean and standard deviation of absolute bias' of estimated relative percentages of Hudson River stock of striped bass in replicated random samples from spawning-stock collec- tions.^ Known percent of Hudson Estimates of absolute bias As-classified Mean SD Iterat ive Adju Mean sted River stock Mean SD SD ^90 23.0 2.40 4.3 2.97 14.4 5 73 80 20.2 5.14 7.4 882 14.3 7.78 75 17.6 3.47 8.4 382 128 5.00 70 13.0 3.32 4.7 452 7.3 4.80 65 10.8 3.11 33 298 7.5 4.07 60 9.0 3.21 53 284 68 5.00 55 7.5 2.95 4.8 3 52 74 4.02 50 5.5 1.99 4.2 266 4.7 3.44 45 2.2 2.17 3.5 1 85 4.7 2 14 40 1.2 1.04 3.2 3.44 43 421 35 2.2 1.45 "4.2 "2 09 44 2.87 30 5.9 3.18 3.3 243 3.4 1.79 25 7.8 3.07 4.0 225 4.2 3.27 20 9.5 2.26 2.8 1.72 1,8 099 15 12.1 4.19 3.5 3.24 33 262 10 15.1 3.80 4.5 3.36 50 3.61 5 17.4 4.03 4.5 3.72 43 385 0 18.1 2.53 2.5 1.73 1.5 1.69 Overall mean 11.0 4.4 6.2 'Absolute value of the difference between the true relative percentage of Hudson River stock in the subsample and the estimated relative percentage based on nine replicates of varying Chesapeake and Roanoke proportions in the subsamples 2Estimates were based on random samples from one-half of spawmng-stock collections which were classified as to area of origin by quadratic functions obtained from the other half of the collections. ^Based on two replicates. ■■Based on eight replicates. tribution « 50^^. The iterative estimate will also be used to estimate Chesapeake and Roanoke stock contributions. 341 FISHERY BULLETIN: VOL. 76, NO. 2 Estimates of Stock Contribution for Oceanic and Overwintering Collections Iterative estimates of relative contribution of Hudson, Chesapeake, and Roanoke stocks indi- cated that the Chesapeake stock was the major contributor to the striped bass fishery along the Atlantic coast while the Hudson and Roanoke stocks were minor contributors (Table 5). The Chesapeake stock predominated in 34 of 35 geo- gi-aphical and temporal strata while the Hudson stock predominated in the remaining stratum. Iterative estimates of Chesapeake contribution to the fishery exceeded SO'^ in all strata not adjacent to the Hudson River. Iterative estimates of the Hudson stock were largest in western Long Island Sound and the New York Bight with values ex- ceeding 20*7^ during some periods. Although itera- tive estimates of Roanoke stock contribution never exceeded 20^7^, they were highest in North Carolina waters (stratum 10) and in strata from Massachusetts to Maine (strata 1, 2). The Hudson stock contribution in strata from Massachusetts north to Maine and from New Jer- sey south to North Carolina (strata 8 to 10) should be low as indicated by iterative estimates (Table 5) and results of tagging studies. Zero estimates in northern waters do not necessarily indicate an absence of Hudson River striped bass since the simulation study has shown that such estimates may be obtained in situations where true con- tribution is low. In fact, data on adult striped bass tagged in the Hudson River during spawning sea- son and recaptured in waters as far north as Bos- ton Harbor, Mass., have indicated a northern mi- gration of a portion of the Hudson stock (Texas Instruments see footnote 4). However, these data support near-zero estimates of Hudson contribu- tion in southern waters since tagged striped bass were not recaptured south of northern New Jer- sey. Data (Chapoton and Sykes 1961) on adult striped bass tagged along the outer coast of North Carolina and recaptured on the spawning grounds of Chesapeake Bay and Albemarle Sound Table 5. — Estimates of relative contribution of Hudson, Chesapeake, and Roanoke stocks of legal-sized striped bass' to 1975 oceanic collections by period and spatial strata. As-cl. = As-classified, Iter. = Iterative, and Adj. = Adjusted estimates. stratum Sample size^ Hudson Chesapeake Roanoke Period As-cl. Iter. Adj. As-cl. iter. Adj. As-cl. Iter. Adj. Jan -Feb. 10 27 25.9 37 6.7 63.0 926 90.7 11.1 3.7 2.6 Mar. -Apr. 5 38 52.6 57.9 54.2 42.1 42.1 458 5.3 0.0 0.0 7 30 23.3 3.3 0.0 73-3 96.7 100,0 33 0.0 00 8 34 235 8.8 08 676 882 992 8.8 29 0,0 9 71 85 0.0 0.0 77,5 97.2 98.9 14.1 28 11 May-June 1 82 110 0.0 00 683 902 885 20.7 98 11,5 2 91 14.3 0-0 00 71 4 95,6 964 14.3 4.4 36 3 60 30.0 3.3 139 600 967 84 5 10,0 0.0 1,6 4 96 21 9 10 0,0 698 990 1000 83 0-0 0.0 5 14 357 286 23.0 57 1 71.4 770 7.1 0,0 0.0 6 89 258 56 5.4 652 933 94.6 9.0 1,1 0.0 7 58 41 4 259 33.7 51 7 707 663 6.9 3,4 0.0 8 113 23.9 0,0 1,5 673 100.0 985 8.8 0.0 0.0 July- Aug 1 58 19.0 00 0,0 67,2 948 95.4 13.8 52 4.6 2 90 7.8 00 0,0 722 96 7 908 20 0 3.3 9.2 3 43 302 2.3 10,3 65 1 977 897 4.7 0.0 0.0 5 15 26 7 0,0 5 1 66 7 100.0 94 9 6.7 0.0 0,0 6 102 225 7,8 1,6 63,7 88.2 927 13.7 3.9 5.7 7 93 333 15,1 13,4 65,6 84.9 86.6 1.1 0,0 0.0 8 28 21 4 0.0 0,0 71 4 100.0 100.0 7 1 0,0 0.0 Sept -Oct 1 74 13.5 0,0 0-0 77,0 98.6 1000 9.5 1,4 00 2 82 12.2 0,0 0,0 58,5 85.4 760 29.3 146 24.0 3 56 250 3,6 7,5 58,9 94.6 833 16.1 1,8 9.2 4 140 16.4 0,7 00 64.3 94.3 886 19.3 5,0 11.4 5 89 41,6 40,4 37,2 46.1 57.3 565 12.4 2.2 6.4 6 86 15.1 0,0 0,0 73.3 96.5 99 8 11.6 3.5 0.2 7 120 233 1,7 29 63.3 950 91.6 13.3 33 5.2 8 73 16.4 0,0 0,0 76.7 986 100.0 68 14 0.0 9 6 16.7 0,0 00 66.7 100,0 922 16.7 00 78 Nov -Dec. 4 99 21.2 00 00 667 980 969 12,1 2.0 3,1 6 106 16.0 00 0,0 698 99 1 959 14,2 0,9 4 1 7 124 21.0 4,8 0,0 76.6 95,2 100 0 24 0,0 00 8 117 21,4 0,0 0.0 72.6 100,0 1000 6.0 0,0 0.0 9 100 8.0 0,0 0.0 80.0 990 100.0 12.0 1,0 0.0 10 24 12.5 0.0 0.0 62.5 83.3 82.0 25.0 16.7 18.0 Overall mean 23.0 6.5 6.6 66.0 90.8 90.2 11.0 2.7 3.2 'Not included are striped bass <406.5 mm FL from New York waters. ^Sample sizes of five specimens or less in any stratum are not included. 342 BERGGREN and LIEBERMAN: RELATIVE CONTRIBUTION OF STRIPED BASS tributaries also support near-zero estimates of Hudson River contribution in waters off North Carolina. Comparison between iterative and adjusted es- timates indicated close agreement for each stock within the 35 strata. The largest difference be- tween estimates was 12.2 percentage points, but differences of <5 percentage points occurred in 809c of the strata for the Hudson stock, 71% of the strata for the Chesapeake stock, and 869r of the strata for the Roanoke stock. The adjusted esti- mates therefore substantiate low iterative esti- mates of contribution of Hudson and Roanoke stocks. Comparison of mean iterative and adjusted es- timates of relative contribution indicated that the two estimates differed by <1 percentage point for each stock. Mean iterative and adjusted estimates were, respectively, 6.5 and 6.6^^ Hudson, 90.8 and 90.2*7^ Chesapeake, and 2.7 and 3.2% Roanoke contribution. The contribution of the Hudson stock to the coastal fishery was greater in strata adjacent to the Hudson River than in the remaining strata. Mean iterative estimates of relative contribution of the Hudson River stock to inner and outer zones were 16.0% (15.0% adjusted) and 2.8% (0.0% ad- TabLE 6. — Mean estimates' of relative contribution of Hudson, Chesapeake, and Roanoke stocks of legal-sized striped bass^ to 1975 oceanic collections within- USNRC zones.^ Inner zone Outer zone Estimate Hudson Chesa- peake Roa- noke Hudson Chesa- peake Roa- noke As-classified Iterative Adjusted 31 7 160 15.0 629 83,1 84.2 55 09 0.8 19.2 2.8 0.0 68.0 94.2 96.4 12.8 3.0 3.6 'Average of five temporal strata since only one striped bass collected in inner zone during period 1 (Jan. -Feb. ). ^Not included are striped bass <406.5 mm FL from New York waters. ^U.S. Nuclear Regulatory Commission inner zone corresponds to study strata 5,7, and 8- 1 : the outer zone corresponds to study strata 1 to 4, 6. 8-2, and 8-3. justed), respectively, for the year (Table 6). Al- though the Chesapeake stock was the predomi- nant contributor to both inner and outer zones, the contribution of the Hudson stock exceeded that of the Roanoke stock in the inner zone but was less in the outer zone. The Hudson stock predominated in collections of sublegal-sized striped bass in western Long Island Sound, the New York Bight, and in collections of specimens overwintering in Croton Bay on the Hudson River (Table 7). Iterative (and adjusted) estimates of the percentage of sublegal-sized fish classified into the Hudson stock in western Long Island Sound (primarily in Little Neck Bay) and the New York Bight were at least 80% , but were less than 40% along the southeastern shore of Long Island (stratum 6) from May through Oc- tober. The iterative (and adjusted) estimated of contribution of the Hudson stock to the overwin- tering population in the Hudson River was greater than 95% . This study has provided additional information in the importance of dominant year classes of striped bass. Approximately 52% of the specimens collected from the coastal fishery in 1975 were from the 1970 year class, and 77% of them were classified as Chesapeake fish. Schaefer (1972) stated that production of young-of-the-year striped bass in Chesapeake Bay during 1970 was the largest ever recorded and that this year class should provide excellent fishing in New York wa- ters for 6 to 8 yr after recruitment. The presence of this dominant year class of Chesapeake fish confirms the rationale used by Merriman (1941) and Schaefer (1968) to conclude that the Chesapeake stock predominates in the coastal fishery. A summary of the occurrence of dominant year classes in the Atlantic coastal fishery has been given by Schaefer (1968). Table 7.— Estimates of relative contribution of Hudson, Chesapeake, and Roanoke stocks of sublegal-sized striped bass' to New York waters by period and spatial stratum and of legal-sized striped bass to the overwintering population in the Hudson River. As-cl. = as-classified, Inter, = iterative, and Adj. = adjusted estimates. Period Stratum Sample size^ Hudson Chesapeake Roanoke Population As-cl. Iter. Adj. As-cl. 23.7 Iter 2.6 Adj. 4.3 As-cl. 0.0 Iter. 0.0 Adj Ovenwintering 76 76.3 97.4 95.7 0.0 Sublegal IVIay-June 5 42 92.9 100.0 100.0 7,1 0.0 0.0 0.0 0.0 0.0 6 8 12.5 0.0 00 500 62 5 64.3 37.5 37,5 3b / 7 11 81 8 81.8 1000 18,2 182 00 0.0 0.0 0,0 July-Aug. 5 85 88.2 100.0 100.0 11 8 00 00 0.0 0.0 0,0 6 17 41.2 35.3 392 41.2 588 47.4 17.6 5.9 13,4 Sept.-Oct. 5 10 80.0 80.0 100.0 20.0 20.0 0,0 0.0 0.0 0,0 6 19 26.3 15.8 20.8 36.8 47.4 41.9 36.8 36.8 3/.3 'Striped bass <406.5 mm FL from New York waters. ^Sample sizes of five specimens or less in any stratum are not included. Three sublegal-sized specimens collected ovenwintering in the Hudson River were classified as Hudson fish. 343 FISHERY BULLETIN: VOL, 76. NO. 2 SUMMARY AND CONCLUSIONS A study was conducted to identify the origin of striped bass collected in the Atlantic coastal fishery and estimate the relative contribution of major stocks to the fishery. Quadratic discrimi- nant analysis was applied to values of five mor- phological characters obtained from Hudson, Chesapeake, and Roanoke spawning-stock speci- mens to determine functions which best separated the stocks. Correct-classification percentages of 76.8, 67.7, 85.9'7f were obtained for the Hudson, Chesapeake, and Roanoke spawning stocks, re- spectively, resulting in an overall correct clas- sification of 74. 47c of the specimens. A simulation study was conducted to investi- gate the bias in as-classified, iterative, and ad- justed estimates of relative contribution due to misclassification error inherent in the discrimi- nant functions. Results indicated that iterative estimates may best approximate the true con- tribution of the Hudson stock in oceanic collec- tions. A stratified sampling design was used during six 2-mo periods in 1975 to collect representative samples of striped bass in the Atlantic coastal fishery from southern Maine to Cape Hatteras. This provided estimates of stock composition by stratum throughout the year. Oceanic samples were classified by discriminant functions and as-classified, iterative, and revised estimates of relative contribution of the major stocks were obtained. Mean iterative estimates of relative contribution for 1975 are 6.59c Hudson, 90.87c Chesapeake, and 2.7% Roanoke stocks. Iterative estimates of Hudson contribution for legal-sized striped bass exceeded 207c only in western Long Island Sound and the New York Bight during certain months. In collections from Western Long Island Sound and the New York Bight, iterative estimates of the percentage of sublegal-sized fish classified into the Hudson stock were at least 807c during the May through October periods. For Hudson River collections of overwin- tering striped bass, an iterative estimate of 97.4% Hudson stock was obtained. The occurrence of a dominant year class was noted. Approximately 52% of the legal-sized specimens collected in the 1975 oceanic sampling program were from the 1970 year class, and 77% of these were classified as Chesapeake in origin. Major conclusions drawn from the study are: 1) the Chesapeake stock is the major contributor to the Atlantic coastal striped bass fishery from southern Maine to Cape Hatteras; 2) the Chesapeake stock is also the major contributor of legal-sized striped bass in the vicinity of the Hud- son River (western Long Island Sound and the New York Bight); 3) sublegal-sized striped bass collected in the vicinity of western Long Island Sound and the New York Bight are predominantly of Hudson origin; and 4l striped bass overwinter- ing in the Hudson River are predominantly of Hudson origin. ACKNOWLEDGMENTS We acknowledge Thurman L. Grove who in- itiated the study and was instrumental in its suc- cessful completion, and George A. Roth who helped gather and process the data and aided in report writing. We also thank Eddie Baldocchi, Dana Grass, Michael Locke, Edwin Manter, Ronald McGratten, Thomas Orvosh, Martin Ot- ter, and Donald Strout for their help in gathering and processing the data; John Bennett and Leanna Pristash for their work in computer pro- gramming; and Dennis DuBose for his help with the Schuermann and Curry methodology. This study was carried out under contract to Consoli- dated Edison Company of New York, Inc., as part of the Hudson River Ecological Survey. LITERATURE CITED ALPERIN, I. M. 1966. Dispersal, migration and origins of striped bass from Great South Bay, Long Island. N.Y. Fish Game J. 13:79-112. AMOS. M. H., R. E. ANAS. AND R. E. PEARSON 1963. Use of a discriminant function in the morphological separation of Asian and North American races of pink salmon, Oncorhynchus gorbuscha (Walbaum). Int. North Pac. Fish. Comm., Bull. 11:73-100. ANAS. R. E., AND S. MURAL 1969. Use of scale characters and a discriminant function for classifying sockey salmon (Oncorhynchus nerka) by continent of origin. Int. North Pac. Fish. Comm., Bull. 26:157-192. ANDERSON, T. W. 1958. An introduction to multivariate statistical anal- ysis. John Wiley & Sons, N.Y., 374 p. CHAPOTON. R. B., AND J. E. SYKES 1961. Atlantic coast migration of large striped bass as evidenced by fisheries and tagging. Trans. Am. Fish. Soc. 90:13-20. CHITTENDEN, M. E., jR. 1971. Status of the striped bass, Morone saxatilis, in the Delaware River. Chesapeake Sci. 12:131-136. 344 BERGGREN and LIEBERMAN: RELATIVE CONTRIBUTION OF STRIPED BASS FUKUHARA, F. M., S. MURAI. J. J. LALANNE, AND A. SRIBHIBHADH 1962. Continental origin of red salmon as determined from morphological characters. Int. North Pac. Fish. Comm., Bull. 8:15-109. Grove, T. L., T. J. Berggren, and D. A. Powers 1976. The use of innate tags to segregate spawning stocks of striped bass, Morone saxatilis. In M. Wiley (editor), Estuarine processes. Vol. 1, p. 166-176. Academic Press, N.Y. HUBBS, C. L., AND K. F. LAGLER. 1958. Fishes of the Great Lakes region. Revised ed. Cranbrook Inst. Sci., Bull. 26, 213 p. KENDALL, M. G., AND A. STUART. 1968. The advanced theory of statistics. Vol. 3. Hafner Publ. Co., N.Y., 557 p. KOO, T. S. Y. 1970. The striped bass fishery in the Atlantic States. Chesapeake Sci. 11:73-93. MANSUETI, R. 1961. Age, growth, and movements of the striped bass, Roccus saxatilis, taken in size selective fishing gear in Maryland. Chesapeake Sci. 2:9-36. MERRIMAN, D. 1941. Studies on the striped bass, Roccus saxatilis, of the Atlantic coast. U.S. Fish Wildl. Serv., Fish. Bull. 50:1- 77. Messieh, S. N. 1975. Delineating spring and autumn herring populations in the southern Gulf of St. Lawrence by discriminant function analysis. J. Fish. Res. Board Can. 32:471-477. MOSTELLER, F., AND J. W. TUKEY. 1968. Data analysis, including statistics. In G. Lindzey and E. Aronson (editors). The handbook of social psychol- ogy, 2d ed., p. 80-203. Addison- Wesley, Reading, Mass. NICHOLS, P. R., AND R. V. MILLER. 1967. Seasonal movements of striped bass, Roccus saxatilis (Walbaum), tagged and released in the Potomac River, Maryland, 1959-61. Chesapeake Sci. 8:102-124. Raney, E. C, and D. p. DE Sylva 1953. Racial investigations of the striped bass, Roccus saxatilis (Walbaum). J.Wildl. Manage. 17:495-509. Raney, E. C, W. S. Woolcott, and a. G. Mehring. 1954. Migratory pattern and racial structure of Atlantic coast striped bass. Trans. 19th North Am. Wildl. Nat. Resour. Conf , p. 376-396. SCHAEFER, R. H. 1968. Size, age composition emd migration of striped bass from the surf waters of Long Island. N.Y. Fish Game J. 15:1-51. 1972. Striped bass. Conservationist 27:27, 46. Vladykov, V. D., AND D. H. Wallace. 1952. Studies of the striped bass, Roccus saxatilis (Wal- baum), with special reference to the Chesapeake Bay re- gion during 1936-1938. Bull. Bingham Oceanogr. Col- lect 14:132-177. WORLUND, D. D., AND R. A. FREDIN. 1962. Differentiation of stocks. In Symposium on pink salmon, p. 143-153. H. R. MacMillan Lectures in Fisheries, Univ. B.C., Vancouver, Can. 345 COPPER SENSITIVITY OF PACIFIC HERRING, CLUPEA HARENGUS PALLASl, DURING ITS EARLY LIFE HISTORY D. W. Rice, Jr and F. L. Harrison' -i ABSTRACT Embryos and larvae of the Pacific herring, Clupea harengus pallasi, were exposed to copper, using a flow-through bioassay system. Herring embryos were exposed continuously from 12 h after fertiliza- tion until hatching, and larvae were exposed from the time of hatching until yolk sac absorption. Embryos were also exposed to 36-h duration pulses of copper in order to evaluate the sensitivy of different developmental stages of herring embryos to copper. Pulsed exposures started at 62, 98, or 136 h after fertilization. The following measurements were taken as indices of the toxic effects of copper: cumulative mortality, percent hatching, and larval length upon hatching. The onset of mortality of herring embryos continuously exposed to copper began 90 h after fertiliza- tion, with deaths occuring over a short interval thereafter (response period). Significant embryo mortalities occurred at a copper concentration as low as 35 /xg/1. Herring larvae continuously exposed to copper showed significant mortality at 300 ju.g/1 copper, with no delay in the onset of mortality. Embryos exposed to 36-h pulses of copper during different developmental stages showed reduced sensitivity when exposed after the response period. Larvae that hatched from eggs exposed to a 36-h pulse of copper before the response period grew significantly less than those hatched from eggs exposed during later developmental stages. Numerous studies have shown that many aquatic animals are adversely affected by increased levels of copper in water; most of the work on fishes has been restricted to freshwater species (Becker and Thatcher 1973; Brungs et al. 1976). Since 90^f of the world's marine fish are taken from the conti- nental shelf and nearshore upwelling areas (Wal- dichuk 1974), increases in copper pollution in coastal aquatic ecosystems are of particular con- cern. The concentration of copper in unpolluted near- shore waters ranges from 0.3 to 3.8 /Lig/1 (Chester and Stoner 1974). Increased concentrations of cop- per in coastal waters have resulted from the re- lease of municipal waste waters (Mytelka et al. 1973; Mitchell and McDermott 1975) and of effluents from power plants ( Hoss et al. 1975; Mar- tin et al. 1977). In polluted waters, concentrations as high as 13,900 /xg/l copper have been reported (Mitchell and McDermott 1975). Examination of the toxic effects of copper on coastal marine fisheries is important for the estab- lishment of water quality standards that will pro- tect fishery resources of coastal zones. Eggs and 'Environmental Sciences Division, Lawrence Livermore Laboratory, University of California, Livermore, CA 94550. Manuscript accepted September 1977. FISHERY BULLETIN; VOL. 76, NO. 2, 1978. larval stages of fish are reported to be the life history stages that are most sensitive to a variety of pollutants (Skidmore 1965; Pickering and Gast 1972; Struhsaker et al. 1974; Christensen 1975). The necessity of conducting toxicity tests during the most susceptible stage in the life history of an organism has been emphasized by Hynes (1970), and the sensitivity of vertebrate embryos to heavy metals has been suggested as a criterion for water quality by Birge and Just (1973). While some work has been done to assess the toxicity of copper to the early life history stages of freshwater fishes (Mount 1968; Hazel and Meith 1970; McKim and Benoit 1971; O'Rear 1972; Gardner and La Roche 1973; Benoit 1975), little assessment has been made of toxic effects of copper on marine fishes. Such studies should be very im- portant since mortalities that occur during the early life history stages of marine fish strongly influence the strength of a given year class offish (May 1974; Bannister et al. 1974; Postuma and Zijlstra 1974; Gushing 1975; Vaughan and Saila 1976). Toxic effects that have an impact upon sur- vival during early developmental stages would also act to reduce the strength of a given year class offish. The embryos and larvae of the Pacific her- ring, Clupea harengus pallasi, represent a useful 347 FISHERY BULLETIN: VOL. 76. NO 2 test organism for evaluating the toxic effect of copper upon the early life history stages of marine fish. The Pacific herring is a commercially impor- tant fish that spawns along both eastern and w^est- ern Pacific coasts (Eldridge and Kail 1973; Hart 1973). Herring spawn great numbers of demersal, adhesive eggs on shallow intertidal substrates. The egg is relatively large, 1.3 to 1.6 mm in diame- ter, and is covered by a thick, three-layered, opaque chorion (Blaxter and Holliday 1963). De- velopment of the embryo is comparatively slow, taking 7 to 9 days at 14 °C (Alderdice and Velsen 1971). The tough chorion permits easy collection and handling, and the slow development of the embryo allows observation time not available in more rapidly developing species. Three bioassays were conducted to evaluate the sensitivity of herring embryos and larvae to cop- per. The first two assays were designed to evaluate the sensitivity of embryos and larvae to continu- ous copper exposure, while the third examined the sensitivity of embryos to brief copper exposures. Since the form of copper to which the herring em- bryos and larvae were exposed may play a sig- nificant role in the toxic response ( Pagenkopf et al. 1974), the partitioning of copper among the com- ponents of the water in the bioassay system was also determined. MATERIALS AND METHODS Collection and Handling of Test Organisms Intertidal collections of Pacific herring eggs were made along the shore of Belvedere Island and the Tiburon Peninsula, San Francisco Bay, Calif. The eggs were collected directly into a 15-gal, in- sulated ice chest containing aerated seawater from the egg collection site and were transported to the laboratory within 2 h after collection. The water temperature at the collection site was 11.0°-11.5°C and upon arrival at the laboratory the temperature of the water increased to no more than 13.5°C. Only eggs deposited in single layers onFucus si).,Laminaria sp., or Gracilaria sp. were chosen for testing. Before placing the eggs into exposure chambers, they were removed from the seaweed by bending the frond and then gently brushing them with a finger into a sorting dish containing seawater kept at 12°C. The eggs were examined with a microscope at 20 x; only viable embryos at the same stage of development were chosen. No more than 51 embryos were placed in any exposure chamber. All transfers of embryos or larvae were carried out with a large-bore, polished glass pipette. Embryos at two different stages of development were used for the tests. The age of the earlier stage embryos, collected 7 February 1975, was esti- mated to be 12 h after fertilization since they were undergoing epiboly (Ahlstrom's stage IV ( Ahlstrom 1943)). These embryos were exposed to copper continuously, each of seven groups being exposed to a different copper concentration. The age of the later stage embryos, collected 26 Feb- ruary 1975, was estimated to be 48 to 50 h after fertilization (Ahlstrom's stage IX). These embryos were divided into four groups. Three groups were exposed for 36 h to the same copper concentration but during different developmental stages. The fourth group was maintained in flowing seawater from the time of collection to within 1 h after hatching, and then continuously exposed to differ- ent copper concentrations as larvae. Bioassay System The organisms were exposed to copper in 5-1 clear plastic bowls (Figure 1). The exposure solu- tion was introduced into each chamber by gravity flow from a mixing chamber into which seawater, at a rate of 11 ml/min, and copper chloride solu- tion, pH 3, at a rate of 1 ml/h, were pumped con- tinuously. Approximately 17 h was required for replacement of 90'7f of the water in the chamber. The height of water in the exposure chambers was maintained by a constant-level out-flow siphon. The diameter of the mouth of the out-flow siphon Seawate Copper Plastic mixing chamber Outflow siphon Clear plastic cover Clear plastic o o ^exposure chamber °^ Figure 1 . — Diagram of the exposure chamber and flow- through dehvery system used to expose Pacific herring embryos and larvae to copper. 348 RICE and HARRISON: COPPER SENSITIVITY OF PACIFIC HERRING was greater than that of the tubing to reduce the flow velocity at the mouth of the siphon. The mouth of the siphon was covered with nylon net- ting (505-)U,m pore size) to prevent the loss of or- ganisms from the chamber. A gentle stream of bubbles delivered to the bottom of the chamber provided aeration and mixing. All exposure chambers were immersed in a water bath whose temperature was monitored. Illumination was provided by the fluorescent lighting in the laboratory and followed the regular ambient photoperiod. The exposure period, the nominal copper con- centration, and the total number of embryos or larvae exposed during a typical experiment are given in Table 1 . Each experiment was repeated at least once. All exposures were initiated by the addition of appropriate amounts of copper chloride to the chambers. The continuous exposure of the test organisms continued until all animals died or, in the case of embryos, until hatching occurred or, in the case of the larvae, until yolk sac absorption occurred. The pulsed exposures were terminated by transferring exposed embryos to an exposure chamber containing control seawater. The following measurements were taken as in- dices of the toxic effect of copper: cumulative mor- tality with time, percent hatching, and larval length at hatching. The embryos or larvae were examined within the exposure chambers at each observation period with a 7 x beam dissection mi- croscope with a 21-cm depth of field. The criterion for embryo death was the lack of heart beat or body movement. Since the embryos were attached in clusters, dead embryos were not removed until the termination of a test. Larvae that hatched from pulse exposed embryos were collected, anes- thesized with a 1% quinaldine solution, and pre- served in 5% Formalin^ in seawater. Measure- ments of the hatched larvae were made with an ocular micrometer and all obvious deformities noted. The criterion for larval death was a failure to respond to a gentle prod with a polished glass rod. Dead larvae were removed during each obser- vation period and preserved in 5% Formalin in seawater. Total copper concentrations were measured two ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table l. — Experimental conditions and median lethal times for bioassays determining the sensitivity of Pacific herring embryos and larvae to copper. Exposure Nominal copper Number of Mean total copper Number of Time to median lethal period concentration organisms concentrations water level (LTso ± 95% confi- Experiment (h) (mq/i) exposed (fxgn ± SD) samples dence interval) (h) Embryos, continuous i 180 Control 150 4.3 ± 1.9 3 (') exposure (12 h atter • 25 48 27 9 ± 7.4 3 V) fertilization 35 49 38.1 ± 9.4 3 2144.9 ± 8.3 through 45 50 44.1 ± 11.6 3 2134.4 ± 5.2 hatching) 55 49 51.2 ± 9.7 3 2134.8 ± 3.0 100 53 127,9 ± 34.6 2 2115.4 ± 2.8 200 49 235.5 ± 47,2 2 298.7 ± 2.1 Larvae, continuous 300 Control 100 2.5 ± 0.8 3 D exposure (Hatching 300 49 274.0 ± 24.1 4 (1,3) through 600 49 572.1 ± 31.5 4 (1,2) yolk sac 1,400 50 1,349.0 ± 247.0 4 241.7 ± 7.3 absorption) 2,000 51 1,969 0 ± 148.0 2 223.8 ± 1.5 2,500 49 2,425.4 ± 89.0 2 220.9 ± 2.4 3,500 51 3,430.5 ± 710.6 2 "15.6 Time to median lethality following termination of pulse ± 95% confidence interval (h) Embryos, pulsed exposures; Pulse 1 36 Control 49 3.0 ± 0.08 2 (') (62 through 98 100 44 93,8 ± 6.3 3 238.4 ± 2.0 h after fertilization) Pulse II 38 Control 93 3.9 ± 1.6 3 D (98 through 136 100 94 111.9 ± 13.7 4 253.9 ± 6.1 h after fertilization) Pulse III 36 Control 46 6,1 ± 0.7 3 0 (136 through 100 48 101.6 ±29.3 3 (') 172 h after fertilization) '50% mortality not achieved at this concentration. 2Slope significantly different from control slope (P<0.01) (Snedecor and Cochran 1967). ^Slope significantly different from control slope (P<0.05) (Snedecor and Cochran 1967). •"Determined according to the method of Litchfield and Wilcoxon (1949). 349 FISHERY BULLETIN: VOL. 76. NO. 2 to four times during each bioassay to determine the actual concentrations to which the organisms were exposed. Water samples were collected in acid-washed polyethylene jars and acidified to pH 2 with concentrated HCl. Total copper was analysed by the APDC-DDDC-MIBK extraction method described by Kinrade and VanLoon (1974). The copper concentration in extracted MIBK solutions was determined with a model 303 Perkin Elmer atomic absorption spectrophotom- eter, using an HGA-2100 graphite furnace with a deuterium background corrector. Since the chemistry of copper in seawater is complex, more than one form of copper may be present in the bioassay water. To examine the form of the copper in the bioassay system water, out-flow samples were collected from the bioassay system before organisms were introduced to de- termine the particulate-bound fractions (>0.45^im), ionic fraction (bound by Chelex-100 resin (Riley and Taylor 1968)), and complexed fraction (not bound by Chelex-100 resin). The analysis scheme is summarized in Figure 2. To monitor the partitioning of copper into each of these fractions, copper-64 was equilibrated with water samples after they were withdrawn from the bioassay system. The partitioning of stable copper in the seawater of the bioassay system was indicated by the percentage of the initial activity recovered in each of the described fractions. Statistical Analysis The measures of toxicity determined in this study were the time to 50^f mortality at each con-'^ centration of copper tested (median lethal time, LT50) and the concentration of copper resulting in 50% mortality over a given time (median lethal concentration, LC50). These toxicity measures were deterimed by performing a weighted linear regression analysis on the sets of cumulative mor- tality data using the logistic function. The straight line transform of the logistic function is: logit P = In PIQ =a + (3x, so that if logit P is plot- ted against X, the points will fall on a straight line with a as the intercept and /3 as the slope (Berkson 1953). In our calculations of LT50, x represented the time from the onset of the reaction period in the case of continuous embryo exposures, from hatching in the case of continuous larval expo- sures, and from the termination of a given pulse in the case of pulsed embryo exposures. In our calcu- lations of LC50, X represented concentration, 350 Seawater APDC-DDDC-MIBK extraction 0.45 u filter Total fraction Chelex 100 ion exchange resin APDC-DDDC-MIBK extraction on waste water Particulate bound fraction Ionic fraction _ Complexed fraction Figure 2. — Analysis scheme for the separation of copper frac- tions recovered from the bioEissay system used to expose Pacific herring embryos and larvae to copper. and our method followed that outlined by the American Public Health Association (1976) with logit analysis used in place of probit analysis. A computer was used to calculate the LC50 and LT50 values, and for each fitted line the program determined: the LT50 or LC50, the 959^ confidence limits associated with the LT50 or LC50, Pearson's rho (p), the slope (/3) and the intercept (a), and the mean square error (EMS); no assumptions of homogeneity were made and the EMS was calcu- lated in every case, rather than assuming an EMS of 1 for homogeneous data (Finney 1964). In the case of embryos that were exposed con- tinuously or exposed to pulses of copper, deaths prior to the delayed reaction period or the onset of the pulsed exposure, respectively, were not used in the data analysis. In no case were mortalities dur- ing these periods greater than 6%. The relationship between time to 50% mortality during continuous exposure of both embryos and larvae and concentration was determined follow- ing the method outlined in the American Public Health Association (1976). The resulting toxicity curve was used to estimate the lethal threshold concentration (incipient LCgo) (Sprague 1969). RESULTS Physical Parameters of the Bioassay System Mean copper concentrations measured during each test are reported in Table 1. The partitioning RICE and HARRISON. COPPER SENSITIVITY OF PACIFIC HERRING of copper-64 among particulate-bound, ionic, and complexed fractions of copper recovered from the bioassay water indicates that the copper was primarily in the ionic form {Table 2). The mean pH of the water in exposure chambers in all tests was 8.08 (SD = ±0.024). The mean temperature for all tests was 13.3°C (SD = ±0.8°C). Table 2. — Percentages of copper-64 in fractions of sea water recovered from the bioassay system used to expose Pacific her- ring embryos and larvae to copper. Nominal copper concentration (^g/i) Particulaie bound Ionic Complexed Total' 10 5.2 83.6 4.3 93.7 50 4.2 88.3 3.7 96.5 100 1.1 89.9 2.1 940 500 0.7 938 1,7 970 1.000 0.6 91,8 1.1 969 2,000 1.0 96.4 1.0 99.9 'Total Includes copper-64 remaining in Chelex-100 resin after elution. Continuous Exposures to Copper Survival of embryos continuously exposed to copper was high at all concentrations of copper tested until 90 h of exposure, at which time dose- related mortalities occurred (Figure 3). The period during continuous exposure when embryo deaths begin is termed the reaction period. Mortalities of developing embryos at copper concentrations of 35 /ug/l and higher were significantly different from controls (P<0.01). Virtually no hatching occurred at concentrations above 45/Ltg/l copper. Develop- mental features observed in the control embryos during the onset of the reaction period included the appearance of eye pigmentation, the onset of coordinated body movements, and the initiation of heart beat. Embryos continuously exposed to cop- per concentrations of 100 /i.g/1 copper and higher developed an opaque cast to the chorion, which was followed later by whitish discoloration of the body. Embryos continuously exposed to 200 /u.g/1 copper developed an opaque change in the chorion at 60-72 h from fertilization, with body discolora- tion, spasmodic contractions, quiverings, and re- duced fin fold development occurring at 84-96 h from fertilization. Herring larvae continuously exposed to copper were many times less sensitive to copper than herring embryos. Larval mortalities differed sig- nificantly from controls at concentrations of 300 /Ltg/l copper and higher (P<0.05) (Figure 4). Prior to death the larvae sank to the bottom of the expo- sure chamber and patches of whitish discoloration were observed over the bodies. Spasmodic quiver- ing and whole body contractions were observed in larvae at concentrations of 1,400 /xg/1 copper and above. The toxicity curves for continuously exposed herring embryos and larvae are shown in Figure 5. Median lethal times for each copper concentration tested and 95% confidence limits are detailed in Table 1. 100 Figure 3. — Percent cumulative mor- tality of Pacific herring embryos con- tinuously exposed to copper (micro- grams per liter). Mortality curves shown are the fitted logit curves used to establish median lethal times. +-> C 0) u s. 4J o > Z3 -H Hatching 0 20 40 60 80 100 120 140 160 180 Hours from fertilization 351 FISHERY BULLETIN: VOL. 76, NO. 2 FIGURE 4. — Percent cumulative mor- tality of newly hatched Pacific herring larvae continuously exposed to copper (micrograms per liter). Mortality curves shown are the fitted logit curves used to establish median lethal times. 4-> c Q) O S- 0) Q. ■4-> S- o > +-> o ^■a n I n I D u °i — \j^ <0^^=^^ Control J \ \ I 0 30 60 90 120 150 180 210 240 270 300 _J Yolk sac L absorbtion ^ Hours from hatching o -M S- o o o E 240 192 144k 96 24 10 T — r ^--1- Hatching completed -^- ■1- Initiation of yolksac absorption EMBRYOS A=LC5o A = LT50 ± 95% confidence limits LARVAE D = LC50 ■ = LT50 ^ 95% confidence limits J L 10 20 30 50 1000 100 300 Copper - pg/£ Figure 5. — Toxicity curves for Pacific herring embryos and larvae continuously exposed to copper. 3000 352 RICE and HARRISON: COPPER SENSITIVITY OF PACIFIC HERRING The toxicity curve for herring embryos is pre- sented for the purpose of discussion only, since the 90 h delay until the onset of mortality, regardless of concentration, biases the toxicity curve for com- parison with other organisms without a reaction period. Sprague (1969) recommended that a con- centration that killed SO^r of the population dur- ing an exposure sufficiently long that acute lethal action has ceased (incipient LC50) be used as the single most useful criterion for toxicity. The inci- pient LC50 is not influenced by the bias introduced by the reaction period. The estimated incipient lethal level for herring embryos was found to be 33 /xg/1 copper. Only larval deaths earlier than 100 h after hatching were considered in the construction of the larval toxicity curve since larvae surviving beyond approximately 200 h after hatching have begun yolk sac absorption, and the apparently synergistic effects of copper stress and starvation can be observed in the larval time vs. percent mortality curves (Figure 4). The estimated incip- ient lethal level for herring larvae was found to be 900 /xg/1 copper. Thirty-six Hour Pulsed Embryo Exposures Pulses of copper exposure for 36 h showed that the sensitivity of herring embryos to copper 100 changed as the embryos developed (Figure 6, Table 1). A 36-h pulse of 100 /xg/l copper delivered during the reaction period (Pulse I) had the greatest effect upon hatching and the length of larvae at hatching (Table 3). A 36-h pulse of 100 fxg/\ copper delivered just before hatching (Pulse III) had a signficant effect on larval length at hatching, but the percentage of embryos hatching was actually greater than controls. Table 3. — Percent Pacific herring embryos hatching and mean larval length at hatching for three groups of Pacific herring embryos exposed to 36-h pulses of 100 /ng/1 copper. Each group received a pulse at a different time during development. Item Mean larval length (mm ± SD) Percent hatching Controls 6.10 ± 0.47 92 Pulse 1 3.77 ± 0.2' 6 Pulse 2 4.23 ± 0.31* 47 Pulse 3 5.75 ± 0.62- 98 •Significantly different from controls (P<0.01) (Snedecor and Cochran 1967). DISCUSSION Several features of the toxic response of herring at various stages of their early life history are of interest. Previous tests examining the sensitivity of other fish embryos and larvae to copper have found that the larval stage is the more sensitive stage (Hazel and Meith 1970; McKim and Benoit 1971, O'Rear 1972; Gardner and La Roche 1973; c a; o 0) Q. 80 - > 60 - (0 •M L. o E > JO Z3 E D o 40 20 - 0 1 1 1 /I A>r* // Continuous—^ f W^- Pulse! J 1 1 /■ / / ■/ / A / ■^M / / Pulse 11-^ 1 / ^Pulse III/ — / \ yf- Control / / x^ i n ^-t-r"^ / / ^^'''^v Uj-*"^"^'^^ 1 1 y ^^..^m^S-^^^m^ 40 80 120 160 Hours from fertilization 200 Figure 6. — Percent cumulative mor- tality of three groups of Pacific herring embryos exposed to 36-h pulses of 100 /j.g/1 copper. Each group received a pulse at a different time during development. The cumulative mortality observed for Pacific herring embryos continuously exposed to 100 /xg/l copper (See Figure 3) is shown for comparison. Pulse I Pulse Pulse 353 FISHERY BULLETIN: VOL. 76, NO. 2 Benoit 1975). Contrary to these findings, we found that embryos of the Pacific herring appear to be the stage that is more sensitive to copper. It should be noted that the fishes examined in previous studies spawn in fresh or brackish waters and cannot be considered true marine species as is the herring. Another interesting feature of the toxic re- sponse of the herring embryos and larvae was that the behavior prior to death was similar to that of adult fish exposed to copper. Jerky, uncoordinated, and spontaneous movements were noted by Baker (19691 in the winter founder, Pseudopleuronectes americanus, acutely exposed to 3,200 and 1,000 juig/1 copper. Bluegill, Lepom/s macrochirus, chronically exposed to 162 /u,g/l copper showed periodic involuntary spasms several weeks prior to death (Benoit 1975). The spasmodic contrac- tions and quiverings noted in herring embryos and larvae prior to death might be of a similar nature. Baker noted that these symptoms are similar to those of Wilson's disease which also manifests spasmodic muscle contractions and quiverings in mammals. Wilson's disease is the result of an in- born error of metabolism that results in an excess of unbound copper in the blood stream (Adelstein and Vallee 1962). Goldfish, Carassius auratus, subjected to doses of 1,000 /xg/1 copper exhibit se- vere neurotoxic symptoms and accumulate copper in nervous tissues at levels similar to those seen in Wilson's disease (Vogel 1959). Some of the toxic effects observed in herring embryos and larvae were similar to those reported for other heavy metals. Striped bass, Morone saxatilis, embryos exposed to copper or zinc (O'Rear 1972) and Baltic needlefish, Belone bel- one, exposed to cadmium (Dethlefsen et al. 1975) developed opaque discoloration of the chorion dur- ing exposure. In the present study, the chorion of the herring embryos became increasingly opaque as exposure to copper continued. Wedemyer (1968) found that in coho salmon, Oncorhynchus kisutch, 70% of the total zinc-65 uptake during exposure was firmly bound to the chorion, 26% was bound in the perivitelline space, and only 2% reached the yolk and 1% reached the embryo. Wedemyer (1968) also demonstrated that copper is bound by the salmon embryo's chorion. The opaque discol- oration noted in herring embryos with continued exposure to copper may well be a reaction result- ing from copper uptake by the chorion. The observation of a reaction period during bioassays with herring embryos has been noted previously. A reaction period for herring embryos continuously exposed to cadmium (Rosenthal and Sperling 1974) and high temperatures and salinities (Alderdice and Velsen 1971) occurred at about the time of the onset of heart beat. The sensitivity of this developmental period in the herring was further borne out by our findings in which 36-h pulses of 100 /xg/1 copper during the reaction period caused higher mortalities than 36-h pulses during later developmental periods. Pacific herring embryos may be vulnerable to toxic effects from effluents now being discharged into coastal environments. A survey of 108 muni- cipal waste effluents on the Atlantic coast showed that 50% of the waste effluents contained >100 H.gl\ copper; some discharges were as high as 5,900 /Ltg/1 copper (Mytelka et al. 1973). A survey of six municipal waste discharges along the southern California coast revealed concentrations ranging from 74 to 13,900 ^xgl\ copper with an average annual mass emission rate of 532 t of copper dur- ing 1971-74 (Mitchell and McDermott 1975). While the amount of copper discharged in the ionic form was not reported, the potential for environ- mental exposure levels approaching the incipient LC50 of 33 /x.g/1 copper found for herring embryos in the present study should be considered in estab- lishing water pollution control standards. Frequently authors conducting bioassays using copper or other heavy metals have not examined the chemical state of the metal in their bioassay system. Such characterizations are important since different chemical forms of metals may have different toxic effects (Lee 1973). The method out- lined in this work for examining the particulate bound fraction, the ionic fraction, and the com- plexed fraction of metals in seawater provides a means of examining the important chemical forms of copper in aquatic bioassay systems. With the use of appropriate isotopes this method could eas- ily be applied to other metals. In the case of the system used to expose Pacific herring embryos and larvae it appears that the ionic form of copper predominated. In freshwater the ionic form of cop- per seems to be the most toxic (Pagenkopf et al. 1974). This is probably also the case for Pacific herring embryos and larvae exposed to copper in seawater. ACKNOWLEDGMENTS We express our appreciation to Richard E. Tul- lis, California State University, Hayward, for re- 354 RICE and HARRISON: COPPER SENSITIVITY OF PACIFIC HERRING viewing the manuscript and for his helpful criti- cisms and suggestions throughout this study. The statistical program used was prepared by John Koshiver, University of California, Lawrence Livermore Laboratory, and we thank him. We also acknowledge the helpful comments regarding the handling of herring embryos and larvae from Maxwell E. Eldridge, Southwest Fisheries Center Tiburon Laboratory, National Marine Fisheries Service, NOAA. We especially thank John Kriegsman, Larry Schramm, and Bob Fountain of the California Department of Fish and Game for their timely information on the location of herring spawn and for their aid in the collection of these eggs. This work was performed under the auspices of the U.S. Energy Research and Development Administration Contract No. W-7405-ENG-48. LITERATURE CITED ADELSTEIN, S. J., AND B. L. VALLEE. 1962. Copper. In C. L. Comar and F. Bronner (editors). Mineral metabolism, an advanced treatise, Vol. II (B), p. 371-401. Academic Press, N.Y. AHLSTROM, E. H. 1943. Appendix. In Studies on the Pacific pilchard or sardine (Sardinops caerulea). 4. — Influence of tempera- ture on the rate of development of pilchard eggs in nature, p. 9-12. U.S. Fish Wildl. Serv., Spec. Sci. Rep. 23. ALDERDICE, D. F., AND F. P. J. VELSEN. 1971. Some effects of salinity and temperature on early development of Pacific herring ( C lupea pallasi). J. Fish. Res. Board Can. 28:1545-1562. American Public Health association. 1976. Standard methods for the examination of water and wastewater. 14th ed. Am. Public Health Assoc, Am. Water Works Assoc., Water PoUut. Control Fed., Wash., D.C., 1193 p. Baker, J. T. P. 1969. Histological and electron microscopical observa- tions on copper poisoning in the winter flounder (Pseudo- pleuronectes americanus) . J. Fish. Res. Board Can. 26:2785-2793. Bannister, R. C. a., D. Harding, and S. J. Lockwood. 1974. Larval mortality and subsequent year-class strength in the plaice ( Pleuronectes platessa L.). In J. H. S. Blaxter (editor), The early life history offish, p. 21-39. Springer- Verlag, N.Y. Becker, C. D., and T. O. Thatcher. 1973. Toxicity of power plants chemicals to aquatic life. Publ. WASH-1249, U.S. At. Energy Comm., 222 p. benoit, d. a. 1975. Chronic effects of copper on survival, growth, and reproduction of the bluegill (Lepomis macro- chirus). Trans. Am. Fish. Soc. 104:353-358. Berkson, J. 1953. A statistically precise emd relatively simple method of estimating the bio-assay with quemtal response, based on the logistic fimction. J. Am. Stat. Assoc. 48:565-599. Birge, w. J., AND J. J. Just. 1 973 . Sensitivity of vertebrate embryos to heavy metals as a criterion of water quality. Ky . Water Resour. Res. Rep. 61, 20 p. Blaxter, J. H. S., and F. G. T. Holliday. 1963. The behaviour and physiology of herring and other clupeids. Adv. Mar. Biol. 1:261-393. Brungs, W. a., J. R. Geckler, and M. Gast 1976. Acute and chronic toxicity of copper to the fathead minnow in a surface water of variable quality. Water Res. 10:37-43. Chester, R., and J. H. Stoner. 1974. The distribution of zinc, nickel, manganese, cad- mium, copper, and iron in some surface waters from the world ocean. Mar. Chem. 2:17-32. Christensen, G. M. 1975. Biochemical effects of methylmercuric chloride, cadmium chloride, and lead nitrate on embryos and ale- vins of the brook trout, Salvelinus fontinalis. Toxicol. Appl. Pharmacol. 32:191-197. Gushing, D. H. 1975. Maine ecology and fisheries. Camb. Univ. Press, Camb., Engl., 278 p. Dethlefsen, v., h. von westernhagen, and H. Rosen- thal. 1975. Cadmium uptake by marine fish larvae. Helgol. wiss. Meeresunters. 27:396-407. Eldridge, M. B., and W. M. Kaill. 1973. San Francisco Bay area's herring resource — a color- ful past and a controversial future. Mar. Fish. Rev. 35(11):25-31. Finney, D. J. 1964. Statistical method in biological assay. 2d ed. Hafner, N.Y., 668 p. Gardner, G. R., and G. LaRoche. 1973. Copper induced lesions in estuarine teleosts. J. Fish. Res. Board Can. 30:363-368. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can., Bull. 180, 740 p. Hazel, C. R., and S. J. meith 1970. Bioassay of king salmon eggs and sac fry in copper solutions. Calif Fish Game 56:121-124. Hoss, D. E., L. C. Coston, J. P. Baptist, and D. W. engel. 1975. Effects of temperature, copper and chlorine on fish during simulated entrainment in power-plant condenser cooling systems. In Environmental effects of cooling sys- tems at nuclear power plants, p. 519-527. Int. At. Energy Agency, Vienna. HYNES, H. B. N. 1970. The biology of polluted waters. Univ. Toronto Press, Toronto, Can., 502 p. Kinrade, J. D., AND J. C. Van Loon. 1974. Solvent extraction for use with flame atomic absorp- tion spectrometry. Anal. Chem. 46:1894-1898. Lee, G. F. 1973. Review paper: Chemical aspects of bioassay tech- niques for establishing water quality criteria. Water Res. 7:1525-1546. Litchfield, J. T., Jr , and F. Wilcoxon. 1949. A simplified method of evaluating dose-efTect exper- iments, J. Pharmacol. Exp. Ther. 96:99-113. Martin, M., M. D. Stephenson, and J. H. Martin. 1977. Copper toxicity experiments in relation to abalone deaths observed in a power plant's cooling waters. Calif. Fish Game 63:95-100. 355 FISHERY BULLETIN: VOL. 76, NO. 2 MAY. R. C. 1974. Larval mortality in marine fishes and the critical period concept. In J. H. S. Blaxter (editor), The early life history offish, p. 3-19. Springer- Verlag, N.Y. MCKIM, J. M., AND D. A, BENOIT. 1 97 1 . Effects of long-term exposures to copper on survival, growth, and reproduction of brook trout (Salvelinus fon- tinalis). J. Fish. Res. Board Can. 28:655-662. MITCHELL, F. K., AND D. J. MCDERMOTT. 1975. Characteristics of municipal wastewater dis- charges, 1974. South. Calif Coastal Water Res. Proj., Annu. Rep. 1975, p. 163-165. Mount, D. I. 1968. Chronic toxicity of copper to fathead minnows (Pimephales promelas, Rafinesque). Water Res. 2:215- 223. MYTELKA, A. I., J. S. CZACHOR, W. B. GUGGINO, AND H. GOLUB. 1973. Heavy metals in wastewater and treatment plant effluents. J. Water Pollut. Control Fed. 45:1859-1864. O'REAR, C. W., Jr 1972. The toxicity of zinc and copper to striped bass eggs and fry with methods for providing confidence limits. Southeast. Assoc. Game Fish Comm., 26th Annu. Conf ,p. 484. PAGENKOPF, G. K., R. C. RUSSO, AND R. V. THURSTON. 1974. Effect of complexation on toxicity of copper to fishes. J. Fish. Res. Board Can. 31:462-465. PICKERING, Q. H., AND M. H. GAST 1972. Acute and chronic toxicity of cadmium to the fat- head minnow (Pimephales promelas). J. Fish. Res. Board Can. 29:1099-1106. POSTUMA, K. H., AND J. J. ZIJLSTRA. 1974. Larval abundance in relation to stock size, spawning potential and recruitment in North Sea herring. In J. H. S. Blaxter (editor), The early life history offish, p. 113- 128. Springer- Verlag, N.Y. RILEY, J. P., AND D. Taylor. 1968. Chelating resins for the concentration of trace ele- ments from sea water amd their analytical use in conjunc- tion with atomic absorption spectrophotometry. Anal. Chim. Acta. 40:479-485. Rosenthal, H., and K.-R. Sperling. 1974. Effects of cadmium on development and survival of herring eggs. In J. H. S. Blaxter (editor). The early hfe history of fish, p. 383-396. Springer- Verlag, N.Y. SKIDMORE, J. F. 1965. Resistance to zinc sulphate of the zebrafish (Brachydanio rerio) at different phases of its life his- tory. Ann. Appl. Biol. 56:47-53. Snedecor, G. W., and W. G. Cochran. 1967. Statistical methods. 6th ed. Iowa State Univ. Press, Ames, 593 p. SpragUE, J. B. 1969. Measurement of pollutant toxicity to fish. L Bioas- say methods for acute toxicity. Water Res. 3:793-821. Struhsaker, J. W., M. B. Eldridge, and T. Echeverria. 1974. Effects of benzene (a water-soluble component of crude oil) on eggs and larvae of Pacific herring and north- em anchovy. In F. J. Vemberg and W, B. Vemberg (editors). Pollution and physiology of marine organisms, p. 253-284. Academic Press, N.Y. Vaughan, D. S., and S. B. Saila. 1976. A method for determining mortality rates using the leslie matrix. Trans. Am. Fish. Soc. 105:380-383. VOGEL, F. S. 1959. The deposition of exogenous copper under experi- mental conditions with observations on its neurotoxic and nephrotoxic properties in relation to Wilson's diseeise. J. Exp. Med. 110:801-810. WALDICHUK, M. 1974. Coastal marine pollution and fish. Ocean Manage. 2:1-60. WEDEMEYER, G. 1968. Uptake and distribution of Zn*^ in the coho salmon egg iOncorhynchus kisutch). Comp. Biochem. Physiol. 26:271-279. \ 356 ESTIMATED ZOOPLANKTON PRODUCTION AND THEIR AMMONIA EXCRETION IN THE KUROSHIO AND ADJACENT SEAS TsuTOMU Ikeda' and Sigeru Motoda^ ABSTRACT Production and ammonia excretion of zooplankton in the Kuroshio and adjacent seas were estimated from field data of biomass, size distribution, and habitat temperature of zooplankton, and from experimental data of respiration and ammonia excretion rates as functions of body size and tempera- ture. Winberg's basic balanced equations were applied to calculate production from respiration data. Further, mortality related to the lifespan and the ratio of herbivores to carnivores in the zooplankton community were estimated from theoretical assumptions. In this study, 18-72% of primary production was grazed by herbivorous zooplankton, and production of herbivorous zooplankton ( = secondary production) was 10-60 mg C/m^ per day. The ecological efficiency between primary and secondary production was 5-22%. Ammonia-nitrogen excreted by zooplankton was 4-24 mg N/m^ per day, which can support 11-44% of the nitrogen requirements of primary production. In marine ecosystems solar energy photosynthet- ically fixed as oi'ganic matter by phytoplankton is channelled through zooplankton to nektonic fishes and crustaceans at higher trophic levels. Impor- tant features of the roles of zooplankton in this scheme are their extremely high conversion ef- ficiency of phytoplankton organic matter (in con- trast with terrestrial ecosystems, see Wiegert and Owen 1971; Steele 1974) and the simultaneous regeneration of nutrients through their excretory activities. The latter role is considered an impor- tant mechanism in maintaining constant primary production levels in the seas, especially in oligo- trophic areas (Ketchum 1962; Corner and Da vies 1971). These dynamic functions of zooplankton have seldom been quantitatively evaluated in the field. One difficulty lies in the fact that the zooplankton community includes animals belonging to a vari- ety of phyla and a number of species which differ geographically. Information from detailed studies on one or a few species is not adequate for this purpose, and collection of all necessary data on each component species in the community is not practical. Therefore, the development of some al- ternative approach is needed to overcome this problem. ^Australian Institute of Marine Science, P.M.B 3, Townsville, MSO, Queensland, Australia. ^Marine Science and Technology, Tokai University, Shimizu, Japan. Manuscript accepted November 1977. FISHERY BULLETIN; VOL. 76, NO. 2, 1978. METHODS In this study, we treat the zooplankton commun- ity as an assemblage of different sizes of animals and use body size-related constant functions for respiration and ammonia excretion from labora- tory experiments to estimate feeding, production, and ammonia regeneration in the Kuroshio and adjacent seas. A systematic survey of the study area had been carried out by Japanese parti- cipants in the CSK (Co-operative Study of the Kuroshio and adjacent region) organized by UNESCO during 1965-67 (Motoda et al. 1970; Irie and Yamazi 1972). Biomass, Habitat Temperature, and Size ( = Weight) Distribution of Zooplankton Zooplankton were sampled vertically from 150 m with a NORPAC standard net (mesh aperture, 0.35 mm) in summer (June-October 1965 and 1966) (Figure lA) and winter seasons (December- April 1965, 1966, and 1967) (Figure 2A). From the average biomass of zooplankton summarized by Yamazi (1971) for 0-150 m, the present study area was divided into four density classes (<10, 10-50, 50-100, and >100 mg wet weight/m^). The isopleth for 100 mg wet weight/m^ shifted northward in the cold season and south- ward in the warm season, especially in the east China Sea (Motoda et al. 1970; Irie and Yamazi 1972). Seasonal difference in the composition of 357 FISHERY BULLETIN: VOL. 76, NO. 2 Figure l. — A. Sampling stations, zooplankton biomass, and isotherms (100-m depth, continuous hnes; 50-m depth, broken lines) during the warm season (June-October) in the Kuroshio and adjacent seas. B. Distribution of estimated secondary production. 358 IKEDA and MOTODA: ZOOPLANKTON PRODUCTION AND AMMONIA KXCRKTION ,0 ZOOPLANKTON BIOMASS (mg wet wt/m-*) >100 50-100 10-50 <10 2S\ 0\ SECONDARY PRODUCTION (mg C/m^/day) & 31-60 11-30 \ ca.10 _L Figure 2. — A. Sampling stations, zooplankton biomass, and isotherms (100-m depth, continuous lines; 50-m depth, broken lines) during the cold season (December- April) in the Kuroshio and adjacent seas. B. Distribution of estimated secondary production. 359 FISHERY BULLETIN: VOL. 76, NO. 2 zooplankton taxonomic groups among stations was less pronounced, with copepods dominant (56-657r of total individual number), followed by Noctiluca (S-lS'/f ), appendicularians (G-T^'r), and chaetognaths (4-5'7f) (Yamazi et al. 1972). Biomass expressed per cubic meter was converted to per square meter by multiplying by depth of sampling. The habitat temperature of zooplankton from 0-150 m was represented by that at 100 m (Japanese Oceanographic Data Center 1967, 1969). In the east China Sea, which is shallower than 150 m, the temperature at 50 m was taken as the habitat temperature (Figures lA, 2A). From data summarized by Yamazi (1971), the biomass of zooplankton per haul was divided by total number of individuals per haul to obtain average body weight. Values thus obtained at all sampling stations were grouped into warm or cold season, and assumed as a general size distribution in each season (Figure 3). The highest frequency was observed in the range 0.1-0.2 mg wet weight/ animal in both seasons. Faunal differences south 50 40 30 ^20 z 10- UJ o u a li. 0 « I •7. 30- hm ft^ -15 -0 5 0 5 LOG »V BODV WT s~^ \- •.'.0 "=30 20- 10- I I""t N:U7 •/, 20- (0-3'/.) m flv •> n n -15 -05 05 LOG AV BODY WT -~Ti~>-^ (2 6'/.) T 1 1 1 1 1 1 1 1 0 0 0501 0 3 0 5 0 7 0 9 11 13 15 17 19 AVERAGE BODY WEIGHT OF A ZOOPLANTER (mg wet wt) Figure 3. — Relative frequency of average size of zooplankton (biomass/number of zooplankton at each sampling station) in warm (June-October) (upper figure) and cold (December- April) (lower figure) seasons in the Kuroshio and adjacent seas. A normalized frequency distribution fitted by logarithmic trans- formation of body weights is superimposed on the right side of each figure. N is number of sampling stations. and north of the subarctic boundary (ca. lat. 40°N) reported by Motoda and Marumo (1965) were ig- nored here, because no systematic difference was found in average body size of zooplankton between these areas. The skewed size distribution was con- verted to a normal distribution curve by logarithmic transformation (base 10). Fitness to the curve was tested primarily by the normal probability paper (Harding 1949) and finally confirmed by chi-square test (warm season: x^ = 17.85,df = 6, P<0.01; cold season: x^ = 7.24, df = 6, 0.25

Mean % ovigerous ? Mean min size ovigerous 9 % 9 >13.8 mm carapace length 7.5 (8,956) 9.7 (6,347 19.4(6,347) 13 8(6,347) 21.1 (6,347) 6.0 (6,951) 7.5 (7,730) 1.1 (7,730) 13.8 (7,730) 2.7 (7,730) DISCUSSION Methods for Measuring Growth The growth rate of crustaceans in nature, though of considerable research interest, has been difficult to measure for several reasons. Primarily, all of the hard parts of the animal are cast off with the molt, making the marking of them all but impossible until recent years. Wenner et al. ( 1974) discussed the problems associated with measuring growth for crustaceans, and their table 1 sum- marized possible patterns of growth for the Crus- tacea. That table stressed the relative contribu- tion of two factors in crustacean growth: molt increment and molt frequency. Both of these may be responsive to different environmental parame- ters, altering the growth pattern of a species. The standard methods for measuring field growth rate for crustaceans (caging, mark and recapture, and analysis of modal size classes with- out corroborative data) are unsatisfactory to com- pare field growth rates for different populations of E. analoga. Since none of these methods alone suffices for this kind of comparative measurement with E. analoga, the instantaneous growth rate approach was used in this comparative analysis. The method has qualities common to some of the other methods mentioned, but avoids some of the inherent problems of those methods. This technique allows direct observation of size-specific molt frequency and molt increment, while minimizing the handling effects normally as- sociated with laboratory impoundment. It is likely that molt frequency estimates by this method are more accurate for the larger crabs, for which the 5-day holding period is a relatively smaller pro- portion of the intermolt period. The method has allowed comparison of growth factors (molt incre- ment and frequency) in detail {orE. analoga and gave repeatable data such as that found for Goleta Bay in 1974 and 1975. Thus a technique for mea- surement of crustacean growth has been de- veloped here which may hold promise for such comparative studies as this, where field caging is impractical. Growth of Emerita analoga The large difference in the growth rate of E. analoga between beaches of Goleta Bay and a Santa Cruz Island bay is remarkable in view of their proximity (about 42 km apart) but not in view of the different environmental conditions found at each beach. The combination of colder water and reduced filterable material in suspen- sion in the water appears to have slowed the growth of .E. analoga on Santa Cruz Island. This difference is evidence of the sensitivity of these two factors of sand crab growth to variation in environmental qualities. It is tempting to construct growth curves from such data on molt increment and molt frequency, having arrived at estimates for these. Both of these factors, however, have been shown to be highly responsive to environmental conditions. In fact, they vary widely in time and space with no clear pattern emerging as yet. A growth curve constructed from these data would apply only under a specific set of environmental conditions. Certainly these large differences in growth rate in nearby beaches precludes the use of modal size classes from several beaches for the determination of growth for£^. analoga, as Efford ( 1967) did ear- lier. 373 FISHERY BULLETIN: VOL. 76, NO. 2 Efford (1967:84, figure 3) presented a growth curve forE. analoga, constructed from data taken from 22 beaches on the Pacific coast between En- senada, Mexico, and Tofino, Canada (a 2,400 km distance). Three-fourths of the data presented were gathered over a period of only 2 mo (between 17 June and 17 August 1961). The remaining data were collected in 1959 and 1963. Where size- frequency data were bimodal, the author assumed that two year classes were present. To construct a growth curve from his data, Efford also had to assume that: 1) growth rate was the same year to year (temporally stable, at least during the grow- ing season); and 2) longshore migration did not take place for E. analoga. Dillery and Knapp (1970) demonstrated that an average E. analoga individual of about 26 mm carapace length travels about 15 m/day alongshore in an easterly direction on local beaches in Santa Barbara. This implies that the individuals in inhabiting a particular location may change from day to day. Barnes and Wenner (1968) suggested that the interpretation of size- frequency data is considerably simplified if sex reversal is assumed for this species, and some di- rect evidence (Wenner 1972) supports this as- sumption. However, recent laboratory data (Fu- saro 1977) suggest that a differential growth rate for males and females between 9 and 14 mm carapace length may account for the observed size-frequency distributions and sex ratio pat- terns, rather than protrandry. Combining data from different beaches, as Ef- ford ( 1967) did, also carries with it the assumption that the growth rate is relatively the same for the various parts of the range o^E. analoga (spatially stable). However, in an analysis of E. talpoida data presented by Wharton (1942), Efford suggested that the growth pattern of this latter species may differ in the southern part of its range. It is likely that temperature was responsible for the difference, as it is likely that temperature has an effect on the growth of £. analoga in different parts of its Pacific coast range. Wenner et al. ( 1974) presented data (their figure 5) which suggested that for E. analoga, even dif- ferent local populations may display different growth patterns, at least as indicated by size at sexual maturity. Data presented in this report imply that molt increment and molt frequency are indeed different in different environmental re- gimes in nature. Growth curves constructed for such different areas would likely differ. To com- bine these sets of data would be to obscure the real differences in growth rate observable in such local, proximate populations. The instantaneous growth rate estimate, though, may be used as an index of how well a population fares under a given set of environmen- tal conditions. Consider this instance. It has been shown that a population oiE. analoga on a beach at Santa Cruz Island grew about one-third as fast as a population in Groleta Bay (molt increment depressed by one-third and molt frequency de- pressed by one-half). Assuming a fixed number of molts to maturity (e.g., Wenner et al. 1974, table 1), the island population would reach maturity at a smaller size and in about twice as long a time. In fact, population structure data (Table 3) shows that sand crabs from the island population reached maturity at about the same size as those from Goleta Bay. If a fixed size at maturity is assumed, the island population sampled would take about three times as long to reach that fixed size. The third possible assumption, that there is a fixed length of time to maturity, is argued against by all available data. In any of these cases, how- ever, the population of sand crabs inhabiting the beach of the Santa Cruz Island bay location was at a distinct disadvantage in terms of reproductive success when compared with the population oiE. analoga inhabiting the Goleta Bay beach. This reproductive disadvantage was brought about at least in part by the large observed differences in molt frequency and molt increment at the two locations. A much smaller percentage of females were of reproductive size, probably due to the dif- ferential growth rate (see Table 3). Cox and Dudley ( 1968) also reported large vari- ations in the size of the smallest egg bearing female £^. analoga found in their collections. Data presented here may account for such previously problematical observations, in that differences in growth rate may affect the size distribution and abundance of mature females. As these data suggest, then, the growth rate of a crustacean population plays an important role in the life history of that species in its particular environmental situation. Of course, when dealing with a species which has pelagic larval stages, it becomes difficult to study local populations under the assumption that they are genetically different. Recruitment patterns are not generally well known for species with pelagic larvae (see Thorson 1950; Efford 1970; Mileikovsky 1971; Strathmann 374 FUSARO: GROWTH RATE OF THE SAND CRAB 1974), thus confounding the issue of reproductive success for a population in a particular habitat. Thus "relative reproductive success" may not be as good a criterion between populations such as these as it is between species. Measurement of differ- ences in such life history factors as growth rate may, therefore, take on added significance in the comparison of two populations or species in differ- ing environments, inasmuch as they do not depend on the assumption of genetic isolation but concern themselves more with the relationship of the population to its particular environmental cir- cumstances. ACKNOWLEDGMENTS I thank A. M. Wenner for consultation through- out the study and for his critical review of the manuscript, and J. Childress, A. Oaten, and J. King for their critical review of the manuscript. I thank C. Stanton of Santa Cruz Island Company for use of the island site studied, and L. Laughrin of the University of California Santa Cruz Island Field Station for the support facilities under his charge. I also thank P. Kearney for assistance in typing the manuscript. LITERATURE CITED Barnes, N. B., and a. M. Wenner 1968. Seasonal variation in the sand crahEmerita analoga (Decapoda, Hippidae) in the Santa Barbara area of CaUfomia. Limnol. Oceanogr. 13:465-475. Cox, G. W., AND G. H. Dudley 1968. Seasonal pattern of reproduction of the sand crab, Emerita analoga, in southern California. Ecology 49:746-751. DILLERY, D. G., AND L. V. KNAPP. 1970. Longshore movements of the sand crab, Emerita analoga (Decapoda, Hippidae). Crustaceana 18:233- 240. Drach. p. 1939. Mue et cycle d'intennue chez les crustaces deca- podes. Ann. Inst. Oceanogr. Monaco 19:103-391. EKFORD. I. E. 1966. Feeding in the sand crab, Emerita analoga (Stimp- son) (Decapoda, Anomura). Crustaceana 10:167-182. 1967. Neoteny in sand crabs of the genus Emerita (Ano- mura, Hippidae). Crustaceana 13:81-93. 1970. Recruitment to sedentary marine populations as exemplified by the sand crab, Emerita analoga ( Decapoda, Hippidae). Crustaceana 18:293-308. FUSARO, C. 1977. Population structure, growth rate and egg produc- tion of the sand crah, Emerita analoga (Hippidae): a com- parative analysis. Ph.D. Thesis, Univ. California, Santa Barbara, 182 p. Green, J. P., and M. R. Neff 1972. A survey of the fine structure of the integument of the fiddler crab. Tissue Cell. 4:137-171. MILEIKOVSKY, S. A. 1971. Tjqies of larval development in marine bottom in- vertebrates, their distribution and ecological significance: a re-evaluation. Mar. Biol. (Berl.) 10:193-213. Neushul, M., W. D. Clarke, and D. W. Brown 1967. Subtidal plant and animal communities of the Southern California Islands. In R. N. Philbrick (editor), Proceeding of the Symposium on the biology of the California Islands, p. 37-57. Santa Barbara Botanical Garden, Santa Barbara. Strathmann, R. 1974. The spread of sibling larvae of sedentary marine invertebrates. Am. Nat. 108:29-44. THORSON, G. 1950. Reproductive and larval ecology of marine bottom invertebrates. Biol. Rev. (Camb.) 25:1-45. WENNER, A. M. 1972. Sex ratio as a function of size in mjirine Crustacea. Am. Nat. 106:321-350. Wenner, A. M., C. Fusaro, and A. Oaten. 1974. Size at onset of sexual maturity and growfth rate in crustacean populations. Can. J. Zool. 52:1095-1106. Wharton, G. W. 1942. A typical sand beach animal, the mole crah, Emerita talpoida (Say). In A. S. Pearse, H. T. Humm, and G. W. Wharton. Ecology of sand beaches at Beaufort, N. C, p. 157-164. Ecol. Monogr. 12. 375 TIME-DEPENDENT SOLUTIONS AND EFFICIENT PARAMETERS FOR STOCK-PRODUCTION MODELS R. Ian Fletcher' ABSTRACT The time-dependent formulations of the Graham-Schaefer and Pella-Tomlinson systems are re- structured so as to accommodate directly the critical-point parameters of their respective governing graphs; the resulting parametric system accounts for the behavior of either model wholly in terms of its management components. The indeterminate exponent and the coefficients of the Pella-Tomlinson equations are uncoupled and the dual formulations associated with the conventional casting of the system are eliminated; the governing equations and corresponding solutions are cast into composite forms and the sign changes of coefficients become automatic. The previously obscure relationships between management parameters and variable graph curvature in the Pella-Tomlinson model are expressly formulated; maximum sustainable yield is shown to be independent of the indeterminacy of the system. Time-delay estimators for both systems are formulated. We analyze here, in a deterministic setting, cer- tain of the transient, nonlinear mechanisms employed in the modelling of stock and yield dur- ing periods of imbalance between fishing removals and stock productivity. The general method of analysis, which appeals primarily to the direct parameterization of critical points, will apply to any nonlinear scheme of exploitation and gross production, but it applies in particular to the Graham-Schaefer hypothesis (Graham 1935; Schaefer 1954) and to the "generalized" model of Pella and Tomlinson ( 1969). Since control of either system rests ultimately with the control of critical points, we restructure the parametric definitions accordingly and the governing equations for both systems are then controlled directly by parame- ters of management significance. Typically, either system reflects the determinis- tic premise that a stock of fishes, otherwise held by exploitation at levels below a prior abundance, will constantly strive to recover its numbers in accord with some innate, self-regulating, and re- peatable mechanism of restoration. Any such res- toration must accrue from the productivity of the stock, and by Graham's hypothesis, the inherent or latent capacity for productivity in a stock of fishes depends jointly on the current size of the stock (in numbers or biomass) and the difference between the current and potentially maximum 'Center for Quantitative Science in Forestry, Fisheries, and Wildlife, University of Washington, Seattle, WA 98195. sizes. Whence, in terms of time-dependent biomass B, and with the proportionality coefficient defined as the ratio of "intrinsic" growth rate k and 6^ , Graham's formula for latent productivity P takes on the familiar form P(B) = kB B '-B\ (1) Manuscript accepted September 1977. FISHERY BULLETIN: VOL. 76. NO. 2. 1978. Of the two expanded terms, the first governs the intrinsic, exponential capacity for growth of the population's biomass, while the negative, non- linear term provides the damping that ultimately slows growth as B(t) approaches its asymptotic maximum B^. The two terms, in their algebraic sum, govern the latent productivity of the stock at any stock size between zero and JS^^- Parameter k, as we shall see, is coupled analytically and phenomenologically to parameter By:, but the de- pendence of k on root Bx in Equation ( 1) can be supressed in favor of the direct parameterization of maximum productivity (which, in the complete exploitation model, we identify with maximum yield rate I. In the Pella-Tomlinson model, the parametric controls for latent productivity exceed by one the total number of such parameters in Graham's formulation, an increase in freedom that comes at considerable cost to tractability, both analytical and statistical. The differential equation that gov- erns latent productivity in the Pella-Tomlinson system has the indeterminate form 377 PiB) = CfB + C2 5", (2) with exponent n the additional parameter, but with the signs of the coefficients now dependent on the range of definition of n. As before, the com- bined terms describe, at any stock size B, the stock's latent capacity for productivity. With n undetermined (its determination being a part of the empirical demonstration), solutions of Equa- tion ( 2 ) constitute infinitely many growth laws. By setting n = 2, and with Cj >0, C2<0, Equation (2) reduces to the Graham equation (Equation (1)). PellaandTomlinson (1969) attribute Equation (2) to Richards ( 1959). For a detailed analysis of (2) as a general growth form, see Fletcher (1975); the anticedents of this analysis appear there. In either of the two systems, exploitation enters the formulation for productivity by the direct dif- • • • ference P - Y, with Y signifying the rate of biomass removal owed to exploitation and P the latent productivity of the stock. Wherefore, in writing B(B) = P(B) - Y(B), (3) we interpret B(B) as being the resultant produc- tivity, at stock size fi, that nets to the stock for its growth. The net may be positive, negative, or zero accordingly as P and Y vary with B. That is P > Y implies B > 0: the stock's latent productivity exceeds the rate of exploitation; a positive net productivity remains to the stock and the stock so tends to a higher level of biomass. • • • P < Y implies B < 0: the rate of biomass re- moval exceeds the stock's capacity for growth; the stock adjusts to the deficit in net productiv- ity by tending to a lower level of total biomass. • • • P = Y implies B = 0: the exploitation rate just balances latent productivity, and biomass trajectory B(t) exhibits an extremum. Should B = 0 over finite time, stock biomass remains stationary and the state called "equilibrium" prevails. Although the detailed time course of any real stock biomass is actually determined by varia- tions in renewal, survival, member growth, and the age- or size-dependent probabilities of capture, such effects are not usually separated in the mod- els of interest here, and yield rate Y customarily takes the form 378 FISHERY BULLETIN: VOL. 76, NO. 2. Y(t) ^F(t)'B(t), (4) with the implication that all fish of the fishable stock are presumed to share, in equal measure, the force of fishing mortality F, irrespective of age or size. By admitting Equation (4) into Equation (3), our general form for net productivity becomes B =P - F'B, (5) where the time variation of F is usually prescribed by average effort f on the assumption that F = qfiT, quantity q being the individual probability of capture per unit of effort and r the averaging in- terval measured in fractions of the dimensional time unit of F. ANALYSIS OF THE GRAHAM SYSTEM Figure 1 illustrates the phase-plane graph of Equation (1), the latent productivity of a Graham stock. Maximum productivity m occurs at stock sizep. And regardless of the conventions employed in the formulation of Equation (2), essential parametric control in the equation resides spe- cifically with its nonzero root By~ and with coordi- nate m of the critical point (p, m). Parameter m and Bx constitute a complete, minimum set of analytically independent parameters for latent ^cxmtrcHJUxbLe parcuneter- m = Pm^ Figure l. — Latent productivity P as a function of stock size B, the Graham model. See Equations (1) and (la). FLETCHER: TIME-DEPENDENT SOLUTIONS AND EFFICIENT PARAMETERS productivity in the Graham system, and they rep- resent the whole extent of available control over the graph of Equation (1). Coordinate p of the critical point has the fixed value B^/2, and the graph of Equation (1) has a fixed curvature of second degi'ee. Wherefore, productivity Equation (1), cast di- rectly in terms of analytical parameters m and B^, takes on the form p=4m[|-]-4m[|-]^ ,la, and intrinsic rate k, as it turns out, bears a propor- tionality dependence on maximum productivity and maximum biomass in the relationship k = i>oo L ^maxj And with the substitution of Equation (la) into Equation (5), the formula for the net productivity of a Graham stock becomes B = 4m ["LJ-^^-ll:] FB. (6) In the integrated, equilibrium versions of the Graham system, maximum latent productivity m becomes maximum sustainable yield (MSY), hence parameter m may be directly interpreted as MSY in any optimization procedure on Equation (6). If we restrict the time-dependence of F to abrupt changes so that any solution of Equation (6) cor- responds on its interval of validity, however brief, to some constant value of F, then the time- dependence of B in Equation (6) becomes productivity (Equation (6)) and the biomass solu- tion (Equation (7)) for cases where F < 4m Boo As indicated in the figure, root fi* becomes the adjustment level to which biomass trajectory B(t) will trend when F is less than critical quantity AmlBy^ (and obviously, Bit) trends to By- in Equa- tion ( 7 ) when F is zero) . The system is governed by the positive branch of Equation (6) when Y

0), and by the negative branch of • • • Equation (6) when Y > P (in which case, B < 0). But this partitioning of F into subranges for nega- tive or positive B is a density-dependent process. Although we must have F < 4m /B^ foJ* positive B*, the values of F on that range that drive the stock either up or down will depend on initial stock size B,,. To insure, for arbitrary B^, that F < P in Equation (6), mortality F must have a value such that 0 < F < 4m Bno [ - tl in which case B(t) increases from initial value B^ towards a higher adjustment level B.;,. But for any value of F on the interval 4m Boo Br Be < F < 4m Boo then Y > P and B(t) decreases from B„ towards a lower adjustment level B.: . Figure 3 illustrates the relationship between net productivity (Equation (6)) and the biomass solution (Equation (7)) when Bit) = B. l + Cne-<*'"/^^-^^>' B. r FBool (7) Be and with initial time tQ set arbitrarily to zero, the integration constant in Equation (7) becomes Co = B, -B( b7~ Figure 2 illustrates the relationship between net F > 4m Boo in which case the adjustment level of biomass cor- responds to the zero root of Equation (6). As indi- cated by the figure, any mortality F so great as to equal or exceed the quantity AmlBy., if main- tained, will fish a Graham stock to extinction. Since Equation (6) governs the relationship be- tween transient biomass and nonequilibrium re- moval, we look to its solution (Equation (7)) for time delays between equilibria. But the asympto- tic behavior of Equation (7) is a minor analytical annoyance to be circumvented here. Let us 379 FISHERY BULLETIN: VOL 76, NO. 2. 6 \ 5<0 5 6, M^c)/ e>>o e>(t) A B Figure 2.— A. Typical phase-plane graph of net productivity B = P - Y, Equation (6). the Graham system, with mortality F constrained to the interval 0 0 and the positive branch applies. B. Typical solution graphs of stock biomass B( t) , Equation ( 7 ) . When Y >P, biomass declines from initial value B^ towards adjustment level S* . When Y < P, biomass increases from initial value S^ towards adjustment level B^. e>(t) r>p A B Figures.— A. Typical graph of net productivity B =P -Y, Equation(6), the Graham system, with mortality F ^AmlBy,. Foranysuch value of F, the zero root of Equation (6) applies, removal rate Y exceeds latent productivity P, and B <0. B. Typical solution trajectory B(t) of Equation (7) when F & Am!B^. Biomass declines from initial value B,, towards extinction level B = 0. 380 FLETCHER: TIME-DEPENDENT SOLUTIONS AND EFFICIENT PARAMETERS presume that no practical technique of estimation will have a precision of resolution better than some assignable percentage of true stock size, and let us reflect that practical uncertainty in our analysis by expanding the asymptotic bound of Equation (7) to a region of radius €'B::, around the analytical value of the bound ( e being the measure of the uncertainty). Whence, with B,j and Sj now signifying initial and adjustment levels, and by supposing thatF changes abruptly at time /„ from value Fq to some new value F^, Equation (7) be- comes (1± 6) = l + Cne~^^'"/^«-'''i^' the plus sign applying when Fj >F„ and the minus sign whenFj 0. F = 2m/Bx; stock size B(t) implies p (p being the biomass level Bx/2 where maximum latent productivity occurs; Figure 1), which implies that Y-^'m. Accordingly, we may identify parameter m, in any of the rate equations here, with MSY (which, we should remember, is it- self a yield rate). Since, by Equation (4), instantaneous removal varies in time as Y(t) = Fit)B(t), then over the course of the adjustment interval that follows an abrupt change in F, yield from a Graham stock will accumulate as 381 FISHERY BULLETIN: VOL. 76. NO. 2. 7(0 = (B^-B^) In B* = B i4mlB^-F ..-1) Boo FB^ Am (9) m ^ 6(1 -n) U1i/(i-"> the plus sign applying to Equation (10) and the minus sign to Equation (11). ANALYSIS OF THE PELLA-TOMLINSON SYSTEM As noted in the foregoing section, the maximum latent productivity m of a Graham stock always occurs at a biomass value exactly one-half the unexploited maximum 5^. In turn, MSY of the equilibrium model must also occur at the stock level B^I2. So as to gain control over the locations of those extrema, Pella and Tomlinson ( 1969) mod- ify the Graham system by writing^the differential equation for latent productivity P essentially in the form of Equation (2), which, by the customary treatment, has a troublesome, dual formulation owing to the sign changes at /? =1 of coefficients c, ,c.^. On the interval Q 1 latent productivity takes on the basic form P = bB aB" (11) (where c\ = b, c.2 = -a, with a and b positive). In either case, the bound B-^, the maximum produc- tivity m, and the ordinate p (which governs the biomass level where m occurs), all depend on the numerical value assigned to exponent n. That is, root By is given by Boo ~ l/(l-n) the ordinate p is determined by P = an b i/(i-M) while maximum productivity m, by the conven- tional casting of the model, must be determined from the formula n > 1 6«/^ n= 1 P(5) m Figure 4.— Typical graph of Equation (12), latent productivity P as a function of stock size B, the Pella-Tomlinson system. Coordinate p. in its location with respect to root By., directly reflects the value assigned to expo- nent n, as indicated by Figure 4. When n takes any value between zero and unity, coordinate p falls on the range between zero and B-^/e ( ~ 0.3679 B^), in which case Equation (10) applies. When n takes any value greater than unity, coordinate p falls on the range between Byle and By, in which case Equation (11) applies. But the coordinate m has no essential dependence on exponent n, and its ap- parent coupling with n ( as indicated by the formu- lation above) is merely an inconvenient artifact of the conventional analysis. With parameters m and n uncoupled (see Fletcher 1975), the Equa- tions (10) and (11) that govern latent productivity in the Pella-Tomlinson system can be consolidated into the single governing equation P = ym [£] --[£]"■ <- with y a purely numerical factor wholly prescribed by n as 382 FLETCHER: TIME-DEPENDENT SOLUTIONS AND EFFICIENT PARAMETERS 7 = n n-1 (13) Be With the coefficients so cast, the sign reversals at turning point n ^ 1 become automatic. In con- sequence, the consolidated interval of definition for n becomes 0 < n < x (the point n ^ 1 being a removable singularity). With parameter m thus separated from n in Equation (12), the undeter- mined exponent n can be defined solely by the fraction pBy. in the relationship Br = n"^' ■n) (14) Consolidated Equation ( 12) now takes on the role in the Pella-Tomlinson system that Equation ( la) takes on in the Graham system. In fact, when n = 2, Equation (12) reducesto Equation (la), in which case y — A and p/B^ = V2. As an interesting aside here, we note that Equation (12), at the turning point n = I, takes on the form P = e m [Boo] [Poo_ (e being Napier's constant), while ratio ( 14), in the limit as n-^l, has the value In fact. Fox (1970) constructed a stock-production model around this special case, but since the ratio p/Bx has the fixed value 1/e, Fox's model "has as rigid a form as the Graham model" (Ricker 1975: 331). Quantities m, p, and B^ constitute a complete, minimum set of independent parameters for la- tent productivity in the Pella-Tomlinson system. Collectively they control the behavior of govern- ing Equation (12), but the influence of any one parameter remains independent of the remaining two. Figure 5 illustrates their separate effects on the graph of Equation (12). By appealing to the same piecewise constraints that enter the Graham productivity equations, we substitute Equation (12) into the general produc- tivity formula (Equation (5)) and net productivity in the Pella-Tomlinson system becomes B = ym [boo] ym B Be FB. (15) And over any time interval, however brief, that mortality F might be presumed to have a fixed value, biomass variable B in Equation ( 15) has the general time-dependent solution B^: unexploited stock level [the nonzero root of Equation ( 12)]. p: biomass level for maximum productivity [the coordinate of S in Equa- tion (12) where m occurs]. m: maximum productivity [the extremum coordinate fmax i"^ Equation (12)]. Figure 5. — The graph of Equation ( 12), latent productivity in the Pella-Tomlinson system, as controlled by independent parameters m, p. and By.. 383 FISHERY BULLETIN: VOL. 76. NO. 2, B{t) 5,^-"+Cexp ((ymlBoo-F)il-n)?j lia-n) (16) B* = [ym-FB^j l/d-M) B. By setting initial time ^^ arbitrarily at zero, the integration constant C in Equation (16) becomes C = B, l-n B. Biomass Equation (16) will apply immediately upon a change in F and remain valid thereafter for the time that F remains constant. Over such time, population biomass will trend up or down in accord with Equation (16) from initial size B^ towards adjustment level B^. Should nonzero root B^. be negative (which is possible only when n > 1), then the adjustment level corresponds to the zero root of Equation (15) and the population tends to extinc- tion by Equation (16). The critical relationships between fishing mor- tality, productivity, and time-dependent yield rate in the Pella-Tomlinson system are considerably more complex than the relationships between F, P, and Y in the Graham system. Figure 6 illus- trates the behavior of P - Y when n < 1, and Figures 7 and 8 illustrate P -Y when n > 1. The ratio ymlB-j- becomes the critical quantity in the Pella-Tomlinson system [AmlB-j. being its coun- terpart in the Graham system). As indicated by Figure 6, the biomass level p where maximum productivity occurs must lie on the range 0

P. And those values of F, for which Sff,* either increases or decreases to B.,.. , depend on the critical ratio ym/B^c and initial biomass value B^. To insure, for arbi- b<0 5 Tttnzshoid &co/& .T"^ — / fZorujc of p / when 1 & 5>0 e>it) F > Ym 1- itnplies Y>Py E>o^£>ifc ddfustmant n—i levcJ 1- ao"-' B, "-' '«» impUcs Y0^ B Figure 6.— a. Typical phase-plane graph of net productivity Equation ( 15), the Pella-Tomlinson system, for values ofn where 0 0 and the positive branch applies. B. Typical solution graphs of stock biomass Br ^J, Equation (16), when 0 < n < 1. Should Y >P, biomass declines from initial value B^ towards adjustment level B+ 384 But when Y < P, biomass increases from B^ towards B^, . FLETCHER: TIME-DEPENDENT SOLUTIONS AND EFFICIENT PARAMETERS trary fig, that Y

0 and the positive branch of Equation (15) applies. Trajectory B(t) then in- creases, in accord with Equation ( 16), from initial value 5p towards a higher adjustment level B... But for any value of F such that F > ym Ban B, n-l -I Be n-l then y > P and the negative branch of Equation (15) applies; trajectory B(t) decreases from B^ to- wards a lower adjustment level B^. Although the sign of B and the consequential course of B(t) is a density-dependent process for given F, we should note here that when n 1 ym Ban (17) then B(t)->'Y and -► m, irrespective of initial con- ditions. Accordingly, we may identify parameter m with MSY in any of the (reformulated) rate equations of the system. As indicated by Figures 7 and 8, the biomass level p where m occurs must lie on the range By^/e

1. And with n so prescribed, root B.., of Equation (15) may have positive or negative values accordingly as F has a value less or greater than the critical ratio ymlB-x_. Figure 7 illustrates the behavior of Equations (15) and ( 16) for the constraints n > 1 0 < F < ym Boo 0 < B* < Boo, in which case, root Bi.. of Equation (15) becomes the adjustment level such that B(t}-^B:i, by Equation (16). But whether Bf^ trends up or down to B:^ depends on the further partitioning of F with respect to initial biomass value B^. To insure, for arbitrary B„, that Y

P, B < 0, and the negative branch of Equation (15) applies; trajectory B(t) decreases from By towards a lower adjustment level B^. as indicated by the upper curve of Figure 7b. Should mortality F equal or exceed the critical ratio ym/By. in a Pella-Tomlinson system where n exceeds unity, the corresponding stock, over sufficient time, will trend to extinction. Figure 8 illustrates the behavior of Equations ( 15) and (16) for the constraints n > 1 ym F > Br B. < 0, in which case the zero root of Equation (15) applies, and we have B < 0 and B(t) -►O, irrespec- tive of initial conditions. By expanding the asymptotic bound of Equation (16) to a region of radius e'Bn:, and by appealing to arguments similar to those that led to the delay estimate (Equation (8)) of the Graham system, we calculate from Equation ( 16) the transition times for a Pella-Tomlinson stock as being Be 4ag (l-n)(7m-FiBoo) In "l-(l±6)^-" "I _l-(Bo/Bi)i-"J (18) where e represents the imprecision of stock- abundance estimates, and where B^, andBj signify initial and adjustment levels as they correspond to mortality values F„ and Fj . Again we suppose that F changes abruptly at zero reference time from value F„ to the new value Fj, the plus sign of Equation (18) applying when Fj > F^ and the minus sign when Fj < F^. By Equation (4) and the assumption that F var- ies in time by taking on fixed values of finite dura- tion, we can write the transient yield rate for the Pella-Tomlinson system in the consolidated form 385 FISHERY BULLETIN: VOL. 76, NO. 2. 5 \ n>i I I Thmshxyldipooje. 5 5(t; to Ym &, -/- an— I o rt—t < F < "67 implies V >P, £>o^&:ic implLes Y\ and F0. Should Y>P, the negative branch of Equation ( 15) applies; should Y\ and FP, biomass trajectory B(t) declines from initial value Bq toward adjustment level B*. Should ,Yi I 1 / b(t) Ex-iinction A B Figures.— A. Phase-plane graph of net productivity Equation (15) when « > 1 andF ^ymlBy-. Forany such combination of n andF, S* < 0 and the zero root of Equation ( 15) applies. B. Typical solution trajectory. Equation ( 16), when « > 1 and F > ymlB^, in which case the stock declines from initial value B„ towards extinction. 386 FLETCHER: TIME-DEPENDENT SOLUTIONS AND EFFICIENT PARAMETERS .1- Y{t) = FB, 1 -(l-(Bo/B*)'"7 exp ((7m/Boo-F)(l-n)/)y^^''"^ (19) which is valid for all values of n save n ^ 1. Owing to the range of definition on exponent 1/1 -n, I have not found a closed form for the general time integral of Equation (19) (although existence is fairly easy to show for n positive and either less or greater than unity). But the usefulness of the analysis does not suffer too greatly for that omis- sion, since one may accommodate Equation ( 19) to a numerical equation solver for finite measures of yield 8Y on associated intervals 8t. When F changes abruptly (as we have assumed throughout), yield rate Y changes abruptly, but the ensuing trends of adjustment are governed, in the Pella-Tomlinson system, by the following rela- tionships: 0 1: F < ym/By,: stock size B(t)->B^ (Figure 7), which implies that Y-^FB^.. F^ymlBy.; stock s\zeB(t) -►O (Figure 8), which implies that Y-^O. n > 0 (both ranges): F = (\-lln)ym!B~f.\ stock size Bit) ->p, which implies that Y -^■m (and we may identify maximum latent productivity m with maximum yield rate in any of the time- dependent formulations of the analysis). The quantity ym/By-, which plays such a promi- nent role in the analysis, can be identified as the "intrinsic growth rate" of the stock whenever ex- ponent n > 1, in direct analogy to quantity k of the Graham system (and, in fact, with n = 2, then y = 4 and Am/By. = k). But as a consequence of the indeterminate power form of the Pella-Tomlinson system and the switching of coefficient signs in the governing equations, the intrinsic growth rate turns out to be density-dependent when n takes on values between zero and unity. That is, by Equa- tion ( 12), the intrinsic rate (if we may call it so) has the form _ JUL 5"-i Boo when n falls on the interval 0 < n < 1 (in which case, 7 < 0). DISCUSSION Any nonlinear stock-production system may be restructured along the lines of the critical-point analysis described in the foregoing sections; such a treatment will generate parametric variables most likely to be those essential to management analysis. A synopsis of the parameters that appear in the restructured Graham and Pella-Tomlinson systems is given by Table 1. Table l. — Parameters of the restructured Graham and Pella- Tomlinson systems as they apply to management components. Control parameters Graham Pella-Tomlinson Management components system system Maximum stock size s^ Boo Maximum productivity (corresponds to MSY) m m Stock size for maximum productivity (the optimum" stock size) e^/2 (fixed) P Ratio p/e^ V2(fixed) n1/(1-n) Fistiing mortality F F General ad|ustment level (consult text for mortality conditions) e.. or 0 S., or 0 Fisfiing mortality for adjustment level p (the "optimum" F) 2/77/e^ (1 - 1/n) ymiBy, Graph curvature fixed n For optimization procedures on the Graham sys- tem, the essential parameters are {F, m, By-} augmented by the auxiliary parameters B^ and B;;:. For the Pella-Tomlinson system we may choose the combination {F, m.p.B,^} or the combi- nation {F, m, n,Bx}, either of which constitutes an essential set of mutually independent parameters. In the first set, p and By. determine n; in the second, n and S^ determine p. The relationships in either case are governed by Equation (14). Although the parametric influence of n is wholly prescribed by the ratio p/Byr, exponent n also determines the curvature of all graphs of the Pella-Tomlinson system. Therefore, when the par- ticularization of the system depends primarily on general curve fitting, the likelihood always exists that ill-determination of parameters will follow, 387 owing to stochastic displacement of datum points at biomass levels remote from locations p and B-^. As revealed by Equation (14), exponent n is quite unstable to small perturbations in the ratio p/B-^. The variational response in n exceeds the pertur- bation in p/B^ by an order of magnitude near n = 1, and the instability increases as p/fi^-^1. But the location of p with respect to fix is far more critical to management analysis than graph curvature and its associated "good fit," since, to the left of p, the stock produces biomass at a positively acceler- ated rate, while to the right of p productivity de- celerates. The trait of degeneracy in the system has been noted by Pella and Tomlinson (1969) and by Fox (1971, 1975), but the exact relationships between exponent n and the quantities m, p, and S^ have been obscured heretofore by the conventional castings of the system. With the restructured gov- erning equations and the explicit formulations of critical parameters, much of the statistical degen- eracy associated with previous routines can be constrained. And since the management parame- ters appear directly in the equations of the system, their variances can be calculated directly in the estimation procedure and appeals to indirect methods are avoided. FISHERY BULLETIN: VOL. 76, NO. 2. LITERATURE CITED Fletcher, R. I. 1975. A general solution for the complete Richards func- tion. Math. Biosci. 27:349-360. 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. 1975. Fitting the generalized stock production model by least-squares and equilibrium approximation. Fish. Bull., U.S. 73:23-37. Graham, M. 1935. Modern theory of exploiting a fishery, and applica- tion to North Sea trawling. J. Cons. 10:264-274. PELLA, J. J., AND P. K. TOMLINSON. 1969. A generalized stock production model. Inter-Am. Trop. Tuna Comm., Bull. 13:419-496. RICHARDS, F. J. 1959. A flexible growth function for empirical use. J. Exp. Bot. 10:290-300. RICKER, W. E. . 1975. Computation and interpretation of biological statis- tics offish populations. Fish. Res. Board Can., Bull. 191, 382 p. SCHAEFER, M. B. 1954. Some aspects of the dynamics of populations impor- tant to the management of the commercial marine fisheries. Inter-Am. Trop. Tuna Comm., Bull. 1:25-56. 388 TROPHIC RELATIONSHIPS IN JUVENILES OF THREE SPECIES OF SPARID FISHES IN THE SOUTH AFRICAN MARINE LITTORAL M. S. Christensen* ABSTRACT The feeding habits of three sparids, Diplodus sargus, D. cervinus, and Sarpa salpa, were studied. Juveniles of these fishes occur commonly in the intertidal and immediately subtidal regions of southeast Africa, while adults were only observed in these zones at high tide. Small juvenile D. sargus feed largely on harpacticoid copepods, amphipods, and, in spring and early summer, cirripede nauplii, chironomid larvae, and an unidentified trochophore larva. Larger individuals mainly take amphipods and green algae. Successive size classes of D. cervinus feed mainly on harpacticoid copepods and chironomid larvae, the shrimp Palaemon pacificus, amphipods, and then polychaetes. Sarpa salpa ingest harpacticoids when small, diatoms and red algae as a large juvenile, and red and green algae as an adult. Corresponding changes in gut length and dentition are reported for S. salpa. Marked ecological separation of the three species was observed. Small juveniles appear at different times of the year and feed on different foods (dietary and temporal separation). Larger juveniles and subadults have different diets or feed in separate parts of the littoral zone (behavioral, dietary, and spatial separation). A brief review of methods of analyzing stomach contents is included and it is suggested that a combination of points and ranking indices would be the most valuable. The method, here termed the comparative feeding index, is described. The food and feeding relationships of fishes in the intertidal zone of South Africa are poorly documented and the results are largely qualita- tive. The most important of these studies deal with one or two species of the families Gobiidae (Pitt- Kennedy^), Sparidae (Hutchings^), Cheilodac- tylidae (Butler^), and Gobiesocidae (Stobbs^). Three species of sparids were investigated in the present survey, Sarpa salpa (Linnaeus 1758), Di- plodus sargus (Linnaeus 1758), and D. cervinus (Lowe 1838). Barnard (1927), Smith (1965), and Hutchings (see footnote 3) described S. salpa (strepie) as being primarily a herbivore, whereas Talbot ( 1954) found the fish to be omnivorous. The 'J. L. B. Smith Institute of Ichthyology, Rhodes University, P.O. Box 94, Grahamstown, 6140, South Africa. *Pitt-Kennedy, S. 1968. A preliminary investigation of feeding in two gobies Coryphopterous caffer (Giinther) and C. nudiceps (C. and V.) with not«s on their sexual maturity. Unpubl. honors proj., 39 p. Zool. Dep., Univ. Cape Town, S. Afr. ^Hutchings, L. 1968. A preliminary investigation into the diets of two littoral teleosts, Sarpa salpa (Linnaeus) and Pachymetopon blochii (Valenciennes), with notes on their biol- ogy. Unpubl. honors proj., 25 p. Zool. Dep., Univ. Cape Tovm, S. Afr. ■•Butler, G. S. 1975. An investigation into the biology of two inter and infratidal species of Cheilodactylidae (Pisces: Teleo- stei). Unpubl. honors proj., 29 p. Zool. Dep., Rhodes Univ., Graham-stown, S. Afr. 'Stobbs, R. E. In preparation. Preliminary investigations into the feeding behaviour and food preferences of Chorisochismus dentex (Pisces: Gobiesocidae). Manuscript accepted August 1977. FISHERY BULLETIN: VOL. 76, NO 2, 1978. latter study was made in an estuary, however, where algae are generally less abundant than in the intertidal region. Diplodus sargus (blacktail) is described as being an omnivore (Biden 1954; Talbot 1954; Smith 1965), as isD. cervinus (zebra). Little, however, has been published on the food and feeding habits of the juveniles of these three species, although they are abundant in the inter- tidal and immediately subtidal regions of this coast. The objectives of this study have, therefore, been to determine: 1) the diet of juveniles of these three species; 2) how feeding changes with age and season; 3) the degree of overlap between species, possibly resulting in competition; 4) recruitment times and approximate growth rates of the fish; and 5) the relationship between dentition, gross gut morphology, and diet. During the course of this study, existing methods of fish feeding analysis were found to be inadequate and a new technique is described which overcomes some of the problems. MATERIALS AND METHODS Fish were collected from February to December 1975 at 2-wk intervals during spring tide, in spite of the possibility of introducing biases, as diving 389 FISHERY BULLETIN: VOL. 76, NO. 2 conditions were most suitable at this time. Hand nets were used in the intertidal pools and multi- prong spear guns in the subtidal area. Hook and line, poison, traps, and gill nets were not used as further biases may be induced to the feeding data (Randall 1967). Fish collected with hand nets were immediately placed in a 10% Formalin^ solution, whereas this procedure was delayed for up to iy2 h in the case of those taken by spear. It was con- cluded that death stops or greatly slows digestion, as the stomach contents were found to be in an equally digested state in both groups on later analysis. This has also been observed by Hobson (1974). The fish were left in Formalin for 10 to 14 days. This time period was maintained throughout to standardize any length and weight changes in- duced by the fixative (up to 5%, Royce 1972). About 10 scales were removed from under the pec- toral fin and cleaned with a camel hair brush after having been soaked overnight in water with a trace of carbolic acid (Pinkas 1966). They were mounted dry and examined over a white background using a low-power binocular micro- scope. Standard lengths to the nearest millimeter were taken and the stomach removed and placed in 45% n-propyl alcohol. The stomach is here defined as that part of the gut between the last gill arch and the gut caecae. The intestines were not examined as some food items are more resistant to digestion than others, with resultant biases as one moves along the gut (Randall 1967; Kionka and Windell 1972; Gannon 1976). Food items were identified to species where possible. Numerous methods have been employed in analyzing the food habits of fishes and volumetric and gravimetric techniques are being more widely used today with the current trend towards greater accuracy (Windell 1971). Both suffer from the same limitation in that digestion of the food both reduces its volume and weight. This has resulted in the use of reconstructed weights and volumes where the live weight and/or volume is back- calculated from a measureable parameter, e.g., carapace length (M. Bruton, pers. commun.). In this particular study some of the fish had fed on diatoms and it is not feasible to determine the volume of such small items (Windell 1971). Simi- larly, the reconstructed weight could not be de- termined as a sample of monospecific, uncontami- nated diatoms is impracticable to obtain and contains an indeterminate number of dead frus- tules which varies from sample to sample (Round 1971). In such cases, the points (Swynnerton and Worthington 1940) and ranking index methods (Hobson and Chess 1973) would appear to be more satisfactory and were initially used in the present study. The points system was modified by Frost (1943) and subsequently by Hynes (1950) to take into account gut fullness, 30 points being allotted when the stomach was distended, 20 when full, 10 when half full, and so on. One, two, four, eight, or sixteen points were assigned to each food item rather than fractions of the total allotted to each stomach in proportion to their volumes. This is an artificial situation and the method was revised as described below. After removal, the stomach is allotted between 0 and 30 points in proportion to its fullness. This is very subjective but overcome to some extent when large numbers of guts are handled. The contents are then sorted, identified, and the percentage volume estimated for each food item with the aid of Data Sheet No. 6 of Geotimes.'' All estimations are made with the organisms spread out to an even depth throughout the microscope field or equiva- lent surface. The total number of points allocated to that stomach is then subdivided amongst the food items in proportion to their percentage vol- umes. The points gained by each food item are summed for the total sample offish and the mean calculated. The values are then scaled down to a percentage to give the dietary composition of the fish examined. In the case of the ranking index (RI), the volume is estimated as above and the mean calculated for each food item per fish. The mean volume is then multiplied by the ratio of the number of fish containing that item to the total sampled. The points method, however, places too much weight on single food items that have been fed on to distension by a few fish, whereas the RI method fails to consider stomach fullness. It is therefore suggested that an alternative, here termed the comparative feeding index (CFI), would be more suitable as it takes into account all three factors, i.e., the volume, fullness, and frequency of occur- rence of each food item. The method involves the ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 390 'Available from the American Geological Institute, 2101 Con- stitution Avenue, N.W., Washing^n, D.C. CHRISTENSEN; TROPHIC RELATIONSHIPS OF SPARID FISHES allotment of points to each food organism as de- scribed above and the mean value per fish is then multiplied by the percentage of the total sample of fish that contain that item. As can be seen, the CFI combines the properties of both points and RI methods and thus reduces to some extent the effect of the problems discussed. The diet of these fish was also determined by the occurrence method (Hynes 1950) as this indicates the feeding preferences rather than the food's vol- umetric value. This is determined as the percent- age of fish in the sample analysed which contain that particular food item. The dietary composition of D. sargus was analysed for the winter (February- July) and summer (August-December) periods, as it was found to be seasonal. This was not done for the other two species, as feeding seasonality is synonymous in these fish with the change in diet with age, as they exhibit discontinuous recruit- ment. All skeletal material was cleared and stained using the trypsin maceration, alizarin stain method of Taylor ( 1967). The gut was dissected out in the same specimens, drawn and measured, and the gut length to standard length ratio (G/S) was calculated as in Weatherly (1972). STUDY AREA The study area is situated about 3.2 km north of Kleinemonde in the eastern Cape and is known locally as Clayton's Rocks (Figure 1). The shoreline consists of a gently shelving, sandy beach with broken rocky areas of varying extent. 33* 15' ' 1 CvT 1 0 Srahomstown \Great Fis^ River 33"X' V\ KleinemondevV'-i;^ Port Alfredjk'-'''''"''^ Clayton's Rocks t SOUTH Z' , ,„^ r^ * 10 km Chart Area X?\< J 26 30 2645 27 00 27 15 Figure l. — The study area in the Eastern Cape Province, South Africa. Adapted from Topographical Chart 3326, Grahamstown. The smaller rocky outcrops are continually cov- ered and uncovered by sand, as the beach is un- stable and backed by large, shifting sand dunes which move at right angles along the coast. The rocky area under study is made up of sandstone which strikes east-west and dips steeply south- wards. This has resulted in the development of gullies and pools partially sheltered from wave action by ridges of resistant rock (Figure 2). The maximum collection depth at the seaward edge of the gullies was 3 m at low tide. Environmental conditions vary greatly and salinities may fall to 25%o at low tide, which is caused by freshwater seepage into the pools from springs in the beach. During the day, at low tide, surface water temperatures have been recorded ranging from 26° (summer) to 15°C (winter) in the intertidal and from 22° (summer) to 14°C (winter) in the open sea. The other major fish species coexisting in the study area are listed below with their general biological characteristics, where known. ARIIDAE T achy sums feliceps: occurs singly, as a juvenile, in crevices. SPARIDAE Lithognathus lithognathus: occurs in small groups of juveniles. Rhabdosargus holubi: juveniles, in small groups. Sparodon durbanensis: juveniles, observed from October to March, either singly or in small groups. CHEILODACTYLIDAE Chirodactylus brachydactylus: as juveniles and subadults, singly, mainly from June to November. MUGILIDAE Unidentified species: occur all year round as juveniles. CLINIDAE Clinus cottoides: a purely intertidal species, lives in weed, juveniles appeared about June/July. Clinus superciliosus: lives in weed, juveniles only observed in November/December. GOBIIDAE Caffrogobius caffer: intertidal species, juveniles seen from June to November. TETRAODONTIDAE Amblyrhyncotes honckenii: singly or in small groups. 391 FISHERY BULLETIN: VOL. 76, NO. 2 ^.■^-'^. Figure 2.— The study site at Clayton's Rocks. RECRUITMENT AND GROWTH Di plod us sargus The monthly and total length-frequency dis- tribution is given in Figure 3. The lumped sample show^s a mode in the 10- to 20-mm size class, indi- cating that larger fish tend to emigrate to deeper water. The juveniles appear in the littoral zone when between 9 and 10 mm standard length (SL), and leave when about 90 mm long. It appears that large fish utilize the intertidal area at high tide as two fish of 107 and 108 mm SL were collected in the intertidal area some 2 h after low tide and another 164 mm long was collected 3 h after low tide. Visibility was <15 cm in the pools of the re- search area during September, October, and De- cember due to flooding of the Fish River. No dives could be made during this period in the subtidal area with the result that fish >40 mm were not collected. Recruitment of the juveniles into the littoral appeared to be relatively constant as no monthly peaks of abundance were found during the survey. This tends to confirm Biden's (1954) suggestion that females of this species spawn throughout the year, though mainly in summer. No monthly modes could be followed over the period of study as it is a continuously recruiting Standard Length (mm) Figure 3. — Monthly and total length- frequency distribution of Diplodus sargus, showing three age-classes (O-t-, 1-I-, and 2 + ). 392 CHRISTENSEN: TROPHIC RELATIONSHIPS OF SPARID FISHES population and so growth rates were not esti- mated. The age of three immature specimens was determined as being 1+ years old (107 and 108 mm SL) and 2+ years old (164 mm SL). Diplodus cervinus The length-frequency distribution is given in Figure 4. The total pooled sample shows that this species first appears in the littoral zone when about 8 mm long and leaves again when about 140 mm long. Large fish were observed to move into the intertidal area after low tide, as was the case with D. Sargus, but no specimens were obtained. Monthly modes were observed, which are age- classes 0-^, 1 + , and 2+ (Figure 4). Recruitment of juveniles into the tide pools is discontinuous, oc- curring between August and November, with a peak in October. Two fish sampled in August ( 132 and 129 mm SL) had just formed the second ring, giving the approximate time of scale-ring forma- tion. The estimated average growth rate as deter- mined for mode 1+ is 45 mm from February to December, which is about 54 mm/yr. 10 5 0 3 5 3 Total n=67 2* ^n fi y^^ Mar- .n n=17 A^ Apr. n = 5 May n=1 ^'*»^'''^^^^ Jun. n = 3 Jul. Jl=ji .-^ iTs^A A Aug n = 8 Sep n=0 Oct. „ n=7 Oh ^/ilA Nov. n=12 ,^> <^^. Dec. n = 5 50 100 Standard Length (mm) 150 Sarpa salpa The length-frequency distribution shows that the majority of the population is from 9 to 45 mm SL (Figure 5). The juveniles appear in the intertid- al when 2=9 mm, and fish >100 mm were never observed in the littoral at low tide; two specimens of age-class 2 -I- were collected 4 h after low tide. Three age-classes were observed, labelled as 0 + , 1 + , and 2 + . Recruitment of juveniles into the tide pools is discontinuous, occurring between June and September. Age-classes H- and 2+ were ap- proximately % and iy2 yr old, respectively, when sampled. The time of scale-ring formation is then likely to have been about June. The average growth rate is estimated to be 45 mm in 5 mo, or 108 mm/yr for the age-class 0-I-. Two fish were obtained in April 1976 with lengths of 68 and 76 mm, which indicates that the estimate may be slightly high, the predicted length being 81 mm. DIETARY COMPOSITION The composition of the diet is illustrated by oc- currence and CFI values scaled down to percent- ages. 30" 20 10 0' 3 3 ■Si 3 IT •b 3 10 10 5 ^^^^ i> Total 2* n=116 Feb. Mar. n=2 Apr May j}=L JiJi n=1 Z^ Jul. J1=S ^>N. Aug il=33 Sep. ^A. ii=i Oct. _Q=]a ^\r^ Nov. n=11 Dec n=A 50 100 Standard Length (mm) 150 Figure 4. — Monthly and total length -frequency distribution of Diplodus cervinus , showing three age-classes ( 0 -(- , 1 -f , and 2 + ). FIGURE 5. — Monthly and total length-frequency distribution of Sarpa salpa, showing three age-classes (0-I-, l-i-, and 2 + ). 393 FISHERY BULLETIN: VOL. 76. NO. 2 Diplodus sargus Winter Feeding The diet is composed mainly of harpacticoid copepods, amphipods, algae, isopods, polychaetes, and ostracods (Figure 6; Table I, n = 88). The diet of the smallest size class (5-15 mm) is composed almost equally of harpacticoid copepods and amphipods, but the percentage consumed of the former increases in the next size class whereas that of the latter decreases. The diet remained similar in the following two size classes. In the 35 - to 50-mm size class, the fish fed little on harpac- ticoid copepods, the diet being largely composed of amphipods. The situation was similar in the largest size class, although algae and polychaetes were increasingly taken. Summer Feeding 8 & 100 80 60 40 20- Length (mm) 5 No. of Fish 9 Figure 6. — Changes in diet with length of Diplodus sargus, collected between February and July 1975, as shown by the comparative feeding index. Food items included in Others are: brachyurans, crab zoaea, diatoms, echinoderms, hydrozoans, leptostracans, molluscs, mysidaceans, Palaemon pacificus. rhydophytan algae, sand, and unidentifiable animal fragments. Although similar to the winter diet, chironomid larvae, diatoms, crab zoaea, and leptostracans are more significant. Cirripede nauplii and an uniden- tified trochophore larva were also commonly taken, and these were not found in winter speci- mens (Figure 7; Table 2, n = 149). The diet of the smallest size class (5-15 mm SL) is composed mainly of harpacticoid copepods. The next size class (15-20 mm SL) fed on a similar diet, although the percentage of harpacticoids taken decreased and that of polychaetes and cirripede nauplii increased. These changes were further magnified in the 20- to 25-mm size class. In the next size classes (25-50 mm), there was a change and the green alga, Ulva sp., contributed sig- nificantly to the diet. Poor diving conditions in September, October, and December reduced the sample size and only nine fish in the 50- to 165-mm size range were analysed. In general, these fish showed an increas- ing tendency to take more amphipods, and less Table 1. — Changes in the percentage composition of the food of Diplodus sargus with length during the period February to July 1975, as assessed by the comparative feeding index (CFI) and occurrence (Occ.) methods. In the case of the former, all values exceeding SC/f have been italicized in order to emphasize those food items which contribute maximally to the diet ( — = absent). Size classes (mm) S^Ts 15-20 20-25 25-35 35-50 50-165 Taxon CFI Occ. CFI Occ. CFI Occ. CFI Occ. CFI Occ, CFI Occ. Chlorophyla ______ 1.3 33.3 1.0 25.0 12.8 41.7 Rhodophyta — — — — — — — — 0.2 8.3 1.6 16.7 Chrysophyta — — 02 118 — — — — — — 0.3 2.8 Polychaefa 2.4 33,3 11.8 76.5 16.7 75.0 3.6 50,0 14 33.3 12,8 47,2 Crustacea Amphipoda 47.4 77,8 5,8 29,4 5,0 25,0 7.8 66,7 80,7 66,6 58,7 72.2 Ostracoda 0.5 22.2 4.2 58.8 1.2 75,0 0.2 16.7 0.1 8.3 — — Harpacticoid copepoda 45.9 88.9 74 2 82,4 69,3 1000 80.9 66.7 1.3 25,0 0,4 11.1 Isopoda 1.2 33.3 1,0 35.3 7,4 75,0 5,0 50,0 3,3 50,0 4,2 33.3 Brachyura (zoaea) — — — — 0,1 12,5 — — — — — — Tanaidacea _ _ _ _ _ _ 0 1 16,7 0,1 8,3 — — Macrura ______ 0,1 16.7 0,2 16,7 — — Mysidacea _ — — — — — 0,2 16.7 — — 0,1 2,8 Insecta _ — — — 0,2 12 5 — — — — 0 1 2 8 Mollusca ________ 01 83 4,1 22,2 Echinodermata — — — — — — — — — — 0.1 2.8 Unidentifiable fragments 2,6 44.4 3.0 52.9 0.1 12.5 0 8 50.0 12.2 66.7 5.4 52.8 No of fish examined 9 17 8 6 12 36 Average no. of points allotted per stomach 18.8 13.4 7.3 20.3 17.2 13.2 394 CHRISTENSEN: TROPHIC RELATIONSHIPS OF SPARID FISHES lOOi 80 60 40 20' Ottiers ■,rww~--| Till IN, irn— --i^ - - ..T^ Length (mm) 5 15 20 25 35 50 165 No of Fish 72 36 K 12 6 9 Figure 7. — Changes in diet with length of Diplodus sargus, collected between August and December 1975, as shown by the comparative feeding index. Food items included in Others are the same as for Figure 8, in addition to the unidentified trochophore larva. algae, diatoms, chironomid larvae, and hydrozoa as they grow larger. Identity of Food Diatoms: two species of the genus Licmophora, predominantly L. pfannkucheae , as well as L. ehrenbergii. Other algae: the chlorophytan algae of the genus Ulua most frequent, although some of the larger size classes also fed on Caulerpa filiformis (50-165 mm: 7.5% in summer and 2.9% in winter), Bryopsis sp., Enteromorpha sp., and Valonia sp. Some rhodophytans taken, including Ceramium sp., Hypnea spicivera, Polysiphonia sp., and Tayloriella spp. Harpacticoid copepods: 12 species, only 4 com- mon, identification was not possible. Amphipods: 28 species — 7 caprellid species in- cluding Caprella danilevskii, C. penantis, C. scaura, and Caprellina longicollis; Cerapus tubularis; Corophium? acherusicum, and C? triaenunyx; Cymadusa sp.; two Gammaropsis species including G. holmesi; Jassa spp.; Lysianassa ceratina; and L. variegata; two Maera species; Paramoera capensis; Parelasmopus suluensis; two Photis species; Temnophlias sp.; Urothoe sp.; and three unidentified species. Isopods: nine species — Cymodocella pustulata, C. siiblevis, Dynamenella huttoni, D. mac- rocephala. Exosphaeroma antikraussi, Gnathia sp., Janiropsis sp., Panathui-a sp., and a Stene- trium species. Polychaetes: Dodecaceria pulchra. E alalia triliueata, two Nereis species, an Onuphis sp., and terebellid tentacles most commonly found in the gut contents, as well as Pista sp.. Pumatoleois kraussi, and Serpula vermicularis. Ostracods were not identified. Table 2. — Changes in percentage composition of the food of Diplodus sargus with length during the period August to December 1975, as assessed by the comparative feeding index (CFI ) and occurrence (Occ.) methods. In the case of the former, all values greater than 30% have been italicized in order to emphasize those food items which contribute maximally to the diet ( — = absent). size classes (mm) 5^15 15-20 20^55 25-35 35-50 50-165 Taxon CFI Occ. CFI Occ. CFI Occ. CFI Occ. CFI Occ. CFI Occ. Chlorophyta 02 42 06 83 2,9 28 6 33 4 50 0 135 33,3 143 55.6 Rhodophyta ______ 0,9 33,3 0,8 33,3 — — Chrysophyta ______ 0,7 25,0 87 33,3 7,4 11,1 Hydrozoa — — — — — — — — — — 11 111 Polychaeta 11 20,8 10,4 47.2 4,0 35 7 14,7 33,3 09 16 7 4.0 22.2 Crustacea Amphipoda 0.3 6 9 17 27 8 14 1 57 1 2 4 25 0 6 5 50 0 62 5 100.0 Ostracoda 6.0 37 5 5.5 55 6 6 4 64 3 0 8 16 7 0 4 16 7 0 1 111 Harpacticoid copepoda 67 8 86.1 58.? 94.4 47.6 100.0 29 6 83.3 176 100.0 — — Isopoda 1.0 16 7 2 1 25 0 3 4 42 9 3 0 25 0 6 9 50 0 9 8 44 4 Brachyura 0.1 5.6 - - 0 2 14.3 0 2 8 3 — — 0 2 111 Cirnpedia 8.8 36 1 139 66 7 215 42 9 13 83 — — - - Leptostraca — — 03 56 — — — Tanaldacea — — — — 0 2 14 3 — — Macrura — — — — 0 7 71 — Insecta 8 8 36 1 4 8 27 8 0 5 214 3 8 33 3 37 2 50 0 — — Moliusca — — — — — — — — — — 0311,1 Trochophore larvae 4.6 30.6 0.3 5.6 16 14 3 12 83 — — — — Unidentifiable fragments 1.3 20.8 2.3 30.6 2.9 28.6 8.1 41.7 7.5 50.0 0.2 11.1 No. of fish examined 72 36 14 12 6 9 Average no. of points allotted per stomach 13.7 17^ 204 12^0 14^8 182 395 FISHERY BULLETIN: VOL. 76, NO. 2 Insects: only the larva of the chironomid Tel- matogeton minor. Brachyurans: Rhyncoplax bovis and the gut of an unidentified crab (in a single case). Molluscs: Gibbula rosea, Helcion pruinosus, Philine aperta, and a rhaciglossid (no shell, so not possible to identify further). Hydrozoans: Symplectoscyphus sp. and Thecocarpus formosus were ingested by one fish in the 50- to 70-mm size class. Echinoderms: Parechinus sp. Tanaidaceans: Leptochelia barnardi most com- monly found, also an Apseudes sp. Mysidaceans: only one species, Mysidops similis, could be identified with any certainty. Di plod us cervintis The diet of Diplodus cervinus is illustrated in Figure 8 and Table 3 (n = 67). The juveniles (10-20 mm) fed mainly on harpacticoid copepods and chironomid larvae. In the next size class (20-35 mm), juveniles of the sand shrimp, Palaemon pacificus, were taken instead of chironomid larvae. This trend continues in the 35- to 50-mm size class, the diet being composed largely of P. pacificus as well as harpacticoid copepods. Polychaetes are more important in this size group, a trend which is maintained in all larger size classes. In the larger fish, there was again a changeover, the percentage of amphipods taken being 65.7% (50-75 mm) and 27.6% (75-100 mm). Unidentifiable crustacean fragments in 100' 80' 60 40 20 Length (nun) 10 No.of Fish 135 Figure 8. — Changes in diet with length in Diplodus cervinus , as shown by the comparative feeding index. Food items included in Others are: chlorophytan algae, cirripede nauplii, coralline al- gae, molluscs, mysidaceans, ostracods, tanaidaceans, and the unidentified trochophore larva. these size classes composed 12.6% and 47.9%, re- spectively, which is partly explained by the fact that 6 of the 40 fish were taken at night and their stomach contents were largely digested and thus indistinguishable. The diet of the largest size class ( 100-135 mm) was made up mainly of polychaetes. Identity of Food The diet was composed of almost all the food species listed for D. sargus, although in differing proportions, as well as the following: Cymodocella eutylos (isopod), Littorina knysnaensis (mollusc), Table 3. — Changes in the percentage composition of the food of Diplodus cervinus with length, as assessed by the comparative feeding index (CFI) and occurrence (Occ.) methods. In the case of the latter, all values exceeding 30% have been italicized in order to emphasize those food items which contribute maximally to the diet ( — = absent). Size classes (mm) 10-20 20-35 35-50 50-75 75-100 100-135 Taxon CFI Occ. CFI Occ. CFI Occ. CFI Occ. CFI Occ. CFI Occ. Chlorophyta — — — — 4.6 12.5 — — — — — — Rhodophyta __________1.9 16.7 Polychaeta 0.2 25 0 3 8 33.3 17 6 50 0 17 2 72 4 11.4 18.0 65.4 82.3 Crustacea Amphipoda 4.1 25 0 10 33.3 0 4 25 0 65 7 74 9 27 6 45 0 5.8 33.3 Ostracoda — — 02 22 2 — — — — 0.6 9.0 — — Harpacticoid copepoda 56.5 100.0 47,6 77 8 16.5 87.5 2.2 51.7 0.1 9.0 — — Isopoda — — 10.0 44.4 4.2 25.0 1.5 20.7 11.0 27.0 1.7 16.7 Cirripedia (nauplii) — — — — — — 0.5 6,9 — — — — Macrura — — 4 7.2 44.4 53.2 50,0 — — 1.0 9.0 — — Tanaidacea — — — — — — 0.2 6.8 — — — — Mysidacea — — — — — — 0.1 3.4 — — — — Insecta 39 0 50.0 2.2 55.6 0 4 12.5 — — — — — — Mollusca — — — — — — — — 0,4 9 0 1.9 16.7 Trochophore larvae 0.2 25.0 — — ___ — ___ — Unidentifiable fragments — — — — 3.1 50.0 12.6 79.3 47.9 36.0 23.3 33.3 No. of fish examined 4 9 8 29 11 6 Average no of points allotted per stomach 8.3 19.3 8.0 10.0 6.4 11.1 396 CHRISTENSEN: TROPHIC RELATIONSHIPS OF SPARID FISHES Phoxostoma sp. (amphipod), and Corallina sp. (rhodophytan alga). Sarpa salpa The percentage composition of the diet of Sarpa salpa is illustrated in Figure 9 and Table 4. The food habits of this species changed from being primarily a carnivore as a juvenile to a herbivore as a subadult. Juveniles in the 10- to 25-mm size classes fed mainly on harpacticoid copepods. In the next size 100- Figure 9. — Changes in diet with length of Sarpa salpa, as shown by the comparative feeding index. Food items included in Others are: bryozoans, cirripede nauplii , crab zoaea, fish muscle, insects, isopods, leptostracans, mysidaceans, ostracods, pMjlychaetes, rhaciglossid molluscs, and tanaidaceans. class (25-35 mm), however, the total animal con- tribution is only 15.1%, diatoms and rhodophytan algae being most important. Diatoms are taken in decreasing amounts from then on and those fish >75 mm SL fed predominantly on chlorophytan and rhodophytan algae. Identity of Food Chlorophytan algae: eight species — Bryopsis sp., Caulerpa filiformis, Chaemaedoris delphini, Cladophora spip.,Enteromorpha sp.,Rhizoclonium sp., and Ulua sp. Rhodophytan algae: Ceramium sp., Champia compressa, Hypnea spicifera, Polysiphonia sp., and Tayloriella sp. commonly taken as well as Ac- rosorium sp., Arthrocardia sp., Centroceras sp., Corallina spp., Polyzonia elegans, and Ptero- siphonia cloiophylla. Chrysophytan algae: three species of diatoms — Isthmia enervis, Licmophora ehrenber- gii, and L. pfannkucheae . Hydrozoans: Gattya humilis commonly ingested by the 75- to 100-mm size class, whereas Ser- tularella sp. and Thecocarpus formosus were un- common. Polychaetes: only two species of Nereis. Isopods: uncommon, but six species were found — Dynamenella huttoni, D. macrocephala, Gnathia sp., Janiropsis sp., Panathura sp., and Stenetrium sp. Table 4. — Changes in the percentage composition of the food of Sarpa salpa with length, as assessed by the comparative feeding index (CFI) and occurrence (Occ.) methods. In the case of the latter, all values exceeding 30% have been italicized in order to emphasize those food items which contribute maximally to the diet ( — = absent). Size classes (mm) 10-15 15-25 25-35 35-50 50-75 75-100 125-150 Taxon CFI Oca ^R Oca "CFJ Oca CFI Occ. CFI Occ. CFI Occ. CFI Occ. Chlorophyta 0.3 8 7 — — 0 9 313 37.2 92 0 34.0 917 30 6 100.0 48.3 100.0 Rhodophyla — — 0.4 9.1 21.3 62.5 323 76.0 28.6 66.7 45.9 83.3 50.7 100,0 Chrysophyta 0.2 87 9.3 24,2 62,7 1000 32 4 68,0 33.4 66.7 _ _ _ — Hydrozoa _______ — 0,1 16.7 20.5 50.0 0.8 50.0 Polychaeta 5.1 17.4 2.6 30,3 _ _ _ _ 0.4 8,3 0,1 167 — — Crustacea Isopoda 17 17.4 0 9 30.3 0,1 6 3 — — — — — — — — Amphipoda 73 30,4 17,8 63,6 4,2 50 0 0.9 16.0 0.1 16.7 0.3 33.3 0.8 50 0 Ostracoda 67 26 1 14 36 4 01 63 01 12 0 ______ Harpacticoid copepoda 77.0 87,0 64 5 72 7 0,5 18 8 1,4 24 0 ______ Cirnpedia (nauplii) 5,5 22,0 2 1 18,2 _____ — _ — — — Brachyura (zoaea) — — 0,1 3,0 0,1 12.5 ________ Leptostraca 0 1 4 4 — — — — — — — — — — — — Insecta 0,1 4 4 0,2 3.0 0.1 12.5 — — — — — — — — Bryozoa ________ — — 01 16 7 — — Mollusca __________ 0.1 16,7 — — Pisces ____ 0,2 6,3 — — — — — — — — Unidentified fragments and sand 2.0 26.1 0.7 24.2 9.8 31.3 1.7 32.0 3.4 41.7 2.4 50.0 — — No, of fisfi examined 23 33 16 25 12 6 2 Average no of points allotted per stomach 7.8 9.5 13.8 15.7 16.5 15.5 28.0 397 FISHERY BULLETIN: VOL. 76. NO. 2 Amphipods: also uncommon except in juveniles which took the following 15 species: Caprella cicur and two other species of^ Caprella, Cerapus sp., Corophium sp., Gammaropsis sp., Lysianassa ceratina, L. uariegata, Jassa sp., Maera sp., Paramoera capensis, Parelasmopus suluensis, Photis sp., and two unidentified species. Harpacticoid copepods: eight species. Insects: larvae of the chironomid Telmatogeton minor. Tanaidaceans: Leptochelia barnardi, uncom- mon. Molluscs: rhaciglossid. DENTITION AND GUT MORPHOLOGY Diplodus ceruinus — there are six upper and four lower incisors which are narrower than those of Z). sargus, and there are fewer molars, the number increasing with age (Figure lOA). This would in- dicate that adult D. cervinus feed on softer foods than D. sargus , which is borne out as the diet of the former consists primarily of polychaetes, whereas the latter took amphipods and molluscs as well. The gut of D. ceruinus is short with a G/S ratio of 0.7 in a 16.5-mm fish and 0.95 in one 74 mm long. Diplodus sargus — there are four stout incisors and three to four rows of fairly large molars in each jaw (Figure lOB), the latter increasing in size and number with age. The teeth are those of a typical omnivore (Weatherly 1972). The G/S ratio was 0.76 in a 16.5-mm fish and this is within the range of omnivores as defined by Nikolsky ( 1963). Sarpa salpa — this species shows a change in dentition correlated with age and diet. The young fish are carnivorous and have short, pointed coni- FlGURE 10.— Dentition. A. Medial view of the left upper and lower jaws oi Diplodus cervinus (MSC 75-36, 94 mm SL). B. Me- dial view of the left upper and lower jaws ofZ). sargus (MSC 75-34, 107 mm SL). C. Lateral view of the upper and lower jaws of a juvenile Sarpa salpa (MSC 75-39, 20 mm SL). D. Lateral view of a single tooth of a subadult S. salpa (MSC 75-37, 39 mm SL). E. Lateral view of the upper and lower jaws of an adult S. salpa (RUSI 74- 323, 99 mm SL). 398 CHRISTENSEN: TROPHIC RELATIONSHIPS OF SPARID FISHES cal teeth to grasp prey (Figure IOC). Multicusped, incisiform teeth begin to break through when the fish are about 20 mm long (Figure lOD). The pointed teeth are completely replaced by the time the fish are 35 mm long, after which they feed predominantly on algae ( approximately 60% CFI). The multiple cusps wear away and the teeth are bicuspid incisiform by the time the fish are 65 to 75 mm long (Figure lOE). These can nip at algae and the diet is composed of 65 to 117c plant matter at this stage. The gut shows a corresponding change in length with diet, a long gut being characteristic of a her- bivore. The G/S ratio increases from 0.86 in a 20-mm fish (Figure llC, typically omnivorous) to 1.36 in a 39-mm fish (Figure IIB), and 2.66 in a 99-mm individual (Figure llA). This latter value is typical of a herbivore (Nikolsky 1963), although not as pronounced as in some other herbivorous fish species. DISCUSSION A number of fish species occur as juveniles or spend their entire life cycle in the eastern Cape intertidal (see description under Study Area). The family Sparidae includes the largest number and the present investigation of the trophic relation- ships of three of these was initiated as competitive interaction is often most vigorous in closely re- lated fish ( Fryer and lies 1972). There is an intense dietary overlap in some cases and the available resources are subdivided in two main ways. Re- cruitment of juveniles of the three species takes Figure ll. — Lateral views ofSarpa salpa with the gut unravel- led and displayed to illustrate the increase in gut length with size. A. 99 mm SL IRUSI 74-323). B. 39 mm SL (MSC 75-37). C. 20 mm SL (MSC 75-39). place at different times of the year and this re- duces competition between those size groups in which the greatest feeding overlap was observed. The remaining size classes were separated as their diets were different. Small juveniles of the three species have the most similar diets of all size classes studied. The resulting competition is reduced by two mechanisms. Firstly, juveniles of Sarpa salpa occur in the tide pools primarily from July to Sep- tember (Figure 5) whereas those of Diplodus cer- vinus were found during October and November (Figure 4) and D. sargus was present throughout the year (Figure 3). Secondly, at the time of maxi- mal competition (July-November), the diet of small D. sargus includes food items not taken at other times of the year, e.g., chironomid larvae and cirripede nauplii (Figure 7). This may be due to either the presence of these prey items only at that time of the year and/or to the effects of com- petition forcing D. sargus to include them in its diet. A combination of both factors would appear to be operative in the case of the chironomids as the larvae were obtained in bottom samples taken in October-November and not in March. No data are available for cirripede nauplii, crab zoaea, and the unidentified trochophore larva as plankton sam- ples were not taken. Competition for food is greatly reduced by the time the three sparids are about 25 to 30 mm long. At this stage, S. salpa feeds mainly on diatoms and red algae (Figure 9); D. cervinus ingests Palaemon pacificus, harpacticoid copepods, and isopods (Fig- ure 8); and D. sargus takes green algae, harpac- ticoids, chironomid larvae, and cirripede nauplii (Figures 6, 7). The separation is equally distinct in subadult fish as S. salpa is then a herbivore, D. cervinus takes polychaetes, some amphipods, and isopods, while D. sargus feeds on amphipods and 399 FISHERY BULLETIN: VOL. 76, NO. 2 green algae. The overlap on amphipods by the latter two species may be partially compensated for by behavioral separation. Diplodus cervinus is a secretive substrate feeder whereas Z). sargus is a more open water fish tending to feed on vertical rock surfaces away from the bottom. The fact that neither species was very common intertidally in these size classes may also contribute towards a reduction in competition. The diet of large juvenile S. salpa is unusual in that it consists mainly of diatoms and epiphytic rhodophytan algae which occur commonly on corallines, Hypnea spicifera and Tayloriella spp. (M. H. Giffen, pers. commun.). The fish must, therefore, selectively separate these food items as few fragments of the algae on which they grow were found in the stomach contents. This is in contrast to Rhabdosargus holubi, also a sparid, which ingests algae for their epiphytic diatoms rather than separating them, even though the algae are not digested (Blaber 1974). The situation may be similar to this in larger S. salpa as the rectal contents appeared to be relatively undi- gested and fewer diatoms were observed on the algae (Figure 12). Temporal separation of juveniles to reduce com- petition has not been reported previously for tide pool fish species as far as I am aware, although it has been observed in two pelagic plankton feeders from the Adriatic, the anchovy and sardine ( Vuce- tic 1975). Large dietary overlaps have been noted in several intertidal fish, including blennies, clinids, gobies, and labrids (Gibson 1968, 1972). These fed predominantly on crustaceans and it is possible that similar mechanisms reduce competi- tive pressure amongst them, although this was not determined as samples were only taken for 2 mo. The data presented indicates that/), cervinus is a carnivore, D. sargus is an omnivore, and S. salpa an omnivore when juvenile and a herbivore when adult. Similar feeding habits were found for adults of the same three species in the Klein River es- tuary (Talbot 1954). The dentition and gross gut morphology changed with size and this was most marked in S. salpa, corresponding with the ob- served diet. Comparable transformations have been reported for other fish species (Nikolsky 1963). Two other sparids, Sparodon durbanensis and R. holubi, cohabit with those studied in the littoral zone. The few specimens examined had fed mainly on harpacticoid copepods as small juveniles, with a resultant overlap with D. sargus as all three oc- curred in the research area in November- December. Large specimens were not examined, but R. holubi appears to be an omnivore as a juvenile in estuaries and a carnivore feeding on molluscs as an adult (Talbot 1954; Blaber 1974). Adult S. durbanensis are carnivores feeding on small fish, molluscs, and crustaceans (Biden Figure 12.— Scanning electron micrograph of the surface of a chlorophytan alga, Ulva sp., removed from the gut ofSarpa salpa (147 mm SL) to show the disappearance of diatoms. A. Oesophageal sample. B. Rectal sample. 400 CHRISTENSEN: TROPHIC RELATIONSHIPS OF SPARID FISHES 1954). Competitive pressure is, therefore, minimized between the five sparid fish species commonly occurring in the littoral zone. ACKNOWLEDGMENTS This study was submitted in partial fulfillment of the requirements for an M.S. degree in the J. L. B. Smith Institute of Ichthyology, Rhodes Univer- sity, Grahamstown. I thank G. S. Butler, B. J. Hill, P. B. N. Jackson, M. van Harten, and R. Winter- bottom for assistance in the field and/or valuable criticisms. I am also grateful to the following, who identified many of the food items: C. Griffiths (am- phipods) and B. F. Kensley (isopods and brachyurans). University of Cape Town; M. H. Giffen (diatoms). Fort Hare University, Alice; and S. C. Seagrief (algae), Rhodes University, Grahamstown. R. Winterbottom provided useful suggestions and criticisms as my supervisor, and carefully reviewed and commented on the manu- script, as did R. N. Gibson, Dunstaffnage Marine Research Laboratory, Argyll, Scotland, and E. S. Hobson, National Marine Fisheries Service, NCAA, Tiburon, Calif. Their assistance was greatly appreciated as was that of J. Pote who typed the manuscript. Finally, I thank the follow- ing for financial assistance: J. L. B. Smith Insti- tute of Ichthyology (Hugh Le May scholarship), CSIR (research bursary), and Rhodes University (Research Council Grant No. 2857, R. Winterbot- tom principal investigator). LITERATURE CITED Barnard, K. H. 1927. A monograph of the marine fishes of South Africa. Ann. S. Afr. Mus. 21(2):419-1065. BIDEN, C. L. 1954. Sea-anghng fishes of the Cape: a natural history of some of the principal fishes caught by sea anglers and professional fishermen in Cape waters. 2d ed. Juta and Co. Ltd., Capetown and Johannesb., 304 p. BLABER, S. J. M. 1974. Field studies of the diet of Rhabdosargus holubi (Pisces: Teleostei: Sparidae). J. Zool. (Lond.) 173:407- 417. Frost, W. E. 1943. The natural history of the minnow, Phoxinus phox- inus. J. Anim. Ecol. 12:139-162. Fryer, G., and T. D. Iles. 1972. The cichlid fishes of the great lakes of Africa: their biology and evolution. Oliver & Boyd, Edinb., 641 p. Gannon, J. E. 1976. The effects of differential digestion rates of zoo- plankton by alewife, Alosa pseudoharengus , on determi- nations of selective feeding. Trans. Am. Fish. Soc. 105:89-95. Gibson, R. N. 1968. The food and feeding relationships of littoral fish in the Banyuls region. Vie Milieu, Ser. A, 19:447-456. 1972. The vertical distribution and feeding relationships of intertidal fish on the Atlantic coast of France. J. Anim. Ecol. 41:189-207. HoBSON, E. S. 1974. Feeding relationships of teleostean fishes on coral reefs on Kona, Hawaii. Fish. Bull., U.S. 72:915-1031. HoBsoN, E. S., AND J. R. Chess. 1973. Feeding oriented movements of the atherinid fish Pranesus pinguis at Majuro Atoll, Marshall Islands. Fish. Bull., U.S. 71:777-786. hynes, h. b. n. 1950. The food of fresh-water sticklebacks (Gasterosteus aculeatus and Pygosteus pungitius), with a review of methods used in studies of the food of fishes. J. Anim. Ecol. 19:36-58. KlONKA, B. C, AND J. T. WINDELL. 1972. Differential movement ofdigestible and indigestible food fractions in rainbow trout, Salmo gairdneri . Trans. Am. Fish. Soc. 101:112-115. NIKOLSKY, G. V. 1963. The ecology of fishes. Academic Press, N. Y., 352 p. PINKAS, L. 1966. A management study of the California barracuda Sphyraena argentea Girard. Calif Dep. Fish Game, Fish. Bull. 134, 58 p. RANDALL, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. (Miami) 5:665-847. ROUND, F. E. 1971. Benthic marine diatoms. Oceanogr. Mar. Biol. Annu. Rev. 9:83-139. ROYCE, W. F. 1972. Introduction to the fishery sciences. Academic Press, N.Y., 351 p. SMITH, J. L. B. 1965. The sea fishes of southern Africa. Cape and Trans- vaal Printers Ltd., Cape Town, 580 p. SWYNNERTON, G. H., AND E. B. WORTHINGTON. 1940. Note on the food of fish in Haweswater (Westmor- land). J. Anim. Ecol. 9:183-187. Talbot, F. H. 1955. Notes on the biology of the white stumpnose, Rhab- dosargus globiceps (Cuvier), and on the fish fauna of the Klein river estuary. Trans. R. Soc. S. Afr. 34:387-407. Taylor, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. U.S. Natl. Mus. 122(3596), 17 p. VUCETIC, T. 1973. Synchronism of the spawning season of some pelagic fishes (sardine, anchovy) and the timing of the maximal food (zooplankton) production in the Central Adria- tic. Pubbl. Staz. Zool. Napoli, 39 Suppl. 1:347-365. Weatherly, a. H. 1972. Growth and ecology of fish populations. Academic Press, Lond., 293 p. WINDELL, J. T. 1971. Food analysis and rate of digestion, /n W. E. Ricker ( editor) , Methods for assessment offish production in fresh waters, p. 215-226. Blackwell Scientific Publications, Oxf. and Edinb. 401 EFFECT OF STARVATION ON THE HISTOLOGICAL AND MORPHOLOGICAL CHARACTERISTICS OF JACK MACKEREL, TRACHURUS SYMMETRICUS, LARVAE Gail H. Theilacker^ ABSTRACT Histological and morphological criteria were developed to assess the nutritional condition of laboratory-reEired jack mackerel, Trachurus symmetricus, larvae. A comparison of the histological features of fed and starved larvae revealed that the digestive tract and its associated glands were the first tissues to be affected by starvation. The extent of cellular deterioration increased with time of starvation. To classify larval condition, histological characteristics of the pancreas and gut were given numerical grades. The histological technique correctly classified 839c of the feeding and starving larvae. The morphometric analysis relied upon a stepwise discriminant analysis that used a combination of five measurements (standard length, head length, eye diameter, body depth at the pectoral, and body depth at the anus) to estimate individual larval condition. The morphometric method was as sensitive as the histological examination in determining whether or not a larva was fed or starved. Ultimately, these histological and morphological criteria may be useful for estimating larval survival in the field by assessing the condition of sea-caught larvae. Fishery scientists generally agree that observed fluctuations in recruitment of young fish to a fish stock may be the consequence of mortality during the larval stage. Because starvation is probably one of the principal causes of mortality (Hunter 1976a), a need exists to develop criteria for detect- ing the incidence of starvation in sea-caught specimens. Several scientists have suggested that the differences in body form between feeding and starving larvae could be used to identify the nutri- tional status of larvae caught in sea surveys. For example, Shelbourne ( 1957) based his assessment of the condition of ocean-caught plaice, Pleuronectes platessa, larvae on their external ap- pearance. Certain morphometric measurements also can be indicative of starvation. A decrease in thickness of the larval fish body has been corre- lated with starvation for several marine and freshwater fish larvae [herring, Clupea harengus, and plaice (Ehrlich et al. 1976); northern anchovy, Engraulis mordax (Arthur 1976); anchovy, E. japonica (Honjo et al. 1959; Nakai et al. 1969); pike, Esox lucius, and carp, Cyprinus carpio (Kos- tomarova 1962)]. Other morphological features (Ehrlich et al. 1976) considered to be indicative of 'Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, La Jolla, CA 92038. Manuscript accepted August 1977. FISHERY BULLETIN: VOL. 76, NO. 2, 1978. starvation in herring and plaice were a decrease in the angle of the pectoral girdle, a change in the ratio of the head to eye height (herring only), and a decrease in the relative condition factor. Coinci- dent with morphometric differences caused by starvation, Ehrlich et al. (1976) described his- tological changes in the gut and liver. The his- tological approach was used to classify yellowtail, Seriola quinqueradiata, larvae into "feeding," "semi-feeding," and "starving" groups by Umeda and Ochiai (1975). This technique was also effec- tive for diagnosing starvation in northern an- chovy larvae (O'Connell 1976). In both species, degeneration of cells of digestive organs was the best indicator for identification of starvation. Sev- eral other studies also have correlated starvation in fish larvae with degeneration of the digestive organs, mainly the gut. Kostomarova (1962) de- scribed a retardation in development of the gut in larvae of starved carp and pike and a reduction in the depth of the epithelial cells lining the gut. Reduced gut cell height was also reported for the larvae of starved yellowtail (Umeda and Ochiai 1975), herring, and plaice (Ehrlich et al. 1976). Morphological criteria are preferable to his- tological ones because they take much less time to determine and require no special preservation techniques. However, histological criteria may be more accurate for classifying individual larva. 403 FISHERY BULLETIN: VOL. 76, NO. 2 The purpose of this study was to develop mor- phological and histological criteria for assessing the nutritional condition of jack mackerel larvae and to evaluate these criteria by comparing their success in identifying fed and starved larvae reared in the laboratory. Ultimately, criteria based on these results may be useful for estimat- ing larval survival in the field by assessing the condition of sea-caught larvae. MATERIALS AND METHODS Jack mackerel eggs were collected by towing a 1-m (mouth diameter, 0.505-mm mesh) plankton net just below the sea surface at various locations between 20 and 200 mi (32 and 320 km) off the coast of southern California in June and July 1975 and in May 1976. The eggs were separated from most of the plankton at sea and then sorted by developmental stage at the Southwest Fisheries Center, La Jolla, Calif. Temperature was main- tained at 15°C during sorting and in the larval rearing containers. The light cycle was 12 h light and 12 h dark. Five hundred normally developing eggs from a single day's spawning were transfer- red into 100 1 black Kydex^ circular rearing tanks containing filtered seawater {5fxm, Cuno filtered). There were three experiments and two treatments in each experiment; larvae in one tank were of- fered food while those in the other were not. The fed larvae were given a diet of a naked dinoflagel- late, Gymnodinium splendens (50/ml), a rotifer, Brachionus plicatilis (30-40/ml), and a copepod, Tisbe sp. (1 or 2/ml). This feeding method has been described (Lasker et al. 1970; Theilacker and McMaster 1971; Hunter 1976b). Histological criteria were developed in the first two experiments. The sampling procedure and the number of larvae sampled differed depending on the requirements of the analysis. Collectively, a total of 152 larvae were examined. In the third experiment, usually 15 larvae were sampled daily for 5 days from the "fed" tank (n =69) and 3 days in the "starved" tank (n = 48). All larvae were examined both histologically and morphologi- cally. No dead larvae were sampled because the postmortem change which takes place in tissues of fish larvae, due to digestion by their own enzymes (autolysis), resembled antemortem destruction caused by starvation. Standard length of each ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 404 larva was measured on a slide; then seawater was removed and replaced by Bouin's fixative. Pre- serving individual larvae in this manner assured that each would be straight and flat, facilitating subsequent morphometric measurements. Five measurements were taken after preservation to monitor daily changes in larval body form and determine effects of starvation: standard length (SL, tip of upper jaw to tip of notochord), head length (HL, tip of upper jaw to cleithrum), eye diameter (ed), body depth at the pectoral (bd-1), and body depth at the anus (bd-2). Standard length shrinkage in Bouin's fixative was 11.5%. Next, measured larvae were prepared for histological examination using standard techniques. Larvae were transferred from Bouin's to 70% ethyl al- cohol after 24 h, dehydrated with an ethyl n-butyl alcohol series in a Fisher Tissuemation, and em- bedded in Paraplast-plus. The Paraplast paraffin blocks were frozen with fluorocarbon spray (Cryokwick) just before the larvae were serially sectioned at 5 fxm in a sagittal plane. The mounted sections were stained with Harris' hematoxylin and eosin and mounted in synthetic resin. Histological Grading System Recently O'Connell (1976) developed a numeri- cal, histological grading system to characterize the nutritional condition of individual northern anchovy larvae. He examined tissues of the larvae microscopically to determine the features of tissue microstructure that were affected by starvation. A grade was assigned to each feature based on the degree of similarity or dissimilarity of the his- tological microstructure between the starved lar- vae and fed larvae. I followed this method and modified it as necessary for the tissues of jack mackerel. Each of the histological characteristics used to assess starvation in jack mackerel larvae was evaluated and assigned a grade. A grade of "3" was given to a characteristic which resembled that in "normal or healthy" larvae, a grade of "2" was given to an intermediate condition, a grade of "1" to the starved condition. Since these criteria were established by comparing actively feeding, seem- ingly healthy larvae with moribund larvae, it was assumed that an average of the 12 graded features examined for each larva classified the larva into the correct nutritional group: "healthy group," average grade = 2.34 to 3.00; "intermediate group" = 1.67 to 2.33; and "starved" = 1.00 to THEILACKER: EFFECT OF STARVATION ON JACK MACKEREL 1.66 (the break points establish three equal groups). Data Analysis In the main, conclusions are based on the results of a stepwise discriminant analysis (SWDA). A discriminant analysis allows one to distinguish between two or more groups, given a set of vari- ables that describe the characteristics in which the individuals in each group are expected to dif- fer. In the stepwise discriminant analysis, all the variables are introduced into a SWDA computer program and the best set of variables, based on the generalized Mahalanobis distance (Rao 1952), is selected. The first variable chosen will usually be the one which gives the best score when classify- ing the individuals into their predetermined groups. The score is equal to the number correctly classified. The selection of each succeeding vari- able improves the score until a subset is chosen which is as good as the full set of variables for discriminating the groups. All variables not in- cluded in the final subset are considered superflu- ous, or not necessary for classification. RESULTS between the round cells. In starved larvae, mitotic activity was arrested and many of the cells were shrunken, which caused large clear areas to ap- pear between densely stained, atrophied cells. Liver (Figures 5-8) In the normal larval jack mackerel liver the hepatic cords were two cells thick. Within each hepatocyte, the nucleus (3) was regular in shape and distinct. The cytoplasm (4) was well dispersed with intracellular spaces, probably an area where glycogen and lipid are stored. Sinusoid areas, where metabolic exchanges take place between the hepatic cords, contained blood cells. After starvation for a few days, the liver atrophied, the cytoplasm condensed, stained darkly, and in- tracellular spaces had disappeared. There were focal degenerative or necrotic areas and accumu- lations of eosinophilic granules and masses. Nuc- lei were often irregular in shape and granules appeared around their periphery; these darkened areas are presumed to be condensed, inactive chromatin (Stein et al. 1975). The gallbladder (5) in starved larvae was always enlarged; normally it discharges its contents under the stimulus of food (Love 1970). Histological The histological condition of yolk-sac and ac- tively feeding larvae ("normal") was compared with that of 3-day starved larvae and many differ- ences were noted in the cells, tissues, and organs. The degree of apparent histological deterioration of 1- and 2-day starved larvae was intermediate between the "normal" and moribund status. This condition was termed "semi-starved" by Umeda and Ochiai (1975) and "intermediate" by O'Con- nell(1976). The following section describes the normal his- tology of actively feedingjack mackerel larvae and that of starving larvae. Twelve histological criteria, which appeared to be indicators of starva- tion, were identified; they are numbered in the test below and are referenced in the photomicrographs (Figures 1-12). Brain (Figures 3, 4) The primitive brain cells exhibited a high inci- dence of mitotic activity (1) in normal larvae and there was relatively little intercellular space (2) Pancreas (Figures 5-8) Cells of the exocrine pancreas were arranged in series, in a circular fashion, as a secretory unit called an acinus. The nucleus (6), clear and dis- tinct, was located in the basal portion of the pyramidal cells. The acinar arrangement of the pancreatic cells (7) was found to be very sensitive to deprivation of food. A breakdown in the sym- metry in the acinus was slight but usually detect- able after 1 day of starvation. Tissue degeneration after 3 days of starvation was extreme. The nu- cleus was irregular and uniformly stained and there was no detectable acinar arrangement. The presence of zymogen, digestive proenzyme se- creted by the acinus and stored as granules at its central apex, was usually associated with starva- tion. Digestive Tract (Figures 7-12) The columnar epithelial cells of the midgut were closely united (8) in a single layer. Microvilli were visible along the border of the lumen giving a brush effect. In starved larvae, the midgut cells 405 FISHERY BULLETIN: VOL. 76, NO. 2 Pancreas Swim bladder Muscle Hindgut Notochord Midgut Brain Liver Heart Pancreas Notochord Swim bladder Hindgut Midgut Liver Gall bladder m^-^ N A RBO :^. itz ^a foodi' i^ BC-- M6 FG . M K W ^ Figure l. — Trachums symmetricus larva, day 8, fed for 3 days. All 12 histological criteria graded as "healthy." 32 x. Figure 2. — Trachums symmetricus larva, day 8, starved for 3 days. All 12 histological criteria graded as "starved." 32 x. Figure 3. — Head of fed larva, graded "healthy." Mitotic activity (1 h) is indicated. Note close proximity of primitive brain cells to each other. 200 x. h = histological grade = healthy. Figure 4. — Head of starved larva, graded "starved." Atrophied and darkly stained primitive brain cells (1 s); large intercellular spaces (2 s). 200 x. s = histological grade = starved. Figure 5. — Day 8 fed larva. Histological features graded "healthy." 200 x . See enlargement. Figure 7. B = brain, BC = blood cells (white or immature red), FG = foregut.I = Islet of Langerhans (endocrine pancreas), K = kidney, L = liver, MG = midgut, M = mus- cle, N = notochord, P = exocrine pancreas, RBC = red blood cells, SB = swim bladder. Figure 6. — Day 8, starved larva. Histological features graded "starved." 200 x. See enlarge, Figure 8. GB = gallbladder, O = oil, Y = yolk, see Figure legend 5 for rest of symbols. Figure 7. — Day 8 fed Trachurus symmetricus larva. Enlargement of Figure 5. Midgut cells in close union (8 h); prominent nuclei (3 h) and large intracellular spaces (4 h) in the liver; pancreatic nuclei distinct (6 h) and cells arranged in a circular unit (acinus) (7 h); gallbladder (GB) "normal." 480 x. h = histological grade = "healthy," L = liver, MG = midgut, P = exocrine pancreas. 406 THEILACKER: EFFECT OF STARVATION ON JACK MACKEREL * etr ^''Wj MG.Sh »^ . L ;^5k 7 M,llh MG EC^ Hi M 10 •'^— ^ II Figure 8. — Day 8, starved Trachurus symmetricus larva. Enlargement of Figure 6. Loss of integrity of midgut cells (8 s); atrophied liver with dark staining and irregular nuclei (3 s); no acinar cellular arrangement in pancreas (7 s); separated muscle fibers (11 s); swollen kidney (K); distended gallbladder (GB); note presence of yolk and oil (Y, O), and eosinophilic mass (EM) in liver. 480 x. s = histological grade = "starved," M = muscle, MG = midgut, P = exocrine pancreas. Figure 9. — Day 8 fed larva. Large eosinophilic inclusions in hindgut (10 h); muscle fibers closely packed (11 h); thin, epithelial integumental cells (EC) are prominent below gut and above trunk musculature. 200 x. h = histological grade ="healthy," HG = hindgut, M = muscle, MG = midgut. Figure id. — Day 8, starved larva. Loss of cellular structure in hindgut; enlarged epithelial integument cells (EC); separated muscle fibers (11 s). 200 x. s = histological grade = "starved," HG = hindgut, M = muscle, N = notochord, SP = spinal cord. Figure ll. — Day 7 larva, starved 2 days and histologically graded "intermediate." No inclusions in hindgut; cellular separation in midgut (MG) and hindgut (HG); midgut cells sloughing (9 s); muscle fibers beginning to separate (11 i). 200 x. i = histological grade = intermediate, s = histological grade = "starved," M = muscle, N = notochord, SP = spinal cord. Figure 12. — Day 7 larva, starved 2 days, graded "intermediate." Abundant intermuscular tissue (12 h); no muscle (M) fiber separation; pancreatic nuclei (6 i) not distinct and cellular acinar arrangement lacking (7 i); midgut cells separating (8 i) and sloughing (9 s). 480 X. i = histological grade = intermediate, s = histological grade = "starved," MG = midgut, P = exocrine pancreas. 407 FISHERY BULLETIN: VOL. 76, NO. 2 began to separate from each other. It appeared that the midgut was extremely vulnerable to a deficiency of food and usually after 1 day of starva- tion, single mucosal cells could be seen sloughed (9) into the lumen. The margin of the lumen con- tinued to lose its integrity as starvation advanced. Cells of the hindgut of feeding larvae exhibited an accumulation of eosin staining inclusions. The in- clusion bodies, which may be the sites of intracel- lular digestion, have been observed in other marine and freshwater fish (Kostomarova 1962; Iwai and Tanaka 1968a, b; Iwai 1968; Umeda and Ochiai 1975; O'Connell 1976). The amount, size, and intensity of the staining (10) of these inclu- sions varied in feeding larvae. They were not pres- ent in starving larvae. Musculature (Figures 9, 10, 12) In feeding jack mackerel larvae, individual muscle fibers were close together (11); they were composed of closely packed, striated, and parallel myofibrils. After a period of starvation, the fibers separated, the fibrils were not distinct, and occa- sionally they lost their parallel structure. Be- tween some fibers there was a granular, basophilic, nucleated substance called "intermus- cular tissue" (12) by O'Connell (1976). In starved larvae, this tissue was usually absent. General Histological Characteristics After jack mackerel larvae had starved for 3 days, signs of depletion were widespread. In addi- tion to changes in major tissues and organs there was a general atrophy and disintegration of all cells and tissues including those of cartilage, kid- ney, endocrine pancreas, and swim bladder. The number of pyknotic nuclei (i.e., darkening and shrinking nuclei, which give the first indication that a cell is dying) increased in all tissues (see eye and brain, Figure 4). Epithelial cells of the in- tegument were hypertrophic, twice as large as normal in 3-day starved larvae (Figure 10), and kidney tubules were swollen (Figure 6). There was always a larger yolk reserve retained by starving larvae (Figure 6). A decrease in yolk absorption in starving larvae was also reported by Kostomarova (1962) for pike and carp and by Umeda and Ochiai (1975) for yellowtail. Histological Grading To determine whether the classification of a jack mackerel larva required the grading of all his- tological features or a lesser number, a group of 27 408 larvae, 14 feeding and 13 starving, was examined and the resulting grades for each criterion were submitted to a SWDA. The experimental treat- ment (fed or starved) was unknown until after all larvae were microscopically examined. The larvae were 7 days old and had been feeding or starving for 2 days. The grading system classified all fed larvae (n = 14) into the healthy group (individual average grade of the 12 histological features ranged between 2.42 and 2.92). The average grades for the 2-day starved larvae were more variable. The larvae were classified, about equally, into each of the three nutritional groups: four had a grade range between 2.35 and 2.54, ranking in the healthy group; four were classified as intermediate, grade range 2.08 to 2.31; and five larvae were ranked as starved with the average grades ranging between 1.15 and 1.54. Results of the SWDA on the above data disclosed that grading only two histological characteristics, the arrangement of the cells in the pancreas (vari- able 7 ) and the sloughing of mucosal cells from the midgut (variable 9), gave the same conclusions as using all 12 features. Therefore, in all subsequent histological assessments, the average grade of these two criteria, variables 7 and 9, was used as the index of larval condition. Morphological The jack mackerel larvae were 2.45 mm SL (pre- served) at hatching and initiated feeding at 3.35 mm, 5 or 6 days after hatching (hatching = day 0, Figure 13A). Atthetimeof first feeding, some yolk and oil were present but the yolk sac was not discernible. The relationship between the five morphological characteristics, measured to de- termine the effects of starvation, and days of star- vation is illustrated in Figure 13B-F. Since no data have been published on the daily growth rate of field-caught jack mackerel larvae, I used length as an estimate of age. When the morphometric measurements were plotted against length, no single measurement was a reliable index of star- vation, as illustrated by pectoral body depth plot- ted for fed and starved larvae (Figure 14). How- ever, some limits can be set from this graph: 1) all larvae <3.30 mm SL that do not have a yolk sac probably are starving (feeding is initiated at 3.35 mm); and 2) larvae with a body depth >0.47 mm are feeding. This leaves the size class between 3.30 and 3.55 mm where the cases cannot be separated. Most individuals in this class (29 fed and 24 THEILACKER; EFFECT OF STARVATION ON JACK MACKEREL 4.2^ A • FED A STARVED 0.60 E E 0.70 X Q- 0.601- UJ o Q 0.50- o iij a. 0.40- 0.30 6 7 8 9 10 h-^-^ 6 7 8 9 DAYS 10 'g 0.35 E ui ijj >- UJ 0.30- < Q 0.25 0.20 8 10 Figure 13. — A. Growth ofTrachurus symmetricus larvae, with means and standard deviations. Sample size on any day was usually 15. B-F. Relationship between five morphological characteristics of T. symmetricus larvae and days of starvation, with means and standard deviations. Sample size was usually 15. starved) have been feeding or starving for 1 or 2 days. To determine whether a set of several mor- phometric variables could predict the condition of larvae in the 3.30 to 3.55 mm size class, a SWDA was run. Eleven morphometric variables, in which the two predetermined groups (fed and starved) were expected to differ, were entered into the SWDA: 1) HL, 2) ed, 3) bd-1, 4) bd-2, 5) HL/SL, 6) ed/SL, 7) bd-l/SL, 8) bd-2/SL, 9) ed/HL, 10) bd-1/ 409 FISHERY BULLETIN: VOL. 76, NO. 2 Figure 14. — Relationship between pectoral body depth and standard length of Trachurus symmetricus lar- vae which were fed (open circles), starved 1 and 2 days (dots), and starved 3 days (x's). 6 E bJ Q >- Q O CD < cr o I- UJ Q. U./'S 070 - 0 0 o 065 — 0 0 060 - 0 00 0 OCD O 0 0 0 0 055 - 0 OCD 0 0 00 050 - X 0 CDOOA O 0 0 Oflftfto 0 CDOOO ^ 0 aco 045 X X X X X >«-perplanei de- rived by describing these regions as multivariate normal density distributions with common variance-covariance matrices i Welch 1939 ». Quadratic discriminant functions have been de- veloped ' Smith 1947 '. The resulting decision sur- faces are nonlinear. The quadratic discriminant function does not require common variance- covariance matrices. Anas and Murai > 1969 > com- pared the classificatory abilities of the linear and quadratic discriminant functions. They found ' in agreement with Isaacson 1954' that even if the assumption that the distributions have common variance-covariance matrices is violated, the linear discriminamt function would still give good results for large sample sizes. But the quadratic function gave slightly bener results. All investigators utilizing discriminant analyses to separate races of Pacific salmon have assumed that the density distributions of mea- surements from a particular class of salmon were multivariate normal. The frequency distributions of scale characters in Major et al. ' 1975 > show that multimodal and skewed distributions occilt for chinook salmon scale characters even in the uni- variate case. In many other cases, the underh-ing distribution functions may be non-Gaussian. Dis- criminant functions based upon non-Gaussian dis- tributions or obtained by distribution-free methods are preferable to those based upon an unrealized assumption of normality. Nearly all of the discriminant function analyses used in the investigations of Pacific salmon have been two-class analyses designed to determine the continent of origin of salmon taken on the high seas. For the two-class situation only one discrim- inant function need be calculated. These two-class problems are a special case of the many-class prob- lems in which a separate discriminant function is calculated for each class. Bilton and Messinger <1975> calculated discriminant functions for each of several runs in a classification study on sockeye salmon. If several stocks of salmon intermingle and are to be classified, analyses of this t>-pe are needed. Spechts 1966' polynomial discriminant method does not require that the underhing den- sity distributions be multivariate normal nor that they have common variance-covariance matrices. Since this method is nonparametric. various scale characteristics may be used for discrimination with no particular regard to the underlying dis- tributions. Thus, the method is flexible and practi- cal. Specht 1966' uses an estimated probability density function of the form described by Parzen 1 1962 1 and extended by Murthy > 1966 • to the mul- tivariate case. The underh"ing multivariate den- sity for each class is modeled by a sum of functions that are multivariate Gaussian in form, one such function for each fish in the learning sample for that class. This set of functions is complete. There- fore, for each class the underhnng continuous probability density. Gaussian or not, may be ap- proximated arbitrarily closely by such a sum. A power series expansion of this estimated denisty then results in a pohTiomial term in the density function, the coefficients of which are functions of the obser\"ations i fish i in the learning sample. One such set of coefficients is computed for each class to be considered. These polynomials determine the nonlinear decision surfaces and are the basis for discrimination. The indi\'idual multivariate Gaussian functions I which when summed model the underlying mid- tivariate distribution for that class' contain a "'smoothing parameter." cr. which appears in the place of a standard error. This parameter is then incorporated in the estimates of the pol\*nomial coefficients. The reader is referred to Specht ' 1966 1 for a discussion of the effect of this smoothing parameter and for the algorithm for the calcula- tion of the sets of pohTiomial coefficients { -Dfe ■ k- ... fefj ( • The f>olynomial discriminant func- tion is: P(X) = Do ^ Z)iXi - D0X2 - DpX, '11^1 D^^X■^~ ' . . . + Df^^f^^Xf^^Xf^^ D^pX\ ^ 1^11 A' + . . . + Dfe^;;2fe3^fei^fe2'^fe3 "^ DpppX p Dk. kh ^fei -^fei 416 COOK and LORD: IDENTIFICATION OF STOCKS OF SOCKE'i'E SALMON where p = dimension of the vector X (set of scale characters ) 1 h,P' (X) for all s =f r where d(X) p'iX) /2; = the decision on an unknown X the classes (origins) the polynomial value for X calculated using the discrimi- nant function for class B; the a priori probability, the uses of which will be de- scribed later. APPLICATION OF THE METHOD Three scale sample sets are required to imple- ment the polynomial discriminant method: learn- ing samples, test samples, and onknowTi samples. The learning and testing samples are collected from each subpopulation when they are segre- gated (i.e.. in the rivers of origin i. Scale characters to be measured in the unknown sample for the required discrimination are determined by evaluating characters measured in the learning samples. The learning samples and the characters selected are used to calculate the coefficients in the pohTiomial discriminant functions. To calculate these coefficients, the value for the smoothing parameter and the point at which the discrimin- ant function should be truncated must be deter- mined. Various circumstances will dictate differ- ent choices. When a smoothing parameter of 1.5 was chosen, all terms in the discriminant function greater than the fourth order contributed negligi- bly to polynomial values and so were truncated in our applications. Often, polynomial discriminant functions of lower order yield adequate results.^ *A poKiiomial discriminant function with six variables and of the fourth order will contain 210 terms. Since our calculations were performed by computer, we chose not to delete the third or fourth degree terms. However, if more than six vso-iables are used, it would be wise to truncate further in order to keep the number of terms down. The fish comprising the test samples are classified to test the effectiveness of the polynomial discrim- inant method and to determine the a priori prob- abilities. 'Each test sample consists offish from one class. I Finally, fish collected from the zone of intermingling are classified to determine the de- gree of intermingling in the area of interest. Appraisal of the method using scale samples of sockeye salmon collected from the 1967 escape- ment in five Bristol Bay rivers showed large per- centages of fish comprising the test samples were correcth- classified. However, misclassified fish in the test group • set of test samples from all rivers being considered i were not assigned to the rivers in proportion to the known relative test sample sizes. To balance these misclassifications. wher- ever a greater number of fish comprising the test group was assigned to a particular river than should have been i according to the relative test sample sizes ). the a priori probability for that river w-as lowered. Corresponding increases were made for those classes with insufficient assignment. By alternatively using the decision procedure of the polynomial discriminant method and adjusting the a priori probabilities, we obtained solutions so that the number offish belonging to a certain river that were misassigned to all other rivers approxi- mately equaled the number of fish misassigned to that certain river from all other rivers. Thus the a priori probabilities were not used in the manner their name suggests, but a priori knowledge may dictate test sample sizes. The relative test sample sizes in the test group may be in the relative pro- portions to be expected in the unknown sample ( i.e.. historical relative run sizes >. The adjustment procedure, then, shifts the nonlinear decision sur- faces between the probability- densities so that the incorrectly identified samples are assigned to the various rivers in the proportions dictated by the test sample sizes in the test group. However, the primary purpose of the adjustment procedtire is not to balance the misclassifications but to maximize the number of correct classifications. As the misclassifications are balanced, the number of correct classifications generally increases. At this point the result is a classification method that maximizes the total number of correct classifica- tions and balances misclassification rates for a test group in which the test sample sizes are in particu- lar proportions. However, it is obvious that the proportions of fish from the various classes in the test group would rareh- be identical to those proportions in 417 FISHERY BULLETIN: VOL. 76, NO. 2 the unknown sample. Thus, imbalance among the misclassified fish will recur, unless the expected accuracy of classification is very good (near 100%). We have devised a method to correct for this. Based upon the results of classification of the known test group, the classification matrix, C, is estimated: C = C\\ Ci2 21 •'22 - C„i C„2 In 2" c where c,j is an estimate of the fraction of fish allocated to class i belonging to class j, such that 2 Cjj = 1.0, Vy. (Note that for each 7 the i = i c,j 's are a set of estimated multinomial prob- abilities and that each test sample size should be adequate.) If the discrimination is error-free, C would be an identity matrix. The adjustment of a priori probabilities causes the initially estimated classification matrix to evolve to the point where CT =Rf such thatT ^R,. The ith component of the vector T is the fraction of fish in the test group from test sample i (class /), and the ith component of the vector J?, is the fraction offish in the test group allocated to class i by the adjusted polynomial discriminant method. The test samples comprising T are not indepen- dent of the classification scheme since they are used to determine the a priori probabilities used in the decision rule. Hence, the estimated prob- abilities in the classification matrix may not be unbiased. However, we did chi-square tests that show elements of the classification matrix are not significantly different when estimated with either the test samples used to determine the a priori probabilities or a second independent test group. Thus, we prefer to use only one test group to de- termine the a priori probabilities and to estimate the elements of the classification matrix because the test sample sizes will be larger (and the var- iance of the Cy's smaller) if we do not subdivide the fish available. Now, let u , be the fraction of fish in a sampled group that belong to the ith class. The vector U is then unknown except for the obvious side condi- tion 2 u,- = 1. The classification matrix now ; = 1 operates on U to give: CU =i?/ where the ith component of/?,, is the fraction of fish in the unknown sample allocated to the ith class. Since C is estimated, R ,, is known and since C is usually nonsingular, we can estimate U by U =C ' R^. Each point estimate iu,) obtained will have some variance. This variability will depend upon the accuracy with which fish from class i are classified, the accuracy with which the elements of C are estimated, and variance due to sampling error en- countered when obtaining the unknown sample. Thus, if any u , is small, then its estimate ( u , ) may be negative. Such solutions are meaningless. In such cases the classes with negative solutions should be dropped (assume such u, ~ 0) and the analyses repeated. We did simulation work to evaluate the classi- fication matrix correction procedure for the two- and three-class situations. Five hundred simu- lated experiments were done for each situation. For the two-class case the average error of the classification results was 0.100 while that of the corrected estimates was 0.055. In 84% of the exper- iments the corrected estimate was closer to the true value than classification result. For the three-class case the average error of the classifica- tion results was 0.127 while that of the corrected estimates was 0.054. In 89% of the experiments the corrected estimate was closest to the true value. The results of these simulations show that the classification correction procedure improves estimates of the true proportion of a class present. This classification matrix correction procedure will reduce to the correction procedure developed for the two-class case by Worlund'' in the following manner: *A similar relationship and a least squares solution technique is given by Worlund and Fredin (1962). ■'Worlund, D. D. 1960. A method for computing the variance of an estimate of the rate of intermingling of two salmon popula- tions. Unpubl. manuscr., 13 p. Bur. Commer. Fish., Biol. Lab., Seattle, Wash. 418 COOK and LORD: IDENTIFICATION OF STOCKS OF SOCKEYE SALMON -1 U = C" R or Generally Since substitution yields u^ = ■"l" 'C22''l - Cl2'-2 C\\Ci2 - ^21^12 "2 Cll'-2 - C2l'"l /11C22 ~ C21C12 = 'i^ii - ^un "« CiiCjj - CjiCij ' 0 = I-'-/, Cji = 1 - Cii , Cjj = 1 - Cij , H<; ;/ . n - Cij Cii which is the correction formula of Worlund and Fredin (1962) (except for differences in notation and terminology) that has been used in many two-class Pacific salmon stock identification studies. Application to Sockeye Salmon Samples Taken in High Seas Sampling A problem of interest to the nations bordering the North Pacific Ocean is the origin of sockeye salmon taken on the high seas. The rivers of origin of sockeye salmon south of the central Aleutian Islands in summer are of particular interest to the United States since an index of their overall rela- tive abundance is used to forecast the numbers of mature fish returning to Bristol Bay in the follow- ing year (Rogers 1975). These fish are primarily of Bristol Bay origin (Hartt 1962, 1966; Hartt et al. 1975). Knowledge of the relative abundance of the various runs of the Bristol Bay stock south of the central Aleutians would be useful for forecast pur- poses and might provide insight into the high seas life history of the various runs. In order to recognize age 2.2 immature sockeye salmon on the high seas in 1976, the freshwater growth patterns of scales from three of the major rivers in Bristol Bay were examined.® Scales from the smolt outmigrations of 1974 for the Kvichak and Naknek Rivers were used as learning and *Age designation indicates fish which migrated to sea after two winters in freshwater and have spent two winters at sea. They are expected to return from the ocean primarily at age 2.3, or after sf)ending three winters at sea. testing samples. For the Egegik River scales from age 2.2 adult fish returning to spawn in 1976 were used as learning and testing samples because smolt scales were unavailable. The freshwater scale patterns offish from these runs were used to classify the sockeye salmon captured south of Adak Island during summer 1976 after having spent two winters in the ocean. The scale patterns were examined under a mi- croprojector of the type described by Dahlberg and Phinney (1968). The widths of the summer, winter, and plus growth zones were measured in terms of circuli counts and distance. The width of the widest circulus was also measured. Each scale character was then ranked over all classes (rivers) and the Kruskal-Wallis statistic (Kruskal and Wallis 1952) calculated. The difference between the average sum of ranks for each pairwise class combination was also calculated. On the basis of these statistics the scale characters providing the best univariate separation were selected for use in the polynomial discriminant method. Highly de- pendent scale characters were not used. By examining the learning samples, six scale characteristics were chosen for use in the polyno- mial discriminant method: 1) The number of the circuli in the first winter growth zone, 2) the number of circuli in the second summer growth zone, 3) the number of circuli in the plus growth zone, 4) the width of the first summer growth zone, 5) the width of the second winter growth zone, and 6) the width of the widest circulus.® Learning sample sizes of 25, 25, and 24 for the Egegik, Kvichak, and Naknek River classes, respectively, were used to calculate the coefficients in the polynomial function for each class. The classi- ficatory ability of these functions was then tested. The relative test sample sizes for each class were determined by examining run size data. According to the average run sizes of age 2.3 salmon for the last 8 yr approximately equal numbers offish from each class were expected to occur in the unknown sample. However, since the Kvichak River test sample size was twice that of the Egegik or Nak- nek River sample size, the fish in the latter test samples were given a weight of 2 when the a priori 'It should be mentioned that all data points were "nor- malized." That is, the mean and standard deviation for each scale character were calculated from the learning samples (all categories combined). All data points were then transformed by subtracting off" the mean and dividing by the standard deviation for the appropriate scale character. This is done for numerical purposes. 419 FISHERY BULLETIN: VOL. 76, NO. 2 probabilities were adjusted. After adjusting the a priori probabilities, we obtained the results given in Table 1. The classification matrix was then es- timated: C = 0.800 0.040 0.167 0.080 0.740 0.208 0.120 0.220 0.625 where the subscripts of the matrix elements ( c,^'s) were 1, 2, and 3 for the Egegik, Kvichak, and Naknek River classes, respectively. Seventy-two percent of the fish in the test group were correctly classified. The fish in the high seas sample were then classified with the adjusted polynomial dis- criminant method. Of the 101 sockeye salmon, 25 were classified as Egegik River fish, 22 as Kvichak River fish, and 54 as Naknek River fish. The resultant vector was: Ru = 0.267 0.222 0.511 The estimated unknown vector was thus: C~' R.. = 1.300 0.037 -0.360 0.267 -0.078 1.498 -0.478 0.222 -0.222 -0.534 1.837 0.511 0.17f 0.067 = U. 0.761 Based upon preliminary data for the 1977 Bristol Bay sockeye salmon run from the Alaska Depart- ment of Fish and Game, the actual unknown vec- tor was: U 0.325 0.061 0.614 The classification matrix correction procedure gave a slightly better estimate than the direct results of the polynomial discriminant method. The differences between the u, 's and the u , 's were due to bias and variability. (We are presently examining methods to reduce the variability of our i/,'s.) A problem with the high seas sample is that some of these sockeye salmon originate in rivers other than those considered. Although the three Table l. — Results of the polynomial discriminant method on a known test group of Bristol Bay sockeye salmon. The a priori probabilities were 0.340, 0.332, and 0.328 for the Egegik, Kvichak, and Naknek River classes, respectively. Calculated Correct decisions Total (all calcu- decisions Egegik Kvichak Naknek lated decisions) Egegik Kvichak Naknek Total (all correct decisions) 40 4 6 50 2 37 11 50 8 10 30 48 50 51 47 148 classes considered will account for nearly all of the age 2.2 sockeye salmon bound for Bristol Bay, some may be non-Bristol Bay fish. When the Bris- tol Bay runs are at a low point in their cycle, up to 20% of the high seas sockeye salmon at Adak Is- land may be non-Bristol Bay fish (Hartt et al. 1975). The possible bias from classifying the non- Bristol Bay fish into the classes established should be considered since 1977 is a low year in the sock- eye salmon run cycle. In conclusion, the polynomial discriminant method can be used to identify certain runs of sockeye salmon on the high seas by differences in freshwater scale growth patterns. Possibly the relative proportions of sockeye salmon that will be returning to inshore areas can be predicted. Even- tually the method will be used to predict one year in advance the relative run sizes to the major Bris- tol Bay rivers by sampling these sockeye on the high seas. Application to Inshore Fishery Stock Separation A problem of interest to the Alaska Department of Fish and Game is the separation of stocks in commercial catches in inshore areas, particularly the separation of Kvichak, Naknek, and Egegik River sockeye salmon. The Division of Commer- cial Fisheries is collecting data on scale measure- ments for growth studies. They are interested in how well these data and the polynomial discri- minant method can separate Bristol Bay sockeye salmon stocks. Scale data from samples of the 1973 spawning escapement were examined. Each of two age- classes was examined separately. Distance and circuli counts to both the freshwater and saltwater annuli were examined for use in the polynomial discriminant method with the Kruskal-Wallis and multiple comparison procedures. The accuracy of classification for age 1.2 and age 2.2 sockeye salm- 420 COOK and LORD: IDENTIFICATION OF STOCKS OF SOCKEYE SALMON on was examined for each age-group with known test groups. The degree of separation for age 1.2 sockeye salmon is shown in Table 2. (Egegik River fish are historically insignificant in this age-class.) The scale characters providing this separation were: 1) the circuli count to the first annulus, 2) the dis- tance to the first annulus, 3) the distance from the first to the second annulus, 4) the distance from the second to the third annulus, 5) the circuli count from the third annulus to the edge of the scale, and 6) the distance from the third annulus to the edge of the scale. Ninety-five percent of the fish in the test group were correctly classified. The degree of separation for age 2.2 sockeye salmon is shown in Table 3. The scale characters providing this separation were: 1) the circuli count to the first annulus, 2) the distance to the first annulus, 3) the circuli count from the first to the second annulus, 4) the distance from the second to the third annulus, 5) the distance from the third to the fourth annulus, and 6) the circuli count from the fourth annulus to the edge of the scale. Seventy-seven percent of the fish in the test group were correctly classified. Thus, the polynomial discriminant method can provide adequate separation with a given data base. The data collected for growth studies provide good separation in some cases. Sockeye salmon from the Egegik, Kvichak, and Naknek Rivers are distinguishable in terms of these scale measure- ments and it should be possible to estimate their relative proportions in catch samples. Table 2 . — Results of the polynomial discriminant method on 1 . 2 age Bristol Bay sockeye salmon from 1973. The a priori prob- abilities were 0.52 and 0.48 for the Kvichak and Naknek River classes, respectively. Calculated Correct decisi ons Total (all calcu- decisions Kvichak Naknek lated decisions) Kvichak Naknek Total (all correct decisions) 18 2 20 0 19 19 18 21 39 T.\BLE 3. — Results of the polynomial discriminant method on 2.2 age Bristol Bay sockeye salmon from 1973. The a priori prob- abilities were 0.342, 0.330, and 0.328 for the Egegik, Kvichak, and Naknek River classes, respectively. Calculated Correct decisions Total (all calcu- decisions Egegik Kvichak Naknek lated decisions) Egegik Kvichak Naknek Total (all correct decisions) 20 1 5 26 3 22 1 26 3 4 14 21 26 27 20 73 COMMENTS AND CONCLUSIONS The key to successful implementation of the polynomial discriminant method is the choice of scale characters that reflect differences between the subpopulations of concern. The scale charac- ters that are most likely different are those that are formed when the populations are geographi- cally separated. Genetic and environmental influences on scale formation probably interact to create these differences. Although it is likely that no single characteristic will provide the required separation, a group of characteristics analyzed with multivariate techniques ( e.g. , the polynomial discriminant method) will often provide this re- quired separation. The polynomial discriminant function technique requires no consideration of the underlying probability density functions for these scale characters because these density func- tions are estimated nonparametrically. Once the characters that provide the best separation are determined (by rank order comparison procedures in this paper) the discriminant function analysis may be implemented. A learning sample is needed to calculate the discriminant function for each subpopulation. These fish comprising these samples must be col- lected before or after the populations intermingle (either as smolts or returning adults in the respec- tive rivers). Learning samples must be taken from the same year class and freshwater age-group as the unknown (mixed) population if the scale characters are known or thought to vary from year to year. Using Specht's (1966) algorithm and the data from these learning samples, the coefficients in the discriminant functions are calculated. The next step is to appraise the effectiveness of these polynomial discriminant functions. By classifying a group of test samples the pro- portion of correctly identified fish and the clas- sification error rates can be determined. The pro- portion of correctly identified fish will likely be low until a good set of a priori probabilities is deter- mined. As the a priori probabilities are adjusted to balance the classification error rates, the propor- tion of correctly identified fish will generally in- crease. The proportion of correctly identified fish, when the classification error rates are satisfactor- ily balanced, gives an indicator of the effectiveness of the polynomial discriminant method. The clas- sification error rates specific to these final a priori probabilities are now estimated so that they may be corrected for when the polynomial discriminant 421 method is applied to the unknown mixed sample. This is done with the classification matrix correc- tion procedure. First, the fish in the unknown mixed sample are classified with the polynomial discriminant method (using the adjusted a priori probabilities). The proportions resulting for each subpopulation and the decision matrix allow simple algebraic solution for the estimated true proportions of the various subpopulations in the zones of interming- ling. Estimates of this type are often needed in par- ticular management situations involving Pacific salmon. By using scale samples and the polyno- mial discriminant method, the proportions of the major classes present in areas where the subpopu- lations mix can be estimated. We have considered only two possible applications in this paper: high seas monitoring for predictive purposes and the analysis of catch samples. Many other possibilities exist for other situations and other salmon species: the timing of inshore runs could be examined in estuarine areas or in river systems, the continent of origin of salmon on the high seas could be examined (for those species or areas not already analyzed), or the intermingling of hatchery and native populations could be analyzed for certain fisheries. Since scale samples are relatively easy to collect and exchange and since computers are readily available to do the necessary calculations, the polynomial discriminant method is a flexible and practical tool for the racial analysis of Pacific salmon, particularly sockeye salmon. ACKNOWLEDGMENTS Many thanks are due to Colin Harris, Allan C. Hartt, and Robert L. Burgner for their editorial advice and guidance. We also wish to thank James B. Scott for his meticulous work with the scale data. Many others providing services to the deep sea tagging project at the Fisheries Research In- stitute deserve thanks. The scale samples and data from the Bristol Bay river systems were col- lected by the Alaska Department of Fish and Game. We are grateful to Paul Krasnowski and other Alaska Department of Fish and Game per- sonnel for providing these vital samples. This re- search was primarily supported by NOAA, Na- tional Marine Fisheries Service, under Contract No. 03-6-208-35470. The inshore stock separation project was supported by the Alaska Department of Fish and Game. FISHERY BULLETIN: VOL. 76, NO. 2 LITERATURE CITED Amos, M. H., R. E. Anas, and R. E. Pearson. 1963. Use of discriminant function in the morphological separation of Asian and North American races of pink salmon, Oncorhynchus gorbuscha (Walbaum). Int. North Pac. Fish. Comm., Bull. 11:73-100. Anas, R. E. 1964, Sockeye salmon scale studies. Int. North Pac. Fish. Comm., Annu. Rep. 1963:158-162. Anas, R. E., and S. Mural 1969. Use of scale characters and a discriminant function for classifying sockeye salmon (.Oncorhynchus nerka) by continent of origin. Int. North Pac. Fish. Comm., Bull. 26:157-192. BILTON, H. T., AND H. B. MESSINGER. 1975. Identification of major British Columbia and Alaska runs of age 1.2 and 1.3 sockeye from their scale characters Int. North Pac. Fish. Comm., Bull. 32:109-129. DAHLBERG, M. L., AND D. E. PHINNEY. 1968. A microprojector for use in scale studies. Prog. Fish-Cult. 30:118-120. Dark, T. A., and B. J. Landrum. 1964. Analysis of 1961 red salmon morphological da- ta. Int. North Pac. Fish. Comm., Annu. Rep. 1962:110- 115. Das Gupta, S. 1973. Theories and methods in classification: A review. In T. Cacoullos (editor), Discriminant analysis and appli- cations, p. 77-137. Academic Press, N.Y. Fisher, R. A. 1936. The use of miltiple measurements in taxonomic problems. Ann. Eugen. 7:179-188. Fukuhara, F. M., S. Murai, J. J. LaLanne, and a. Sribhibhadh. 1962. Continental origin of red salmon as determined from morphological characters. Int. North Pac. Fish. Comm., Bull. 8:15-109. Hartt, A. C. 1962. Movement of salmon in the North Pacific Ocean and Bering Sea as determined by tagging, 1956-1958. Int. North Pac. Fish. Comm., Bull. 6, 157 p. 1 966. Migrations of salmon in the North Pacific Ocean and Bering Sea as determined by seining and tagging, 1959- 1960. Int. North Pac. Fish. Comm., Bull. 19, 141 p. Hartt, A. C, G. E. Lord, and D. E. Rogers. 1975. Monitoring migrations and abundance of salmon at sea. Int. North Pac. Fish. Comm., Annu. Rep. 1973:73-79. ISSACSON, S. L. 1954. Problems in classifying populations. In O. Kempthome, T. A. Bancroft, J. W. Gowen, and J. L. Lush (editors), Statistics and mathematics in biology, p. 107- 117. Iowa State Coll. Press, Ames. KONOVALOV, S. M. 1971. Differentiation of local populations of sockeye salm- on Oncorhynchus nerka (Walbaum). (Translated by L. V. Sagen, 1975. Univ. Wash. Pub. Fish., New Ser., 6, 290 p.) KRUSKAL, W. H., AND W. A. WALLIS. 1952. Use of ranks in one-criterion variance analysis. J. Am. Stat. Assoc. 47:583-621. Major, R. L., S. Murai, and J. Lyons. 1 975. Scale studies to identify Asian and western Alaskan 422 COOK and LORD: IDENTIFICATION OF STOCKS OF SOCKEYE SALMON Chinook salmon. Int. North Pac. Fish. Comm., Annu. Rep. 1973:80-91. Mason, J. E. 1966. Sockeye salmon scale studies. Int. North Pac. Fish. Comm., Annu. Rep. 1964:117-118. Ml'RTHY, V. K. 1966. Nonparametric estimation of multivariate densities with applications. In P. R. Krishnaiah (editor), Mul- tivariate analysis, p. 43-56. Academic Press, N.Y. Parzen, E. 1962. On estimation of a probability density function and mode. Ann. Math. Stat. 33:1065-1076. Patrick, E. a. 1972. Fundamentals of pattern recognition. Prentice- Hall, Englewood Cliffs, N.J., 504 p. ROGERS, D. E. 1975. Forecast of the sockeye salmon run to Bristol Bay in 1976. Univ. Wash., Fish. Res. Inst. Circ. 76-1, 49 p. Smith, C. A. B. 1947. Some examples of discrimination. Ann. Eugen. 13:272-282. SPECHT, D. F. 1966. Generation of polynomial discriminant functions for pattern recognition. Stanford Univ., Tech. Rep. 6764-5, 127 p. Welch, B. L. 1939. Note on discriminant functions. Biometrika 31:218-220. Worlund, d. D., and r. a. FREDIN. 1962. Differentiation of stocks. In Symposium on pink salmon, p. 143-153. H. R. MacMillan Lectures in Fisheries, Univ. B.C., Vancouver, Can. 423 EFFECTIVENESS OF ESCAPE VENT SHAPE IN TRAPS FOR CATCHING LEGAL-SIZED LOBSTER, HOMARUS AMERICANUS, AND HARVESTABLE-SIZED CRABS, CANCER BOREALIS AND CANCER IRRORATUS^ Jay S. Krouse^ ABSTRACT During 1976 a study was conducted to find an escape vent that would select similar sized lobsters as the rectangular vent, yet retain Cancer crabs s90 mm carapace wddth. Analysis of the size composition of research and commercial catches from experimental traps revealed that circular (58 mm in diameter) and rectangular (44.5 x 152.4 mm) vents release shorts and retain legal lobsters ( 3^81 mm carapace length) equally well, and decidedly more marketable-sized crabs were captured in traps with circular vents. Length-width relationship shows that crabs 3=90 mm carapace width have lengths 3^58 mm, thus precluding the possibility of marketable-sized crabs exiting through an opening 58 mm in diameter. Escapement studies for lobsters confirm that with the present minimum legal size of 3^/i6 in, a 58-mm diameter vent vnll select legals and allow most of the sublegals to escape. Accordingly, the Maine Department of Marine Resources recommends that either circular ( 358 mm in diameter) or oblong (3=44.5 x 152.4 mm) escape vents be incorporated in all crab and lobster traps along the Maine coast. Although rectangular escape vents are a very beneficial type of savings gear for the lobster fishery (Templeman 1939; Wilder 1945, 1948, 1954; Krouse and Thomas 1975; Krouse 1976), this vent does not retain marketable-sized rock crab, Cancer irroratus, and Jonah crab, C. borealis. Since these commercially important crab species are often caught incidental to lobsters, I undertook the present study to find an escape opening that would retain harvestable-sized crabs and have similar fishing selectivities for the lob- ster, Homarus americanus, as the rectangular vent. In designing a trap to catch crabs and exclude lobsters, Stasko (1975) observed in laboratory tests that circular holes retained commercial- sized crabs yet allowed small lobsters to escape; however, the effectiveness of escape holes was not tested in the field. Jow (1961) demonstrated the advantages of circular escape openings in the trap fishery for Dungeness crab, C. magister. In this paper I evaluate the relative efficiency of 'This study was conducted in cooperation with the U.S. De- partment of Commerce, National Marine Fisheries Service, under Public Law 88-309, as amended, Commercial Fisheries Research and Development Act, Project 3-153-R. ^Maine Department of Marine Resources, West Boothbay Harbor, ME 04575. circular and rectangular vents by examining data from: 1) commercial and research catches com- piled from vented and nonvented traps; 2) studies of escapement from traps; and 3) certain mor- phometric relationships of crabs and lobsters. METHODS From November 1976 through March 1977 a commercial fisherman recorded and provided me with catch data from traps with circular vents (58-mm diameter) fished alongside traps without vents. This experimental gear was arranged into two groups with four trawls [series of six traps spaced about 6 fathoms ( 11.0 m) apart with a sur- face buoy at either end] per group. In each group half the traps in a trawl had no vents, while the remainder had either single (end of trap) or paired (side of trap) vents depending upon the group (Fig- ure IB, D). Every time the fisherman hauled these traps he recorded the following information: 1 ) number of days traps were set between hauls; and 2) number of lobsters &81 mm carapace length, CL (keepers), and <81 mm CL (shorts) caught in the vented and nonvented traps for each trawl string. From July through November 1976, project per- sonnel fished commercial lobster traps near Manuscript accepted July 1977 FISHERY BULLETIN: VOL. 76, NO. 2, 1978. 425 FISHERY BULLETIN: VOL. 76, NO. 2 7>\ ^ D Figure l. — Lobster traps having a rectangular vent positioned vertically (A) and horizontally (C) and single (B) and paired (D) circulai vents. Boothbay Harbor, Maine, with: 1) circular [58-mm (2.3 in) and 61-mm (2.4 in) diameter] vents; 2) vertical and horizontal rectangular [44.5 mm ( 1.8 in) X 152.4 mm (6.0 in)] vents; and 3) traps with- out vents. Carapace length of lobsters was mea- sured from posterodorsal edge of eye socket to posterior margin of carapace and carapace width (CW) of crabs, distance between the two m^ost pos- terior notches on the anterolateral border of the carapace, to the nearest millimeter. Trap escapement was studied by placing lobsters of known sizes in traps with circular open- ings of 58, 60, and 61 mm in diameter. Side en- trances of each trap were closed so escapement had to be via the vents. Traps were secured to the laboratory dock and usually checked daily for es- capement for about a week. To determine whether or not a crab or lobster could pass through a round opening of a given size, we correlated carapace length of lobsters with carapace height (CH), and the carapace width of crabs with carapace length . For 2 1 7 lobsters ( sexes combined), ranging from 70 to 98 mm CL, carapace height was determined by positioning the lobster's ventral surface on a flat board and then measuring the greatest perpendicular dis- tance from the board to the top of the carapace. Carapace length of crabs was measured from the anterior margin of the frontal region to the poste- rior border of the intestinal region. Measurements for the two Cancer species were treated separately due to the species disparities in body shapes. We recorded carapace length for 103 male rock crabs (females were excluded due to commercial unim- portance) ranging from 90 to 122 mm CW, and 96 Jonah crabs (sexes combined) ranging from 96 to 132 mm CW. RESULTS AND DISCUSSION Lobsters in Research Gear There are marked differences in size composi- tion and number of lobsters caught in nonvented 426 KROUSE: EFFECTIVENESS OF ESCAPE VENT SHAPE Table l. — Lobsters caught with nonvented and various types of vented traps from July through November 1976. Catch Catch effort Mean Vent type Total no. Sublegals; legals carapace length (mm) Standard error Legals per trap haul No of trap hauls Months fished Nonvented 749 4.3:1 76.2 ±0.28 0.53 265 July-Nov. Horizontal 198 0.6:1 839 ±062 0.54 229 July-Nov. Vertical 107 0.5:1 84.3 ±1.05 0.56 129 July- Sept. Circular: 58 mm 25 0.6:1 828 ±1.35 0.30 53 Oct. -Nov. 61 mm 42 0.4:1 85.9 ±1.35 0.47 66 Sept. and vented traps (Table 1). Vented traps caught fewer sublegal lobsters per trap-haul than non- vented traps (t-test, P<0.01). The ratio of sublegal to legal lobsters did not differ among the four types of vents (^-test, P>0.1), with the exception of 61-mm circular vents which caught fewer sublegals than horizon- tal vents (P<0.01). As will be discussed later, the 61-mm hole is slightly oversize for a minimum size of 81 mm CL, thus some smaller legal lobsters and most shorts escape. Nevertheless this information suggests that circular openings are as effective as the rectangular vent ( Krouse and Thomas 1975) in permitting escapement of short lobsters. To further assess the relative efficiencies of the various vents, catch-effort values (numbers of lobsters per trap haul set over day, CPUE) were calculated and plotted for legal-sized and all-sized lobsters combined for each vent type (Figure 2). For this figure, 58- and 61-mm circular vent data were pooled because of the small sample size and similar catch values. Figure 2 graphically shows that the CPUE for legal-sized lobsters was similar for all vent types; however, for combined catches of legals and sublegals, the CPUE for nonvented traps was several fold greater. Thus, this indicates that all traps tested were about equally efficient in capturing legal lobsters; but, as to be expected, nonvented traps caught substantial numbers of short lobsters which probably would have escaped from vented traps. Most importantly, these data support an earlier conclusion that circular vents select about the same size lobsters as do rectangu- lar vents. Lobsters in Commercial Gear Catch data provided by a local lobsterman were compiled according to the following categories of gear: 1) end vented traps with a single circular hole of 58 mm diameter (Figure IB); 2) side vented traps with paired round openings of 58 mm diame- 4.2 r 3.9 3.6 3.3 _l =) 3.0- < ^ 2.71- a. < 2.4 CE >- 2-1 S 1.8 a. 1.5 90 mm CL had ^58 mm CH. Based on this relationship alone, it appears that many lobsters ranging from 8 1 to 89 mm CL would be able to squeeze through a 58-mm diameter hole; however, this is refuted by the previous sections on the commercial and research catches of lobsters with circular vented traps and the following dis- cussion of escapement studies. Lobster escape- ment through a round opening cannot be accu- rately determined by carapace height alone since this measurement excludes the walking legs which contribute to the lobster's overall height or depth. Whether or not a lobster is successful in passing through a round hole will be determined not only by the lobster's greatest transverse di- mension (carapace height plus protruding legs) but also by the lobster's ability to maneuver through a tight opening. Obvious limitations with the aforementioned morphometric relationship caused me to seek an alternate approach to assess escapement. Thus, I decided to determine the largest size lobster that could be manually passed through a 58-mm diameter hole. Lobsters 81 mm CL passed through the hole rather easily following careful manipula- tion of the walking legs and 82-mm CL lobsters required considerable force, often causing bodily harm, while larger lobsters ( >82 mm CL) could not pass through the opening. Patterns of escapement for lobsters ranging from 78 to 84 mm CL from traps with 58-, 60-, and 61-mm diameter vents varied decidedly as de- picted by retention curves in Figure 6. Only the 58-mm vent retained all legal-sized lobsters and still had reasonably high escapement of sublegals; whereas, the other vents which were merely 2 or 3 mm larger allowed legal-sized lobsters to escape. These data emphasize the importance of accu- rately producing the 58-mm opening, else the vent's desired effect will be lost. Crabs Carapace width-length relationships for C. borealis and C. irroratus graphically show that crabs >90 mm CW (commercially harvested size) have carapace lengths (dimension limiting es- capement) which exceed 58 mm (Figures 7, 8). Accordingly, commercial-sized crabs of either species cannot egress through a circular opening 58 mm in diameter. In fact, if the vent diameter 77 78 79 80 81 CARAPACE LENGTH (mm) 83 84 Figure 6. — Retention curves for lobsters placed in lobster traps with circular vents of 58, 60, and 61 mm in diameter. were increased to as large as 65 mm (certainly, an over estimate) to accommodate an upward shift in the lobster minimum size (Maine Department of Marine Resources recommends an increase from 3^/i6 to 3y2 in CL by Vi6-in increments annually over a 5-year period) this would have little or more likely no effect on catches of marketable crabs. RECOMMENDATIONS In view of the findings of this study and past investigations (Krouse and Thomas 1975; Krouse 1976), all lobster and crab traps fished in Maine waters should have a rectangular escape vent not less than 1.75 in (44.5 mm) by 6 in ( 152.4 mm) or at least two circular escape vents not less than 2.28 in (58 mm) in diameter. To insure maximum es- capement of sublegal lobsters, vents should be in- stalled next to the sill on the side or end of the trap's parlor section. Although fishermen should certainly have the option to fabricate their own vents, provided that the prescribed dimensions are adhered to, the use of synthetic, prefabricated vents is highly recom- mended (Krouse and Thomas 1975). Recently, a plastics manufacturer assured me that vents could be produced and retailed for about 200 each. At this low price and with today's high price of laths (about 50 each), if a synthetic vent replaces two laths every 3 yr, then after 6 yr the original cost of 430 KROUSE: EFFECTIVENESS OF ESCAPE VENT SHAPE 90r 95 /oPredtction Intervals 95 ''o Confidence Intervals 110 120 CARAPACE WIDTH (MM) 30 Figure 7. — Carapace width-carapace length relationship for male rock crabs with 95^^ confidence and prediction intervals. 90- 5 X O § 80 LU U < < < 70 60f- 90 --♦a 95% Prediction Intervals 100 110 120 CARAPACE WIDTH(MM) I3O Figure 8. — Carapace width-carapace length relationship for Jonah crabs with 'dWc confidence and prediction intervals. 431 the vent will be defrayed by the replacement cost of the laths, resulting in a cost savings. Therefore, those fishermen interested in captur- ing only lobsters and, perhaps, minimizing their crab catches, would be encouraged to use rectan- gular vents, while fishermen interested in both lobsters and crabs or solely the latter should employ circular vents. ACKNOWLEDGMENTS I thank David A. Libby for his assistance in field collections, data compilation, and figure prepara- tion; and Charles Begin, a commercial fisherman, who often at an inconvenience to himself, fur- nished me with invaluable catch information. LITERATURE CITED Jow, T. 1961. Crab trap escape-opening studies. Pac. Mar. Fish. Comm. 5:49-71. KROUSE. J. S. 1972. Some life history aspects of the rock crab, Cancer irroratus, in the Gulf of Maine. J. Fish. Res. Board Can. 29:1479-1482. 1973. Maturity, sex ratio, and size composition of the natural population of American lobsters, Homarus FISHERY BULLETIN: VOL. 76. NO, 2 americanus, along the Maine coast. Fish. Bull., U.S. 71:165-173. 1977. Incidence of cull lobsters, Homarus americanus, in commercial and research catches off the Maine coast. Fish. Bull, U.S. 74:719-724. KROUSE, J. S., AND J. C. THOMAS. 1975. Effects of trap selectivity and some population parameters on size composition of the American lobster, Homarus americanus, catch along the Maine coast. Fish. Bull., U.S. 73:862-871. Miller, R. J. 1976. North American crab fisheries: Regulations and their rationales. Fish. Bull., U.S. 74:623-633. Stasko, a. B. 1975. Modified lobster traps for catching crabs and keep- ing lobsters out. J. Fish. Res. Board Can. 32:2515-2520. TEMPLEMAN, W. 1939. Investigations into the life history of the lobster (Homarus americanus) on the west coast of Newfound- land, 1938. Newfoundland Dep. Nat. Resour., Res. Bull. (Fish.) 7, 52 p. Thomas, J. C. . 1973. An analysis of the commercial lobster (Homarus americanus) fishery along the coast of Maine, August 1966 through December 1970. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-667, 57 p. Wilder, D. G. 1945. Wider lath spaces protect short lobsters. Fish. Res. Board Can., Atl. Biol. Stn. Circ. G-4, 1 p. 1948. The protection of short lobsters in market lobster areas. Fish. Res. Board Can., Atl. Biol. Stn. Circ. G-11, 1 p. 1954. The lobster fishery of the southern Gulf of St. Law- rence. Fish. Res. Board Can., Gen. Ser. Circ. 24, 16 p. 432 SCHOOL STRUCTURE OF THE SQUID LOLIGO OPALESCENS Ann C. Hurley' ABSTRACT The squid Loligo opalescens forms schools which are similar in many respects to those of obligate schooling fishes. These schools are marked by parallel orientation of individuals and strong cohesive- ness. Laboratory experiments indicate that the main sensory modality regulating schooling is vision. Squid on opposite sides of a clear rigid Plexiglas barrier readily schooled. The structure of schools of six squid depended on size of individuals and was modified by environmental disturbance. Parallel orientation was weaker in schools of smaller squid (ca. 7 cm dorsal mantle length) than it was in larger ones (ca. 12 cm). In the field, L. opalescens schools are composed of uniformly sized individuals. Laboratory experiments designed to determine whether this was due to actual size selection were inconclusive, but they did suggest mechanisms which might be important in determining squid position in the school. Considerable effort has been spent in understand- ing the schooling behavior of fish in terms of physiological mechanisms and possible survival value and ecological consequences. (See reviews by Radakov 1973 and Shaw 1970, 1978.) Virtually no work has been done on schooling behavior of invertebrates which occur in the same environ- ments as schooling fish. The most evident school- ing invertebrates in the pelagic environment are the squid. Squid and fish play very similar ecologi- cal roles and the two groups of organisms possess a large number of similarities. (See Packard 1972, for a discussion of convergent evolution.) Loligo opalescens is common off the west coast of North America with a reported range from Baja California to lat. 55°N (Fields 1965; Bernard 1970). Relatively little is known of the behavior or general ecology of L. opalescens in spite of the fact that there is a fishery for this species in California. The fishery is based primarily upon the tendency of squid to spawn in large aggregations in shallow water (McGowan 1954; Fields 1965). Very little is known about the distribution or location of newly hatched squid as well as squid in later stages of life. Attempts to catch the juveniles have often been unsuccessful (Okutani and McGowan 1969) and only recently have attempts been made to catch nonspawning adults. Even though field data are difficult to obtain, it is possible to keep both juvenile (Hurley 1976) and adult L. opales- cens alive in the laboratory. Schooling in the laboratory was examined to provide insights about the function of schooling in squid. METHODS The squid used in the behavioral studies were obtained either by dipnetting them after they had been attracted to an underwater light or by pur- chasing them from a local bait dealer. In the laboratory, the squid were maintained in a 3-m diameter circular tank with rapidly circulating seawater. They were fed irregularly on small fish (either mosquitofish, Gambusia affinus, or goldfish, Carassius auratus). Mosquitofish were taken much more readily than were the goldfish. Occasionally, the squid could be trained to take dead food. This was accomplished by first getting them to accept live fish and by then throwing dead fish in along with the live ones. In this manner, the squid could also be coaxed to accept pieces of frozen northern anchovy, Engraulis mordax. If the squid were undamaged when they arrived at the laboratory and there was an abundant supply of small fish available, it was relatively easy to keep them for over a month. Experiments designed to examine various as- pects of schooling behavior were run in a 2 x 3 m rectangular Plexiglas^ tank which was filled to a depth of 0.4 m. The tank was painted flat white and the primary source of lighting (in addition to general room illumination) was provided by 'Moss Landing Marine Laboratories, P.O. Box 223, Moss Landing. CA 95039. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Manuscript accepted November 1977. FISHERY BULLETIN: VOL. 76. NO. 2. 1978. 433 FISHERY BULLETIN: VOL. 76, NO. 2 fluorescent lights placed around the perimeter of the tank which shone through the walls. This pro- vided even, diffuse light in the tank. The water in the tank was noncirculating. Schooling behavior was recorded on Tri-X film using a 35 mm camera with motor drive and a variable setting automatic timer. A mirror was placed above the tank and pictures were taken of the squid by photographing the surface image reflected in the mirror. A black plastic barrier surrounded the experimental tank. A small hole in the barrier allowed observations to be made of the squid without disturbing them. Exact methods, timing of pictures, etc. varied with the experiment and will be described in the appro- priate section. After the films were developed, they were analyzed using a Scientific Data Products data tablet (Graf-Pen) coupled with a PDP 11-45 com- puter. The data tablet is a set of microphones placed at right angles which record the sound pro- duced by an electrical spark made by a special marking pen. The x and y coordinates of a point were relayed to the computer by pressing the pen down on the tablet at that point. This device al- lowed the recording of large amounts of squid posi- tion data. In each frame, the tip of the tail, the tip of the outstretched arms, and a point midway be- tween the two eyes were recorded for each squid. Other information, such as the position of barriers, was recorded in the same way. Mea- surements taken from the photographs were sub- sequently converted to real distances by multipli- cation by appropriate scale factors. Students of schooling have examined school geometry both as a two-dimensional system on a horizontal plane (Breder 1959; Williams 1964; John 1964; Hunter 1966, 1968; Van 01st and Hunter 1970) and as a three-dimensional struc- ture (Cullen et al 1965; Symons 1971a, b; Pitcher 1973). Since squid schools do have a three- dimensional structure in nature, a three- dimensional analysis will eventually be necessary to determine all of the structural details of the school. A three-dimensional analysis, however, is much more difficult than a two-dimensional analysis. It was felt that a two-dimensional analysis would suffice to examine certain aspects of squid schooling behavior. In these experiments, the squid were very nearly confined to a two- dimensional plane by the shallow water depth in the experimental tank. Observations of small schools (up to six squid) in a deeper tank (1 m depth) indicated that the two-dimensional struc- ture observed in the experimental tank was not uncommon. Three indices were chosen to quantify the angu- lar orientation of individuals in a school, the over- all shape of the school, and the distance between neighboring individuals in a school. These indices were proposed by Hunter ( 1966) and he includes a detailed discussion of their properties. The three indices are: 1. Mean separation distance: An average of the horizontal distances separating each squid from every other squid in the school. It is influenced by school shape, distance between neighboring squid, and number of squid in the school. Distances be- tween all possible pairs of squid are measured and these values are averaged. Distance is measured between the two closest points on the midline of the bodies ( including outstretched arms) of the two squid. 2. Mean distance to nearest neighbor: An aver- age of distances from each squid in the school to its nearest neighbor. The measurement is made be- tween the two closest points on the midline of the bodies (including outstretched arms) of the two squid. The same measurement is used twice if two squid are closer to each other than to any other squid. 3. Mean angular deviation: This is a measure of the differences in orientation among squid within a school. The heading of each squid is determined and the resultant direction of the school is com- puted by assigning each squid a value of one and adding the headings vectorally. The mean number of degrees individual squid deviated from this re- sultant direction of the school was calculated as the index of orientation. One difference between squid and most of the schooling fish which have been studied is that squid readily swim both forward and backward. Thus, a squid with an orientation that was 180° out of phase with the rest of the school might still be swimming with the rest of the school. For this reason, one orientation measure was calculated which regarded the squid as a line segment rather than as a directed vector and measured the small- est angular deviation between line segments. Such measurements were rarely different from measurements made considering the orientation of the squid and therefore will not be considered further in this paper. 434 HURLEY: SCHOOL STRUCTURE OF LOLIGO OPALESCENS Where measurements of squid length are given, they are of dorsal mantle length from tip of the pen to the tip of the tail. The total length of the squid (including arms but excluding tentacles) is about 1.3-1.5 times the dorsal mantle length (Fields 1965). RESULTS Response to Disturbance One set of factors that caused changes in school- ing can be grouped under "external disturbances." These included introducing objects (such as a net) into the water near the squid or tapping on the side of the tank. The typical response was for the squid to group more tightly and. in cases where it was not already marked, to increase the degree of parallel orientation. The amount of change in schooling behavior and the temporal characteris- tics of this change depended upon the nature and intensity of the disturbance and upon its duration. One attempt at quantifying the stimulus in- volved placing an aquarium air stone in the tank. Pressurized air delivered to this air stone in differ- ing amounts and duration produced a stream of bubbles which could be used as a disturbance stimulus of varied intensity and duration. A small stream of bubbles produced little squid reaction, while vigorous water action due to the bubble stream produced marked changes in behavior. Figure 1 shows the changes in three of the school- ing indices in response to a moderate disturbance caused by turning on the air bubble stimulus. The degree of parallel orientation, which was already pronounced, did not change appreciably. But the squid did draw noticeably closer together. Schooling Structure as a Function of Squid Size Six squid of nearly equal size were haphazardly taken from the holding tank and placed in the experimental tank. The squid swam in this tank for an hour before measurements were made. With the exception of one experiment, a picture was taken of the squid every minute for approximately 1 h. During this other experiment, a picture was taken every 10 s for 10 min. This set of experi- ments was conducted during the daylight hours of two different days. All of the squid used in this set of experiments had been captured on the same night. There was a decrease in the mean angular de- viation as the size of the squid increased (Table 1). Since small values of the mean angular deviation index are associated with increased parallel orien- tation, the degree of parallel orientation is greatest in schools composed of large individuals. Even in the case of the small individuals, however, the value of the index does not approach what would be expected if the squid were each orienting in a random direction. In a simulation of 1 million values for six randomly oriented fish. Hunter (1966) found that the mode of the frequency dis- tribution was 69°. Although the average values for mean angular deviation do give a measure of average departure from parallel orientation for a whole experiment, they do not give an indication of how variable a particular group of squid is in its orientation over time. For example, an experiment of 30 pictures and an average value of the mean angular devia- tion index of 20° could have had all of the 30 values close to 20°. This would indicate a consistent mod- erate degree of parallel orientation over time. On the other hand, such an average value could also come from a situation where the squid had strong parallel orientation part of the time and were much more loosely oriented the rest of the time (e.g., the index value could have been 10° on 15 frames and 30° on 15). This kind of difference can be detected if a measure of the variability of the mean angular deviation index for each experi- ment is calculated. The variability (standard de- viation, SD, Table 1) increased with decreasing squid size, indicating that not only do the smaller squid not orient on the average in as parallel a manner as larger squid, but they are also more temporally variable in their orientation. This dif- ference can also be seen if individual experiments are examined. Figure 2 shows the values for mean separation distance and mean angular deviation Table l . — Relationship between average size of Loligo opales- cens and parallel orientation and separation of individuals in the six-squid experiments. Each index was calculated for each frame. Group number Mean mantle length (cm) No. frames examined Mean angular deviation in- dex (degrees) Mean separation distance Index (cm) X SD X SD 1 7.5 44 32.0 18.4 32.3 26.5 2 7.6 58 290 15.3 25.6 8.0 3 7.7 60 18.1 10.7 162 4.5 4 9.7 19 18.5 5.7 14.0 4.2 5 97 62 16.2 62 18.7 5.2 6 102 69 17.2 6.6 20.9 5.0 7 11.9 65 11.1 56 18.5 4.0 8 120 55 9.6 28 15.3 2.7 9 12.0 46 9.1 4.2 13.8 3.0 435 FISHERY BULLETIN: VOL. 76. NO 2 30 20 r¥m^. 20 CD C/1 FiGL'RE 1. — Values of schooling indices for a school of six Loligo opalescens be- fore and after disturbance (turning on bubbler). Dashed line indicates when air was turned on. Pictures were taken every minute for 64 min. 30 --^ 20 LU •—< s:° 10 for two (Groups 3 and 7) of the experimental runs summarized in Table 1. The parallel orientation is stronger and the variability less in the larger squid (Figure 2C, D). The mean separation dis- tance index is not as clear a function of size (Table 1; Figure 2A, B). Schools in the Ocean Very little is known of the natural behavior of Loligo opalescens when it is not in large mating schools. In many areas, there often is a large con- centration of squid in the vicinity of the deep- scattering layer (C. Recksiek, Moss Landing Marine Laboratories, Moss Landing, CA 95039, 436 TIME pers. commun., October 1976) and large layered concentrations of L. opalescens have been reported by those involved in submersible exploration (A. Flechsig, Sea Grant Marine Advisory Service, University of California at San Diego, La Jolla, CA 92093, 1973). There is evidence to indicate, however, that L. opalescens is often found in much smaller schools and that these schools contain a narrow size range of individuals. Fields (1965) presents data on the uniformity of size of young squid taken from the same fish catch (presumably the same squid school ) and speculates that the size ranges in the schools he observed represent ap- proximately one-half or less than one-half of a year's growth. HURLEY: SCHOOL STRUCTURE OF LOLIGO OPALESCENS 30 A 20 . 2 10 V / V" .' S-. \/\ \y \/ 30 . 20 10 - ^/\/\ A.; WW n 10 ~T — 20 30 MINUTES ~T — 40 — I- 50 60 •a MINUTES Figure 2. — Values of mean separation distance and mean angular deviation for two (Groups 3 and 7) of the experiments presented in Table 1. Mean size ofLoligo opalescens in the experiment represented in A and C was 7.7 cm mantle length. Mean size of squid in the experiment represented in B and D was 11.9 cm mantle length. 437 FISHERY BULLETIN: VOL. 76, NO. 2 I also obtained data on the uniformity of size in individuals of the same school. The squid were caught during a 1-wk period in August in locations ranging from San Diego to Santa Catalina Island, Calif. A night-light was placed off the stern of the ship in the center of an L-shaped 3-m long mesh net. Squid were attracted to the light and would rush into the net. The net was then raised and the squid could be removed with dip nets. The "schools" were all of the squid which swam into the net at the same time. Squid caught during this period ranged from 5.8 to 17.3 cm dorsal mantle length. But for a given school, they were much more uniform in length. The average size range for 29 schools of 2 to 32 individuals was 2.5 cm. Maintenance of School Structure and Orientation Experiments in the laboratory have indicated that vision is sufficient sensory input to mediate schooling behavior. Squid on different sides of a clear, rigid Plexiglas barrier will readily school with each other and they appear to maintain the same type of parallel orientation that is present in normal schooling behavior. Preliminary experi- ments using such Plexiglas barriers were run to try to elucidate the mechanisms by which spacing is maintained. Two-Squid Experiments Experiments were run to determine whether squid would school in the same manner with or without a clear Plexiglas barrier in place. Mea- surements were obtained for squid swimming to- gether and for the same squid swimming on oppo- site sides of a Plexiglas barrier which divided the tank into two compartments. The order of the treatment was randomized for each pair of squid. Squid ranged in size from 7 to 13 cm mantle length. For a given experiment, the two squid were of similar length. Pictures were taken of the squid in each treatment every 10 s for 3 min after they first came together and again every 10 s for 3 min after the squid had been left undisturbed for 15 min. If the squid did not come together to within at least 0.5 m within 1 min, the experiment was terminated. Table 2 shows the results of five such experi- ments. The first 3-min periods have been compared with each other, as have the later runs. This was to see if the pattern of schooling changed after the Table 2. — Median nearest distances and median separation angles for two-squid (Loligo opalescens) experiments. With Without Item barrier barrier Difference' Nearest dis- 20.6 16.2 P = 0 05 barrier greater tance (cm) 113 6.6 P<0.01 barrier greater first 3 mm 38.6 11.65 P<0 01 barrier greater 14.9 7.9 P<0.01 barrier greater 13.8 6.4 P<0.01 barrier greater Nearest dis- 24.2 15.8 P<0.01 barrier greater tance (cm) 18.4 6.3 P<0 01 barrier greater second 3 23.9 18.35 P<0.05 barrier greater mm 14.35 13.1 P<0.05 barrier greater 12.5 8.2 P<0.01 barrier greater Separation 16.4 52.7 P<0.01 barrier less angles (de- 24.3 11.2 P<0 05 barrier greater grees) first 75.1 21.0 P<0 01 barrier greater 3 mm 21.2 12.2 NS^ 17.0 30.9 P<0.05 barrier less Separation 28.2 15.4 P<0.05 barrier greater angles (de- 15.4 13.6 NS grees) sec- 19.1 18.0 NS ond 3 mm 25.4 18.3 NS 11.1 24.9 P <0.05 barrier less 'Significance of difference in medians from IVlann- Whitney U-test. ^NS = no significant difference. P 0.05 squid became more adapted to the experimental regime. This table presents results for the median nearest distance between the two squid for each run and for the median separation angle for these same runs. Separation angle for each frame is simply a measurement of the angle between the two squid and is a measure of orientation (0° sep- aration angle indicating parallel alignment facing the same direction). The barrier has an effect upon the separation distance between the two squid. In all cases, there was a significant difference between the distance between squid with and without the barrier. When the Plexiglas barrier was present, the squid tended to space themselves farther apart. There is not a clear relationship between angular separa- tion and the presence of the barrier. Of the six runs showing significant differences, three had greater median separation angles with the barrier in place and three had greater median separation angles when the barrier was not present. Three-Squid Experiments The experimental tank was divided crosswise into three equal compartments (1 x 2 m each) by clear Plexiglas partitions. A squid was chosen from the holding tank and was placed in the cen- tral compartment. Then a squid for each of the outer compartments was selected. These squid were assigned at random to each of the outer com- partments. The squid were allowed to adapt to the experimental situation for 15 min and then were filmed for 5 min (one picture every 10 s). The two 438 HURLEY. SCHOOL STRUCTURE OF LOLIGO OI'ALESCENS outer squid were then switched from one outer compartment to the other and the squid were again allowed to adapt for 15 min. They were then filmed for 5 min (once every 10 s). Squid in these experiments ranged from 9.2 to 15.3 cm mantle length. It was hoped that this experimental design would indicate whether the center squid, if given a choice, would choose to school with a larger or smaller squid or one closer to its own size. One way to determine whether such a choice is being made would be to determine whether the center squid spends more of its time closer to one outer squid than to the other. Each 5-min run was considered as a unit and each frame was scored according to which outer squid the center squid was nearest. For each run, the data were compared with a binomial distribution which assumed that the center squid had an equal probability of being closest to either outer squid. Of the 17 runs, 16 showed a significant deviation from the expected binomial distribution (Ps:0.05 for 1; P^O.Ol for 15). These 16 runs were now grouped according to whether the center squid was closest to the larger or smaller outer squid. In 8 of the 16 cases, the center squid was nearest the larger outer squid, while in the other 8 cases, it was nearest the small- er squid. There is no evident preference for large versus small squid. The data can also be arranged to determine whether the center squid spent most of its time near the squid closer to its own size. There were 14 runs for which it was possible to say that the center squid was closer in size to one of the outer squid. Of these 14 runs, the center squid was significantly nearer to the squid closer to its own size 9 times and nearer to the squid farther from its own size 5 times. These experiments may be viewed in another way by looking at the absolute position of the squid in the tank. The nearest distances of the squid to the Plexiglas barriers were calculated for each frame. These data are summarized in Figure 3 for the 17 runs. The side squid usually are very near the barrier which separates them from the center compartment, while the center squid varies his position within the center compartment, but approaches the Plexiglas barriers much less often. DISCUSSION Pelagic fish and squid represent a striking case of convergent evolution, not only morphologically (Packard 1972), but behaviorally as well. One as- pect of behavior where this is particularly appa- rent is schooling. Since many of the same ecologi- cal pressures exist for both pelagic groups, it is not surprising that some sort of schooling behavior would have developed in both fish and squid. What is surprising, given the very different physiology and mode of locomotion, is that so many aspects of this behavior are the same. Loligo opalescens fits Breder's (1967) definition of obligate schoolers. Single L. opalescens are rarely caught in the field, and they immediately come together when placed in a tank in the laboratory. As has been reported for many species offish (Radakov 1973), L. opalescens schools con- sist of individuals of approximately the same size. It has been suggested that the reason that fish school in such groups has to do with swimming speed. Small and large individuals would not swim at the same speed and thus would not nor- mally stay together. This is possibly also true for squid, but data on the swimming speed of large and small L. opalescens are not available to sub- stantiate the argument. For several reasons, the swimming speed hypothesis seems less plausible for squid than for fish. In schools offish which show parallel orientation, the fish continually maintain forward motion and thus swimming speed is likely to be an important factor. But field and laboratory observations have indicated that individuals in squid schools spend much of their time hovering in the same position in the water column with only 16 14 12 10 8 6 . 0 2 . 0 m- Figure 3. — Histograms of mean nearest distance between Loligo opales- cens and barrier in the 17 three-squid experiments. Distances are broken up into 10-cm intervals. From left to right: left outer squid to left barrier, center squid to left barrier, center squid to right barrier, right outer squid to right barrier. 439 FISHERY BULLETIN: VOL 76, NO. 2 slight backward and forward motion caused by jets of water from the siphon. Even when disturbed, the squid do not make long extended swims which would tend to sort out those of differing swimming ability. In the field, the most common response of a squid school to a disturbance (the presence of a scuba diver or a shark) is to clump closer together and move off a slight distance. On one occasion when I was diving in a large spawning school (several thousand individuals), the squid executed the same type of maneuver that has been reported for fish schools. Instead of moving off, the school completely enclosed me, leaving a spherical space of approximately 3 m radius around the "pred- ator." One other piece of evidence suggests that it is not differences in swimming speed alone which cause the squid to school according to size. While diving in the Bahamas in the Hydrolab underwa- ter habitat, we observed a school of squid which routinely visited the habitat. This school was composed of Doryteuthisplei, a species which quite closely resembles L. opalescens and presumably has similar swimming ability. This school con- sisted of seven squid and, in this case, was not composed of individuals of the same size, the largest individual being at least two times the length of the smallest individual. We chased this school several times but were never able to force them to separate. The smallest squid maintained the same swimming speed as the largest squid. It is possible that squid maintain schools of in- dividuals of a fairly narrow size range because of social factors. Generally, workers studying school- ing have assumed that all of the fish in a school may be treated as equivalent individuals in the production of the behavior and that there is no social structure in the schools. In fact, some work- ers have suggested that schooling is really just a modified form of individual cover-seeking be- havior (Williams 1964; Hamilton 1971). This as- sumption of equality of individuals may be an untenable one for squid schools. In the field, Hochberg and Couch ( 1971 ) observed signaling by some members of a school of Sepioteuthis sepioidea which they felt prevented other squid from joining the school. Furthermore, in the laboratory, I have observed complicated agressive interactions in L. opalescens which certainly demonstrate that all squid cannot be considered behaviorally equiva- lent individuals at all times (Hurley 1977). One aspect of schooling in fish which has been emphasized by many workers is that the structure of schools may change as a function of the age or the physiological condition of the fish. Van 01st and Hunter ( 1970), for instance, found that in five species of marine fishes, schools of young fish were less compact and showed greater differences in angular headings than did schools of adult fish. In addition. Hunter ( 1966) showed that distances be- tween jack mackerel tended to increase with food deprivation, while Keenleyside (1955) noticed that sticklebacks were more densely packed in a school when well fed than when starved. I attempted to determine whether similar phenomena were observable in squid schools. Schools of small squid ( 7-9 cm mantle length) gave the impression of being less cohesive than schools of larger squid ( 13-15 cm mantle length). This was supported to some extent by the quantitative mea- surements, particularly those of angular orienta- tion. The variability was also higher for all of the indices for the smaller squid. It has been suggested for fish (Van 01st and Hunter 1970) that the ob- served change with size could have been an adap- tation to the higher food requirements of the juvenile fish. This speculation is supported by ob- servations that a number of species school less cohesively under conditions of food deprivation. The same explanation may also hold for squid, but my existing data do not support it. I ran two exper- iments in which schools of six squid were filmed before and after feeding. In one experiment, there were no significant differences in the schooling indices before and after feeding, while in the other, there was significantly less school cohesion and parallel orientation after feeding. In any event, it is not possible to guess which factors are instru- mental in this increased cohesiveness and consis- tent geometry. As is the case for fish, vision seems to be the primary sensory system used in squid schooling. Squid will readily school across a clear, rigid Plexiglas barrier, although they tend to stay somewhat farther apart than they normally would. Investigators dealing with fish also found that the presence of a clear, rigid barrier caused abnormalities in the spacing between individuals, in some cases increasing the fish-to-fish distance ( Cahn 1972) and in some cases decreasing it (Shaw 1969). These workers speculated that this change was due to lack of lateral line input and a resultant loss of information concerning the position of the adjacent fish. Squid do not have a similar exten- sive vibration-sensitive system, although they may be able to detect vibrations with their stato- 440 HURLEY; SCHOOL STRUCTURE OF LOLIGO OPALESCENS cysts. In the case of L. opalescens, the most likely explanation for this change in spacing is that the presence of the barrier physically limits the extent to which each squid can compensate for the other individual's movements. In the experimental tank, squid seemed to differ in their motivation to school. When the barrier was not present, a squid with a strong tendency to school could always maintain proximity to another squid. But if the barrier were present, that squid could only follow another squid as far as the barrier and had to remain there until the other squid returned. I had hoped that the experiments with the three squid separated by Plexiglas barriers would give some clue as to whether the squid actively chose to school with individuals of the same size, but the results were inconclusive. The results did indi- cate, however, a possible mechanism for mainte- nance of spacing within a school. The center squid tended to stay toward the middle of the compart- ment, while the side squid maintained positions very near the Plexiglas barrier. It is possible that the center squid was attempting to equalize the visual angle subtended by the squid on each side, while the outer squid were attempting to get into positions with squid on each side. The measure- ments of visual angle which I can get from my photographs are not accurate enough to determine whether this is happening. If outer squid are con- tinually trying to achieve a position where they have squid on either side of them, individuals in a school should be continually shifting positions. Casual observations have indicated that this does happen some of the time; but at other times, the individuals maintain the same positions relative to one another. An area where a comparison of squid and fish schooling may be useful is in the speculation con- cerning the evolution of the schooling behavior and its possible advantages. Many recent papers have concentrated on the hydrodynamic advan- tages offish schooling (e.g., Breder 1976) and base their explanations of many of the details of school structure on the fish mode of tail-flip locomotion and the vortices which are subsequently created. Van 01st and Hunter ( 1970) suggest that the typi- cal nearest neighbor distance in fish schools is about one-half a body length and that this distance may be explained by considering the amplitude of the tail beat in swimming. It is interesting that in squid, with their very different mode of locomotion (jet propulsion as opposed to tail flips), the spacing between nearest neighbors is still maintained be- tween one-half and one body length in undis- turbed squid. Other investigators have speculated that a primary function of schooling is as a defense against predation. ( See reviews by Shaw 1970 and Radakov 1973.) Squid have many of the same reactions to disturbance that fish do. They both clump more closely together as a result of distur- bance and both have been seen to surround their predators. Further evidence which suggests that predator defense may be an important function of squid schooling comes from the development of the behavior in juvenile squid. In the course of rearing L. opalescens (Hurley 1976), I made observations on schooling behavior. The newly hatched squid appeared to have no attraction to each other, but after 6 or 7 wk schooling was occasionally ob- served. This schooling was only evident in re- sponse to disturbance (tapping on the tank or put- ting a net into the water). When the squid were feeding undisturbed, there was no obvious school- ing behavior. ACKNOWLEDGMENTS I would like to thank P. Hartline, J. Hunter, and G. D. Lange for their assistance during this study; E. Shaw, G. Cailliet, and J. Nybakken for valuable comments and suggestions; and Rosemary Keegan for typing the manuscript. This work was supported by NIH grant NS-09342 and NSF grant GH-41809 to the laboratory of G. D. Lange, Uni- versity of California, San Diego, and the South- west Fisheries Center, La Jolla Laboratory, Na- tional Marine Fisheries Service, NOAA, while I held a NOAA associateship at that laboratory. LITERATURE CITED BERNARD, F. R. 1970. A distributional checklist of the marine molluscs of British Columbia, based on faunistic surveys since 1950. Syesis 3:75-94. BREDER, C. M., JR. 1959. Studies on social groupings in fishes. Bull. Am. Mus. Nat. Hist. 117:393-482. 1967. On the survival value of fish schools. Zoologica (N.Y.) 52:25-40. 1976. Fish schools as operational structures. Fish. Bull., U.S. 74:471-502. CAHN, p. H. 1972. Sensory factors in the side-to-side spacing and posi- tional orientation of the tuna, Euthynnus affinis, during schooling. Fish. Bull., U.S. 70:197-204, 441 FISHERY BULLETIN: VOL. 76, NO 2 CULLEN, J. M., E. Shaw, and H, A. Baldwin. 1965. Methods for measuring the three-dimensional struc- ture of fish schools. Anim. Behav. 13:534-543. FIELDS, W. G. 1965. The structure, development, food relations, repro- duction, and life history of the squid Loligo opalescens Berry. Calif Dep. Fish Game, Fish Bull. 131:1-108. Hamilton, W. D. 1971. Geometry for the selfish herd. J. Theor. Biol. 31:295-311. HOCHBERG, F. G., AND J. A. COUCH. 1971. Biology of cephalopods. In J. W. Miller, J. G. Van- Derwalker, and R. A. Waller (editors), Tektite 2. Scien- tists in the sea, p. VI-221-VI-228. U.S. Dep. Inter., Wash., D.C. Hunter, J. R. 1966. Procedure for analysis of schooling behavior. J. Fish. Res. Board Can. 23:547-562. 1968. Effects of light on schooling and feeding of jack mac- kerel, Trachurus symmetricus . J. Fish. Res. Board Can. 25:393-407. Hurley, a. C. 1976. Feeding behavior, food consumption, growth, and respiration of the squid Loligo opalescens raised in the laboratory. Fish. Bull., U.S. 74:176-182. 1977. Mating behavior of the squid Lo/tgoopa/escens. Mar. Behav. Physiol. 4:195-203. JOHN, K. R. 1964. Illumination, vision, and schooling of Astyanax mexicanus (Fillipi). J. Fish. Res. Board Can. 21:1453- 1473. Keenleyside, M. H. a. 1955. Some aspects of the schooling behavior of fish. Be- havior 8:183-248. McGowan, J. A. 1954. Observations on the sexual behavior and spawning of the squid, Loligo opalescens, at La Jolla, California. Calif Fish Game 40:47-54. Okutani, t., and J. A. McGowan. 1969. Systematics, distribution, and abundance of the epiplanktonic squid (Cephalopoda, Decapoda) larvae of the California Current, April, 1954 - March, 1957. Bull. Scripps Inst. Oceanogr. 14:1-90. Packard, A. 1972. Cephalopods and fish: the limits of convergence. Biol. Rev. Cambr. Philos. Soc. 47:241-307. PITCHER, T. J. 1973. The three-dimensional structure of schools in the minnow, Phoxinus phoxinus (L.). Anim. Behav. 21:673-686. RADAKOV, D. V. 1973. Schooling in the ecology of fish. John Wiley and Sons, N.Y., 173 p. Shaw, E. 1969. The duration of schooling among fish separated and those not separated by barriers. Am. Mus. Novit. 2373, 13 p. 1970. Schooling in fishes: critique and review. In L. R. Aronson, E. Tobach, D. S. Lehrman, and J. S. Rosenblatt (editors). Development and evolution of behavior, p. 452- 480. W. H. Freeman and Co., San Franc. 1978. Schooling fishes. Am. Sci. 66:166-175. SYMONS, P. E. K. 1971a. Spacing and density in schooling threespine sticklebacks (Gasterosteus aculeatus and mummichog {Fundulus heteroclitus). J. Fish. Res. Board Can. 28:999-1004. 1971b. Estimating distances between fish schooling in an aquarium. J. Fish. Res. Board Can. 28:1805-1806. VAN OLST, J. C, AND J. R. HUNTER. 1970. Some aspects of the organization of fish schools. J. Fish. Res. Board Can. 27:1225-1238. Williams, G. C. 1964. Measurement of consociation among fishes and comments on the evolution of schooling. Publ. Mich. State Univ. Mus. 2(7):351-383. 442 NORTHERN ANCHOVY SCHOOL SHAPES AS RELATED TO PROBLEMS IN SCHOOL SIZE ESTIMATION James L. Squire, Jr.' ABSTRACT Horizontal fish school profiles of the northern anchovy, Engraulis mordax, taken from day aerial photographs and video tapes of school bioluminescence at night were examined to determine the percentage of school area within a circular field of view and the school length and width ratios. Schools observed during the day had an average length to width ratio of 2.09:1, at night the ratio was 2.53:1. The percent coverage of the school's area in relation to a circle drawn tangent about the school averaged 42. 19c during the day and 29. 2^^ during the night. The effect of school shape on estimation of individual school area as observed with a side-looking sonar was determined. School width measurements, similar to that obtained by the sonar, were used to determine school area and indicated a possible average overestimate of the actual school area of 1.72:1. The relation of school length and width to the error was determined, indicating the gfreater the length to width ratio the greater the error. Profiles offish schools as viewed and photographed in the horizontal plane from an airborne platform have been published by numerous authors. Radakov (1972), in his review offish schooling, described the characteristic horizontal shapes of fish schools in nature as being very diverse and extremely changeable. He stated that a spherical shape of a school is the rarest of all and also that a school's shape, size, or density is a result of the interaction between the fish and the physical and biological environment. School shape and behavior in nature have been studied with such techniques as aerial observa- tion, hydroacoustic measurements, and underwa- ter observation. Each of these methods has limita- tions. Underwater visual observations are subject to restrictions due to illumination and restricted visibility. Aerial observation is limited in the day by water transparency, illumination, and reflec- tance from the water surface resulting from wind and wave action. Visual observation of school shape at night, as outlined by bioluminescent or- ganisms, is limited to the moon's dark cycle or to periods of no moon, and is affected by water trans- parency and the density of bioluminescent or- ganisms present in the water. Both day and night observations are limited by the school's proximity to the surface in relation to the factors affecting water visibility. Hydroacoustic observations using lower fre- 'Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. quency sounders of the type used commonly in sonar fish surveys give imprecise images of fish schools in the form of echograms that must be interpreted. Greater resolution of fish school shapes, but with limited range, can be obtained with ultrasonic scanning equipment (Voglis and Cook 1966). All of these observation techniques may alter the environment and in many cases may result in modification of fish school behavior. Fish school behavior is affected when in close proximity to ships, submersibles, and divers, and aircraft (noise, shadow) could possibly modify the school, though this is not documented. Surveys and research studies using variations of these three observation techniques are cur- rently in use for direct biomass estimation offish populations by observation of individual schools, school groups, and the internal structure of the school. Hydroacoustic research on schooling fish is cur- rently being conducted by the Southwest Fisheries Center (Smith 1970; Hewitt et al. 1976). Coastal, hydroacoustic surveys are conducted by the State of California (Mais 1974) to determine a relative abundance estimate of the northern anchovy, En- graulis mordax. These surveys are conducted dur- ing the daylight hours, as comparative tests indi- cate an increased probability of detection during this period (Smith 1970). Aerial observations by commercial fish spotters, in the form of school counts and estimates of total tonnage, are being used by the Southwest Manuscript accepted August 1977. FISHERY BULLETIN: VOL. 76. NO. 2. 1978. 443 FISHERY BULLETIN: VOL. 76, NO. 2 Fisheries Center to calculate indices of apparent abundance for several coastal species, including the northern anchovy (Squire 1972). To aid in the detection and quantification of pilchard, Sar- dinops ocellata, shoal occurrence off South Africa, Cram (1974) used an airborne, low-light-level, electron image intensifier to view the ocean's sur- face, detecting the bioluminescence offish schools. During these night aerial surveys the school's horizontal surface area was interposed on the in- strument's circular field of view, and running es- timates of the percentage of coverage were made. These percentage estimates were then used in the computation of biomass estimates. The intensifier used by the Sea Fisheries of South Africa has been used by the author off the southern California coast on an experimental basis. Due to the highly variable school shapes encountered, making estimates of the percentage of school coverage in the circular field of view are difficult. Experience indicated that examination of aerial color photographs and night low-level video tapes of anchovy school shapes for determi- nation of the percentage of school coverage within a circle would be useful, particularly if in the fu- ture, surveys were to be conducted at night using this method for the development of biomass esti- mates for the northern anchovy and other near- surface schooling pelagic species. In addition, an analysis of anchovy horizontal school shapes may assist hydroacoustic research- ers in determining error parameters for computa- tion of sonar biomass estimates. Hydroacoustic surveys currently conducted for the northern an- chovy use both side- and vertical-looking sonar to detect and measure fish schools and school groups during the day along a predetermined survey track line. The acoustic "beam" used in these sur- veys varies according to the unit and is of ±5° to 10°. When detecting the school, the side-looking sonar measures the maximum dimension in one aspect of the school, either normal to or parallel to the ship's track. For the purpose of calculating horizontal area, in contrast to the aircraft's verti- cal view of the actual horizontal school area, the echogram school width is assumed to be elliptical (Smith 1970). Preliminary attempts at biomass estimation from sonar surveys have used the sim- ple assumption that a series of estimates of the width of an elliptical school from random aspects will result in an unbiased estimate of school hori- zontal area. In a side-looking sonar the school width is measured and provides two points of ref- erence with the orientation of these points about the school's profile being unknown. If an ellipse is fitted randomly between these two points, the re- sulting average area will equal a circle, a condi- tion that was not observed in aerial photographs of anchovy schools. METHODS To examine the shape of northern anchovy schools as observed during day and night and to determine what percentage the school occupies of a circle tangent to two points along the school's edge and containing the school inside the circle, a circle was drawn about school profiles ob- tained from a series of 20 day oblique aerial color photographs (from the photographic files of the Southwest Fisheries Center) and of 20 night photographs of fish school bioluminescence. The bioluminescent anchovy school shapes were photogi'aphed from a television monitor as it pro- jected video tapes recorded from an airborne low- light-level television camera used during anchovy resource surveys off northwestern Mexico. The night photographs were made available through the courtesy of Zapata, Inc.^ (Zapata Fisheries), Houston, Tex. The night surveys using low-light- level television were conducted at elevations of up to 1,828 m (6,000 ft) and this survey technique is effective because the northern anchovy commonly migrates to the near-surface area during hours of darkness (Squire 1972). The actual area of the schools observed in the photographs are unknown due to lack of data on the aircraft's altitude, camera angle, and camera geometry; however, all were taken from angles approaching vertical. However, all area calcula- tions are expressed in percentages of a circle drawn tangent about the school's edge. The day school profiles were further analyzed to determine what the school area would be if the width measurement were considered to be equal to the school's diameter and what the area would be if viewed systematically from six points 30° arc) about an arc of 180° around the school (based on school width or diameter as determined similar to the measurements made from a hydroacoustic sounder). These area data calculated from the six points of observation to determine school width were then compared to the actual school area. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 444 SQUIRE: ANCHOVY SCHOOL SHAPES School length to width ratios were determined for both day and night schools. The school area, as expressed in terms of percentage of the circle drawn about it, was de- termined by tracing the school profile upon paper, cutting the school out of the circle and weighing both the school profile and the nonschool portion of the circle on a sensitive laboratory balance. School areas for the six points of school width (diameter) measurement about the 180° arc were computed statistically. RESULTS Night Observations Figure 1 illustrates the profiles of schools re- corded by the airborne low-light-level television system. On each figure are given the area percent- age of a tangent circle that the fish school occupies and the length to width ratio. The average percent coverage of a night an- chovy school, in relation to the circle tangent to the school, is 29. 27^ The average ratio of school length to width is 2.53:1. Day Observations Figure 2 illustrates profiles of schools observed during the day. On each figure are given the per- centage of a circle tangent to the school that is occupied by the fish school, the length to width ratio, and the 30° arc points that were used to determine the school's estimated diameter and area, simulated as randomly viewed by a side- looking sonar. The ratio of the actual school area to the calculated area of the school's average, high and low estimate, as viewed every 30° of a 180° arc based on simulated sonar measurements of width, is given in Table 1. The average percent coverage of a day anchovy school to the tangent circle about the school is \2.V7c. The average length to width ratio for all day schools is 2.09:1. The ratio of estimate of the area of all day schools, as calculated from mea- surements from 30° arc points about the school, to the actual area of the school is 1.72:1. The ratios, length to width, and estimate of school area to actual school area were compared and Figure 3 graphs the relationship. The graph displays the variables plotted on log-log paper showing two main points: One, that the variance is changing proportionally to the mean. This is expected as there should be more variation as the school Table l. — Ratio of the actual anchovy school area to the aver- age area based on six points of observation as viewed every 30° of a 180° area, and the high and low ratio. School Average High Low ID 2.61 1:3 95 1:0 83 2D 96 1:2 91 090 3D 22 1:1 56 0 85 4D 2 12 1:3,71 063 5D 13 1:1.41 080 6D 2 73 1:4.53 1.18 7D 50 1 2.35 060 8D 32 1 2 00 054 9D 49 1:1 88 1.19 10D 31 1:1 60 1 04 11D 2 65 1:4 82 062 12D 33 1:2.02 0.85 13D 47 1:2 35 0.77 14D 68 1:2 54 049 15D 97 1.3 30 065 16D 60 1:2 69 0.44 17D 38 1:2 10 0.84 180 38 1:2 00 068 19D 99 1:3 52 0.42 20D 77 1:2.90 0.66 length to width increases. Two, the plotted regres- sion line indicates that more bias (higher esti- mated actual school ratio) is introduced as the school length to width ratio increases. The line is significant at the 95'7f confidence interval as proven by the ^test ( 1.98 <2.298). The confidence limits are from 0.0545 to 0.734. SUMMARY AND COMMENTS The data on day/night school length to width ratios support what is commonly known about the schooling shapes of the northern anchovy. They are more common in the near-surface area at night, generally in large elongate thin surface schools. These elongate schools tend to group to- gether in the early morning hours and descend to depth to form more compact schools during the day (Mais 1974). Studies by Squire ( 1972) of aerial fish spotter data show anchovy schools to be more fre- quently observed, and observed in larger quan- tities at night, when compared with day observa- tions. The schools percent area coverage of the tangent circle at night is 12.9*^ less than its coverage dur- ing the day and the length to width ratio is greater by 0.44. In addition, analysis of school length to width ratios compared to the ratio of estimated school size to actual size (Figure 3) shows that as the length to width ratio increases a greater error in school area estimate will occur. Schools with a length to width ratio of 2:1 have an estimated to actual error of about 1.5:1 while more elongate schools of a ratio of 3:1 have an estimated error of about 1.75:1. 445 FISHERY BULLETIN: VOL. 76, NO. 2 NIGHT I NIGHT 2 NIGHT 3 ^'^^T "^ 1195 14 98% 1420 1840% 1138 2967% '338 12 12% NIGHTS ^'GHT6 NIGHT 7^ ^ NIGHTS^ ^ e (1) e €) 1456 1284% 1263 2159% 12 28 3425% 1135 4326% NIGHT 9 NIGHT 10 NIGHT II NIGHT 12 (/) (J) e e 1381 2125% 1321 2868% 1297 2180% ||73 4050% NIGHT 13 NIGHT 14 NIGHT 15 NIGHT ^6^ 1300 2160% 1191 42 35% 12 51 2787% ' ^ ^^ 31 96 /« NIGHT 17 NIGHT 18 NIGHT 19 NIGHT 20 1153 25 38% 12 00 2739% 1207 4075% 1126 6654% Figure l. — Profiles of anchovy schools observed at night off southern California, indicating the width to length ratio and the percentage of a tangent circle about the school. 446 SQUIRE: ANCHOVY SCHOOL SHAPES DAY I DAY 2 I 2.05 DAY 5 3021% DAY 6 DAY 3 I 139 DAY 7 6671% I i 39 6978% DAY 4 12.20 3152% 12.29 'K)84% I 1.92 4918% DAY 10 DAY II I 127 57.59% I 1.28 6240% DAYJ2 I 3.54 2074% 1:1.59 50 36% DAY 13 DAY 14 I 1.78 41.92% DAY 15 DAY 16 1252 12.70 28.80% h2.76 3721% DAY 17 DAY 18 I 1,64 4760% DAY 19 DAY 20 1285 2801% 1236 3392% FIGURE 2. — Profiles of anchovy schools observed during the day off southern California, indicating the width to length ratio and the percentage of a tangent circle about the school. Measurements of school width were taken at the six points (long arrow shaft) indicated about a 180° arc. 447 FISHERY BULLETIN: VOL. 76, NO. 2 1.0 1.5 2.0 RATIO- SCHOOL 2.5 3.0 4.0 5.0 LENGTH TO WIDTH Figure 3. — Regression plot for the ratio estimated anchovy school size to actual size compared with the school length to width ratio. For simulated sonar observations of school widths used in the calculation of school area, the preliminary examination of these data indicates a possible 1.72:1 average overestimate ofarea due to school shape deviations from a circle or ellipse. Fish schools, being highly variable in horizontal profile, are probably equally complex in vertical structure; the relationship of horizontal complex- ity to vertical complexity is not known. Also un- known is the question of whether the individual school's axis is oriented in the same general direc- tion within a group of schools, a possible factor which, if it occurs, could provide a source of sub- stantially higher or lower school area error esti- mates from sonar track line surveys. The problem of accurately estimating the per- centage of school area within view of a low-light- level viewer is difficult, as the examples of school shapes within the target circle would indicate. Parameters of human viewing error could be es- tablished for this survey technique. However, the conduct of surveys using a low-light-level televi- sion system where the video signal can be recorded and later electronically analyzed with the aid of an image analyzer, should result in a higher degree of survey accuracy. School shapes were taken from photographs randomly selected from an aerial photo file. Many of the photos were taken in the nearshore areas. There is the possibility that schools may be slightly more elliptical in shape over deep water than in the nearshore areas, but this is not documented. If this were true the error estimate would be reduced. This and other aspects of school profile and orientation should be investigated further and estimates of length to width ratios from aerial surveys, done in conjunction with each acoustic survey, may be useful for determination of a correction factor for the acoustic data. ACKNOWLEDGMENTS The suggestions of Reuben Lasker and Paul Smith and the assistance of Jim Zweifel in the calculation of the weighted linear regression are appreciated. LITERATURE CITED Cram, D. L. 1974. Rapid stock assessment of pilchard populations by airrraft-bome remote sensors. Proc. 9th Int. Symp. on Remote Sensing. Ann Arbor, 15-19 April, p. 1043-1050. Hewitt, R. P., P. E. Smith, and J. C. Brown 1976. Development and use of sonar mapping for pelagic stock assessment in the California Current area. Fish Bull, U.S. 74:281-300. Mais, K. F. 1974. Pelagic fish surveys in the California Cur- rent. Calif Dep. Fish Game, Fish Bull. 162, 79 p. RADAKOV, D. V. 1972. Schooling in the ecology offish. [In Russ.] Izdatel. "Nauka," Moscow. (Engl, transl., 1973. 173 p. Isr. Pro- gram Sci. Transl. Publ, John Wiley and Sons, N.Y.) Smith, P, E. 1970. The horizontal dimensions and abundance of fish schools in the upper mixed layer are measured by so- nar. In G. B. Farquhar (editor), Proc. International Symposium on Biological Sound Scattering in the Ocean, p. 563-591. Maury Cent. Ocean Sci., Dep. Navy, Wash., D.C. Squire, J, L., jr. 1972. Apparent abundance of some pelagic marine fishes off the southern and central California coast as surveyed by an airborne monitoring program. Fish. Bull., U.S. 70:1005-1019. VOGLIS, G. M., AND J. C. COOK 1966. Underwater applications of an advanced acoustic scanning equipment. Ultrasonics 4:1-9. 448 SYNERGISTIC EFFECTS OF ENVIRONMENTAL VARIABLES ON THE METABOLISM OF THE COPEPOD EUTERPINA ACUTIFRONS FROM TWO DIFFERENT AREAS OFF THE COAST OF THE STATE OF SAO PAULO, BRAZIL^ G. Scares Moreira^ and W. B. Vernberg^ ABSTRACT The combined efFects of temperature and salinity on the respiratory rate of two populations of the copepod Euterpina acutifrons have been determined. One population was taken from a nonpolluted area, Sao Sebastiao Channel, and the other from a polluted area, Santos Bay, both off the coast of the State of Sao Paulo, Brazil. Four groups of copepods were used in the experiments: 1) Sao Sebastiao animals kept in Sao Sebastiao water (35%« salinity); 2 1 Santos animals kept in Santos water (28%o salinity); 3) Sao Sebastiao animals kept in Santos water; and 4) Sao Sebastiao animals kept in diluted Sao Sebastiao water (28%o salinity). Results showed that Sao Sebastiao copepods in either full strength seawater (35%«) or lower salinity seawater (28%o) could metabolically regulate over a wider range of salinities than could Santos copepods in Santos water or Sao Sebastiao copepods maintained in Santos water. It was concluded that the water quality of the Santos Bay was responsible for changes in the metabolic regulatory capacity of the copepods exposed to Santos water. The planktonic harpacticoid Euterpina acutifrons (Dana) is distributed in the warm waters of the world between lat. 66°N and 40°S (Haq 1972). It is a euryhaline species and has been reported in salinities ranging from 8%o (Cananeia Estuary, southern Brazil, Tundisi 1972) to 39%o (Mediter- ranean Sea, El-Maghraby 1965). Laboratory studies have shown that reproduction can occur over a salinity range of 15 to 45%o (Moreira and Yamashita 1975). Euterpina acutifrons is an im- portant link in the marine trophic web serving as food source for both adult and larval fishes (Pouchet and de Guerne 1887; Lebour 1918; Blin 1923; Carvalho 1945; Marques 1951; Thayer et al. 1974). In an earlier paper, Moreira (1975) reported that salinity tolerances for Brazilian populations of £. acutifrons from Santos were very different from those of populations of this species from Sao Sebastiao. This in itself is not surprising since the salinity regimes of the two areas are different. The salinity in the Santos Estuary varies widely from 17 to 30%o depending on the tide and season of the 'Contribution no. 214 of the Belle W. Baruch Institute for Marine Biology and Coastal Research. ^Physiology Department, Institute of Biosciences, and Insti- tute of Marine Biology, University of Sao Paulo, Brazil. ^School of Pubhc Health, Belle W. Baruch Institute for Marine Biology and Coastal Research and Department of Biology, Uni- versity of South Carolina, Columbia, SC 29208. Manuscript accepted September 1977. FISHERY BULLETIN: VOL. 76, NO. 2, 1978. year, while in Sao Sebastiao Channel the salinity is approximately 35%o throughout the year. Water temperatures in both areas are essentially the same, ranging from 19° to 30°C depending upon season. It was not determined, however, if the observed differences in salinity tolerances of the two populations were genetically or environmen- tally induced. Subsequently, a study was initiated to resolve this question by measuring metabolic response patterns of specimens from both popula- tions to different thermal-salinity regimes. It soon became apparent that environmental parameters other than temperature and salinity were factors in determining the metabolic response patterns of these copepods. A detailed chemical analysis of the water in Santos Bay is not available, but great numbers of tankers and other vessels continuously operate near shore, discharging ballast water and con- taminating seawater and adjacent regions with petroleum. In addition, there are a large number of industries that discharge wastes directly into the water. One sample analysis of Santos Bay seawater was found to contain 270 ppb lead and 200 ppb nickel (unpublished data). Furthermore, to minimize the effects of human waste or degra- dation products, approximately 400 tons of chlorine are added monthly near shore. The data presented in this paper demonstrate that 449 FISHERY BULLETIN: VOL 76, NO. 2 metabolic response patterns of £. acutifrons to the normal fluctuations found in estuarine systems were significantly altered by the water quality of Santos Bay water. MATERIALS AND METHODS The copepods were collected in two fixed loca- tions off the coast of the State of Sao Paulo, one in Santos Bay (lat. 23°59'S; long. 46"19'W), the other in Sao Sebastiao Channel (lat. 23°50'S; long. 45°25'S). Collections were made with a nylon plankton net (20 m/x) during the winter when water temperatures averaged approxi- mately 21°C. All samples were brought im- mediately into the laboratory whereupon E. acutifrons were sorted from the plankton using a mouth pipette under a binocular microscope. The copepods were placed in 2-1 crystallizing dishes, 20 cm in diameter and 15 cm high. In the first two series of experiments, the copepods were placed in the water obtained from the collection points, i.e., copepods from Sao Sebastiao Channel were placed in Sao Sebastiao water (35%o) and copepods from Santos were placed in Santos water (28%o). In the last two series of experiments, copepods from Sao Sebastiao were placed either in Santos water or in Sao Sebastiao water diluted to 28%o. The copepods were maintained under temperature and photo- period regimes approximating field conditions: 19°-24°C and 11 L:13 D. The copepods were fed with Phaedactylum and Platymonas daily and kept in the laboratory at least 1 wk before being used in the respiration experiments. Oxygen uptake was determined using Carte- sian diver respirometers (Holter 1941), which have a total volume of 8-13 />tl. Only nongravid females were used. Two or three copepods were placed in each diver, depending upon the salinity/ temperature regime of the experiment. The oxy- gen uptake was determined during a 2-h interval. The first 30-min reading was discarded; after this initial reading, uptake rates remained constant. Oxygen uptake rates were measured under the following environmental conditions: Sao Sebas- tiao animals maintained in Sao Sebastiao water, Santos animals in Santos water, and Sao Sebas- tiao animals in Santos water, 15°, 20°, 25°, 30°, and 32°C at 15, 25, 35, 45, and 55%o salinities. The oxygen uptake of Sao Sebastiao animals main- tained in Sao Sebastiao water diluted to 28%o was determined at 15°, 25°, and 30°C over the same salinity ranges used in the other experiments. Ten determinations were made under each set of en- vironmental conditions. Distilled water or freeze- concentrated brine was added to filtered seawater to attain the desired salinities. Salinities were determined by titrating against silver nitrate (Harvey 1955). Dry weights for the copepods were obtained using a TorbaP torsion balance, 0.01 mg sensitiv- ity. The copepods were rinsed with distilled water and dried at 70 °C for 24 h before they were weighed. Three replicates of 200 nongravid fe- males from each area of collection were used. Re- sults were expressed as microliters of oxygen per milligram per hour. Significant difference of means was calculated by the method of Simpson et al. (1960) for small samples. The metabolic data obtained in the first three series of experiments were analyzed statistically using multiple regression techniques. The basic experimental design used in this study is usually referred as a factorial design. Specifically, the plan was a 5 X 5 factorial using five levels of tempera- ture and five levels of salinity, making in all 25 combinations of experimental conditions. Since 10 determinations of oxygen uptake were made in each combination, a total of 250 observations were made in each series. Thus, the 250 observations may reasonably be considered as continuous re- sponses of a function of the two factors and interac- tions. Oxygen uptake data were analyzed as percent- age of oxygen consumption relative to that at 25°C and 35%o salinity for Sao Sebastiao animals in Sao Sebastiao water and at 25°C and 25%o salin- ity for Santos animals in Santos water and Sao Sebastiao animals in Santos water, i.e., the rate under these "standard" conditions was assumed to be lOO'/f , and rates obtained under other regimes were calculated as the percent deviation from that rate. Since the observations are treated as percent- age measurements generated by data from bino- mial populations, the transformation Y = arc sin v'jc, where x is observed percent respiration, is appropriate to stabilize variances (Mendenhall 1968). Analysis of variance for this data indicated which of the factors (temperature, salinity, or temperature-salinity interactions) had significant effects on the metabolism of the copepods. The program was run on an IBM 360 computer. ''Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 450 MOREIRA and VERNBERG: SYNERGISTIC EFFECTS OF ENVIRONMENTAL VARIABLES RESULTS Animals From Sao Sebastiao in Sao Sebastiao Water (35%o) The metabolic rate ofEuterpina acutifrons from Sao Sebastiao which were maintained in water from Sao Sebastiao Channel was less influenced by changes in salinity than were the other groups o{ E. acutifrons (Figure 1). Rate did, however, in- crease with increasing temperature up to 30°C; at 32 °C, rates either leveled offer decreased. Great- est increases in respiration rates were observed between 15° and 20°C. These increases are reflected in the relatively high metabolic rates within this thermal range obtained from this group of animals. At 20°C metabolic rates were not significantly different over the entire salinity range of 15-55%o, and at 15°C the copepods were able to regulate their metabolism over a range of 25-55%o. At higher temperatures (25°, 30°, and 32°C) rates generally were lower at the salinity extremes (15, 45,55%o) and highest at 35%o. Rates varied from a minimum of 4.87 to a maximum of 22.36 /xl/mg h ^^ dry weight (Figure 1). Statistical analysis indicated that 66% of the observed variability in the rates could be explained by the temperature-salinity combina- tions (Table 1), although the linear effect of tem- perature was the single significant factor (1% level). The linear effect of salinity, the quadratic effects of the temperature and salinity, and the temperature-salinity interaction did not contri- bute significantly to the observed changes in res- piration rates. Figure 2A shows the response sur- face contours fitted over the experimental design. Table l. — Analysis of variance of data for Euterpina acutifrons (Sao Sebastiao animals in Sao Sebastiao water, 35%o). T = tempera- ture, S = salinity. Variable r2 Significance level T 054953 1% 72 0,65769 Not significant S 0.65796 Not significant S2 0.65885 Not significant Tx S 065933 Not significant Animals From Santos in Santos Water (28%o) In the Santos animals maintained in Santos wa- ter, the rate of oxygen uptake also increased over the temperature range to 30°C for the entire salin- ity range. In most of the salinities, the largest metabolic increase occurred between 25° and 30°C. This contrasts with Sao Sebastiao copepods which exhibited the largest increase between 15° and 20°C. The Santos animals did not show the metabolic regulation observed in the Sao Sebas- tiao animals maintained in Sao Sebastiao water, and tended to have low metabolic rates at the salinity extremes, i.e., 15, 45, and 55%o. Highest rates occurred at salinities of 25-35%o. The rates varied from a minimum of 7.97 to a maximum of 28.20 Atl/mg h^Mry weight (Figure 1). Statistical analysis indicated that only 46% of the observed variability in the respiration rates of these animals could be explained by the temperature-salinity combinations (Table 2). The significant factors were the quadratic effects of temperature and salinity (0.05% level) and the linear effect of salinity (0.05% level). The temperature-salinity interaction was not a sig- nificant factor. Figure 2B shows the response sur- face contours fitted over the experimental design. Table 2. — Analysis of variance of data for Euterpina acutifrons (Santos animals in San- tos water 28%o). T = temperature, S = salin- ity. Variable Significance level T2 S^ S T X S 024350 0,28955 046193 046198 0.05% 0.05% 0.05% Not significant Animals From Sao Sebastiao in Santos Water Transfer of Sao Sebastiao copepods into water from Santos markedly altered their metabolic re- sponses, especially their response to salinity. Res- piration rates increased with temperature up to 25°C at the extreme salinities (15, 45, 55%o) and up to 30°C at salinities of 25 and 35%o, before leveling off or decreasing. The copepods which were transferred to Santos water did not regulate metabolically at any temperature at the salinity extremes. Lowest rates were obtained at salinities of 15 and 55%o, and the highest rates were ob- served at 25%o at 15° and 25°C. At 30° and 32°C, peak metabolic rates occurred at 35%o (Figure 1). Rates varied from a minimum of 6.80 to a maximum of 37.23 /xl/mg h^ dry weight (Figure 1). 451 FISHERY BULLETIN; VOL. 76. NO. 2 15 1 10 > 9 ■c 8 r 7 y- y / ,1 ,) ..- ■.. 1 1 ■c 6 4 , 15 25 35 45 55 Salinity %o 20*'C - 30 , ^ ^ 5 25 -5 20 . ,'-'' ^% en E ' M. / V j: 15 • / JJ 3. l_/ ■ """^ ^1^^ ""h 10 ■ / ' ^^ 9 15 25 35 45 Salinity %o 55 35 30 ^ 25 ?15 20 ■ 10 9 8 25''C 15 25 35 Salinity %o 45 55 40 35 D § 30 > ■c 25 ai P 10 ^ CN O lb 3. o\j \^ r /'-^ / ^ ' K ^\y-^^>^' \ •I \ / yy^ 1 ■■■.\ K \ /y^ X ■ L> M ■■■^r~~"vi *f 1 -.j:J|i 15 25 35 45 55 Salinity %o 40 35 25 ■ I 30 ■a 20 E ■^ 15 O 10 9 8 32*0 15 25 35 45 Salinity %o 55 Figure l. — Euterpina acutifrons: metabolic rate of animals from Sao Sebastiao maintained in Sao Sebastiao water (solid line), ani- mals from Santos maintained in Santos water (dashed line), ani- mals from Sao Sebastiao maintained in Santos water (dot-dash line), and animals from Sao Sebastiao maintained in Sao Sebastiao diluted water (dotted line), in different temperature and salinity regimes. 452 MOREIRA and VERNBERG: SYNERGISTIC EFFECTS OF ENVIRONMENTAL VARIABLES 15 20 25 Temperature °C 30 15 20 25 Temperature °C 15 20 25 Temperature "C Figure 2. — Response-surface estimation of percent metabolic rate of Euterpina acutifrons maintained at 25 different tempera- ture and salinity combinations. A. Copepods from Sao Sebastiao maintained in Sao Sebastiao water. B. Copepods from Santos maintained in Santos water. C. Copepods from Sao Sebastiao maintained in Santos water. Table 3. — Analysis of variance of data for Euterpina acutifrons (Sao Sebastiao animals in Santos water, 28%o). T = temperature, S = salinity. Variable /•2 Significance level T 0-17164 0.05% S2 0-32500 0.05% 8 046808 0.05% T2 0 59715 0.05% T ■ S 062298 0.05% All of the analyzed factors, i.e., linear tempera- ture and salinity, as well as the quadratic effect of these factors, and the temperature-salinity in- teraction, contributed significantly (0.057f level) to the observed variability in respiration rates (Table 3). A total of 62% of the variability could be explained by these various factors. Figure 2C shows the response surface contours fitted over the experimental design. Animals From Sao Sebastiao in Sao Sebastiao Diluted Water In this series the respiration experiments were run at three temperatures to test whether or not the results obtained for Santos animals were the result of acclimation to a lower salinity. The respi- ration rates at 28%o were essentially the same as those for animals maintained in undiluted Sao 453 FISHERY BULLETIN: VOL. 76, NO. 2 Sebastiao water. At 15°, 25°, and 30°C, metabolic rates were not significantly different over the range of 15-45%o. The rates varied from a minimum of 4.32 to a maximum of 19.17 ix\/mg h ' dry weight (Figure 1). DISCUSSION In Brazilian waters, populations of £■. acutifrons thrive over a wide range of salinities and variable salinity alone does not seem to be a limiting factor in their distributional patterns (Tundisi 1972; Moreira and Yamashita 1975). Indeed, of the vari- ous environmental variables tested, temperature alone significantly affected the metabolic rates of these copepods. The present data demonstrate that specimens of copepods from the unpolluted Sao Sebastiao Channel have the capability of metabolic regulation over a wide range of salinities when tested using Sao Sebastiao water. On the other hand, marked diminution in the capability to regulate metabolically at salinity ex- tremes was noted in E. acutifrons from the Santos population and specimens from Sao Sebastiao maintained in Santos water. For both groups of animals salinity, as well as temperature, proved to exert a statistically significant effect at the 5% level (or less) on their oxygen uptake rates. These marked changes in metabolic control in the copepods taken from or exposed to Santos water compared with that of copepods from Sao Sebas- tiao are depicted in Figure 2. While we did not measure population densities of the E. acutifrons in our two study areas (Santos Bay and Sao Sebastiao Channel), there is some indication in the literature that population size is sensitive to polluted waters. Gabriel et al. (1975) reported a decrease in abundance of this species in the Milford Haven Estuary following its de- velopment into the largest oil port in the United Kingdom in the 1960's, and there are several examples that indicate that pollutants can affect the survival of copepods and planktonic larvae. Barnes and Stanbury ( 1948) have studied the toxic action of copper and mercury salts on the copepod Nitocra spinipes and verified that mercuric chloride is a very effective poison; in contrast, these animals are very resistant to copper. D'Agostino and Finney (1974) have found that copper and cadmium inhibit growth and develop- ment of the copepod Tigriopus Japonicus at 0.064 mg/1 and 0.044 mg/1, respectively. Heinle (1969) suggested that the high mortality rate ofAcartia tonsa in a power plant effluent was due to the chlorination of the cooling water, correlating the apparent periodicity in the mortality rate with the chlorination schedule. Latimer et al. (1975) studied the toxicity of 30-min exposures of re- sidual chlorine to two species of copepods, Lim- nocalanus macrurus and Cyclops bicuspidatus thomasi. The predicted "safe" concentrations were 0.9 mg/1 for L. macrurus and 0.5 mg/1 for C. b. thomasi. Roberts et al. (1975) studied the acute toxicity of chlorine to some estuarine species, in- cluding molluscan larvae, copepods, shrimps, and fishes. They found that molluscan larvae and Acartia tonsa were the most sensitive species tested, with 48-h TLg,, values at chlorine levels <0.005 ppm. Gray (1974) demonstrated that lead (Pb(N03)2) at 0.3 ppm reduced the growth rate of the marine ciliate protozoan Cristigera by 11.7% and at 0.15 ppm by 8.46%. Mercury was found to have an effect on survival, metabolism, and be- havior of the planktonic larvae of Uca pugilator (DeCoursey and Vernberg 1972; Vernberg et al. 1973). Generally, larvae are much more sensitive to toxicants than are adults and very low concen- trations of a toxicant can interact with environ- mental factors to cause increased mortalities among larvae (Vernberg 1975). Detailed chemical analyses of Santos water ob- viously are needed, but the very high concentra- tion of lead and nickel which were found in one sample, plus the oil and other industrial effluents that are being discharged, leave little doubt that the Santos Estuary is highly polluted. Data pre- sented in this paper strongly suggest that speci- mens living in the Santos Estuary do so at a high cost energetically. This high metabolic cost for survival following exposure to salinity extremes would almost certainly be a factor limiting the distribution of £. acutifrons in polluted estuaries, since fluctuating salinity regimes are characteris- tic of this environment. Results obtained in this study highlight the fact that the physiological re- sponses of marine organisms may be markedly modified if test animals are taken from or exposed to polluted waters. LITERATURE CITED Barnes, H., and F. A. Stanbury. 1948. The toxic action of copper and mercury salts both separately and when mixed on the harpacticid copepod, Nitocra spinipes (Boeck). J. Exp. Biol. 25:270-275. 454 MOREIRA and VERNBERG: SYNERGISTIC EFFECTS OF ENVIRONMENTAL VARIABLES BLIN. [F.) 1923. Note sur ralimentation de la Sardine. Euterpes et Sardines. Bull. Soc. Zool. Fr. 48:99-105. CARVALHO, J. DE PAIVA. 1945. Copepodos de Caioba e Taia de Guaratuba. Arq. Mus. Parana. 4:83-116. D'AGOSTINO, A., AND C. FiNNEY. 1974. The effect of copper and cadmium on the develop- ment ofTigriopusjaponicus. In F. J. Vernberg and W. B. Vemberg (editors). Pollution and physiology of marine organisms, p. 445-463. Academic Press, N.Y. DECOURSEY, p. J., AND W. B. VERNBERG. 1972. Effect of mercury on survival, metabolism and be- haviour of larval Uca pugilator (Brachyura). Oikos 23:241-247. El-Maghraby, a. M. 1965. The seasonal variations in length of some marine planktonic copepods from the eastern Mediterranean at Alexandria. Crustaceana 8:37-47. Gabriel, P. L., N. S. Bias, and a. nelson-Smith. 1975. Temporal changes in the plankton of an indus- trialized estuary. Estuarine Coastal Mar. Sci. 3:145- 151. Gray, J. S. 1974. Synergistic effects of three heavy metals on growth rates of a marine ciliate protozoan. In F. J. Vernberg and W. B. Vemberg (editors). Pollution and physiology of marine organisms, p. 465-485. Academic Press, N.Y. Haq, S. M. 1972. Breeding of Euterpina acutifrons, a harpacticid copepod, with special reference to dimorphic males. Mar. Biol. (Berl.) 15:221-235. Harvey, H. W. 1955. The chemistry and fertility of sea waters. Camb. Univ. Press, N.Y., 224 p. Heinle, D. R. 1969. Temperature and zooplankton. Chesapeake Sci. 10:186-209. Holter, H. 1941. Technique of the Cartesian diver. C. R. Trav. Lab. Carlsberg (Chim.) 24:399-478. Latimer, D. L., A. S. Brooks, and A. M. Beeton. 1975. Toxicity of 30-minute exposures of residual chlorine to the copepods Limnocalanus macrurus and Cyclops bicuspidatus thomasi. J. Fish. Res. Board Can. 32:2495-2501. LEBOUR, M. V. 1918. The food of post-larval fish. J. Mar. Biol. Assoc. U.K. 11:433-469. Marques, E. 1951. Copepodes encontrados no conteudo gastrico de algun clupeideos da Guine Portuguesa. Anais Junta In- vest. Colon. 6:11-18. MENDENHALL, W. 1968. Introduction to linear models and the design and analysis of experiments. Wadsworth, Belmont, Calif., 465 p. Moreira, G. S. 1975. Studies on the salinity resistance of the copepod Euterpina acutifrons (Dana). In F. J. Vemberg (editor), Physiological ecology of estuarine organisms, p. 73- 79. Belle W. Baruch Libr. Mar. Sci. 3, Univ. S.C. Press, Columbia. Moreira, G. S., and C. Yamashita. 1975. Influencia de la salinidad en la reproduction y desar- rollo de Euterpina acutifrons (Dana). In Memorias I Simposion Latino-americano sobre Oceanografia Biologica (Mexico), p. 236-245. POUCHET, G., AND J. DEGUERNE. 1877. Sur la nourriture de la Sardine. C. R. Acad. Sci., Paris 104:712-715. ROBERTS, M. H., Jr., R. J. DIAZ, M. E. BENDER, AND R. J. HUGGETT. 1975. Acute toxicity of chlorine to selected estuarine species. J. Fish. Res. Board Can. 32:2525-2528. Simpson, G. G., a. Roe, and R. C. Lewontin. I960. Quantitative zoology. Revised ed. Harcourt, Brace and Co., N.Y., 450 p. THAYER, G. W., D. E. HOSS, M. A. KJELSON, W. F. HETTLER, JR., AND M. W. LaCROIX. 1974. Biomass of zooplankton in the New^port River es- tuary and the influence of postlarval fishes. Chesapeake Sci. 15:9-16. TUNDISI, T. M. 1972. Aspectos ecologicos do zooplankton da jegiao la- gunas de Cannaneia com especial referenda aos Copepoda (Crustacea). Ph.D. Thesis, Sao Paulo Univ., Sao Paulo. Vernberg, W. B. 1975. Multiple factor effects on animals. In F. J. Vem- berg (editor), Physiological adaptation to the environ- ment, p. 521-537. Intext Educational Publishers, N.Y. Vernberg, W. B., p. J. deCoursey, and W. J. Padgett. 1973. Synergistic effects of environmental variables on larvae of Uca pugilator. Mar. Biol. (Berl.) 22:307-312. 455 DESCRIPTION OF LARVAE OF A HIPPOLYTID SHRIMP, LEBBEUS GROENLANDICUS, REARED IN SITU IN KACHEMAK BAY, ALASKA Evan Haynes^ ABSTRACT Larvae of Lebbeus groenlandicus , a hippolytid shrimp, were reared in situ in Kachemak Bay, Alaska, from the first zoea (Stage I) through themegalopa (Stage III). Each of the three stages is described and illustrated, and then compared with descriptions of larvae of Lebbeus spp. given by other authors. Information on the larval stages of the genus Leb- beus is meager. Pike and Williamson (1961), in their summary of the generic characteristics of Spirontocaris and related genera, note that the only larva o^ Lebbeus known for certain is a larva of L. polaris dissected from a well-developed egg. During studies on rearing larvae of pandalid shrimp for descriptive purposes (Haynes 1976, 1978), I succeeded in rearing larvae of L. groen- landicus to the megalopa stage. This report de- scribes and illustrates each of the two zoeal stages and megalopa of L. groenlandicus , and compares the stages obtained from rearing in situ and from plankton in Kachemak Bay with provisionally identified larvae of L. groenlandicus reported by other authors. METHODS A complete discussion of rearing technique, methods of measurement, techniques of illustra- tion, and nomenclature of gills and appendages is given by Haynes (1976). Briefly, the rearing technique consists of obtaining Stage I zoeae from known parentage in the laboratory and then rear- ing the zoeae to postlarvae in 500-ml flasks sus- pended upright beneath the surface of the sea. Cast skins and larvae removed from the flasks were examined in the laboratory to determine sequence and morphology of each stage. Larval stage was also verified using larvae from plankton reared in the same manner as larvae obtained in the laboratory. In the illustrations (Figures 1-3), for clarity, setules on setae are usually omitted but spinulose 'Northwest and Alaska Fisheries Center Auke Bay Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 155, Auke Bay, AK 9982L Manuscript accepted November 1977. FISHERY BULLETIN: VOL. 76, NO. 2, setae are shown. The terms are defined as follows: setose - set with bristles (setae) spinose - bearing many spines spinous - spinelike spinulose - set with little spines. The figures are in part schematic and represent typical setal counts. STAGE I ZOEA Total length of Stage I (Figure lA) 6.9 mm (range 6.4-7.4 mm; 10 specimens). Live specimens characterized by bright orange color extending along ventral surface of body from antennules to fourth abdominal segment, orange gut, small orange chromatophore at anus, and greenish in- ternal thoracic organs; remainder of zoea trans- lucent. Rostrum slightly sinuate, without teeth, about two-thirds length of carapace. Carapace with dorsal rounded prominence at base of ros- trum and near posterior edge; no supraorbital spines. Usually at least two minute spinules occur along ventral margin of carapace immediately posterior to pterygostomian spine. ANTENNULE (FIGURE IB).— First antenna, or antennule, consists of an unsegmented cylindrical basal portion and two distal conical projections; largest conical projection bears four aesthetascs of various lengths; smallest conical projection bears a single heavily plumose seta. ANTENNA (FIGURE IC).— Consists of inner flagellum (endopodite) and outer antennal scale (exopodite). Flagellum two-segmented, about twice length of scale; distal segment styliform and terminating in narrow projection. Two simple 457 1978. FISHERY BULLETIN: VOL. 76, NO. 2 J L J 0. 5 mm 0.25 mm Figure l. — stage I zoea ofLebbeus groenlandicus: A, whole animal; B, antennule; C, antenna; D, mandibles (right and left). 458 HAYNES: LEBBEUS GROEXLANDICUS LARVAE 1 . 0 mm Figure l. — Stage I zoea of Lebbeusgroenlandicus: E, maxillule; F, scaphognathite of maxilla; G, first maxilliped; H, second maxilliped. setae occur at joint. Antennal scale distally di- vided into four segments (proximal joint often in- complete) and fringed with 11 heavily plumose setae along terminal and inner margins. A small seta often occurs proximally near lateral margin. Protopodite bears two simple spines ven- trally, one at base of flagellum and one at base of scale. MANDIBLES (FIGURE ID).— Without palps; 459 FISHERY BULLETIN: VOL 76, NO. 2 0.5 mm 1 . 0 mm Figure L — Stage I zoea of Lebbeus groenlandicus: I, third maxilliped; J, first pereopod; K, third pereopod, L, first pleopod; M, second pleopod; N, telson. well developed. Incisor process of left mandible bears five teeth, one of them located near movable premolar denticle (lacinia mobilis), in contrast to triserrate incisor process of right mandible. Both mandibles bear well-developed denticles along terminal margin. Truncated end of molar process of right mandible formed into curved lip. Only left mandible bears a subterminal process. 460 HAYNES: LEBBEUS GROENLANDICUS LARVAE MAXILLULE (FIGURE IE).— First maxilla, or maxillule, bears coxal and basial endites and an endopodite. Proximal lobe (coxopodite) bears 15 setae, most of them spinulose. Median lobe (basipodite) bears 24 spines terminally, 9 of them spinulose; and 2 spines subterminally, 1 of them plumose and the other simple. A series of fine hairs occurs in vicinity of the simple spine. Endopodite originates from lateral margin of basipodite and bears three terminal and two subterminal spinulose setae. No evidence of outer seta on maxillule. MAXILLA (FIGURE IF).— Bears platelike exopodite (scaphognathite) with 33 long, plumose setae along outer margin, and a longer, thick seta at the proximal end. Endopodite not segmented; setae spinous, setation formula 2, 2, 1, 2, 3. Both basipodite and coxopodite bilobed. Basipodite bears 29 setae, 14 on distal lobe and 15 on proxi- mal lobe. Coxopodite bears 23 setae, 5 on distal lobe and 18 on proximal lobe. FIRST MAXILLIPED (FIGURE IG).— Protopodite segmented; bears 27 setae on distal segment and 8 on proximal segment, most of them spinulose. Endopodite four-segmented; setation formula 4, 3, 3, 7. Basal segment of endopodite bears conspicuous setulose spine. Exopodite seg- mented at base; bears four natatory setae. Epipo- dite distinctly bilobed. SECOND MAXILLIPED (FIGURE IH).— Protopodite not segmented; bears nine setae, five of them spinulose. Endopodite five-segmented; fourth segment expanded somewhat laterally; terminal segment tipped by five setae and bears single seta subterminally; basal segment bears conspicuous setulose spine like that on basal seg- ment of endopodite of first maxilliped; setation formula 6, 4, 2, 3, 4. Exopodite about three times longer than endopodite; bears five natatory setae. Epipodite present but not bilobed. THIRD MAXILLIPED (FIGURE II).— Protopodite not segmented; bears three setae. En- dopodite five-segmented; as long as exopodite; number of setae somewhat variable. Exopodite bears five natatory setae. No epipodite. FIRST PEREOPOD (FIGURE IJ).— Endopodite relatively short, wide, and partially segmented; chela partially formed; dactylopodite bears three simple spines. Exopodite a small lobe. SECOND PEREOPOD.— Similar in shape to first pereopod except narrower, exopodite smaller, and chela more deeply cleft. THIRD (FIGURE IK) TO FIFTH PEREO- PODS. — Each pair essentially identical except that they decrease slightly in size from third to fifth. No exopodites. PLEOPODS.— First pleopod (Figure IL) slightly cleft, without joints or setae. Second pleopod (Fig- ure IM) bilobed; outer lamella segmented; inner lamella usually only partially segmented but bears bud of appendix interna. Third to fifth pleopods essentially identical to second pleopod except both lamellae distinctly segmented. ABDOMEN AND TELSON (FIGURES lA, IN). — Abdomen consists of five segments and tel- son (somite six is fused with telson in Stage I). Fourth and fifth abdominal segments each with pair of posterolateral spines nearly as long as segments themselves. Telson slightly emar- ginated distally; bears 19-21 densely plumose setae; small spinules occur between bases of all setae except two outermost pairs. Enclosed uropods visible. Anal spine present. STAGE II ZOEA Total length of Stage II 8.3 mm (range 8.1-8.7 mm; 8 specimens). Color similar to Stage I zoea but more diffuse. Rostrum (Figure 2 A) arched upward; slightly blunter than in Stage I; without teeth. Carapace bears supraorbital, antennal, and pterygostomian spines in addition to several spinules along anteroventral margin. ANTENNULE (FIGURE 2B).— Shows considera- ble change from Stage I. Largest conical projection segmented at tip; terminal segment bears three setae of different lengths; proximal segment bears six groups of five aesthetascs each in addition to row of four aesthetascs laterally and single seta distally. Smallest conical projection bears three nonplumose setae, one long and two short. Pedun- cle of antennule rounded laterally, not segmented, and bears five plumose setae that originate ven- trally. ANTENNA (FIGURE 2C).— Flagellum of an- tenna still two-segmented, but slightly stouter and projection at tip smaller than in Stage I; a few 461 FISHERY BULLETIN: VOL 76, NO 2 0. 5 mm 0. 5 mm Figure 2.— Stage II zoea of Lebbeus groenlandicus: A, carapace; B, antennule; C, antenna; D, first pereopod; E, second pereopod; F, second pleopod. 462 HAYNES: LEBBEUS GROENLANDICUS LARVAE small setae occur along lateral margin. Antennal scale distally divided into two segments and fringed with 29 or 30 thin, plumose setae along terminal and inner margins; distal outer projec- tion a stout spine. Protopodite bears two stout spines, one at base of flagellum and other at base of scale. MANDIBLES, MAXILLULE, AND MAX- ILLA.— Essentially identical to Stage I except scaphognathite of maxilla usually bears 35 setae along outer margin, in addition to the longer and thicker seta at proximal end, and proximal cleft slightly deeper. MAXILLIPEDS.— Essentially identical to Stage I except exopodites of first, second, and third maxil- lipeds bear 5, 16, and 16 natatory setae, respec- tively. FIRST PEREOPOD (FIGURE 2D).— Segmented; without exopodite; chela functional. SECOND PEREOPOD (FIGURE 2E).— Adult in shape; chela functional; ischiopodite articulates somewhat laterally with basipodite. No exopodite. THIRD TO FIFTH PEREOPODS.— Similar to Stage I except ischiopodite articulates somewhat laterally with meropodite and basipodite. PLEOPODS.— First pleopod slightly more de- veloped than in Stage I but still only about one- third length of second pleopod and without appen- dix interna. Second pleopod ( Figure 2F) larger and narrower than in Stage I; outer lamella about one-fourth longer than inner lamella; both lamel- lae and appendix interna fully segmented at their bases. Third to fifth pleopods essentally identical to second pleopods. ABDOMEN AND TELSON.— Posterolateral spines on fourth and fifth abdominal somites still present, those on fourth somite being only slightly shorter in relation to length of somite than in Stage I. Telson essentially identical to Stage I except segmented from sixth abdominal segment and bears 20 or 21 densely plumose setae. Uropods still enclosed. STAGE III (MEGALOPA) Total length of Stage III 7.5 mm (range 7.4-7.6 mm; two specimens). Antennal spine of carapace larger and pterygostomian spine smaller than in Stage II; no evidence of minute spinules along anteroventral margin. Rostrum (Figure 3 A) short; bears single tooth at base in addition to dorsal protuberance. Antennules similar in shape to adult; outer flagellum six-segmented; inner flagel- lum five-segmented; peduncle three-segmented, Figure 3. — Stage III (megalopa) of Lebbeus groenlan- dicus: A, carapace; B, telson. 0.5 mm 463 FISHERY BULLETIN: VOL 76, NO. 2 lateral spine of proximal segment well developed. Antennal flagellum with at least 30 segments; about four times length of scale. Mandibles with unsegmented palps bearing four or five short teeth. Endopodite of maxillule reduced. Maxil- lipeds shaped as in adult, exopodites reduced. Dac- tylopodites of first and second pereopods well developed; carpopodite of second pereopod six- (sometimes seven-) segmented. Lateral margins of pleopods fringed with setae; appendix internae with minute cincinnuli. Posterolateral spines on abdominal segments four and five remnant or lacking. Telson (Figure 3B) rectangular in shape; bears two pairs of spines terminally and one pair laterally (one or two additional spines may occur centrally on terminal margin). Uropods exposed; fully developed except transverse hinge not com- plete. COMPARISON OF LARVAL STAGES WITH DESCRIPTIONS BY OTHER AUTHORS Under the name "Spirontocaris-larva No. lA," Stephensen (1935) included four specimens that were morphologically identical to zoeae provi- sionally identified by him as Stage I S. polaris (= Lebbeus polaris (Sabine)) except that they dif- fered by lacking spines on abdominal segments four and five, exopodites on any pereopods, or free uropods. He regarded these four zoeae as belong- ing to either Spirontocaris groenlandica ( = L. groenlandicus), S. gaimardii (= Euahis gaimar- dii (H. Milne Edwards)), or S. spinas (Sowerby). Pike and Williamson (1961) have shown that the absence of spines on abdominal segments four and five eliminates the zoeae from being either E. gaimardii or S. spinus. They agree with Stephen- sen that his specimens oC'Spirontocaris-larwa No. lA" are closely allied to zoeae he tentatively de- scribed earlier (Stephensen 1917, 1935) as S. polaris (= L. polaris). They suggest, therefore, that Stephensen's "Spirontocaris -larva No. lA" probably belongs to the genus Lebbeus and spe- cifically to L. groenlandicus. Comparison of my zoeae of L. groenlandicus with the descriptions given by Stephensen for "Spirontocaris-larva No. lA" shows that "Spirontocaris -larva No. lA" are not zoeae of L. groenlandicus. My Stage I zoeae bear remnant exopodites on the first and second pereopods and lateral spines on abdominal segments four and five, but Stephensen's Stage I zoeae bear neither the exopodites nor the spines. My Stage I zoeae do not bear supraorbital spines, the peduncle of the antennule is without joints or a ventral spine, and there is no indication of the carpopodite of the second pereopod being jointed; Stephensen's Stage I zoeae bear supraorbital spines, the peduncle of the antennule is three-jointed and bears a distinct ventral spine, and the carpopodite of the second pereopod is partially jointed. In addition, the chelae of the first and second pereopods are not as well formed in my Stage I zoeae as they are in Stephensen's Stage I zoeae. Several of the morphological characteristics de- scribed by Stephensen as pertaining to "Spiron- tocaris-larva No. lA" are typical of later stage zoeae, a fact already noted by Pike and Williamson ( 1961) in their discussion of the morphology of the zoeae of L. polaris and which prompted them to suggest that Stephensen's zoeae were actually in the second, or penultimate, zoeal stage. Even if Stephensen was mistaken in identifying his zoeae as Stage I rather than Stage 11, the morphological differences between my zoeae and his are too great to consider them identical species. My Stage II zoeae bear spines on abdominal somites four and five and the telson is segmented from the sixth abdominal somite, whereas Stephensen's zoeae do not bear spines on abdominal somites four and five and the telson is not segmented from the sixth abdominal somite. Also, in my Stage II zoeae the peduncle of the antennule does not bear a ventral spine and is unsegmented but in Stephensen's zoeae the peduncle bears a ventral spine and is segmented. I have no further evidence on the identity of Stephensen's "SpjVontocans-larva No. lA." Of the three members of the genus recorded from Green- land waters, L. polaris, L. groenlandicus, and L. microceros (cf. Holthuis 1947; Squires 1966), L. microceros was not recorded by Stephensen. Ap- parently it is rare and its larvae have not been described. Also, the advanced development of Stephensen's "Spirontocaris-larva No. lA" makes it unlikely that it belongs to another genus of the spirontocarid group (cf. Pike and Williamson 1961). Apparently Stephensen's "Spirontocaris- larva No. lA" is either the zoea of L. microceros or that of another species of Lebbeus not yet recorded from Greenland waters. On the basis of descriptions of "Spirontocaris- larva No. lA" by Stephensen (1935) and a late stage embryo of Hippolyte polaris ( = L. polaris ) by Kr^yer (1842), Pike and Williamson (1961) 464 H'WNES LEBBEUS GROEM.ANDICUS LAKVAE characterized larvae of the genus Lehbeus as hav- ing two (or three) zoeal stages, five-segmented pereopods, and a small rostrum in Stage I, and pereopods without exopodites in the last zoeal stage. My description of larvae oiL.groenlandicus confirms the generic characteristics for Lebbeus larvae as given by Pike and Williamson. As noted by Pike and Williamson, however, larvae are de- scribed for only a few species of hippolytids, in- cluding the genus Lebbeus, and further confirma- tion of the generic characteristics of the larvae is desirable. LITERATURE CITED Haynes, E. 1976. Description of zoeae of coonstripe shrimp, Pandalus hypsinotus, reared in the laboratory. Fish. Bull., U.S. 74:323-342. 1978. Description of larvae of the humpy shrimp, Pan- dalus goniurus, reared in situ in Kachemak Bay, Alas- ka. Fish. Bull., U.S. 76:235-248. HOLTHIUS, L. B. 1947. The Decapoda of the Siboga Expedition. Part IX. The Hippolytidae and Rhynchocinetidae collected by the Siboga and Snellius Expeditions with remarks on other species. Siboga Exped. 140, Monogr. 39a», 100 p. KR0YER. 1842. Monografisk Fremstilling af Slaegten Hippolyte's nordiske Arter. Med Bidrag til Decapodernes Udvik- lingshistorie. K. Dan. Vidensk. Selsk. Naturv. Math. Afh. Kbh. 9:209-360. (This work has not been seen by the author.) Pike, R. b., and D. I. Williamson. 1961. The larvae of Spirontocaris and related genera ( De- capoda, Hippolytidae). Crustaceana 2:187-208. Squires, H. J. 1966. Distribution of decapod Crustacea in the northwest Atlantic. Ser. Atlas Mar. Environ., Am. Geogr. Soc. Folio 12. STEPHENSEN, K. 1917. Zoogeographical investigation of certain fjords in Southern Greenland, with special reference to the Crus- tacea, Pycnogonida and Echinodermata, including a list of Alcyonaria and Pisces. Medd. Gr0nl. 53:229-378. 1935. Crustacea Decapoda. The Godthaab Expedition, 1928. Medd. Gr0nl. 80:1-94. 465 PREDICTING ABUNDANCE OF STRIPED BASS, MORONE SAXATILIS, IN NEW YORK WATERS FROM MODAL LENGTHS * Herbert M. Austin^ and Clarence R. Hickey, Jr.^ ABSTRACT The abundance of cohorts for any given year class of striped bass, Morone saxatilis, prior to their leaving Chesapeake Bay is inversely related to the modal length offish in that year class 2 yr later in New York waters. The modal length of bass in their third year migrating into the New York area is a reliable index of the abundance of that year class. When back extrapolated modal lengths at the end of the second year of life are considered for the dominant year classes in the New York fishery (ages ni-VI), a high degree of inverse correlation is found between age II and modal length and reported landings suggesting that this is an effective method of predicting the abundance of the stock for the fishery. In discussing natural fluctuations in fish popula- tions, Royce (1972) posed the question, "... can we forecast their occurrence to take maximum advantage of periods of high abundance and pro- tect populations during periods of scarcity?" This question is pertinent to the striped bass, Morone saxatilis, stocks of the Atlantic coast of the United States. The Atlantic coast commercial catch of this species, while following a pattern of fluctuations, has been in an upward trend in recent years, ap- parently as a result of an increasing abundance of fish (Koo 1970; McHugh 1972). This increasing abundance has been reflected by an increased commercial harvest in the State of New York. Concurrently, although not as well documented, is an increase in the number of recreational fisher- men utilizing the resource. Both phenomena necessitate the gathering of management infor- mation while the resource is still in good condi- tion. Most ( >807c ) of the New York commercial har- vest of striped bass occurs in the waters of eastern Suffolk County (Figure 1) where the major fisheries are primarily with haul seine and pound net. Fish taken in this region are predominantly of Chesapeake Bay origin (Neville et al. 1939; Alper- in 1966; Schaefer 1968, 1972; Koo 1970; Austin CONNECTICUT ATLANTIC OCEAN Figure l. — Location of Long Island, N.Y., and the southeastern tip near Montauk Point where striped bass were collected during 1972 and 1974. and Custer 1977; Austin and Hickey;'* Texas In- struments, Inc.^). This study was designed as one phase of a pro- gram to tag and monitor "short" or prerecruit striped bass (less than the legal, 406-mm New York State limit). As stated by Talbot ( 1966), little is known of these fish outside of their nursery areas. Monitoring of these fish, then, permits study of the next year's catch, a segment of the striped bass population often overlooked in fishery investigations. Prerecruit striped bass in New York waters of eastern Long Island are predominantly 2- and ■New York Ocean Science Laboratory, Contribution No. 82. Funded by a grant to the New York Ocean Science Laboratory from the State of New York, Project No. BR74-17F. ^Division of Fisheries Sciences and Services, Virginia Insti- tute of Marine Science, Gloucester Point, VA 23062. ^Environmental Specialists Branch, U.S. Nuclear Regulatory Commission, Washington, DC 20555. Manuscript accepted September 1977. FISHERY BULLETIN: VOL. 76, NO. 2, 1978. ■» Austin, H. M., and C. R. Hickey, Jr. 1974, Migration and mortality of striped bass tagged in eastern Long Island, p. 11- 16. Proc, Am. Littoral Soc./N.Y. Ocean Sci. Lab. Fish Tag Seminar, Dec. 1974, Montauk, N.Y. ^Texas Instruments, Inc. 1976. Report on relative con- tribution of Hudson River striped bass to the Atlantic coastal fishery. Unpubl. rep., 110 p. Texas Instruments, Inc., Dallas, Tex. 467 FISHERY BULLETIN: VOL. 76, NO. 2 3-yr-old fish which are making their first annual migration from their Chesapeake Bay nursery areas to the northern summer feeding ground (Austin and Hickey see footnote 4). The concen- trated study of one age-group of fish permits monitoring of the cohorts for successive years starting with first departure from their home grounds and, thereby, permits a description of dif- ferences or variations in migration and abundance on an annual basis, as well as an accurate evalua- tion of year class mortality in successive years. METHODS AND MATERIALS Prerecruit striped bass were randomly removed from the catches of commercial haul seine and pound net fishermen in the waters of East Hampton on the southeastern end of Long Island, N.Y. (Figure 1). Samples were collected during May and June 1972 and April-June 1974, thus the age II fish were of the 1970 and 1972 year classes, respectively. Fork lengths were measured in the field to the nearest millimeter and scale samples were removed for age determination. The fish were then tagged (Floy*^ FD-69B anchor tags) and released. The initial purpose of the study was tagging of prerecruit fish to monitor the seasonal migration and mortality of cohorts as they reached legal size in the different states. The feasibility study was focused on the 1970 and 1972 year classes. Large differences in the modal size of the fish in their third year (11 + ) existed between the two year classes ( Figure 2 ) . The smaller sized 1970 year class of fish were from the most abundant Chesapeake Bay year class on record (Schaefer 1972). Examination of the literature shows that the length of cohorts may be inversely propor- tional to the abundance or density of the fish (Ste- vens 1977; Texas Instruments'' ), suggesting to us that the length of the striped bass, when they first appear in New York waters, could be an indicator of year class strength and subsequently a means of predicting stock abundance in local waters. Con- sequently the focus of the study was redirected towards examination of these differences. Schafer (1968, 1972) stated that most commer- cially harvested striped bass in New York are of four age-classes, III- VI. Based on this, Schaefer ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ■'Texas In.struments, Inc. 1975. First annual report for the multiplant impact study of the Hudson River estuary. Unpubl. rep., vol. 1, p. VIII-8-VIII-12. Texas Instruments, Inc., Dallas, Tex. r-l YC 1970 (in 1972) I YC 1972 (in 1974) I I i_J 350 FORK LENGTH IN MILLIMETERS Figure 2. — Length-frequency distribution of age II striped bass captured by commercial fishing gear near eastern Long Island, N.Y., during 1972 and 1974. (1972) related the New York harvest to a 4-yr mean brood production (year class strength; ex- pressed as annual mean number of juveniles per standard seine haul in Chesapeake Bay, Md.) 3 to 6 yr prior to the harvest. He concluded that ap- proximately 70% of the variability in annual New York landings could be explained by annual fluc- tuations in year class strength in Maryland wa- ters of Chesapeake Bay. We hypothesized that the growth rate of striped bass, and, therefore, the body length at the end of the 2-yr residence time in the Chesapeake Bay nursery grounds, is a density dependent function with the length inversely proportional to the year class abundance (number offish). This hypothesis was tested via a correlation analysis using modal lengths at age 11+ (our data combined with pub- lished data of Alperin 1966 and Schaefer 1968) and year class abundance indices supplied by the Maryland Department of Natural Resources. The analyses were performed using a Hewlett- Packard Model 9100B programmable calculator with an X-Y plotter, which provided both a regres- sion line and a correlation analysis and coefficient. The relationship resulting from the above analysis suggested that the density dependent hypothesis is true. Since Schaefer ( 1972) described a relationship between the annual New York har- vest (reported commercial landings) of striped bass and the Chesapeake Bay year class abun- dance, and since we have described a probable relationship between year class abundance and modal length at age 11+ in New York waters, it seemed reasonable to test the correlation between 468 AUSTIN and HICKEY: PREDICTING ABUNDANCE OF STRIPED BASS the New York harvest and the modal length at age 11+ via Model II correlation analysis. These analyses were performed in an effort to describe a method for predicting the commercial harvest (and therefore the apparent abundance) of striped bass in New York waters. As each of the several steps in the analyses were dependent on the results of those previously calculated, they are discussed in more detail along with the results below. The reliability of the suggested technique for predicting the abundance of striped bass in New York waters is dependent on several assumptions: 1 ) The Chesapeake Bay stock of fish is the major contributor of striped bass to the New York commercial fishery, as suggested by the several authors noted above; 2 ) The annual relative contribution of the several Atlantic coastal breeding stocks to the coastal stock of fish and, therefore, to the New York fishery remains constant or that it fluctuates or cycles in a consistent manner; 3) The commercial fishery for striped bass in New York effectively collects representative "sam- ples" of the Chesapeake Bay stock of fish; this assumption appears to be valid based upon the relationships described by Schaefer (1972), Texas Instruments, Inc. (see footnote 5) and those described herein, and based upon our ob- servations and those of Schaefer (1972) that many size classes of fish are present in the commercial catch — small age II prerecruits to large mature fish >16 kg total weight; 4) The forecast of commercial striped bass land- ings is based upon past historical landings in relation to past life history events of the species (year class abundance and length at age 11+) and does not reflect changes in commercial fishing effort or any changes in the contribu- tions to the reported landings by recreational fishermen; we have assumed a constant fishing effort, as did Schaefer (1972), and thus com- pared our results with his; while we recognize the weakness in this assumption, there is no alternative as there is no estimate of effort. RESULTS AND DISCUSSION Year Class Strength and Modal Size The lengths of striped bass at age 11+ near Long Island are probably related to ecological cir- cumstances encountered by the fish during their first 2 yr of residence in the rivers of Chesapeake Bay (density, competition, amount of available food). Similarly, Cushing ( 1968) found a close rela- tionship between the mean length of age III Atlan- tic herring, Clupea harengus, and the density of their food source in the sea, and Clark (1967) de- scribed reduced growth rates for sunfish, Lepomis, due to overcrowding, excessive competition, and reduced food supply. It has also been demonstrated by Anthony ( 1 97 1 ) that the growth of young ( age I and II) Atlantic herring is inversely related to their abundance, and Wagner (1969) has stated that in most fishes the growth rate per individual is inversely related to their density. If an inverse relationship exists between the abundance of a year class of striped bass and the cohort length at age II + , a similar relationship should exist between the commercial harvest (as an index of abundance) and the length at age II + , assuming that fishing effort remains approxi- mately constant. To test this hypothesis, age 11 + modal length data for year classes 1970 and 1972 (Figure 2) were combined with other published modal length data for year classes at age 11+ in New York waters from Alperin (1966) and Schaefer ( 1968 ) (Table 1 ), providing a total of eight annual data points. A correlation analysis was performed between these eight annual modal lengths of age 11+ fish and their respective Chesapeake Bay year class strengths 2 yr earlier (Figure 3). The year class strength data (supplied by the Maryland Department of Natural Re- sources) are expressed as the annual mean Table l. — Comparison of observed and computed modal fork lengths for age II striped bass in New York waters. Observed modal Computed modal Year Year class length at age II length at age 11^ class strength' (mm) (mm) 1954 5.2 ^31 3 318 1958 18 1 "285 278 1959 1,3 3335 330 1960 6.4 3310 314 1961 14.4 3290 289 1962 12.2 3300 296 1970 26.8 =245 251 1972 8.5 5295 307 Mean 297.9 ±53.0 297.9 ±50.4 Standard deviation 26.5 25.2 tr = 6.99 n =8 P<0.001 'Courtesy Joseph Boone. Maryland Department of Natural Resources, An- napolis, Md,, data expressed as annual mean number of ageO+ juveniles per standard seme haul. ^Based on the relationship y = 333 - 3X, where Vis the modal length of age 11+ fish and X is the strength of the year class (Figure 3). ^Extrapolated from Schaefer (1968), ■•Extrapolated from Alperin (1966). ^Data from the present investigation. 469 FISHERY BULLETIN: VOL. 76, NO. 2 400-^ ^1959 i" -I — I — I — I — r- ^1962 *I972 ~~~~-~_J96l .__^58 1970 a 10 15 20 25 3 STRENGTH OF YEAR CLASS (NO /SEINE HAUL 1 Figure 3.— Modal size (mm fork length) of age II striped bass from Long Island waters as a function of Chesapeake Bay year class strength. Year classes are indicated. number of age 0+ juveniles per standard seine haul near the Maryland shores of Chesapeake Bay. These are the same data used by Schaefer (1972). The relationship (Y = 333 - 3X) yielded a correlation coefficient of -0.95 ir^ = 0.90), suggesting that 909c of the annual variation in modal length at age 11+ for striped bass in New York waters can be explained by annual fluctua- tions in year class abundance in the waters of Chesapeake Bay. Modal Size and the New York Commercial Harvest The equation described above (Y = 333 - 3X) was used to calculate (and thus to esti- mate) modal lengths of age II -I- fish for those 8 yr for which actual modal lengths exist. A ^test com- parison between the observed age II modal sizes and those computed using the correlation formula above showed no significant difference at the 0.001 probability level (Table 1). Since no significant difference existed between the observed and calcu- lated modal lengths, the assumption was made that reliable modal lengths could be calculated for years in which no actual measurements exist. The equation described above was, therefore, used to estimate modal lengths of age 11+ striped bass for all years between 1954 and 1972, using the corre- sponding year class abundance data. A correlation analysis was then performed (similar to that done by Schaefer 1972) between the New York land- ings of striped bass ( Y) and a 4-yr mean of the computed modal lengths of age 11+ fish 1 to 4 yr prior to harvest (X). The relationship (Y = 15,205,309 - 46,859X) (Figure 4A) yielded a correlation coefficient of -0.86 (r^ = 0.74) sig- nificant at the 0.001 probability level it, = 6.06, n = 13). This expression permits the hindcasting of New York landings as well as a forecast 1 yr in advance, with 95'7r confidence limits. The hind- casts and 1-yr forecasts (for 1975) are superim- posed on the actual New York landings in Figure 5A. As stated by Schaefer (1968, 1972), the New York harvest is predominantly fish of ages III-VI. Close examination of his catch data for 1962, how- ever, revealed that age VII fish, although <27c of the catch in number, could constitute a significant proportion of the catch by weight. Schaefer's (1968) age-frequency distribution shows that in 1962 the age III fish outnumbered the age VII by about 10:1. Using the mean age- weight relation- ships of Mansueti (1961) as 1.8 lb at age III and 12.5 lb at age VII, the age III fish in Schaefer's ( 1968) 1962 catch thus outweighed the age VII fish by less than 1.5:1. Similarly, the age VI fish ( mean 295 300 305 MEAN MODfiL Size I - 5 YEARS PRIOR TO HARVEST Figure 4.— Relationship of New York commercial landings of striped bass to the mean modal size at age II; A) 1 to 4 yr prior to harvest; B) 1 to 5 yr prior to harvest. 470 AUSTIN and HICKEY PREDICTING ABUNDANCE OF STRIPED BASS 2000- flCTUflL LANDINGS CALCULATED LANDINGS ^ 1500- ^ f\" \ \ \ \ ^ \ ' r h \ 1 V 1000- \ 7 /X / 1 500- / _-' 400- ^'\/\y 300- A ACTUAL LANDINGS CALCULATED LANDINGS 1500- f< * v' / A \( \ I f '^- \ // u 1000- !•• k / /i \ I \ 1 1 \ '/\ ^ t % /y u\ 500- /U/ 400- ^^ /\ / 300- B Figure 5. — Actual New York commercial landings of striped bass from 1954 through 1974 with calculated landings through 1975 sujjerimposed using: A) 4-yr mean modal sizes of age II fish; B) 5-yr mean modal sizes of age II fish. weight 8.1 lb) outnumbered the age VII fish by 2:1, but outweighed them by only 1.3:1. It was apparent that during some years the New York harvest of striped bass may be dominated by five age-groups rather than four, as suggested by Schaefer (1972). Another correlation analysis was, therefore, performed between the New York landings (Y) and a 5-yr mean of the computed modal sizes of age II + fish 1 to 5 yr prior to harvest iX). This 5-yr function was expressed as a linear relationship (7 = 17,315,491 - 53,810X) (Figure 4B) with a correlation coefficient of -0.83 (r^ = 0.69), significant at the 0.001 probability level it, =5.05, n =12). Although this coefficient was reduced slightly from that of the 4-yr function above (r = -0.86), the fit of estimated-to-actual landings (with 95^f confi- dence limits) was better for many years (Figure 5B) and was closer to the actual landings than the calculated predictions of Schaefer (1972) (Table 2). Size, Age, and Migration As stated, age II + modal sizes may be computed. Another method of size determination at age II + is by back calculation of scale radii from larger, older fish. Although no age II modal sizes determined by this method were used in the predictive models, our attempts to do so produced some interesting information. Mansueti (1961) described the body length-scale length relationship of striped bass as an allometric linear function, permitting the back calculation of size at each year of age using the scale radii method. Scales from 142 age III striped bass captured in eastern Long Island waters dur- ing 1973 (year class 1970) were made available to the authors by the New York State Department of Environmental Conservation. The ages were re- checked and the fork lengths at age II determined by back calculation from body length:scale radii ratios. The length-frequency distribution of back- calculated data was bimodal, with equal peaks at 205 mm and 235 mm. The second peak was 4.1% lower than the observed unimodal size of 245 mm. Although the back calculated values were slightly lower than the observed, the fit suggests that back calculations may be used for obtaining age II sizes of striped bass during years when these data are lacking. Unpublished length-frequency data for age II striped bass of the year classes 1968, 1969, and 1971 taken in the Virginia rivers of the Chesapeake Bay System were made available to the authors by John V. Merriner of the Virginia Institute of Marine Science. These data showed bimodal distributions similar to that of the back- calculated age II lengths above (Merriner pers. commun.). Merriner suggested that multimodal frequencies occurred because the fish were from different river systems. Merriner's data were from Virginia rivers while our data (Austin and Hickey Table 2. — Comparison of actual and calculated commercial landings of striped bass in the State of New York 1972-75. Calculated Calculated Calculated Year Actual landings' landings 4-yr function^ landings 5-yr function^ landings by Schaefer (1972) 1972 818.150 926,903 852,860 908.000 1973 1,673,984 1,447,975 1,496,965 1.455,000 1974 1,378,529 1,592,301 1,477.594 1 .607,000 1975 1,137,074 1,639.160 1,500.732 — 'Courtesy Fred Blossom, National Marine Fisheries Service, NOAA, Patch- ogue. NY ^Forecasts using the linear regression formulae discussed in the text. 471 FISHERY BULLETIN; VOL 76, NO. 2 see footnote 4) suggest that the bass we examined for back calculation of length were from both Maryland and Virginia rivers, which could ex- plain the differences in the results of the back calcu- lations and the observed lengths. These data suggest that the size-frequency distribution of age II striped bass on Long Island could be bimodal rather than unimodal. The fact that they were not may be due to striped bass migrating by size rather than by age. Two observations of prerecruit striped bass near Long Island lend support to this theory: 1) lOO'/r of the small 454 sublegal fish tagged in 1972 were age-group II, and 2 1 only 28'7( of the 696 sublegal fish tagged in 1974 were age- group II (year class 1972), the remaining 72*7^ were age-groups III (659<^ ) and IV {T7( ). Those fish probably were the larger 1972 and the smaller 1971 and 1970 fish. This large overlap in length ranges permitted an intermingling of the age- classes during the migration of 1974. Management Implications The size increments between different year classes at the same age, and the size differences of individuals within the same year class have sev- eral implications: 1) Faster growing large individuals of any given year class or a less abundant year class of larger individuals are subject to earlier exploi- tation in Chesapeake Bay and along the entire Atlantic seaboard; 2) Slower growing individuals or small individu- als of a large year class may be recruited sev- eral months later than normal in Chesapeake Bay, but perhaps not until a full year later among the northern Atlantic States; a late re- cruitment in the Chesapeake area might result in more available fish to the fisheries in the other coastal states when the fish migrate out of the bay; 3) Projecting sizes offish on the basis of age or vice versa may be invalid, e.g., age II fish in 1972 compared with age II fish of Merriman ( 1941 ) or Mansueti (1961). The use of a mean 4- or 5-yr modal function for prediction of landings treats all year classes equally. A weighted mean providing greater rep- resentation to more abundant year classes might result in more accurate predictions of landings. Such a method could be used by any State simply 472 by monitoring the spring catches of age 11+ pre- recruit fish taken by commercial fishermen. This would require, in New York for example, annual monitoring of the spring run with measurements of sublegal fish. The use of observed modes rather than computed modes for prediction of landings will probably result in more accurate estimates, as suggested in Table 3. Table 3. — Comparison of actual New York commercial landings of striped bass with those calculated using computed and observed age II modal values, for years in which sufficient empirical data exist.' Item 1964 1965 Actual N Y landings 925.500 702,935 4-yr function Landings calculated using: Computed modes 1,021,090 807,881 Empirical modes 913,314 618,102 5-yr function Landings calculated using. _ Computed modes 799.588 1,098,771 Empirical modes — 849,631 'Computed and observed age II modal values are those on Table 1. The eastern New York commercial harvest of striped bass is primarily dependent upon the year class abundance of the Chesapeake Bay stock. The harvest is influenced not only by the larger and older individuals, but also by the annual recruit- ment of age III fish, especially when dominant year classes are present. Knowledge of Chesapeake Bay year class strength or age 11+ modal sizes in New York wa- ters offers a means of forecasting the New York commercial harvest, and thus the apparent abun- dance of striped bass in New York waters. If, as suggested, the level of the New York har- vest is primarily related to the Chesapeake Bay stock of fish, than the former can be used as a qualitative measure of the latter. Such predictive tools as those discussed should be flexible to allow for the occurrence of more than four age-groups of fish in the catch. This may be especially important when dominant year classes are present for several years. Necessary, then, is the annual monitoring of the prerecruit fish in the commercial catch by age or year class, length, and weight. Age- weight data are especially important as commercial landings are recorded by weight of catch and not by numbers offish. Differences noted between calculated and observed landings may be due to environmental variability, changes in fishing effort, the dominance of a particular year class in the fishery, and the fluctuation in the relative contributions offish from the several At- AUSTIN and HICKEY: PREDICTING ABUNDANCE OF STRIPED BASS lantic coastal breeding grounds. Future research and managment efforts should take these into consideration. ACKNOWLEDGMENTS We acknowledge the help of the following: Joseph Boone, Maryland Department of Natural Resources, Annapolis, Md., for unpublished Mary- land year class data; John V. Merriner, Virginia Institute of Marine Science, Gloucester Point, Va., for his unpublished Virginia data; Bryon H. Young, New York State Department of Environ- mental Conservation, Stony Brook, N.Y., for the use of unpublished data; and Fred Blossom, Na- tional Marine Fisheries Service, Patchogue, N.Y., for data on New York landings. J. L. McHugh, State University of New York, and Jack P. Wise and Richard H. Schaefer, National Marine Fisheries Service, reviewed the manuscript. Their comments are appreciated. LITERATURE CITED ALPERIN, I. M. 1966. Dispersal, migration and origins of striped bass from Great South Bay, Long Island. N.Y. Fish Game J. 13:79-112. AUSTIN, H. M., AND O. CUSTER. 1977. Seasonal migration of striped bass in Long Island Sound. N.Y. Fish Game J. 24:53-68. ANTHONY, V. C. 1971. The density dependence of growth of the Atlantic herring in Maine. Rapp. P.-V. Reun Cons. Int. Explor. Mer 160:197-205. Clark, G. L. 1967. Elements of ecology. John Wiley and Sons, N.Y. , 560 p. CU,SHINC., D. H. 1968. Fisheries biology. Univ. Wis. Press, Madison, 200 p. KOO, T. S. Y. 1970. The striped bass fishery in the Atlantic States. Chesapeake Sci. 11:73-93. MANSUETI, R. J. 1961. Age, growth and movements of the striped bass, Roccus saxatilis, taken in size selective fishing gear in Maryland. Chesapeake Sci. 2:9-36. MCHUGH, J, L. 1972. Marine fisheries of New York State. Fish. Bull., U.S. 70:585-610. Merriman, D. 1941. Studies on the striped bass {Roccus saxatilis) of the Atlantic Coast. U.S. Fish Wildl. Serv., Fish. Bull. 50: 1- 77. Neville, W. C, C. L. Dickinson, and J. R. Westman. 1939. Striped bass (Roccus saxatilis). In A biological survey of the salt waters of Long Island, 1938. Part I, p. 107-113. N.Y. State Conserv. Dep., Suppl. 28th Annu. Rep., 1938, No. 14. ROYCE, W. F. 1972. Introduction to the fishery sciences. Academic Press, N.Y., 351 p. Schaefer, R. H. 1968. Size, age composition and migration of striped bass from the surf waters of Long Island. N.Y. Fish Game J. 15:1-51. 1972. A short-range forecast function for predicting the relative abundance of striped bass in Long Island waters. N.Y. Fish Game J. 19:178-181. Stevens, D. E. 1977. Striped bass (Morone saxatilis) year class strength in relation to river flow in the Sacramento-San Joaquin estuary, California. Trans. Am. Fish. Soc. 106:34-42. Talbot, G. B. 1966. Estuarine environmental requirements and limit- ing factors for striped bass. In A symposium on es- tuarine fisheries, p. 37-49. Am. Fish. Soc. Spec. Publ. 3. Wagner, F. H. 1969. Ecosystem concepts in fish and game manage- ment. In G. M. Van Dyne (editor). The ecosystem con- cept in natural resource management, p. 259-307. Academic Press, N.Y. 473 NOTES OBSERVATIONS ON A WHITE-SIDED DOLPHIN, LAGEMORHYNCHUS ACUTUS, PROBABLY KILLED IN GILL NETS IN THE GULF OF MAINE On 20 July 1976, a white-sided dolphin, Lageno- rhynchus acutus, was observed floating with its beak out of the water on Jeffreys Ledge, Maine (lat. 43°09'N, long. 70°04'W). The 201-cm long female weighed 113.2 kg and was freshly dead, still bleeding freely from symmetrical injuries to the left and right sides of both the upper and lower jaws and the flippers. The lungs contained foamy materials and were mottled white, indicating drowning as the immediate cause of death. Many gill nets were present in the area, and the sym- metrical nature of the injuries indicated that the animal had become entangled in the mesh, drowned, and perhaps been freed or discarded dur- ing hauling of the net. A humpback whale, Megaptera novaengliae, was entangled in a gill net for 2 h before freeing itself on the same day in the same general area. Gross autopsy revealed several cysts in the ab- dominal muscles of the lower left side and a 5 cm x 7.5 cm yellow, pussy abscess 15 cm anterior and dorsal to the right mammary gland, perhaps caused by a bladderworm stage (plerocercoid) of Monorygma grimaldi (Geraci et al.M. No other parasites were found, although all major organs except the brain were inspected. Tissue and organ weights are shown in Table 1. The length and 'Geraci, G., S. A. Testaverde, D. J. St. Aubin, and T. H. Loop. 1976. A mass stranding of the Atlantic white-sided dolphin, Lagenorhynchus acutus: a study into pathobiology and life his- tory. Unpubl. manuscr., 166 p. submitted to Marine Mammal Commission by New England Aquarium, Boston. Table l. — Tissue and organ weights of a Lagenorhynchus acutus from Jeffreys Ledge, Maine. Weights not corrected for blood loss. Weight Weight Tissue or organ (kg) Tissue or organ (kg) Muscle 69.0 Lett kidney 0.473 Blubber and tins 245 Right adrenal 0.011 Gastrointestinal tract 6.8 Left adrenal 0.010 Liver 3.2 Spleen 0.084 Heart 0.959 Right ovary 0.0051 Right lung 1.117 Left ovary 0.0047 Left lung 1.149 Bones' 5.4 Right kidney 0.476 Total 113.2 weight of this animal fit well on a regression line developed for this species (Geraci et al. see foot- note 1). From a length-age relationship (Geraci et al. see footnote 1), it is likely that this female was between 2y2 and 3 yr old, and was immature. The stomach contained 980 g of food, including four 25- to 30-cm herring, Clupea harengus, three partly digested (total weight 340 g) and one skele- ton (15 g); and one partly digested short-finned squid, Illex illecebrocus (anterior mantle length 17.5 cm, weight 90 g), with remains of 10 other squid of this species (represented by 5 complete pairs of beaks plus 5 single anterior beaks). Mean length of anterior beaks (± SD) was 1.42 ± 0.03 cm, corresponding to mantle lengths from 17 to 19 cm (Testaverde^). Also 10 left and 11 right otoliths from silver hake, Merluccius hilinearis, (mean size ± SD = 1.15 ± 0.06 cm) indicated consumption of at least 11 fish of 22-26 cm fork length (Nichy 1969). From these data and the literature it appears that C harengus and/, illecebrocus are staples in the summer diet of white-sided dolphins. A 161-kg female collected on 14 September 1954, off Cape Cod contained 12 fresh herring, digested fish (ap- parently herring), and squid (Schevill 1956). A 180-cm long male driven ashore with pothead whales, Globicephala melaena, in Newfoundland on 30 July 1954, contained herring and short- finned squid (Sergeant and Fisher 1957). Short- finned squid was the most common food in the stomachs of white-sided dolphins which mass- stranded at Lingley Cove, Maine, on 6 September 1974; however, no herring were found despite the fact that "brit" herring were present in the cove (Geraci et al. see footnote 1). Smelt, Osmerus mor- dax, remains, were found in five individuals; silver hake had been eaten by one individual; and un- identified crustacean remains were found in another stomach. Schools of white-sided dolphins were unusually common in the Gulf of Maine in 1976, perhaps because squid were abundant, possibly as a result of this year's unusually high sea temperatures 'Bones were bleached and dried in the laboratory. ^Testaverde, S.A. 1975. An informal discussion concerning the cestode Phyllobothrium sp. in squid, Illex illecebrocus illece- brocus and its possible relationship to marine mammals. Un- publ. manuscr., 20 p. 475 (Anonymous 1976; Prescott and Moore 1976). Silver hake, normally of variable abundance here (Bigelow and Schroeder 1953) was also abundant during 1976. On several different occasions, groups of 6-30 white-sided dolphins were seen by one of us (SKK) swimming close to pods of either finback whale, Balaenoptera physalus, or humpback whale, Megaptera novaeangliae , and apparently feeding with them. Acknowledgment We are grateful to the crew of the MV Exxon Bay State for procuring this specimen; to John E. Fitch, California Department of Fish and Game, for otolith identification; and to Fred Nichy, North- east Fisheries Center, National Marine Fisheries Service, NOAA, Woods Hole, Mass., for providing data on otolith growth. Literature Cited ANONYMOUS. 1976. Rise in squid Down East traced to warmer waters. Ellsworth (Maine) American, August 26, 1976, p. 1. BIGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildi. Serv., Fish Bull. 53, 577 p. NiCHY, F. 1969. Growth patterns on otoliths from young silver hake, Merluccius bilinearis (Mitch.). Int. Comm. Northwest Atl. Fish. Res. Bull. 6:107-117. PRESCOTT, R., AND M. MOORE. 1976. Cape Cod squid influx and pelagic bird gathering. Event 93-76, The Center for Short-Lived Phenomena, Re- view of Annual Events. Camb., Mass., Nov. 19, 1976. SCHEVILL, W. E. 1957. Lagenorhynchus acutus o^ Cape CoA. J. Mammal. 37:128-129. Sergeant, D. E., and H. D. Fisher. 1957. The smaller Cetacea of eastern Canadian waters. J. Fish. Res. Board Can. 14:83-115. College of the Atlantic Bar Harbor, ME 04609 10 Riggs Street Gloucester, MA 01930 103 Riverdale Drive Orono, ME 04473 Steven K. Katona Salvatore a. Testa VERDE Bradley Barr reciprocal hybridization between the california and gulf of california grunions, leuresthes tenuis and leuresthes sardina (atherinidae) The California grunion, Leuresthes tenuis, and the Gulf of California grunion, L. sardina, are the only fishes that temporarily leave the water during spring high tides to deposit their eggs in beach sand (Walker 1952). The eggs develop in the nearly dry sand and hatch when uncovered and agitated by the surf of the next series of high tides. The grunions have an allopatric distribution. The California grunion ranges from Monterey Bay, Calif., to Bahia Magdalena, Baja California Sur. The Gulf grunion is endemic to the Gulf of California, ranging from Bahia Concepcion, Baja California Sur, and Guaymas, Sonora, Mexico to the mouth of the Rio Colorado (Moffatt and Thom- son 1975). Recent comparisons show that morphological, physiological, and behavioral differences exist be- tween the grunions. Morphologically very similar, the most diagnostic characterictics distinguishing them are lateral scale row counts; the mean number in L. tenuis is 75 and in L. sardina is 55. Gulf grunion adults are also significantly longer, more slender, have a smaller eye diameter, and are more lightly pigmented than those of the California grunion (Moffatt 1974; Moffatt and Thomson 1975). Gulf of California grunion have wider embryonic and larval thermal tolerances, a higher larval preferred temperature, and wider larval salinity tolerances (Reynolds and Thomson 1974a, b, c; Reynolds et al. 1976, 1977; Moffatt 1977). Light response remains positive in Gulf grunion through adulthood, whereas the response shifts from positive in the larvae to negative in the adults of the California grunion (Walker 1952; Reynolds and Thomson 1974c; Reynolds et al. 1977). In response to the shorter wave period in the northern Gulf of California, the duration of the spawning act of the Gulf grunion females is much briefer than that of the California grunion females (Thomson and Muench 1976; Muench 1977). Only recently has the congeneric status of the grunions been recognized (Moffatt 1974; Moffatt and Thomson 1975). Evidence to date indicates that the California grunion, the less primitive of the two species, has adapted to the less fluctuating tidal and thermal regimes of the California coast, following isolation from an ancestral type by the 476 Baja California peninsula (Moffatt and Thomson 1975; Moffatt 1977). Hybridization and hybrid survival experiments have been widely used as indices of divergence and have made valuable contributions as a tool in the definition of phylogenetic relationships (Hubbs 1967, 1970). In an attempt to further illuminate the relationship between the grunions, we made artificial and reciprocal crosses and we report on the first successful reciprocal hybridization of Leuresthes tenuis and L. sardina. Materials and Methods Adult grunions, although easily obtained in large numbers, are difficult to maintain and transport alive. On 18 March 1976 (2330 PST), milt from six California grunion males was col- lected at Scripps Beach, La Jolla Calif., mixed in the beaten yolks of two hen eggs (Bratanov and Dikov 1961), and transported to El Golfo de Santa Clara, Sonora, Mexico. The milt-yolk mixture, maintained between 16° and 20 °C, was used to fertilize the eggs from 8 to 10 Gulf grunion females obtained at El Golfo on the following day ( 19 March) during a spawning run which began about 1700 MST. During this same run Gulf grunion milt was collected from 6 to 7 males, transported in the same manner and used to fertilize California grunion eggs from about 10 females obtained dur- ing a run that night at La Jolla at 0115 PST (20 March). One prior and four subsequent attempts to hybridize the grunions were made during the 1975, 1976, and 1977 spawning seasons, but these were unsuccessful because one or both grunions failed to spawn. The female grunions were rinsed thoroughly in clean seawater before their eggs were stripped directly into the milt-yolk mixture. The mixture was diluted slightly with fresh seawater to in- crease sperm motility, gently agitated, and kept cool until the end of the spawning run. The eggs were then strained, rinsed with seawater, and placed in plastic refrigerator containers between moist paper towels for transport and incubation. Conspecific control embryos of each species were obtained by mixing eggs and milt in a bucket of seawater, one-third full, and did not involve the transportation or preservation of milt in hen yolk. When spawning individuals were plentiful, as at the Gulf grunion run on 19 March, six to nine males were stripped per one female in order to achieve maximum fertilization levels (Moffatt 1977). Both sets of hybrid fertilized eggs and the con- specific controls of L. tenuis (18 March) and L. sardina (19 March) were transported from San Diego, Calif., to the University of Arizona at Tuc- son aboard commercial airlines. Upon arrival (22 h postfertilization in L. sardina x L. tenuis and L. sardina controls; 13 h in L. tenuis x L. sardina; and 32 h inL. tenuis controls) each set of eggs was inspected. Their development was monitored daily thereafter. Both California grunion spawning runs were sparse at La Jolla. Therefore, the greatest portion of eggs and sperm available were devoted to the hybridization experiments and a low conspecific L. tenuis sample size resulted. Consequently, the de- velopmental and hatching data reported herein for these embryos are a compilation of these few controls and egg sets obtained on other occasions, incubated at 20°C from 12 h postfertilization (Mof- fatt 1977). Yolk-sac larvae of the two hybrids and the con- specific controls were placed in separate tanks containing artificial seawater and raised on newly hatched Artemia, freeze-dried marine zooplank- ton, commercial staple food, and frozen Artemia nauplii. Larvae of the hybrids and controls were maintained for nearly 5 mo although initial mor- tality rates (first 2 mo) in all groups were high ( >90% ). On 19 August, 141 days posthatching, the aquaria air lines were fouled by compressor oil and the few remaining hybrids and controls died. Only two L. tenuis x L. sardina and nine L. sardina x L. tenuis individuals survived to a size ( >12 mm) at which the scale rows could be counted. This is not to imply that scales might not have been pre- sent prior to this time, merely that no attempt was made to count them. Results At 22 h postfertilization, cleavage had pro- gressed to the gastrula stage in L. sardina x L. tenuis embryos as it had in theL. sardina controls. The L. tenuis x L. sardina hybrids had reached a 32-cell blastodic stage at 13 h postfertilization as do L. tenuis embryos. Artificial fertilization levels in the conspecific controls fell between 85 and 99'^ during the peaks of their spawning seasons when male to female ratios of 6 or 9:1 were available. The fertilization 477 levels of both hybrids ranged from 60 to 70%. These diminished levels in the hybrids may have resulted from a combination of several factors such as: the low male to female ratios used ( <1:1); decreased sperm motility in the viscous hen-yolk medium; high sperm mortality due to time, star- vation, t^imperature shock, handling, etc. or par- tial reproductive isolation between the species in the form of mild fertilization block to non- conspecific spermatozoa. The grunions, L. tenuis and L. sardina, showed similar developmental rates (Moffatt 1977). De- velopment proceeded normally in the hybrids and at about the same rate as the controls. No unusual embryonic mortality was observed in the hybrids, evidence that these embryos were not gynogenetic hybrids (Moore 1955). Preliminary trials showed that hen's yolk and seawater alone will not initiate cleavage in Gulf grunion eggs. Precautions were taken to prevent conspecific milt contamination. Preliminary examination of cellular nuclei smears of develop- ing embryos immersed in colchichine revealed somatic chromosome numbers of about 2n = 40 in all four sets of embryos (controls and hybrids), further evidence that these embryos were true diploid hybrids. Grunion embryos will hatch after vigorous agi- tation in seawater. On 31 March at 284 h (11.8 days) postfertilization, 65.6% of the L. sardina x L. tenuis embryos hatched and 66.5% of the L. tenuis x L. sardina embryos hatched at 272 h (1 1.4 days). These hatch times are similar to those of the controls. Hatching can be induced in both grun- ions at 10.2 days postfertilization when embryos are incubated at 20° C (Moffatt 1977). Newly hatched L. tenuis larvae are typically more darkly pigmented; they have a larger eye diameter; they are stronger swimmers; and they are more capable of escaping net capture than newly hatched L. sardina larvae (Moffatt 1977). Leuresthes tenuis larvae are 10% longer (mean total length = 7.70 mm) than those of L. sardina ( mean total length = 6.93 mm). The greater length of the California grunion yolk-sac larvae occurs in the postanal region as in the adults. California grunion larvae are also 52%. heavier (mean dry weight = 0.340 mg) whereas, the mean dry weight ofL. sardina equals 0.223 mg (Moffatt 1977). The greater length and weight of the California grun- ion at hatching may be attributable to the 4.10 times greater ovum volume (Moffatt 1977; Moffatt and Thomson in press). These differences which distinguish the prolar- vae of L. tenuis and L. sardina were also observed in the hybrids. In most characteristics the L. tenuis x L. sardina larvae were not visibly distin- guishable from the maternal controls (L. tenuis), e.g., size, pigmentation, and swimming ability. However, the L. sardina x L. tenuis larvae ap- peared to be somewhat intermediate to the con- trols in extent of pigmentation and swimming ability. At 2 wk after hatching the length and pigment differences between the larvae were more pronounced. Premaxillary teeth were visible in the L. sardina x L. tenuis larvae but not in the reciprocal hybrids. Again, hybrids closely resem- bled the maternal controls. Gulf grunion adults typically have much stronger dentition than do the adults of the California grunion (Moffatt and Thomson 1975). As previously mentioned, the most diagnostic differences between the adult grunions are the lateral scale row counts. Scale counts of the 141- day-old controls were essentially the same as those of the adults (Table 1). The counts of the hybrids were intermediate and significantly dif- ferent from each other. Those shown by both hy- brids were significantly different from those of both parental species. The lateral scale rows of the hy- brids were closer in number to those of the mater- nal controls. Mean counts of L. sardina x L. tenuis were 32% closer to those of L. sardina; and L. tenuis x L. sardina were 20% closer to L. tenuis than to those of the paternal parents, L. tenuis and L. sardina. respectively.^ The intermediate counts indicate paternal genome influence and that these are indeed diploid hybrids. 'A mean hybrid count greater or less than 65 (the midvalue between the parental species) indicated the affinity to one parent or the other. The numerical affinities (percentages) were calcu- lated as the ratio of the differences between 65 and the hybrid count and between 65 and the adult counts. Table l. — Means, ranges, n, and P values of lateral scale row counts observed in 141-day-old hybrids and controls and adults of the grunions, Leuresthes tenuis and L. sardina. Parents : L tenuis . L sardina Juveniles Adults P Juveniles Adults P L tenuis L. sardina X = 75 5 (75-76) n = 2 X = 67.0* (66-68) n =2 P--0.01 X = 74 6 (69-80) n = 143 •0 6 X = 61.8 (61-63) n = 9 X = 55.1 (54-57) n = 9 P< 0.001 X = 55 3 (51-60) n = 177 >0.7 'Student's f-test comparison of the lateral scale rovi( counts between the two hybrids P- 0 001. 478 Discussion Natural hybridizations are reportedly more common among freshwater fishes than among marine fishes (Hubbs 1955). Hubbs (1970) stated that, "teleost hybrids are relatively easily pro- duced and if the parental morphology is similar the hybrids are easily reared." The results of natural and artificial amphibian and teleost crosses have been widely employed for estimating degrees of phylogenetic divergence, revealing sys- tematic patterns, and explaining mechanisms controlling development and differentiation. Davidson (1968) reports that the closer the phylogenetic relationship between the species hybridized, the less likely the hybrid genome con- trol will be displayed early in development. This is because the mechanical aspects of early develop- ment tend to be similar in closely related species and may be primarily under the control of mater- nal RNA accumulated in the egg prior to fertiliza- tion. Davidson believes this, at least in part, ac- counts for the commonly observed resemblance of hybrids in early developmental stages to the mat- ernal parent. The genetic influence of the paternal genes in the hybrid genome may not be apparent phenotypically until long after the onset of dif- ferentiation (Davidson 1968). It is possible that such mechanisms account for the maternal resemblance pattern observed in the grunion hybrids as well. The hybrids resembled the maternal parents in overall size and body proportions, coloration, swimming ability, net- escape capability, and dentition until long after hatching. Only when the lateral scale rows were counted at 141 days after hatching did the influence of paternal genes become visibly and quantitatively apparent. The numerous artificial and two natural hy- bridizations (interspecific and intergeneric) re- ported among the Atherinidae are reviewed by Hubbs and Drewry (1959), Rubinoff (1961), and Hubbs (1970). Natural hybrids reported between Menidia menidia and M. beryllina along the At- lantic coast of Florida (Gosline 1948) exhibit in- termediate counts (i.e., scales and fin rays). Most of the experimental crosses between these atherinids resulted in low developmental success and low survival rates except those of M. beryllina 9 X M. menidia 6 (Rubinoff 1961). Rubinoff did not report whether any intermediate characteris- tics existed in these hybrids nor was the reciprocal cross attempted. Geographically isolated species forms adapt to their respective environments by the evolution of appropriate gene complexes. Then, if sympatry reccurs and hybridization takes place, hybrid in- dividuals will usually be selected against (Mayr 1963; Ford 1964). Hybrids not selected against will usually be successful over only a narrow geo- graphical range, since in animals, natural hybrid- ization is commonly associated with environ- mental perturbation (Mayr 1963; Manwell and Baker 1970). The Menidia species are sympatric and hybrid- ization does occur in northern Florida, a very nar- row portion of the overlap in their ranges (Gosline 1948). Like these species, grunions are marine fishes with similar, but not identical, ecological preferences. However, the grunions are allopatric and natural hybridization is not possible. According to Mayr (1963), some investigators argue that renewed sympatry with hybridization is required as a process of speciation in order to "perfect isolating mechanisms," and, therefore, unlike the Menidia species, the heterospecific status of the grunions may be questioned, espe- cially in light of the hybridization success reported herein. We conclude that, despite our success at hybridizing L. tenuis and L. sardina, the mor- phological, physiological, and behavioral distinc- tions between them warrant their continued rec- ognition as separate species. Acknowledgments We thank O. M. Moffatt, V. J. Moffatt, and D. Dutcher for accompanying and assisting the senior author on the collecting excursions. We especially thank O. M. Moffatt for saving the pre- served milt of the Gulf grunion from being swept away by an unusually high wave at La Jolla. We acknowledge the time and efforts of P. C. Cook in preparing the preliminary somatic-cell smears, and we appreciate the recommendations concern- ing the preparation of this manuscript given by W. W. Reynolds, C. D. Ziebell, E. A. Stull, and J. S. Frost. Literature Cited BRATANOV, C, and v. DIKOV. 1961. Sur certaines particularites du sjserme chez les pois- sons. Proc. Int. Congr. Anim. Reprod. 4:895-897. Davidson, E. H. 1968. Gene activity in early development. Academic Press, N.Y., 375 p. 479 FORD, E. B. 1964. Ecological genetics. John Wiley & Sons, Inc., N.Y., 335 p. GOSLINE, W. A. 1948. Speciation in fishes of the genus Menidia. Evolu- tion 2:306-313. HUBBS, C. L. 1955. Hybridization between fish species in nature. Syst. Zool. 4:1-20. HUBBS, C. 1967. Analysis of phylogenetic relationship using hy- bridization techniques. Bull. Nat. Inst. Sci. India 34:48- 59. 1970. Teleost hydridization studies. Proc. Calif Acad. Sci. 38:289-298. HUBBS, C, AND G. E. DREWRY. 1959. Artificial production of an intergeneric atherinid fish hybrid. Copeia 1959:80-81. Manwell, C, and C. M. a. Baker. 1970. Molecular biology and the origin of species, heterosis, protein polymorphism and animal breed- ing. Univ. Wash. Press, Seattle, 394 p. MAYR, E. 1963. Animal species and evolution. Belknap Press of Harvard Univ. Press, Cambr., 797 p. MOFFATT, N. M. 1974. A morphometric and meristic comparison of the Gulf grunion, Leuresthes sardina (Jenkins and Evermann), and the California grunion, Leuresthes tenuis ( Ayres). MS Thesis, Univ. Arizona, Tucson, 36 p. 1977. Thermal effects on the survival and development of embryonic grunions, Leuresthes sardina and L. tenuis. Ph.D. Thesis, Univ. Arizona, Tucson, 88 p. MOFFATT, N. M., AND D. A. THOMSON. 1975. Taxonomic status of the Gulf grunion (Leuresthes sardina) and its relationship to the California grunion (L. tenuis). Trans. San Diego Soc. Nat. Hist. 18:75-84. In press. Tidal influence on the evolution of egg size in the grunions (Leuresthes). Environ. Biol. Fishes. MOORE, J. A. 1955. Abnormal combinations of nuclear and cytoplasmic systems in frogs and toads. Adv. Gen. 7:139-182. MUENCH, K. A. 1977. Behavioral ecology and spawning periodicity of the Gulf of California grunion, Leuresthes sardina. Ph.D. Thesis, Univ. Arizona, Tucson, 92 p. REYNOLDS, W. W., AND D. A. THOMSON. 1974a. Tem|)erature and salinity tolerances of young Gulf of California grunion, Leuresthes sardina (Atherini- formes: Atherinidae). J. Mar, Res. 32:37-45. 1974b. Ontogenetic change in the response of the Gulf of California grunion, Leuresthes sardina (Jenkins & Ever- mann), to a salinity gradient. J. Exp. Mar. Biol. Ecol. 14:211-216. 1974c. Responses of young Gulf grunion, Leuresthes sar- dina, to gradients of temperature, light, turbulence and oxygen. Copeia 1974:747-758. REYNOLDS, W. W., D. A. THOMSON, AND M. E. CASTERLIN. 1976. Temperature and salinity tolerances of larval California grunion, Leuresthes tenuis (Ayres): a com- parison with Gulf grunion, L. sardina (Jen- kins & Evermann). J. Exp. Mar. Biol. Ecol. 24:73-82. 1977. Responses of young California grunion, Leuresthes tenuis, to gradients of temperature and light. Copeia 1977:144-149. 480 RUBINOFF, I. 1961. Artificial hybridization of some atherinid fishes. Copeia 1961:242-244. THOMSON, D. A., AND K. A. MUENCH. 1976. Influence of tides and waves on the spawning be- havior of the Gulf of California grunion, Leuresthes sar- dina (Jenkins and Evermann). Bull. South. Calif Acad. Sci. 75:198-203. WALKER, B. W. 1952. A guide to the grunion. Calif Fish Game 38:409- 420. Nancy M. Moffatt Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NOAA P.O. Box 271, La Jolla, CA 92038 DONALD A. Thomson Department of Ecology and Evolutionary Biology University of Arizona Tucson, AZ 85721 TYCHOPLANKTONIC BLOODWORM, GLYCERA DIBRANCHIATA, IN SULLIVAN HARBOR, MAINE The bloodworm, Glycera dibranchiata, is distri- buted from the Gulf of St. Lawrence to the Gulf of Mexico and from central California to lower California and Mexico. It occurs from intertidal water to 402 m depth (Pettibone 1963), but it is more abundant in shallow coastal water. In Maine and Nova Scotia the worms are dug commercially along the coast from the upper layers of the inter- tidal sand-silt-clay strata (Dow and Creaser 1970; Anonymous 1974; Glidden^). Spawning bloodworms are briefly pelagic occur- ring in large numbers as they swarm in the after- noon. Creaser ( 1973) observed swarming in Maine during June. Simpson (1962) reported swarming both in June and November-December, suggest- ing a biannual spawning in Maryland. Klawe and Dickie (1957) did not observe swarming by blood- worms in Nova Scotia, although other evidence indicated that the worms spawned in mid-May. They suggested that the worms had a short noc- turnal swarming period making them difficult to observe. Simpson (1962) checked this possibility •Glidden, P. E. 1951. Three commercially important poly- chaete marine worms from Maine: Nereis (Neanthes) virens, Glycera dibranchiata, Glycera americana. Rep. to Maine Dep. Sea Shore Fish., Augusta, Maine. in Maryland by making 40 observations with a night-light between June and November. No worms appeared at the surface under the light. Individual bloodworms occasionally are pelagic when not spawning. Pettibone (1963), when not- ing the sightings of others, reported a bloodworm swimming at the surface of Eel pond. Woods Hole, Mass., on the evening of 17 August 1943; another at the surface perhaps at the same pond on 28 January 1876; and another in Delaware Bay on 29 January 1957. No time was given for the two January sightings. On 2 October 1969, E. P. Greaser, Jr. sighted a bloodworm at the surface near a dock on McKown Point, Boothbay Harbor, Maine. The large nonspawner was observed at noon swimming during a flood tide. We have found that nonspawning bloodworms may also occur as fairly abundant members of the tychoplankton — bottom dwellers that are either swept upward with tidal currents or migrate upward at night. This study was originally designed to sample lar- val Atlantic herring, Clupea harengus harengus Linnaeus, and these results will be presented la- ter. The implications of a large incidental catch of bloodworms prompted our writing this note. Materials and Methods The site of this investigation, Sullivan Harbor, is an embayment along the eastern coast of Maine. It is divided into northern and southern sectors by a constriction formed by an island, point of land, and ledges (Figure 1). The southern sector opens onto Frenchman Bay, which in turn opens onto the Gulf of Maine. At its upper end, the northern sec- tor constricts into a tidal falls. A narrow channel extends north of the falls eventually bifurcating into broad extensive shallows. Only small streams enter these shallows about 5 km north of the highway bridge. Sullivan Harbor is thus rela- tively saline (31-32%o). Six sampling stations were located within the northern sector of the harbor; two in the landward end of the channel (No. 3, 4), two in the seaward end (No. 1,2), and one at each seaward entrance to the subtidal flats (No. 5, 6). At each station within the channel, four lines of buoyed and anchored nets were set (Graham and Venno 1968). On each line one net fished near the surface and a second at 3 m just above the edge (4 m) of the subtidal chan- nel (Figure 1). A third net fished below the edge at 10 m and a fourth near the bottom (12-20 m). At the entrance to the subtidal flats, one net was suspended near the surface and another at 3 m just above the bottom. The nets were set at each station at dusk and retrieved at dawn, fishing approximately one tidal cycle. Calibrated meters centered within the nets determined the amount of water strained. The contents of the nets were preserved in the field using a 5% Formalin^ solution. The sexes of the worms were determined at a later date by inspec- tion of the coelomic contents. Since variable shrinkage of the worms made length measure- ments unreliable, dry weight was obtained for each worm. Results The nets strained 72 bloodworms from tidal cur- rents during 6 of 10 cruises in autumn and winter 1974-75. During 1974, the nets captured 2 worms on 14 October, 7 on 1 1 November, 2 on 5 December, 51 on 10 December, and 1 on 19 December. During 1975, the nets captured nine worms on 2 De- cember. Only five worms were immature; their weights varied from 0.02 to 0. 1 1 g. Mature females outnumbered mature males about two to one (41:24). The mean weights of the two sexes were similar, 0.57 g, and their range varied from 0. 1 1 to 1.47 g. Bloodworms were dispersed throughout the water column and over both the channel and sub- tidal flats. Nets at all stations and depths captured worms. The average number netted was three and ranged from one to seven. Of the 72 worms from all cruises, nets set in the channel contained 58 worms and those over the subtidal flats held 14. Their numbers decreased vertically: 33 near the surface, 17 at 3 m, 15 at 10 m, and 7 near the bottom. An exceptionally large catch per unit effort was obtained on 10 December. During the 10 cruises the nets strained approximately 8,000 to 20,000 m^ of tidal water per cruise. Five of the sets yielded catch rates varying from 0. 1 to 0.7 worm/1,000 m^. A sixth set (10 December) yielded 3.38 worms/ 1,000 m^. This catch rate was sufficiently large to permit comparison of synoptic catch rates with location and depth. The four lines of nets in the channel strained 39 worms from 10,194 m^, yield- ing a catch rate of 3.8 worms/1,000 m^ Those nets ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 481 68 15' Longitude W 43 32 43 30- FlGURE 1— Sampling stations (1-6) for larval Atlantic herring with sets of buoyed and anchored nets in Sullivan Harbor, Maine. 43 28- -43 32 -43 30 43 26 in the flats strained 12 worms from 4,889 m^, yielding 2.4 worms/1,000 m^. Shallow nets, above the channel edge and those near the surface of the flats, captured 37 worms by straining 9,614 m^ for a catch rate of 3.8 worms/1,000 m^. Deep nets, below the channel edge and those near the bottom of the flats, captured 14 worms by straining 5,469 m^ for a catch rate of 2.6 worms/1.000 m^. The numbers of worms captured in the nets were few when compared with the numbers of smaller tychoplankters, such as amphipods. Each indi- vidual weight, however, was relatively large com- pared with those individuals of more numerous taxa and suggested that a large biomass of blood- worms sometimes enters the water column of the harbor. Discussion The mature bloodworms captured during winter in buoyed nets at Sullivan Harbor were not free- swimming spawners. Creaser (1973) sampled a small worm flat at Wiscasset, Maine, from No- vember 1967 to August 1969. During that time, among the many worms dug, only three spawners occurred during winter. Analysis of his collections showed that egg diameters increased somewhat during December and January but ceased growth during the colder months of February and March. Spawning was triggered in June by formation of the epitoke, the growth of eggs to the spawning "range" and a water temperature of at least 13°C. These conditions were not found in the present 482 study. Also, we did not detect any morphological changes that accompany formation of the epitokes as described by Simpson (1962). Swimming bloodworms at night have also been reported for two other Maine inshore waters. Dean^ saw 22 bloodworms during observations made between 24 January and 29 March 1977 on 33 nights. The worms were present during five nights in March and 15 were collected under a night-light in the Damariscotta River, Maine — 8 on 11 March and 7 on 12 March. The gametes of the worms were not sexually mature and the pre- sence of the worms near the surface at night was not related to spawning. Dean also reported that buoyed and anchored nets set in Montsweag Bay and the Sheepscot estuary between 1970 and pre- sent captured 22 glycerids, some of which were G. dibranchiata. In contrast, the senior author of this paper did not capture bloodworms in buoyed and anchored nets set in the Sheepscot estuary over the same time period and in the same vicinity. Possibly, the swimming of bloodworms at night is sporadic. A recent study of residual currents in Sullivan Harbor suggested that the relatively shallow nets above the edge of the channel (Figure 1) and at the surface over the tidal flats strained a residual seaward flow transporting tychoplankters and the relatively deep nets strained a residual landward flow. Distribution of bloodworms throughout the water column would, therefore, insure their wide dispersal by horizontal tidal currents, and it is unlikely that after a tidal cycle they would regain the location of their original burrows. We hope to study further the bloodworms of Sullivan Harbor and do not wish to speculate on their origin or fate at this time. Rather, it is our purpose to suggest that researchers investigating bloodworms within their bottom habitat should also examine their possible role as tychoplankters for two reasons: populations of this important commercial species in separate flats may become intermixed, introducing problems in their man- agement; and the reestablishment of worm popu- lations previously destroyed by pollution or other environmental catastrophe might proceed more rapidly in those areas where there is winter trans- port of mature worms, as well as the "normal" dispersion of late spring larvae. Acknowledgements We thank C. Adams and D. Clifford of the Maine Department of Marine Resources for collecting the worms, sometimes under severe winter condi- tions, and for processing the worms in the labora- tory. We thank D. Dean for permitting us to cite his unpublished manuscript. Literature Cited ANONYMOUS. 1974. Environmental inventory, benthic invertebrates. A socio-economic and environmental inventory of the North Atlantic Region, Vol. 1, p. 71-73. Res. Inst. Gulf Maine. Greaser, E. p., Jr. 1973. Reproduction of the bloodworm (Glycera dibran- chiata) in the Sheepscot Estuary, Maine. J. Fish. Res. Board Gan. 30:161-166. Dow, R. L., AND E. P. Greaser, Jr. 1970. Marine bait worms, a valuable inshore resource. Atl. States Mar. Fish. Gomm. Leafl. 12, 4 p. Graham, J. J., and P. M. W. Venno. 1968. Sampling larval herring from tidewaters with buoyed and anchored nets. J. Fish Res. Board Can. 25:1169-1179. KLAWE, W. L., AND L. M. DICKIE. 1957. Biology of the bloodworm, Glycera dibranchiata Ehlers, and its relation to the bloodworm fishery of the Maritime Provinces. Fish. Res. Board Gan., Bull. 115, 37 p. PETTIBONE, M. H. 1963. Marine polychaete worms of the New England re- gion. U.S. Natl. Mus., Bull. 227, 356 p. Simpson, M. 1962. Reproduction of the polychaete Glycera dibran- chiata at Solomons, Maryland. Biol. Bull. (Woods Hole) 123:396-411. JOSEPH J. Graham EDWIN P. Greaser, Jr. Maine Department of Marine Resources Research Laboratory West Boothbay Harbor, ME 04575 'Dean, D. The swimming of bloodworms (Glycera spp.) at night. Unpubl. manuscr. SIMULATED FOOD PATCHES AND SURVIVAL OF LARVAL BAY ANCHOVY, ANCHOA MITCHILLI. AND SEA BREAM, ARCHOSARGUS RHOMBOIDALIS Survival rates of laboratory-reared marine fish larvae often are directly related to prey concentra- tion. Best survival usually has been reported when prey are available at concentrations > 1,000/1 ( O'Connell and Raymond 1970; Laurence 483 1974, 1977). Houde (in press) recently demon- strated that survival of three species of marine fish larvae from hatching to metamorphosis was 109c or higher when mean prey concentrations were only 34-130/1. But, he also found enhanced surviv- al when food concentrations were increased. For significant numbers of larvae to survive the tran- sition stage from yolk nutrition to active feeding, some researchers believe that dense patches of prey must occur in the sea (O'Connell and Raymond 1970; Hunter 1972). Such patches might occur at densities of 10 to 1,000 times above the mean prey density. Lasker (1975) has discussed the dense patches of the dinoflagellate Gym- nodinium splendens, which serves as prey for lar- val northern anchovy, Engraulis mordax, in the California Current and their possible relationship to larval survival. Hunter and Thomas (1974) de- monstrated that larval northern anchovies were able to remain in patches of G. splendens that were artificially created in laboratory experiments. In two series of laboratory experiments we have examined the effect of two simulated patches of prey on survival in the bay anchovy, Anchoa mitchilli, and the sea bream, Archosargus rhom- boidalis. Patches were simulated during the first 6 days after hatching, when these larvae are most susceptible to starvation mortality. The purpose of the experiments was to determine if prey at high density that were offered for more than some minimum period would result in survival rates of larvae that approached those obtained at a high, constant prey concentration. This would indicate that the larvae were able to obtain a daily ration suitable for maintenance and growth by increas- ing their feeding rate during the period of expo- sure to the patch concentration of prey. At the low prey concentrations usually found in the sea, a relatively great expenditure of energy would be required by larvae to obtain the minimum daily ration for maintenance and growth. Such larvae might weaken or fail to grow and thus be more susceptible to starvation or predation. Methods Larvae were hatched from fertilized eggs that were collected in plankton nets from Biscayne Bay, Fla. In each experimental trial 140 sea bream eggs were stocked (2.0/1) and 280 bay anchovy eggs were stocked (4.0/1) in a 76-1 glass aquarium. Lar- vae were reared for 10 days at 26±1°C. Salinities ranged from 30.0 to 32.5%o for bay anchovy and 484 33.0 to 33.5%o for sea bream. Lighting was pro- vided at 2500-2800 Ix by 40-W, cool-white fluores- cent tubes. A 13 h light-11 h dark schedule was maintained. Tanks were isolated in a black plastic enclosure and all light was extinguished during the dark periods. Sea bream and bay anchovy lar- vae do not feed in the dark. At the end of experi- ments, survivors were preserved in 5% Formalin^ and measured using an ocular micrometer. Prey were the nauplii and copepodid stages of copepods, approximately 50-100 /xm in diameter, that were collected in 53-/Ltm mesh plankton nets. Prey concentrations were determined by counting organisms in 100- to 200-cm'^ aliquots from the rearing tank (Houde 1975, 1977) several times per day during the 13-h feeding period. Background (i.e., nonpatch) prey levels were set at 25-50/1; this concentration was maintained when patch concentrations were not offered from 2-6 days after hatching and continuously from 7-10 days after hatching. The patch concentration was 500 prey/1. Patches were provided for periods rang- ing from 1.5 to 11 h (Tables 1,2). Both Oh, at which no patches were provided, and 13h, at which a constant 500/1 prey concentration was main- tained, also were included in the series of experi- ments for each species. The patch schedules were maintained for only the first 5 days of active feed- ing because larvae that survived that period had greatly increased their searching ability and were less dependent on high prey concentrations for successful feeding. Patches were created by adding prey to obtain the 500/1 concentration. After larvae had fed at the patch concentration for the desired period, prey were reduced to 25-50/1 by siphoning them out of the system through a 280-/>tm mesh screen and replacing the siphoned water with 26°C filtered seawater from a 150-1 header tank. Sea bream larvae had no difficulty avoiding the siphon and its screen during water exchanges, but precautions were necessary for bay anchovy larvae. A 280-)U,m mesh partition was used to "herd" anchovy larvae toward one end of the tank prior to each siphoning procedure. Siphoning procedures and water ex- changes also were carried out in the 0-h and 13-h patch period experiments to insure that those lar- vae were exposed to the same procedural distur- bances as larvae in experiments where prey con- centration was being varied. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Larvae were exposed to the patch concentration twice during each 13-h feeding period to obtain the total desired time at the patch level. For example, for a 6-h patch exposure, the prey concentration was adjusted to 500/1 from 0800 to 1100; it was then quickly reduced to 25-50/1, where it was maintained until 1800, when the prey concentra- tion was readjusted to 500/1 for the remaining 3 h of the light cycle. Results Table 2. — Survival and standard lengths of Ardiosargus rham- boidalis larvae at 10 days after hatching, based on 140 eggs and variable patch exposure time. A patch is a prey concentration of 500/1. Nonpatch levels were 25-50/1. Patch conditions were pre- sented to larvae on days 2-6 after hatching. Patch exposure time (h) Survival Standard l( sngth (mm) Percent No. Mean SD '0.0 3.57 5 4,21 0.44 1.5 22.00 31 404 0.56 3.0 32 14 45 3 77 0.56 6.0 59.29 83 387 0.40 9.0 41.43 58 3.31 0.41 11.0 6643 93 4.18 028 2130 42.00 59 4.21 0.31 Anchoa mi t chilli 'Food concentration was tneld constant at 25-50,'l during days 2-6. ^Food concentration was held constant at 500/1 dunng days 2-6, then re- duced to 25-50/1 during days 7-10 Percent survival ranged from 0.36% at 0-h patch exposure to 22.86% at 13 h (Table 1). The steady increase in survival as patch exposure time was increased was described by an exponential function, Y = 0.3038 e034i9x^ where Y = percent survival and X = hours at 500/1 prey concentration (coefficient of determination, r^ = 0.98). For the 500/1 patch concentration, there was no minimum time of exposure above which larval anchovy sur- vival increased sharply or equalled the survival obtained when larvae were exposed throughout the day to the 500/1 prey concentration. Surviving bay anchovies at 10 days after hatch- ing differed significantly in mean standard lengths (Table 1) among patch exposure times (analysis of variance, P<0. 001). Mean lengths at 3-, 6-, and 9-h patch exposure times were sig- nificantly greater than those at 11 and 13 h (Student-Newman-Keuls test, P <0.05). Table l. — Survival and standard lengths of Anchoa mitchiUi lar- vae at 10 days after hatching based on 280 eggs and variable patch exposure times. A patch is a prey concentration of 500/1. Nonpatch levels were 25-50/1. Patch conditions were presented to larvae on days 2-6 after hatching. Patch exposure Survival Standard length (mm) time (h) Percent No Mean SD '0.0 0.36 1 6.75 — 3.0 0.70 2 7.79 0.17 6.0 1 79 5 7.50 0.68 9.0 929 26 7.18 064 11.0 13.93 39 6.41 0.76 213.0 2286 64 6.58 0.58 'Food concentration was held constant at 25-50/1 during days 2-6. ^Food concentration was held constant at 500/1 during days 2-6. then re- duced to 25-50/1 durino davs 7-10. Archosargus rhoviboidalis Survival ranged from 3.57 to 66.43% for sea bream larvae over the range of patch exposure times (Table 2). The relationship between percent survival and patch exposure time was described by a power function, 7 = 25. 07395s: 0 2878^ where Y = percent survival and X = hours at 500/1 prey con- centration. Although the power function described the relationship reasonably well (coefficient of de- termination, r^ = 0.94), an asymptotic regression might be better to describe the relationship be- cause sea bream larvae exposed to a 500/1 patch density for between 3 and 6 h daily apparently survived as well as when the 500/1 prey concentra- tion was offered throughout the day. The power function is retained here because fits to the data by asymptotic regressions gave lower coefficients of determination, due to the relatively high variabil- ity in observed survival as patch exposure times increased. Mean lengths of survivors at 10 days (Table 2) differed significantly among patch exposure times (analysis of variance, P <0.001 ), but there was no clear relationship between the mean lengths that differed significantly (Student-Newman-Keuls test, P<0.05) and the time of exposure. Discussion There was a marked difference in response of bay anchovy and sea bream larvae to the simu- lated patch conditions. Sea bream survival im- proved greatly when larvae were presented with prey at 500/1 for more than 3 h/day, the observed survival then equaling that when they were of- fered a constant 500/1 prey concentration. Bay an- chovies were less successful in using the patch conditions to improve their survival, although in- creased survival rates did occur when larvae were exposed for more than 6 h to the patch concentra- tion. Results imply that first feeding bay anchovy may require a high and stable prey density to attain best survival in the sea, but that sea bream 485 are better adapted to survive under fluctuating food conditions. Survival observed in these experiments can be compared with that reported previously (Houde in press), when survival was related to prey den- sities that were held constant from day 2 to day 16. Predicted survivals at constant prey densities of 25-50/1 and 500/1 were 0.72-3.86% and 29.31%, respectively, for bay anchovy larvae; and 5.94- 16.61% and 70.45% for sea bream larvae. Ob- served survivals at 0-h and 13-h patch exposures (Tables 1, 2), which correspond to the 25-50/1 and 500/1 constant prey concentrations, were only slightly lower than those reported in the constant prey level experiments (Houde in press). The small differences probably were caused by the siphoning and water exchange procedures which did subject larvae to some stress. The similarity of results in the two reports indicates that the patch simulation procedure was effective in demonstrat- ing the impact of patches on larval survival. Growth results were inconclusive. Significant differences in mean lengths were observed among patch exposure times for both species (Tables 1,2). In sea bream there was no clear relationship be- tween mean lengths and patch exposure times, but, unexpectedly, bay anchovy mean lengths were smallest at the longest patch exposure times. Presumably only the hardiest larvae survived when patches were presented for only a short time, and these larvae also may have had a relatively great potential for growth. At the long exposures to patch densities, survival was better, but no im- provement in growth was noted, possibly because some larvae with relatively poor growth potential survived, or because of density-dependent effects on growth that have been previously observed (Houde 1975, 1977). Another compensating factor was that patches were only presented on day 2 to day 6 of the experiments, the prey concentrations in all experiments being held constant at 25-50/1 from day 7 to day 10. Only one possible patch regime was used in these experiments. It is possible that other patch densities or exposure schedules might alter re- sults or conclusions. An infinite number of possi- ble patch conditions could be simulated but future experiments should be delayed until the temporal and spatial scales of patchiness of organisms con- sumed by marine fish larvae are better known. Conditions that were simulated in these experi- ments do not discount the possible ability of larvae in the sea to maintain themselves within prey 486 patches that retain their integrity for days or weeks. Hunter and Thomas (1974) demonstrated that northern anchovy larvae could maintain themselves within small patches of Gymnodinium splendens in laboratory tanks. Lasker (1975) found that feeding northern anchovy larvae were relatively more abundant in the chlorophyll maximum layer of the Los Angeles Bight, where G. splendens was abundant, than in surface wa- ters, and he suggested that larvae might be able to maintain themselves in this rich source of food. Bay anchovy larvae in our experiments derived small benefits from the patch regime that we pro- vided, but there may be stable patch conditions in the sea which could greatly increase their poten- tial for survival. Acknowledgments This study was supported by the National Science Foundation, Biological Oceanography Program, Grant OCE 74-18141. John Hunter re- viewed and criticized an early draft of the manu- script. Scott Siddall and A. Keith Taniguchi as- sisted with the experiments. Literature Cited Houde, e. d. 1975. Effects of stocking density and food density on sur- vival, growth and yield of laboratory-reared larvae of sea bream Archosargus rhomboidalis (L.) (Sparidae). J. Fish Biol. 7:115-127. 1977. Food concentration and stocking density effect on survival and growth of laboratory-reared larvae of bay anchovy Anchoa mitchiUi and lined sole Achirus lineatus. Mar. Biol. (Berl.) 43:333-341. In press. Critical food concentrations for larvae of three species of subtropical marine fishes. Bull. Mar. Sci. HUNTER, J. R. 1972. Swimming and feeding behavior of larval anchovy, EngrauUs mordax. Fish. Bull., U.S. 70:821-838. HUNTER, J. R., AND G. L. THOMAS. 1974. Effect of prey distribution and density on the search- ing and feeding behaviour of larval anchovy EngrauUs mordax Girard. In J. H. S. Blaxter (editor). The early life history offish, p. 559-574. Springer- Verlag, N.Y. Lasker, R. 1975. Field criteria for survival of anchovy larvae: The relation between inshore chlorophyll maximum layers and successful first feeding. Fish. Bull., U.S. 73:453-462. Laurence, G. C. 1974. Growth and survival of haddock [Melanogrammus aeglefinus) larvae in relation to planktonic prey concen- tration. J. Fish. Res. Board Can. 31:1415-1419. 1977. A bioenergetic model for the analysis of feeding and survival potential of winter flounder, Pseudopleuronectes americanus, larvae during the period from hatching to metamorphosis. Fish. Bull, U.S. 75:529-546. O'CONNELL, C. P., ANn L. P. RAYMOND. 1970. The effect of food density on survival and growth of early post yolk-sac larvae of the northern anchovy iEn- graulis mordax Girard) in the laboratory. J. Exp. Mar. Biol. Ecol. 5:187-197. Edward D. Houde Richard C. Schekter Division of Biology and Living Resources Rosenstiel School of Marine and Atmospheric Science University of Miami 4600 Rickenbacker Causeway, Miami FL 33149 DISCOVERY OF JUVENILE PACIFIC SALMON (COHO) IN A SMALL COASTAL STREAM OF NEW BRUNSWICK Three juvenile Pacific salmon (Figure 1) were dis- covered in a small coastal stream in southern New Brunswick (Figure 2) in October 1976 while young Atlantic salmon, Salmo salar, were being col- lected for laboratory experiments. The Pacific salmon were not recognized by the electrofishing team, and their presence among the Atlantic salm- on was not realized until the fish were sorted in the laboratory some days or weeks later. Iden- tification as either coho salmon, Oncorhynchus kisutch, or chinook salmon, O. tshawytscha, was later confirmed by W. B. Scott of Huntsman Marine Laboratory, St. Andrews, N.B. Positive identification to species of these juvenile fish was not possible, but they were almost certainly coho salmon because of recent introductions of this species to the Atlantic coast. Coho salmon are not native to the Atlantic, and no populations reproducing in natural streams of the Atlantic coast are known. Two aquaculture operations using coho salmon are under way in Maine, and coho salmon smolts have been released in streams in New Hampshire and Massachusetts since 1969 and 1971, respectively (Figure 2, inset). Presumably, the parents were from one or more of these four operations. No adults have been re- ported from New Brunswick streams. When the coho calmon were recognized, further trips were made to obtain an estimate of their numbers in the stream, their size, and habitat preference in comparison with Atlantic salmon and brook trout, Salvelinus fontinalis , which were also present. The stream, known locally as Frost Fish Creek, drains into the estuary of the Digdeguash River about 250 m from the Digdeguash Falls. It is a small stream approximately 3 m wide in the lower kilometer where all fishing took place. Its drain- age area is approximately 570 ha. Discharge dur- ing low summer fiow reaches as little as 80 1/s (Symons and Harding 1974). The lowermost 0.25 km is steep with cascades and pools. The stream here is either open to the sky or overhung with alders. Most of the Atlantic salmon yearlings occur in this portion of the stream. Through the next 0.25 km upstream the gradient decreases; occasional riffles are separated by pools and slow- flowing water. Bankside cover consists of conifer- ous softwoods partially clearcut. Atlantic salmon yearlings and underyearlings occur in the riffles of this section while the pools and quieter water are inhabited by brook trout. Above this section the Figure l.— Underyearling coho salmon captured on 28 October 1976 in Frost Fish Creek, N.B. 487 Figure 2. — Streams of southern New Brunswick and their access. Dots, sites of spot checks; star, site where coho salmon were captured. Inset, province of New Brunswick, Canada, and northeastern United States showing location of aquaculture opera- tions (Maine) or release sites (N. H. and Mass.) of coho salmon (squares) with respect to location of underyearling coho salmon discovered in New Brunswick (star). stream gradient becomes lower, the surrounding area is swampy, the stream is choked with alders and inhabited almost exclusively by brook trout. Coho salmon occurred through the middle riffle- pool section and extended in diminishing numbers into the swamp area upstream. To estimate the numbers of young coho in the creek, two equal-effort electrofishings were per- formed through the riffle-pool section and approx- imately 50 m into the swampy section. The lower fast section was fished separately during the first trip (28 October), but since it contained no coho salmon it was omitted on the second 20 days later. Although some coho salmon might have moved downstream in the period between the two fishings, coho salmon were scarce in most up- stream areas on both occasions, suggesting there were few above the point where fishing ceased. Twelve coho salmon were caught on the first trip and five on the second. The total population esti- mated by the depletion method ( Seber and Le Cren 1967) was 21. Three coho salmon had been cap- 488 tured during the collection trip on 9 October, so that the total estimated population of coho salmon in the stream was 24. During electrofishing, particular note was taken of the kind of habitat in which coho, Atlantic salmon, and brook trout were captured. Coho salm- on were found where there was immediate or nearby overhead cover in the form of overhanging banks, tree roots, or fallen trees or brush, and where the water current was slow ( <30 cm/s). This kind of habitat was also frequently occupied by brook trout. On at least one occasion, a brook trout and coho salmon were captured together. Atlantic salmon were scarce above the lowermost steep sec- tion of the stream. However, in October and November three or four Atlantic salmon were cap- tured in slow water where they had never been seen in summer (Symons and Harding 1974). These observations suggested that summer habitat requirements of coho salmon were more similar to those of brook trout than of Atlantic salmon, although the latter may utilize brook trout-coho salmon habitat in winter. All captured coho salmon were retained and taken to the laboratory for measuring and weigh- ing. The average fork length of all coho salmon captured was 89 mm, ranging from 75 to 100 mm. There was no statistical difference between aver- age lengths in October (89 mm) and in November (91 mm). Examination of scales revealed that these coho salmon were underyearlings. They were considerably larger than underyearling At- lantic salmon (60-70 mm fork length) and under- yearling brook trout (40-60 mm) captured at the same time. The coho salmon were retained for use in laboratory experiments through the winter, and the 1 0- 1 5 that survived were returned to Frost Fish Creek the following April. To investigate whether coho salmon might be present elsewhere, spot checks were made in 17 nearby locations (Figure 2) between 28 October and 17 November. Spot checks consisted of 10-35 min of electrofishing with most effort being ex- pended in parts of streams having habitat similar to that in which coho salmon were caught in Frost Fish Creek. No coho salmon were found at any of these sites. Brook trout were caught in all streams, and Atlantic salmon were caught in streams where they were known to occur. Brown trout, Salnjo trutta, were caught in Frost Fish Creek (2 individuals), Burns Brook ( 1 ), and Sorrel Ridge Brook (1), all tributaries to the Digdeguash River. Brown trout were introduced to the Dig- deguash as early as 1921 (MacCrimmon and Mar- shall 1968), and they continue to exist there in small numbers. In sum, an estimated population of 24 under- yearling coho salmon was found in Frost Fish Creek in fall 1976. No coho were discovered in neighboring streams during a cursory search. Al- though adult coho salmon are known to spawn in small, gravelly coastal streams (Scott and Crossman 1973), spawning may not have occurred in the creek. Atlantic salmon apparently do not spawn there despite the presence of young which are thought to arrive from the main Digdeguash River, having descended the falls into the estuary and then reentering the nearest available fresh- water. The young coho salmon may have arrived by the same route. Regardless of the exact location in which coho salmon spawned, should they estab- lish a run in the river system, it would probably be revealed by continued sampling of fish in the creek. Acknowledgments Constructive criticisms of an earlier draft of the manuscript were made by J. W. Saunders and R. L. Saunders, whom we thank. Literature Cited MacCrimmon, H. R., and T. C. Marshall. 1968. World distribution of brown trout, Sa/mo^rw«a. J. Fish. Res. Board Can. 25:2527-2548. Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Fish. Res. Board Can., Bull. 184, 966 p. SEBER, G. A. F., AND E. D. LE CREN. 1967. Estimating population parameters from catches large relative to the population. J. Anim. Ecol. 36:631- 643. Symons, p. E. K., and G. D. Harding. 1974. Biomass changes of stream fishes after forest spray- ing with the insecticide fenitrothion. Fish. Res. Board Can., Tech. Rep. 432, 47 p. + append. PHILIP E. K. Symons Department of Fisheries and Environment Fisheries and Marine Service Biological Station. St. Andrews, N.B. Present address: Pacific Biological Station Nanaimo, B.C. V9R 5K6 James D. Martin Department of Fisheries and Environment Fisheries and Marine Service Biological Station. St. Andrews, N.B. EOG 2X0 489 SUBSAMPLER FOR ESTIMATING THE NUMBER AND LENGTH FREQUENCY OF SMALL, PRESERVED NEKTONIC ORGANISMS' When many samples, containing large numbers of organisms, must be processed it is often necessary to take subsamples and assume that they are rep- resentative of the total sample. Frequently sub- samples are taken in some arbitrary fashion which is described in such terms as "100 fish were randomly selected." However, it is doubtful whether any selection can be adequately random. Therefore, numerous devices have been designed in attempts to secure more representative sub- samples and to increase the speed and efficiency of subsampling. Most subsamplers have been designed for use with plankton, small benthos, and invertebrate drift samples and are generally unsuitable for larger organisms. However, Lewis and Garriott ( 1971) modified a Folsom plankton splitter for use on meter net samples containing larval fish up to 19 mm long, and Hightower et al. ( 1976) described a subsampler specifically designed for use with nektonic organisms. In the present paper I describe the design, oper- ation, and efficiency of a subsampler originally built for research on estuarine nekton (Herke 1971). The subsampler proved to be useful for es- timating the number and length frequencies of small nektonic organisms such as the bay an- chovy, Anchoa mitchilli, tidewater silverside, Menidia heryllina, and brown shrimp, Penaeus az- tecus, as well as young of larger species such as gulf menhaden, jBreyoor^/apafronus, and Atlantic croaker, Micropogon undulatus. Although differ- ent from most subsamplers, the design is fairly similar to that described by Hightower et al. (1976); it bears some similarities to those de- scribed by Hewitt and Burrows ( 1948) for subsam- pling live hatchery fish, by Gushing (1961) for plankton, and by Sodergren (1974) and Hickley (1975) for benthos. My sampler differs from that of Hightower et al. (1976) in at least four respects: 1) it has fewer moving parts; 2) fewer water jets are required to achieve through mixing of the sample; 3) a cen- tral pillar or cylinder prevents organisms from clumping in the center; and 4) the total sample is 'Contribution no. 24 of the Louisiana Cooperative Fishery Research Unit: Louisiana State University, Louisiana Wildhfe and Fisheries Commission, and U.S. Fish and Wildhfe Service cooperating. subdivided by raising vanes through the mixed sample, rather than allowing the sample to settle into baskets. Also, spin-dry weighing is required, but it takes <1 min to complete the subsampling process (after the organisms are placed in the sub- sampler), rather than several minutes as required for the subsampler described by Hightower et al. I have made no comparative tests between the two subsampler designs, however; individual cir- cumstances may determine which would be most practical in any given situation. Subsampler Construction My subsampler can be constructed of various materials, and the same general design can be used for large and small models. A small Plexi- glas^ version ( Figure 1 ) has an outside diameter of 305 mm, and Herke (1971) also illustrated one with a 580-mm outside diameter that utilized part of a 208-1 steel drum for the outer cylinder, a 19-1 bucket for the inner cylinder, and plywood for the false floor. The subsampler in Figure 1 was constructed primarily of Plexiglas about 6 mm thick. Plexiglas joints were bonded with solvent (methylene chloride and trichlorethylene). The major parts and their functions are as follows (numbers refer to the parts labeled in Figure 1): 1. Base. 2. Brass hinge for attaching base to edge of table top. 3. Outer cylinder bonded to base; in addition to solvent, a suitable cement may be required to ensure a watertight seal to the base. 4. Gentral pillar of Plexiglas tube bonded to base at exact center of circle formed by the outer cylinder (3). 5. Rubber stopper in (4) to prevent material from falling inside the piller. 6. Inner cylinder, which slides smoothly up and down over (4). 7. Locking pin for holding (6) in the raised posi- tion. Rubber bands around (6) and over a peg through the shaft of (7) hold the pin in place (these are omitted from the diagram to avoid cluttering). 8. Vane bonded to (6); the outer edge almost touches the outer cylinder. In the raised posi- ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 490 Figure l. — Basic design of the nektomc subsampler; see text for explanation. tion shown, the three vanes subdivide the sample into portions approximating 0.2, 0.2, and 0.6 of the total. 9. False floor consisting of three sections bonded to the inside of the outer cylinder (3) at a height so that the upper surface is exactly even with the upper edges of the vanes when (6) is lowered to the base. The vanes move up and down through the slits left between the sections of the false floor. Enough space must be left between the inner edges of the false floor and the inner cylinder (6) and its at- tached vanes to allow the inner cylinder and vanes to move freely up and down. Con- versely, the clearance must be small enough to prevent organisms from falling into the space below the false floor. Omitted from the diagram are braces extending from the base to near the inner edges of the two smaller sec- tions of false floor. 10. Hinged door for removing subsampled or- ganisms. 11. Latch holding the other door closed. 12. One side of a spout into which the water and organisms pour when doors are opened; the organisms are collected in a sieve below the spout. 13. Rubber tubes to carry water; the middle one enters the cylinder (3) beneath the false floor (9). 491 14. Copper tube with outlets for each rubber tube ( 13). Water to operate the subsampler comes through a large-diameter garden hose and pistol-type hose nozzle (not shown) attached to this tube. 15. Overflow tube attached to the outside of the cylinder (3). The cut edges of a longitudinal section of Plexiglas tubing are bonded to the cylinder from the overflow intake to the bot- tom of the base (1). Below the base the tube is not sectioned (i.e., left intact) so a drain hose can be attached to it. 16. Aluminum window screen covering overflow intake; bottom of intake opening is level with the top of the vanes (8) when they are raised. 17. Rubber stopper in drain hole below spout. Also omitted from the diagram are "stops" on the bottom edges of the vanes (which prevent the vanes from pulling through the false floor) and spongy, foam gaskets attached to the doors with rubber cement. Subsampling Procedure In subsampling, one pushes the inner cylinder (6) with the attached vanes down until it rests on the base; in this position the tops of the vanes are even with the top of the false floor so that the vanes and floor form a single flat surface. The entire sample is then placed on the false floor. The hose nozzle trigger is squeezed fully open, squirting water rapidly through the rubber tubing. (Nor- mally, the space below the false floor is still filled with water from previous use.) Some of the water rises through the three vane slits in the false floor, thereby inhibiting downward passage of the smaller specimens; most of the water squirts out of the upper tubes, causing the water above the false floor to swirl rapidly. Turbulence thoroughly mixes the sample as both sample and water re- volve. When the water almost reaches the bottom of the overflow intake, the inner cylinder (6) and attached vanes are quickly raised as far as possi- ble so that the locking pin (7) slides farther through its hole in (6) and over the top edge of (4); simultaneously, the hose nozzle trigger is re- leased. The sample has now been divided into three parts equal to about 0.2, 0.2, and 0.6 of the whole. The entire subsampler is next tilted on its hinges (2) in preparation for emptying. If a 0.2 subsample is desired, only one door is opened and 492 the contents of that compartment flow through the spout (12) into a sieve. (To avoid bias, the user should always open the same door first. Occasion- ally fish balance on top of the vanes; the user can avoid personal bias by always pushing the fish so it falls headfirst.) Opening both doors produces a 0.4 subsample and the remainder of the material in the subsampler constitutes a 0.6 subsample. The 0.6 subsample is removed by first taking out the 0.4 subsample and then lowering the vanes as far as they will go. The 0.6 subsample may then be washed into a sieve below the spout. ( When remov- ing any subsample, it is easier to wash the or- ganisms out of the subsampler than to pick or push them out.) A wide variety of subsample ratios can be obtained by sequentially subsampling subsam- ples (e.g., 0.8 x 0.2 x 0.2 = 0.032). Small organisms do occasionally fall through the vane slits into the space between the base and the false floor. Such losses are normally insig- nificant compared with the total number being subsampled, but they are noticeable through the Plexiglas. These organisms may be recovered by washing them out through the drain hole plugged by the rubber stopper (17). No special leveling of the subsampler is re- quired for proper operation; it may be mounted on any reasonably level surface such as a table top or laboratory bench. Discussion The subsampler is useful for estimating both total numbers in a sample and the total length- frequency distribution. If the total sample is not first separated by species, one should at least make a thorough scan of the sample, before subsam- pling, to remove any unusually large or odd spec- imens. As stated by Hightower et al. ( 1976), these can later be added to the total estimate, which is derived by extrapolating the subsample results. However, subsampling can give erratic results for inconspicuous species present in small numbers. Therefore, I think it usually is best to first sepa- rate the total sample into individual species, and subsample only the abundant ones. For each of these species, a subsample is first taken, and its weight and that of the remainder are obtained by the spin-dry method described by Herke (1973). (In contrast to plankton, preserved fishes and many crustaceans can be easily and precisely weighed without damage by using the spin-dry method.) All organisms in the subsample are then counted and the number in the total sample is estimated on the basis of the weights of the sub- sample and total sample. Since the estimate is based on weight rather than volume, the three vanes need not divide the subsampler into exactly 0.2, 0.2, and 0.6 segments. If a length-frequency estimate is desired, the subsample can be further subsampled. Since the number in the first subsample is now knovra, any desired number for the length-frequency subsam- ple can be closely approximated by selecting the proper sequence of subsamples. For instance, sup- pose the first subsample contains 3,371 anchovies and a length frequency is desired from approxi- mately 100 fish; 3,371 x 0.2 x 0.4 x 0.4 = 108. Therefore, subsamples taken in this sequence should produce the desired number for measuring. The consistency with which the desired number is obtained may be judged (Table 1) by comparing the "theoretical" and "actual" numbers obtained in 20 successive trials. The two subsamplers used in these trials had a tendency to slightly exceed the desired number; one or both of the smaller compartments in each subsampler probably con- tained a bit more than 0.2 of the whole. However, the increased subsample size actually improves the probability of obtaining an accurate length- frequency estimate. Also, with use, one soon learns whether the tendency is to obtain more or fewer than the theoretical number and can select the subsampling sequence accordingly. How well the length-frequency estimates de- rived from subsampling groups of anchovies and menhaden represented the true length frequen- cies of the groups was examined by using the Kolmogorov-Smirnov one-sample, two-tailed test, which is a test of goodness of fit. The test involves comparing the observed cumulative frequency distribution from a subsample with the cumula- tive frequency distribution of the total sample. It is sensitive to any kind of difference between the two distributions — differences in location (central tendency), in dispersion, in skewness, etc. Accord- ing to Siegel ( 1956) the Kolmogorov-Smirnov test is definitely more powerful than the chi-square test when samples are small, and may be more powerful in all cases. The cumulative length-frequency distribution for only one subsample was significantly different (oc = 0.05) from its corresponding total sample (Table 1 ). In the other 19 tests, the probability was greater than 0.20 that a divergence of the observed magnitude would occur if the observations were really a random subsample from the total sample (0.20 is the highest probability listed in Siegel's table). Table L— Results of 20 tests to determine the correspondence between: 1) the theoretical and actual number of bay anchovies or gulf menhaden in the subsample, and 2) the cumulative length frequency distribution of fish in the subsample and in the corresponding total sample. Subsamples were returned to the total sample after each trial. The cumulative distribution shown in italics (in the same row with the number in the total sample) was the true distribution obtained by measuring every fish in the sample. Number in total sample Subsample sequence Final subsample no. Standard length in millimeters' Theoretical Actual 15 20 25 30 35 40 45 50 55 60 3,371 0.539 .481 0 772 .797 0B2^ 835 0.867 880 0914 .947 0.957 .970 0.984 .977 0.995 992 7 000 anchovies (0.2) (0.4) (0.4) 108 133 1,000 124 .524 .758 838 863 .911 960 .976 1 000 134 2.418 .739 .791 858 932 962 1.000 141 .489 .709 .773 822 894 .950 .979 993 1.000 146 .479 .740 781 .856 925 938 986 1.000 1.505 . 0.367 .347 553 .551 .643 .571 .774 673 %46 734 908 .795 964 .917 .997 .958 998 999 .999 anchovies (0.2) (0.2) 60 49 59 .322 .576 .661 729 814 848 .933 967 1.001 77 .338 520 559 .676 793 832 949 988 1 001 70 .357 528 628 728 799 885 .956 985 999 71 .389 .570 .598 667 .778 .875 .958 1 000 Same (0.4) (0.2) 120 128 .328 586 672 766 836 .914 992 1.000 1,505 125 .272 .544 600 .776 864 .920 .976 1 000 anchovies 133 .353 .556 .654 789 842 .880 .940 985 .993 1 001 134 .306 .507 589 .768 .843 .903 970 1.000 152 .283 .526 .598 710 .780 881 934 ,987 1 000 1.221 .020 000 273 .278 756 .656 .980 1.000 .998 7.000 menhaden (0,4) (0,2) 98 90 116 .026 .198 .733 1.000 115 .017 252 .765 991 1.000 128 .031 .242 664 969 984 1.000 113 .027 .345 .796 1.000 'Measured in 5-mm increments: i.e.. 15 = 15.0-19.9, 20 = 20.0-24.9, etc. ^The probability of a divergence this large in a random subsample from the total sample was between 0.05 and 0.01 The probability for the 1 9 other subsamples was >0.20 493 Number in subsample : 133 124 134 141 146 3,371 (total) FIGURE 2.— Length-frequency distri- bution of a total sample of 3,371 bay anchovies, and of each of five subsam- ples taken from the total. (From Herke 1971.) If) o (M ID CM in o in O in o in ro ^ 'J in in ID s Stondard length m millimeters It is difficult to visualize, from inspection of the cumulative length-frequency distributions, how well the percentage of fish in each subsample length group represents the percentage in the cor- responding length group in the total sample. Therefore, this comparison is shown graphically (Figure 2) for the first five subsamples listed in Table 1. Literature Cited Gushing, C. E., Jr. 1961. A plankton sub-sampler. Limnol. Oceanogr. 6:489-490. Herke, w. H. 1971. Use of natural, and semi-impounded, Louisiana tidal marshes as nurseries for fishes and crustaceans. Ph.D. Thesis, Louisiana State Univ., Baton Rouge, 264 p. University Microfilms, Ann Arbor, Mich. (Diss. Abstr. 32:2654-B.) 1973. Spin-drying of preserved fishes and macroinverte- brates. Trans. Am. Fish. Soc. 102:643-645. HEWITT, G. S., AND R. E. Burrows. 1948. Improved method for enumerating hatchery fish populations. Prog. Fish-Cult. 10:23-27. HICKLEY, P. 1975. An apparatus for subdividing benthos samples. Oikos 26:92-96. HIGHTOWER, G. M., K. T. KIMBALL, AND C. A. BEDINGER, JR. 1976. A water-powered mechanical device for accurately subsampling large numbers of nektonic organisms. Trans. Am. Fish. Soc. 105:509-513. LEWIS, S. A., AND D. D. GARRIOTT. 1971. A modified Folsom plankton splitter for analysis of meter net samples. Proc. Annu. Conf. Southeast. Assoc. Game Fish Comm. 24:332-337. SIEGEL, S. 1956. Nonparametric statistics for the behavioral sci- ences. McGraw-Hill, N.Y., 312 p. SODERGREN, S. 1974. A simple subsampler for stream-bottom-fauna sam- ples. Arch. Hydrobiol. 73:549-551. WILLIAM H. HERKE Louisiana Cooperative Fishery Research Unit Louisiana State University Baton Rouge, LA 70803 494 ERRATUM Fishery Bulletin, Vol. 76, No. 1 Berrien, Peter L., "Eggs and larvae o( Scomber scombrus and Scomber japonicus in continental shelf waters between Massachusetts and Florida," p. 95-115. 1) Page 99, left column, line 9, correct line to read: or seven pterygiophores in the 2d through 5th (deleting the zero) INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the following instructions. These are not absolute requirements, of course, but desiderata. CONTENT OF MANUSCRIPT The title page should give only the title of the paper, the author's name, his affiliation, and mailing address, including Zip code. The abstract should not exceed one double- spaced page. In the text, Fishery Bulletin style, for the most part, follows that of the U.S. Government Printing Office Style Manual. Fish names follow the style of the American Fisheries Society Special Publi- cation No. 6, A List of Common and Scientific Names of Fishes from the United States and Canada, Third Edition, 1970. 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 photog- raphy to 5% inches (for single-column figures, allowing for 50% reduction in printing), or to 12 inches (for double-column figures). The maximum height, for either width, is 14 inches. Photo- graphs should be printed on glossy paper. Do not send original drawings to the Scientific Editor; if they, rather than the photographic re- ductions, are needed by the printer, the Scientific Publications Office will request them. Each table should start on a separate page. Consistency in headings and format is desirable. Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be numbered sequentially in arabic numerals. To avoid confusion with powers, they should be placed to the left of numerals. Acknowledgments, if included, are placed at the end of the text. Literature is cited in the text as: Lynn and Reid (1968) or (Lynn and Reid 1968). All papers re- ferred 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 literature cited is the re- sponsibility of the author. Abbreviations of names of periodicals and serials should conform to Bio- logical Abstracts List of Serials with Title Abbrevi- ations. (Chemical Abstracts also uses this system, which was developed 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 mea- sure 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 car- bon copies. The sequence of the material should be: TITLE PAGE ABSTRACT TEXT LITERATURE CITED APPENDIX TEXT FOOTNOTES TABLES (Each table should be numbered with an arabic numeral and heading provided) LIST OF FIGURES (Entire figure legends) FIGURES (Each figure should be numbered with an arabic numeral; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. Jay C. Quast, Scientific Editor Fishery Bulletin Northwest and Alaska Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA P.O. Box 155, Auke Bay, AK 99821 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 HAYNES, EVAN. Description of larvae of a hippolytid shrimp, Lebheus groenlan- dicus, reared in situ in Kachemak Bay, Alaska 457 AUSTIN, HERBERT M., and CLARENCE R. HICKEY, JR. Predicting abundance of striped bass, Morone saxatilis, in New York waters from modal lengths 467 Notes KATONA, STEVEN K., SALVATORE A. TESTAVERDE, and BRADLEY BARR. Observations on a white-sided dolphin, Lagenorhynchus acutus, probably killed in gill nets in the Gulf of Maine 475 MOFFATT, NANCY M., and DONALD A. THOMSON. Reciprocal hybridization between the California and Gulf of California grunions, Leuresthes tenuis and Leiiresthes sardina ( Atherinidael 476 GRAHAM, JOSEPH J., and EDWIN P. CREASER, JR. Tychoplanktonic blood- worm, Glycera dibranchiato, in Sullivan Harbor, Maine 480 HOUDE, EDWARD D., and RICHARD C. SCHEKTER. Simulated food patches and survival of larval bay anchovy, Anchoa mitchilli, and sea bream, Archosargus rhomboidalis 483 SYMONS, PHILIP E. K., and JAMES D. MARTIN. Discovery of juvenile Pacific salmon (coho) in a small coastal stream of New Brunswick 487 HERKE, WILLIAM H. Subsampler for estimating the number and length frequency of mall preserved nektonic organisms 490 •i: GPO 796-049 ,* .<< °>^. •at ^o Fishery Bulletin National Oceanic and Ati noi ^^ATES O^ ^ ^M^&^dr^lAMiS^^ iMMlMke Fisheries Service LIBRARY NOV 0 0 1978 WHHfJSf HRI§: M;!c:c; Vol. 76, No. 3 July 1978 CLARKE, THOMAS A. Diel feeding patterns of 16 species of mesopelagic fishes from Hawaiian waters 495 FLETCHER, R. IAN. On the restructuring of the Pella-Tomlinson system 515 RIVARD, D., and L. J. BLEDSOE. Parameter estimation for the Pella-Tomlinson stock production model under nonequilibrium conditions 523 LEIS, JEFFREY M. Systematics and zoogeography of the porcupinefishes iDi- odon, Diodontidae, Tetraodontiformes), with comments on egg and larval develop- ment 535 HEARD, WILLIAM R. Probable case of streambed overseeding — 1967 pink salmon, Oncorhynchus gorbuscha. spawners and survival of their progeny in Sashin Creek, southeastern Alaska 569 YOUNG, RICHARD EDWARD. Vertical distribution and photosensitive vesicles of pelagic cephalopods from Hawaiian waters 583 VENRICK, E. L. Systematic sampling in a planktonic ecosystem 617 PEARCY, WILLIAM G. Distribution and abundance of small flatfishes and other demersal fishes in a region of diverse sediments and bathymetry off Oregon . . . 629 PEARCY, WILLIAM G., and DANIL HANCOCK. Feeding habits of Dover sole. Microstomas pacificus; rex sole, Glyptocephalus zachirus; slender sole, Lyopsetta exilis; and Pacific sanddah, Citharichthys sordidus, in a region of diverse sediments and bathymetry off Oregon 641 BARKLEY, RICHARD A., WILLIAM H. NEILL, and REGINALD M. GOODING. Skipjack tuna, Katsuwoniis pelamis, habitat based on temperature and oxygen requirements 653 QUINN, WILLIAM H., DAVID O. ZOPF, KENT S. SHORT, and RICHART T. W. KUO YANG. Historical trends and statistics of the Southern Oscillation, El Nino, and Indonesian droughts 663 FIEDLER, PAUL C. The precision of simulated transect surveys of northern an- chovy, Engraulis mordax, school groups 679 Notes PEREZ FARFANTE, ISABEL. Intersex anomalies in shrimp of the genus Penaeop- sis (Crustacea: Penaeidae) 687 BONE, QUENTIN, JOE KICENIUK, and DAVID R. JONES. On the role of the different fibre types in fish myotomes at intermediate swimming speeds 691 (Continued on back cover) Seattle, Washington 0 U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Richard A. Frank, Administrator Terry L. Leitzell, Assistant Administrator for Fistieries NATIONAL MARINE FISHERIES SERVICE Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. n EDITOR Dr. Jay C. Quast Scientific Editor, Fishery Bulletin Northwest and Alaska Fisheries Center Auke Bay Laboratory National Marine Fisheries Service, NOAA P.O. Box 155, Auke Bay, AK 99821 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. Bruce B. Collette National Marine Fisheries Service Dr. Edward D. Houde University of Miami Dr. Merton C. Ingham National Marine Fisheries Service Dr. Reuben Lasker National Marine Fisheries Service Dr. Sally L. Richardson Oregon State University Kiyoshi G. Fukano, Managing Editor The Fishery Bulletin is published quarterly by Scientific Publications Otflce. National Marine Fisheries Service. NOAA, Room 450, 1107 NE 45th Street, Seattle. WA 98105. Controlled circulation postage paid at Tacoma, Wash. Although the contents have not been copyrighted and may be reprinted freely, reference to source is appreciated. The Secretary of Commerce has determined that the publication of this periodical Is necessary In the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget through 31 December 1978. I Fishery Bulletin CONTENTS Vol. 76, No. 3 July 1978 CLARKE. THOMAS A. Diel feeding patterns of 16 species of mesopelagic fishes from Hawaiian waters 495 FLETCHER, R. IAN. On the restructuring of the Pella-Tomlinson system 515 RIVARD. D., and L. J. BLEDSOE. Parameter estimation for the Pella-Tomlinson stock production model under nonequilibrium conditions 523 LEIS, JEFFREY M. Systematics and zoogeography of the porcupinefishes (Di- odon, Diodontidae. Tetraodontiformes), with comments on egg and larval develop- ment 535 HEARD, WILLIAM R. Probable case of streambed overseeding — 1967 pink salmon, Oncorhynchus gorbuscha, spawners and survival of their progeny in Sashin Creek, southeastern Alaska 569 YOUNG, RICHARD EDWARD. Vertical distribution and photosensitive vesicles of pelagic cephalopods from Hawaiian waters 583 VENRICK, E. L. Systematic sampling in a planktonic ecosystem 617 PEARCY, WILLIAM G. Distribution and abundance of small flatfishes and other demersal fishes in a region of diverse sediments and bathymetry off Oregon . . . 629 PEARCY, WILLIAM G., and DANIL HANCOCK. Feeding habits of Dover sole, Microstomus pacificus: rex sole, Glyptocephalus zachirus; slender sole, Lyopsetta exilis; and Pacific sanddab, Citharichthys sordidus. in a region of diverse sediments and bathymetry off Oregon 641 BARKLEY, RICHARD A.. WILLIAM H. NEILL, and REGINALD M. GOODING. Skipjack tuna, Katsuwonus pelaf7}is. habitat based on temperature and oxygen requirements 653 QUINN, WILLIAM H., DAVID O. ZOPF, KENT S. SHORT, and RICHART T. W. KUO YANG. Historical trends and statistics of the Southern Oscillation, El Nino, and Indonesian droughts 663 FIEDLER, PAUL C. The precision of simulated transect surveys of northern an- chovy, Engraulis mordax, school groups 679 Notes PEREZ FARFANTE, ISABEL. Intersex anomalies in shrimp of the genus Penaeop- sis (Crustacea: Penaeidae) 687 BONE, QUENTIN, JOE KICENIUK, and DAVID R. JONES. On the role of the different fibre types in fish myotomes at intermediate swimming speeds 691 (Continued on next page) Seattle, Washington For sale by the Superintendent of Documents, U.S. Government Prmting Office. Washington, DC 20402 — Subscription price per year; $12.00 domestic and $15.00 foreign. Cost per single issue: $3.00 domestic and $.3.75 foreign. Contents-continued JEWETT, STEPHEN C. Summer food of the Pacific cod, Gadus macrocephalus, near Kodiak Island, Alaska 700 LEMING. THOMAS D., and HILLMAN J. HOLLEY. A computer software system for optimizing survey cruise tracks 706 Notices NOAA Technical Reports NMFS published during the first 6 mo of 1978 715 Data on fisheries subjects accessioned through NMFS by NODC during the first 6 mo of 1978 715 Vol. 76, No. 2 was published on 29 June 1978. 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. DIEL FEEDING PATTERNS OF 16 SPECIES OF MESOPELAGIC FISHES FROM HAWAIIAN WATERS Thomas A. Clarke^ ABSTRACT Diel patterns of stomach fullness, as percent of dry weight, were determined for 16 species of mesopelagic fishes. Nine species of myctophids and one melamphaid, all vertical migrators, appeared to feed solely or principally at night in the upper layers. These species encountered higher tempera- tures and prey concentrations at night. Four species of stomiatoid fishes appeared to feed during the day regardless of the extent of their migration or the absence thereof Prey concentrations encountered by the stomiatoids during the daytime appeared to be higher than or similar to those encountered at night. One myctophid and one gonostomatid showed no diel pattern; diel changes in the environmental factors considered were relatively small in spite of the fact that both species undertook limited vertical migrations. Crude estimates of instantaneous evacuation rate and daily ration were made from data for four species. These indicated that evacuation rate was increased at night in the upper layers and that daily rations of species which migrated into the upper layers were similar to values for shallow-living zooplanktonivores, while rations of deeper living species were lower. Thus while the adaptive value of upward migration in the species which feed at night is obviously related to feeding activity, the upward ascent by the daytime feeders may allow processing of larger daily rations than if they remained at low temperature all day. The extensive diel vertical migrations of certain mesopelagic fishes have been well documented in a vanety of oceanographic situations. While a number of theories have been proposed to explain the adaptive value of the behavior — in both fishes and migrating invertebrates as well — data to support any of them are few. One of the most frequently proposed hypotheses (e.g., Marshall 1960) is that the organisms ascend at night to feed in the upper layers where food is presumably at higher concentrations and descend during the day to avoid predation while the upper layers are well lighted. Several studies of mesopelagic fishes (to be cited below) have considered the relationship between feeding chronology and vertical distribu- tion in an effort to support at least one-half of the hypothesis, but the results have for the most part been rather equivocal. Apparent diel trends in stomach fullness or details thereof are often ques- tionable owing to low numbers of specimens examined, insensitive methodology, or incomplete diel coverage. Furthermore, all such studies on mesopelagic fishes, with the exception of Merrett and Roe (1974), have been conducted in high 'University of Hawaii, Department of Oceanography and Hawaii Institute of Marine Biology, P.O. Box 1346, Kaneohe, HI 96744. Manuscript accepted March 1978. FISHERY BULLETIN: VOL. 76, NO. 3, 1978. latitude or neritic situations and have dealt with only one or, at most, three species. This study considered the feeding chronology of 16 species from 5 families of mesopelagic fishes from the north central Pacific Ocean. Vertical dis- tribution and certain other apsects of the ecology of these fishes are covered in Clarke (1973, 1974) and Clarke and Wagner (1976); results from re- lated investigations in the same study area are summarized in Maynard et al. ( 1975). Comparison of diel patterns of stomach fullness and diel changes in temperature and prey concentration allow consideration of adaptive value of the verti- cal migrations undertaken by most of these species. In four species, rough calculations of daily ration are possible using equations similar to those presented by Eggers (1977). MATERIALS AND METHODS Field Sampling Specimens for this study were all collected with a 3-m Isaacs-Kidd midv/ater trawl ca. 20 km west of the island of Oahu, Hawaii (ca. lat. 2r20-30'N; long. 158"20-30'W) in waters 2,000-3,000 m deep. In order to reduce the concentration of zoo- plankton in the cod end of the net and thus 495 FISHERY BULLETIN: VOL. 76, NO. 3 minimize bias due to fishes' feeding after capture, the net terminated in a 1-m diameter cone of ca. 3-mm (Vs-in) knitted nylon mesh instead of the "normal" plankton netting. Specimens were taken in oblique tows which sampled vertically migrating species at nine dif- ferent periods of the 24-h cycle. At night, cable was paid out in increments over a period of 1.5 h such that the trawl fished roughly equal amounts of time at all depths between the surface and ca. 350 m. The trawl was retrieved immediately after- wards for a total towing time of about 2 h. Four such tows were made between last light at dusk and first light at dawn. During the day, 1,200 m of wire were paid out initially. This placed the trawl at ca. 350-400 m. Subsequently, cable was paid out in increments such that the trawl fished between this depth and ca. 1,100-1,200 m over a period of 1.5 h and then retrieved for a total fishing time of ca. 2.5 h below 350-400 m. Three such tows were made during the day. At dusk, 1,500 m of cable were paid out initially, placing the trawl at ca. 500 m. Cable was then retrieved in increments such that the trawl fished between 500 m and the sur- face over 1 .5 h. The trawl reached maximum depth just before sunset and was on deck shortly after last light. At dawn, the process was reversed, and the trawl shot before first light, and fished from the surface to ca. 500 m over 1.5 h such that it reached maximum depth ca. 1 h after sunrise. Ship speed was ca. 1 m/s (2 kn) for cable retrieval and ca. 2 m/s (4 kn) for all other phases. In order to collect sufficient numbers of speci- mens for as many species as possible, three 24-h series of nine tows each (dusk, four at night, dawn, and three during the day) were made 27-30 Au- gust 1973. These dates were chosen to bracket new moon (August 28) and minimize avoidance of the trawl at night (Clarke 1973). One day tow of this series was fouled and could not be repeated until 13 September 1973. The total range of time fished by equivalent tows of each series (Table 1) overlapped — considerably so for the night series due to one night's fishing proceeding ahead of schedule. The overlap was effectively less than shown in Table 1 since most of the fishes analyzed were probably taken below 50 m (based on previ- ously cited studies of vertical distribution) and not during the first 15 min or the last 5 min of each tow when the trawl was shallower than 50 m. Con- sequently, equivalent tows from each 24-h series were considered replicates and specimens were combined for data from each period. The nine 496 sampling periods will subsequently be designated as follows: SS for sunset and SR for sunrise; Nl, N2, N3, N4 for the four night periods in chronolog- ical sequence; and Dl, D2, D3 for the three day- time periods in sequence. Danaphos oculatus, a nonmigrating species, was not taken in the shallow night tows described above. Nighttime data for this species were based on specimens from three night series of three tows each taken 30 August-1 September and 13-14 Sep- tember 1973 (Table 1) using the same towing schedule described above for daytime (ca. 400- 1,000 m). Thus only eight periods of the diel cycle were considered. The three nighttime periods for D. oculatus were designated dNl, dN2, and dN3. In order to obtain more specimens of three species of stomiatoids, I utilized specimens taken 24-25 May 1974 in seven tows at the same location with the same net, and with the same procedure and timings as Nl-4 and Dl-3. The numbers of specimens used from this series will be noted in the results. All specimens of the other species came from 1973 collections. The catch was immediately preserved in 4-5% formaldehyde in seawater. The specimens re- mained in this solution for up to 2 yr before pro- cessing, but since all specimens of a given species were processed within a period of 2-3 wk, any between-sample differences in weight loss due to leaching can be considered negligible. Laboratory Analyses The ratio of the dry weight of the stomach con- tents to that of the fish as percent was used as an index of stomach fullness. Where sufficient speci- mens of a given species were available, 20 from each of the nine sampling periods were examined. If possible, the least damaged specimens (or perhaps more appropriately — equivalently dam- aged specimens) were selected from a narrow size range. For many species, however, it was neces- sary to use specimens damaged to various degrees and of all sizes between recently (but fully) metamorphosed juveniles and mature adults. In cases where a specimen was damaged beyond loss of scales or fin rays, i.e., where tissue was missing, I used the median dry weight of other specimens of the same standard length. Each fish was briefly rinsed with tapwater and gently blotted; standard length was measured to the nearest millimeter. The stomach (anterior end of the esophagus to the pyloric valve) was removed CLARKE: DIEL FEEDING PATTERNS OF MESOPELAGIC FISHES Table L— Towing times (Hawaiian Standard Time) for three 24-h series of oblique tows to sample vertically migrating mesopelagic fishes and three all night series for deep-living nonmigratory fishes. Times given for dusk (SS), dawn (SR), and shallow night (Nl-4) are for the entire tow; those for day tows (Dl-3) and deep night tows (dNl-3) are for the time the trawl fished below ca. 350-400 m. Period 27-28 Aug. 28-29 Aug. 29-30 Aug. Midpoint Period 30-31 Aug. 31 Aug.-l Sept. 13-14 Sept. Midpoint SS 1754-1956 1815-1955 1820-2023 1910 dNI 2040-2300 2005-2233 2000-2235 2130 N1 2045-2240 2001-2155 2040-2233 2120 dN2 2352-0215 2325-0154 2330-0155 0050 N2 2315-0110 2207-0005 2255-0045 2340 dN3 0308-0540 0250-0515 0255-0525 0415 N3 0120-0320 0015-0210 0113-0310 0150 N4 0330-0520 0220-0420 0318-0515 0350 SR 0535-0742 0515-0725 0533-0745 0630 D1 0822-1047 0807-1034 0820-1045 0930 D2 1143-1425 1125-1350 1135-1410 1300 03 1510-1740' 1448-1710 1515-1740 1615 'The D3 tow for 28 August was fouled; time given is for tow made on 13 September. and its contents, if any, placed on a clean glass slide. The fish including the empty stomach was placed in a preweighed aluminum pan. After examination, the stomach contents were rinsed into a second preweighed pan using distilled wa- ter. The stomach contents were examined only casually. A rough estimate of fullness was made and degree of digestion noted. Prosome length (PL) of copepods and total length (TL) of other prey were recorded from intact items. Intact prey items could usually be identified to genus, but no serious attempt was made to determine composition of the diet from these samples. The remarks below on types of prey include only the most frequently encountered items and are not meant to be taken as detailed analyses of diets. Both fish and stomach contents were dried at 60°C for 24 h ( somewhat longer for a few large fish) and allowed to cool under partial vacuum before weighing. The pans with stomach contents were weighed to the nearest 0.01 mg on a microbalance, and the content weight determined by subtrac- tion. Both control pans and reweighing of several pans with dried stomach contents after a second period in the drying oven or desiccator indicated that the weighing and handling error was of the order of ±0.02 mg. There was no indication that error was proportional to the amount of material in the pan. Pans with fish were weighed on a semimicro balance; the reading was recorded to 0.01 mg on small fish and to 0.1 mg on those over ca. 100 mg. Based on changes in weight of control pans and reweighing offish after a second period of desiccation, the error was <1% of the fish weight. While the weighing and handling error was such that estimates of stomach fullness were af- fected only to the fourth or possibly third decimal place, other errors or biases inherent in the mate- rial should be mentioned. As noted above, an un- known fraction of the material was lost due to leaching. Damage to the fish positively biased the ratios since there was some loss of skin, scales, or fin rays in almost all specimens. Such errors were unrelated to the time of collection and were more likely to increase variability and thus to obscure rather than cause diel trends in the data. The intestinal contents, which were dried and weighed with the fish, may have varied with time and thus introduced a systematic error in fish weights. Based on visual examination, however, largest amounts of materials in the intestine were almost certainly <1% of the total fish weight, and, con- sequently, affected the stomach fullness index by <0.1%. The 2-3 h durations of the tows were a possible source of bias and high variability. Bias in stomach fullness could result from evacuation of stomach contents between capture and death (Eggers 1977). It is likely that this was negligible since the fishes considered here were probably dead soon after capture by the net. The 2-3 h possi- ble differences in capture time for fishes from the "same" period of the diel cycle almost certainly contributed to the variability in stomach fullness — particularly during periods when the latter was changing rapidly. Stomach fullness could possibly be biased nega- tively by regurgitation after capture or positively by feeding in the net. (Either type of bias would tend to obscure rather than cause diel differences in stomach fullness.) Regurgitation apparently occurred infrequently in all species considered ex- cept Lampanyctus nobilis. Except for the latter (see below), specimens with partially digested food remains in the mouth or everted stomachs were not used. Hopkins and Baird (1975) showed that feeding in the net is an unimportant source of error even when a fine-mesh cod end is used, and there was little indication of net feeding in the present study. Zooplankton in good condition, usually crustaceans with appendages erect and extended, were infrequently found in the mouth. These were assumed to have lodged there during 497 FISHERY BULLETIN: VOL. 76, NO. 3 capture and were not counted, but the fish and any other contents were used. Items part way down the esophagus with appendages flattened against the body or the body folded were assumed to have been eaten before capture and were included. I consi- dered such "esophagus" items unlikely to have been eaten after capture because concurrent analyses of diet (Clarke in prep.) on the same species collected by the same net indicate that there is no difference in species composition be- tween such items and items clearly in the stomach and partially digested. Stomach fullness values for a single species and single time period were rarely distributed nor- mally. Usually the values were skewed to the left, but variably so — the mean being sometimes close to the median and sometimes close to the 75th percentile. Consequently, the entire set of stomach fullness values for each species were ranked and tested for between-period differences by the Kruskal-Wallis nonparametric equivalent of analysis of variance (//-test). The test is mainly sensitive to differences in position (Tate and Clel- land 1957), and significance implies differences among the medians for the separate time periods but does not single out which sets of data are different. Each adjacent (in time) pair of data sets was tested for differences in the median with the Mann-Whitney or Rank sum test (Tate and Clel- land 1957); however, because of multiple testing on the same data, the significance levels from this cannot be taken rigorously. Neither test used is sensitive to differences in variability, and no separate testing was done. Some idea of differences in frequency distribution can be gleaned from relative position of the mean and median. Other gross differences, e.g., bi- modality vs. unimodality, will be pointed out in the results. Likewise, I did not test for possible correlations between sex or size of the fish and stomach fullness. The data from each period were, however, ranked and compared (by inspection) with sex and rank in length; no obvious correla- tions were found. RESULTS A total of 15 vertically migrating species (10 myctophids, 4 stomiatoids, and 1 melamphaid) and 1 nonmigrating stomiatoid were investigated. These included species for which 20 individuals were collected at most of the nine periods sampled plus a few, less frequently taken species selected to 498 give broader coverage with respect to systematic position or vertical distil jution pattern. In addi- tion to graphical presentations (cited specifically below), ancillary data for all species are sum- marized in Table 2. In the subsequent presenta- tion, stomachs were considered "empty" if stomach fullness was <0.1'7r. This included both visually empty stomachs and those with only a trace of digested remains in the pyloric end of the stomach. Types of prey organisms, state of diges- tion, and other aspects not obvious from the figures or Table 2 are considered in individual species accounts below. Comments on vertical distribution of prey items are based on preliminary analyses of opening- closing plankton tows taken in the study area and their general agreement with data in the litera- ture for the same or closely related species in other central water mass localities. The plankton tows — 16 taken in September 1973 and 20 in November 1974 — covered the depth ranges of the fishes considered both day and night. Euphausiids from all samples have been counted and identified, and copepods either counted (shallow night sam- ples) or sufficiently examined to at least roughly determine the depth ranges of the important prey species. The apparent depth ranges agree gener- ally with those given by Brinton (1967) and Roe (1972). These two important types of prey can, with a high degree of certainty, be classified as shallow nonmigrators (above 200-300 m both day and night), vertical migrators (above 200-300 m at night and below this depth by day), and deep living (below 300 m day and night). Similar statements cannot be made for ostracods, the other important crustacean group, nor for other taxa of zooplank- ton. Myctophidae Bentbosema suhorhitale (Figure 1) The //-test indicated highly significant (P <0.005) differences in stomach fullness over the diel cycle. The data from SS and Nl were charac- terized by low averages, narrow percentile limits, and high proportions of empty stomachs. Sub- sequently stomach fullness generally increased until SR and decreased throughout the day. The most frequent prey items were copepods of the genera Pleuromamma, Candacia, and Paracan- dacia. Euphausia spp. and occasionally small de- capods contributed significantly to the weight of CLARKE: DIEL FEEDING PATTERNS OF MESOPELAGIC FISHES Table 2. — Summary of data for each of the 16 species of fishes examined from each of the periods of the diel cycle. In each species/time block, the first line gives the number of specimens examined and the number with stomach fullness <0.1% of fish dry weight in parentheses; the second line, the size range of the specimens in millimeters standard length; and the third, the range of stomach fullness values in percentage offish dry weight. All values of stomach fullness are rounded to the nearest 0.1%. Species SB N1 N2 N3 N4 SR D1 D2 D3 Benthosema suborbitale 20(12) 24-30 0-09 20( 8) 15-32 0-2.8 20( 3) 18-31 0-60 20( 1) 19-29 0.1-2.7 20( 5) 22-31 0 1-28 19( 2) 17-27 0-2.6 20( 2) 16-30 0.1-3.6 20( 3) 18-32 0-3.8 20( 2) 17-30 0-2.2 Bolinichthys longipes 20( 7) 24-46 0-1.4 20( 2) 21-46 0.1-2.0 20( 2) 20-51 0.1-2.5 20( 0) 17-49 02-15 20( 1) 18-43 0.1-2.4 20( 0) 21-32 0.3-4.6 20( 0) 16-50 0.2-48 20( 0) 17-46 0 1-15 20( 0) 19-48 0 1-11 Ceratoscopelus warmingi 20( 9) 22-53 0-28 20( 1) 18-45 0 1-98 20( 2) 24-45 0-81 20( 2) 23-50 0-50 20( 0) 25-47 0.1-7.6 5( 0) 18-23 0.9-4.9 20( 1) 19-48 0.1-2.4 20( 0) 19-59 0.1-4.5 20( 2) 18-52 0.1-7.4 Diaphus schmidti 20( 2) 25-39 0 1-23 20( 2) 20-41 0.1-1.7 20( 0) 19-41 0.3-1,4 20( 0) 19-40 02-60 20( 0) 17-38 03-38 20( 0) 20-38 0.1-22 20( 0) 14-37 0.2-2.3 20( 0) 15-38 0.2-2.7 20( 2) 19-38 0-2.3 Hygophum proximum 20(14) 26-33 0-1.0 20( 0) 19-32 0.2-67 20( 2) 18-38 0-5.8 20( 1) 18-42 0 1-3.6 20( 0) 18-37 0.2-7.0 20(14) 19-43 0-0 5 20(16) 19-42 0-0.4 20(13) 19-40 0-13 18(11) 18-46 0-1.3 Lampanyctus niger 20(14) 65-84 0-1.9 20(11) 65-85 0-1.0 20( 7) 63-83 0-1.7 20( 8) 64-85 0-13 20( 4) 51-85 0-1.5 0 20(10) 52-87 0-4.0 20(11) 68-85 0-3.7 20(10) 64-85 0-3.9 Lampanyctus nobilis 9( 4) 24-84 0-62 20( 3) 29-81 0-3.3 20( 2) 27-88 0-7.5 20( 4) 24-80 0-8.5 20( 1) 26-98 0-5.5 0 14( 2) 25-94 0-3.6 14( 4) 25-90 0-103 18( 6) 25-94 0-20 Lampanyctus steinbecki 20( 2) 25-39 0-27 20( 4) 22-48 0-3,7 20( 3) 22-41 0-4.4 20( 1) 21-39 0-35 20( 0) 22-42 0.2-4.0 20( 2) 18-36 0 1-9.8 20( 0) 22-42 0.1-4.5 20( 4) 23-44 0-2.1 20( 2) 27-42 0-5.5 Notolychnus valdiviae 20( 0) 19-24 0.2-3.4 20( 1) 20-23 01-2.9 20( 3) 19-23 0-3.0 20( 0) 19-23 0.3-4.7 20( 0) 18-23 0.3-3.6 20( 0) 19-24 0.3-35 20( 2) 20-23 0-1.7 20( 1) 17-22 0-25 20( 0) 19-24 0.1-3.5 Triphoturus nigrescens 20( 3) 18-34 0-5.9 20( 6) 17-31 0-36 16( 3) 15-34 0-9.0 15( 3) 15-35 0-6.6 19( 0) 16-33 1.0-14.3 20( 0) 15-36 07-9.9 20( 4) 16-34 0-16,0 20( 3) 18-33 0-6,7 20( 4) 18-34 0-4.5 Melamphaes danae 3( 0) 16-20 04-0.6 12( 3) 16-21 0-32 14( 2) 16-21 0-30 12( 0) 16-21 0.4-3.5 9( 0) 16-22 05-4.0 0 3( 0) 19-21 0.3-1 5 2( 0) 16 1.1-1.5 14( 1) 16-22 0-1.6 Gonostoma atlanticum 4( 0) 48-58 11-3.3 20( 2) 36-62 0-6.3 20( 1) 32-65 0-45 20( 2) 31-64 0.1-4.3 17( 4) 26-62 0-3.5 7( 2) 48-65 0.1-1.6 9( 0) 22-64 0.3-2.4 20( 1) 50-68 0.1-4.6 8( 0) 48-57 0.4-11.2 Gonostoma elongatum 20( 2) 26-112 0.1-2.6 20( 2) 31-126 0-13.1 20( 1) 30-135 01-59 20( 9) 30-132 0-5.0 20( 3) 32-79 0-2.0 8( 0) 30-150 02-80 20( 4) 24-125 0-166 14( 3) 30-149 0-5.2 10( 1) 29-87 0.1-7.3 Vinciguerria nimbaria 20( 0) 21-34 0.7-8.9 20( 0) 24-35 0.2-8.9 20( 3) 24-36 0-7.9 20( 3) 25-32 0-3.1 20(13) 23-34 0-0.6 20(15) 24-33 0-1.8 20( 9) 20-35 0-12.7 20( 1) 20-33 0.1-6.3 20( 2) 20-30 01-14.2 Danaphos oculatus 15( 0) 27-40 0.5-25 9( 0) 11( 0) 11( 0) 27-39 29-40 29-36 0.6-2.3 0.3-10 0.3-18 15( 2) 28-42 0-0.8 20( 2) 31-40 0-1.0 4( 0) 28-38 0.6-1.7 10( 0) 27-39 0.2-2.3 Valenciennellus tripunctulatus 6( 0) 25-30 1.2-4.5 20( 0) 22-32 1.2-3.6 12( 0) 23-32 0.9-2.2 9( 2) 21-32 0-2.2 10( 1) 21-32 0.1-1.0 3( 2) 21-31 0-0.2 7( 2) 25-30 0-0.6 5( 0) 21-30 0.9-2.5 10( 0) 22-33 1.0-3.5 food. The food from specimens taken by day sam- ples was generally well digested; except for the thoracic spots or "buttons" from Pleuromamma, prey was rarely recognizable beyond general category. Bolinichthys longipes (Figure 1) Analyses of B. longipes were complicated by the frequent presence in the stomachs of digenetic trematodes. These were 1-10 mm long (most were 1-5 mm) and occurred in 419f of the stomachs examined. They were mingled with the food and appeared to have been fixed while wrapping around or holding to items. As a probable conse- quence, whole prey were rarely found inB. longipes' stomachs. The parasites were, however, easily separated from the food; they were not included with either the fish or stomach content weight. The number of trematodes was roughly a func- tion of size of the fish. Fish < ca. 30 mm SL usually had 0-2 individuals while several > 40 mm con- tained 10-20. Since there was little between- period difference in size composition of the fish examined, there was no apparent correlation of trematode number with time of day. Also there 499 FISHERY BULLETIN: VOL. 76. NO. 3 i !.U 3. suborbitole - o * r -1 r ' ' ^ r " - -| r 1 .0 : ^ : T o — ( • ( I 1 1 i t _ X ( 1 '- J L T -1 L — X 1 J L -J L l/l 0 i * '- UO 1 1 r 2. Or 1.0- B. longipes < > i it o O -1 r * 1 1 1 1 1 1 I V 1 1' i 1 1 1 1 T 1 1 1 1 ■ 1 ; 1 ( 1 " ' 1 i I D3 SS N) N2 N3 N4 SR Dl D2 D3 2 0 f— "" C. warmingi »« 3.05 20r 1.0 ~1 1 — \ 1 1 1 r~ D3 SS Nl N2 N3 N4 SR Dl D2 D3 I I \ 1 1 1 1 1600 2000 0000 0400 0800 1200 1600 TIME OF DAY D3 SS Nl N2 N3N4 SR Dl D. schmidti D2 D3 h * 0 ' 1 o 1 o 1 . , 1 1 1 i. 1 . 1 ( 1 < 1 f — I 1 1 — I — I — I 1 1 1 1 — D3 SS Nl N2 N3 N4 SR Dl D2 D3 I 1 1 1 1 1 1 1600 2000 0000 0400 0800 1200 1600 TIME OF DAY Figure l. — Medians (dots), means ( x's), and ranges between 25th and 75th percentiles (solid vertical lines) of stomach fullness as percentage of body weight throughout the diel cycle for four species of myctophids: Benthosema suborbitale, Bolinichthys longipes, Ceratoscopelus warmingi, and Diaphus schmidti. Values are positioned at the midpoint of each sampling period (Hawaiian Standard Time). Dashed vertical lines indicate significant differences between adjacent pairs (circle — 0.052'^ ). Because of the latter, ranges and percentile limits were broad, and means were much higher than me- dians. Ceratoscopelus warmingi fed on a wider variety of taxa and sizes of prey than did the other species covered here. The most frequent items were copepods, ostracods, and small euphausiids, but heteropods, siphonophores, and other zooplankton also occurred. Intact items of such relatively small prey were recorded mostly from specimens col- lected at night; remains from day-collected speci- mens were usually well digested. Ceratoscopelus warmingi also took items up to 10'7( of bodily weight; squid, other fishes, and large euphausiids or decapods occurred in specimens >35-40 mm. Such single large items accounted for nearly all the fish with high values of stomach fullness, and intact prey of this size occurred at all times of the day. Most such items were vertically migrating species that could have been taken at night, but remains of nonmigrating Cyclothone spp., which could have only been encountered between dawn and dusk, were found in 1 1 specimens. Thus, while the overall trend of the data indicates that C. warmingi feeds principally on small zooplankton in the upper layers at night, it probably takes large prey whenever encountered. Diaphus schmidti (Figure 1) Diel differences in stomach fullness for D. schmidti were highly significant (P<0.005), and the trend was similar to that of the preceding myctophids except for timing; the maximum value occurred at Dl instead of SR. Empty stomachs occurred only in a few specimens from D3, SS, and Nl. Diaphus schmidti took a large variety of prey items; the dominant taxa were small crustaceans (ca. 0.5-3.0 mm PL or TL): ostracods, copepods, and larval and juvenile malacostracans. Heteropods, pteropods, polychaetes, and chaetognaths were also noted. Excepting chaetognaths, few items were >4-5 mm. Frequency of intact items was highest at SR, and lowest at D3 and SS. Hygophum proximum (Figure 2) Diel differences in stomach fullness for//, prox- imum were highly significant (P<0.005), and the trend quite different from those of the other species examined here. Most stomachs were empty, and even 75th percentile values were zero or nearly so between SR and SS; the peak value occurred at N2. Hygophum proximum fed princi- pally on medium-sized copepods (1-3 mm PL) and occasionally other crustaceans. Less than 109f of the stomachs were empty for any of the night periods, but intact items were found frequently only in stomachs from Nl. By N2 most of the prey were unrecognizable, and only six items were rec- ognizable to even general category in all the other samples. Larnpanyctus niger (Figure 2) This species, one of three forms of the L. niger- complex which occur near Hawaii, has minute pectoral fins and lower AO counts than the others; it was designated as "Form B" in Clarke (1973). Zahuranec^ has recently identified the form as L. niger (sensu stricto). There was evidence from deep night tows taken during the same sampling period that a fraction of the population ofL. niger did not vertically migrate; consequently, some of the day-caught specimens may not have ascended to the upper layers the previous night. (Such "non-migration" was also recorded in previous studies, see Clarke 1973.) The //-test indicated no significant diel differ- ences in stomach fullness (P>0.10), and none of the adjacent pairs differed significantly. The me- dians from nighttime show a trend similar to that of other myctophids, but the means were highest during the day. No specimens were available from SR. Values of stomach fullness were overall much lower than observed in other species. Stomach fullness exceeded 19c in only 21 of the 160 speci- mens, and over 509c of the stomachs were empty at all periods except N2, N3, and N4. The most frequent food items were large copepods of the familes Metridiidae, Euchaetidae, and Aetideidae and small (<10-15 mm TL) euphausiids. Occasionally small fishes were found. Intact prey items were found in stomachs from all periods. Deep-living copepods such as Metridia and Pseudochirella were noted in day- caught specimens indicating that at least some feeding occurs during the day. ^B. J. Zahuranec. Oceanic Biology Program, Office of Naval Research, Arlington, VA 22217. Personal communications, June 1977. 501 FISHERY BULLETIN: VOL. 76. NO. 3 A Or- !i proximum 3 0 2.0 1.0 :: II 1 1 1 r 1.0 1 1 o 1 •— « 1 111 1 -h-^ — f-+ D3 SS N1 N2 N3 N4 SR Dl D2 D3 3.U L. nobilis -1 r ■ 2 0 * T r o 1.0 0 1 i ~ X 1 " I I' J 1 1 1 1 1 1 1 1 1 1 1 1 1/1 Z D3 SS Nl N2 N3N4 SR Dl D2 D3 ^30r N.valdiviae U- 2.0 iin -1 1 1 1 — I — I 1 1 I r~ D3 SS Nl N2 N3 N4 SR Dl D2 D3 D3 SS Nl N2 N3 N4 SR Dl D2 03 I I I \ \ 1 1 1600 2000 0000 0400 0800 1200 1600 TIME OF DAY l.U L. niger J "■ : ~ X X 0.5 - r 1 r . y- ., :: X ■■ " T T T "■ 1 1 I 1 1 ^ 0 ~ 1 . < 1 -L 1 1 > 1 . D3 SS Nl N2 N3 N4 SR Dl D2 D3 3. Op L. steinbecki 2.0 10 60 mm. The small specimens had eaten mostly copepods and amphipods 1-3 mm and Euphausia spp. <10 mm, while the large ones had taken large copepods ( >3 mm PL) and euphausiids, mysids, sergestiids, and fishes 10-30 mm long. Intact prey were found frequently in night specimens and oc- casionally in those caught by day. The latter were, with the exception of a single Lophothrix humili- frons (apparently a deep-living copepod), migrat- ing species that could have been taken at night. One specimen from N4 contained; among the re- mains of a euphausiid, crab megalopa, and copepods; a partially digested insect (probably a hymenopteran). Lampanyctus steinbecki (Figure 2) Stomach fullness values for L. steinbecki dif- fered significantly (P<0.005) over the diel cycle. The medians generally increased from SS to SR and thereafter stayed at about 1% until a sharp decrease between D3 and Nl. The percentage of fish with empty stomachs was low for all periods. The principal prey ofL. steinbecki were copepods > ca. 2 mm PL — mostly aetideids, Pleuromamma, and Candacia — and euphausiids. A few intact items were found in specimens from Dl and D2 but all were shallow-living or migrating species that could have been taken the previous night. With the exception of a s\r\g\ePareuchaeta sp. (probably a deep-living nonmigrator), the prey from D3 and SS were all well digested. Notolychnus valdiviae (Figure 2) The //-test indicated highly significant (P<0.005) diel differences in stomach fullness for N . valdiviae. Median values were low early in the night and increased to a peak at N4. The minimum value at Dl was slightly below the early night values. Stomach fullness increased slightly until SS and then decreased at SS-N 1 . The percentage of fish with empty stomachs was low at all periods. The positions of the 75th percentiles indicated higher percentages of fish with relatively full stomachs at N3, N4, SS, and D3. Notolychnus valdiviae had taken a wide variety of sizes (ca. 0.5-4.0 mm PL) and species of copepods, but the bulk of the food in terms of weight was made up by large (relative to the weight of N. valdiviae) items such as Pleuromamma xiphias, Candacia longimana, and 2-4 mm aetideids. Intact prey were more fre- quently noted in specimens from N3 and N4 than in those from the apparent "secondary peak" in stomach fullness at D3 and SS. Considering only those specimens with stomach fullness >2% (whose numbers distinguish the peak periods from others), only three of the nine from D3 and SS contained intact or partially intact Pleuromam- ma. The other six contained remains that were either unrecognizable or barely so. In contrast, of the 15 specimens from N3 and N4, 12 contained 1-3 intact items, while only 3 contained unrecog- nizable remains. This plus the absence of any ap- parent significant differences associated with the D3/SS peak indicate that the latter was due to a chance collection of a few more specimens that had taken large meals the previous night rather than to extensive daytime feeding. Triphoturus nigrescens (Figure 2) Overall diel differences in stomach fullness were highly significant (P<0.005). Both medians and means rose from low values at SS and Nl to a peak at N4 and then, except for a slight increase at D3, declined until SS. Due to the broad overlap in 503 ranges and percentiles for most pairs, only the large increase between Nl and N2 was even mar- ginally significant. The percentage of empty stomachs was highest at Nl and zero at and just after the peak at N4. Triphoturus nigrescens fed principally on Pleuromamma and Euphausia spp. Intact prey were recorded more frequently during N2-N4 than in other periods. As in the case of N. valdiviae, the apparent peak at D3 was due to a few fishes' con- taining large amounts of well-digested material rather than freshly taken items. Melamphaidae Melamphaes danae (Figure 2) Few M. danae were taken at any period. None were taken at SR, and only two or three were taken at SS, Dl, and D2. The data indicate a diel trend similar to that of several myctophids, but the//-test indicated that diel differences were only marginally significant (P = ca. 0.10). If the data from SS, Dl, and D2 were not included, the //-test indicated significance atP = ca. 0.05 and the N4 and D3 values differed atP<0.05. This latter, and statistically dubious, manipulation indicates that the apparent trend in the data is real, but that more specimens would be needed to confirm it properly. Melamphaes danae fed on a wide variety of zoo- plankton including polychaetes and chaetognaths as well as crustaceans — mostly small copepods and ostracods. The copepods identified were all either vertical migrators or shallow-living, non- migrating species. Intact items were present in nighttime specimens; those from daytime con- tained remains barely identifiable to general tax- on. Gonostomatidae Gonostoma atlanticum (Figure 3) Relatively few G. atlanticum were available from four periods even though 23 additional specimens from the May 1974 collections were in- cluded. Still there were significant (P<0.05) dif- ferences in stomach fullness over the diel cycle. Median values rose steadily from SR to D3, re- mained at ca. 2% between D3 and N2, and then dropped sharply between N2 and N3. Though the median for N4 was slightly higher than that for 504 FISHERY BULLETIN; VOL. 76, NO. 3 either N3 or SR, the percentage of empty stomachs was highest at N4 and SR, indicating an overall trend for decrease during the late night. Gonostoma atlanticum fed on large copepods — mostly Pleuromamma xiphias, Can- dacia longimana, and aetideids and scolecithricids of several genera — and small (<10-15 mm) euphausiids. Intact prey were found in stomachs from all periods, but were mostly from the period between Dl and N2. The majority of the contents from N3 and N4 were well digested. Gonostoma elongatum (Figure 3) Relatively few G. elongatum were available from three periods and the size range of individu- als used was extremely broad (26-150 mm). There is evidence from past studies that fractions of the population occasionally do not migrate (Clarke 1974), but catches from deep night tows taken during the same sampling period did not clearly indicate whether or not this occurred during this study. The //-test indicated marginally significant dif- ferences (0.05« 1.0- ihi 1 { D3 SS Nl N2 N3 N4 SR Dl D2 D3 V. tripunctulotus n ~~l \ 1 1 1 T T D3 SS Nl N2 N3 N4 SR Dl D2 D3 D3 SS dNl dN2 dN3 SR Dl D2 D3 I 1 1 1 1 I I 1600 2000 0000 0400 0800 1200 1600 TIME OF DAY 70r 60- 5.0- 4.0- 1.0- - G. e ongatom ~ -1 * 1 -I r - - 1 r 1 ■» : 1 -r 1 ■\ " 1 r T • 1 J L , 1 :, " '• I ' -'- 11 1 1 r r r 3.0- T 2.0- i.\J D. oculatus o T : 1 -1 1 ., 1 -^ 1' , ■ O o 1 1.0 1— 1 1 J 1 ^ -1 [ '■! ' 1 . 1 L J 0 III 1 i 1 ( 1 1 1 1 1 D3 SS Nl N2 N3N4 SR Dl D2 D3 6.U V. nimbaria 5.0 ♦ o 4.0 - 1 1 ■ 1 ♦ 3.0 - - ' 0 o * ( ' ' 1 1 2 0 - ' 1 1 ' 1 1 1.0 0 1 1 1 1 ( 1 1 1 1 1 1 . !' 1 II II I 1 D3 SS Nl N2 N3 N4 SR Dl D2 D3 I 1 1 1 1 I I 1600 2000 0000 0400 0800 1200 1600 TIME OF DAY Figure 3. — Stomach fullness throughout the diel cycle for five species of stomiatoids: Gonostoma atlanticum, Gonostoma elongatum, Valenciennellus tripunctulatus , Danaphos oculatus, and Vinciguerria nimbaria. Symbols and format as in Figure 1. rather than the medians, because there were marked differences in frequency distribution for these periods — differences to which the median is not sensitive. At D2 the data were skewed to the left with most values <1% and very few full stomachs. At D3 the data were bimodal; 7 values were higher than the mean of ca. 2.757f and 11 <1%. By SS, the data were again unimodal and skewed slightly to the right with 16 values >2% and only 2 <1%. Thus the trend between D2 and SS was one of a gradual change in percentages of the fish with very full stomachs, and the abrupt increase in median values between D3 and SS occurred as the high values became the dominant mode. The percentages of empty stomachs showed a trend opposite to that of the average values, i.e., an increase during the night and a decrease be- tween SR and D2. 505 Vinciguerria nimbaria fed upon a wide variety of sizes and taxa of prey. Small ( < ca. 2 mm PL) copepods and ostracods were most frequent, but larger copepods and small euphausiids occurred regularly. Both the number of prey items and ab- solute values of stomach fullness for the peak period were higher than for most of the the other species examined here; in several instances the remains of 20-40 prey items were found in a single stomach. Intact items were most frequent at SS and common in specimens from day samples. Some intact items were noted from Nl and a few from N2, but stomachs from N3, N4, and SR contained practically nothing but well-digested remains. Sternoptychidae Datictphos oculatus (Figure 3) Few D. oculatus were available for any period except Dl, and numbers were particularly low for D2. Nine of the specimens used came from the May 1974 series. In spite of this, there was an evident and highly significant (P<0.005) diel trend in stomach fullness. Median values rose steadily from a minimum at SR to a maximum at SS and declined nearly constantly throughout the night. There were a few empty stomachs at SR and Dl and none at other periods. Danaphos oculatus fed almost exclusively on Pleuromamma xiphias, Euchaeta media, and similar-sized juveniles and adults of several aetideid species. Intact items were most frequently noted in D3 and SS speci- mens; some were found in those from Dl and D2. Almost none of the night specimens contained any but well-digested remains. Valenciennellus tripumtulatus (Figure 3) Few V. tripunctulatus were available from any period except Nl; 31 of the total examined came from the May 1974 collections. Still, like D. oculatus, V. tripunctulatus showed a clear and highly significant iP <0.005) diel trend in stomach fullness. Medians rose from zero at SR to a maximum at SS and declined throughout the night. The principal prey items were P. xiphias, P. abdominalis, E. media, and similar-sized aetideids. The stomachs from D2 to SS were nearly uniformly packed with intact prey while those from late night and SR were either empty or con- tained only traces of well-digested remains. FISHERY BULLETIN: VOL. 76, NO. 3 DISCUSSION Feeding Chronology Interpretation of data on stomach fullness is limited because observed fullness is a function of two rate processes — feeding rate and stomach evacuation rate. Diel changes in stomach fullness indicate that one or both rates vary over the diel cycle, but without independent estimates of one or the other, the only certain statements that can be made are that feeding exceeds evacuation during periods when fullness increases, the opposite when fullness decreases, and that both rates are zero when the stomach is empty. Notes on state of digestion of stomach contents are helpful, but must be interpreted with caution. Absence of in- tact items indicates that feeding rate is zero, but presence of intact items does not necessarily mean feeding rate was positive during a given period since some items may remain intact for an un- known time after feeding ceases. Still, it is possi- ble within these limits to qualitatively consider changes in the two rates and to relate them to environmental changes which the fishes en- counter over the diel cycle. The species considered here undergo diel changes in numerous environmental factors, some of which are likely to affect either feeding or stomach evacuation rate in a qualitatively pre- dictable manner. The migrating species encounter higher temperatures at night. Diel temperature changes for each species (Table 3) were deter- mined using temperature-depth profiles from the study area (Maynard et al. 1975 give profiles from several seasons of three different years) and depth ranges of the fishes (Clarke 1973, 1974; Clarke and Wagner 1976). Because all species considered occur below the steepest part of the thermocline during the day, the magnitude of the diel tempera- ture change is mostly a function of nighttime depth range and not day depth or absolute range of migration. For the same reason, juveniles, which occur shallower than adults in most species (Clarke 1973), incur greater temperature change than adults of the same species. The migrating species also encounter lower pressures and higher oxygen concentrations at night (oxygen-depth profiles for a site near the study area are given in Gordon 1970). Unless the fishes are able to regulate metab- olism over the range of diel changes, the day-night 506 CLARKE: DIEL FEEDING PATTERNS OF MESOPELAGIC FISHES Table 3. — Depth ranges, estimated diel changes in tempera- ture, and probable day-night differences in prey concentration for the 16 species of fishes considered. (See text for sources of estimates.) Depth range (m) Temperature Night change ("C) Prey density Species Day Night-day Night vs. day Benthosema 0-100 18-19 N =D suborbitale 500-600 Bolinichthys 50-150 16-19 N >D longipes 500-700 Ceratoscopelus 0-150 16-20 N > -D warmingi 600-1,000 Diaphus 0-75 17-19 N>D schmidti 500-600 Hygophum 0-150 15-19 N>D proximum 500-700 Lampanyctus 75-200 10-17 N>>D steinbecki 600-1,000 Lampanyctus 50-150 16-20 N>>D nobilis 600-1,200 Notolychnus 80-150 15-16 N>D valdiviae 500-650 Triphoturus 25-100 17-19 N >D nigrescens 550-750 Melamphaes 75-200 10-18 N>>D danae 750-1,200 Danaphos 450-650 0 N<D niger 650-900 differences in temperature and oxygen concentra- tion both predict lower rates of metabolic proces- ses in general and in particular lower feeding or stomach evacuation rate during the day. Childress (1975) and Childress and Nygaard (19731 indi- cated that mesopelagic organisms can regulate over a wider range of oxygen partial pressures than these fishes encounter off Hawaii. Thus temperature changes are more likely to affect rate processes. Teal ( 1971 1 showed that increased pres- sure can stimulate metabolic rates and thus mediate or cancel effects of temperature; however, it seems likely that temperature effects are pre- dominant for the species considered here since these fishes migrate through a much stronger thermocline than did the shrimps studied by Teal. As a consequence of vertical migration — by the fishes and by many of their prey — the fishes en- counter diel differences in prey concentration, with which feeding rate is likely to be positively correlated. As noted above, the depth distributions of all prey species in the study area are not known in detail; however, general, qualitative features were evident from the available plankton samples (see above). Most of the important prey species were either shallow-living nonmigrators that oc- curred above ca. 200 m day and night or were vertical migrators with maximal concentrations at ca. 300-450 m by day. Some important genera, e.g. Euphausia, Pleuromamma, and Euchaeta, oc- curred as deep as 600 m during the day but not at high densities. At night, most copepods and many of the euphausiids occurred at highest densities above ca. 150-200 m. Many prey species occurred between 200 and 300 m at night, but except for a few euphausiid species, concentrations were much lower than in the upper 200 m. Below ca. 600 m by day and below ca. 300 m at night, total zoo- plankton concentration was low and that of impor- tant prey species nearly zero. Based on the above features and the fishes' depth ranges, qualitative estimates of day-night differences in prey concen- tration were made for each species (Table 3). Nine species of myctophids and probably Melamphaes danae had similar diel patterns in that median values of stomach fullness were min- imal at or near dusk and increased only at night, but details of the patterns were variable. Six species, Benthosema suborbitale, Bolinichthys longipes, Ceratoscopelus warmingi, Diaphus schmidti, Lampanyctus steinbecki, and L. nobilis (Figures 1, 2), had two periods of increasing stomach fullness during the night separated by a decline. Maximum stomach fullness occurred at or near dawn, and the fish reached day depth with relatively full stomachs. Stomach fullness ap- peared to decrease during the day in some species and showed no clear trend in others, but in most there was a significant decrease at or near dusk. In Notolychnus valdiviae, Triphoturus nigrescens, and possibly Melamphaes danae (Figure 2), me- dian stomach fullness appeared to increase stead- ily throughout the night to a peak value just before dawn. In the first two of these species, stomachs were partially evacuated by the time they reached day depth. In Hygophum proximum (Figure 2) median fullness reached a peak value early in the night, and stomachs were completely evacuated by dawn. For most of the above species there was no evi- dence of significant feeding at depth during the day. Intact items were more frequent at night, and stomach contents of day-caught fish were usually 507 FISHERY BULLETIN: VOL 76, NO. 3 well digested. Lampanyctus nobilis and L. stein- becki occasionally take deep-living copepods dur- ing the day, and C. warmingi apparently takes large items whenever it encounters them. Still, the instances of definite day feeding were so few in even the latter three species that the medians and, therefore the diel patterns, were only marginally affected. All of these myctophids undergo diel changes in temperature and prey concentration (Table 3) that correlate with the observed pattern of feeding sole- ly or mostly at night while in the upper 200 m. All are at much higher temperatures at night. Al- though some species occur as shallow as ca. 500- 600 m during the day and thus partially overlap the daytime depth ranges of certain of their prey, all occur below daytime maxima of prey concen- trations and almost certainly encounter higher concentrations at night. Certain details of the pat- terns of stomach fullness indirectly indicate that stomach evacuation rate may be lower during the day as predicted by temperature differences. In many species, stomach fullness did not clearly de- crease during the day; since feeding rate was ap- parently zero then, the evacuation rates must have been low or zero. The sharpest declines in stomach fullness occurred at or near dusk in most species, near dawn in N. valdiviae and T. nigres- cens, and during the night in H. proximum — not during periods when the fishes remained within their day depth ranges. In all cases except H. prox- imum, however, something related to vertical migration itself, e.g., activity, could be responsible for the apparent increases in evacuation rates. Four species of stomiatoids, Gonostoma atlan- ticum, Danaphos oculatus, Valenciennellus tripunctulatus, and Vinciguerria nimharia, fed only during the day. The last three species occur somewhat shallower by day then do the myc- tophids and are consequently at or near depths of maximum concentration of their prey then. The upward migration of V. nimbaria is similar in extent to that of its prey. Thus this species en- counters little or no diel change. Danophos oculatus does not migrate, and Valenciennellus tripunctulatus migrates less than do its prey. Con- sequently, both species occur below high concen- trations of prey at night. The adults of G. atlan- ticum (as were most specimens used here) occur near the lower depth limits of most prey species both day and night, and the day-night difference is probably minor. Thus in these species, the day depth ranges, rather than the occurrence or up- 508 ward extent of migration, seem more related to observed feeding pattern. All four species feed at nearly the same, low temperature. Diel temperature change is zero for D. oculatus, and relatively small for V. tripunctulatus and large G. atlanticum because they penetrate only part way through the ther- mocline. Vinciguerria nimbaria undergoes a change similar to that of the myctophids. The temperature changes or lack thereof obviously have no effect on feeding periodicity; however, the steepness of the nighttime decline in stomach fullness seems roughly correlated with nighttime temperature indicating an effect on stomach evacuation rates. This trend is considered in more detail below. Lampanyctus niger and G. elongatum, the two species which showed no diel pattern in stomach fullness, do not undergo large diel changes in either temperature or prey concentration in spite of the fact that they migrate. The large individuals of both species I as were all theL. niger and mostG. elongatum) undergo a relatively small tempera- ture change. Likewise, only the smallest juveniles of either species encounter markedly higher prey concentrations at night. The relatively low values of stomach fullness in both species and the pres- ence of deep-living, nonmigrating zooplankton in L. niger indicate that these two species feed at a low rate whenever and wherever they encounter prey. Relationship to Previous Studies Comparison of the present results with those of previous studies is restricted because method- ology in all cases was different from that of the present study and in many cases equivocal or probably not sensitive enough to discern diel trends or lack thereof. With the exception of the study by DeWitt and Cailliet (1972), appropriate statistical testing was not done, and it is impossi- ble to do so from the published data. The most directly comparable study is that by Holton (1967) on Lampanyctus (= Triphoturus) mexicanus. Using 10 fish from each of eight periods of the day, he determined dry weights, but for some unknown reason weighed the entire alimentary canal with the food. The minimal val- ues observed, presumably from empty stomachs, indicate that his "% nutrition" values should be decreased by about 2.5-3 to make them roughly comparable to those of the present study. Though CLARKE: DIEL FEEDING PATTERNS OF MESOPELAGIC FISHES the ranges and standard deviations of the data are broad relative to the differences in means, the diel trend in the latter is similar to that observed here for H. proximum, i.e., peak value was reached early in the evening and then dropped to low val- ues and probably zero before the dawn descent. Most previous studies have used visual esti- mates of fullness with a scale of 3-5 ranks. Because of the lack of "intercalibration" between inves- tigators, only the rank for "empty" can be com- pared unequivocally, and it is not certain in what manner the ranks might correlate with percent- ages of the fishes' dry bodily weight. Finally the validity or absence of trends and details thereof are questionable because scales of only 0-3 or 0-4 are rather insensitive. (Had only visual estimates of fullness been used for the present study, only in a few cases, e.g.,//. proximum or Valenciennellus tripunctulatus, would the diel trends have been obvious.) Anderson's ( 1967) data on T. mexicanus indicate a peak in stomach fullness just before sunrise, but his data on degree of digestion indicate that fresh food items were most frequent between sunset and midnight. His data for Bathylagus stilbius (car- diac portion of the stomach only) indicate two separate periods of increasing fullness at night and the sharpest decrease prior to ascent at dusk. This pattern correlates with frequency of less- digested prey items and is very similar to that observed for several myctophids in this study. Similar indices were used in the studies of four species of high latitude myctophids: Benthosema glaciale (Gjosaeter 1973) and Stenobrachius leucopsarus, Diaphus theta, and Tarletonbeania crenularis (Tyler and Pearcy 1975). Both studies examined large numbers of specimens from each of a few, very broad time periods. Their data indi- cated highest percentages of full or nearly full stomachs at night and highest percentages of low values during the day. The occurrence of some full stomachs during the day led Gjosaeter to conclude that diel variation in feeding was not great and Tyler and Pearcy to conclude that there was no evidence against diurnal feeding. Both studies noted a higher degree of digestion during the day. These results are, however, consistent with the possibility that like many of the myctophids in the present study, their species descended at dawn with full stomachs and did not evacuate them completely until the dusk ascent. The latter may well have not been detected in these studies due to the broad time periods used. Data on myctophids from recent studies by Mer- rett and Roe (1974) and Baird et al. (1975) are consistent with nocturnal feeding but are equivocal to varying degrees due to low numbers of specimens, incomplete diel coverage, or methodology. Both studies based stomach fullness estimates on counts of identifiable prey items. Ap- parently, the presence of a single resistant part, e.g., a Pleuromamma button, was counted the same as an intact, whole individual of the same taxon. Because of this and the likelihood that some prey taxa or parts of prey are digested — and con- comitantly rendered unrecognizable — at different rates (e.g., Pandian 1967; Gannon 1976), such counts seem to be insensitive or possibly biased estimates of gut fullness — especially so when the counts are used to back-calculate dry weight as done by Baird et al. Furthermore, neither study corrected the fullness estimate for fish weight, which (using standard length ranges given by these authors and assuming that weight is roughly porportional to the cube of the length) varied by factors of ca. 7-15 in the myctophids covered by Merrett and Roe and ca. 2.75 in D. taaningi, the species studied by Baird et al. Merrett and Roe's data for L. cuprarious indi- cated peak fullness in the middle of the night and a decrease before the dawn descent — a pattern simi- lar to that of//, proximum. Their data (or Lobian- chia dolfleini and A^. valdiviae include no samples between dusk and near dawn, but show fuller stomachs at dawn. Data of Baird et al. for D. taaningi are also similar to that for H . proximuni . The rise in fullness from empty or nearly empty stomachs in the afternoon to fairly high values in early evening is evident and based on 39 and 9 specimens, respectively; however, the subsequent decline is based on a single specimen from late night and 4 from just after dawn ( 1 which con- tained a fair amount of food). Fewer stomiatoids have been examined elsewhere, but much of the data available is con- sistent with diurnal feeding. Perhaps the most convincing data (because of good diel coverage and numerous specimens) presented by Merrett and Roe (1974) is that for Valenciennellus tripunctulatus, which does not migrate in their study area. The pattern is clearly similar to that observed for the Hawaiian specimens. Hopkins and Baird ( 1977) cited their own unpublished data also indicating diurnal feeding for the same species. Merrett and Roe (1974) interpreted dusk peaks of numbers of items/nonempty stomach as 509 FISHERY BULLETIN: VOL. 76, NO. 3 an indication of dusk feeding activity in two species of A rgy rope lee us; however, the data for A. hemigymnus seem to me more consistent with in- creasing stomach fullness throughout the day and a nighttime decline. Except for high dawn values (based on only three specimens from two tows), A. aculeatus shows a similar trend. DeWitt and Cailliet ( 1972) found no diel trend in feeding ofCyclothone signata, but, based on fewer empty stomachs in fish caught in the upper part of the depth range, proposed that this species, al- though it does not undertake diel vertical migra- tions, may ascend irregularly to levels of higher prey concentration to feed. Their data also indi- cated that a deeper living species C. acclinidens, had a higher percentage of empty stomachs by day; as noted by the authors, the latter seems to defy any reasonable explanation. Legand et al. ( 1972) considered feeding chronol- ogy of 14 species of mesopelagic fishes from the South Pacific. Though trends in stomach fullness of some species are similar to those noted here, e.g., that for Triphoturus microchir (which almost certainly = T. nigrescens) is very similar to that for T. nigrescens near Hawaii, a number of species show patterns quite different from those reported by either the present or other studies. Interpreta- tion of the validity of such "exceptions" is difficult owing to the sparse presentation of Legand et al. Though total numbers of specimens are fairly high, it is not clear that they were equitably dis- tributed among diel periods, from the same area, or from the same season, etc. The percent fullness values are obviously based on wet weights — an imprecise measurement, particularly for stomach contents — and it is not clear whether all fish and stomach contents were weighed or some sort of averaging or regression procedure was employed. The feeding patterns shown by previous studies cannot be compared in detail with those presented here; however, there is general agreement in data on the two dominant groups of mesopelagic fishes. Myctophids feed mostly at night, while stom- iatoids tend to feed by day. My interpretations indicate that near Hawaii, the differences are at least partially related to different diel relation- ships of the fishes to vertical distributions of their prey. Other interpretations are obviously possible, e.g., the feeding patterns may prove to be charac- teristic of the two taxa regardless of relationship to prey distribution. It would be of particular in- terest to investigate myctophids with vertical dis- tribution patterns similar to those of the stomiatoids, i.e., with shallow day depth ranges at or near high daytime concentrations of zoo- plankton. (Certain Myctophym and Diciphus spp. from Hawaii meet this criterion [Clarke 1973], but were not captured in sufficient numbers to be in- cluded in this study.) The diel feeding patterns of mesopelagic fishes could well be related to light rather than (or in addition to) temperature and prey concentration. No data on diel light changes near Hawaii are available; however, data of Kampa ( 1970) from a similar area of clear oceanic water in the North Atlantic show that during full moon the diel change in depths of relevant isolumes is of the order of 300-350 m. Even allowing for considerable differences in extinction coefficients between Hawaii and Kampa's study area, the diel change in isolumes at new moon (when the present sam- ples were taken) off Hawaii is probably at least 300-350 m and could be as great as 500 m. The absolute diel change in depth for most of the myc- tophids is over 500 m while that for the 4 day- feeding stomiatoids is ca. 400 m or less (Table 3). Thus it is possible that feeding in both groups occurs when higher light levels are encoun- tered— at night for the myctophids and by day for the stomiatoids. Estimation of Rates As mentioned previously, neither feeding rate nor stomach evacuation rate can be considered quantitatively without an independent estimate of the other. Because of the difficulty in keeping mesopelagic fishes alive for grazing or evacuation experiments, it will likely be a long time before independent estimates are available. For a few species considered here it is, however, possible to derive "quasi-independent" estimates of evacua- tion rate given certain plausible assumptions. These allow, with further assumptions, rough es- timates of feeding rate and daily ration. For any period where feeding rate is zero, changes in stomach fullness are due to evacuation alone, and, if temperature, pressure, etc., remain essentially constant during that period, the rate of evacuation can be assumed to be proportional to the amount of food in the stomach (Kjelson and Johnson 1976; Eggers 1977). The change in stomach fullness would then be described by: dSldt = -kS orS, = So^*' (1) 510 CLARKE: DIEL FEEDING PATTERNS OF MESOPELAGIC FISHES where S is stomach fullness as percentage of fish weight; So and S,, the values at the beginning and end of a period oft hours; and/?, the instantaneous evacuation rate in per hour. For most of the species considered here, there is no extended period of decline in stomach fullness where the above assumptions are met, but a rough estimate of /? is possible for//, prox^'mum and three species of stomiatoids. Hygophum proximum ap- parently ceases feeding early in the night, and stomach fullness declines from N2 to SR under essentially constant conditions, i.e., the fish re- main in the upper layers. Stomach fullness de- clines from SS to SR in Vinciguerria nimbaria, Valenciennelliis tripunctulatus, and Danaphos oculatus, and except for relatively brief periods of migration in the first two species, they remain at the same temperature, etc., for this period. The values of ^ for these four species were calcu- lated by simply using the integral form of Equa- tion (1) and the median values of S for the begin- ning and end of the periods mentioned above (Table 4). (Other fitting procedures, such as least square methods, require that a number of ques- tionable statistical assumptions be made.) The values of k are inversely correlated with night depth and thus positively with temperature being lowest for D. oculatus, highest for Vinciguerria nimbaria and H. proximum, and intermediate for Valencienellus tripunctulatus. For each of the four species, prey concentration and temperature, pressure, etc., were essentially constant throughout the period when feeding oc- curred (SS to N2 for H. proximum and SR to SS for the stomiatoids). It is not unreasonable to assume, as a first approximation, that feeding rate was constant during the periods of increasing stomach fullness. Changes in fullness would then be de- scribed by: dS/dt ^ F - k'S (2) where /j ' is the instantaneous evacuation rate dur- ing the period of feeding, andF is the feeding rate in percentage bodily weight per hour. Integrating and rearranging gives an equation forF in terms of ^', the duration of the feeding period /' in hours, and median fullness at the beginning (Sq ' ) and end (S, ') of the feeding period: F = k' (S/ X S„'e-''-' 1 -*'t' (3) (In some cases, there were a few relatively high values of stomach fullness among the data for a given period; consequently, the feeding rate of some individuals may have been lowered due to satiation. Such values had little effect on the me- dian, and thus the assumption of constant feeding rate is probably not seriously violated as long as medians are used in the calculations.) Estimates of feeding rate and daily ration ( = Ft') were calculated (Table 4) using median values of stomach fullness at SR and SS as So' and S,', respectively, for the stomiatoids and, similarly, SS and N2 for H. proximum. Since both D. oculatus and H. proximum feed at the same temperatures as those under which the instantaneous evacua- tion rates were estimated above, /? ' in Equation (3) was assumed equal to/? calculated from Equation ( 1 ). The daytime or "feeding" temperatures of Vin- ciguerria nimbaria and Valenciennellus tripunc- tulatus are lower than those under which k was estimated from Equation (1). During the day both species occur at nearly the same temperature as does D. oculatus both day and night. Con- sequently, for each of the two migrating stomiatoids, two values of feeding rate and daily ration are given in Table 4 — one calculated using Table 4. — Estimates of instantaneous stomach evacuation rates, feeding rates, and daily rations for four species of mesopelagic fishes based on changes in median stomach fullness over the dial cycle. The first three columns give the sampling periods (Table 1 ) between which feeding rate was assumed to be zero, the duration of this interval it), and the calculated instantaneous stomach evacuation rate ik). The last five columns give the sampling periods between which feeding rate was assumed constant and positive, the duration of this interval (f), the instantaneous stomach evacuation rate assumed for the feeding periods (k '), and calculated feeding rate (F in % of bodily weight per hour) and daily ration (/? = Ft' in '7c of bodily weight per day). For both Valenciennellus tripunctulatus and Vinciguerria nimbaria, two values of /e ', F, and R are given: the higher values under the assumption of constant stomach evacuation rate night and day ik' = k), the lower under the assumption that stomach evacuation rate during the feeding period was lower and equal to that estimated for the nonmigrating, deep-living, Danaphos oculatus. See text for formulae and further explanation. Species Nonfeeding period t (h) k (h-') Feeding period r (h) k' (h- F(%h) ,;d) Hygophum proximum N2-SR 6.8 -0 52 SS-N2 45 -0.52 1.26 5.7 Danaphos oculatus SS-SR 11.3 -0.10 SR-SS 127 -0.10 015 1.9 Valenciennellus SS-N4 8.7 -0.22 SR-SS 127 -0.22 0.57 7.3 tnpunctulatus -0.10 0.34 4.3 Vinciguerria SS-SR 11.3 -0.38 SR-SS 127 -0.38 1.42 18.1 nimbaria -0.10 0.51 6.5 511 FISHERY BULLETIN: VOL. 76, NO. 3 k' = k from Equation (1) and the other using k ' = 0.10, the value for D. oculatus. The estimated ration for Vinciguerria nimbaria seems inordinately high (18%) if/?' is assumed equal to k, the nighttime estimate of evacuation rate. Such values have been estimated for very young, rapidly growing zooplanktonivorous fishes, e.g., A/osa aestivalis (Burbidge 1974) and Oncorhynchus gorbuscha (Parsons and LeBras- seur 1970). Data from Kjelson and Johnson's 1 1976) study of postlarval Lagodon rhomboides and Leiostomus xanthurus feeding rates on zoo- plankton yield estimates of daily ration of only 9.4 and 8.69c , respectively, in terms of wet weight ( my calculations from their data). The estimated ra- tion for V. nimbaria using the low value for /? ' and that for//, proximum lie within the range of val- ues observed for larger individuals in the first two studies cited above and for Morone chrysops juveniles feeding on zooplankton (Wissing 1974). The daily ration of the California sardine, Sar- dinops caerulea, which is a larger zooplanktoni- vore, is apparently slightly lower; judged from Lasker's (1970) estimates of metabolic and grovvd:h requirements, the daily ration is probably about 3-4% ( in terms of calories) for the sizes considered. The above comparisons are admittedly stretched and ignore, among other things, possible differences due to environmental temperature, but the similarity of estimated daily rations of//. proximum and V. nimbaria to those of shallow- living planktonivores is not entirely unexpected. Childress and Nygaard (1973) have shown that the chemical composition of mesopelagic fishes which migrate to the upper layers at night is more similar to that of epipelagic species than to non- migrating, deep-living forms. Differences between estimates for the three stomiatoids are correlated with the extent of verti- cal migration. Nighttime stomach evacuation rate is highest in V. nimbaria, lowest in D. oculatus, and intermediate in Valenciennellus tripunctulatus. Feeding rate and daily ration es- timates show the same trend regardless of whether or not daytime stomach evacuation rates are assumed lower. The absolute values of stomach fullness at the end of the feeding period (Figure 3) are also highest in Vinciguerria nim- baria and lowest in D. oculatus. These trends in- dicate a possible adaptive value for the upward migrations of some stomiatoids. The higher tem- peratures encountered at night by migrators could allow processing of larger meals and presumably faster growth, turnover, etc., rates than for species which remain at depth day and night. ACKNOWLEDGMENTS I thank the captain and crew of the RV Teritu for their cooperation on their vessel's final cruise; the many people who assisted in collection of the samples; P. J. Wagner and G. L. Hoff for careful and competent sorting of the fishes and dry weight determinations, respectively; and K. Gopala- krishnan for identification of euphausiids and de- capods. This research was supported by NSF GA- 38423 and the State of Hawaii, Hawaii Institute of Marine Biology. LITERATURE CITED Anderson, R. 1967. Feeding chronology in two deep-sea fishes off Cal- ifornia. M.S. thesis, Univ. South. Calif., Los Ang., 11 p. Baird, R. C, T. L. Hopkins, and D. F. Wilson. 1975. Diet and feeding chronology of Diaphus taaningi (Myctophidae) in the Cariaco Trench. Copeia 1975:356-365. BRINTON, E, 1967. Vertical migration and avoidance capability of euphausiids in the California Current. Limnol. Oceanogr. 12:451-483. BURBIDGE, R. G. 1974. Distribution, growth, selective feeding, and energy transformations of young-of-the-year blueback herring, Alosa aestivalis (Mitchill), in the James River, Virgin- ia. Trans. Am. Fish. Soc. 103:297-311. CHILDRESS, J. J. 1975. The respiratory rates of midwater crustaceans as a function of depth of occurrence and relation to the oxygen minimum layer off Southern California. Comp. Biochem. Physiol. 50A:787-799. Childress, J. J., and M. H. Nygaard. 1973. The chemical composition of midwater fishes as a function of depth of occurrence off southern California. Deep-Sea Res. 20:1093-1109. Clarke, T. A. 1973. Some aspects of the ecology of lantemfishes (Myc- tophidae) in the Pacific Ocean near Hawaii. Fish. Bull., U.S. 71:401-434. 1974. Some aspects of the ecology of stomiatoid fishes in the Pacific Ocean near Hawaii. Fish. Bull., U.S. 72:337-351. Clarke, T. A., and P. J. Wagner. 1976. Vertical distribution and other aspects of the ecology of certain mesopelagic fishes taken near Hawaii. Fish. Bull., U.S. 74:635-645. DeWITT, F. a., JR , AND G. M. CaILLIET. 1972. Feeding habits of two bristlemouth fishes, Cv- clothone acduudens and C. signata (Gonostomatidae). Copeia 1972:868-871. EGGERS, D. M. 1977. Factors in interpreting data obtained by diel sam- 512 CLARKE: DIEL FEEDING PATTERNS OF MESOPELAGIC FISHES pling of fish stomachs. J. Fish. Res. Board Can. 34:290- 294. Gannon, J. E. 1976. The effects of differential digestion rates of zoo- plankton by alewife, Alosa pseudoharengus. on determi- nations of selective feeding. Trans. Am. Fish. Soc. 105:89-95. GJOSAETER, J. 1973. The food of the myctophid fish, Benthosema glaciale (Reinhardt), from western Norway. Sarsia 52:53-58. Gordon, D. C. 1970. Chemical and biological observations at Sta. Gol- lum, an oceanic station near Hawaii, January 1969 to June 1970. Rep. Hawaii Inst. Geophys., Univ. Hawaii, H.I.G.-70-22,44 p. HOLTON, A. A. 1969. Feeding behavior of a vertically migrating lan- temfish. Pac. Sci. 23:325-331. HOPKINS, T. L., AND R. C. BAIRD. 1975. Net feeding in mesopelagic fishes. Fish. Bull., U.S. 73:908-914. 1977. Aspects of the feeding ecology of oceanic midwater fishes. In N. R. Andersen and B. J. Zahuranec (editors). Oceanic sound scattering prediction, p. 325-360. Plenum, N.Y. KAMPA, E. M. 1970. Underwater daylight and moonlight measurements in the eastern North Atlantic. J. Mar. Biol. Assoc. U. K. 50:397-420. KJELSON, M. a., and G. N. JOHNSON. 1976. Further observations of the feeding ecology of post- larval pinfish, Lagodon rhomboides, and spot, Leiostomus xanthurus. Fish. Bull., U.S. 74:423-432. LASKER, R. 1970. Utilization of zooplankton energy by a Pacific sar- dine population in the California current. In J. H. Steele (editor). Marine food chains, p. 265-284. Univ. Calif. Press, Berkeley. Legand, M., p. Bourret, p. FOURMANOIR, R. GRANDPERRIN. J. A. GUEREDRAT, A. MICHEL, P. RANCUREL, R. REPELIN, AND C. ROGER. 1972. Relations trophiques et distributions verticales en milieu pelagique dans I'Ocean Pacifique intertropi- cal. Cah. O.R.S.T.O.M., ser. Oceanogr. 10:303-393 Marshall, N. B. I960. Swimbladder structure of deep-sea fishes in relation to their systematics and biology. Discovery Rep. 31:1- 122. MAYNARD, S. D., F. V. RIGGS. AND J. F. WALTERS 1975. Mesopelagic micronekton in Hawaiian waters: Faunal composition, standing stock, and diel vertical migration. Fish. Bull., U.S. 73:726-736. Merrett, N. R., AND H. S. J. Roe 1974. Patterns and selectivity in the feeding of certain mesopelagic fishes. Mar. Biol. (Berl.) 28:115-126. PANDIAN, T. J. 1967. Transformation of food in the fish Megalops cyp- rinoides. I. Influence of the quality of food. Mar. Biol. (Berl.) 1:60-64. Parsons, T. R., and R. J. LeBrasseur. 1970. The availability of food to different trophic levels in the marine food chain. In J. H. Steele (editorl. Marine food chains, p. 325-343. Univ. Calif. Press, Berkeley. Roe, H. S, J. 1972. The vertical distributions and diurnal migrations of calanoid copepods collected on the SOND Cruise, 1965. 1. The total population and general discussion. J. Mar. Biol. Assoc. U.K. 52:277-314. Tate, M. w., and R. C. Clelland. 1957. Nonparametric and shortcut statistics in the social, biological, and medical sciences. Interstate Printers and Publishers, Danville, 111., 171 p. TEAL, J. M. 1971. Pressure effects on the respiration of vertically mi- grating decapod Crustacea. Am. Zool. 11:571-576. Tyler, H. R., Jr.. and W. G. Pearcy. 1975. The feeding habits of three species of lanternfishes (family Myctophidae) off Oregon, USA. Mar. Biol. (Berl.) 32:7-11. WISSING, T. E. 1974. Energy transformations by young-of-the-year white hass Moronechrysops (Rafinesque) in Lake Mendota, Wis- consin. Trans. Am. Fish. Soc. 103:32-37. 513 ON THE RESTRUCTURING OF THE PELLA-TOMLINSON SYSTEM R. Ian Fletcher* ABSTRACT The time-dependent analysis of an earlier work is extended to the equilibrium case of the Pella- Tomlinson system, and the relationships between the equilibrium and nonequilibrium versions of the restructured system are developed. The dual formulations of the conventional analysis are avoided and maximum sustainable yield is separated from the indeterminacy of the system. All arbitrary coefficients are eliminated and the management components incorporated directly into the system equations. The source of the statistical degeneracy in the model is revealed and explicitly formulated, and in the companion article by D. Rivard and L. J. Bledsoe (this issue of the Fishery Bulletin) the restructured model is treated by a new statistical method that subdues the estimation failures associated with past treatments of the Pella-Tomlinson system. Because the equilibrium versions of all stock- production models follow from steady-state inte- grations, the strategy of fishery regulation be- comes a strategy of accommodation, so to speak, as determined by a pattern of balanced model states where removals just equal the productivities otherwise surplus to the maintenance needs of the stock. Population status usually enters the process in the simple, robust form of integrated numbers or biomass, and the removals of fishing constitute direct fractions of the whole fishable stock without reference to age or weight distributions. Since the appearance of Schaefer s work (Schaefer 1954) the strategy has been applied to the management of many fisheries. Schaefer devised a rational, linearized method for estimating the parameters of Graham's equilibrium model (Graham 1935) from the actual nonequilibrium yields and effort expenditures of a fishery, a contribution that is often misunderstood. In applying Schaefer's method or like schemes of synthesis, it is not so much that one hopes to observe a pattern of equilibrium levels in a fishery or even expects them to come about, but rather, by knowing the response history of a stock to various exploitation pressures, one might then be guided by the model in bringing a stock, through a sequence of man- agement actions, into a state where some desired level of sustainable yield most likely abides. The philosophy is widely accepted in fisheries man- agement but its application is often censured. 'Center for Quantitative Science in Forestry, Fisheries and Wildlife, University of Washington, Seattle, WA 98195. either on economic or biological grounds (see, for example, Larkin 1977) The exploitation model of Pella and Tomlinson ( 1969), as it is customarily thought of, extends the more "basic" model of Graham from a system of second degree in nonlinearity to a flexible or more "general" system of indeterminate degree. The in- creased flexibility comes into the Pella-Tomlinson model through the addition of a single exponential parameter, but the analytical peculiarities that accompany the improvement often lead to paradoxical ends since the equations of the system then permit the simultaneous generation of good data fits and poor parameter estimates (see the commentary of Ricker 1975:323-326 and the treatments of Fox 1971, 1975). This disturbing trait of the statistical model arises from the conflict between the variable (or parametric) cur- vature of the analytical model and the coupling of that curvature, in the conventional formulations, with all the coefficients of the system. As shown in a prior work (Fletcher 1978), those effects may be separated in the time-dependent analysis by re- structuring the system equations so as to accom- modate directly the critical-point coordinates of the system graphs. In this paper we extend the analysis to the equilibrium version of the Pella- Tomlinson system, and we show the relationships between the equilibrium model and the (restruc- tured) time-dependent equations. For a stock of mixed age classes, the most difficult problem in applying any equilibrium model will lie, essentially, in the interpretation of time-dependent transitions between idealized states (however momentary, long-enduring, or Manuscript accepted March 1978. FISHERY BULLETIN: VOL. 76. NO 3. 1978. 515 FISHERY BULLETIN: VOL. 76, NO. 3 unobserved such states may be), since the stock will include simultaneously the young and the old, the older having accumulated a probabilistic his- tory of mortality, fecundity, and growth which may differ considerably from the current schedule that affects both. Various tactics for adjusting the parametric mechanics of stock-production models to such long-term, delayed influences are given by Gulland (1969), Fox (1975), Walter,^ and others, but in the case of the Pella-Tomlinson system the difficulties have been compounded by artifacts of the conventional analysis and by an instability inherent to the mathematical indeterminacy of the system itself. With the critical-point analysis, most of those impediments will convert to tracta- ble relationships or vanish altogether. We can suppress the troublesome dual formulations as- sociated with the conventional casting of the sys- tem, we can uncouple the indeterminate exponent and the coefficients of the governing equations, and we can make explicit the relationships be- tween parametric graph curvature and the man- agement components of the system. THE REFORMULATED GOVERNING EQUATIONS Stock-production models, as they are usually defined, arise from the common premise that a fish stock, when reduced by exploitation to a level below some prior abundance, will always strive to recover its former size in accord with some latent, self-regulating mechanism of restoration. Irre- spective of the compensatory details, any such re- covery must accrue directly from the productivity of the stock, and in the conventional representa- tion of the Pella-Tomlinson system, the latent capacity for biomass production in a stock of fishes is given the dual formulation P(B) = ±aB" + b B. la) lb) • P(B) being the production rate of the stock at stock sizeB. Equation ( la) applies when exponents falls on the range 01. In either case, all the critical compo- nents of the system — maximum stock size, maximum productivity, the stock level where maximum productivity occurs — depend in some ^Walter, G. G. 1975. Non-equilibrium regulation of fisheries. Int. Comm. Northwest Atl. Fish. Res. Doc. 75/IX/131, 12 p. 516 way on the numerical value assigned to exponent n. That is, root B^c is given by lll—n B the critical ordinate p (which corresponds to the stock level where maximum productivity occurs) is determined by P = 1/1—?? while extremum coordinate m (which corresponds to productivity P ma.x ' must be determined from the formula m n \b ) the plus sign applying to Equation (la) and the minus sign to Equation (lb). Although exponent n controls the graph curva- tures of Equations ( la) and ( lb), the nonzero roots and extrema are controlled hy By- and the coordi- nate pair (p, m). As shown by Fletcher (1975), coordinate m has no essential dependence on ex- ponent n, and with the appropriate transforma- tions the dual formulation (Equations ( la, b)) may be suppressed. In consequence, either of the parametric sets {m, p,Bx}or{m,/7.5x}will consti- tute a complete set of independent governing parameters for latent productivity in the Pella- Tomlinson system, and the dual formulation (Equations (la, b)) converts to the single differen- tial equation for latent productivity P = ym it) -HI)"- <^» with y a purely numerical factor wholly prescribed by n as ,n/n — l y 1 ■ (3) With the coefficients so cast, the sign reversals at turning point /; = 1 become automatic, and the consolidated interval of definition for n becomes 00 and the stock increases. Should • • • Y = P, then B = Q and biomass trajectory Bit) exhibits an extremum, which is the necessary condition for equilibrium fishing. Yield rate Y cus- tomarily takes the form With initial time /„ set at zero, the integration constant C in Equation (8) becomes ^0^ - B. The quantity B ;, , when positive in Equation (8), becomestheadjustment level such thatBf^^ -► B* over time. When, for certain ranges of n and F, quantity B*<0, then the zero root of Equation (7) applies and B(t) -► 0. When mortality F takes the value MSY ■(^) ■ym bZ (9) irrespective of the value of parameter n, then B(t) -►p and Y ->m (which are the conditions, in the equilibrium limit, for maximum sustainable yield). In terms of the parameter set { m, p, B^^ }, Equation (9) becomes, simply, F. MSY m P Figure 1 gives a summary of the general con- straints on the time-dependent system; for a more detailed treatment of system behavior, see Flet- cher (1978). THE RESTRUCTURED EQUILIBRIUM SYSTEM Yit) = F{t) ' B{t) (6) with the assumption that all fish of the fishable stock share equal probabilities of capture. By ad- mitting Equations (2) and (6) into Equation (5), the differential equation that governs net produc- tivity in the restructured system becomes B = 7m B B ym ©" FB (!) and over any time interval, however brief, that mortality F might be presumed to have a fixed value, biomass variable B in Equations (6) and (7) has the general time-dependent solution B(0 = (b,i-" + Cexp ((7m/B„ * [ym-FB^) By Equations (2) and (5), the time-varying rate of yield in the reformulated Pella-Tomlinson sys- tem takes the form y = ,„ (y _ ,„ (y _ s. ,10, and when, for given F and n, governing Equation (7) exhibits a positive root, then B(t) -► B^ and B -► 0 in Equation (10), and yield rate Y, over sufficient time, approaches a constant value. In the steady-state tor "equilibrium") limit, yield then accumulates as F)(l -n)t)y"'-''\ (8) 1 /(!-«) B. 517 FISHERY BULLETIN: VOL. 76. NO. 3 00 Y — Fb^ A 6, 5- n-i\ r>p MSV n >-/' E>oo Y/77 ' MSV Figure l. — Time-dependent response of the Pella-Tomlinson system to parametric variations of exponent n and mortality F. The upper diagram summarizes system response when n falls on the range 0 < /! < 1 . The adjustment level of biomass is never zero for this range of n however great the value ofF, and mortality F ms-, has no absolute constraints; such a stock cannot be fished to extinction. The lower diagram summarizes system behavior when n falls on the range n > I. Mortality F ^sv 's then constrained to the interval indicated by the diagram. WhenF exceeds the critical value ym/By^, then the stock, over sufficient time, trends to extinction. f (lY = ym and for any such equilibrium interval t, the inte- grated yield rate iY^h) takes on the parametric formulation = 7m (11) with maximum latent productivity m of the time- dependent system becoming the maximum sus- tainable yield rate (the MSY) of the equilibrium system. With B^ as the parametric variable in Equation (11), a zero left endpoint exists for Y^.Ij when /? > 1 and F = ymlB^. Should F exceed the critical value ymlBy- when exponent n >1, no equilibrium state exists; such conditions in the 518 time-dependent system correspond to extinction trends. But when n has any value on the range 00 no matter how great the value of F. That is, when 01, however, the corresponding stock can have non- zero adjustment levels B^ only whenF1 and when fishing mortality ex- ceeds the critical value ym/Bx, the "adjustment" level corresponds to extinction and Equation (12) does not apply. Upon the substitution of Equation (12) into Equation (11), the direct relationship between equilibrium yield and equilibrium fishing mortal- ity becomes 1 T V ym J B. (13) and the fishing mortality that maximizes Equa- tion ( 13) is given by Equation (9). That is, with the substitution of F^gy into Equation (13) then Y*/t = m. Under the equilibrium conditions, the conven- tional quantity U (which signifies accumulated catch per unit of fishing effort as a function of fishing intensity f/r) can be cast into the restruc- tured form \ °° jm T I (14) which eliminates the explicit appearance of catch- ability coefficient g, permitting instead the direct quantification of maximum sustainable yield m. Quantities U and J7x have the customary mean- ings Y U = —f' ^Y* being the yield accumulated over time interval r as a consequence of ef- fort f). U^ ^ qB^ (q being the individual probability of capture per unit of fishing effort f). Should the accumulation interval t be a year, the variable U becomes annual CPUE (catch per unit of effort) and the variable [It becomes effort per annum. With exponents >1 in the Pella-Tomlinson system, no steady-state CPUE exists for a fishing intensity in excess of critical value ym /U^.. But if n 519 FISHERY BULLETIN: VOL. 76. NO. 3 has any value on the interval 0 Inionmcubion itouo > PlOQiam contAol ilow Vautjo. -input Vzcla-ion aZgoiyCthm Calculation algo- lithm an analytical form for the uncertainties in the final values of the parameters (Bevington 1969). By letting SiO) be the weighted residual sum of squares for the final parameter estimates, how- ever, the variance-covariance matrix of the esti- mates (Bard 1974) can be approximated by imation in the neighborhood of O is appropriate. A necessary and sufficient condition for the F-distribution to be appropriate here is that dif- ferences in true and estimated parameter values are independent and approximately normally dis- tributed with zero mean and equal variance. Ve = j^ J S(e)/(r-5). (12) Some idea of the joint variability of the parame- ters can be obtained by evaluating the ellipsoidal confidence region, based on the assumption that the linearized form has validity around B ( Draper and Smith 1966). The confidence region is then given by [e-B] J'^ J[B-B]' F(5,r-5,l-a), (13) where F(5, r-5, l-«) is the standard tabulated F-statistic. The ellipsoid is not a true confidence region, of course, since the dependent variable, Y, is a nonlinear function of B. The intervals ob- tained are valid to the extent that a linear approx- DETERMINATION OF STARTING VALUES In order to reduce the number of iterations re- quired to minimize Equation (8), reasonably accu- rate starting values should be employed. Starting values can be calculated from a linearization and simplification of the basic model. STEP 1. By using y,,y2, .. . , Y, and /■,, /a, . . . fr, find an estimate of g from the Delury technique. Note that this procedure generally underesti- mates (? (see Ricker 1975). Correction forq will be provided in step 4. STEP 2. Find estimates of B, + i from the equa- tion B ..1 = (^M.l^^..l,..2)/2' (14) 526 RIVARD and BLEDSOE: PARAMETER ESTIMATION FOR PELLA-TOMLINSON MODEL where (by assuming ft_t^\ constant over the inter- val t,t + \) mates of m and B x- Finally, Bo is approximated by B t.t+\ i.t+i'^ 't.t + \ (15) Note that Y,j + y and /",, ^i correspond to y, and/', of Equations (6) and (5). STEP 3. Let /z= 2, as in the Graham-Schaefer model, and estimate m andB-^ by fitting the linear model (16) where y = - Tjn—l t B^dt + qf^, x^ = B';-\ Equation (16) is derived from Equations (1) and (3). However, Equation (16) requires an estimate of the relative growth rate dB,/Bidt, say i?,. As suggested by Causton ( 1969), the mean value of 7? between / and t+2 is given by ^M.2 = (lnB^^,-lnB^)/2. (17) For the purpose of fitting Equation ( 16), quantity /?,,+2 n^^y t>6 considered an estimate ofRi + i, which corresponds to B, + ^. Whence. Equation (16) provides estimates of/?? andSx as m = h %"] (l/l-n) 7 B = ct. d. 1/1— ^7 (18) (19) STEP 4. Steps 2 and 3 are repeated iteratively for increasing values of q. The value of g which provides the minimum residual sum of squares [1 (Y, - y, )2] is accepted as the appropriate start- ing value for q. STEP 5. Step 3 is repeated iteratively for in- creasing values of n, parameter q being kept con- stant. The value of n which provides the minimum residual sum of squares [1 (y, - Y,)^] is accepted as the appropriate starting value for n . In the last iteration, Equations (18) and (19) provide esti- ^0 = ^0,1 a (20) where B, andB,, , are estimated by Equations ( 14) and (15), respectively. Steps 1 through 5 provide a set of starting values for the optimization algorithm (11). Usually the starting values are near the solution and few iter- ations will be needed. Of course, it would be possi- ble to derive algorithms for more accurate starting values, but our purpose here is to find a rough estimate for each coefficient and to let the iterative procedure (11) converge to the minimum. Some- times, by experience or by prior information, it is possible to provide starting values as satisfactory as those provided by the algorithm given above. MONTE-CARLO SIMULATIONS The parameter values that we chose to generate the data of Table 1 (deterministic model) were recovered exactly by the estimation procedure. Results of fitting 18 stochastic versions of the de- terministic model are also included in Table 2. Based on our simulation results, there do not ap- pear to be any serious problems with bias of parameter estimates. The bottom line of Table 2, which gives the coefficients of variation of the parameter estimates, reveals that estimates of the three parameters of principal interest to the man- ager have the smallest variability. Those parameters are maximum sustainable yield (m, C.V. = 147f ), optimal effort level(/MSY> C.V. = 67r ), and yield per unit of effort at optimum effort (t/ivjsY' C.V. = 9%). Our results confirm the ob- servations of Fox (1971) and Pella and Tomlinson (1969) on the robustness of m and /"msy with re- spect to error in the measurement of the yield data. From Table 2, we can also compare variance estimates from Equation ( 12 » with variance of es- timates for 10 replicates at o" = 0.200. Equation (12) appears to give (approximately) unbiased es- timates of the variance of the sampling distribu- tion of G. Also, out of the 19 cases considered, the true parameter value lay outside the arbitrary ±2 (SD) confidence interval twice form and only once each forB^,/?, andB,,. Although we did not employ an extensive Monte-Carlo simulation, our results suggest that the normal approximation to the sampling distribution of Q is an acceptable ap- proximation, at least for management purposes. 527 FISHERY BULLETIN: VOL. 76, NO. 3 In a few additional simulations (replicates 13 and 15), parameters obtained by the five-para- meter procedure were ill determined. A parameter is considered ill determined if its estimated value responds unreasonably to seemingly insig- nificant variations in the data (Bard 1974). The basic difficulty is that the model is extremely gen- eral and capable of several types of behavior over the space of 0. In the Pella-Tomlinson system, ill determination often occurs whenever an itera- tion of the algorithm (11) gives an estimate of O such that the point (m , /msy* o^ ^he yield-effort plane lies outside the concentration of data. In such a circumstance the exponent n takes on small- er and smaller values in the successive iterations and the solution of system (1) and (3) degenerates to an exponential form for which only four parameters are required for uniqueness. That is, as/? ->0. in Equation (1), then (B/By.)" -►I and y -► - 1 . The five-parameter procedure then over- prescribes the system, which in turn predisposes the coefficient estimates to extremely large var- iances. The ultimate irony here is the fact that wholly unrealistic parameter estimates still gen- erate good fits to the catch-effort history (i.e. small residuals). For example, in Figure 2 the fitted five-parameter curve predicts /"jv^gy near infinity while in the true model /'msy actually corre- sponds to 174,000 units of effort. However dif- ferent the equilibrium curves are, the five- parameter procedure still generates a good fit to the catch history (S(0) = 1.10). Incompleteness of information over a wide range of effort values, as well as excessive noise in the catch-effort data, will tend to bring about such pathological condi- tions. To overcome these difficulties, reformulation of the estimation problem is necessary. By the fol- lowing considerations, the five-dimensional parameter space can be reduced to three dimen- sions. First, we will approximate £„ by Equation ( 20). Furthermore, if the data contain information on the yields under low exploitation, we may define Sx as B = MAXiYJq f.) r. (21) By using Equations (20) and ( 21 ),fi, I andfixcanbe deleted from G, leaving only m, q, and n as the independent parameters requiring estimation. It is important to understand at this point that B^ and fix are not fixed; they are reevaluated by Equations (20) and (21) at each iteration, along with the parameters tu ,q, and n . In fact, the solu- tion of Equations ( 1) and (3), as well as Equations (20) and (21), specify a new model with unknowns O =[m,q, nY. By this restructuring, much of the degeneracy associated with the model can be eliminated. As shown in Figure 2, this procedure also provides a closer correspondence between the "estimated" and the "true" equilibrium model. Furthermore, the three-parameter procedure still generates an adequate nonequilibrium catch his- tory (S(G) = 1.40). In a Monte-Carlo simulation study, parameter estimates obtained by using these transformations fell within reasonable FlOURE 2. — Comparison of the "true" model with the models obtained by using the estimation procedure on three and five parameters respectively- Solid lines show equilibrium yield curves; data points show nonequilibrium simu- lated (dots) yields and predicted (circles) yield values from the three-parameter approach. Dashed vertical lines indi- cate the magnitude of residuals. UJ 528 RIVARD and BLEDSOE: PARAMETER ESTIMATION FOR PELLA-TOMLINSON MODEL hounds (Table 3). Out of the 20 cases considered, the true parameter vahie lay outside the arbitrary ±2 (SDl confidence interval only once tor ru and/?. Also, variance estimates were comparable with the variance estimates of the five-parameter pro- cedure (compare Tables 2 and 3). Table 2. — Estimated parameters for the deterministic model and for 18 stochastic replicates. The Levenberg-Marquardt algorithm is employed in a five-dimensiona! parameter space i m ,By.,n,q,Bg). For each parameter and replicate, the parameter estimate ± its estimated standard deviation from Equation ( 12) are tabulated. Replicates 13 and 15 have been e.xcluded due to degeneracy of the model, as discussed in the text. Repli- m Sx q Bo 'msv cate (f = S (6) (r-3). ^For 12 replicates with -r = 0 200 ^Overall standard deviation of parameter estimates for 12 replicates with u = 0.200. 529 FISHERY BULLETIN: VOL. 76. NO. 3 CASE STUDIES We applied the three-parameter method to the catch-efibrt data of the yellowfin tuna fishery of the eastern tropical Pacific, 1934 through 1967 [the same data that were analyzed by Pella and Tomlinson ( 1969) and by Fox ( 1971)]. Table 4 gives a comparison of results, and our final equilibrium model is shown by Figure 3. As indicated by Table 4, the parameter estimates of the Levenberg- Marquardt method are comparable with the esti- mates that Fox obtained with his search al- gorithm. Pella and Tomlinson also employed a searching algorithm but their minimization criterion was an unweighted least-squares func- tion. Our standard deviation estimate is very small for w( MSY) but relatively large f or B^,n,q, and fi„, which is a consequence of insufficient in- formation in the yellowfin tuna data on yield at high fishing rates. With such limited information, one can anticipate that neither the shape nor the location of the descending portion of the equilib- rium curve (dashed in Figure 3) could be deter- mined with much accuracy, and the large variance estimates on the system coefficients reflect this situation. Of course, the variance estimates for /^j^Y ^^^ ^MSY ^^^ always be calculated by the delta method, and to avoid the complex deriva- tions that accompany the presence of covariance terms, an alternative would be to define a new parameter space so as to estimate /"msy oi' ^^msy directly. The variance-covariance matrix for the coefficients would then provide the desired infor- mation on the variability of those parameters. Our final example is based on the data from the Pacific halibut fishery in International Pacific Halibut Commission Area 2, as given in Ricker (1975, table 13.1). To analyze these data, Ricker derived an estimate of c/ from the age composition of the catch. Then he obtained parameter esti- mates for a Graham-Schaefer model by regressing Ye/B against B and Yg/f against / (Ricker 1975, examples 13.5 and 13.6). In both cases, Ricker employed GM and Nair-Bartlett regression. The results Ricker obtained by fitting the Graham- Schaefer model were compared with the results we obtained from fitting the generalized stock pro- duction model by our three-parameter version of the Levenberg-Marquardt method (Table 5). The latter provided estimates of m, q, and n with rela- tively small variance estimates. Furthermore, the estimate of n appears to be significantly different from 2.00, which validates the use of the Pella- Tomlinson model. Nevertheless, estimates of m are not significantly affected by the choice of the wrong model, while estimates of /msy are slightly Table 4. — Comparison of parameter estimates obtained by Pella and Tomlinson (1969), by Fox ( 1971) and by the Levenberg-Marquardt algorithm for the yellowfin tuna in eastern Pacific Ocean. Values that follow the ± signs are the standard-deviation estimates for each parameter. Bx q 8o m Reference P (10«) n (10--) (108) (108) 'MSV ^MSV Residuals Pella and Tomlinson (1969, table 5) — — 1 40 450 — 1 826 35.300 5,173 1 78 ■ 10^8 Fox (1971. table 4) — 1 427 2 10 8 10 1 206 1 926 32.700 5.890 0,736 Levenberg-Marquardt — 1 448 208 8 01 1 192 1.924 32.700 5.884 0735 algorithm ±0 890 ±0 75 ±4 9 ±1 24 ±090 Levenberg-Marquardt algorithm 0 27 1 274 2 30 9 08 1 079 1 962 32.170 6.097 0641 (correlated error) -025 *0 653 ±0 55 ±4 7 ±0553 ±0 106 Table 5.— Comparison between the estimates of Ricker ( 1975) for the Pacific halibut ( Interna- tional Pacific Halibut Commission Area 2) and those obtained by the Levenberg-Marquardt algorithm. Values that follow the ± signs are the standard-deviation estimates for each parameter. B , q m Reference P MO'') n (10-') (10^) 'msv ^MSY Ricker (1975. example 13 5) GM regression — 204 2.00 907 31 2 337 926 Nair-Bartlett regression — 195 2 00 9 07 31 0 3 50 88 6 Ricker (1975. example 13 6) GM regression — 256 2 00 9 07 33 0 2 84 1162 Nair-Baniett regression — 239 2 00 9 07 31 8 294 108 2 Levenberg-Marquardt algorithm — 187 1 28 1445 31 6 2 83 1117 ±18 ±0 09 ±1 36 ±083 Levenberg-Marquardt algorithm 0 33 188 1 28 14 33 31 8 2 84 1120 (correlated error) -0,16 ±22 ±0 12 ±1 67 ±10 530 RIVARn and RLEDSOE PARAMETER ESTIMATION FOR PELLA-TOMLINSON MODEL 2 5 0 2 0 0- 1 5 0 Q 10 0- 111 >- 5 0 Figure 3. — Equilibrium stock produc- tion model for yellowfin tuna data, from 1934 through 1967, as determined by the Levenberg-Marquardt algorithm. EFFORT (ilO boil-liri) > overestimated in most applications of the Graham-Schaefer model (note also that Ricker's treatment assumes equilibrium). In contrast to the yellowfin tuna data, analysis of the equilib- rium model for halibut data indicates that fishing effort has been concentrated slightly to the right of /msy (compare Figures 3 and 4). In the preceding case studies, the Levenberg- Marquardt algorithm gave estimates with rela- tively small coefficients of variation. In both cases, Figure 4. — Equilibrium stock produc- tion models for Pacific halibut (Interna- tional Pacific Halibit Commission Area 2), from 1910 through 1957, as deter- mined by Ricker (1975, examples 13.5 and 13.6, Nair-Bartlett regressions) and by the Levenberg-Marquardt al- gorithm (three-parameter version). 10 11 EFFORT (ilO ikitll) 531 FISHERY BULLETIN: VOL 76, NO 3 e.g.. the coefficients of variation for the estimates of maximum sustainable yield {rh ) were below G^c . It is questionable, however, whether the data can justify such precision. Variability of the exploited population due to migration, to changes in fishery regulations over time, and to expansion of the fishing areas, as well as variability of q due to learning by fishermen and to technological de- velopments, are important factors underlying the complexity of events influencing the serial catch- effort information. In future research, alternative forms of the model in which g is a variable parameterized with respect to time will be explored. Furthermore, in a randomly fluctuating environment, equilibrium population levels (and MSY, by extension) are not constant and the equilibrium points are instead described by a probabilistic cloud representing the equilibrium probability distribution (May 1974). The knowl- edge of this equilibrium probability distribution would give us some idea of the probability of achieving the desired management goal ( MSY, for instance). W = (r-iy Y.Y^P Y,Y,P y,yp y.' : Kyy-'' YYp 1—1 f-< (r-l) Y.Y^p (r-2) Parameter p, constrained between 0 and 1, is a measure of the importance of lags and can be esti- mated along with the parameters of the differen- tial equations (1) and (3). Itcanbeseen that Equa- tion (9) is a particular case of Equation (22), where the off-diagonal elements of W are null. The Levenberg-Marquardt algorithm, as formu- lated in Equation (11), is designed to minimize directly a sum of squares of residuals as given by Equation (9). In order to minimize Equation (22) by using Equation (11), we must scale W by the transformation DISCUSSION ON ERROR STRUCTURE In the preceding examples, we found runs in the time sequence plot of residuals. Those runs indi- cate correlations among the residuals. Serial cor- relation, as we usually find in applying production models to catch data, indicates that the real sys- tem is working differently than the presupposed model and that some minor effects have been neg- lected (such as age composition or environmental factors). But as indicated by Draper and Smith (1966), the effects of correlation can be ignored when the ratio {r - p )/r tends to unity (p being the number of estimated parameters). In certain situ- ations, of course, this ratio is likely to become small (tending to zero as r approaches p) and we may want to consider weights (W,) which account for both the inequality of variance and the correla- tions. In our estimation procedure, the assumption of uncorrelated error can be relaxed by writing Equation (9) in the more general form (J. J. Pella, pers. commun.) X = D~^ W D (23) SiO,p) = [Y-Y] W-i [Y-Yl' (22) where Y is the row vector of observed yields, Y is the row vector of predicted yields, and W is the symmetric, positive definite matrix 532 where D is a diagonal matrix having elements D, 1 r), and write W as y, (I W = D U A U'T D. (24) where U.\U^ is the eigenvalue and eigenvector decomposition ofX. Note thatX is actually the correlation matrix of errors. Therefore Equation (22) becomes sie,p) = [Y-Y] D-i U A~i U'^ D~^ [Y Y]''. (25) Then Equation (25) has the same form as Equation (9), where the weights iW,) are the square roots of the eigenvalues of X and where the residuals are given by [Y— Y J D ' U. Such a procedure requires, however, diagnoalization of an r by r matrix. Moreover, diagonalization must be repeated at least p times for each iteration. This procedure produces a 10-fold increase in computing time. Although an exhaustive study of all possible stochastic effects on the model was not attempted, some simulations were done to determine the magnitude of error in parameter estimates due to serial correlations of the e,. Results are given on Tables 4 and 5. For the yellowfin tuna data, p = RIVARD and BLEDSOE: PARAMETER ESTIMATION FOR PELLATOMLINSON MODEL 0.27. For the Pacific halibut data, p = 0.33. In either case, p exhibits a relatively large coefficient of variation when compared with the elements of O. One could anticipate such results since p reflects the "persistence" of fluctuations in popula- tion size, and the estimation of p would therefore require a longer catch history in order to achieve a greater precision. But more importantly, the val- ues of G and Var[OJ were not significantly altered by the inclusion of the additional parameter. And while the errors of any particular catch history might indeed by correlated, the minimization criterion (9) will provide satisfactory estimates of B despite the fact that correlations do not enter into its formulation. The limited results contained herein suggest that serial correlation can be safely ignored when the ratio (r - p)/r is near unity. Under such a condition the estimation procedure is robust with respect to the assumption of inde- pendence of errors in actual data. CONCLUSION modifications of the model to incorporate such hypothetical effects as migration or stock interac- tions can be made easily. Of course, to the extent that the esti mation procedure must rely strictly on catch-effort data, it will be subject to the same information uncertainties as any other method. But within the basic estimation procedure, we can combine the catch-effort data with prior informa- tion and thereby reduce the uncertainties in our estimates. The prior information can be any in- formation on a state variable, s\xch.asB(t), or even any prior knowledge of the coefficients as express- ed by B ± Var(B). Suppose, for example, that we have information from independent surveys on stock density (acoustic surveys, indirect estima- tion from knowledge of larval densities, or even virtual population analysis from catch records). Such surveys would then provide us with esti- mates B^ each having a variance ViB, ), let us say, at various times t. We can easily introduce such information into the estimation procedure by defining the new objective function The purpose of this paper has been to examine a version of the Levenberg-Marquardt algorithm as an alternative method for estimating the coefficients of the generalized stock production model. The parameter values obtained by this pro- cedure are close to those obtained by previous studies on yellowfin tuna and Pacific halibut. Ob- viously, data requirements are such that a full range of effort values (ranging over low and high exploitation rates) are necessary to insure con- vergence in the estimation procedure and to pro- duce estimates with small variability. Our simu- lations reveal that with the Levenberg-Marquardt method both the estimates of coefficients and the estimates of variances remain approximately un- biased when white noise is considered. If present, such bias is sufficiently small as to be obscured in the variability associated with catch error. The simulations also showed the range of variability in parameter estimates that might be expected for given levels of normally distributed error in catch data. Because the parameters of interest appear explicitly in the system equations, the estimation procedure for the parameters also produces the variance estimates directly. Moreover, the method has a reliability and an efficiency of computation somewhat greater than previous methods. And since the estimation procedure relies on a numeri- cally integrated system of differential equations. s(B) = s w.iY.-yy + s ill I j V -l^D _DX2 {B.-B.y. (26) Introduction of the second term in the objective function constrains the optimization and thereby improves convergence. If the prior information has extremely large variance, then this informa- tion is of no value; the second term of Equation (26) will tend to zero and the objective function then reduces to Equation (9). In general, the alteration permits the simultaneous employment of the two state variables. Therefore, the final coefficients are no longer based solely on catch and effort data; their determination includes our knowledge of previous stock densities. As observed here in a statistical setting, and by Fletcher (1978a, b) in the exact analysis, the Pella-Tomlinson system exhibits internal in- stability in its parametric relationships. That behavior arises from the variable nature of the system's nonlinear! ty, which would not be particu- larly detrimental if our problems were limited strictly to the geometric syntheses of data by curve fitting. But for the purposes of management and preservation of stocks, the subject is elevated partly at least to the status of parameter estima- tion "where we look for procedures to obtain val- ues of the parameters that not only fit the data well, but also come on the average fairly close to the true value" (Bard 1974). Although the Pella- 533 FISHERY BULLETIN; VOL. 76, NO. 3 Tomlinson system exhibits a convenient flexibil- ity with a minimum number of coefficients, the peculiar coupling of the coefficients to the non- linearity of the system often provides more flexi- bility than we care to have, and a conventional least-squares statistic may not be sufficient to con- trol the system in the estimation procedure. In consequence, many constraints have to be imposed on the system in order to obtain convergence in the estimation procedure and to insure reliability in the coefficient values thus estimated. ACKNOWLEDGMENTS We would like to thank R. I. Fletcher, J. J. Pella, and C. G. Walters, each of whom reviewed the manuscript and offered many helpful suggestions. This research was supported by NORFISH, a marine research project of the University of W^ashington Sea Grant Office and the National Marine Fisheries Service (Grant 04-7-158-44021, Office of Sea Grant, National Oceanic and Atmos- pheric Administration, U.S. Department of Com- merce). Financial support was also provided by the National Research Council of Canada and by the Minister of Education of Quebec. Contribution No. 492 of the College of Fisheries, University of Washington. LITERATURE CITED B.AKD, Y. 1974. Nonlinear parameter e.stimation. Academic Press, N.Y.. 341 p. Bevington. p. R. 1969. Data reduction and error analysis for the physical sciences. McGraw-Hill Book Co., N.Y., 336 p. Brown, K. M., and J. E. Dennls 1972. Derivative free analogues of the Levenberg- Marquardt and Gauss algorithms for nonlinear least squares approximations. Numer. Math. 18:289-297. C.M^STON, D. R. 1969. A computer program for fitting the Richards func- tion. Biometrics 25:401-409. CONWAY. G. R., N. R. GLA.S.S. AND J. C. WiLCOX. 1970. Fitting nonlinear models to biological data by Mar- quardt's algorithm. Ecology 51:503-507. Draper. N. R., and H. Smith 1966. Applied regression analysis. John Wiley & Sons Inc., N.Y.,407 p. Fletcher, R, I. 1975. A general solution for the complete Richards func- tion. Math. Biosci. 27:349-360. 1978a. Time-dependent solutions and efficient parameters for stock-production models. Fish Bull., U.S. 76:377- 388. 1978b. On the restructuring of the Pella-Tomlinson sys- tem. Fish. Bull., U.S. 76:515-521. FOX, W. W., JR 1971. Random variability and parameter estimation for the generalized production model. Fish. Bull., U.S. 69:569-580. 1975. Fitting the generalized stock production model by least-squares and equilibrium approximation. Fish. Bull., U.S. 73:23-37. LEVENBERG, K. 1944. A method for the solution of certain non-linear prob- lems in least squares. Q. Appl. Math. 2:164-168. May, R. M. 1974. Stability and complexity in model ecosystems. 2d ed. Princeton Univ. Press, Princeton, N.J., 265 p. MARQUARDT, D. W. 1963. An algorithm for least-squares estimation of non- linear parameters. J. Soc. Ind. Appl. Math. 11:431-441. Pella, J, J., and P, K. Tomlinson. 1969. A generalized stock production model. Inter-Am. Trop. Tuna Comm., Bull. 13:419-496, RICKER, W. E. 1975. Computation and interpretation of biological statis- tics of fish populations. Fish. Res. Board Can., Bull. 191, 382 p. SCHNUTE, J, 1977. Improved estimates from the Schaefer production model: theoretical considerations. J. Fish. Res. Board Can. 34:583-603. WALTER. G. G. 1975. Graphical methods for estimating parameters in simple models of fisheries. J. Fish. Res. Board Can. 32:2163-2168. 534 SYSTEMATICS AND ZOOGEOGRAPHY OF THE PORCUPINEFISHES (DIODON, DIODONTIDAE, TETRAODONTIFORMES), WITH COMMENTS ON EGG AND LARVAL DEVELOPMENT^ Jeffrey M. Leis^ ABSTRACT The porcupinefish genu.s Diodon is composed of five species: D. hystrix Linnaeus and D. eydouxii Brissout de Bameville are closely related species, each of which has a relatively elongate body, spines on the caudal peduncle, and high dorsal and anal fin ray counts; D. holocanthus Linnaeus and D. Uturosus Shaw form a second species pair, each of which has a round body, no caudal peduncle spines, and moderate dorsal and anal fin ray counts; D. nicthemerus Cuvier is a round-bodied species but differs from D. holocanthus and D. Uturosus in meristic characters and spination. Diodon hystrix. D. holocanthus. and D. eydouxii are distributed circumtropically. The Atlantic population of D. holocanthus has diverged from the Indo-Pacific (including eastern Pacific) popula- tions. Diodon eydouxii is pelagic, and both D. hystrix and D. holocanthus have pelagic juvenile stages. Diodon Uturosus is found in the Indo-West Pacific, and D. nicthemerus is limited to Tasmania and southern Australia. It is not known whether the latter species have pelagic stages. The egg and larval stages of D. hystrix and D. holocanthus (the latter identified by rearing) are similar. The pelagic eggs are 1.6-2.1 mm in diameter and hatch in about 5 days at 25°C. The larvae metamorphose to spiny juveniles at ca. 4 mm in about 3 wk. Both species have pelagic juvenile stages of long duration: D. hystrix remains pelagic to ca. 200 mm standard length, thus providing ample time for dispersal. Eggs and larvae of the other species are unknown. The identities of the species of the genus Diodon have been confused since the time of Linnaeus. The most recent description of a valid "new" species was in 1846, but, unfortunately, time has done little to clarify the situation. Twenty-eight nominal species attributable to Diodon have been described since 1758, and most contemporary au- thors recognize two or three species. However, Le Danois (1959), in the only recent review of the genus as a whole, recognized six species. The present study grew out of attempts to iden- tify juvenile Diodon that resulted from rearing of pelagic eggs taken in Kaneohe Bay, Oahu, Hawaii (Watson and Leis 1974). These juveniles could not be identified using existing keys. While current literature recognized only two species of Diodon in Hawaiian waters, examination of museum speci- mens revealed that three were present there. This discovery, together with the encouragement of J. E. Randall of the Bernice P. Bishop Museum, led to the present study clarifying the identities of all of the species of Diodon and the description of their 'Hawaii Institute of Marine Biology Contribution No. 548. ^Department of Oceanography, and Hawaii Institute of Marine Biology, University of Hawaii, Kaneohe, Hawaii; pres- ent address: Marine Ecological Consultants, 533 Stevens Av- enue, Solana Beach, CA 92075. Manuscript accepted December 1977. FISHERY BULLETIN: VOL. 76, NO. 3, 1978. development. An attempt was made to obtain in- formation on existing type-specimens and this, along with the examination of a large number of specimens, has led to the conclusion that the genus is composed of five species, three of which are dis- tributed circumtropically. Further, it is shown that the present taxonomic confusion is attributa- ble to inadequate original descriptions, reliance on poor characters for differentiation, the close similarity of several of the species, and unusual aspects of the life histories of the species of Diodon . All of the nominal species could be distinguished with some certainty with two exceptions: the type of Diodon echinii.s Rafinesque 1810 could not be located and the original description provides no clue to its identification; the holotype of Trichocy- clus ennaceus Giinther 1870 (BMHN 1976.2.23.1) is a small fish in especially poor condition, giving the appearance of having been obtained from a stomach of some predator, and, while it is cer- tainly a Diodon, more specific identification could not be made. Diodon dussumieri Bibron (see Le Danois 1959, 1961) is a nomen nudum, but exami- nation of the "type" (MNHN 1306) by J. E. Randall of the Bernice P. Bishop Museum indicates that Le Danois was correct in placing D. dussumieri in synonomy with D. holocanthus. 535 FISHERY BULLETIN; VOL. 76, NO. 3 Although basically shorefishes, the diodontids (at least Dwdon and Chilomycterus) are strongly tied to the pelagic environment through pelagic eggs and well-developed pelagic juvenile stages. In Diodon these juveniles remain pelagic for weeks or months ( judging from size) and are often found far from shore. In fact, juvenile Diodon spp. are commonly encountered in the stomachs of such pelagic predators as dolphins (Gibbs and Collette 1959), and one species, D. eydouxii, is apparently pelagic throughout its life cycle. The eggs of diodontids are poorly known. Nichols and Breder (1926) described the unfer- tilized eggs of Chilomycterus schoepfi from New Jersey as demersal, nonadhesive, transparent, and about 1.8 mm in diameter. However, Breder and Clark (1947) suggested that the eggs of C. schoepfi may be normally pelagic. The pelagic eggs of D. holocanthus and D. hystrix from Hawaii were briefly described as Diodon sp. and "diodontid II," respectively, by Wat,son and Leis (1974). Sanzo ( 1930) described the development of what are ap- parently the pelagic eggs of D. hystrix (identified as Crayracion sp.?) from the Red Sea. Wolfsheimer (1957) reported an aquarium spawning of D. holocanthus (identified by him as D. hystrix), but provided little descriptive information on the eggs. The eggs mentioned by Wolfscheimer sank, but did not adhere, to the bottom of the aquarium. They did not develop, .so it is likely that they were not fertilized. Larval and juvenile Diodon are no better known than the eggs. Blanco and Villadolid (1951) illus- trated a juvenile "Diodon bleekeri" but this fish is clearly a juvenile tetraodontid. Many juvenile tet- raodontids have prominent spines, particularly on the ventral surfaces. Fowler (1928) illustrated a juvenile Diodon, identified as D. hystrix, but the figure does not show spines on the caudal peduncle (see below), so this identification is apparently incorrect (assuming the drawing is accurate). No locality or other descriptive data are given by Fowler, so a specific identification cannot be made. Sanzo ( 1930) illustrated two larvae that resulted from rearing of his D. hystrix eggs and a juvenile Diodon captured in a plankton tow. The illustra- tion of this latter fish shows no peduncle spines, but in other respects it resembles D. hystrix. Mito ( 1966) illustrated a larval and a juvenile Diodon, both identified as D. holocanthus. The pigmenta- tion and the relatively small eye .shown in Mito's illustrations more closely resemble the specimens of D. hystrix studied here. At least four species of 536 Diodon occur in Japanese waters, and Mito's specimens could be any of these, because he gives no information as to how the identifications were made. Nishimura (1960) reported on juvenile Diodon cast ashore in the Sea of Japan, but did not provide specific identifications. MATERIALS AND METHODS Measurements and counts are as defined by Hubbs and Lagler (1958:19-28) unless otherwise stated. Measurements routinely were made with needle point dividers to the nearest 0.5 mm. Fish <10 mm and all eggs were measured under a dis- secting microscope to the nearest division of the ocular micrometer ( ±0.02 mm at 50 x, the power normally used). All measurements are from pre- served specimens. Unspecified lengths are in millimeters standard length. Caudal peduncle length was measured from the posterior base of the dorsal fin to the end of the hypural plate. Head width was measured immediately behind the eyes. Body width was measured at the base of the pectoral fin. Width of the eye was taken horizontally across the clear cornea. Measurements are given as proportions of standard length. Dorsal and anal fin ray counts included all rays, branched and unbranched. The last two rays were counted separately because they have separate bases. Pectoral fin ray counts excluded the upper ray. This ray, although well developed in small ( <30 mm) juveniles, is a rudiment in adults and is often not visible because it is embedded. In large specimens, the fin bases are especially fleshy and accurate fin ray counts are difficult to make with- out dissection or radiography. Body measurements are given as range, mean (I), and standard deviation (SD). The sample size for the measurements is given in parentheses at the beginning of the description of each species. Morphometries are included only from individuals >50 mm. Fin ray counts are included for all specimens on which counts could be made (Table 1). In most cases, rays in both pectoral fins were counted. Fin rays were not counted on specimens with fin damage or on specimens that had rays obscured due to the thick bases of the dorsal and anal fins. Radiography was tried unsuccessfully to obtain vertebral counts: the dermal spines and their bases obscured the vertebrae, and made ac- curate counts impossible. The vertebral counts LEIS: SYSTEMATICS AND ZOOGEOGRAPHY OK THE PORCUPINEFISHES Table l. — Fin ray counts of Djodon species. Dorsal fin rays 12 13 14 15 16 17 18 X D eydouxii Atlantic 3 1 17.25 Indo-Paciflc 4 25 6 17.06 D hystnx Atlantic 2 7 1 15.90 Indo-Pacific 2 20 8 15.20 D holocanthus Atlantic 22 4 14.15 Indo-Paclfic 9 39 7 13.96 0. liturosus 1 12 15 15.50 D nicthemerus 1 10 12.91 Anal fin rays 12 13 14 15 16 17 18 X D eydouxii Atlantic 3 1 17.25 Indo-Pacific 2 22 11 17.26 D. hystnx Atlantic 3 7 15.70 Indo-Pacific 1 15 6 15.23 D holocanthus Atlantic 9 18 13.67 Indo-Pacific 26 27 1 13.54 D liturosus 10 17 1 14.68 D nicthemerus 3 5 2 12.80 Pectoral fin rays 19 20 21 22 23 24 25 X D eydouxii Atlantic 5 3 21.38 Indo-Pacific 7 58 17 20.12 D. hystnx Atlantic 1 10 10 1 23.50 Indo-Pacific 2 16 34 6 22.76 D holocanthus Atlantic 7 31 20 22.22 Indo-Pacific 1 17 57 27 5 22.17 D liturosus 2 11 24 17 2 23.11 D. nicthemerus 2 9 9 20.35 given for D. holocanthus were made on cleared and stained material. The dermal spines require special terminology and measurements, as given below. Measure- ments, except for shaft length, were taken on dis- sected spines (Figure 1). The spine shaft is that portion bearing the pointed tip, but excluding the shaft extension. The length of the spine (= shaft length) was taken from the lower portion of the lateral arm to the tip of the shaft. The starting point for this measure- ment can be found most easily by probing around the base of the spine. The shaft extension is the portion of the shaft extending past the lateral arms of the base, and its length was measured from the lower portion of the lateral arm to the tip of the extension. The lateral arms of the base are the subdermal portions of the spine upon which the spine pivots during erection. The length of the spine base was the straight line distance from tip to tip of the lateral arms. The frontal spines are those of the anteriormost row on the head between the eyes. The pectoral axil spines are the spines immediately posterior to the base of the pectoral fin. Figure l. — Typical Diodon body spine: (A) spine (or shaft) length, (B) length of the shaft extension, (C) length of the spine base. The tip of the spine shaft points caudad. The number of spines in a longitudinal row over the dorsum from the snout to the dorsal fin base (S-D spines) and the spines in a longitudinal row over the ventrum from the lower jaw to the anus (S-A spines) were counted. These rows of spines are irregular and difficult to follow, so the counts should be considered approximate. With practice, repeated counts of ±1 can be achieved consis- tently. The numbers of spines between pectoral fins, both over the dorsum (P-D-P spines) and ven- trum (P-V-P spines), were also counted, but these counts are even less reproducible than the lon- gitudinal counts. Repeated reference is made to the spines on the caudal peduncle. In some species the only spines in the region of the caudal peduncle are some rather large spines associated with the dorsal and anal fin bases. Although these spines extend over the peduncle, their subdermal bases (lateral arms and shaft extension) are at least partially anterior to a line between the base of the posteriormost rays of the dorsal and anal fins, and they are considered not to be on the peduncle. In other species, there are relatively small spines which are wholly pos- terior to the line defined above on the dorsal and dorsolateral surfaces of the peduncle; these spines are considered to be on the peduncle (Figure 2). Larvae were obtained from plankton samples (field specimens) and rearing experiments using eggs from plankton tows (reared specimens). All eggs and larvae were captured around the Hawaiian island of Oahu. Rearing took place in 537 FISHERY BULLETIN: VOL. 76, NO. 3 B Figure 2. — Semidiagrammatic lateral view of the caudal peduncle and posteriormost spines of (A) a slender-bodied, long peduncled species (Diodon eydouxii) and (Bl a round-bodied, short peduncled species (D. holocanthus) . the laboratory under ambient temperature (ca. 25°C) and a variety of conditions. Generally, the eggs were hatched in unaerated 4-1 beakers filled with seawater from the collection area. Hatched larvae were transferred to 10-20 1 containers and provided with overhead illumination. The con- tainers were wrapped in black plastic. Wild zoo- plankton (ca. 60-200 ixm) from a plankton pump were added on alternate days; this was later supplemented with Artemia nauplii. Water was changed twice a week and .specimens were re- moved periodically for preservation. Many rear- ing attempts were made, but since fewer than 20 eggs usually were available per attempt, few of the attempts were successful. Some larvae were cleared and stained using the KOH-alizarin red method of Hollister (1934). Measurements and definitions of stages generally follow those of Leis (1977), unless otherwise noted. All drawings of eggs and larvae were made with the aid of a camera lucida. The institutions housing the examined speci- mens are as follows: Academy of Natural Sci- ences of Philadelphia (ANSP); Australian Museum, Sydney (AMS); Bernice P. Bishop Museum, Honolulu (BPBM); British Museum (Natural History) (BMNH); California Academy of Sciences (CAS); Gulf Coast Research Labora- tory and Museum (GCRL); George Vanderbilt Foundation (GVF), deposited in CAS; Hawaii In- stitute of Marine Biology (HIMB); Los Angeles County Mu.seum of Natural History (LACM); Museum National d'Histoire Naturelle, Paris (MNHN); National Marine Fisheries Service, Honolulu, Hawaii (NMFS H), La Jolla, Calif. (NMFS LJ), and Miami, Fla. (NMFS M); Naturhistorisches Museum, Vienna (NMV); J. L. B. Smith Institute of Ichthyology at Rhodes Uni- versity, South Africa (RUSI); Scripps Institution of Oceanography (SIO); Tulane University (TU); National Museum of Natural History, Smithso- nian Institution (USNM); University of Arizona (UA). A catalog number is given when available; many GVF specimens were uncataloged and therefore the register or station number is given. The synonymies include all known original usage of names. In addition, references of sys- tematic or zoogeographic interest are included. If the identification of a nominal species is question- able, it is preceded by a question mark (?). Pre- Linnaean literature is cited in the text if appro- priate, but is omitted from the synonymies. GENUS DIODON LINNAEUS Diodon Linnaeus 1758:334, after Artedi 1738. Type-species D. hystrix Linnaeus by subsequent designation of International Com- mission on Zoological Nomenclature, opinion 77. Paradiodon Bleeker 1865:49. Type-species D. hystrix Linnaeus by original designation. Trichodiodon Bleeker 1865:49. Type-species D. pilosus Mitchill by original designation. Trichocyclus Giinther 1870:316. Type-species T. erinaceus Giinther by monotypy. Diagnosis. — Body rotund, width 0.25-0.54, depth varies greatly depending on degree of inflation. Eyes large, 0.05-0.17. Swim bladder bilobed. Teeth in each jaw fused into a single beaklike unit without a median suture dividing upper or lower jaws into right and left halves. Gill opening a short, vertical slit immediatelv anterior to the 538 LEIS: SYSTEMATICS AND ZOOGEOGRAPHY OF THE POKCT'PINEKISHKS pectoral fin base. Approximately 20 vertebrae. Dorsal and anal fins usually rounded, set far back on body, with 12-18 rays. Caudal rounded, with 9 rays (there are no secondary rays). Pectoral fin slightly emarginate, with 19-25 rays, the upper- most ray (not counted) greatly reduced in adults. No pelvic fins. Body covered with long spines, all but a few (around the gill opening, dorsal fin base, and caudal peduncle! of which are erectile. Erec- tile spines consisting of a long pointed shaft, tw'o subdermal lateral bases lying in nearly the same plane as the shaft, and usually a shaft extension which is shorter than the shaft. The shaft exten- sion may be greatly reduced. Nasal organs consist- ing of a short tentacle with a pair of lateral open- ings near the tip. In larger individuals of some species the tissue closing the end of the tentacle may be absent, giving rise to a bifid nasal tentacle without nostrils. Both species whose reproductive habits are known (D. hystnx and D. holocanthus) spawn pelagic spherical eggs of 1.6-2.1 mm in diameter. Ren2arks. — Only Bleeker's (1865) proposal of Paradiodon for the species here considered to be- long in Diodon (because of page priority, he be- lieved Diodon should apply to those species usu- ally referred to Chilomycterus) has disturbed the stability of the usage of the name Diodon. Trichodiodon and Trichocyclus are names applied to juvenile stages oi Diodon. Although subgeneric status seems unwar- ranted, Diodon can be broken into two groups on the basis of body width, caudal peduncle length, and squamation. The species of the slender-bodied group, D. eydouxii and D. hystrix, have a rather narrow body (Figure 3, Table 2), long caudal peduncle (Figure 3, Table 2), and several small spines in the dorsal and dorsolateral surfaces of the peduncle. The species of the round-bodied group, D. holocanthus, D. liturosus, and D. nicth- emerus, have a wider body, shorter caudal pedun- cle (Figure 3), and lack spines on the caudal peduncle (although there are strong spines, pro- jecting over the peduncle, at the base of the dorsal and anal fins). Upon inflation, the dorsal and anal fins are engulfed by the expanding skin. In the round-bodied group, the caudal peduncle and fin are also largely obscured in inflated specimens and the large spines mentioned above provide added protection. In the slender-bodied group, the peduncle remains largely uncovered and is pro- tected only by the relatively small spines on its upper surfaces. Diodon nivthemerus, although clearly a member of the round-bodied group, ap- pears to have undergone a reduction in spine number and base size, and is thus separable from D. holocanthus and D. liturosus. 70-1 E 60- E -LITUROSUS -EYDOUXII -HYSTRIX -NICTHEMERUS -HOLOCANTHUS X 100 150 200 Figure 3. — Plotted regression lines of (top) caudal peduncle length vs. standard length and (bottom) body width vs. stan- dard length for the five species of Diodon. Lines plotted only- over size range of specimens examined. The line with arrow- head for D. hystrix extends to 571 mm SL. Regression data in Table 2. Table 2. — Regression equations for caudal peduncle length (PL) and body width (BW) vs. standard length (SL) in the five species of Diodon (see also Figure 3). Species Regression equation r 'slope df D, hystrix PL = 0.189 SL - 2.79 0.97 21.16 31 D eydouxii PL = 0.226 SL - 4.79 0.95 17.44 33 D. liturosus PL = 0.159 SL- 2.40 096 17.88 26 D. nicthemerus PL = 0.151 SL- 1.38 0.94 785 8 D- holocanthus PL = 0.152SL+ 4.69 0.90 1586 61 D. hystrix BW = 0.338 SL+ 6.01 097 23.41 31 D. eydouxii BW = 0.262 SL + 5 27 0.88 10.45 33 D. liturosus BW = 0.333 SL + 10.29 0.90 10.58 26 D. nicthemerus BW =0 313SL ^ 13.11 0.93 6.44 6 D holocanthus BW = 0.368 SL ^ 6.29 0.96 2575 62 539 FISHERY BULLETIN; VOL. 76. NO. 3 KEY TO THE SPECIES OF THE GENUS DIODON la Two or more small spines wholly on the dorsal or dorsolateral surfaces of the caudal peduncle (Figure 2A); color pattern of adults dominated by small (smaller than eye) spots; at least D, P, and C fins of adults with dark spots 2 lb No spines w^holly on the caudal peduncle (Figure 2B); color pattern of adults dominated by large dorsal and lateral bars or blotches; fins of adults without spots except in some cases at base 3 2a P 19-22. both D and A 16-18; D and A of adults falcate; S-A spines^U; head width less than 30'y SL D. eydouxii (circumtropical, oceanic) 2b P 22-25 (rarely 21), D 14-17, A 14-16; D and A of adults rounded; S-A spines&14; head width greater than 307^ SL D. hystrix (circumtropical, shore fish but juveniles pelagic) 3a No small, fixed, tribase spine immediately above the gill opening; no small, flat spines on the anterior border of the depression surrounding the gill opening (Figure 4); S-A spines ^11; adult color pattern dominated by four large lateral bars, dorsum uniformly dark D. nicthemerus (Australia) Figure 4. — Head of Dwdon nicthemerus (AMS 1.16990-004) showing arrangement of spines in the region of the gill opening. Note that spines anterior to gill opening are not flattened. Also note tubular nostril. 3b One or two small, fixed, tribase spines above the gill opening; three or four small, flat spines forming the anterior border of the depression surrounding the gill opening ( Figure 5); S-A spines S3l2; adult color pattern dominated by several dorsal blotches 540 LEIS: SYSTEMATICS AND ZOOGEOGRAPHY OF THE PORCUPINEFISHES Figure 5.— Head of Dwdon liturosus (CAS 30967) showing arrangement of spines in the re- gion of the gill opening. Note that spines on an- terior border of opening are short and flattened. Also note the small dow-nward-pointing spine below the anterior border of the eve. Longest Frontal Spine/ SL — I © 1 D. liturosus 1 ® 1 D. hystrix 1 0 1 D. eydouxii At! D. holocanthus ^ ^ — i la-Pac. D. holocanthus 1 1© — D nicthemerus ^ H FlGURE 6. — Ratio of frontal spine length to standard length for the five species of Diodon. Line indicates range, circle and bar indicate mean, and vertical bars alone denote ±1 SD. Note dif- ference of spine length between Atlantic and Indo-Pacific specimens of Z). holocanthus . Number of specimens given in de- scription for each species. I 1 1 r .04 .08 .12 -! 1 — .16 .20 — I — .24 4a Frontal spines 0.04-0.10 (Figure 6i, much shorter than pectoral axil spines; 17-22 S-A spines; a small downward-pointing spine below the anterior margin of the eye present; dorsal blotches with a distinct light border; a dark gular band from eye to eye under the lower jaw D. liturosus (Indo-Pacific) 4b Frontal spines 0.13-0.28 (Figure 6), slightly shorter to much longer than pectoral axil spines; 12-15 S-A spines; a small downward-pointing spine below the anterior margin of the eye absent (Indo-Pacific specimens) or present (most Atlantic specimens); dorsal blotches without a distinct light border; no gular band D. holocanthus (circumtropical) 541 FISHERY BULLETIN VOL 76. NO 3 DIODON EYDOUXII BRISSOUT DE BARNEVILLE Pelagic Porcupinefish (Figure 7) Diodon cyduuxii Brissout de Barneville 1846:142 (eastern Pacific); Troschel 1847:364; Dumeril 1855:278. Diodon melanopsis Kaup 1855:228 (no locality given). Diodon spinosissimu.s (not of Cuvier): Giinther 1870:307 (Cape of Good Hope, Siam). Diagnosis. — A slender-bodied Diodon , head width 0.25-0.30, peduncle length 0.16-0.22. Caudal peduncle armed dorsally with short spines. Body spines long and slender, moderate in number, S-D spines 13-17, S-A spines 10-14. Pectoral axil spines 0. 11-0.16, usually longer than longest fron- tal spines. A short, fixed tribase spine im- mediately above gill opening. D 16-18, A 16-18, P 19-22. Nasal tentacle with a pair of lateral open- ings. No barbels or fleshy tentacles. Dorsal and anal fins falcate (rounded in juveniles). Color pat- tern dominated by small (ca. = to pupil) dark spots dorsally and laterally. These often as- sociated with the spine axils. A dark gular band starting from below the eyes and continuing under the chin, usually with a branch extending dorsally between eye and gill opening. Descn'p^/o/;.— (35 specimens) D 16-18, A 16-18, the first two or three rays unbranched; P 19-22. Head width 0.25-0.30 (.v = 0.27; SD - 0.01), body width 0.25-0.35 (.V = 0.30; SD = 0.02), peduncle length 0.16-0.22 (.V = 0.19; SD = 0.02), eye 0.05-0.10 (X = 0.08; SD = 0.01). Dorsal and anal fins falcate, not rounded. Nasal tentacles with a pair of lateral openings. S-D spines 13-17, S-A spines 10-14, about 12 spine rows over the dorsum between pectoral fin bases, about 21 spine rows over the ventrum be- tween pectoral fin bases. Four or five frontal spines. Longest frontal spine 0.07-0.15 ix = 0.11; SD =0.05), pectoral axil spines 0.11-0.16 ix = 0.14; SD = 0.01). Pectoral axil spines usually the longest on the body, 0.61-1.03 ix = 0.78; SD = 0. 1 1 ) in frontal spines. Spines long and slen- der. Frontal, middorsal, and ventral spines of about the same length. Pectoral axil spines and those dorsolateral spines from over eye to over pectoral fin among the longest on body (ca. 0.8 in frontal spines). Spines on caudal peduncle short (ca. 1.5 in frontal spines) and fixed due to a rather long shaft extension (ca. 2 in shaft). Shaft exten- sion on other spines reduced, never more than 159f of the shaft length. Subdermal bases moderate in extent, and, except for spines around fin bases and caudal peduncle, always shorter than shaft. No spines markedly reduced other than on caudal peduncle; the latter spines generally arranged in one or two bilateral pairs along the dorsolateral edge of the peduncle. Approximately 40^^ (14 of 36) of the specimens examined also possess a single dorsomedial spine on the caudal peduncle. A short, fixed tribase spine immediately above the gill opening and a second slightly posterior to it above the pectoral base. Three short, flat spines Figure l.— Diodon eydouxii, 128 mm SL, central Pacific (NMFS H CHG 55-71) 542 LEIS: SYSTEMATICS AND ZOOGEOGRAPHY OF THK PORCHPINEFISHES with broad lateral bases form the anterior border of the gill opening. No spines on the snout. No barbels or fleshy tentacles. Dorsally the ground color is light grey to brown grading to white ventrally. Dorsal and lateral sur- faces marked with dark ovoid spots ( 1.5 mm) and numerous oil droplets serves to distinguish the eggs of D. holocanthus from those of all other pelagic eggs except those of other tetraodontiform species. The eggs of the molid Ramania laevis have been described by Leis ( 1977). Ramania laevis eggs may be distin- guished from D. holocanthus eggs by the former's smaller size (1.4-1.65 mm) and by the extensive pigment which develops on the ven- tral surface of the yolk sac of R. laevis in the middle stage. Hawaiian ostraciid eggs (Ostracion and Lac- toria) may be distinguished by their slightly oblong shape, fewer oil droplets ( <10), but most reliably by a patch of bumps on the chorion surrounding the micropyle. This "rough patch" is easilv overlooked. FISHERY BULLETIN; VOL. 76. NO, 3 Diocion hystrix eggs are the only other Diodon eggs known ( see section on D. hystnx). They can be distinguished from those of D. holocanthus by their larger size ( >1.9 mm), greater number of oil droplets ( >30), and the orange (rather than red) pigment. Larval Development: Fifteen reared and 12 field-collected larvae in good enough condition for descriptive purposes were available. Mor- phometric data are summarized in Table 4. The newly hatched larva has well-developed, apparently functional eyes, jaws, and gas blad- der (Figure 22). The pectoral fins are quite large, although no rays are formed. The larvae are 1.9-2.1 mm SL at hatching and the body is rotund. Development in reared larvae is slow. Dorsal and anal fin anlagen form by day 10 (2.4 mm, Figure 22); the olfactory pit also forms by this time and the eyes have become proportion- ally larger. The oldest reared larva available was 16 days old, but it was smaller than the Table 4. — Morphometric and meristic data for larval and juvenile Diodon holocanthus (measurements in mm). ? indicates individuals of unknown age, from plankton samples; X indicates damaged. Age (days) Notochord Snout Fin ray count;:; of reared or standard to anus Width Head Head Mouth — fish length length of eye length width width D A P Larvae 1 2.0 1-5 03 09 1,1 0,5 0 0 0 1 2.1 15 03 0,8 1,2 05 0 0 0 1 2.0 15 03 0,8 12 04 0 0 0 1 2.0 1.4 03 08 1 2 04 0 0 0 1 1.9 1,4 03 07 12 05 0 0 0 ? 1.9 1,5 03 0,8 1,1 0,6 0 0 0 ? 1.9 1 4 03 0,9 — 0,4 0 0 0 ? 1.« 1 6 04 09 1,2 06 0 0 0 ? 1.9 1 0 03 06 — — 0 0 0 7 2.0 1,6 04 08 1,1 06 0 0 0 ? 2.0 1 6 03 07 1 1 03 0 0 0 5 1.8 14 03 08 12 0,6 0 0 0 6 1.8 14 04 0,8 1 0 06 0 0 0 7 1.9 14 03 0.8 1,0 05 0 0 0 8 2.1 1 4 04 0,9 1 1 06 0 0 0 8 2.2 1 7 0 5 09 14 07 0 0 0 8 2.1 15 04 0,8 1,1 06 0 0 0 9 2.0 1 4 0,4 0,8 10 0,5 0 0 0 ? 2.2 1 5 03 0.8 1,2 0,5 0 0 0 ? 2.3 2,0 05 0,8 1,5 06 0 0 0 ? 2.3 1,5 03 0,8 1,1 0,5 0 0 0 10 2.4 20 05 0,9 1 5 08 0 0 0 10 2.2 1 7 05 09 13 06 0 0 0 ? 2.5 2 1 05 07 — 04 0 0 0 ? 2.6 2,0 0 5 0,7 08 04 0 0 0 ? 2.7 22 05 07 — — 0 0 0 16 11.9 . 1,5 0,5 10 1,2 0,7 0 0 0 Juveniles: ? 3.8 34 08 1,9 2,6 1,0 22 25 4.8 40 1 0 23 35 1,9 14 14 23 ? 5.5 4,8 1,1 2,8 3,3 16 — — — ? 6.0 5,3 12 3,3 3,5 18 — — — 33 6.7 5,2 1,4 29 43 1,7 15 . 21 ? 7.2 5,5 1 5 3,4 37 1 7 14 14 23 ca 30 8.1 67 1 8 39 4,8 1,8 14 14 21 7 11.0 9,0 20 55 63 2,7 — — — ? 14.1 109 2,7 63 7,6 32 15 14 23 'Fish in emaciated condition. 560 LEIS: SYSTEMATICS AND ZOOGEOGRAPHY OF THE PORCrPlNEFISHES Figure 22. — Reared larvae ofDiodon holocanthus: (top) newly hatched larva 2.0 mm, (middle) 10-day-old larva 2.4 mm, and (bottom) dorsal view of 10-day-old larva with pigment omitted. day-10 larvae and appeared emaciated. There are incipient fin rays and bases visible in the fins of the 16-day-old larva, but it otherwise is not obviously advanced over the 10-day-old specimen. There is no sign of development of the caudal fin complex. The largest larva avail- able is a 2.7-mm field-collected specimen which is no more advanced than the day- 16 larva. The dermal sac is inflated in young larvae (Figure 22), but the subdermal space is virtually gone by day 10 (Figure 22). The larvae are more or less uniformly pig- mented with scattered melanophores on the dorsal surfaces at all stages. The pigment spreads laterally, but there is little below the level of the pectoral fin and the ventral surfaces remain devoid of melanophores until metamor- phosis. The newly hatched larvae have no melanophores posterior to the anus (Figure 22), but by day 10 postanal pigment has spread to the middle of the dorsal fin anlage. In life, the newly hatched larva is covered with widely scattered red chromatophores on the dermal sac and fins. The red pigment persists through the larval stage and on about day 2 it is supplemented by a yellow background pigment covering all the body surfaces (not the dermal sac), but being most obvious ventrally due to a lack of melanophores there. A 2.0-mm field-collected specimen was cleared and stained. The only ossified struc- tures were the cleithrum, coracoid, and six branchiostegals. Juvenile Development: Metamorphosis appar- ently occurs at ca. 3 mm at an age of about 3 wk. The smallest juvenile available is 3.8 mm and resembles Mito's (1966) illustration of a 3.7-mm juvenile except that Mito's fish had smaller eyes. The caudal, dorsal, anal, and pec- toral fins are all formed as are the teeth, and the body is covered with small spines. The spines do not appear to be erectile, but the fish is capable of inflation. The spines are covered with a sheathlike tissue. They elongate rapidly with growth and by 4.8 mm SL (Figure 23) they are obviously erectile. The nostrils are formed in the 3.8-mm flsh, although the nasal tentacle with two lateral openings is not formed until 4.8 mm SL, and in fish as large as 6.0 mm. it may be open at the ends. The 4.8-mm fish is in all respects a miniature adult with all external structures formed and functional. External changes to the adult stage involve only changes in proportion; the spines in particular elongate, the body becomes less rotund and the eye rela- tively smaller. Morphometric and meristic data are summarized in Table 3. A 33-day-old juvenile of 6.7 mm was cleared and stained. The vertebral column and skull are incompletely ossified but all other structures are ossified. The vertebral formula is 12 + 9 = 21 and the vertebral column is strongly arched. There are 1 1 dorsal and 1 1 anal pterygiophores which are associated with ver- tebrae 12-16 and 13-17, respectively. At metamorphosis, pigment changes radi- cally. The background color in live material is still predominantly yellow with scattered red chromatophores but this does not persist. Dor- sally, the melanophores are scattered fairly uni- formly, with a concentration at the pectoral base and very little pigment on the caudal peduncle. 561 FISHERY BULLETIN. VOL. 76, NO. 3 Ventrally, however, a number of distinct spots have formed that cover the belly (Figure 23). The spots (pelagic spotting) are at first close together but become less numerous and propor- tionately larger, aligning in rows with growth (Figure 19). Dorsal spotting (always more dif- fuse than ventral spotting) begins to form at around 10 mm and the characteristic dorsal blotch pattern is generally visible by 30 mm, although in pelagic specimens the contrast with the background color is not great. The pelagic spotting is retained in all pelagic individuals examined (to 86 mm) and in some specimens collected inshore. The fins remain unpigmented except for a few melanophores along the fin rays of the dorsal fin. Identification of Larvae and Juveniles: Diodon- tid larvae are likely to be confused only with the rotund, heavily pigmented, sac enclosed ceratioid larvae and other tetraodontiform lar- vae. Reference to Bertelsen's (1 95 1 ) work should allow ceratioid larvae to be distinguished as such. Rotund tetraodontiform larvae may be distinguished from diodontid larvae as follows: molids by their body spination and early form- ing pectoral rays; ostraciids by their pigmenta- tion and early forming pectoral rays; tetraodon- tids by their relatively more elongate body shape and early forming fin rays. Diodon larvae are heavily pigmented only on dorsal surfaces, do not develop fin rays until near or at metamorphosis, have very wide heads and bodies ( >body depth), and have very wide mouths. The larvae of D. hnlocanthus can be distin- guished from the putative D. hystrix larvae, the only other larval diodontid known, by the less Figure 23.— Reared juvenile of Diodon holocanthus, 4.8 mm SL, 25 days old. Note pelagic spotting. Hawaiian material. well-developed condition at hatching of the lat- ter (see section on D. hystrix). In addition, D. hystrix larvae are predominantly orange upon hatching while those of D. holocanthus are yel- low. Melanophores of D. holocanthus do not extend onto the postanal myomeres past the middle of the dorsal and anal fin anlagen; the postanal myomeres of D. hystrix are moderately pigmented. Lastly, the eyes of Z). hystrix larvae are smaller than those of D. holocanthus larvae (Tables 2,3). Once the spines form, the lack of caudal peduncle spination, fin ray counts and spine placement serve to distinguish D. holocanthus from all other Diodon species (see Key). The duration of the pelagic stage is unknown, but judging from reared specimens, metamor- phosis occurs about 3 wk after hatching at about 4 mm SL. The largest individual captured pelag- ically was 86 mm while the smallest captured inshore was 60 mm. A certain amount of plastic- ity in the duration of the pelagic stage is indi- cated, but its length clearly must be measured in terms of months. No special adaptations for pelagic life are evident in these juvenile stages except, perhaps, in color. In the tetraodontiform fishes (except the molids) the larval stage is short and relatively unspecialized, while a rela- tively unmodified pelagic juvenile stage may be quite long (see Remarks under D. eydouxii). This strategy (for dispersal?) is in marked con- trast to that in many advanced perciform shorefishes (e.g., Acanthuridae, Chaetodon- tidae) where bizarrely modified and long-lived larval and pelagic prejuvenile stages are de- veloped which subsequently undergo marked (and rapid) metamorphosis upon becoming benthic. 562 LEIS: SYSTEMATICS AND ZOOGEOGRAPHY OF THE PORCUPINEFISHES Diodon holocanthus eggs and larvae have been found in Hawaiian waters from February through September, with an apparent peak in abundance in May-June, although they are never common. Larvae usually occurred singly in plankton tows ( volume filtered 200- 1 ,000 m-'^). Although as many as 30 eggs 1.000 m'^ have been taken, 1-5 eggs/1,000 m^ were more usual, and most tows contained none. Eggs were usu- ally found close to shore, but larvae rarely were found closer than 1 km from shore (pers. ob- serv.). Holotype. — No holotype or type-series is known to exist. Linnaeus based his description on that of Artedi (1738). Distribution . — Diodon holocanthus is circumtrop- ical in distribution, but is seemingly absent in the southwest and central Pacific east of the andesite line (the separation of continental from oceanic rocks, Figure 14). However, it reappears in Hawaii, Pitcairn, and Easter Islands. Cuvier's holotype of D. quadrimaculatus was allegedly col- lected by Peron in Tahiti (see Le Danois 1961). Inasmuch as it is known that much of the locality data accompanying Peron's specimens are incor- rect (associated with a shipwreck, see Whitley 1931:25) this record is questionable. There is evi- dence of divergence of the Atlantic population! s) from those of the Indo-Pacific (see Remarks). Remarks. — I follow the spelling holocanthus (rather than holocanthus of many authors) which was used consistently by both Linnaeus and Ar- tedi (see also Bailey et al. 1970), and is thus not considered to be a misprint as maintained by Jor- dan and Evermann (1891). Linnaeus' description is brief; the only useful information being the statement that the spines are terete and ex- tremely long on the head and nape. However, this can apply only to D. nicthemerus or D. holocan- thus. Assuming that "Habitat in India" means India as understood today, and not the entire Indo-Pacific, D. nicthemerus is eliminated. How- ever, even if "Habitat in India" means the entire Indo-Pacific, it is unlikely that specimens of D. nicthemerus, a species apparently confined to southern Australia, could have reached Artedi by 1738. In any case, subsequent usage and stability demand that the name D. holocanthus apply to the species described above. Diodon pilosus is synonymized with D. holocan- thus on the basis of Mitchill's observation that no spines were present between the dorsal and caudal fins of his small (ca. 38 mm) New York specimen. Diodon holocanthus is the only Atlantic species that lacks peduncle spines. Mitchill distinguished D. pilosus on the basis of its flxible spines, but this is the usual condition in small specimens. No holotype is known to exist. Cuvier's types are extant. Information and photographs of these specimens (catalog numbers and other information are given by Le Danois 1961) provided by M. L. Bauchot (pers. commun., MNHN, 20 May 1975) clearly establish D. novemmaculatus, sex?naculatus . quadrimac- ulatus, and multimaculatus (all of Cuvier) as junior synonyms of D. holocanthus . Inasmuch as Cuvier's (1818) descriptions are relatively clear, only his D. novemmaculatus requires comment. The holotype of D. novemmaculatus (MNHN A. 9928, 107 mm) is D. holocanthus, apparently from the Atlantic (no locality data are available for this specimen). A spine is present below the anterior margin of the eye and the eye bar is dis- continuous over the interorbital. Unfortunately, Cuvier's figure resembles D. liturosus as much as D. holocanthus (the figure shows the frontal spines shorter than they actually are). This probably led Bleeker (1865) to apply the name D. novem- maculatus to D. liturosus. Diodon maculifer Kaup ( 1855) is included here with some questions. Kaup's description is of little help, and no type material can be found in the British Museum where it would be expected to reside. The holotype may have been part of Kaup's lost personal collection (A. C. Wheeler, pers. com- mun.). Examination of one of the South African (Kaup's type-locality) specimens of "Diodon maculifer" listed by Giinther (1870) (BMNH 1845.7.3.103, 100 mm, loaned by A. C. Wheeler) reveals it to be an inflated, dried D. holocanthus. In this specimen, inflation is so great ( an artifact of stuffing and drying?) that the subdermal spine bases project through the dried skin. Thus, the base of the spines appear to be expanded and transversely compressed. The only characteristic feature of Kaup's description is the compressed nature of the spines, and it seems likely that his description was based on a dried, inflated D. holocanthus. Steindachner's Atopomycterus bocagei can be placed in the synonomy of D. holocanthus on the basis of information on the holotype (NMV 63848) 563 FISHERY BULLETIN; VOL. 76, NO. 3 provided by P. Kahsbauer (pers. commun., NMV, 1975). Steindachner's (1866) description is essen- tially correct and unquestionably refers to D. holocanthus. The placement of this specimen in Atopumycterus was apparently based on the split nasal tentacle (see section on D. nicthemerus). A single split nasal tentacle was present on only 3 of the more than 100 specimens of D. holocanthus examined, so this condition is rare but not unpre- cedented. Both D. liturosus Shaw and D. maculatiis Lacepede (the Latinized version of Le Diodon Tachete) have been incorrectly applied to D. holocanthus by various authors (see section on D. liturosus). For about the past 50 yr the chief sources of confusion on the identity of D. holocanthus have been confusion with D. histrix by some (mostly American) authors and the lumping of D. liturosus under D. holocanthus by nearly all authors. The latter problem is discussed under D. liturosus. The confusion between D. hystnx and D. holocanthus stems primarily from three sources. Many authors (e.g., Gosline and Brock 1960) have conjectured that D. holocanthus is the young of D. hystrix because the former does not reach a large size, and few, if any, small specimens of the latter were available. However, as discussed under D. hystrix, this species is pelagic to ca. 200 mm and is thus unavailable to inshore collecting. Inasmuch as D. holocanthus does not commonly exceed 200 mm, the confusion was perhaps understandable. Second, many early descriptions are poor and keys often rely solely on the size of frontal spines relative to the pectoral axil spines to distinguish the two species. Especially in Atlantic specimens of D. holocanthus, the frontal spines are likely to be approximately the same size or even shorter than the pectoral axil spines. Finally, as noted by Clark andGohar( 1953) (see alsoBagniset al. 1972:225), living/), hystrix ohen display a dorsal blotch pattern not unlike that of D. holocanthus. I have not observed this color pat- tern in preserved D. hystrix. The apparent divergence of the Atlantic and Indo-Pacific populations of D. holocanthus men- tioned above is of interest. At present, since D. holocanthus is apparently absent from the Red Sea and the Mediterranean, gene flow could occur only around southern Africa. Evidence that this is ap- parently not happening comes from the Indian Ocean specimens which lack a snout spine and have very long frontal spines in contrast to the Atlantic specimens (Table 5). In addition, Poll's ( 1959) description (as D. hystrix ) of a west African specimen is typical of the specimens from the western Atlantic examined by me. The apparent increase in frontal spine length from the Atlantic to the Pacific to the Indian Oceans is curious. Based on studies of other groups (Ekman 1967) affinities might be expected between the Atlantic and eastern Pacific populations, but no extension to Hawaii and Easter and Pitcairn Islands would be expected. The lack of the snout spine in all but the Atlantic population and one Hawaiian speci- men may indicate that the Atlantic population is distinct. Fin ray counts are of little help in resolv- ing this question. Because all the characters which appear to differ between the Atlantic specimens and those from other areas are rather variable (although some are significantly differ- ent in a statistical sense), I choose not to distin- guish formally the populations nomenclaturally at the subspecific level. If future study shows this split to be desirable, the proper name for the At- lantic specimens would be Diodon holocanthus pilosus Mitchill. Le Danois ( 1954 ) reported sexual dimorphism in D. holocanthus, but her illustration of a female D. holocanthus (p. 2355:fig. 3) appears to be D. liturosus. Material examined. — 141 specimens, 5-289 mm. EASTERN PACIFIC: NMFS LJ (1;18.5) 18°56'N, 104 = 10 "W; NMFS LJ D31-133.25 ( 1:64. 5l 26°04.5'N, 112°48.0'W; NMFS LJ TO-5801 ( 1:85.5) 5°29.5'N, 77°57'W; NMFS LJ ( 1:73.5) "350 mi. west of Costa Rica"; NMFS LJ B-5011 157.40 (2:41-41.5) 21°32.5'N, 111°14.5'W; UA 66-39-18 (1:242) San Agustin Bay, Sonora, Mexico; UA 69-35-25 ( 1:245) Guaymas, Sonora, Mexico; Table 5. — Comparison of selected characters of Diodon holocanthus from five regions (see also Figure 6). n = number of individuals examined for snout spine. No. with Frontal Fin rays (X) Area n snout spine spine/SL D P Interorbital bar Atlantic 58 52 0.146 14 15 22 15 Usually discontinuous E. Pacific 11 0 0.154 1380 22 10 Usually continuous Hawaii. Pitcairn. and Easter Is. 24 1 0.174 14 44 22.57 Usually continuous W Pacific 29 0 0.205 14.11 22.17 Usually continuous Indian 6 0 0.200 13.80 21.92 Usually continuous 564 LEIS: SYSTEMATICS AND ZOUGECHIKAl'HV OF THE Pt)KtlPINEFlSHES UA 7 1-63-8 (1 : 1 45 ) Puerto Vallarta, Jalisco, Mexico; UA 7 1-65-9 (1:126) IslaJaltemba, Jalisco, Mexico; SIO 59-373 (l:ca. 200) La Jolla, Calif.; SIO 63-82 (l:ca. 90) Cape Marco, Colum- bia. HAWAIIAN IS.: HIMB (3:135-289), HIMB 67-58 (1:67) Kaneohe Bay, Oahu; HIMB (1:181) Punaluu, Oahu; BPBM 10635 ( 1:63), BPBM 6977 (1:167) Diamond Head, Oahu; BPBM 5124 (1:129) French Frigate Shoals; NMFS H TC32-6,9,11,14 (6:12.5-30) 21°22'N, 158°14'W; NMFS H TC32-23 (1:14) 2r00'N, 158°30'W; NMFS H TC32-73 (1:7.0) 19°31'N, 156°06'W. SOUTHEAST PACIFIC: (all BPBM) 16459 (2:144- 168), 13251 (1:135), 16455 (1:122) Pitcairn I.; 6797 (1:150.5), 6798 ) 1:185), 6799 (1:158), 6800 ( 1: 156.5) Easter I. WESTERN PACIFIC: GVF stn HK91 (2:85-109) 19"38'N, 111°30'E; GVF 2269 ( 1:128) Gulf ofThailand; CAS 29126 (1:32) Ternate, Moluc- cas; CAS 6987 (1:41) Misaki, Japan; CAS 6752 (3:100-114) Wakanoura Kii, Japan; CAS 53402 (1:225) Hachijo I., Japan; CAS 15849 (10:90-125) Taiwan Strait. AUSTRALIA: AMS 1.17228-001 (10:67-91) New South Wales. INDIAN OCEAN: RUSI 2782 (1:47.5) Knysna, South Africa; RUSI 3709 (1:60.5) East Cape, South Africa; RUSI 3710 ( 1:65) Inhaca, Mozambique: BPBM 19022 (2:173-188) Negombo, Ceylon; BPBM 20255 ( 1:195) Wolmar, Mauritius. WESTERN ATLANTIC OCEAN: CAS 4761 (1:150) Jamaica; CAS 54039 (1:94) Havana, Cuba; CAS 18182 (2:50.5-57.5) 29°14'N, 88°19'W; CAS 17184 (1:91) Pine I., Fla.;GCRLVTS:11184( 1:113) San Bias, Panama; LACM 1463 ( 1:84.5) Key Biscayne, Fla.; LACM 6281, 6282, 6283, 6284, 5781, 5872 (23:64-159) southern Jamaica; NMFS LJ Gill 3-64 (1:59) 33°29'N, 76^40 'W; NMFS LJ Silver Bay 3458 (1:60) 29°03'N, 78°04'W; NMFS M Oregon 77-72-39-144 (1:12.5) 23°34'N, 82°22'W, 39-73 (1:13) 21°31'N, 86°14'W, 39-50 (1:24) 16°50'N, 80°13'W, 39-48 (1:24.5) 17°26'N, 79°26'W, 39-58 ( 1:30) 21°01'N, 80°14'W, 39-63 (2:10-40) 19°41'N, 84°13'W, 39-01 (1:23) 13°00'N, 60°00'W, 39-39 (1:45) 18°00'N, 73°00'W, 39-11 (1:56.5) 17°25'N, 63°00'W; NMFS M Bowers 75-126-8 (1:28) 26°00'N, 79°30'W; NMFS M Oregon 77-76-66-19786 (2:23-32) 18°18'N, 75°22'W, 66-19789 (2:20-30) 18°49'N, 74°44'W, 66- 19790 (6:27-34) 19°22 'N, 75°44'W, 66-19791 ( 18:19-33) 17°50'N, 74°47'W. Note. — Since this paper was accepted for publica- tion, NMFS H and most HIMB specimens were transferrred to BPBM. ACKNOWLEDGMENTS I am grateful to the following individuals for information on and loans of specimens: D. G. Smith, Marine Biomedical Institute, University of Texas; M. M. Smith and R. Winterbottom, RUSI; T. Potthoff, NMFS M; S. J. Karnella, USNM; M. L. Bauchot, NMHN; J. Moreland, Na- tional Museum of New Zealand; E. H. Ahlstrom and B. Y. Sumida, NMFS LJ; P. M. Sonoda and W. N. Eschmeyer, CAS; D. A. Thomson, UA; P. Kahsbauer, NMV; J. M. Dixon, National Museum of Victoria; A. C. Wheeler, BMNH; C. E. Dawson, GCRL; D. F. Hoese and J. R. Paxton, AMS; R. J. Lavenberg, LACM; R. H. Rosenblatt, J. Copp, and J. Pulsifer, SIO; W. F. Smith-Vaniz, ANSP; and A. Suzumoto, BPBM. Special thanks go to J. C. Tyler (NMFS M) and C. Baer (HIMB) for critically reading the manuscript and W. I. Follett and L. Dempster (CAS) for valuable ad- vice on nomenclature. J. E. Randall (BPBM) pro- vided the initial stimulus for this study, and his help and encouragement are gratefully acknowl- edged. L. Y. Maluf (UA) gave valuable assis- tance in the final stages of manuscript prepara- tion. W. Waston (HIMB) drew Figure 2L D. Hashimoto (HIMB) conducted the rearing exper- iments that produced two of the metamorphosed juveniles of D. holocanthus. LITERATURE CITED Artedi, p. 1738. Ichthyologia (Lugduni Batavorum). Pars III Genera Piscium. 1962 reprint, Historae Naturalis Classica 15, J. Cramer, Weinheim, p. 203-294. BAGNIS, R., p. MAZELLIER, J. BENNETT, AND E. CHRISTIAN. 1972. Fishes of Polynesia. European Book Co., San Franc, 368 p. Bailey. R. M., J. E. Fitch, E. S. Herald, E. A. Lachner, C. c. LiNDSEY, C. R. Robins, and W. B. Scott 1970. A list of common and scientific names of fishes from the United States and Canada. 3d ed. Am. Fish. Soc. Publ.6, 150p. Bertelsen,E. 1951. The ceratioid fishes. Ontogeny, taxonomy , distribu- tion and biology. Dana-Rep. Carlsberg Found. 39, 276 p. Blanco, G. J., and D. V. Villadolid. 1951. The young of some fishes of Luzon. Philipp. J. Fish. 1:67-93. BLEEKER, P. 1865. Atlas Ichthyologique des Indes Orientales Neer- landaises, Vol. V. Baudroies. Ostracions, Gymnodontes, Balistes. Amst.,152p. BLOCH.M.E. 1785. Naturgeschichte der Auslandischen Fische. 2 vol.: 1:1-136,2:1-160. Berl. Bloch, M. E., and J. G. Schneider. 1801. Systema ichthyologiae. Berl., 584 p. Breder, C. M., Jr., and E. Clark. 1947. A contribution to the visceral anatomy, develop- ment, and relationships of the Plectognathi. Bull. Am. Mus. Nat. Hist. 88:287-319. Breder, C. M., and D. E. Rosen. 1966. Modes of reproduction in fishes. Nat. Hist. Press, Garden City, N.Y., 941 p. BRIS.SOUT DE BARNEVILLE, M. L. 1846. Note sur les diodoniens. Rev. Mag. Zool. Pure Appl. 9:136-143. Castelnau,F. 1872. Contribution to the ichthyology of Australia. Proc. Zool. Acclim. Soc. Victoria 1:29-247. Clark, E., and H. A. F. Gohar. 1953. The fishes of the Red Sea: Order Plectog- nathi. Mar. Biol. Stn. Al Ghardaga, Publ. 8, 80 p. Coleman, N. 1974. Australian marine fishes in colour. A. H. and A. W. Reed, Sydney, 108 p. 565 FISHERY BULLETIN: VOL. 76, NO. 3 CURTISS, A. 1938. A short zoology of Tahiti. Priv. publ., 193 p. CUVIER, M. G. 1818. Surlesdiodons, vulgairementorbesepineux. Mem. Paris Mus. Natl. Hist. Nat. 4:121-138. Day, F. 1878. The fishes of India; being a natural history of the fishes known to inhabit the seas and fresh waters of In- dia, Burma and Ceylon. Priv. publ., Lond., 778 p. De Beaufort, L. F. 1962. The fishes of the Indo- Australian archipelago, Vol. XI. E.V. Brill, Leiden, Neth., 481 p. DeKAY,J.E. 1842. Zoology of New York; or. The New York fauna. In Natural History of New York Geological Survey, Vol. I, Part 4, Fishes, 415 p. DUMERIL, A.H. A. 1855. Note sur un travil inedit de Bibron relatif aux pois- sons plectognathes Gymnedontes (Diodons et Tetro- dons). Rev. Mag. Zool. Pure Appl., 2d Ser., 7:274-282. EgeR,W. H. 1963. An exotoxin produced by the puffer Aro^/jro^ his- pidus, with notes on the toxicity of other plectognath fishes. M.S. Thesis, Univ. Hawaii, Honolulu, 88 p. ElGENMANN, C. H. 1885. A review of the genera and species of Diodontidae found in American seas. Ann. NY. Acad. Sci. 3:297-311. EK\L-\N,S. 1967. Zoogeography of the sea. Sidgwick & Jackson, Lond., 417 p. EYDOUX, J. F. T., AND F. L. A. SOULEYET. 1841. Voyage de la Bonite, 1836-1837. Zool. Vol. I. Poissons p. 157-215. Paris. Fowler, H.W. 1928. The fishes of Oceania. Mem. Bemice P. Bishop Mus. 10, 540 p. Fraser-Brunner, a. 1943. Notes on the plectognath fishes. — VIII. The clas- sification of the suborder Tetraodontoidea , with a snyop- sis of the genera. Ann. Mag. Nat. Hist., Ser. 11, 10:1-18. GIBBS, R. H., AND B. B. COLLETTE. 1959. On the identification, distribution, and biology of the dolphins, Coryphaena hippurus and C. equiselis. Bull. Mar. Sci. Gulf Caribb. 9:117-152. GOSLINE, W. a., AND V. E. BROCK. 1960. Handbook of Hawaiian fishes. Univ. Hawaii Press, Honolulu, 372 p. Grant, EM. 1972. Guide to fishes. Dep. Primary Ind, Brisbane, 472 P- GUNTHER, A. C. L. G. 1870. Catalogue of the fishes in the British Mu- seum. Vol. 8, 549 p. Lond. 1910. Andrew Garrett's Fische der Siidsee, Vol. IX. J. Mus. GodefTroy 17:389-515. HALSTEAD, B. W. 1967. Poisonous and venomous marine animals of the world. Vol. 2: Vertebrates. U.S. Gov. Print. Off., 1070 p. HERRE,A. W. C.T. 1924. Poisonous and worthless fishes. An account of the Philippine plectognaths. Philipp. J. Sci. 25:415-511. HOBSON,E.S. 1974. Feeding relationships of teleostean fishes on coral reefs in Kona, Hawaii. Fish. Bull., U.S. 72:915-1031. 566 Hollister,G. 1934. Clearing and dyeing fish for bone study. Zoologica (N.Y.) 12:89-101. HUBBS, C. L., AND K. F. LAGLER. 1958. Fishes of the Great Lakes region. (Revised ed.) Cranbrook Inst. Sci., Bull. 26, 213 p. JENKINS. O. P. 1904. Report on collections of fishes made in the Ha- waiian Islands, with descriptions of new species. Bull. U.S. Fish. Comm. 22:417-511. JORDAN, D. S., AND B. W. EVERMANN. 1891. The fishes of North and Middle America. Bull. U.S. Natl. Mus. 47(2):1241-2183. Jordan, D. S., and C. H. Gilbert. 1883. List of fishes collected at Panama by Captain John M. Dow, now in the United States National Museum. Proc. U.S. Natl, Mus. 5:373-378. Jordan, D. S., and J. O. Snyder. 1902. A review of the gymnodont fishes of Japan. Proc. U.S. Natl. Mus. 24:229-264. KAUP.J. J. 1855. Ubersicht iiber die Species einiger Familien der Sclerodermen. Arch. Naturgesch. 21( l):215-233. Klunzinger, C.B. 1871. Synopsis der Fische des Rothen Meeres, II Theil. Verb. Zool. Bot. Ges. Wien 21:441-688. Lacepede,B. G. E. 1798. Histoire naturelle des poissons. Vol. 2. Plassan, Paris, 632 p. LeDanois, Y. 1954. Sur le dimorphisme sexuel des poissons de la famille des Diodontides. C. R. Hebd. Seances Acad. Sci. (Paris) 238:2354-2356. 1959. Etude osteologique, myologique et systematique des poissons du sous-ordre des Orbiculates. Ann. Inst. Oceanogr. (Monaco), New Ser., 36:1-273. 1961. Catalogue des types de poissons orbiculates du Museum National d'Histoire Naturalle. II. Families des Tetraodontidae, Lagocephalidae, Colomesidae, Diodon- tidae et Triodontidae. Bull. Mus. Natl. Hist. Nat., Ser. 2,33:462-478. Leis, J. M. 1977. Development of the eggs and larvae of the slender mola, Ranzania laevis (Pisces: Molidae). Bull. Mar. Sci. 27:448-466. Linnaeus, C. 1758. Systema Naturae. 10th ed. Holmiae, Impensis Direct Laurentii Salvi, Tomus 1, 824 p. Marshall, T.C. 1965. Fishes of the Great Barrier Reef and coastal waters of Queensland. Livingston Publ. Co., Narberth, Pa., 566 p. Masuda, H., D. araga, and T. Yoshino. 1975. Coastal fishes of southern Japan. Tokai Univ. Press, Tokyo, 379 p. Meek, S. E., and S. F. Hildebrand. 1928. The marine fishes of Panama. Field Mus. Nat. Hist., Zool. Vol. 15, Part 3, 1045 p. MITCHILL, S. L. 1815. The fishes of New York, described and arranged. Trans. Lit. Philos. Soc. N.Y. 1:355-492. MITO.S. 1966. Fish eggs and larvae. [In Jpn.] In S. Motoda (editor), Illustrations of the marine plankton of Japan, Vol . 7 , 74 p. Soy 6-Sha, Tokyo. LEIS: SYSTKMATICS AM) /.OOCKOC.KAl'H Y OF IHK I'OKCCPIXKFISHES NlCHOI-S. J. T., AND C. M. BHKDKH, JH. 1927. The marine fishes of New York and southern New England. Zoologica (N.Y.) 9:1-192. NlSHIMl'RA.S. 1960. Some aspects of the natural history of porcupine puffers migrating to the Japanese waters. I. Spawning and migration. [In Jpn., Engl, summ.l Jpn. Ecol. 10:6-11. Orsi,J. J. 1974. A check list of the marine and freshwater fishes of Vietnam. Publ. Seto Mar. Biol. Lab. 21:153-177. PARRA.D. A. 1787. Descripcion de diferentes plezas de historia natural, las mas del rama maritimo. representadas en setenta y cinco laminas. Havana, 543 p. POLL,M. 1959. Expedition oceanographique beige dans les eaux cotieres Africaines de I'Atlantique Sud. ( 1948-1949). Vol. IV. Fascicule 3B. Poissons V. Teleosteen acanthoptery- giens (deuxieme partie). Bruxelles, 417 p. RAFINESQUE.C.S. 1810. Indiced'ittiologiaSiciliana, Messina, 70 p. RANDALL, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. (Miami) 5:665-847. 1968. Caribbean reef fishes. T. F, H. Publ., Inc., Nep- tune City, 318 p. Sanzo, L. 1930. Ricerche biologiche su material! raccolti dal Prof. L. Sanzo nella campagna idrografica nel Mar Rosso della R. N. Ammiraglio Magnaghi 1923-1924. VII. Plectog- nathi. Mem. Com. Talassogr. Ital. 167:1-111. SCOTT, E. O. G. 1957. Observations on some Tasmanian fishes: Part VIII. Pap. Proc. R. Soc. Tasmania 91:145-156. Scott, T. D. 1962. The marine and freshwater fishes of South Austra- lia. Gov. Print, Adelaide, 338 p. SHAW,G. 1804. General zoology or systematic natural his- tory. Pisces 5(2): 1-463. Lond. STEINDACHNER, F. 1866. Zur Fischfauna von Port Jackson in Austra- lien. Sitzungsber. Akad. Wiss. Wien 53:424-481. TORCHlo,M. 1963. Accertata presenza di un rappresentante della famiglia Diodontidae in Mediterraneo (Osteichthyes Tet- raodontiformes). Atti Soc. Ital. Sci. Nat. Mus. Civ. Stor. Nat. Milano 102(3):l-7. TR0SCHEL,F.H. 1847. Bericht iiber die Leistungen in der Ichthyolgie wahrend des jahres 1846. Arch. Naturgesch. 13:343- 367. waite, E. R. 1923. The fishes of South Australia. Gov. Print., Adelaide, 243 p. Watson, W., and J. M. Leis. 1974. Ichthyoplankton of Kaneohe Bay, Hawaii. A one- year study offish eggs and larvae. Univ. Hawaii, Hon- olulu. Sea Grant Program TR-75-1, 178 p. Whitley, G. P. 1931. Studies in ichthyology. No. 4. Rec. Aust. Mus. 18:96-133. 1933. Studies in ichthyology. No. 7. Rec. Aust. Mus. 19:60-112. WOLFSCHEIMER,G. 1957. A spawning of procupine puffers. Aquarium (Phila. 126:288-290. 567 PROBABLE CASE OF STREAMBED OVERSEEDING— 1967 PINK SALMON, ONCORHYNCHUS GORBUSCHA, SPAWNERS AND SURVIVAL OF THEIR PROGENY IN SASHIN CREEK, SOUTHEASTERN ALASKA William R. Heard' ABSTRACT The 1967 escapement of 38,067 pink salmon, Oncorhynchus gorhuscha. to Sashin Creek, southeastern Alaska, was the largest since 1942. Studies on distribution and density of spawners and freshwater survival of their progeny indicated that deposition of excessive numbers of eggs caused a severe compensatory mortality of alevins during winter. Spa wner density was 1.7, 1.6, and 1.2 females/m^ in upper, middle, and lower study areas respectively. The greater density of spawners in the upper area in the odd-numbered years may be determined by genetic factors like timing of escapements and by greater marine survival of fry from the upper area. Based on the previously consistent relation between timing of adult entry and resulting freshwater survival, 1967 spawners should have produced 8 million fry rather than the 3 million that were produced. Mortality of eggs and alevins was high during spawning, low between spawning and hatching, and high between hatching and emergence. Between 1 December 1967 and 25 March 1968, 11.1 million eggs or alevins, 10.7 million of which were alive on 1 December, disappeared within the streambed. Initial mortality of these progeny probably occurred in the early alevin stage from oxygen privation, whereas disappearance was probably related to rapid decomposition and invertebrate scavenging. A "snowball effect" is postulated whereby alevins that die shortly after hatching place increasing demands on available oxygen, causing accelerated mortality. A review of historical patterns of fry production in Sashin Creek indicates that streambed overseeding occurred in 1967. Studies of pink salmon, Oncorhynchus gorbuscha, in Sashin Creek, Baranof Island, southeastern Alaska, have shown that certain factors markedly affect freshwater survival. These factors include: 1) seasonal timing of spawning (Skud 1958); 2) density and distribution of adults on the spawning grounds relative to ecological characteristics of the stream, especially gradient (Merrell 1962); and 3) quality of the intragravel environment, including oxygen content of intragravel water and amount of silt and fine particulate material in streambed gravels (McNeil 1966, 1968). Other fac- tors of significance, but believed to be of less influence on freshwater survival in Sashin Creek, include predation on eggs and alevins (McLarney 1967), stream discharge during spawning (Ellis 1969), and incubation (McNeil 1968). The spawning ground of Sashin Creek extends from the head of tidewater to an impassable falls 1,200 m upstream and includes 13,629 m^ of streambed. Ninety-six percent (13,084 m^) of this ground comprises three distinct ecological areas 'Northwest and Alaska Fisheries Center Auke Bay Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 155, Auke Bay, AK 99821. Manuscript accepted February 1978. FISHERY BULLETIN: VOL. 76. NO. 3. 1978. that differ in gradient and size of particles in the substrate. McNeil (1966) called the areas upper, middle, and lower and described them briefly as follows: upper (2,945 m^) — relatively steep gra- dient (0.79'f ) and coarse streambed gravel; middle (4,067 m^) — intermediate gradient (0.3*^) and medium-sized streambed gravel; and lower (6,072 m^) — low gradient (0.1%) and relatively fine streambed gravel. The remaining 4% (545 m^) of spawning ground is located in a short section of stream between the counting weir and the lower area and is not treated in this paper. Pink salmon spawners entering Sashin Creek (the escapement) have been counted at a weir at the mouth of the creek since 1934, and the result- ing numbers of fry from these escapements have been determined since 1940. During this time, the number of spawners varied from as few as 8 to more than 90,000 and the number of fry produced varied from 50 to almost 6 million. The percentage of freshwater survival, based on the estimated po- tential egg deposition, ranged from 0.06 to 21.75% (Table 1). The high escapement of 38,067 pink salmon spawners in 1967, following a long series of rela- tively low escapements, gave me an opportunity to 569 FISHERY BULLETIN: VOL 76, NO. 3 Table l. — Number of adult pink salmon, potential egg deposi- tion, number of fry produced, and freshwater survival in Sashin Creek, 1934-67. (Modified from McNeil 1968.1 Brood year Number of adults Potential egg deposition' Number of fry produced Percentage freshvi^ater survival 1934 7.917 —- — — 1935 6.323 — — — 1936 5.364 — — — 1937 9,085 — — — 1938 6.467 — — — 1939 16,830 — — — 1940 53.594 52.858.000 3.399,900 8.43 1941 84,303 88.678.000 1.024.300 1.16 1942 92.085 78.894.000 674.000 0.85 1943 14,883 14.980,000 227.800 1.52 1944 4,050 3,904,000 105.600 2.71 1945 5,465 5,062,000 43.100 0.85 1946 933 736.000 1.200 0.16 1947 1,486 1.330.000 27.600 2.07 1948 597 516.000 9.100 1.76 1949 4,902 4,800,000 176.200 3.67 1950 112 86,000 50 0.06 1951 4,366 4,062,000 412,500 10.15 1952^ 45 — 740 — 1953 1,164 1.284.000 95.400 7.43 1954 21 12.000 660 548 1955 9,267 10,286,000 266,200 12,31 1956 933 1.018.000 5,040 050 1957 2.834 2.588.000 562,900 21,75 1958 217 174.000 10.700 6 13 1959 35.391 40.379,000 5.332.400 13.21 19602 162 — 480 — 1961 28.759 29,425,000 5,940,300 20.19 1962^ 8 8,000 100 1 20 1963 16,757 16,640,000 3.256.300 19-57 1964 ^2.193 2,230,000 ■"SIO.OOO 1391 1965 14,833 12,668,000 "2.235.000 17.92 1966 5,761 6.255.000 "744,000 11.99 1967 38.067 44.384.000 3,007,200 6.78 'Based on 2,000 eggs female except when actual fecundity was calculated ^An attempt was made to destroy the spawners or their progeny. ^Natural returning adults (327) were supplemented by the introduction of 1.866 adults taken from Bear Harbor. Kuiu Island "Fry weir not operated: figures are estimates of live alevins in the gravel |ust before start of emergence study the effects of a large spawning population on freshwater survival. For the 1967 escapement I studied 1 ) timing of entry into the stream and the distribution and density of pink salmon on the spawning grounds, and 2) survival of progeny by time periods in the three ecological areas of the stream and the overwinter disappearance of eggs and alevins from streambed gravels. In this paper I present all the available data on escapament size and production of fry in Sashin Creek and develop the hypothesis that streambed overseeding occur- red in 1967. As Ricker (1962:186) pointed out, detailed knowledge on the effects of overseeding is important in understanding why pink salmon populations fluctuate. He stated, "Because it [overseeding] happens rarely nowadays, no chance should be lost to make such a study if one occurs." Simply stated, overseeding can be defined as an egg density in spawning bed gravels that leads to a significantly greater freshwater mortal- ity than a lesser density would cause. As discussed more fully later, it is a complex and dynamic in- teraction between egg density, streambed ecology, and specific climatic conditions. TIMING OF ENTRY AND DISTRIBUTION AND DENSITY OF SPAWNERS The timing of stream entry was determined from daily counts of the adults at the Sashin Creek weir. This timing apparently influences the freshwater survival of the progeny. An inverse relation between time of stream entry of spawners and survival of progeny is usual in Sashin Creek (Skud 1958; Merrell 1962; McNeil 1968; Ellis 1969): high survival has been associated with early spawning and low survival with late spawn- ing. Merrell ( 1962) further pointed out that pink salmon spawn in Sashin Creek an average of 12 days earlier during odd years than even years. ^ As a result, freshwater survival is usually higher among progeny of spawners from odd years than among those from even years (Table ll. In 1967, 507f of the spawners had entered Sashin Creek by 20 August, the second earliest date on record. The early entry indicated that sur- vival of eggs and alevins would be high, but this did not prove to be the case. Throughout the run, random lots of females were tagged at the weir, and the distribution and density of spawners were determined from daily counts of both tagged and untagged females on the spawning grounds. This technique, described by McNeil (1968) and used by Ellis (1969), provides two methods of estimating the numbers of females spawning in the upper, middle, and lower areas. One method assumes that tagged females distrib- ute themselves among the three sections the same as untagged females. In the other method, the summed daily count of all females in each area is divided by the average longevity on the spawning grounds. The results from the two methods were generally in agreement, except for the upper area where estimates based on distribution of tagged females were considerably higher than those based on total females. The difference may reflect the difficulty in making accurate counts on spawn- ing riffles where densities of spawners are high; in such a situation an observer might count small ^The date when 50% of the escapement to Sashin Creek had entered the stream has been commonly used as an index of time of spawning. 570 HEARD: PROBABLE CASE OF STREAMBED OVERSEEDING numbers of tagged females more accurately than large numbers of untagged females. Because the relative accuracy of the two methods is unknown, I averaged them to arrive at mean estimates of den- sities of females in the three areas (Table 2). Spawner density in Sashin Creek is usually un- equal in the three study areas, depending in part on the total number of spawners. Densities in 1967 were the highest recorded for specific areas-'^ of the stream (Table 3). Merrell (1962) noted that in years when many spawners were present, they utilized all of the available spawning grounds, and in years when few were present, they spawned mostly in the lower portion of the stream. When the upper area was used, survival of eggs and alevins in that area was higher and the number of fry produced was proportionally much greater than in the middle and lower areas (Merrell 1962). In addition, the sediment content and water qual- ity of the stream in the upper area were better than in the other two areas (McNeil 1966, 1968). Sashin Creek thus presents an apparent paradox — the least favorable areas are used in years of relatively few spawners, and the best areas are used only during years of great numbers of spawners. Merrell ( 1962) thought that che greater use of the upper area was related primarily to density- dependent spawner interactions. In 1967, how- ever, the heavy use of the upper area was appar- ently not the result of high densities downstream forcing spawners into upstream areas: spawner ^Although the total number of spawners entering the stream has been recorded since 1934 (Table 1), detailed studies on the distribution of spawners in the upper, middle, and lower areas of the stream have been available only since 1961. Table 2. — Estimated densities of female pink salmon spawning in three areas of Sashin Creek, 1967. Females per square meter Area Based on counts of tagged females only Based on counts of tagged and untagged females Mean Upper Middle Lower 1.90 1.49 1.19 1.59 1.76 1.15 1.74 1.62 1.17 Tables.- —Estimated densities of female pink salmon spawning in three areas of Sashin Creek, 1961-67. Females per square meter Area 1961' 19632 1 9643 1 9654 19665 1967 Upper Middle Lower 1.00 1.00 1.00 0.59 089 059 0-01 0 58 0.09 062 0.13 0.44 0.04 0.27 0.28 1.74 1.62 1.17 'Extrapolated from subjective estimate (McNeil et al. 1964). ^Adjusted from McNeil (1966). ^McNeil et al (1969). "McNeil (1968). ^Ellis (1969). densities in the upper area built up rapidly before spawning reached significant levels in the middle and lower areas. Although the upper area contains only 22'7f of the combined spawning grounds of the three areas, in 1967, 62^/ of the first group of female pink salmon tagged at the weir spawned in the upper area (Table 4). In general, the intensity of spawning in 1967 progressed to downstream areas from the upper area rather than the reverse. McNeil (1966, 1968) noted similar downstream shifts in spawning in Sashin Creek in 1963 and 1965. Although McNeil (1966) felt that the shift occurred because of heavy rainfall during the spawning period, he later noted ( McNeil 1968) the same phenomenon during an unusually dry year. It appears that the upper area in Sashin Creek is not necessarily used because of spawner overflow but because of more complex factors. Two interre- lated factors could account for the spawner dis- tributions observed in recent years: 1) migratory behavior associated with timing of the escape- ment, and 2) a genetic tendency for odd-year spawners to use upstream areas. Odd-year spawn- ers enter the stream earlier than even-year spawn- ers. A characteristic of early stream entry in anadromous fishes may be a tendency to migrate farther upstream than spawners associated with late stream entry (Briggs 1955). In addition to early entry and use of the upper area, odd-year spawners for the past 9 or 10 generations have consistently had higher escapements and, except for 1967, higher freshwater survival of progeny than even-year spawners (Table 1). Natural selec- tion may be operating, in recent odd-year genera- tions, to encourage progeny produced in the upper area to spawn in that area. Wells and McNeil ( 1970) showed that fry produced in the upper area of Sashin Creek were larger and presumably of better quality than those produced in the downstream areas. Differential marine survival Table 4. — Dates of tagging and percentage of total escapement counted through weir, numbers of female pink salmon tagged, and spawning distribution of tagged females in three areas of Sashin Creek, 1967. Percentage of total escape Tagged females Percentage of tagged fisti accounted for Date of ment cour ited Females observed Upper Middle Lower tagging' througfi weir tagged spawning area area area 10 Aug. 3 49 40 62 25 12 12 Aug. 13 50 40 22 37 40 17 Aug 26 50 42 21 23 55 20 Aug. 54 50 50 22 42 36 5 Sept. 98 50 40 22 30 48 'Females tagged on each date received color-coded tags that differentiated them from females tagged on other dates. 571 FISHERY BULLETIN; VOL. 76, NO. 3 that favored fry produced in the upper area over those produced in the downstream areas could ac- count for the greater escapements of odd-year spawners in recent years. SURVIVAL OF EGGS AND ALEVINS Survival of eggs and alevins from the 1967 brood year was estimated in Sashin Creek for four time periods: 1) from stream entry to end of spawning, 2) from end of spawning to hatching, 3) from hatching to shortly before fry emergence, and 4) from shortly before emergence to emergence and downstream migration of fry. The estimates of survival were based on esti- mates of the potential egg deposition of female spawners and estimates of the surviving eggs and alevins in the three study areas. Potential egg deposition was estimated by multiplying the number of females by average fecundity. Densities of eggs and alevins were determined after spawn- ing, during hatching, and before fry emergence by sampling randomly selected 0.1-m- points in the streambed with a hydraulic sampling technique described by McNeil (1964a). The number of fry migrating from the stream were estimated on the basis of daily counts of fry migrating through a weir at the stream mouth. Numbers of females entering the stream and average fecundity were derived from counts and samples taken at the weir. Of the 38,067 pink salmon spawners entering Sashin Creek in 1967, 19,639 (52'7f ) were females. Total counts of mature eggs from each of 35 females selected at random from the run ranged from 810 to 2,954 (average 2,260) eggs/female (907r confidence limit of mean fecundity was ±115 eggs). The percentage of eggs available for deposition that are actually buried in the streambed is partly dependent on the density of spawners. McNeil ( 1964b) discussed the role of redd superimposition and showed that at spawner densities approaching 3 to 4 females/ m^ of spawning ground, an upper asymptotic limit on the density of eggs in the streambed is reached. Factors other than spawner density that may influence egg deposition include loss of adults in the stream before spawning and retention of eggs in the female's body (Neave 1953), type and characteristics of the spawning substrate (McNeil 1966), streamflow during spawning (Ellis 1969), and loss of eggs to verte- brate predators during the spawning process (Moyle 1966; McLarney 1967; Reed 1967). The efficiency of egg deposition of pink salmon spawners in Sashin Creek is highly variable, from 37 to 829^^ of the potential egg deposition (Ellis 1969). In 1967 the number of pink salmon eggs potentially available for deposition was 44.4 mil- lion, with 19.9 million of these (459^ of the poten- tial) estimated to be in the streambed after spawn- ing. The efficiency of egg deposition was 47'7f in the upper area, 50% in the middle area, and 387f in the lower area. Although spawner densities were high in 1967 (Table 3), the ability of pink salmon to void most of their eggs during spawning did not seem to be affected. Egg retention is characteristically low in Sashin Creek, usually less than 57c of fecundity ( McNeil 1966; Ellis 1969). In 1967, 1 examined the body cavities of 402 spent female pink salmon (about 2'7c of the total) and found that average egg retention was 1.59^ of average fecundity. The proportion of eggs actually deposited that were alive at the end of the spawning period in 1967 was highest (93%) in the upper area, inter- mediate (83%) in the middle area, and lowest (74% ) in the lower area (Table 5). This high survi- val in the upper area is consistent with that of previous years. The ratio of live to combined live and dead eggs and alevins was usually higher in the upper and middle areas than in the lower area through hatching to the beginning of fry emergence (Table 5). Survival of eggs and alevins varied among the three time periods (during spawning, between end of spawning and hatching, and between end of hatching and emergence). Survival within each time period for each area was higher between spawning and hatching than during spawning or between hatching and emergence (Table 6). As previously discussed, survival during spawning was related primarily to the ability of females to successfully deposit their eggs because a high per- centage of the eggs buried were alive shortly after spawning. Survival between spawning and hatch- ing and between hatching and emergence pertains to survival of eggs and alevins within the streambed. The densities of live preemerged fry in the streambed of Sashin Creek in late March 1968 were 382, 260, and 108/m- in the upper, middle, and lower areas, respectively. From these den- sities I estimated a population of 2.9 million fry in the entire stream. Operation of the fry weir began just after the late March streambed sampling was 572 HEARD: PROBABLE CASE OF STREAMBED OVERSEEDING Table 5. — Potential egg deposition, number of live and dead eggs and alevins, ratio of live to combined live and dead eggs and alevins, and estimated survival of 1967 brood year pink salmon in three areas of Sashin Creek. Area Potential egg deposition per square meter Mean 3,947 3,672 2.644 90°o confidence limits of mean Period beginning 10 Aug and ending Combined live and dead eggs and alevins per square meter QO^o confidence Mean limits of mean Percentage of live to combined live and dead eggs and alevins Mean 90°o confidence limits of mean Percentage calculated survival Upper Middle Lovi/er ::;201 :187 :136 1 Oct. 1.863 ±254 93 ± 1 43 1 Dec. 1.714 ±295 86 ± 4 37 25 Mar. 647 ±138 59 ±13 10 1 Oct. 1,826 ±218 83 ±12 41 1 Dec. 1,591 ±226 70 ± 7 30 25 Mar. 702 ±147 37 ± 2 7 1 Oct. 1,015 ±120 74 ±17 28 1 Dec. 989 ±116 72 ± 2 27 25 Mar. 350 ±70 31 ±10 4 Table 6. — Percentage of estimated survival of 1967 brood year pink salmon eggs and alevins for three time periods in three areas of Sashin Creek and for the entire stream, 1967. Percentage survival Between end of Between end of During spawning and fiatching and Area spawning fiatcfiing emergence Total Upper 43 85 26 10 Middle 41 73 23 7 Lower 28 95 15 4 Entire stream' 37 83 22 68 ' Data weigfited and adjusted to include spawning grounds not included in tfie three study areas completed. Relatively few fry migrated downstream through the weir until mid-April; the daily fry migrations increased steadily through late April, reached a peak in early May, then de- clined rapidly, and were essentially completed by early June (Figure 1). The total number of pink salmon fry estimated to migrate from Sashin Creek from the 1967 brood year spawners was 3 million. Similar close agreements between esti- mates based on densities of preemerged fry and those based on number of fry counted at the weir have occurred in previous years (McNeil 1968). The 3 million fry migrating from Sashin Creek in the spring of 1968 represent a total freshwater survival of 6.8'''!^ of the 44.4 million potential egg deposition. This is the lowest freshwater survival in the odd-year line of pink salmon spawners in Sashin Creek since 1949 (Table 1). I will sub- sequently attempt to show that this reduced sur- vival was primarily due to excessive seeding of the streambed during spawning. DISAPPEARANCE OF EGGS AND ALEVINS To determine the number of eggs and alevins that disappeared from the streambed, I compared the potential egg deposition with the numbers of live and dead eggs at the end of spawning and the number of eggs and alevins at the time of hatching and just before emergence. In 1967, 55% of the potential egg deposition disappeared during spawning. The fate of these eggs is unknown, but they were probably removed from the stream dur- ing the spawning period by predators, scavengers, or turbulent streamflow. McLarney (1967) and McNeil ( 1968) discussed the roles offish predators (especially sculpins) and water turbulence in re- moving eggs from Sashin Creek during spawning and between spawning and hatching. McNeil ( 1968) found that eggs and alevins of the 1963 and 1965 brood years disappeared at differ- ent rates in the upper, middle, and lower study areas of Sashin Creek. Most of the 1963 brood year progeny disappeared during spawning, and most of the 1965 brood year progeny disappeared be- tween hatching and emergence (over the winter). I will examine closely the possible fate of eggs and alevins during this period (December to March) because the factors that caused a reduced freshwa- ter survival of 1967 brood year progeny prevailed during this period. The estimated percentages of the potential egg deposition that disappeared in the upper, middle, and lower areas of Sashin Creek were similar within each of the three periods. This disappear- ance varied greatly between periods: 55% of the progeny (eggs or alevins) had disappeared by 1 October, 4% between October and December, and 25% between December and March (Table 7). The disappearance between hatching and emergence (December and March) appears more significant when expressed in terms of numbers present in December; 56-65% of the eggs and alevins in the upper, middle, and lower areas of Sashin Creek on 1 December had disappeared by 25 March (Table 8). 573 FISHERY BULLETIN: VOL. 76. NO. 3 Figure l. — Daily number of 1967 brood year pink salmon fry counted through Sashin Creek weir in spring 1968. i ' ^ 364,000 325 - 300 - 275 - 250 - Q 225 z < c/) O 200 - u. — NUMBER OF tn O - 100 - 75 - 50 - 25 - {III, 1 lllli... ...... u^ I •■■|"'"|""|""|' i'|i II 1 1 25 5 15 25 MARCH APRIL 15 MAY 25 5 15 JUNE Mortality Patterns in the Streambed Mortality of eggs and alevins within the streambed at Sashin Creek is evident in two ways: 1 ) as a reduction in the total population of eggs and alevins within the streambed, i.e., they disappear, and 2) as an increase in the number of dead eggs and alevins in the streambed and a decrease in the number of live eggs and alevins. In the first in- stance, some factors that can cause eggs and ale- vins to disappear are turbulent streamflow, streambed scouring, predation, and scavenging. Excluding predation, these same factors can also cause dead eggs and alevins that may have died for other reasons to disappear from the streambed. In the second instance, factors causing an increase in the number of dead eggs and alevins within the streambed are generally relatable to desiccation, freezing, and the quality of intragravel water. In addition, dead eggs and alevins may disap- pear because of at least two factors that do not affect live eggs and alevins: biochemical decom- position and consumption by intragravel inver- tebrate scavengers. Thus, factors that cause eggs 574 HEARD: PROBABLE CASE OF STREAMBED OVERSEEDING Table 7. — Percentage of potential egg deposition of 1967 brood year pink salmon that disappeared from three areas of Sashin Creek and from the entire stream by 1 October 1967, 1 December 1967, and 25 March 1968. Estimated percentage ot potential egg deposition disappearing Area By 1 Oct. 1 Oct. to 1 Dec. 1 Dec. to 25 Mar. Upper Middle Lower Entire stream' 53 50 62 55 4 6 1 4 27 24 24 25 ' Data weighted and adjusted to include spawning grounds not included in the three study areas. T.\BLE 8. — Estimated densities of all eggs and alevins (live and dead) in three areas of Sashin Creek on 1 December 1967 and 25 March 1968, and percentage that disappeared between the two dates. Number of eggs and alevins per square meter on Area 1 Dec. 25 Mar. Percentage that disappeared between dates Upper Middle Lower 1,714 1,591 989 647 702 350 62 56 65 or alevins to disappear from streambed gravels may or may not have been the initial cause of death. It is unlikely that turbulent streamflow, streambed shifting, or predation were the reasons that 1967 brood year eggs or alevins disappeared between early December and late March, Streamflows were generally low, and although an intermittent ice cover was present on Sashin Creek during January, February, and March, there was no indication of streambed shifting be- cause of ice scouring. A series of short metal stakes driven into the streambed throughout the stream in November to mark coho salmon, O. kisutch, redds was still in place in March, indicating that no streambed shifting had occurred. Most fish in Sashin Creek that could eat pink salmon eggs and alevins (juvenile coho salmon; rainbow trout, Salmo gairdneri; Dolly Varden, Salvelinus malma\ and coastrange sculpin, Cottus aleuticus) are essentially dormant during the winter when water temperatures are low. Chap- man and Bjornn ( 1969) have shown that resident stream salmonids may disappear into the sub- strate when water temperatures fall below 4.4 °- 5.5 °C. I have observed similar behavior in Sashin Creek. Stream temperatures in Sashin Creek were below 4.4 °C from 13 December 196-7 to 20 April 1968. Water ouzels, Cinclus mexicanus, and mergan- sers, Mergus merganser, are occasionally present on Sashin Creek during the winter and could ac- count for the disappearance of some eggs and ale- vins during open water periods. In considering the magnitude of the disappearance of 1967 brood year eggs and alevins in Sashin Creek between December and March, it is unlikely that the maximum possible loss to these sources is sig- nificant. This conclusion is based on the small numbers of mergansers and ouzels present, the amount of time the stream was covered with ice, and the large number of eggs or alevins that dis- appeared. Between two and five mergansers were noted in the vicinity periodically. When present, these birds spent much of their time in the inter- tidal portion of Sashin Creek or in the adjacent estuary. Only four of the smaller and territorial ouzels normally occur along the upper, middle, and lower areas of Sashin Creek in winter, and during periods of ice cover these birds go elsewhere. Based on periodic observations and temperature records, I estimate the stream was covered with ice approximately half the 1967-68 winter. The estimated population of live and dead pink salmon eggs and alevins in Sashin Creek was 18.3 million on 1 December 1967 and 7.2 million on 25 March 1968 (Table 9), making a loss of 11.1 mil- lion eggs and alevins between the two dates. Be- cause there is little evidence that the loss was caused by external factors that physically re- moved eggs or alevins from the streambed, the loss was likely due to factors within the intragravel environment. The disappearance of 11.1 million eggs and ale- vins from the streambed between hatching and emergence led me to examine the relation between live eggs and alevins and dead eggs and alevins in the streambed. The densities of dead eggs and alevins in the upper, middle, and lower areas (Ta- ble 10) indicated that the numbers of dead eggs and alevins remained relatively stable between the time periods. This does not necessarily indi- cate that dead eggs in the streambed during one sampling period were still there during a later sampling period. Dead eggs can disappear at any time for many reasons, but can persist in a T.ABLE 9. — Estimated population of live and dead pink salmon eggs and alevins in Sashin Creek on 1 October 1967, 1 December 1967, and 25 March 1968, Millii ons of eggs and alevins in streambed Sample date Live Dead Total 1 Oct. 1967 1 Dec. 1967 25 Mar. 1968 16.5 13.7 3.0 34 19.9 46 18.3 4.2 7.2 575 FISHERY BULLETIN: VOL. 76, NO. 3 streambed for as long as 18 mo ( McNeil et al. 1964). Once hatching is completed, no new dead eggs can be added to the streambed. Because hatching of live eggs was well underway on 1 December ( about 35% completed), many of the dead eggs present in March had already died by 1 December (Table 10). Most of the eggs and alevins that disappeared over the winter (about 11 million; Tables 7, 8, 9) were individuals that had been alive on 1 De- cember because the number of dead eggs and ale- vins was essentially unchanged from December (4.6 million) to March (4.2 million) (Table 9). Mor- tality (in the form of disappearance) of live eggs and alevins in the streambed between 1 December and 25 March was 74% in the upper area, 77% in the middle area, and 85% in the lower area (Table 11). Of the 11.1 million pink salmon eggs and alevins that disappeared within the streambed be- tween 1 December and 25 March, 10.7 million were alive on 1 December. As previously mentioned, the cause of the dis- appearance of dead eggs and alevins in the streambed may differ from the cause of their deaths. This apparently occurred with the 1967 brood year pink salmon progeny in Sashin Creek, and I offer the following theoretical sequence to explain the major overwinter disappearance of eggs and alevins. The greatest number of fry produced in Sashin Creek since 1940 was 5.9 million (Table 1). On 1 December 1967, 13.7 million live pink salmon eggs and alevins were in the Sashin Creek streambed (Table 9), a number that appears to Table 10. — Estimated densities of dead pink salmon eggs and alevins in three areas of Sashin Creek on 1 October 1967, 1 December 1967, and 25 March 1968. Dead eggs and alevins per square meter Upper area Eggs Alevins Middle area Loviier area Date Eggs Alevins Eggs Alevins 1 Oct, 1967 1 Dec. 1967 25 Mar. 1968 129 0 223 7 196 69 310 0 459 18 334 108 264 266 199 0 11 43 Table ll. — Estimated densities of live pink salmon eggs and alevins in three areas of Sashin Creek on 1 December 1967 and 25 March 1968 and disappearance of live eggs or alevms between the two dates. Area Upper Middle Lovi/er Live eggs and alevins per square meter 1 Dec. 1967 25 Mar 1968 Alevins Eggs Alevins Eggs 899 769 463 Percentage of live eggs or alevins that disappeared between dates exceed the capacity of the streambed for pink salmon fry production. I postulate that the high initial density of eggs led to a severe mortality of embryos in the early alevin stage, probably be- cause of widespread oxygen privation or a combi- nation of oxygen privation and a buildup of toxic metabolites. The rate of oxygen consumption by embryos increases steadily with development (Wickett 1954, 1962) and coincides with the gen- eral lowering of streamflows during the late fall, followed by stabilization of streamflows at near the normal winter levels.^ This combination of conditions permitted the embryo population to survive up to, but not much beyond, the hatching period. These recently hatched dead alevins then apparently disappeared rapidly within the streambed through the combined action of biochemical decomposition and intragravel inver- tebrate scavenging. As I will show later, the rapid disappearance of recently hatched dead alevins in the streambed seems consistent with this hypothesis. Although no intragravel water quality data are available from Sashin Creek during or shortly after hatching to support the above theory, a com- parison of the rates of oxygen consumption by pink salmon embryos of various ages indicates that oxygen requirements do steadily increase during the hatching period. The rates of oxygen consump- tion reported for early stage eggs (7-26 days old) have ranged from 0.0003 mg O2 /egg per h ( Wickett 1954) to 0.0005 mg 02/egg per h (Brickell 1971). Brickell found that the rate of oxygen consump- tion by 35-day-old pink salmon eggs was 0.0018 mg Oj/egg per h, almost four times the rate he mea- sured for 7-day-old eggs. Faintly eyed 38-day-old eggs had an oxygen consumption rate of 0.002 mg Og/egg per h (Wickett 1962) while 7-day-old ale- vins had a consumption rate of 0.01 mg 02/alevin per h (Wickett 1954). x 575 345 249 382 260 108 74 77 85 ""Seasonally, stream discharge in Sashin Creek is usually highest in fall and lowest in summer. Discharge in winter months may also be low, but is normally above summer levels. Because unseasonably low winter discharge could reduce oxygen delivery to embryos below the normal seasonal pattern, I com- pared the low monthly discharge during December, January, February, and March for 1967-68 with low discharge patterns in the same months for the period 1951-52 to 1966-67. The low mean monthly discharge from Sashin Creek during December, January, February, and March ranged from 18 to 62 ft^/s and averaged 33 ft^/s for the 16-yr period. The mean minimum monthly discharge during these same 4 mo in 1967-68 was 30 ft^/s (U.S. Geological Survey 1969), suggesting that low streamflow levels during these months in 1967-68 were near normal. 576 HEARD: PROBABLE CASE OF STREAMBED OVERSEEDING In addition to the increasing oxygen require- ments due to growth and development of Hve em- bryos, Brickell (1971) found that rates of oxygen consumption by dead intact pink salmon eggs ex- ceeded those of early stage live eggs fourfold: 0.0018 mg Og/whole dead egg per h versus 0.0004-0.0005 mg Oa/T-day-old live egg per h. He noted even gi'eater oxygen consumption for dead eggs when the chorion was pierced or slit or the egg was fragmented: mean oxygen consumption of fragmented dead eggs in constant-flow cylinders was 0.017 mgOs/eggper h, which exceeds the rate Wickett (1954) found for 7-day-old live alevins. It follows that alevins that die shortly after hatch- ing, because of their soft, exposed, and readily oxidizable tissue, would have higher rates of oxy- gen consumption than whole intact dead eggs, live eggs, or early stage live alevins. These increases in oxygen consumption upon death of developing pink salmon embryos are the rationale for suggesting a "snowball effect" — rapidly increasing deaths of embryos once lethal oxygen concentrations were approached. With high densities of live embryos already placing ex- cessive demands on the oxygen and each death increasing the demand, the resulting heavy mor- tality could have caused fry production to plunge below that expected from lower initial egg densities — an excellent example of Neave's ( 1953) theory of compensatory mortality. Disappearance of Dead Eggs Versus Disappearance of Dead Alevins To test the hypothesis that dead alevins disap- pear within the streambed more rapidly than dead eggs, I conducted a small study in Sashin Creek in the winter of 1968-69 to consider the relative per- sistence of dead eggs and alevins in the streambed. A series of Vibert boxes (small plastic perforated containers), each containing a mixture of streambed gravel, 10 dead eggs, and 10 dead ale- vins (all from 1968 brood year pink salmon) were buried in Sashin Creek on 14 December 1968. The boxes were buried about 20.3 cm deep across a riffle in the middle study area. At irregular inter- vals, pairs of the boxes were removed from the streambed and the contents were preserved for examination. Alevins disappeared from the Vibert boxes at a much faster rate than eggs (Table 12). Fewer than half of the original number of alevins were still recognizable at the end of 2 wk; after 37 days only Table 12. — Contents of Vibert boxes with dead pink salmon eggs and alevins buried in Sashin Creek streambed between 14 December 1968 and 14 April 1969.' Eggs recovered Alevins recovered Invertebrates recovered No of days buried Insect larvae^ Planarian worms^ 0 20 20 0 0 9 20 10 14 4 16 20 4 37 20 24 20 2 40 34 30 20 1 27 144 37 20 4 36 11 44 20 0 55 5 51 19 0 70 3 71 19 0 196 4 86 20 2 72 9 96 20 0 135 6 109 20 0 149 29 121 19 0 102 36 'Each box originally contained 10 dead eggs and 10 dead alevins. Two boxes were removed on each sample date and the contents combined tor reporting. ^Of all insect larvae recovered, 80°o were Plecoptera. 16°o Diptera, 3% Trichoptera. and 1°o Ephemeroptera, ^Tentatively identified as Polycelis borealis. a species that Kenk (1953) commonly found in clear cold streams in southern parts of Alaska, one box contained identifiable alevins. Although the dead alevins disappeared rapidly, only a few of the dead eggs disappeared. In a study to determine whether certain stonefly numphs were predators or scavengers on salmon eggs and alevins, Ellis (1970) found in one experiment that dead pink salmon alevins buried in Vibert boxes in a stream essentially disappeared within a 2-wk period. Concurrently with the rapid disappearance of dead alevins from the buried boxes was a rapid buildup of invertebrates in the boxes. Although invertebrates are commonly found with salmon embryos (Briggs 1953; McDonald 1960; Nicola 1968), it is frequently impossible to determine if predation or scavenging is occurring. Although some groups of stonefly nymphs are known to at- tack live salmon embryos (Stuart 1953; Claire and Phillips 1968), Ellis ( 1970) concluded that nymphs of the carnivorous genus Alloper la were basically scavengers rather than predators. In addition to various insect larvae, a planarian worm tentatively identified as Polycelis borealis was commonly found in the boxes buried in Sashin Creek (Table 12). Little is known on the biology or life history of this planarian, but under favorable conditions it appears to rapidly increase its num- bers in the streambed, and thus may be particu- larly important in removing dead alevins. I have observed successive seasonal increases in the rela- tive abundance of planarians in samples taken from the Sashin Creek streambed with the hy- draulic sampler in the fall, winter, and spring. Total counts of planarians removed from the streambed with the hydraulic sampler are not pos- 577 FISHERY BULLETIN: VOL. 76, NO. 3 sible,^ but partial counts indicated that by March the densities of planarians in some parts of Sashin Creek commonly reached several thousand per square meter. A similar seasonal increase in streambed populations of planarians concomitant with the seasonal occurrence of sockeye salmon, O. nerka, embryos has been noted elsewhere.^ In Sashin Creek there is little doubt that high planarian populations are related to the presence of salmon eggs and alevins, because planarians are scarce in streambed gravels above the impass- able falls where salmon do not spawn. However, the precise role of these organisms in the ecology of spawning beds is unknown. To learn something about the role of planarians, I conducted tests with various combinations of planarians and live and dead salmon eggs and alevins in experimental containers. In these tests planarians did not prey on and were not toxic to live embryos, nor did they feed on dead eggs unless the chorion was broken and the egg contents exposed. EVIDENCE OF OVERSEEDING In assessing the probability of streambed over- seeding in Sashin Creek in the 1967 brood year, it is most useful to compare fry production in 1967- 68 with production in other years. Since 1940, ^When large numbers of planarians are excavated with the hydraulic sampler, many elongate their bodies and pass through the meshes of the collecting net. 6W. L. Hartman, W. R. Heard, and C. W. Strickland. 1962. Red salmon studies at Brooks Lake biological field station, 1961. Unpubl. manuscr. on file, NWAFC Auke Bay Lab. NMFS, NOAA, P.O. Box 155, Auke Bay. Alaska. production has varied from 50 fry to almost 6 mil- lion fry; corresponding parent escapements varied from 8 to 92 ,085 ( Table 1 ) . Only three escapements exceeded that of 1967, and only one of these ( 1940) produced more fry (about 0.4 million more) (Table 1). When the numbers of fry are plotted against potential egg deposition, a dome-shaped fry pro- duction curve is derived for Sashin Creek (McNeil 1969). The relative position of fry production for the 1967 brood year falls near the descending limb of the curve; fry production from the 1941 and 1942 brood years indicates a continuing decrease in fry production as escapements increased (Fig- ure 2). Data collected since 1961 on the density of eggs in the three study areas at the end of spawning provide a means of more precisely defining the fry production potential of the stream. Plotting fry production as a function of actual egg deposition for each area produces curves that suggest the potential maximum fry production in the upper and middle areas is around 500 fry/m^ and the potential in the lower area is about half that number (Figure 3). In 1967 the actual density of eggs deposited considerably exceeded twice the theoretical maximum fry production in all three areas and the fry production was considerably below the maximum; it appears that overseeding occurred in 1967. Until 1967 the timing of entry of adults into Sashin Creek had usually been an accurate indi- cator of the freshwater survival of progeny (Mer- rell 1962; McNeil 1968; Ellis 1969). The presumed biological basis for the correlation between early spawning and high survival was that embryos de- 500 Figure 2. — Production of pink salmon in Sashin Creek, 1940-67. The shaded area in the lower left is shown on a larger scale in the upper right corner. Nine genera- tions of the even-year line 1946-62 were excluded because fry productions were all v ^14 _ \^^ ■ — Q > \(.76) >. z \ O 12 _ \ 5 ^ _l < \ ^0 - v. 31) \ z \( 42) 0. \ uj 3 _ (.15) \ > \ -J ^\..,.^ \ 6 - — -U" \( 26) \ 4 - (~07)^^'^;;;;--V^ 2 - ^0- 0 Aiir. 'SFPT ' nn ' ' MnV ' HF r ' lAM ' FFR ' MAD ' AP D ' MAV ' Figure 5. — Number of live pink salmon in Sashin Creek at the beginning and end of three periods in freshwater for 1963, 1965, and 1967 brood years. Numbers in parentheses show instan- taneous monthly mortality coefficients from the egg through alevin stages. Mortality during fry migration (April and May) for the 1963 and 1967 brood years was negligible when measured as the difference between streambed and weir estimates of total fry production. The dotted extension of the 1965 brood year assumes no mortality during this period. 1963 and 1967 (0.76 and 0.71); the lower mortality during spawning in 1965 (0.31) probably reflects efficient spawning during the low streamflow con- dition prevailing that year (McNeil 1968). Mortal- ity from spawning to hatching was similar, but mortality from hatching to emergence was strik- ingly different in each of the 3 yr (Figure 5). McNeil (1968) suggested the increase in over- winter mortality of 1965 brood year progeny (0.26) over 1963 brood year progeny (0.07) might have been related to a delayed mortality from low con- centrations of dissolved oxygen in early embryo development during drought conditions in late summer and early fall in 1965. The number of spawners in 1967 was more than double the number in 1963 and 1965 (Table 1). The over- winter mortality of 1967 brood year progeny (0.42) was considerably higher than the high mortality of the 1965 brood year (0.26). The heavy over- winter mortality experienced by the 1967 brood year progeny may also have been caused by low dissolved oxygen concentrations. However, be- cause no drought conditions existed while the progeny were in the gravel, these poor oxygen conditions probably resulted from the high density of eggs and alevins in the streambed. SUMMARY 1. In 1967, 38,067 pink salmon spawned in Sashin Creek on Baranof Island, Alaska. Fifty- two percent of the spawners ( 19,639) were females; mean fecundity was 2,260 eggs/female and the potential number of eggs available for deposition totaled 44.4 million. 2. Entry of spawners into the stream was the second earliest on record; based on the previously consistent relation between time of entry and freshwater survival, the production of fry should have been greater than any previously recorded, but the 3 million fry produced were less than half the predicted number. 3. Mean female densities on the spawning grounds were 1.74/m2 in the upper area, 1.62/m2 in the middle area, and 1.17/m2 in the lower area. Densities were higher in the upper area at the beginning of spawning before significant levels of spawning occurred in the middle or lower areas. The tendency for spawners in the odd-year line to utilize the upper area of Sashin Creek may be due to genetic factors, including timing of escape- ments, and possibly differential marine survival favoring fry produced in the upper area. 580 HEARD: PROBABLE CASE OF STREAMBED OVERSEEDING 4. Survival of progeny of the 1967 spawners was determined a) from stream entry to end of spawning, b) from end of spawning to hatching, c) from hatching to shortly before fry emergence, and d) from shortly before emergence to emergence and downstream migration of fry. In general, sur- vival in each of these time periods was greatest in the upper area, lowest in the lower area, and in- termediate in the middle area, a pattern consis- tent with previous survival studies at Sashin Creek. Total freshwater survival from potential egg deposition to preemerged fry was 107f , 79c , and 49c in the upper, middle, and lower areas, respec- tively, and 6.8% for the entire stream. The total number of migrating fry agreed closely with the estimates of preemerged fry in the streambed in late March. 5. Mortality of eggs and alevins was high dur- ing spawning, low between spawning and hatch- ing, and high between hatching and emergence. Between 1 December 1967 and 25 March 1968, 11.1 million eggs or alevins disappeared within Sashin Creek streambed; 10.7 million of these were alive on 1 December. The high densities of eggs and alevins in the streambed after spawning and at hatching are believed to exceed the streambed capacity for fry production. High over- winter mortalities appear to have occurred shortly after hatching, probably from critical levels of dis- solved oxygen in intragravel water. Critical oxy- gen levels apparently developed under average winter streamflow conditions due to the high biochemical oxygen demand placed on the streambed by high egg and alevin densities. 6. Recently hatched dead alevins disappear rapidly within the streambed because of biochem- ical decomposition and invertebrate scavenging. In comparison with dead alevins, dead eggs disap- pear slowly. In Sashin Creek, insect larvae and a planarian, probably Polycelis borealis, may be particularly important in removing dead salmon eggs and alevins from the streambed. 7. Several aspects of the historical patterns of pink salmon fry production in Sashin Creek suggest that streambed overseeding occurred in 1967. Fry production from the 1967 spawners falls on the descending limb of the fry production curves, both for the stream as a whole (since 1940) and for the individual stream areas (since 1961). From the historical pattern of time of adult entry and resulting freshwater survival, freshwater survival of 1967 brood year progeny should have been around 18% (or a production of 8 million fry). Survival of progeny during spawning and between spawning and hatching was adequate to reach these predicted levels. Overwinter mortalities (be- tween hatching and emergence), however, were higher than any previously recorded. Compensa- tory losses during this period were probably due to the presence of too many eggs and alevins in the gravel for existing environmental conditions — streambed overseeding. 8. Overseeding does not invariably occur at some precise density of eggs, but rather is a dynamic interaction between densities of eggs and alevins in the gravel, certain ecological charac- teristics that define the fry production capability of the streambed, and the prevailing climatologi- cal features during the 6- to 8-mo period eggs and alevins reside in the streambed. ACKNOWLEDGMENTS The following people assisted with field studies on the 1967 spawners and freshwater survival of the progeny: David Brickell, Robert Coats, Richard Crone, Calvin Fong, Henry Kopperman, Derek Poon, and Roger Winchester. LITERATURE CITED Brickell, D. C. 1971. Oxygen consumption by dead pink salmon eggs in salmon spawning beds. M.S. Thesis, Univ. Alaska, Col- lege, 53 p. BRIGGS, J. C. 1953. The behavior and reproduction of salmonid fishes in a small coastal stream. Calif. Dep. Fish Game, Fish Bull. 94, 62 p. 1955. Behavior pattern in migratory fishes. Science (Wash., D.C.) 122:240. Chapman, D. W., and T. C. Bjornn. 1969. Distribution of salmonids in streams, with special reference to food and feeding. In T. G. Northcote ( editor), Symposium on salmon and trout in streams, p. 153-176. H. R. MacMillan Lect. Fish. Inst. Fish. Univ. B.C., Van- couver. CLAIRE. E. W., AND R. W. PHILLIPS. 1968. The stone^y Acroneuria pacifica as a potential pred- ator on salmonid embryos. Trans. Am. Fish. Soc. 97:50- 52. Ellis, R. J. 1969. Return and behavior of adults of the first filial gen- eration of transplanted pink salmon, and survival of their progeny, Sashin Creek, Baranof Island, Alaska. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 589, 13 p. 1970. Alloperla stonefly nymphs: Predators or scavengers on salmon eggs and alevins? Trans. Am. Fish. Soc. 99:677-683. 581 FISHERY BULLETIN: VOL 76. NO. 3 KKNK, R. 1953. The fresh-water triclads (Turbellaria) of Alaska. Proc. U.S. Natl. Mus. 103:163-186. McDonald. J. G. I960. A possible source of error in assessing the survival of Pacific salmon eggs by redd sampling. Can. Fish Cult. 26:27-30. McL.XKNEY. W. 0. 1967. Intra-stream movement, feeding habits, and popula- tion of the coastrangesculpin,Co»(/sa/e!^?;cKS. in relation to eggs of the pink salmon, Oncorhynchus gorbuscha. in Alaska. Ph.D. Thesis, Univ. Michigan, Ann Arbor, 131 P- McNeil, W. J. 1964a. A method of measuring mortality of pink salmon eggs and larvae. U.S. Fish Wildl. Serv., Fish. Bull. 63:575-588. 1964b. Redd supenmposition and egg capacity of pmk salmon spawning beds. J. Fish. Res. Board Can. 21:1385-1396. 1966. Distribution of spawning pink salmon in Sashin Creek, southeastern Alaska, and survival of their prog- eny. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 538, 12 P- 1968. Migration and distribution of pink salmon spawners in Sashin Creek in 1965, and survival of their prog- eny. U.S. Fish Wildl. Serv., Fish. Bull. 66:575-586. 1969. Survival of pink and chum salmon eggs and alevins. In T. G. Northcote (editor). Symposium on salmon and trout in streams, p. 101-117. H. R. MacMillan Lect. Fish. Inst. Fish. Univ. B.C., Vancouver. McNeil. W. J., S. C. Smedley. and R. J. Ellis. 1969. Transplanting adult pink salmon to Sashin Creek, Baranof Island, Alaska, and survival of their prog- eny. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 587, 9 p. McNeil. W. J., R. A. Wells, and D. C. Brickell. 1964. Disappearance of dead pink salmon eggs and larvae from Sashin Creek, Baranof Island, Alaska. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 485, 13 p. MERRELL, T. R., Jr. 1962. Freshwater survival of pink salmon at Sashin Creek, Alaska. In N. J. Wiiimovsky (editor). Symposium on pink salmon, p. 59-72. H. R. MacMillan Lect. Fish. Inst. Fish. Univ. B.C., Vancouver. Moyle. p. 1966. Feeding behavior of the Glaucous-winged Gull on an Alaskan salmon stream. Wilson Bull. 78:175-190. NEAVE. F. 1953. Principles affecting the size of pink and chum salm- on populations in British Columbia. J. Fish. Res. Board Can. 9:450-491. Nicola. S. J. 1968. Scavenging by Alloperla (Plecoptera: Chloroper- lidae) nymphs on dead pink (Oncorhynchus gorbuscha I and chum lO. ketat salmon embryos. Can. J. Zool. 46:787-796. Reed. R. J. 1967. Observation of fishes associated with spawning salmon. Trans. Am. Fish. Soc. 96:62-67. RICKER. W. E. 1962. Regulation of the abundance of pink salmon popula- tions. In N. J. Wiiimovsky (editor). Symposium on pink salmon, p. 155-201. H. R. MacMillan Lect. Fish. Inst. Fish. Univ. B.C., Vancouver. SKUD, B. E. 1958. Relation of adult pink salmon size to time of migra- tion and freshwater survival. Copeia 1958:170-176. Stuart. T. A. 1953. Spawning migration, reproduction and young stages of loch trout (Sa/wo truttaL.) . Scott. HomeDep., Fresh- water Salmon Fish. Res. 5, 39 p. U.S. GEOLOGICAL Survey 1969. Water resources data for Alaska 1968. Part 1. Sur- face water records. U.S. Geol. Surv., Water Resour. Div., 155 p. Wells. R. a., and W. J. McNeil. 1970. Effect of quality of the spawning bed on growth and development of pink salmon embryos and alevins. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 616, 6 p. WICKETT, W. P. 1954. The oxygen supply to salmon eggs in spawning beds. J. Fish. Res. Board Can. 11:933-953. 1962. Environmental variability and reproduction poten- tials of pink salmon in British Columbia. In N. J. Wiiimovsky (editor). Symposium on pink salmon, p. 73-86. H.R. MacMillan Lect. Fish. Inst. Fish. Univ. B.C., Vancouver. 582 VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES OF PELAGIC CEPHALOPODS FROM HAWAIIAN WATERS Richard Edward Young * ABSTRACT Vertical distribution data were obtained for 47 species ofpelagiccephalopods off Oahu, Hawaii. Peaks in species richness occurred at 500-800 m during the day and in the upper 300 m at night. Over SO^f of the individuals occurred in the upper 250 m at night. Approximately 60% of the species underwent diel vertical migration, and most of these migrated into the upper 250 m. In five of nine groups of closely related species, clear differences in habitat were found. Deepwater spawning appeared to occur in a variety of cephalopods. Two of the bathypelagic octopods brooded their young at or above the upper limit of the remaining adult population. In doing so, the extent of the upward migration of newly hatched individuals was reduced. Photosensitive vesicles occurred in all species. These organs probably detect downwelling daylight for regulating vertical migration and counterillumination. The vesicles also appeared to form an elaborate system for monitoring bioluminescent light from the animal's own photophores, from within the mantle cavity, and from other animals located outside the visual field. Cephalopods must occupy a wide variety of ecolog- ical roles in the pelagic realm of the open ocean: the highest diversification of families and genera is found in this environment. In order to under- stand these roles, the vertical distribution of these animals must be determined. A number of papers have treated various aspects of the vertical dis- tribution of oceanic cephalopods (e.g., Pearcy 1965; Clarke 1969; Roper 1969; Gibbs and Roper 1971; Clarke and Lu 1974, 1975; Lu and Clarke 1975a, b; Roper and Young 1975). Their vertical habitats, however, remain poorly known. Data on the vertical distribution of cephalopods is difficult to obtain: many species are uncommon, and some avoid small trawls. In this study an opening-closing net (modified Tucker trawl) pro- vided unambiguous depth data, and a slightly larger open net (3-m Isaacs-Kidd midwater trawl) added considerable additional data; nevertheless, fast-swimming species were poorly sampled. Extraocular photoreceptive organs, the photo- sensitive vesicles, were examined in each species for clues that would indicate the role of light in regulating vertical distribution patterns. The or- gans in squid, known as the parolfactory photo- sensitive vesicles, lie near the brain within the confines of the cephalic cartilage. In octopods the organs, known as epistellar photosensitive vesi- cles, lie within the mantle cavity adjacent to the 'Department of Oceanography, University of Hawaii, Hono- lulu, HI 96822. stellate ganglia. The photosensitive vesicles are paired organs. Each organ, as the name implies, is generally composed of a large number of small vesicles. The individual vesicles contain photo- sensitive cells similar to those of the retina, and their photoreceptive nature has been well estab- lished ( Nishioka et al. 1966; Mauro and Baumann 1968; Mauro 1977). The specific functions of the photosensitive vesicles are unknown in both neri- tic and oceanic cephalopods although many suggestions have been made (see Packard 1972). Several papers discussing the relationship of vertical distribution to eye structure, biolumines- cence and/or development of photosensitive vesi- cles in selected species have already appeared (Young 1972a, 1973, 1975a, c, d, 1977). Some data on distribution taken during the initial phases of this program have been published by Roper and Young (1975). This paper examines the vertical distribution of all pelagic cephalopods taken off Hawaii and the morphology and orientation of their photosensitive vesicles. MATERIALS AND METHODS Specimens were collected off the island of Oahu in the Hawaiian archipelago at long. 158°20'W, lat. 21 = 20 'N over depths between 1,500 and 4,000 m. Collections were made from September 1969 to November 1974 primarily from the RV Teritu. Over 3,300 specimens were taken in horizontal tows during about 1,000 h of trawling time. Manuscript accepted January 1978. FISHERY BULLETIN; VOL" 76. NO. 3, 1978. 583''^ FISHERY BULLETIN: VOL. 76. NO. 3 Cephalopods were collected primarily with two types of nets: a 3-m opening-closing modified Tucker trawl and a 3-m Isaacs-Kidd midwatej- trawl (IKMT). Details of the trawling with the Tucker trawl are given by Walters (1976). When the Tucker trawl failed to close or close com- pletely, the trawl was considered an open tow. Tows usually were made at 5 to 6 km/h for a period of 3 h. Twilight periods were generally avoided. Tows made with net closed indicated the catch contained almost no contamination. Contamina- tion from previous tows was minimized by care- fully cleaning the net after each tow. The trawl tended to wander vertically when open; this was most severe in deep tows. During the latter part of the program, acoustic depth telemeters allowed trawl depths to be continuously adjusted and greatly reduced wandering. The distribution figures indicate the extent of this wandering. Trawl depths usually were determined with a time-depth recorder attached to the trawl. Clarke ( 1973) discussed trawling methods with the IKMT. The trawl was lowered quickly then towed horizontally at 5 to 6 km/h for usually 2 h. Retrieval was rapid with the ship moving slowly ahead. Vertical wandering of the net was not as serious as with the Tucker trawl. All specimens captured with the IKMT were assumed to come from the modal trawling depth of the net, or if no clear mode was present, from the midpoint of the effective vertical range of the tow. The occasional capture of a specimen during setting or retrieval of a net results in an anomalous depth record below the animal's normal habitat. Contamination of the catch by animals from previous tows occasion- ally occurred with the IKMT. This contamination is especially serious as the error may be impossible to detect. IKMT data for a few of the most abundant species are presented both as catch per trawling effort and as actual catch figures (Table 1). The remaining distribution figures are designed to show animal size vs. depth relationships and to indicate the precision and reliability of the data (e.g., fishing range of the tow, open or opening- closing tow). As a result, corrections in the data for unequal sampling at various depths could not be made. This bias was especially critical at depths <400 m during the day and at depths > 1,000 m during the day and night where sampling was low. The magnitude of this error can be determined from Table 2, which lists sampling time in each 100-m depth interval. Depth data for most species taken over the en- tire trawling period have been combined. There- fore, short-term variation in depth distributions may be obscured. Where sufficient data exist to determine general distribution patterns based on Tucker trawls alone, these data are presented separately. For species with insufficient data, data from both trawls are combined in the figures. In most cases larvae, which usually have a different vertical distribution than adults, have been excluded from the distribution figures and the Table l. — Depth distribution, capture rates, and numbers of the most abundant cephalopod species captured by the Isaacs-Kidd midwater trawl. Day captures for Pterygioteuthis giardi are included in Figure 4. R = capture rate in numbers per 1,000 m^ of water sampled. N = actual number captured. ND = no data. Day Night m o CO X CO CO CO n ,co X CO 3 am Oi b Q.CD O . Q) e 3 PTi 1^ o 5) ^ O o 0) o en O CO 1^ I 0) 51 S u. i: Cl> 5: E u. 5: Depth (m) R N R N R N R N R N R N R N 0-50 ND ND ND ND ND ND 5.4 14 15.2 39 1.5 4 5.8 15 50-100 ND ND ND ND ND ND 17.9 52 62,1 180 6.5 19 7.9 23 100-150 ND ND ND ND ND ND 3.2 8 23,3 57 7.3 18 2.0 5 1 50-200 0 0 0 0 0 0 5,2 14 14,1 38 15.9 43 2.2 6 200-250 0 0 0 0 0 0 1,0 1 0 0 3.1 3 0 0 250-300 ND ND ND ND ND ND 1.0 1 1-0 1 6.5 6 4.3 4 300-400 3.0 4 2.3 3 6.9 9 0 0 0 0 0 0 0 0 400-500 10.9 14 42.3 54 14.1 18 49 3 4.9 3 4.9 3 0 0 500-600 7.4 9 72.2 87 20.7 25 3,4 2 6,9 4 5.1 3 0 0 600-700 22.8 26 8.7 10 14,9 17 0 0 1,4 1 0 0 0 0 700-800 1.0 2 5.3 10 3.2 6 28 3 0 0 0 0 0 0 800-900 3.6 3 8.5 7 1.2 1 0 0 0 0 0 0 0 0 900-1,000 2.1 2 1.0 1 2.1 2 0 0 0 0 0 0 0 0 1,000-1,100 1.5 1 20.1 13 1.5 1 0 0 0 0 5.4 1 0 0 584 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES Table 2. — Trawling time in minutes. Since trawls of two different sizes were used, a correction factor of 0.6 was applied to the trawling times of the Tucker trawl to compute the adjusted total trawling time. This factor represents the approximate difference in effective mouth areas of the two nets. Open tows = timeofeach tow was assigned to one depth. Opening-closing tows = time of each tow was apportioned among depth zones traversed by the trawl. IKMT = Isaacs-Kidd midwater trawl. Tucker trawl 1 IKMT Total adjusted to IKMT Open Opening-closing Day Night Open Depth (m) Day Night Day Night Day Night 0-50 1.091 939 2,562 3,780 50-100 180 1,112 614 2.897 108 3,932 100-150 781 27 591 2.442 16 3,265 1 50-200 144 530 84 870 130 2,689 267 3,529 200-250 536 186 320 177 944 289 1,458 250-300 146 31 871 911 106 1,434 300-350 180 203 514 452 605 682 913 350-400 180 180 460 407 839 552 1,223 904 400-500 360 502 1,529 1,413 1,276 605 2,409 1,754 500-600 376 1,748 927 1,204 577 2,253 1,359 600-700 133 1.683 1,420 1,139 714 2,149 1,646 700-800 313 220 1.638 838 1,862 1,052 3,033 1,687 800-900 1,244 519 820 179 1,566 490 900-1,000 180 30 709 184 917 179 1,450 307 1,000-1,100 182 464 64 646 182 924 330 1,100-1,200 195 230 156 180 195 435 289 1,200-1,300 200 300 380 180 528 180 1.300-1,400 10 256 234 98 146 212 1,400-1,500 67 106 40 64 1.500-1,600 30 18 1.600-1,700 8 38 5 23 1,700-1,800 104 267 62 160 1,800-1.900 84 228 50 137 1.900-2,000 48 28 29 17 2.000-2.100 15 27 9 16 2.100-2.200 72 29 43 17 2,200-2.300 43 33 26 20 2.300-2.400 8 5 text. Specimens captured during twilight periods, with a few exceptions, have also been excluded from the charts. Species examined are listed in Table 3. Larvae or juveniles of several additional species were cap- tured but are not included in this study. These are: Tremoctopus violaceus, Argonauta sp., Cranchia scabra, Thysanoteuthis rhombus, Onykia sp. One pelagic species reported from Hawaii by Berry (1914), Iridoteuthis iris, was not taken. This species belongs in the genus Nectoteuthis and probably lives in association with the ocean floor. ^ Photosensitive vesicles of most species were sectioned. Material was fixed either in glutaraldehyde-osmium tetroxide or Bouin's so- lution and was embedded in Epon 812^ or paraffin. All vesicles sectioned contained cells with photo- sensitive processes and, therefore, appeared to be functional. In only a few cases did the general histology of the organs add to our understanding of ^Roper, C, and R. Young. Review of the Heteroteuthinae. Unpubl. manuscr. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. their function. As a result, histological details are not included for most species. In order to quantify the size of vesicles, an at- tempt was made to obtain dry weights. Many types of vesicles proved difficult to remove and clean completely. Vesicles from a series of similar-sized Pyroteuthis addolux, which are easily removed, were weighed and found to vary by a factor of 1.5. Because of the large individual variations and in- accuracies due to difficulties in isolating many types of vesicles, this method of quantification was abandoned. As a result, camera lucida drawings of photosensitive vesicles provide the only measure of organ size: their relative size can be approxi- mately determined by comparison with the brain size. RESULTS Pyroteuthis addolux Young 1972 Vertical Distribution During the day, 39 specimens captured by the Tucker trawl indicate a vertical range for this 585 FISHERY BULLETIN: VOL. 76, NO. 3 Table 3. — Species of cephalopods considered. Order Teutholdea Family Enoploteuthidae Pyroteuthis addolux Young 1 972 Pterygioteuthis microlampas Berry 1913 Pterygioteuthis giardi Fischer 1895 Abralia tngonura Berry 1913 Abralia astrosticta Berry 1909 Abraliopsis sp. A. n.sp being described by L, Burgess Abraliopsis sp. B, n.sp. being described by L, Burgess Abraliopsis sp. C, n.sp. being described by L. Burgess Enoploteuthis sp. A. n.sp, being described by L. Burgess Enoploteuthis sp B, n.sp. being described by L. Burgess Thelidioteuthis alessandnnii (Verany 1851) Family Ommastrephidae Symplectoteuthis oualaniensis (Lesson 1830) Hyaloteuthis pelagicus (Bosc 1802) Notolodarus hawaiiensis (Berry 1912) Family Histioteuthidae Histioteuthis dofleini (Pfeffer 1912) Histioleuthis celetana pacifica (Voss 1962) Histioteuthis sp, under study by N. Voss Family Neoteuthidae Neoteuthis sp Family Bathyteuthidae Bathyteuthis abyssicola Hoyle 1885 Family Ctenopterygiidae Ctenopteryx siculus (Verany 1851) Family Onychoteuthidae Onychoteuthis compacta (Berry 1913) Family Octopoteuthidae Octopoteuthis nielseni Robson 1948 Family Cycloteuthidae Cycloteuthis serventyi Joubin 1919 Discoteuthis laciniosa Young and Roper 1969 Family Brachioteuthidae Brachioteuthis sp. Family Chiroteuthidae Chiroteuthis n.sp., being described by Roper and Young Chiroteuthis picteti Joubin 1 894 Chiroteuthidae ngen , n.sp. being described by Roper and Young Planktoteuthis lippula (Chun 1908) Gnwatditeuthis bomplandi (Verany 1837) Family Mastigoteuthidae Mastigoteuthis famelica (Berry 1909) Mastigoteuthis inermis Rancurel 1972 Family Joubiniteuthidae Joubiniteuthis portieri (Joubin 1912) Family Cranchiidae Liocranchia valdiviae (Chun 1906) Liocranchia reinhardti (Steenstrup 1856) Leachia pacifica (Issel 1908) Phasmatopsis fishen (Berry 1909) Taonius pavo (LeSueur 1821) Sandalops melancholicus Chun 1906 Helicocranchia beebei Robson 1948 Bathothauma lyromma Chun 1 906 Order Octopoda Family Bolitaenidae Eledonella pygmaea Verrill 1884 Japetella diaphana Hoyle 1885 Family Amphitretidae Amphitretus pelagicus Hoyle 1885 Family Vitreledonnelidae Vitreledonnella nchardi Joubin 1918 Order Vampyromorpha Family Vampyroteuthidae Vampyroteuthis inlernalis Chun 1903 Order Sepioidea Family Sepiolidae Heteroteuthis hawaiiensis (Berry 1909) species of 375 to 510 m; most captures came from 450 to 500 m (Figure 1). IKMT data lumped into 100-m increments show most day captures be- tween 400 and 700 m (Table 1 ). At night, 38 of the 41 specimens captured by the Tucker trawl indi- cate a vertical range of about 110 to 225 m; most specimens camp from 150 to 200 m. Three speci- mens were captured during the night in opening- closing tows near their day habitat at depths be- tween 360 and 480 m. Each of these three speci- mens was taken in a separate tow during a cruise in November 1972 within a few days of new moon. Although the upper 200 m was not sampled on this cruise, these captures indicate that at least part of "1 — I — I — I" 200 400 Q600 SOD 1000 I I I I r T r X 1 A O ■ '■O O ji ■ A oo ' o ' o ,iO O 'OOO. >0O O O;; 0|0* :t t ft o o o 9 9 90 o 10 22 26 30 34 MANTLE LENGTH ,m m 38 42 50 Figure. L— Vertical distribution of Pyroteuthis addolux. Symbols for Figure 1 and subsequent figures: open circles represent day captures; closed circles represent night captures. A bar with a circle indicates an opening-closing tow with the bar representing the depth range of the tow and the circle the most likely depth of the capture (the modal depth, or if no clear mode is present, the midpoint of the vertical range of the tow). A circle without a bar indicates a capture in an open tow, A bar without an associated symbol indicates an open oblique tow. Such bars do not always intersect the zero depth 1 ine as some gradual oblique tows were made between specific depths. Solid bars represent night captures. Dashed bars represent day captures, A small dot represents a presumed contaminant. 586 YOUNG; VKRTirAl. DISTRIBLTION AND PHOTOSKNSITIVK VESICLES the population was not migrating during this period. IKMT data lumped into 50-m increments show that most night captures were made between 50 and 200 m with peak catches between 150 and 200 m. Photosensitive Vesicles (Figure 2 A) The organs are very similar to those described by R. E. Young (1977) \n Pterygioteuthis micro- lampas. Pyroteuthis addolux has two sets of or- gans. The dorsal organs (the more dorsal set) lie embedded in the posterodorsal wall of the cephalic cartilage and adjacent to the optic lobes of the brain. Each ventral organ lies deeply embedded in the posteroventral surface of the cephalic carti- lage. Except for a thin medial extension on each ventral organ, all organs are thick, compact, and approximately circular to square in outline. The histological structure of the dorsal and ventral organs is similar. The integument adjacent to both the dorsal and ventral organs lacks pigment and thereby forms distinctive "windows" for the pas- sage of light. Nerves from both dorsal and ventral organs enter the peduncle complex of the brain and their fibers disperse in the base of the peduncle lobe near its broad junction with the olfactory lobe. Pterygioteuthis microlampas Bcrr\ I91.'^ X'ertical Distribution (Figure 3) The vertical distribution has been described by R. E. Young ( 1977). During the day, 48 specimens captured with the Tucker trawl indicate a depth range of 450 to 575 m; 857r of the specimens were taken between 450 and 500 m. IKMT data lumped into 100-m increments (Table 1) show most day captures between 400 and 600 m. At night 56 specimens taken by the Tucker trawl indicate a depth range of 25 to 180 m; nearly 85*^ of the captures were made between 50 and 105 m. IKMT data lumped into 50-m increments indicate a range of 0 to 200 m with a strong peak in the 50- to 100-m depth zone. The night distribution was not affected by moonlight (R. E. Young 1977). D PV POST PS MASS FIGURE 2.— A. Photosensitive vesicles oiPyroteuthis addolux. This illustration and most subsequent drawings show a side view of the bram. The optic stalk has been cut (as indicated by cross-hatching) and the optic lobe removed. The esophagus can be seen passmg through the bram. Three major subdivisions of the bram are apparent (i.e., the supraesophageal mass, the posterior subesophageal mass, and the middle subesophageal mass.) A large nerve tract which extends anteriorly to the anterior subesophageal mass was cut (indicated by dotted line) and the latter portion of the brain is not shown. B. Photosensitive vesicles of A 6ra/iops;s sp. B. Abbreviations for Figure 2 and subsequent figures of photosensitive vesicles: AV. PV.— Anteroventral photosensitive vesicles; C. PV.— central photosensitive vesicles; DOR.— dorsal; D. PV.— dorsal photosensitive vesicles; ES.— esophagus; GILL— gill; H. RET, M,— head retrac- tor muscles; INT,— intestine; M. S, MASS-middle subesophageal mass of the brain; MV, PV.— midventral photosensitive vesicles; N.-nerve; OP. ST.-optic stalk; PED. L.-pedunclelobe; PIG. S.-pigment screen; POST.-posterior; P. PV. -posterior photosensitive vesicles; P. S. MASS— posterior subesophageal mass of the brain; PV.— Photosensitive vesicles; V. PV.— ventral photosensitive vesicles; S. MASS — supraesophageal mass of the brain; VENT. — ventral. 587 FISHERY BULLETIN: VOL 76. NO. 3 200- 400 E r a. LU Q600 800 1000 9? 12 "^I — ' — ^ — WMVrry y^r-- ^ ' * Mt^ nm* TTTT* ITTT ^T TT* T 13 •lo 00900 ^oo^oj ' ' ' ■ " I' '• ■ 4 _, looo 14 IS 16 17 18 MANTLE LENGTH. mf i? ^^ J L-v-l- 20 21 24 Figure 3.— Vertical distribution of Pterygioteuthis microlampas. From R. E. Young (1977). Symbols as in Figure 1. Photosensitive Vesicles The organs have been described by R.E. Young (1977). They are essentially the same as in Pyroteuthis addolux. Pterygioteuthis giardi Fischer 1895 Vertical Distribution (Figure 4) During the day 30 specimens captured from both trawls indicate a depth range from about 390 to 525 m; over 907^ of the captures were made between 390 and 450 m. One IKMT tow captured eight badly damaged specimens at 630 m, well below their zone of maximum abundance. The previous tow had captured six specimens at 390 m that were in excellent condition; specimens from the deeper tow probably are contaminants. The depth distribution of this species may be biased by the relatively low sampling effort between 350 and 400 m. At night, 39 captures with the Tucker trawl indicate a depth range of 15 to 180 m; over Ih'^c of the specimens came from 15 to 50 m. IKMT data lumped into 50-m increments (Table 1) show the maximum abundance at depths of 0 to 100 m. Photosensitive Vesicles The organs are essentially the same as in Pyroteuthis addolux. Ahralia trigotiura Berry 1913 Vertical Distribution (Figure 5) Fifty specimens were captured by both trawls. Excluding presumed contaminants, the day cap- tures were made between 390 and 650 m with nearly 809f between 450 and 560 m. At night cap- tures were made between 30 and 200 m with over 759^ between 50 and 100 m. 800 - Figure 4. — Vertical distribution of Pterygioteuthis giardi . Sym- bols as in Figure 1. 588 200 1000 T-' — r T r- 'O o o o _l I ■ . I 22 26 30 34 MANTLE LENGTH, mm 42 Figure 5. — Vertical distribution oiAbralia trigonura. Symbols as in Figure 1. YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES Photosensitive Vesicles (Figure 6) The arrangement of organs is similar to Ab- raliopsis sp. described by Young ( 1973). Four sets of organs are present; all lie adjacent to the cephalic cartilage. One set is located dorsally, one posteriorly, and two ventrally. Each dorsal organ is situated in a concavity of the cephalic cartilage at the posterodorsal edge of the head. The organ is compact, dorsoventrally flattened and circular to triangular in outline. The posterior organs are located on the posterior surfaces of the optic lobes. Each is approximately elliptical and very flat. The posterior organs have a strong yellow pigment which does not fade after fixation. Other organs contain an orange pigment that is lost after fixa- tion. The posterior organs lie immediately an- terior and lateral to the opaque liver and directly anterior to the attachment zone of the transparent head retractor muscles (Figure 6B). The ventral organs on each side consist of two, narrowly joined, flattened lobes. One of these, the anteroventral lobe, is located somewhat anterior and medial to the other. The anteroventral lobe has its medial and anterior ends in a deep depression of the cephalic cartilage. The anteromedial edge of each lobe nearly makes contact with its counterpart of p PV Figure 6. — A. Photosensitive vesicles oiAbralia trigonura . B. Ventral view of A. trigonura with portion of mantle removed. This illustration shows the relationship between the posterior photosensitive vesicles, the opaque liver (L), and the mantle cavity. Abbreviations as in Figure 2. the opposite side. The more posterior of the two lobes, the midventral lobe, lies between the ven- tral surface of the optic lobe and the cephalic carti- lage. Circular windows, similar to those described in Abraliopsis sp. (Young 1973), and characterized by a reduced number of chromatophores, are pres- ent above each dorsal organ. A large ventral win- dow, totally lacking chromatophores, lies on the ventral surface of the head above the funnel and below the ventral vesicles. Ahralia astrosticta Berry 1909 Vertical Distribution No specimens were captured during the present program. However, one specimen was taken in a gill net set overnight by T. Clarke on the bottom in 180 m. The National Marine Fisheries Service, NOAA, has captured 44 juveniles (7-38 mm ML (Mantle length)) in pelagic trawls between 10 and 130 m at night and 10 adults in a benthic shrimp trawl at 110 m at night near Hawaii. The type was captured in a bottom dredge between 354 and 650 m, presumably during the day. Roper and Young (1975) indicated that this animal lives near the ocean floor even when migrating. Photosensitive Vesicles The organs and associated windows are basi- cally the same as in A. trigonura. Abraliopsis sp. A Vertical Distribution (Figure 7) Sixty-seven specimens were captured. Exclud- O 600 - 800 - -r^ •• •* •• • • • •• • • • ' 1 • [ ■ — r • • • I I - • • • • • •• • - ■ - - o ■ 0 o _ o o o 0 0 o o o _ ■ o o 0 o o o 0 " - - - i- I 1 1 1 1 i 22 26 30 34 MAN TIE LENGTH, mm Figure 7. — Vertical distribution of Abraliopsis sp. A. Symbols as in Figure 1. 589 FISHERY BULLETIN: VOL. 76. NO. 3 ing presumed contaminants, specimens captured during the day came from depths of about 475 to 700 m; 80% were taken between 550 and 700 m. At night, captures were made between about 20 and 200 m; nearly 80% were taken in the upper 100 m. Photosensitive Vesicles The organs and associated windows have been described in detail by Young ( 1973); they are simi- lar to those of Abralia trigonura. Ahraliopsis sp. B Vertical Distribution (Figure 8) During the day, 23 specimens taken by the Tucker trawl indicate a depth range of 500 to 650 m; most captures came from 500 to 600 m. IKMT data lumped into 100-m increments indicate most specimens came from 400 to 700 m. At night, 19 specimens from the Tucker trawl probably came from depths between 50 and 100 m. IKMT data lumped into 50-m increments show a strong peak in the 50- to 100-m interval. A few IKMT captures were made as shallow as 15 m. Photosensitive Vesicles (Figure 2B) The dorsal and anteroventral organs are similar to other species of Abraliopsis and Abralia. The posterior lobes, however, are absent, and the mid- ventral lobes are enlarged and extended dorsally. In addition, a thin string of vesicles extends from each midventral lobe dorsally between the brain and the optic lobe to join with the dorsal lobe. The structure of this string is slightly variable, and in some specimens the vesicles found about at mid- T T ' ,( • } ' ^\ 1.'. 111. T — r- • — 1 1 • • 200 - - 400 - _ 600 o 66 o6 '■''0'' o b 6 o 6 0 f o o 9 too innn . 1 1 1 ' 1 1 1 ■ 1 1 brain level are slightly enlarged and elongate (Figure 2B). The yellow pigment characteristic of the posterior lobes in related species does not occur in any of the lobes of this species. Ahraliopsis sp. C Vertical Distribution (Figure 9A) Only 12 specimens were captured. During the day, five specimens were taken between 500 and 600 m. At night, all of the captures were made in the upper 100m. o 600 - * 1 •• ' — r ' 1 t \ - 0 D - - O - - > 1 1 . 1 - 20 30 to 50 MANTLE LENGTH, m m m ANTIE It NGTH MANTLE LENGTH, «« Figure 8. — Vertical distribution oi^ Ahraliopsis sp. B, Symbols as in Figure 1. Figure 9. — A, Vertical distribution ofEnoploteuthis sp. A (cir- cles ) and Enoploteu this sp.B( squares ) . B . Vertical distribution of Ahraliopsis sp. C (circles) and Thelidioteuthis allesandrinii (triangles). Symbols as in Figure 1. Photosensitive Vesicles The organs are similar to Abralia trigonura and Abraliopsis sp. A except that the posterior lobe is smaller, slightly more medially located, and con- tinuous with the midventral organ. This latter connection, however, does not have the yellow pigment that the posterior organ possesses. Also, a few scattered vesicles lie on the posterior margin of the nerve from the dorsal organ. Erioploteuthis sp. A Vertical Distribution (Figure 9B) During the day, two captures were made be- tween 500 and 600 m; and at night, three captures were made in the upper 100 m. Photosensitive Vesicles (Figure lOA) The organs are similar to Abralia trigonura 590 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES ^^ - — - D PV Photosensitive Vesicles (Figure I OB) Three sets of organs, dorsal, posterior, and ven- tral, consist primarily of a loose association of var- iously shaped, mostly independent vesicles (Young 1977). The organs are broad, flat struc- tures, with the greatest concentration of vesicles along the lateral margins of the organs. The or- gans lack yellow pigment. Family Ommastrephidae Symplectotenthis oualaniensis (Lesson 1830) PPV Figure lO.— A. Photosensitive vesicles oi Enoploteuthis sp. A. B. Photosensitive vesicles of Thelidioteuthis alessandrinii . Abbreviations as in Figure 2. with the following exceptions. The midventral organ has a more irregular shape, is less compact, and has a narrow connection with the posterior organ. The posterior organ is continuous with the dorsal organ via a strand of vesicles that extends over the optic lobe. Except for a short segment adjacent to the dorsal organ, this strand contains yellow pigment as does the posterior lobe. Enoploteuthis sp. B Vertical Distribution (Figure 9B) One specimen was captured during the day at 515 m and three were taken at night between 50 and 150 m. Photosensitive Vesicles The vesicles are the same as those of Enop- loteuthis sp. A except for some differences in the posterior organ. The posterior organ in Enop- loteuthis sp. B is more elongate, more medially located on the optic lobe, and lacks yellow pig- ment. Thelidioteuthis alesandrinii (Verany 1851) Vertical Distribution (Figure 9A) During the day, one specimen was taken in an opening-closing tow between 720 and 780 m. At night, three specimens were taken in open tows between 80 and 100 m. Vertical Distribution Except for larvae, only one specimen was cap- tured in the midwater trawls. This specimen was taken at night by the IKMT which fished at 100 m. This fast-swimming squid normally avoids our trawls. Members of this species are commonly seen at the surface at night around the night-light and a number have been dipnetted. Little is known, however, of their day distribution, al- though Young (1975b) had assembled evidence which indicates that they live in the upper few hundred meters but may descend on occasion to great depths. Photosensitive Vesicles (Figure IIA) Three sets of organs are present: a dorsal, cen- tral, and ventral set. The ventral organ lies within the cephalic cartilage at the posterior end of the head and immediately above the posterolateral portion of the funnel. It consists of a series of flat, D PV Figure ll. — A. Photosensitive vesicles of Symplecoteuthis oualaniensis. B. Photosensitive vesicles of //ya/o24 mm ML and all specimens captured above 600 m during the day. One hundred fourteen specimens are plotted. Although only a few shallow day captures were made, the vertical distribution pattern is clear: animals between 5 and 15 mm ML predominate in 599 FISHERY BULLETIN: VOL. 76, NO. 3 'tttt ♦ ? 9 ,o 4 i o •••o'w P? • o "" . 9 o o otPO° • 900 O 10 20 30 10 50 60 MANTLE LENGTH, mm Figure 29, — Vertical distribution of Liocranchia valdiviae. Symbols as in Figure 1. the upper few hundred meters. Descent to adult depths begins within the 5- to 15-mm ML size range or occasionally larger. Most specimens 15 to 25 mm ML are captured between depths of 500 and 700 m, while most animals >25 mm ML are found deeper than 700 m with progressively larger specimens found at progressively greater depths. Diel vertical migration does not occur. Five large specimens captured at depths of 40 to 525 m at night, however, indicate that some specimens oc- casionally wander into the upper depths at night. Mature specimens were not captured. Photosensitive Vesicles (Figure 30A) Liocranchia valdiviae has a single set of small organs. Each organ is elongate and extends along the posterior side of the optic stalk. Each organ usually consists of three elongated vesicles. A strip of dark brown screening pigment with ir- regular margins extends along much of the an- teromedial edge of the ventral half of the organ. The broad dorsal vesicle either lacks screening pigment or has only a trace of it. The slender middle vesicle has a narrow, often discontinuous strip of pigment which widens ventrally. The ven- tral vesicle, which is the largest, has a broad, con- tinuous layer of screening pigment. The vesicles of L. valdiviae grow allometrically. At 30 mm ML the vesicles are small, and screening pigment consists of a single small patch on the ventral vesicle. In the largest specimen ( 102 mm ML), the pigment screen is very extensive and covers much of the anterior surface of the dorsal as well as ventral portions of each organ. The dorsal and ventral vesicles in each organ are somewhat broader in this specimen, making the organ more dumbbell shaped. Liocranchia reinhardti (Steenstrup 1856) Vertical Distribution (Figure 31) All 12 juvenile specimens were captured at night. Ten of the 12 specimens were taken in the upper 100 m; the other 2 came from 150 to 200 m. A single mature specimen was captured at 775 m in a Tucker trawl that failed to close on retrieval. This specimen was a female that had recently spawned: remnants of what appeared to be sperm reservoirs were attached to the inner right wall of the mantle near the base of the funnel; the nida- mental glands were gelatinous and extremely swollen; the ovary was depleted; and the muscular tissue of the mantle, fins, head, and arms was flaccid. Unfortunately, there are no data on the day distribution of this species in Hawaiian waters. • • •• 1 1 1 1 •• T I 1 • 200 . _ 400 - - 600 - - 800 1 1 . 1 1 • 1 1 Figure 30. — A. Photosensitive vesicles of Liocranchia val- diviae. B, Photosensitive vesicles of L. reinhardti. Abbrevia- tions as in Figure 2, 90 130 170 MANTLE LENGTH, mm Figure 31, — Vertical distribution of Liocranchia reinhardti. Symbols as in Figure 1, 600 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES However, in the tropical North Atlantic, two specimens of L. reinhardti (44 and 48 mm ML) were captured during the day in opening-closing tows between 510 and 600 m, while a 75-mm ML specimen was taken in an open tow that fished between 390 and 800 m (C.C. Lu pers. commun.). Also, M. Clarke ( 1969) reported specimens of 46 and 69 mm ML from depths between 450 and 810 m during the day in the Atlantic. Photosensitive Vesicles (Figure 30B) The organs of L. reinhardti have been described by Messenger (1967a). This species has a single large set of organs lying along the posteromedial surface of the peduncle lobe. Each organ consists of a linear array of 20 to 25 tightly packed vesicles. The vesicles are elongated in a transverse direc- tion except for those at the dorsal and ventral ends which are nearly circular. The vesicles are sepa- rated from one another by a heavy brown pigment screen which also covers most of the convex an- terior side of the organ. Most vesicles within each organ thus form elongate cups which presumably admit light only from one surface. The dorsal vesi- cle, however, lacks screening pigment from the posterior lateral and dorsolateral surfaces. The ventral vesicle is larger, with photosensitive pro- cesses twice as long as those of the dorsal vesicle. It lacks pigment on its ventral and anterior surface. The curvature of the organ and the arrangement of screening pigment allows light to enter dif- ferent vesicles from a wide range of angles. Specimens ^47 mm ML have no screening pig- ment on the vesicles while those 2*70 mm ML exhibit pigment as described above. The vesicles in the largest specimen (spent female) are slightly larger (especially the ventral vesicle) relative to the brain size than in smaller specimens. Leachia pacifica (Issel 1908) Vertical Distribution (Figure 32) The vertical distribution of L. pacifica has been described elsewhere (Young 1975a). This species reaches about 809^ of its maximum length in near-surface waters. Large specimens (45-60 mm ML) are found throughout the water column be- tween 30 and at least 1,800 m with those taken from progressively deeper water exhibiting pro- gressively greater sexual maturity. Gravid females were taken at depths > 1,300 m. 200 400- 600 800 )000 6 )200 I t— a. Q )400 16001- 1800 2000 22001- 2400, ■ ■>■•' ' ' _ ; °8 0 • • o • 0 "■ 8 • °0 o 0 0 o • 1 • o 1 0 m * • — 1 — 1 1 1 1 1 1 1 1 1 1 1 1 10 20 30 40 50 60 MANTLE LENGTH, mm Figure 32.— Vertical distribution of Leachia pacifica. From Young ( 1975a). Symbols as in Figure 1. Photosensitive Vesicles (Figure 33A) Leachia pacifica has a single set of organs lo- cated on the posteroventral surface of the peduncle lobe. Each organ consists of 4 or 5 cup-shaped vesicles that are closely packed into a small oval organ. A dark brown screening pigment covers Figure 33. — A. Photosensitive vesicles of Leachia pacifica . B. Photosensitive vesicles of Sandalops melancholicus. Abbrevia- tions as in Figure 2. 601 FISHERY BULLETIN: VOL. 76, NO. 3 much of the anterior and slightly dorsal surface of each organ. This pigment is also found in the walls between some vesicles, tending to isolate them from one another. Although the vesicles are mi- nute in the larva and very small in the adults, a small positive allometric growth of the vesicles seems to occur. Screening pigment first appears on the vesicles between about 20 and 30 mm ML. In the adult, the screening pigment is most extensive and covers the entire anteromedial surface of the organ. Phasmatopsis fisheri (Berry 1909) Vertical Distribution (Figure 34) Over 300 specimens of P. fisheri were captured but most were larvae. Metamorphosis occurs at a size of 40 to 50 mm ML. During the day, six larvae were captured be- tween 150 and 250 m. Seventeen juvenile and adult specimens were captured between about 625 and 800 m; most captures were made between 650 and 775 m. At night, larvae «30 mm ML were taken primarily in the upper 50 m; larvae 31 to 40 mm ML were found throughout the upper 200 m. Four- teen juveniles and adults were taken at night be- tween 90 and 225 m; most captures were made between 100 and 200 m. Photosensitive Vesicles (Figure 35) Phasmatopsis fisheri has a single set of large organs. Each organ consists of a broad, elongate vesicle that extends from the optic gland on the dorsal surface of the optic stalk ventrally over the posterior surface of the peduncle complex onto the side of the ventral subesophageal mass, where it A PIG PIG.S Figure 35. — Photosensitive vesicles oiPhasmatopsis fisheri. A. Larva, 35 mm ML. B. Juvenile, 70 mm ML. C. Adult, 130 mm ML. Abbreviations as in Figure 2. bends slightly dorsally. Each organ is thick later- ally and medially. Most of the anterior, medial, and lateral surfaces of each organ are covered by dark brown pigment screen. The dorsal tip of each organ lacks screening pigment on its lateral por- tion and has limited pigment screen on its medial portion. The curvature of each organ allows light to enter various parts from a wide range of angles. The anterior wall of each organ consists of little more than a membrane backed by dense pigment. The posterior wall and the walls of the dorsal and ventral ends of each organ contain 4 or 5 layers of sensory-cell bodies. Sensory processes are longer (215 /Ltm) and thinner (inner diameter 2 to4/>tm) in the ventral parts of the organ than in the dorsal 200 - 4 00 - Figure 34.— Vertical distribution of Phasmatopsis fisheri. Symbols as in Figure L 0 600 - 800 m^//^)9om'////A^^ »!■ — ' — I — ' — I — ' — I — ' — t — ' — I — ' — r^'-T-'^ , • •«( • lOOo' I'll — ^- J J L 10 20 30 40 50 60 70 MANTLE LENGTH, mm 80 90 I *f \ \ I L. 100 130 170 602 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES parts (length 155 /xm, inner diameter 3 to 6 ju,m). The processes are long and slender and organized in a straight, parallel alignment. Each organ is small in larvae and lacks screen- ing pigment. At 35 mm ML, the vesicles form a narrow strip along the posterior surface of the peduncle lobe (Figure 35A). The largest larvae have relatively small organs without screening pigment; the youngest juveniles have large organs that are heavily pigmented. In the juvenile and adult stages, the organ exhibits positive allomet- ric growth (compare Figure 35B and C). In the adult stages, the organ exhibits positive allomet- ric growth (compare Figure 35B and C). In the brain. The ventral half of the organ is particularly enlarged and the organs on each side of the brain contact broadly (but do not fuse) below the ventral midline of the brain. Taonius pavo (LeSueur 1821) Vertical Distribution (Figure 36) The vertical distribution of T. pavo has been described by Young ( 1975d). Larvae probably live in the upper 400 m, although only one capture was made. Juveniles were found primarily between 600 and 650 m, and adults were captured between 725 and 970 m. Diel vertical migration does not occur. u -, ^ 1 1 (— ' T ' T ' I 1 ' < I 200 - • " eioo - • • , - ^=600 - b - o " 0 b o - 80C \r\ne\ 1 1 ■ , 1 1 1 1 1 1 i 0 o - 40 60 80 100 IJO HO 160 180 200 220 WANTIE LENGTH, mm Figure 36. — Vertical distribution ofTaoniuspavo. From Young U975d.). Symbols as in Figure 1. F1GL!RE 37. — Photosensitive vesicles of Taonius pavo . A. Lar- va. B. Juvenile, 140 mm ML. C. Adult, 220 mm ML. lumen is unoccupied. Sensory processes occupy about V5 (i.e., about 230 ^im) of the lumen diame- ter on the anterior, posterior, and dorsal sides and slightly more (about 300^t.m) on the ventral sides. The processes are loosely packed and intertwined to a large extent dorsally and more tightly packed ventrally. Inner diameters of the processes vary greatly from about 3 to 30 ^im. The wall on the dorsal half of the organ contains about two layers of sensory-cell bodies compared with about three layers ventrally. In a 140-mm ML juvenile, the dissected vesicle was also hollow, with an even thinner region of the lumen occupied by sensory processes. The organs exhibit positive allometric growth (Figure 37). Sandalops melancholicus Chun 1906 Vertical Distribution (Figure 38) The vertical distribution in this species has been reported by Young ( 1975d). Larvae were found in the upper 400 m. Juveniles were captured between 450 and 674 m, and two adults were captured near 800 and 1,075 m. Diel vertical migration does not occur. Photosensitive Vesicles (Figure 37) Taonius pavo has a single set of organs located on the posteroventral side of the peduncle com- plex. Each organ consists of a single oval vesicle. No screening pigment is present. The large size of the vesicle in a 220-mm ML specimen belies its internal structure. The large central region of the Photosensitive Vesicles (Figure 33B) Sandalops melancholicus has a set of organs located along the ventral surface of the peduncle complex. Each organ consists of a single bilobed vesicle (R. E. Young 1977). Slight positive allo- metric growth of the vesicles occurs between the juvenile and adult stages. 603 FISHERY BULLETIN VOL 76, NO 3 200- 400- -600- 800- 1000- 1200 - i. * \ 6 'c 1 'III i 0 OQO • ooj 0 • , 1 , 1 ' 1 1 1 • , 1 , 1 1 1 20 40 60 80 MANTLE LENGTH, mm 100 Figure 38. — Vertical distribution of Sandalops melancholicus. From Young ( 1975d). Symbols as in Figure 1. Helicocrarichia heehei Robson 1948 sively greater depths, although the relationship of size to depth is not very precise. The deepest cap- ture was probably at 1,200 m. Mature specimens were not captured. Photosensitive Vesicles (Figure 40A) One set of organs is present. Each organ consists of a single small oval vesicle located on the poste- rior surface of the peduncle complex. No screening pigment is present. Very slight, if any, positive allometric increase in the size of the vesicles oc- curs from juveniles to the largest specimens. Bathothanma lyromma Chun 1906 Vertical Distribution (Figure 41) Although only 12 specimens were captured, a general pattern of ontogenetic descent is evident. Vertical Distribution (Figure 39) Including larvae, 47 specimens were captured. Although day and night captures are not well in- termingled in Figure 39 (due largely to sampling inequities), the data indicate that this species does not migrate. Rather, it seems to undergo on- togenetic descent. The youngest specimens were captured between 100 and 200 m. Progressively larger specimens were generally taken at progres- FlGURE 40. — A. Photosensitive vesicles of Helicocrarichia beebei. B. Photosensitive vesicles of Bathothauma lyromma. Abbreviations as in Figure 2. 200- 400 6 600 X a. S 800- 1000- 1200- 1400 "1 1 1- I '^ —tfi* . • 4 .^i « • 10 20 30 40 50 60 MANTLE LENGTH, mm 1200- 20 40 60 80 100 MANTLE LENGTH, mm 140 Figure 3.9. — Vertical distribution of Helicocrarichia heehei. Symbols as in Figure 1. Figure 41. — Vertical distribution of Bathothauma lyromma. Symbols as in Figure 1. 604 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVK VESICLES Day and night captures were in the same depth range, indicating that diel vertical migration does not occur. The three specimens captured at the greatest depths were gravid females. The speci- men captured at 910 m had sperm receptacles im- bedded in the back of the head and in the an- terodorsal surface of the mantle. The nidamental and oviducal glands were greatly enlarged and the entire visceropericardial coelom was packed with large eggs. The muscular tissue was slightly flabby. The specimen captured at 1,125 m exhi- bited almost identical features. The specimen cap- tured at about 1,100 m had similarly placed sperm reservoirs, less extensively enlarged nidamental and oviducal glands, and lacked eggs (apparently due to damage during capture*. This specimen exhibited no sign of muscular degeneration. The mantle cavity of this specimen had two very long arms from another specimen (presumably a male) attached to the inner wall of the mantle. The largest specimen was an immature female. Its size was largely due to its fixation in a relaxed state. In this species, the pen is extraorinarily delicate and accurate measurements of contracted, crumpled specimens are nearly impossible. Photosensitive Vesicles (Figure 403) Bathothauma lyrommQ has a single set of or- gans. Each organ consists of a flat oval vesicle located on the posteroventral surface of the pedun- cle complex. No screening pigment is present. Slight positive allometric growth of the vesicles occurs from juveniles to adults. Galiteuthis pacifica (Robson 1948) Vertical Distribution (Figure 42) The 27 specimens captured indicate a broad ver- tical range for this species. Fourteen of the 19 captures of specimens >20 mm ML came from depths of 700 m or more. The data indicate that diel vertical migration does not occur. Photosensitive Vesicles The vesicles of this species are similar to those of G. phyllura described by Young ( 1972a). A single set of organs is present. Each organ consists of a large oval vesicle attached to the posteroventral surface of the peduncle complex. Considerable positive allometric growth of the vesicles occurs. 0| i r 200 400- E x" ■600 800- 1000- 1200 1 — ' — r 6oo o o 9 •■ o o 1350m .J I . 1 I 10 20 30 40 50 260 MANTLE LENGTH, mm FiGL'RE 42. — Vertical distribution of Galiteuthis pad fica . Sym- bols as in Figure 1. Order Octopoda Family Bolitaenidae Eleciouella pygmaea Verrill 1884 N'ertical Distribution (Figure 43) Eighty specimens were captured. Day and night captures were in the same depth range (except above 300 m where day trawling was minimal), indicating that diel vertical migration does not occur. Most specimens between 5 and 15 mm ML \^ o • • L». ,,.,,, J' f» -I 1 1 r -| r "nil '9 J 0, nt, 20 ;5 30 M AN T I E le NGTM, Figure 43. — Vertical distribution oi Eledonella pygmaea. Cir- cles with crosses represent brooding females. Double circle rep- resents a gravid female. Otherwise symbols as in Figure 1. 605 FISHERY BULLETIN: VOL. 76, NO. 3 were captured either around 200 m or below 600 m. Apparently the size at which young begin their descent to adult depths is rather variable. The deep captures exhibit a clear pattern of ontogene- tic descent. At 25 mm ML or larger all specimens (excluding brooding females) were captured be- tween depths of 975 and 1,425 m. Four females, apparently brooding, were captured between about 800 and 870 m. The pigmentation of the female changes as she becomes gravid: the chromatophores over the mantle and especially over the aboral surface of the arms and web become more numerous, and the oral surfaces of the arms and web develop an even denser pigmentation. Nearly all iridophores are lost. At the same time the arms and the web be- come thicker. The web between the dorsal six arms becomes more extensive, and the web between the two ventral arms is reduced. These dark octopods spawn and apparently brood their young (Young 1972b). Five specimens taken from horizontal tows exhibited this increased pigmentation. In four cases, the ovary was depleted, and in the fifth, captured at 1,400 m, the eggs were not fully ma- ture, but were considerably larger than in an im- mature female of approximately the same size. In two cases egg strings with developing embryos were found in the same trawl with dark and pre- sumably brooding females. No mature males were taken. However, judging from the development of the hectocotylus, the penis, and the spermatophore glands, two speci- mens captured at 1,200 and 1,425 m were nearly mature. Another slightly less mature specimen was taken at 1,325 m. Three still less mature specimens were taken between 1,175 and 1,200 m. while a large male taken at 1,025 m was the least developed of all. Photosensitive Vesicles (Figure 44 A) The photosensitive vesicles consist of a single pair of organs; each organ is a spherical vesicle attached to the posterior margin of the stellate ganglion. Japetella diaphana Hoyle 1885 Vertical Distribution (Figure 45) Seventy-four specimens were captured. Diel vertical migration does not occur. Specimens <20 mm ML were captured mostly in two regions, be- tween 170 and 270 m and between 500 and 800 m. Oj 1 r— — I— — r- <^ • -1 1— I 1 1 r ,t. Wo .w? % * o 9 II I . I 0 ® 20 30 40 50 60 70 MANTLE LENGTH, m • Figure 45. — Vertical distribution of Japetella diaphana. Cir- cles with crosses represent brooding females. Double circles rep- resent gravid females. Otherwise symbols as in Figure 1. *"^. \ STG STG Figure 44. — A. Section through the photosensi- tive vesicle of adult Eledonella pygmaea. B. Photosensitive vesicles ofAmphitretuspelagicus. ST. G. — Stellate ganglion. Otherwise symbols as in Figure 1. B 606 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES where they exhibited an ontogenetic descent. The depth range for specimens 5^20 mm ML was 725 to 1,065 m; nearly 909^ of the animals occur between 700 and 950 m and nearly 60'7f between 750 and 850 m. Gravid and brooding females were found at the extremes of this range. Two gravid females were captured at 1,050 and 1,065 m while three spent and presumably brooding females were taken between 725 and 800 m. As in£'. pygmaea, the gravid and spent females have a very heavy pigmentation and lack most of the iridophores present in younger specimens. Five such females were captured in horizontal tows. One gravid female with a sperm mass embedded in the gelatinous tissue between the second and third arms was taken at 1,050 m. Another taken at 1,065 m had been gutted in the trawl but had not spawned: the musculature was firmer than in spent females, and the catch contained a large number of octopod eggs which undoubtedly came from the ruptured ovary. Three specimens taken between 725 and 800 m probably had spawned: two had depleted ovaries and the third was gutted but had deteriorated musculature. In the same tow with the last specimen were four newly hatched larvae, presumably from the brood of the female. One large, heavily pigmented female taken in an oblique tow had the remnants of an egg string dangling from one of the large suckers of the third arm. Two eggs were completely en- gulfed by the sucker, while a third dangled from the broken egg string extending from the sucker. No mature males were taken. 0 • 1 . 1 1 300 - I - 400 - - • - 600 " ° i - ■ • • 00 - • ■ • 1000 , — 1 — , — 1 1 MANTiE LENGTH 0 > 1 ' I ' I - 200 - — 400 - - 600 - - 8 00 _ O • _ . o . 1000 • o - 1200 _ • ■ ,i\ , 1 1 - 20 40 60 MANTLE \.l NGTH.m m Figure 46. — A. Vertical distribution of Amphitretus pelagicus (squares) and Vitreledonella richardi (circles). B. Vertical dis- tribution ofVampyroteuthis infernalis . Half-closed circles repre- sent a twilight capture. Otherwise symbols as in Figure 1. Photosensitive Vesicles The vesicles are as in E. pygmaea. Family Amphitretidae Amphitretus pelagicus Hoyle 1885 Vertical Distribution (Figure 46A) Two specimens were taken at night in the upper 350 m. Photosensitive Vesicles (Figure 44B) Amphitretus pelagicus has one set of organs. They lie on the stellate ganglia immediately an- terior to the entry points of the pallial nerves. Each organ consists of a large complex of a dozen or more generally circular vesicles which cover most of the anterior wall of the ganglion. Family Vitreledonnelidae Vitreledonnella richardi Joudin 1918 Vertical Distribution (Figure 46A) Four specimens were captured. One small specimen was taken in an oblique twilight tow between the surface and 400 m. Two other small specimens were taken between 600 and 650 m during the day. One large specimen was captured at 775 m during the night. Photosensitive Vesicles An organ consisting of a single spherical vesicle is located on the posterior margin of each stellate ganglion. Order Vampyromorpha Family Vampyroteuthidae Vampyroteuthis iuferualis Chun 1903 Vertical Distribution (Figure 46B). Eleven specimens were captured. Ten of the 11 were taken between depths of 800 and 1,200 m. The remaining specimen came from an open ob- lique tow that fished between 1,100 and 1.900 m. Diel vertical migration does not occur. 607 FISHERY BULLETIN: VOL. 76, NO. 3 Photosensitive Vesicles The vesicles have been described in detail by Young (1972a). One set is present. They are lo- cated in the dorsal wall of the mantle cavity at the base of the funnel. Each organ consists of a small cluster of spherical vesicles. Order Sepioidea Heteroteuthis hauaiiensis (Berry 1909) Vertical Distribution (Figure 47) The distribution of this species has been discuss- ed by R. E. Young ( 1977). During the day, speci- mens ssl7 mm ML were taken between 250 and 350 m; larger specimens were taken between 375 and 650 m. At night, most specimens <17 mm ML came from depths between 150 and 200 m; larger specimens were taken between depths of 110 and 550 m. Males and females mature at about 15-16 mm ML. Photosensitive Vesicles (Figure 48) Two sets of organs are present (R. E. Young 1977). The more dorsal set lies on the posterior margin of the peduncle complex and consists of a short and narrow string of tiny vesicles. An even 200 E =E 400 a. uj 600 SOOh ' 1 rau. ^10 •• -1 — ' — 1 — ' — r ' 1 ' 1 -1 — 9 1 ' • T 6 9 -r- . • • 6 -r 0 ' — T" ' • • o 00 1 ' 9 1 • . O 0 * ^ • O CD OOO 0 . . . % - " ' 1 1 , .1. . 1 , 1 .,. 1 1 1 i 1 1 1 1 ■ 1 1 1 10 12 U 16 18 MANTLE LENGTH, mm 20 22 Figure 47.— Vertical distribution of Heteroteuthis hawaiiensis. From R. E. Young (1977). Symbols as in Figure 1. 24 26 Figure 48. — Photosensitive vesicles oi^ Heteroteuthis hawaiiensis. In this figure the outline of portions of the head and mantle are superimposed to give a clear perspective of the peculiar arrangement of vesicles in this species. EYE— eye; FUN.— funnel; MAN.— mantle. Otherwise abbreviations as in Figure 2. MAN 608 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVK VESICLES narrower string of tiny vesicles extends ventrally around the eyes and joins a rather large but ex- tremely thin, loosely associated group of vesicles that lies over the lateral base of the funnel. DISCUSSION Vertical Distribution The numbers of cephalopod species taken in different 100-m depth zones for the upper 1,400 m showed a broad peak between 500 and 800 m dur- ing the day (Figure 49). An abrupt increase in the number of species near 400 m was obscured by the method of analysis: eight species occurred for the first time between depths of 375 and 450 m. To indicate faunal change, the number of species found for the first time in each zone (i.e., depth zones containing species upper range limits) were compared with zones where species found in lesser depths were absent for the first time (i.e., depth zones immediately below the lower range limits) (Figure 49). The peak at 700-800 m in the summed plot of species added and species lost indicates that many species dropped out in the 600-700 m zone and many were added in the 700-800 m zone. The chart also indicates that only two species were encoun- tered for the first time at 800 m or below. One was the poorly sampled Brachioteuthis sp. and the other was deep-living Vampyroteuthis infernalis. The data indicate peak species richness in the upper few hundred meters with relatively little change between 300 and 1,000 m during the night (Figure 49). Numbers of individuals in different depth zones in the upper 1,400 m (exclusive of young individu- als, captures in oblique tows, and contaminants) were also examined (Figure 50). During the day, the greatest abundance of individuals occurred between 400 and 700 m. This peak reflects the dominance of the enoploteuthids, especially Pyroteuthis and Pterygioteuthis spp. The high rate of capture in the 300- to 400-m zone was due in part to a few species whose upper limits extended slightly above 400 m. Nevertheless, an abrupt in- crease in number occurred in the 400- to 500-m zone. The rates of capture below 1,000 m were unreliable due to the small amount of trawling. The night data in the upper 400 m were lumped into 50-m increments due to greater control over trawling depths in near-surface waters. The largest catches at night were made in the upper 200 m. In this region two peaks were apparent (Figure 50). The peak in the 50- to 100-m zone was largely due to Pterygioteuthis microlampas, the NO A NO L A«L TOTAL NO NIGHT NO A NO L A' L TOTAL NO E 600 200 E- 500 0 10 0 10 0 20 0 20 NO SPECIES 0 10 0 10 0 10 0 20 NO SPECIES Figure 49. — Numbers of species versus depth. The histograms were based on species ranges from midwater trawl data. Data were lumped into 100-m depth increments and were not cor- rected for unequal trawling times at different depths. Data for some species were very meager. Young stages found in near- surface waters that can be distinguished by an abrupt change in habitat or by a metamorphosis have been eliminated from the figures. No. A — number of species added (i.e., found for the first time in a given depth zone). No. L. — number of species lost (i.e., absent from a given depth zone but present in the shallower zone). A + L — sum of two previous histograms. Total No. — total number of species in each depth zone. -1 — I — I — I — ' — I 0 40 80 120 0 40 80 120 NO. SPECIMENS/ 1000 MIN. TRAWLING Figure 50.— Total catch rate of numbers of cephalopod speci- mens from both trawls. 609 FISHERY BULLETIN: VOL. 76, NO. 3 most abundant species in the collection. The peak in the 150- to 200-m zone was largely due to Pyroteuthis addolux and Heteroteuthis hawaiien- sis; young Histioteuthis dofleini also contributed considerably. YoungEledonella were also found in this zone although they were excluded from the figures as "larvae." The zone between 200 and 700 m was sparsely inhabited at night. The peak between 700 and 1,000 m represented the deep nonmigrating popu- lation. The capture rate in this region was almost identical to the day capture rate at the same depths: deep-living migrators were few. The total rate of capture for the water column during the day was 459 specimens/1,000 min of trawling. Surprisingly, the total capture rate at night was only 309 specimens/1,000 min. This dif- ference was largely due to smaller-than-expected catches at night in the upper 200 m of the few most abundant species. The reason for the low night catches is unknown. Another estimate of the number of animals in the upper 250 m at night was obtained by assuming that the day peak from 300 to 700 m (minus the night catch at these depths) shifted into this upper zone at night (see below). On this basis, nearly 807c of the individuals occur- red in the upper 250 m at night. If one considers also the abundant ommastrephids which avoided midwater trawls but occurred in near-surface waters at night, then only a small percent of the total number of individuals would remain below 250 m at night. In many species, most of the population shifted upward at night. Such day-night differences existed in at least 25 of the 47 species examined, based on present data and literature records. Adequate data were lacking for ommastrephids and Neoteuthis sp. Therefore, where the vertical ranges are known, nearly 60% of the species ex- hibited diel shifts in habitat. Species not exhibit- ing diel migrations belonged primarily to the Cranchiidae ( seven species ) and the Octopoda ( five species). At least 18 of the 25 species that exhibited diel migrations occurred almost exclusively in the upper 250 m at night. These included all 11 enop- loteuthids, Liocranchia reinhardti, Phasmatopsis fisheri, Ctenopteryx siculus, Octopoteuthis neilseni, Brachioteuthis sp., Chiroteuthis sp., Onychoteuthis compacta, and young Heteroteuthis hawaiiensis. Two species for which the data were incomplete (Cycloteuthis serventyi and Chiroteuthis picteti) probably belonged to this category as well. Therefore, at least 80% of the migratory species occurred in the upper 250 m at night. Amesbury ( 1975) examined vertical zonation of midwater fishes during the day off Hawaii. He concluded that the water column could be divided into three regions: epipelagic, mesopelagic, and bathypelagic zones. The boundary between the epipelagic and mesopelagic zones occurred at about 400 m and was marked by a sharp increase in the numbers of individuals. This boundary ap- peared to apply equally well to cephalopods. The boundary between the mesopelagic and bathypelagic zones occurred at about 1,200 m. This boundary was marked by a noticeable de- crease in fish numbers and represented the greatest day depths of vertically migrating fish. This lower boundary was not applicable to cephalopods; there was no comparable decrease in numbers of individuals; and this depth seemed to be well below the range of vertical migrators. Amesbury further divided the mesopelagic zone into upper and lower zones with the boundary at about 650 to 700 m. Cephalopods exhibited maximum species turnover at about this depth, as well as changes in light-related adaptations in some species (Young 1975d). Although fish and cephalopod distributions differed with respect to the lower boundary, the distribution of cephalopods generally supported Amesbury's zo- nation. In spite of the rather small size of the collection, some evidence of vertical habitat separation among closely related species emerged. Three of the more abundant species belong to the subfamily Pyroteuthinae: Pyroteuthis addolux, Pterygio- teuthis microlampas. and P. giardi. In general body proportions and armature, P. microlampas was more similar to Pyroteuthis addolux than to its congener. During the day, these two species occupied the same depths around 500 m. At night, their populations peaked at distinctly different depths: Pterygioteuthis microlampas occurred primarily between 50 and 100 m, while Pyroteuthis addolux occurred primarily between 150 and 200 m. Although the data were less clear for Pterygioteuthis giardi, this species seemed to center around 400 m during the day and in the upper 100 m at night with about half of the popula- tion in the upper 50 m. Thus this species was shallower than its two relatives during the day and tended to be shallower at night, although broad overlap occurred with its congener. 610 YOUNG; VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES Two species of the genus Abralia occurred off Hawaii. A 6ra//a trigonura was a common vertical migrator in the area sampled of the open ocean, while A. astrosticta was never taken there. Ab- ralia astrosticta seems to be a vertical migrator that moves in close proximity to the ocean floor (Roper and Young 1975). Two species of Mastigoteuthis were taken. Both species shared the same day habitat at 700 to 800 m. At night, the population of M. inermis spread upward in the water column 400 or 500 m. Al- though the data were few, M. famelica appeared to spread little or not at all. One of the clearest cases of habitat separation of congeners occurred in Liocranchia . Liocranchia valdiviae was taken in lower mesopelagic depths during the day and it did not migrate. Liocranchia reinhardti was taken in near-surface waters at night and apparently occurred in upper mesopelagic depths during the day. Although the octopods Japetella diaphana and Eledonella pygmaea are placed in separate genera, they are very closely related. Both were taken in deep waters and did not migrate. The adults (ex- cept brooding females) were taken at distinctly different depths: E . pygmaea occupied depths from 975 to 1,425 m while J. diaphana occupied depths primarily from 700 to 950 m. Young stages prior to descent were found in near-surface waters. In this habitat E. pygmaea was captured primarily at 200 m or just above while J. diaphana was captured primarily below 200 m. Young stages of both species in the process of descent occupied depths of about 400 to 800 m or more. The data indicated, however, that at any given size, except for those just beginning descent, the young of £■. pygmaea occupied greater depths than the young of J . diaphana. In the geruis Abraliopsis three species were tak- en: Abraliopsis sp. A and Abraliopsis sp. C form the most closely related species pair. The available data show no obvious habitat differences. Al- though the more common Abraliopsis sp. A reached a considerably larger size than species C (43 mm ML vs. 33 mm ML), young individuals of species A, however, apparently cooccurred with species C of the same size. The day and night habitats of Abraliopsis sp. B were not separable from its two congeners. Three other groups of congeners were taken (i.e., in Enoploteuthis , Histioteuthis, and Chiroteuthis). No differences in habitats were found within these groups; however, the data were extremely sparse. Reproduction Young ( 1972b) presented evidence for brooding in Eledonella pygmaea (incorrectly reported as Bolitaena microcolyla) and suggested that brood- ing occurs in all pelagic octopods. Additional evi- dence from the present study substantiated the brooding habit for E. pygmaea. In addition, evi- dence indicating brooding in the octopod Jape^e//a diaphana was found. This species underwent changes at maturity similar to E. pygmaea. Further, newly hatched young have been found in the same trawl with spent females, and in one case the remnants of an egg string was found attached to an arm sucker of such a female. In both species, gravid or near-gravid females were taken only at the lower limits of the species' vertical range. Although mature males were not taken, those nearest maturity were also taken in the lower parts of the depth range. Apparently, mating takes place at the lower depth limits of the population. Brooding females, on the other hand, were found only at the upper limit of the adult population in J. diaphana and only well above the upper limit of the remaining adult population in E. pygmaea. The brooding females of both species occurred around 800 m. Presumably the females migrate upward to around 800 m either just before or just after spawning. The increased risk of pre- dation above 800 m probably pi-events the female from further upward movement: the numbers of fishes increase greatly above 800 m (Amesbury 1975), and visual detection of the large silhouette presented by a brooding female should be possible above about 750 to 775 m (Young and Roper 1977 ). The upward movement must be unrelated to feed- ing since brooding females do not feed (Young 1972b). The upward migration may simply de- crease the distance the young must travel after hatching to their larval habitat near 200 m. A number of cephalopods may spawn at the lower end of their depth range. Evidence for deep- spawning was found in several vertically migrat- ing species. A single spent female of Liocranchia reinhardti was captured at 775 m at night, well below its normal night habitat in the upper 200 m. A single gravid, mated female of Brachioteuthis sp. was captured at 1,125 m at night; its normal night habitat is in the upper 200 m. Heteroteuthis hawaiiensis migrated vertically and exhibited 611 FISHERY BULLETIN: VOL. 76. NO. 3 narrow vertical ranges day and night until sexual maturity was reached; a poorly defined ontogene- tic descent then ensued. Unfortunately, no other clues to spawning depth are known. The nonmigrating species that exhibit a gradual ontogenetic descent would be expected to spawn at the lower end of their range. Indeed, this appeared to happen in Bathothauma lyromma. The most dramatic example occurred in Leach i a pacified. Young (1975a) demonstrated that this species descends near the time of sexual maturity from near-surface waters to depths of 1,000 to 2,000 m to mate and spawn. Larvae of most pelagic oceanic cephalopods occur in near-surface waters. Upward migration of larvae would seem to be a formidable task for species that spawn at great depths. The deep- living octopods apparently carry their young partway up presumably to lighten this task. Perhaps squid egg masses are positively buoyant and float to the surface. There are a number of observations of egg masses of pelagic cephalopods floating at or near the ocean surface (see Clarke 1966). However, these have yet to be shown to belong to a deep-spawning species. Photosensitive Vesicles These vesicular organs were present in all Hawaiian pelagic cephalopods and they occurred in a great variety of shapes, sizes, and locations. In many squids, the organs were subdivided into as many as four sets of separate organs. In squid, the organs were always found within the confines of the cephalic cartilage and were located either on the optic stalk ( central organs ) or dorsal, posterior, or ventral to the optic stalk (dorsal, posterior, and ventral organs, respectively). The separate organs often faced different directions (i.e., their broadest surface faced a dorsal, posterior, or ventral direc- tion). These separate organs were frequently as- sociated with distinctive "windows" in the overly- ing skin bearing few if any chromatophores. Such windows seem to be unnecessary since most cephalopods can become quite transparent by con- traction of their chromatophores. The windows in combination with the more pigmented surround- ing skin, however, may restrict light to specific receptors and thereby improve the directionality of the organs. In a few cases (e.g., Phasmatopsis fisheri and Ctenopteryx siculus ), the organ was not subdivided but elongate and curved, allowing dif- 612 ferent portions of a single organ to face various directions. A directional response of each portion was insured either by heavy pigment (e.g., P. fiaheri) or silvery iridophores (C siculus) which shielded one surface of the organ. Not all species, however, had vesicular organs that could dis- criminate between dorsal, posterior, and ventral sources of light. Some species had undivided cen- tral organs (e.g., Sandalops melancholicus, Taonius pavcj ) without apparent screening devices which therefore are nondirectional. In others, the total area surveyed by a nondirectional organ was restricted by its cryptic position (e.g., Vam- pyroteuthis). Clearly not all cephalopods use these organs in the same way. General trends between organ size and habitat depth during the day occurred in these animals. Teuthoids and sepioids found in the upper 400 m (neritic species and young Heteroteuthis hawaiiensis) generally had small organs. Species found primarily between 400 and 700 m generally had large, complex organs. These included most enoploteuthids, histioteuthids, probably Dis- coteuthis laciniosa, Liocranchia reinhardti, and young Taonius povo. Between 700 and 800 m, species with large, complex organs (i.e., Cteno- pteryx siculus, Phasmatopsis fisheri, The- lidioteuthis allessendrinii, Cycloteuthis serven- tyi, probably large Octopoteuthis nielseni, adult Taonius pavo, and Galiteuthis pacificus, and Bathyteuthis abyssicola) cooccurred with species which had small organs (i.e., chiroteuthids, Mas- tigoteuthis, Grimalditeuthis bomplandii, large Liocranchia valdiuiae, and probably Neoteuthis sp.). Many of the small-vesicle species had ranges that extended well below 800 m, where they were joined by other small-organ species (i.e., Vam- pyroteuthis infernalis and probably Joubiniteuthis portieri). The relationship of organ size to habitat depth was especially marked in young Phasmatopsis fisheri. The epipelagic larvae of P. fisheri, which may grow to 40 and 50 mm ML had small central vesicles. Upon metamorphosis and descent to the adult day habitat (650-775 m), the organs became greatly enlarged (Figure 34). As growth con- tinued, however, a gradual positive allometric growth of the organs occurred without a clear in- crease in habitat depth. A number of species did not follow these general trends. Several cranchiids exhibited gradual on- togenetic descent; one of these {Helicocranchia beebei) had small organs, while the others {San- YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES dalops melancholicus andBathothauma lyromma) had large organs. Leachia pacifiva, which had small organs, spent most of its life in epipelagic waters and then descended to depths > 1,000 m. Onychoteuthis compacta seemed to range widely during the day and had rather large organs (its habitat, however, is poorly known). Brachioteuthis had similar organs but probably occurred below 800 m during the day. The ommas- trephids had a complex arrangement of organs, yet these animals were primarily epipelagic. In juveniles of many species (e.g., enoploteuthids), the size of the organs (relative to the size of the brain) may be large; yet their absolute size was small when compared with adults occupying the same depths. Compared with squid, all octopods had small organs. With the probable exception of the tubular eyed Amphifretus pelagicus, octopods probably do not occupy depths between 400 and 700 m during the day except as juveniles in transit to greater depths. Amphifretus pelagicus is the only pelagic octopod that exhibited clear modification of its or- gans. In contrast to the small organs, each consist- ing of a single vesicle, of other octopods, this ani- mal has a larger organ composed of many separate vesicles. Presumably the general trends with depth were related to depth gradients in both downwelling daylight and bioluminescent light. Downwelling daylight decreases exponentially with depth. Bioluminescent activity should increase from 400 to 600-800 m then decline rapidly if numbers of midwater fishes at various depths (see Amesbury 1975) provide an index to bioluminescent activity during the day. While many vesicles may detect both downwelling skylight and bioluminescent light, we will examine evidence for these two func- tions separately. The eyes of some mesopelagic animals can prob- ably detect silhouettes at depths of 750 to 775 m (Young and Roper 1977). Presumably some photo- sensitive vesicles are at least as sensitive as the eyes, especially when we consider the large size and apparent high pigment density of some (see Young 1972a). The large dorsal organs of squid were positioned so they are exposed to downwel- ling daylight. Large central organs appeared to be exposed to this light in species lacking dorsal or- gans. Some experimental evidence indicates that midwater cephalopods detect downwelling light with these vesicular organs. A number of cephalopods have been seen to conceal themselves with bioluminescent light (Young and Roper 1977). This counterillumination requires that the intensity of downwelling light is precisely determined by the animal, and the photosensitive vesicles seem the likely photoreceptor (Young 1973, 1977). Recently R. E. Young, C. F. E. Roper, and J. Walters (in manuscr.) covered the dorsal organs oi Abraliopsis sp. B while it was counteril- luminating and recorded a 909f drop in its luminescence. They concluded that the dorsal or- gans detect downwelling light. Since animals can detect downwelling light with these organs for counterillumination, they may use this photic in- formation for other purposes as well. Vertical migration in many midwater animals is closely associated with changing light levels (Boden and Kampa 1967). Since cephalopods mi- grate during twilight periods, light cues received by the vesicular organs may serve to trigger or regulate their migrations. This view is supported by three sources of evidence. First, nerves from the vesicles pass into the peduncle complex of the brain. Messenger (1967b) suggested, on the basis of experimental evidence in Octopus, that this complex is part of a visuomotor system: visual information from the eyes enables this complex to exercise control over locomotion. Secondly, ex- perimental evidence on the function of the photo- sensitive vesicles in neritic Octopus strongly suggests that these organs regulate diurnal activ- ity patterns (R. Houck pers. commun.). Finally, most migrating cephalopods have large vesicular organs positioned to detect downwelling light. The only exceptions are species of Mastigoteuthis and Chiroteuthis , whose migration patterns are not as distinct as in other species. If dorsal and central organs function primarily in the detection of downwelling light, we may have a clue to the peculiar arrangement of vesicular organs in ommastrephids. The ommastrephids were the only squids that had central organs on the dorsal surface of the optic stalk as well as dorsal organs. In Nototodarus hawaiiensis, these two organs differ morphologically (the small cen- tral organ has large component vesicles and the large dorsal organ has small vesicles), but the organs are adjacent to one another. The structural differences suggest separate functions for the or- gans, yet their close proximity indicates that both will be exposed to the same source of light. The same argument holds for these organs in other ommastrephids, although the two organs are 613 FISHERY BULLETIN: VOL. 76, NO 3 somewhat further separated. The ommastrephids have the unusual habit of usually living in epipelagic waters but occasionally descending to great depths (Roper and Young 1975). Perhaps the central organs function in epipelagic waters while the dorsal organs operate only in deep water. Cer- tainly there are considerable problems associated with a single organ functioning over such a wide range of light intensities. Certain photosensitive vesicles appear to detect bioluminescent light rather than downwelling skylight. The small vesicular organs in the deep- living Vampyroteuthis infernalis are shielded from downwelling skylight (Young 1972a). Ves- icular organs are present in the blind bathypelagic octopod Cirrothauma murayi (J. Z. Young in Packard 1972) which lives in depths where detect- able surface light is absent. Certain photosensi- tive vesicles of many other species were shielded from downwelling light. Such organs presumably detect bioluminescent light. Young (1973, 1977) demonstrated that certain vesicular organs in some species were directly exposed to the animal's photophores, presumably for counterillumination purposes. The detection of bioluminescence is not limited, however, to the animal's own photophores. With only a few exceptions all species examined had some means of "viewing" various parts of the man- tle cavity with their vesicular organs. In many cases, the organs seemed precisely placed for this purpose (see Figure 6). In most species with large opaque livers (e.g., ommastrephids, enop- loteuthids, histioteuthids, bathyteuthids, cy- cloteuthids, octopoteuthids, and mastigoteuthids), some organs extended laterally or ventrally past the liver, enabling a "view" of the mantle cavity. In other species with the liver far back in the mantle cavity (e.g., cranchiids, Brachioteuthis, Ctenopteryx), only central organs were present. In Vampyroteuthis infernalis, the organs lay within the mantle cavity and could only be exposed by light originating within this cavity, the funnel, or at the mantle opening. This animal, like most other cephalopods, had no photophores in these locations. The photosensitive vesicles in octopods were also located within the mantle cavity. The view of the mantle cavity is obscured only in Onychoteuthis, Chiroteuthis , Joubiniteuthis, and Chiroteuthidae gen. sp., although most of these species could still detect light from within the fun- nel and at the entrance to the mantle cavity. Young (1972a) suggested that the photosensi- tive vesicles in Vampyroteuthis detect small glow- ing organisms that are carried into the mantle cavity with the respiratory current. In the deep sea, a glowing organism within the mantle cavity could reveal the squid's location and have disas- terous consequences. J. Z. Young ( 1977) extended this idea to octopods. Nevertheless, this sugges- tion seems unlikely to have broad application in explaining the consistent relationships between vesicle location and mantle cavity "visibility"; however, no alternative function has been found. Some squid may detect bioluminescent light originating outside the animal. The large vesicu- lar organs in the deep-living Bathyteuthis abys- sicola are not exposed to its own photophores and probably detect bioluminescence from animals lo- cated outside its restricted visual field (Young 1972a). In Ctenopteryx siculus, the elongate ves- icular organs joined in the midventral line over the funnel and were there shielded dorsally and laterally by a thick layer of iridophores. The ven- tral part of this organ would detect light originat- ing within the funnel. Yet, the high organization and sophisticated structure of the vesicles seem overly matched for such a task. This organ proba- bly "looks" ventrally through the funnel to the area below the squid. Similar arguments could be made for certain lobes in other squids. The photosensitive vesicles in many cephalo- pods apparently form an elaborate system for monitoring bioluminescent light from their own photophores, from within the mantle cavity, and from the immediate vicinity of the animal that lies outside the visual field The great variety of photosensitive vesicles found among the species of pelagic cephalopods off Hawaii presumably reflects a variety of functions for these organs associated with the detection of both downwelling and bioluminescent light. The morphology and placement of these organs have provided some clues to these functions. A full un- derstanding of this complex sensory system, how- ever, must await experimental studies on living animals. ACKNOWLEDGMENTS I wish to thank the Captain and the crew of the RV Teritu and the many people that participated in the "Teuthis" cruises. Of these, I especially thank John Walters, Steven Amesbury, Sherwood Maynard, and Fletcher Riggs. Most of the 614 YOUNG: VERTICAL DISTRIBUTION AND PHOTOSENSITIVE VESICLES cephalopods taken by the IKMT came from a sam- pling program by T. A. Clarke, University of Hawaii. I thank J. Walters and S. Maynard for their comments on the manuscript. This study was supported by grants GB 20993 and GA 33659 from the National Science Foundation. Hawaii Insti- tute of Geophysics Contribution No. 916. LITERATURE CITED Amesbury, S. S. 1975. The vertical structure of the midwater fish commun- ity off leeward Oahu, Hawaii. Ph.D. Thesis, Univ. Hawaii, Hoholulu, 106 p. Berry, S. S. 1914. The Cephalopoda of the Hawaiian Islands. Bull. U.S. Bur. Fish. 32:255-362. BODEN, B. P., AND E. M. KAMPA. 1967. The influence of natural light on the vertical migra- tions on an animal community in the sea. Symp. Zool. Soc. Lond. 19:15-26. Clarke, M. R. 1966. A review of the systematics and ecology of oceanic squids. Adv. Mar. Biol. 4:91-300. 1969. Cephalopoda collected on the SOND cruise. J. Mar. Biol. Assoc. U.K. 49:961-976. Clarke, M. R., and C. C. Lu. 1974. Vertical distribution of cephalopods at 30°N 23°W in the North Atlantic. J. Mar. Biol. Assoc. U.K. 54:969- 984. 1975. Vertical distribution of cephalopods at 18°N 25°W in the North Atlantic. J. Mar. Biol. Assoc. U.K. 55:165- 182. Clarke, T. A. 1973. Some aspects of the ecology of lantemfishes (Myc- tophidae) in the Pacific Ocean near Hawaii. Fish. Bull., U.S. 71:401-434. GIBBS, R. H., Jr., AND C. F. E. ROPER. 1971. Ocean Acre: Preliminary report on vertical distribu- tion of fishes and cephalopods. In G. B. Farquhar (editor). Proceedings of an International Symposium on Biological Sound Scattering in the Ocean, p. 1 19-133. Dep. Navy, Washington, D.C., Maury Cent. Ocean Sci. Rep. 005. LU, C. C, AND M. R. Clarke. 1975a. Vertical distribution of cephalopods at 40°N, 53°N and 60°N at 20°W in the North Atlantic. J. Mar. Biol. Assoc. U.K. 55:143-163. 1975b. Vertical distribution of cephalopods at 11°N, 20°W in the North Atlantic. J. Mar. Biol. Assoc. U.K. 55:369- 389. Mauro, a. 1977. Extra-ocular photoreceptors in cephalo- pods. Symp. Zool. Soc. Lond. 38:287-308. MAURO, A., AND F. BAUMANN. 1968. Electrophysiological evidence of photoreceptors in the epistellar body o^Eledone moschata. Nature (Lond.) 220:1332-1334. Messenger, J. B. 1967a. Parolfactory vesicles as photoreceptors in a deep- sea squid. Nature (Lond.) 213:836-838. 1967b. The peduncle lobe: a visuo-motor centre in Oc- topus. Proc. R. Soc. Lond., Ser. B, Biol. Sci. 167:225-251. NISHIOKA, R. S., I. YASAMASU, A. PACKARD, H. A. BERN, AND J. Z. YOUNG. 1966. Nature of vesicles associated with the nervous sys- tem of cephalopods. Z. Zellforsch. Mikrosk. Anat. 75:301-316. PACKARD, A. 1972. Cephalopods and fish: the limits of con- vergence. Biol. Rev. Camb. Philos. Soc. 47:241-307. PEARCY, W. G. 1965. Species composition and distribution of pelagic cephalopods from the Pacific Ocean off Oregon. Pac. Sci. 19:261-266. ROPER, C. F. E. 1969. Systematics and zoogeography of the world wide bathypelagic squid Bathyteuthis (Cephalopoda: Oegopsi- da). U.S. Natl. Mus. Bull. 291, 210 p. ROPER, C. F. E., AND R. E. YOUNG. 1975. Vertical distribution of pelagic cephalo- pods. Smithson. Contrib. Zool. 209, 51 p. Walters, J. F. 1976. Ecology of Hawaiian sergestid shrimps (Penaeidea: Sergestidae). Fish. Bull., U.S. 74:799-836. Young, J. Z. 1977. Brain, behaviour and evolution of cephalopods. Symp. Zool. Soc. Lond. 38:377-434. Young, R. E. 1972a. Function of extra-ocular photoreceptors in bathypelagic cephalopods. Deep-Sea Res. 19:651-660. 1972b. Brooding in a bathypelagic octopus. Pac. Sci. 26:400-404. 1973. Information feedback from photophores and ventral countershading in mid-water squid. Pac. Sci. 27:1-7. 1975a. Leachia pacifica (Cephalopoda, Teuthoidea): Spawning habitat and function of the brachial photo- phores. Pac. Sci. 29:19-25. 1975b. A brief review of the biology of the oceanic squid, Symplectoteuthis oualaniensis (Lesson). Comp. Biochem. Physiol. 52B:141-143. 1975c. Function of the dimorphic eyes in the midwater squid Histwteuthis dofleini. Pac. Sci. 29:211-218. 1975d. Transitory eye shapes and the vertical distribution of two midwater squids. Pac. Sci. 29:243-255. 1977. Ventral bioluminescent countershading in midwa- ter cephalopods. Symp. Zool. Soc. Lond. 38:161-190. Young, R. E., and C. F. E. Roper. 1977. Intensity regulation of bioluminescence during countershading in living midwater animals. Fish. Bull., U.S. 75:239-252. Note added in proof: The correct name for the species listed here as Phasmatopsis fisheri is Megalocranchia fisheri (N. Voss. In press. Studies on the cephalopod family Cranchiidae. A revision of the genera, with a key for their determination. Bull. Mar. Sci.) 615 SYSTEMATIC SAMPLING IN A PLANKTONIC ECOSYSTEM E. L. Venrick' ABSTRACT Two sampling studies, computer simulation and field, investigated the consequences of applying restricted systematic sampling (at predetermined depths) to estimate total chlorophyll in the water column. Comparison was made with stratified random designs with one and two samples per strata. Systematic sampling appeared more accurate than most stratified random designs. However, when repeated over restricted spatial or temporal intervals, systematic designs tended to produce biased estimates. In the central Pacific, an interval of several days, or 100-200 km, appeared necessary for natural population fluctuations to average out the bias inherent in a restricted systematic sampling design. Underlying sampling theory is the assumption of random collection of samples. This is the only satisfactory method of assuring a representative sample from an unknown population. In pelagic ecology (and undoubtedly in other fields) this as- sumption is generally neglected and surveys are conducted at fixed geographic positions, at fixed spatial or temporal intervals, and/or at fixed depths, without recourse to randomization. The implicit assumption is that the natural complex variability of pelagic populations provides the necessary element of randomization. Two types of sampling strategies are frequently called systematic. The present study is concerned with the situation in which the sampling positions are fixed according to some pattern determined by the investigator and are not necessarily at equal intervals; this will be termed restricted systematic sampling (RSS) to distinguish it from the strategy in which only the sampling interval is fixed and the location of the first sample in the first interval is determined at random (randomly located sys- tematic sampling; Yates 1948). Among the alter- nate sampling strategies which provide the requi- site randomization, unrestricted random and stratified random sampling ( SR ) have received the most attention. In unrestricted random sampling, samples are selected individually from the entire population by some random process, such as by numbering all sampling units and selecting from them by means of a random numbers table. In SR, the population is first divided into subpopulations from each of which one or more samples are 'Scripps Institution of Oceanography, University of Califor- nia, San Diego, La JoUa, CA 92093. selected at random. SR is useful because it ensures that the samples are distributed throughout the entire population. Three characteristics of sampling designs are of interest (Figure 1): 1) bias, any consistent devia- tion between the true population parameter and repeated estimates based on the same sampling design; 2) precision, the variability of successive estimates about their mean when a sampling de- sign is repeated on the same population; and 3) PRECISION BIAS >- <_) -z. UJ o 3IASED BUT PRECISE MORE ACCURATE UNBIASED BUT IMPRECISE LESS ACCURATE Figure l. — Normal frequency distributions used to illustrate: a) precision, the spread of observations about their mean value (x); b) bias, the deviation of the mean of repeated observations from the true parameter ( d)\ c) a distribution which is biased but precise; and d) a distribution which is unbiased but imprecise. Distribution c will be more accurate than distribution d, in spite of the bias, if the average deviation of observations from B is smaller. Manuscript accepted Februar\' 1978. FISHERY BULLETIN: VOL '76. NO. 3. 1978. 617 — FISHERY BULLETIN: VOL. 76, NO. 3 accuracy, a concept including both freedom from bias and high precision and which, in the absence of bias, is equivalent to precision. The practical determination of precision, in its strictest sense, is restricted to quasi-static populations in which the population remains unchanged between collec- tions of replicate samples (forests or soil types or mussel beds, etc.). In the case of RSS, the concept of precision has no meaning in this type of popula- tion because successive application of the same sampling design to the same population will give identical results. Such static populations do not exist in a planktonic system because spatial and temporal variability produce continual change. Thus, in the present study, the concept of a popula- tion is expanded to incorporate spatial and tem- poral fluctuations in which case the precision of RSS has a real value. Theoretical aspects of systematic random sam- pling strategies have been considered by many (e.g., Yates 1946, 1953; Doming 1950; Cochran 1963; Sukhatme and Sukhatme 1970). Empirical investigations have been restricted to terrestrial systems, particularly to surveys of vegetation types or timber volumes ( e.g., Hasel 1938; Osborne 1942; Finney 1948b, 1950; Numata and Nobuhara 1952; Bourdeau 1953; Milne 1959). The results from these studies indicate that randomly located systematic sampling often gives more accurate es- timates than other procedures (Hasel 1938; Os- borne 1942; Madow 1946; Yates 1946, 1948; Fin- ney 1948a; Bordeau 1953; Milne 1959; Grieg- Smith 1964) especially when the sampled popula- tion has positive correlation between neighboring units (Cochran 1946; Milne 1959; Sukhatme and Sukhatme 1970). Because of the greater precision and gi-eater convenience of systematic sampling, some workers have recommended its use for ter- restrial surveys (Hasel 1938; Yates 1946; Milne 1959). On the other hand, it has been shown that irregular distributions or pronounced patterns of variation, especially periodicity or linear trends, may cause systematic designs to give biased esti- mates or estimates of reduced precision (Madow and Madow 1944; Finney 1950; Bourdeau 1953; Sukhatme and Sukhatme 1970); nor does the pre- cision necessarily improve with increasing sample size (Madow 1946; Bordeau 1953). Of the random designs, SR generally offers greater precision than unrestricted random sam- pling (Yates 1953; Milne 1959) and, with a con- stant number of samples, this precision increases as the number of strata increases (Yates 1953). The most precise design is one with one sample per strata, but this (like a systematic sample) offers no internal estimate of error (Finney 1948a, b). The success of systematic sampling clearly de- pends upon the nature of the sampled population. If individuals or properties in a population are distributed at random, all strategies will be equi- valent. Pronounced pattern, however, may in- crease or decrease the effectiveness of systematic designs. Thus, quite aside from the theoretical objections to systematic sampling, uninformed application of any systematic sampling is to be discouraged. Although Strickland (1968) warned that dis- crete samples may give a poor representation of the vertical distributions of highly stratified sub- stances, such as chlorophyll, a thorough study of the consequences of systematic sampling in the ocean has not been conducted, even though most populations have marked gradients, especially along the vertical axis. This may be attributed to the logistical difficulties of enumerating an oceanic population in its entirety, in contrast to a timber stand in which every individual may be observed, counted, measured, and mapped. The present study is restricted to the conse- quences of applying RSS in the vertical direction. The distribution investigated is that of chlorophyll in an oligotrophic oceanic environ- ment. Total chlorophyll in the water column is a frequently used index of plant crop and it is most often estimated from a series of restricted sys- tematic samples. The major question is whether such sampling produces any bias in the estimate of total chlorophyll, or whether the temporal and spatial heterogeneity of the chlorophyll distribu- tion is sufficient to average out the biases of indi- vidual determinations. Of secondary concern is whether there is a significant difference in preci- sion or accuracy between estimates derived from RSS and those derived from SR. The area of study is the North Pacific Central Gyre in the vicinity of lat. 28°N, long. 155°W. The region is one of relatively low spatial and temporal variability (Venrick et al 1973; Gregg et al. 1973; McGowan and Williams 1973; Eppley et al. 1973; Haury 1976). Thus, it is an environment in which any adverse characteristics of RSS are expected to be magnified. The general features of the distribu- tion of chlorophyll in the North Pacific Central Gyre have been summarized (Venrick et al. 1973). Most of the year, surface concentrations are low (0.02-0.06 mg/m^), and there is a narrow subsur- 618 VENRICK: SYSTEMATIC SAMPLINC IN KCOSYSTKM face maximum layer (0.10-0.20 mgm-M centered between 90 and 120 m. The present study was conducted in two parts. In part A, a computer was used to sample nine semiartificial populations derived from continu- ous vertical profiles of chlorophyll fluorescence. Changes in the fluorescence per extractable chlorophyll unit with depth iKiefer 1973) and smoothing of small-scale features during the pumping procedure result in a profile which repre- sents only the grosser features of the true distribu- tion. From the vertical profiles, the total popula- tion along the vertical axis was calculated, allowing the accuracy of various sampling strate- gies to be determined directly. Study B was con- ducted in the field where restricted systematic and stratified random samples were collected simul- taneously from the population. In this study, a real population was studied but the total population could only be approximated. METHODS Analytical Procedures offset the increase in fluorescence per unit of ex- tractable chlorophyll with depth, one conversion equation was used down to and including the chlorophyll maximum and another below the maximum. The conversion factors were deter- mined by analysis of chlorophyll extracted from discrete water samples collected periodically dur- ing the cruise. The surface value of each continu- ous profile was set to 0.03 mg/m-^ and the minimum value below the maximum to 0.01 mg/m-^ these were the mean values of extracted chlorophyll ob- served at the surface and at 200 m, respectively. The horizontal scale was adjusted to bring the mean maximum value of all profiles to 0.156 mg/m'\ the average maximum of the discrete sam- ples. A typical adjusted profile is presented in Fig- ure 2. These semi-artificial populations were sampled with four stratified random designs ( Table 1 ). The success of SR depends upon the extent to which the strata can be made internally homogeneous. In an attempt to achieve this, the stratum boundaries of SR-1 and SR-2 were determined as much as possi- ble by the hydrographic, biological, and chemical Chlorophyll a was determined fluorometrically according to the procedure of Yentsch and Menzel (1963) as modified by Holm-Hansen et al. (1965). Water for discrete, extracted chlorophyll samples was obtained with Nansen bottles. Water for con- tinuous vertical profiles was obtained with the seawater pumping system described by Beers et al. (1967) and was passed through a fluorometer equipped with a flowthrough door. Stud) A The chlorophyll fluorescence profiles were taken during September 1968, on 9 consecutive days during which time the ship followed two drogues which were set at 10 m depth to follow the mixed layer. These were launched at lat. 27°00'N, long. 155°18'W and moved in a northwesterly direction at speeds between 0.5 and 1.5 kn covering 345 km in 9 days. The profiles were not made at the same time of day. The closest two profiles were sepa- rated by 13 h, the most distant by 40 h. Additional aspects of these profiles and accessory data have been published (Scripps Institution of Oceanog- raphy 1974). The fluorescence profiles were read at 1-m in- tervals and translated into units of approximate chlorophyll down to a depth of 180 m. In order to 50 a^ioo Q 15 TEMPERATURE (°C) 20 25 1 1 1 \ I I I I 1 I 1 1 1 T" CHLOROPHYLL (mg/m3) 05 iO 15 .20 .25 .30 "TT - A - A - A A A A A I M I I [ I M I I [ I I I CHLOROPHYLL M M M M I M I TEMPERATURE " 150 L- hAh ■ Figure 2.— a typical population of chlorophyll values derived from a continuous profile of fluorescence (27 September 1968) and sampled in study A, together with the temperature values from the associated hydrocast. Triangles indicate the location of samples in restricted systematic design 3; bars represent the boundaries of strata used in stratified random design 1. 619 FISHERY BULLETIN: VOL. 76, NO. 3 Table l. — Systematic and stratified random sampling designs used in studies A and B. Systematic: RSS-1 RSS-2 RSS-3 RSS-4 Stratified random: SR-1 SR-2 SR-3 SR-4 0 0 0 0 0 0 0 0 20 10 20 45 35 75 15 45 40 25 40 65 55 95 45 90 Sample 60 35 60 80 Stratum 75 105 75 110 depths 80 50 80 90 boundaries 85 125 90 130 (m) 100 60 90 100 (m) 95 180 100 180 120 75 100 110 105 110 140 100 110 120 115 120 160 125 130 137 125 130 180 180 180 180 150 180 150 180 characteristics of the environment (Figure 2), and larger strata were assigned to the layers in which environmental gradients were small and several narrow strata were placed in the region of the chlorophyll maximum. The 35-m boundary marked the average depth of the mixed layer; 95 m was the approximate depth of penetration of 19( of the surface radiation, and 125 m represented the beginning of the nutricline. Design SR-1 consisted of 10 strata, each with one sample; in design SR-2, adjacent strata were lumped giving five strata with two samples in each. Designs SR-3 and SR-4 were those used in study B (below) and were thus based on environmental characteristics observed at that time. Each of the nine populations was sampled 20 times with each of the stratified ran- dom designs. To facilitate comparison with the systematic samples, for which there was only one cast of each design per profile, it was desirable to examine a series of unreplicated stratified random samples. For this purpose, 10 subsets were selected at random from the replicate casts, each subset containing nine stratified random casts, one from each population. Total chlorophyll was calculated from the mean (arithmetic) concentra- tion per strata times the width of that strata, summed over all strata. This is the classical proce- dure for summarizing data collected by SR. Four RSS designs were employed: RSS-1, one sample at the surface and every 20 m thereafter; RSS-2, the design actually employed in September 1968 in which the cast was partially determined by standard hydrographic depths; RSS-3, a design which was based upon complete knowledge of the vertical distributions which were being sampled and which was derived from application of the general rules of sample allocation, i.e., samples were concentrated in the region of maximum var- iability (the chlorophyll maximum layer); RSS-4, a design based on stratified random design 1 (and therefore more strictly comparable to it) with a sample at the top of the upper stratum (0 m) and bottom of the lowest stratum (180 m) and at the center of all intermediate strata. Two methods of calculating total chlorophyll from systematic samples were investigated. In the first, the layer between adjacent samples was rep- resented by the arithmetic mean of the two sam- ples (equivalent to integration with linear inter- polation). In the second, the layer was represented by the geometric mean of adjacent samples. This latter procedure is sometimes recommended when the population exhibits large, nonlinear changes between adjacent samples. A comparison of the two procedures was made in study A, on the basis of which the method using geometric means was rejected. Study B Study B, conducted in June 1977, combined two 10-sample designs, one restricted systematic (RSS-1) and one stratified random (SR-3 or SR-4) into a single 20-bottle cast. The strata boundaries were primarily determined from two preliminary 18-bottle casts which defined the regions of chlorophyll gradients and from a single STD trace which defined hydrographic strata (Figure 3). As in study A, narrower strata were established at the depths of maximum gradients of chlorophyll (the region of the maximum layer). The major differences between designs SR-1 and SR-2 and designs SR-3 and SR-4 were due to a shallower mixed layer and broader, deeper maximum layer observed in June 1977. Over a period of 21 days, a total of 18 casts were made, 9 employing RSS-1 and SR-3 and 9 employ- ing RSS-1 and SR-4. All casts were located within a rectangle bounded by lat. 28°21.6'N and 28°45.9'N, and by long. 155°14.0'W and 155°33.5'W. Fourteen casts were taken in con- junction with another program between the hours of 2200 and 0300 (with one exception, delayed by winch failure until 0550). Twelve casts were 620 VENRICK: SYSTEMATIC SAMPI.IN'd IN ECOSYSTEM TEMPERATURE (°C) 15 20 25 I — I — I — I — I — I — I — \ — \ — \ — \ — \ — I CHLOROPHYLL (mg/m3) ■05 .10 .15 .20 .25 I I I I I I I I I M I I I I I M I I I I I I I value and may be measured as a percent of the true value ( 0): 50 CL 100 UJ Q ,- kAh - A - A - A - A hAH hAH - A 150 hAH Figure 3. — Chlorophyll values observed in two 18-bottle casts preliminary to study B, together with the temperature trace from the associated STD lowering. Triangles indicate location of samples in restricted systematic design 1; bars represent the boundaries of strata used in stratified random design 3. paired, taken within a few hours and within 3 n.mi. of each other. These have been considered replicate casts. When the combined systematic-stratified ran- dom design called for bottles to be spaced more closely than 3 m, it was necessary to use a mes- senger heavier than the standard Nansen mes- senger (such as a Niskin bottle messenger) in order that it develop enough momentum to trip the second bottle. When both sampling designs called for the same depth, the extra bottle was arbitrarily positioned, usually filling in the largest gap in the region of the chlorophyll maximum layer. This "free" sample was used only in the calculations of total chlorophyll in the water column. Statistical Procedures Bias is evaluated by the consistency with which n observations (.v, , / = 1, n) from a given sampling design fall above or below the true population 2 iXi-9) n X 100% le. Precision is measured by the variance of a series of n observations about their mean (x): ^j ^X j' X) In an analogous way, accuracy is measured by the mean square deviation of a series of observations from the true population total: 2 {x,-Qy n Both accuracy and precision are inversely related to their statistical measures, increasing as the numerical value of the measure decreases. Since most scientists are used to thinking in terms of variances and sums of squares, it did not seem desirable to invert these measures to achieve di- rect correspondence. In the analysis of the results, limited use was made of the parametric analysis of variance. Most statistical tests were nonparametric tests which make few assumptions about the characteristics of the data (e.g., Dixon and Massey 1957; Tate and Clelland 1957; Conover 1971; Hollander and Wolfe 1973). Unless stated otherwise, the prob- abilities associated with conclusions in the text are derived from the binomial distribution withp = y-i. In several analyses in these studies, the problem of multiple testing arose, as when all four sys- tematic designs were tested for bias. Unfortu- nately, the tabulation on most nonparametric procedures is not sufficiently complete to allow correction for multiple testing to be made without making the tests extremely conservative. Since this was deemed undesirable, the probabilities given for the statistical tests are uncorrected. It is unlikely that this makes any real difference in the outcome of these studies which gain most of their force from the similarity of results in the two ap- proaches. 621 FISHERY BULLETIN: VOL 76, NO. 3 RESULTS Study A Integration of values The results of study A are summarized in Table 2. The total chlorophyll values derived from the four systematic sample designs were calculated by integration with linear interpolation (i.e., using the arithmetic mean of adjacent samples to repre- sent the average chlorophyll in the stratum be- tween them). Use of the geometric mean ir. this calculation resulted in the true total being under- estimated 27 out of 36 times (P = 0.01). Nor was there any increase in accuracy (the resultant ac- curacies, based on use of the geometric mean, were 0.538, 0.987, 0.488, and 0.752 for RSS-1 through RSS-4). The use of the geometric mean in the cal- culation of total chlorophyll does not appear to be justified. Bias The biases observed in the eight sampling strategies are summarized in Table 3. Of the four restricted systematic designs, only RSS-2 gave no signs of bias. Design RSS-3, the "best informed" design, overestimated the true population total in eight of the nine trials (P<0.05). RSS-1 overesti- Table 3. — Bias of systematic and stratified random sampling designs, Study A. Date Systematic designs Stratified random designs (1968) RSS-1 RSS-2 RSS-3 RSS-4 SR-1 SR-2 SR-3 SR-4 19 Sept. - - + - -1- _ _ - + 20 Sept. - - + -»- - - -t- '0 - 21 Sept - - + - + - -^ '0 + 22 Sept. + + + + - - + '0 + 23 Sept. - - -1- - - - - 24 Sept. + - -1- + -1- -1- -1- 25 Sept. + + -1- '0 - - + 26 Sept. + - - - -1- h - + 27 Sept. + + -1- + -1- — -I -1- - ' Estimate = = true value. mated the population only fivir- out of nine times, but the overestimates were clustered toward the end of the series and the underestimates toward the beginning. This temrjral trend lies just out- side the usual level of significance (run test; P<0.10) but it indicates that the time period necessary for the population to provide "random" variability of sufr.cient magnitude to eliminate bias may be of tht- order of several days or 100-200 km. The magnitudes of the biases were —4.09^ for the period 19-21 September and +3.77c for the period 24-27 September. Similarly, the bias intro- duced by using RSS-3 to estimate the true popula- tion total for 19-25 September was -3.69c. The peculiar periodicity of bias seen in RSS-4 also indicates a nonrandom interaction between the sampling design and the sampled population Table 2. — Restilts of study A, a computerized simulation sampling study. The estimated parameter ( 6) is total chlorophyll above 180 m; units are milligrams per square meter; time is local time. Date Time True value {0) Systematic designs One cast each Means and Stratified random designs variances (in parentheses) of 20 replicates (1968) RSS-1 RSS-2 RSS-3 RSS-4 SR-1 SR-2 SR-3 SR-4 19 Sept. 1719 8.50 8.00 8.00 8.62 8.39 8.52 8.38 8.39 8.59 20 Sept. 2312 10.31 9.91 9.98 11.12 10.71 (0.345) 10.29 (2.460) 10.18 (0.642) 10.31 (0.050) 10.24 21 Sept 2335 7.35 7.21 6.69 7.57 7.33 (0.244) 7.36 (0.987) 7.16 (0.182) 7.35 (0.513) 7.36 22 Sept. 2351 6.89 7.23 7.72 7 45 718 (0.059) 6.85 (0.950) 6.88 (0.215) 6.89 (0.466) 6.92 23 Sept. 2025 8.83 7.97 7.51 9.03 8.47 (0.072) 8.62 (0.151) 8.68 (c.oei)i F,.8.2 (0.129) 8.78 24 Sept. 0900 9.68 9.70 9.20 9.94 10.56 (0.801) 9.81 (1.780) 9.57 (0.341) 9.85 (1.227) 9.79 25 Sept. 0800 11.00 11 84 12.08 11.08 11.00 (0.178) 10.86 (1.490) 10.90 (0.406) 10.92 (1.841) 11.16 26 Sept 0830 13.85 14.23 13.14 13.36 13.06 (0.263) 13.92 (1.436) 13.23 (0.443) 1365 (0.687) 13.90 27 Sept. 2400 13.90 14.47 14.56 14.94 14.78 (1.123) 14.17 (0490) (5.323) 13.60 (3,676) (2.289) 14.06 (1.409) (3.954) 13.83 (3.348) Accuracy n - 1 0.308 0.695 0.308 0.320 '0.574 '1.464 '0.779 '1.662 Precision 6.441 8.113 7.729 6.537 6.725 '8.004 '6.312 '7.036 '8.540 n - 1 'Mean values from 10 sets of unreplicated casts. 622 VENRICK: SYSTEMATIC SAMPLING IN ECOSYSTEM (run test, P = 0.10). With the sampling interval employed here, the biases of individual estimates average out over the entire study. Had the inter- val been twice as large, a consistent overestimate or underestimate would have resulted, with re- spective magnitudes of -i-5.8'7f and -1.9*^, until 25 September when the phase relationship ap- pears to have shifted. Tables 2 and 3 also present the results of the four stratified random designs, based upon the means of 20 replicates. The consistent underestimates resulting from SR-2 were sufficiently unexpected that a second series of 20 SR-2 samples were drawn from each population. This series showed no evidence of bias and, thus, it appears that the initial results were the product of random chance. Precision Precision, in its strictest sense, could only be examined in the case of the stratified random de- signs, for which replicates were available. The designs employing 10 strata, each with one sam- ple, SR-1 and SR-3, offered greater precision than designs with fewer strata. However, there was a highly significant concordance (Kendell coefficient, P<0.01i between the precisions of all designs with respect to the profiles giving the most precise result. Examination of the individual profiles indicated that the precision of the results was inversely related to the strength of the chlorophyll maximum and to the amount of small-scale variability along the vertical axis, or, in other words, to the structural complexity of the population. Later, the accuracy of the systematic designs (discussed below) was found to show the same relationship. For all stratified random designs, the variance between replicates was trivial compared with the variance between the nine populations. Analyses of variance gave /"s ^g ratios ranging from 54 to 344 (all P«0.01). When all nine profiles were consi- dered to be replicates of the same population, the variance between the nine estimates from each systematic cast could be compared with the var- iance between single stratified random casts, one from each population (Figure 4A). On this scale, there were no differences in precision between any of the sampling designs. The large variation be- tween populations masked any difference in per- formance. Thus, when the concept of the sampled population is expanded to include spatial and temporal variations, RSS appears to offer neither o CL cr 60 V I- < Z5 O o < 40 20 30 140 120 100 80- X A Precision: I(xrx) = \2 _ TRUE S^ BETWEEN POPULATIONS > < cr Z) o o < 20 V Z(x,-0)^ B Accuracy: — —, — 4^ !V r> ^ fv ^ ^ ^ Figure 4. — The results of the computer simulation sampling study, study A, showing the relative precisions and accuracies of the four restricted systematic sampling designs (RSS) and four stratified random designs (SR). advantages nor disadvantages with respect to pre- cision of estimates. Accuracy The accuracy of the various designs was also compared using sets of unreplicated stratified random casts (Figure 4B). The greater accuracies of stratified random designs SR-1 and SR-3 rela- tive to SR-2 and SR-4 undoubtedly reflected their greater precision; and perhaps the greater accu- racy of SR- 1 relative to SR-3 was due to selection of more appropriate strata. The systematic designs were generally more accurate than the stratified random designs. Only stratified random design SR-1 achieved the accuracy of the systematic de- signs. 623 FISHERY BULLETIN: VOL 76. NO. 3 Most of the chlorophyll work in the central Pacific has been based upon 12 or more sampled depths. Thus, it was encouraging to find that as few as 10 depths, regardless of the sampling strategy, gave a generally satisfactory picture of the amount of chlorophyll in the water column. Of nearly 400 estimates from individual casts, IGOi fell within ±109? of the true value. This percent increased to 859? for stratified designs SR-1 and SR-3 and to 949r for the 36 systematic casts. How- ever, to the extent that these fluorescence profiles underestimate the structural complexity of the true chlorophyll distribution, these results proba- bly overestimate the accuracies of the designs. Study B The results of the field study were remarkably similar to those of the computer study (Table 4). Bias and accuracy were investigated by assuming that the entire population was exactly represented by the 20 samples in one cast (systematic samples plus stratified random samples plus "free" sam- ples). The results of study A indicate that the dis- crepancy is not likely to be severe. Table 4. — Results of study B, a field sampling study. The esti- mated parameter is total chlorophyll above 180 m and the true value ( 6) is estimated from the 20 combined samples of the two designs; units are milligrams per cubic meter. Date Local time H X, bias (1977) RSS-1 SR-3 SR-4 5 June" 6 June_ 2345 1887 18.73- 18.25- 0220 1768 17.10- 18.69 + 8 June 0033 15.54 14.06- 18.21 + 9 June 0236 18 11 17.23- 17.55- 9 June" 2241 16.70 16.22- 17.19 + 10 June_ 0550 15.47 14.75- 16.68 + 13 June" 2208 13.26 1329^ 12.31- 14 June^ 0035 13.42 13.47 + 13.33- 15 June 2203 14.53 15-9W 11.19- 19 June" 20 June. 2149 11 50 10 91- 9.03- 0100 1029 10.78 + 10.78 + 21 June 2246 14,25 14.73 + 11.42- 22 June 1107 10.67 11.00 + 10.17- 23 June" 2333 1360 13.41- 12.97- 24 June. 0133 1252 12.16- 12.09- 24 June" 1505 1683 17.11 + 18.11 + 24 June. 1632 1698 17.69 + 15.72- 26 June 0822 13.32 13.58 + 13.10- Accuracy: n - 1 0.45 283 2.31 Precision (6/5-6/15) 0492 0.808 0.249 MS (6/19-6/25) 0442 0.319 1.591 w/in pa rs : indicates pair of replicate casts. Acciiracx Bias When the 18 casts are considered in chronologi- cal sequence, it is evident that RSS tended to de- viate from the true value in the same direction on adjacent casts. The direction of bias was the same within five of the six pairs of replicate casts (0.0563-697. sand (2-7) >28% sand (23) >84-99% sand (22-15) (see also Table 2). The same trend with sediments is indicated (P<0.05) for slender sole weight per square meter, but because catches per square meter were more variable or mean dif- ferences were less, differences were not significant atP<0.01. 635 FISHERY BULLETIN: VOL 76, NO. 3 Table 4. — Results of an analysis of variance (ANOVA) of the numbers and weights of fishes caught per square meter or per tow'; * indicates P<0.05; ** P<0.01. R^ (coefficients of determination) values are given below. df Alisp ecies Rex sole Dover sole Slender sole Pacific No sanddab No. Wt. No. Wt, No. Wt. No Wt. Wt Item m^ Tow m^ Tow m^ Tow m^ Tow m^ Tow m^ Tow m^ Tow m^ Tow m-^ Tow m^ Tow Year 2 .. .. •• • •• ■• Sediments 3 * ** ■ ■ Seasons 1 * ' Sediment ■ season 3 * • * " Depth 1 * Depth ■ season 1 .. .. Error 95 Total 106 R^ (tow) 0 18 020 023 0.16 0 17 0.16 036 0.33 028 0.34 ft2 (m2) 039 0,51 047 0.40 0.46 043 024 0.28 036 037 'An ANOVA using a square root transformation for the data on all species combined per tow and per square meter gave similar significance effects The ANOVA was unbalanced because of unequal numbers of observations per station, season, depth, etc Effects were tested using the extra sum of squares principle (Searle 1971) The biomass of Pacific sanddab. on the other hand, showed opposite trends (P 0.05) and was large on sandy sediments and small on silt or clay sediments ( see also Day and Pearcy 1968; Barss et al. see footnote 8). Since the effect of sediment was not significant for total fish catch by numbers or weight sandy stations with low percent organic carbon, apparently did not support a markedly lower abundance of demersal fishes (Table 2). Al- though adult Dover sole show a strong preference for mud or silt bottom (Barss et al. see footnote 8; Demory see footnote 6), this trend was not appar- ent for the small Dover sole caught inshore of Heceta Bank in this study. A sediment-season effect was indicated for slen- der sole. They were caught in larger numbers per tow and weight per square meter (P<0.05) at the stations with a low percentage of sand (6-8) in the winter than the summer. Depth Effects The slope of the regression between depth and number and weight per tow of Pacific sanddab was significant (P<0.01) and negative. Catches per square meter on a number and weight basis gave the same trends (P<0.05). Sanddab were most abundant in shallow water. Weight of rex sole per square meter and per tow and total fish numbers and weight per tow also tended to decrease (P<0.05) with depth. Depth-season interactions were significant on a square meter basis for all species combined (number and weight) and for numbers of rex and Dover soles. These effects were caused by appreci- ably larger catches in deep water in winter than summer. Seasonal differences were small in shal- low water. This trend for lower catches on the outer edge of the continental shelf during summer than winter was obvious for Pacific sanddab. They were completely absent from the deep stations (2, 6, 8) during the summer but were present at all stations during winter. Seasonal bathymetric migrations, with spawning migrations into deep water in the winter and return to relatively shal- low depths in the summer, have been described for Dover sole and rex sole by Hagerman (1952), Harry (1956), Alverson (1960), and Demory (1971). Such movements could explain these depth-season effects. Seasons No significant seasonal differences were de- tected, indicating little seasonal variation in catches of these species when all stations are com- bined. Year Effect On the basis of numbers and weight per square meter and per tow, more fishes were captured in 1969 and 1968 than in 1970 at all stations (Figure 3). This trend was significant (P<0.01) for all species combined and for rex sole, Dover sole, and Pacific sanddab. Year effects were also indicated for slender sole (P<0.05). I have no cogent expla- nation for these large annual variations. They could represent actual variations in abundance or availability, due to natural events or increased fishing activity, or to undetected changes in sam- pling efficiency. Dominant year classes have been reported for these flatfishes off Oregon (Demory and Robinson see footnote 9), which may contri- bute to these annual differences, though changes in length-frequency distributions were not obvi- ous over this 2-yr period. 636 PEARCY niSTRIRl'TION AND ABUNDANCE OF SMALL FLATFLSHES FiGL'RE 3. — Variations in the total num- bers of fishes caught at each of the seven stations, 1968-71. The two tows at each station for each samphng period were av- eraged. OCT 1968 MAY AUG 1970 The amount of variability explained by the re- gression {R'^) of all effects on catches ranged from 0.16 to 0.51 (Table 4). Values were larger for the analysis based on catches per square meter than catches per tow, except for slender sole. These low values indicate that most of the variability was not accounted for by the variables of sediment, depth, year, and season. Large residual mean squares indicate that sampling variability as- sociated with catches at individual stations is ap- preciable. Oviatt and Nixon (1973) completed a multiple regression analysis of biomass and num- bers of benthic fishes in Narragansett Bay, R.I., with 14 environmental variables. Depth and sed- iment organic content contributed significantly to the regression for total fish numbers and fish biomass. But an R^ of only 0.21 was found. In both of these studies, only a small fraction of the total variability was explained by the environmental factors included. Size-Frequency Distributions Differences in length-frequency distributions were sometimes obvious among the stations lo- cated at different depths or sediment types. For example, the main length mode of rex sole at the 100- and 102-m stations was 125 mm, but at 190- and 195-m depth there was a distinct bimodal dis- tribution with peaks at 45 and 215 mm (Figure 4). These differences imply that young-of-the-year 10 U 0 F,20 10 ff£X SOLE 100- 102 m n=t004 — 1 1 r 1 1 — 50 100 150 200 250 STANDARD LENGTH (mm) 300 FIGURE 4.— Length-frequency data for rex sole at 100-102 m stations (above) and 190-195 m stations (below). 637 FISHERY BULLETIN VOL 76, NO 3 (<50 mm) and age-groups III- V (200-250 mm) rex sole (see Hosie and Horton 1977 for age-length data) preferentially inhabit deep waters on the outer edge of the continental shelf while inter- mediate sizes (75-150 mm) inhabit shallower wat- ers of the inner shelf. The peak of young-of-the- year rex sole at 200 m corroborates the conclusion of Pearcy et al. (1977) about the depth of larval settlement and the nursery ground for early benthic life. They concluded that rex sole larvae settle to the bottom mainly on the outer continen- tal shelf during the winter when they are >50 mm SL. Powles and Kohler (1970) and Markle (1975) believed that the nursery grounds ofGlyptocepha- lus cynoglossus are also in deep waters off the east coast of the United States. Small Pacific sanddab (<70 mm) composed a larger proportion of the catch at 102 m where sand was 28'7f of the sediment (Station 23) than at 74 and 102 m where sand made up over 647f of the sediment (Stations 22, 15) (Figure 5). Young sanddab appear to inhabit deeper water with finer sediments in early life and then aggregate on sandy bottom areas in shallow water where they often dominate the demersal fish fauna. Hence, this trend of decreasing depth with increasing age is similar to that found for rex sole. SUMMARY 1. Demersal fishes were sampled at seven sta- tions on Oregon's central continental shelf at vari- 15 PACIFIC SANDDAB Stations 15822 84-99% Sand 10 ^ /^^ 74 -102m 1 \ /7 - 138/ — 5 - / - 1- z UJ o ^ r^ I ' 1 1 ^ 01 Q. 100 150 200 STANDARD LENGTH (mm) 250 Figure 5.— Length-frequency data for Pacific sanddab at .sta- tions with 84-99'7f sand (above) and 28% sand (below). ous seasons of the year during a 2-yr period. A fine-meshed, 3-m beam trawl was used in order to quantitatively sample small flatfishes. The sta- tions ranged from 74 to 195 m deep and had sedi- ment types ranging from nearly lOO'/r sand to clayey-silts with about 3''/< sand. 2. Stations were selected in an attempt to sepa- rate the effects of depth and sediment on the as- semblages of fishes and abundances of common species. Three station-pairs were recognized that had similar sediment types but were located at different depths. Separation of sediment and depth effects was complicated however by differ- ences in measured (and possibly unmeasured) fac- tors between station pairs. 3. Two general assemblages of fishes were re- cognized on the basis of species composition of fishes by numbers, biomass per square meter of dominant species, and similarity indices among the seven stations. These were a shallow water (74-102 m) assemblage dominated numerically by Pacific sanddab, and a deepwater ( 148-195 m) as- semblage dominated by slender sole. 4. Species diversity iH) varied between 1.6 and 2.5 except at the shallow, sand station where it was only 0.7. Dominance was pronounced at this station: 86''^^ of all the individual fishes captured were Pacific sanddab. The largest number of spe- cies (34 or 35) was recorded for the three deep stations. These values of H are similar to others for temperate, demersal fish communities. 5. Similarity indices of the species composition of fishes were high for two of the three station pairs with similar sediments. However, indices were also high among the four shallow stations of differ- ing sediment types. Stations that were near each other geographically were similar, indicating the possibility of a proximity effect, but high similar- ity was also found among deep stations, one of which was over 65 km from the others. 6. An analysis of variance of the number and weight per square meter and per tow of Dover, rex, and slender soles. Pacific sanddab, and all species combined indicates some effects of sediments and depth. Largest catches of slender sole were at the clayey-silt station pair, and largest catches of Pacific sanddab were on sandy sediments. Catches of Pacific sanddab were significantly larger at the shallow stations. Catches of rex sole and all species combined also tended to decrease with in- creasing depth. 7. Differences in the length-frequency distribu- tions of Pacific sanddab and rex sole were corre- 638 PEARCY: niSTRIBUTION AND ABUNDANCE OF SMALL FLATFISHES lated with depth or sediment type. Small sanddab predominated on the silty-sand station, whereas large sanddab preferred sandy sediments. Young-of-the-year rex sole were concentrated on the outer edge of the continental shelf ( 190- 195 m). 8. Catches were sometimes larger in the winter than the summer, especially at the deep stations. This trend, which was noted for all four flatfishes and for all species combined, is probably the result of seasonal bathymetric movements. 9. A pronounced decrease in the catches of most species and total catch per square meter occurred during the 2 yr of this study. Reasons for this decline are unknown. 10. The biomass of benthic fishes ranged from 0.9 to 2.4 g m "- at the seven stations. Biomass was not appreciably lower at the pure sand stations, which had about O.l'^^ organic carbon in the sedi- ment. This is related to the fact that the Pacific sanddab, the predominant species at this station, is a pelagic feeder (see Pearcy and Hancock 1978). 11. The weight of fishes per square meter caught in the 3-m beam trawl was several times lower than that estimated from larger otter trawls with coarser meshes. Although the beam trawl caught many small flatfishes, large fishes and nektobenthic species effectively avoided this small beam trawl, resulting in low biomass esti- mates. ACKNOWLEDGMENTS This research was sponsored by NOAA Office of Sea Grant, No. 04-5-158-2. I am especially grate- ful to D. L. Stein and G. L. Bertrand for their help at sea; to D. L. Stein for identifying the fishes; to F. L. Ramsey, W. L. Gabriel, and R. G. Petersen for analysis of data; and to J. Dickinson, A. G. Carey, Jr., and J. C. Quast for helpful comments on the manuscript. LITERATURE CITED ALTON. M. S. 1972. Characteristics of the demersal fish fauna inhabit- ing the outer continental shelf and slope off the northern Oregon coast. In A. T. Pruter and D. L. Alverson (edi- tors). The Columbia River estuary and adjacent ocean waters, p. 583-634. Univ. Wash. Press, Seattle. ALVERSON, D. L. 1960. A study of annual and seasonal bathymetric catch patterns for commercially important groundfishes of the Pacific Northwest coast of North America. Pac. Mar. Fish. Comm.. Bull. 4, 66 p. Ai.vERSON. D. L., A. T. Pruter, and L. L. Ronholt. 1964. A study of demersal fishes and fisheries of the north- eastern Pacific Ocean. H.R. MacMillan Lect. Fish., Inst. Fish. Univ. B.C., Vancouver, 190 p. Bertrand, G. A. 1971. A comparative study of the infauna of the central Oregon continental shelf Ph.D. Thesis, Oregon State Univ., Corvallis, 113 p. Carey, a. G., Jr., and h. Heyamoto. 1972. Techniques and equipment for sampling benthic or- ganisms. In A.T. Pruter and D.L. Alverson (editors). The Columbia River estuary and adjacent ocean waters, p. 378-408. Univ. Wash. Press, Seattle. Day, D. S., .and W. G. Pearcy. 1968. Species associations of benthic fishes on the continen- tal shelf and slope off Oregon. J. Fish, Res, Board Can, 25:2665-2675. DEMORY, R, L. 1971, Depth distribution of some small flatfishes off the northern Oregon-southern Washington coast. Res, Rep, Fish, Comm, Oreg, 3:44-48, Edwards, R, and J, H, Steele, 1968, The ecology of O-group plaice and common dabs at Lock Ewe. I. Population and food. J. Exp. Mar. Biol. Ecol. 2:215-238. Emery, K. O. 1938. Rapid method of mechanical analysis of sands. J. Sediment. Petrol, 8:105-111, Gunther, F, J, 1972, Statistical foraminiferal ecology from seasonal samples, central Oregon continental shelf Ph,D, Thesis, Oregon State Univ,, Corvallis, 128 p, Haedrich, R, L,, and S, O. Haedrich. 1974. A seasonal survey of the fishes in the Mystic River, a polluted estuary in downtown Boston, Massachu- setts. Estuarine Coastal Mar. Sci. 2:59-73. Haedrich, R. L., G. T. Rowe, and P. T. Pollonl 1975. Zonation and faunal composition of epibenthic popu- lations on the continental slope south of New Eng- land. J. Mar, Res, 33:191-212, Hagerman, F, B, 1952, The biology of the Dover sole, M/crostomwspaci/zcws ( Lockington), Calif Dep, Fish Game, Fish Bull. 85, 48 p. Harry, G, Y„ III 1956, Analysis and history of the Oregon otter-trawl fishery, Ph,D, Thesis. Univ, Washington, Seattle, 329 p. HOSIE, M, J,, AND H, F, HORTON, 1977, Biology of the rex sole, Glyptocephalus zachirus, in waters off Oregon, Fish, Bull,, U,S, 75:51-60, INMAN, D, L, 1952, Measures for describing the size distribution of sed- iments, J, Sediment, Petrol. 22:125-145, KRUMBEIN, W, C, AND F. J. PETTLJOHN. 1938. Manual of sedimentary petrography. Appleton- Century-Crofts, Inc., N.Y., 549 p. KUIPERS, B. 1975. On the efficiency of a two-metre beam trawl for juvenile plaice iPleuronectes platessa). Neth. J. Sea Res. 9:69-85. KULM, L. D., R. C. ROUSH, J. C. HARLETT, R. H. NEUDECK, D. M. Chambers, and E. J. Runge. 1975. Oregon continental shelf sedimentation: interrela- tionships of facies distribution and sedimentarj' proces- ses. J. Geol. 83:145-175. 639 FISHERY BULLETIN: VOL. 76, NO. 3 M.AKGALEF, R. 1968. Perspectives in ecological theor>-. Univ. Chicago Press, Chicago, 111 p. M.^KKLE, D. F. 1975. Young witch flounder. Glyptocephalus cynoglossus, on the slope off Virginia. J. Fish. Res. Board Can. 32:1447-1450. MClNTIRE. C. D., AND W. W. MOORE. 1977. Marine littoral diatoms: Ecological considera- tions. In D. Werner (editor). Biology of diatoms, p.333- 371. Blackwell Scientific Publ. Oxf. Merriman, d., and H. E. WARFEL. 1948. 5. Studies on the marine resources of southern New England. VII. Analysis of a fish population. In A sym- posium on fish populations, p. 131-164. Bull. Bingham Oceanogr. Collect., Yale Univ. 11(4). Oviatt, c. a., and S. W. Nixon. 1973. The demersal fish of Narragansett Bay: an analysis of community structure, distribution and abun- dance. Estuarine Coastal Mar. Sci. 1:361-378. Pearcy. W. G. and D. Hancock. 1978. Feeding habits of Dover sole, Microstomus pacificus; rex sole, Glyptocephalus zachirus; slender sole, Lyopsetta exilis; and Pacific sanddab, Citharichthys sordidus, in a region of diverse sediments and bathymetry off Ore- gon. Fish. Bull., U.S. 76:641-651. PEARCY, W. G., M. J. HOSIE, AND S. L. RICHARDSON. 1977. Distribution and duration of pelagic life of larvae of Dover sole. Microstomus pacificus: rex sole, Glyptoce- phalus zachirus; and petrale sole, Eopsetta Jordani, in waters off Oregon. Fish. Bull., U.S. 75:173-183. POWLES, P. M., AND A. C. KOHLER. 1970. Depth distributions of various stages of witch floun- der {Glyptocephalus cynoglossus) off Nova Scotia and in the Gulf of St. Lawrence. J. Fish. Res. Board Can. 27:2053-2062. Richards, S. W. 1963. The demersal fish population of Long Island Sound. I. Species composition and relative abundance in two localities, 1956-1957. Bull. Bingham Oceanogr. Collect., Yale Univ. 18(2):5-31. RILEY, J. D., AND J. CORLETT. 1966. The numbers of 0-group plaice in Port Erin Bay, 1964-66. Mar. Biol. Stn. Port Enn Annu. Rep. 78:51-56. ROUSH, R. C. 1970. Sediment textures and internal structures: a com- parison between central Oregon continental shelf sedi- ments and adjacent coastal sediments. M.S. Thesis, Oregon State Univ., Corvallis, 59 p. Sanders, H. l., and r. r. hes.sler. 1969. Ecology of the deep-sea benthos. Science (Wash., D.C.) 163:1419-1424. Searle, S. R. 1971. Linear models. John Wiley and Sons, N.Y., 532 p. Shannon, C. E., and W. Weaver. 1963. The mathematical theory of communica- tion. LTniv. 111. Press, Urbana, 117 p. Thorson, G. 1957. Bottom communities (sublittoral and shallow shelf). In J. W. Hedgpeth (editor). Treatise on marine ecology and paleoecology. Vol. 1, p. 461-534. Geol. Soc. Am. Mem. 67. 640 FEEDING HABITS OF DOVER SOLE, MICROSTOMUS PACIFICUS; REX SOLE, GLYPTOCEPHALUS ZACHIRUS; SLENDER SOLE, LYOPSETTA EXILIS; AND PACIFIC SANDDAB, CITHARICHTHYS SORDIDUS, IN A REGION OF DIVERSE SEDIMENTS AND BATHYMETRY OFF OREGON William G. Pearcy and Danil Hancock* ABSTRACT The feeding habits of the Dover sole and rex sole (mainly juveniles) and of slender sole and Pacific sanddab were investigated at seven stations on the continental shelf off central Oregon. Dover sole had a catholic diet, feeding on a large variety of infaunal and epifaunal invertebrates. The composition of the diet varied among stations of different depth and sediment type indicating opportunistic feeding. Pelecvpoda were the most important prey on a weight basis at the shallow station (74 m) of well-sorted sand where they were the dominant macrofaunal invertebrate. Ophiuroids, sea pens, anemones, and pelecypods were the most important prey at 100-102 m stations of silty sand or sandy silt. Polychaetes composed over 90^^ of the diet at the deep stations ( 148-195 m) of clayey silt or silty sand. The average standing stocks per square meter of Dover sole caught in beam trawl collections and polychaetes in grab samples were positively correlated among stations. Similarity of the food habits of Dover sole on the basis of food weight or frequency of occurrence was generally higher among stations of similar depth than of similar sediment texture. Similar trends were noted for assemblages of benthic fishes and invertebrates. Dover sole collected during the winter had the highest percentage of empty stomachs, the fewest prey taxa. and often the lowest frequency of occurrence of prey taxa within a size group. Because seasonal variations were not observed in abundance of macrofaunal food in the sediments, availability of prey may change with season, or more likely, Dover sole feed more intensely and less selectively during summer. Small ( <150 mm standard length) rex sole fed mainly in amphipods and other crustaceans. Large (150-450 mm standard length) rex sole preyed chiefly on polychaetes. The diet of rex sole was less diverse than that of the Dover sole and overlap of diet between the two species was not large. Both the Pacific sanddab, numerically the most common species offish at the shallow sand station, and the slender sole, the most common species at the three deep, soft-sediment stations, preyed principally on pelagic crustaceans such as euphausiids, shrimps, and amphipods. Although the biomass of mollusks in the sediments was large at the shallow sand station, they were not consumed by Pacific sanddab. Fish were occasionally an important food for the sanddab. The objectives of this study were: 1) to describe the food habits of the four species of flatfishes that are common in trawl catches on the central continen- tal shelf off Oregon: Dover sole, Microstomas pac/- ficus. rex sole, Glyptocephalus zachirus, slender sole, Lyopsetta exilis, and Pacific sanddab, Citharichthys sordidus; 2) to evaluate the possible effects of depth and sediment, size of fish, and season of capture on their food habits; and 3) to compare the biomass and composition of fish food from grab samples with feeding habits of fishes. These species are among the most abundant flatfishes in demersal communities of this region of the Pacific Ocean ( Alverson et al. 1964; Day and 'School of Oceanography, Oregon State University, Corvallis, OR 97331. Pearcy 1968; Alton 1972; Demory and Hosie^). They dominated the fish catches at the stations where they were captured for this study ( Pearcy 1978). In order to know more about the role of these fishes in their ecological communities, in- cluding competitive-predatory relationships, more data are required on their food habits. Hagerman (1952) listed food items found in Dover sole caught in California waters. Pearcy and Vanderploeg ( 1973) listed general taxonomic groups preyed upon by Dover, rex, and slender soles and Pacific sanddab. Kravitz et al. (1977) gave a detailed account, including species of prey Manuscript accepted .January 1978. FISHERY BULLETIN: VOL. 76 NO. .3. 1978. ^Demory, R. L., and M. J. Hosie. 1975. Resource surveys on the continental shelf of Oregon. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Commer. Fish. Res. Dev. Act, Annu. Rep. July 1, 1974 to June 30, 1975, 9 p. 641 FISHERY BULLETIN: VOL. 76, NO. 3 consumed by the rex sole and Pacific sanddab caught in a single collection off the central Oregon shelf. This study is, to our knowledge, the most complete study of the food habits of these four species. METHODS Fishes were collected during 115 tows with a 3-m beam trawl at seven stations on the continen- tal shelf off central Oregon between August 1968 and August 1970. These stations are classified by four depth categories and the percentage of sand in the sediments in Figure 1 . Details on methods and descriptions of the stations are given by Pearcy (1978). All fishes were preserved at time of capture with Formalin,-' and the body wall of large (>150-200 mm SL) fishes was incised to insure preservation of stomach contents. Fishes were identified and measured (standard length, SL) in the shore laboratory. Stomachs were removed from 326 Dover sole represented in the catches at all seven stations; and from 614 rex sole, 1,109 slender sole, and 723 Pacific sanddab captured at two or three stations where each of these species was most common. Stomach contents were removed and empty stomachs were noted. Food organisms were iden- tified to species when possible. Annelids, crusta- ceans, mollusks, echinoderms, coelenterates, and remaining taxa (major taxa) were weighed to the ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. DEPTH (m) 100- 190- 74 102 148 195 < CO u u UJ 20 60 75 100 6 8 23 7 2 22 15 Figure l, — Classification of the seven stations, each indicated by station number, according to depth of water and percent of sand (0-20, 20-60, 60-75, and 75-100^^ ). The stations with similar sediment types but different depths are 6 and 8, 7 and 2, and 22 and 15. (See Pearcy 1978 for additional information.) nearest 0.01 g (wet-preserved weight). Usually these weights were obtained for the contents of a single stomach, but when the contents were in- sufficient for accurate weighing, taxa from the stomach contents of several fish of the same species and size, and from the same tow, were combined and weighed together to constitute an observation. The number of observations for Dover, rex, and slender sole and Pacific sanddab were 325, 374, 607, and 392, respectively. Results are reported as the a) percent that each major food taxa constitutes of the total wet weight of food found in stomachs for all seasons combined and for winter and summer seasons separately; and b) the frequency of occurrence (FO) of princi- pal prey, i.e., species or taxa found in 59^ or more of the observations for a species or size group of a species for all seasons combined. RESULTS General Food Habits Two general feeding types are indicated by dif- ferences in the weights of major food taxa (an- nelids, crustaceans, mollusks, echinoderms, coelenterates, and other taxa) found in the stomach contents of the four species (Table 1). Dover and rex soles fed largely (64^ ) on annelids, while slender sole and Pacific sanddab fed mainly on crustaceans (75'7r ). Within these two apparent feeding types, differences occurred among the proportions of prey taxa of secondary importance. For example, crustaceans were more abundant in the diet of rex than Dover sole (31% vs. 11%), whereas mollusks were more abundant in Dover than in rex sole ( 18% vs. I9t ). Annelids composed more of the stomach contents of slender sole than Pacific sanddab (159^ vs. 7% ). Based on the average frequency of occurrence of principal prey (F0>5% ) from all sizes offish and from all stations (Table 2), it is obvious that the food habits within these two feeding types (Dover sole-rex sole vs. slender sole-Pacific sanddab) are not as similar as shown by Table 1. Principal prey of Dover sole, for example, included 11 different identified polychaetes. Rex sole preyed mainly on three identified species of polychaetes. Only one principal prey species of polychaete was common to the diet of both Dover and rex soles. The shrimp Pandalus Jordani, pelecypods, and ophiuroids were principal prey in the food of Dover but not rex sole, whereas crab larvae, cumaceans, and Oiko- 642 PKARCY and HANCOCK: FEEDING HAHITS OF FOUR FLATFISH SPECIES Table l. — Percent by weight that major food taxa composed of the diet of the four flatfishes, all stations and seasons combined. Slender Pacific Taxa Dover sole Rex sole sole sanddab Annelida 644 64,8 15.6 7.2 Crustacea 11,2 31 0 75.6 74,8 Mollusks 183 1,4 0.7 0 Echinoderms 34 0,1 <0.1 0 Coelenterales 26 0 0 0 Other taxa 0 2.8 8.2 18.0 T.\BLE 2. — Average frequency of occurrence of principal prey (those occurring in S'c or more of the observations) in the four species, all stations combined. Rex Dover Slender Pacific Prey sole sole sole sanddab Polychaeta: Sternaspis fossor 12 14 Myriochele heeri 15 Nothna geophiliformis 9 Goniada brunnea 10 Aricidea neoseucica 6 Haploscoloplos elongatus g Chaetozone setose 6 Terebellides stroemii 6 Rhodine bitorquala 5 Typosyllis hyalina 9 Lumbrinens sp. 6 Amphicteis sp. 5 Glyceridae 5 Unidentified 25 19 7 Crustacea: Pandalus jordani 6 7 Euphausia pacifica 9 7 Crab larvae or juvenile 7 13 Gammand amphipods 29 • 24 Copepods 12 Cumaceans 8 Unidentified 16 12 21 24 Mollusca Pelecypoda 12 Ecfiinodermata Optiiuroidea 12 Miscellaneous: Oikopleura spp. 7 No, observations 347 325 607 392 No, fisfi 614 326 1.109 723 pleura were principal prey of rex but not Dover sole. Gammarid amphipods occurred frequently in stomachs of both rex and Dover soles; had they been identified to species, the overlap between the diets of the Dover and rex soles would appear even small- er. Overlap in the diets of Dover and rex soles of various size groups were also found to be small. Gammarid amphipods, the major crustacean prey of Dover and rex soles, were not principal prey for slender sole and Pacific sanddab. Slender sole and Pacific sanddab fed largely on pelagic crustaceans, as indicated by the occurrence of euphausiids in stomachs of both species, by crab larvae and calanoid copepods in sanddab, and by Pcimkdusjordani in slender sole. Slender sole and Pacific sanddab are chiefly pelagic feeders. Pan- dalus jordani is known to migrate off the bottom at night (Pearcy 1970) and hence could have been consumed on the bottom or in midwater by slender sole. We have occasionally caught both slender sole and Pacific sanddab in midwater trawls at night. Barss"* caught in midwater trawls at night Pacific sanddab that had been feeding heavily on northern anchovy, Engraulis mordax. The only good evidence for benthic feeding by either of these two species is the presence of annelids in their diets (Table 1). Differences Among Stations The proportions (by weight) of the major taxa in the diet of Dover sole were sometimes markedly different among stations (Table 3). Prey composi- tion and availability may be functions of sediment and/or depth. Annelids constituted over 907f of the diet on a weight basis at the three deepest stations (2, 6, and 8), but <13'^ at the shallowest station (22) where the sediment was well-sorted sand. At this shallow station, mollusks and crustaceans were the major food items in the diet. Coelenter- ates ( feeding polyps of sea pens and anemones) and echinoderms (brittlestars) were minor food taxa for Dover sole at all stations except Stations 15 and 23 ( 102 m depth), where together they com- posed over one-half the diet. The proportion offish with food in their stomachs was also higher at these two stations than at any of the other sta- tions. To illustrate the similarities of the food habits of Dover sole among these stations, we constructed a station-station matrix (Table 4) using an index ( C ^) that Horn ( 1966) recommended for comparing overlap in exploitation of alternative food sources ''Barss, W. H. (compiler). 1976. The Pacific sanddab. Oreg. Dep. Fish Wildl., Inf Rep. 76-5, 5 p. Table 3. — The average percentage composition of stomach con- tents of Dover sole on a weight basis at each of the stations. Station (depth in meters i in parentheses) 22 7 15 23 6 2 8 Taxa (74) (100) (102) (102) (148) (190) (195) Annelids 12,6 59 9 250 305 91,9 93.4 91.0 Crustaceans 296 14 1 3,3 3,1 3 7 3.5 6.5 Mollusks 57 9 23,7 13,5 12,4 30 2.2 0.3 Ecfiinoderms 0 11 33-8 30-1 1,4 1 1 2-2 Coelenterates 0 1 2 24,3 238 0 0 0 Otfier taxa 0 0 0 0 1 0 0 0 No, fisfi examined 65 38 22 10 49 91 51 No, fisfi witti stomacfi contents 28 19 20 10 22 61 32 643 FISHERY BULLETIN: VOL, 76, NO. 3 Table 4. — Similarity iCj^) in the diets of Dover sole at the seven stations based on the percentage of major taxa in their diets on a weight basis (below diagonal) and frequency of occurrence of principal polychaete prey (above diagonal). Stations are ar- ranged by depth. Stations 22 Frequency of occurrence 15 23 6 22 7 15 23 6 2 8 0-04 0 0 0.09 0.06 0 058 064 0.57 0.24 0.54 005 0.34 0.56 0.59 0.21 0.38 0 0 34 0.64 0.99 0.09 0.42 0 0.22 0.88 044 052 037 042 0.21 0.87 043 0.52 1.00 0 45 0.21 088 043 0.53 1.00 1 00 within the same habitat. This measure of similar- ity varies from 0 to 1.0. The percentages by weight of the major taxa (lower half of Table 4) were identical at the three deepest (148-200 m) Sta- tions: 2, 6, and 8. Stations 15 and 23, both located at 102 m depth, were also very similar. Station 7 at 100 m was fairly similar ( C^^ > 0.87) to Stations 2, 6, and 8 located in deeper water. The percent of major taxa in the diet of Dover sole at the shallow, sand location (Station 22) was not very similar to any other station (C^^O.58). The frequency of occurrence of principal prey of Dover sole (Table 5) indicates fairly low similarity among different stations for species of polychaetes. Most species occurred at only one or two stations and the assemblage of polychaetes eaten by Dover sole appears to be different at each station. As one would expect, similarity is higher when higher taxa such as gammarid amphipods or pelecypods are considered as a group. For this reason comparisons of similarity among stations should be confined to prey identified to the same taxonomic level. To examine differences in prey species among stations we calculated the overlap in diet (Cj based on polychaetes alone, the most common and speciose prey animals of Dover sole (and the food gi-oup that one of us (Hancock) was familiar with taxonomically). The range in overlap of diets based on frequency of occurrence of individual taxa of polychaetes at these stations (Table 5) was appreciably lower than that based on weight per- centage of major taxa (upper half of Table 4). Sta- tions 2, 6, and 8, which were very similar on the basis of the weight of major taxa in the Dover sole stomachs, overlapped only moderately on the basis of frequency of occurrence of polychaetes ( C^^ = 0.37 - 0.45). Stations 15 and 23, similarly based on major taxa, were the most similar iC , = 0.64) stations based on frequency of occurrence of polychaetes. Station 23 was the next highest in Table 5.— Frequency of occurrence of principal prey of Dover sole from the seven stations. Taxa Stn 22 7 23 15 6 2 8 Polychaeta: Sternaspis tossor 23 24 6 32 Myriochele heeri 31 14 6 8 27 Nothna geophiliformis 9 14 13 Chloeia pinnate 36 16 Melmna cnstata 14 Goniada brunnea 7 8 14 Ancidea uschakowi 5 9 Aricidea neosuecica 9 23 Haploscoloplos leongatus 25 18 8 Owenia fusilormis 10 18 Maldane sarsi 5 6 Chaetozone setosa 8 6 8 9 Terebellides stromenii 11 10 Tharyx multifilis 6 8 Rhodine bitorquata 13 Typosyllis hyalina 25 6 Ammonlrypane aulogasler 5 Nephtys cornuta 11 Anaitides groenlandica 10 Lumbnnens sp 12 6 Ammonlrypane spp 9 Amphitteis 9 18 Ancidea spp 5 Pherusa papillata 7 Tharyx spp. 9 Nephtys spp 5 8 8 Terebellidae 9 Ampharetidae 5 Maldanidae 9 Lumbrineridae 11 Glyceridae 10 8 Spinonidae 10 5 6 9 Hemipodus borealis 10 Laonice cirrata 6 Megelona sp. 5 Crustacea. Pandalus lordani 30 7 16 23 8 gammarid amptiipod 70 22 22 24 14 45 26 copepod 20 cumacean 7 8 Diastylis spp. 8 Valvilera spp 10 Ostracoda 10 14 5 unidentified Crustacea IVIollusca Gastropoda 7 Solenogasters spp 7 8 Pelecypoda 20 20 10 16 8 Yoldia ensifera 8 Lucina sp 9 Megacrenella columbiana 10 10 Tellina salmonea 10 Acila castrensis 10 Ectiinodermata: Ophiuroidea 11 27 16 23 sea pen 12 Miscellaneous *"X^ Nematoda 18 No. observations 10 91 51 49 64 22 38 No. fish 10 91 51 49 65 22 38 similarity with Stations 7 and 15. Thus, the polychaete prey of Dover sole were most similar among these three 100-102 m stations. Because stomachs of the other flounders were examined for only two or three stations, few sta- tion comparisons could be made. As with Dover sole, the percentage of major taxa in the diets of rex sole at Stations 2, 6, and 7 were similar, and food habits were almost identical at Stations 2 and 644 PEARCY and HANCOCK; FEEDING HABITS OF FOUR FLATFISH SPECIES Table 6. — The average percent by weight that major taxa com- posed of the diet of rex sole, slender sole, and Pacific sanddab at Stations 7, 6, and 2. Rex sole Slender sole Pacific sanddab Taxa Stn 7 6 2 7 6 2 7 6 Annelida 58.7 69.7 64.5 1.2 18.5 13.1 8.2 0.1 Crustaceans 392 24.8 29.6 92 2 72,3 77,8 71.3 99.9 Mollusks 2.1 10 05 09 0.8 0 0 0 Echinoderms 0 0 0 0 0 0 0 0 Coelenterates 0 0 0 0 0 0 0 0 Other taxa 0 4.5 55 58 8.5 9 1 20,5 0 No. fish examined 376 210 28 68 844 197 690 33 No. fish with stomach contents 262 160 25 35 403 83 478 13 6 (Table 6). A larger percentage of crustaceans (and the lowest percentage of annelids) was found at Station 7 than at 2 or 6 for both rex and slender soles. But crustaceans were more abundant in the diet of sanddab at Station 6 than at Station 7. Fishes (included as other taxa) were an appreci- able part of the sanddab diet at Station 7. Again, differences in availability of food taxa apparently occurred among stations for the same predator species, and different trends in the importance of food taxa are evident for different species offish at the same stations. The principal prey were most similar for rex sole at Stations 2 and 6, as were the major taxa by weight. The polychaete Nothria geophilifonnis and the larvacean Oikopleura occurred in over 59^ of the observations only at these two stations. It is curious that the planktonic Oikopleura was so fre- quent in the diet of this primarily benthophagus fish. Other prey common at all these stations in- cluded the polychaete Goniada brunnea, uniden- tified polychaetes, gammarid amphipods, and cumaceans. PandaluH jordani was a principal prey species for slender sole at Stations 6 and 7 but not at Station 2. The shrimp Sp/ro/Jtoca/v's hiapinosa and unidentified fish were found in over 5r/c of the fish only at Station 6. Copepods were common only at Station 7. Euphausia pacifica was a principal prey for Pacific sanddab at Stations 6 and 7. Pandalusjor- dani occurred in 2&7( of the fish at Station 6, but was uncommon at Station 7. Decapod crab larvae and copepods, on the other hand, were common prey only at Station 7. Variations With Seasons or Size of Fish Changes in the relative proportions of the major taxa of food consumed by different sizes of the four species of flatfishes are shown for "summer" (May-September) and "winter" (October- April) in Figures 2-5. Because food habits as well as sizes of fishes vary among stations (Tables 3, 6; Pearcy 1978), geogi'aphic effects are confounded in these figures. Annelids usually dominated the diet of all size groups of these juvenile Dover sole during both seasons (Figure 2). Crustaceans appeared to de- crease in importance with increasing size of fish during the winter season, but reached peaks in the summer. Mollusks iSolegasters spp., Yoldia ensif- era, and unidentified pelecypods) attained peaks in the diet of intermediate-sized (200-300 mm) Dover sole, and echinoderms (mainly ophiuroids) attained a peak at a larger size of fish. 100 NUMBER OF FISH rej (3) (12) (27) 136)120) (11) (I) (7) (10) (5) (15) (22) (14) (I) (I) I I I 0 100 200 300 400 1 I r (WINTER) / . v6 / \ ^-. 2/ / ; / \ .1 V / \ ' \ / V Figure 2. — The percent by wet weight of the major food taxa for different length groups of Dover sole for summer and winter. 1 = crusta- ceans, 2 = annelids, 3 = other taxa, 4 = mollusks, 5 = echinoderms, and 6 = coelenterates. 100 200 300 400 STANDARD LENGTH (mm) 645 FISHERY BULLETIN; VOL. 76. NO. 3 The largest difference between seasons was for coelenterates. Anemones and the feeding polyps of sea pens were unimportant constituents of the food during the summer (<29^ of diet by weight) but were sometimes a major food >30'^ by weight) during the winter. Anemones and sea pens are probably available as prey during both seasons but for some reason only consumed in significant quantities during the winter. Seasonal differences in the intensity of feeding were also indicated by the higher frequency of empty stomachs in winter than in summer (Table 7). The number of principal prey occurring in the diet of Dover sole was consistently larger during summer than winter regardless of fish size. Al- though the smaller number of stomachs with con- tents during the winter reduces sample size, and hence the number of taxa found, the frequency of occurrence of many of the individual taxa of polychaetes, crustaceans, and mollusks (taxa listed in Table 5) was higher in summer than winter. Bertrand (1971) found no evidence for sea- sonal variations in the numbers or biomass of infauna sampled with grabs at these stations. Therefore a more diverse assemblage of prey was probably available to Dover sole during the sum- mer or fish were usually less selective during the summer than during the winter. The summer is the season of most active growth of Dover sole (Demory 1972) when intraspecific and possibly in- terspecific competition for food may be most in- tense. Decreased prey selectivity is known to occur under conditions of low food abundance or avail- ability (Ivlev 1961; Schoener 1971). The number of principal prey taxa generally increased with size of Dover sole (Table 7). This trend may be related to sample size (number of stomachs with food) and to the ability of large fish to consume a larger range of prey sizes than small fish. The less diverse diet of small fish resulted from ingestion of only a few species of polychaetes. Table 7. — Frequency of empty stomachs (no. empty stomachs/ no. fish) and the number of principal taxa (occurring in 5% or more of at least 10 observations) of prey for different sizes of Dover sole collected during summer and winter seasons. Frequency of Standard empty stomachs Number of taxa length (mm) Summer Winter Summer Winter 51-100 6/12 15/22 9 4 101-150 — 19/29 — 8 151-200 4/16 10/15 24 12 210-250 7/34 1 9/34 25 18 251-300 11/47 20/42 37 17 301-350 4/24 7/21 21 15 351-400 5,16 — 41 — Those prey types eaten by a broad size range (50- 400 mm SL) of Dover sole include: Myrochele heeri, Typosyllis hyalina, Lumhrlneris sp., Glyceridae, gammarid amphipods, pelecypods, Megacrenella Columbiana, ophiuroids, unidentified polychaetes, and unidentified crustaceans. Annelids and crustaceans were the major food items for rex sole (Figure 3). (Most of the rex sole represented here are juveniles.) Annelids in- creased in importance with an increase in the sizes of rex sole, up to 150-250 mm. This increase was associated with a decrease in the proportion by weight of crustaceans, the dominant food item for small rex sole during both seasons. Euphausiids, decapod crab larvae, copepods, and ostracods were only found as principal prey of rex sole of <200 mm. Mollusks formed only a minor portion of the diet. Differences in the FO of principal prey were not pronounced. Some polychaetes (Sternaspis fos- sor, Myriochele heerie ,Nothria geophiliformis , and Chuelia pinnata) were found more frequently in large (220-300 mm SL) rex sole. Some seasonal differences in the diet of rex sole were evident. Euphausiids were principal prey only during the summer. Cumaceans and Oiko- pleura were more common during the winter. Principal prey that were commonly ingested by all or most size groups during both seasons were: Sternaspis fossor, Goniada brunnea, unidentified polychaetes, gammarid amphipods, and uniden- 100 NUMBER OF FISH f/9J (90) (33 J (41) (2) (6) (48) (63) (50) (82) (13) 1 I \ \ (SUMMER) I no 200 300 0 100 200 STANDARD LENGTH (mm) 300 Figure 3. — The percent by wet weight of the major food taxa for different length groups of rex sole for summer and winter. 1 = crustaceans, 2 = annelids, 3 = other taxa, and 4 = mollusks. 646 PEARCY and HANCOCK: FEEDING HABITS OK FOUR FLATFISH SPF.CIKS tified crustaceans. Kravitz et al. ( 1977) listed Not- hria spp. as frequently occurring polychaetes and Ampclisca macrocephola, Hippumedon wecomus, Paraphoxus epistumusi?), and P. ubtusidens as frequently occurring amphipod prey for rex sole. Crustaceans composed the bulk of the diet of all sizes of slender sole during both seasons (Figure 4). Annelids and "other taxa" were most important in the diet of intermediate-sized (101-200 mm) slender sole during either summer or winter. Pelagic crustaceans such as copepods, eu- phausiids, and crab larvae occurred frequently in the diet of small (<150 mm) slender sole, whereas polychaetes, the shrimps P. Jordan i and S. bi- spinusa, and fishes were important for large slen- der sole (>150 mm). Again, a larger number of principal prey taxa occurred during the summer than winter. Crustaceans also were the most important taxa in the diet of the Pacific sanddab, except for five 201-250 mm individuals during the summer, when fishes composed 959^ of the food by weight (Figure 5). Kravitz et al. (1977) found that all C. sordidiis (90-377 mm total length) collected in May off Oregon had been feeding intensively on northern anchovy. Barss (see footnote 4) reported NUMBER OF FISH ^2 J (15) (70) (41) (6) (7) (64) (179) (130) (7) lOOn X o >- GQ UJ O LJ CL (WINTER) STANDARD LENGTH Figure 4. — The percent by wet weight of the major food taxa for different length groups of slender sole for summer and winter. 1 = crustaceans, 2 = annelids, 3 = other taxa, 4 = mollusks, and 5 = echinoderms. (3) (96)(252)(55) (5) lOOr- NUMBER OF FISH (32) (46) (4) (17) (I) (WINTER) A , 100 200 300 STANDARD LENGTH (mm Figure 5. — The percent by wet weight of the major food taxa for different length groups of Pacific sanddab for summer and winter. 1 = crustaceans, 2 = annelids, and 3 = other taxa. that sanddab eat small fishes, squids, and oc- topuses. Crustaceans were the predominant prey during both seasons and for most sizes of sanddab. Euphausiids, copepods, and cumaceans occurred more frequently in small than large individuals. Pandaliis jordani, crangonids, and fishes were most common in the diet of large Pacific sanddab. DISCUSSION The four common flatfishes caught in this study compose two generalized feeding types. Dover and rex soles feed almost exclusively on benthic inver- tebrates, mainly polychaetes and amphipods, while slender sole and Pacific sanddab prey mainly on pelagic crustaceans. The food habits of these two types are related to mouth structure and digestive morphology. Flatfishes that feed on benthos usually have asymmetrical jaws, small stomachs, and long intestines, whereas pelagic feeders have longer, symmetrical jaws with sharp teeth and long serrated gill rakers, adaptations for grasping and retaining animals that swim in midwater (Hatanaka et al. 1954; Groot 1971). Dover and rex soles belong to the benthos-feeding type and sanddab and slender sole to the pelagic- feeding type. Kravitz et al. (1977) also recognized these two feeding types among five flatfishes off 647 FISHERY BULLETIN: VOL. 76. NO. 3 Oregon, and included rex sole as a benthophagus species and Pacific sanddab as a piscivorous- pelagic feeder. Rae (1956, 1969) .studied the feeding habits of the lemon sole, Microstomas kitt, and the witch, Glyptocephalus cynoglossus, off Scotland. Some of his results are remarkably similar to ours for the congeneric Dover sole, M. pacificus, and rex sole, G. zochirus. Both the witch and lemon sole, like the Dover and rex soles, feed predominantly on polychaetes. Crustaceans were next in importance followed by other phyla such as mollusks, echinoderms, and coelenterates. Ophiuroids and anthozoans were also eaten by both lemon sole and the witch. These similarities in diets indicate common feeding specializations within pleuronec- tid genera. Although the major food of the lemon sole and witch were very similar, these two species, like the Dover and rex soles, preyed on different families or different genera of the same family so that food overlap, and presumably competition, are rare (Rae 1956, 1969). As pointed out by Rae, these differences in feeding habits reflect behaviorial differences of the fishes as well as differences in the composition of the benthic communities of which these fishes are a part. The habitats of the lemon sole and witch often differ, the lemon sole preferring hard, rocky bottoms, and the witch soft, muddy bottoms. Both the lemon sole and witch fed most heavily during the summer. Regional differences were also marked. Polychaetes decreased in importance as prey for the witch in shallow water ( < 100 m), as they did in our study for Dover sole (Table 3). Rae (1939, 1956) also believed that differences in the types and quantities of food available between one area and another resulted in different growth rates of lemon sole. Sedentary polychaetes were most common as prey in areas of rapid growth. One of the objectives of this study was to learn if differences in the availability of prey for flatfishes occurred and how it may be related to sediment types and water depth at our stations. The compo- sition of prey of Dover sole clearly varies among stations. Polychaetes were the main food at the three deepest stations; echinoderms, coelenter- ates, and polychaetes were similar on a weight basis at the two 102-m stations; polychaetes, fol- lowed by mollusks, were most important at the 100-m station; and mollusks and crustaceans were most abundant at the 74-m station (Table 3). Based on the percentage by weight of major food taxa, higher similarities occurred among stations at similar depths rather than with similar sedi- ment types: Stations 15 and 23 at 102 m and the deep stations 6, 2, and 8 at at 148-195 m (Table 4). Sediment texture at Stations 15 and 23 were dis- similar. (See Figure 1 for summary of depth and sediments for the stations.) Although Station 2 had an average sediment texture that differed from Stations 6 and 8, a thin layer of silt overlaid coarse sand at Station 2, hence the surface sedi- ment of Stations 6, 2, and 8 were probably more similar than indicated in Figure 1. The occurrence of individual species of poly- chaetes consumed by Dover sole is probably a more sensitive indicator of station differences than the biomass of major taxa. Stations 7, 15, and 23, at 100-102 m, but with different sediment types, were most similar in polychaete prey. Stations 2, 6, and 8 in deep water, at 148-195 m were again similar. Thus, these similarities in prey for these two groups of stations seem to be correlated with depth. However, polychaete prey at Station 2 ( 190 m) was similar to that of Station 7 (100 m), which had similar sediment type, as well as that at Sta- tions 15 and 23 ( 102 m) with different sediments. Stations 22 and 15 with similar sediment types, but at different depths, had low similarity of polychaete prey. Based on 82 species of mollusks, cumaceans, and ophiuroids sampled in O.l-m^ Smith-Mclntyre grabs, Bertrand ( 197 1 ) calculated the similarity of the fauna among the same seven stations included in this study. He also found that Stations 2, 6, and 8 formed a deep-water group of high similarity. Stations 7 and 23 (at 100-102 m) were similar, as were Stations 7 and 8, with different depths and sediment types. Gunther (1972) also calculated similarities among these same stations based on living benthic foraminifera and found that strong faunal affinities crossed depth and sediment boundaries. Again, Stations 2, 6, and 8 formed one group. Stations 2 and 7, and 15 and 22, station pairs based on sediments, were not very similar. Similarities among the fishes caught were also strong among Stations 2, 6, and 8. The remaining stations (7, 15, 22, and 23) formed another group of high affiinity (Pearcy 1978). These two species associations agree with those described by Day and Pearcy (1968) for the continental shelf off central Oregon. They found a shallow (42-73 m) water association on a sand bottom dominated by Pacific sanddab and English sole, Parophrys vet- iilus, and an association at 119-159 m on a silty- 648 PEARCY and HANCOCK: FEEDING HABITS OF FOUR FLATFISH SPECIES sand bottom dominated by slender sole and rex sole. Shallow-water and deep-water associations are therefore evident at these stations, based on pre- vious studies of benthic invertebrates and verte- brates, as well as the composition of the diet of Dover sole in this study. Because surface sed- iments were fairly similar at our three deep stations, sediment vs. depth effects could not be separated here. The lack of precise similarities of sediment types for station pairs also weakens this part of our study. Nevertheless, stations with the most similar sediment types often had low similar- ity of benthic fauna. We conclude that depth- related factors may have greater influence than sediment type on the composition of benthic fishes, fish food, and invertebrate fauna within the boundaries of our study area. This conclusion must be tempered, however, by the realization that other sediment parameters besides texture and percent organic matter may be important, and we simply did not study the proper sediment characteristics. We agree with Peterson ( 1918): "It is clear then that the character of the bottom is of fundamental importance for the presence or ab- sence of epifauna. Nevertheless, the succession of the various types of epifauna and of the com- munities belonging to the level bottom cannot be explained by the character of the bottom alone." Bertrand ( 1971) estimated the "edible" biomass of infauna ( >1.0 mm) for demersal fishes (i.e., all infauna less holothurians, echinoids, echiurids, and burrowing anemones) at these stations from 0. 1-m'^ Smith-Mclntyre grab samples taken on the same cruises. He detected no seasonal variations in the wet or ash-free dry weight of this biomass fraction. The ash-free dry weights per square meter for polychaetes, mollusks, and crustaceans given by Bertrand for the seven stations are shown in Table 8. Crustacean biomass was consistently low at all stations, probably because of the ineffec- tiveness of the grab to sample epibenthic and motile amphipods, major food items of Dover and rex soles. There was no direct or consistent rela- tionship between the biomass per square meter of T.\BLE 8. — Ash-free dry weights in grams per square meter of macro-infaunal fish food at the seven stations (from Bertrand 1971). Taxa Stn 22 23 15 8 Polychaetes Mollusks Crustaceans Total 0.04 0 17 0.30 008 0.17 0.14 0.19 4.50 1.10 1.74 209 0.20 0.16 0.07 0.006 0.005 0 001 0004 0.008 0.004 0.003 4 55 1.28 2.04 2.17 0 38 0.30 0 26 "edible" fish food and the biomass of all fish or Dover sole. Stations with similar standing stocks of infaunal food had widely different standing stocks of benthic fishes. Station 22, the beach sand station — with the lowest organic carbon in the sediment of all stations — supported a fairly low biomass of fish, but the largest biomass of edible fish food, 4.55 g/m^, and the largest biomass of invertebrate mac- robenthos. Conversely, Wigley and Mclntyre (1964) found the largest biomass in finer sedi- ments off Massachusetts, and Lie and Kisker (1970) found that the shallow-water sand com- munities off Washington had a lower average standing stock of infauna than deeper com- munities on the shelf. The large biomass at Sta- tion 22 is composed primarily of the bivalves Acila castrefhsis and secondarily of Tellina salnwnea. Both of these mollusks were principal prey of Dover sole only at Station 22 (Table 5). Although the frequency of occurrence of these two mollusks in Dover sole stomachs was only 107^ , mollusks composed 58^f by weight of the food of Dover sole at this station. Thus Dover sole are versatile pred- ators, changing their diets opportunistically in re- sponse to changes of prey availability. The dominant fish at Station 22 was Pacific sanddab, primarily a pelagic feeder. Mollusks were not principal prey. Acila and other burrow- ing animals are unavailable as food for fishes adapted for pelagic feeding, illustrating a basic reason for the lack of any direct relationships be- tween edible fish food and fish biomass. The average biomass of Dover sole was directly related to the biomass of their principal food, polychaetes, at the seven stations (Figure 6). Sta- tion 22, where Dover sole consumed principally mollusks, had the lowest biomass of both polychaetes and Dover sole; intermediate values of biomass of both fish and food are found at the three deep stations (2, 6, 8). The three stations at about 100 m (2, 7, and 15) differed markedly in standing stocks of both polychaetes and Dover sole. This positive correlation (r = 0.73) of stand- ing stocks of predator and prey implies that Dover sole selected habitats within our study area where their prinicpal preferred food was most abundant regardless of depth and bottom type. Of more fun- damental interest is the fact that standing stocks of polychaetes may indicate the amount of food available to Dover sole, and perhaps the produc- tion rates of polychaetes at the different stations. Similar direct relationships between standing 649 FISHERY EU'LLETIN: VOL. 76, NO. 3 CM I E 2 4 a> o > o Q 85'7f of their time in water warmer than 20° C and < 10% in water col- der than 18° C (Dizon et al. in press). In the western South Pacific, skipjack tuna have been caught in water near 15° C, off Tasmania (Robins 1952) and off eastern Australia (G. I. Murphy, Division of Fisheries and Oceanography, CSIRO, New South Wales, Australia, 1977, pers. commun.). These fish probably belong to a differ- ent subpopulation (Fujino 1972) than fish found in Hawaii. Lower Dissolved Oxygen Limit Gooding and Neill (see footnote 4) examined the effects of low dissolved oxygen concentrations on skipjack tuna. Their animals, habituated in open tanks with circulating, essentially saturated sea- water (4.5 ml 0.>/l, or 6.4ppm), were transferred to tanks in which the concentration of dissolved oxy- gen could be maintained at a preselected constant subsaturation level. Temperatures in both sets of tanks were ambient, 23° to 24° C. Dissolved oxy- gen concentrations down to 1.0 ml/1 ( 1.4 ppm) were used. Resistance times and swimming speeds were measured, and general behavior was observed for up to 4 h in each experiment. Under their experi- mental conditions, Gooding and Neill concluded that hypoxic stress was first manifest, through changes in swimming behavior and speed, at about 2.8 ml/1 (4.0 ppm), a value fairly typical for fish. Lethal oxygen levels, leading to death in 4 h or less, were found to be higher than those for any other freshwater or marine fish thus far studied. Only one fish (out of six) survived 4 h at 2.5 ml/1 (3.5 ppm), and none survived as long as 2 h at still lower concentrations. At higher oxygen values, above 2.5 ml/1, all skipjack tuna tested in this study survived at least 4 h. Because we sought to estimate the lowest dis- solved oxygen concentrations that skipjack tuna can tolerate indefinitely without significant stress, we have chosen a conservative value of 3.5 ml/1 (5 ppm) as the lower limit to the skipjack tuna's habitat, where temperature and other vari- ables are not limiting. Upper Temperature Limit The case for an upper temperature limit to the skipjack tuna's habitat is somewhat less direct. Three small (30-35 cm) individual skipjack tuna maintained in water warmed l°C/day survived BARKLEY ET AT: SKIPJACK TUNA HABITAT until the temperature reached 33 °C, when two died; the other lived until the water reached 34° C (Dizon et al. 1977). Skipjack tuna have a high metabolic rate and a countercurrent heat ex- changer in their circulatory system which dramat- ically restricts heat loss through the gills. This accounts for the fact that freshly caught wild skip- jack tuna can have red muscle core temperatures as high as 11° C above that of the surrounding water (Stevens and Fry 1971). Temperature excesses of this magnitude could lead to dangerously high muscle temperatures if they occur in the warmer parts of the ocean. To examine this possibility we use a heat balance model developed for skipjack tuna by Neill et al. (1976) which yields an estimate of temperature excess in the red muscle core as a function of size and metabolic activity (Figure la). Actual muscle core temperatures are found by adding the values shown in this figure to the temperature of the surrounding water, for any given size fish. Clearly, large skipjack tuna in surface waters of the tropics must either tolerate high muscle core temperatures, or reduce their metabolic activity substantially below the 3 mg Og g"^ h-^ level. But skipjack tuna appear to avoid heating their muscle tissue much above 35° C (Stevens and Fry 1971 ). This upper limit must place a similar upper limit on the water temperatures which skipjack tuna can inhabit, unless they can thermoregulate physiologically or behaviorally. In Figure lb, 35°C is taken as the upper limit for the red muscle core, and temperature excesses of Figure la are subtracted from that value to arrive at an estimate of the upper temperature limits for the habitat of skipjack tuna, as a function of size. If the values thus obtained are valid, these fish should be able to live anywhere in the ocean when they are small, but they should be limited to lower and lower environmental temperatures as they grow. The largest known skipjack tuna, weighing approxi- mately 16 kg, would — if active enough — be confined to water temperature near 18°C, which is also their approximate lower limit. SKIPJACK TUNA HABITAT HYPOTHESIS We hypothesize that skipjack tuna of the central and eastern Pacific Ocean occupy a primary habitat — a volume of water whose properties they can tolerate indefinitely — which is 18°C or warmer, but cooler than the upper limits for nor- mally active animals shown in Figure lb, provided that the dissolved oxygen concentration is at least 3.4 ml/I (5 ppm). Skipjack tuna can presumably 20 u a: IT hJ a. o v> 3 S to 15 10 ^ "^ 1 ^.^ ,^ "^ ■^ ^ i 1 1 ' 1 ' ^ K ^ ^ ^3mg«( H ^^ |x 1 A r ! i Y' J Y" / / . — / ' — / f , — 1 ■^ ' ' / ■"^ ^r,P» lmg«02g''hr' 1 / ..^ J r^ > 10 15 WEIGHT OF SKIPJACK TUNA (Kg) 20 Figure la. — Calculated excess of internal temperature, over that of the surrounding water, in red muscle for skipjack tuna of all known sizes. Values are shown for a measured minimum (anesthetized) level of metabolic activity (lower line) and our estimate of the mean metabolic activity for normally active animals (upper line), triple the minimum level. (From Neill et al. 1976.) 655 FISHERY BULLETIN. VOL, 76, NO. 3 35 30 iij Q. 2 25 20 15 N — MUSCLE BREAK- DOWN TEMPERATURE (?) \ \ r — - p_ ^Tm = 1 ma*0, a'hr"' \ 1 ^ / \, I 1 r— ______ s s, ■ 1 \ 1 ^ 1 1 N ' 1 1 X X, ^ ; 1 ^3 mg'Og ^-< 1 1 "V. ■-^ ■^, ^-^ ^ ,^^ "^ 10 15 WEIGHT OF SKIPJACK TUNA (kg) 20 FIGL'RE lb. — Maximum tolerable water temperature for skipjack tuna as a function of size for the same two rates of metabolic activity illustrated in Figure la. Based on calculated internal temperature excesses and the assumption that damage to red muscle tissue occurs above 35 "C. 200 la'CU-SSml/LOg-i — «— I8°C^"^ -3.5ml/L02- -I8°C- 5° S 200 Figure 2. — Upper panel: Temperature and dissolved oxygen (selected isopleths only) along long. 119' W, eastern Pacific Ocean, August 1967 (Love 1972). Lower limits of the skipjack tuna habitat are assumed to be either 18° C or 3.5 ml/1 dissolved oxygen, as indicated. The hatched layer should be warm enough for these fish, but oxygen deficient. Lower panel: Hypothesized habitat layers for skipjack tuna of two sizes, 4 kg ( entire hatched area ) and 9 kg ( cross hatched area only) in the same section. Fish <4 kg could presumably live anywhere between the sea surface and the lower limits, 18°C or 3.5 ml/1 of dissolved oxygen. 656 BARKLEY ET AL: SKIP.JACK TUNA HA15ITAT leave their primary habitat only for limited periods of time without suffering thermal or oxy- gen stress. Prolonged excursions to colder water would require increased activity and thus more dissolved oxygen and food. Skipjack tuna could stay in the warm upper layers only if they would reduce their physical activity, or tolerate over- heating. Hypothetical consequences of these con- ditions are illustrated in Figure 2 which shows the hypothetical layers along long. 119°W for skip- jack tuna of two sizes. The 4-kg fish are those most abundant in catches by the eastern Pacific fishery; 9-kg fish are the largest normally found there, and then only in certain areas such as the Revil- lagigedo Islands (ca. lat. 17° N, long. 112°W). The deeper limit of the habitat should be the same for skipjack tuna of all sizes. The upper limit is deeper and more restrictive for larger fish, which find essentially no habitable water between lat. 5° and 12° N, a distance of more than 700 km or 400 n.mi. Larger fish also have much less con- tinuous access to the sea surface than those weigh- ing 4 kg. Only skipjack tuna of the smallest size commonly found in this area (<4 kg) could inhabit all of the water above the lower limits in Figure 2. Figures 3 to 6 are maps of a hypothetical skip- jack tuna habitat for the entire central and east- ern Pacific Ocean, based on oceanographic station data used in preparing the Oceanographic Atlas of the Pacific Ocean (Barkley 19681. For these maps, Figure 3.— Hypothetical maximum depth (meters) of the skipjack tuna habitat in the eastern Pacific Ocean, as determined by the depth of the 18°C isotherm (hatched area) or the 3.5 ml/1 (5 ppm) isopleth of dissolved oxygen (cross hatched area). Contour interval is 50 m except for a few areas near the coast, where a 25-m contour interval is used. 657 FISHERY BULLETIN, VOL. 76, NO. 3 Figure 4. — Hypothetical minimum depth of the skipjack tuna habitat in the Pacific Ocean east of long. 180°, for fish weighing about 6.5 kg ( 14 lb) which are limited to water cooler than 24° C. Contours show the depth, in meters, of the 24 °C isotherm. station data were averaged within 2° areas of latitude and longitude, for all months, to approxi- mate annual mean conditions. Figure 3 shows the hypothetical "floor" of the skipjack tuna habitat, i.e., maximum depths (in meters) for this species. In Figure 3, the unhatched areas off the Pacific coast of the Americas, and at latitudes higher than about 40° in both hemi- spheres, indicate water which, at all depths, is colder on average than 18° C; presumably skipjack tuna would not normally be present. Figure 4 shows the minimum habitat depth or ceiling for 6.5-kg fish. Outside of the hatched area annual mean surface temperatures are <24°C, and 6.5-kg fish would normally have access to all of the water column above the habitat floor (Fig- ure 3), up to and including the sea surface. Figure 5 shows the hypothesized habitat layer thickness for 6.5-kg skipjack tuna. In some areas, there is no water cooler than 24° C with more than 3.5 ml/1 dissolved oxygen, so in these areas there is no habitat for 6.5-kg or larger fish; such areas are double hatched on Figure 5. Extensive regions around these areas have habitat layer thickness of 10 m or less, and large areas of the equatorial Pacific Ocean have <25 m of habitat layer thick- ness. This rather thin layer can lie beneath as much as 1 50 m of water warmer than 24 ° C ( at lat. 4°N, long. 170°W, e.g.). North of the Hawaiian Islands, the opposite situation is present: a 150-m thick habitat layer lies under 25 to 50 m of water 658 BARKI.EY KT AI.: SKIPJACK TUNA HABITAT Figure 5.— Thickness of the hypothetical habitat layer (meters), for 6.5-kg skipjack tuna in the eastern and central Pacific Ocean. Contours were obtained by subtracting depths of the upper habitat limit (Figure 4) from the lower one (Figure 3). In the crosshatched areas off Mexico and Peru, there should be no habitat suitable for fish of this or larger sizes. In the hatched area, water warmer than 24°C is present above the habitat layer. Immediately beyond this area the habitat (dashed depth contours) extends to the sea surface. Outside of the 18° C surface isotherm, the water is probably too cold for skipjack tuna. warmer than 24 C. Clearly, the South Pacific Ocean offers the roomiest habitat to large skipjack tuna, and in fact some of the largest known indi- viduals of this species were caught slightly south of Tahiti (lat. 17"S. long. 150"W), according to catch records of longline fishing boats (R. A. Skillman, Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA. Honolulu, HI 96812, 1977, pers. commun.). An interesting feature of the habitat map for 6.5-kg fish I Figure 5) is a channel of relatively cool and adequately oxygenated water some 200 km off the coast of Mexico. This channel should allow skipjack tuna as large as 6.5 kg to pass from the Baja California fishery to the equatorial fishery, or vice versa, when fish as small as 4 kg would find stressful conditions for hundreds of kilometers on either side of that channel. Seasonal and year-to- year closure or shifting of this channel could readi- ly explain puzzling variations in the distribution of skipjack tuna catches in the eastern Pacific. Figure 6 shows areas of presumably stressful environment (zero habitat thickness) in the east- ern tropical Pacific for fish of various sizes and therefore temperature limitations. Skipjack tuna >11 kg should find no habitat at all within the shaded area. Fish 4 kg in size would find no habitat within the smallest contour (temperatures above 659 FISHERY BULLETIN, VOL. 76, NO. 3 26° C). Fish weighing <4 kg should find some thickness of habitable water, within and just below the upper mixed layer, everywhere in the eastern tropical Pacific. DISCUSSION Although we anticipate that temperature and dissolved oxygen will prove to be primary deter- minants of the habitat of skipjack tuna in all oceans, it is possible that limiting values of these variables may differ from one population or region to another. The lowest temperature (ca. 15° C) at which skipjack tuna are caught in Australian waters (Robins 1952) is considerably lower than in the eastern Pacific. The fish caught off Australia may also differ in their ability to tolerate warm or low-oxygen water. The gross features of the distribution of skipjack tuna in the eastern tropical Pacific, where only small skipjack tuna are found in large numbers (Williams 1970), agi-ee with the hypothesis. Those areas where large skipjack tuna do occur (Ma- tsumoto 1975) are outside of the hatched area in Figure 6: the Revillagigedo Islands, e.g., are just north of the hatched area, and Tahiti is well south of it. The hypothetical habitat proposed here ex- plains why skipjack tuna leave the northern fishery of the eastern Pacific when they reach a certain size. To find cooler, better oxygenated water as they grow, these fish must move out of the eastern tropical Pacific toward higher latitudes in the central Pacific. Also, they must then spend less time at or near the sea surface, since the thermo- cline, where they live, is generally much deeper in the central Pacific, and the water above the ther- mocline is too warm to permit normal activity. This size-specific movement in response to the en- vironment is consonant with Rothschild's (1965) migration model for the eastern Pacific skipjack tuna population. It also suggests a mechanism for the evolution of migratory processes, an important topic in marine ecology. For several reasons, existing fishery data are inadequate for making a refined judgment of our skipjack tuna-habitat hypothesis: 1) Commercial fishery data generally include neither information Figure 6. — Average water temperature in the eastern Pacific Ocean at those depths where the concentration of dissolved oxygen is 3.5 ml/1. Deeper water is cooler and lower in oxygen, shallower water is warmer and has more oxygen. See Figure lb for the relationship between skipjack tuna size and upper temperature limits. 660 BARKI.EY ET AL: SKIPJACK TUNA HABITAT on individual sizes of skipjack tuna composing the catch nor synoptic information on the vertical dis- tribution of temperature and dissolved oxygen in the fishing area. 2) The degree to which catch per unit effort measures fish abundance may vary greatly with gear type, fish size, and environmen- tal conditions. For example, the habitat hypothesis implies that purse seines (which fish the upper 50 m or so of the water column ) should be most effective in those parts of skipjack tuna habitat with a shallow floor, hi fact, the eastern Pacific purse seine fishery operates almost en- tirely in waters with a skipjack tuna habitat-floor at depths of 50 m or less (c.f. our Figure 3 and fig. 1 of Matsumoto 1975). Efforts to catch skipjack tuna by purse seining in Hawaiian waters — where the habitat-floor lies at depths near 200 m (Figure 3) — have been ineffectual (Murphy and Niska 1953 k Green ( 1967) reported strong positive corre- lations between the success of purse seining for eastern Pacific skipjack tuna and yellowfin tuna, Thunnus albacares, and the presence of shallow ( «60 m to top), steep ( >0.55°C cm ' ) thermoclines. 3) Commercial fishermen naturally fish only where they expect to find and catch fish; thus, fishing effort tends to be very unevenly distri- buted. A partial test of the habitat hypothesis might be achieved through experimental fishing in and near the hatched areas of Figure 6; fishing effort, the sizes of captured skipjack tuna, and vertical distributions of temperature and dissolved oxygen would need to be measured at each fishing loca- tion. But, experimental fishing — even if con- ducted in a thorough and systematic fashion — might not yield a conclusive test of the hypothesis, because there would still be no guarantee that catch per unit effort accurately reflected flsh abundance (reason 2 above). We advocate, instead, the application of ul- trasonic telemetry to test the skipjack tuna- habitat hypothesis. Because skipjack tuna tagged with ultrasonic transmitters tend to stay with their school (Yuen 1966) and because skipjack tuna schools tend to be homogeneous with respect to fish size (Brock 1954), the track of a single tagged fish could be taken as representative of a large number of normally behaving, similarly sized fish. Pressure-sensitive ultrasonic transmit- ters ( like that described by Luke et al. 1973 ) would permit continuous monitoring of fish position in all three spatial dimensions through time. Spatial-temporal coordinates offish could then be compared with synoptic data on vertical distribu- tions of temperature and dissolved oxygen. A few dozens of such comparisons, for fish of various sizes in waters with diverse vertical distributions of temperature and dissolved oxygen, would consti- tute a valid and sufficient test of the habitat hypothesis. Toward this end, preliminary tele- metry work is now underway at this Laboratory. SUMMARY Work with captive skipjack tuna at this Laboratory has yielded information on the tem- perature and dissolved oxygen requirements of this species. If these laboratory results apply to skipjack tuna in nature, they provide new insight into the evolution of migration in skipjack tuna populations, make it possible to account for the geographic distribution of skipjack tuna on the basis of environmental conditions, and provide means for predicting their movements in major fisheries such as those of the eastern tropical Pacific. In particular, we suggest that only young skip- jack tuna can inhabit tropical surface waters, and that the habitat of adult skipjack tuna in the tropics is the thermocline and not the warmer surface layer, as has generally been thought. Since the thermocline in many areas is too oxygen-poor to support these active fish and the well-oxygenated surface layer is too warm for adult skipjack tuna, only heat-tolerant young skipjack tuna can live in those areas. As they grow, these fish are forced to move into areas where well-oxygenated water of the proper tem- perature is more readily available. Up to now, it has not been possible to trace the movements of migrating skipjack tuna largely be- cause they move through areas of many millions of square miles, at unknown depths. Knowledge of their temperature and dissolved oxygen require- ments dramatically reduces the scope of the prob- lem: the fish should be in a well-defined layer of water, of directly and easily measured thickness, whose geographic extent can be sharply defined with either historical or current oceanographic observations. ACKNOWLEDGMENTS The physiological studies on which this paper is based were supported, in part, by the University of Wisconsin Brittingham Foundation. 661 FISHERY BULLETIN, VOL. 76, NO. 3 We thank John J. Magnuson (University of Wisconsin) and Garth I. Murphy (CSIRO, Aus- traha) for reading this manuscript and providing valuable suggestions. LITERATURE CITED Bakkley. R. a. 1968. Oceanographic atlas of the Pacific Ocean. Univ. Hawaii Press, Honolulu, 20 p., 156 fig. BROCK, V. E. 1954. Some aspects of the biology of the aku, Katfiuwonus pelamix. in the Hawaiian Islands. Pac. Sci. 8:94-104, DIZON, A, E. 1977, Effect of dissolved oxygen concentration and salinity on swimming speed of two species of tunas. Fish. Bull., U.S. 75:649-653. DIZON. A. E., R. W. BRILL. .AND H. S. H. YUEN In press. Correlation between environment, physiology, and activity and its effect on thermoregulation in skipjack tuna, Katsuu'onus pelamis. In G. D. Sharp and A. E, Dizon (editors). The physiological ecology of tunas. Academic Press, N.Y. DIZON, A. E., W. H. Neill, and J. J. Magnuson 1977. Rapid temperature compensation of volitional swimming speeds and lethal temperatures in tropical tunas (Scombridae). Environ. Biol. Fishes, 2:83-92. FU.JINO, K, 1972. Range of the skipjack tuna subpopulation in the western Pacific Ocean, In K. Sugawara (editor), The Kuro.shio II, p. 373-384. Saikon Publ. Co. Ltd., Tokyo. Green, r. e. 1967. Relationship of the thermocline to success of purse seining for tuna. Trans. Am. Fish. Soc. 96:126-130. Love, C. M. (editor). 1972. EASTROPAC atlas. Vol. 5. Physical oceanographic and meteorological data from principal participating ships. Second survey cruise, August-September 1967, U.S, Dep. Commer., NOAA Tech. Rep. NMFS CIRC-330. Luke, D. McG., D. G, Pincock, and a. B. Stasko. 1973. Pressure-sensing ultrasonic transmitter for track- ing aquatic animals. J, Fish, Res. Board Can. 30:1402- 1404. MATSUMOTO, W. M. 1975. Distribution, relative abundance, and movement of skipjack tuna, Katsuwonus pelaniis, in the Pacific Ocean based on Japanese tuna longline catches, 1964-67. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-695, 30 p. MURPHY, G. I., AND E. L. NISKA. 1953. Experimental tuna purse seining in the central Pa- cific. Commer. Fish. Rev. 15(4):1-12. Neill. w. h., r. k. c. Chang, and a. e. dizon 1976. Magnitude and ecological implications of thermal inertia in skipjack tuna, Katsuwonus pelamis (Lin- naeus). Environ. Biol. Fishes. 1:61-80. Robins, J. P, 1952, Further observations on the distribution of striped tuna., Katsuwonus pelamis L., in eastern Australian wa- ters, and its relation to surface temperature. Aust. J. Mar. Freshwater Res. 3:101-110. Rothschild, B. J. 1965. Hypotheses on the origin of exploited skipjack tuna (Katsuwonus pelamis) in the eastern and central Pacific Ocean. U.S. Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 512, 20 p, 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, Williams, F, 1970. Sea surface temperature and the distribution and apparent abundance of skipjack ( Katsuwonus pelamis) in the eastern Pacific Ocean 1951-1968. Inter-Am. Trop. Tuna Comm. Bull. 15:229-281. Yabe, H., Y. Yabuta. and S, Ueyanagi 1963. Comparative distribution of eggs, larvae and adults in relation to biotic and abiotic environmental factors. /n H. Rosa, Jr. (editor). Proceedings of the World Scientific Meeting on the Biology of Tunas and Related Species, p. 979-1009. FAO Fish. Rep. 6. YUEN, H. S. H. 1970. Behavior of skipjack tuna, Katsuwonus pelamis. as determined by tracking with ultrasonic devices. J. Fish. Res, Board Can. 27:2071-2079. 662 HISTORICAL TRENDS AND STATISTICS OF THE SOUTHERN OSCILLATION, EL NINO, AND INDONESIAN DROUGHTS William H. Quinn, David O. Zopf. Kent S. Short, and Richard T. W. Kuo Yang' ABSTRACT A 116-yr Southern Oscillation index record was used in conjunction with environmental data and reports from various authors on disturbances to the anchoveta fishery, marine bird life, etc. off the Peruvian coast, to infer the occurrence of past El Nino type events and their intensities. The resulting long time history substantiates our earlier report that certain Southern Oscillation index features are excellent precursors of subsequent El Nino type events. We suggest that statistics derived from this time history could be useful in the management of the Peruvian anchoveta fishery and for providing long-range outlooks on El Nino type activity. Anomalously heavy precipitation in the central and western equatorial Pacific and Indonesian droughts were closely associated with El Nino type events. In recent years the world demand for fishmeal has continued to increase, as has the world population. The Peruvian anchoveta fishery, which ordinarily provides over half the world's supply of fishmeal, has become a critical resource; and anything that affects the output of this fishery is of world-wide significance. Johnson and Seckel (1977) reported that the catch in this fishery declined from a high of over 12 million tons (about V.-> of the total world catch of all fish) in 1970 to about 2 million tons in 1973. Although overfishing in 1970-71 may have contributed heavily to this decrease in anchovy catch, the strong El Nino of 1972-73 was undoubt- edly also a major cause for the precipitous decline in catch (Figure 1). However, the 1975 catch was still only about 257c of the record 1970 catch, the 1976 catch remained low, and the target for 1977 has now been reduced to 2 million tons of an- choveta and other fish such as sardines and hake. Apparently the unfavorable environmental condi- tions caused by the very weak event of early 1975 and the moderate El Nino of 1976-77 have not only contributed to the delay in recuperation of the fishery, but also are causing a further degradation of it. In early October 1977 the Fisheries Ministry of Peru said (according to a Reuters wire service report) that the stocks were believed to be so low that the anchoveta fishing, which was suspended in May 1977, would not resume until the second half of 1978. 'School of Oceanography, Oregon State University, Corvallis, OR 97331. Statistical information pertaining to the histor- ical occurrence of El Nino type events is presented to: 1) aid in long-term fishery assessment (Peru- vian anchoveta fishery); 2) provide a basis for speculative long-range outlooks on event occur- rence (beyond a year in advance); and 3) guide long-range predictions ( 1-12 mo in advance). Rela- tionships between El Nino type events. Southern Oscillation index trends, index component trends, and Indonesian droughts are shown and discussed. LD 14 Li_ IP _ 0 '"^^ CO 1 10 _1 _) ^ 81- X o I- < > o X o 6- 4 - 0 I I I ± J \ \ \ I I 1965 1970 YEARS 1975 Manuscript accepted January 1978. FISHERY BULLETIN: VOL. 76, NO. 3. 1978. Figure l. — The Peruvian anchovy catch for the period 1962-76 as obtained from the Industrial Fishery Products Market Review and Outlook for June 1977 'National Marine Fisheries Service 19771. The 1976 figure is a preliminary value. The 1977 figure is the Peruvian Fishery Ministry target value for anchovies and other species such as sardine and hake, as reported by Reuters wire service on 19 October 1977. 663- FISHERY BULLETIN: VOL 76, NO. 3 Wooster (1960), Idyll (1973), Miller and Laurs ( 1975), and Caviedes ( 1975) furnished background information on El Nino; Quinn (1974) discussed monitoring and prediction; and Berlage (1957, 1966). Troup (1965), and Quinn (1971, 1976) pro- vided background information on the Southern Oscillation and how it relates to phenomena dis- cussed in this paper. Definitions for terms frequently used in this paper follow: The Southern Oscillation was origi- nally identified by Walker (1924). It was loosely defined by Berlage (1966) as a fluctuation in the intensity of the intertropical general atmospheric and hydrospheric circulation over the Indo-Pacific region. The fluctuation is dominated by an ex- change of air between the South Pacific subtropi- cal high and the Indonesian equatorial low. The differences in sea level atmospheric pressure between sites representing the South Pacific sub- tropical high and sites representing the Indone- sian equatorial low are used as indices to repre- sent the Southern Oscillation (Quinn 1974). The El Nino type event refers to the appearance of anomalously warm sea surface temperatures and abnormally heavy rainfall in the equatorial Pacific and an invasion of anomalously warm sur- face water off the coast of Peru and southern Equador. This event, which is brought about by relaxation from a prolonged period of strong southeast trades, is represented by falling and low Southern Oscillation indices (Quinn 1974). The magnitude of the interannual relaxation and its timing with relation to the regular seasonal relax- ation (Southern Hemisphere summer) appear to determine the strength of the El Nino invasion along the Peruvian coast. Heavy central and west- ern equatorial Pacific precipitation usually starts a few or more months after El Nino initially sets in, but this may not always be the case. By using the term "El Nino type" we avoid arguments over what is and what is not an El Nino and can then account for events that evolve in a similar manner but vary in timing, intensity, and extent. The anti-El Nino refers to the contrasting situa- tion when a strengthening and strong southeast trade system prevails (represented by rapidly ris- ing and high Southern Oscillation indices). At such times we can expect strong upwelling (due to the divergent equatorial flow under the influence of strong southeast trades and equatorial easter- lies), anomalously low sea surface temperatures, and abnormally low amounts of rainfall over the equatorial Pacific. Also, off the coast of Peru, we find strong coastal upwelling, low sea surface temperatures, lower than average sea level, and generally favorable physical environmental con- ditions for biological productivity (due to the up- welling of nutrient-rich water from lower levels). METHODS Data Processing Atmospheric pressure and much of the rainfall data before 1961 were obtained from the World Weather Records (Clayton 1927, 1934; Clayton and Clayton 1947; U.S. Department of Commerce 1959, 1968). Data for 1961-76 were obtained from Monthly Climatic Data for the World (U.S. De- partment of Commerce 1961-76). We were primar- ily interested in the large-scale interannual changes. Therefore, we eliminated regular oscilla- tions from the data, such as the diurnal cycle, by using monthly mean values (or monthly amounts, for rainfall), and the seasonal or annual cycle by subtracting long-term average or normal monthly values from the actual monthly values. Data so processed show no particular regularity and no apparent cycle (Panofsky and Brier 1965). The filtered and unfiltered monthly anomalies were used to detect, identify, and evaluate any unusual changes that took place. Our interests were focused on fluctuations of an intermediate scale (Southern Oscillation), with periods ranging between about 1 and 6 yr. The remaining short period fluctuations in the anomalies were eliminated by filtering with a low pass filter. At the other end of the time scale, there m.ay be a gradual change of the variate over many years which is part of oscillations that are long compared with the record. These extremely long, gradual changes were not a factor in our study. In earlier papers (e.g., Quinn 1974, 1976) the 12-mo running mean was applied directly to monthly values of pressure, pressure differences (indices), rainfall, etc. as a low pass filter. This filter not only smoothed the data to some extent but also eliminated the annual cycle. To more clearly define the interannual fluctuations (Southern Oscillation), we recently switched to the use of the triple 6-mo running mean filter on the monthly anomalies, which requires three suc- cessive passes of the 6-mo running mean over the data. It results in smoother plots and more clearly defined peaks and troughs, which are of particular assistance in establishing long-term trends. The 664 QUINN ET AL.: SOUTHERN OSCILLATION. EL NINO. AND INDONESIAN DROUGHTS loss of 3 mo time with each application of the 6-mo running mean is a drawback to its use in forecast- ing, so we also use the 3-mo running mean and monthly plots of anomalies for locating inflection points and evaluating trends on a more immediate basis in support of forecasts. Anomaly trends for several indices were main- tained in time section plots (Figure 2a, b) to evaluate the Southern Oscillation and its expected effects on the southeast trade system. Although these limited records ( 25-30 yr) clearly showed the close association of low indices with El Nino type activity, and high indices with anti-El Nino condi- tions (Quinn 1974, 1976), it was essential to extend the study over a much longer period to determine how frequently these climatic extremes occurred The World Weather Records were searched for the longest and most complete atmospheric pres- sure records which could be used to extend our study into the past. Madras, India (1841-1976); Bombay, India (1847-1976); Djakarta, Indonesia (1866-1974); and Darwin, Australia (1882-1976) were within the area noted by Berlage ( 1957, 1966) to reflect Southern Oscillation-related pres- sure changes in the Indonesian equatorial low pressure cell. Santiago, Chile ( 1861-1976) had the only long pressure record that could possibly rep- resent Southern Oscillation-related pressure changes affecting the South Pacific subtropical high pressure cell. Although Santiago is generally to the east of the subtropical high, it does reflect these pressure changes (Berlage 1957, 1966). Correlations were run between the Tahiti- Darwin index and the Santiago-Darwin index on data for 1935-76 to further substantiate use of the Santiago-Darwin index for representing the Southern Oscillation and related El Nino type ac- tivity. The Tahiti-Darwin index was used for this comparison since it and the Santiago-Darwin index showed similar amplitudes in their interan- nual fluctuations. The similarity was due to the fact that Tahiti and Santiago are separated by analogous distances from the usual core of activity in the subtropical high ( see fig. 10 in Berlage 1957, or fig. 10 in Bjerknes 1969). At zero lag the correla- tion coefficient between the two indices was 0.88. The maximum correlation was 0.89 when the Tahiti-Darwin index led the Santiago-Darwin index by 1 mo. Figure 3a-h shows the triple 6-mo running mean plots of pressure anomalies for Madras (1841-1976). Bombay (1847-1976), Djakarta (1866-1974), and Darwin (1882-1976). They also show similar plots of pressure index anomalies for Santiago-Bombay ( 1861-81) and Santiago-Darwin (1882-1976). The anomaly plots were used along with other data in the evaluation of El Nino type events reported over the past 135 yr. Classification of Events The classification of El Nino type events by in- tensity is highly subjective since no two cases are exactly alike with regard to time of onset, dura- tion, areal extent, thermal departure, degree of devastation, etc. Determinations concerning event occurrence and intensity were primarily based on: 1) reported disruptions of the anchoveta fishery and marine bird life off the coast of Peru; 2) scientific reports which discussed events that af- fected the coastal regions of Peru and southern Ecuador [e.g., Eguiguren ( 1894), Frijlinck ( 1925), Murphy (1926), Hutchinson (1950), Sears (1954), Schweigger (1961)1; 3) hydrological data for the Peruvian coastal region; 4) sea-surface tempera- ture data along the coasts of Peru and southern Ecuador; 5) rainfall at coastal stations in Peru and southern Ecuador; 6) height of preevent peaks and depth of relaxation troughs in Southern Oscilla- tion index trends; 7) related indications from index component trends (when pressure compo- nents from only one core of the Southern Oscilla- tion were available); 8) sea-surface temperatures over the equatorial Pacific; 9) rainfall data for islands in the central and western equatorial Pacific. We categorized events as strong, moderate, weak, or very weak, depending on the intensity of the activity and the time of year that it occurred. The true El Nino sets in during the first half of the year. A symptom which is common to El Nirios is the presence of anomalously high sea-surface temperatures off the coasts of southern Ecuador and Peru. Other frequently mentioned features include a southward coastal current, heavy rain- fall, red tide (aguage), invasion by tropical nekton, and mass mortality of various marine organisms including guano birds, sometimes with sub- sequent decomposition and release of hydrogen sulfide (known as El Pintor) (Wooster 1960). Strong El Ninos are recognized as such by all investigators; they involve positive sea-surface temperature anomalies along the coast in excess of 3°C, they display most of the aforementioned fea- tures, and the anchoveta fishery is seriously 665 FISHERY BULLETIN: VOL 76, NO. 3 I9«8 1909 1950 1951 195? 1953 1954 1955 1956 1957 1956 1959 I960 1961 1962 I96i 5 _ •■■■•. 1. -..1 - 5 2 - .••■■■ ........... ..•••■—] *••••., .• •_ 2 JUAN FERNANDEZ - DARWIN ( tibl 0 -2 2 - ■•■ ..... * • 0 -2 2 -•...• .••■•••. ., ... ..•••■■•• • ^ ■': 0 '. .•• .*■ • .••■■• 1 0 EASTER -DARWIN (mtJl ..' ,.."• ..*' .... •■ -2 - ••..••■' .. .• -2 -3 2 - r -3 2 TOTEGEGIE- DARWIN 0 .•■■ ■•. .■■■■•■■. .. 1 0 -1 (mbl u ..•■ •. ..••■ "...C -2 2 - -2 2 RAPA- DARWIN 0 -1 - •• .■•••.. .' *'*** h*. 0 (mbl - •■'"■•.. .•■"■•.-. ...••' '■•••' -2 2 _ _.•■ .•*" '■•..•' ..- -2 2 0 ~ *.. ... .- ..••-.......• -:.- 0 (mbl •-. .. ..• ••••■ -2 - -2 -3 - -3 EN(W) EN(M) EN(S) » EN(W) 1955 1956 YEARS 1957 1953 Figure 2a. — Triple 6-mo running mean plots of anomalies of the difference in sea level atmospheric pressure ( millibars) between Juan Fernandez Is. (33°37'S, 78°50'W) and Darwin, Australia 1 12°26'S, 130"52'El, between Easter Is. (27°10'S, 109°26'W) and Darwin, between Totegegie i23°06'S, 134°52'W) (Gambler Is.) and Darwin, between Rapa (27°37'S, 144°20' W) (Austral Is.) and Darwin, and between Tahiti ( 17°33 'S, 149°20'W) (Society Is.) and Darwin for 1948-63. El Nino t.ype events (EN) are indicated in strong (S), moderate (M), and weak or very weak (W) intensity. i96a 1965 1966 1967 1968 1969 1970 1971 1972 1973 197a 1975 1976 1977 3 - - 5 2 - ... . - 2 JUAN FERNANDEZ - 1 ".* •••■ - 1 DARWIN (mb) 0 -1 -2 ..•" • 0 -2 - •.. - 2 2 1 0 r---.. ... .■ - 1 EASTER-DARWIN (mb) -1 - ..■ -1 -1 -2 - — -2 -3 - •:.■• ■-. .... - -3 2 .... • • - 2 TOTEGEGIE -DARWIN 1 0 F •■ '•• • " (mb) : . -2 - ■•■.... •' ... ■-...• ' .. ■•. .•■■■• •.•■ - -2 2 2 1 — .■• *.. . '. • — 1 RAPA- DARWIN 0 -1 •'—'•- ." ■... . .. (mb) ,-.....* ••** •' . _ -1 -2 : '•..•■ ■••" "•...•■ .•• ..•■••• •. .• — -2 2 2 i — ,••• *•. , ■..._ . - I 0 -.• — • '"'■*. '.•—-. . • ■. . • (mb) -1 j- ", .•■ '; '.-' .' — -1 -2 - ■-•■ - -2 -3 - - -3 EN(M1 EN(W) EN(S) > EN(W) EN(M) I96« 1965 1966 1967 1968 1969 1970 1971 YEARS 1972 1975 I97fl 1975 1977 1978 Figure 2b. — Triple 6-mo running mean plots of anomalies of the difference in sea level atmospheric pressure (millibars) between Juan Fernandez Is. and Darwin, between Easter Is. and Darwin, between Totegegie and Darwin, between Rapa and Darwin, and between Tahiti and Darwin for 1 964-76. El Nino type events (EN) are indicated in strong (S) , moderate ( M) , and weak or very weak ( W ) intensity. 666 QUINN ET AL.: SOUTHERN OSCILLATION. EL NINO, AN!) INDONESIAN DROUGHTS IB4I 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 2 1 BOMBAY 0 (mb) .1 -2 - - 2 - - -1 •2 2 1 MADRAS 0 (mb) .1 -2 1841 ......■•■■ *••< 1850 1851 - 2 1 0 -1 •2 1842 1843 IS44 EN(S)- 1845 1846 1847 1848 1849 1852 2 1 BOMBAY 0 (mb) -1 -2 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 2 - ••.•• ■•••■" _-l 2 2 MADRAS 0 (mb) -1 -2 - •■• •• ...•••■■■■ ,.. ■— ^._ •**r. .••■■ ■••-...-H 2 1 0 -1 ■2 - 2 SANTIAGO - 0 BOMBAY ^ (mb) ■2 : .■••■■■• - 2 1 - EN(S) *■*. -1 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 YEARS Figure 3a. — Triple 6-mo running mean plots of sea level atmospheric pressure anomalies fmillibars) for Bombay ( 18'54'N, 72''49'E), India ( 1847-65) and for Madras ( 1300'N, 80"11'E), India ( 1841-65); also triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago ;33°27'S, 70°42'W), Chile and Bombay (1861-65). El Nino type events (EN) are indicated in strong (S) or moderate (M) intensity. 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 2 _ 2 BOMBAY 0 (mb) ., T. •■■ '■■-.. .— •.. •■ •• '■•. ..••••■■ **■•• ~ 1 _ *■ ..••*" *' *•„ -1 -2 - 2 2 _ 2 1 MADRAS 0 (mb) -1 - ,•' ', .•* •-. - 1 _ 1 '*•••• " "•. ^. ,.•• -2 ^ -2 2 1 ..■••■■••• .. •..^ 2 t ••• •, ,.•* (mb) .1 - ••■ *• ..•••* •1 -2 - ■2 2 1 - ■••. .. 2 SANTIAGO- 0 BOMBAY -1 *.. ••••. EN(S)- -••— ••.. •••. (mb) -2 - *••••••* * 2 -3 - EN(M) EN(M) ■••....•■ EN(M) 3 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1379 1880 1881 YEARS Figure 3b. — Triple 6-mo running mean plots of sea level atmospheric pressure anomalies (millibars) for Bombay and Madras, India, and Djakarta i06'irS, lOG'Sl'E), Indonesia (1866-81); also, triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago and Bombay ( 1866-81). El Nino type events (EN) are indicated in strong (S) or moderate (M) intensity. 667 FISHERY BULLETIN; VOL. 76. NO. 3 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 2 _ - 2 BOMBAY 0 (mb) .1 - ..•■■ .■■••■• — _ ■•* - -1 -2 _ - -2 2 - 2 MADRAS 0 - '-. .■ '*•.. •■*' (mb) -1 - -1 ■2 _ -2 2 2 DJAKARTA 0 - ■••. n ' ,..-• ...•' (mb) -1 _ --' -2 ^ --2 2 1 1 , ^2 1 - _.'••■■. 1 DARWIN 0 •••' (mb) -1 — .... ■1 -2 - -2 2 - ..... 2 1 SANTIAGO - ~ 0 DARWIN ■ •■....-• .1 (mb) ^ - EN(S)- EN(M)- — EMS) EMM) 2 1682 1883 1884 1885 1886 1887 I88S 1889 1890 1891 1892 1893 1894 1895 1896 1897 Figure 3c. — Triple 6- mo running mean plots of sea level atmospheric pressure anomalies (millibars) for Bombay, Madras, Djakarta, and Darwin (12°26'S, 130"52'E), Australia (1882-97); also triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago and Darwin ( 1882-97). El Nino type events (EN) are indicated in strong (S) or moderate (M) intensity. 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 2 1 - •— 2 1 (mb) -1 -2 - -2 2 MADRAS 0 (mb) -1 •2 - 2 1 0 - 2 2 DJAKARTA 0 (mb) -1 -2 - 2 1 - - •2 2 1 DARWIN 0 (mb) .1 •2 - - ...•:! 2 1 0 -...••■■ ...-• -1 -2 2 SANTIAGO - 0 DARWIN -1 (mb) -2 •3 r 2 - EN(S) — .•■' EN(M) EMM) EMS)— 2 3 1898 1899 ' 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 YEARS Figure 3d. — Triple 6-mo running mean plots of sea level atmospheric pressure anomalies (millibars) for Bombay, Madras, Djakarta, and Darwin ( 1898-1913); also, triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago and Darwin (1898-1913). El Niifio type events (EN) are indicated in strong (S) or moderate (M) intensity. 668 ca'INN ET AL : SOUTHERN OSCILLATION. EL NINO. AND INDONESIAN DROUGHTS 1914 1915 1916 1917 1918 1919 19^0 1921 1922 1923 1924 1925 1926 1927 1928 1929 2 BOMBAY 0 (mb) -1 -2 - *****. •••^ - 2 1 '••..... ■■■■■••■..• ■**' - ■2 2 MADRAS 0 (mb) .1 ■2 - "., .«•••• •"*•., ..••■•■ - 2 1 0 - ...... '"••. •..•• .***•• *' '■*. ^ .1 2 2 DJAKARTA 0 (mb) -1 -2 ••,.^ , *'*'"*" .••"*■• — ■•••** — ^— 2 1 0 - *'"'*'. **••.. 2 2 1 - ...■• -••** ,., •-.. '**•*. -■•■^^^ 2 1 0 DARWIN 0 (mb) -1 -2 - '' -1 -2 2 SANTIAGO- - ..• .■•■■•• ••••■■••. ,, ....- •.. ,,•** EN(M)-^ 2 0 DARWM (mb) 2 EN(M) / EN(S)— ..••"'* ■■■ •...• EN(S)- .•••■" 2 1914 1 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 YEARS Figure 3e.— Triple 6-mo running mean plots of sea level atmospheric pressure anomalies (millibars) for Bombay, Madras, Djakarta, and Darwin ( 1914-29); also, triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago and Darwin ( 1914-29). El Nino type events (EN) are indicated in strong (S) or moderate (M) intensity. 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 2 - - 2 1 (mb) -1 •2 - •■..••■*'* •■' -2 2 MADRAS 0 (mb) -1 -2 - — •• - 2 0 - ..•**'**• ■•■■ -^ 2 2 — - 2 0 DJAKARTA 0 (mb) -1 -2 - ' *..^ — - •2 1 _ ••■—••■• - 2 0 DARWIN 0 (tub) -1 ■2 _ - -2 2 SANTIAGO - - .••■ ^ - 2 0 DARWIN (mb) ^ EN(M) EN(M) EN(S) -1 ■2 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 YEARS Figure 3f. — Triple 6-mo running mean plots of sea level atmospheric pressure anomalies (millibars) for Bombay, Madras, Djakarta, and Darwin ( 1930-45); also, triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago and Darwin < 1930-45). El Nino type events (EN) are indicated in strong (S) or moderate (M) intensity. 669 FISHERY BULLETIN: VOL. 76, NO. 3 1906 I9«7 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 z 1 BOMBAY 0 (mb) .1 -2 - ,.....:i^ - 2 1 0 -1 2 -■•••... "••■"■ *** 2 1 - - 2 1 MADRAS 0 (mb) .1 -2 _ - ■ 1 2 2 DJAKARTA 0 (mb) -1 -2 - - 2 - '^ .1 2 2 DARWIN 0 (mb) .1 2 ~ 2 - - -2 2 1 SANTIAGO - 0 DARWIN (mb) '^ - - 2 -■• • EN(M) EN(S)- : - -3 1946 ,947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 YEARS Figure 3g. — Triple 6-mo running mean plots of sea level atmospheric pressure anomalies (millibars) for Bombay, Madras, Djakarta, and Darwin ( 1946-61); also, triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago and Darwin (1946-61). El Nino type events (EN) are indicated in strong (S) or moderate (M) intensity. 1962 1963 1964 1965 1966 1967 1968 1969 1970 197! 1972 197 3 1974 1975 1976 1977 2 - 2 1 BOMBAY 0 ..* '"—'',— 0 -1 (mb) -1 ••*'* -2 - 2 2 1 - 2 1 MADRAS 0 (mb) .1 .,•* =^^ 0 _ '■—•■■ -2 - 2 2 - 2 DJAKARTA 0 (mb) -1 — 0 -1 ■■■ ■•••• •■•• ■2 - 2 2 - 2 1 DARWIN 0 (mb) -1 — .■■■■ •, .•■' _, '*. .* 0 \- *•..•*' *,. .-• ■2 - 2 2 1 - .._ .••• .••■■"• 2 1 SANTIAGO- 0 DARWIN ■■... .• • EN(S)- ■• ► (mb) - EN(M)' EN(M) 2 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 YEARS Figure 3h. — Triple 6-mo running mean plots of sea level atmospheric pressure anomalies (millibars) for Bombay, Madras, Djakarta, and Darwin (1962-76); also, triple 6-mo running mean plot of difference in atmospheric pressure anomalies between Santiago and Darwin (1962-76). El Niiio type events (EN) are indicated in strong (Sj or moderate (M) intensity. 670 (JUINN ET AI. : SOUTHERN OSCILLATION. KL NINO. AND INDONESIAN DROL'OHTS affected (e.g., the 1957-58 and 1972-73 cases of re- cent years). Moderate cases are recognized as El Ninos by most investigators, and display typical El Nino features to a lesser degree; maximum monthly sea-surface temperature anomalies along the coast usually peak in the 2.0°-3.5°C temperature range (e.g., the 1953, 1965-66, and 1976-77 cases of recent years). The effects of a moderate El Nino on the anchoveta fishery are considerable, but less serious than for the strong category. Weak events may or may not be recognized as El Nirios by investigators; maximum monthly sea- surface temperature anomalies along the coast usually peak in the 1.0°-2.5°C temperature range, but may appear relatively late in the year ( e.g., the 1951 and 1969 cases of recent years). Very weak events are not considered to be El Nirios; maximum sea-surface temperature anomalies, if they penetrate into the coast, are in the 0°-2°C range (e.g., the 1963 and 1975 events). The weak and very weak categories are included in this dis- cussion because the difference between weaker and stronger events depends not only on the height of the preevent index anomaly peak and the subsequent degree of relaxation reflected in the southeast trade strength, but also on the timing of this interannual relaxation. If the timing is in phase with the regular annual relaxation (South- ern Hemisphere summer and early fall), a moder- ate or strong event is likely to occur; if they are out of phase, a weak or very weak event is likely. Relaxation troughs that occur near the end of the year are usually associated with high Peruvian coastal sea temperature anomalies in the latter half of the year. The weak and very weak events may not be of significance to the Peruvian an- choveta fishery, but they do show up in the west- ern equatorial Pacific rainfall and their larger scale aspects may be significant from the standpoint of associated global fluctuations. Fig- ure 4 shows an example of how the recent events were reflected in the Tarawa rainfall. The weaker events were included as EN(W) in Figure 2a, b, since we have a fairly large amount of evidence available from 1950 on. They were not included in Figure 3a-h due to the decreasing availability of evidence as we reach further back in time. However, these weaker events, ascertain- ed to the best of our ability from availAle data 1946 19 J7 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1 1958 1959 i960 150 _ - 150 100 - .... .. .". .••.. • ' - 100 TARAWA RAINFALL 50 _ •, •.,.• •.. • - 50 ANOMALY (mm) '. , * ■. 6MRM TRIPLE PASS ° *....* ■. .■ . •....• ■•.. -50 - . * ...1 -50 -100 - *, .••"■•• ...•■ - -100 -ISO — •••■' - -150 3 3 2 — .••■ ■•. - 2 RAPA-DARWIN PRESSURE , .■ . -J ANOMALY (mb) . ...' '*. 6 MRM TRIPLE PASS ° •■ *. ....... -2 - ■•...••■ - -2 350 _ ••• - 550 300 _ , - 300 250 _ , - 250 200 _ . - 200 TARAWA RAINFALL ANOMALY (mm) 6 MRM TRIPLE PASS '°° - -■ ■-. . • . .•'- ISO 100 50 • ■". •' •. . ■ ■•. . ... • • ■■ ~ 50 - *. 0 -50 r"""'- ^ ; •. .... .• ■-. *. \^'' !• -SO -100 - ••■ ■•.•' ...1 Iff •^ - -too 2 • RAPA-DARWIN PRESSURE ~~ ■■■•••.. ..*'. .•"*. *. • .• ^ ANOMALY (mb) '~, T' • . ..... • 1 , 6 MRM TRIPLE PASS -1 - • ........ . • -2 - **...* •. .• "....■ 1961 1962 1963 1961 1965 1966 1967 1968 1969 19'0 1971 1972 1975 9^4 ■975 9'6 YEARS Figure 4. — Triple 6-mo running mean plot of anomalies of the difference in sea level atmospheric pressure (millibars) between Rapa ( 27°37 'S, 144 °20 ' W) ( Austral Is. ) and Darwin ( 1 2°26 'S, 130°52 'E), Australia compared with a similarly filtered plot of Tarawa ( 0r2 1 'N, 172°55'E) (Gilbert Is.) rainfall anomalies (millimeters). 671 FISHERY BULLETIN: VOL^ 76, NO. 3 and literature, are included in Table 1; and, the typical index and index component trends as- sociated with them can be noted in Figure 3a-h. Eguiguren's (1894) data were evaluated to ex- tend the record back even further. He classified rainfall at Piura, Peru (lat. 5°5'S, long. 80°38'W) into five categories: dry (0), light ( 1), moderate (2l, good (3l, and extra ordinary (4). Considering the distribution of events over the period 1891-1976 in relation to Eguiguren's rainfall category distribu- tion for 1791-1890, it appeared that we could re- late his category 4 to a strong El Nino, category 3 to a moderate El Nino, and categories 2 and 1 to weak and very weak events respectively. One must realize that for this Peruvian desert area long-term average rainfall values have little meaning, since averages combine data from the more-frequent drought years with data from the smaller number of event years when significant rainfall may occur. A year when an average amount of rain fell is likely to have been a year when an event occurred. Whereas the categories 3 and 4 rainfall situations were likely to have been associated with El Nino, there is no assurance that the categories 1 and 2 rainfalls were associated with the oceanic events. A study of presumed event occurrences (based on Eguiguren's information) in relatiion to trends of the Southern Oscillation indices and index components for a period when overlapping data records were available (1841-90) showed a high degree of compatibility. Table 1 lists years in ac- cordance with Eguiguren's rainfall classification as well as our interpretation of event intensity after considering his indications, the index and index component trends, and the various data and information sources listed early in this section. After 1790 and prior to 1841, when no pressure records were available for a cross-check, we avoid- ed the weak event category, but accepted Eguigu- ren's stronger categories 3 and 4 events. Events occurring prior to 1791, as reported by Frijlinck ( 1925), were considered to be of the strong variety. Sources for each event are listed in Table 1. STATISTICAL STUDY OF EL NINO TYPE EVENTS From a practical standpoint, we were concerned with the question of when the next event is likely to occur and what its intensity might be. There- fore, this study referred to the onset times for separate events and the interval between onset times. When the year immediately following the year of onset reflected an event of equal or lower intensity, it was assumed that the initial event extended into this next year or that the effects of the initial event held over into the early part of the next year; and, the whole situation was treated as a single event. However, when a weaker initial event preceded a stronger event in a following year they were treated as two separate events, Table l. — Year of onset of El Nino type events, 1726-1976, as classified according to event intensity by Eguiguren ( 1894), left, and the present authors, right (events below intensity 3 were not accepted prior to 1841 when pressure data became available). Numbers refer to event intensity: 1, very weak; 2, weak; 3, moderate; and 4, strong. Asterisks indicate onset of events considered separate (see text). Event Key to Event Key to Event Key to Event Key to Year intensity source Year intensity source Year intensity source Year intensity source •1726 (3) G •1852 2 (2) B, PC •1896 (3) PI. R •1943 (2) G. T, R •1728 (4) A, G 1854 2 B •1899 (4) D. PI, R 1944 (2) PI, T •1763 (4) A •1855 (2) PC 1900 (3) PI, R •1946 (1) PI, R ■1770 (4) A •1857 2 (2) B. PC •1902 (3) PI. R ■1948 (1) PI,T, R •1791 4 (4) A, B 1862 2 B •1905 (3) PI. R ■1951 (2) G, PI,T •1803 2 B •1864 4 (4) A. B, PI •1911 (4) F, E, PI •1S53 (3) G, E, T •1804 4 (4) B 1866 2 B 1912 (3) F. PI •1957 (4) G, L, PI •1814 4 (4) B ■1868 1 (3) B, C. PI •1914 (3) G, PI. R 1958 (4) G, L, PI •1817 3 (3) B •1871 4 (3) A, B, C ■1917 (2) H •1963 (1) PI, R •1819 3 (3) B •1873 (2) PI ■1918 (4) C, D. PI •1965 (3) M, PI, T •1821 3 (3) 8 •1875 1 (1) B, PC 1919 (3) D, PI, R •1969 (2) PI, T, R •1824 3 (3) B •1877 4 (4) A, B, PI •1923 (2) H. PI ■1972 (4) N, 0, PI •1828 4 (4) A. B 1878 4 (4) A, B. PI ■1925 (4) D. E. PI 1973 (4) N, 0, T 1829 1 B •1880 2 (3) B. PI 1926 (4) 1. PI. T ■1975 (1) P, PI, T •1832 3 (3) B •1884 4 (4) A, B, PI •1929 (3) G, 1. PI ■1976 (3) Q, PI,T •1837 3 (3) B 1885 (3) PI 1930 (3) G. PI. T •1844 3 (2) B, PC •1887 2 (3) B. PI ■1932 (2) E, 1, J •1845 4 (4) B. C. PC 1888 2 (3) . B, PI ■1939 (3) E, J, T 1846 2 (3) B. PC 1889 1 (1) B. PI 1940 (2) C, PI, T •1850 2 (2) B. PC •1891 (4) D. E, PI ■1941 (4) E, K, T Key Source Key Source Key Sour ce A Frijlinck (1925) H Lavalle (1917. 1924) O Caviedes (1975) B Eguiguren (1894) 1 SInepard (1930. 1933) P Wyrtki et al (1976) C Hutchinson (1950) J Mears (1944) Q Quinn (1 976) D Murphv (1923. 1926) K Lobell (1942) R Rainfall (equatorial and;or Peruvian) E Sears (1954) L Wooster (1960) T Sea-surfi ice temperature off Peru F Forbes (1914) M G jillen (1967) PC Pressure component of G Schweigger (1961) N Idyll (1973) Souttie rn Oscillation ir dex PI Souttiern Oscillation pressure index 672 gl'lNN KT AL SOUTHERN OSCILLATION, KL NINO, AND INDONESIAN DKOIIGHTS since an additional contribution was introduced in the following year. The foregoing assumptions were based on findings from a study of the South- ern Oscillation index trends and associated events over recent decades when more data were avail- able for case history studies. Our study of strong events was limited to the period 1763-present (Table 2, Figure 5), since the break of 35 yr between 1728 and 1763 was 14 yr longer than the longest subsequent break between events, and there was no way of eliminating the possibility that one or more strong events might have gone unreported over the 35-yr gap. These data indicate that given a strong El Niiio, there is a 3d7c probability of having another strong event in 7-8 yr, and an 82^7^ probability of having one within the next 15-16 yr. Considering all available data, the time between onsets of separate strong events was never <7 yr. For strong and moderate events (Table 3, Fig- ure 5) the record was limited to the period 1791-present when data for both categories were available. For strong, moderate, and weak events (Table 4, Figure 5) the record was limited to 1842-present, so we would have at least one index component trend available for cross-checking the less prominent weak events. (Madras pressure data became available in 1841.) With the addition of very weak events (Table 5, Figure 5), we limited our record to 1862-present in order to have an 0 50 3 o > Slrong Events .(1763-19721 Slrong , Moderate and Weak Events 0.50- 0 0 0 3 3 > I cvj I ^ T il ool ol *^ (NJ I ^ ! U3 fo >n r^ Strong and Moderate (1844-19? - IS 9^ ■ 4 4 ' ,. 'V C\J ^ U3 00 Events 0 25 - (1791-1976) (1864-1976) 18 0 50 It - 0 25 - J_ 4 4 0 4 ,;i- T^/il/u. 0 0 Jti,0 . . / CM 1 in CD oIojI^Iu)I(dIoIc\jI 7 T T T 7 *^ f^ 01 = Ki .n t cr> - OJ ro ID CD 1 in ^- Strong , Moderate, Weak and Very Weali Events CLASS LIMITS (years) Figure 5. — Histograms of frequency distributions for El Nino type events by intensity. Number of occurrences within class intervals is indicated. index trend available for cross-checking the more obscure very weak events. (The Santiago-Bombay index became available in 1861.) Cases were noted where relaxation from a large preevent index anomaly peak appeared to be a two or more stage process. [This type development was Table 2. — Strong El Ninos, with intervals between events from onset to onset. Years Years Onset Onset between Onset Onset between year year onsets year year onsets 1763 1770 1791 1804 1814 1828 1845 1864 1877 1770 7 1884 1891 7 1791 21 1891 1899 8 1804 13 1899 1911 12 1814 10 1911 1918 7 1828 14 1918 1925 7 1845 17 1925 1941 16 1864 19 1941 1957 16 1877 13 1957 1972 15 1884 7 209 (cumulative years between onsets) - 17 (number of intervals) = 12.3 yr, average time interval between onsets of strong El Ninos. Table 3. — Strong and moderate El Nines with intervals be- tween events from onset to onset. Years Years Onset Onset between Onset Onset between year year onsets year year onsets 1791 1804 13 1887 1891 4 1804 1814 10 1891 1896 5 1814 1817 3 1896 1899 3 1817 1819 2 1899 1902 3 1819 1821 2 1902 1905 3 1821 1824 3 1905 1911 6 1824 1828 4 1911 1914 3 1828 1832 4 1914 1918 4 1832 1837 5 1918 1925 7 1837 1845 8 1925 1929 4 1845 1864 19 1929 1939 10 1864 1868 4 1939 1941 2 1868 1871 3 1941 1953 12 1871 1877 6 1953 1957 4 1877 1880 3 1957 1965 8 1880 1884 4 1965 1972 7 1884 1887 3 1972 1976 4 185 (cumulat ve years between onsets) - 34 (nu Tiber of intervals) - 5.4 yr. average time interval between onsets. Table 4. — Strong, moderate, and weak El Ninos with intervals between events from onset to onset. Years Years Onset Onset between Onset Onset between year year onsets year year onsets 1844 1845 1850 1852 1855 1857 1864 1868 1871 1873 1877 1880 1884 1887 1891 1896 1899 1902 132 (cumulat average time 1845 1850 1852 1855 1857 1864 1868 1871 1873 1877 1880 1884 1887 1891 1896 1899 1902 1905 1 5 2 3 2 7 4 3 2 4 3 4 3 4 5 3 3 3 ive years between onsets) interval between onsets. 1905 1911 1914 1917 1918 1923 1925 1929 1932 1939 1941 1943 1951 1953 1957 1965 1969 1972 36 (number 1911 1914 1917 1918 1923 1925 1929 1932 1939 1941 1943 1951 1953 1957 1965 1969 1972 1976 of intervals) 6 3 3 1 5 2 4 3 7 2 2 8 2 4 8 4 3 4 3.7 yr. 673 FISHERY BULLETIN: VOL. 76, NO. 3 Table 5. — Strong, moderate, weak, and very weak El Ninos with intervals between events from onset to onset. Onset year Onsel year Years between onsets Onset year Onset year Years between onsets 1864 1868 1871 1873 1875 1877 1880 1884 1887 1891 1896 1899 1902 1905 1911 1914 1917 1918 112 (cumulati average time 1868 1871 1873 1875 1877 1880 1884 1887 1891 1896 1899 1902 1905 1911 1914 1917 1918 1923 1923 1925 2 1925 1929 4 1929 1932 3 1932 1939 7 1939 1941 2 1941 1943 2 1943 1946 3 1946 1948 2 1948 1951 3 1951 1953 2 1953 1957 4 1957 1963 6 1963 1965 2 1965 1969 4 1969 1972 3 1972 1975 3 1975 1976 1 ive years between onsets) nterval between onsets 35 (number of intervals) 3 2 yr. first mentioned in Quinn and Zopf (1975).] In some cases there was an initial fall from a large preevent (primary) peak which was not fully in phase with the seasonal relaxation (Southern Hemisphere summer and early fall) and the result was a relatively weak event; then, there was the rise to a smaller secondary peak followed by relax- ation to a secondary trough which was in phase with the seasonal relaxation and resulted in a stronger event. The length of time between the two troughs was generally 18-22 mo and it is our opinion that situations of this type may account for many of the event-to-event intervals that fall in the short 1-2 yr category. Examples of such developments can be noted in 1950-53, 1962-65, and 1973-76 (Figure 2a, b). Preevent peaks occur- red in 1950, 1962, and late 1973-early 1974. The first relaxation troughs following these peaks oc- curred in late 1951, late 1963, and late 1974-early 1975, and weak or very weak events resulted in all three cases. Then, there were rises to secondary peaks by mid-1952, mid-1964, and late 1975, fol- lowed by falls to troughs by early to mid-1953, mid- 1965, and mid-1976, resulting in moderate El Ninos for these latter years. We must be aware that these situations can arise and should be par- ticularly wary when a large preevent peak is fol- lowed prematurely by a weak or very weak event. (One must not lose sight of the fact that these interannual fluctuations in the index anomaly trends were used to represent the interannual fluctuations in southeast trade and equatorial easterly strength as affected by the Southern Os- cillation,) Figure 6 demonstrates the similarity of the three two-stage developments discussed above; a particularly obvious index trend was selected to represent each case. The index trend between late 1872 and 1877 indicates a possible three stage development (Fig- ure 3b), with a weak event in 1873, a very weak event in 1875, and a strong event in 1877 (Table 1 ), It is noteworthy that Indonesian droughts, which are usually associated with El Nino, occurred in 1873, 1875, and 1877 (Berlage 1957), The preevent index anomaly peak has been re- ported to be a reliable indicator for subsequent El Nino type activity, and our long index record sub- stantiates this viewpoint. We compiled statistics on the climb time from trough to peak and fall time from peak to trough from our long index anomaly record to provide some general guidance for event predictions. Figure 7 shows the applicable statis- tics. Events usually set in while the index is fall- ing and prior to the index trough inflection point. Therefore, the contents of Table 6 and Figure 8, which pertain to time between index peak and subsequent event onset, can be used to further refine event predictions. We assumed a March onset time for all cases in arriving at values in the column headed "Peak to event onset" (Table 6), This assumption was made since month of onset was not available for most of the early cases, and a study of recent cases showed onset times to range from January to May, INDONESIAN DROUGHTS What happens over Indonesia relates to the Southern Oscillation (Berlage 1957) and is, there- fore, an integral part of the activity affecting the E-D 1949 JF-D 1961 R-D 1973 3- 1950 1962 1974 1951 1963 1975 1952 1964 1976 1953 1965 1977 if) — in < CO Q. LlI _i Q. cr 2- - R-p - - ■■ / 1 : / 1 •■' / / \ *•' \ -^ V * ■ \ ^ f\ 7 T J E-D \ YEARS Figure 6. — Recent examples of two stage developments using triple 6-mo running mean plots of: 1) Easter-Darwin (E-D) index anomalies (1949-53); 2) Juan Fernandez-Darwin (JF-D) index anomalies (1961-65); 3) Rapa-Darwin (R-D) index anomalies (1973-77). 674 QUINN ET AL : SOUTHERN OSCILLATION, EL NINO. AND INDONESIAN DROUGHTS equatorial Pacific and the oceanic region off northwestern South America. In general, years when the index is low and El Nino type activity occurs are also years of drought in Indonesia (par- ticularly during the east monsoon season, May- October). Using sea salt production on the Island of Mad- ura (near Java), which is a very sensitive indicator of drought and precipitation, as well as some ad- ministration reports from Java estates, compiled by Van Bemmelen ( 1916), Berlage ( 1957) drew up a complete series of east monsoons drier than normal, from 1830 to 1953. Although 937( of the drought periods occurred during years when El Niiio type events were under way (Table 7), only 77*^ of the periods when El Nino type events were underway were also designated as periods of east monsoon drought (Table 8). Nevertheless, the as- sociation between occurrences of these two FIGURE 7. -Frequency distributions of rise time from trough to phenomena and changes in the Southern Oscilla- pealt and fall time from peak to trough lin months) for the triple i- • j i. j i u- -iu *- ^ . , r o .u /-> 11 ^- ■ J tion mdex trends are close enough in either case to 6-mo running mean plots of Southern Osculation index " anomalies (see text). The number of cases falling within a class indicate common relationships With the large- interval is entered at the top of the relevant histogram element, scale ocean-atmosphere changes over the Indo- 0 40 _ Pea* to Trough 13 (1862-19761 0 30 - 10 a 0 20 ^ 0 10 , } >. .-. 3- 0 ^0 (\j m ^1 o " '^ tr: 2) ;, A > 0 40 Trough to Peak o ^^ (IRfi4-iq7'i> ^0 30 3 020 — " *; 0 10 0 \— J Ifl t\i C las > L imi S ( mo iths ) Table 6.— Time (in months) between index peak and El Nino type event onset (assuming onset is in March of indicated year), and time between index peak and associated Indonesian drought onset (assuming onset is in May of associated year). Pressure indices (see text) used to determine time of preevent peak were: S-B, Santiago-Bombay; S-D, Santiago-Darwin; JF-D, Juan Fernandez- Darwin; E-D, Easter-Darwin; T-D, Totegegie-Darwin; and R-D, Rapa-Dai-win. In last column ND indicates no associated drought. El Nino type event Preevent index peak Peak to event onset No. of Monttis Peak to drougfit onset Year of onset Month of peak Index used No. of months 1864 Aug -Sept 1862 S-B 18.5 20.5 1868 Mar-Apr 1867 S-B 115 ND 1871 Jan -Feb 1870 S-B 13.5 ND 1873 Jan 1873 S-B 20 4.0 1875 Aug -Sept. 1874 S-B 6.5 8.5 1877 Feb. 1876 S-B 13.0 15.0 1880 Nov 1878 S-B 16.0 30.0 1884 May 1882 S-B 22.0 12.0 1887 July-Aug 1886 S-B 7.5 21.5 1891 Nov.-Dec,1889 S-B 15,5 17.5 1896 Oct 1893 S-D 29 0 31.0 1899 Nov, -Dec. 1897 S-D 155 ND 1902 May 1901 S-D 10.0 12.0 1905 Sept.-Oct 1903 S-D 145 16.5 1911 Jan -Feb. 1910 S-D 13 5 ND 1914 Sept -Oct 1912 JF-D 175 7.5 1917 Aug. -Sept 1916 S-B 6.5 ND 1918 Aug. -Sept. 1917 S-D 65 8.5 1923 Aug. 1921 JF-D 190 21.0 1925 May-June1924 S-B 95 11.5 1929 Dec -Jan 1928 29 JF-D 25 4.5 1932 Aug -Sept 1931 S-D 6.5 8.5 1939 June-July 1938 JF-D 8.5 22.5 1941 Jan -Feb. 1940 JF-D 13.5 15.5 1943 July-Aug 1942 JF-D 7,5 21.5 1946 Dec-Jan. 1944 45 S-B 14,5 4.5 1948 Aug. 1947 JF-D 7.0 ND 1951 Apr-May 1950 E-D 105 ND 1953 Apr -May 1952 R-D 105 12.5 1957 July-Aug 1955 E-D 195 t 1963 May 1962 JF-D 10.0 1965 June 1964 JF-D 9.0 Data not 1969 Mar -Apr. 1967 T-D 23.5 available 1972 Oct -Nov. 1970 R-D 16.5 1 1975 Nov -Dec. 1973 R-D 15.5 1976 Sept.-Oct. 1975 R-D 5.5 7.0 675 0 40 0 30 Peak to Event Onset 15 (1862 -1976 > o 1 ig I FISHERY BULLETIN: VOL. 76, NO. 3 Pacific region. Based on the association between El Nino type events, drought years and index fea- tures, and also an assumption that the drought will set in during May of involved drought years, we arrived at values in the column headed "Peak to drought onset" (Table 6). Figure 8 shows the resulting statistics which could be applied to In- donesian drought predictions. 040 ~ Peak to Drougtit Onset 0 30 (1862-1953 a 1976) f^J fO fO Figure 8. — Frequency distributions of time (in months) be- tween preevent peaks in triple 6-mo running mean plots of Southern Oscillation index anomalies (see text) and: 1) the onset of subsequent El Nino type events (assuming onset is in March of involved years); 2) the onset of associated Indonesian droughts (assuming onset is in May of involved years). The number of cases falling within a class interval is entered at the top of the relevant histogram event. Class Limits (months) Table 7. — Association of east monsoon droughts in Java with El Niiio type events. Drought years El Nino type event years Notes Drougtit years El Nino type event years Notes 1844 1844 1913» 1914( 1845 1845-46 1850 1850 1918»^ 1853 None Event in 1852 1919' 1855 1855 1923 1857 1857 1925\ 1926/ 1864 1864 1873 1873 1929 1875 1875 1932 1877 1877-78 1935 1881 1880 Index low 1880-81 1940 1883 1941 1884) 1884-85 1944 1885 1945* 1888 1887-89 1946/ 1891 1891 1953 1896 1896 Drougt 1902 1902 1976 1905 1905 28 (separate events) - 30 (east monsoon drought situations) = = 0,93, 1914 1918-19 1923 1925-26 1929-30 1932 None 1939-40 1941 1943-44 1946 1953 Slight lowering of index Drought data unavailable 1954-75 1976 93% of east monsoon droughts can be associated with El Nino type events Table 8.— Association of El Nino type events with east monsoon droughts in Java. El Nitio type event years Drought years Notes El Nino type event years Drought years 1844 1845-46 1850 1852 1855 1857 1864 1868 1871 1873 1875 1877-78 1880 1884-85 1887-89 1891 1896 1899-1900 1902 1844 1845 1850 None 1855 1857 1864 None None 1873 1875 1877 1881 1883-85 1888 1891 1896 None 1902 Drought in 1853 Index low 1880-81 1905 1905 1911-12 None 1914 1913-14 1917 None 1918-19 1918-19 1923 1923 1925-26 1925-26 1929-30 1929 1932 1932 1939-40 1940 1941 1941 1943-44 1944 1946 1945-46 1948 None 1951 None 1953 1953 Drought data unavailable 1954-75 1976 1976 28 (east monsoon drought situations) - 36 (separate events) '^ 0 78, 78°o of El Nirio type events can be associated with east monsoon droughts. 676 QUINN ET AL : SOUTHERN OSCILLATION. EL NINO, AND INDONESIAN DROUGHTS DISCUSSION Over the past 1 16 yi- ( 1861-1976), for which we have adequate data on the occurrence of El Nino type events of weaker intensity, there were de- cades of minimal activity (e.g.. 1901-10, 1931-40, 1961-70), but no decade without such activity. There is no reason to expect any significant change in the amount of El Nino activity in the foresee- able future over that experienced in the past cen- tury. Therefore, it would appear that our data (e.g.. Tables 2-5) might eventually be used in con- junction with associated catch data and biological findings for effective long-range planning in the management of the anchoveta fishery. For exam- ple, assessment of maximum sustainable yields under various environmental conditions ranging from the favorable extended anti-El Nino condi- tion (when there are two or more consecutive years with high Southern Oscillation indices) to the El Nino situation (when there are rapidly falling and low Southern Oscillation indices) might prove use- ful for determining the optimum size and flexibil- ity of the fishing fleet and fish processing facilities. A key element to such assessments will be a know- ledge of the required biological recuperation time following cessation of an unfavorable physical en- vironmental condition. Such data could also be used for speculative long-range outlooks. For example, if we had just experienced a strong El Nino, our results suggest that there is a near-zero probability that we would experience another strong event in < 7 yr after the onset of the recent situation. However, there would be an 86Vir probability that an event in the very weak, weak, or moderate category would occur within 3-4 yr after the strong El Nino onset. Considering the current situation, and recogniz- ing that a moderate event set in during 1976 and held over into early 1977, there is a dAVc probabil- ity (based on our data) that another event of un- known intensity would set in during 1980. It would not be reasonable to go beyond statistical estimates until we find we are approaching a peak in the Southern Oscillation index anomalies. When we are nearing a preevent peak and can assess its height and time of occurrence, then we can use peak to trough statistics (e.g., Figure 7) to advantage in forecasting onset time and likely intensity of the coming El Nino type event. The intensity would be based on the height of the index anomaly peak and the time of year when the sub- sequent trough was expected to occur. Event onset time can be further refined by considering Figure 8 statistics. It is also essential in the prediction procedure to realize that some developments may involve two or more stages. In cases of this type, forecast lead times for the separate stages will often be greatly reduced (to 1-6 mo in advance), unless historical analogies lead to pattern recog- nition as the situation evolves. ACKNOWLEDGMENTS We thank the Chief of the Naval Weather Ser- vice and the Director of the Hydrographic Insti- tute of the Armada de Chile; the Director of the Civil Aviation Service and Chief of the Meteorological Service of Polynesie, Francaise; the Director of the Meteorological and Geophysi- cal Institute, Djakarta, Indonesia; the President of the Instituto del Mar del Peru; Ramon Mugica, Cniversidad de Piura, Peru; the Director of the Australian Bureau of Meteorology; and, the Na- tional Climatic Center, Environmental Data Ser- vice, NOAA, for their support of this study. We are indebted to Forrest R. Miller of the Inter- American Tropical Tuna Commission and Richard Evans of the Southwest Fisheries Center, Na- tional Marine Fisheries Service, NOAA, for their timely information on sea temperatures and weather conditions over the eastern tropical Pacific. We also thank Clayton Creech for his sup- port in data processing. Support by the National Science Foundation under the North Pacific Ex- periment of the International Decade of Ocean Exploration through NSF Grant No. OCE 75- 21907 AOl, and under the Climate Dynamics Program of the Division of Atmospheric Sciences through NSF Grant No. ATM77-00870 is grate- fully acknowledged. LITERATURE CITED Berlage, H. p. 1957. Fluctuations of the general atmospheric circulation of more than one year, their nature and prognostic val- ue. K. Ned. Meteorol. Inst., Meded. Verh. 69, 152 p. 1966. The Southern Oscillation and world weather. K. Ned. Meteorol. Inst., Meded. Verh. 88, 152 p. BJERKNES. J. 1969. Atmospheric teleconnections from the equatorial Pacific. Men. Weather Rev. 97:163-172. Caviedes. C. N. 1975. El Nino 1972: Its climatic, ecological, human, and economic implications. Geogr. Rev. 65:493-509. Clayton. H. H. 1927. World weather records. Smithson. Inst. Misc. Col- lect. 79, 1199 p. 677 FISHERY BULLETIN: VOL 76. NO. 3 1934. World weather records. 1921-1930. Smithson. Inst. Misc. Collect. 90, 616 p. Ci-.-xYToN. H. H.. .•wn F. L. Cl.wton 1947. World weather records, 1931-1940. Smithson. Inst. Misc. Collect. 105, 646 p. ECUKU'REN. D. V. 1894. Las Nuvias de Piura. Bol. Soc. Geogr. Lima 4(7- 9):241-258. FoKBE.S. H. O. 1914. Notes on Molina's pe\ican< Pelecanus thagus). Ibis, Ser. 10, 2:403-420. FRI.ILI.NCK. C. p. M. 1925. Bijdrage tot het probleem der Klimatwis- selingen. Nat. Utr. 45:372-374. Guillen. O. 1967. Anomalies in the waters off the Peruvian coast dur- ing March and April 1965. Stud. Trop. Oceanogr. (Miami) 5:452-465. HUTCHINSON. G. E. 1950. Survey of existing knowledge of biogeochemistry 3. The biogeochemistry of vertebrate excretion. Bull Am. Mus. Nat. Hist. 96, 554 p. Idyll, C. P. 1973. The anchovy crisis. Sci. Am. 228(6):22-29. Johnson, J. H„ and G. R. Seckel 1977. Use of marine meteorological observations in fishery research and management. U.S. Dep. Commer., NOAA, Environ. Data Serv., p. 3-12. L.av.alle. Y Garcia. J. A., De. 1917. Informe preliminar sobre la causa de la mortalidad anormal de las aves ocurida en el mes de marzo del pre- sente ano. Mem. Cia. Adm. Guano, Lima 8. 1924. Estudio de la emegracion y mortalidad de las aves guaneras. Mem. Cia. Adm. Guano, Lima 15. LOBELL, M. G. 1942. Some observations on the Peruvian coastal cur- rent. Trans. Am. Geophys. Union 23:332-336. MEARS. E. G. 1944. The ocean current called "The Child." Annu. Rep. Smithson. Inst., 1943, p. 245-251. MILLER, F. R., AND R. M. LAURS. 1975. The El Nino of 1972-73 in the eastern tropical Pacific Ocean. Inter-Am. Trop. Tuna Comm. Bull. 16:403-448. Murphy, r. C. 1923. The oceanography of the Peruvian littoral with ref- erence to the abundance and distribution of marine life. Geogr. Rev. 13:64-85. 1926. Oceanic and climatic phenomena along the west coast of South America during 1925. Geogr. Rev. 16:26- 54. National Marine Fisheries Service. 1977. Industrial fishery products, market review and out- look, June 1977. U.S. Dep. Commer.. NOAA, Natl. Mar. Fish. Serv., Curr. Econ. Anal. 1-29, 27 p. Panofsky. H. a., and G. W. Brier. 1965. Some applications of statistics to meteorology. Pa. State Univ. Press, University Park, 224 p. QUINN, W. H. 1971. Late Quaternary meteorological and oceanographic developments in the equatorial Pacific. Nature (Lond.) 229:330-331. 1974. Monitoring and predicting El Nifio invasions. J. Appl. Meteorol. 13:825-830. 1976. Use of Southern Oscillation indices to assess the physical environment of certain tropical Pacific fisheries. In Proceedings of the NMFS/EDS Workshop on Climate and Fisheries, Columbia, Mo., April 26-29, 1976, p. 50- 70. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv./ Environ. Data Serv. QUlNN, W. H., AND D. O. ZOPF 1975. The Southern Oscillation, equatorial Pacific anomalies and El Nino. Geofis. Int. 15:327-353. SCHWEICCER, E. H, 1961. Temperature anomalies in the eastern Pacific and their forecasting. Soc. Geogr. Lima. Bol. 78:3-50. Sears. M. 1954. Notes on the Peruvian coastal current. 1. An intro- duction to the ecology of Pisco Bay. Deep-Sea Res. 1:141-169. SHEPARD. G. 1930. Notes on the climate and physiography of south- western Ecuador. Geogr. Rev. 20:445-453. 1933. The rainy season of 1932 in southwestern Ecuador. Geogr. Rev. 23:210-216. Troup, A. J. 1965. The 'southern oscillation." Q. J. R. Meteor. Soc. 91:490-506. U.S. DEPARTMENT OF COMMERCE. 1959. World weather records, 1941-1950. Wash., D.C., 1361 p. 1961-76. Monthly climatic data for the world. Vol. 14- 29. Wash., D.C., var. pag. 1966-68. World weather records, 1951-1960, Vol. 3, 4, and 6. Wash., D.C., var. pag. Van Bemmelen. W. 1916. Droogte-jaren op Java. Nat. Tijdschr. Ned. -Indie 75:157. Walker. G. T. 1924. World Weather II. Mem. India Meteor. Dep. 24:275-332. WoOSTER, W. S. 1960. El Nino. Calif Coop. Oceanic Fish. Invest. Rep. 7:43-45. Wyrtki. K., E. Stroup. W. Patzert, R. Williams, and W. QUINN. 1976. Predicting and observing El Nino. Science (Wash., D.C.) 191:343-346. 678 THE PRECISION OF SIMULATED TRANSECT SURVEYS OF NORTHERN ANCHOVY, ENGRAULIS MORDAX, SCHOOL GROUPS Paul C. Fiedler' ABSTRACT Simulated transect surveys of model anchovy populations were compared in terms of precision and efficiency. The precision of systematic surveys varies inversely with the distance between transects. Systematic surveys give more precise population estimates than random surveys, due to the large positive correlation between closely spaced transects. The precision of stratified systematic surveys is not significantly different from that of the unstratified surveys when the school groups are randomly distributed in the survey area. However, stratified systematic surveys are more precise when the school groups are clumped in one end of the survey area. The results of the simulations show that the patchy distribution of anchovy schools can be a major source of error in population estimates. Any sampling program intended to estimate the size of a population is subject to a variety of errors which may reduce the accuracy or precision of the estimate. Precision is the reciprocal of the varia- tion of replicate estimates. Successful manage- ment of a northern anchovy fishery in California will require the monitoring of changes in the popu- lation size. Acoustic survey techniques are cur- rently being developed to obtain population and biomass estimates independent of the fishery (Hewitt et al. 1976). As in the study of any biologi- cal population, it will be important to avoid con- fusing the variation of a series of estimates due to sampling error with true fluctuations in the popu- lation size. Precision may be affected by 1) the manner in which the sampled population is distributed in space, and 2) variations within the sampling method itself Several studies have shown that the patchy distribution of individuals in a population may cause considerable variation in replicate population estimates and that the variation is re- lated to sample design. Winsor and Clarke ( 1940) studied the variation of catches in series of plankton net tows. Although they did not separate the components of between-tow variation due to factors (1) and (2), it was observed that oblique tows were more precise than vertical or horizontal tows. Barnes and Marshall (1951) took an exten- sive series of replicate pump samples and attri- buted the considerable variation observed to the ■Scripps Institution of Oceanography, Universitv of Califor- nia, San Diego, La Jolla, CA 92093. nonrandom distribution of the zooplankton since the volumes filtered were known accurately. Taft ( 1960) analyzed the variance of sardine egg counts in a grid of closely spaced stations. The distribu- tion of eggs was extremely patchy (the densities between samples ranged over more than four or- ders of magnitude) and the relative 95^}^ con- fidence limits for an estimate of the egg population in the area of the grid from a single sample were represented by a factor of 62. A simulation study by Wiebe (1971) showed that the precision of zoo- plankton population estimates depends both on the sampling design (net size and tow length) and the distribution of the population (size and loca- tion of patches). Similar studies have investigated the precision of sampling fish populations. Taylor ( 1953) discuss- ed the implications of the patchy distribution of fish for the optimum design of trawl surveys to estimate population size. Cram and Hampton ( 1976) demonstrated that the patchy distribution of pilchard schools can cause imprecision suf- ficient to render a population estimate useless for management. The anchovy population is patchy on two levels: individual fish are aggregated in schools and schools themselves tend to be aggregated in school groups. This patchiness, or nonrandomness, is ex- pected to be a major source of variation in popula- tion estimates. The present study simulated sur- veys of model anchovy populations to determine the effect of patchiness on the precision of popula- tion estimates. Three transect survey designs were compared: systematic, random, and strat- ified systematic. These are merely different Manu.-icnpt accepted .January 1978. FISHERY BULLETIN: VOL. 76. NO. 3. 1978. 679 FISHERY BULLETIN: VOL. 76. NO. 3 methods of selecting transects, or allocating sam- pling effort. The three types of simulated survey, with a range of sample sizes (numbers of tran- sects), were run on 15 model anchovy populations. METHODS Anchovy populations were modeled as arrays with each element representing 1 n.mi.-. The ari-ay dimensions were 180 x 75, approximately the dimensions, in miles, of the Los Angeles Bight. Since a school is the population unit detected in an acoustic or aerial survey, the units of the model populations were schools. One hundred fifty thousand schools were distributed in the array resulting in a mean density of 11.1 schools mi^. Four acoustic surveys by the California Depart- ment of Fish and Game- in 1975 and 1976 yielded estimates ranging from 88,887 to 319,878 an- chovy schools off southern California in the area of the bight. Mais (1974) gave a range of 21,920- 343,070 ix = 150,996) schools off southern California and northern Baja California, most of which were within the bight. The schools were placed in circular school groups located at random. Schools were distrib- uted uniformly within a school group. School group radii and densities were chosen randomly and independently from log-normal approxima- tions of observed frequency distributions based on 52 school groups from six California Department of Fish and Game Sea Survey acoustic surveys (MacCall et al.-M (Figure 1). There was no sig- nificant correlation between the density of schools within a school group and the size of the school group in these observations. Where school groups overlapped, the densities were simply added to- gether, although this effectively increased both the mean radius and density. In one model popula- tion illustrated in Figure 2, 16 school groups con- taining 150,303 schools covered about 149^ of the survey area. Fifteen model populations were used, each with the same total number of schools, but different locations, sizes, and densities of school groups. ^S.J.Crooke. 1975. Cruise reports 75- A- 1 and 75- A-6. K. F. Mais. 1976. Cruise reports 76-A-3 and 76-A-9. State of California - The Resources Agency, Department of Fish and Game, Marine Resources Region, Long Beach, CA 90802. ■■'MacCall, A., P. E. Smith, G. Stauffer, J. Squire, J. Zweifel, and S. Crooke. Report of CalCOFI anchovy workshop working group on methods of estimating anchovy abundance. Unpubl. manuscr. Southwest Fisheries Center, National Marine Fisheries Service, NOAA, La Jolla, CA 92038. >- o 5 030 o Ul tr - / / / / ^ 020 / x > / \ 5 QIO UJ q: ■>.. . - ^ ^ 0 "* T r 1 1 1 2.2 2.6 3.0 3.4 38 4.2 4.6 5.0 54 Lege DENSITY (schools /miles^) 0.40 V o / \ S O30 3 - / / / / \ \ \ o Ul tr \ \ ^ Q20 - / / \ Ul > . / \ \ h- / j 0.10 Ul cc - / ' \ / \ / \ N ^ 0 r I 1 1 I I I -2.0 -1.0 0 1.0 2.0 3.C 4.0 5.0 Log e RADIUS (miles) Figure l. — Comparison of distributions of northern anchovy school group density and size observed in the California Current during California Department Fish and Game surveys (solid lines) to log-normal approximations used in simulations (dashed lines). A simulated survey consisted of a series of transects across the survey area. There were 180 possible transects, each 1 mi wide. Acoustic sur- veys currently run by the Southwest Fisheries Center, National Marine Fisheries Service, NOAA, used a transect width of 0.14 mi (250 m). Aerial transect widths were typically 0.2 to 0.5 mi. A larger transect width was used in the simula- tions to hold the model population array down to a reasonable size. We assumed that the general re- sults of the simulations would not change by using a smaller transect width. Since all schools were coimted within a transect, the only source of error in the survey estimate was the large variance in the number of schools per transect. For instance, in the model population in Figure 2, the mean number of schools per transect was 835.0, while SD was 920.4 (variance = 8.47 xlO^). Systematic surveys were simulated by counting the schools within a series of transects separated by a constant transect interval. A population es- timate was calculated simply by dividing the survey count by the fraction of the survey area covered by the transects. Transect intervals of 2, 3, 680 FIEDLER f'KECISION OF SIMl'LATEn TRANSECT Sl'RVFYS N 180 Figure 2. — A model northem'anchovy population. Densities of school groups in schools per square mile. Simulated survey transects are oriented horizontally. The numbers on the axes are the coordinates of the array and the dimensions, in miles, of the survey area it represents. 4. 5. 7, 10, 12, 15, 20, 25, 30, and 40 mi were used. A survey with a transect interval of d miles con- sisted of 180 c/ transects. For each transect inter- val, 20 replicate surveys were run by randomly choosing the initial transect from the first d tran- sects in the survey area. The replicate survey esti- mates were used to calculate an unbiased mean population estimate and a coefficient of variation (standard deviation of the replicate estimates di- vided by the mean), which is a measure of the precision of the estimate (Wiebe 1971). This pro- cedure was repeated on the 15 different model populations. Random surveys were simluated in an analo- gous manner to allow a direct comparison of sam- pling errors. For each of the transect intervals (c/i of the systematic surveys, 20 replicate surveys were run consisting of 180 (/ transects chosen at random without replacement la transect was not repeated within a survey). Coefficients of varia- tion were calculated as a measure of precision. Stratified systematic surveys were simulated after dividing each of the model populations into four 45-mi wide strata. The schools along three transects in each of the four strata were counted to obtain a preliminary estimate of relative popula- tion sizes. Then a total of 60, 36, 18, 12, 9, 7, or 6 transects were divided among the strata according to the estimated population fractions. For exam- ple, if a stratum contained one-half of the schools counted in the preliminary survey, one-half of the total number of transects was allocated to that stratum for the stratified survey. At least one transect was allocated to each stratum to avoid biasing the final population estimate. For each total number of transects, 20 replicate systematic surveys were run by randomly choosing the initial transect and simulating a systematic survey within each stratum with the allocated number of transects. Once again, coefficients of variation of the replicate population estimates were calculated for each of the 15 model populations. RESULTS The results of the systematic survey simula- tions indicate that the sampling error represented e 5 F, IT o q: 09 000 O ^ CO CO o o o S 9 i" o f^ — iC o o o O O o 8° c\j o 1 I 11 — 1 1 — I i I I 1^ 2 34 5 7 10 12 15 20 25 30 TRANSECT INTERVAL-MILES 40 Figure 3. — Results of the simulations of systematic surveys of the 15 models for northern anchovy populations. Relative ef- ficiency is proportional to precision (the reciprocal of the coefficient of variation) divided by relative cost (see text). Aver- ages and 95% confidence limits are given for coefficients of varia- tion. 681 FISHERY BULLETIN; VOL. 76. NO. 3 by the coefficient of variation increased as the transect interval increased and sample size de- creased (Figure 3). The cost of a survey was assum- ed to be proportional to the distance covered along the transects plus 360 mi to and from port. Relative efficiency was 10-' times the reciprocal of the pro- duct of the coefficient of variation (C.V.) and rela- tive cost, i.e., precision divided by cost. Efficiency generally decreased as the transect interval in- creased, but peak efficiency was obtained at a transect interval of 3 mi. By interpolation, it can be seen that a population estimate may range 10 and 2U7c (2 x C.V.) from the true population size when surveys are run with transect intervals of 8.5 and 16 mi. respectively. Systematic sampling gave a consistently lower coefficient of variation, or greater precision than random sampling (Figure 4). The variability be- tween model populations, indicated by the confidence limits on the mean coefficient of varia- tion, was greater for the random sampling error (F,,^i^, ^5.44,P<0.05 12)for8ofthe 12 sample sizes. Also represented in Figure 4 are the ex- pected coefficients of variation for random sampl- ing calculated from the model population parameters (o-^ and yu.) by the following equation with a finite population correction: C.V. = \k- (^^) [i^ n \ N / where a- - the average variance of the number of schools per transect in the 15 model populations =1,154,636 At = the mean number of schools per tran- sect = 835.3 n = the number of transects in the survey N = total number of transects in the survey area = 180. 0.5C 0.20 STRATIFIED SYSTEMATIC 1 I M [— 2 34 5 7 90 45 26 60 36 10 12 18 15 20 9 25 30 6 — I 1 — TronseCt 40 Interval-miles 5 Number of Transects Figure 4. — Comparison of the results of the simulations of random, systematic, and stratified systematic surveys. For ran- dom surveys, the curve represents the expected coefficients of variation calculated from the parametric variance of the model populations (see text). For 1 1 of the 12 sample sizes, the expected value was within 95''^ confidence limits of the mean ob- served coefficient of variation. This close agree- ment supports the validity of the method used to obtain the coefficients of variation in the simula- tions. There was apparently no significant differ- ence between the coefficients of variation of the systematic and stratified systematic surveys (Figure 4). This was confirmed by analysis of var- iance (Table 1,P>0.25 that there was no added variance due to survey design). However, there were significant interaction effects between sur- vey design and model population (P<0.01 that there was no added variance from this source) and between survey design and the number of trans- ects (P<0.05). An attempt was made to elucidate the interac- tions involving survey design by performing analyses of variance on subsets of the data. It was found that for large sample sizes (transect interval ^15 mi, or number of transects ^12) unstratified Table l. — Analysis of variance of the coefficients of variation from simulated systematic and stratified systematic surveys of model northern anchovy populations. This is a mixed model analysis of variance (Sokal and Rohlf 1969): survey design and number of transects are fixed treatment effects and model population is a random effect. Source of variation SS df MS Significance level Ivlain effects; Sun/ey design Number of transects Ivlodel population interactions; Design-number of transects Design-population Number of transects-population Error Total 0.0157 1 00157 1 2872 6 0,2145 05270 14 00376 0 0748 6 00125 02418 14 00173 03179 84 00038 0.3988 84 0.0047 2.8633 209 (6 841 = ' M4S4) , = 0.908. P -0 25 56.681. P<<0.001 = 8 000. P- 0.001 F (6,MI F 114 84) '164 84 = 2.627. P = 3.681. P-: = 0 819. P ;0-05 0,001 ■0 50 682 FIEDLER: F^RECISION OF SIMll.ATED TKANSECT SIRVEYS systematic surveys were significantly more pre- cise than stratified systematic surveys iP<0.025). although there is still a significant interaction between survey design and model population (P<0.001). For smaller sample sizes, there was no significant difference between the precision of the two designs. In the model populations, school groups were located randomly within the survey area. How- ever, the distribution of schools between strata was never random because of the wide range of school group sizes and the small number of school groups in a population. The 15 model populations were divided into three groups (low, intermediate, and high nonrandomness) based on the index of dispersion of the number of schools per stratum. Analysis of variance revealed that for highly non- random populations, stratified systematic surveys were significantly more precise than unstratified surveys (P<0.025). On the other hand, there was no significant difference between the survey de- signs for populations of intermediate or low non- randomness. The effect of the nonrandomness of the populations, in the limited sense used here, is illustrated more dramatically below. In summary, these results indicate that both the number of transects and the spatial distribution of the population can affect the precision of a survey estimate. The effect of survey design involves complex interactions with the other two factors. These factors should be considered, if possible, when choosing the optimum design for a survey. DISCUSSION In general, systematic sampling may result in considerable gains or losses in precision compared with simple random sampling. The greatest in- crease in precision occurs when there is a high degree of correlation between adjacent sampling units and the correlation decreases as the interval between units increases. In this situation, sys- tematic sampling resembles stratified sampling. On the other hand, precision may be greatly re- duced when there is a periodic variation in the population and the sampling interval is equal to this period or a multiple of it (Hansen et al. 1953). Correlograms between sampling units (tran- sects) in five of the model anchovy populations indi- cated that transects <10 mi apart had a high posi- tive correlation, while the correlation tended to be slightly negative at distances >20 mi (Figure 5). This autocorrelation structure was due to the fre- FlGURE 5. — Autocorrelation of transect counts in five mode! northern anchovy populations. quency distribution of school group sizes. The mean distance at which the autocorrelation func- tion passed through zero was 15.0 mi, while the mean diameter of the individual school groups in the five model populations was 11.8 mi. Distribu- tion of school groups within the model populations was random. However, real populations are likely to be nonrandom in this respect and additional correlations would be expected from this factor. The strong positive correlation between transects separated by short distances explains why sys- tematic surveys with small transect intervals were more precise than random surveys with an equivalent number of transects. As the transect interval increased, the correlation between tran- sects decreased to near zero and the imprecision of systematic sampling approached that of random sampling (Figure 4). In order to reduce total sampling error, a com- mon strategy is to allocate effort proportional to the sampling error within parts of a sampling program. The variation observed in the population estimates of the simulated surveys was caused by the large variance in the number of schools per transect. It can be shown in the model populations, as in many biological populations, that the stan- dard deviation was positively correlated with the mean number of schools per transect in a stratum. Therefore, it was thought that the stratified sys- tematic surveys would reduce the total sampling error by allocating more transects where the var- iance was large. The simulations failed to show any gains in precision from this strategy. This result was not expected, but is possibly due to the random distribution of school groups. The model populations may have been ideal in this sense, but we had relatively little information on the dis- tribution of school groups within the range of the 683 FISHKRY Hl'ia.KTIN: Vol, 7(S, NO. 3 northern anchovy. As slated above, stratified sys- tematic sui'veys were significantly more precise than unstratified surveys for the five model popu- lations with the most nonrandom distribution of schools between strata. If the school groups them- selves are aggregated, it is reasonable to expect an increase in precision by stratifying the survey. To test this possibility, the simulations were repeated on model populations in which the school groups were limited to only one-half of the survey area. An analysis of variance (Table 2) indicated in this case that the stratified systematic surveys were more precise than the unstratified systema- tic surveys (P<0.005 that there was no added var- iance due to survey design). The overall mean coefficients of variation were 0.095 and 0.133, re- spectively. However, there were significant in- teraction effects involving survey design, indicat- ing that the advantage of stratifying the survey will depend on the number of transects and the spatial distribution of the population. The addi- tional cost of the preliminary survey in the strat- ified design must also be considered when compar- ing it with the unstratified design. The results of the simulated systematic surveys showed that the patchy distribution of schools was an important source of error in estimates of the anchovy population size. Acoustic surveys run by the Southwest Fisheries Center have used tran- sect intervals of 6.6 and 40 mi. The simulations gave evidence that the population estimates from these surveys could be expected to range at least 8 and 90''^ (2 x C.V.), respectively, from the true population size. The most efficient simulated sam- pling, in terms of precision per unit cost, occurred at a transect interval of 3 mi. This would require a cruise grid of 4,860 mi, equivalent to a 34-day acoustic survey at 12 kn and 12 h per day, to reduce the coefficient of variation (due to the patchy distribution of schools) to 1.4^7^ . Maximiz- ing efficiency is not a valid goal, however, when the precision gained is greater than that required for the problem of managing the fishery, when other sources of error become more important, and when there are absolute limits on cost. Anchovy population estimates within 1W( of the true value might be considered sufficient for management, at least to allow confidence that a consistent change observed over several years is real (pers. commun., P. E. Smith, Southwest Fisheries Center, National Marine Fisheries Service, NOAA, La Jolla, Calif., Oct. 1977). As stated before, the anchovy population is patchy on two levels: individuals are aggregated into schools and schools are aggregated into school groups. The simulations have quantified the sam- pling error due to the second level of patchiness only. Although little is known about the distribu- tion of anchovy school groups, it was also de- monstrated that their aggregation is potentially an important consideration in designing a survey. The acoustic survey methods currently used by the National Marine Fisheries Service and the California Department of Fish and Game do little more than count the number of anchovy schools (Hewitt et al. 1976; Mais 1974). The Department of Fish and Game calculates a biomass estimate by multiplying the observed school area by a constant factor thought to represent an average biomass per unit area. More sophisticated methods of es- timating biomass from the acoustic signal re- ceived from a school are now being explored at the Southwest Fisheries Center. For these reasons, the problem of sampling error due to a varying number offish per school (the first level of patchi- ness) was not addressed here. Many sources of error may be involved in an anchovy biomass estimate. Patchiness is impor- tant in any type of sampling program. Other sources of error that may be important in an Table 2. — Analysis of variance of the coefficients of variation from simulated systematic and stratified systematic surveys of model northern anchovy populations when the model popula- tion school groups are clumped in one-half of the survey area. Source of variation SS df MS Significance level Main effects Survey design 0 0732 1 0 0732 F ,,,„ = 12.908, P<0.005 Number of transects 1 4995 6 02499 f ,6641 = 119.403, P<<0.001 Model population 0 1327 14 00095 '^,,4.64, = ''318, p. 0 001 Interactions Design-number of transects 0 0476 6 00079 F,684, = 3,589, P<0,005 Design-population 0 0794 14 00057 Fi,4„, = 2,591, P< 0.01 Number of transects-population 0.1758 84 0,0021 F,e4e4, = 0,955, P 0 50 Error 0 1858 84 00022 Total 2.1941 209 684 FIEDLER: PRECISION OF SIMULATED TRANSECT SURVEYS acoustic survey are as follows (Instituto del Mar del Peru 1974; Cram and Hampton 1976; P. E. Smith pers. commun.): 1) Failure to discriminate between anchovy schools and other acoustic targets. 2) Unschooled fish and small schools not detected. 3) Vessel avoidance. 4) Inability to survey in shallow inshore waters. 5) Movement of school groups relative to the sur- vey grid. 6) Fish in the top surface layer missed by the acoustic beam. 7 ) Errors in the factor for conversion of the acous- tic signal information to a biomass estimate. 8) Effect of varying hydrographic conditions on the acoustic signal. 9) Blocking of signal to and from fish far from the ship by fish nearer to the ship. The magnitude of the error caused by these fac- tors can now only be roughly estimated. They may affect either or both the precision and accuracy of a population estimate. Corrections to reduce the biases are conceivable. The present study has de- monstrated the magnitude of the error associated with the patchiness of the anchovy population. Although the model population distributions may be crude approximations to the real distribution, the general conclusions reached here are not likely to be changed by adding further levels of complexity to the model. The sampling error due to patchiness can be reduced by properly designing a survey, but never eliminated. Temporal and spa- tial differences in population estimates must be interpreted with an awareness that the error exists. ACKNOWLEDGMENTS I thank Paul E. Smith for his guidance in com- pleting this project, as well as Alec McCall, Keith Parker, Jim Zweifel, and Roger Hewitt for valuable discussions. Ian Hampton and two anonymous reviewers provided helpful comments on earlier drafts of the manuscript. This work was carried out under a State of California Marine Research Corpmittee contract at the Southwest Fisheries Center, La Jolla, Calif. LITERATURE CITED Barnes, H., and S. M. Marshall 1951. On the variability of replicate plankton samples and some applications of "contagious" series to the statistical distribution of catches over restricted periods. J. Mar. Biol. Assoc. U.K. 30:233-263. CRAM, D. L., AND I. Hampton 1976. A proposed aerial/acoustic stragety for pelagic fish stock assessment. J. Cons. 37:91-97. Hansen, M. H., W. N. Hurwitz, and W. G. Madow. 1953. Sample survey methods and theory. Vol. 1. John Wiley and Sons, N.Y., 638 p. Hewitt, R. P., P. E. Smith, and J. C. Brown. 1976. Development and use of sonar mapping for pelagic stock assessment in the California Current area. Fish. Bull., U.S. 74:281-300. Instituto del Mar del Peru 1974. Report of the fourth session of the panel of experts on stock assessment on Peruvian anchoveta. Bol. Inst. Mar Peru (Callao) 2:605-723. Mais. K. F. 1974. Pelagic fish surveys in the California Cur- rent. Calif Dep. Fish Game, Fish Bull. 162, 79 p. SOKAL. R. R., AND F. J. ROHLF 1969. Biometry. W.H. Freeman and Co., San Franc, 776 p. TAFT B. a. 1960. A statistical study of the estimation of abundance of sardine (Sardinops caerulea) eggs. Limnol. Oceanogr. 5:245-264. TAYLOR, C. C. 1953. Nature of variability in trawl catches. U.S. Fish Wildl. Serv., Fish. Bull. 54:145-166. WIEBE, P. H. 1971. A computer model study of zooplankton patchiness and its effects on sampling error. Limnol. Oceanogr. 16:29-38. WiNSOR, C. P., AND G. C. Clarke 1940. A statistical study of variation in the catch of plankton nets. J. Mar. Res. 3:1-34. 685 NOTES INTERSEX ANOMALIES IN SHRIMP OF THE GENUS PliSAHOPSIS (CRUSTACEA: PENAEIDAE) While examining a relatively large collection of Penaeopsis (40 lots containing 196 specimens) taken by the U.S. steamer Albatross during the Philippine Expedition. 1907-10, I found three specimens having external characteristics of both males and females. Each specimen had a fully developed thelycum, a moderately well-developed petasma (about two-thirds the length of the petasma of males of corresponding sizei, small ap- pendices masculinae. and genital apertures on the coxae of the fifth pair of pereopods. The shrimps were poorly preserved — the exoskeletons were soft, rostrums and telsons broken, and internal organs macerated; however, most of the features of the carapace were clearly distinct and the external genitalia intact. The discovery of these shrimp elicits several questions: to which species do they belong? What is their functional sex? Do they represent a transi- tional stage in a protandrous hermaphroditic species? If not, have their intersex-appearing anomalies resulted from parasitism? Although none of these questions is answered definitively, all are discussed following a brief description of the external genitalia of the shrimp. The specimens are deposited at the National Museum of Natural History: USNM 170581. 23 mm cl (carapace length), Bohol Strait, between Bohol and Cebu, 291 m, 25 March 1909, Albatross stn 5418. USNM 170582. 24.5 mm cl, Bohol Strait, 320 m, 25 March 1909, Albatross stn 5419. USNM 170583, 19 mm cl. Gulf of Davao, SE Mindanao, 247 m, 18 May 1908, Albatross stn 5247. Description Petasma (Figure IB-C) with length about two- thirds that of petasma of males off*, rectaciita of comparable size, and twice as long as endopod of first pair of pleopods in females of P. rectaciita (Figure 1 A) and all other congeners. Dorsomedian lobule with distinct distomedian projection and well-formed proximal plate. Dorsolateral lobule with supporting rib (in two of three specimens) ending proximally in subelliptical process. Ven- tral costa tapering distally, forming free, inwardly excavate, blunt projection directed dorsomesially at broadly obtuse angle with shaft of petasma. FIGURE 1.— A, Penaeopsis rectacuta, USNM 170586, 9 23.5 mm cl (carapace length), off Palompon. Leyte, Philippines, endopod of left first pereopod, dorsal view. B, Penaeopsis sp. USNM 170582, 24.5 mm cl, Bohol Strait, Philippines, Albatross stn 5419, petasma, dorsal view of left half. C, Ventral view of same. D. Left appendix masculina, dorsal view, same specimen. E, Penaeopsis rectaciita, USNM 170587, 6 13 mm cl, off Mindanao, Philippines, A /6a?ross stn 5518, petasma, dorsal view of left half F, Ventral view of same. G, Right appendix masculina, dorsal view, same specimen. 0.5 mm indicated. 687 Appendix masculina (Figure ID) minute, somewhat oval, bearing distal patch of setae; thickening at its base, inconspicuous. Thelycum (Figure 2A-C) with anterolateral borders of plate of sternite XIV varying from slightly concave to convex, and separated by pos- teromedian projection of sternite XIII; plate strongly slanting dorsomesially, its surface flat or each side biconvex ventrally. Lateral borders slightly concave, strongly converging posteriorly, not reaching posterior ridge but separated from it by deep depression, latter extending anteriorly adjacent to median rib and merging with an- teromedian depression; median rib broadest bas- ally, gently tapering toward, hut not reaching, posteromedian projection of sternite XIII. Median plate of XIII trilobed to cordiform, slightly to pro- nouncedly elongate, covered with setae (most lost in specimen illustrated in Figure 2Cl except for naked central concavity, setae directed anteriorly except on base of posteromedian projection where directed caudally; posteromedian projection short, with caudal margin straight or shallowly emargi- nate. Sternite XII bearing small posteromedian tooth and pair of sharp ridges, extending pos- terolaterally from base of tooth. D isCLis.sion In several features the three specimens are markedly similar to members of P. rectacuta ( Bate 1881). The rostrum (Figure 3) is straight and its second tooth is located in line with the orbital margin, the anteroventral angle of the carapace is approximately 90°, and the moderately long bran- chiocardiac carina is conspicuous, its anterior ex- tremity not nearly reaching the posterior end of the hepatic sulcus. Also the relative length of the pereopods and — in the two smaller animals — the shape of the median plate of sternite XIII are simi- lar to those in P. rectacuta. Furthermore, the three shrimps were collected together with specimens of the latter species, all three in localities where the only other Penaeopsis occurring in the area (P. eduardoi Perez Farfante 1977) was not taken — another indication that these shrimp probably be- long to P. rectacuta. The petasma and the thelycum of these shrimp are different from those of other species of Penaeopsis, including those of P. rectacuta. The ventral costae of the petasma (Figure IB-C), ta- pering distally into a short projection disposed at an obtuse angle to the shaft, differ from those of adult males of P. rectacuta in which the ventral costae turn abruptly at right angles and bear a thin marginal border that is bent inward. In the three specimens the petasma somewhat resembles that of large juveniles (with a carapace length of about 13 mm) of P. rectacuta (Figure lE-F); how- ever, in the latter the distomedian projections are less distinct than they are in my specimens or in adult P. rectacuta. Also, in juveniles of P. rectacuta the ventral costae do not taper distally into free projections, instead the tips are broad and turned Figure 2.~Penaeopsis sp. Thelyca. A, USNM 170582, 24.5 mm cl, Bohol Strait. A/iofro.ss stn 5419. B, USNM 170581, 23 mm cl. Bohol Strait, A /6a ^rosi- stn 5418. C, USNM 170580, 19 mm cl, Gulf of Davao, Mindanao, Philippmes, Albatross stn 5247. 2 mm indicated. 688 Figure 3. — Penaeopsis sp, USNM 170582, 24.5 mm cl, Bohol Strait, Philippines, Albatross stn 5419. Cephalothorax, lateral view. 5 mm indicated. at right angles to the shaft. The appendices mas- culinae (Figure ID) are considerably less well de- veloped in the present specimens than in juvenile males of P. rectacuta, in which they are circular (Figure IG). and bear only marginal setae. The thelyca of the three specimens differ from those of P. rectacuta in that the lateral borders of the plate of sternite XIV converge strongly (rather than gradually) toward the posterior thoracic ridge and are separated from the ridge by a deep groove, a unique characteristic; the median plate of sternite XIII, although trilobed in one specimen, is cor- diform in the other two, the latter resembling that of P. rectacuta. The functional sex of the three specimens can- not be ascertained because their gonads had disin- tegrated. It is unlikely that they were hermaphro- dites for each bears only one pair of gonopores. Because they have a completely developed thelycum one would expect ovipores to occur on the coxae of the third pair of pereopods, but whereas these coxae are similar in outline to those of female P. rectacuta. they lack openings and are covered by a hardened cuticle (Figure 2A-C) like those of males; instead, the coxae ofthe fifth pair of pereopods exhibit a membraneous cuticle with an opening on the proximomesial border. Although the latter aperture is situated on the last pereopod, it occurs on the coxa (Figure 4A) rather than on the bulging articular membrane as it does in typi- cal males (Figure 4B). Furthermore, no terminal ampullae — the ectal muscular region of the vasa deferentia — appear to have been present, even though the skeletal muscles are rather well pre- served. Many anomalies of the secondary sexual characters of decapod crustaceans have been re- coi'ded, e.g. in lobsters (Chace and Moore 1959, among others) and crayfishes (Turner 1924, 1929, 1935). Recently Zongker (1961) described many sexually aberrant individuals within a population o{ Caniharus niontanus acuminatus Faxon 1884. Among the aberrant individuals she found were females (sex identified by examination of the gonads) lacking ovipores on the coxae ofthe third pair of pereopods, but with "male openings" on those ofthe fifth, an anomaly similar to that exhi- bited by my specimens. In these shrimp, the aper- tures are not typical of penaeoid males because of their location on the coxae rather than on the articular membranes. Being present on the coxae, they resemble female openings; however, ovipores are typically subcircular rather than slitlike and, furthermore, they are characteristically situated on the mesial surface of the coxa, dorsal to the coxal plate, instead of on the ventral face as in my specimens. Individuals ofthe superfamily Penaeoidea bear- ing both a thelycum and a petasma have not been recorded previously in the literature. Based on size distribution and characters ofthe endopod ofthe first pair of pleopods in females, Heegaard ( 1967, 1971, 1972) suggested the possibility that protan- drous hermaphroditism occurs in Solenocera memhranacea (Risso 1816) and also in Perjaeu.'i kerathurus (Forskal 1977), but no individuals with both petasma and thelycum were found by him. The external genitalia in my three specimens causes one to suspect that they might be transi- tional forms and that therefore at least some members ofthe genus Penaeopsis exhibit protan- drous hermaphroditism (protandrous because at their size the thelyca are fully developed whereas the petasmata are relatively small). Their rather 689 Figure 4.— Right fifth pereopod. A, Penaeupsis sp. USNM 170582, 24.5 mm cl. Bohol Strait, Phihppines, A/ftarross stn 5419. B.Penaeopsis rectacuta , USNM 170586, 5 24.5 mm cl, off Palompon, Leyte, Philippines, A /6a?ross stn 5403. 1 mm indicated. large size, however, makes it unlikely that they are in a transitional stage. Furthermore, herma- phroditism has not been recorded in any species of the genus Penoeopsis, consequently its occurrence in these specimens would be exceptional. Effects of parasitism in a species of Metapenaeopsis, a genus closely allied io Penaeop- sis, were reported by Hiraiwa and Sato (1939). These authors observed conspicuous changes in the petasmata of males and the gonopores of males and females in the shrimp Penaeopsis akayebi Rathbun 1902 i = Mctapenaeopsis barbata de Haan 1850) parasitized by the bopyrid isopod Epipenaeon japonica Thielemann 1910. In males, the petasmata were considerably smaller than those in normal individuals of corresponding size and their two parts were unjoined; the gonopores were barely noticeable, and the papillae, at the tips of which the gonopores are situated in normal individuals, were lacking. In the females, the ovi- pores were obscured, but the thelycum was not apparently affected by the presence of the para- site. In the extensive material examined, how- ever, none of the specimens bore both a petasma and a thelycum. Among the specimens of P. rec- tacuta collected in the waters of the Phillippines, I found a few that were parasitized by bopyrids (one of them was taken at Bohol Strait, in a locality near those at which two of my three individuals were obtained). The parasitized specimens had normal external genitalia, thus lending no sup- port to an assumption that the anomalies in the genitalia of these three individuals were induced by a bopyrid parasite. Nevertheless, parasitism offers the only clue as to the possible origin of the anomalies present in these shrimp. Acknow ledgments I am grateful to Horton H. Hobbs, Jr., for valu- able suggestions during the study, and to Thomas E. Bowman and Fenner A. Chace, Jr., for review- ing the manuscript. My thanks are also due Maria M. Dieguez for preparing the figures. Literature Cited CHACE, F. A., JR., AND G. M. MOORE. 1959. A bicolored gynandromorph of the lobster, //omari^s americanus. Biol. Bull. (Woods Hole) 116:226-231. Heegaard. p. 1967. On behaviour, sex-ratio and growth of Solenocera memhranacea (Risso) (Decapoda, Penaeidae). Crus- taceana 13:227-237. 1971. Penaetis kerathurus Forskal, a protandric her- maphrodite. Bull. Inst. Oceanogr. Peche iSalambo) 2:257-266. 1972. Sexual dimorphism in some Mediterranean penaeids and their spawning grounds. Thalassia Jugosl. 8:5-13. 690 HiRAiwA, Y. K., AND M. Sato. 1939. On the effect of parasitic Isopoda on a prawn, Penaeopsis akayebi Rathbun, with a consideration of the effect of parasitization on the higher Crustacea in gener- al. J. Sci. Hiroshima Univ., Ser. B, Div. 1, 7:105-124. Thielemann, M. 1910. Beitrage zur kenntnis der Isopodenfauna Ostasiens. Beitr. Naturgesch. Ostasiens. Ahb. Math-Phys. Kl. K. Bayer. Akad. Wiss. II. Suppl.-Bd. 3 Abh., 109 p. TURNER, C. L. 1924. Studies on the secondary sexual characters of crayfishes. I. Male secondary sexual characters in females of Camharus propinquus. Biol. Bull. (Woods Hole) 46:263-276. 1929. Studies on the secondary sexual characters of crayfishes, IX. Females of Cambarus with aberrant female characters. Biol. Bull. (Woods Hole) 56:1-7. 1935. The aberrant secondary sex characters of the crayfishes of the genus Cambarus. Am. Midi. Nat. 16:863-882. ZONGKER, J. 1961. Monoecious tendencies in a population o{ Cambarus montanus acuminatus Faxon. M.A. Thesis, Univ. Vir- ginia. Charlottesville, 30 p. ISABEL PEREZ FARFANTE Northeast Fisheries Center Systematics Laboratory National Marine Fisheries Service, NOAA National Museum of Natural History Washington. DC 20560 ON THE ROLE OF THE DIFFERENT FIBRE TYPES IN FISH MYOTOMES AT INTERMEDIATE SWIMMING SPEEDS In most fishes the myotomal locomotor muscula- ture is made up of two main fibre types: a super- ficial layer of red fibres overlies the white fibres which form the main mass of the myotome. A spectrum of such differences as mitochondrial con- tent, enzyme activities, blood supply, and innerva- tion (as well as color) distinguishes these two fibre types. The electrophysiogical properties of the two fibre types have only been investigated in a few species, but in all of these the white fibres have been found to propagate muscle action potential, whereas only local nonpropagated activity is seen from red fibres (which are invariably multiply in- nervated). In many (but not all) fishes, there are also other less abundant fibre types in the myotomes, in some respects intermediate between the red and the white fibres (e.g., Patterson et al. 1975). There is general agreement that at low sus- tained swimming speeds only the red fibres are employed and that the white fibres are active dur- ing short bursts of maximum speed, which cannot be long sustained. However, agreement has not yet been reached about which fibres are active during sustained swimming at speeds above the minimum cruising speed. Indirect evidence from a number of teleost species (e.g., Greer Walker and Pull 1973) indicated that the white fibres are ac- tive at these intermediate swimming speeds, as did the direct electromyographic investigations of Hudson (1973). More recently, several workers have suggested that fibres of intermediate type are recruited as swimming speed rises from the minimal cruising speed, before white fibres are activated and the fish attains its maximal sustained speed. In this note, we report elec- tromyographic observations on various teleosts swimming at controlled speeds in a tunnel res- pirometer, which show that the activity of the myotomal fibre types during sustained swimming is different in different fishes. Material and Methods We studied herring, carp, and trout. Juvenile Pacific herring, Clupea harengus pallasi Valen- ciennes, 15-17.5 cm FL (fork length) were caught by seining in the Georgia Straits, B.C., and held in circulating seawater at the Department of Zool- ogy, University of British Columbia, until swum in a tunnel respirometer (Brett 1964). Herring are delicate fish and did not settle quietly in the res- pirometer at flow lengths below 2-3 body lengths per second (BL/s). Instead, they darted upstream, and fell back again in an irregular manner, so that it was necessary to force them to swim at such speeds from their first entry to the apparatus, without the acclimation period usual when work- ing with other fishes. Varnished copper wire ( 40 standard wire gauge) electrodes bared at the tips were placed in the postanal myotomes. The fish were anaesthetized with MS-2221 (Sandoz) and the electrodes sutured to the dorsal surface before being led downward and backward to enter the myotomes. After recov- ery for 30 min or so in a bucket of seawater, the fish were introduced to the respirometer and muscle potentials recorded on a Gould Brush 220 pen re- corder via Tektronix 122 preamplifiers. It proved difficult to record from electrodes whose tips lay amongst the white muscle fibres, but activity from ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 691 these fibres was picked up by electrodes whose tips lay in the thin lateral red muscle strip. Relative proportions of red and white muscle fibres were determined by dissection, and their innervation patterns were examined in Formalin-fixed mate- rial. Cryostat sections stained for lipid and for succinic dehydrogenase by routine methods were used to distinguish different fibre types. Similar studies were carried out on carp, Cy- prinus carpio Linnaeus, 25-30 cm FL, caught by seining in the Fraser Valley, B.C., and rainbow trout, Salmo gairdneri Richardson, 17-30 cm FL obtained from a commercial supplier. Before trial in the respirometer both species were held in cir- cular tanks around which water was pumped to give a constant flow of 30-40 cm/s around the cir- cumference where the fish normally swam. For these freshwater fish, much improved sig- nal : noise ratios were obtained by adding small amounts of sea water to the freshwater in the res- pirometer; this did not affect the behavior of the fish, which were allowed to acclimate for 18-20 h in the apparatus before testing. Results Pacific Herring The records of muscle activity shown in Figure lA-C are from electrodes with their tips lying in the lateral strip of red muscle. These records show that bursts of irregular potentials around 200-300 /uV peak to peak are recorded from the red fibres during slow sustained swiming, becoming more synchronous and shorter as swimming speed in- creases and tail beat frequency rises. At sustained ■••^M^^^^^ B ^Mf^^y^'^r^'^r^^ ^^«/^|4.^^^ D 4/ftW^H4i ii^f^fJ^aXt^W|^ Figure l, — Records from lateral red muscle of a juvenile Pacific herring ( 16.4 cm fork length) swimming at different speeds. A: 1.8 body lengths per second; B: 2.8 body lengths per second; C: 3.1 body lengths per second; and D: 4.3 body lengths per second. Note gradually increasing synchrony of potentials as swimming speed and tail beat frequency rise (A-C), and the appearance of large potentials in D < picked up from underlying white fibres) when the fish struggled to maintain position at high flow speed. Vertical bar: 500 ^V; time marker: seconds. 692 swimming speeds up to 4-5 BL/s, these are the only kind of potentials recorded; no larger potentials are observed. Because herring are delicate fish, velocity/endurance experiments in respirometers or water tunnels are likely to underestimate their real capabilities, for it is probable that they slowly deteriorate during their sojourn under experimen- tal conditions. Our limited series of measurements of sustained swimming speeds (Figure 2) showed that juvenile herring were able to maintain speeds around 4 BL/s for periods of at least 5 h, a per- formance about double that previously observed by Boyar ( 19611, but similar to that seen in large circular tanks by Hempel (in Blaxter 1969). Boyar's study was very much more extensive than ours; some of his results are plotted in Figure 2 for comparison, where it can be seen that the form of the velocityendurance curves we obtained is similar to those found for other fish ( e.g. , Hunter 1971 ). It seems probable that 4-5 BL/s represents a sensible upper value for continuous sustained cruising by herring of this size. If the speed of flow in the respirometer is in- creased above this speed, or if the fish becomes progressively exhausted, it intersperses periods of steady swimming, as before, (during which it slowly falls back to the downstream electrified grid) with a few rapid tail beats, which drive it upstream, and the cycle is repeated. During these rapid beats (Figure ID), large potentials around 1 mV are observed. Similar potentials are seen when the fish is struggling, and there can be no doubt that (as in dogfish. Bone 1966) the electrodes in the red fibre layer pick up these large potentials from the underlying white fibres. White fibres in o 5 - tn 4 — I t- 3 - o o m - o t • o« - o - o 1 1 1 o 1 o , 5 10 MINUTES 30 1 2 5 HOURS FIGL'RE 2. — Swimming speeds (in body lengths per second) plot- ted against the time that the speeds were sustainable (abscissa). Dots: present observations; open circles: data plotted for compar- able fish from Boyar (1961). Note the different forms of the velocity/endurance curves given by the two sets offish, probably the consequence of damage to Boyar's fish in his apparatus. herring are similar to those of dogfish in that they are focal ly innervated (Bone 1964) and they must therefore propagate action potentials. The white fibre system in herring was rapidly exhausted, for the fish could not swim at velocities above 5 BL/s for more than 1-2 min (as indicated in Figure 2). Thus there is good accord between our elec- tromyographic observations and the values ob- tained for maximum sustained swimming vel- ocities: in herring only red muscle fibres are employed during sustained cruising. Histologically, the red and white fibres are dif- ferent from each other. The red fibres are of more or less uniform diameter, are multiply innervated, and lipid and succinic dehydrogenase (SDH) posi- tive. In contrast, the white zone of the myotome contains both large fibres, and much smaller fibres arrayed around them in a sort of lattice. Both types contain little lipid, are SDH negative, and there are no intermediate fibres either in the juvenile herring which we examined in the res- pirometer, or in adults. These histological ar- rangements are summarized in Figure 3A. Carp The carp used were much more robust and larger fish than the Pacific herring and it proved possible to make simultaneous recordings of activ- ity within white and red portions of the myotomes. The results obtained were entirely different from those seen in the herring. At speeds between 0.5 BL/s (the lowest speed at which the fish would swim reliably) and the maximum speed used, around 4 BL/s, electrical activity was always de- tectable from both sets of electrodes in red and white zones of the myotomes (Figure 4). As speed increases from the lowest values, the bursts of activity from each zone became more synchronous and shorter and their amplitude increased. Occa- sional spikes of greater amplitude were observed from the white muscle zone (Figure 4B), these were faster events than those composing the re- mainder of the motor bursts. When the fish was swimming near the maximum speed sustainable in the respirometer ( Figure 4C ), these rapid poten- tials formed the larger part of the motor bursts and were always seen on both red and white record- ings, though smaller from the former. Presuma- bly, they represent spikelike activity from the white zone of the myotome, picked up (as in her- ring) by electrodes in the red zone. Since the red and white electrodes did not lie in the same 693 ^""^^"i^^^ r\ r^ — . - — ■'^ B FIGURE 3. — Diagram summarizing the organization of white muscle fibres in the myotome of Pacific herring ( A), compared with those of rainbow trout and carp (B). It is not known whether a single axon may supply both large and small white fibres in herring (though both are focally innervated), nor is it known which of the two alternative innervation patterns for rainbow trout and carp are actually present. An overlap of innervation between small and large fibres seems most likely. Note the presence ofintermediate fibres between red and white zones of the myotome in carp and rainbow trout; they are absent m herrmg. (B partly after Patterson et al. 1975 and Johnston et al. 1977.) myotome (though on the same side of the fish and fairly close to each other), the appearance of occa- sional spikelike potentials in the white zone was not always reflected directly in the record from the red. At lower swimming speeds, when the elec- trodes in the white zone did not pick up spikelike potentials at every tail beat, higher recording speed (Figure 4D) showed the variety of response from the white zone of the same myotome at suc- cessive tail beats. Spikelike potentials were present (although usually <0.5 mV in amplitude) and were often reflected at lesser amplitude by the electrodes in the red portion of the myotome, but there were also much smaller irregular potentials from the white region of the myotome, resembling the smaller irregular potential bursts from the multiply in- nervated red fibres. In carp, both red and white muscle fibres are multiply innervated and there 694 are intermediate fibres lying between red and white fibre zones ( Figure 3B ). The electrodes in the white portion of the myotome were placed close to the spinal column so that they did not lie near the intermediate zone recently described by Johnston et al. (1977). Our results clearly indicated that the white fibres were active even at low swimming speeds, and that the activity at these speeds did not re- semble the spikelike muscle potentials observed when the fish are swimming faster. Rainbow Trout Rainbow trout were examined last of the three fish studied and, to our surprise, gave results com- parable with those from the herring, although in salmonids the white fibres are multiply inner- vated, as they are in carp. At speeds below 2 BL/s, no activity was detectable from the white (mosaic) w ->-**- -Mm- <— -itU ¥i»f -Hth- -^ R w -^^Hf — f 'if "t 'j -^ t -^ — -^--f-^ -^ 4 '^ — ^ fl 'f B R ^^-^#-^^Hik^^-^4*^^-4^4~-^f^^^ 4 w f/'^ -A' - -^ -^^'v/^A/VV'^AMl Wl'' ... . — 'r^^l ■ff^r- D " 1 Figure 4. — Records of activity from red (R) and white ( W) regions of a carp myotome at different swimming speeds. A: 0.75 body length per second; B: 0.9 body length per second; C: 1 .26 body lengths per second; and D: 2.0 body lengths per second. Note that the white fibres are active even at low swimming speed and that there is spikelike activity at each tail beat in C. The lowest record (D) taken at higher chart speed shows the appearance of irregular bursts containing spikelike potentials. Note pick-up of these events by the overlying red fibre electrode. Vertical bar; 500 /nV; time marker; seconds. zone of the myotome: 1-200 /xV potentials of the usual kind were obtained from the lateral red musculature (Figure 5A). When startled, a few rapid tail beats drove the fish forward and, under these conditions, larger spikelike potentials around 0.5 mV peak to peak were recorded from the white zone of the myotome corresponding to the rapid tail beats. As can be seen from Figure 5B, these events were picked up at lower amplitude by the electrodes whose tips lay in the lateral red muscle layer. After a few rapid tail beats, the fish coasted forward before dropping back and resum- ing regular swimming: the normal rhythm of the red fibre system was inhibited for a few cycles. 695 w P — >V' . i> .ly ■ wi~^. .^^ ^ fi/ «l/.--~i\-. — .jjV . .^f. -y. ..^ ..ij — "V^"^' w ■MMM«MlM«M««M^i^^MMaaM*M«**i>*MMll«i«*M«M««*«^»MWWM>M4 B R ,y^,.«^,y..^vy^^^Hi,HVi^WHv-^^ ,«,♦W^-^^^^^-^-^^V,^A--M^"-'V'V-^*^^ w R -M~^ P" 'l*»W^**<^j f^Hk^4"Hfrf Figure 5. — Records of activity from red (R) and white (W) regions of the myotome of different rainbow trout swimming at various speeds. A: 1 .8 body lengths per second; B: 4.6 body lengths per second; C: 2.2 body lengths per second. Note that during regular sustained swimming involving red muscle activity, no activity is detected from the white zone of the myotome. Occasional spikelike activity from the white zone of the myotome is seen in B and C (also picked up by the electrode lying in the overlying red zone of the myotome), sometimes inhibits the red muscle bursts (B) and sometimes does not (C). Vertical bars: 1 mV; time marker: seconds. A single rapid tail beat (to the right of the rec- ord) interrupted the red muscle for a single cycle. At higher sustained speeds, above 2 BL/s (as in Figure 5C) this inhibition of red activity following single rapid movement no longer took place. Rather, the behavior was similar to that of the herring in that the fish fell gradually back despite the regular activity of the red system, until driven forward again by a few rapid beats; to drop back again and repeat the cycle until the white system was exhausted. Under these conditions, the fish did not "coast" following rapid tail movements. No electrical activity was observed from the white zone of the myotome (the so-called mosaic zone) apart from the spikelike potentials shown in Figure 5B and C, although particular pains were taken to ensure that the electrodes were recording satisfactorily. All the fish recorded fiom gave this same result. We conclude from our observations that this part of the motor system is not active at 696 speeds below 2-2.5 BL/s. Figure 3B summarizes the structure of the system. Discussion Our observations have shown once again that the lateral red musculature is used by fish for sustained slow cruising, and that rapid move- ments of the tail are brought about by the activity of the white motor system, during which spikelike potentials can be recorded from the white zone of the myotome. At intermediate speeds, there are manifest differences between different fishes. The simplest situation is shown by the Pacific herring, where sustained activity depends only on the activity of the red motor system of the myotome: the white fibres play no part in any activity except rapid movements of short duration. It is true that such movements can "top up," as it were, the sustained activity of the red motor sys- tern, but this process cannot be long continued: in the respirometer flow velocities which overload the red system and involve occasional activity from the white system soon exhaust the fish. Pre- sumably this artificial situation, where the fish are forced to swim at such speeds, is not found in nature. The taxonomic position of clupeids is not yet agreed upon (see Greenwood etal. 1966), but in the organization of their myotomal motor system they show the primitive pattern of focal innervation of the white fibres (Bone 1970) found also in elas- mobranches, Agnatha, and dipnoi, but in few other teleosts. We may surmise that in all fish where the white motor system is innervated in this way, sustained swimming will be the responsibility of the red system alone, as it is in herring and dogfish. It is important to notice that this is not to say that gradation may not take place separately within either system. For example, there are five fibre types in the dogfish myotome (three slow and two fast) distinguishable by histochemical and ultra- structural criteria, and it is entirely reasonable to suppose that the two fast fibre types are recruited for movements of different rapidity as Kry vi and Totland (1977) have suggested. At present, our preliminary ultrastructural and histochemical investigations of young and adult herring myotomal fibres have only shown one type of red fibre and two types of white fibre. The two white fibre types may operate at different stages during rapid swimming, but there is no direct evidence for this assumption, and it may be more reasonable to interpret the smaller white fibres as growth stages in the development of the larger (see Bone in press). In carp, the situation during sustained swim- ming at all speeds is entirely different. There is inevitably some ambiguity in the interpretation of electromyographic records since the position of the electrode tip may not be certainly known, and the records obtained may be from nearby small elec- trical events or from distant larger ones, but it certainly does not seem probable that the small events recorded from the carp white muscle at slow sustained swimming speeds can have been picked up from the distant red muscle system. To judge from our records taken deep within the white muscle, as far as possible from the lateral red strip, some fibres within the white zone are active even at the slowest sustained speeds, and this activity increases as the fish increases its swimming speed. This kind of electrical activity at the slower sustained speeds is very similar to that of the red motor system, and presumably repre- sents the activity of fibres which are not propagat- ing muscle action potentials. Such records could not, naturally, be obtained from the white system of fish where the white fibres are focally inner- vated, and in fact are not seen in herring or dogfish. At higher sustained speeds, or when the carp is disturbed, much larger rapid potentials are observed from the electrode within the white zone. Plainly, two alternative explanations are possible for the variety of electrical response from a single recording site within the white muscle. Either the electrode tip lies close to fibres of two different types, one of which is capable of propagating mus- cle spikes and the other is not. In this situation, the potentials observed simply reflect the fact that the former system is only activated at higher speeds, the latter operating during slow swim- ming and so resembling the red motor system. In other words, in the carp myotome, the arrange- ment is essentially a mosaic one, in which red fibres are intermingled with the usual fast fibres of the white zone. Or, alternatively, the white zone contains only a single muscle fibre type, which is capable of local contractions not involving muscle action potentials, but can also be stimulated to twitch rapidly and, in this state, propagates mus- cle action potentials. As pointed out earlier (Bone 1975) this would be an ingenious way of ensuring for a single muscle fibre that it always operated at the flattened upper part of the power curve, con- tracting at very different rates whilst swimming slowly and rapidly. Our electromyographic records do not allow us to distinguish between these two alternatives but there is no evidence from the histochemical studies by Patterson et al. (1975), or the recent excellent paper by Johnston ( 1977), that there are "red" fibres in the white zone of the carp myotome. These authors have demonstrated clearly, how- ever, that there is a zone of intermediate fibres between the lateral red and deep white fibres of the carp myotome. They have also shown that these three fibre types are active at different swimming speeds. At 1 BL/s only red fibres were found to be active; at 1.3-1.5 BL/s both red and pink fibres were active, whereas at 2.0 BL/s and above, electrical activity appeared from the white zone of the myotome. These results clearly indi- cated the sort of recruitment of intermediate fibres at intermediate sustained swimming speeds 697 which was implied by their accompanying biochemical studies. Interestingly enough, Johnston et al. ( 1977) observed the same kind of electrical activity from the white zone of the myotome that we observed at low speeds, and it seems therefore extremely probable that such ac- tivity (around 75 fxV in their records at 2.0 BL/s) is indeed generated by muscle fibres in the white zone. They did not observe spikelike activity from the white zone, presumably because their fish were not swimming sufficiently fast, i.e., they in- vestigated only the lower sustained swimming speed range. It is then still an open question whether indi- vidual fibres in the white zone can sometimes op- erate producing only local potentials, at other times generating muscle action potentials; or whether there are two different fibre types in the white zone, as yet not distinguishable histochemi- cally. We incline to the former opinion, but to settle the matter evidence from intracellular studies will be essential. In rainbow trout, our results were again differ- ent. We obtained no evidence for activity of the mosaic zone of the myotome during sustained ac- tivity even at 4.5 BL/s (the maximum speed at which we could swim the smaller fish). Consider- ing Hudson's (1973) electromyographic evidence from the same species, where he observed activity from the mosaic zone at speeds above 3.0BL/s, this seemed at first rather surprising. However, the fish Hudson used came from a stock of notoriously poor swimming performance (see Webb 1971), and it is therefore quite possible that we never attained the critical speed at which the mosaic muscle became active in our fish. The main muscle mass in rainbow trout consists of a mosaic of small reddish fibres scattered amongst larger pale fibres (Johnston et al. (1975) have studied them histochemically), and it is thus un- clear whether the low-level electrical activity which Hudson (1973) recorded from this region (similar to that which we found in carp white muscle) comes from the same fibres as those generating muscle action potentials during burst swimming. In other words, the two kinds of electrical re- sponses from the rainbow trout mosaic muscle may result from the activity of two different kinds of muscle fibres. Fish are so diverse, and their patterns of life so varied, that it is hardly surprising that there should be differences on their locomotor muscula- ture. We perhaps ought rather to be surprised at 698 the general uniformity of design of the locomotor system imposed by the aquatic medium. It seems probable, from the distribution of patterns of in- nervation amongst different fish groups, and in- deed amongst the teleosts alone, that focally in- nervated, twitch fibres operating by anaerobic glycolysis for short bursts of swimming represent the primitive arrangement of the aquatic fast motor system (see Bone 1970). This fast-motor system contrasts markedly with the universally found multiply innervated nontwitch red fibre system for sustained move- ment that operates aerobically. However, his- tological and biochemical investigations of the white myotomal zones of some specialized teleosts such as tuna (Guppy et al. in press) or carp (Johnston et al. 1977) have shown a definite aerobic capacity in the white fibre system, and the original simple dichotomy between anaerobic white fibres and aerobic red fibres rather naively suggested from elasmobranch studies (Bone 1966) is plainly not a good description of the operation of the myotome in all teleosts. On the whole, it seems reasonable to assume that in most teleosts where the white portion of the myotome is multiply innervated, there will be aerobic intermediate fibres for use during fast sus- tained cruising, and that at the maximum cruis- ing speed at least some fibres in the white zone of the myotome will also be active aerobically. This seems to be the situation in rainbow trout, and it probably also obtains in most scombrids. The situation in carp is less clear. The work of Smitetal. ( 1971) has shown that goldfish (close to carp) are able to sustain high speeds in a res- pirometer apparently using the white muscle system anaerobically. In line with this, Driedzic and Hochachka ( 1975) were unable to detect other energy sources than anaerobic glycolysis in carp white muscle, and Johnston et al. (1977) only found low values of aerobic enzymes in this sys- tem. We have provided clear evidence that the white motor system is operating over a wide speed range, from the lowest speed at which the fish will swim in the respirometer, and it seems bizarre that a relatively inefficient anaerobic metabolism should drive sustained activity. At low sustained swimming speeds carp might keep in overall aerobic balance by transferring lactate from the white zone to other regions of the body, where lactate could be completely metabolized (Bone 1975). Driedzic and Hochachka found only low lactate levels in the white zone after severe hypoxic stress, and suggested that this could be explained by lactate transfer out of the system. It is very hard to believe that such a process could account for the extremely interesting results of Smit and his colleagues (Driedzic and Hochachka entitled their paper "The unanswered question of high anaerobic capabilities of carp white muscle"), and we agree with Johnston et al. ( 1977) in their conclusion that carp would appear to be an ideal species for studying the relationship between muscle design and locomotor function. Ackn<)\\ ledgements We are grateful to D. J. Randall (Department of Zoology, University of British Columbia, Van- couver) for procuring the Pacific herring for us and for allowing us to use his respirometer. One of us (Q. B.) did part of this work during the award of a Nuffield-NRC visiting lectureship, which is grate- fully acknowledged, as is support (to D. R. J.) by research grants from the National Research Council of Canada and Fisheries Research Board of Canada. Literature Cited BLAXTER. J. H. S. 1969. Swimming speeds offish. In A. Ben-Tuvia and W. Dickson (editors), Proceedings of FAO conference on fish behaviour in relation to fishing techniques and tactics, p. 69-100. FAO Fish. Rep. 62 BONE. Q. 1964. Patterns of muscular innervation in the lower chor- dates. Int. Rev. Neurobiol. 6:99-147. 1966. On the function of the two types of myotomal muscle fibre in elasmobranch fish. J. Mar. Biol. Assoc. U.K. 46:321-349. 1970. Muscular innervation and fish classification. In A. de Haro (editor), I Simposio Intemacional deZoofilogenia, p. 369-377. Alamanca V.P. 1975. Muscular and energetic aspects of fish swim- ming. In T. Y-T. Wu, C. J. Brokaw, and C. Brennen (editors), Swimming and flying in nature, Vol. 2, p. 493- 528. Plenum Press, N.Y. In press. Locomotor muscle. In W. S. Hoar and D. J. Ran- dall (editors), Fish physiology, Vol. 7. Academic Press, N.Y. BOYAR. H. C. 1961. Swimmingspeedofimmature Atlantic herring with reference to the Passamaquoddy Tidal Project. Trans. Am. Fish. Soc. 90:21-26. BRETT. J. R. 1964. The respiratory metabolism and swimming per- formance of young sockeye salmon. J. Fish Res. Board Can. 21:1183-1226. DRIEDZIC. W. R.. AND P. W. HOCHACHKA 1975. The unanswered question of high anaerobic capabilities of carp white muscle. Can. J. Zool. 53:706- 712. Greenwood. P. H., D. E. Rosen, S. H. Weitzman, and G. S. Myers 1966. Phyletic studies of teleostean fishes, with a provi- sional classification of living forms. Bull. Am. Mus. Nat. Hi.st. 131:339-455. Greer Walker. M., and G. Pull. 1973. Skeletal muscle function and sustained swimming speeds in the coalfish Gadus virens L. Comp. Biochem. Physiol. 44A:495-501. GuppY. M., W. C. Hulbert. and p. W. Hochachka In press. The tuna power plant and furnace. In G. D. Sharp and A. E. Dizon (editors). Physiological ecology of tuna. Academic Press, N.Y. Hudson, R. C. L. 1973. On the function of the white muscles in teleosts at intermediate swimming speeds. J. Exp. Biol. 58:509- 522. HUNTER. J. R. 1971. Sustained speed of jack mackerel, Trachurus sym- metricus. Fish. Bull., U.S. 69:267-271. Johnston, I. A. 1977. A comparative study of glycolysis in red and white muscles of the trout iSalmo gairdneri) and mirror carp iCyprinus carpio). J. Fish Biol. 11:575-588. Johnston. I. A., W. Davison, and G. Goldspink 1977 Energy metabolism of carp swimming muscles. J. Comp. Physiol. 114:203-216. Johnston, I. A., P. S. Ward, and G. Goldspink 1975. Studies on the swimming musculature of the rain- bow trout. I. Fibre types. J. Fish. Biol. 7:451-458. Kryvi, H., and G. K. TOTLAND 1977. Histochemical studies with microphotometric de- terminations of the lateral muscles in the sharks Etmop- terus spinax andGaleus melastomus. J. Mar. Biol. Assoc. U.K. 57:261-271. Patterson S., I. A. Johnston, and G. Goldspink. 1975. A histochemical study of the lateral muscles of five teleost species. J. Fish. Biol. 7:159-166. SMIT. H., J. M. AMELINK-KOUTSTAAL, J. Vl.rV'ERBER. AND J. C. VON Vaupel-Klein 1971. Oxygen consumption and efficency of swimming goldfish. Comp. Biochem. Physiol. 39A:l-28. WEBB. P. W. 1971. The swimming energetics of trout. I. Thrust and power output at cruising speeds. J. Exp. Biol. 55:489- 520. Quentin Bone Plymouth Laboratory Marine Biological Association of the United Kingdorri Citadel Hill Plymouth PLl 2PB, England JOE KICENIUK David r. Jones Department of Zoology University of British Columbia Vancouver, B.C., Canada V6T 1W5 699 SUMMER FOOD OF THE PACIFIC COD, GADiS MACROCEPHALLS. NEAR KODIAK ISLAND, ALASKA 1,2.3 The Pacific cod, Gadiis macrocephalus Tilesius, was the target of the earliest United States com- mercial fishery in the North Pacific (Buck'*). Its fleet, organized in spring 1865 (Bean 1887), began to fish along the Alaska Peninsula and the Aleu- tian Islands and eventually expanded into the Bering Sea (Cobb 1916). Dwindling stocks and poor market prices ultimately resulted in the col- lapse of this fishery shortly after World War II (Ketchen 1961). Growing pressures in recent years on domestic fishing stocks, in addition to increased worldwide protein demand, improved technological skills and readily available investment capital, have re- sulted in renewed interest in Pacific cod in the United States (Jones 1977). A bottomfish survey off the coast of Kodiak Island and throughout Shelikof Strait by the National Marine Fisheries Service in 1973 showed the Pacific cod to be one of the most abundant fishes inhabiting the area and the standing stock was conservatively estimated to be about 36,363 t (Hughes and Parks 1975). A small experimental trawl fishery for the Pacific cod and other bottom fishes has been proposed for the Kodiak region by Jones (1977). Preliminary examination of G. macr^ocephalus stomach contents by Alaska Department of Fish and Game (ADF&G) biologist Guy C. Powell and the author during ADF&G crab investigations off Kodiak Island indicated a high frequency of oc- currence of the commercially important snow crab, Chionoecetes bairdi. In view of the probable predation pressure on existing snow crab popula- tions by G. macrocephalus and in view of the po- tential commercial importance of the Pacific cod, the summer food habits of this fish in the Kodiak area were examined by me. Ancillary goals in- cluded a comparison of food data from pot- and trawl-captured cod. 'Contribution No. 339, Institute of Marine Science, University of Alaska, Fairbanks, AK 99701. ^This study was partially supported under contract 03-5-022- 56 between the University of Alaska and NOAA, U.S. Depart- ment of Commerce through the Outer Continental Shelf En- vironmental Assessment Program to which funds were provided by the Bureau of Land Management, U.S. Department of In- terior. ■'Based on a thesis submitted in partial fulfillment of the re- quirements for the M.S. degree. University of Alaska ■•Buck.E. H. 1973. Alaska and the law of the sea. National patterns and trends of fishery development on the North Pa- cific. Alaska Sea Grant Rep. No. 73-4, 65 p. 700 Methods Specimens were taken near Kodiak Island, Alaska, (Figure 1) in conjunction with the crab- assessment studies of ADF&G and the surveys of the International Pacific Halibut Commission. Fishing gear consisted of commercial king crab pots, measuring 203 x 203 x 76 cm (inside) and weighing 340 kg; baited with chopped, frozen her- ring. Webbing was #72 tarred nylon thread with mesh stretched to 7.6 cm. The gear used on the halibut-survey vessels in July 1975 and July 1976 was a standard 400-mesh Eastern otter trawl (Greenwood 1958). Sampling by pots was from 26 June to 3 August 1973, 28 June to 31 July 1974, and 30 June to 27 July 1975. Stations usually consisted of 4-12 pots in a straight line, equally spaced every 0.46 km. Gear was pulled every 18-24 h except when weather conditions prolonged fishing time. A haphazard sample of 3,933 of Pacific cod was taken from 10,857 cod caught in pots (the number sampled was contingent on the shipboard time available for analysis of stomach contents). Food items were identified to the lowest taxon practical aboard ship, and unidentifiable contents were pre- served for later laboratory examination. Analysis of stomach contents was carried out using the fre- quency of occurrence method in which the prey organisms are expressed as the percent of stomachs containing various food items from the total number of stomachs analyzed. Cod were ar- bitrarily divided into 33-52 cm, 53-72 cm, and 73-92 cm size (total length) groups for analysis. The frequency of occurrence method was also used for food data from trawl-caught Pacific cod. The stomachs of 344 cod were examined from 24 trawl stations, which were located in the same general area as the pot stations (Figure 1). Results and Discussion As determined from the pot data, the summer diet of G. macrocephalus was fishes, crabs, shrimps, and amphipods, in decreasing order of occurrence (Table 1). The most frequently occur- ring fish was walleye pollock, Theragra chalco- gramma. Flatfishes (Pleuronectidae) and Pacific sand \a.nce, A in modytes hcxapterus, were also fre- quent. Suyehiro (1942:233-236), Moiseev (1953, 1960), and Mito (1974) also reported that Pacific cod feed on these fishes. Figure l. — Stations near Kodiak Is- land. Alaska, where Pacific cod were collected by pots and trawls during summers of 1973-75. ... • ^ .. .. . . -(11.9) "1 O.ll- (9.2) 53 1 1 1 1 1 2 2 2 1 1 3 1 34 0.1 0.1 0.1 '♦(0 3 1 3) 9 4.5] 74 3.6] 0.9^.(5.4) 24 I.2L. ■(4.8) 0.1 0.1 0.2 0.1 0.1 1 0.1 1 0.1 1 0.1 1 0.1 2 0.1 1 0.1 1 0.1 4 0.2 36 1.8 5 0.2 1 0.1 3 0.1 1 0.1 1 0.1 0.2 2.6 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.1 0.1 0.2 0.2 0.1 1.7 164 8.0 127 49 l-(10.3) 0.08 0.03 0.2 L*(0.31) 1.2 l- 0.03 (4.4) 2 0.05 1 0.03 5 0.1 0.08 0.03 0.05 0.03 0.03 0.03 0.05 0.03 0.03 0.03 0.05 0.08 0.05 0.03 0.08 0.1 109 2.7 0.03 12 0.3 0.05 0.03 0.03 0.1 0.03 0.03 0.03 0.2 0.05 0.03 11 0.3 106 2.7 0.03 0.03 0.05 0.03 0.05 0.01 0.05 0.05 0.08 0.03 0.08 0.03 0.03 0.08 0.03 0.1 0.05 0.03 0.2 0.03 0.03 0.01 0.03 61 1.5 -(10.2) 326 8.3 702 TABLE 1. — Continued. FOOD ITETIS 1973 N=689 7. Freq. Freg. 1974 N= 1183 % Freq Fr€ q 34 2. 9 4 0. 3 195 16. 5 118 10. cT 4 0. 3 1975 N=2061 % Freq . Freq . TOTAL 1973-75 N=3933 X Freq. Freq. Malacostraca Euphausiacea (krill) and Mysidacea (mysids) Isopoda (pill bugs) Amphipoda (sand fleas) A^pelisaa macroaephala Decapoda Pandalidae (shrimps) Pandalus hovealis Pandalopsis dispar Pandalus goniurus Pandalus hypsinotus Pandalus montagui tridens Pandalus platyseros Crangonidae (shrimps) Argis arassa Scleroemngon boreas Hippolytidae Spivontoaar^is sp. Unidentified shrimps Paguridae (hermit crabs) Etassoahirus oavinanus Elassochirus tenuinanus Lithodidae (crabs) Paralithodes camtsahatioa Placetron wosnessenskii Fhinolithodes uosnessenskii Galatheidae (crabs) Munida quadrispina Cancridae (crabs) Cancer sp. Telmessus cheiragonus Pinnotheridae (pea crabs) Pinnixa sp. Hajidae (spider crabs) Chionoecetes bairdi Hyas lyratus Oregonia gracilis Unidentified crabs Echinodermata Asteroidea (sea stars) Ctenodiscus cnspatus Echinoidea (sea urchins) Holothuroidea (sea cucumbers) Ophiuroidea (brittle scars) Ophiura sarsi Vertebrata Osteichthyes Clupeidae (herrings) Clupea harengus pallasi Osmeridae (smelts) Mallotus villosus Gadldae (codfishes) Theragra chalaogramma Gadus macrocephatus Zoarcidae (eelpouts) Lyaodes brevipes Scorpaenidae (rockflshes) Hexigrammidae (greenlings) Pleurogramus monopterigius Cottidae (bullheads) Dasyaottus se tiger Hemilepidotus jordani Gyrmocanthus sp . Agonidae (poachers) Bathymasteridae (ronquils) Bathymaster signatus Trichodontidae (sandfishes) Trichodon trichodon Cyclopteridae (lumpsuckers) Pleuronectldae (flatfishes) Atheresth.es storrias Hippog lossoides e lassodon Hippoglossus stenolepis Ammodytidae (sand lances) Arrmodytes hexapterus Stichaeidae (pricklebacks) Crypacanthodidae (wrymouths) Lynaoneotes aleutensis Unidentified fishes Stomachs empty 20 3 192 67 77 131 24 281 13 12 12 7 29 1 22 20 14 9 256 2.9 0.4 27.8 9.7 11.1 1 95 19.0L-(39.8) 82 3.41 21 0.3 0.5 0.1 0.7 40.7 1.8 1.7 I- 36 0.1 0.1 0.3 0.8 0.4 1.7 1.0 4.2 0.1 1.1 5 3 •-(0.4) - 0.1 3.2 2.9 2.0 1.3 37.1 1.6 32 13 9 27 4 1 21 20 0.1 8.0 6.9^^(25.3) 171 1.8 0.1 0.8 0.1 0.1 3.0 428 36.2 44 3.7 3 0.3 "(49.2) 3 0.3 2 0.2 23 •-(55.9) 476 59 0.4 0.3 0.1 0.2 2.7 0.9 0.8 0.1 2.3 0.3 0.1 0.3 0.1 1.8 1.7 0.3 40.2 5.0 '-(0.9) 1 1 1 10 3 2 181 8.8 235 10 0.5 17 407 19. "s] 794 52 2.5^(22.3) 52 185 166 23 4 7 8 4 458 3 5 166 8.1 19 0.9 4 0.2 7 0.3 8 0.4 3 0.2 286 13.9 3 0.2 5 0.2 55 2.7 2 0.1 3 0.2 31 1.5 2 0.1 1 0.1 1 0.1 13 0.6 0.2 5 a.3_'-(32.9) 384 100 2 4 42 3 1 1 14 4 3 64 0.1 1.1 735 35.6 42 2.0 6 0.3 -(46.4) 4 0.2 0.1 0.1 0.1 0.5 0.2 0.1 Hi.i) 4 1 2 17 6 2 2 0.1 4 0.2 1 0.1 109 5.3 3 0.2 7 0.3 3 0.2 2 0.1 6 0.3 2 0.1 1 0.1 6 0.3 17 0.8 0.1 2 0.1 5 0.2 40 1.9 2 0.1 12 0.6 2 0.1 9 0.4 10 0.5 6.0 0.4 20 0.2~l 1.3 L.<21. 5) 4.7 4.2 0.6 0.1 0.2 0.2 0.1 11.6 0.08 0.1 0.1 9.8 '-<31.8) 2.5 0.05 0.1 1.1 0.08 0.03 0.03 0.4 0.1 0.08 1.6 1444 36.7 99 2.5 9 0.2 K44.7) 19 0.5 '-<46.0) 0.1 0.03 0.05 0.4 0.1 0.05 ■-(0.73) 9 0.2 9 0.2 1 0.03 153 3.9 23 0.6 45 1.1 3 0.08 2 0.05 2 0.05 41 1.0 2 0.05 1 0.03 6 0.1 20 0.5 0.05 6 0.1 7 0.2 83 2.1 2 0.05 12 0.3 2 0.05 49 1.2 24 0.6 4 0.2 17 0.4 -(51.9) 655 31.8L-(44.1) 1387 35.3 '-(48.2) 184 8.9 251 6.4 703 Table 2. — The importance of the snow crab, Chionoecetes bairdi, in the summer diet of Pacific cod. Analysis based on specimens from pots. Crab incidence is given for total number of cod examined; incidence as a percent of feeding cod given in parentheses. Cod examined Feeding cod Incidence of crabs Crabs Average crab occurrence Sampling date (no.) (%) Number Percent (no.) in cod feeding on crabs 26 June-3 August 689 98.8 281 40.7 1,022 3,6 1973 (41.3) 28 June-31 July 1,183 95.0 427 36.2 1.033 2.4 1974 (38,0) 30 June-27 July 2.061 91.0 734 35.6 2,682 3.6 1975 (39,1) Total 3.933 93.6 1.442 36,7 (39.2) 4.737 3.3 Table 3. — Frequency and percent frequency of occurrence of food items in stomachs ofGadus mac- rocephalus collected July 1975 and 1976 by otter trawl near Kodiak Island, Alaska. N = number of stomachs examined. Subtotals in parentheses. Ju ly July Total 1975 1976 1975- -1976 N = 150 N = = 194 N = 344 Food items Freq. _%_ Freq Freq %_ Freq Freq. % Freq Annelida Polychaeta 2 1.3 3 1.5 5 1.4 Mollusca Pelecypoda and Gastropoda 17 11.3 10 5.1 27 7.8 Cephalopoda 3 2.0 8 4.1 11 3.2 Arthropoda Crustacea Euphausiacea and Mysidacea 13 8.6 10 5.1 23 6.7 Isopoda - - 3 1.5 3 0.9 Amphipoda 14 9.3 15 7.7 29 8.4 Decapoda Pandalidae 16 10.7 24 12.4' 40 11.6 Crangonidae 37 24.7 37 19.1 74 21.5 Unidentified shrimps 18 12.0 •(47 4) 24 12.4 -(43 9) 42 12.2 -(45 3) Ma j idae Chionoecetes bairdi 55 36.7 82 42.T 137 39.8 Unidentified crabs 13 8.7 ^(45 4) 23 11.9 -(54 2) 36 10.5 -(50 3) Echinodermata 1 0.6 - - 1 0.3 Vertebrata Osteichthyes Cadidae Theragra ahalaogranma 6 4.0 7 3.6 13 3.8 Pleuronec tidae 5 3.3 4 2.1 9 2.6 Ammodytidae Armodytes hexapterus 20 13. 3 13 6.7 33 9.6 Unidentified fishes 66 44.0 -(64 6) 70 36.1 ^(48 5) 136 39.5 - (55 5) Stomachs empty 7 4.7 13 6.7 20 5.8 Table 4. — Comparison of percent frequency of occurrence of summer food groups in male and female Gadua macrocephalus caught by pots and trawls in the Kodiak Island area. Percent frequency of occurrence in Pot-caught cod Trawl- caught cod Food groups Males Females Males Females Fishes 21 8 24.2 26.3 24,8 Crabs 22,0 19.3 24,2 20,9 Sfirimps 15,1 14,2 15,4 24,7 Amphipods 10,0 14,3 4,1 4,3 Gastropods and pelecypods 5,0 4,7 3,3 4,5 Cephalopods 36 4,7 2,3 09 Eupfiausiids and mysids 2,1 4,0 4,0 2,7 Polychaetous annelids 14 3,1 0,3 1,1 Echinoderms 0.4 0.4 0,1 0,2 Isopods 0.2 0.2 0,5 0,4 Empty stomachs 4.4 2.0 2,8 3,0 Stomachs examined (no.) 2,106 1,827 188 156 other studies on Gadiformes (e.g., Romans and Vladykov 1954; Wigley 1956; Powles 1958; Wigley and Theroux 1965). A significant difference ( x~, a = 0.05) was found for occurrence of food groups between years for each size group (Figure 2). The only similarity was among 33-52 cm fish between 1973 and 1975 and among 73-92 cm fish between 1974 and 1975. Some trends in frequency of food groups by cod size were apparent (Figure 2). Fishes and cephalopods increased in frequency with increasing cod size over all years while amphipods and polychaete worms decreased. Daan (1973) investigated the relative size of food items (crustaceans and fishes) used by the Atlantic cod, G. morhua, and found 704 1973 1974 1975 100 50 "" 100 CL IT D U " 50 0 100 -I 1 r- 73 92 cm -I ^ 1 r- 50 33 52 cm N 66 -■ 1 1 1 1 1 1 1 r- 53 72 cm N - 424 —I \ \ r- «= 199 S. fja. 100 50 0 100 50 100 73 92 cm - 53 72 cm n 1 \ — 33 52 cm 50 — -\ 1 r- «= 125 N = 719 N = 339 100 50 100 50 100 -I 1 1 r 73 92 cm 53 72 cm yn N = 212 T 1 r- N = 1047 ." 2 e O O D. a E < Figure 2. Food Items -Percent frequency of occurrence of summer food items within three size groups of pot-caught Pacific cod by year of coUection — 1973-75 — near Kodiak Island, Alaska. that smaller crustaceans were more commonly found in small cod while a gradual shift to a mixed diet of larger prey (primarily fishes) was noted for the larger fish. Arntz ( 1974) examined juvenile G. morhua, and found the most important food to be small crustaceans, mainly cumaceans (35.6% by weight of the total food consumed); fishes ac- counted for only 15.3% by weight of the total food consumed. This trend of large cod being more pis- civorous than small cod has also been dem- onstrated by Powles (1958) and Rae (1967). Acknowledgments I am especially indebted to the ADF&G and Guy C. Powell for their assistance in collection of data. Special thanks to Howard M. Feder, University of Alaska, for his editing suggestions. Thanks to John R. Hilsinger, University of Alaska, for allow- ing me to use his Pacific cod feeding data obtained during International Pacific Halibut Commission surveys, and to George Mueller and Kenneth Vogt, both of the Marine Sorting Center, Univer- sity of Alaska, for their taxonomic assistance. Mol- lusc identifications were made by Rae Baxter, ADF&G. Literature Cited Arntz, W. E. 1974. The food of juvenile cod iGadus morhua L.) in Kiel Bay. [In Germ., Engl, summ.] Ber. dtsch. wiss. Komm. Meeresforsch. 23:97-120. BARR, L. 1970. Alaska's fishery resources — the shrimps. U.S. Fish Wildl. Serv., Fish. Leafl. 631. 10 p. BEAN, T. H. 1887. The cod fishery of Alaska. In G. B. Goode and staff of associates, Fishery and fishery industries of the United States, Sec. V. Vol. 1, p. 198-226. Wash. BROWN, R. B., AND G. C. POWELL. 1972. Size at maturity in the male Alaskan tanner crab, Chionoecetes bairdi, as determined by chela allometry, reproductive tract weights, and size of precopulatory males. J. Fish. Res. Board Can. 29:423-427. COBB, J. N. 1916. Pacific cod fisheries. Rep. U.S. Comm. Fish., 1915, append. 4, 111 p. (Doc. 830.) DAAN, N. 1973. A quantitative analysis of the food intake of North Sea cod, Gadus morhua. Neth. J. Sea Res. 6:479-517. 705 Greenwood, M. R. 1958. Bottom trawling explorations of southeastern Alaska. 1956-1957. Commer. Fish. Rev. 20( 12):9-21. ROMANS, R. E. S., AND V. D. VLADYK JV. 1954. Relation between feeding and the sexual cycle of the haddock. J. Fish. Res. Board Can. 11:535-542. HUGHES, S. E., AND N. B. PARKS. 1975. A major fishery for Alaska. Natl. Fisherman 55(13):34-40 JONES, W. G. 1977. Emerging bottomfish fisheries - potential ef- fects. Alaska Seas Coasts 5:1-5. KETCHEN, K. S. 1 96 1 . Observations on the ecology of the Pacific cod (Gadus macrocephalus ) in Canadian waters. J. Fish. Res. Board Can. 18:513-558. MITO, K. 1974. Food relation in demersal fishing community in the Bering Sea - walleye pollock fishing ground in October and November 1972. Master's Thesis, Hokkaido Univ., Hakodate, 86 p. MOISEEV, P. A. 1953. |Cod and flounders of far-eastern waters.] Izv. Tikhookean. Nauchno-Issled. Inst. Rybn. Khoz. Okeanogr. 40:1-287. (Tranl. 1956, Fish Res. Board Can. Transl. Ser. 119, 576 p.) 1960. On the habits of the cod-fish Gadus morhiia mac- rocephalus Tilesius in different zoogeographical re- gions. [In Russ., Engl, summ.] Zool. Zh. 39:558-562. POWLES, P. M. 1958. Studies of reproduction and feeding of Atlantic cod (Gadus callarias L.) in the southwestern Gulf of St. Law- rence. J. Fish. Res. Board Can. 15:1383-1402. Rae, B. B. 1967. The food of cod in the North Sea and on the west of Scotland grounds. Dep. Agric. Fish. Scotl, Mar. Res. 1967(1), 68 p. RONHOLT, L. L. 1963. Distribution and relative abundance of commer- cially important pandalid shrimps in the northeastern Pacific Ocean. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 449, 28 p. SUYEHIRO, Y. 1942. A study on the digestive system and feeding habits of fish. Jpn. J. Zool. 10:1-303. WIGLEY, R. L. 1956. Food habits of Georges Bank haddock. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish 165, 26 p. WiGLEY, R. L., AND R. B. THEROUX. 1965. Seasonal food habits of Higlands Ground had- dock. Trans. Am. Fish. Soc. 94:243-251. Stephen C. Jewett Institute of Marine Science University of Alaska Fairbanks, AK 99701 A COMPUTER SOFTWARE SYSTEM FOR OPTIMIZING SURVEY CRUISE TRACKS' Since 1972, the Southeast Fisheries Center, Na- tional Marine Fisheries Service, NOAA, has been conducting resource assessment surveys for groundfish in the northern Gulf of Mexico. Ran- dom sampling stations were selected and cruise tracks plotted by hand requiring several man- days of effort without assurance than an optimum cruise track had been chosen. Consequently, a computer routine was developed at the NMFS Na- tional Fisheries Engineering Laboratory, Bay Saint Louis, Miss., to satisfy two requirements: Generate a set of randomly selected sampling sta- tions from a preestablished station grid and minimize the distance the vessel must travel to sample each station once. This paper presents the resultant routine, a comparison of results with actual cruises, and a discussion of other possible applications of the program. Background The problem of determining the optimum cruise track to sample a given set of stations can be re- stated as, "determining the shortest route from one point to another which allows a vessel to visit every station once." This problem is similar to one in the field of operations research generally refer- red to as "the traveling salesman problem." The original formulation of the problem was to minimize the time required by a traveling sales- man to visit a number of cities and return home (Bellmore and Nemhauser 1968). Several al- gorithms have been developed which solve the problem exactly; however, computer storage and running time increase exponentially with the number of points to be visited. Because the groundfish surveys normally deal with station numbers in excess of 100, an heuristic method of solving the problern was selected. Lin and Ker- nighan ( 1973) at the Bell Telephone Laboratories (BTL) developed an approximate procedure for solving traveling salesman problems with large number of visitation points which appeared applicable to cruise track optimization. ^ The Na- tional Fisheries Engineering Laboratory obtained 'Contribution No. 78-19F from the Southeast Fisheries Center, National Marine Fisheries Service, NOAA, NSTL Sta- tion, MS 39529. MARMAP Contribution No. 154. ^To develop a feeling for the complexity of these problems, it should be noted that for a given number of stations, n, there are 706 a Fortran program from BTL and converted it to operate on a Univac^ 1108 system at the National Aeronautics and Space Administration Computer Complex, Slidell, La. Modifications to the BTL algorithm were made to satisfy requirements of the groundfish survey program. Most internal modifications were fairly general so that the program could be used for other areas and purposes. Specifics of grid locations and random selection requirements were stored on magnetic tape in a separate master file. The pro- gram, as presently configured, can handle up to 150 stations; however, 300 stations could be han- dled using extended core storage. Algorithm Description Assume a number of stations (n) have been selected, either randomly or specifically. There are a total ofnin - l)/2 links between the n stations. The object is to find an n -subset of these links such that (a) each station is sampled exactly one time, and (b) the total distance traveled is a minimum. A sequence of links satisfying (a) is called a tour; if it also satisfies (b), it is the optimum tour. The optimization algorithm begins by comput- ing all nin - l)/2 distances and storing them in a matrix. A completely random tour is generated to use as a starting point. An attempt is then made to find two sets of links A' = .Vj , .V2 • • • ^'/,. o''<^ Y - v, , y.2 . . . V;, such that if the links in X are replaced with the links in Y, the result gives a tour of a shorter distance. This is done by identifying^ j and yi as the "most-out-of-place" pair, setting them aside, then proceeding with .t^ and y2, x^ and y^, and so on. A criterion is then used to determine how many pairs of links are to be exchanged. This criterion can be explained as follows: Let the length of.r, and y, be dx, and dy,, and g, = dx, - dy^. This deter- mines the gain (shorter distance) by exchanging x, withy,. After examining a sequence of proposed exchanges x^ , X2 ■ • . a:^ and y 1 , Vg . . . y^ with their corresponding gains §1,^2 • • -Sk^ the actual value of ^ that defines the number of sets to exchange is the one for which §, + g2 + . . . +g^ is always zero or negative. This indicates the solution is a local (n - 1 ) factorial possible cruise tracks that satisfy the criterion of sampling all stations once and returning to starting position (e.g., if n = 101, the number of possible solutions is 9.3326 x 10'"). ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. optimum based on the fact that if a sequence of numbers has a positive sum, there is a cyclic per- mutation of these numbers such that every partial sum is positive. Hence, the algorithm looks for sequences ofg/s whose partial sum is always posi- tive, reducing the number of sequences that need to be examined. This means that the value of ^, which gives the number of links to be exchanged, IS determined when G* = S g-^Q, i.e. when the partial sum of the gains fails to remain positive. These links are then exchanged and the process of selecting new links to be exchanged begins again ati = 1. When all possibilities have been tried, the tour length is recorded. The program generates a new random initial tour and the entire process begins again. Eleven distinct solutions are pro- duced in this manner, and the tour with the short- est length is considered the optimum solution. Program operation can best be understood by a simple example. Assume that n stations are selected and a ran- dom tour generated (Figure la5. The black dots represent the stations and the circle represents the random tour. Any station S 1 is selected and S2 is designated as an adjacent station in the tour. The link connecting the two stations is designated Figure 1. — Example of the algorithm operation (modified from Lin and Kemighan 1973). 707 as .\\ as shown schematically in Figure lb. The station closest to S., is designated as S3 and Vj is the link joining S^ and S;^. The link .\', is not permitted to be either of the links already connected to S.,- The gain cri- terion is then calculated as g^ = c/.v, - rA',. If this is negative, S.2 is designated as the other neighbor of S, in the tour. If^, is positive, S4 is designated as one of the tour neighbors of S;j as shown in Figure Ic. If y2 were chosen to join S^ with Si, the result would be a tour. The gain criter- ion is then calculated asg? ^ dx2 - dy^. If^i + ^2 >0, the original tour could be improved by ex- changing Xi and .V.2 with .v'l and y2, respectively. This potential improvement, which results from closing up the tour immediately (G'-' = g^ +^2'' is then stored. Now S5 is chosen as the nearest neighbor of S4, and _V2 is designated as the link connecting the two stations. Station S5 is not per- mitted to be either of the stations already con- nected to S^. Figure Id shows there is only one choice for station Se and the link X3 such that if Sg is connected toSj, a tour remains. If S^ were chosen as the other neighbor of S5 in the original tour, closing up Sg to Sj would result in a tour of two disconnected pieces (Figure le). The gain as- sociated with closing up immediately (connecting Sg with S, ) is then compared with that obtained by joiningS4toS] (G* ). The link connecting Sg to S, is designated asy j. The gain criterion is then calcu- lated as §3 = dx-f, - dy:i. Ifg^ + g-z + gs^G'' G*, however, a new station S^ and link .V4 are selected and the process is continued. A limited backtracking feature of the program is included for the case when G- = 0 (i.e., no im- provement can be made). The link ^2 was chosen ( Figure Id ) to join S-, to S4 as the closest station to S4. When no improvement is made at some stage (G* = 0), new links y2 are considered in order of increasing length to a maximum of five choices. If still no improvement is found, the fiveV] links are examined in order of increasing length. When G" cannot be improved, and the valued determined, a new initial station S, is selected and the process repeated. The procedure ends when all n stations have been examined. A new random tour is gener- ated, and each station is examined as an Sj again in the same manner. This limited backtracking significantly increases program effectiveness. The computational procedure has other features that improve the calculations and reduce running time; such as limited foresight to the next links to be broken, allowance for nonsequential link ex- changes, and elimination from computation of those links previously recorded in good tours. For a more complete description of the algorithm, see Lin and Kernighan (1973). Results Station Description Separate station grids lere used for areas east and west of the Mississippi River Delta. A station consisted of a rectangle, lat. 2'30" by long. 2'30", within which three trawl tows were made. Sta- tions were identified and located at the center point of the rectangle. The station grid for the West Delta area con- sisted of an area extending from long. 89°30'W to 91°30'W (Figure 2). The station gind for the East Delta area consisted of a primary and secondary zone extending between long. 88°00'W and 89^^30'W and long. 79°30'W and 88"00'W (Figure 3). Each area was limited by the 9.2-m (5-fm) and 92-m (50-fm) depth contours. Stations were excluded from random selection in both areas be- cause of navigation and trawling hazards, and areas of known low groundfish densities. Random Selection Station number, latitude, and longitude were stored in a master grid file for each area. Input to station selection for the West Delta region was the number of stations to be sampled. This region had 780 stations. For the East Delta area, the number of stations must be specified separately for the primary and the secondary zones — there were 555 stations in the primary zone and 139 in the sec- ondary region. Station selection was performed by a random number generator which selected sta- tions based on the number required for each area. Crujse Track Optimizatit)n Requirements for an optimized cruise track were different for the areas east and west of the delta. A round-trip track was desired for the West Delta area, while a one-way calculation was de- sired for the East Delta area. The latter consisted of the shortest route from a designated starting point near Pascagoula. Miss., through each selected station and ending at a point near the mouth of the Mississippi River. 708 Figure 2.— Master station grid for groundfish survey sampling in northern Gulf of Mexico - West Delta area. Dot labeled 00-00 is start and end point for round-trip cruise track optimization. Since the grid used in the calculations was square, a coefficient was included to account for differences in absolute distance for one unit of longitude vs. one unit of latitude. The coefficient used for optimizing groundfish survey tracks is 52.10/59.85, which is the ratio of the distance in nautical miles for 1° of longitude to that for 1° of latitude at lat. 30°N. All longitudinal Cartesian coordinate distances were multiplied by this coefficient before calculations began. For the West Delta area, the cruise track was optimized from a point located just east of the primary survey area (Pascagoula station number 00-00) through all randomly selected stations, re- turning to the starting point. The optimization program computed 11 solutions and the best route in terms of the shortest distance was selected. Output consisted of a listing of stations in proper sampling order, and a plot of the stations and optimum cruise track with every fifth station labeled. The starting point of the cruise track was south of Pascagoula for the East Delta area. Optimiza- tion was done for a cruise track that visited all 709 randomly selected primary and secondary stations and terminated at a point near the Mississippi River Delta designated 99-99 (Figure 3). Outputs were the same as for the West Delta except for treatment of the stations randomly selected which appear in blocks 45, 46, 47, and 48. These were not included in the optimized cruise track, but were listed at the end of the optimized cruise track list- ing. The stations in these blocks were added to the end of the optimized cruise track and plotted as individual points labeled with their Pascagoula number. Figure 3. — Master station grid for groundfish survey sampling in northern Gulf of Mexico - East Delta area. Dot labeled 00-00 is starting point and dot labeled 99-99 is end point for one-way cruise track optimization. Primary and secondary areas are indicated by arrows at top of figure. 710 30.0 29.5 29.0 28.5 28.0 92.0 91.5 WEST ORIGINAL TOUR 91.0 90.5 90.0 89.5 89.0 Figure 4. — Actual cruise track followed for West Delta area, FRV Oregon II cruise 55. Every fifth station is labeled. 30.0 29.5 29.0 28.5 28.0 92.0 91.5 WEST TOUR 10 91.0 90.5 90.0 89.5 89.0 Figure 5.— Optimized cruise track for West Delta area, FRV Oregon II cruise 55. Ever>- fifth station is labeled. 711 Test Case and Sample Products The optimization program was tested to com- pare computational results with a cruise track actually fol lowed during a survey — FR V Oregon II cruise 55, 5-29 November 1974. West Delta The 126 stations sampled during cruise 55 for the West Delta area were entered in the order they were sampled (Figure 4), and the total dis- tance (in grid units) was calculated to be 254. Each grid unit was equivalent to approximately 4.6 km; thus, the total distance was about 1,176 km. Eleven computations were performed on these stations by the optimization program, and a minimum length of 233 grid units (approximately 1,078 km) occurred three times. It can be said with confidence the optimum tour (Figure 5) rep- resented an 8.3% improvement over the actual cruise track. Distances were calculated from the center of each subsquare; therefore, the actual dis- 30.5 30.0 •04 29.5 29.0 28.5 89.5 EAST ORIGINAL TOUR 89.0 88.5 88.0 87.5 Figure 6.— Actual cruise track followed for East Delta area, FRV Oregon II cruise 55. Every fifth station is labeled. Station numbers listed at lower left are those not included in optimization calculations. 712 tance would be decreased by the vessel cutting corners of the subsquares. Calculations for the 126 stations on the Univac 1108 system used about 60K of core storage and required 2 min of Central Processing Unit (CPU) time. East Delta Cruise 55 was used to test the program for the East Delta area also. Of 1 16 stations sampled, 105 were included in the computation of an optimum one-way cruise track. The other 11 stations were located in blocks 45, 46, 47, and 48. They were, however, added to the end of the optimized listout, plotted, and labeled on the cruise track plot. The actual cruise track distance for the 105 stations was 229 grid units (approximately 1,061 km) (Figure 6). The optimized one-way path was calcu- lated to be 216 grid units ( 1,000 km), an improve- ment of 5.8'7f (Figure 7). Calculations for the 105 stations used 60K of core storage and required 66 s of CPU time. 30.5 30.0 29.5 29.0 28.5 89.5 EAST TOUR 10 89.0 88.5 88.0 87.5 Figure 7. — Optimized cruise track for East Delta area, FRV Oregon II cruise 55. Every fifth station is labeled. Station numbers listed at lower left are those not included in optimization calculations. 713 Discussion The basic optimization program has the capabil- ity and inherent versatility to be utilized for a wide range of applications. The round-trip capa- bility can be modified to a one-way path calcula- tion as was done for the East Delta portion of the groundfish survey by manipulating the distance matrix. Cartesian integrity of the start-stop points is kept intact but the distance between the two stations is set equal to zero in the distance matrix. The program then calculates the optimum tour as if the start-stop points were very close together when, in fact, they are not. There is no requirement that distance be the optimization parameter. Factors such as cost, time, or suitable weighted combinations of other variables could be used to compute a cruise track considered optimum for specific user require- ments. Also, there is no requirement that the prob- lem be symmetric or Cartesian in nature. For example, the distance (cost, time, etc.) in going from station A to station B need not be equal to that from station B to station A. Applications of these characteristics and other distance matrix manipulations include: 1) The "cost" in going from station to station in the presence of strong currents, such as the Gulf Stream, could be adjusted. "Downstream" directions from station to station would be given preferential status for computing the op- timum cruise track. 2) In some situations, it may be desirable to group selected stations to be sampled preferentially as a subset or subsets of the total station pat- tern. This might occur if certain sampling areas had a higher priority than others because of biological and/or environmental considera- tions. 3) Actual curvilinear distances between stations could be entered into the distance matrix when sampling in areas near the coast. This would be done for station pairs connected by a straight line that passes across land. 4) If the number of stations exceeds the present 150 maximum allowable (300 with extended core storage), and it is possible to divide them into subgroups, the problem is limited only by CPU restrictions. Many variations of the optimum cruise track theme could be solved with this program and the requirements are usually unique to a particular problem or investigation.'* The examples demon- strate the types of problems that could be solved. Simple problems, such as those solved for the groundfish survey, can be improved about 7% over manually produced cruise tracks. Improvements obtained using the optimized cruise track for the cited application are not dramatic, but would be significant over a long time period and/or extensive cruising distance. The program eliminates selecting stations from ran- dom number tables and hand plotting the cruise track, which may require several man-days Literature Cited Oper. BELLMORE, M., AND G. L. NEMHAUSER. 1968. The traveling salesman problem: A survey. Res. 16:538-558. Lin, S., and B. W. Kernighan. 1973. An effective heuristic algorithm for the traveling- salesman problem. Oper. Res. 21:498-516. Thomas D. Leming hillman j. holley Southeast Fisheries Center National Fisheries Engineering Laboratory National Marine Fisheries Service, NOAA NSTL Station, MS 39529 ■^Inquiries regarding possible uses and applications of this system should be directed to the Director, Southeast Fisheries Center National Fisheries Engineering Laboratory, National Space Technology Laboratories, NSTL Station, MS 39529. 714 Notices NOAA Technical Reports NMFS published during the first 6 mo of 1978. Circulars 409. Marine flora and fauna of the northeastern United States. Copepoda: cyclopoids parasitic on fishes. By Ju-Shey Ho. February 1978. iii + 12 p.. 17 fig. 410. The 1976 Cerotium tripos bloom in the New York Bight: causes and consequences. By Thomas C. Malone. May 1978. iv + 14 p. 17 fig., 1 table. 411. Systematics and biology of the tilefishes (Per- ciformes: Branchiostegidae and Malacanthidael, with descriptions of two new species. By James K. Dooley. April 1978, v + 78 p., 44 fig., 26 tables. 412. Synopsis of biological data on the red porgy, Pa- grus pagrus (Linnaeus). By Charles S. Manooch III and William W. Hassler. May 1978, iii + 19 p., 12 fig., 7 tables. Also FAO Fisheries Synopsis No. 1 16. For sale by the Superintendent of Documents. U.S. Govern- ment Printing Office. Washington, DC 20402 Stock No. 003-017-00418-0. 413. Marine flora and fauna of the northeastern United States. Crustacea: Branchiura. By Roger F. Cressey. May 1978, iii -i- 10 p., 15 fig. For sale by the Superin- tendent of Documents. U.S. Government Printing Office, Washington, DC 20402 Stock No. 003-017- 00419-8. Special Scientific Report — Fisheries 719. Seasonal description of winds and surface and bot- tom salinities and temperatures in the northern Gulf of Mexico, October 1972 to January 1976. By Perry A. Thompson, Jr. and Thomas D. Leming. February 1978, iv -t- 44 p., 43 fig., 2 tables. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402 Stock No. 003-017-00414-7. 720. Sea surface temperature distributions obtained off San Diego, California, using an airborne infrared radiometer. By James L. Squire, Jr. March 1978, iii -i- 30 p., 15 fig.. 1 table, 90 app. fig. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402 Stock No. 003-017-00415-5. 721. National Marine Fisheries Service survey of trace elements in the fishery resource. By R. A. Hall, E. G. Zook, and G. M. Meaburn. March 1978, iii + 313 p., 5 tables, 3 app. fig., 1 app. table. 722. Gulf menhaden, fireroor//a patron ;/s, purse seine fishery: catch, fishing activity, and age and size com- position, 1964-73. By William R. Nicholson. March 1978, iii -H 8 p., 1 fig., 12 tables. 723. Ichthyoplankton composition and plankton vol- umes from inland coastal waters of southeastern Alaska, April-November 1972. By Chester R. Mattson and Bruce L. Wing. April 1978, iii + 11 p., 1 fig., 4 tables, 1 app. table. 724. Estimated average daily instantaneous numbers of recreational and commercial fishermen and boaters in the St. Andrew Bay system, Florida, and adjacent coastal waters, 1973. By Doyle F. Sutherland. May 1978, iv + 23 p., 31 fig., 11 tables. 725. Seasonal bottom-water temperature trends in the Gulf of Maine and on Georges Bank, 1963-75. By Clar- ence W. Davis. May 1978, iv + 17 p., 22 fig., 5 tables. NOAA Technical Reports NMFS are available free in limited numbers to Federal and State government agencies. They are also available in exchange for other scientific and technical publications in the marine sciences. Individual copies, if available, may be obtained by purchase from the Superintendent of Documents or by writing to User Services Branch (D822), Environmental Science Information Center, NOAA, Rockville, MD 20852. Data on fisheries subjects accessioned through NMFS by NODC during the first 6 mo of 1978. Drift bottle, northwestern Gulf of Mexico, February 1962 to December 1963. Fifty data sheets in manuscript form. By R. F. Temple and John R. Martin, Gulf Fisheries Center, NMFS. Ref: NAPIS 78-0035. This material is available from the National Oceanographic Data Center (D7514), National Oceanic and Atmospheric Administration, Washington, DC 20235. 715 ERRATA Fishery Bulletin. Vol. 76, No. 2 Fletcher, R. Ian., "Time-dependent solutions and efficient parameters for stock-production models," p. 377-388. 1) Page 377, right column, line 4, correct line to read: growth rate/? and B^^., Graham's formula for latent 2) Page 378, left column. Equation (2), correct equation to read: P{B) = c^B + c^B\ (2) 3) Page 378, left column, line 14, correct line to read: antecedents of this analysis appear there. 4) Page 378, right column, line 7, correct line to read: by average effort f on the assumption that F = 5) Page 379, left column, the equation that immediately follows Equation (la), correct equation to read: k = Am 'b~ Br 6) Page 381, right column, line 4, correct line to read: Equation (6), B = 0,F =Fj and B =B,. If we now 7) Page 381, right column, line 28, correct line to read: F = 2mlB~^\ stock size Bft) -^ p (p being 8) Page 383, right column. Equation (15), correct equation to read: B = ym ~ B — ym B~ FB. (15) 9) Page 383, Figure 5, caption under right figure, line 3, correct line to read: Pniax in Equation (12)]. 10) Page 385, left column, line 14, correct line to read: then Bit)-^p and Y-*m, irrespective of initial con- 716 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 i mailing address, including Zip code. The abstract should not exceed one double- spaced page. In the text, Fishery Bulletin style, for the most part, follows that of the U.S. Government Printing Office Style Manual. Fish names follow the style of the American Fisheries Society Special Publi- cation No. 6, A List of Common and Scientific Names of Fishes from the United States and Canada, Third Edition, 1970. 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 photog- raphy to 5% inches (for single-column figures, allowing for 50% reduction in printing), or to 12 inches (for double-column figures). The maximum height, for either width, is 14 inches. Photo- graphs should be printed on glossy paper. Do not send original drawings to the Scientific Editor; if they, rather than the photographic re- ductions, are needed by the printer, the Scientific Publications Office will request them. Each table should start on a separate page. Consistency in headings and format is desirable. Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be numbered sequentially in arable numerals. 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 re- ferred 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 literature cited is the re- sponsibility of the author. Abbreviations of names of periodicals and serials should conform to Bio- logical Abstracts List of Serials with Title Abbrevi- ations. (Chemical Abstracts also uses this system, which was developed by the American Standards Association.) Common abbreviations and symbols, such as mm, m, g, ml, mg, °C (for Celsius), %, %o and so forth, should be used. Abbreviate units of mea- sure 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 pap)er. Please triple space above headings. We would rather receive good duplicated copies of manuscripts than car- bon copies. The sequence of the material should be: TITLE PAGE ABSTRACT TEXT LITERATURE CITED APPENDIX TEXT FOOTNOTES TABLES (Each table should be numbered with an arabic numeral and heading provided) LIST OF FIGURES (Entire figure legends) FIGURES (Each figure should be numbered with an arabic numeral; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. Jay C. Quast, Scientific Editor Fishery Bulletin Northwest and Alaska Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA P.O. Box 155, Auke Bay, AK 99821 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 JEWETT, STEPHEN C. Summer food of the Pacific cod, Gadus macrocephalus , near Kodiak Island, Alaska 700 LEMING, THOMAS D., and HILLMAN J. HOLLEY. A computer software system for optimizing survey cruise tracks 706 Notices NOAA Technical Reports NMFS published during the first 6 mo of 1978 715 Data on fisheries subjects accessioned through NMFS by NODC during the first 6 mo of 1978 715 ■is GPO 796-049 ^ V Fishery Bulletin ^o ^^ATES O^ ^ Vol. 76, No. 4 SINDERMANN, CARL J. Pollution-associated diseases and abnormalties of fish and shellfish: a review 717 FROST, BRUCE W., AND LAWRENCE E. McCRONE. Vertical distribution, diel vertical migration, and abundance of some mesopelagic fishes in the eastern subarc- tic Pacific Ocean in summer 751 SMITH, T. D., and T. POLACHECK. Analysis of a simple model for estimating historical population sizes 771 GORE, ROBERT H. Larval development of Galathea rostrata under laboratory conditions, with a discussion of larval development in the Galatheidae (Crustacea Anomura) 781 LENARZ, WILLIAM H., and JAMES R. ZWEIFEL. A theoretical examination of some aspects of the interaction between longline and surface fisheries for yellowfin tuna, Thunnus albacares 807 PARRACK, MICHAEL L. Aspects of brown shrimp, Penaeus aztecus, growth in the northern Gulf of Mexico 827 EHRLICH, KARL F., J. MYRON HOOD, GERALD MUSZYNSKI, and GERALD E. McGOWEN. Thermal behavioral responses of selected California littoral fishes 837 BROWN, B. E., J. A. BRENNAN, and J. E. PALMER. Linear programming simula- tions of the effects of bycatch on the management of mixed species fisheries off the northeastern coast of the United States 851 Notes ROBERTS, JOHN L., and JEFFREY B. GRAHAM. Effect of swimming speed on the excess temperatures and activities of heart and red and white muscles in the mackerel. Scomber japonicus 861 EHRLICH, KARL F., JOHN S. STEPHENS, GERALD MUSZYNSKI, and J. MYRON HOOD. Thermal behavioral responses of the speckled sanddab, Citharichthys stigmaeus: laboratory and field investigations 867 SHULTZ, CYNTHIA D., and BERNARD M. ITO. Mercury and selenium in blue marlin, Makaira nigricans, from the Hawaiian Islands 872 PERSCHBACHER, PETER W., and FRANK J. SCHWARTZ. Recent records of Callinectes danae and Callinectes marginatus (Decapoda: Portunidae) from North Carolina with environmental notes 879 (Continued on back cover) Seattle, Washington U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Richard A. Frank, Administrator Terry L. Leitzell, Assistant Administrator for Fislieries NATIONAL MARINE FISHERIES SERVICE Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as docimients through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. EDITOR Dr. Jay C. Quast Scientific Editor, Fishery Bulletin Northwest and Alaska Fisheries Center Auke Bay Laboratory National Marine Fisheries Service, NOAA P.O. Box 155. Auke Bay, AK 99821 Editorial Committee Dr. Elbert H. Ahlstrom Dr. Merton C. Ingham National Marine Fisheries Service National Marine Fisheries Service Dr. Bruce B. Collette Dr. Reuben Lasker National Marine Fisheries Service National Marine Fisheries Service Dr. Edward D. Houde Dr. Jerome J. Pella University of Miami National Marine Fisheries Service Dr. Sally L. Richardson Gulf Coast Research Laboratory Kiyoshi G. Fukano, Managing Editor The Fishery Bulletin is published quarterly by Scientific Publications Office. National tVIanne Fisheries Service. NOAA. Room 450, 1107 NE 45th Street, Seattle. WA 98105. Controlled arculation postage paid at Tacoma, Wash. Although the contents have not been copyrighted and may be repnnted freely, reference to source is appreciated. The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the public business required by 'aw of this Department Use of funds for printing of this periodical has tjeen approved by the Director of the Office of Management and Budget through 31 December 1978. Fishery Bulletin CONTENTS Vol. 76, No. 4 SINDERMANN, CARL J. Pollution-associated diseases and abnormalties of fish and shellfish: a review 717 ^^ FROST, BRUCE W., AND LAWRENCE E. McCRONE. Vertical distribution, diel vertical migration, and abundance of some mesopelagic fishes in the eastern subarc- tic Pacific Ocean in summer 751 SMITH, T. D., and T. POLACHECK. Analysis of a simple model for estimating historical population sizes 771 -Vr~ GORE, ROBERT H. Larval development of Galathea rostrata under laboratory conditions, with a discussion of larval development in the Galatheidae (Crustacea Anomura) 781 "^ LENARZ, WILLIAM H., and JAMES R. ZWEIFEL. A theoretical examination of some aspects of the interaction between longline and surface fisheries for yellowfin tuna, Thunnus albacares 807 PARRACK, MICHAEL L. Aspects of brown shrimp, Penaeus aztecus, growth in the northern Gulf of Mexico 827 -V EHRLICH, KARL F., J. MYRON HOOD, GERALD MUSZYNSKI, and GERALD E. McGOWEN. Thermal behavioral responses of selected California littoral fishes 837 BROWN, B. E., J. A. BRENNAN, and J. E. PALMER. Linear programming simula- tions of the effects of bycatch on the management of mixed species fisheries off the northeastern coast of the United States 851 Notes ROBERTS, JOHN L., and JEFFREY B. GRAHAM. Effect of swimming speed on the excess temperatures and activities of heart and red and white muscles in the mackerel. Scomber Japonicus 861 EHRLICH, KARL F. , JOHN S. STEPHENS, GERALD MUSZYNSKI, and J. MYRON HOOD. Thermal behavioral responses of the speckled sanddab, Citharichthys stigmaeus: laboratory and field investigations 867 SHULTZ, CYNTHIA D., and BERNARD M. ITO. Mercury and selenium in blue marlin, Makaira nigricans, from the Hawaiian Islands 872 PERSCHBACHER, PETER W., and FRANK J. SCHWARTZ. Recent records of Callinectes danae and Callinectes marginatus (Decapoda: Portunidae) from North Carolina with environmental notes 879 (Continued on next page) Seattle, Washington 1979 For sale by the Superintendent of Documents. U.S. Government Printing Office. Washington. DC 20402— Subscription price per year: $12.00 domestic and $15.00 foreign. Cost per single issue: $3.00 domestic and $3.75 foreign. Contents-continued STOUT, VIRGINIA F., and F. LEE BEEZHOLD. Analysis of chlorinated hydrocar- bon pollutants: a simplified extraction and cleanup procedure for fishery products . 880 RENSEL, JOHN E., and EARL F. PRENTICE. Growth of juvenile spot prawn, Pandalus platyceros, in the laboratory and in net pens using different diets .... 886 LAURENCE, GEOFFREY C. Larval length-weight relations for seven species of northwest Atlantic fishes reared in the laboratory 890 CRADDOCK, DONOVAN R. Effect of thermal increases of short duration on sur- vival of Euphausia pacifica 895 MAY, ROBERT C, GERALD S. AKIYAMA, and MICHAEL T. SANTERRE. Lunar spawning of the threadfin, Polydactylus sexfilis, in Hawaii 900 PEEBLES, JOHN B. The roles of prior residence and relative size in competition for shelter by the Malaysian prawn, Macrobrachium rosenbergii 905 COLTON, JOHN B., JR., WALLACE G. SMITH, ARTHUR W. KENDALL, JR., PETER L. BERRIEN, and MICHAEL P. FAHAY. Principal spawning areas and times of marine fishes. Cape Sable to Cape Hatteras 911 WADE, LAWRENCE S., and GARY L. FRIEDRICHSEN. Recent sightings of the blue whale, Balenoptera musculus, in the northeastern tropical Pacific 915 BAGLIN, RAYMOND E., JR. Sex composition, length- weight relationship, and reproduction of the white marlin, Tetrapturus albidus, in the western North Atlan- tic Ocean 919 BECKER, C. DALE, and DENNIS D. DAUBLE. Records of piscivorus leeches (Hirudinea) from the central Columbia River, Washington State 926 SMIGIELSKI, ALPHONSE S. Induced spawning and larval rearing of the yellow- tail flounder, Limanda ferruginea 931 YOUNG, DAVID R., and TSU-KAI JAN. Trace metal contamination of the rock scallop, Hinnites giganteus , near a large southern California municipal outfall . 936 INDEX, VOLUME 76 941 Vol. 76, No. 3 was published on 16 November 1978. 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. POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES OF FISH AND SHELLFISH: A REVIEW Carl J. Sindermann' ABSTRACT The relationship of disease and environmental stress is becoming increasingly well established with time. Human activities — particularly those that result in chemical additions to the coastal/estuarine environment — have increased the potential stresses on fish and shellfish inhabiting those areas. Circumstantial evidence for associations of pollutants with certain fish and shellfish diseases and abnormalities is accumulating. This paper attempts to review and evaluate existing information about associations of diseases and marine environmental degradation. Emphasis has been placed on: diseases caused by contaminant stress and related facultative pathogens; stress-provoked latent infections; environmentally induced abnormalities; genetic abnormalities associated with mutagenic and other properties of contaminants; experimentally induced lesions; contaminant effects on resistance and immune responses; and pollutant-parasite interactions. There are several diseases, particularly fin erosion and ulcers in fish and shell disease in crustaceans, for which a relationship with pollution seems evident, and there are a number of other diseases or abnormalities (such as certain neoplasms and skeletal anomalies) for which a relationship with pollution is indicated. Furthermore, there is some evidence that certain latent viral infections may be provoked into patency by environmental stress. Disease is a constant concomitant of life for any species, normally removing individuals from the population continuously. Marine animals are, of course, subject to a wide spectrum of diseases of infectious or noninfectious etiology ("disease" can be defined in the broad sense as "any departure from normal structure or function of an animal" or as "the end result of interaction between a noxious stimulus and a biological system"). Disfunction and death due to the activity of in- fectious agents constitute the narrower, but often predominant concept of disease. Infectious diseases — caused by viruses, bacteria, fungi, pro- tozoa, and other pathogenic organisms — are usu- ally prime suspects in searches for causes of mor- talities, often to the exclusion of other possible causes. Noninfectious diseases include such phenomena as environmentally induced skeletal anomalies, genetic abnormalities, physiological malfunctions caused by chemical environmental factors, metabolic disorders resulting from nutri- tional deficiencies, many forms of neoplasia, and a host of others (Sparks 1972). In many instances, it is probably the combination of an infectious agent 'Northeast Fisheries Center Sandy Hook Laboratory, Na- tional Marine Fisheries Service, NOAA, Highlands, NJ 07732. and environmental stress that eventually causes mortality. The distinction between "infection" and "dis- ease" must be kept in mind. Most organisms are constantly hosts to potentially pathogenic micro- organisms, but disease results from imbalance of the interactive system which includes virulence of the pathogen, resistance of the host, and effects of environmental stresses. Infectious disease usually exists in an enzootic form, weakening or disabling individuals and rendering them more susceptible to predators or other environmental stresses. Occasionally, though, epizootics and mortalities comparable to the great plagues of the Middle Ages may sweep through animal populations. In marine species we have seen such massive epizootics result in the great herring mortalities of the mid-1950's in the Gulf of Saint Lawrence (Sindermann 1958), and the extensive oyster mortalities of the 1960's in the Middle Atlantic states (Sindermann 1968). These epizootics are triggered by a complex interplay of pathogen, environment, and host pop- ulation. Considering only the environmental as- pects of such outbreaks, any departure from nor- mal conditions produces a degree of stress on the population, and may contribute to an increase in prevalence of a pathogen, or of facultative invad- Manuscript accepted Mav 1978. FISHERY BULLETIN; VOL. 76, NO. 4, 1979. 717 ers. Some of these environmental factors are dras- tic changes in temperature, lack of adequate food, or overcrowding. Resistance of the host animal to the disease is, of course, intimately related to these stresses (Snieszko 1974). Environmental stresses have been implicated in a number of fish and shellfish diseases, but are difficult to quantify. Even a definition of stress can be elusive. Selye ( 1950, 1952) defined stress as the sum of all the physiological responses by which an animal tries to maintain or reestablish a normal metabolism in the face of a physical or chemical force. Brett ( 1958) defined it as "A state produced by any environmental or other factor which ex- tends the adaptive responses of an animal beyond the normal range, or which disturbs the normal functioning to such an extent that, in either case, the chances of survival are significantly reduced." Another definition which identifies stress as the product and not the cause of homeostatic change is thatofEschetal. (1975): "Stress is the effect of any force which tends to extend any homeostatic or stabilizing process beyond its normal limit, at any level of biological organization." Human activity has introduced or has increased environmental stresses for fish in estuarine and coastal waters. We have, for instance, added pes- ticides and other synthetic chemicals which can, even in low concentrations, drastically affect the physiology of fish and shellfish, and with which the species may have had no previous evolution- ary experience. We have added heavy organic loads, in the form of sewage sludge and effluents, which can produce anaerobic or low-oxygen envi- ronments and which are often accompanied by other contaminants such as heavy metals, that can interfere with enzymes of the fish and the food organisms they consume. During the past decade, several diseases and abnormalities offish and shellfish have been de- scribed that seem associated with pollutant stres- ses. These can be categorized and discussed as: 1. Diseases caused by contaminant stress and related pathogens; 2. Stress-provoked latent infections; 3. Environmentally induced abnormalities; 4. Genetic abnormalities associated with mutagenic and other properties of contam- inants; 5. Experimentally induced lesions; 6. Contaminant effects on resistance and im- mune response; and FISHERY BULLETIN: VOL. 76, NO. 4 7. Pollutant-parasite interactions. In the first and second categories a synergistic activity of chemical contaminants (or other form of pollutant stress) and an infectious agent seems to be a plausible explanation for at least some of the observed effects. In categories three and four, it is sometimes difficult to determine conclusively whether environmental contaminants act directly on target tissues or biochemical pathways, or if the genetic material is first affected, with subsequent changes in structure and/or function. During the past several years there have been signs of increasing interest in relationships be- tween marine fish and shellfish diseases and en- vironmental pollution. Several conferences have been held recently, including the 1974 Symposium on Tumors in Aquatic Animals, held in Cork, Ire- land; the 1975 Symposium on Sublethal Effects of Pollution on Aquatic Organisms, held as part of the 13th Pacific Science Congress in Vancouver, B.C.; and the 1976 Conference on Aquatic Pollu- tants and Biological Effects with Emphasis on Neoplasia, held in New York. The amount of rel- evant literature available for consideration within the title "pollution-associated diseases and abnormalities of fish and shellfish" is somewhat overwhelming. Even the list of books containing pertinent material is impressive (Dawe and Harshbarger 1969; Snieszko 1970; Ruivo 1972; Vernberg and Vernberg 1974; Koeman and Strik 1975; Ribelin and Migaki 1975; Dawe et al. 1976; Lockwood 1976; Kraybill et al. 1977; Vernberg et al. 1977). Additionally, significant recent reviews have appeared, for example, Rosenthal and Alder- dice (1976) and Mclntyre^. This paper attempts to summarize the present state of knowledge about possible associations of fish and shellfish diseases (infectious and nonin- fectious) with estuarine and coastal pollution. Much of the evidence for such associations is still circumstantial and is presented as such. The orig- inal literature on this subject, as for any pollution-related subject, is voluminous. The ref- erences cited here constitute only a small but, I hope, a representative fraction of the published information available. It should also be pointed out here that this paper does not consider ^Mclntyre, A. D. (Convenor). 1976. ICES working group on pollution baseline and monitoring studies in the Oslo Commis- sion and ICNAF areas. Report of the subgroup on the feasibility of effects monitoring. Int. Counc. Explor. Sea, Doc. CM1976/ E:44, 36 p. 718 SINDERMANN: POLLUTION-ASSOCIATED DKSEASES AND ABNORMALITIES physiological and behavioral disorders, which might be included in a broad definition of disease. Finally, in these introductory comments, it should be noted that to make any firm association of a disease with environmental pollution there are several basic requirements: 1) knowledge of the history of occurrence of the disease in a par- ticular species in the geographic area of concern; 2) knowledge of the history of occurrence and levels of particular pollutants in that area; 3) a review of the biology, life history, and occurrence of the disease in other areas, in other species, and under different environmental conditions; 4) an intensive baseline survey of the current disease and pollution situation, with attention to statisti- cal reliability of sampling; 5) laboratory and field experimentation with the principal objective of reproducing the disease by exposure to known levels of contaminants; and 6) resurveys of the disease and pollution levels over several years, looking for changes or trends. As will become ap- parent in this paper, these requirements have been fully satisfied for few if any of the diseases discussed. DISEASES CAUSED BY CONTAMINANT STRESS AND RELATED FACULTATIVE PATHOGENS Fin Erosion Probably the best known but least understood disease offish from polluted waters is a nonspecific condition known as "fin rot" or "fin erosion" (Fig- ures 1, 2), a syndrome which seems rather clearly associated with degraded estuarine or coastal en- vironments. Fin rot has been reported from the New York Bight (Mahoney et al. 1973; Ziskowski and Murchelano 1975; Murchelano 1975), California (Young 1964; Southern California Coastal Water Research Project-"^; Mearns and Sherwood 1974), Puget Sound (Wellings et al. 1976), Biscayne Bay and Escambia Bay in Florida (Couch 1974a; Sindermannetal. 1978), the Gulf of Mexico ( Overstreet and Howse 1977 ), the Irish Sea (Perkins et al. 1972), and the Japanese coast (Nakai et al. 1973). Fin rot seems to occur in at least two types: one ^Southern California Coastal Water Research Proj- ect. 1973. The ecology of the Southern California Bight: Im- plications for water quality management. Ref. No. SCCWRP TR 104, El Segundo, Calif. in bottom fish, where damage to fins seems site- specific and related to direct contact with con- taminated sediments, and another in pelagic nearshore species, characterized by more generalized erosion, but with predominant in- volvement of the caudal fin. Recent quantitative surveys along the Middle At- lantic coast have disclosed high prevalence (up to 38*^ ) of fin rot in samples of trawled marine fishes from the New York Bight. Thus far, 22 affected species have been found. While bacteria of the genera Vibrio, Aeromonas, and Pseudomonas were frequently isolated from abnormal fish, a definite bacterial etiology has not been estab- lished. Fin rot disease was significantly more abundant in the New York Bight Apex, the area of greatest environmental damage, than in any com- parable coastal area from Block Island, R.I., to Cape Hatteras, N.C. (Murchelano and Ziskowski 1976). An association between high fin rot preva- lence and high coliform counts in sediments is emerging (Mahoney et al. 1973), as is an associa- tion between high fin rot prevalences and high heavy metal levels in sediments (Carmody et al. 1973). The disease signs can be produced experi- mentally by exposure offish to polluted sediments. Fin erosion has also been observed in striped bass, Morone saxatilis, overwintering in heated efflu- ents of power plants in the Middle Atlantic States. The histopathology of fin erosion in winter flounder, Pseudopleuronectes americanus, from the New York Bight was examined by Murchelano (1975). Significant descriptive findings were epidermal hyperplasia accompanied by dermal fibrosis, hyperemia, and hemorrhage. Bacterial in- fections were not found, nor was pronounced in- flammatory response. However, reference was made to acute fin lesions seen in summer flounder, Paralichthys dentatus, in which bacteria were readily demonstrable. The absence of pronounced inflammatory response in either species of floun- der led Murchelano to suggest that the necrotic process is not primarily microbial and that ac- tivities of a chemical irritant may be involved. Another histopathological and bacteriological study of fin rot in winter flounder from Narragan- sett Bay, R.I., by Levin et al. (1972) described acute ulcerative lesions as well as fln erosion, thought to be produced by Vibrio anguillarum. Acute inflammatory response was observed, and ulcerations were reproduced in fish exposed ex- perimentally to V . anguillarum isolates. It is pos- sible that several poorly defined disease entities or 719 FISHERY BULLETIN: VOL. 76. NO. 4 Figure l. — Site-specific fin erosion concentrated in the midportion of fins in winter flounder (anterior dorsal fin is folded over in this picture). Note melanism in areas of erosion. (Photograph courtesy of J. O'Reilly, Northeast Fisheries Center Sandy Hook Laboratory, NMFS, NOAA, Highlands, N.J.) generalized disease signs (one of which is fin ero- sion) may be responsible for the disparate nature of histopathological findings in this report, as compared with those of Murchelano ( 1975). Fin rot, with associated mortalities, was re- ported by Couch and Nimmo (1974b) in Atlantic croaker, Micropogon undulatus, and spot, Leio- stomus xanthurus, from Escambia Bay, Fla. The disease syndrome and mortalities were observed for several years during periods of high tempera- ture and low dissolved oxygen. Escambia Bay has been polluted by the PCB (polychlorinated biphenyl), Aroclor^ 1254, for a number of years (Duke et al. 1970). Information from southern California (South- ern California Coastal Water Research Project, see footnote 3) also indicates an association of fin rot with degraded habitats; relevant statements are: "The incidence of fin erosion was high in areas with high concentrations of waste water con- stituents in the sediments . . . ." "Although there is a definite association between fin erosion and waste water discharges, the causal factors are un- known." "Nearly half of the 72 species caught off the Palos Verdes Peninsula were affected with this syndrome" (eroded fins). It is interesting that a histopathological study of fin erosion in Dover sole, Microstomus pacificus, from the California coast (Mearns and Sherwood 1974; Klontz and Bendele^) produced findings similar to those of Murchelano (1975) — hyperplasia, fibrosis, ab- sence of inflammation, and absence of microbial infection. Some species either seem more resistant to fin erosion or are exposed differentially to toxic sub- stances in water or sediments. A recent study by Wellings et al. (1976) in a heavily polluted arm of Puget Sound (the Duwamish River) in which over 6,000 fish of 29 species were examined, disclosed fin erosion only in starry flounder, Platichthys stel- latus, and English sole, Parophrys vetulus. Av- erage incidences were 8 and 0.5% respectively. Histopathological findings were similar to those for east coast and California flatfishes — epidermal hyperplasia, fibrosis, resorption of fin rays, aggre- gation of melanophores, mucus cell changes, and absence of bacterial invasion. The authors de- scribed briefly what may be highly relevant obser- vations of liver pathology in starry flounder from the area where fin erosion was common. His- topathology included increased fat deposition in hepatic cells, fibrosis, and vascular distension. Recent Japanese publications have mentioned fin erosion in fish from polluted bays. Nakai et al. ■"Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ^Klontz, G. W..andR. A.Bendele. 1973. Histopathological analysis of fin erosion in southern California marine fishes. Southern Calif Coastal Water Res. Proj., El Segundo, Calif, Rep. TM203, 8 p. 720 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES Figure 2. — Generalized fin erosion in weakUsh, Cynoscion regalis, (above) and in summer flounder (belowl. Note that in the weakfish the anal, caudal, and pelvic fins are eroded, while the dorsal fins are not usually damaged. In contrast the summer flounder shows erosion of wide areas of the fin fringes. ( 1973) found that as many as 60^^ of all stargazer, Uranoscopusjaponicus, sampled from Suruga Bay had evidence of disintegration of caudal and pec- toral fins. Six other species also had abnormal fins. An increase in occurrence of fin erosion and other epidermal lesions (ulcers and lymphocystis) in flatfish from the Irish Sea since 1970 was re- ported by Perkins et al. (1972). Fin damage, unknown before 1970, was observed in plaice, Pleuronectes platessa, and dab, Limanda limanda, taking the form of erosion or total loss of caudal and lateral fins. Ulcers were described that "did not have the typical appearance of bacterial ulcers . . . ." The authors pointed to ocean dump- ing of toxic wastes, particularly of PCB's, as a possible factor contributing to observed preva- lences of epidermal lesions, but no clear relation- ship was demonstrated. Another study conducted in the Irish Sea in 1972 (Shelton and Wilson 1973) did not identify fin erosion in plaice or dab, but did find a low incidence of "healed fin damage (proba- bly caused by previous capture and rejection or by passage through the cod-end mesh)." The possible role of environmental chemical contamination in the etiology of fin erosion emerges more clearly as additional studies are reported. Fish from the New York Bight, reported in studies by Mahoney et al. (1973), Murchelano (1975), and Ziskowski and Murchelano (1975), exist in a highly contaminated area, with chemi- cals such as heavy metals and petroleum residues in sediments far above background levels. In California, McDermott and Sherwood*^ fount DDT to be significantly higher in fish with fin erosion, and PCB levels slightly higher in such fish than in normal individuals. Both contaminants were sig- nificantly higher in Palos Verdes fish than in fish ''McDermott, D. J., and J. Sherwood. 1975. Annual re- port. Dep. Fish. Mar. Fish. Program, Coastal Water Res. Proj., El Segundo, Cahf , p. 37. 721 FISHERY BULLETIN: VOL. 76, NO. 4 from a distant control area ( Dana Point). Wellings et al. (1976) found abnormally high concentra- tions of PCB's in English sole and starry flounders from the Duwamish River in Washington. Several authors have postulated that fin erosion in flatfish may be initiated by direct contact of tissues with contaminated sediments. Mearns and Sherwood (1974) and Sherwood and Mearns (1977), for example, suggested that toxic sub- stances (sulfides, heavy metals, chlorinated hy- drocarbons, etc.) could remove or modify the pro- tective mucus coat and expose epithelial tissues to the chemicals. Sherwood and Bendele^ reported that Dover sole from the California coast with severe fin erosion produced much less mucus than normal fish. It seems quite likely that the "fin erosion" syn- drome in fish includes chemical stress, possibly acting on mucus and/or epithelium; stress result- ing from marginal dissolved oxygen concentra- tions, possibly enhanced by a sulfide-rich envi- ronment; and secondary bacterial invasion in at least some instances. Some recent experimental information tends to support this hypothesis. A series of experiments at the Gulf Breeze (Fla.) Environmental Research Laboratory of the U.S. Environmental Protection Agency, using the spot, resulted in experimental production of fin rot dis- ease following exposure to 3-5 ^tg/l of Aroclor 1254 (Couch 1974a). Mortalities of up to 80% were re- ported. Minchew and Yarbrough (1977) exposed Mugil cephalus in brackish water ponds ( 12%o) to 4-5 ppm crude oil and found that fin erosion developed in most of the exposed fish within 6-8 days. Lesions were often hemorrhagic, and a tentative Vibrio sp. was isolated consistently from surfaces of diseased fish, but was rarely found systemically. Fin regen- eration characterized most experimental fish 2 mo after exposure. This experiment should be re- peated and extended. Experimental induction of fin erosion has fol- lowed exposure to several other contaminant chemicals. Chronic exposure of fingerling rainbow trout, Salmo gairdneri, to lead caused a variety of grossly visible abnormalities, including fin ero- sion (Davies and Everhart^); and chronic exposure of minnows (Phoxinus phoxinus) to zinc and cad- ■'Sherwood, M. J., andR. A. Bendele. 1975. Mucous produc- tion in Dover sole. Annu. rep., Coastal Water Res. Proj., El Segundo, Calif., p. 51. 8Davies, P. H.,andJ.H. Everhart. 1973. Effects of chemi- cal variations in aquatic environments. III. Lead toxicity to mium resulted in similar abnormalities (Bengtsson 1974, 1975). A recent report by Overstreet and Howse ( 1977) pointed to fin erosion and other abnormalities as indicators of gradually increasing pollution stress on the Mississippi gulf coast. Among other disease conditions noted by Overstreet and Howse was "red sore," characterized by hemorrhagic lesions beneath scales, occasional hyperplasia, and ac- companying ciliate (Epistylis sp.) infestation of the body surfaces. The authors indicated that red sores now occur in many of the fish in some fresh- water and low salinity areas of the gulf coast of Mississippi, a striking similarity to recent obser- vations in Biscayne Bay, Fla., where many fish of many species now exhibit hemorrhagic lesions be- neath the scales, a condition which was unknown a decade ago (Sindermann 1976). Red sores and associated mortalities have also been described by Rogers (1970, 1972) and Esch et al. (1976) from centrarchid fishes in freshwater reservoirs of the southeastern United States. The disease condition in freshwater seems clearly related to Epistylis infestation, probably abetted by secondary bacte- rial infections, particularly by Aeromonas, al- though there is still some question about which organism is the primary invader. It seems likely that generalized disease signs, such as fin rot and red sores (and probably other epidermal lesions such as ulcerations, papillomas, and lymphocystis), may be characteristic of fishes resident in degraded habitats, where environmen- tal stresses of toxic chemicals, low dissolved oxy- gen, and high microbial populations exist. The extent and nature of these external manifesta- tions are probably variable with resistance of the particular species and the extent and nature of environmental degradation. Ulcers Next to fin erosion, probably the commonest ab- normality reported from fish taken in polluted waters can be identified as "ulceration of bacterial etiology," even though precise bacterial etiology has not been demonstrated in every case. Where bacterial isolations have been made from ulcer- ated tissue. Vibrio anguillarum has been by far the most predominant organism, with pseudo- monads and aeromonads in lesser abundance. rainbow trout and testing application factor concept. EPA- R3-73-011C, 80p. 722 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES The report on ulcerations and fin rot in winter flounders from Narragansett Bay, by Levin et al. (1972) has been mentioned in the previous section. The acute ulcerative lesions were thought to be caused by V. anguillarum infections, and the ul- cerative phase was reproduced in fish exposed ex- perimentally to cultured V. anguillarum isolates. A more recent report by Robohm and Brown (1977) described systemic bacterial infections and ulcerative lesions of the tail and dorsal muscles in summer flounder from Connecticut waters. A highly pathogenic Vibrio sp. was isolated, and ex- perimental infections were produced by sub- cutaneous inoculation and by seeding holding tanks with bacteria at levels of 360/ml. Ulcers at the inoculation site and subcutaneous hemor- rhages along the bases of fins characterized ex- perimental infections (Figure 3). These observa- tions resemble those of Levin et al. (1972) in winter flounder. Ulcerations, probably of bacterial etiology, have been reported in fish of several species from the Irish Sea. Perkins et al. (1972) and Shelton and Wilson (1973) reported ulcers from European flounders (Platichthys flesus), dab, and plaice. Prevalences were low (1-49^) in most instances. An "ulcer syndrome" in cod, Gadus morhua, from Danish coastal waters has been studied for several years and seems associated with localized areas of severe pollution (Jensen and Larsen 1976, 1977; Larsen and Jensen 1977a, b). Vibrio anguil- larum and anAeromonas species have been impli- cated (S^rensen 1977). Ulcerations or external lesions on fish may, of course, have a number of causes other than bacte- rial infection. They may be due to net damage or other surface abrasions, or to predator attacks. Some protozoa (Myxosporida and Microsporida) can infect muscle or skin tissue and multiply to produce gross cysts. These infections mature to produce many characteristic microscopic spores, and in the process the overlying epidermis may be sloughed, producing ulcers with usually smooth borders (Figure 4). However, it seems to be a Figure 3. — Ulcers and fin erosion in summer flounder produced by experimental inoculation oiVibrio sp. (Photograph courtesy of R. Robohm, Northea.st Fisheries Center Milford Laboratory, NMFS, NOAA, Milford, Conn.) 723 FISHERY BULLETIN; VOL. 76. NO. 4 Figure 4. — Ulcer with smooth margins in Atlantic herring, resulting from infection by the myxosporidan protozoan, Kudoa clupeidae. reasonable generalization that many of the infec- tions that produce grossly visible ulcerations in fish are bacterial, and are often due to pathogens of the genera Vibrio, Pseudomonas, or Aeromonas (Lamoletetal. 1976). Ulceration often begins with scale loss or formation of small papules, followed by sloughing of the skin, exposing the underlying muscles, which may also be destroyed. Bacterial ulcers may have rough or raised irregular mar- gins, and will often be hemorrhagic. Ulcers may or may not be associated with fin erosion. Shell Disease of Crustacea Also associated with badly degraded estuarine and coastal waters is a disease condition in Crus- tacea commonly referred to as "shell disease" or "exoskeletal disease" or "shell erosion." This can be considered in some ways as the invertebrate counterpart of fin erosion. Homarus americanus and rock crabs (Cancer irroratus ) from grossly polluted areas of the New York Bight were found to be abnormal, with ap- pendage and gill erosion a most common sign, by Young and Pearce (1975). Skeletal erosion occur- red principally on the tips of the walking legs, ventral sides of chelipeds, exoskeletal spines, gill lamellae, and around areas of exoskeletal articu- lation where contaminated sediments could ac- cumulate. Gills of crabs and lobsters sampled at the dump sites were usually clogged with detritus, possessed a dark brown coating, contained localized thickenings, and displayed areas of ero- sion and necrosis. Similar disease signs were pro- duced experimentally in animals held for 6 wk in aquaria containing sediments from sewage sludge or dredge spoil disposal sites. Initial discrete areas of erosion became confluent, covering large areas of the exoskeleton, and often parts of appendages were lost. The chitinous covering of the gill fila- ments was also eroded, and often the underlying tissues became necrotic. Dead and moribund crabs and lobsters have been reported on several occasions by divers in the New York Bight Apex, and dissolved oxygen con- centrations near the bottom during the summer often approach zero (Pearce 1972; Young 1973). Low oxygen stress, when combined with gill foul- ing, erosion, and necrosis, could readily lead to mortality. In a related study, Gopalan and Young (1975) examined "shell disease" in the caridean shrimp, Crangon septemspinosa, an estuarine and coastal food chain organism common on the east coast of North America and important in the diets of bluefish, weakfish, flounders, sea bass, and other economic species. Examinations of samples of Crangon from the New York Bight disclosed high prevalences (up to 15*7^ ) of eroded appendages and blackened erosions of the exoskeleton. The disease condition was only rarely observed at other col- lecting sites (Beaufort, N.C., and Woods Hole, Mass.). Histological examination of diseased specimens produced findings similar to those of 724 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES Young and Pearce ( 1975) with crabs and lobsters. All layers of the exoskeleton were eroded; affected portions were brittle and easily fragmented; cracking and pitting of calcified layers occurred; and underlying tissues were often necrotic. Laboratory experiments using seawater from the highly polluted inner New York Bight resulted in appearance of the disease in 509f of individuals. Erosion was progressive, crippled individuals were cannibalized, and eroded segments of append- ages did not regenerate after ecdysis. No disease signs developed in control animals held in arti- ficial seawater. A German study of the effects of industrial wastes on the brown shrimp, Crangon crangon (Schlotfeldt 1972), disclosed high prevalence of so-called "black spot disease," with signs very similar to those seen in C. septemspinosa from the New York Bight. Juvenile and adult shrimps from the Fohr Estuary had black areas of erosion on the carapace and appendages, with necrosis of under- lying tissues, and, frequently, missing terminal segments of appendages. The disease condition varied in prevalence seasonally, with a peak of 8.9^f in summer. Lesions persisted and worsened after ecdysis, and experimental exposure to deter- gent accelerated the course of the disease. Shell disease of Crustacea has been observed in many species and under many conditions, both natural and artificial (Rosen 1970; Sindermann 1970). Actual shell erosion seems to involve activ- ity of chitinoclastic bacteria, with subsequent secondary infection of underlying tissue by facul- tative pathogens. Initial preparation of the exo- skeletal substrate by mechanical, chemical, or mi- crobial action probably is significant; thus high bacterial populations and the presence of contam- inant chemicals in polluted environments, as well as extensive detrital and epibiotic fouling of gills, could combine to make shell disease a common phenomenon and a significant mortality factor in crustaceans inhabiting degraded environments. There is much room for study in this cloudy territory at the boundary between infectious and noninfectious disese processes, as exemplified by fin and shell erosion. This is the area where en- vironmental stress and facultative microor- ganisms exert their impacts; where high bacterial populations in eutrophic waters interact with ex- posed, or injured, or chemically modified surface membranes; where epibiotic fouling organisms can assume pathogenic roles; and where nonspecific lesions such as fin rot and skeletal erosions can occur in epizootic proportions. Lymphocystis While fin erosion, ulcers, and shell disease seem to have reasonable associations with degraded en- vironments, it is difficult to find additional good examples in the category of "Diseases caused by facultative pathogens." Probably the most likely candidate (in an obviously poor field) would be lymphocystis, a virus disease which causes ex- treme hypertrophy of fibroblast cells in a large number of freshwater and marine fishes, and which has been postulated to be associated with environmental stresses. Perkins et al. (1972) found in a 1971 survey that three diseases — lymphocystis, epidermal ulcers, and fin erosion — were abundant in plaice and dab from the Northeast Irish Sea. Lymphocystis infection levels in individual trawl catches ranged from 0 to 2b^( in plaice and from 0 to 17^^ in dab. The au- thors pointed out that the Irish Sea has been used recently for dumping of toxic wastes, particularly PCB's, but their concluding statement is ". . . there is insufficient evidence to be certain whether the increased incidence of the diseases noted in 1971 is the result of an outbreak of epidemics of purely biological origin or if the dumping of toxic wastes is responsible." Another survey of lymphocystis in the Irish Sea, this one in 1972, was reported by Shelton and Wilson ( 1973). They found lymphocystis to be the most abundant of observable pathological condi- tions, with highest prevalence ( 14.6*^ ) in flounder, Platichthys flesus, and lesser prevalences in other flatfish (1.97f in plaice and 1.1% in dab). Unlike Perkins et al. ( 1972), Shelton and Wilson consid- ered recent pollution of the Northeast Irish Sea to be the least likely explanation for high .levels of lymphocystis — pointing out that the disease has been known from that area for 70 yr, having been described early in the century by Woodcock ( 1904) and Johnstone ( 1905) from flounders taken in the Irish Sea. Van Banning ( 197 1 ) studied lymphocys- tis in North Sea plaice (Figure 5) and also con- cluded that pollution was not a likely cause of high prevalences. A recent lymphocystis epizootic with over 50% prevalence was reported from flatfish in the North Sea by Mann (1970) and earlier epizootics have occurred in Europe (Weissenberg 1965). Temple- man (1965) reported an epizootic in American 725 FISHERY BULLETIN; VOL^ 76. NO. 4 Figure 5. — Lymphocystis in European plaice, Pleuronectes platessa. (Photograph courtesy of P. Van Banning, Rijkinstituut voor Visserijonderzoek, IJmuiden, Netherlands.) plaice, Hippogiossoides platessoides, from the Grand Banks of Newfoundland. He suggested sev- eral possible explanations for the outbreak, in- cluding the possibility that the disease is enzootic in the population and may increase in intensity periodically. Earlier, Awerinzew (1911) found an- nual lymphocystis prevalences of 11% in P. flesiis from the Murmansk coast, and Nordenberg ( 1962) found infections as high as 12% in the same species from the Oresund, with some indication of higher prevalence in the warmer months of the year. None of these outbreaks seems to have any ap- parent association with environmental contami- nation. Lymphocystis has been reported recently in Baltic herring (C/;//xY/ harengus var. membras ) by Aneer and Ljungberg (1976). Of the 2,629 indi- viduals examined, 14 had gross signs of the dis- ease. The authors pointed out that a number of infections were slight and might easily have been overlooked. It is quite likely that this is the case with other species also. The presence of lymphocystis cells in the viscera of herring was noted by Aneer and Ljungberg, and there are several other reports of systemic lym- phocystis infections, particularly that of Dukes and Lawler (1975) in which lymphocystis cells were found in and behind the eyes and in the kidney, spleen, liver, heart, ovaries, and mesen- teries of silver perch, Bardiella chtysiira, from the Mississippi coast. Lymphocystis has also been recognized in 4.3% ofyellowfin sole, Limanda aspera, sampled in the Bering Sea by Alpers et al. (1977a) and in 68% of winter flounder sampled in 1975 from Casco Bay in the Gulf of Maine (Murchelano and Bridges 1976). Despite inconclusive attempts to relate lym- phocystis epizootics in flatfish to specific environ- mental factors, including pollutants, there are re- cent observations of the disease in fishes of the Gulf of Mexico that reopen the issue. Christmas and Howse (1970) found lymphocystis in Atlantic croaker and sand seatrout, Cynoscion arenarius, from the Mississippi coast of the Gulf of Mexico and observed that "The pollution load was much gi'eater in estuarine systems where lymphocystis was encountered." However, only 12 infected fish were found in a 10-mo trawling survey with monthly collections at 35 stations, which is not overwhelming evidence for a relationship of the disease to pollution. In a later study, Edwards and Overstreet (1976) reported marked increases in lymphocystis incidences in Atlantic croakers from the Mississippi coast, with as high as 50% infected fish in some trawl collections. Increased preva- lences of another strain of lymphocystis were also observed in silver perch. In a later paper Over- 726 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES street and Howse (1977) stated that (with refer- ence to the silver perch strain) "prevalence appears to relate to rainfall, suggesting that toxi- cants, salinity, or enriched water could play a major role in infections." Lymphocystis in striped bass, Morone sa.xatilis, on the U.S. east coast seems to have some tenuous association with heated effluents. Recent unpub- lished observations by staff members of the Sandy Hook Laboratory (J. S. Young, Fishery Biologist, Northeast Fisheries Center Sandy Hook Labora- tory, National Marine Fisheries Service, NOAA, Highlands, NJ 07732. Pers. commun., September 1975), pointed to high prevalence of lymphocystis disease (Figure 6) in limited samples of striped bass overwintering in the heated effluent of a Long Island generating station (Northport, N.Y.). This disease is considered rare in striped bass (Anonymous 1951; Krantz 1970), and its unusual abundance in a localized population may well be related to the abnormally high winter tempera- ture regime in which the population exists, or to abnormal crowding, with consequent increase in stress and ease of transfer of the pathogen. The high temperature may promote survival or trans- fer of the pathogen, or lower resistance of the host, or provoke latent infections into patency, result- ing in grossly recognizable stages of infection. Lymphocystis is considered to be highly infec- tious; initial lesions often develop where injuries to the fish have occurred; and lymphocystis virus reaches peak infectivity when water temperatures are high (Midlige and Malsberger 1968). Some or all of these factors may be important in fostering the high prevalences observed in striped bass. An important concern about fish diseases such as lymphocystis in populations overwintering in heated effluents is that a focus of infection will be provided for incoming spring migrants. STRESS-PROVOKED LATENT INFECTIONS A number of microbial diseases offish have been shown to be provoked into patency by environ- mental stress (Wedemeyer 1970; Snieszko 1974). This seems to be true for kidney disease and furunculosis of salmonids, which often exist in carrier or latent states that can develop into active infections if fish are stressed. It is also probably true for anaerobic bacterial (Eubacterium sp.) in- fections of mullet and 10 other species offish from Biscayne Bay (Udey et al. 1977). A report of vib- riosis in eels held in freshwater (Reidsaether et al. 1977) suggested that latent infections with Vibrio anguillarum produced disease and mortalities when eels were exposed experimentally to 30-60 ixgl\ copper for 50 days in freshwater. Similarly an epizootic oi Aeromonas liquefaciens (= A. hydro- phila) in Atlantic salmon, Salmo salar, and the sucker, Catostomus commersoni, in the Miramichi River, Canada, seemed to be related to combined stresses of copper and zinc pollution and high water temperatures (Pippy and Hare 1969). Figure 6. — Lymphocystis disease in striped bass from heated effluent of a power plant. 727 FISHERY BULLETIN: VOL. 76. NO 4 Snieszko ( 1962) stated, concerningA. liquefacicns that ". . . fish may have latent infections that flare up when the flsh are exposed to stress." There are recent published accounts of two viral diseases of marine invertebrates which also indi- cate that latent infections may be provoked into patency by environmental stress. One, a Baculovirus infection of pink shrimp, Pcnaeus duororum, was first recognized in stressed laboratory populations (Couch 1974b, 1976). The other, a herpes-like viral infection of oysters, was discovered in a population held in a heated power plant effluent in Maine (Farley et al. 1972). An association of shrimp virus disease and low- level chronic exposure to pollutant chemicals is being explored at the Gulf Breeze Environmental Research Laboratory of the U.S. Environmental Protection Agency (Couch 1974a, 1978). In this work a virus disease of pink shrimp caused by B. penaei reached patent levels and caused mor- talities of dO-SO'Vo in shrimp exposed to the PCB Aroclor 1254 and to the organochlorine insecticide Mirex (Couch and Nimmo 1974a, b; Couch 1974a, b, 1976). Other experiments in which the shrimp were crowded, but not exposed to chemicals, re- sulted in similar enhancement of virus infections, indicating that environmental stress may be an important determinant of patent infections. The virus infection has been found subsequently in brown and white shrimp (Overstreet and Howse 1977; Couch 1978). Couch and Courtney (1977) have recently pro- posed an elaborate and unique conceptual scheme to utilize the shrimp virus for interactive bioas- says for chronic sublethal effects of contaminants. The authors point out that there are a number of possible interactions of host, pathogen, and chem- ical stressors — change in resistance of shrimp to the virus, enhancement of widespread latent in- fections in the shrimp population, change in viru- lence of the virus, and losses of diseased shrimps by cannibalism. Criteria developed by Couch and Courtney for interaction include increased viral prevalence in stressed populations (as indicated by numbers of inclusion bodies), increased infec- tion intensity in stressed individuals, increased mortality in stressed populations, and greater cytopathic effects in infected and stressed indi- viduals. The shrimp virus infection has great po- tential for elucidating effects of pollutants on host-pathogen relationships. An association of high environmental tempera- tures with high disease prevalence (or disease en- hancement) in molluscan shellfish sampled from thermal effluents has been made recently. Farley et al. (1972) described a lethal herpes-type virus disease of oysters held in heated discharge water in Maine. The disease, which apparently existed at a low enzootic level in oysters growing at nor- mal low environmental temperatures (12°-18°C summer temperatures), seemed to proliferate in oysters maintained at elevated temperatures (28°-30°C) and to produce mortalities in those populations. Intranuclear inclusion bodies, con- taining viral particles, characterized advanced in- fections. Mortalities of oysters held at higher temperatures were correlated with greater preva- lence of the viral inclusions. Elevated water tem- peratures were considered by the authors to favor spread of the infection or to activate latent infec- tions, or both. This evidence for a possible role of environmen- tal stress in activating latent viral infections could hardly be termed overwhelming, since it is possi- ble that new infections produce the effects discus- sed. However, the two viral diseases may provide an insight into the total effect of pollutant and other environmental factors on disease prevalence and disease-caused mortalities. The carrier state is often difficult to diagnose, but it may play a much larger role in the epizootiology of marine disease than can be demonstrated at present. ENVIRONMENTALLY INDUCED ABNORMALITIES Neoplasms (Tumors) The terms "neoplasia" and "neoplasms," par- ticularly as they concern lower animals, are difficult to define precisely. The Oxford Dictionary definition of neoplasm is "a new formation in some part of the body; a tumor." More elaborate defini- tions exist. Warren and Meissner (1971) defined a neoplasm as "a disturbance of growth charac- terized primarily by an unceasing, abnormal, and excessive proliferation of cells." Prehn (1971) defined neoplasia as "that form of hyperplasia which is caused, at least in part, by an intrinsi- cally heritable abnormality in the involved cells." Although neoplasia has been studied most exten- sively in humans and laboratory mammals, the existence of tumors in fish and shellfish has been recognized for almost a century (the first oyster tumor, for example, was reported by Ryder in 728 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES 1887, and Bonnet mentioned thyroid hyperplasia in fish due to iodine deficiency in 1883). Circumstantial evidence associating environ- mental contamination with neoplasms (tumors) in fish has accumulated from a number of studies: 1. Lucke and Schlumberger (1941) described 166 catfish (Ameiurus nebulosus) with epitheliomas of lips and mouth, taken from the Delaware and Schuylkill Rivers near Philadelphia. The rivers were grossly pol- luted. Tumors of this type may result from mechanical, infectious, or chemical irrita- tion. Catfish from other areas did not have a high prevalence of tumors. The authors did not exclude the possibility that the lesions were induced by chemical carcinogens in the water. The lesions developed into epidermoid carcinomas, some of which were invasive. 2. Russell and Kotin (1957) found 10 of 353 white croakers, Genyonemus lineatus, from Santa Monica Bay, Calif., with papillomas of lips and mouth. Fish were taken 2 m from an ocean outfall. No tumors were found in 1,116 croakers from unspecified nonpolluted wa- ters 70 km away. 3. Cauliflower disease (epidermal papilloma) has been increasing in prevalence in eels (Anguilla anguilla) from the Baltic since 1957. The pattern of spread and high preva- lence indicates an infectious process (viral arrays have been seen) or progressive ac- cumulation of industrial contaminants such as fuel oil and smelter wastes (known to con- tain carcinogenic hydrocarbons such as ben- zopyrene and heavy metals such as arsenic). 4. Cooper and Keller ( 1969) reported that 12'7f of nearly 16,000 English sole from San Fran- cisco Bay had epidermal papillomas, with as many as 33 tumors per fish. Incidence of tumorous fish in the northern part of the Bay was twice that in the southern part. The greatest concentration of industrial waste discharge, especially petrochemicals, existed in the northern part of the Bay. A later sur- vey (Kelly 1971) failed to confirm the areal difference in tumor abundance. 5. Young (1964) found many small (10-15 cm) Dover sole from Santa Monica Bay with tumors. Fish above 15 cm did not have tumors. According to Young, numerous white croakers from Santa Monica and Los Angeles-Long Beach were found with papil- lomas of the lips, and papillomas were ob- served on tongue soles, cusk eels, and Pacific sanddabs. Such tumors were not seen by Young on fish from unpolluted areas, but Dover sole with epidermal papillomas have since been collected off Baja California as far south as Cedros Island (Sherwood and Mearns 1976). The prevalence of lip tumors in white croakers from Santa Monica and the Palos Verdes shelf has been <¥'/< since 1970 (Mearns and Sherwood 1977). 6. Carlisle (1969) found "growths" frequent on white croakers and Dover sole from Santa Monica. 7. Sindermann (1976) found wartlike tumors histologically resembling fibromas in Mugil cephalus from Biscayne Bay in 1969-70 (Fig- ure 7). Other fibrous tumors have been re- ported since then by Lightner (1974) and Edwards and Overstreet (1976) in mullet from the Gulf of Mexico. From the foregoing, it is apparent that much attention has been given, and continues to be giv- en; to the common occurrence of epidermal papil- lomas in a number of Pacific flatfishes ( Wellings et al. 1964, 1965; Wellings 1969a, b). The tumors of English sole from the Pacific coast, for example, have been studied for almost half a century (Pacis 1932; McArn et al. 1968; Good 1940; Angell et al. 1975). Stich et al. (1976) in their review offish tumors and sublethal effects of pollutants, found highest prevalences to occur in young-of the-year fish. Maximum prevalences reported in the litera- ture were 587c in English sole (Stich and Acton 1976); 55^7^ in starry flounder (McArn and Wel- lings 1971); 15% in Ratheadsole, Hippoglossoides elassodon (Miller and Wellings 1971); and over 40'7f in sand sole, Psettichthys melanostictus (Nig- relli et al. 1965). A relationship of high frequen- cies of such papillomas with coastal pollution is still uncertain. Stich et al. (1976) stated "There seems to be a higher skin tumor frequency among English sole inhabiting areas of urban contamina- tion (Vancouver) than among fish populations in regions remote from human activities . . . ." In an extension of this study, Stich et al. (1977) reported prevalences of skin neoplasms in 1-yr-old English sole of from 20 to 70% in samples taken near eight cities on the Pacific coast, while preva- lences did not exceed 0.17c in several samples taken on the British Columbia coast more distant from cities. However, Oishi et al. (1976) examin- 729 FISHERY BULLETIN: VOL. 76, NO. 4 Figure 7. — Wartlike fibrous tumors in Mugil cephalus from Biscayne Bay, Fla. ing prevalences of similar epidermal papillomas in flatfish from relatively unpolluted waters of northern Japan felt that a possible association existed between high tumor occurrence ( up to 20% in certain samples) and parasitization of the flesh by a nematode, Philometra marine, but then they suggested that the involvement of naturally oc- curring chemical contaminants as well as man- made pollutants must be considered in the etiol- ogy of flatfish neoplasms. Wellings et al. (1977) found 1.0% of rock sole, Lepidopsetta bilineata, sampled in the still- unpolluted Bering Sea, with epidermal papil- lomas. Infections were widely distributed geo- graphically, mostly in older individuals. The age distribution of infection was quite different from that in Puget Sound flatfishes, where predomi- nantly younger fish are involved. The etiology of skin tumors in English sole from the Pacific coast of North America was reviewed in a recent paper by Angel et al. (1975), with the conclusion that the cause is unknown, and may be multifactorial. Three stages of tumorigenesis were described in young-of-the-year English sole, beginning with angioepithelial nodules and pro- gressing to epidermal papillomas and an- gioepithelial polyps. No conclusive role of an en- vironmental carcinogen has been demonstrated; there seem to be subpopulation differences in dis- ease prevalences; and electron microscopy has dis- closed the presence of viruslike particles in cells of papillomatous fish (Wellings and Chuinard 1964), but attempts to isolate a viral agent have been unsuccessful. To further complicate the story, an unknown cell type, called an "X cell," has been found in all three tumor stages in English sole. The cells may be parasitic, as was suggested by Brooks et al. (1969), and Alpers et al. (1977b), or they may be transformed host cells, analogous to lymphocystis cells, as was suggested by Angell et al. (1975). Angell et al. concluded by stating that "given the pervasiveness of certain pollutants, experimental evidence and further field studies will be neces- sary to clarify the relationship between tumorous flatfishes and pollution." Another observation on the possible relation- ship of flatfish tumors and pollutants has been supplied by Mearns and Sherwood (1974). The dis- tribution and abundance of skin tumors and fin 730 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES erosion were studied simultaneously in Dover sole from the California coast. Fin erosion was more common in specimens collected near major sewer outfalls, whereas tumorous fish were distributed more evenly throughout southern California coastal waters. The authors concluded that "The spatial and temporal distribution of tumor- bearing Dover sole suggest that initiation of the disease was not related to [municipal] wastewater discharges [in southern California]." A recent report of neoplasms in the Atlantic hagfish, Myxine glutinosa, by Falkmer et al. (1977) suggested a possible relationship of PCB contami- nation and tumor prevalence. During a 5-yr (1972-76) study in Gullmar Fjord, Sweden, neo- plasm prevalences, particularly hepatomas, de- creased from 5.8 to 0.6'^ . PCB levels in livers of hagfish were appreciable (5 ppm), but the use of PCB was prohibited in 1971. Liver PCB levels in hagfish caught inside the Qord were five times higher than in those caught outside. However, the association of PCB contamination with liver tumors must be considered to be tenuous. Earlier reports of neoplasms in hagfish (Fange etal. 1975; Falkmer et al. 1976) described remarkably high frequencies in Gullmar Fjord, but only low con- centrations (0.5-1.0 ppm) of PCB in livers, and low environmental levels of PCB and other contami- nants. The role of environmental chemical factors in induction of neoplasms in shellfish is even less clear than for fish, but there is some limited infor- mation. Yevich and associates (Barry and Yevich 1975; Yevich and Barszcz 1976, 1977) have for a number of years examined the occurrence of neo- plastic growths in the soft-shell clam, Mya arenaria, in relation to petroleum contamination. Gonadal and hematopoietic neoplasms were ob- served in animals collected from two chronically contaminated sites on the Maine cost, with pre- valences up to 29% in certain samples. Yevich and Barszcz (1976) stated that "no tumors similar to those described [from the petroleum contaminated area] have been encountered in animals collected from any other area." They described the scope of their study as "several thousand animals from all coastal areas of the United States." Additional samples of clams from a number of other coastal locations are needed, as is a more precise descrip- tion and confirmation of the neoplastic condition. It is interesting that a counterpart study of soft-shell clams from Rhode Island and Mas- sachusetts (Brown et al. 1976) reported occur- rences of neoplasia, apparently of hematopoietic origin, in up to 26% , with the highest frequency in samples from a 1975 oil spill area near Bourne, Mass. A later report (Brown et al. 1977) included additional samples from other geographic areas. Neoplasms of gonadal origin, similar to those re- ported by Yevich and associates, were found in clams from an oil-contaminated site at Searsport, Maine. The highest prevalence of neoplasms of hematopoietic origin was 64% , in a small sample from Bourne. The authors pointed out, however, that clams from some oil-contaminated sites had no neoplasms, and stated that "These results suggest that the type and degree of hydrocarbon pollution are possibly related to the frequency of neoplasms and other lesions in Mya, but they are by no means the only causative factors." Other types of cellular abnormalities have been reported from soft-shell clams. In earlier studies by Yevich and associates (Barry et al. 1971) atypical epidermal hyperplasia in gills and kidney was reported in up to 40% of clams sampled near Providence, R.I. Lesions occurred more frequently in large individuals, and seasonal changes were not observed. Lower prevalences were found in limited samples from Maine, Maryland, and California. Unlike the oil spill studies, no associa- tion with environmental factors was made by the authors. Yevich and associates (Yevich and Barry 1969; Barry and Yevich 1972) have also described gonadal neoplasms in quahogs, Mercenaria mer- cenaria, from Narragansett Bay. Samples col- lected in 1968, 1969, and 1970 had tumor frequen- cies of 0.2, 2.3, and 2.7% respectively. Epizootic neoplasms with a possible environ- mental etiology were reported from several mol- luscan species of Yaquina Bay, Oreg. (Farley 1969b; Farley and Sparks 1970; Mix et al. 1977). Blue mussels, Mytilus edulis, native oysters, Ostrea lurida, and two species of Macoma were affected, and winter mortalities were associated with the disease. Neoplasms have not been found in bivalve molluscs sampled elsewhere on the Oregon coast (Mix et al. 1977). In another study (Christensen et al. 1974) simi- lar epizootic neoplasms (up to 10% prevalence) were found in a localized population of the clam Macoma balthica from a tributary of Chesapeake Bay. The neoplasms were invasive and systemic, with initial foci in the gill epithelia. Holding ex- periments indicated that the disease was usually 731 FISHERY BULLETIN: VOL. 76. NO 4 fatal. The authors suggested, but did not dem- onstrate, an environmental contaminant etiology, possibly associated with bottom detritus. Other bivalve molluscs in Chesapeake Bay contain neo- plasms. American oysters, Crassostrea uirginica, were found with hematopoietic neoplasms (Farley 1969a; Couch 1969, Frierman 1976), and indi- vidual oysters have been reported to contain other types of neoplasms (Pauley 1969; Couch 1970). Much of the evidence associating certain neo- plasms offish and shellfish with pollutants should be considered as circumstantial but provocative (Rentchnick 1976). Many of the neoplasms have been reported from bottom-feeding fish and de- tritus or filter-feeding bivalves, as was pointed out by Harshbarger.^ Chemical carcinogens such as certain heavy metals and hydrocarbons can be concentrated in surficial layers of bottom sedi- ments and can thus be readily available to ani- mals inhabiting that zone. It should be noted, though, that a number of recent studies of neo- plasms in fish and shellfish have found no obvious relationship between neoplasms and specific en- vironmental factors. Skeletal Anomalies Skeletal anomalies, particularly those of the spinal column, are commonly observed in fish and are the subject of an extensive literature (see Rick- ey 1972, for a recent summary and Dawson 1964, 1966, 1971, and Dawson and Heal 1976 for a com- plete bibliography). Such anomalies may be genetic, resulting from mutations or recombinations; epigenetic, acquired during embryonic development; or postembryonic, acquired during larval development, at metamor- phosis, or during juvenile development (Hickey 1972). Spinal flexures and compressions, as well as vertebral fusions, have been observed in many teleost species, as have head and fin abnor- malities. Evidence exists for a hereditary basis for some skeletal anomalies (Gordon 1954; Rosenthal and Rosenthal 1950), but other evidence points to effects of environmental factors such as tempera- ture, salinity, dissolved oxygen, radiation, dietary deficiencies, and toxic chemicals. For example, in- creased percentages of abnormal embryos and lar- vae of Atlantic herring, Clupea harengus, resulted 'Harshbarger, J.C. 1974. Activities report (of the) registry of tumors in lower animals 1965-1973. Smithson. Inst., Wash., D.C., 141 p. from experimental exposures to sulfuric acid waste water (Kinne and Rosenthal 1967) and to the algicides 2,4- and 2,5 dinitrophenol ( Rosenthal and Stelzer 1970). Recently, increased prevalences of skeletal de- formities and anomalies, considered to be pollution-associated, have been recognized in a few fish species from southern California, the British Isles, and Japan. In studies carried out in California, skeletal deformities occurred with greater frequency in samples from areas with sig- nificant pollutant stress (Valentine and Bridges 1969; Valentine et al. 1973). Exposure of fry to very low concentrations of DDT ( <1 ppb) produced anomalies in fin rays (Valentine and Soule 1973). Probably the most convincing observational evidence for environmental influences on induc- tion of skeletal abnormalities in marine fish is that presented by Valentine (1975). Examining samples of barred sand hass, Paralabrax nebulifer, Valentine found significantly higher prevalences of anomalies, particularly gill raker deformities, in fish from the southern California coast (Los Angeles and San Diego) than from the Baja California coast. The anomalies increased in fre- quency and severity with increasing size of the fish and an association with disturbed calcium metabolism was suggested. The author pointed to the high chlorinated hydrocarbon and heavy metal levels which characterize the California coastal area (Schmidt et al. 1971; Galloway 1972), but emphasized that a causal relationship with increased prevalence of anomalies had not been established. However, Valentine's suggestion of a possible causal relationship between high en- vironmental levels of chlorinated hydrocarbons and heavy metals, both of which are known to interfere with calcium metabolism, and skeletal anomalies in fish seems reasonable, in view of experimental evidence from a wide range of ver- tebrates (Ferm and Carpenter 1967; Lehner and Egbert 1969; Peakall and Lincer 1970; Pichirallo 1971; McCaull 1971; Galloway 1972). Valentine (1975) referred briefly to additional observations on two other Pacific coastal species — California grunion, Leuresthes tenuis, and barred surfperch, Amphistkhus argenteus — in which gill raker anomalies increased in fre- quency with age, and were "virtually restricted to [samples from] fishes from Southern California." This finding in three species reduces the likelihood that frequency differences could be attributable to 732 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES inherited subpopulation differences in one of the three species studied. While the deformed gill rakers were the most prevalent anomalies observed in southern California barred sand bass by Valentine, other abnormalities (pugheadedness, cranial asymme- tries, deformed vertebrae, and fin anomalies) oc- curred and were associated directly in frequency and severity with gill raker deformity. An analysis of vertebral deformities in herring taken in waters around the British Isles (van de Kamp^") indicated a slight but significant increase in prevalences from 1960 to 1975. The predom- inant abnormality was a cluster of two or three incomplete vertebrae located near the pelvic fins or anus. The highest percentages of abnormalities were found, according to the author, in areas "which probably had the highest degree of pollu- tion." It was in these areas where prevalences also showed slight increases during the study period, supporting the author's hypothesis that vertebral deformities in herring can be related to "unusual substances" in the environment. However, van de Kamp concluded by stating that more experimen- tal work on the causal relationship between pollu- tion and deformities will be required. Several reports from Japan refer to high and increasing occurrences of skeletal anomalies in fish. Komada (1974) and Ueki and Sugiyama ( 1976) observed increasing numbers of malformed sweetfish or ayu, Plecoglossus altivelis, in rivers and culture farms. Skeletal abnormalities in mul- let and eight other species from the Inland Sea of Japan were reported by Matsusato (1973). Deformed fin rays (Figure 8) and associated skeletal abnormalities have been observed re- peatedly in winter flounders from the highly pol- luted waters of the New York Bight (Ziskowski et al. in press), and a summarization of observations on skeletal anomalies and related developmental defects has been published recently (Sindermann et al. 1978). There is some evidence from studies of a few other fish species for an involvement of various kinds of environmental stress in the occurrence of skeletal anomalies. Gabriel (1944) noted anomalies in vertebrae of Fundulus heteroclitus due to temperature changes, and Mottley (1937) found anomalies in vertebral numbers of trout due '"van de Kamp, G. 1977. Vertebral deformities in herring around the British Isles and their usefulness for a pollution monitoring programme. Int. Counc. Explor. Sea, Fish. Improv. Comm.. Doc. CM1977 E:5, 9 p. to temperature (and possible oxygen). Hubbs (1959) found high prevalences of vertebral abnor- malities in mosquitofish, Gambusia affinis, from Texas warm springs and concluded that the high temperature was responsible. There is also an appreciable literature con- cerned with induction of skeletal injuries in fish by exposure to contaminants. Vertebral damage following experimental exposure to aquatic con- taminants has been reported for a number of freshwater fishes (Bengtsson 1975). Long-term (10 wk) exposure of minnows {Phoxiniis phoxinus) to sublethal concentrations of zinc and cadmium resulted in hemorrhaging, spinal curvatures, and vertebral fractures, particulary in the caudal region, in up to 70^^ of individuals. Spinal curva- tures and muscle atrophy were produced in rain- bow trout by chronic exposure to lead. It is in- teresting that caudal fin erosion was also observed in these experiments. In earlier studies, sum- marized by Bengtsson, exposure to sublethal con- centrations of the chlorinated hydrocarbon Toxa- phene as well as to Malathion, parathion, and certain other organophosphorus pesticides pro- duced vertebral damage or spinal flexures in sev- eral fish species. Vertebral damage was consid- ered to have a neuromuscular origin, or, in the case of long-term exposure, to be a consequence of demineralization. John Couch and associates at the Gulf Breeze Environmental Research Laboratory of the U.S. Environmental Protection Agency are developing experimental evidence for induction of skeletal abnormalities by exposure to environmental con- taminants. Couch et al. (1977) reported severe scoliosis and associated pathology in the sheeps- head minnow, Cyprinodon variegatus, exposed to the organochloride pesticide Kepone. The authors concluded that scoliosis was a secondary effect of Kepone toxicity, with the nervous system or cal- cium metabolism as the primary target. Couch and associates (J. A. Couch, Research Pathologist, Environmental Research Labora- tory, U.S. Environmental Protection Agency, Gulf Breeze, PL 32561. Pers. commun., June 1977) have also found that trifluralin (Treflan) induced extensive osseous hyperplasia in vertebrae of sheepshead minnows when life history stages from zygote to 28-day juveniles were exposed to 25-50 ppb trifluralin. Centra of vertebrae, thick- ened by active osteoblasts and fibroblasts, in- creased in size up to 10-30 times their normal dimensions — a striking sublethal effect. 733 FISHERY BULLETIN: VOL. 76, NO. 4 Figure 8. — Deformed fin rays in winter flounder from New York Bight. External appearance (above) and radiograph (be- low i (From Sindermann et al. 1978. GENETIC ABNORMALITIES The mutagenic properties of a number of chem- ical contaminants including heavy metals, pes- ticides, and petroleum-derived polycyclic hydro- carbons have been demonstrated in experimental studies with terrestrial animals (Huberman 1975; Longwell"). Fish eggs can be vulnerable to con- taminant effects from the body burden of the par- ent female and from exposure to contaminants in surface water and/or sediments (depending on where in the water column spawning and de- velopment occur). Sperm cells are sensitive to contaminants, and eggs are especially sensitive during meiosis and early cleavage stages. Fur- thermore, chemical mutagens can reduce the rate ll "Longwell, A. C. 1975. Mutagenicity of marine pollutants as it could be affecting inshore and offshore marine fisheries. 734 Middle Atl. Coastal Fish. Cent., Natl. Mar. Fish. Serv., Inf. Rep. 79, 72 p. SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES of cell division and can damage the mitotic spindle apparatus. Pelagic eggs may be most severely damaged, since the surface film of the ocean has been found to contain high concentrations of con- taminants such as petroleum components, halogenated hydrocarbons, and heavy metals (Maclntyre 1974). Some experimental evidence is available. Fish larvae incubated in cadmium-polluted water ac- cumulated the metal (Westernhagen et al. 1974; Rosenthal and Sperling 1974), and eggs incubated in as little as 1 ppm cadmium produced low per- centages of viable larvae (Westernhagen et al. 1975; Westernhagen and Dethlefsen 1975). Some relevant experimental research on radionuclide-induced mutagenesis (Romashov and Belyayeva 1966; Ivanov 1967; AEC-TR- 7299'"^) has disclosed that many fish embryos with severe chromosomal damage died during the transition from blastula to gastrula. Abnormal postgastrula embryos contained higher numbers of chromosomal aberrations than normal em- bryos, and the abnormal embryos had high mor- tality just before hatching. However, even the normal-appearing embryos with radiation expo- sure (and consequent genetic disturbances) had low viability and high mortality at hatching and subsequent to hatching. Recently, Longwell (1976a, b) reported high prevalences of chromosomal anomalies in Atlantic mackerel , Scorn ber scotfi brus , eggs and embryos in certain samples taken from the New York Bight. All degrees of chromosomal damage were found, including failure to align at the metaphase plate, incomplete spindle formation, translocation bridges, chromosomal "stickiness," losses of por- tions of chromosomes and "pulverization." Eggs with at least one chromosome or mitotic abnor- mality varied from 13 to 79^r . Higher percentages seemed associated generally with degrees of en- vironmental degradation. In addition to chromosomal anomalies, one station (the one with highest prevalence of anomalies) was also charac- terized by significant {26'7c ) egg mortality. The techniques developed by Longwell (1976b) permitted examination of historical collections of eggs and embryos for chromosomal damage. A limited collection taken in 1966 from the same geographic area disclosed a lower incidence of cytogenetic abnormalities than that found in the 1974 collection. Samples examined to date from normal and de- graded waters are still insufficient, as Longwell ( 1976b) pointed out, to make definitive statements about the relationship of pollutants and extent of damage to genetic material, but the data pre- sented so far indicates that such a relationship may exist. Because of the implications of these findings in survival and abundance of economic marine species, it is particularly important that this kind of research be pursued vigorously. It may well be that a new and significant mortality factor for estuarine and coastal populations — increased genetic damage — may have been introduced with increasing chemical pollution. It is likely that marine organisms will respond to mutagens in species-specific ways and with dif- fering sensitivities. Some indication of this can be found in a recent paper by Vandermeulen and Lee^-^ in which cultures of the alga Chlainydomo- nas reinhardtii were exposed to crude and refined oils (Kuwait crude, Saran Gach crude, diesel 25, and bunker C). No enhanced mutation rates (as detected by streptomycin resistance) were found after 3 wk of exposure (40-50 generations), a sur- prising finding, since the alga is susceptible to cer- tain other known mutagens and since the test oils contain various polycyclic aromatics which are known mutagens. No cytological examinations were reported. The authors pointed out that con- centrations of mutagenic components in the test oils may be low compared with concentrations used in cell and tissue culture to elicit enhanced mutation rates, and that extrapolation of labora- tory results to the marine environment should be done very conservatively. An indirect test for the presence of mutagens in the marine environment has been reported re- cently by Parry et al. (1976). Mytilus edulis were sampled from polluted and unpolluted waters of the United Kingdom, and extracts of their tissues were tested for ability to induce genetic changes in bacterial and yeast cultures. Significant increases in mutation rates for specific gene loci charac- terized cultures exposed to extracts of mussels from polluted waters, but not those from clean waters — providing evidence for the presence of mutagens that had been concentrated in the tis- ■2AEC-TR-7299. Marine radioecology. 1972. (Distrib- uted by NTIS, U.S. Department of Commerce, 5285 Port Royal Road, Springfield, VA 22151.) i^andermeulen, J. H., and R. W. Lee. 1977. Absence of mutagenicity due to crude and refined oils in the alga Chlamydomonas reinhardtii. Int. Counc. Explor. Sea, Plankton Comm., Doc. CM1977/E:69, 5 p. 735 FISHERY BULLETIN: VOL. 76. NO. 4 sues of mussels from polluted areas. The chemical nature of the mutagens was not identified, except that the mussels came from areas with heavy in- dustrial pollution. EXPERIMENTALLY INDUCED LESIONS There is a vast and almost unmanageable amount of published information about the induc- tion of various lesions in fish by experimental ex- posure to chemical contaminants (see for example Ribelin and Migaki 1975). A "lesion" may be defined generally as "any localized abnormal structural change in the body." Such a definition obviousl)' includes too much, so the term can be reduced to encompass "those cellular and tissue changes, demonstrable histologically, that result from a disease process." Histopathology offish and shellfish is still a developing science, and as such it still draws from human and veterinary (mamma- lian) patholog}' for its concepts and much of its terminology. Histopathology has been a basic tool in human medicine for some time, and a large amount of information is available about cellular responses to toxicants. A similar core of knowl- edge is being developed for fish and shellfish — relating cell and tissue changes to kinds and amounts of contaminants. Early experimental exposures of estuarine and marine animals to contaminants usually had the purpose of determining lethal dosages, either from acute or chronic exposures. More recently, atten- tion has been redirected to sublethal toxic effects — expressed in behavioral, physiological, or cytological responses to specific contaminants. An extensive literature exists concerning cell and tis- sue damage resulting from experimental exposure to contaminant chemicals. Generalizations that can be made are almost predictable: 1) increas- ing dosages, beyond a threshold level, produce in- creasingly severe tissue abnormalities; 2) particu- lar contaminants often exert effects on specific target tissues; 3) principal target tissues seem to be gill epithelium, liver (or in the invertebrate, the hepatopancreas), and neurosensory cells; 4) specific lesions cannot usually be described as characteristic of any group or class of chemicals; and 5) effects that may be of chemical origin can be obscured by stress-provoked infections with facul- tative pathogens. Some information about ex- perimental induction of fin erosion and skeletal abnormalities has been included in earlier sec- tions of this paper, but because of the sheer volume 736 of published information about other types of ex- perimental lesions, it seems worthwhile to sum- marize some of the observations here. Couch (1975) published a recent and excellent review of the histopathological effects of pesticides and related chemicals on the livers of fishes. The liver and fatty tissues of fish from natural waters are known to accumulate a number of chlorinated hydrocarbons (Duke and Wilson 1971), and ex- perimental exposures offish to pesticides result in high concentrations and greatest effects on the liver (Johnson 1968; Eisler and Edmunds 1969; Hansen et al. 1971; Eller 1971). Some of the ob- served liver histopathology includes: Chlorinated hydrocarbon pesticides: Focal areas of parenchymal cell vacuolation and de- generation (Eller 1971), inflammation, and loss of glycogen and fat (Lowe 1965). Clorinated hydrocarbon herbicides: Increase in connective tissue, massive focal necrosis (Cope et al. 1969), and loss of glycogen (Cope et al. 1970). PCB's: Focal degenerative regions, paren- chymal cell vacuolation and pleomorphism (Eller''*), lipid accumulation in hepatic cell vacuoles, and leucocytic infiltration (Couch 1975). Organophosphates: Edema, hyperemia, vacuo- lation, and necrosis of parenchymal cells (El- ler, see footnote 14). Carbamates: Hypertrophy and vacuolation of acinar cells (Couch 1975). It should be noted that not all experimental exposures to pesticides, even for prolonged periods, necessarily caused demonstrable tissue pathology, but in many instances additional expo- sure experiments are needed (even though the lit- erature as summarized by Couch (1975) seems voluminous). Couch pointed out that over 900, commercial pesticide formulations are in general use, and of these fewer than 30 have been tested for pathological effects on livers of fishes. Pesticides can, of course, affect fish tissues other than liver. A summarization of general his- topathological effects of pesticides on fish was pub- lished by Walsh and Ribelin (1975). Data from their own studies with coho salmon, Oncorhyn- chus kisutch, and lake trout, as well as from other '••Eller, L. L. 1970 and 1971. Annual reports. U.S. Bur. Sport Fish. Wildl., Fish Pestic. Lab., Columbia, Mo. SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES published work, led them to the conclusion that tissue changes observed as a result of exposure to an array of common pesticides were largely nonspecific, and therefore of limited diagnostic value. Their attempts to identify specific lesions as characteristic of any group or class of pesticides were described by them as "futile," but the amount of histopathological information presented in the paper is substantial, and their summarization of pathology pi'oduced by exposure to widely used pesticides is instructive. DDT: Necrosis of hepatic cells; lymphocytic infiltration of intestinal lamina propria; pos- sible degeneration of kidney tubules. Carbaryl (Sevin): Intramuscular hemor- rhages adjacent to vertebral column; atrophy of the lateral line musculature; myxomatous degeneration of fat; vacuoles within the optic tectum of the brain. Malathion: Subcutaneous hemorrhages at the bases of pectoral fins. Endosulfan (Thiodan): Hyperemia of intestine and brain; adrenal cortical hyperplasia. 2,4-D: Striking degree of brain hyperemia; hyperemia of intestine. Atrazine: Marked edema of all tissues; changes in skin pigmentation. It is interesting that Walsh and Ribelin (1975) (unlike Couch 1975) found liver changes in fish exposed to pesticides ". . . minimal and diagnosti- cally unimportant . . . ." They also considered gill epithelial hyperplasia, gill hemorrhages, and lymphocyte reduction in the spleen to be nonspecific responses to stress and/or infection. They further pointed out that rapid autolysis of fish tissues after death rather than direct effects of pesticides might account for some reported his- topathological findings. These are all points of im- portance in evaluating histological findings after exposure to any contaminant. There are still other histopathological studies of the effects of pesticides on fish that disclose dam- age to neurosensory tissue. Epithelial necrosis was found in lateral line canals of killifish, F»/?- dulus heteroclitus. that survived 96-h exposures to the chlorinated hydrocarbon methoxychlor at 25 mg/1 (Gardner 1975). No damage to the mechanoreceptors was evident, but the radius of the canal lumina was reduced. Pesticides can produce tissue pathology in in- vertebrates as well. Oysters exposed chronically to 3 ppb DDT, Toxaphene, and parathion exhibited variable lesions, including leucocytic infiltration or hyperplasia of the gonadal germinal epithelium, necrosis of digestive tubule epithelium, and edema (Lowe et al. 1971). In another study, chronic exposure of oysters to 5 ppb PCB produced atrophy of digestive epithelium, leucocytic in- filtration, and degeneration of vesicular connec- tive tissue (Lowe et al. 1972). Gill edema and pro- gressive necrosis of filaments in the crustacean Gammarus oceanicus resulted from exposure to sublethal concentrations of PCB (Wildish 1970). Examination of pink shrimp, exposed experimen- tally to PCB's, disclosed a variety of nonspecific tissue changes, especially in the hepatopancreas (Couch et al. 1974). Histological changes included lysis of hepatopancreatic epithelium, nuclear pye- nosis, vacuolization of secretory cells, and a vari- ety of ultrastructural changes in absorptive cells. The literature on experimentally induced le- sions in estuarine/marine fish caused by exposure to heavy metals was reviewed recently by Gardner ( 1975) in a paper which also presented significant new information. His general conclusion was that sensory organ systems of some species are vulner- able to copper, mercury, and silver. Short-term exposure of the killifish to sublethal concentra- tions of copper resulted in degeneration of anterior lateral line and olfactory sensory tissues (Gardner and LaRoche 1973). Prolonged exposure to copper (copper chloride) resulted in hyperplasia or necro- sis of sustentacular epithelium of the olfactory organs and necrosis of the epithelial lining of ol- factory pits. Mercury (mercuric chloride) also pro- duced severe degenerative changes in cells of the lateral line canals and olfactory organs of killifish, but without associated necrosis of supporting tis- sues. Exposure to silver produced histopathologi- cal changes very similar to copper. Cadmium (cadmium chloride), however, did not seem to af- fect the sensory tissues discussed above, at least in terms of causing demonstrable tissue changes. Cadmium exposure did result in transient thyroid hyperplasia and altered blood cell ratios in long- term exposures. Experimental exposure of the cunner, Tautogolabrus adspersus, to cadmium caused pathological changes in kidney, intestine, hemopoietic tissue, epidermis, and gills (Newman and MacLean 1974). Necrosis of tubular epithelium of the kidney, sloughing of intestinal epithelium, hypertrophy and hyperplasia of gill epithelium, and decrease in mucus secretion were 737 FISHERY BULLETIN: VOL. 76. NO. 4 the principal histopathologic findings. Mortality following acute exposures was attributed to renal failure. These results were similar in most re- spects to cadmium-induced pathology in killifish, reported earlier by Gardner and Yevich (1970). Experimental cadmium exposures can cause gill lesions in shrimp, as reported in recent papers by Nimmo et al. (1977) and Couch (1977). Expo- sure of pink shrimp to 763 /tg/l of cadmium for 15 days resulted in a "black gill" condition charac- terized by necrosis of all cell types in the distal gill filaments, with coincident appearance of black granules in the cytoplasm, and some hemocyte infiltration at the bases of the necrotic filaments. Couch suggested that the black deposits could be a metallic sulfide or even cadmium. He further pointed out that the distal filament tissue has been postulated to have detoxifying, as well as os- moregulatory and respiratory functions, so that cell death could result from cadmium filtration and accumulation as part of a detoxification pro- cess. Zinc has been shown to be toxic for fish (see reviews by Skidmore 1964, 1970). Gill tissues can be destroyed in acute exposures, while chronic levels induce stress which may result in mortality and may also produce severe degenerative changes in the liver and kidneys (Crandall and Goodnight 1963). Synergistic activity of zinc with a wide range of environmental variables — other contaminant heavy metals, low dissolved oxygen, and temperature — has been demonstrated for a number of fish species (Doudoroff 1957; Lloyd 1960, 1961a, b). Resistance to zinc poisoning var- ies with individuals, with age, with degree of acclimatization, and with species (Jones 1938, 1940). Histopathological effects of sublethal concen- trations of copper on the winter flounder were de- scribed by Baker ( 1969). At dosages of 1,000-3,200 /Lig/1, the kidney hemopoietic tissue became necro- tic; gill epithelium became disoriented; chloride cells increased in number and size; gill lamellar fusions occurred; and fatty metamorphosis of the liver was observed. Experimental concentrations were far above those levels expected in most marine environments (concentrations in polluted waters have been reported to reach 300/Ltg/l by Fujiya (I960)). An interesting study of pathology in American lobsters was made following disclosure of severe yellow phosphorus industrial contamination of Placentia Bay, Newfoundland (Aiken and Byard 738 1972). Experimental lobsters, exposed to phos- phorus contaminated sediments in aquaria, ex- hibited degenerative changes in antennal glands and in all cell types in the hepatopancreas, as well as massive coagulation of hemolymph. Experimental exposure to petroleum compo- nents and residues may also induce histopatholog- ical changes in fish. Hyperplasia of the olfactory sustentacular epithelium and degeneration of the olfactory mucosa of the Atlantic silverside, Menidia menidia, resulted from exposure to crude oil (Gardner 1975). Additionally, degeneration of the ventricular myocardium of the heart and pseudobranch secretory cells was seen. Soluble components of the crude oil also caused epithelial metaplasia, replacing the sensory epithelium of the olfactory organs by poorly defined cell types ( Gardner 1975). Liver damage occurred in fish fed cyclopropenoid fatty acids (Malevski et al. 1974), but Brocksen and Bailey (1973) found no his- topathology in chinook salmon and striped bass exposed to sublethal concentrations of benzene. Histopathological effects of petroleum on bivalve molluscs are varied in the extreme. Ef- fects, particularly on gill epithelium, have been observed by Barry et al. (1971), Jeffries (1972), LaRoche (1972), Clark et al. (1974), and Gardner et al. ( 1975). Fries and Tripp ( 1976) found damage to gill epithelium in hard (hard-shell) clams, Mer- cenaria mercenaria, exposed to as little as 1 ppm phenol. Vaughan^'', however, found little his- topathology after chronic exposures of oysters to No. 2 fuel oil. Stainken ( 1975) found that exposure of soft-shell clams to No. 2 fuel oil at winter seawa- ter temperatures (4°C) for 28 days had little his- topathological effect, beyond signs of starvation (glycogen depletion and vacuolization of digestive diverticula cells), and a generalized leucocytosis, even at 100 ppm. No mortalities occurred, and exposure concentrations dropped rapidly, possibly because much of the oil was trapped in mucus as part of the mucociliary feeding mechanism, and ejected from the clam. Experimental lesions are instx'uctive in iden- tifying target organs and tissues for particular contaminants, but they have numerous flaws when attempts are made to relate experimental findings to events in the natural (polluted) envi- ronment: 1) dosage levels are often beyond '^Vaughn.B.E. (editor). 1973. Effects of oil and chemically dispersed oil on selected marine biota - a laboratory study. Am. Pet. Inst. Publ. 4191. SINDKRMANN: fOLLl'TION-ASSOCIATKl) OISEASKS AND ABNOKMAI.ITIKS maximum observed environmental levels; 2) usu- ally single chemicals are tested, ignoring possible synergisms and antagonisms; 3) tests are often static acute rather than chronic exposures in flow-through systems; and 4) experimental ani- mals are often under stress from the mere act of confinement. These and other limitations of experimental studies degrade the evidence obtained to cir- cumstantial when attempts are made to extrapo- late findings to natural populations in polluted habitats. Despite this handicap, there is a large and useful literature on experimental lesions in fish and shellfish produced by chemicals which occur as contaminants in the coastal environment. The presence of specific pollutants cannot be recognized by the occurrence of specific lesions, but a general description of pathological responses can be useful. Categories of pathological responses which should be considered in experimental studies are; 1 (inflammation (acute and chronic); 2) degeneration (including edema, necrosis, and metaplasia); 3 ) repair and regeneration (prolifera- tion, hyperplasia, and scar formation); 4) neo- plasia (including consideration of cell origin, stage, and type — whether benign or malignant); and 5) genetic derangement (including chromosomal changes and skeletal abnor- malities). CONTAMINANT EFFECTS ON RESISTANCE AND IMMUNE RESPONSES Suppression of immune responses by toxicants such as heavy metals and pesticides has been demonstrated repeatedly in mammals (Kolom- iitseva et al. 1969; Hemphill et al. 1971; Khan and Hill 1971; Jones et al. 1971; Roller 1973; Street and Sharma 1975). Therefore, it might be expected that environmental pollutants could influence the ability of fish and shellfish to resist infection by reducing the effectiveness of external and internal defense mechanisms, and indeed there is some evidence that this is so. Changes in the principal external defenses — mucus secretion offish and the epicuticle of Crustacea — have already been men- tioned in connection with fin erosion and exo- skeletal erosion. Some specific information is available about contaminant influences on inter- nal defenses, principally through suppression of immune responses. Environmental stress from contaminants can affect internal resistance to in- fection in fish by causing a decrease in phagocytic activity (Wedemeyer 1970) or a decrease in anti- body synthesis (Goncharov and Mikyakov 1971). Both mechanisms have been demonstrated ex- perimentally. One of the best pieces of supporting infor- mation about suppression of host responses was derived from a recent multidisciplinary experi- mental study of the effects of short-term sublethal exposures to cadmium on the teleost Tautogolab- rus adspefsus (Calabrese et al. 1974). The study included chemical analyses of tissue uptake, physiological and biochemical effects, his- topathological changes, and effects on the immune system. Robohm and Nitkowski ( 1974), who were responsible for the immunology, found that expo- sure offish to 12 ppm cadmium affected phagocyte response to foreign antigen, but not the humoral response. The rate of bacterial uptake in phago- cytes of liver and spleen was increased, but the rate of bacterial destruction within the phagocytes was decreased significantly. No change was ob- served in the antibody response of immunized con- trol and experimental fish as determined by hemagglutination techniques. The authors postu- lated that cadmium may prevent delivery of lysosomal substances to the phagocytic vacuole, or may inhibit the action of these substances on bac- teria, but that cadmium does not seem to inhibit antibody synthesis by lymphocytic cells. The au- thors further suggested that cadmium and possi- bly other pollutants may affect fish populations by causing phagocytic dysfunction, reducing the re- sistance offish to facultative and other pathogens. The effect of sublethal copper exposure on the immume response of juvenile coho salmon, On- corhynchus kisutch, was examined by Stevens.'^ At copper levels of 18 )U,g/l, agglutinin titers in fingerlings injected intraperitoneally with Vibrio anguillarum bacterin were significantly lower than those of controls. Copper exposure also re- duced survival of coho salmon fingerlings during saltwater acclimation. Reduction in immunological competence may well have been involved in observed outbreaks of vibrosis (V . anguillarum ) in eels exposed to copper (R0dsaether et al. 1977) and in epizootics of Aeromonas liquefaciens (= A. hydrophila) in salmon and suckers exposed to copper and zinc pollution (Pippy and Hare 1969), although in '^Stevens, D. G. 1977. Survival and immune response of coho salmon exposed to copper. Environ. Prot. Agency - 600/3- 77-031, 37 p. 739 FISHKRY BULLKTIN: VOL. 76. NO. 4 neither instance were antibody titers determined. In the latter instance, A. Uquefaviens is an ubiquitous water bacterium, but only causes dis- ease and mortalities in fish with lowered resis- tance (Snieszko 1962). Reduction in antibody response to injected virus was demonstrated by Perlmutter et al. (1973) in blue gourami, Trkhogaster trichopteri/s, to result from overcrowding. The authors postulated that stressed fish released a pheromonelike im- munosuppressive factor under crowded condi- tions. It is reasonable to expect that other types of environmental stresses could result in a similar response. Among the invertebrates, indirect evidence for reduction of disease resistance caused by conta- minant exposure is available and has already been discussed in previous sections on crustacean shell disease and shrimp virus disease. Direct experi- mental evidence however, is scarce. Fries and Tripp (1976) exposed hard (hard-shell) clams to phenol and found damage to gill and digestive tract epithelia — tissues which are considered im- portant components of internal defense mechanisms. The authors suggested, but did not demonstrate, that phenol-treated clams may be more susceptible to microbial infections than normal ones. In other studies with invertebrates, Telford (1968, 1974) demonstrated that environ- mental stress affected blood glucose levels in Homarus americanus and crayfish, Cambarus clarkii. POLLUTANT-PARASITE INTERACTIONS Much has been said and much documentation exists about the role of environmental stress in induction, severity, and persistence of disease. Some of the best information about stress and dis- ease in fish comes from studies concerned with aquaculture — where environmental factors such as temperature, oxygen, water quality, salinity, and diets clearly influence the course of disease and the impact of disease on cultured populations. There is also a developing body of information, from experimental work as well as from field ob- servations and surveys, about the possible rela- tionship of parasitism and pollution. The relation- ship is not simple, and in essence involves a double-edged phenomenon, in which pollutant stress may result in an increase (or in some in- stances decrease) in the prevalence of certain 740 parasites, or in which parasitization may decrease host resistance to toxic pollutants. Subsidiary is- sues quickly emerge however, such as the effects of pollutants on intermediate or alternate hosts in parasite life cycles, possible effects of pollutants on free-living life cycle stages of parasites, and effects of pollutants on host defenses against parasite in- vasion. Thus far in this review, the role of microbial infectious agents, principally viruses and bac- teria, has been emphasized, but there is some li- mited evidence that environmental pollution may change the relationships among animal parasites and their fish hosts (Esch et al. 1975). Looking first at the influence of parasites on host susceptibility to contaminants, several recent papers (principally from studies in freshwater) offer significant insights. Boyce and Ydmada (1977) found in laboratory experiments that sock- eye salmon, Oncorhynchus nerka, smolts with preexisting parasitization by the intestinal pseudophyllidean cestode Eiihothrium salvelini were more susceptible to zinc poisoning than un- parasitized siblings. Similarly, Pascoe and Cram (1977) found that survival times of the threespine stickleback, Gasterosteiis aciileatus, exposed to various concentrations of cadmium, were much shortened if the fish were parasitized by the larval cestode Schistocephalus solidiis. Perevozchenko and Davydov (19741 found that juvenile carp parasitized by the intestinal cestode Both rioceph- alus goivkongensis were more susceptible to DDT poisoning than were nonparasitized individuals. These results are not surprising, since fish already weakened by parasites would undoubtedly be less able to tolerate other environmental stresses. The nature and degree of parasitization offish clearly must be considered in bioassays and in studies of effects of contaminants on fish and shellfish species. Looking next at the reverse viewpoint, the influence of contaminants on parasite prevalence, definitive information is less readily available for marine species, but some information is available for freshwater species. Thermal loading was as- sociated with changes in the distribution and abundance of two larval trematodes in mos- quitofish (Aho et al. 1976). Similarly, thermal loading from a nuclear power plant was directly correlated with incidence of the ciliate Epistylis sp. and the bacterium Aeromonas liquifaciens ( = A. hydrophila) in six species of centrarchids in South Carolina (Esch et al. 1976). Effects of ther- SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES mal effluents on parasitism of largemouth bass, Micropterus salmoides, by the acanthocephalan Neoechinorhynchus cylindratus were examined by Eure and Esch ( 1974). Parasite densities were sig- nificantly higher in fish from heated water during the winter months, a possible reflection of greater densities of larval parasites and intermediate host populations in the effluent. River pollution from domestic and industrial sources was considered to be a contributing factor in increased parasite bur- dens found in fish from areas of heaviest pollution in Poland (Dabrowska 1974). For marine species, good evidence relating pol- lutants with changes in parasite abundance is scarce. Results of an extensive survey of external parasites and disease conditions in North Sea fish (Moller''') did not disclose clear-cut relationships between parasitism and pollution, although the higher prevalence of vibriosis and lymphocystis in southern sectors which are most polluted indi- cated a possible influence of pollution. Other fac- tors seemed responsible for differential abun- dances reported for the larger external parasites. Several parasites of estuarine fishes from the Gulf of Mexico were examined by Overstreet and Howse (1977) in a search for associations with environmental pollution. Samples of Atlantic croaker were collected in 1970-72, and again in 1975. Large variations in prevalences of helminth parasites occurred, but clear associations with pol- lutants and changes in pollutant levels were not established. A myxosporidan protozoan seemed to be more promising. Infections of sheepshead min- nows by Myxobolus lintoni were very abundant in one polluted bayou of Mississippi, but were absent in seemingly healthy habitats. The stalked peritrich ciWdiie Epistylis sp., men- tioned in an earlier section in connection with fin erosion and red sores, seems to be related to high organic content and possibly other stresses in freshwater and brackish water habitats. The ciliate, together with secondary bacterial invaders (principally Aeromonas liquifaciens (= A. hy- drophila), produces a hemorrhagic hyperplastic condition beneath the scales that is referred to as red sore (Overstreet and Howse 1977). The ciliate infests a wide range offish species in low salinity waters of Mississippi, especially centrarchids, "Moller, H. 1977. Distribution of some parasites and dis- eases of fishes from the North Sea in February, 1977. Int. Counc. Explor. Sea, Fish. Improv. Comm., Doc. CM1977/E:20, 16 P- sheepshead, and black drum (the drum is a marine invader in brackish water). Secondary bacterial infections associated with the ciliate m.ay become systemic, and mortalities may result. In addition to field observations, there is some experimental evidence for a causal relationship between specific pollutant chemicals and fungus parasitization offish and shellfish. In one study, oysters exposed to pesticides (DDT, Toxaphene, and parathion) became infected with a mycelial fungus that caused lysis of the mantle, gut, gonads, gills, visceral ganglion, and kidney tubules (Lowe et al. 1971). None of the control oysters became infected, indicating a role for one or several of the pesticides in altering the host- parasite relationships of the oysters and the fun- gus. Presence of fungus infections made it difficult to differentiate histopathological effects of pes- ticide exposure from those due to the parasite. CONCLUSIONS In considering pollution-associated diseases of fish and shellfish, a number of conclusions seem warranted: 1. Environmental stress from pollutants seems to be an important determining factor in sev- eral fish and shellfish diseases. Effects in- clude direct chemical-physical damage to cell membranes or tissues, modification of physiological and biochemical reactions, in- creased infection pressure from facultative microbial pathogens, and reduced resistance to infection. 2. The multifactorial genesis of disease in marine species is becoming apparent, involv- ing environmental stress, facultative patho- gens, resistance of hosts, and latent infec- tions. 3. Some circumstantial evidence for the role of environmental carcinogens in the etiology of neoplasms offish and shellfish is accumulat- ing, but at present definitive conclusions are not justified. 4. The presence of marginal or degraded estuarine/coastal environments may be sig- nalled by the appearance of, or the increase in prevalence of a number of diseases, includ- ing fin erosion, "red sores," ulcers, and possi- bly lymphocystis in fish; by "shell disease" in crustaceans; and by certain neoplasms in bivalve molluscs, but an absolute cause and 741 FISHERY BULLETIN: VOL. 76, NO. 4 effect relationship has not yet been de- monstrated for most of these diseases. 5. Among the most severe and persistent prob- lems in establishing pollutant-disease rela- tionships are: the absence of baseline in- formation about the organisms and their habitats prior to pollution, the existence of multiple pollutants in many badly degraded waters, and the circumstantial nature of much of the evidence linking pollution and disease. 6. A number of viruses have been found in crus- taceans and molluscs in recent years, and the pathogenic role of two of them (shrimp Baculovirus and oyster Herpesvirus) has been demonstrated by exposure to increasing environmental stress. Other latent virus in- fections of invertebrates may be identified by similar experimental methods. The evidence for an association of pollution and disease presented in this paper (except for results of experimental studies) is largely circumstantial. When confronted with the hard question "Can you state positively that the disease condition seen in natural populations is caused by specific environ- mental contaminants?", the answer at present has to be "No." However, the weight of this cir- cumstantial evidence, particularly for diseases such as fin erosion and ulcers, is such that it leads to the conclusion that associations do exist be- tween pollutants and disease. ACKNOWLEDGMENTS A number of people read drafts of this paper, and many changes and additions have resulted from their comments. I would like to acknowledge the assistance of R. Overstreet, A. Sparks, M. Sher- wood, R. Wolke, J. Couch, J. Pearce, A. Farley, A. Rosenfield, and R. Murchelano — without neces- sarily implying their agreement with any or all of the interpretations and conclusions in this paper. I would also like to thank K. Melkers for maintain- ing the continuity and accuracy of the manuscript through a series of revisions. LITERATURE CITED Aho, J. M., J. w. Gibbons, and g. w. esch. 1976. Relationship between thermal loading and parasitism in the mosquitofish. In G. W. Esch and R. W. McFarlane (editors), Thermal ecology II, p. 213-218. Technical Information Center, Energy Research and De- velopement Agency (CONF-750425). Aiken, D. E., and E. H. Byard. 1972. Histological changes in lobsters {Homarus americanus) exposed to yellow phosphorus. Science (Wash., D.C.) 176:1434-1435. ALPERS, C. E., B. B. McCain, M. S. Myers, and S. R. Well- INGS. 1977a. Lymphocystis disease in yellowfin sole (Limanda aspera) in the Bering Sea. J. Fish. Res. Board Can. 34:611-616. ALPERS, C. E., B, B. McCain, M. Myers, S. R. Wellings, M. POORE, J. BAGSHAVv', and C. J. DAWE. 1977b. Pathologic anatomy of pseudobranch tumors in Pacific cod, Gadus macrocephalus. J. Natl. Cancer Inst. 59:377-398. Aneer, G., and O. Ljungberg. 1976. Lymphocystis disease in Baltic herring iClupea harengus var. membras L.). J. Fish. Biol. 8:345-350. angell, C. L., b. S. Miller, and s. R. wellings. 1975. Epizootiology of tumors in a population of juvenile English sole (Parophrys vetulus) from Puget Sound, Washington. J. Fish. Res. Board Can. 32:1723-1732. Anonymous. 1951. Diseased stripers in Connecticut are safe to eat. Salt Water Sportsman, June 22, p.l. AWERINZEW, S. 1911. Studien liber parasitische Protozoen. V. Einige neue Befunde aus der Entwicklungsgeschichte von Lymphocys- tis johnstonei Woodc, Arch. Protistenkd. 22:179-196. Baker, J. T. P. 1969. Histological and electron microscopical observa- tions on copper poisoning in the winter flounder (Pseudo- pleuronectes americanus). J. Fish. Res. Board Can. 26:2785-2793. BARRY, M. M., AND P. P. YEVICH. 1972. Incidence of gonadal cancer in the quahaug {Mer- cenaria mercenaria). Oncology 26:87-96. 1975. The ecological, chemical and histopathological evaluation of an oil spill site. Part III. Histopathological studies. Mar. Pollut. Bull. 6:171-173. Barry, M. M., P. P. Yevich, and N. H. Thayer. 1971. Atypical hyperplasia in the soft-shell clam Mya arenana. J. Invertebr. Pathol. 17:17-27. BeNGT-SSON, B.-E. 1974. Vertebral damage to minnows Phoxinus phoxinus exposed to zinc. Oikos 25:134-139. 1975. Vertebral damage in fish induced by pollutants. In J.H. KoemanandJ. J. T. W. A. Strik (editors), Suble- thal effects of toxic chemicals on aquatic animals, p. 23- 30. Elsevier Sci. Publ. Co., Amsterdam. Bonnet. R. 1883. Studien zur Physiologie und Pathologie der Fische. Bayer. Fisch. Z. 8:79-81. Boyce, N. p., and S. B. YAMADA. 1977. Effects of a parasite, Eu both rium salvelini (Cestoda: Pseiidophyllidea), on the resistance of juvenile sockeye salmon, Oncorhynchiis nerka ,tozmc. J. Fish. Res. Board Can. 34:706-709. Brett, J. R, 1958. Implications and assessments of environmental stress. In P. A. Larkin (editor), The investigation of fish-power problems, p. 69-97, H. R. MacMillan Lee. Fish., Univ. B.C., Vancouver. 742 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES Brocksen, R. W., AND H. C. Bailey. 1973. The responses of juvenile chinook salmon and striped bass exposed to benzene, a water-soluble compo- nent of crude oil. Proceedings of the 1973 Conference, Prevention and Control of Oil Spills, Wash., D.C., p. 783- 791. Brooks, R. E., G. E. McArn, and S. R. Wellings. 1969. Ultrastructural observations on an unidentified cell t}T)e found in epidermal tumors of flounders. J. Natl. Cancer Inst. 43:97-109. Brown, R. S., R. E. Wolke, and S. B. Saila. 1976. A preliminary report on neoplasia in feral popula- tions of the soft-shell clam, Mya arenaria: prevalence, histopathology and diagnosis. Proc. First Int. Colloq. Invertebr. Pathol., Kingston, Ontario, Can., p. 151-159. BROWN, R. S., R. E, Wolke, S. B. Saila, and C. W. Brown. 1977. Prevalence of neoplasia in 10 New England popula- tions of the soft-shell clam {Mya arenaria). Ann. N.Y. Acad. Sci. 298:522-534. Calabrese, a., R. S. Collier, and J. E. Miller. 1974. Physiological response of the cunner.Taf/to^o/aftr;/.? adspersus, to cadmium. I. Introduction and experimental design. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-681, p. 1-3. Carlisle, J. G., Jr. 1969. Results of a six-year trawl study in an area of heavy waste discharge: Santa Monica Bay, California. Calif. Fish Game 55:26-46. Carmody, d. J., J. B. Pearce, and W. E. Yasso. 1973. Trace metals in sediments of New York Bight. Mar. Pollut. Bull. 4:132-135. Christensen, D. J., C. A. Farley, and F. G. Kern. 1974. Epizootic neoplasms in the clam Macoma balthica (L.) from Chesapeake Bay. J. Natl. Cancer Inst. 52:1739-1749. Christmas, J. Y., and H. d. howse. 1970. The occurrence of lymphocystis in Micropogon un- dulatus and Cynoscion arenarius from Mississippi es- tuaries. Gulf Res. Rep. 3:131-154. Clark, R. C, Jr. J. S. Finley, and G. G. Gibson. 1974. Acute effects of outboard motor effluent on two marine shellfish. Environ. Sci. Technol. 8:1009-1014. Cooper, R. C, and C. a. Keller. 1969. Epizootiology of papillomas in English sole, Paro- phrys vetulus. Natl. Cancer Inst. Monogr. 31:173-185. Cope, O. B., J. P. McCraren, and L. Eller. 1969. Effects of dichlobenil on two fishpond environments. Weed Sci. 17:158-161. Cope, O. B., e. m. wood, and g. H. Wallen. 1970. Some chronic effects of 2,4-D on the bluegill iLepomis macrochirus). Trans. Am. Fish. Soc. 99:1-12. Couch, j. a. 1969. An unusual lesion in the mantle of the American oyster, Crassostrea virginica. Natl. Cancer Inst. Monogr. 31:557-562. 1970. Sarcoma-like disease in a single specimen of the American oyster. [Abstr.] In R. M. Butcher (editor). Comparative leukemia research 1969. p. 647. Bibl. Haematol. 36. 1974a. Free and occluded virus, similar to Baculovirus, in hepatopancreas of pink shrimp. Nature (Lond.) 247:229-231. 1974b. An enzootic nuclear polyhedrosis virus of pink shrimp: Ultrastructure, prevalence, and enhance- ment. J. Invertebr. Pathol. 24:311-331. 1975. Histopathologic effects of pesticides and related chemicals on the livers of fishes. In W. E. Ribelin and G Migaki (editors). The pathology of fishes, p. 559- 584. Univ. Wis. Press, Madison. 1976. Attempts to increase Baculovirus prevalence in shrimp by chemical exposure. Prog. Exp. Tumor Res. 20:304-314. 1977. Ultrastructural .study of lesions in gills of a marine shrimp exposed to cadmium. J. Invertebr. Pathol. 29:267-288. 1978. Diseases, parasites, and toxic responses of commer- cial penaeid shrimps of the Gulf of Mexico and south Atlantic coasts of North America. Fish. Bull., U.S. 76:1-44. Couch, J. A., and L. Courtney. 1977. Interaction of chemical pollutants and virus in a crustacean: a novel bioassay system. Ann. N.Y. Acad. Sci. 298:497-504. Couch, j. a., and d. R. Nimmo. 1974a. Ultrastructural studies of shrimp exposed to the pollutant chemical, polychlorinated biphenyl (Aroclor 1254). Bull. Pharmacol. Environ. Pathol. 11:17-20. 1974b. Detection of interactions between natural patho- gens and pollutants in aquatic animals. In R. L. Am- borski, M. A. Hood, and R. R. Miller (editors), Proc. Gulf Coast reg. symp. diseases of aquatic animals, p. 261- 268. La. State Univ., Cent. Wetland Resour. Publ. No. LSU-SG-74-05. Couch, J., G. Gardner, J. C. Harshbarger, M. R. Tripp, and P. P. Yevich. 1974. Histological and physiological evaluations in some marine fauna. In Marine Bioassays Workshop Proceed- ings 1974, p. 156-173, Mar. Technol. Soc, Wash., D.C. Couch, J. A., J. T. Win.stead, and L. R. Goodman. 1977. Kepone-induced scoliosis and its histological con- sequences in fish. Science (Wash. D.C.) 197:585-587. Crandall, C. a., and C. j. Goodnight. 1963. The effects of sublethal concentrations of several toxicants to the common guppy, Lebistes re- ticulatus. Trans. Am. Microsc. Soc. 82:59-73. Dabrowska, H, 1974. An attempt to evaluate the state of health of fish from the Lyna and Walsza Rivers in connection to their pollution. [In Pol] Przegl.Zool. 18(3):390-395. Abstract by S.F.Snieszko,1977,/« Fish Health News 6(3):156-157. Dawe, C. j., and j. C. Harshbarger (editors). 1969. Neoplasms and related disorders of invertebrate and lower vertebrate animals. Natl. Cancer Inst. Monogr. 31, 772 p. Dawe, C. J., D. G. Scarpelli, and S. R. Wellings (editors). 1976. Tumors in aquatic animals. Prog. Exp. Tumor Res. 20,411 p. Dawson, C. E. 1964. A bibliography of anomalies of fishes. Gulf Res. Rep. 1:308-399. 1966. A bibliography of anomalies of fishes. Suppl. 1, Gulf Res. Rep. 2:169-176. 1971. A bibliography of anomalies of fishes. Suppl. 2, Gulf Res. Rep. 3:215-239. Dawson, C. E., and e. Heal. 1976. A bibliography of anomalies of fi.shes. Suppl. 3, Gulf Res. Rep. 5:35-41. 743 FISHERY BULLETIN: VOL. 76, NO. 4 DOUDOROFF, P. 1957. Water quality requirements of fishes and effects of toxic substances. In M. E. Brown (editor), The physiol- ogy of fishes, Vol. II, p. 403-4.30. Acad. Press, N.Y. DUKE, T. W., .AND A. J. Wilson, Jr. 1971. Chlorinated hydrocarbons in livers of fishes from the northeastern Pacific Ocean. Pestic. Monit. J. 5:228-232. Duke, T. W., J. I. Lowe, and A. J. Wilson, Jr. 1970. A polychlorinated biphenyl (Aroclor 1254*) in the water, sediment, and biota of Escambia Bay, Flori- da. Bull. Environ. Contam. Toxicol. 5:171-180 DUKES, T. W., AND A. R. LAWLER. 1975. The ocular lesions of naturally occurring lym- phocystis in fish. Can. J. Comp. Med. 39:406-410. Edwards, R. H., and R. M. Overstreet. 1 976. Mesenchymal tumors of some estuarine fishes of the northern Gulf of Mexico. I. Subcutaneous tumors, proba- bly fibrosarcomas, in the striped mullet, Mugil cephalus. Bull. Mar. Sci. 26:33-40. Eisler, R., and p. H. Edmunds. 1966. Effects of endrin on blood and tissue chemistry of a marine fish. Trans. Am. Fish. Soc. 95:153-159. ELLER, L. L. 1971. Histopathologic lesions in cutthroat trout (Salmo clarki) exposed chronically to the insecticide en- drin. Am. J. Pathol. 64:321-336. EscH, G. W., J. W. Gibbons, and J. E. Bourque. 1975. An analysis of the relationship between stress and parasitism. Am Midi. Nat. 93:339-353. E.SCH, G. W., T. C. HAZEN, R. V. DIMOCK, JR., AND J. W. GIB- BONS. 1976. Thermal effluent and the epizootiology of the ciliate Epistylis and the bacterium Aeromonas in association with centrarchid fish. Trans. Am. Microsc. Soc. 95:687- 693. EURE, H. E., AND G. W. ESCH. 1974. Effects of thermal effluent on the population dynamics of helminth parasites in largemouth bass. In J. W. Gibbons and R. R. Sharitz (editors). Thermal ecol- ogy, p. 207-215. U.S. At. Energy Comm. Symp. Ser. (Conf. 730505). Fange, R., C. Rappe, and S. FALKMER. 1975. Liver tumours in the Atlantic hagfish (Myxine glutinosa L.). In J. H. Koeman and J. J. T. W. A. Strik (editors). Sublethal effects of toxic chemicals on aquatic animals, p. 103-109. Elsevier Sci. Publ. Co.. Amster- dam. FALKMER, S., S. O. EMDIN, Y. dSTERBERG, A. MATTISSON, M. L. JOHANSSON SJOBECK, AND R. FANGE. 1976. Tumor pathology of the hagfish, Myxine glutinosa, and the river lamprey, Lampetra fluviatilis. A light- microscopical study with particular reference to the oc- currence of primary liver carcinoma, islet-cell tumors, and epidermoid cysts of the skin. Prog. Exp. Tumor Res. 20:217-250. FALKMER, S., S. MARKLUND, P. E. MATTSSON, AND C. RAPPE. 1977. Hepatomas and other neoplasms in the Atlantic hagfish (Myxine glutinosa): a histopathologic and chemi- cal study. Ann. N.Y. Acad. Sci. 298:342-355. Farley, C. A. 1969a. Probable neoplastic disease of the hematopoietic system in oysters, Crassostrea virginica and Crassostrea gigas. Natl. Cancer Inst. Monogr. 31:541-555. 1969b. Sarcomatoid proliferative disease in a wild popula- tion of blue mussels iMytilus edulis). J. Natl. Cancer Inst. 43:509-516. Farley, C. A., and A. K. Sparks. 1970. Proliferative diseases of hemocytes, endothelial cells and connective tissue cells in mollusks. In R. M. Dutcher (editor). Comparative leukemia research 1969, p. 610-617 Bibl. Haematol. 36. Farley, C. A., W. G. Banfield, G. Kasnic, Jr., and W. S. Foster. 1972. Oyster herpes-type virus. Science (Wash., D.C.) 178:759-760. Ferm, V. H., AND S. J. Carpenter. 1967. Developmental malformations resulting from the administration of lead salts. Exp. Mol. Pathol. 7:208- 213. Frierman, E. M. 1976. Occurrence of hematopoietic neoplasms in Virginia oysters (Crassostrea virginica). Mar. Fish. Rev. 38(10):34-36. Fries, C, and M. R. Tripp. 1976. Effects of phenol on clams. Mar. Fish. Rev. 38(10):10-11. FUJIYA, M. 1960. Studies on the effects of copper dissolved in sea water on oyster. Bull. Jpn. Soc. Sci. Fish. 26:462-468. Gabriel, M. L. 1944. Factors affecting the number and form of vertebrae in Fundulus heteroclitus. J. Exp. Zool. 95:105-147. Galloway, J. N. 1972. Man's alteration of the natural geochemical cycle of selected trace-metals. PhD. Thesis, Univ. California, San Diego, 151 p. Gardner, G. R. 1975. Chemically induced lesions in estuarine or marine teleosts. In W. E. Ribelin and G. Migaki (editors), The pathology of fishes, p. 657-693. Univ. Wisconsin Press, Madison. Gardner, G. R., and G. LaRoche. 1973. Copper induced lesions in estuarine teleosts. J. Fish. Res. Board Can. 30:363-368. Gardner, G. R., and P. P. Yevich. 1970. Histological and hematological responses of an es- tuarine teleost to cadmium. J. Fish. Res. Board Can. 27:2185-2196. Gardner, G. R., P. P. Yevich, and P. F. Rogerson. 1975. Morphological anomalies in adult oyster, scallop, and Atlantic silversides exposed to waste motor oil. In Proceedings of the 1975 Conference, Prevention and Con- trol of Oil Pollution, San Francisco, Calif., p. 473- 477. Am. Pet. Inst. Goncharov, G. D., and V. R. Mikyakov. 1971. The effect of low concentrations of phenol on anti- body formation in carp, Cyprinus carpio. Probl. Aquatic Toxicol. 1970, p. 171-175. (Transl. U.S. Bur. Sport Fish. Wildl., Wash., D.C.) GOOD, H. V. 1940. A study of an epithelial tumor of Parophrys vet- ulus. M.S. Thesis, Univ. Washington, Seattle, 48 p. GOPALAN, U. K., AND J. S. YOUNG. 1975. Incidence of shell disease in shrimps in the New York Bight. Mar. PoUut. Bull. 6:149-153. 744 SIN'nERMANN POLUTION-ASSOCIATED DISEASES AND ABNORMALITIES GORDON, M. 1954. The genetics offish diseases. Trans. Am. Fish. Soc. 83:229-240. Hansen, D. J., P. R. Parrish. J. I. Lowe, A. J. Wil.son. Jr., AND p. D. Wilson. 1971. Chronic toxicity, uptake, and retention of Aroclor " 1254 in two estuarine fishes. Bull. Environ. Contam. Toxicol. 6:113-119. Hemphill, F. E., M. L. Kaeberle, and W. B. Buck. 1 97 1 . Lead suppression of mouse resistance to Salmonella typhimurium. Science (Wash., D.C.) 172:1031-1032. HICKEY, C. R.. JR. 1972. Common abnormalities in fishes, their causes and effects. N.Y. Ocean Sci. Lab. Tech. Rep. 0013, 20 p. HUBBS, C. 1959. High incidence of vertebral deformities in two natural populations of fishes inhabiting warm springs. Ecology 40:154-155. HI'BERMAN, E. 1975. Mammalian cell transformation and cell-mediated mutagenesis by carcinogenic polycyclic hydrocarbons. Mutat. Res. 29:285-291. IVANOV, V. N. 1967. Effects of radioactive substances on the embryonic development of fish. In Problems of biological oceanog- raphy, p. 185. AEC-TR-6940. Jeffries, H. P. 1972. A stress syndrome in the hard clam, Mercenaria mercenaria. J. Invertebr. Pathol. 20:242-251. Jensen, N. J., and J. L. Larsen. 1976. Forskningsprojet vedrorende marine vib- riopopulationers for hold til dratforurening i marine biotoper. Dans. Vet. Tidsskr. 59:521-524. 1977. Sarsygdom hos torsk i danske kystnaere omrader med forskellig vandkvalitet. Nord. Veterinaermed. 29(Suppl. 11:4. JONES, J. R. E. 1938. The relative toxicity of salts of lead, zinc and copper to the stickleback iGasterosteus aculeatus L.I and the ef- fect of calcium on the toxicity of lead and zinc salts. J. Exp. Biol. 15:394-407. 1940. A study of the zinc-polluted river Ystwyth in north Cardiganshire, Wales. Ann. Appl. Biol. 27:368-378. JONES, R. H., R. L. WlI.l.IA.MS. AND A. M. JONES. 1971. Effects ofheavy metal on the immune response. Pre- liminary findings for cadmium in rats. Proc. Soc. Exp. Biol. Med. 137:1231-1236. JOHNSON, D. W. 1968. Pesticides and fishes — A review of selected litera- ture. Trans. Am. Fish. Soc. 97:398-424. JOHNSTONE, J. 1905. Internal parasites and disease conditions of fishes. Rep. Lanes. Sea-Fish. Labs. 1904. No. 13, p. 98-120. Kelly, D. L. 1971. Epidermal papilloma in the English sole. Parophrys vetulus. Ph.D. Thesis, Univ. California, Berkeley. 139 p. KHAN, A., AND J. M. HILL. 1971. Immunosuppression with (/.s-platinum (II) diam- minodichloride: effect on antibody plaque-forming spleen cells. Infect. Immun. 4:320-321. KiNNE. O., AND H. Rosenthal. 1967. Effects of sulfuric water pollutants on fertilization, embryonic development and larvae of the herring, Clupea harengus. Mar. Biol. (Berl.i 1:65-83. KOEMAN, J. H., AND J. J. T. W. A. STRIK (editors). 1975. Sublethal effects of toxic chemicals on aquatic ani- mals. Elsevier Sci. Publ. Co., Amsterdam, 234 p. Koi.LER. L. D. 1973. Immunosuppression produced by lead, cadmium, and mercury. Am. J. Vet. Res. 34:1457-1458. Kolomiitseva, M. G., F. M. Voznesenskaya, and E. a. I.SAEVA. 1969. Immunobiological reactivity of the animal organism in association with different copper content in the ra- tions. J. Microbiol. Epidemiol. Immunobiol. 46(8i:99- 104. Komada, N. 1974. Studies on abnormality of bones in anomalous "Ayu", Plecoglossus altivelis Gyobo Kenkyu 8:127-135 (In Jpn., Engl. Summ.). KRA.NTZ, G. E. 1970. Lymphocy.stis in striped bass, Roccus saxatilis, in Chesapeake Bay. Chesapeake Sci. 11:137-139. Kraybill, H. F., C. J. Dawe, J. C. Harshbarger, and R. G. TARDIFF (editors). 1977. Aquatic pollutants and biologic effects with em- phasis on neoplasia. Ann. N.Y. Acad. Sci. 298, 604 p. La.molet, L., C. Chepeau, and B. Rea. 1976. Etude sur les pollutions le long du littoral francais de sud de la merdu nord. Inst. Sci. Tech. Peches. Marit., Dunkerque, 81 p. L.\R0CHE, G. 1972. Biological effects of short-term exposures to hazard- ous materials. In Proc. 1972 Natl. Conf Contr. Hazard- ous Mater. Spills, U.S. Environ. Prot. Agency, p. 199-206. Larsen, J. L., and N. J. Jen.sen. 1977a. An Aeromonas species implicated in ulcer-disease of the cod iGadus morhua\. Nord. Veterinaermed. 29:199-211. 1977b. De mikrobiologiske forholdved "ulcus-syndromet" hos torsk og relationeme til mikrobiologien i kystnaere omrader. Nord. Veterinaermed. 29(Suppl. 1):5. LEHNER, p. N., and A. EGBERT. 1969. Dieldrin and eggshell thickness in ducks. Nature (Lond.) 224:1218-1219. Levin, m. A.. R. E. wolke, and V. J. Cabelli. 1972. Vibrio anguillaritm as a cause of disease in winter flounder, iPseudopleuronectes americanus). Can. J. Mi- crobiol. 18:1585-1592. Lk;htnkr, D. V. 1974. Case reports of ossifying fibromata in the striped mullet. J. Wildl. Dis. 10:317-320. LOOGKWOOD, A. P. M. (editor). 1976. Effects of pollutants on aquatic organisms. Camb. Univ. Press, Camb. 193 p. Lloyd. R. 1960. The toxicity ofzinc sulphate to rainbow trout. Ann. Appl. Biol. 48:84-94. 1961a. Effect of dissolved oxygen concentrations on the toxicity of several poisons to rainbow trout {Salmo gairdnerii Richardson). J. Exp. Biol. 38:447-455. 1961b. The toxicity of mixtures of zinc and copper sul- phates to rainbow trout {Salmo gairdnerii Richardson). Ann. Appl. Biol. 49:535-538. Longwell, a. C. 1976a. Chromosome mutagenesis in developing mackerel eggs sampled from the New York Bight. Am. Soc. Lim- nol. Oceanogr., Spec. Symp. 2:337-339. 745 FISHERY BULLETIN: VOL. 76, NO. 4 1976b. Chromosome mutagenesis in developing mackerel eggs sampled from the New York Bight. U.S. Dep. Commer., NOAA Tech. Memo, ERL-MESA-7, 61 p. LoWK, J. I. 1966. Some effects of endrin on estuarine fishes. South- east Assoc. Game Fish Comm., Proc. 19th Annu. Conf , p. 271-276. Lowe, J. 1., P. R. P.xrrish, J. M. Patrick. Jr., .^nd J. Fores- ter. 1972. Effects of the polychlorinated biphenyl Aroclor" 1254 on the American oyster, Crastiostrca virgini- ca. Mar. Biol. (Berl.) 17:209-214. Lowe, J. L,P. D. Wilson, A. J. Rick, and A. J. Wilson, Jr. 1971. Chronic exposure of oysters to DDT, Toxaphene and Parathion. Proc. Natl. Shellfish. Assoc. 61:71-79. LUCKE, B., AND H. SCHLUMBEROER. 1941. Transplantable epitheliomas of the lip and mouth of catfish. J. Exp. Med. 74:397-408. MacIntyre, F. 1974. The top millimeter of the ocean. Sci. Am. 230(5):62-77. MAHONEY, J. B., F. H. MlDLIOE, AND D C. DEUEL. 1973. A fin rot disease of marine and euryhaline fishes in the New York Bight. Trans. Am. Fish. Soc. 102:596-605. MALEVSKI, Y., J. H, WALES, AND M. W. MONTGOMERY. 1974. Liver damage in rainbow trout {Salmogairdneri I fed cyclopropenoid fatty acids. J. Fish. Res. Board Can. 31:1397-1400. MaNN, H. 1970. Uber den Befall der Plattfische der Nordsee mit Lymphocystis. Ber. Dtsch. Wiss. Komm. Meeresforsch. 21:219-223. MATSUSATO, T. 1973. On the skeletal abnormalities in marine fishes. \. The abnormal marine fishes collected along the coast of Hiroshima Pref Bull. Nansei Reg. Fish. Res. Lab. 6, p. 17-58. MCARN, G. E., AND S. R. WELLINGS. 1971. A comparison of skin tumors in three species of flounders. J. Fish. Res. Board Can. 28:1241-1251. McArn, G. E., R. G. Chuinard, B. S. Miller, R. E. Brooks, and s. r. wellings. 1968. Pathology of skin tumors found on English sole and starry flounder from Puget Sound, Washington. J. Natl. Cancer Inst. 41:229-242. McCaull, J. 1971. Building a shorter life. Environment (St. Louis) 13(7):2-15. Mearns, a. J., AND M. J. Sherwood. 1974. Environmental aspects of fin erosion and tumors in southern California Dover sole. Trans. Am. Fish. Soc. 103:799-810. 1977. Distribution of neoplasms and other diseases in marine fishes relative to the discharge of waste wa- ter. Ann. N.Y. Acad. Sci. 298:210-224. MiDLiGE, F. H., Jr., and R. G. Malsberger. 1968. In vitro morphology and maturation of lymphocystis virus. J. Virol. 2:830-835. Miller, b. s., and s. R. Wellings. 1971. Epizootiology of tumors on flathead sole (Hippoglos- soides elassodon) in East Sound, Orcas Island, Washington. Trans. Am. Fish. Soc. 100:247-266. Minchew, C. D., and J. D. Yarbrough. 1977. The occurrence of fin rot in mullet (Miigi! cephalus) associated with crude oil contamination of an estuarine pond-ecosystem. J. Fish. Biol. 10:319-323. MI.X, M. C, H. J. PRIBBLE, R. T. RILEY, AND S. P. TOMASOVIC. 1977. Neoplastic disease in bivalve mollusks from Oregon estuaries with emphasis on research on proliferative dis- orders in Yaquina Bay oysters. Ann. N.Y. Acad. Sci. 298:356-373. Mottle Y, C. McM. 1937. The number of vertebrae in trout (Sa/77io). J.Biol. Board Can. 3:169-176. MURCHELANO, R. A. 1975. The histopathology of fin rot disease in winter floun- der from the New York Bight. J.Wildl.Dis. 11:263-268. MURCHELANO, R. A., AND D. W. BRIDGES. 1976. Lumphocystis disease in the winter flounder, Pseudopleuronecten amencanus. J. Wildl. Dis. 12:101- 103. MURCHELANO, R. A., AND J. ZlSKOWSKI. 1976. Fin rot disease studies in the New York Bight. Am. Soc. Limnol. Oceanogr., Spec. Symp. 2:329-336. NAKAI,Z.,M. KOSAKA,S. KUD0H,A.NAGAI,F. HAYASHIDA.T. KUBOTA, M. OGURA, T. MIZUSHIMA, AND I. UOTANI. 1973. Summary report on marine biological studies of Suruga Bay accomplished by Tokai University 1964-72. J. Fac. Mar, Sci. Technol. Tokai Univ. 7:63- 117. NEWMAN, M. W., AND S, A. MACLEAN. 1974. Physiological response of the cunner, Tautogolabrus adspersus, to cadmium. VI. Histopathology. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-681, p. 27-33. NIGRELLI, R. F,, K. S. KETCHEN, AND G. D. RUGGIERI. 1965. Studies on virus diseases of fishes. Epizootiology of epithelial tumors in the skin of flatfishes of the Pacific coast, with special reference to the sand sole iPsettichthys melanosticus) from northern Hecate Strait, British Co- lumbia, Canada. Zooligica (NY.) 50:115-122. NiMMo, D. R., R. R. Blackman, a. J. Wilson, Jr., and J. FORESTER. 1971. Toxicity and distribution of Aroclor *1254 in the pink shrimp, Penaeus diiorarum. Mar. Biol. (Berl.) 11:191-197. NiMMO, D. R., D. V. LIGHTNER, AND L, H. BAHNER, 1977, Effects of cadmium on the shrimps, Penaeus duorarum, Palaeomonetes pitgio and Palaemonetes vul- garis. In F. J, Vernberg, A, Calabrese, F. P. Thurberg, and W. B. Vernberg (editors). Physiological responses of marine biota to pollutants, p. 131-183. Acad. Press, N.Y. NORDENBERG, C. 1962. Das Vorkommen der Lymphocystiskrankheit bei Scholle und Plunder in Oresund. K. Fysiogr. Sallsk. Lund Forh. 43:17-26. OisHi, K., F. Yamazaki, and T. Harada. 1976. Epidermal papillomas of flatfish in the coastal wa- ters of Hokkaido, Japan. J. Fish. Res. Board Can. 33:2011-2017. OVERSTREET, R. M., AND H. D. HOWSE. 1977. Some parasites and diseases of estuarine fishes in polluted habitats of Mississippi. Ann. N.Y. Acad. Sci. 298:427-462. PACIS, M. R. 1932. An epithelial tumor of Parophrys vetulus. M.S. Thesis, Univ. Washington, Seattle, 94 p. Parry, j. m., d. j. Tweats, and m. a. j. al-mossawi. 1976. Monitoring the marine environment for mutagens. Nature (Lond.) 264:538-540. 746 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES Pascoe, D., AND P. Cram. 1977. The effect of parasitism on the toxicity of cadmium to the three-spined stickleback, Gasterosteus aculeatus L. J. Fish. Biol. 10:467-472. Pauley, G. B. 1969. A critical review of neoplasia and tumor-like lesions in mollusks. Natl. Cancer Inst. Monogr. 31:509-539. Peakall, D. B., and J. L. Linger. 1970. Polychlorinated biphenyls, another long-life wide- spread chemical in the environment. Bioscience 20:958-964. PEARCE, J. B. 1972. The effects of solid waste disposal on benthic com- munities in the New York Bight. In M. Ruivo (editor), Marine pollution and sea life, p. 404-411. Fishing News (Books) Ltd., Lond. Perevozchenko, L L, and C. N. Davydov. 1974. DDT and its metabolites in some cestodes in fishes. [In Russ.] Gidrobiol. Zh. 10(61:86-90. Perkins, E. J., J. R. S. Gilchrist, and O. J. Abbott. 1972. Incidence of epidermal lesions in fish of the North- East Irish Sea area, 1971. Nature (Lond.) 238:101-103. PERLMUTTER, a., D. a. SAROT, M.-L. YU, R. J. FlLAZZOLA, AND R. J. Seeley. 1973. The effect of crowding on the immune response of the blue gourami, Trichogaster trichopterus, to infectious pancreatic necrosis (IPN) virus. Life Sci. 13:363-375. Pichirallo, J. 1971. PCB's: Leaks of toxic substances raises issue of ef- fects, regulation. Science (Wash., D.C.I 173:899-904. PiPPY, J. H. C, AND G. M. Hare. 1969. Relationship of river pollution to bacterial infection in salmon (Salmo salar) and suckers iCatostomus com- mersoni). Trans. Am. Fish. Sec. 98:685-690. PREHN, R. T. 1971. Neoplasia. In M. F. LaVia and R. B. Hill, Jr. (editorsi. Principles of pathobiology, p. 191-241. Oxford Univ. Press, N.Y. RENTCHNICK, P. (editor). 1976. Neoplasms in aquatic animals as indicators of en- vironmental carcinogens. UICC (Union Int. Cancer) Bull. 14(2):4-5. RIBELIN, W. E., and G. MIGAKI (editors). 1975. The pathology of fishes. Univ. Wis. Press, Madi- son, 1004 p. ROBOHM, R. A., AND C. BROWN. 1977. A new bacterium (presumptive Vibrio species) caus- ing ulcers in flatfish. Abstracts of the 2d Annual Eastern Fish Health Workshop, Easton, Md., Abstr. No. 3, 2 p. ROBOHM, R. A., AND M. F. NITKOWSKI. 1974. Physiological response of the cunner, Tautogolabrus adspersus, to cadmium. IV. Effects on the immune sys- tem. U.S. Dep. Commer., NCAA Tech. Rep. NMFS SSRF-681, p. 15-20. R0DSAETHER, M. C, J. OLAFSEN, J. RAA, K. MYHRE, AND J. B. Steen. 1977. Copper as an initiating factor of vibriosis {Vibrio anguillarum) in eel (Anguilla anguilla). J. Fish. Biol. 10:17-21. Rogers, w. a. 1 970 . A summary of fish disease cases received a five-year period at the Southeastern Cooperative Fish Disease Laboratory. Proc. 23rd Annu. Conf Southeast Assoc. Game Fish Comm., p. 353-358. 1972. Disease in fish due to protozoan Epistylis (Ciliata: Peritricha) in the Southeastern U.S. Proc. 25th Annu. Conf. Southeast Assoc. Game Fish Comm., p. 493-496. ROMASHOV, D. D., AND V. N. BELYAYEVA. 1966. About constancy of radiation damage chromosomes in fish embryogenesis. Genetika 4:4-15. Rosen, B. 1970. Shell disease of aquatic crustaceans. In S. F. Snieszko (editor), A symposium on diseases of fishes and shellfishes, p. 409-415. Am. Fish. Soc. Spec. Publ. 5. Rosenthal, H., and D. F. Alderdice. 1976. Sublethal effects of environmental stressors, natural and pollutional, on marine fish eggs and lar- vae. J. Fish. Res. Board Can. 33:2047-2065. Rosenthal, H., and K.-R. Sperling. 1974. Effects of cadmium on development and survival of herring eggs. In J. H. S. Blaxter (editor). The early life history offish, p. 383-396. Springer-Verlag, Berl. Rosenthal, H., and R. Stelzer. 1970. Wirkungen von 2,4- und 2,5-dinitrophenol auf die Embryonalentwicklung des Herings Clupea haren- gus. Mar. Biol. (Beri. I 5:325-336. Rosenthal, H. L., and R. S. Rosenthal. 1950. Lordosis, a mutation in the guppy. J. Hered. 41:217-218. RUIVO, M. (editor! 1972. Marine pollution and sea life. Fishing News (Books) Ltd., Lond., 624 p. Russell, F. E., and P. Kotin. 1957. Squamous papilloma in the white croaker. J.Natl. Cancer Inst. 18:857-861. Ryder, J. A. 1887. On a tumor in the oyster. Proc. Acad. Nat. Sci. Phila. 39:25-27. Schlotfeldt, H.-J. 1972. Jahreszeitliche Abhangigkeit der "Schwarzflecken- krankheit" bei der Garnele, Crangon crangon L. Ber. Dtsch. Wiss. Komm. Meeresforsch. 22:397-399. Schmidt, T. T., R. w. Risebrough, and F. Gress. 1971. Input of polychlorinated biphenyls into California coastal waters from urban sewage outfalls. Bull. Envi- ron. Contam. Toxicol. 6:235-243. SELYE, H. 1950. Stress and the general adaptation syndrome. Br. Med. J. 1:1383-1392. 1952. The story of the adaptation syndrome. Acta. Inc., Montreal Que., 225 p. Shelton, R. G. J., AND K. W. Wilson. 1973. On the occurrence of lymphocystis, with notes on other pathological conditions, in the flatfish stocks of the north-east Irish Sea. Aquaculture 2:395-410. Sherwood, M. J., and A. J. Mearns. 1976. Occurrence of tumor-bearing Dover sole (Micros- tomas pacificus ) off Point Arguello, California and ofFBaja California, Mexico. Trans. Am Fish. Soc. 105:561-563. 1977. Environmental significance of fin erosion in south- em California demersal fishes. Ann. N.Y. Acad. Sci. 298:177-189. SINDERMANN, C. J. 1958. An epizootic in Gulf of Saint Lawrence fishes. Trans. 23rd North Am. Wildl. Conf, p. 349-360. 1968. Oyster mortalities, with particular reference to Chesapeake Bay and the Atlantic coast of North America. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 569, 10 p. 747 FISHERY BULLETIN: VOL. 76, NO. 4 1970. Principal diseases of marine fish and shellfish. Acad. Press, N.Y., 369 p. 1976. Effects of coastal pollution on fish and fisheries — with particular reference to the Middle Atlantic Bight. Am. Soc. Limnol. Oceanogr., Spec. Symp. 2:281- 301. SINDERMANN,C. J., J. J. ZISKOWSKI, AND V.T.ANDERSON, Jr. 1978. A guide for the recognition of some disease condi- tions and abnormalities in marine fish. Natl. Mar. Fish. Serv., Northeast Fish. Cent., Tech. Ser. Rep. 14, 65 p. SKIDMORE, J. F. 1964. Toxicity of zinc compounds to aquatic animals, with special reference to fish. Q. Rev. Biol. 39:227-248. 1970. Respiration and osmoregulation in rainbow trout with gills damaged by zinc sulphate. J. Exp. Biol. 52:481-494. SNIESZKO, S. F. 1962. The control of bacterial and virus diseases of fishes. In Biological problems in water pollution, 3rd Semin., p. 281-282. U.S. Pub. Health Serv. Publ. 999-WP-25. 1974. The effects of environmental stress on outbreaks of infectious diseases of fishes. J. Fish. Biol. 6:197-208. SNIESZKO, S. F. (editor). 1970. A symposium on diseases of fishes and shellfishes. Am. Fish. Soc. Spec. Publ. 5, 526 p. S0RENSEN, U. S. 1977. Praeliminaere serologiske unders0gelser over Aeromonas-og V;6no-stammer associeret med ulcus syn- dromeincod. Nord. Veterinaermed.29(Suppl. 1):10-12. SPARKS, A. K. 1972. Invertebrate pathology. Noncommunicable dis- eases. Acad. Press, N.Y., 387 p. STAINKEN, D. M. 1975. The effect of oil pollution on the physiology of A/va arenaria, with special reference to the effect upon respira- tion and mode of accumulation. Ph.D. Thesis, Rutgers Univ., Newark, N.J. 160 p. Stick, H. F., and A. B. Acton. 1976. The possible use of fish tumors in monitoring for carcinogens in the marine environment. Prog. Exp. Tumor Res. 20:44-54. Stich, H. F., a. B. Acton, and C. R. Forrester. 1976. Fish tumors and sublethal effects of pollutants. J. Fish. Res. Board Can. 33:1993-2001. Stich, H. F., A. B. Acton, K. Oishi, F. Yamazaki, T. Harada, T. HIBINO, AND H. G. MOSER. 1977. Systematic collaborative studies on neoplasms in marine animals as related to the environment. Ann. N.Y. Acad. Sci. 298:374-388. STREET,J. C, AND R. P. SHARMA. 1975. Alteration of induced cellular and humoral immune responses by pesticides and chemicals of environmental concern: quantitative studies of immunosuppression by DDT, Aroclor 1254, carbaryl, carbofuran, and methyl- parathion. Toxicol. Appl. Pharmacol. 32:587-602. Telford, M. 1968. The effects of stress on blood sugar composition of the lobster, Homarus americanus. Can. J. Zool. 46:819- 826. Telford, M. 1974. Blood glucose in crayfish — II. Variations induced by artificial stress. Comp. Biochem. Physiol. 48A:555-560. TEMPLEMAN, W. 1965. Lymphocystis disease in American plaice of the eastern Grand Bank. J. Fish. Res. Board Can. 22:1345- 1356. Udey, L. R., E. Young, and B. Sallman. 1977. Isolation and characterization of an anaerobic bac- terium, £t;6ac?en(/w tarantellus sp.nov., associated with striped mullet (Mugil cephalus) mortality in Biscayne Bay, Florida. J. Fish. Res. Board Can. 34:402-409. Ueki, N., and T. Sugiyama. 1976. On abnormality of bones in artificially produced ayu, Plecoglossus altevelis (Preliminary Note). [In Jpn.] Bull. Fish. Exp. Stn. Okayama Pref. 50:360-362. Valentine, D. W. 1975. Skeletal anomalies in marine teleosts. In W. E. RibelinandG. Migaki (editors). The pathology of fishes, p. 695-718. Univ. Wis. Press, Madison. Valentine, D. W., and K. W. Bridges. 1969. High incidence of deformities in the serranid fish, Paralabrax nehulifer, from southern California. Copeia 1969:637-638. Valentine, D. W., and M. E. Soule. 1973. Effect of p,p'-DDT on developmental stability of pec- toral fin rays in the grunion, Leuresf/iesfenuis. Fish. Bull., U.S. 71:921-926. Valentine, D. W., M. E. soule, and p. Samollow. 1973. Asymmetry analysis in fishes: A possible statistical indicator of environmental stress. Fish. Bull., U.S. 71:357-370. Van Banning, p. 1971. Wratziekte bij platvis. Visserij 24:336-343. VERNBERG, F. J., AND W. B. VERNBERG (editors). 1974. Pollution and physiology of marine organisms. Acad. Press, N.Y., 492 p. VERNBERG, F. J., A. CALABRESE, F. P. THURBERG, AND W. B. VERNBERG (editors). 1977. Physiological responses of marine biota to pollu- tants. Acad. Press, N.Y., 462 p. Walsh, a. H., and W. E. Ribelin. 1975. The pathology of pesticide poisoning. In W. E. Ribelin and G. Migaki ( editors). The pathology of fishes, p. 550-594. Univ. Wis. Press, Madison. Warren, S., and W. a. Meissner. 1971. Neoplasms. 7n W. A. Anderson (editor), Pathology. (6th ed.) Vol. 1, p. 529-561. C. V. Mosby Co., St. Louis. WEDEMEYER, G. 1970. The role of stress in disease resistance of fishes. In S. F. Snieszko (editor), A symposium on diseases of fishes and shellfishes, p. 30-35. Am. Fish. Soc, Spec. Pub. No. 5. WEISSENBERG, R. 1965. Fifty years of research on the lymphocystis virus disease of fishes (1914-1964). Ann. N.Y. Acad. Sci. 126:362-374. WELLINGS, S. R. 1969. Environmental aspects of neoplasia in fishes. In O. W. Neuhausand J. E. Halver (editors). Fish in research, p. 3-22. Acad. Press, N.Y. 1969. Neoplasia and primitive vertebrate phylogeny: Echinoderms, prevertebrates, and fishes — A re- view. Natl. Cancer Inst. Monogr. 31:59-128. WELLINGS, S. R., C. E. ALPERS, B. B. MCCAIN, AND B. S. MIL- LER. 1976. Fin erosion disease of starry flounder (Platichthys stellatus) and English sole (Parophrys vetulus) in the es- 748 SINDERMANN: POLLUTION-ASSOCIATED DISEASES AND ABNORMALITIES tuary of the Duwamish River, Seattle, Washington. J. Fish. Res. Board Can. 33:2577-2586. Wellings, S. R., C. E. Alpers, B. B. McCain, and M. S. Myers. 1977. Fish disease in the Bering Sea. Ann. N.Y. Acad. Sci. 298:290-304. Wellings, S. R., and R. G. Chuinard. 1964. Epidermal papillomas with virus-like particles in flathead sole, Hippoglossoides elassodon. Science (Wash., D.C.) 146:932-934. Wellings, S. R., R. G. Chuinard, and M. Bens. 1965. A comparative study of skin neoplasms in four species of pleuronectid fishes. Ann. N.Y. Acad. Sci. 126:479-501, Wellings, S. R., R. G. Chlunard, r. T. Gourley, and R. a. Cooper. 1964. Epidermal papillomas in the flathead sole, Hippo- glossoides elassodon, with notes on the occurrence of simi- lar neoplasms in other pleuronectids. J. Natl. Cancer Inst. 33:991-1004. Westernhagen. H. von, and V. Dethlefsen 1975. Combined effects of cadmium and salinity on de- velopment and survival of flounder eggs. J. Mar. Biol. Assoc. U.K. 55:945-957, Westernhagen, h. von, V. Dethlefsen, and H. Rosenthal. 1975. Combined effects of cadmium and salinity on de- velopment and survival of garpike eggs. Helgol. wiss. Meeresunters. 27:268-282. Westernhagen, H. von, H. Rosenthal, and K.-R. Sperling. 1974. Combined effects of cadmium and salinity on de- velopment and survival of herring eggs. Helgol. wiss. Meeresunters. 26:416-433. WlLDISH, D. J. 1970. The toxicity of polychlorinated biphenyls (PCB) in sea water to Gammarus oceanicus. Bull. Environ. Con- tam. Toxicol. 5:202-204. WOODCOCK, H. M. 1904. Note on a remarkable parasite of plaice and floun- ders. Rep. Lanes. Sea-Fish. Labs. 1903, No. 12:63-72. YEVicH, P. P., AND M. M. Barry. 1969. Ovarian tumors in the quahog, J, Invertebr. Pathol. 14:266-267. YEVICH, P. P., AND C. A. BARSZCZ. 1976. Gonadal and hematopoietic neoplasms in Mya arenana. Mar. Fish. Rev. 38( 10):42-43. 1977. Neoplasia in soft-shell clams {Mya arenaria) col- lected from oil-impacted sites. Ann. N.Y. Acad. Sci. 298:409-426. Young, J. S. 1973. A marine kill in New Jersey coastal waters. Mar. Pollut. Bull. 4:70. Young, J. S., and j. b. pearce. 1975. Shell disease in crabs and lobsters from New York Bight. Mar. Pollut. Bull. 6:101-105. Young, p. h. 1964. Some effects ofsewer effluent on marine life. Calif Fish Game 50:33-41. ZiSKOwsKi, J., V. Anderson, and R. Murchelano. In press. A bent fin disease of winter flounder (Pseudo- pleuronectes americanus) from Sandy Hook-Raritan Bays, New Jersey, and Lower Bay, New York. J. Fish. Res. Board Can. ZISKOWSKI, J., AND R. MURCHELANO. 1975. Fin erosion in winter flounder. Mar. Pollut. Bull. 6:26-28. 749 I I VERTICAL DISTRIBUTION, DIEL VERTICAL MIGRATION, AND ABUNDANCE OF SOME MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC OCEAN IN SUMMER* Bruce W. Frost and Lawrence E. McCrone^ ABSTRACT Vertical distributions of myctophid fishes and other components of the mesopelagic micronekton were determined during the summers of 1973-75 at two stations in the eastern subarctic Pacific Ocean. Stratified samples were collected with a multiple net Tucker trawl so that the entire water column extendmg to between 385 and 460 m could be sampled during a daytime or nighttime period; two to four day and night vertical series of samples were obtained each summer. Four species of myctophids made up 87^^ of the total fish catch: Stenobrachius leucopsarus and Diaphus theta, which performed diel vertical migrations of 300 m vertical extent: andProtomyctophum thompsoni andS. nannochir, which exhibited only slight diel variation in vertical distribution. Populations of each myctophid species tended to be vertically stratified by age or size with larger individuals occurring in samples taken progressively deeper. Two other major components of the micronekton were euphausiids and decapod shrimps, chiefly Euphausia pacifica and Sergestes stmilis; both species were conspicuous diel vertical migrators. Samples collected in horizontal hauls immediately following sunset showed that three migratory species, the two migratory myctophids and E. pacifica. were closely associated with the single migratory sound-scattering layer (12 kHz); S. similis lagged the ascent of the migratory scattering layer. A single, deep, nonmigratory sound-scattering layer corresponded closely to the distribution of P. thompsoni during both day and night. As in other subpolar oceanic waters, abun- dance and standing stock of myctophids were high — 0.9 fish/m^ and 0.37 g dry weight/m^. In 1973 we began a field study of some small mesopelagic fishes of the family Myctophidae, commonly known as lanternfishes or myctophids, in the eastern subarctic Pacific Ocean. The objec- tives of the study were to determine the vertical distribution and migration characteristics of the numerically dominant species, to document their feeding behavior, and to ascertain if the distribu- tions of fish were in any way influenced by the distribution of their preferred prey. Myctophids are major components of the mesopelagic fauna throughout the world ocean, and in most areas they are sufficiently abundant and stratified in the water column to cause deep sound-scattering layers (Baird et al. 1974; McCartney 1976). In- deed, study of these fishes has been heavily oriented toward aspects of their distribution in relation to sound-scattering layers (e.g.. Tucker 1951; Barham 1966; Taylor 1968; Holton 1969; Farquhar 1971; Baird et al. 1974), although some investigations emphasized aspects of biological 'Contribution No. 1039 from the Department of Oceanog- raphy, University of Washington. Seattle, WA 98195. ^Department of Oceanography, University of Washington, Seattle, WA 98195. Manuscript accepted .Mav 1978. FISHERY BULLETIN: VOL. 76, NO. 4. 1979. and ecological significance, such as individual growth rates, seasonal changes in abundance, and association among species (e.g., Pearcy and Laurs 1966; Harrisson 1967; Lavenberg and Ebeling 1967; Smoker and Pearcy 1970; Badcock 1970; Clarke 1973; Pearcy et al. 1977). Much of the research on myctophids has, in addition, stressed description of the prominent diel vertical migra- tions which are apparently undertaken by almost all species. In the few species studied in detail, both the occurrence and pattern of vertical migration vary with age or ontogeny. Larval myctophids are nonmigratory, spending day and night in near- surface waters (Ahlstrom 1959). Diel vertical migration is first evident at or shortly after metamorphosis and usually persists throughout the remaining life of the fish, although in very old fish, migrations may differ substantially in char- acter from those of younger fish and may even be supressed (Nafpaktitis 1968). Apart from this variation with age, diel vertical migrations of myctophids seem to be relatively regular, on a day-to-day basis, and exhibit little or no seasonal variation (Pearcy and Laurs 1966; Halliday 1970; Pearcy et al. 1977). Among some species, however, 751 therfe may be a portion of the population which does not migrate, while other members of similar size and age do migrate (Clarke 1973; Badcockand Merrett 1976; Pearcy et al. 1977). Virtually nothing is known about the biological causes or consequences of these diel vertical mi- grations, either with respect to the myctophids or their environment. Marshall (1954) suggested that myctophids migrate into the surface layer each night in order to feed on zooplankton, which is usually most abundant in surface waters (Vino- gradov 1968). As pointed out above, larval myc- tophids spend both day and night in the zooplankton-rich surface layer, but as the larvae grow they perhaps become more conspicuous to visual predators and, after metamorphosis, they descend to greater depths, returning to the surface layer only at night, if at all. Vertical migrations may indeed have evolved as a means of avoiding or minimizing predation, but it is unlikely that this hypothesis can be tested in the ocean. On the other hand, it is practicable to investi- gate the feeding ecology of myctophid fish in rela- tion to their migrations; for example, what types of prey the fish utilize, when and where in the water column the fish feed, and whether the vertical distributions of the fish are affected by the vertical distribution and abundance of their preferred prey. As necessary background for such a study, in this paper we present details of the vertical dis- tributions of the numerically dominant species of myctophids in the eastern subarctic Pacific Ocean. METHODS Study Area We conducted the investigation during three summer cruises in areas centered at lat. 50°N, long. 145°W (July-August 1973 and July-August 1975; Station P in Figure 1 ) and at lat. 51°N, long. 137°W (July 1974; Station Q in Figure 1). These stations lie within the hydrographic province des- ignated the Central Subarctic Domain by Dodimead et al. (1963). We chose the subarctic region for ease of sampling and identifying the fish and zooplankton. For example, in an earlier meridional cruise from Kodiak, Alaska, to Hon- olulu, Hawaii (August-September 1972), we found that deep sound-scattering layers are fewer in number, shallower, and more intense in the sub- arctic region than in transition and subtropical waters (Frost unpubl. data). Apparently related to 752 FISHERY BULLETIN; VOL. 76, NO. 4 1 M'^.c^^^ TRANSITIONAL _j I I ; ; I L Figure l. — Sampling stations in the eastern subarctic Pacific Ocean. Representative hydrographic domains for summer condi- tions after Dodimead et al. ( 1963). this, the subarctic myctophid fauna is a simple one; only a few species are abundant, and they are relatively shallowly distributed in the daytime (Taylor 1968). Further, the study area is an open ocean environment, outside the potentially com- plicating influences of coastal and transitional waters (cf. McGowan 1971) and is roughly in the middle of the latitudinal range of several species of myctophids. Finally, the zooplankton assemblage in subarctic waters is also less diverse than in lower latitudes, it is well known taxonomically, and relatively few species are abundant. Sampling Gear Nekton samples were collected with a modified Tucker trawl ( Tucker 1951) described by Frost and McCrone (1974). Briefly, the trawl had a rigid rectangular mouth with a A-vcr area when inclined forward at a 45° angle from vertical, and carried five separate nets ( 6. 35-mm stretch mesh, knotless nylon ace netting) stacked one on top of another (much like fig. 4 in Harding et al. 1971). The net shape followed the design of Clarke (1969). The trawl carried an electronics package containing a strain gage pressure transducer (range 0-1,500 Ib/in^) for determination of depth and a precision pendulum-type tilt transducer (range 0°-90° from vertical) for determination of angle of inclination of the trawl mouth. A TSK (Tsurumi-Seiki Kosakusho)-' flowmeter fitted with a magnetic ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. FROST and McCRONE: MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC reed switch was mounted on the top heam of the trawl. The trawl was towed on a two-conductor coaxial cable and its depth, angle of inclination of the mouth, and revolutions of the flowmeter were monitored continuously during trawling by means of a shipboard display unit. The nets were opened and closed at the mouth by means of a net-tripping assembly which was controlled electronically from the ship. The bottom net ( without cod end ) was left open during deployment to eliminate kiting of the trawl when a net was first opened (Clarke 1969; Badcock and Merrett 1976); thus, four sequential samples could be collected in one haul. The volume of water filtered by each net (assuming 10O7f fil- tration efficiency) was calculated from flowmeter revolutions and average angle of inclination of the net mouth. To determine vertical distributions of fish and other components of the nekton, we towed the trawl obliquely and collected samples on the up- ward leg of a haul. We monitored the speed of the ship during trawling by reference to a Doppler ship's speed indicator. Recordings of deep sound-scattering layers were obtained using an Edo-Western transducer operating at 12 kHz (pulse length 3-10 ms, beam width about 33°) and a Precision Depth Recorder (PDR) operating on the 0-400 fm (0-732 m) depth range scale. At the beginning of each cruise, echo- sounder characteristics (pulse length, power out- put) and recorder gain were set to give optimal resolution of the sound-scattering layer and were not varied thereafter. Sampling Program Taylor (1968) found a close correlation between distribution of abundant species of myctophid fishes and distribution of deep sound-scattering layers in the eastern subarctic Pacific near the Queen Charlotte Islands. Relying on this correla- tion, at each station we designed our sampling program after observing the depth and migrations of deep sound-scattering layers. The number, depth, and migration of scattering layers were virtually identical at Stations P and Q. We ob- served no differences between years at Station P, and our observations do not differ substantially from those of Bary ( 1967) who also used a 12-kHz echosounder in summer at Station P. In the day- time, a single, diffuse, sound-scattering layer ex- tended from about 275 to 375 m depth (Figure 2 A). In the late afternoon and early evening, this scat- tering layer became broader, chiefly by upward movement of the top of the layer, and it persisted with relatively little further change throughout the night. At about 2130 h (local time), a single, upwardly migrating layer became evident, and within half an hour it merged completely with the surface reverberation (Figure 2B). This migratory scattering layer began descent at about 0530 h and merged with the deep nonmigratory layer shortly after 0600 h. Slight variations in times of ascent B 0 -I 100 H 200- 300 -I 400 - 1200 2100 2130 2200 Figure 2. — 12-kHz echograms typical ofthesummer period (July-August) in the sampling areas in the northeastern Pacific Ocean. A. Daytime record about noon, local time. B. Evening record taken on the same day showing the upward movement of the migratory sound-scattering layer and the persistence of the nonmigratory layer at depth. Local time, depth in meters. The dark areas above 100 m are due to surface reverberation. 753 FISHERY BULLETIN: VOL. 76, NO. 4 and descent of the migratory sound-scattering layer depended on weather conditions; also, year- to-year differences are attributed to slight varia- tions in time of cruises. We usually set the lower limit of nekton sampling at least 50 m below the depth of the deep nonmigratory scattering layer. With the exceptions noted below, nighttime sam- pling was confined to the time period between as- cent and descent of the migratory scattering layer. Somewhat different sampling programs were carried out in different years (Table 1 ). At Station P in 1973 the objective was to obtain information on vertical distributions offish and zooplankton to aid in developing an optimal sampling strategy for studying diet and feeding behavior of myctophids. The 0-440 m water column was sampled in 55-m depth strata, and seven successive vertical series of samples, 4 night and 3 day series, were obtained. A shallow haul (0-220 m) and a deep haul (220-440 m) were required for each complete vertical series. The first nighttime series was not completed before the descent of the migratory sound-scattering layer. In order to obtain both the shallow and deep hauls during one night, the hauls were of relatively short duration, and con- sequently the nekton samples were relatively small. At Station Q in 1974, the objectives were to confirm the vertical distributions found in 1973 at Station P and to document the feeding chronology of the common myctophids. The sampling for ver- tical distributions ( Table 1 ) extended from the sur- face to between 400 and 460 m, depending on the depth of the deep sound-scattering layer, and usu- ally included one sample collected below the scat- Table 1 . — Sampling data for vertical series of nekton samples in the northeastern Pacific. The three lower entries for Station P (1975) represent data for; first, the routine day-night vertical series (0-400 m); second, a single shallow (0-60 m) night vertical series; and, third, a single deep (440-782 m) daytime vertical series. Mean (range) Mean (range) Ship duration volume filtered speed of samples per sample Stn. Dates (km h) (min) (m^) P 5-9 Aug. 1973 7.2 = 0.5 29(16-45) 8.412 (4,863-13,357) O 18-22 July 1974 6.2±0.5 45(24-66) 14.229 (6,833-20,361) P 26-28 July 1975 6,9 = 0.5 32(19-43) 10.984 (7.664-14,731) 27 July 1975 6.9=0.5 21 (16-29) 7.638 (6,184-10,052) 31 July 1975 6.9±0.5 68(58-79) 23,232 (20,012-28.614) tering layer. Complete daytime vertical series (both shallow and deep hauls) of samples were collected on 2 days; as no fish were collected in the upper 200 m, two additional daytime vertical series were made with only one haul extending from below the depth of the sound-scattering layer to about 200 m. In order to achieve adequate sam- ple size, the duration of each haul was long, but because there were so few hours of darkness, only one nekton haul could be made each night. Data from a shallow haul and a deep haul on successive nights were therefore combined to give a single, complete, night vertical series; two such night series were obtained, and all sampling was per- formed between ascent and descent of the migra- tory sound-scattering layer. In addition to vertical series of nekton samples taken at Station Q, we utilized two types of hori- zontal hauls. To identify components closely as- sociated with the migratory sound-scattering layer, on two evenings the trawl was launched near sunset and towed horizontally at 125 m. The 12-kHz echosounder was operated continuously during the hauls. Approximately 30 min before the scattering layer began to ascend from its day- time depth, a trawl net was opened and sampling began. The second net was opened just as the scat- tering layer reached 125 m, and the net was closed after the layer had passed that depth. The trawl was towed at 125 m for an additional 30 min, taking a third sample, then closed and retrieved. As part of the study of diel variations in feeding intensity of myctophids, a series of three horizon- tal hauls, each yielding four samples of 30 min duration, was made in the upper layer (40-m depth) throughout one night. The sampling program at Station P in 1975 was similar to that at Station Q, although the nekton sampling for vertical distributions (Table 1) was much less extensive than in the previous two cruises. We obtained only one complete nighttime vertical series (400-0 m) and two deep daytime vertical series (385-220 m) to check on vertical distributions of myctophids. We collected one shal- low vertical series in the 0-60 m layer in 15-m depth strata to examine the vertical distribution of myctophids within the surface layer at night, and we took one very deep daytime vertical series (782-440 m) to determine the distribution of myc- tophids below our usual sampling depths. All samples obtained with the nekton trawl were preserved in a 47f formaldehyde seawater solution buffered with sodium borate. 754 FROST and McCRONE; MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC' PACIFIC Analysis of Samples All organisms in the nekton samples were counted. Fish were identified from descriptions in Hart (1973) and Wisner (19761. The standard length (SL) (distance from the tip of the snout to the end of the vertebral column) of each fish was measured to the nearest millimeter. Among the invertebrates collected in nekton samples, only euphausiids and decapod shrimps were consis- tently captured in substantial numbers. The numbers offish and shrimp were standardized to number per 10,000 m-' of water. Myctophid fish, preserved for about 3 yr, were sorted from samples for determination of body length and dry body weight. Intact, undamaged specimens were dried to constant weight at 65°C on glass slides in a drying oven. Drying usually took 3-4 days, but up to 10 days for some of the largest fish. Dried fish were weighed to the nearest milligram. The relationship between standard body length and dry body weight for each species of myctophid was determined by linear regression analysis of logarithmically transformed mea- surements. RESULTS The vertical series of nekton samples collected at Stations P and Q yielded nine species of myc- tophids, one abundant species of chauliodontid, and one relatively raremelanostomiatid. All other families combined made up only 2-Wi of the total catch by number (Table 2). In addition to fish, the samples contained considerable numbers of euphausiids and decapod shrimps. Other inver- tebrate groups, such as siphonophores and squids, were only sporadically captured. The more com- mon fishes had similar relative abundances at the two stations. More than 80^- of the myctophids consisted of three species (Stenobrachius leucop- sarus, Protoniyctophum thompsuni , and Diaphus thetu ) whose vertical distributions were generally well bracketed by the sampling. The other species of myctophids were either rare or appeared to be distributed below the usual range of sampling; therefore, emphasis in this study was placed on the above three, abundant, relatively shallowly distributed species. Vertical Distribution of Fish The most abundant fish in our samples was S. leucopsarus. Its only congener, S. nannuchir. was rarer (Table 2). As discussed later, with the excep- tion of the very deep vertical series ( 782-440 m ) in 1975, only small specimens ( <35 mm) of S. nan- nochir occurred in the vertical series. The fish caught at Station P in 1973 were in such poor condition that it was not possible to discriminate the smaller specimens of the two species of Steno- brachius. However, there is evidence (presented below) that S. nannochir was extremely rare at Station P in 1973, much rarer than at Station Q or Station P in 1975 (Table 2). Redesign of the cod ends, after the 1973 cruise, provided us with good specimens which permitted discrimination of the two congeners. T.ABLE 2. — Composition of the total fish catch in vertical series of nekton samples in the northeastern Pacific. Data for each station combine all vertical series. Family and species Myctophldae: Stenobrachius leucopsarus S nannochir Protomyctophum thompsoni Diaphus Iheta Tarletonbeania crenulans Lampanyclus regalis L. ritteri Notoscopelus japonicus Symbolophorus californiense Chauliodontidae: Chauliodus macouni Melanostomiatidae: Tactostoma macropus Others Totals Station P, 1973 Station Q, . 1974 Station P. 1975 No, O. No. °o No. % 720 638 1.038 454 461 49.3 (') (') 268 11.7 92 9.8 125 11.1 413 18.1 210 22.5 111 9.8 304 13.3 61 65 16 1.4 26 1.1 3 0.3 9 0.8 11 0.5 0 0 3 0.3 11 05 0 0 1 0.1 0 0 0 0 1 0.1 0 0 0 0 57 1.129 5.0 147 6.4 2.288 84 935 9.0 21 19 15 0.7 3 03 64 5.7 55 2.4 21 22 'Due to the poor condition of the fish caught at Station Pin 1973. it was impossible to discriminate the smaller specimens of the two species of Stenobrachius It is possible that some of the fish listed here as S leucopsarus •Mere in fact S, nannochir. but for reasons described in the text, we do not believe this to be the case. 755 FISHERY BULLETIN; VOL 76, NO, 4 At Station P in 1973, S. Icmopsarus occurred in largest numbers in the surface layer (0-55 m) at night and at 275-330 m during the day (Figure 3A). This pattern could occur if most specimens undertook a diel vertical migration over a depth range of 250-300 m. To be certain that the data reflect a vertical migration and not simply light- aided avoidance of the net by fish in the daytime, it was necessary to compare day and night total catches of fish integrated over the water column sampled (Table 3 ). Assuming that the entire verti- cal range of S. leiicopsarus was sampled (this as- sumption is qualified below), then it is clear that, since the total catches for day and night series were statistically indistinguishable (Table 3), there was no evidence of daytime avoidance of the trawl by fish. Further, judging from the results of replicate sampling of zooplankton with nets (Wiebe and Holland 1968), the day and night to- tals in Table 3 are well within the range of vari- ability expected for repeated samples from a pelagic population. Thus, the observed diel differ- n Nl 01 N2 D2 N3 D3 N4 c B FIGLIRE 3. — Vertical distribution of Stenobrachius leucopsarus at Stations P and Q in the northeastern Pacific Ocean. A. 1973, Station P: four night (Nl and three day (D) vertical series. B. 1974, Station Q: four day and two composite night vertical series. C. 1975, Station P: two composite night and two day vertical series. The profiles for each year are presented in the chronologi- cal order in which they were taken. Scales represent 100 individuals/10'' m^. D 1 D 2 N 1.2 D 3 N 3,4 D 4 Table 3. — Day and night total catches for the water column sampled, of selected fish and crustacean species at Stations P and Q in the northeastern Pacific; means and ranges (parentheses) as number/100 m^. Ratios given of largest to smallest estimate of abundance for a station. Species Station Day Nigtit Ratio Stenobrachius leucopsarus Diaphus Iheta Protomyctophum thompsoni Chauliodus macouni Euphausia pacifica Sergestes similis p, Q, P. 1973 1974 1975 677 101 1 28 0 (28,6-820) (48 4-153 6) (20 3-358) 59 8 62 1 37,6 (47,3-73,7) (56,5-67 7) (29 2-46 1) 2.9 3.2 23 P. Q, P. 1973 1974 1975 6,8 226 36 (1 1-13,8) (15,2-34,4) (3 1-4.0) 12,4 '11.5 54 (3,9-18.2) (11.0-12.0) (3.6-10.4) 16.5 3.1 34 P, Q, P 1973 1974 1975 156 26 7 173 (12 7-20 9) (21 6-39 1) (17 2-17 3) 7.9 23.6 47,0 (0.0-13.8) (20.0-27.3) (28.9-65.2) 24.8 2.0 38 Q, 1974 10,1 (3,2-15,7) 8,2 (7.5-90) 4.9 P, Q. P 1973 1974 1975 1697 4174 3225 (242-323,9) (272 0-531,5) (223 8-421,3) '1.075 3 154 0 423 7 (214 5-5.825 0) (175-2905) (294 5-532 9) 2407 304 25 P, Q P. 1973 1974 1975 19,8 330 10,2 (0-5-31,3) (14,0-47,6) (9,3-11.0) '50.3 '72,9 111 (28 0-80.8) (57.5-883) (9.6-12.6) 161.6 6.3 1.4 'Day and night abundance significantly different (P- ^Estimate based on smallest nonzero catcfi. 0.1) by rank test (Tate and Clelland 1957). 756 FROST and McCRONE: MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC ences in the vertical distribution of S. leucupsarus (Figure 3A) indicate that the majority of individu- als do perform a diel vertical migration. The occurrence of some S. leucopsarus in the deepest samples at night (Figure 3A) indicates that the entire population was not participating in the vertical migration described above. Differ- ences in migratory pattern appear to be largely a function of size or age of individual fish. The length-frequency histogram for our entire catch of S. leucopsarus at Station P in 1973 (Figure 4A) indicates that several size classes of fish were sampled. Since S. leucopsarus metamorphoses to the juvenile stage at 18 mm (Smoker and Pearcy 1970), the abundant 19-35 mm size class (Figure 4A) probably represented the youngest juvenile fish. The largest specimens caught at Station P attain the maximum size expected for S. leucop- sarus, 85-111 mm (Kulikova 1957; Smoker and Pearcy 1970). To determine the effect of size on vertical migra- tion, we examined the three obvious size classes of S. leucopsarus: 19-35 mm, 38-82 mm, and 90-112 mm. The smallest size class ( 19-35 mm) performed a clear diel vertical migration from 275-330 m in the daytime to 0-55 m at night (Figure 5A). The anomalously low density of fish on the third day must be attributed to horizontal patchiness offish. Note especially that only on one night ( N3, Figure 5A) was one small-sized Stenobrachius captured below 275 m. The medium size class (38-82 mm) shows a similar migration (Figure 5B), though these fish seemed to be more dispersed vertically, both at night and in the daytime, than the small- est size class. The high density of medium-sized fish at 275-330 m on the first night, not apparent on the other three nights, probably reflected the fact that this sample was collected between 0554 and 0613 h, a time period when the migratory sonic scatterers, and presumably myctophids, were descending. Some of the medium-sized fish probably had already descended into the 275-330 m layer at the time this sample was collected. The largest size class of S. leucopsarus (90-112 mm) had a pattern of vertical distribution totally dif- ferent from those of the two smaller size classes. Individuals of the largest size class were not cap- tured at all in the first two daytime series and were caught only in the two deepest samples in the third daytime series ( Figure 5C ). They were captured in all four night series, but never in the surface layer (0-55 m), and in three of the four night series, the greatest density of large-sized fish occurred be- tween 330 and 440 m. It is tempting to conclude from these data that the individuals of the largest size class also perform a diel vertical migration, moving from daytime depths below our lower limit of sampling (440 m) into our sampling range at night. Of course, a similar vertical distribution pattern could be obtained if the largest fish avoid the trawl in the daytime, although it seems un- likely that all fish of this size class could effec- tively do so. Nevertheless, with the data from Station P (1973), it is impossible to discriminate between these two possibilities for the largest fish. Stenobrachius leucopsarus had a very similar pattern of distribution and diel vertical migration at Station Q (Figure 3B). The length-frequency distribution of the species was strongly skewed to juvenile fish 1 19-31 mm), which made up 88. 79^ of the total catch of the species. There was only one relatively distinct secondary mode, consisting of very large fish (81-108 mm), which composed 3.9% of the total catch (Figure 4B). Fish in the smallest mode and also the rarer intermediate sizes offish (32-79 mm) were clear vertical migrants, closely following the pattern described above for Station P, and there was no difference in vertical distribu- tion between the small- and medium-sized fish. Also, as at Station P, representatives of the largest size class offish were captured, with the exception of 1 fish ( out of 45 caught ) in the deepest sample on day 4, only in the night hauls and almost always (43 out of 44 fish) below 50 m. At Station P in 1975, the same patterns of diel vertical migration ( Figure 3C ) and size-dependent variation in vertical distribution of S. leucopsarus were evident, though far fewer fish were collected, both because of the fewer vertical series taken and an apparent decrease in abundance of the species compared with the previous 2 yr (Table 3). This decrease appears due partly to reduced abundance of the smallest size class (17-32 mm) which made up only 47.2% of the total catch in 1975 (Figure 4C ), compared with 62.6% at Station P in 1973 and 88.7% at Station Q. In the one deep daytime verti- cal series (782-440 mm) at Station P, large S. leucopsarus were captured between 440 and 740 m (Table 4), thus supporting our earlier hypothesis that the largest fish caught at night above 440 m migrated in the daytime below our usual range of sampling. However, extensive day and night sam- pling over the entire vertical range of the large fish is required to completely rule out daytime avoidance of the trawl. 757 FISHERY BULLETIN; VOL. 76, NO. 4 70 -1 60- _i < > 40 - 30 - LlJ I 20 Z) 10 A Mbjy Uhl MmE 50 60 70 LENGTH (mm) lruD_, 7 80 90 irm rin T 100 no 200 - < 5l50 > z 100 - cc UJ CD 5 3 50 B n r 20 30 In I n.l n hji-i n ip — cam a " "^ — I" ■ — • f ^ " n -r-TTl-w-t I - 40 50 60 70 80 LENGTH (mm) 90 100 n — I no 35 30 - < g 25H > 5 20-1 0 15 H cr UJ 1 10 H 3 Z c M 20 r 30 Myrm ^[]mn n ^mlllNJlTTi I ' I ' I ' ' 1 1'" I" I ""I I I — I — I 40 50 60 70 80 90 100 110 LENGTH (mm) Figure 4. — Length-frequency distributions of Stenobrachius leucopsarus from all vertical series. A. 1973, Station P.N = 720. B. 1974, Station Q, TV = 1,038. C. 1975, Station P,iV = 461. 758 FROST and McCRONE: MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC fl Nl 01 N2 D2 N3 D3 N4 B Nl 01 N2 02 N3 03 N4 c Nl 01 N2.D2 N3 03 N4 Figure 5. — Vertical distribution of three sizes of Steno- brachius leucopsarus at Station P, 1973. A. 19-35 mm, scale represents 100 individuals/lO^m^. B. 38-82 mm, scale represents 50 individuals/lO'' m^. C. 90-112 mm, scale represents 25 individuals/10'* m^. Sequence of vertical series as in Figure 3A. The one shallow night vertical series (60-0 m) at Station P indicated that S. leucopsarus were dis- tributed throughout the surface layer but were concentrated between 15 and 30 m (Figure 3C). Examination of sizes of fish caught in this series suggested very fine-scale vertical stratification by age or size (Table 5). Recall that in all other verti- cal series taken at night very large fish ( >80 mm) were always captured (except for one fish) below 50 m. Because we took only one such shallow ver- tical series, we cannot evaluate the frequency of occurrence or temporal persistence of this appar- ent stratification offish by age in the surface layer at night. The third most abundant myctophid in our sam- ples, Diaphus theta, also performed a diel vertical migration (Figure 6); there was no consistent dif- ference between day and night catches (Table 3). At night, D. thcta ranged over the upper 1 65 m but was concentrated near the surface (0-55 m), while during the day most of these fish were collected below 275 m. As stated above for 5. leucopsarus, the occurrence of D. theta at 275-330 m the first night at Station P ( 1973 ) is misleading because the sample was probably collected after the downward vertical migration of myctophids had begun. The size range for the total catch of D. theta was 36-88 mm in 1973 and 46-84 mm in 1975 at Station P, and 33-76 mm at Station Q. The size-frequency distributions were similar in all 3 yr. Considering only Station Q, for which we have the largest col- lection, the size-frequency distribution (Figure 7A) was quite different from S. leucopsarus (Fig- ure 4B). Small and large fish were rare and the samples contained primarily intermediate sizes (45-58 mm). Distinguishing, somewhat arbi- trarily, three classes in the size-frequency dis- tribution, there is indication of size-dependent Table 5. — Vertical distribution of size classes ofStenobrachius leucopsarus in the shallow night vertical series at Station P (1975) in the northeastern Pacific, as number/10,000 m^. Data based on a single haul with a single sample at each depth. Depth (m) 17-32 mm 37-82 mm •85 mm 0-15 37,2 16 0 15-30 1405 136 0 30-45 7.8 468 0 45-60 15.9 42,8 1.0 Total no captured 138 89 1 Table 4. — Deep daytime vertical distribution of selected species of micronekton at Station P (1975) in the northeastern Pacific, as number/10,000 m^. For Stenobrachius leucopsarus, the numbers in parentheses are abundances of large fish (91- 1 12 mm SL). Data based on a single haul with a single sample at each depth. Stenobrachius Stenobrachius Prolomyctophum Lampanyctus Chauliodus Sergestes Depth (m) leucopsarus nannochir thompsoni ritten macouni similis 440-540 4.5 (3,5) 49,6 0.7 0 2.4 10.8 540-640 8,5 (7.5) 70 0 0 0 7.5 640-740 4,4 (1,3) 4,9 0 0.9 0.4 0 740-782 0 2,7 0.5 0.5 0 0 Total no captured 40 (28) 173 3 3 8 45 759 FISHERY BULLETIN: VOL. 76. NO. 4 n Nl Dl N2 D2 N3 D3 N4 B 0 1 D 2 N 1 .2 D 3 N 3.4 D 4 U 100 : "^ T n 200 X li 1 300 1 (_) 1 4UU t;nn c N 1 N 2 D 1 D 2 FIGURE 6.— Vertical distribution oWiaphus theta. A. 1973, Sta- tion P. B. 1974, Station Q. C. 1975, Station P. Scales represent 25 individuals/10'' m^. Sequence of vertical series as in Figure 3. variation in vertical distribution and vertical mi- gration. The smallest sizes ( 35-44 mm ) offish were consistently shallower than larger sizes both dur- ing the day and at night (Table 6); although the numbers offish are small, they do indicate a possi- ble trend. Diaphus theta was not captured in the very deep (782-440 m) daytime vertical series at Station P in 1975. The second most abundant myctophid, Pro- tomyctophum thompsoni , did not perform an ex- tensive die! vertical migration similar to that of S. leucopsarus or D. theta; it remained below about 200 m both day and night (Figure 8). Neverthe- less, the species tended to be somewhat more shal- lowly distributed at night than in the daytime. This is best demonstrated by the data from Station Q where the largest catches of this species were made. At Station Q,P. thompsoni ranged from 16 30-, 25- 20- 15 - jR A 30 1 ' r 50 60 LENGTH ( mm ) 80 70 -1 50 - 30 - 20 - 3 10 - r B r Jl^nrrdV " flr^^ 1 1 0 20 1 1 ' ' 1 30 40 50 6 LENGTH (mm ) Figure 7. — Length-frequency distributions of Diaphus theta {A),N = 304, andProtomyctophum thompsoni {B),N = 413, from all vertical series at Station Q, 1974. to 53 mm SL and the length-frequency distribu- tion of the population was bimodal (Figure 7B). Calculations of mean depths of the two size classes showed that the smaller fish were always slightly more shallowly distributed that the larger fish (Table 6). Moreover, both size classes tended to be deeper in the daytime than at night, although the average change in depth (30-40 m for both size classes) was relatively small (Table 6). Protomyc- tophum thompsoni was rare below 440 m at Sta- tion P in 1975 (Table 4). The size range of the species at Station P was 18-51 mm (1973) and 16-50 mm ( 1975), and the size-frequency distribu- tion was similar to that of Station Q. The above three species of myctophids had ver- tical distributions which were, with the possible exception of the rare large specimens of S. leucop- sarus, well bracketed by our vertical series of samples. Two other relatively abundant species of fish seemed to have vertical distributions which 760 FROST and McCRONE: MESOPKLACMf FISHES IN THE EASTERN SUBARCTR- PACIFIC Table 6. — Mean depth (meters) of size classes ofDiaphus theta and Protomyctoph urn Ihompsoni in day (D1-D4) and night (Nl. N2) vertical series at Station Q in the northeastern Pacific. Mean depth/) was calculated from the equation D = i;!,Z,/i/!, , where n I is the population density (number/ 10,000 m^) of a size class in sample; andZ, is the midpoint of the depth range of sample i. Size class (mm) D1 D2 D3 D4 Nl N2 Total no captured Diaphus theta 35-44 342 357 325 336 25 25 40 45-58 394 382 377 351 32 30 224 58 390 400 404 344 76 69 40 Protorrtyctophum thompsoni: 16-35 332 316 330 301 294 257 36-53 381 338 340 340 307 309 354 59 fl Nl Dl N2 D2 N3 D3 N4 B 0 2 N 1.2 D 3 N 3.4 0 4 E 200 c 300 400 500 N 2 N 1 D 1 D 2 Figure 8. — Vertical distribution of Protomyctophum thompsoni. A. 1973, Station P. B. 1974, Station Q. C. 1975, Station P. Scales represent 25 individuals/10^ m^. Sequence of vertical series as in Figure 3. extended deeper than our usual range of sampling. Stenobrachius nannochir was only captured below 275 m in the routine vertical series at Stations Q (1974) and P (1975). As noted earlier, due to the poor condition of the catch, small specimens of S. nannochir and S. leiaupsarus were not distin- guished in samples from Station P ( 1973). Half of the total catch ofS. nannochir was from below 400 m, and all of the specimens caught above 440 m were <35 mm SL. It is for this reason that we think that the species must have been extremely rare in the 0-440 m layer at Station P in 1973, for we caught almost no small Stenobrachius in the deep samples at night (Figure 5A). The virtual restriction of catches of S. nannochir to our deepest samples, day and night, indicates that its distribution probably extended below our range of sampling. Indeed, it was the most abundant fish in the one very deep daytime vertical series at Sta- tion P ( 1975); it occurred down to 782 m and was concentrated in the 440-540 m layer (Table 4). Furthermore, an interesting vertical stratifica- tion by size was evident in this series, with the smallest fish dominating the shallowest sample and largest fish dominating the deepest two sam- ples (Table 7). Note that we captured only small specimens ( <35 mm) in all of the other, shallower vertical series. Stenobrachius leucopsarus and S. nannochir of similar body size tended to be verti- cally well separated in the water column at all times (Tables 4, 7; Figure 5). Table 7. — Vertical distribution of size classes of Stenobrachius nannochir in the deep daytime vertical series at Station P ( 1975) in the northeastern Pacific, as number/10,000 m^. Depth (m) 22-37 mm 38-70 mm 85-113 mm 440-540 37.0 12.2 0.3 540-640 0.5 65 0 640-740 0 22 2.7 740-782 0 09 1.8 Total no. captured 107 55 11 The only other moderately abundant fish was the chauliodontid Chauliodus macouni, and only at Station Q was it captured in sufficient numbers to warrant description. Chauliodus macouni al- ways occurred below 150 m, and there was no conclusive evidence of change in its vertical dis- tribution during the day-night cycle (Figure 9, Table 3). However, in contrast to P. thompsoni, whose range of vertical distribution apparently was well sampled day and night ( Figure 8B, Table 4), it appears from the abrupt truncation of histo- grams in Figure 9 that the deepest portion of the population of C. macouni was not sampled either in the daytime or at night. Indeed, in the very deep vertical series at Station P ( 1975), a number of C. macouni were captured in the 440-540 m layer 761 FISHERY Bl'LLETIN: VOL. 76, NO. 4 500 D 1 D 2 N 1,2 D 3 N 3.4 D 4 Figure 9. — Vertical distribution ofChauliodus macouni at Sta- tion Q, 1974. Scale represents 25 individuals/10'* m^. Sequence of vertical series as in Figure 3B. Table 8. — Abundance fnumber'10,000 m^) of Stenohrachius leucopsarus and Diaphus thcta in three series of half-hour sam- ples collected in horizontal tows at 40-m depth during one night at Station Q in the northeastern Pacific. Sampling commenced after the migratory scattering layer had merged with the surface reverberation and terminated after the scattering layer had descended below the surface reverberation (0425). Time is when net was opened. S D S. 0. Time leucopsarus theta Time leucopsarus theta 2200 157 31 0130 262 42 2230 89 15 0200 329 36 2300 102 10 0300 188 72 2330 109 8 0330 178 66 0030 80 21 0400 146 28 0100 197 54 0430 8 0 (Table 4), indicating that the distribution of this fish probably extended below the normal limit of sampling in the routine vertical series. At Station Q, specimens of C. macouni ranged from 29 to 189 mm SL. Very large fish ( >100 mm) were usually captured at night in the deepest samples, but for fish <100 mm there was no clear trend of size- dependent variation in vertical distribution. Other fish species (Table 2) occurred sporadi- cally in the samples and were caught primarily at night: the only daytime catches were below 300 m (e.g., Lampanyctus ritteri in Table 4). Included in the category "Others" in Table 2 were members of the families Bathylagidae, Gonostomatidae, Melamphaeidae, Opisthoproctidae, Paralepidi- dae, and Scopelarchidae. Variability in Abundance of Myctophids in Replicated Samples With a few exceptions, the estimates of abun- dance of myctophids integrated over the water column sampled did not vary by more than a factor of 4 between vertical series within cruises (Table 3). At Station Q, three series of half-hour horizon- tal hauls were made at 40 m throughout one night (Table 8). Excluding the sample (0430) collected after the scattering layer had descended, concen- trations of S. leucopsarus varied by a factor of about 4, those for D. theta by a factor of about 9. For both species, there was a significant trend (P = 0.05, run test, Tate and Clelland 1957) toward increased abundance during the night, and their abundances were strongly correlated (rank differ- ence correlation coefficient 0.74, P ~ 0.01, Tate and Clelland 1957). Myctophids were abundant in the surface layer until the migratory scattering layer descended. Estimated Abundance and Standing Stock of Fishes Our data for mean abundance of all fishes cap- tured for the 3 yr ranged from 0.78 to 1.61/m^ for the water column extending to between 385 and 460 m (Table 9). The three most abundant species of myctophids combined accounted for 77-859r by number of all fish collected. There was no consis- tent difference between day and night estimates of concentrations offish. Equations for the regression of dry body weight on body length (Table 10) were used in conjunction with the lengths and abundance offish from each sample to calculate the population standing stocks of S. leucopsarus, D. theta, and P. thompsoni for Table 9. — Estimated mean abundance and standing stock of mesopelagic fishes at Stations P and Q in the eastern subarctic Pacific Ocean. Myctophids includes only the three most abun- dant species, Stenohrachius leucopsarus, Diaphus theta, and Protomyctnphum thompsoni . Estimated mean abundance and standing stock are based on average of all day and night vertical series; values in parentheses are means for night vertical series only. Abundance (no 'm^) Standi ig stock (g dry wt/m^) Station Myctophids All fishes Myctophids P, 1973 Q. 1974 P, 1975 0,85 1.24 0.61 1-00 1.61 0.78 0.53 (0.77) 0-.27 (0.39) 0.34 (0.55) Table 10. — Equations for the regression of dry body weight, W (grams), on body length, L (centimeters), for three species of myctophids. Regression Range of Species equation SL (cm) N Stenohrachius leucopsarus W = 0,00125/.^"* 2.0-11-8 92 Diaphus theta W = 0.00537 1^*'^ 3.0-7,4 79 Protomyctophum thompsoni W = 0.00212/."" 1.7-4.9 54 762 FROST and McCRONE: MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC each vertical series of nekton samples at each sta- tion. A slight (90 mm), specimens of S. leucopsarus which were caught chiefly at night. Vertical Distribution of Crustaceans The most abundant organisms in the vertical series of nekton samples were euphausiids, pre- dominantly large individuals (>12 mm total length). At Station P in both 1973 and 1975, Euphausia pacifica made up more than 80% of the euphausiid catch by number. At Station Q, 51% of the total euphausiid catch was E. pacifica; other species were Thysanoessa spinifera (30%), Tes- sarabrachion occulatum (9%), Thysanoessa lon- gipes (8% ), and Stylocheiron maximum (2% ). All of these species also occurred at Station P, but were rare. Consequently, only the data for E. pacifica are presented here. During the day, large E. pacifica occurred in greatest concentration between 275 and 400 m, while at night they were usually concentrated in the upper 55 or 60 m (Figure 10). No consistent difference between day and night total catches was evident, but sporadic, extraordinarily large or small catches of £. pacifica were obtained in both 1973 and 1974. Variations such as these are com- mon in euphausiid catches (Brinton 1962b) and are usually attributed to horizontal patchiness. Our ranges of estimated abundances were con- sequently very large (Table 3). The other four species of euphausiids were too rare to draw definite conclusions about their distributions. The penaeid decapod shrimp, Sergestes similis, was the only other abundant invertebrate in our nekton samples. At Station P (1973) and Station 0 100- 5 200- 6 300 f 400 500 0 100 5 200 n fc 300 CD 400 500 ■ 1 L i 1 n Nl Dl N2 02 N3 03 N4 + 1 B 0 1 0 2 N 1.2 0 3 N 3.4 0 4 500 c N 1 N 2 0 1 D 2 Figure lO.— Vertical distribution of Euphausia pacifica. A. 1973, Station P. Scale represents 1,000 individuals/lO" m^. (The 0-55 m sample on the fourth night represents 10,447 individuals/lO-* m^*.) B. 1974, Station Q. Scale represents 500 individuals/ia» m^. C. 1975, Station P. Scale represents 500 individuals/10^ m^. (The 15-30 m sample on the first night repre- sents 3,086 individuals/10^ m^*.) Sequence of vertical series as in Figure 3. 763 FISHERY BULLETIN; VOL. 76, NO. 4 Q, the species appeared to be performing an exten- sive diel vertical migration (Figure 11 A, B); how- ever, the average daytime catches at both stations were a bit less than half the average nighttime catches, though only in 1973 and 1974 were there statistically significant differences (Table 3). Ex- cept for the largest size class of Stenobrachius leucopsarus (Figure 5C), Scrgestes similis is the only species for which we found such a prominent, repeated, day-night difference in catches. Either S. similis is a diel vertical migrator and descends below our usual range of sampling in the daytime or it is capable of avoiding the nekton trawl in the daytime. Our very deep daytime vertical series taken at Station P (1975) bears on this question. Although the species seemed considerably less abundant in 1975 (Table 3), this was probably n Nl Dl N2 02 N3 D3 N4 B 01 02 N1.2 03 N3.4 04 100 1 E 200 1 n S: 300 a II 1 1 400 son c N 1 N 2 0 1 0 2 Figure ll.— Vertical distribution ofSergestes similis. A. 1973, Station P. B. 1974, Station Q. C. 1975, Station P. Scales repre- sent 50 individuals/10' m''. Sequence of vertical series as in Figure 3. partly due to the shallower depth to which the routine vertical series extended in the daytime. In the very deep daytime vertical series, S. similis occurred in considerable numbers between 440 and 640 m (Table 4 ). Thus it probably was a migra- tor and in the daytime ranged well below the greatest depth of sampling on routine vertical series. At both stations, S. similis tended to be rather broadly distributed over the 0- 1 50 m layer at night and often was more abundant below 50 m than above ( Figure 11). In this respect its diel migration differs from that of the two migratory myctophid fishes and E. pocifica, which tended to aggregate strongly above about 60 m at night. In addition to S. similis several other types of malacostracans were collected in the samples: the caridean decapods Hymenodora frontalis, Noto- stomiis japoniciis, and Pasiphaea sp.; the penaeid decapod Bentheogennema borealis; and the my- sids Gnathophausia gigas, Boreomysis sp., and Eucopia sp. All were rare, were collected only at night, and almost always occurred below 200 m. Micronekton Associated With Sound-Scattering Layers In the daytime, the position of the scattering layer corresponded closely with the daytime depth of occurrence of the smaller size classes of Steno- brachius leucopsarus and the populations of D. theta and Protomyctophum thompsoni (Figure 12A). For example, in the profiles shown in Figure 12A, the 300-400 m stratum contained an average concentration of 136 fish/10,000 m^ of the three species combined. Sergestes similis is distributed too broadly and deeply in the daytime to contrib- ute to the observed scattering layer (Figure IIB, Day 3). Excluding euphausiids, in our samples no other potential sound-scattering organism (e.g., physonect siphonophores) consistently had its center of abundance between 275 and 400 m in the daytime. The large E. pacifica collected with the nekton trawl had a pattern of vertical distribution (Figure lOB, Day 3) very similar to that of the migratory myctophid fishes. Comparison of the vertical distribution and diel migration of Stenobrachius leucopsarus with the echosounder trace indicates a correlation between the fish and the migratory sound-scattering layer (Figures 2, 3). The correlation is best for individu- als of the small and medium size classes (Figure 5). Similarly, the vertical distribution and diel mi- 764 FROST and McCRONE: MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC 300 + #T ,■,.;, v*; On 100- 200- B T 200tish/IO''m3 S leucopsarus D theta f p. thompsont Figure 12. — Vertical distribution of three species of myctophids relative to a sound-scattering layer recorded with the 12-kHz echosounder. A. Midday distribution of fish and an echogram showing the position of the scattering layer at the time of sam- pling (Day 3 at Station Q). B. Nighttime distribution offish and an echogram showing the position of the scattering layer at the time of sampling (Nights 3, 4 at Station Q). gration of D. theta closely parallel the behavior of the migratory sound-scattering layer (Figures 2, 6). To examine this relationship more closely, at Station Q two series of three horizontal samples each were collected at 125 m in the periods preced- ing, during, and after ascent of the migratory sound-scattering layer past that depth. In the first series (Table 11, 17 July) bothS. leucopsarus and D. theta were most abundant in the sample col- lected as the scattering layer was passing 125 m. Euphausiids (predominantly E. pacifica) and Sergestes similis were also abundant in the sam- ples; however, maximum concentrations of each were obtained either in the sample collected before or after the scattering layer had passed 125 m (Table 11). Results from the second series (Table 11, 18 July) were similar except that euphausiids were not as abundant in the first sample of the series, and Stenobrachius leucopsarus was most abundant in the sample collected after the scatter- ing layer had passed 125 m. The results, therefore, indicate that both migratory myctophids and euphausiids are associated with the migratory sound-scattering layer, whereas sergestid shrimps are not. Table ll. — Occurrence of migratory myctophids and crusta- ceans (number/10,000 m'') in two series of three samples col- lected in horizontal hauls at 125 m depth before, during, and after ascent of the migratory sound-scattering layer past that depth. 17 July 18 July Species Before During After Before During After Stenobrachius leucopsarus 0 38 1 0 13 27 Diaphus theta 1 37 5 0 25 0 Euphausia pacifica 87 64 1 5 68 2 Sergestes similis 3 13 53 0 4 82 Position of the nonmigratory portion of the deep sound-scattering layer which was present at night was strongly correlated with the distribution of P. thompsoni, particularly the small size class. The scattering layer was broader and more diffuse at night, and so was the distribution of P. thompsoni (Figures 2, 8, 12B). Over the 200-300 m stratum, the average concentration of P. thompsoni was 16.3 fish/10,000 m^ for the profile shown in Figure 12B. The day-to-night persistence of the nonmi- gratory scattering layer (Figure 2) cannot be explained by reference to the distribution of either S. leucopsarus or D. theta. The two smaller size classes of S. leucopsarus and all D. theta have migrated into the surface layers at night, and the largest S. leucopsarus are not only rare but broadly distributed over 50-450 m. There are no other abundant potential sound-scattering or- ganisms concentrated in the 200-300 m stratum at night. DISCUSSION Previous work on myctophids in open waters of the subarctic Pacific dealt chiefly with systematics and biogeography (Wisner 1976). However, Aron (1962) and Taylor (1968) considered aspects of the distribution of myctophids in eastern subarctic waters. Aron's (1962) results are qualitative due to the nature of the sampling gear used (unme- tered, nonclosing nets of variable mesh size). Dif- ferences between results of our study and those of Taylor's (1968) comprehensive investigation are probably attributable to the different sampling gear employed rather than to fundamental varia- tions in behavior of fish in different parts of the subarctic Pacific. For example, Taylor's use of very course-meshed nets probably accounts for both his finding of different relative abundances of myc- tophid species and for somewhat different patterns 765 FISHERY BULLETIN: VOL 76, NO. 4 of vertical distribution of species. Thus in Taylor's study, carried out not far from Station Q, P. thonipsojti and D. thcUi were more abundant than S. leiHopsarus, but this was probably because Taylor's net either did not efficiently catch juvenile ( <35 mm) S. lei/copsarus which were the numerically dominant size class of that species in our samples (Figure 4), or they were much less abundant during the time he sampled. Further, Taylor obtained some of the largest catches of D. theta and S. Icucopsariis below 90 m at night. This probably also reflects the sampling bias of his net for larger sizes offish, which at night tend to be more broadly spread over the water column than smaller fish (Figure 5C: Table 6, night series). Unfortunately, Taylor did not report the sizes of fish captured. Except for probable sampling bias toward larger sizes offish, Taylor's results on ver- tical distribution of the nonmigratory P. thompsoni and other species of fish agree with ours. Pearcy et al. ( 1977) described patterns of verti- cal distribution of mesopelagic fishes and crusta- ceans off the coast of Oregon. The mesopelagic assemblage there is essentially subarctic in faunistic affinity and the vertical distributions of species are similar to those observed at Stations P and Q. The only notable departure from our re- sults was the finding by Pearcy et al. that sig- nificant numbers of all sizes of S. leucupsarus did not participate, at least on a regular basis, in the diel vertical migrations. Our observations at both Stations P and Q indicate that virtually all S. leucopsarus smaller than about 80 mm performed extensive diel vertical migrations (Figure 5). However, in our studies, S. leucopsarus was also very rare below 400 m (Figure 5, Table 4), whereas Pearcy et al. found large concentrations below that depth. Thus there may be major differences in the vertical distribution and migration behavior of S. leucopsorus in different parts of its geograph- ical range (Paxton 1967). Significantly, Pearcy et al. ( 1977) detected no seasonal variations in verti- cal distributions and migrations for any species, which may also be true for subarctic waters to the north (Taylor 1968). Perhaps the most remarkable feature of the mesopelagic fauna of the area sampled was its simplicity. Only four species of myctophid fishes were abundant in the upper 700 m. Two of these species, S. leucopsarus and D. theta, undertook diel migrations of substantial vertical extent; the other two, P. thompsoni andS. nannochir, did not. 766 Other taxonomic groups also showed low diver- sity. Among the micronektonic crustaceans there were single species of abundant euphausiid, E. pacifica, and decapod shrimp, Sergestes similis, and both were vertical migrators. The contrast between this relatively simple mesopelagic mi- cronekton fauna and that, for example, in the sub- tropical North Pacific (Brinton 1962a; Clarke 1973; Walters 1977) or subtropical North Atlantic (Badcock 1970; Foxton 1970a, b) is striking, but not atypical. Low taxonomic diversity of the mesopelagic micronekton is found in other subpo- lar oceans, such as the Boreal Atlantic (e.g.. Back- us et al. 1971; Zahuranec and Pugh 1971). Associated with the taxonomic simplicity of the mesopelagic fauna herein reported, was a rela- tively simple structure of the sound-scattering layers. Generally, both the number and depth of sound-scattering layers change with latitude in the deep ocean; fewer and shallower layers are found in subpolar oceans than in tropical- subtropical oceans (Haigh 1971; Cole et al. 1971; Donaldson and Pearcy 1972; Tont 1976). Our un- published observations on deep sound-scattering layers ( 12-kHz echosounder), taken in September 1972 along long. 155°W between Alaska and Hawaii, showed this trend. Subarctic waters had the relatively simple sound-scattering structure illustrated in Figure 2, with single migratory and nonmigratory layers occurring shallower than 400 m. In the subtropical waters near Hawaii, at least three sound-scattering layers were observed in the daytime at depths ranging from 260 to 625 m, and three to four migratory layers were re- corded. It is unlikely that the correlation between taxonomic diversity of the mesopelagic micronek- ton and complexity of the sound-scattering struc- ture in the water column was fortuitous. Attempts to causally relate deep sound-scattering layers to aggregations of mesopelagic organisms were stimulated by hypotheses advanced more than three decades ago (for a review see Hersey and Backus 1962). However, field studies based on net samples taken simultaneously with echosounder records tend to be inconclusive for a variety of reasons. A major difficulty is that different taxonomic groups tend to occur together at the same depths and may even show similar migra- tory behavior. For example, all four of the migra- tory mesopelagic species in our study (Stenobrach- ius leucopsarus , D. theta , E. pacifica , and Sergestes similis) ascended towards the surface layer after FROST and McCRONE: MKSOPEI.AGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC sunset, and only from fine temporal spacing of samples did it become apparent that some species were more closely associated with the scattering layer than others (Table 11). Similarly, in the day- time some of these migratory species cooccurred at the depth of the sound-scattering layer together with the nonmigratory P. thonipsoni, and any or all could have contributed to the daytime sound- scattering layer. Despite extensive cooccurrence of several types of potential sound-scattering or- ganisms, the most reasonable hypothesis is that myctophids were primarily responsible for both the migratory and nonmigratory sound-scattering layers in the eastern subarctic Pacific. Taylor (1968), also working in the subarctic Pacific, found the best correlation between deep sound-scattering layers and those mesopelagic fish which possessed gas-filled swim bladders. Al- though Taylor grouped Stenuhrach ius leucopsarus and D. theta among fish with fat-invested swim bladders, gas is present in the swim bladders of immature individuals ( <30 mm SL) of both species (Capen 19671. Taylor made no mention of the size of the fish caught in his study; however, in view of the very coarse-meshed nets he used, it is probable that he did not quantitatively sample immature fish. At Stations P and Q, some indi- viduals of S. leucopsarus and D. theta were theoretically the right size to resonate at 12 kHz while at their daytime depths (Capen 1967), and the abundance of either species was probably sufficient to produce deep sound-scattering in the daytime (Hershey and Backus 1962). This pre- sumably holds also for P. thompsoni, which has a gas-filled swim bladder throughout life (Taylor 1968: Butler and Pearcy 1972). Indeed, concentra- tions of either D. theta or P. thompsoni alone in Figure 12 were comparable with the concentration of D. taaningi, which Baird et al. (1974) believe was responsible for the migratory sound- scattering layer over the Cariaco Trench. As pointed out above, Sergestes si m His may be excluded as a potential sound scatterer; it was distributed too broadly and deeply in the daytime and lagged the ascent of the migratory sound- scattering layer at sunset (Figure 11, Table 11). Although E. pacifica (Figure 10) was about five times more abundant in the depth of the daytime sound-scattering layer than all myctophid fishes combined, it did not approach concentrations necessary for it to be an effective scatterer of 12-kHz sound (Hersey and Backus 1962; Bary 1966; Beamish 1971). In conclusion, we suggest that the nonmigratory deep sound-scattering layer (Figure 2B) in the vi- cinities of Stations P and Q in the eastern subarc- tic North Pacific was caused by P. thompsoni, and that the migratory sound-scattering layer (Figure 2B) recorded the migrations of smaller size classes of Stenohrachius leucopsarus and D. theta. Pro- tomyctophum thompsoni may have been largely responsible for the deep scattering layer observed in the daytime, with possible lesser contributions from the two migratory myctophid species. Pearcy (1977) found similar general correspondence be- tween vertical distributions of the same three species of myctophids and deep sound-scattering layers off Oregon, but he pointed out that quan- titative correlation between abundance of poten- tial sound-scatterers and distribution of volume scattering was not always strong. A more defini- tive analysis, similar to that of Baird etal. (1974), is required; that is, simultaneous observations should be obtained on distribution of volume scat- tering and abundance and acoustical properties of suspected sound-scatterers. In single hauls, we observed concentrations of myctophids, all species combined, which regularly exceeded 100 fish/10'* m'^ in the region of the deep sound-scattering layer in the daytime and in the surface layer at night. Similar concentrations of myctophids are found in other oceans (e.g., Kash- kin 1967). Further, the maximum concentrations of myctophids observed by us in the surface layer at night (365 fish/ 10-* m^ Table 8) and at depth in the daytime (874 fish/ 10"* m'K horizontal haul at 327-333 m. Station Q) equal or exceed maximum concentrations inferred from the apparently high catch rates of single hauls reported by Halliday (1970) and Backus et al. (1971) for the western Boreal Atlantic, where one species of myctophid, Benthosema glaciale , predominates. The very low concentrations of myctophids found by Pearcy et al. (1977), using a 2.4 m Isaacs-Kidd midwater trawl, are puzzling and seem to indicate that myc- tophids are about Vio as abundant off the Oregon coast as in the open subarctic Pacific. However, the data of Pearcy et al. ( 1977) differ from the earlier results of Pearcy and Laurs (1966), in which re- ported concentrations of myctophids were much higher and similar to concentrations observed by us; the difference could be due to year-to-year var- iability (Pearcy 1977). There is relatively little variability between years in our estimates of abundance of myctophid fishes (the three most abundant species. Table 2) 767 FISHERY BULLETIN: VOL. 76, NO. 4 in the water column extending to 385-460 m. We estimate 0.61-1.24 myctophid fish/m^ based on av- eraged day and night series (Table 10). No quan- titative study comparable to ours has been made in the open subarctic Pacific, but Pearcy and Laurs (1966) provided data on the abundance of mesopelagic fish near the Oregon coast. In two cruises (August 1963), Pearcy and Laurs found about 0.78 myctophid fish/m^ in the 0-500 m water column at night; this estimate is based upon the three numerically dominant myctophid fish cap- tured (Pearcy and Laurs 1966, fig. 4), two of which ranked 1 and 3 in abundance among myctophids in our study. The average standing stock of all mesopelagic fish found by Pearcy and Laurs ( 1966) was 2.9 g wet weight/m^ in the 0-500 m water column at night. Using a factor of 0.3 to convert wet weight to dry weight, the average nighttime standing stock is 0.87 g/m^, a value probably not significantly different from our estimates based on night samples (Table 9), especially since the Pearcy and Laurs estimate is based on all mesopelagic fish captured. Similar concentrations of myctophids (about 0.6-0.8 fish/m^) are found in the subtropical Pacific near Hawaii (Clarke 1973; Maynard et al. 1975). However, many more species of myctophids (47) occur there, and the standing stock of myctophids (about 0.3-0.7 g wet weigh t/m^) is somewhat less than our estimates (0.23-0.53 g dry weight/m^, Table 9), probably be- cause the fish are considerably smaller in average size (Clarke 1973). With regard to sampling bias, we found no evi- dence of light-aided avoidance of the nekton trawl by either myctophids or other types of micronek- ton occurring in the upper 385-460 m during the daytime (Table 3). Consistent day-night differ- ences in catches of organisms, such as those ob- served for the largest size class ( >80 mm SL) of S. leucopsarus and for Sergestes similis, were proba- bly due to migration of these organisms below the depth range of daytime sampling. The results of the single very deep vertical series at Station P (Table 4) support this interpretation. Further- more, very deep vertical migrations of both species are well documented in other parts of their geo- graphical ranges in the North Pacific (Omori et al. 1972; Pearcy et al. 1977). In addition to determining vertical distributions and vertical migrations of myctophid fishes, on each cruise we also sampled zooplankton with a smaller trawl (Frost and McCrone 1974). Analyses of the zooplankton samples, together with data on stomach contents of the three most abundant myc- tophids, are the subject of a report (in preparation) on the feeding behavior of myctophids in relation to their vertical distribution and the vertical dis- tribution of their zooplankton prey. ACKNOWLEDGMENTS We extend special thanks to David Thoreson for his assistance in all phases of the research. Bruce Davies participated in the cruises and was primar- ily responsible for nearly flawless operation of the trawl. We were fortunate to obtain help and advice on systematics of myctophid fishes from Richard McGinnis. Karl Banse made many useful sugges- tions on the manuscript. Others whom we wish to thank for participating in the cruises or assisting with the analysis include Gene Anderson, Arthur Griffiths, Louise Hirsch, Jeffrey Napp, Bruce Nes- set, Mary Nirini, Layne Nordgren, Scott Ralston, Wesley Rowland, Gary Shigenaka, and Steve Spyker. This research was supported by the Office of Naval Research (Contract N00014-75-C-0502, Project NR 083-012). Early development of the trawl was supported by National Science Founda- tion Grant GA-25385. 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. ARON, W. 1962. The distribution of animals in the eastern North Pacific and its relationship to physical and chemical con- ditions. J. Fish. Res. Board Can. 19:271-314. Backus, R. H., J. E. Craddock, R. L. Haedrich, and D. L. Shores. 1971. The distribution of mesopelagic fishes in the equato- rial and western North Atlantic Ocean. In G. B. Far- quhar (editor), Proceedings of an international sym- posium on biological sound scattering in the ocean, p. 20-40. Maury Center for Ocean Science, Wash., D.C. BADCOCK, J. 1970. The vertical distribution of mesopelagic fishes col- lected on the SOND Cruise. J. Mar. Biol. Assoc. U.K. 50:1001-1044. BADCOCK, J., AND N. R. MERRETT. 1976. Midwater fishes in the eastern North Atlantic - I. Vertical distribution and associated biology in 30°N, 23°W, with developmental notes on certain myctophids. Prog. Oceanogr. 7:3-58. Baird, R. C, D. F. Wilson, R. C. Beckett, and T. L. Hopkins. 1974. Diaphus taaningi Norman, the principal component of a shallow sound-scattering layer in the Cariaco Trench, Venezuela. J. Mar. Res. 32:301-312. 768 FROST and McCRONE: MESOPELAGIC FISHES IN THE EASTERN SUBARCTIC PACIFIC BARHAM, E. G. 1966. Deep scattering layer migration and composition: observations from a diving saucer. Science (Wash., D.C.) 151:1399-1403. BARY, B. MCK. 1966. Back scattering at 12 kc/s in relation to biomass and numbers of zooplanktonic organisms in Saanich Inlet, British Columbia. Deep-Sea Res. 13:655-666 1967. Diel vertical migrations of underwater scattering, mostly in Saanich Inlet, British Columbia. Deep-Sea Res. 14:35-50. Beamish, P. 1971. Quantitative measurements of acoustic scattering from zooplanktonic organisms. Deej>Sea Res. 18:811- 822. BEAUCHAMP, J. J., AND J. S. OLSON. 1973. Corrections for bias in regression estimates after logarithmic transformation. Ecology 54:1403-1407. BRINTON. E. 1962a. The distribution of Pacific euphausiids. Bull. Scripps Inst. Oceanogr., Univ. Calif 8:51-269. 1962b. Variable factors affecting the apparent range and estimated concentration of euphausiids in the North Pac- ific. Pac. Sci. 16:374-408. BL'TLER, 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 Elec- tron. Lab., Res. Rep. 1447, 32 p. CLARKE, M. R. 1969. A new midwater trawl for sampling discrete depth horizons. J. Mar. Biol. Assoc. U.K. 49:945-960. CLARKE, T. A. 1973. Some aspects of the ecology of lantemfishes (Myc- tophidaei in the Pacific Ocean near Hawaii. Fish. Bull., U.S. 71:401-434. Cole, H. p., G. M. Bryan, and A. L. Gordon. 1971. The deep scattering layer: patterns across the Gulf Stream and the Sargasso Sea. In G. B. Farquhar (editor). Proceedings of an international symposium on biological sound scattering in the ocean, p. 281-293. Maury Center for Ocean Science, Wash., D.C. Dodimead, a. J., F. Favorite, and T. Hirano. 1963. Salmon of the North Pacific Ocean. Part II. Review of oceanography of the subarctic Pacific region. Int. North Pac. Fish. Comm. Bull. 13, 195 p. Donaldson, h. a., and w. G. pearcy. 1972. Sound-scattering layers in the northeastern Pa- cific. J. Fish. Res. Board Can. 29:1419-1423. Farquhar, G. B. (editor). 1971. Proceedings of an international symposium on biological sound scattering in the ocean. Maury Center for Ocean Science, Wash., D.C, 629 p. FOXTON, P. 1970a. The vertical distribution of pelagic decapods [Crus- tacea: Natantia] collected on the SOND Cruise 1965. I. The Caridea. J. Mar. Biol. Assoc. U.K. 50:939-960. 1970b. The vertical distribution of pelagic decapods [Crus- tacea: Natantia] collected on the SOND Cruise 1965. II. The Penaeidea and general discussion. J. Mar. Biol. Assoc. U.K. 50:961-1000. Frost, B.W., and l. e. mcCrone. 1974. Vertical distribution of zooplankton and myctophid fish at Canadian weather station P, with description of a new multiple net trawl. Proceedings of the Interna- tional Conference on Engineering in the Ocean Environ- ment, Vol. I, p. 156-165. Halifax, N.S. haigh, K. K. R. 1971. Geographic, seasonal, and annual patterns of mid- water scatterers between latitudes 10° and 68° North in the Atlantic. In G. B. Farquhar (editor). Proceedings of an international symposium on biological sound scatter- ing in the ocean, p. 268-280. Maury Center for Ocean Science, Wash., D.C. Halliday, R. G. 1970. Growth and vertical distribution of the glacier lan- temfish, Benthosema glaciale, in the northwestern Atlan- tic. J. Fish. Res. Board Can. 27:105-116. Harding, D., E. Shreeve, D. S. Tungate, and D. Mummery. 1971. A net-changing mechanism for the Lowestoft mul- tipurpose sampler. J. Cons. 33:483-491. Harrisson, C. M. H. 1967. On methods for sampling mesopelagic fishes. Symp. Zool. Soc. Lond. 19:71-126. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can., Bull. 180, 740 p. Hersey, J. B., AND R. H. Backus. 1962. Sound scattering by marine organisms. In M. Hill (editor). The sea: Ideas and observations on progress in the study of the seas. Vol. 1, p. 498-539. Interscience, N.Y. Holton, a. a. 1969. Feeding behavior of a vertically migrating lan- temfish. Pac. Sci. 23:325-331. KASHKIN, N. I. 1967. On the quantitative distribution of lantern fishes (Myctophidae s.l.) in the Atlantic Ocean. In T. S. Rass (editor). The pelagic and bathypelagic fishes of the world oceans, p. 206-264. Tr. Inst. Okeanol. Akad. Nauk SSSR 84. (Translated from Russ. by the Translation and In- terpretation Division, Institute of Modern Languages, Inc., 1969.) KULIKOVA, E. B. 1957. Growth and age of deep- water fishes. /nB. N.Niki- tin (editor), Marine biology, p. 284-290. Tr. Inst. Okeanol. Akad. Nauk SSSR 20. (Translated from Russ. by Am. Inst. Biol. Sci., 1959.) LaVENBERG, R. J., AND A. W. EBELING. 1967. Distribution of midwater fishes among deep-water basins of the southern California shelf. In R. N. Phil- brick (editor), Proceedings of the symposium on the biol- ogy of the California Islands, p. 185-201. Santa Barbara Botanic Garden, Inc., Santa Barbara, Calif. Marshall, N. B. 1954. Aspects of deep sea biology. Hutchinson's Scien- tific and Technical Publications. Lond., 380 p. MaYNARD, S. D., F. V. RIGGS, AND J. F. WALTERS. 1975. Mesopelagic micronekton in Hawaiian waters: Faunal composition, standing stock, and diel vertical migration. Fish. Bull., U.S. 73:726-736. McCartney, b. s. 1976. Comparison of the acoustic and biological sampling of the sonic scattering layers: R.R.S. "DISCOVERY" 769 FISHERY BULLETIN: VOL. 76, NO. 4 SOND Cruise, 1965. J. Mar. Biol. Assoc. U.K. 56:161- 178. MCGOWAN, J. A. 1971. Oceanic biogeography of the Pacific, /n B. M. Fun- nell and W. R. Riedel (editors), The micropaleontology of oceans, p. 3-74. Camb. Univ. Press, Camb. NAFPAKTITIS, B. G. 1968. Taxonomy and distribution of the lantemfishes, genera Lobianchia an&Diaphus, in the North Atlantic. Dana Rep. Carlsberg Found. 73, 131 p. Omori, M., A. Kawamura, and Y. AIZAWA. 1972. Sergestes similis Hansen, its distribution and impor- tance as food of fin and sei whales in the North Pacific Ocean. In A. Y. Takenouti (editor). Biological oceanog- raphy of the northern North Pacific Ocean, p. 373-391. Idemitsu Shoten, Tokyo. Paxton, J. R. 1967. A distributional analysis for the lantemfishes (Fam- ily Myctophidae) of the San Pedro Basin, California Copeia 1967:422-440. PEARCY, W. G. 1977. Variations in abundance of sound scattering ani- mals off Oregon. In N. R. Anderson and B. J. Zahuranec (editors), Oceanic sound scattering prediction, p. 647-666. Plenum Press, N.Y. PEARCY, W. G., E. E. KRYGIER, R. MESECAR, AND F. RAMSEY. 1977. Vertical distribution and migration of oceanic mi- cronekton off Oregon. Deep-Sea Res. 24:223-245. PEARCY, W. G., AND R. M. LAURS. 1966. Vertical migration and distribution of mesopelagic fishes off Oregon. Deep-Sea Res. 13:153-165. Smoker, W., and W. G. Pearcy. 1970. Growth and reproduction of the lantemfish Steno- brachius leucopsarus. J. Fish. Res. Board Can. 27:1265-1275. Tate, M. W., and R. C. Clelland. 1957. Nonparametric and shortcut statistics in the social, biological, and medical sciences. Interstate Printers and Publishers, Inc., Danville, 111., 171 p. Taylor, F. H. C. 1968. The relationship of midwater trawl catches to sound scattering layers off the coast of northern British Colum- bia. J. Fish. Res. Board Can. 25:457-472. Tont, S. a. 1976. Deep scattering layers: patterns in the Pacif- ic. Calif. Coop. Oceanic Fish. Invest., Rep. 18:112-117. TUCKER, G. H. 1 95 1 . Relation of fishes and other organisms to the scatter- ing of underwater sound. J. Mar. Res. 10:215-238. Vinogradov, m. E. 1968. Vertical distribution of the oceanic zooplank- ton. Inst. Okeanol. Akad. Nauk SSSR, Moscow, 339 p. (Translated from Russ. by A. Mercado and J. Salkind, Isr. Program Sci. Transl., Jerusalem, 1970.) Walters, J. F. 1977. Ecology of Hawaiian sergestid shrimps (Penaeidea: Sergestidae). Fish. Bull., U.S. 74:799-836. WiEBE, P. H., and W. R. Holland. 1968. Plankton patchiness: effects on repeated net tows. Limnol. Oceanogr. 13:315-321. Wisner, R. L. 1976. The taxonomy and distribution of lantemfishes (Family Myctophidae) of the eastern Pacific Ocean. Navy Ocean Res. Dev. Act., Rep.-3, U.S. Gov. Print. Off., Wash., D.C., 229 p. Zahuranec, B. J., and W. L. Pugh. 1971. Biological results from scattering layer investiga- tions in the Norwegian Sea. In G. B. Farquhar (editor). Proceedings of an international symposium on biological sound scattering in the ocean, p. 360-380. Maury Center for Ocean Science, Wash., D.C. 770 ANALYSIS OF A SIMPLE MODEL FOR ESTIMATING HISTORICAL POPULATION SIZES T. D. Snqth' and T. Polacheck=^ ABSTRACT Estimates of historical abundance of animal populations are important in many management deci- sions. Historical estimates based on a simple model of population growth have been made for several populations of dolphin involved with the yellowfin tuna purse seine fishery. We used the data for the bridled dolphin, Stenella attenuata. to investigate the behavior of the model by which these historical estimates were calculated. For populations with low net reproductive rates, the effect of bias in the estimates of the input parameters on the estimated historical abundances was approximately linear and additive. When all the input parameters were independently estimated, the variances of the historical abundance estimates were dominated by the variance of the initial abundance estimate and the coefficient of variation of the historical estimate was less than the largest coefficient of variation of any parameter. Many decisions about the management of animal populations are based on the estimates of abun- dance of the population relative to its historical or preexploitation size. These estimates are basic to any application of the theory of maximum sus- tained yield as incorporated in several interna- tional marine mammal management agreements such as the North Pacific Fur Seal Treaty and the International Whaling Convention. Similarly, the concept of "optimum sustainable populations" as specified in the recent Marine Mammal Protection Act of 1972 (MMPA) has been defined in terms of comparing the present size of a population with its original size (Southwest Fisheries Center^). Schools of dolphin of several species (primarily Stenella attenuata and S. longirostris) have been used by purse seine fishermen in the eastern tropi- cal Pacific to locate yellowfin tuna, Thunnus alba- cares, since 1959, as described by Perrin (1969). Significant numbers of dolphin have been killed by becoming entangled in the purse seines. In order to make management decisions under the MMPA about these dolphin populations, the Na- tional Marine Fisheries Service (NMFS) needed 'Department of Zoology, University of Hawaii, Honolulu, Hawaii; present address: Southwest Fisheries Center, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. ^Department of Zoology, University of Hawaii, Honolulu, Hawaii; present address: Department of Biology, University of Oregon, Eugene, OR 97477. ^Southwest Fisheries Center. 1976. Report of the Workshop on Stock Assessment of Porpoises Involved in the Eastern Pacific Yellowfin Tuna Fishery. Southwest Fish. Cent. La Jolla Lab., Natl. Mar. Fish. Serv., NOAA, Admin. Rep. LJ-76-29, 53 p. estimates of the preexploitation abundance of the various populations. The NMFS convened a work- shop of scientists to obtain the estimates based on a simple model of population change (see footnote 3). This paper evaluates the behavior of estimates of abundance obtained from their approach. This is important in order to be able to evaluate the degree of confidence to be placed in such estimates, and hence in management plans based on them. METHODS AND MATERIALS The model used to estimate preexploitation abundance is based on a common discrete model of population growth: AT^^j = N-K^HN-K^) (b-d) (1) where N^. = the abundance at time r b = the birth rate d = the natural death rate K^ = the number of animals killed, as- sumed to occur at the beginning of time interval t N^^ 1 = the abundance 1 time unit later. Reversing the procedure (i.e., solving the above equation for N,. ) results in the expression ^ l+i?_ ^ (2) Manuscript accepted May 1978. FISHERY BULLETIN; VOL. 76. NO. 4, 1979. where Nr now is the estimate of abundance 1 yr 771 FISHERY BULLETIN: VOL. 76. NO. 4 earlier andi?. is the net reproductive rate ib-d). The above model was modified in the procedure used by NMFS to account for situations when the kills occur throughout the time interval instead of instantaneously at the end of the interval, as: r+i "• T (3) This equation can be repeatedly applied to give estimates any number of years it) into the past. When rearranged to explicitly display the popula- tion size t years earlier, and relabeling so that the initial abundance is N„, one obtains 1974 and the annual incidental kills and repro- ductive rates from 1959 to 1974. Several se- quences of estimated annual kills and reproduc- tive rates were considered, incorporating the uncertainty in the data. In the present paper the sequences of annual kills and net reproductive rates given in Table 1 are used to illustrate several general aspects of the behavior of Equation (4). These correspond to the "high kill" and "central reproductive rate" se- quences for the bridled dolphin, Stenella at- tenuata, in the Workshop report. The estimate of 1974 abundance used by us and the Workshop was 3.5 million. A^„ A^. = t n (1+^p t X.(l+i?./2) (4) Note in this form that the time-index t runs back- wards from zero. As is apparent in this form, the estimation of abundance t years earlier involves 2t+\ parameters. The sequences of annual kills and net reproductive rates can be termed the kill and the net reproductive rate vectors, each com- posed of t elements. The data used here to explore this estimation procedure is from the report of NMFS Workshop discussed above (see footnote 3).^ From existing unpublished data and reports the Workshop par- ticipants used estimates of the population size in ''It should be noted that the estimates used here are based on a number of assumptions currently under investigation and that these estimates are subject to significant change in the near future (I. Barret, Director, Southwest Fisheries Center, La Jolla, CA 92038, pers. commun. April 1978). Estimation of Bias A sensitivity analysis was done to examine the effects of biased parameter estimates on the backcalculated abundance. A new population size 1 yr earlier, from Equation ( 3 ), when each parame- ter is changed by a specified amount is N^{n, k, r) = N^{l+n)+0.bK^{l+k) and in general for t years earlier. (5) iV;(n, k, r) = N^il+n) U^ (l+/2.(l+r)) , KiUk) (l-H(/2 (l+r)/2)) -H_2— ^-^ (6) S.(l+/?.(l+r)) Table l. — Estimates used for kill and reproductive rate vectors of Stenella attenuata in the eastern Pacific. Kill f Year (thousands) Net reproductive rate 1 1973 120 0,040 2 1972 273 .040 3 1971 185 .040 4 1970 308 .036 5 1969 331 .032 6 1968 164 .028 7 1967 194 .024 8 1966 281 .020 9 1965 297 .016 10 1964 255 .012 11 1963 133 .008 12 1962 106 .004 13 1961 446 .000 14 1960 534 .000 15 1959 129 .000 where Nq, R^, and K^ are defined as above, and n = the proportion that A^q deviates from its estimate r = the proportion that all elements of the net reproductive vector deviate from their es- timates k = the proportion that all elements of the kill vector deviate from their estimates. N'j in,k,r) was then compared with A^, from Equa- tion (4) or equivalently N ', (0,0,0). As a measure of the sensitivity of the basic model, S, {n,k,r) is defined to equal the percent that A^', (n,k,r) de- viates from A^^ 772 SMITH AND POLACHECK: ANALYSIS OF SIMPLE MODEL S^ (n, k, r) = N^ (n, K r)-N^ N, • 100 . (7) Estimation of Variance The variance of the backcalculated estimate of A^, from Equation (4) was approximated using the delta method (Sober 1973). This method is based upon a Taylor series expansion for a function in which quadratic and other higher order terms are ignored. If/" is a function of the random variables x^, X2,X2 . . . ,Xn then the expression for the vari- ance of/" by the delta method is v(r(x^,x,,X3...,xj)= iv(xp(^]' + 2S2 Gov (X.,X.) I-J^ ■ -^1 . (8) i^j ^i' v\ax. bx.i 1 J/ In applying this expression to Equation (4), it is necessary to be able to define which of the parameters should be considered as random vari- ables, and to give reasonable estimates for value of the variances and covariances of these variables. For the purpose of exploring the behavior of Equa- tion (4), we assumed that the estimates of all the parameters in Equation (4) are independent ran- dom variables. The covariance terms in Equation (8) are then zero. This approach provides a picture of the variance of the back estimate of abundance if in fact independent estimates of the kills and the net reproductive rates were available for each year. A generalized expression for the variance using this approach is V(iV^) V(iV,) dN. ^2 3A^. t + S V{K.) I dN. dN. + 2 V(i?.) y=i ^ r 'dN. (9) where all parameters are defined as for the basic model [Equation (4)]. For detailed expressions for each of the right hand terms see Appendix I. As noted the method used for approximating the variance of a function depends on the higher order terms in the Taylor's series expansion being small. The higher order terms in the delta method ex- pression for the variance of A/^, are composed of the second and higher order derivatives of N, with respect to A^n- K^, and Rf, and the higher order central moments of the probability distributions of the estimates of N„, K,, and R, (i.e., skewness, kurtosis, etc.). The second and higher derivatives with respect to A^^ andKf are zero. Thus the terms involving/?, are the only higher order terms not equal to zero. The higher order derivatives of Nf with respect to i?, involve i?,+ j to increasing nega- tive powers. The three higher order moments ofR, are always decreasing since /?, is much less than one. Thus each of the higher order terms in the delta method expression for the variance of N, are each less than the first order term in R; (iii of Appendix I). The contribution of this first order term in Rt to the variance ofN, is small, as shown below. Thus the error induced by ignoring the higher order terms in the Taylor's series appears small. The objective in doing the variance calculations was to understand the behavior of the variance of the population size when estimated by the basic back projection model [Equation (4)]. Thus a range of variances was calculated for a range of reason- able values of the variances of the estimated parameters. However, in our example of bridled dolphin estimates of the variance of many of the parameters were not available. Many of the kill estimates were not independently estimated and hence have large unknown covariances (Smith and Polacheck^). Estimates of net reproductive rate were obtained by extrapolation from other populations and fi'om assumptions about density dependence. It is not clear that the uncertainty in these estimates can adequately be described by the notion of variance. Thus, the variances that we used and that we calculated for N, should not be interpreted as actual estimates of variance for this population. RESULTS Bias The results of the sensitivity analysis of the basic model will be presented by examining the effects of varying each of the variables «, k, and r of Equation (7), separately, and then in combina- tions. The sensitivity of the back projected estimates ^Smith, T. D., and T. Polacheck. 1977. Uncertainty in estimat- ing historical abundance of porpoise populations. Contract Rep. MM 7A C006, 39 p. Marine Mammal Commission, 1625 Eye Street, Washington, DC 20006. 773 FISHERY BULLETIN: VOL. 76, NO. 4 (S,) for a fixed number of years / into the past is linear with respect to A? or/? (Figure D.ThisHnear- ity can be seen in Equation (6) since n and k enter only as linear terms in the numerator. Positive 30i Figure l.— Sensitivity of the model Sffn,;fe,H in 1959 for a range of deviations in the initial number (n ), for a range of deviations in the kills (k), or for a range of deviations in the net reproductive rate (r), for Stenella attenuata in the eastern tropical Pacific. values of either n ov k yield positive deviations in the back estimates. However, the farther back the population is projected in time, the smaller the contribution of N,, to the back estimate becomes relative to the contribution of the kills. Thus the effect of bias in the estimate of the initial numbers (n) becomes progressively smaller the farther back in time the population is projected, while the consequence of a consistent bias in the kill esti- mates (k) becomes larger (Figure 2). Since the annual kills have no simple relationship to time, the effect of a particular value of « or ^ over time (Figure 2) cannot be described by any simple func- tion. This trade off in the sensitivity of the back projected estimates between n and k is exact in the sense that for any decrease over time in the slope of S with respect to n there is an equivalent in- crease in the slope of S with respect to k. This can be seen by evaluating the partial derivates of S with respect to n and with respect to k and noting that they sum to 1. The effects of bias in the estimates of the net reproductive rate vector are more complicated than for the other two factors. Positive deviations in the net reproductive rates (r) yield negative deviations in the back projected estimate (Figure 2). The effect of r tends to increase over time (Fig- ure 2). S approaches being linear with respect to r for any particular year, but unlike the relation- ship for k and n, this result is not exact ( Figure 1 ). The approximate linearity of the sensitivity ofN^ £. 10- .^ 5 -5- ,o- .- 0-' .-O' ,o' .-cr ,0' * A- ^ — *- *■ —A— 0 1 I 2 1 3 1 4 1 5 r 6 1 T 7 8 Time (t) I 9 10 1 11 — I — 12 r ■ 13 14 15 1973 1971 1969 1967 Year 1965 1963 1961 1959 FIGURE 2.— Sensitivity of the model St (n,k,r} over time to a 30% deviation in the initial number (n = 0.3), in the kill vector (k = 0.3), and in the net reproduc- tive rate vector (r = 0.3) when all factors are held constant (or Stenella attenuata in the eastern tropical Pacific. 774 SMITH AND POLACHECK; ANALYSIS OF SIMPLE MODEL to r appears to be a general feature of this proce- dure when r is small. This can be seen by examin- ing S, expressed as a function of r, which can be obtained explicitly by substituting the definitions of N, [Equation (4)] and iV',[Equation (6)] into Equation (7) and simplifying. The consequences of having two factors varying simultaneously are shown in the series of contours of equal values of S from Equation (7) (Figures 3-5). These contour plots present a visual picture of the sensitivity of the back projection to the dif- ferent factors. From this set of contour maps, it can be seen that the surface generated by S [Equa- tion (7)] tends to be nearly linear. Since S has no nonlinear terms with respect to n and k, the sur- face described by S in these two dimensions is simply a plane (Figure 4). There are nonlinear effects between the net reproductive rate and both initial abundance and the sequence of kills. For the example examined here, the nonlinearity be- tween k and r is insignificant. For instance, if r and k both equal 0.50, S deviates from a linear model by 60%), the CV of the back estimate does not exceed the CV of A^q. 1 + and the basic model [ Equation (3)] for the dolphin population examined here is given in Table 6. The simpler model always gives a slightly higher es- timate for the size of the back projected population but the increase in the estimate is always <1%. The sensitivities of the two models are nearly equivalent. When the values for the parameters in these models deviate as much as SO'/t the differ- ence between sensitivities of the two models is <1%. The approximate variances of the back es- timates of the two models are also similar. That the difference between the original and the simpler model is small can be shown by analyti- cally comparing the two models. If the projections are made only 1 yr into the past, the ratio of the estimate from Equation (2) to the estimate from Equation (3) is 1 + O.bR^K^ N^+K^+O.bR^K^- Only if the value ofR^K^ is large relative to Ao +^i can this ratio deviate significantly from 1. This is only possible if /?i is relatively large. The general formula for the ratio of the two models is S 0.5KJi.( fl (l+R, J) N^+ 1 0.5K. (n (l+R^ J)+ 2 O.bK.in (l+RJ) Comparison of Equations (2) and (3). A comparison of the estimated back abundance as calculated by the simpler model [Equation (2)] As in the case for projecting back only 1 yr, it can be seen that unless the RjKj terms are large rela- tive to Nq and unless the net reproductive rate is also large, the ratio of the two models will be close to 1. Table 6. — Comparison of the back estimate of the abundance of bridled dolphin as calculated by the basic model [Equation (3)] and the simpler model [Equation (2)]. Simple model Basic model Year ('10«) (xlO^) Simple/basic 1974 3.500 3.500 1.000 1973 3.485 3.483 1.001 1972 3.624 3.617 1.002 1971 3.670 3.659 1.003 1970 3.850 3.835 1.004 1969 4.062 4.0416 1.005 1968 4.115 4.093 1.005 1967 4.214 4 190 1.006 1966 4.412 4.386 1.006 1965 4.640 4.612 1.006 1964 4.840 4.811 1.006 1963 4.934 4.905 . 1.006 1962 5.021 4.991 1.006 1961 5.467 5.437 1.005 1960 6.001 5.971 1.005 1959 6.130 6,100 1.005 DISCUSSION AND CONCLUSIONS The results of this analysis indicate that errors in the input parameters do not compound in this procedure for estimating historical abundance. In fact, a systematic bias in the procedure for the estimation of a single set of parameters (either A^o or R^'s or K,'s) always induces a bias in the back projected estimate which is less than the bias of the estimated parameters. This conclusion follows directly from the linear or near linear relation between St and n, k, or /• with small rates of change. Moreover, the effects of bias in two or more sets of parameters are nearly additive. The interaction effects of bias in estimates of kills, net reproductive rates, and the initial number tend to 777 FISHERY BULLETIN: VOL 76. NO. 4 be small or nonexistent. This will be globally true for the relationship between k and n, but will be true for the relationship between /?, r, and /; only when the net reproductive rate is small. The rela- tive importance of bias in A'/s, /?/s, or 7V„ on A^, depends upon the actual values of the parameter. In the bridled dolphin example, after 15 yr, the back estimates were most sensitive to bias in the kill estimate, slightly less sensitive to bias in N„, and considerably less sensitive to bias in the net reproductive rate. However, the importance of bias in A/^o will diminish with the number of years in the back estimate with a proportionate increase in the importance of bias in the kills. The sensitivity analysis developed in this paper will include the extremes of a complete sensitivity analysis of the model. The values forS/ (0,/?,0) are limiting values to a complete sensitivity analysis of the individual elements of the kill vector on N,. Similarly S, ( 0,0,/') is a limit to complete sensitiv- ity analysis of the individual elements of the net reproductive rate. Given the additivity of S, with respectto«,r, and^, the surface S^ («,/?,r) contains the extremes of a sensitivity analysis in all 2t+\ dimensions. If in fact the elements within the kill vector and within the reproductive vector are highly interdependent (as is the case for the data used here), then the sensitvity analysis used to look at the effects of bias in this paper approaches a total sensitivity analysis of the back projected estimate given these constraints. The variance approximations also indicate that variability in the parameter estimates does not result in compounding uncertainty in the back projected estimates. When estimates of the parameters are independent and the net reproduc- tive rate is low, the CV of the back estimate will be smaller than the CV of the input parameters. In our example if all the CV's were equal, the vari- ance of A^o would make the largest contribution to the estimated variance oiN,. In general this will be true as long as the kills in any one year do not approach the initial abundance. This is a direct consequence of the basic additivity of the model when the net reproductive rate is small. In Smith and Polacheck (see footnote 4), an al- ternative probability structure was considered in which the elements within the kill vector and within the net reproductive rate vector were highly interdependent. In this situation, the vari- ance of Nt is not completely dominated by the variance of N^^. The variances of N, calculated using this interdependent probability structure are larger than the variances presented here in which all the parameters are assumed indepen- dent. However, the CV of TV, for the dolphin data within this interdependent probability structure is still less than the CV of the parameters if all parameters have equal C V. It appears that even in the situation in which a high degree of inter- dependence exists within the kill estimate or the net reproductive estimates, the variability in the parameter estimates does not induce compound- ing uncertainty in the back projected estimate. The comparison of the results from the basic model [Equation (3)] with the simpler model [Equation (2)] indicate that there are no sig- nificant differences between the two models as long as the net reproductive rate is small. Thus it appears that there is no reason to favor the more complex model over the simpler. In conclusion, it appears that this back projec- tion procedure (either model) has reasonable statistical properties, at least when the net repro- ductive rates are small. However, Equation ( 1 ) is a simplified description of how the abundance of a population changes through time, especially in not accounting for changes in age structure. The authors feel that caution should be used in apply- ing estimates from this procedure to the manage- ment of long-lived species since changes in the age structure for long-lived species are likely to be important. ACKNOWLEDGMENTS Financial support for this study was supplied by the U.S. Marine Mammal Commission (Contract MM74C006). We wish to acknowledge the journal editor and an anonymous reviewer for their help- ful comments. LITERATURE CITED Perrin, W. F. 1969. Using porpoise to catch tuna. World Fish. 18(6):42-45. Seber, G. a. F. 1973. The estimation of animal abundance and related parameters. Hafner Press, N.Y., 506 p. 778 SMITH AND POLACHECK: ANALYSIS OF SIMPLE MODEL Appendix I. — Expressions for the variance components of N,. Expression for the right hand terms of Equation (9) are: ^ »/ \.n (i+R,), 2 V(X)(^ = 2 V(if) -^ ^ (ii) ;?, ^<«>' br = ,?, v<«/' — -i-> — ) • ""> \ '/ \(i+R.)^^n.^^(i.R,)/ Appendix II. — Coefficient of variation of a sum of random variables. The following is a proof that the coefficient of variation of a sum of two independent random variables is smaller than the greatest CV for either of the random variables if the expected value of the random variables is greater than zero. If A and 5 are independent random variables such that E(A) = a>Q E(B) = &>Oand then CV(A) = ^^^>«=CV(B) V(A) ^ V(B) V(A)(62 + 2ab) > YiB)a^ , V{A)ib^ + 2ab) + YiA)a^ > YiB)a^ + Y{A)a^ , V(A)(a + bf > [V(B) + V(A)]c2, V(A) Y{B) + V(A) ^ V(A + B) a^ (a + 6)2 [E(A + B)f ' CV(A) > CV(A + B) . 779 LARVAL DEVELOPMENT OF GALATHEA ROSTRATA UNDER LABORATORY CONDITIONS, WITH A DISCUSSION OF LARVAL DEVELOPMENT IN THE GALATHEIDAE (CRUSTACEA ANOMURA)^ Robert H. Gore^ ABSTRACT The complete larval development of the western Atlantic anomuran crab, Galathea rostrata, consists of four or five zoeal stages, and a single megalopal stage, based on larvae cultured under laboratory conditions. Variation in the duration and number of zoeal stages appears to be temperature-dependent, with larvae reared at 15°C developing through five zoeal stages and attaining megalopa in 52 days, whereas larvae cultured at 20°C passed through four or five zoeal stages, reaching megalopa in 18 or 23 days, respectively. At 20°C some third stage zoeae molted to a "regular" fourth zoeal stage, without pleopods, which was followed by a subsequent fifth stage before reaching megalopa. Other zoeae molted to an "advanced" fourth stage, possessing pleopods, which subsequently molted directly to megalopa, bypassing stage V completely. The variation noted in larval development in other galatheid genera is briefly discussed, and a provisional s3Tiopsis of morphological characters of systematic value is pro- vided for their identification. The anomuran crab genus Galathea is presently represented in the western North Atlantic by two species, G. agassizii andG. rostrata (A. Milne Ed- wards 1880). Galathea agassizii, primarily tropi- cal and insular in distribution, is a deepwater species known from 166 to 490 fm (304-897 m) off St. Augustine, Fla., and from Cuba, St. Vincent, St. Lucia, and Barbados in the Caribbean Sea. In the eastern Atlantic the species is found from 82 to 898 fm (150-1,643 m) in the vicinity of both the Cape Verde and Canary Islands, and off northwestern Africa (Chace 1942; Miyake and Baba 1970). Contrarily, G. rostrata appears to be a warm- temperate or tropical/subtropical species, primar- ily continental in distribution. The species is re- corded from the North American continental shelf at Cape Hatteras, N.C., to southeastern Florida, and in the Gulf of Mexico from western Florida, the Mississippi Delta, and southward to Islas Jol- bos, north of the Yucatan Peninsula. There is a questionable record from off Rhode Island (Wil- liams 1965). Galathea rostrata is also found in shallower water than G. agassizii and has been collected from 10 to 50 fm (18-92 m), with the exception of the possible depth record of 1,178 fm 'Scientific Contribution No. 100, from the Smithsonian Institution-Harbor Branch Foundation, Inc., Scientific Consor- tium, Link Port, Ft. Pierce, Fla. This report is Article IX. Studies on Decapod Crustacea from the Indian River Region of Florida. ^Smithsonian Institution, Ft. Pierce Bureau, Ft. Pierce, FL 33450. Manuscript accepted June 1978. FISHERY BULLETIN: VOL. 76, NO. 4, 1979. (2,156 m) from off Rhode Island. The only distribu- tional record of the species for the entire eastern Florida coast was that of Haig (1956) who reported a single specimen collected from 21 fm (38 m) off Hillsboro Lighthouse (Broward County) in south- eastern Florida. However, recent collections show that the species is not uncommon in the Indian River region of the central eastern Florida coast, especially on deeper water (60+ m) coquinoid limestone ledges and reefs of the ivory tree coral, Oculina varicosa Leseuer. The few studies made on the larval development of new world galatheid crabs (e.g., Rayner 1935; Boyd 1960; Fagetti 1960; Boyd and Johnson 1963; Fagetti and Campodonico 1971) have all been made on eastern Pacific species, and the larvae of Atlantic American galatheids, including the genus Galathea, remain undescribed. This paper provides the first description and illustration of the complete larval development of G. rostrata, as well as the first report on any species of Galathea reared totally under labora- tory conditions, from hatching to megalopal stage. The larvae and postlarvae are compared with lar- val stages known from other members of the Galatheidea throughout the world, and shared features are briefly summarized. MATERIALS AND METHODS Eight ovigerous females of G. rostrata were ob- tained on 15 April 1977 by lockout diver from the 781 FISHERY BULLETIN: VOL. 76, NO. 4 Research Submersible Johnson-Sea-Link II, of the Harbor Branch Foundation, Inc. The adult galatheids inhabited a large clump of ivory tree coral which grew in 80 m of water on Jeffs Reef, lat. 27°32.8'N, long. 79°58.8'W, located about 17 n. mi. (27 km) northeast of Ft. Pierce Inlet, Fla. The entire coral colony was collected and returned to the surface inside of a 500-/u,m mesh cloth bag. Ambient seawater temperature on Jeffs Reef was 12°C at time of collection. The galatheids were immediately placed in compartmented plastic trays containing recently collected neritic seawa- ter previously chilled to 10°C. Upon return to the laboratory each adult specimen was transferred to individual 100 x 80 mm covered glass laboratory dishes filled with approximately 340 ml of seawa- ter previously chilled to 15°C. Each isolated female was maintained at this temperature, pro- vided a change of chilled seawater, and fed freshly hatched Artemia salina nauplii, daily. All speci- mens were exposed to a 12-h light-12-h dark il- lumination program in a controlled temperature unit (CTU) until hatching occurred. Five females survived in this regimen and yielded larvae over a period from 16 April to 6 May 1977. Seven larval series were initiated. Using methodology previously described by Gore (1968), five such series were cultured in 24- compartmented plastic trays. These consisted of two series of 8 and 24 larvae, held in the CTU at 15°C ( ±0.5°C), and three series of 24 larvae each, maintained at cool laboratory room temperature (ca. 20°C, ±1°C). Two mass culture series of about 30 larvae each were also established in individual 100 X 80 mm glass dishes at cool laboratory room temperature, which was controlled by reverse- cycle air conditioning, and was monitored daily with a 7-day recording thermometer. Fresh surfzone seawater (35.5-36%o) was collected week- ly, filtered through glass wool, stored in 14-gal (ca. 56-1) polypropylene carboys, and used throughout the rearing period. All larval series were checked daily, and any molts or dead individuals were recorded and pre- served in IWc ethanol. Specimens were examined microscopically, slides prepared, and drawings made as described in previous studies by Gore (1968). Measurements given below are the arith- metic average of all specimens examined in any particular stage. A complete series of larvae, or their molts, is deposited in the National Museum of Natural History, Washington, D.C. (USNM 170862); the Allan Hancock Foundation, Univer- 782 sity of Southern California, Los Angeles (AHF 1028-01); the British Museum (Natural History), London (BMNH 1978:103); and the Rijksmuseum van Natuurlijke Historie, Leiden (D 31735). RESULTS AND DISCUSSION OF THE REARING EXPERIMENT Galathea rostrata passes through four or five morphologically distinct zoeal stages and a single megalopal stage, before completing development in the laboratory. Culture temperature undeni- ably affects duration of development, and perhaps larval survival as well. While the duration of the zoeal and megalopal stages differed at each rear- ing temperature, it was nevertheless generally consistent within each of the temperature series, as will be discussed below. At 15°C five morphologically distinct zoeal stages were observed for those larvae surviving to metamorphosis. The minimum time required to pass through these stages and attain megalopal stage was 52 days. Most larvae remained in each zoeal stage approximately 9-11 days through the first four stages. Only two stage V zoeae survived, and they remained as such 14 and 16 days before molting to megalopa. However, neither of these specimens survived longer than 6 or 7 days as megalopae, so the mean duration of the postlarval stage at 15°C remains unknown (Table 1). With the minimum noted period of 6-7 day duration for megalopae at this temperature, completion of de- velopment and metamorphosis to first crab stage Table l. — Duration of larval and postlarval development in Galathea rostrata under laboratory conditions at the indicated temperatures. Temp and Days required to attain next stage stage Mm Mean Mode Max 15°C: 1 10 10.8 10 14 II 8 9.7 10 M6 III 9 10.5 9 17 IV 8 9.4 9 11 V 14 — — 16 Mg 6-7 (Both megalopae died in stage) 20 C: 1 5 5.8 6 7 II 4 4.2 4 5 III 4 4.1 4 5 IV (regular) 5 5.7 6 6 V (regular) 5 5.8 6 6 IV (advanced) 3 33 3 4 Mg (combined)^ 12 12.6 13 13 III- IV (inter- mediate) 7 — — 8 (Died in stage) 'One zoea remained 30 days in stage II, dying 13 days later in stage III. ^Combined megalopae data include stages obtained from both IV (ad- vanced) and IV (regular). GORE: LARVAL DEVELOPMENT OF GMATHEA ROSTRATA is conservatively estimated to take well over 60 days (Figure 1). At 20°C either four or five morphologically dis- tinct zoeal stages occurred. The minimum time required to complete larval development and reach megalopa was 18 days. Most larvae re- mained in each zoeal stage from 3 to 6 days and as megalopae from 12 to 13 days. The total duration of development from hatching to first crab stage spanned a minimum of 30 days at 20°C, if only four zoeal stages were required, but took at least 37 days with five zoeal stages (Figure 1, inset). The larvae generally fared well at both culture temperatures. Although the larvae at 15°C took longer to complete their development, they ini- tially appeared to survive better than their coun- terparts at 20°C (Figure 1). At 15°C, at least 50^7^ larval survival occurred through stage IV, before a rapid decrease occurred in stage V and megalopa. Larvae reared at 20°C exhibited a steep decline after stage 1, to about 35'??^ survival, and showed a continual decline thereafter. The precipitous de- cline in larval survival at this temperature from stage I to stage II was the result of an almost complete mortality in one culture tray, for un- known reasons. At 15°C ecdysis in the earlier zoeal stages (I-III) generally was a less critical period than at 20°C, although the larvae at the latter temperature were still able to complete most molts. The larvae at 20°C attained subsequent stages more rapidly than did those at 15°C, and some were able to complete zoeal development, although overall lar- val mortality was relatively higher. On the other hand, at 15°C larval survival may have been en- hanced by lower temperature, but the major difficulty then seemed to be the attainment of stage V and megalopa. Only two megalopae were obtained in the 15°C program and neither was able to molt to the succeeding first crab stage. In con- trast, four megalopae survived at 20°C, and molted to crab stage I; three of these specimens were maintained in the laboratory to crab stages XII and XIV. Ecdysial and Sequential Variation in Galathea rostrata Two modes of developmental variation were noted in G. rostrata at 20°C. In one mode, some zoeae III molted to an instar which, for purposes of discussion, is labelled "regular" stage IV. This stage was characterized, among other features, by a reduced number of antennular aesthetascs and was always without well-developed pleopod buds on the abdominal somites. Zoeae remained in this stage for 3-4 days before molting to stage V, an instar possessing distinct, well-developed, pleopod buds and an increased number of antennular aes- thetascs. The duration of stage V lasted 5-6 days and was followed by the molt to megalopa. One of these postlarvae subsequently molted to first crab stage. In the second mode of variation, some zoeae III molted to an "advanced" stage IV, with some, but not all, of the features as noted above for stage V. Zoeae remained longer in the advanced stage (5-9 days) before molting directly to megalopa. Three of these megalopae went on to attain first crab stage. The two types of development are compared in the inset of Figure 1. Two other stage III zoeae, which remained in stage III 7-8 days (instead of the usual 4-5), molted to what appeared to be an intermediate stage IV. These zoeae exhibited some stage V zoeal features in size, maxillipedal setae numbers, and in posses- sing pleopod buds, although the latter were only rudimentaiy. A reduced number of antennular aesthetascs similar to that of regular stage IV zoeae was also seen. The two specimens survived only 4-5 days in this stage before dying. This mode of variation was not considered as important as the previous two modes and will not be discussed further. Remarks The regular and advanced fourth stages cannot be equated to an early and late fourth stage, nor to substages IVa and IVb, because no molt occurred from one fourth stage or substage to another. If the molt from stage III was to regular stage IV, this was invariably followed by an ecdysis to stage V, and then a subsequent molt to megalopa. If the molt from stage III produced an advanced stage IV, this in turn molted directly to megalopa, skip- ping stage V altogether. At 15°C the regular stage IV and stage V appear to be necessary plateaus in larval development, whereas at 20°C development may proceed in some zoeae without resorting to either of these instars. The regular fourth zoeal stage (as defined above), therefore, appears to be a true sequential stage of development, inasmuch as it was seen in larvae at both 15°C and 20°C programs. However, it is also a stage which can occasionally be skipped 783 FISHERY BULLETIN: VOL. 76, NO. 4 100 5 10 DAYS IN STAGE 784 Figure l.— Percentage survival and duration of stages in larvae of Galathea rostrata reared under laboratory conditions at 15°C (upper 2 graphs) and 20°C (lower graph). Inset at 20°C gives duration and survival of regular (IVr, dashed line) and advanced (IVa, solid line) stages; number of days the same as in larger graphs. GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA by some 20°C larvae, and thus could be thought of as an intercalated stage, if the advanced stage IV be considered more indicative of the developmen- tal sequence. Other features shared between the advanced fourth and regular fifth zoeal stages (besides the presence of well-developed pleopod buds noted earlier) include increased numbers of antennular aesthetascs, a remarkable elongation of the antennal endopodite, the appearance of a mandibular palp, and slight changes in setae number on maxillulae, maxillae, maxillipeds, and telsonal uropods (see section on Description of the Larvae). Moreover, the advanced stage IV zoeae were always larger than the regular stage IV zoeae. It will probably remain a question of semantics whether the regular stage IV is considered an in- tercalated stage or one that occasionally may be skipped. It could just as well be asked whether the advanced stage IV was an intercalated stage be- cause it embodies many of the features of regular stage IV, plus some seen only in stage V zoeae in the developmental sequence. What is of more im- portance in the development of G. rostrata is that the substitution of an advanced stage IV and the subsequent elimination of the regular stages IV and V allows earlier postlarval metamorphosis. The resultant early benthic crab stages may be reached in a shorter period of time by the species, thereby reducing the time spent in" the plankton. Discussion It is, of course, conjectural as to whether the larvae of G. rostrata skip stages in their develop- ment in the natural environment or are ever sub- ject to constant low (e.g., 15°C) or intermediate (20°C) seawater temperatures. The adults of the species, found in deeper continental shelf waters, presumably are often exposed to cool seawater temperatures, as was noted, e.g., during the time the adult females for this study were collected. It is not unreasonable to assume that developmental stages may occasionally be subjected to relatively constant cool temperatures as well, either im- mediately after hatching or just prior to postlarval metamorphosis when the megalopae settle to the sea floor. In addition, should the larvae become entrained in cyclonic cold core rings of Gulf Stream origin (see Richardson 1976; Wiebe 1976; Wiebe et al. 1976), they would presumably be sub- jected to relatively constant cold water (at least 17°C) for at least part of their developmental period. Delayed metamorphosis provides an alter- native hypothesis against the more traditional "stepping-stone" idea, to account for the rather extensive distribution of the species along the Middle and North American continental shelves. There is some evidence that larvae of other species of Galathea may skip stages in the plankton (Lebour 1930, 1931) and that other galatheids may intercalate substages (e.g., Boyd and Johnson 1963). For example, the larvae of four of the five British galatheids described by Lebour, viz. Munida rugosa (Fabricius 1775 [as M. banffica = M. bamffica (Pennant 1777)]), Galathea inter- media Lilljeborg 1851, G. squamifera Leach 1814, and G. strigosa (Linnaeus 1767) developed through four zoeal stages, whereas G. dispersa Bate 1859, exhibited four or five stages. Lebour ( 1930) considered five stages in the latter species as "probably normal" but pointed out that the megalopa could be obtained from the fourth [numerical] stage, and "the normally fifth [numerical] stage has been seen to emerge from the third stage." She stated that the fourth or fifth stage may therefore be omitted in G. dispersa, but made no mention of intercalated stages or sub- stages. The developmental situation in G. dispersa is quite similar to that noted in this report for G. rostrata, in which an advanced fourth stage re- places the regular fourth and fifth stages, thereby causing them to be omitted from the developmen- tal sequence. Lebour's (1930) "fifth stage. . . from third" is probably equivalent to what is termed in this report the advanced fourth stage. Her state- ment that long, unjointed pleopods appear in the "last" stage of G. dispersa indicates that either the fourth stage (or advanced) or fifth stage (or regu- lar) possess these appendages, depending on whichever stage is "last." It also indicates that the molt to megalopa does not occur without the ap- pearance of pleopods in the "last" larval stage. However, North Sea species of Galathea differ from G. rostrata in possessing pleopod primordia "in the third stage" which are "long but unjointed in the last stage" (Lebour 1930). In addition, Sars (1889) had also noted and illustrated pleopod de- velopment in the "last" stage of larvae attributed to G. intermedia, Munida rugosa, and Munidopsis [as Galathodes]tridentata (Esmark 1857). The lat- ter species will be considered further below. Rayner (1935), using planktonic stages from Argentinian waters, described the larvae he at- tributed to Munida gregaria (Fabricius 1793) and 785 FISHERY BULLETIN: VOL. 76. NO. 4 M. subrugosa (White 1847). Rayner did not note any substages or skipped stages in the five instars he descritjed for the two species, and was not cer- tain whether additional stages followed. By anal- ogy with M. rugosa [as M. bamffica] he thought it possible that the next stage would be postlarval. In this he was probably correct, but it seems strange in retrospect that Rayner did not attach impor- tance to the well-developed pleopods on the larvae before him, a feature by which he earlier charac- terized the fifth zoeal stage. These appendages in other galatheid larvae are quite obviously de- veloped at stage V (see Lebour, Sars, Boyd and Johnson, and others), and Sars (1889) even drew attention to them when describing his "last zoeal stage." Intercalation of substages, however, is known in the genus Pleuroncodes, as was specifically discus- sed by Boyd and Johnson ( 1963) in the larvae of P. planipes Stimpson 1860.^ Five zoeal stages had been initially noted in this species (Boyd 1960), but a sixth stage, apparently unnatural and not known to occur in the plankton, could be induced in the laboratory. Boyd and Johnson thought this stage was due to the presence of penicillin pills or to the CaCOg buffer in the pills, used to control bacterial growth in the cultures. These authors also stated that numerical stage IV could be sub- divided into a complex of from four to nine sub- stages, each represented by a molt, all without pleopods, but otherwise morphologically similar to each other. Although no sequential substages were skipped (e.g., a molt from substage IVa to IVh), one or more substages could be omitted ter- minally, with a subsequent molt to the morpholog- ically discrete stage V, which possessed pleopods (Boyd 1960). Boyd and Johnson suggested that in P. planipes the number of substages in stage IV was probably influenced by temperature, with higher culture temperatures (e.g., 16°-20°C) pro- ducing faster development but causing more sub- stages to occur before the molt to stage V. They noted, however, that other factors such as food supply or crowding of larvae might also exert an effect on the number of substage instars, but ne- glected to consider the possibility that the large number of induced substages in stage IV might also be due to the use of antibiotics in the cultures, as suggested by Fagetti and Campodonico ( 1971 ). ^'Both Stimpson (1860) in his original description of Pleuron- codes planipes and Haig (1955) have suggested that the species may prove to be only a northern Pacific form of the Chilean P. Monodon. 786 Figure 2. — Galathea rostrata, zoeal stages in lateral and dorsal view: (A, a) First zoea; (B, b) second zoea; (C, c) third zoea; (D, d) fourth zoea (regular); (E, e) fifth zoea. Scale line equals 1.0 mm. The Chilean congener, Pleuroncodes monodon Milne Edwards 1837, also was found to have inter- calated substages (Fagetti and Campodonico 1971). At 15°C, substage IVa-d were followed by a molt to stage V, possessing pleopods; at 20°C a fifth substage (IVe) was attained instead of zoeal stage V. Whether stage IVe would be followed by ecdy- sial stage V is not known because all larvae in stage IVe died. However, the lack of pleopods in stage IVe implies that stage V should occur, with pleopods, before the molt to megalopa takes place. Whether such substages occur in the plankton is conjectural , but they certainly would present some difficulty in separation because of their great simi- larity to each other in samples collected from the plankton. Abbreviated larval development is also known to occur in at least two galatheids. Sars ( 1889), in describing the prezoel, "first" and "last" zoeal stages of Munidopsis tridentata from Norwegian waters suspected that development time was shorter than that seen in Galathea , but came to no conclusion as to the total number of stages. He commented on the remarkably advanced features exhibited in the early zoea, an observation later supported by Samuelsen (1972). Samuelsen de- termined that only three zoeal stages exist for M. tridentata and further suggested that the megalopal stage followed stage III because the latter stage was in the same relative state of de- velopment as some fourth zoeae which preceded the megalopae in other galatheids. Samuelsen noted that the presence of a mandibular palp, pleopod primordia, antennular aesthetascs, an- tennal setae, and scaphognathite setae in the early zoeal stages were all advanced features usu- ally restricted to later zoeae in other galatheid larvae. The relatively nonsetose feeding append- ages and endopodites of the natatory appendages indicate that the larvae may not feed, although they can swim well. Al-Kholy (1959) described and figured larvae attributed to a "Galathea sp." which apparently developed through only three zoeal stages. How- ever, no methodology was given, nor indication as to whether the larvae were cultured in the laboratory or collected from the plankton. It is doubtful whether the species will ever be identi- GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA 787 FISHERY BULLETIN: VOL. 76, NO. 4 fiable based on his incomplete descriptions and rather stylized illustrations. Advanced development is implied in but one galatheid, the cave-dwelling Munidopsis polymorpha Koelbel 1892. This species is pres- ently known only from a littoral cave formed by lava tunnels which connect to the sea in the Ca- nary Islands (Fage and Monod 1936). These au- thors never found more than five, extremely large (1.5-1.8 mm in diameter) eggs on an individual female. No larval stages were described, but it was hypothesized that the young Munidopsis was well advanced in development inside the egg and prob- ably hatched into a form nearly like the adult. Given the rather unique habitat for a Munidopsis, advanced development in M. polymorpha would not be surprising. The vast majority of other species of Munidopsis are deep-sea forms, most of which occur below 500 m (Mayo 1974) in the At- lantic Ocean, although some species occur in shal- lower waters on the continental shelf. In summary, it is apparent that larval develop- ment in the Galatheidae is quite diverse, includ- ing advanced development (i.e., with imminent metamorphosis) in the cave dwelling M. polymorpha , abbreviated development with as few as three larval stages (M. tridentata, Al-Kholy's Galathea sp.?), to "normal" development of four- five zoeal stages (e.g., Munida, Galathea). Sub- stage intercalation is known in the genxx^Pleuron- codes, but seems to be restricted to the fourth, or penultimate, ecdysial stage. Intercalation of a sixth zoeal stage, perhaps only a laboratory ar- tifact, is also known in one species of this genus. Skipped stages appear only in two species of Galathea, and perhaps one of Munida, at present, and these result in the elimination of regular zoeal stages IV and V and their replacement by an ad- vanced stage IV which subsequently molts di- rectly to megalopa. Developmental variation such as that just dis- cussed allows some interesting speculation as to its evolutionary consequences in view of the fact that the phylogenetically closely related anomu- ran family Porcellanidae generally undergo a re- ■"The term "direct" development is restricted in this paper to those larvae which hatch from the egg in a form morphologically similar to the adult and undergo no further metamorphosis. Larvae exhibiting "advanced" development usually hatch in the penultimate or ultimate zoeal stage and thus may undergo addi- tional ecdysis prior to metamorphosis. Larvae with "ab- breviated" development hatch as early zoeae (often with a pre- zoeal or first zoeal stage present), but may dispense with one or more intermediate stages in completing their larval develop- ment. duced developmental sequence of usually no more than two zoeal stages. These stages appear to be morphologically equivalent in most respects to Galathea stages I and IV, sensu lato. Substages have been postulated for some porcellanid larvae, notably Indo-Pacific species, but are not positively known to occur in Atlantic and eastern Pacific species. Previously postulated substages in Atlan- tic species have been shown to be the result of accelerated morphological development without an ensuing molt and have been seen primarily in larvae collected from the plankton (Gore 1968 and others). However, the larvae of the western Pacific genus Petrocheles apparently do reflect their galatheid ancestry by undergoing five zoeal stages during development. Morphological features of the telson, uropods, and antennal scale in these larvae all resemble, to a greater or lesser degree, their counterparts in larvae of Galathea and Munida (Wear 1965). Further studies along these lines should be most interesting and productive. DESCRIPTION OF THE LARVAE First Zoea Carapace length: 1.0 mm. Number of specimens examined: 10. Carapace: (Figure 2A, a). Typically galatheid, somewhat inflated; rostral spine horizontal, little expanded proximally, straight, extending to level of scapherocerite spine, or slightly beyond, about 0.5 X carapace length(CL), unarmed; posterolat- eral carapace margins armed with a series of about 15 small denticles placed before large, pos- terior spine; latter slightly more than 0.1 x CL; dorsomedial carapace margin excavated, with about 13 small denticles along sinus margin. Two small setae medially above eyes; latter sessile. Antennule: (Figure 3A). A simple rod, both en- dopodite and exopodite fused to protopodite; former with 1 elongate plumose seta, latter with 3 aesthetascs and 3 setae. Antenna: (Figure 3B). Endopodite rodlike, about 0.4 X scaphocerite length, fused to protopo- dite, a single distinct spine at its tip, plus a long plumose seta; scaphocerite usually with 9 setae along margin, tip produced into long daggerlike spine about 0.3 x total scale length; protopodite 788 GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA Figure 3. — Galathea rostrata, first zoeal appendages: (A) Antermule; (B) antenna; (C) mandibles, lower view rotated interiorad to zoea to show dentition; (D) maxillule; (E) maxilla; (F) maxilliped 1; (G) maxillped 2; (H) maxilliped 3; (I) telson. Scale lines total 0.3 mm. 789 FISHERY BULLETIN: VOL. 76, NO. 4 with sharply pointed spine ventrally, armed along either side with distinct acute spinules; this spine falling short of distal tip of endopodite; scattered setae basally on protopodite. Mandibles: (Figure 3C). Asymmetrical den- tate and spined processes, as shown. Maxillule: (Figure 3D). Endopodite seg- mented, 3 terminal, 1 subterminal seta. Basal en- dite with 2 large, widely separated strong spines, plus 3 setae; coxal endite with 4 spines, 3 strong setae. Maxilla: (Figure 3El. Endopodite setae, pro- gressing subterminally, 3-4. 3. plus 3 laterally, and additional fine hairs as illustrated. Basal en- dite proximal and distal lobes each with 3 regular and 1 spinelike seta; coxal endite proximal and distal lobes with 8, and 4 spinelike setae, respec- tively. Scaphognathite with 4 lateral, 1 stout elongate apical seta. Maxilliped 1: (Figure 3F). Coxopodite with 2 setae. Basipodite setae formula progressing dis- tally 2, 3, 3, 3. Endopodite five-segmented, setae progressing distally 3, 2, 1, 2, 4 +1 (Roman nu- meral denotes dorsal setae); all endopodal and basipodal setae heavy, spikelike. Exopodite two- segmented, 4 natatory setae. Maxilliped 2: (Figure 3G). Coxopodite naked. Basipodite setae 1, 2, progressing distally. En- dopodite four-segmented, setal formula 2, 2,2,4-1- I; all spikelike. Exopodite two-segmented, 4 natatory setae. Maxilliped 3: (Figure 3H). A small, unseg- mented amorphous bud. Pereiopods: Appear as small and undifferen- tiated buds, gradually enlarging as stage progres- ses. Abdomen: (Figure 2A, a). Five somites; last 2 with large lateral spines; somites 2-5 each with paired setae dorsally, plus a series of small distinct spinules along posterior margin of somite; somite 6 fused to telson. Pleopods absent. Telson: (Figure 31). Setal formula on margin 7 + 7; all plumose setae (= processes 3-7) with small, hooklike spinules progressing down their 790 length; other setae and hairs as illustrated. Anal spine absent. Color: Zoea transparent; frontal region of carapace diffused with orange, brighter orange dorsally on midgut region. Chromatophores as fol- lows: orange on protopodite of antennule, faintly orange on scaphocerite of antenna; red-orange around inner oral region; mandibles and labrum outlined in red, interiorly orange; basipodites of maxillipeds 1 and 2 red-orange along dorsal and ventral margins; red spiderlike chromatophores dorsally in longitudinal line on abdominal somites 3-5; orange chromatophores ventrally placed in a similar manner. Eyes black, with bluish high- lights in reflected light. Second Zoea Carapace length: 1.2 mm. Number of specimens examined: 8. Carapace: (Figure 2B, b). More inflated; ros- tral spine more or less knifelike in lateral view, noticeably expanded proximally in dorsal view; about 0.5 X CL, overreaching distal tip of antennal scaphocerite spine in several specimens, unarmed; posterolateral margins of carapace with about 14 small denticles or spinules, dorsomedial margin possessing only scattered nubs or with denticles totally absent; posterior spines remain slightly more than 0.1 x CL; eyes now stalked. Antennule: (Figure 4A). Incipient segmenta- tion seen at junction of exopodite with protopodite; former usually carrying 4 aesthetascs and setae, with 4 small thick setules on junction with pro- topodite. Endopodite retains single long plumose seta. Antenna: (Figure 4B). Endopodite thickened, drawn into point distally, appearing conical, about 0.3 X scaphocerite length, incompletely fused to protopodite, now lacking elongate plumose seta seen in first stage. Scaphocerite usually with 10 marginal setae, plus numerous small spinules ventrally along outer margin; distal spine about Figure 4. — Galathea rostrata, second zoeal appendages: (A) Antennule; (B) antenna; (C) mandibles; (D) maxillule; (E) maxilla; (F) maxilliped 1; (G) maxilliped 2; (H) maxilliped 3; (I) telson. Scale lines total 0.3 mm. GORE; LARVAL DEVELOPMENT OF GALATHEA ROSTRATA A-H y 791 0.2 X scale length. Protopodite now carries second sharp spine ventrally, armed as first along outer margins; larger ventral spine now shorter than endopodite; ventral spinulelike setae inconspicu- ous or lacking. Mandibles: (Figure 4C). Dentition now larger, more complex. No palp. Maxillule: (Figure 4D). Endopodite unchanged from stage I. Basal endite with 4 large spines, 3 setae: coxal endite processes stronger, but number unchanged, from stage I, appearing to be 5 spines, 2 strong setae. Maxilla: (Figure 4E). Endopodite setal formula progressing subterminally 4, 2, plus 3 lat- erally, and fine hairs as shown. Basal endite proximal and distal lobes with 4-5, 6 processes, respectively, former as 3-4 spinelike and 1 thin seta, latter as 1 strong and 1 regular spine, 4 thin setae. Coxal endite distal lobe with 3 spines, 1 strong seta, proximal lobe with 8 spines or strong spinelike setae. Scaphocerite with 6 lateral, plus usual elongate apical seta. Maxilliped 1: (Figure 4F). Coxopodite and basipodite setae unchanged from stage I. Endopo- dite setal formula now 3, 2 -^ I, 1 + I, 2, 4 + I. Exopodite remains two-segmented throughout later development, now with 7 natatory setae. Maxilliped 2: (Figure 4G). Coxopodite and basipodite as in stage I. Endopodite setal formula 2, 2 -f- 1, 2, 5 + I. Exopodite as above and for later stages, carrying at this stage 7 natatory setae. Maxilliped 3: (Figure 4H). Remarkably de- veloped; incompletely two-segmented exopodite with 6 natatory setae; endopodite poorly calcified, originating about half way up basipodite, two- segmented, with 2 terminal setae. Pereiopods: (Figure 2B). Undifferentiated, but enlarging buds throughout stage. Abdomen: (Figure 2B, b). Five somites, sixth still fused to telson; lateral spine on somite 5 distinct, that of somite 4 reduced, even vestigial; paired dorsal setae on posterior dorsal margins of somites 2-5 remain, and are present throughout later zoeal stages; posterior marginal spinules much reduced in size and number. 792 FISHERY BULLETIN: VOL. 76, NO. 4 Telson: (Figure 41). Marginal setal formula 8 + 8, additional pair added in medial sinus; latter reduced from distinct U-shaped notch seen in stage I. Armature on plumose processes as before, but distal tips with hooklike processes more dis- tinct; other setae and hairs as shown. Color: Similar to stage I, but with less diffusion of orange frontally; internal midgut region, man- dibles and maxillipedal basipodites retain red- orange color, mandibles showing noticeable red outline, maxillipedal color appearing more dif- fused than stage I; abdominal somites 4-5 with red dorsal and lateral chromatophore lines, plus orange line ventrally, all connecting to single orange ring of spiderlike chromatophores around each anterior margin of somites 4 and 5. Eyes electric blue to black in reflected light. Third Zoea Carapace length: 1.3 mm. Number of specimens examined: 8. Carapace: (Figure 2C, c). Proximal margins of rostral spine more developed laterally when seen dorsally, in this and subsequent stages: length remains about 0.4-0.6 x CL, distal tip reaches to about tip of scaphocerite spine or slightly beyond; posterolateral margins of carapace with denticles much reduced, becoming irregular nubs; dor- somedial margin with only poorly developed, rag- ged nubs, almost totally obsolete; posterior carapace spines considerably shortened, less than 0.1 X CL. Eyes much enlarged, basal peduncles elongate. Antennule: (Figure 5A). Exopodite segmented from protopodite, bearing 2 lateral aesthetascs in addition to 3 terminal, plus 3 or 4 setae, at tip. Endopodite slightly enlarged, retaining long plumose seta. Protopodite carries single long lat- eral seta distally, plus 2 short fine setae, placed medially, and basally, and 4 short stout setae dis- tally. Antenna: (Figure 5B). Endopodite continues to develop, but remains incompletely segmented from protopodite, now about 0.5-0.6 x scaphocerite length, a thin seta just below spinous tip. Scapho- cerite with 9-11 marginal setae, number some- what variable on left and right appendages in same specimen, plus additional shorter spinules GORE; LARVAL DEVELOPMENT OF GALATHEA ROSTRATA A-H ► Figure b.—Galathea rostrata, third zoeal appendages: (A) Antennule; (B) antenna; (C) mandibles; (D) maxillule; (E) maxilla; (F) maxilliped 1; (G) maxilliped 2; (H) maxilliped 3; (I) telson. Scale lines total 0.3 mm. 793 FISHERY BULLETIN: VOL. 76, NO. 4 along ventral outer margin; distal spine shortened to about 0.1 X scale length. Protopodite retains 2 sharp ventral spines, larger about 0.6 x endopo- dite length, smaller about 0.3 x length of larger. Mandibles: ( Figure 5C ). Incisor and molar pro- cesses more developed; no palp. Maxillule: (Figure 5D). Endopodite un- changed. Basal and coxal endites both with 5 spines, 3 setae. Maxilla: (Figure 5E). Endopodite unchanged. Numbers and form of processes on either endite little changed from earlier stage, with exception of basal endite distal lobe; latter now with 1 spine, 4 strong setae, 1 thin seta. Scaphognathite with 10 marginal setae and usual thick apical seta. Maxillipeds 1 and 2: (Figures 5F, G). Coxal, basipodal, dorsal and ventral endopodal, and exopodal natatory setae as in previous stage. Maxilliped 3: (Figure 5H). Endopodite bud now subequal to basipodite length, incipient seg- mentation more prominent in some specimens than others; exopodite with 7 natatory setae. Pereiopods: (Figure 2C, and detailed in- set). More developed, many with incipient seg- mentation; partial chelation of protochela often visible. Abdomen: (Figure 2C, c). Six somites present, sixth divided from telson; distinct lateral spine remains only on somite 5; spinules on dorsal poste- rior margins vestigial, ragged and irregular. In some specimens small, amorphous swellings occur ventrally on somites 2-5, signifying future posi- tion of pleopod buds. Telson: (Figure 51). Marginal setal formula remains 8-1-8; fourth pair of processes elongate spines fused to telson; processes 3, 5-8 retain noticeable hooklike spinules distally; other dorsal setae as illustrated. Uropods present at junction of abdominal somite 6 and proximal margin of tel- son; exopods of same well developed, with variable number of marginal plumose setae (usually about 8); endopods, if present, merely foreshortened naked buds. Color: More distinctly colored than stage II. 794 Orange chromatophores: dorsally on interior margin of eyestalks, a single orange spot on carapace laterally, just above each maxilliped 1, another small grouping laterally on abdominal somite 2; diffused orange on antennular peduncle, ventrolaterally on carapace below eyes, interiorly on mouthparts and within gut region, and on en- dopodites of maxillipeds 1-3. Red chromatophores: on cutting edge of mandibles, dorsomedially and laterally on abdominal somite 4, laterally on so- mite 5, the latter appearing as if small drops of blood. Fourth Zoea (Regular) Carapace length: 1.4 mm. Number of specimens examined: 10. Carapace: (Figure 2D, d). Rostral spine with noticeably raised lateral margins, greatly ex- panded basally at point of attachment to carapace, slightly overreaching scaphocerite spine and an- tennular exopodite; carapace posterolateral and dorsomedial margins unarmed; posterior spines quite short, hooked downward. Eyes large, on elongated stalks. Antennule: (Figure 6A). Exopodite with three rows of lateral aesthetascs, numbering distally 2-3, 3, 2-3, in addition to usual 3, plus 3 setae, at tip. Endopodite about 0.5 x length of exopodite, plumose seta absent. Protopodite retains distal lateral setal, plus usual 4 stout setae at junction of exopodite; 3 medial, 2 basal setae now also pres- ent. Antenna: (Figure 6B). Endopodite elongate sub- equal to scaphocerite length; latter bearing 10-12 (numbers variable on left and right appendages in same specimen) marginal setae plus numerous ventral spinules on outer margin. Larger pro- topodal ventral spine about 0.3 x endopodite length, smaller about half size of larger; armature of both remains as in earlier stages. Figure 6. — Galathea rostrata, fourth zoeal appendages: (A) Antennule , regular stage; ( a ) same , advanced stage; ( B ) antenna, regular stage; (b) same, advanced stage; (C) mandibles, regular stage; (c) same, advanced stage; (D) maxillule; (E) maxilla; (F) maxilliped 1; (G) maxilliped 2; (H) maxilliped 3; (I) telson; all regular stage; (i) detail, fifth telsonal process, 40 x objective. See test for discussion. Scale lines total 0.3 mm. GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA A-H ^ 795 FISHERY BULLETIN: VOL. 76, NO. 4 Mandibles: (Figure 6C). Molar and incisor pro- cesses acutely spinous, otherwise unchanged from earlier stages; no palp. Maxilliile: (Figure 6D). Endopodite un- changed, but may have small seta at base of seg- ment. Basal endite with 7 stout spines, 3 setae; coxal endite with 5 or 6 long spines, 3 strong setae, and occasional small tooth. Maxilla: (Figure 6E). Endopodite unchanged. Basal endite distal lobe with 6 spines, 2 setae, proximal lobe with 7 or 8 spines and strong setae intermixed; coxal endite distal lobe with 4 termi- nal spines, 1 lateral seta, proximal lobe with 11 spines, placed 5 terminally, 4 subterminally, 2 laterally. Scaphognathite with 17-20 marginal setae, including 2 or 3 anteriorly near base, plus usual long apical plumose seta, as shown. Maxilliped 1: (Figure 6F). Basipodite adds a single small seta proximally, ventral formula now 3, 3, 3, 3. Endopodal and coxal setae unchanged; 8 exopodal natatory setae. Maxilliped 2: (Figure 6G). Coxal and basipodal setae unchanged. Endopodite setal formula 2,2 + I, 2 + I, 5 + I. Exopodal natatory setae 8. Maxilliped 3: (Figure 6H). Endopodite over- reaches basipodite, bearing 3 setae. Exopodite with 8 natatory setae. Pereiopods: (Figure 2D). Chelation and segmen- tation more or less apparent, progressing rapidly throughout stage; entire pereiopodal mass hangs from beneath posterolateral carapace region in later stage. Abdomen: (Figure 2D, d). Lateral spine present only on somite 5; dorsal spinulation on posterior margins of somites nearly absent; paired dorsal setae remain. A short, sharp spine on pos- terolateral margin of somite 6, just above inser- tion of uropodal basipodite. Pleopod primordia may be present in some specimens, but develop- ment is weak and occurs slowly, if at all, through- out stage. Telson: (Figure 61, i). Uropods completely de- veloped, exopodite distal outer tip produced into long spine, 8-11 long marginal setae present; en- dopodites with 4 or 5 setae; with shorter setae on 796 both rami. Telson marginal setal formula 8 + 8, fused fourth process now heavily spinulose, other movable processes (except process 2, which, as in other anomurans, remains a simple seta) carry distinctive, sharp, separated, spinules along their length (Figure 6i; 40 x objective), these spinules much more hooklike distally, more spinous prox- imally. Other dorsal and ventral setae on telson as illustrated. Color: Similar to stage III; quite developed and noticeable along anterior and internal margin of eyestalks; interiorly on midgut, and bases of maxillipeds; single red chromatophores now bas- ally on antennular protopodite, on posterior mar- gin of maxillipedal basipodites, and laterally on abdominal somites 3-5; eyes blue, reflecting green highlights. Remarks: This stage, with limited aesthetasc numbers, reduced antennular endopodite and un- segmented protopodite, lacking mandibular palps, and with developing pereiopods and usually only pleopodal primordia, always molted to zoeal stage V. Fourth Zoea (Advanced) Carapace length: 1.6 mm. Number of specimens examined: 12. Carapace: Differs little from regular stage IV except being larger, more inflated; armature simi- lar to regular stage zoea. Antennule: (Figure 6a). Endopodite about 0.75 X to nearly equal to length of exopodite; latter with four rows of aesthetascs laterally, as 1, 3, 3, 2, plus usual 3, plus 3 setae terminally. Other setae as illustrated. Antenna: (Figure 6b). Endopodite distinctly overreaches (1.2x) scaphocerite; latter with 12-14 marginal setae; larger ventral propodal spine about 0.25 X endopodite length, remaining half again as long as smaller propodal spine. Mandibles: (Figure 6C). Large, heavily toothed processes, distinguished now by undivided simple palp. Maxillule: May add one more process on basal endite; tooth on coxal endite usually distinct. GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA Pereiopods: Well formed, completely segmented and chelated, protruding almost totally from under posterolateral carapace margins. Abdomen: Somites 2-5 each with a pair of undi- vided pleopod buds, gradually lengthening as stage progresses, but never becoming bifid. Color: Similar to regular stage IV zoeae. Remarks: The zoeae in this stage are much more developed morphologically, possessing a different arrangement of antennular aesthetascs, a well- developed antennal endopodite, mandibular palps, segmented and chelated pereiopods, and distinct (but undivided) pleopod buds. These zoeae molt directly to megalopae, bypassing stage V completely. Fifth Zoea Carapace length: 1.6 mm Number of specimens examined: 8. Carapace: (Figure 2E, e). Rostral spine with lateral margins appearing somewhat embossed at posterolateral angle of zoeal orbit; carapace lat- eral margins deeply rounded, convex posterolat- erally, unarmed; posterior spine recurved ventral- ly in some specimens, nearly straight in others, inner margin of same curving regularly inward to deeply excavated dorsomedial margin of carapace; latter entirely without armature. Eyes large, ovoid, on well-developed elongate stalks. Antennule: (Figure 7A). Exopodite with five rows of aesthetascs laterally: 2, 3, 3, 3, 2, plus 3 and 3 setae at tip. Endopodite from about 0.75 x to just subequal in length to exopodite. Protopodite segmented into elongate basipodite and truncated coxopodite; former with a single long plumose seta distally, plus 4 stout setae terminally at exopodite junction, 3 more medially; latter with 2 stout setae ventrally near line of segmentation. Antenna: (Figure 7B). Endopodite very notice- ably longer than scaphocerite ( 1.3-1.4 x); latter bearing 12-14 plumose marginal setae plus addi- tional ventral marginal spinules as in earlier stages. Larger propodal ventral spine less than 0.2 X endopodite length, smaller remains about half the size of larger, both armed similarly as illustrated. Toward end of larval stage trans- parent endopodite reveals distinctly segmented megalopal antennal flagellum within endopodal sheath. Mandibles: (Figure 7C). Noticeably dentate, each with simple, distinct palp. Maxillule: (Figure 7D). Endopodite un- changed from regular stage IV; basal setule may not be present. Basal endite with 8 stout spines, 3 setae; coxal endite with 6 long spines, 3 strong setae, and small tooth, placed as illustrated. Maxilla: (Figure 7E). Endopodite unchanged. Basal endite distal lobe with 8 spines and strong setae, 2 thin setae terminally, one regular seta laterally; proximal lobe with 6 terminal, 2 sub- terminal, 2 lateral processes, most appearing to be strong setae and spines. Coxal endite distal lobe with 2 spines, 2 strong apical setae, 2 thinner subapical or lateral setae; proximal lobe with about 13 spines and strong setae, progressing ter- minally to laterally as 7, 4, 2. Scaphognathite with about 22-25 marginal setae, including enlarged plumose seta apically; 2 small setules present, positioned laterally. Maxilliped 1 and 2: (Figures 7F, G). Little changed from previous stage. Maxilliped 3: (Figure 7H). Little changed in form from previous stage, except endopodite now much larger, longer, extending well past distal margin of basipodite; 3 setae as before. Pereiopods: (Figure 2E). Extremely large, ap- pearing to be nearly functional, protruding be- neath, and forcing posterolateral margins of carapace, outward; walking leg segmentation and cheliped chelation distinctly visible. Abdomen: (Figure 2E, e). Lateral spine on so- mite 5, and that on posterodistal angle of somite 6, the only armature. Pleopods present as well-developed, bifid, buds. Telson: (Figure 71, i). Uropods well-developed, both endopodite and exopodite with variable number of marginal setae, usually 8-10, and 10-13 or occasionally 14, respectively. Telsonal fused and movable processes as illustrated; fourth pro- cess distinctly spinulose; occasionally an extra 797 FISHERY BULLETIN: VOL. 76, NO. 4 798 GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTHATA FlClRE 7. — Galathea rostrata. fifth zoeal appendages: (A) An- tennule; (B) antenna; iCi mandibles, lower view rotated ex- teriorad ofzoea to show dentition; (Dl maxillule; (E) maxilla; (F) maxilliped 1; iGi maxilliped 2; (H) maxilliped 3; (Ii telson; (ii detail i,40 x objective), fifth telsonal process. Scale lines total 0.3 mm. plumose process appears making telson marginal setal formula 8 + 9 as shown. Color: Chromatophores as follows: Red, on an- terior margin of eyestalks, paired on carapace dor- sally just behind eyes on frontal region, single ventrolaterally beneath each eyestalk just above mandibular region, ventrally on both antennular and antennal peduncles at junction with carapace, laterally on carapace above insertion of maxil- liped 2; interiorly on mouth region on outer mar- gin of mandible, posterior to mandible on maxil- lule, and on midgut; abdominal somites 3-5 with several groups laterally, plus a reddish-orange line above hindgut of same. Orange chromatophores in elongate streaks longitudi- nally on basipodite of maxillipeds 1-3, more dif- fused on maxillipedal endopodites, and in lateral groupings on abdominal somites 3-5. Eyes blue- green in reflected light, corneas dark, probably black. Remarks: This stage followed the regular stage IV and invariably molted to megalopal stage. Megalopa Carapace length x width: 1.7 x 1.2 mm. Number of specimens examined: 4. Carapace: (Figure 8A, B). Resembling minia- ture adult; rostrum triangular proximally, drawn into sharp point distally, armed along lateral margins with 4 distinct spines, some smaller spinules occasionally interspersed; frontal region with additional spinules as illustrated; 2 elongate thickened setae on gastric region, plus other setae and spinules as shown; lateral margins with 4 large spines, including 1 at epibranchial angle, 2 placed about equidistant behind, and the fourth at junction with cervical groove; a variable number, usually 3, smaller spines laterally between larger spines; a fifth large spine on posterolateral mar- gin, followed by another, smaller, dorsally and posteriorly. Numerous small setae scattered over entire carapace; eyes each with 2 large, feathery setae on anterodorsal margin. AnU'iuutle: ( Figure 9A, a). Biramous; peduncle large, three-segmented; basal segment inflated, with 2 large forward-directed spines dorsally, another, smaller, distoventrally; other setae as shown; remaining two segments nearly smooth, sparsely setose. Lower ramus three-segmented, tip with 2 spinules (see detail. Figure 9a), other setae as shown. Upper ramus six- or occasionally indistinctly seven-segmented; aesthetascs on segments two through five in the following se- quence of rows and numbers: one row (2), two rows (3, 3, + 2 setae), two rows (3, 2,-1-1 seta), one row < 2 ); sixth segment with a single elongate terminal seta plus other smaller setae. Antenna: (Figure 9B). Peduncle three- segmented, heavily spined; flagellum with 2 or 3 fused segments plus a variable number ( about 24) shorter segments each bearing 5 or 6 setae dis- tally; terminal segment with 7 longer setae, as illustrated. Mandible: (Figure 9C). Symmetrical, scoop- shaped process, chitinized along leading margin; a three-segmented palp, basal segment of which bearing 2 short, spinelike setae, third segment with about 13 or 14 stout, toothed spines. Maxillule: (Figure 9D). Endopodite now pos- sessing but a single short, terminal seta. Basal endite with 4 strong terminal setae, followed by 16 short, stout spines, 4 subterminal and 3 lateral setae; a single seta basally as shown; coxal endite lower portion extended into elongate, weakly chitinized, lobe fringed with fine hairs; 3 basal setae, 3 lateral setae, followed by 11 stout spines and 8 strong setae terminally. Maxilla: (Figure 9E, e). Endopodite with a single, long subterminal seta. Coxal and basal en- dites heavily spinose and setose, numbers and po- sition difficult to discern, but approximately as follows: basal distal lobe, about 14 terminally, 4 + 2 subterminally, 2 laterally; proximal lobe, about 6 terminally, 3 + 1 subterminally, 1+2 laterally; coxal distal lobe, about 3 each, terminally and subterminally, 2 + 8 in irregular row laterally; proximal lobe, about 11 placed more or less termi- nally, 8 subterminally, 22 in a row encircling lobe laterally, 1+2 beneath these; for exact position- ing refer to outer (Figure 9E) or inner (Figure 9e) views of lobes. Scaphognathite with about 40 799 FISHERY BULLETIN; VOL. 76, NO. 4 yj ff:l // ,/V/\ \ V: 'r^ ^ y/ _^ rf K J /■ / 800 B Figure 8. — Galathea rostrata, megalopal stage: (A) lateral view; (B) dorsal view. Scale line equals 1.0 mm. GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA Figure 9. — Galathea ros^rato, megalopal appendages: (A) An- tennule; (a) detail, tip of lower ramus; (B) antenna; (C) mandible; (D) maxillule; (E) maxilla, exterior surface; (e) same, interior surface showing only inner setation; (F) maxilliped 1; (G) maxil- liped 2; (H) maxilliped 3; (I ) pereiopod 4; (J) cheliped, pereiopod 5; (K) pleopod 1 (right) and 4 (left); (L) tail fan. Scale lines total 0.3 mm. AAi.iai- a.C-H.J I- 801 FISHERY BULLETIN: VOL. 76, NO. 4 marginal setae plus finer setules on either side of upper lateral surface. Maxilliped 1: (Figure 9F). Exopodite and en- dopodite weakly chitinized; former two- segmented, with 3 more or less terminal setae; latter naked. Protopodite with about 27 and 15 setae on basal and coxal endites, respectively, placed as illustrated. Maxilliped 2: (Figure 9G). Exopodite two- segmented, 8 terminal, plus other setae, as shown. Endopodite four-segmented, proximal two with 4 and 2 setae, respectively, distal two each with about 12 processes, including 5 daggerlike spines terminally. Setae on basipodite and coxopodite as illustrated. Maxilliped 3: (Figure 9H). Exopodite two- segmented, 8 terminal setae. Endopodite five- segmented; ischium and merus each with strong, sharp triangular spine, plus a shorter spine at anterodistal angle; ischium also with prominent crista dentata; last three segments (carpus, prop- odus, dactylus) with 3, about 15, about 18 long daggerlike spines plus numerous longer setae in- terspersed among them. Several setae on coxopo- dite and basipodite. Pereiopods: (Figures 8 A, B; 91, J). Chelipeds rounded, equal, elongate, heavily spined, covered with long, stiff bristlelike setae, these more prom- inent in gape of fingers and on outer surface of manus; fingers of each hand trifid at tips. Merus and carpus with marginal spines. Walking legs thin, elongate; merus, carpus, and propodus cov- ered with setae plus small spinules ventrally along margins, these often difficult to discern ex- cept under higher (40x objective) magnification; propodus with 2 larger spinules ventrally; dac- tylus with 3 large movable spinules plus one fixed triangular tooth on ventral margin; a second, very small, almost vestigial tooth may appear about midway between larger fixed tooth and dactylar tip. Pereiopod 5 chelate, 1 long serrated seta, 3 scythelike pectinate setae quite noticeable, plus additional numerous setules on manus; 2 vei'y small, spinulelike teeth on distal tip of dactylus. Pleopods: (Figures 8B, 9K). Occur on somites 2-5; biramous, greatly elongate; exopodal setae progressing toward telson 8, 8, 8, 7, with minor variation of 1 or 2 occasionally seen on left or right 802 side in same specimen; endopodites not as long as exopodites, thin, naked, but each with appendix interna of 2 or 3 small hooks developed at tip. Tail Fan: (Figure 9L). Telson with 6 or 7 long plumose setae, and several shorter marginal setae interspersed among these, numbers of latter in- consistent in same specimen; 1 or 2 small toothlike spines laterally, as shown. Uropods biramous, each with 4 widely separated marginal spines, that on outer lateral margin of endopodite the strongest; exopodite with about 18-20, endopodite blade with 11-14 plumose marginal setae, num- bers again variable in same specimen. Smaller setae on dorsal and ventral surfaces of tail fan as illustrated. Color: Megalopa beautifully colored. Carapace and abdomen overall red-orange, dorsolateral carapace margins and spines darker red; an ir- regular longitudinal white, or semitranslucent stripe extends dorsally from just behind frontal region along entire length of carapace and ab- domen; this stripe bordered with darker orange- red chromatophores along its length; a similar white stripe appears laterally, below which carapace becomes translucent, but covered with numerous red spiderlike chromatophores; a third stripe appears ventrally on sternum, extending to junction of abdomen. Numerous pale blueish- white dots interspersed over dorsal surface of carapace, especially on either side of previously noted longitudinal stripe. Eyestalks orange-red, with regular white band longitudinally, this meet- ing second band which encircles distal margin of eyestalk just before cornea; latter black, overlain with dark red maculations. Chelipeds with distal margin of mei'us, entire carpus, propodus, and all but distal tip of dactylus ivory white; merus prox- imally red-orange; cheliped finger tips orange. Walking legs translucent, speckled with many red-orange chromatophores, these coalescing to form irregular bands on outer segments; dactyli of latter clear, or light horn color. DISCUSSION In the western North Atlantic Ocean the family Galatheidae is represented by four genera: Galathea (2 species), Munida (31 species), Munidopsis (48 species), and Phylladiorhynchus (1 species). With the exception of the present re- port, the larval development of the remaining 81 GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA galatheid species known to occur in the western Atlantic is unknown. The majority of our knowledge on galatheid lar- vae comes from studies conducted on species from the eastern Atlantic and Pacific Oceans, and as- sociated seas such as the Red, Mediterranean, or North Seas. Lebour ( 1930) first characterized lar- vae in the family Galatheidae, and Gurney ( 1942) was the first to provide a synopsis of larval fea- tures based on Lebour's work and studies he made on western Pacific galatheid larvae. As might be expected, only some of the characters considered important by Gurney in 1942 remain valid today, and the lack of detailed descriptions in earlier studies on galatheid larval morphology prevents comparative statements to be made among most of the species for which the larvae are known. Nevertheless, morphological differences in ros- tral, carapacial, antennal, abdominal, and tel- sonal features continue to be of some value in distinguishing the larvae of at least five galatheid genera. In general, the larvae so far described for species of Munida share several features with those known from Galathea and Pleiironcodes, and are thus somewhat indicative (as seems true for the adults) of close relationships among the three gen- era. Pleuroncodes, an eastern Pacific genus, is morphologically very similar to Munida in several larval features, more so than are larvae of Galathea as presentlydescribed. As noted in the following synopses, the larvae of the three genera can be easily separated. The adults, based on pres- ent taxonomic criteria, are distinct and generic status is undoubtedly warranted. The genus Munidopsis, on the other hand, is a heterogeneous grouping of forms, some adults bearing little resemblance to others in the taxon (see Mayo 1974, for discussion). The larvae from the sole species so far described, however, are cer- tainly distinctive and do not resemble those from other genera. The genus Miinidopsis, as presently constituted, would seem to provide an ample example of a taxon wherein the relationships among the various species (and perhaps their ele- vation to generic status) might be clarified on the basis of morphological relationships among their larvae. The first zoeal larvae of the eastern Pacific species Ceruimunida johni (Fagetti 1960) are quite spinose but could perhaps be confused with either Munida or Pleuroncodes larvae (Fagetti and Campodonico 1971). It remains to be seen whether later larval stages would be more distinc- tive. The presence of a single ventral antennal spine (instead of two as seen in other genera) is of limited value, because Galathea and Munida exhibit a single spine in stage I and two spines in later stages (see below). In the genus Galathea, larvae are principally known from northeastern Atlantic species de- scribed by Lebour (1930, 1931) and Sars (1889). Live specimens of these species may be separated from larvae of G. ro.strata by chromatophore color and position, but unfortunately no further de- tailed comparison is possible until the former species are completely redescribed and illustrated. This holds true for most of the studies by the 19th and early 20th century authors which were listed in Gurney (1942). The "Galathea sp." briefly de- scribed and illustrated by Al-Kholy (1959) from the Red Sea agrees in several respects with "typi- cal" Galathea larvae, but differs in others. Whether it may be equated with Gurney's (1938) G. longimana remains uncertain as the brief de- scriptions and illustrations of both authors pro- hibit meaningful comparison between the two studies, and those on other Galathea larvae. In order to facilitate comparison between the two western Atlantic Galathea species a summary of larval features exhibited by G. rostrata is pro- vided in Table 2. These may be applied both to G. agossizii, when its larvae become known, and to other Galathea larvae when expanded or more complete descriptions are provided. In addition, a provisional synopsis of larval characters for the five genera discussed above is also presented. The summaries have been extracted from the more reliable larval descriptions, as so noted, and may allow distinction among the more typical larvae in each genus. As our knowledge increases further modification may be required. SYNOPSES OF GALATHEID LARVAE In the following section, emphasis is placed on the setal-spinal formulae of the larval telson. Conventionally, this formula may be expressed thusly: 8+8, indicating that eight telsonal pro- cesses, consisting of fixed and movable spines, setae, and thin hairs, occur on each side of the telsonal midline. It is apparent now that the type of these processes may provide a useful reference feature in distinguishing between the various galatheid larvae. Accordingly, spines (whether movable or fixed) are herewith denoted by Roman 803 FISHERY BULLETIN: VOL. 76, NO. 4 Table 2. — Summary of zoeal features in the larval stages oi Galathea rostrata. Zoea 1 Zoea II Zoea III Zoea IV (regular) Zoea IV (advanced) Zoea V Rostral spine Not expanded Expanded Expanded Raised lateral As in regular Expanded proximally proximally proximally margins stage proximally Posterior cara- pace spines Elongate Elongate Reduced Slightly hooked Slightly hooked Hooked Eyes Sessile Stalked Stalked. Stalked, Stalked. Stalked, greatly enlarged elongate elongate developed Antennule Simple rod. As in stage 1 Exopodite seg- 3 lateral rows 4 lateral rows 5 lateral rows no lateral Endopod more mented. 2 lat- aesthetascs aesthetascs aesthetascs aesthetascs developed eral aesthetascs Endopod "2 exo- Endopod subequal Endopod subequal Endopod reduced Protopod lacks Protopod with 1 pod length to exopod to exopod lateral seta lateral seta Protopod unseg- mented Protopod segmented Antenna Exopod with seta Exopod lacks seta Exopod "2 scapho- Exopod subequal Exopod longer Exopod much longer Scaphocerile Scaphocerite cerite length. to scaphocerite, than scaphocerite than scaphocerite spine elongate spine reduced with apical seta with apical seta with apical seta with apical seta Mandibles Without palp Without palp Without palp Without palp Palp present Palp present Maxillipeds Endopod (1) 3.2,1,2,4 + 1 3.2 + 1.1 4 1.2.4 . 1 As in previous stage and thereafter Endopod (2) 2.2,2,4 + 1 2.2 + 1.2.5 + 1 2.2 + 1.2.5 + 1 2,2 + 1.2 + 1.5+1 As in previous stage and thereafter Endopod (3) Bud Less than basipod. Subequal to basi- Longer than basi- As in regular Much longer than 2 seta pod, 2 seta pod. 3 seta stage basipod, 3 seta Exopods 1 , 2) 4 7 7 8 8 8 3) 0 6 7 8 8 8 Pereiopods Amorphous buds Developing buds Developing buds Well-formed buds Segmented, che- lated buds Large, nearly func- tional Alxlomen 5 somites, later- 5 somites, spine 6 somites, spine 6 somites, spine 6 somites 6 somites al spines 4, 5 on 4 reduced on 4 vestigial on 4 absent Pleopods absent Pleopods absent Pleopods absent Pleopod primordia may be present Pleopods present, undivided Pleopods present, bifid Uropods Absent Absent Exopods present Endopods rudimen- tary 8 + 8 setae Exopods and endo- pods developed As in previous Stage and thereafter Telson 7 + 7 setae 8+8 setae 8 + 8 setae 8 + 8 setae 8+8 setae 4th process 4th process 4th process As in previous stage and thereafter movable movable fused numerals, setae by Arabic numerals, and fine hairs by lower case Roman numerals. It should also be remembered that previously movable setae may, in a subsequent stage, become fixed spines and the setal formulae will change accordingly. Thus, a setal configuration proceeding medially of a fixed spine (I), a thin hair ( ii ), a regular seta (3), a previously movable seta now a fixed spine (IV), followed by four movable setae (5-8) results in the telsonal formula of I + ii + 3 + IV + 5-8. While somewhat more ponderous than the previously used formula of 8 -1-8, it does provide a clearer picture of the type of processes and their changes throughout subsequent larval development. Cervimunida (Fagetti I960) Rostrum elongate, needlelike, noticeably den- ticulate; carapace posterolateral and posterior margins dentate; posterior spines extremely elon- gate, reaching fifth abdominal somite; antennal scaphocerite elongate, aciculate, distinctly spined along outer margin, and upper surface, basal seg- ment with a single dorsal spine, unarmed ven- trally (thus differing noticeably from other galatheids where the situation is exactly the re- verse); abdominal somites spined dorsally, so- mites 4 and 5 with large lateral spines; telson deeply bifurcate, furcae heavily armed; setal for- mula I -H ii + 3-7 (based on first stage zoeae). Presumably four or five larval stages. Galathea (Sars 1889; Lebour 1930, 1931) Rostrum acute, often expanded at base, may be armed distally; carapace posterolateral margin usually spinulate or denticulate, posterior spine rarely exceeding third abdominal somite; anten- nal scaphocerite broad, flattened, basal segment with single spine ventrally in stage I, two spines in all other stages; posterodorsal margins of abdomi- nal somites minutely denticulate, but may become unarmed in later stages; distinct posterolateral spines on somites 4, 5, or both but may be absent later; no median dorsal spine on somite 6; telson triangular, not deeply bifurcate in early stages, becoming more elongate and truncately triangu- lar in later stages; lateral spines may be denticu- late; marginal setal formula in stages I and II of I + ii -(- 3-7, 3-8, respectively, and in all later stages I -I- ii -f 3 -I- IV + 5-8. Four or five larval stages, pleopods present in last stage, as primordia in penultimate stage on occasion. Munida (Sars 1889; Lebour 1930, 1931; Rayner 1935) Rostrum elongate, needlelike, spinulate on dis- 804 GORE: LARVAL DEVELOPMENT OF GALATHEA ROSTRATA tolateral margins and tips in early stages, but may be unarmed in later stages; a serrated posterolat- eral carapace margin with noticeable posterior spine, latter often extending to about fourth ab- dominal somite; antennal scaphocerite elongate, thin or even noticeably aciculate, often spined; basal segment with a single ventral spine in first stage, 2 in later stages; abdominal somites 2-5 with two or more spines or spinules dorsally, mar- gin of somite 6 with a single larger median spine from stage III onward; telson originally deeply bifurcate in early stages of development, but be- coming more triangularly truncate later, thus ap- pearing similar to that in Galathea in later stages; telson furcae often spined; telson setal formula I + ii + 3-7, 3-7 or -8 in stages I and II and I + ii + 3 + IV + 5-9, 5-10, 5-11 or -12 in stages III-V, respec- tively. Four of five larval stages, pleopods present in last stage. Munidopsis (Sars 1889; Samuelsen 1972) Rostrum broad, flattened, nearly spatulate in all zoeal stages, profusely armed about outer mar- gins; carapace with a large, forward-directed spine on anterolateral margin; entire ventral and pos- terolateral margins noticeably spinulate, posterior margin rounded, lacking elongate posterior spine otherwise typical of larvae in the family; antennal scaphocerite a flattened blade, two spines ven- trally; posterodorsal margins of abdominal so- mites unarmed, a posterolateral spine present on somite 5; telson broadly spatulate, posterior mar- ginal setal formula 1 + 2 -l-iii + IV -I- 5-15 in stage I, and I + 2 -t- iii -h IV + 5-15 in stages II and III; other smaller hairs interspersed among setae 5-15. Three larval stages, pleopods present in each. Pleuroncodes (Boyd I960; Fagetti and Campodonico 1971) Rostrum flattened basally, expanded similarly to that oi Galathea, distal portion acute, margins noticeably spinulate, especially at tip; posterolat- eral carapace margins serrated, elongate posterior spines usually extending to fourth abdominal so- mite; antennal scaphocerite narrow, not as acicu- late as in some Munida , basal segment with either 1 or 2 ventral spines; abdominal somites 1-5 heav- ily spined dorsally on posterior margins, becoming somewhat reduced in spination in later stages; somite 6 with a median dorsal spine in stage III and later; telson deeply bifurcate in stages I and II, becoming more truncately triangular in later stages as in Munida and Galathea; furcae may be denticulate; marginal setal formula I -I- ii + 3-7, 3-8 in stages I and II, and I + ii + 3 + IV + 5-9, 5-10, and 5-12 in stages III-V, respectively. Five larval stages, including up to eight substages in stage IV; stage VI in laboratory culture; pleopods present in stage V. The genus is presently restricted to the eastern Pacific Ocean. ACKNOWLEDGMENTS I thank my laboratory assistants Kim A. Wilson and Nina Blum for their aid in field collections and in laboratory culture of the larvae. Robert M. Avent and the Coral Biology Section of the Harbor Branch Science Foundation Laboratory obtained the ovigerous females which provided the larvae used in this study. Susan Bass and Karen Rodman recorded data and helped in the laboratory during their tenure on Harbor Branch Foundation and Smithsonian Institution Fellowships. LITERATURE CITED Al-Kholy. a. a. 1 959. Larval stages of three anomuran Crustacea (from the Red Sea). Publ. Mar. Biol. Stn., Al Ghardaqa 10:84-89. Boyd, C. M. 1960. The larval stages ofPleuroncodesplanipes Stimpson (Crustacea, Decapoda, Galatheidae). Biol. Bull. (Woods Hole) 118:17-30. Boyd. C. M., and M. w. Johnson. 1963. Variations in the larval stages of a decapod crusta- cean, Pleuroncodes planipes Stimpson (Galatheidae). Biol. Bull. (Woods Hole) 124:141-152. CHACE, F. a., Jr. 1942. Reports on the scientific results of the Atlantis ex- peditions to the West Indies, under the joint auspices of the University of Havana and Harvard University. The Anomuran Crustacea. I. Galatheidea. Torreia (Havana) 11, 106 p. Fage, L., and Th. MONOD. 1936. Biospeologica. LXin. La faunne marine due jameo de agua. Lac. souterrain de I'lle de Lanzarote (Canaries). Arch. Zool. Exp. Gen. 78:97-113. Fagetti, E. I960. Huevos y el primer estadio larval del langostino (Cervimunida johni Porter 1903). Rev. Chil. Hist. Nat. 55:33-42. Fagetti, E., and L Campodonico. 1971. Larval development of the red crab Pleuroncodes monodon (Decapoda Anomura: Galatheidae) under laboratory conditions. Mar. Biol. (Berl.) 8:70-81. Gore, R. H. 1968. The larval development of the commensal crab Polyonyx gibbesi Haig, 1956. (Crustacea: Decapo- da). Biol. Bull. (Woods Hole) 135:111-129. 805 FISHERY BULLETIN: VOL. 76, NO. 4 GURNEY. R. 1938. VIII. The larvae of Galathea longimana Paul- son. In Notes on some decapod Crustacea from the Red Sea— VI.-VIII., p. 82-84. Proc.Zool. Soc. Lond., Ser. B, 108. 1942. Larvae of decapod Crustacea. Ray Soc. (Lond.) Publ. 129, 306 p. HAIG, J. 1955. Reports of the Lund University Chile Expedition 1948-49. 20. The Crustacea Anomura of Chile. Lands Univ. Arsskr. (N.F. Avd. 2) 51(121:1-68. 1956. The Galatheidea ( Crustacea Anomura) of the Allan Hancock Atlantic Expedition with a review of the Porcel- lanidae of the western North Atlantic. Rep. Allan Han- cock Atl. Exped. 8:1-45. LEBOL'R, M. V. 1 930. The larvae of the Plymouth Galatheidae. I. Munida hanffica, Galathea strigosa and Galathea dispersa. J. Mar. Biol. Assoc. U.K. 17:175-187. 1931. The larvae of the Plymouth Galatheidae. II. Galathea squamifera and Galathea intermedia. J. Mar. Biol. Assoc. U.K. 17:385-390. Mayo, B. S. 1974. The systematics and the distribution of the deep-sea genus Munidopsis (Crustacea, Galatheidae) in the west- em Atlantic Ocean. Ph.D. Thesis, Univ. Miami, Miami, 432 p. MIYAKE, S., AND K. BABA. 1970. The Crustacea Galatheidae from the tropical- subtropical regionof West Africa, with a list of the known species. Atl. Rep. 11:61-97. RAYNEK. G. W. 1935. The Falkland species of the crustacean genus Muni- da. Discovery Rep. 10:209-245. Richardson. P. 1976. Gulf Stream rings. Oceanus 19(3):65-68. SAMUELSEN, T. J. 1972. Larvae of Munidopsis tridentata (Esmark) (De- capoda, Anomura) reared in the laboratory. Sarsia 48:91-98. Sars, G. 0. 1889. Bidrag til Kundskaben om Decapodernes For- vandlinger. II. Lithodes — Eupagurus — Spiropagurus — Galathodes — Galathea — Munida — Porcellana — Neph- rops). Arch. Math. Naturvidensk. 3:133-201. STIMPSON, W. 1860. Notes on North Amercian Crustacea in the museum of the Smithsonian Institution. No. 11. Annals Lyceum Nat. Hist. N.Y. 7:176-246. Wear, R. G. 1965. Larvae of Petrocheles spinosus Miers, 1876 (Crus- tacea, Decapoda, Anomura) with keys to New Zealand porcellanid larvae. Trans. R. Soc. N.Z., Zool. 5:147-168. WIEBE. P. 1976. The biology of cold-core rings. Oceanus 19(3):69-76. WiEBE, P. H,, E. M. HURLBURT, E. J. CARPENTER, A. E. JAHN, G. P. KNAPP III, S. H. BOYD, P. B. ORTNER, A.ND J. L. COX. 1976. Gulf Stream cold core rings: large-scale interaction sites for open ocean plankton communities. Deep-Sea Res. 23:695-710. WH.LIAMS, A. B. 1965. Marine decapod crustaceans of the Carolinas. U.S. Fish Wildl. Serv., Fish. Bull. 65:1-298. 806 A THEORETICAL EXAMINATION OF SOME ASPECTS OF THE INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES FOR YELLOWFIN TUNA, THUNNUS ALBACARES William H. Lknarz' and James R. Zweifel^ ABSTRACT This paper explores several aspects of a dual fishery (surface and longline) on yellowfin tuna, Thunnus albacares. The work is exploratory in nature and results, though indicative, are not conclusive for any specific fishery. Our results indicate that the yield per recruit is higher for the longline fishery than for surface gear if all fish are available to both gears and higher for the combined gears than for either gear fishing alone. Theeffect of fishing by one gear on yield to the other gear and the effect of the fishery on stock fecundity is shown to be greater for the often assumed 1:1 sex ratio than for the ratios usually observed. A simulation model was used to examine the interrelations of pattern of movement offish, pattern of recruitment, and fishing strategy. It was assumed that movements were random and recruitment occurred either only along the coast or throughout the fishing area. The results indicated that either of these patterns of recruitment could allow for increased catch as the surface fleet moved offshore. However, location or pattern of recruitment is shown to be important when measuring natural mortality and for examining the potential of a localized fishery, primarily on younger fish, relative to a fisherj' exploiting the full range of the stocks or to one taking primarily older fish. Tagging and fecundity studies are suggested for further investigation of the questions examined in this paper. An unsolved problem common to many of the tuna fisheries of the world is the nature of the interac- tion between longline and surface (i.e., seining, pole and line, and occasionally trolling and shal- low handline) fisheries for the same species. Fisheries for yellowfin tuna, Thunnus albacares; albacore, T. alalunga: bluefin tuna, T. thynnus; southern bluefin tuna, T. maccoyii; and bigeye tuna, T. obesus, are prosecuted by both types of gear in the Pacific, Atlantic, and Indian Oceans. Although there can be considerable overlap of sizes of fish taken by the two types of gear, in general, longline gear takes larger (older) fish. Exploitation of a tuna stock by the two types of gear presents management with the problems of determining the effect of various combinations of fishing effort by the two gears on both yield per recruit to the two gears and recruitment to the stocks. In order to make these determinations, it is necessary to estimate 1 ) availability of the stock at each age to each of the two gears [The available portion of the stock is subject to both other mortal- ' Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, La Jolla, Calif; present ad- dress: Southwest Fisheries Center Tiburon Laboratory, NMFS, NOAA, 3150 Paradise Drive, Tiburon, CA 94920. ^Southwest Fisheries Center La Jolla Laboratory, NMFS, NOAA, P.O. Box 271, La Jolla, CA 92038. Manuscript accepted Mav 1978. FISHERY BULLETIN: VOL. 76. NO. 4. 1979. ity (any mortality not caused by gear of concern) and fishing mortality caused by the gear of con- cern. The unavailable portion of the stock is sub- ject only to other mortality.], 2) fishing mortality of the available portion of the stock caused by each gear, 3) natural mortality, 4) growth, 5) fecundity, and 6) the relationship between egg production and recruitment. The aim of this paper is to examine the interac- tions between longline and surface fisheries for yellowfin tuna and to determine the impact such interactions may have on the assumptions often made in assessment of yellowfin tuna fisheries and thus on the assessment calculations themselves. The paper is divided into three major sections. The first section examines the relationship between availability of the stock(s) of yellowfin tuna to surface and longline fishing and yield per recruit to the two gears. This is an important, and to our knowledge unexamined, aspect of all tuna fish- eries exploited by both types of gear; the subse- quent sections examine two asepcts of the biology of tuna that can affect the catch by each type of gear. The second section examines the effect of age specific sex ratios of yellowfin tuna on yield per recruit to the two types of gear and on egg produc- tion. The third section examines the effect of 807 FISHERY BULLETIN: VOL. 76, NO. 4 random migration or dispersal and location of re- cruitment of yellowfin tuna on estimates of mor- tality and yield per recruit to each gear. We have restricted our analysis to yellowfin tuna but be- lieve that the concepts that we develop apply to the other species as well. MATERIALS AND METHODS While stocks of yellowfin tuna are subjects of important fisheries in all tropical oceans, infor- mation on vital parameters is sketchy and nonuniform. For example, tagging information available in the Pacific is lacking for the Atlantic stocks. On the other hand, regulation of the Pacific fishery makes interpretation of the catch informa- tion more difficult. Hence it is necessary to pick and choose from the available information that which is most relevant to the problems at hand. Although the parameters are likely to differ for fish from different oceans, if not fish from different areas of the same ocean, few studies have conclu- sively demonstrated that such differences exist. In addition, several (e.g., Lenarz et al. 1974) have found that conclusions from studies such as de- scribed in this paper are often insensitive to the likely range of values of parameters such as natural mortality, fishing mortality, and growth. In the first and second sections, we have used data primarily from the eastern Atlantic because his- torically catches have been more equally shared by longline and surface fisheries than in the east- ern Pacific; in the third section we have modelled the eastern Pacific since information on migration patterns is more extensive. In both instances, the results are intended to be general rather than specific. Data extracted from one area and used in another is thought to be the best available and the question of real differences is left for further inves- tigation. With a noted exception, the growth equation L = 194.8 X (1 - e-0 42u 0.67)) estimated by Le Guen and Sakagawa ( 1973) and length-weight equation W = 0.0000214L2^^36 estimated by Lenarz (1974) are used for yellowfin tuna where L is fork length in centimeters, t is age in years, and W is weight in kilograms. Unless otherwise stated, we assumed that the annual instantaneous coefficient of natural mortality (M) is 0.8 (Hennemuth 1961). We estimated age-specific fecundity from two indi- ces derived by Hayasi et al. ( 1972) (Table 1). Their index I was obtained from longline data and their index II was obtained from surface data. The Table 1. — Indices of fecundity of yellowfin tuna as interpolated from Hayasi et al. ( 1972), for fish caught in the Pacific calculated by multiplying average ova counts by percentage of mature female fish for each age and then dividing each product by the product calculated for age 3 fish. Midpoint of size interval Fecundity Fecundity (cm) index 1 Index II 80 0.04 0.07 85 0.04 0.14 90 0.05 0.21 95 0.08 0.27 100 0.15 0.36 105 0.23 0.42 110 0.33 0.51 115 0.42 0.61 120 0.55 0.70 125 0.70 0.81 130 0.88 0.92 135 1.12 1.04 140 1.40 1.15 145 1.80 1.26 150 2.30 1.37 155 2.77 1.50 160 3.20 1.62 165 3.57 1.76 170 4.05 1.91 175 4.42 2.06 180 4.82 2.23 5.01 2.43 fecundity indices were calculated by Hayasi et al. ( 1972) for fish caught in the Pacific by multiplying mean ova counts by percentage of mature female fish for each age and then dividing each product by the product calculated for age 3 fish. For much of our work, we used estimates of the 1967-71 aver- age size (age) composition of the Atlantic yellowfin tuna fishery made by Lenarz et al. (1974) (Table 2). Use of length-age key assumes that length and age are equivalent. Sex composition shown in Table 2 is based on data from the Pacific. Estimates of the size- (age-) specific instantane- ous coefficient of fishing mortality (F,) on an an- nual basis were made using the Gulland (1965) and Murphy (1965) method. The computer pro- gi-am COHORT, written by W. W. Fox, Jr., of the Southwest Fisheries Center, was used to obtain estimates of F, for each 5-cm size interval, begin- ning at 32.5 cm. The estimation procedure was initiated with a trial value of F, for the largest size interval (Input F). Estimates of F, were obtained from the average 1967-71 catch composition data (Table 2) as was done by Lenarz et al. (1974). When feasible it is more desirable to estimate F, from individual cohorts. This was not done because of the small number of years in the data series and belief that estimates from the average composition would adequately reflect conditions of the fishery. In a latter study, Fonteneau and Lenarz (1974) esti- mated F; for individual cohorts from a longer time 808 LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES Table 2. — Composite catch in numbers of yellowfin tuna by gear, sex, and size. Length composition by gear is based on data from Lenarz et al. ( 1974) on the Atlantic fishery. Sex composition is based on data from the Pacific (Murphy and Shomura 1972). Age at IVlidpoint beginning of interval of size interval (cm) IVIale Female (yr) Surface Longline Total Surlace Longline Total 1 0579 35 1.179 0 1,179 1,179 0 1,179 1 1325 40 14.528 0 14,528 14,528 0 14,528 1 2039 45 61.563 0 61,563 61,563 0 61.563 1 2888 50 186.611 4 186,615 186.611 4 186.615 1.3710 55 237.622 11 237,633 237,622 11 237,633 1 4562 60 210.711 226 210,937 210,711 226 210,937 1.5445 65 121.824 324 122,148 121,824 324 122.148 1 6363 70 137.389 1,076 138,465 137,389 1,076 138,465 1.7317 75 102.046 2,718 104.764 102.046 2.718 104,764 1 8310 80 90,710 2,847 93,557 90.710 3.847 93,557 1 9348 85 67,060 6,013 73,073 67.060 6,013 73,073 20432 90 52.541 6,525 59,066 52,541 6,525 59,066 2 1568 95 51,366 5,833 57,199 51,366 5,833 57,199 22761 100 56.714 7,537 64.251 56,714 7,537 64,251 24017 105 52.752 17.036 69.788 52.752 17,036 69.788 25343 110 51,497 20.105 71,602 51.497 20,105 71.602 26748 115 35,981 22,017 57,998 35,981 22,017 57.998 28240 120 26.167 21,430 47,597 26.167 21,480 47,597 2 9832 125 30,779 28,679 59,458 30.779 28,679 59,458 3 1538 130 26,001 29,272 55,273 26,001 29,272 55,273 33376 135 21.975 22,345 44.320 21,975 22.345 44,320 35368 140 16.749 26.035 42,784 16,749 26.035 42,784 37542 145 26.919 38,782 65,701 11,661 16.800 28,461 39935 150 31,942 36,099 68.041 8,450 9,549 17.999 4 2595 155 24,727 33.933 58.665 3,767 5,170 8,937 45590 160 18,701 22,644 41.345 1,524 1.845 3,369 49017 165 14,497 13.140 27.637 573 519 1,092 53021 170 5.621 6,162 11,783 94 103 197 57838 175 3,703 240 3,943 21 1 22 6 3883 180 1,836 55 1,891 3 0 3 Total 1,781,711 371.093 2,152,804 1.679.858 254.020 1,933.878 span and obtained results similar to Lenarz et al. (1974). The computer program MGEAR, written by W. H. Lenarz, was used to obtain estimates of yield per recruit using the Ricker ( 1958) yield equation. A description and listing of MGEAR is available from its author. The program was slightly mod- ified to calculate indices of egg production using the following equation Ef,.,, = 0.5 (t2 - t,)N,^ {FI>^ + FI,fi -'^'. ^ '^'i'"^-''') where E,^,,^ = index of egg production between age ^1 and t.2, FI,^ = index of fecundity for age ^. , A^,_ - number of females in population of age ^, F,^ = coefficient of instantaneous fish- ing mortality between age t^ and age t2, and Mi^ = coefficient of instantaneous natural mortality between age ^i and age ^2- For this equation it is assumed that the estimates of FI are proportional to egg production per female, which is assumed to be continuous, and that the rate of egg production is linear over the interval (^i, ^2*- A computer program MIGR was written by J. R. Zweifel to perform the calculations used for the third section of this paper. Since new methodology is developed, a description of the calculations will be given in that section. AVAILABILITY OF THE STOCK(S) OF ATLANTIC YELLOWFIN TUNA TO SURFACE AND LONGLINE GEAR In previous works on yield per recruit, yellowfin tuna of all ages in either the entire Atlantic (e.g., Hayasi and Kikawa 1970; Wise 1972; Hayasi et al. 1972; Lenarz et al. 1974), or in the eastern Atlan- tic (e.g., Fonteneau and Lenarz 1974) were as- sumed to be equally available to both longline and surface gear. However, since the surface fishery for yellowfin tuna occurs very close to the west African coast (Fox and Lenarz 1973) while the longline fishery for yellowfin tuna is distributed throughout the tropical Atlantic, it seems possible that the longline fishery is exploiting some fish that are not available to the surface fishery. It is 809 FISHERY BULLETIN; VOL. 76. NO. 4 also possible that some stock(s) which are avail- able to surface fishing are never available to the longline fishery. Since significant tagging efforts have begun only recently in the Atlantic and the results of these studies have not been published, data are not available to evaluate the availability of yellowfin tuna to both gears. However, there is evidence from the Pacific that yellowfin tuna are not equally available to long- line and surface gears. With the permission of W. H. Bayliff of the Inter-American Tropical Tuna Commission (lATTC), we examined yellowfin tuna tag return data from the eastern Pacific dur- ing 1963-66 in an attempt to evaluate the avail- ability offish to both gears in that area. We tabu- lated the number of tag returns for fish larger than 100 cm at return by 10-cm size groups (Table 3). All of the fish hadbeenatliberty for at least 10 mo. Although all of the tagged fish were measured when released, not all were measured when recov- ered. Bayliff recommended the relationship 167 (1 - e -0,fii/-0.833i - Again at the suggestion of Bayliff, we estimated the expected return of tags from longline-caught fish when all fish are equally available to both gears. Assuming tag recoveries were independent of each other, recovered tags were reported at the same rate by both components of the fishery, and tagged fish were equally available to both gears: then the expected returns of tagged fish of size / by gear / in year k is given by E(R„,) =R,.,N„,JN,^k (1) '1, when size is between 101 and 110 cm ,6, when size is between 151 and 160 . _ fl, when fish are caught by surface gear l2, when fish are caught by longline gear '1, when fish are caught in 1963 , 2, when fish are caught in 1964 3, when fish are caught in 1965 ,4, when fish are caught in 1966 estimated by Davidoff ( 1963) for growth of yellow- fin tuna in the eastern Pacific as the best equation to estimate the size of unmeasured fish. All of the returns were surface-caught fish, even though longliners captured a considerable number of yel- lowfin tuna in the eastern Pacific (east of long. 130"^ W) (Kume and Joseph 1969). In fact for many of the 10-cm size groups, the longliners caught more yellowfin tuna than the surface gear operators (Table 4). Table 3. — Number of returns of tagged yellowfin tuna from the eastern Pacific Ocean by size interval and year iW. H. Bayliff, pars, commun.). Size interval (cm) 1963 1964 1965 1966 101-110 111-120 121-130 131-140 141-150 151-160 16 7 0 0 0 0 where /?,;/; = number of returns and N ,ji; = number of fish caught. A dot in the position of a subscript signifies sum- mation of the variable over the subscript, e.g., X, .^ 2 = 1 X, ./ = 1 ,jh ■ Forty fish were returned by the surface gear during 1963-66 (Table 3). Using the statistics of Tables 3 and 4 and the three assumptions, a return of 5.4 of these tags would have been expected from the longline fishery and 34.6 from the sur- face fishery. The chi-square value, corrected for discontinuity, for the observed and expected re- turns (Equation 1) is 5.13, with probability slightly less than 0.025. The power of the test of the hypothesis of independence, equal reporting rate, and equal availability was reduced because we combined the year and size strata to avoid Table 4. — Catch of yellowfin tuna from the eastern Pacific Ocean (east of long. 130°W) in hundreds of fish by size and gear iKume and Joseph 1969). Size interval (cm) 1963 1964 1965 1966 Surface gear Longline gear Surface gear Longline gear Surface gear Longline gear Surface gear Longline gear 101-110 111-120 121-130 131-140 141-150 151-160 653 473 508 237 240 212 336 455 390 751 541 144 4,082 2.245 720 448 320 102 173 465 1.078 804 469 104 3.386 2.211 1.895 905 498 194 30 93 444 758 466 205 2.926 2.044 1.312 718 536 204 54 116 304 515 575 200 810 LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES strata with low expected values. The probability under Equation ( 1) of a returned tag being from a surface-caught fish (P,ik) is P.u = N.u/Nik (2) The exact probability of all returns during the 1963-66 period being from surface-caught fish, given the distribution of returns among year and size categories, is p.i. = n n (Pnk)"'" 1=1 * = 1 (3) Our estimate of P j is 0.00152, which is very low and indicates that Equation (1) does not hold. Thus we may conclude that 1) tag returns are not independent (e.g., fish that were captured from a school and tagged may remain in the same school until recaptured), and/or 2) longline recoveries are reported at lower rates than surface recoveries, and/or 3) the fish were not equally available to both gears. Since all fish were at liberty for more than 10 mo before being recovered, the assump- tion of tag returns being independent seems likely to be valid. The independence of tag returns would seem to be a desirable subject for further research since the assumption is so often made in analyses of tag returns. A considerable number of southern bluefin tuna have been recovered and returned by longliners (Shingu 1970), indicating longline fishermen do cooperate in tagging programs. Dur- ing the period of the study, the surface fishery was only beginning to move offshore (Calkins and Chatwin 1971), while the longline fishery was dis- tributed throughout the area (Kume and Joseph 1969). Also, the fish that were released were caught by surface gear, tagged, and released in nearshore areas. Thus, tagged fish were probably more representative of fish exploited by the sur- face fishery than those that were exploited by the longline fishery, if two groups offish existed. Thus it seems plausible that the tagged fish were not equally available to longline and surface gears. This is further evidence of unequal availability of yellowfin tuna to the two gears in the Pacific. Previously, Hisada (1973) showed that yellowfin tuna caught near the surface using handlines were of the same size as those caught by longliners at the same time and in the same area of the western Pacific. However, the surface-caught fish tended to be more sexually mature except in areas in which the 26°C isotherm occurred at depths fished by longliners. He attributed this phenome- non to a preference for warmer waters by sexually mature fish and noted that larvae of yellowfin tuna tend to be found at water temperatures ex- ceeding 26°C. Thus, some yellowfin tuna evidently behave in a fashion that makes them available to surface fishing but not to longline fishing. Further evidence along these lines is provided by Shingu and Tomlinson (Patrick K. Tomlinson, Inter- American Tropical Tuna Commission, La Jolla, Calif. Pers. commun., 1974) who found that the length-weight relationship estimated by Lenarz (1974) for surface-caught yellowfin tuna in the Atlantic was more representative of the longline catch in the eastern Pacific than was the relation- ship estimated by Chatwin (1959) for surface- caught yellowfin tuna in the eastern Pacific. With the above in mind, we considered three hypothetical stock structures for the Atlantic yel- lowfin tuna fishery: 1) the same stock(s) are equally available to both gears, 2) half of the catch of the longline fishery comes from stock(s) not available to the surface fishery, and 3) the entire catch of the longline fishery comes from stock(s) not available to the surface fishery. The effects of the three hypotheses on estimates of fishing mor- tality and yield per recruit to the gear were examined. Using the data in Table 2, we estimated the vector F of size-specific instantaneous mortality rates F, under the three hypotheses which are identified by the proportion = 0.5, the surface catch plus 50% of the longline catch was used and for (t> - 0.0 only the surface catch was used for estimating F. When (/> - 0, an additional F vector was estimated for a longline fishery operating without the presence of a surface fishery by using only the longline catch. The F vectors were then used to calculate yield per recruit to the two gears. Estimation of a vector of size-specific F requires an estimate of natural mortality and size-specific F for one size category. In all instances, we chose to use an estimate of size-specific F for the fish > 177.5 cm. This estimate will be referred to as Input F. The final value of size-specific F was set at 0.2 following Lenarz et al. (1974). The estimates (Figure 1) indicate that values of F for large fish are directly related to the portion of the longline catch that comes from the stock(s) exploited by the 811 FISHERY BULLEITN: VOL. 76. NO. 4 80 100 120 140 FORK LENGTH (cm) Figure l. — Estimates of size- specific fishing mortality of Atlan- tic yellowfin tuna as a function of porportion of catch ((/>) by longline fishery that comes from stock(s) exploited by surface fishery. surface fishery. The relative values of yield per recruit within a hypothesis are not significantly affected by the portion of the longline catch that comes from the stock(s) exploited by the surface fishery (Figure 2). Therefore, the three hypothet- ical stock structures do not seem to have much bearing on decisions concerning minimum size regulations. Estimates of yield per recruit were also plotted as functions of fishing effort (mortality), size at recruitment, and portion of longline catch that comes from stock(s) exploited by the surface fishery. Again the relative values of the results are not significantly influenced by the stock structure (Figure 3a, b). We note that Figure 3 is in agree- ment with the conclusion of Fox and Lenarz (1974), ". . . that the Atlantic yellowfin fishery is approaching or has obtained a plateau where sub- stantially increased sustainable average yield of yellowfin tuna will not be obtained by increasing fishing effort without some concomitant change in the constitution of the fishery. . . ." They used the production model approach under the alternative assumptions that either the longline or surface gear exploits the same or separate stock(s). The effect of the surface fishery on the longline fishery was examined by estimating yield per re- 5 4r 9 i 5.0 q: o 111 a: 4.6 4.2 3.8 J L J L J. J L 32.5 52.5 72.5 92.5 112.5 132.5 FORK LENGTH AT RECRUITMENT (cm ) Figure 2. — Yield per recruit ( kilograms) ofAtlantic yellowfin to surface and longline gear as a function of size at recruitment and proportion of catch () by the longline fishery that comes from stock! s ) exploited by surface fishery . The vector of fishing mortal- ity is equal to the value at the time of study. cruit to the longline fishery in the presence and in the absence of a surface fishery (Figure 4). The results suggest that if the two gears exploit the same stock(s), the surface fishery reduces the po- tential yield per recruit to the longline fishery by about twofold at the position of the fishery during the study period (i.e., multiplier of effort = 1) and about fivefold for a threefold increase in effort. The same procedure was used to examine the effect of the longline fishery on the surface fishery (Figure 5). The results indicate that at the level of fishing effort at the time of study, the yield per recruit to the surface fishery would be increased by 25% if the longline fishery ceased. Although the presence of each fishery reduces the yield per recruit of the other, the yield per 0 04 08 12 16 20 2'» MULTIPLIER OF EFFORT 04 08 12 16 20 24 28 MULTIPLIER OF EFFORT Figure 3. — Yield per recruit of Atlantic yellowfin tuna as a function of fishing effort and proportion of catch i4>) by longline fishery that comes from stock(s) exploited by surface fishery: (a) size at recruitment is 32.5 cm, (b) size at recruitment is 77 cm. 812 LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES tr. o u o 0.8 1.2 16 2.0 MULTIPLIER OF EFFORT Figure 4. — Estimates of yield per recruit of Atlantic yellowfin tuna to the longline fishery as a function of effort and presence ( 4> = 1.0) or absence (({> = 0.0) of a surface fishery. o llJ q: o ^^^— ^ •— • — • • Jj^ • — • o — o 0 = 0.0 - Jd o — 0 f \ \ 1 0 = LO 1 1 1 1 1 J 1 0.4 0.8 1.2 1.6 2.0 2.4 MULTIPLIER OF EFFORT 2.8 Figure 5. — Estimates of yield per recruit of Atlantic yellowfin tuna to the surface fishery as a function of effort and presence (<^ = 1.0) or absence {^ = 0.0) of longline fishery. recruit of the combined fisheries is higher than the yield per recruit of either fishery alone. The re- sults suggest that if a catch quota system is im- posed on the Atlantic yellowfin tuna fishery, all components should be included unless it is shown that different stock(s) are being exploited by the gear. The above results (Figures 4, 5) suggest that a stock of yellowfin tuna will produce a potentially higher yield per recruit to a longline fishery than to a surface fishery, if the fish are equally avail- able to the two gears. However, until the question of availability is settled, it is not possible to predict the potential production to the two gears. We point out here that gear-specific availability is not well known for any tuna fishery and would be difficult to determine. Thus, we are faced with the prospect of probably being forced to determine empirically the production potential for each gear in each fishery. After a fishery is established, an analysis of the type conducted on the Atlantic yellowfin tuna fishery could be used to examine the effects of availability to the two gear types, and a tagging study could be designed to provide the required answers. EFFECTS OF AGE-SPECIFIC SEX RATIOS OF ATLANTIC YELLOWFIN TUNA ON YIELD PER RECRUIT TO THE TWO TYPES OF GEAR AND STOCK FECUNDITY While a number of authors have noted that the ratio of females to males appears to be less than 1:1 for catches of larger tunas, none to our knowledge has incorporated these observations into calcula- tions of yield per recruit or stock fecundity. Beardsley ( 1971) reported that the ratio of female to male Atlantic longline-caught albacore was 233:365 during the December 1969-September 1970 period. Males increasingly dominated at sizes >100 cm. Females slightly outnumbered males between 92 and 100 cm. One explanation for the catch curves estimated by Beardsley is a 1:1 sex ratio at small sizes, a slightly slower growth for females for fish >90 cm, and beyond 100 cm, either a higher rate of natural mortality for females or a change in behavior that makes females less available than males to longline fishing. Other explanations exist, e.g., a combina- tion of low sex ratio and slow growth of females throughout their life. Sakamoto ( 1969) noted for Atlantic bigeye tuna, ". . . males predominated in areas of higher water temperature. Proportion of females increase as the water temperature gets lower." His data indicate that as size increases the proportion of females decreases and females may grow slower than males in waters between lat. 30° to 50°N, but not in equatorial waters. Data pre- sented by Kikawa (1964) indicate that southern bluefin tuna >150 cm are predominantly males, while females often outnumber males at smaller sizes. Thus, female southern bluefin tuna may grow more slowly than males. Since there is considerable evidence for age- specific changes in the sex ratio of tunas, we be- lieve that the effects of such changes on estimates of yield per recruit to each gear type and fecundity should be investigated. We have assumed sex ratios to be the same as with Pacific yellowfin tuna because no extensive studies of age-specific sex ratios for Atlantic yellowfin tuna have been pub- lished. We used results from a study by Murphy and Shomura (1972), who found that beyond 140 cm male yellowfin tuna greatly outnumbered females (Figure 6). The data in Figui'e 6 do not show a large excess of females in any size interval and thus no evidence of sex-specific growth is exhibited. Using their data and the age-length 813 FISHERY BULLETIN: VOL. 76, NO. 4 60 70 80 90 100 110 120 130 140 150 160 170 LENGTH (cm) Figure 6. — Length distribution by sex of longline-caught yel- lowfin tuna in the central Pacific Ocean (Murphy and Shomura 1972). relationship of LeGuen and Sakagawa (1973), we estimated that beyond 140 cm ]nR = 6.74 - 1.96^ (4) where R = ratio of females to males t = age in years. One interpretation of the above result (assum- ing that males have a coefficient of instantaneous natural mortality of 0.8 on an annual basis as do all fish <145 cm) is that female yellowfin tuna >140 cm have a coefficient of apparent natural mortality of 2.76. Assuming that the results of Murphy and Shomura apply to the Atlantic and that all yellowfin tuna are equally available to both gears, we separated the catch of yellowfin tuna into males and females using Equation (4) and Table 2, and estimated F for the males using Input F values of 0.2 and 0.8 for fish >177.5 cm (Lenarz et al. 1974). An alternative method would be to use the same Input F for the three hypotheses at the smallest size interval. This was attempted and resulted in either estimates of F, which, based on the results of other studies, appeared to be too low under the 1:1 hypothesis or too high under the other hypotheses. The estimates of size- specific F are similar except for very large yel- lowfin tuna (Figure 7). Since the deviations in sex ratio from 1:1 occurs only at large sizes, we used both sets of estimates of F. o 35 55 75 95 115 SIZE (cm) T — r 135 155 175 HIGH M Female BEH Female 1 — f~— I — r SIZE (cm) Figure 7. — Estimates of size and sex specific coefficient of in- stEintEineous fishing mortahty on annual basis (F) for Atlantic yellowfin tuna for 1:1, BEH and HIGH M hypotheses (see text): (a) low Input F, B) high Input F. For females, three hypotheses were examined for estimating F: 1 ) the observed differences in sex ratios are artifacts, and consequently females have the same values of F and M as males (de- noted 1:1); 2) females >140 cm have a higher natural mortality rate than males but are exploited at the same rate as males for all sizes (denoted as HIGH M); and 3) females have the same natural mortality rate as males but become less subject to fishing mortality beyond 140 cm (denoted as BEH for behavior changes). Under the BEH hypothesis, F, for females >140 cm is equal to the ratio of the catch of females to the catch of males times F, estimated for males. The alterna- tive hypotheses considerably affected the esti- mates of size-specific F (Figure 7). In the following analyses, we found that the BEH and HIGH M hypotheses produce similar results. To save space, we refer to only the one hypothesis that produced results which showed the greatest difference from the 1:1 hypothesis. Also, when not specifically indicated, size of re- cruitment and effort are assumed to be those at the time of the study, i.e., 1967-71 where the multi- plier of effort is equal to unity. Estimates of yield per recruit as a function of fishing effort are shown in Figure 8. The choice of Input F has little effect on the relative values of 814 LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES ENTIRE FISHERY ENTIRE FISHERY HIGH M 3 o UJ q: Q -BEH -BEH -1 — I — r 4 8 12 ~ 1 I I I I I I r 16 20 24 28 32 1 3 6 MULTIPLIER OF EFFORT Figure 8. — Estimates of yield per recruit of Atlantic yellowfin tuna at size of recruitment at time of the study as a function of fishing effort and sex hypothesis: (a) high Input F, (b) low In- put F. yield per recruit. Yield per recruit is closer to the maximum under high Input F than low Input F. The curves are considerably more dome-shaped when a 1:1 sex ratio is assumed than under the other two hypotheses. Under high Input Fand the 1:1 hypothesis only a 3^c increase in yield per recruit could be obtained by increasing fishing effort. Under the BEH hypothesis, a 20% increase in yield per recruit could be obtained by doubling the effort. Estimates of yield per recruit as a function of size at recruitment are shown in Figure 9. Again the choice of Input F has little effect on the relative values of yield per recruit. A slightly greater de- pendence of yield per recruit on minimum size is obtained when the high Input F is used. Under high Input F, and the 1:1 hypothesis a 107c in- crease in yield per recruit could be achieved by increasing size at recruitment. Under the BEH hypothesis, only a d'^c increase would occur. Eumetric fishing occurs when size at recruitment is raised from the current 32.5 to 82.5 cm under the 1:1 hypothesis and 72.5 cm under the BEH hypothesis. Estimates of yield per recruit as a function of fishing effort were also calculated for each gear H M 1 — I — 1 — \ r J2 5 52 5 72 5 92 5 112 5 132 5 32 5 52 5 72 5 92 5 112 5 132 5 MINIMUM SIZE (cm) Figure 9. — Estimates of yield per recruit of Atlantic yellowfin tuna at level of fishing effort at the time of the study for 1:1, BEH and HIGH M hypotheses as a function of size at recruitment: (a) high Input F, (b) low Input F. (Figure 10). The results show that the curves are more dome-shaped for the longline fishery than for the surface fishery under all three hypotheses. Furthermore, the longline fishery is more sensi- tive to fishing effort under the 1:1 hypothesis than under the other two. The curves for the surface fishery are dome shaped under the 1:1 hypothesis, but appear to approach an asymptote under the other two. We also estimated yield per recruit for each gear when the other gear is not exploiting the stock (Figure 11). A comparison of Figures 10 and 11 reveals that yield per recruit to the longline fishery would increase by about 115*^ if surface fishing were eliminated under high Input F and the 1:1 hypothesis and 76*?^ under high Input F and the BEH hypothesis. Yield per recruit to the surface fishery would increase by about 30% if the longline fishery were eliminated under high Input F and the 1:1 hypothesis and 229c under the BEH hypothesis. Thus, the nature of age-specific sex ratio has a greater effect on that of the longline fishery than on the relative success of the surface fishery. The curves for a longline fishery in the presence of a surface fishery are dome-shaped (Figure 10), while the curves in the absence of a surface fishery are not (Figure 11). This again points out the importance of not treating the two fisheries as separate entities unless it is shown that they exploit separate stocks. Stock fecundity (egg production per recruit) relative to an unfished stock was estimated as a function of fishing effort. Stock fecundity was con- siderably affected by the choice of fecundity index 815 FISHERY BULLETIN: VOL. 76, NO. 4 4i- OC_l I I I I I I I I I I I — I — I — I 1 1 — I 0 04 08 12 16 20 24 2 8 3 2 36 MULTIPLIER OF EFFORT Figure 10. — Estimates of yield per recruit of Atlantic yellowfin tuna when both gear fish at size of recruitment at the time of the study as a function of sex ratio hypothesis, fishing effort, and gear: (a) surface gear with high Input F, (b) longHne gear with high Input F, (c) surface gear with low Input F, and (d) longline gear with low Input F. and sex ratio hypothesis but only slightly affected by the choice of Input F (Figure 12). At the level of fishing effort at the time of the study under high Input F and 1:1 hypotheses, the relative fecundity is 0.28 when the fecundity index I is used and 0.39 when fecundity index II is used. Under the HIGH M hypothesis, relative fecundity is 0.55 when fecundity index I is used and 0.61 when fecundity index II is used. Thus, at the level of fishing effort at the time of the study, the choice of fecundity has a 10 to 30% effect on estimates of relative fecun- dity, while the choice of sex ratio hypothesis has a 30 to 50% effect. The two choices, fecundity index and sex ratio hypothesis, also have considerable effect on relative fecundity when plotted as a func- tion of size at recruitment (Figure 13). The relationship between stock fecundity and recruitment has not been demonstrated for any tuna. As shown above, one of the difficulties in demonstrating such a relationship is obtaining a reasonably accurate estimate of stock fecundity. Even if stock fecundity could be accurately deter- mined, the recruitment process is likely to be so complex that much more research would be re- quired before a reliable predictor of recruitment could be developed. It is interesting to note that similar estimates of yield per recruit and relative fecundity are ob- tained under the HIGH M and BEH hypotheses. Thus it appears that research should be directed toward determining whether or not the 1:1 hypothesis or one of the other two are valid rather than distinguishing between the HIGH M and BEH hypotheses. This research should be a fairly simple matter. The choice of fecundity index is also of significance for estimating relative fecun- dity. The difference between the two indices is caused mainly by different maturity schedules (Hayasi et al. 1972). The surface-caught fish ap- peared to mature at an earlier age than longline- caught fish, and could be an artifact related to the phenomenon noted by Hisada ( 1973); i.e., mature fish tend to prefer warm water. It should also be a fairly simple matter to determine the cause of the difference between the two indices. SIMULATION MODEL OF PATTERNS OF DISPERSAL AND RECRUITMENT OF YELLOWFIN TUNA Factors that could cause groups of tuna to not be available to all components of a fishery include nonrandom movements, random movements but nonrandom distribution of fishing gear or effort, and recruitment that is nonrandom in a geo- graphical sense. Extensive tagging experiments have not pro- duced any clear-cut evidence of a definite migra- tion pattern for yellowfin tuna in the eastern Pacific. Bayliff and Rothschild (1974) recently found evidence for both random dispersal and di- rected movements. They were not able to remove the effects on their data of lack of fishing effort in some time-area strata and of the coastal boundary. The evidence for directed movements indicated that such movements were generally parallel to the coast, suggesting that the presence of the coast influenced their results. Fink and Bayliff (1970), in a synthesis of extensive tagging data, proposed that recruitment to the nearshore surface fishery is not random in a geographical sense, but tends to take place off Mexico and in the Panama Bight. 816 LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES 5- 0 1 1 4- BEH 3- /y/^ 2- 1/ 1— 1 - 0- "T r r r 1 1 — 1 — 1 — 1 — 1 — 1 — 1 — r — -I — I — I — 1 — \ 32 36 MULTIPLIER OF EFFORT Figure ll. — Estimates of yield per recruit at size at recruitment at the time of the study as a function of fishing effort, sex ratio h>'pothesis, and fishing gear when only one gear is fishing: (a) high InputF and surface gear, (b) high InputF and longline gear, (c) low Input F and surface gear, and (d) low Input F and longline gear. 3 < HIGH M MULTIPLIER OF EFFORT Figure 12. — Estimates of relative stock fecundity at size at recruitment at the time of the study as a function of fishing effort, fecundity index, and sex ratio hypothesis: (a) high Input F and fecundity index I, (b) high Input F and fecundity Index II, (c) low Input F and fecxmdity index I, and (d) low input F and fecundity index II. With the above results in mind, we developed a computer simulation model to examine the inter- relationships of: 1 ) patterns of movement offish; 2 ) patterns of recruitment (i.e. by area), and 3) fishing strategy for two gear types (surface and longline) fishing alone or together on the same population. The model is general in that it allows the user to 817 FISHERY BULLETIN: VOL 76, NO. 4 325 525 725 92 5 112 5 132 5 32 5 62 5 72 5 92 5 1125 1325 SIZE AT RECRUITMENT (cm) Figure 13. — Estimates of relative stock fecundity at level of fishing effort at the time of study as a function of size at recruit- ment, fecundity index, and sex ratio hypothesis: (a) high Input F and fecundity index I, (b) high Input F and fecundity index II, (c) low Input F and fecundity index I, and (d) low Input F and fecundity index II. specify the nature of movements, locations of re- cruitment, parameters of growth, and natural fishing mortality. We crudely represented the eastern Pacific Ocean with the grid of 5° square areas shown in Figure 14. The number offish of a specific age in each cell at time t is given by the vector N, = AS,N,_, (5) where A^, (112 x 1) has elements (n,), equal to the numberof fishincelU attimef, S;(112 x 112) is a diagonal matrix with elements (s,,); equal to the survival rate offish in cell ;' from time ^ - 1 to time t, A (112 X 112) is a probability transfer matrix with elements (Oy) equal to the probability of a fish incellj moving to cell J, and where A'^o^ 112 x l)has elements {n,)^ equal to the number of recruits in cell i. Five consecutive year classes are in the system at a time. For our work we specified A, the transfer ma- trix, by the assumption that for any cell the prob- abilities of fish remaining stationary and moving to each of eight adjacent cells is the same, i.e., 1/9. Any other transfer has zero probability. This gen- eral rule is modified as follows: 1) Probabilities of remaining stationary in cells adjacent to the shore are augmented by the sum of probabilities of those movements which would otherwise put fish on land and the probability of occurrence on land is zero. 818 1 2 3 4 5 COLUMN 6 7 8 9 10 // 12 13 14 / 1 2 3 4 5 6 7 8 9 10 II 12 13 14) 2 15 16 17 18 19 20 21 22 23 24 25 26 " *' \\\\\ 27 28; ^\~^~- 3 29 30 31 32 33 34 35 36 37 38 39 40 41 42; s\S\\N 5 4 O Q: 5 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 ;70; \\\\N 6 71 72 73 74 75 76 77 78 79 80 81 82 83 84 7 85 86 87 88 89 90 91 92 93 94 95 96 97 ^98^ 8 99 100 101 102 103 104 105 106 107 108 109 no III Figure 14. — Representation of eastern Pacific Ocean. Each cell represents a 5° square area. Hatched cells represent land. Col- umn 1 is western boundary and Column 14 is eastern boundary. Row 1 is northern boundary and row 8 is southern boundary. 2) Probabilities projecting beyond the northern and southern edges are similarly absorbed on the boundaries. 3) In cells of rows 2 and 7, probabilities of mov- ing toward rows 1 and 8 are decreased by half with the probability of remaining stationary increased by a like amount. This is an attempt to simulate a stock encountering increasingly marginal condi- tions as the northern and southern boundaries are approached. 4) Probabilities of remaining stationary on the western edge are augmented by the probability of returning from beyond the boundary in a single time interval. The remainder of the fish that move beyond the western boundary are lost to the sys- tem. The speed of dispersion is controlled both by A and the time interval. The time interval was 3 mo for this study. The combination of A as defined and time interval of 3 mo allows a fish to travel a maximum of 1,200 mi in a year. Only 1 out of 820 surviving fish that begin the year in the center of the grid travel 1,200 mi in a year. These relatively slow random movements seemed reasonable, based on the results shown in Bayliff and Rothschild (1974) and recent results of lATTC tagging studies (Inter- American Tropical Tuna Commission^). Two alternative recruitment models were examined. For the first, denoted as inshore re- ^Cited with permission of M. Clifford Peterson, Acting Direc- tor of the Inter- American Tropical Tuna Commission. From the Inter-American Tropical Tuna Commission Bi-monthly Report, March- April 1976:8-13. LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES cruitment, recruits are divided equally among the five cells 51, 52, 69, 83, and 84, which resemble the recruitment areas proposed by Fink and Bayliff (1970). For the other alternative, denoted as uni- form recruitment, recruits are divided equally among all cells except those on the boundaries or on land. Total annual recruitment is 100 fish. We assumed 1 ) that fish are 1 yr old when recruited, 2 ) growth proceeds according to the von Bertalanffy curve of LeGuen and Sakagawa (1973), and 3) the coefficient of instantaneous natural mortality is 0.8 on annual basis and is independent of time and location. Fish >6 yr old (175 cm) were removed from the system. Consequently, under constant conditions the fishery reaches equilibrium in 5 yr. The system was always run for 5 yr before an experiment was begun. We first examined the effects of sampling loca- tion, dispersal, and location of recruitment on age distribution and the resulting apparent rate of natural mortality obtained from unbiased sam- ples from an unfished population. Mortality was estimated with the standard linear regression model {\n Nt = In Nq - Mt) from the age distribu- tion offish in each cell. It is assumed that mortal- ity is constant after full recruitment, and that the modal age represents first age of full recruitment. The results reveal that M is usually overestimated as would be expected when fish emmigrate from a sampled area (Figure 15). Estimates of M tend to be relatively high near areas of spawning with inshore recruitment. In the case of uniform re- cruitment, estimates of M tend to be highest on the western boundary where fish are lost to the sys- tem. Modal age tends to increase in a westerly direction for inshore recruitment and stay rela- tively constant for uniform recruitment (Figure 15). The modal size of actual catches of surface- caught yellowfin tuna in the eastern Pacific in- creases in a westerly direction (Figure 16). Al- though the surface fishery probably does not take an unbiased sample of the size distribution of the population, the data are suggestive of reduced re- cruitment in the western areas. We simulated a 20-yr hypothetical yellowfin tuna fishery to examine interactions among a longline fishery, inshore surface fishery, ocean- wide surface fishery, and ocean-wide surface fishery that does not heavily exploit young fish as follows: 0--0- -0--0-- . MODAL AGE ROW ROW COLUMN COLUMN Figure 15. — Estimates of coefficient of instantaneous natural mortality on annual basis ( M) and modal age of yellowfin tuna by row and column: (A) inshore recruitment, and (B) uniform recruitment. 819 FISHERY BULLETIN: VOL. 76. NO. 4 S^ o J ■'4 ^ o i£ < - UJ f^ < ro q: io i K U> < in / 5-OD X 1 O " ^ * / UJ * z „ O cfl ^ >- OS U 3 c 1 (u 0 +j 0 0 c 0) c 0 *-> to 01 01 Tf J3 C8 CO B E 2 CO CO (1) u ca s- ^^ a< 01 ^ nt-i >H C) T3 OJ C n 0) CO .M JS .*-> « 0 u c 3 4-^ c CO 0 OJ fTl C be 1 ^ lo 0 -*-> m n 01 >> c 01 E 0 c 0 01 u 01 •E -M *J m r/l , r ca 3 .«-! 0) -Q k. C8 CS ■ — ■ 0> to ^ J= ■o t^ >> fT) C u T—t X C 3) r. 3 0 •s 1) m to i E 6 01 0 3 c 0) c 3 0) c •—1 00 (Tl u 0 c Tl H ca 3 4^ c (M to 0 < § Ih f-H ID < CO 0 .2 CO 0. E 0 CO 0 be o> N 0) m E 0 1 rn 0 H 0 01 1 y= ^ 1 :* 3 T CO .— t fc 01 w e 0) CO & Iw^ ffl V Uh 01 -a 820 LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGLINE AND SURFACE FISHERIES 1) For the first 5 yr only longliners fished and only in rows 5 to 8. 2) For the next 5 yr, this longline fishery was augmented with surface gear in all cells adjacent to the coast. 3) Next, exploitation by the surface gear was expanded to include all cells for 5 yr. 4) Finally, for the last 5 yr, age specific surface fishing mortality was reduced by 75% for fish <2.5 yr of age because much of the surface catch of yellowfin tuna in offshore areas of the eastern Pacific comes from schools associated with por- poise. Typically, porpoise schools contain few yel- lowfin tuna <2.5 yr of age (Calkins 1965). Steps 1,2, and 3 resem.ble the sequence of events in the eastern Atlantic fishery for yellowfin tuna. Yellowfin tuna first were exploited in a significant fashion by longliners in a 10° band along the equator, then a nearshore surface fishery became significant, and in recent years some exploitation by surface gear in offshore areas has occurred. To our knowledge, step 4 has not occurred in the At- lantic. Age-specific fishing mortality rates similar to those by surface gears estimated by Lenarz et al. ( 1974) for the Atlantic yellowfin tuna fishery were used (Table 5). The Ricker yield equation was used to calculate yield for each time-area stratum. Total yields per recruit were calculated and are shown in Figure 17. Yields per recruit are quite similar for both recruitment models except near shore, where yield per recruit was considerably higher for the inshore recruitment model than for the uniform recruitment model. The difference in yield per recruit betweeti the two models decreases slightly as time increases. Yield per recruit closely approached equilibrium yield within 3 yr after a change was made in the fishery. Total equilibrium yield per recruit with an inshore surface fishery Table 5. — Estimates of age-specific F on an annual basis used as baseline for simulation. See text for modifications of mortality rates during simulation. Age Surface gear with (yr) Longline gear Surface gear reduced F 1.0 0.00 0.30 0.08 1.5 0.00 0.30 0.08 2.0 0.05 0.22 0.06 2.5 0.15 0.20 0.20 3.0 0.25 0.18 0.18 3.5 0.35 0.30 0.30 4.0 0.45 0.35 0.35 4.5 0.40 0.42 0.42 5.0 0.40 0.27 0.27 5.5 0.20 0.20 0.20 6.0 0.05 0.15 0.15 6 00- \ 400- J^ 2 00- :v_^:rH 1 1 1 1 1 1 1 1 1 1 Uniform RecruitmenI Inshore Recruitment 04- 0 200- FlGURE 17. — Yield per recruit of hypothetical yellowfin tuna fishery: (a) total, (b) longliners in all areas, (c) surface gear in all areas, (d) longliners in cells 71 and 85, (e) surface gear in cells 71 and 85, (f) longliners in cells 69, 84, and 97, and (g) surface gear in cells 69, 84, and 97. and longline fishery was about 179c higher than with a longline fishery alone, 54% higher with a uniform surface fishery than with only a longline and inshore surface fishery, and increased by 9% when F for small fish was reduced by 75% . Under the assumption that the catchability coefficient is independent of area, the surface fishery increased its equilibrium yield per recruit about fourfold by increasing its effort about 12-fold when it ex- panded into offshore waters. The same action de- creased yield per recruit to the longliners by about 55%r. We next examined the potential yield per re- cruit to longliners in rows 5, 6, 7, and 8 by starting a longline fishery with the age-specific F vector multiplied by the scalar 0.3 and then multiplying by 1.3 each year afterward. Yield per recruit ap- pears to approach an asymptote of about 6 kg for inshore recruitment and 5 kg for uniform recruit- ment (Figure 18). The reduction in catch per re- cruit per effort by fishing is not significantly af- fected by choice of recruitment model. Even though catch per recruit per effort at high levels of effort was only about 20% of that at the beginning of exploitation, overfishing in a yield-per-recruit sense did not occur. Average size of fish in the catch was not significantly affected by the re- cruitment model, and decreased from about 50 to 30 kg with increased fishing effort (Figure 18). A simulation for an inshore surface fishery indi- cated an asymptotic production curve with a 821 FISHERY BULLEITN: VOL. 76, NO. 4 c 2 3 4 5 6 7 MULTIPLIER OF EFFORT 2 3 4 5 6 MULTIPLIER OF EFFORT Figure 18. — Yield per recruit, shield per recruit per effort, and average size of catch for hypothetical longline fishery: (a) total yield per recruit, (b) total yield per recruit per effort, (c) yield per recruit per effort in cells 7 1 and 85, (d) yield per recruit per effort in cells 69, 84, and 97, and (e) average size in all squares. maximum yield per recruit of about 1.4 kg for uniform recruitment and 2.2 kg for inshore re- cruitment (Figure 19). Catch per recruit per effort was reduced by about 75% under both alterna- tives. The ratio of maximum yield per recruit for a longline fishery to an inshore surface fishery was about 2.7 for inshore recruitment and 3.4 for uni- form recruitment. Average size offish in the catch was about 2 kg higher for uniform recruitment than for inshore recruitment and decreased from 16 or 18 kg to 8 or 11 kg with increased fashing effort (Figure 19). Simulation of a uniform surface fishery revealed that choice of recruitment model had an insig- nificant effect on yield per recruit, catch per re- cruit per effort, and average size of catch, except that catch per recruit per effort in the nearshore area was relatively high for inshore recruitment (Figure 20). A 75% reduction in F for fish <2.5 yr old had considerable effect on the results. Maxi- mum yield increased from about 5. 1 to 6.9 kg when F was reduced. Both yield curves are dome- shaped. Catch per recruit per effort became rela- tively higher at high levels of effort when F was reduced. As expected, average size was consider- ably higher for reduced F. With inshore recruitment, maximum yield per recruit changes from about 2.2 kg for an inshore fishery (Figure 19) to about 5.1 kg for a uniform „ 4 t 3 3 o K 2 >- o Z3 O 10 08 06 04 02 UNIFORM RECRUITMENT INSHORE RECRUITMENT c rT 30 j£ •—^ UJ M 20 - CO ^i^ZT — — UJ in ^ '" — " °~-— m- o oc X — " LlJ > < 1 1 1 1 1 — 1 1 1 1 01 23456789 MULTIPLIER OF EFFORT Figure 19. — Yield per recruit, yield per recruit per effort, and average size of catch for hypothetical inshore surface fishery: (a) yield per recruit, (b) yield per recruit per effort, and (c) average size of fish in catch. 822 LENARZ and ZWEIFEL: INTERACTION BETWEEN LONGUNE AND SURFACE FISHERIES 0.5- CE O ^^ 0.4 H 5 0.3- q: o UJ . (2b) The Gompertz function S^ = S.exp[G(l-exp[-^(a-c.)])] (3a) becomes S^. = S.[expG][S,J(S.expG)]^''P[-^<^°>l (3b) A linear function of size upon age S = b + ka (4a) a is written S = S +k{Aa) . (4b) r rn Definitions of symbols employed above are: Sa = size at age a, Sr = size at recapture, S,„ = size when marked. Si = size of the smallest animal in the data, a, = age of the smallest animal in the data, Table l. — Brown shrimp mark- recapture experiments, northern Gulf of Mexico. Length range (mm) Length range (mm) Number Release area Release date of released shnmp of recovered shrimp recovered Galveston Estuary May 1967 66-175 71-124 13 50 mi east of Galveston, Tex. June 1967 83-147 86-178 301 60 mi southeast of Freeport, Tex Sept. 1967 122-181 124-196 40 Biloxi Bay, Miss. May 1 968 90-122 90-181 4,218 40 mi southeast ot Freeport, Tex. Feb-Mar 1969 109-168 136-185 69 Galveston Estuary June-July 1969 90-128 91-182 257 50 mi southeast of Freeport, Tex. Nov 1969 145-203 141-213 593 828 PARRACK: ASPECTS OF BROWN SHRIMP GROWTH Sx= an equation parameter, the asymptotic size, b = an equation constant related to the size at birth, Aa = a^ - o„, = time at large, Gr - age of an individual on the date recap- tured, a,„ = age on the date marked and released, and G, g, and k are equation parameters. Equation parameters S^, k, G, and g were esti- mated by utilizing the Marquardt algorithm to minimize the residual sum of squares: y:iS'-s f \ f J./ where n = the number of individuals marked and recaptured, S'r = the observed size at recapture, and Sr = the size at recapture as estimated by the growth equation. The remaining equation constants, b in Equa- tions (la) and (2a) and a, in Equation (3a), are respectively computed: 6 = (sjs^)-i b = is^-s^)is a. In 1- In (S^ IS.) g (5a) (5b) (5c) where S^ is the size at birth and other symbols are as before. The parameter b in Equation (4a) is simply the size at birth. Studies of the early development of brown shrimp indicate that newly hatched larvae are 0.35 mm total length (Cook and Murphy 1971). Estimates of the equation constant b in the logistic and von Bertalanffy models were based on that length at birth. Shrimp eggs are 0.26 mm in diameter (Cook and Murphy 1971) and about the density of water (Cook and Lindner 1970) so that the weight at hatching is about 0.000009 g. Brown shrimp un- dergo metamorphosis 11 to 15 days after hatching (Cook and Lindner 1970) and are 0.0008 g at that time (Wheeler 1969). The weight at birth was cal- culated as the midpoint between that weight and the egg weight. Calculations of b and a, in the various models were based on that weight at birth. RESULTS Growth in Length In anticipation that differences in growth be- tween sexes may exist, equations were fit for males and females separately. Estimated equa- tion parameters (Table 2) are quite different be- tween sexes. The fitted models indicate that females are much larger than males of the same age. The estimates of the growth coefficient k do not differ greatly between sexes for both the lo- gistic and the von Bertalanffy models; the 90% probability support plane confidence intervals (Conway et al. 1970) extensively overlap for both models. The estimates of asymptotic length are, however, greatly different and such confidence in- tervals on those estimates are very disjoint. Pool- ing all data together to estimate overall growth functions for both sexes combined was therefore judged unrealistic. The relative abilities of the von Bertalanffy, logistic, and linear models to correctly reflect growth was judged by comparing residual sums of squares (Table 3). The von Bertalanffy function produced the smallest residual and the linear model the largest. The residual sum of squares for the linear model was well over three times that of the von Bertalanffy and logistic models for both males and females. The difference between the two nonlinear models was much smaller; the re- sidual of the logistic was but 8% larger than that of Table 2. — Growth models for brown shrimp. Lengths (L) in millimeters, weights {W) in grams, and ages (a) in months. Model Males Females Logistic L = 162.8/(1 + 464.14296 ^0.56643) L = 187.5/(1 + 534.71436 "0-6^ ^^3) von Bertalanffy Linear Gompertz L L W = 168.7(1 - 0.9979e"° 33573) = 0.35 + 4 21813 = 5.07(exp( 1.9996(1 - exp[-0.3735(a - 4.6688)])]) L L W = 193.6(1 - 0.99826 -0-33633) = 0.35 + 7.82093 = 3.55(exp[2.8359(2 - exp[-0.4410(3 - 3.2549)])]) Monomolecular Linear w w = 43.51(1 - 0.99996^°^^"*^^) = 0.0004045 + 1.8018a W W = 74.32(1 - 0.99996 ~°^'*^63) = 0.0004054 + 3.9013 829 FISHERY BULLETIN: VOL 76. NO. 4 the von Bertalanffy in the case of males and 5% larger for females. Table 3. — Residual sums of squares for six brown shrimp growth models. Model Males Females Length: Von Berlalantfy Logistic 44.161 65 47.661.96 155.797.40 163,278.00 Linear 162.661 15 599,677.13 Weight: Monomolecular 5.548.57 33,930.69 Gompertz Linear 7,027.72 12,335.42 38.751.26 67,526.07 Number of observations 1.536 3.588 The difference in growth between sexes and the ability of the von Bertalanffy model to fit the ob- servations is visible from plots of the observed lengths about the growth models. Data points were plotted by first calculating the age at release from the fitted model, adding time at large to com- pute the age at recapture, then plotting that age and the recapture size. The plots (Figure lA, B) show that sex specific growth does exist and that the differences are of significant magnitude. Further, the von Bertalanffy model does visibly fit the observed data. Although the observed data do o C\J CNJ 00 • ■-.^^^^■^-r^'^r'^ .■■••" ■. .J^i^r^r^ ;■• ^^ 'j/^- - ". ' ID- ^^^-^^""^"^""^ ./■M.-'. ■ . -^ • •• ■/^:".- ■■ f ' CM ro- ■■^ ■ ■ M- S ■'.f': ■ ■ . ■■/■■• -2- ^ m-- LENGTH 88 1 f-y A p: B / / / / C\J CNJ" o- 1- r 1 » — -r F 1 ? 1 > 7 10 12 14 AGE (MONTHS) 7 10 12 14 AGE (MONTHS) Figure L— Brown shrimp growth models. A) von Bertalanffy growth model, males; B) von Bertalanffy growth model, females; C) monomolecular model, males; D) monomolecular model, females. 830 PARRACK: ASPECTS OF BROWN SHRIMP GROWTH not in general fall close to the modeled line, the scatter is not severe. Growth in Weight The Marquardt algorithm (Marquardt 1963) was employed to estimate parameters of weight- length relations used to transform individual re- lease and recapture lengths into weights so that growth in weight could be modeled. Plots of the estimated relations (Figure 2) indicate them to be sex specific. Support plane confidence intervals (Conway et al. 1970) on equation parameters (90% probability) for males did not overlap those for females further indicating that the functions dif- fer between sexes. In addition the data were log- ged to linearize the relation and covariance analysis techniques applied to test for differences between sexes. The probability that the linearized functions are the same is small (P,. <0.00 1 ) further indicating the sex specificity of these relations. Further inspection of the plots shows the scatter of observations to be restricted and that the models effectively fit. These sex specific models were therefore employed to transform the data. The magnitude of residual sums of squares (Ta- ble 3) indicates the monomolecular model is the best predictor of weight at age and the linear model the poorest. The residual term for the linear fit is about twice as large as that for the monomolecular model and about 1.8 times that of the Gompertz for both sexes. The reduction in re- siduals of the monomolecular model as compared with the Gompertz was much smaller, 25*^ in the case of males and 14% in the case of females. As in the case of growth in length, estimated growth parameters indicate that growth in weight is sex dependent. Both the Gompertz and the monomolecular model estimate females to be much larger than males of the same age. Asymp- totic weight (monomolecular model) is estimated to be 75 g for females and 46 g for males; support plane confidence intervals (90% probability) on these estimates do not overlap. Estimates of the parameter k in the monomolecular model appear to be about the same for both sexes and in fact the support plane confidence interval for males com- pletely includes that interval for females. The differences in growth between sexes and the degree of fit of the monomolecular model is shown in Figure 1. Although appreciable scatter is ap- o (NlT o. o MALES WEIGHT = 3.931587 x IQ-^ LENGTH 3. 152658 0 24 48 72 96 120 144 168 192 LENGTH (MM) 0 24 48 72 96 120 144 168 192 216 240 LENGTH (MM) Figure 2. — Weight-length relationships for brown shrimp. 831 FISHERY BULLETIN: VOL. 76, NO. 4 parent, systematic departure of the observed points from the model is not evident so that the model does reflect the data. CONCLUSIONS The relative abilities of prediction of the differ- ent models can be judged by comparison of their residual sum of squares. The comparison strongly suggests that the linear function was by far the poorest model of brown shrimp growth both in length and weight. Although the size-age relation does appear linear for small young individuals, the rate of increase in size decreases with age, a phenomenon documented for many organisms both terrestrial and aquatic. A nonlinear function is therefore required to model brown shrimp growth throughout their entire life span. The residual sum of squares for the von Ber- talanffy equation was smaller than the logistic equation when modeling weight; however, these differences were not large. It is therefore not com- pletely evident that the von Bertalanffy equation is vastly superior to the logistic and Gompertz in the modeling of brown shrimp growth. The von Bertalanffy equation did, however, constantly fit these data best for both sexes in the modeling of both length and weight. This study does therefore show the von Bertalanffy model to be slightly superior to the logistic and the monomolecular model superior to the Gompertz for both sexes. The difference in the size-age function between sexes was found to be large. This phenomenon was previously reported for brown shrimp in the southern Gulf of Mexico (Chavez 1973) and northwest Atlantic (McCoy 1972) and for many other marine organisms. This study indicates that male brown shrimp apparently grow to approxi- mately only three-fifths the weight and five-sixths the length of females; however, the coefficients of growth, as indexed by /j in the monomolecular and von Bertalanffy models, are roughly equivalent. It is interesting to note that the rate of increase in size tends to fall off at an earlier age for males than for females (see Figure 1). Since, in general, a decrease in that rate roughly conforms to the age of maturity and sexual activity, it is not unreason- able to assume that males mature at a younger age than do females. Comparison of growth functions derived herein with those generated by other workers indicate that brown shrimp growth in the northern Gulf of Mexico is very different than that in the southern gulf and in U.S. Atlantic coastal waters. Growth functions derived from populations off Mexico (Chavez 1973) demonstrated a faster and pro- longed growth compared with growth observed in this study. That trend was consistent for both males and females. Studies off North Carolina (McCoy 1972) showed growth in Atlantic waters to be very rapid although a smaller asymptotic size was realized. As before, that trend was the same for both sexes. The kinds of data used and the methods employed to fit the growrth models dif- fered in all three studies; therefore, some dis- agreement in results may be expected. The magnitude of the differences observed, however, indicated truly different rates of growth may well exist in the three geographical locations. The growth of wild populations of white shrimp, Penaeus setiferus, a similar species, is correlated with water temperature (Gaidry and White 1973) in the shallow estuarine and nearshore areas they inhabit throughout their entire life span. Since the temperature of seasonally homothermic deep offshore waters where brown shrimp spend their adult life may be assumed to increase with de- creasing latitude, the differences in growth be- tween northwest Atlantic, northern gulf, and southern gulf brown shrimp populations are likely positively correlated with gross water tempera- ture. ACKNOWLEDGMENTS The staff at the Southeast Fisheries Center Gal- veston Laboratory, National Marine Fisheries Service, NOAA, provided assistance in this study. Susan Brunenmeister and Edward Klima con- tributed helpful advice in the reviewing of this manuscript. Patricia Phares and Scott Nichols provided valuable advice as to applicable statisti- cal procedures. LITERATURE CITED BERTALANFFY, L. VON. 1938. A quantitative theory of organic growth (Inquiries on growth laws. II). Hum. Biol. 10:181-213. ChAVEZ, E. a. 1973. A study on the growth rate of brown shrimp (Penaeus aztecus aztecus Ives, 1891) from the coasts of Veracruz and Tamaulipas, Mexico. Gulf Res. Rep. 4:278-299. CLARK, S. H., D. A. EMILIANI, AND R. A. NEAL. 1974. Release and recovery data from brown and white shrimp mark-recapture studies in the northern Gulf of 832 PARRACK: ASPECTS OF BROWN SHRIMP GROWTH Mexico, May 1967 - November 1969, U.S. Dep. Com- mer., NOAA, Natl. Mar. Fish. Serv., Data Rep. 85, 152 p. CONWAV, G. R., N. R. Glass, and J. C. Wuxox. 1970. Fitting nonlinear models to biological data by Mar- quardt's algorithm. Ecology 51:503-507. COOK, H. L., AND M. J. Lindner. 1970. Synopsis of biological data on the brown shrimp Penaeus aztecus aztecus Ives, 1891. FAO Fish. Rep. 57:1471-1497. COOK, H. L., AND M. A, MURPHY. 1966. Rearing penaeid shrimp from eggs to postlar- vae. Proc. 19th Annu. Conf Southeast Assoc. Game Fish. Comm., p. 283-288. 1971. Early developmental stages of the brown shrimp, Penaeus aztecus Ives, reared in the laboratory. U.S. Fish. Wildl. Serv., Fish. Bull. 69:223-239. Fabens, a. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Fontaine, C. T., and R. A. Neal. 1971. Length- weight relations for three commercially im- portant penaeid shrimp of the Gulf of Mexico. Trans. Am. Fish. Soc. 100:584-586. Gaidry, W. J., and C. J. White. 1973. Investigations of commercially important penaeid shrimp in Louisiana estuaries. La. Wildl. Fish. Comm. Tech. Bull. 8, 154 p. George, M. J. 1962. Preliminary observations of the recruitment of post- larvae and growth of juveniles of the brown shrimp Penaeus aztecus Ives in Barataria Bay. La. Wildl. Fish. Comm. 9 Bienn. Rep., p. 160-163. GOMPERTZ, B. 1825. On the nature of the function expressive of the law of human mortality, and on a new mode of determining the value of Life Contingencies. Philos. Trans. R.Soc. Lond., Ser. B, Biol. Sci. 115(11:513-585. Jacob, j. S. 1971. Observations on the distribution, growth, survival and biomass of juvenile and subadult Pe^iaews aztecus in southern Louisiana. M.S. Thesis, Louisiana State Univ., Baton Rouge, 68 p. Knudsen, E. E., W. h. Herke, and J. M. MACKLER. 1977. The growth rate of marked juvenile brown shrimp, Penaeus aztecus. in a semi-impounded Louisiana coastal marsh. Gulf Caribb. Fish. Inst., Proc. 29th Annu. Sess., p. 144-159. LOESCH, H. 1965. Distribution and growth of penaeid shrimp in Mobile Bay, Alabama. Publ. Inst. Mar. Sci., Univ. Tex. 10:41-58. McCoy, E. G. 1972. Dynamics of North Carolina commercial shrimp populations. N.C. Dep. Nat. Econ. Res. Spec. Sci. Rep. 21, 53 p. Marquardt, D. W. 1963. An algorithm for least-squares estimation of non- linear parameters. Siam. J. Appl. Math. 11:431-441. Medawar, p. B. 1945. Size, shape, and age. In W. E. Le Gros Clark and P. B. Medawar (editors). Essays on growth and form pre- sented to D'Arcy Wentworth Thompson, p. 157-187. Oxf. Univ. Press, Lond. Pearl, R., and L. J. reed. 1920. On the rate of growth of the population of the United States since 1790 and its mathematical representa- tion. Proc. Natl. Acad. Sci. U.S.A. 6:275-288. RINGO, R. D. 1965. Dispersion and growth of young brown shrimp. U.S. Fish Wildl. Serv., Circ. 230:68-70. Rose, C. D., a. H. Harris, and B. Wilson. 1975. Extensive culture of penaeid shrimp in Louisiana salt-marsh impoundments. Trans. Am. Fish. Soc. 104:296-307. St. Amant, L. S., J. G. Broom, and T. B. Ford. 1966. Studies of the brown shrimp, Penaeus aztecus. in Barataria Bay, Louisiana, 1962-1965. Gulf. Caribb. Fish. Inst., Proc. 18th Annu. Sess., p. 1-17. St. Amant, L. S., k. C. Corkum, and j. G. Broom. 1963. Studies on growth dynamics of the brown shrimp, Penaeus aztecus. in Louisiana waters. Gulf Caribb. Fish. Inst., Proc. 15th Annu. Sess., p. 14-26. Silliman, R. p. 1967. Analog computer models of fish populations. U.S. Fish Wildl. Serv., Fish. Bull. 66:31-46. Welker, B. D., S. H. Clark, C. T. Fontaine, and R. C. Ben- ton. 1975. A comparison of Petersen tags and biological stains used with internal tags as marks for shrimp. Gulf Res. Rep. 5(1): 1-5. WENGERT, M. W. 1972. Dynamics ofthe brown shrimp, Penaeus oz/ecMs Ives 1891, in the estuarine area of Marsh Island, Louisiana in 1971. M.S. Thesis, Louisiana State Univ., Baton Rouge, 93 p. Wheeler, R. S. 1969. Culture of penaeid shrimp in brackish-water ponds, 1966-67. Proc. 22d Annu. Conf Southeast Assoc. Game Fish. Comm., p. 387-391. 833 FISHERY BULLETIN: VOL. 76, NO. 4 APPENDIX The linear, logistic, and Gompertz functions were expressed in terms of size at release age, size at recapture age, and change in age (time at large) following the rationale presented by Fabens ( 1965) for the von Bertalanffy function (as follows). Each individual was of some unknown age (a,„) upon the date marked (t,„) and released. Upon the recapture date it^) the individual was of an unknown older age (a,) so that the difference between the release and recapture date (AO is equivalent to the increase in age (Aa) of that individual: /sa = /st = t — t = a —a . (Al) r m r m That equality can be substituted into the von Bertalanffy function when expressed in terms of the size at recapture (S^) and the age at recapture. Therefore the von Bertalanffy equation S^ = S^[l-6exp(-/2a^)] (A2) becomes S = S^{1 — b exp{—ka ) exp[—k{Aa)]). (A3) The equation, when expressed in terms of the size when marked (S,„ ) with rearrangement, is: bexp{-ka ) = 1-(S /S ). (A4) That expression is substituted into Equation (A3) to yield the required function: S = S -{S -S )e-fe(^«). (A5) r "o ^ CO ui ' That form can then be employed to estimate the equation parameters k andS from mark-recapture data. The final parameter ib) can be calculated directly by first rearranging terms of the original function: S = S {1-be-'''') (A6) a °° SO that b = [l-{SJSJ]le-'''' (A7) where S,, is the size at age a. If the size at birth, i.e., at age 0, is known, then: b = 1-iSJSJ (A8) where S-^_ is estimated from Equation (A5) and the size at birth (S/, ) is derived from life history studies. That same rationale was applied to the logistic function. The size of a recaptured individual is expressed: S^ = S^/[l + b exp{-ka^)] . (A9) Since Ur = a + a,n substitution gives: S^ - S^I{l + bexp[-k(Aa)]bexpi-ka^)]. (AlO) Expressing the logistic equation in terms of the size marked and rearrangement of terms gives: bexpi-kaj + [{S^-SJISJ. (All) 834 PARRACK: ASPECTS OF BROWN SHRIMP GROWTH Substitution yields: S = S /(I + ((e~''^'''^)(S -S )IS )]. (A12) Since S,., S,„, and Aa were all directly observable from mark-recapture data, the logistic equation parameters S^ and k may be estimated from the data set. The remaining equation constant was calculated by rearrangement of terms: b = {(SJSJ-l) le-"^"'> . (A13) From life history studies the size at birth, Sf,, was determined. Since at birth age is zero (a = 0) the expression can be written: b = {SJSJ-1. (A14) The Gompertz function was likewise expressed in terms of the mark-recapture data. From Equation (3a), the size at the time of marking is: S^ = S. exp[G(l-exp[-^(a^-G.)])] (A15) and by substitution becomes S^ = S. exp[G -G(exp[-^(a^^^ -c.)]) (exp[-^(Aa)])] . (A16) Writing in terms of the size at recapture and rearrangement of terms gives: exp(-G exp[^(a^ -a.)]) = S^ /(S. exp G). (A17) Substitution yields the expression required to estimate the constants G and^ from the mark-recapture data: S^. = [S.exp(G)][S^/(S. exp(G))J^''P[ '^^"^^ (A18) where S, was the smallest size observed in those data. The remaining equation constant, a,, was then calculated by writing Equation (3a) in terms of the size at birth: S^ = S. exp[G( l-exp[-g(a-a.)])]. (A19) Since at birth age is zero (a = 0) the expression can be written as: a. = ln(l-[ln(S^/S.)/G]/^) (A20) where S , the size at birth, was determined from natural history studies. The linear function: S = b + ka (A21) a requires a much simpler derivation. Expressed in terms of the size at recapture: S = b + ka . (A22) r r Substitution gives: S = b + k {t.a+a ) . (A23) r ni 835 FISHERY BULLETIN: VOL. 76. NO. 4 The function expressed in terms of the size at release can be rearranged to ^,n = ^ -^ f'^.n ^A24) °m ^ ^^m ~^^/^ erated on a counter-current principle. Alternation of hot and cold water entering the experimental chamber between replicate runs eliminated any potential rheotactic interference. The inner experimental trough ' 1.75 m long x 5 cm diameten 0.8 mm wall thickness) contained DEFINITIONS The term "preferred temperature" has been used in various contexts in the literature 'e.g.. Brett 1952: Javaid and Anderson 1967: McCauley and Tait 1970: Tatyankin 1972; McCauley and Read 1973 ». Much of the variation in the use of this T0= /EW * h ®6 ^ ■ 7- V ^^~^ V csOST VIEW WITH COVER *'^ri_ ®-^)LJ-^^ # i- ^ Figure 2. — &nall experimental dbamber for temperature selection measurements in larval fishes: aj experimental chamber, b) water jaiiet, c) air line, d) drains, ei seawater input line, f > freshwater input line, gi daylight-simulating fluorescent light, h; light diffuser, i; viewing slits, jj thermistor probes, ki door on lightproof cabinet, 1> supports for water jacket, m; water flow control valves, n> nylon screen on ends of e^)erimental chamber. 840 EHRLICH ET AL THERMAL BEHAVIORAL RESPONSES OF FISHES term can be attributed to different species exhibit- ing various behavioral patterns in a gradient tank, just as they do in their natural habitats. This makes it impossible to use only one procedure to determine the preferred range for all species under all conditions. We determined the preferred range and final perferendum by evaluating, on a case-by-case basis, the behavioral responses of a species subjected to known conditions such as ac- climation temperature, feeding patterns, and cap- tivity environment. Experiments with larvae lasted 5-6 h, but juveniles and adults were tested for approxi- mately 7-8 h. An "experiment," in this study, con- sisted of a set of individual runs, with each "run" being the observation of the position and water temperature selected by each fish in the gradient at a fixed point in time. We employed constant time intervals between runs for any experiment: 5 min for larvae and 15 min for juveniles and adults. Run selected temperatures were the primary data source, and we calculated their mean, mode, and variance prior to combining them with data from other runs to determine the preferred tempera- ture. DATA ANALYSES The frequency of occurrence of all experimental temperatures was not uniform due to the shifting of the gradient as well as having a variable number of degree intervals per run and generally fewer than the 21 compartments. This caused a bias in the number offish observed at a particular temperature when summed over an entire exper- iment. To compensate for this, prior to calculation of the mean and modal selected temperatures, we adjusted the data by using the number offish per total occurrence of a particular temperature in all experimental compartments rather than the ac- tual number of fish at each temperature. We defined the "initial selected temperature" as that chosen by the fish immediately following es- tablishment of a gradient of lO'^C. The modal selected temperature was determined from the percent occurrence frequency distribution derived from adjusted data. After an initial time of appar- ent searching and testing of water conditions the experimental animals selected a temperature or range of temperatures at which they remained for the duration of the experiment. We called this the final selected temperature (or temperature range) and determined it from plots of selected tempera- ture against time. The mean selected temperature was derived by methods presented in Appendix Table 1. Reynolds (1977) reported that skewness of temperature preference frequencj' distributions required a complete description of the distribu- tion. We examined the following parameters to delineate thermal behavioral responses, the ini- tial, mean, modal, and final selected tempera- tures, standard deviation about the mean, coefficients of skew^ness and kurtosis, and coefficient of dispersion. The first four parameters defined the preferred temperature range. The standard deviation about the mean selected temp- erature quantified movement through a range of temperatures and gave a measure of the degree of eury- or stenothermal preference. We used coefficients of skewness and kurtosis (Sokal and Rohlf 1969) in testing for normality and then to help define the shape of the temperature-specific fish frequency of occurrence distribution and to refine interpretation of behavioral types. The coefficient of dispersion quantified the tendency of a species to aggi-egate or school and gave the per- centage of use of the experimental chamber by all fish within one standard deviation of the run selected temperature. EXPERIMENTAL TECHNIQUES AND BEHAVIORAL RESPONSES Our experimental techniques and data in- terpretation methods are useful for a wide variety of behavioral tj-pes. There are three salient fea- tures of this methodology- 1) the shifting and re- versal of the temperature gradient to partition position preference from thermal preference, 2) the extended duration of the experimental period and its relationship to the thermal histon,- of the test organisms, and 3) the criteria for behavior evaluation. Shifting and Reversal of Temperature Gradient Hasler ( 1956) pointed out that fishes in experi- mental gradients can position themselves accord- ing to small deformities in the tank structure. We employed two methods to eliminate this factor: shifting the position of a given isotherm in the gradient during an experiment, and reversing the hot and cold ends between replicate experiments. 841 FISHERY BULLETIN: VOL. 76, NO. 4 Shifting the isotherm position during an exper- iment required the fish to thermoregulate ac- tively, similar to those studied by Beitinger ( 1976, 1977) in his temporal gradient. This technique demonstrated that the fish could follow an isotherm and did not necessarily arbitrarily select a position in the experimental tank. The precision with which a group of fish followed an isotherm varied between species and was related to the size of their preferred temperature range. Juvenile surfperch, Danialichthys vacca, for example, which preferred a narrow range of temperatures. closely followed an isotherm ( approximately 1 1°C) (Figure 3). In contrast, juvenile topsmelt, Atheri- nops affinis, after initially selecting approximate- ly 22°C, remained within that compartment, and shifting the gradient did not cause them to move until the temperature reached 26°-27°C. This isotherm was then tracked (Figure 3). Topsmelt are physiologically eurythermal, at least during embryonic stages, and the upper limit for hatching of topsmelt eggs is 26.8°C (Hubbs 1965). Brett (1956) suggested that the preferred temperature may not be a strong enough directing force to move I I I 1 I I I I T I • r 180 240 TIME (minutes) Figure 3. — Changes in fish and isotherm position in the experimental gradient. Juvenile Damalichthys vacca followed the 10°- 1 1°C isotherms. Juvenile Atherinops affinis remained in the position they initially selected and moved very little until the temperature reached 26°-27°C. They then followed the temperature range of 25°-27°C. The small numbers indicate isotherms. Large dots indicate the mean temperature selected by nine individuals. 842 EHRLICH ET AL: THERMAL BEHAVIORAL RESPONSES OF FISHES fish with wide temperature tolerances from a par- ticular area until stress-inducing conditions are reached. Temperatures Selected and Their Relationship to Thermal History We classified the behavioral responses of the 16 species surveyed into four groups based on changes in temperatures selected throughout an experiment: 1) immediate response — no general shift in selected temperature over time, 2) fast response — a shift in selected temperature not ex- ceeding the first 2 h of the run, 3 ) slow response — a shift in selected temperature over more than 2 h, and 4) positioned response — a broad preference and a tendency to remain in a given position in the gradient until conditions become extreme. Shiner surfperch, Cymatogaster aggregata; pile surfperch, Damalichthys vacca; black surfperch, Embiotoca jacksoni; and black croaker, Cheilo- trema saturnum, showed the first behavioral pat- tern of immediate response (Figure 4). These fishes moved most quickly from their preexperi- mental acclimation temperature to their final selected one or range, remaining there for the du- ration of the experiment. These fishes generally had the narrowest selected temperature ranges and also aggregated tightly (Table 1). Fishes with a fast response to the temperature gi'adients included speckled sanddab, Citharich- thys stigmaeus; seiiorita, Oxyjulis californica; spotted sand bass, Paralabrax maculatofasciatus; 25 20 — 15 Ol5 o II). I • ♦ * ♦ 4 • ♦ Cymatogaster aggregota ,.l..()t,'*l I • ' • I ' • ' I I • « • I I I • I UJ |_20H < cr I5H .♦♦*♦♦♦♦•♦♦.♦♦♦ 4 ♦♦ ♦ Damalichthys vocco T 4 4 I I I UJ • + 4 Embiotoca jacksoni I • • « I 1 M • ♦ . . I I T T 25 20 4 * ♦ . 4 ' N ♦ * 4 4 ..t Cheilotrema saturnum • I ' ' ' I ' ' • I •••'i.'t'* '"'" ' 1 ' ■ • 1 ' ' ' 1 ' ' ' 1 ' ' ' 1 ' ' ' ' 1 ' •-n 60 120 ISO 240 300 TIME (minutes) 360 420 Figure 4. — Immediate response to temperature change. These species showed no trend in selected temperature with time. Dots are mean selected temperatures and vertical lines are 1 SD about the means. Results are for duration of one experiment. and sculpin Scorpaena guttata (Figure 5). These species required up to 2 h to home in on a selected temperature and also generally did not aggregate as tightly, nor select as narrow a temperature range as those fishes that showed an immediate response to the temperature gradients (Table 1). All larvae studied responded slowly under ex- perimental conditions. These included topsmelt, Atherinops affinis; California grunion, Leuresthes tenuis; rockpool blenny, Hypsoblennius gilberti; and painted greenling, Oxylebius pictus (Figure 6). Four older fishes also showed this behavior: kelp bass, Paralabrax clathratus; olive rockfish, Sebastes serranoides; California halibut. Para- TABLE 1. — Behavioral responses of larval and juvenile fishes in temperature selection experiments. Initial and final selected temperatures are taken from Figures 4-6 and other similar experimental data. Mean Mean Coeffi- Coeffi- Coeffi- Standard Acclimation Selected temperatures ( = C) cient of cient of cient of Experimental date No of length (mm) temperature skewness kurtosis disper- Species fish Mean SD Mode Initial Final (g.) (92) sion (%) Immediate response: Cymatogaster aggregata 12 June 1975 8 109 18.2 19.9 2.1 19 18.7 21 0.22 2.87 24 Damalichthys vacca 6 June 1975 9 69 18.1 10.5 0.9 10 11.4 11 0.41 4.6r 12 Embiotoca jacksoni 13 Dec 1974 7 118 16.7 18.0 1.6 18 17.0 18 -0.61- 4.54- 5 Cheilotrema saturnum 6 Oct. 1975 7 42 17.0 27.6 2.0 28 26.6 28 -0.79- 3.12 13 Fast response: Citharichthys stigmaeus 22 Dec, 1975 9 90 18.9 10.1 2.6 10 14.8 10 0.43 3.44 33 Oxyjulis californica 28 May 1975 6 120 17.2 15.5 1.9 15 15.0 16 -069- 5.44- 11 Paralabrax maculatofasciatus 31 July 1975 6 179 20-6 24.2 3.1 27 21 2 25 -0.83- 2.70 29 Scorpaena guttata 24 Nov. 1975 6 64 17.6 17.5 4.2 19 17.2 17 -0.63- 4.01" 35 Slow response: Atherinops affinis (larvae) 31 July 1975 6 14.5 21.5 252 2.9 27 21.9 27 -1 07- 4.12- 26 Leuresthes tenuis (larvae) 9 May 1975 6 8.1 16-5 25.2 4-1 26 19.2 27 -0.12 2.86 37 Hypsoblennius gilberti (larvae) 2 July 1975 6 4.4 19-4 222 3-1 19 19.7 26 0.39 2.10 36 Oxylebius pictus (larvae) 14 May 1975 6 3.4 16.0 26.8 33 27 19.7 29 082- 3.24 30 Paralabrax clathratus 28 July 1975 6 196 21.0 13.5 3.1 14 17.2 15 0.13 2.70 47 Sebastes serranoides 11 Dec 1974 8 82 17.0 180 1.3 18 16.2 17 -0.21 2.44 4 Paralichthys californicus 15 Oct. 1975 5 94 20.5 20.8 6.6 24 20.3 22 -0.10 1.91- 65 Pleuronichthys coenosus 3 Dec. 1975 4 134 10.0 7.5 2.5 7 10.8 7 1.40" 5.15- 23 Positioned response: Atherinops affinis 14 Jan. 1976 9 60 15-0 23.3 32 26 16.4 26 -0.71- 2.41 4 •P<0.05. 843 15 10' \\ Cithorichthys stiqmaeus 1 1 " ■' I M I H i f * * M * ♦ M H H I 5*11 20i I I I — I I I I I I • — r- 3 • ihl. O 10 I • • ♦ 4 • f ♦ I I I I I I I I I I Oxyjulis colifornica T T T tt: 15 UJ 35 a. 2 30 UJ t-25 20 15 HI W lii^ Porolabrox maculatofasciotus I I I I I I I I T T T Scorpoena guttata l|tiMlf|||) 10 ' I I I I I I I I I I I I I I I I 1 I I — I I • « I I • • I 60 120 180 240 300 360 420 TIME (minutes) Figure 5. — Fast response to temperature change. These species changed their selected temperatures over the first 2 h of the experiment only- Symbols as in Figure 4. lichthys californicus; and C-0 turhot, Pleuronich- thys coenosus (Figure 6). Members of this group required more time to stabilize their response than either the immediate or fast responders. The temperature selection acuity and aggregating tendencies of these fishes were similar to those of the second group (Table 1). Juvenile topsmelt were the only species ob- served that showed a positioned response (Figure 3). Ehrlich et al. (in press) discussed this behavior in detail. California grunion are closely related to topsmelt, and we have observed them together, in the field, throughout larval, juvenile, and adult stages. Possibly juvenile California grunion, which were not tested, may show similar re- sponses. Preexperimental acclimation temperatures showed the greatest effect on thermal selection of the fishes during the first 2 h after establishment of the gradient ( Figures 4-6). The short duration of the influence of thermal history on temperature selection has also been reported by Doudoroff (1938). Clearly, trying to determine a preferred temperature for these species or others with simi- lar responses, during the transition period, would make data interpretation difficult. After this ini- tial period, the fishes, in most cases, chose a final selected temperature, which may be synonymous with what Fry (1947) termed the "final preferen- dum." He defined this as the temperature range 30 25- 20 • 35 1 30 25 20 FISHERY BULLETIN: VOL. 76, NO. 4 Atherinops offinis larvae .. 1^ M p* t f I M JM* *H 15 25' 20 O 30' o ""' 25 LU (T 20 Z) I Hilt (■■It'ltMil Leurestl\as tenuis larvae ' ' ' I .,|nHtt).|||t-*t*'t't' Hypsoblennius gilbert i larvae T T T .1 tllH ttt)tHtlM(tt*l Oxylebius pictus larvae < 20-j UJ 15- a. Ill iiiiiii| I • ' ' I M|,j|MtH|llt(.Mli.n*' Poralabrax ctothrotus 20' 15 ■ 10' 35' 30' 25' 20' 15' I I I i I I I I I I I I I I I I I I I — I I I I ♦ . • ♦ ♦ ^ M • ♦ Sebastes serranoides -r-^ \.^ Paralichttiys californicus n ,1 4 t " 10* I I I I I I I I I I I I I I I I — I I I I ■ — I — I I I I I 10 Pleuronictithys coenosus *''l]IlT'*»»**+***»»***»* 1 I I I I I I I I I I I I I I I — I I I I I — I I I I I I I 60 120 180 240 300 360 4 20 TIME (minutes) Figure 6. — Slow response to temperature change. These fishes changed their selected temperature over more than 2 h in the experiments. Symbols as in Figure 4. that the fish would eventually select, independent of their acclimation temperatures. Topsmelt, however, did not show this pattern, for their initial selected temperature gave a good indication of their preference and was independent of their ac- climation temperature (Ehrlich et al. in press). Doudoroff (1938) also found that fishes did not select the temperatures to which they had been acclimated but rather selected a common range of temperatures, which he suggested must have some physiological significance. Figures 4-6 show that the final preferendum was reached within several hours after the estab- lishment of the gradient. This is considerably less time than the approximately 24 h reported by 844 EHRLICH ET AL: THERMAL BEHAVIORAL RESPONSES OF FISHES Reynolds and Thomson (1974) or Reynolds and Casterlin (1976). The differential, however, be- tween the acclimation and the final preferendum must be considered. Reynolds and Thomson ( 1974 ) tested fish acclimated 17°C below their final pref- erendum. Crawshaw (1975) used a range of ac- climation temperatures from 22°C below to 3°C above the final preferendum and found that as the temperature differential diminished so did the time required to reach the final preferendum. Dif- ferentials of 5°C required as little as 1 h and 3°C only 0.5 h (Crawshaw 1975). Based on the temper- ature differences between acclimation and the final preferendum (Table 1), Reynolds' and our results generally fit the pattern described by Crawshaw. Behavioral Criteria Most studies pertaining to behavioral responses of fishes to thermal gradients have been concerned with only one factor: the final preferendum. Addi- tional information, however, can be obtained from examination of parameters associated with the frequency distribution of the selected tempera- tures, particularly: skewness (degree of distortion from symmetry) and kurtosis (peakedness). Ivlev and Leizerovich (1960) compared the percent of the area under the curve of number of fish per temperature against the mode of the distribution as well as the percentage of the curve on either side of the mode. Reynolds and Casterlin (1976) and Reynolds (1977) discussed the relationship between various measures of central tendency (mean, mode, and median) and skewness. They also improved descriptions of thermal behavioral responses by quantifying skewness but did not state the statistical significance of the skewness. Sokal and Rohlf (1969) stated that the absolute value of coefficients of skewness and kurtosis have little meaning and that they must be tested for statistical significance. We identified distinct be- havioral types with respect to the frequency dis- tribution of selected temperatures by examining skewness and kurtosis. The responses were, in part, species-specific but also varied with on- togenetic stage and nutritive condition. Reynolds and Casterlin (1976) showed that skewness also varied diurnally. Kurtosis can be used to assess whether the test organisms display eury- or steno- thermal behavioral responses (Ivlev and Leizero- vich 1960). A narrow preferred temperature range will be overly peaked about the mean (leptokur- tic), and a broad range of preferred temperatures will show no obvious mode or only a very slight one (platykurtic). The coefficient of kurtosis is particu- larly useful for quantifying the strength of the temperature selection response in populations that are not normally distributed where normal parameters such as mean and standard deviations are inappropriate. A normal bell-shaped frequency distribution is representative of species with a wide preferred temperature range that is not close to lethal or avoided temperatures. Speckled sanddabs dis- played this type of behavior (Table 1, Figure 7). Newly hatched larvae, however, of species such as California grunion showed little temperature selection acuity and preferred an even wider range of temperatures (^i = 0.003, 0.50.05) nor lepto- or platykurtic (^2 ~ 2-5, 0.2 19 0 0 0 5 9 9 0 0 0 STEP 7. Determine the mean n MST = (0.20) (12°C) + 6 10 29 0 0 0 selected temperature MST E P i/lOOT, (0.60)(13°C) + 7 11 10 0 0 0 (MST), uhen TjtR. 1=1 (0.20)(14''C) = 13.0°C 8 12 f R 30 29 0.97 20% (See Data Analyses for 9 13 10 30 3.00 60% Artificial Data Set.) 10 U 12 13 14 15 16 X 30 10 , 25 30 1.00 0 0 0 0 20% 0 0 17 7 0 - 14 18 13 0 - 15 19 3 0 - 16 2n - 0 if .=4.97=f ARTIFICIAL DATA SET RUN SELECTED TEMP.(°C) MEAN SD TIME COMPARTMENTS cd 21 (min) 1 2 3 ^ 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 000 TEMP. (°C) 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 1.0 11.5 12.0 12.5 13.0 13.5 14.0 14.5 15.0 NO. OF FISH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 3 2 1 0 0 13.0 0.9 8 015 TEMP. ( C) 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11. n 11.5 2.0 12.5 13.0 13.5 14.0 14.5 15.0 15.5 16.0 NO. OF FISH 0 0 0 0 0 0 0 0 0 0 0 0 1 2 3 2 1 0 0 0 0 13.0 0.9 8 030 TEMP. (°C) 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 U.O 11.5 12.0 12.5 3.0 13.5 14.0 14.5 15.0 15.5 16.0 16.5 17.0 NO. OF FISH 0 0 0 0 0 0 0 0 0 0 1 2 3 2 1 0 0 0 0 0 0 13.0 0.9 8 045 TEMP. (°C) 8.0 8.5 9.0 9.5 10.0 10.5 U.O 11.5 12.0 12.5 13.0 13.5 4.0 14.5 15.0 15.5 16.0 16.5 17.0 17.5 18.0 NO. OF FISH 0 0 0 0 0 0 0 0 12 3 2 1 0 0 0 0 0 0 0 0 13.0 0.9 8 060 TEMP. (°C) 9.0 9.5 10.0 10.5 u.o 11.5 12.0 12.5 13.0 13.5 14.0 14.5 5.0 15.5 16.0 16.5 17.0 17.5 18.0 18.5 19.0 NO. OF FISH 0 0 0 0 0 0 1 2 3 2 10 0 0 0 0 0 0 0 0 0 13.0 0.9 8 075 TEMP. (°C) 10.0 10.5 11.0 11.5 12.0 12.5 13.0 13.5 14.0 14.5 15.0 15.5 6.0 16.5 17.0 17.5 18.0 18.5 19.0 19.5 20.0 NO. OF FISH 1 0 0 0 0 2 3 2 10 0 0 0 0 0 0 0 0 0 0 0 13.0 0.9 9 090 TEMP. (°C) 9.0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 13.0 13.5 K.O 14.5 5.0 15.5 16.0 16.5 17.0 17.5 18.0 18.5 19.0 NO. OF FISH 0 0 0 0 0 0 1 2 3 2 10 0 0 0 0 0 0 0 0 0 13.0 0.9 8 105 TEMP. C°C) 8.0 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 13.0 13.5 4.0 14.5 15.0 15.5 16.0 16.5 17.0 17.5 18.0 NO. OF FISH 0 0 0 0 0 0 0 0 12 3 2 1 0 0 0 0 0 0 0 0 13.0 0.9 8 120 TEMP. (°C) 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 U.O 11.5 12.0 12.5 3.0 13.5 14.0 14.5 15.0 15.5 16.0 16.5 17.0 NO. OF FISH 0 0 0 0 0 0 0 0 0 0 1 2 3 2 1 0 0 0 0 0 0 13.0 0.9 8 135 TEMP. (°C) 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 U.O 11.5 2.0 12.5 13.0 13.5 14.0 14.5 13.0 15.5 16.0 NO. OF FISH 0 0 0 0 0 0 0 0 0 0 0 0 1 2 3 2 1 0 0 0 0 13.0 0.9 %_ 849 LINEAR PROGRAMMING SIMULATIONS OF THE EFFECTS OF BYCATCH ON THE MANAGEMENT OF MIXED SPECIES FISHERIES OFF THE NORTHEASTERN COAST OF THE UNITED STATES B. E. Brown, J. A. Brennan, and J. E. Palmer^ ABSTRACT We evaluated the results of using historic bycatch (incidental catch) ratios in adjusting fishing regulations by linear programming techniques. We used both 197 1 and 1973 bycatch ratios separately to assess the sensitivity of the results to the reported changes in bycatch ratios in estimating the total 1975 catch of countries fishing in the northwest Atlantic. For 4 of the 1 1 countries for which data were examined, the difference between the percentage of a country's species total allowable catches (i.e., those catches allowed a country by regulation) using the 1971 and 1973 bycatch ratios, was at least 20% . Only four countries were predicted to catch at least 807f of their species total allowable catches. The predicted total catches of all countries and all species was only 60% of the total species quotas. The simulated directed fisheries constituted only 70% of the total catch using 1971 bycatch ratios and only 73% using 1973 bycatch ratios. Examination of the reported 1975 catches indicated that the total allowable catches for herring were most frequently limiting a country's catch. Except for U.S.S.R., the differences between reported and simulated catches were less than 50 metric tons, with the difference less than 10 metric tons for 6 of the 11 countries. There was little difference in reported versus simulated catches between the schemes using the 1971 and 1973 bycatch ratios. The control of fishing mortality by means of indi- vidual species catch quotas is difficult in a mixed fishery, i.e., where a significant proportion of the fishing mortality on a given species is generated as a result of the incidental catch, or bycatch, of that species in fisheries directed toward other species. Moreover, if a country is allowed to catch a spec- ified amount of a given species by means of a di- rected fishery for that species, the total species catch may exceed that amount because of the as- sociated bycatch of that species in the other fisheries. The International Commission for the North- west Atlantic Fisheries (ICNAF) modified its regulatory measures several times in attempts to account for bycatches of species under quota re- strictions. The initial haddock quota regulations (Subarea 5 and Division 4X, Figure 1) stated that the directed fishery should cease when the ac- cumulated catch (directed catch plus bycatch) re- ported to ICNAF biweekly reached 809f of the quota, anticipating that the catch after closure (a bycatch by definition) would be 209f of the quota (ICNAF 1969). When yellowtail flounder was added to the list of species under quota, the closure 'Northeast Fisheries Center Woods Hole Laboratory, Na- tional Marine Fisheries Service, NOAA, Woods Hole, MA 02543. procedures were changed. The Assessments Sub- committee of ICNAF estimated the expected monthly bycatch after closure of directed fisheries and the decision to cease directed fishing was then made when the accumulated total catch reported to ICNAF on a biweekly basis plus the expected bycatch during the remainder of the year equalled the quota (ICNAF 1970). With the introduction of national quota allocations in 1972, the procedure again changed, requiring each country to control its directed fishery so that the sum of its directed catch and the estimated bycatches would not ex- ceed its quota allocation (ICNAF 1972a). The bycatch problem was acknowledged by ICNAF in its decision to establish a TAC (total allowable catch, i.e., that catch allowed a country by regulation) for all species combined that was less than the sum of the individual species TAC's for 1974 and 1975 (ICNAF 1974a). Linear pro- gramming simulations utilizing bycatch ratios from directed fisheries for all countries combined substantiated this policy (Brown et al. 1973; An- thony and Brennan 1974). Since 1974, TAC's were set for all species (either singly or in groups) and for national catches (IC- NAF 1974a, 1975a). Under this regime, it was possible to utilize linear programming more realistically to investigate the extent to which the Manuscript accepted July 1978. FISHERY BULLETIN: VOL. 76. NO. 4, 1979. 851 FISHERY BULLETIN: VOL 76. NO 4 64°40 A Figure l. — Northwest Atlantic Ocean partitioned into ICNAF areas. regulations in ICNAF were adequate to account for the bycatch. Simulations of 1975 catches were made utilizing bycatch ratios from both 1971 and 1973 to assess the sensitivity of the technique to differences in historic bycatch ratios. Brennan ( 1975) found little evidence of a decline in bycatch ratios when examined on a country-gear level over the years 1970-73. We compared the simulated catches and the reported catches on a species basis and on a country basis and examined the results to determine for which countries and species the simulations were successful. METHODS AND MATERIALS Data Base Almost all countries fishing in Subarea 5 and Statistical Area 6 (Figure 1) submitted data on nominal catch (i.e., that reported landed (adjusted to live weight) by the country, not necessarily that actually caught — it is the term used in the ICNAF Statistical Records following standard United Na- tions Food and Agricultural Organization proce- dures) and effort for main species (or a species) sought. These data are published each year in tables 4 and 5 in the annual ICNAF Statistical Bulletins. The data of 1971 and 1973 (ICNAF 1972b, 1975b) were the sources of the bycatch ratios. Data of these years were reported according to the species categories given in Table 1. The nominal catches do not include fish caught and discarded at sea. The nominal catch and effort (days fished) for 1971 and 1973 for finfish were summed over months for each target fish of the fishery (the "main species sought") categories reported in ta- bles 4 and 5 of the ICNAF Statistical Bulletin (1972b and 1975b, respectively). Catches made with fixed gear as well as catches of Atlantic menhaden, Atlantic halibut, and large pelagic fishes, i.e., tunas, billfishes, and sharks (other 852 BROWN ET AL.: LINEAR PROGRAMMING SIMULATIONS than dogfishes), were excluded. Most of these were not covered by the regulations and have <1 1 (met- ric ton) per 100 t of directed species caught. In in- stances where no "main species sought" category was indicated or where landings were attributed to a mixed fishery, the monthly landings by vessel classification and gear were assigned to "species sought" categories according to the species which formed a simple plurality of the catch. The United States of America often reported mixed fisheries on groundfish species. The Union of Soviet Socialist Republics (U.S.S.R.), Poland, Japan, and German Democratic Republic (G.D.R.) typi- cally reported their pelagic and/or squid fishery catches as mixed. The term "fishery" as used in this paper refers to the vessels and associated catch on these "main species sought" categories. The term "species" re- fers to both individual species and species groups. All reported landings were thus identified by two factors: species and fisheries. Such tabulations were prepared for all nations for which data were available. For Romania, which has had an Atlan- tic herring fishery but did not report a directed Atlantic herring fishery in 1973, bycatch ratios for 1972 (ICNAF 1974b) were used for that species fishery. The only countries with an allocated na- tional quota for which 1971 and 1973 data were not available and thus could not be analyzed were Italy (1971 and 1973) and France (1971). In this paper, all catch restrictions described below will all be referred to as "quotas." To apply linear programming techniques to the bycatch problems restraints on the total catches for each species by country need to be set. For countries and for species categories reported in ICNAF Statisti- cal Bulletins, we used restraints in linear pro- gramming (ICNAF 1974a). For countries and/or species for which ICNAF had not set specific quota allocations (but for which the quota was included in, say, "other countries" under ICNAF regula- tions— a country not giv^ a specific catch quota could fish in competition with other similar coun- TabLE 1. — Species categories as reported to ICNAF, 1971 and 1973. 1971 1973 1973 Atlantic cod Atlantic cod Yellowtall flounder Haddock Haddock Other flounder Redfish Redfish Atlantic herring Atlantic halibut Silver hake Atlantic mackerel Silver hake Red hake Other pelagic Atlantic herring Pollock Other groundfish Other pelagic Amencan plaice Other fish Other groundfish Witch flounder Squids Other fish plus squids tries from an "other country" allocation or "other flounder" category), we estimated these re- straints by the following procedures. These were chosen so that the categories of quota allocations matched the species categories (Table 1) by which the catches were reported. We proportioned the "others" allocation category for each individual species to countries based on the 1973 nominal catch for each particular species and the catch of that species of all of the countries that did not have a national quota for the species. We proportioned the quota for "other groundfish" and "other pelagic" from the "other fish" TAC for each coun- try. The quotas for American plaice and witch flounder were subtracted from the "other floun- der" TAC for each individual country. Since the quota for pollock was set by ICNAF for Division 4VWX plus Subarea 5, national quota allocations were estimated as an average percent of the nomi- nal pollock catches during 1971, 1972, and 1973 in Subarea 4VW and 5. Analysis Methods Linear programming is a optimization method for which the effectiveness of an allocation scheme distributed over several variables is measured by the maximum or minimum value of some linear function of those variables, when those variables are subject to linear constraints. The problem con- sidered here was to determined = (Xj, X2, . . . , x„) such that 1 = 1 (1) is maximized, where for each /, c, was the weight- ing coefficients of the variable x^. In the present context, X, = catch of species / to be taken in directed fishery for species /, = catch of species / in all fisheries divided by catch of species i taken in directed fishery for species i (c, ^ 1.00), = number of directed fisheries considered, and z = total catch of all species. Solutions (Xj, X2, . . . , Xn ) of Equation (1) were sub- ject to the constraints for each / c, n (2) 853 FISHERY BULLETIN: VOL. 76, NO. 4 0 (3) where d,^ = catch of species j taken in directed fishery for species //catch of species i in directed fishery for species i bi = constraint on total catch of speciesj, for J = 1 . . . m. The estimates of d^ for each country for 1973 are presented in Appendix Table 1. Analogous tables for the 1971 data are in Brown et al. (1973). The solution used in this paper was devised by using the Simplex Algorithm (Hadley 1963: 132f) which was computed by using a Honeywell^ com- puter program LINPRO; a description of this use of linear programming is given in appendix II of Brown et al. (1973). In this analysis the linear constraints were that no country would exceed its national allocation for any species (6,). The output of the LINPRO program includes the vector X of directed catches of the species along with the re- sultant total catches of the species and the overall total catch. RESULTS AND DISCUSSION The results of each country's simulation are given in Appendix Table 2. In each case the sum of the species quota allocations exceeded the coun- try's maximum possible catch (without violating single species constraints) as determined by the linear programming model. Table 2 lists the ratios of the simulated catches to the TAC's using 1973 and 1971 bycatch ratios. For 4 countries (Bul- garia, Canada, G.D.R., and Japan) of the 11, the percentages derived from 1971 bycatch ratios dif- fered from those derived from 1973 fishing pat- terns by at least 0.20. More detailed reporting of catches (i.e., by species rather than groups) in 1973 than in 1971 and, therefore, in the analysis contributed to this change. Poland, United States, France, and Federal Republic of Germany (F.R.G.) were the only countries which could have taken >80'^ of the sum of their species TAC's based on 1971 or 1973 bycatch rates. The United States, however, has a significant discard of fish which is not taken into consideration in this analysis. Of the other countries considered, the effect of unre- TABLE 2. — Comparison of maximum catches from linear pro- gramming simulation using 1971 and 1973 bycatch ratios, with sum of species "quotas" for the ICNAF area. Maximum catch — sum of species quota using: 1 973 bycatch 1971 bycatch Country ratios ratios Bulgaria 0.64 0.83 Canada .54 .78 France .52 — Federal Republic of Germany .97 82 German Democratic Republic .40 .64 Japan .57 .17 Poland .94 .93 Romania .08 .05 Spam .72 .72 USSR. .25 .35 United States .90 .93 ported discard would be expected to be greatest in the Spanish squid fisheries. Closer inspection of Appendix Tables 2 and 3 reveals the species which were the limiting factors in a country's inability to take the sum of its species quotas at present. These are the species which were caught in significant amounts as bycatch and directed catch and for which a species quota was met. The species whose catch was most frequently limiting was herring, when either 1971 or 1973 bycatch ratios was used. The next major species using 1973 ratios were pollock and "other pelagic" and using 1971 ratios were "other fish," "other pelagic," and haddock. Pollock was less limiting when 1971 ratios were used because it was combined with the "other groundfish" cate- gory, which had not been limiting. The sum of the linear programming estimates over countries using 1971 and 1973 data are pre- sented in Tables 3 and 4, respectively. In each case the sum of the expected maximum catches deter- mined by the linear programming runs was only about 60^^ of the sum of the species quota. The simulated directed fisheries catch levels composed only 707c using 1971 bycatch ratios and 73% of the Table 3. — Sum of individual country's linear programming simulation of 1975 catches in the ICNAF area, maximizing total catch (1,000 t) and using 1971 bycatch ratios. Catches of France assumed to be those using 1973 bycatch ratios. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Total allowable Directed Total Species sought catch restraint catch catch Atlantic cod 45.00 1.7 18.53 Haddock 6.00 0.0 5.23 Redfish 25.00 6.60 22.20 Silver hake 175,00 43.65 62.68 Flounders 41.00 1.32 36.25 Other groundfish 152.00 64.08 84.49 Atlantic herring 175.00 140 14 176.69 Other pelagic 311.90 18907 210.48 Other fish plus squids 127.40 26.08 67.25 Total 1,058.30 482.64 683.81 854 BROWN ET AL : LINEAR PROGRAMMING SIMULATIONS Table 4. — Sum of individual country's linear programming simulation of 1975 catches, maximizing total catch (1,000 1), and using 1973 bycatch ratios for the ICNAF area. Total allowable Directed Total Species sought catch restraint catch catch Atlantic cod 45.00 16.39 31.48 Haddock 6.00 000 5.25 Redfish 25.00 1824 22.25 Silver hake 175.00 74 69 85.72 Red hake 65.00 11.83 26.51 Pollock 21.30 9.57 20.28 American plaice 2.70 — 1.15 Witch flounder 4.30 — 1.70 Yellowtail flounder 1600 11.02 15.06 Other flounder 18.00 — 6.54 Other groundfish 65.70 27 38 40.96 Atlantic hernng 175.00 107.38 120.01 Atlantic mackerel 28500 127.51 150.60 Other pelagic 26.90 16.97 26.45 Other fish 56.40 9.33 33.35 Squids 71.00 25.93 40.30 Total 1 ,058.30 456.24 626.75 total using 1973 bycatch ratios, the rest being taken as bycatch. The highest percentage of TAC's, which were caught in directed fisheries, were for other pelagics (9(K^), Atlantic herring (797r), other groundfish (76%), and redfish (75%) using 1971 bycatch ratios, and for Atlantic her- ring (89% ), silver hake (87% ), Atlantic mackerel (85%), and redfish (82%) using 1973 bycatch ratios. Referring to the individual country linear pro- gramming output tables in the Appendix, it is obvious that under 1971 and 1973 bycatch ratios, national patterns ran the gamut from almost a total mixed fishery by the U.S.S.R., and to a some- what lesser extent by the G.D.R., to very specific fisheries of the F.R.G. and Poland. As noted earlier, the species which was most frequently limiting to the total reported 1975 catch was Atlantic herring (6 out of 11 countries), and the countries which had the most limiting species TAC's were United States (5) and U.S.S.R. (4). Except for the catches of U.S.S.R., United States, G.D.R., and Poland, there was little differ- ence in reported total catch minus simulated re- ported catch, when 1971 and 1973 bycatch ratios were used. Moreover, only for U.S.S.R were these differences > 50,000 t, and for six of the countries the differences were < 10,000 t for both schemes. The species for which the simulated and reported total catches differed most varied by country. At- lantic herring and Atlantic mackerel were the species most frequently differing in simulated vs. reported catches, but Atlantic mackerel and silver hake contributed most in metric tons to the differ- ences. In general, and in view of the findings of Brennan ( 1975), the differences between schemes using 1971 and 1973 bycatch ratios were minimal, and more likely due to the different grouping of the data. A summary of the 1975 TAC's, the 1975 re- ported catches, and the linear program estimates of total catch by country, is presented in Table 5. It is obvious that the overall TAG of 850,000 t for 1975 would not be attained without exceeding cer- tain species TAC's unless bycatch was reduced, according to the simulations. The expected catches of 626,750 t using 1973 bycatch ratios and of 681,050 t using 1971 bycatch ratios are only 74% and 80% , respectively, of the 1975 total TAG On a country basis, and using the results derived from the 1973 bycatches, it can be seen that the country total TAC's were set for 1975 at approximately appropriate levels for France and Spain (based on Table 5. — Comparison of linear programming estimates of maximum total catch by overall country's total allowable catches (TAC's) in the ICNAF area. Figures in 1,000 t. 1973 nominal Sum of Linear programming estimate of Actual 1975 catch of species species 1975 total catch nominal catch of regulated by TACs total 1973 bycatch 1971 bycatch species regulated on total TAG Country the total TAG for 1975 TAG ratios ratios Bulgaria 37.29 34.40 24.65 2222 28.74 24.69 Canada 16.80 26 32 26.00 14.24 20.51 14.00 France 3.62 529 2.95 276 2.76 3.36 Federal Republic of Germany 38.28 30.89 24.85 3005 2531 25.10 German Democratic Republic 150 85 100 98 82.85 40.52 64.17 82.74 Italy 3 92 4.15 (') (') 4.40 Japan 32 90 45.35 21.25 26.05 7.59 20.84 Poland 190 55 153.94 129.25 144.87 144.37 127.05 Romania 7.14 5.71 3.85 0.46 0.27 1.80 Spain 22.20 20 98 14.80 15.06 15.10 14.65 U.S.S.R. 449.04 36664 301.80 93.10 127.02 313.78 United States 203.09 26237 211.60 237.42 245.21 221.04 Total 1,155.68 21,052.87 3850.00 62675 "681.05 853.45 'No estimate available. ^Six thousand metric tons of other species not prorated to other species. ^Includes 2,000 t allocated to others. "Due to the absence of bycatch ratios for 1971 data, estimate of France's total catch is derived from the 1973 bycatch ratios. 855 nSHERY BLLLETTN: VOL. 76. NO. 4 reported statistics', too low for the F.R.G.. Japan. Poland, and United States, and too high for the other countries. In fact, summing the national total TAC's rather than the linear program esti- mate-; • "'■ c • : - - r>- catch, when the former are limit- ing. : " overall estimated catch, results in an expe :atch of 575.000 t. only 68^ of the TAG. The analogous expected total cater, derived from 1971 bycatch ratios was 627.470 t. only 74'~c of the overall TAG. Bycatch may be reduced through actions initiated by fishing fleets or by regulations such as the closure to bottom trawling by larger vessels in the south- em New England. Middle Atlantic, and Georges Bank areas IGNAF 1975' for 1975 and by the similar closure on Georges Bank for 1976. The reduction of the overall TAG to 650.000 t in 1976 IGXAF 1976 ' and 525.000 1 in 1977 > IGXAF 1977 1 was designed to reduce the bycatch problem. It should be noted, however, that despite the above potential for change as well as the in- a^T _ -lies of the rep>orting to IGXAF. which may c:~ : -r -.ore than one directed fisherv" under a mixed category-, there were other factors which worked in the opposite direction. The first was the inadequate recording of bycatch noted during in- ternational inspections. Some of this was dis- carded and not rep>orted. and some was apparently utilized but not accurately reported on logbooks. Both the lack of reporting and any underestimates of bycatch can cause the bycatch ratios used in this anal\-sis to be underestimated. In mixed species fisheries, bycatch mxist be con- sidered in the allocation of quotas to species and to elements of the fisherj- 'in this example the ele- ments are countries, but under different cir- cumstances they cotild be otherwise — e.g.. ports i. Lack of attention to attendant bycatch may result in an unexpected overhar\"est of selected species or conversely the wastage of large quantities of pro- tein depending on whether or not the directed fishen.' ceased when a small amount of bycatch had been taken. Linear programming pro%ides a suitable technique for examing this problem. However, to have a refined analysis, accurate statistics as to main species sought and the com- position of the bycatch including discards must be available. Lacking these, the inferences as in this pap>er. are directional. The specific indi\-idual es- timates can be interpreted for policy decisions only when the user has the understanding of the fishen.- to qualitatively account for the appropriate re- porting inadequacies. LITERATURE CITED ANTHONY. V. C. .^NT> J. .\. BRENN.^N. 1974. An example of the by-catch piroblem on directed fisheries for 1975. Annu. Meet. Int. Comm. Northwest Atl. Fish.. Summ. Doc. 7-i 47 'Re%is€d i. Ser. No. 3386. 5 p. BREN'N.\N. J. A. 1975. By-catch trends of selected fisheries operating in ICNAF Subareas 5 and 6. Annu. Meet. Int. Comm. Northwest Atl. Fish.. Res. Doc. 75 70. Ser. No. 3554, 14 p. BROWN. B. E.. J. A. BREN-N.\N. E. G. HEYERD.\HL, .4NT) R. C. Hen-n-emlth. 1973. Effect of by-catch on the management of mixed specie fi.gheries in Subarea 5 and Statistical area 6. Int. Comm. Northwest Atl. Fish.. Redb. 1973. Part HI. p. 217- 231. H.\DLEY. G. F. 1962. Linear pn^ramming. Addison- Wesley, Reading. Mass.. 520 p. INTERNATIONAL COESOSSION FOR THE NORTHWEST ATIAN- TTC FISHERIES. 1969. Int. Comm. Northwest Atl. Fish. Annu. Proc. 19. 55 P- 1970. Int. Comm. Northwest Atl. Fish. Annu. Proc. 20. 47 P- 1972a. Int. Comm. Northwest Atl. Fish. Annu. Proc. 22, 94 P- 1972b. Int. Comm. Northwest Atl. Fish.. Stat. Bull. 21, 135 P- 1974a- Int. Comm. Northwest Atl. Fish.. -Annu. Proc. 24. 128 p. 1974b. Int. Comm. Northwest Atl. Fish., Stat. Bull. 22, 239 P- 1975a. Int. Comm. Northwest AtL Fish. Annu. Rep. 25. 116 p. 1975b. Int. Comm. Northwest Atl. Fish.. Stat. Bull. 23. 277 P- 1976. Int. Comm. Northwest Atl. Fish. Annu. Rep. 26, 139 P- 19 / / . Int. Comm. Northwest Atl. Fish. Annu. Rep. 27, 84 P- 856 BROWN ET AL.: LINEAR PROGRAMMING SIMULATIONS •E CO tC C bo 3 o v a. JZ u "S 0) S5 c a> o < 1s 2 iS '^ Q. ce X •^ 0) CO u -*^ o in a, ^ -^ C 3 o be C c ca CO c E o c CO Ed < E- X 3 z w a. a. < (uro y O) = 0) OB Og ^1 2 o a) o §g J58 CO o CD p d I I I I I I I I II o o o O I CD O in I o o I I o in 5 I 8 b C3 -■ O o I C^ O o cb 6 ^ cj) r- 38 CD d en o ■* o CM o - in 1- o o o o d 6 II I I CO I I I 8 I d I I I 8 I un <3^ o CM in O I I - o o I I d o -^ <- -t a> o q 5 I I d d d o p II I d o c\j T in o o o I I o o o I I d d d CO in cy o o o I I d d d •^ (D CM GO CM r-^ I I O - 'J- I I d o CO I I I I o ° I S ■* S I I o o) c6 I I p o o odd CO o o 00 I CO >- o r- I o o o ■■- o o ^' o I I CO CD 2 o o o f^ CM in ^ f^ o CO •- r^ CO ^ o p p 00 CO Tt o d d d d d '-^ GO CM in o *- CO CD m o r^ p o p o o d d ^ f~ c^ r^ CO o f^ m in -- o o ^ O CO O T CM d d -^ d d I I CM CM o o o o o »- o O) CD r^ o r^ CO m CO in CO o o> O CM o o <- o '- d o d d o T- d - I ■- d d d ^ -^ CO CM •^ 2 I o 1 o op I o IP d d d d 00 O CM ^- T- I I in o in ■»- CO I I cu O CO CO CM d y- d d d f^ o in in CO o ■^ I o r^ CM CO CO p I o o o o o d -^ d o d d CM CM cj> rs. r^ I I ^ ^ CO »- in I I o o o o o d d d d d I I I I MM II II I I I I I II II II II II I I M II II II MM II I II I II I I I II II I I I II Mill M MM I II II O -- CD O O I O o o I O ■^ d d CO d I II I II ■- o CO o o o o o o MSI CM O d II I 8 CO '- CD I O O O I o p O O CM O . CM ^ CM O I O O O I 5 I I I I CM r~ ■■j- CO CM Tt I 2 CO I O O O d I I CD -- O O O O d d C35 CM 1^ O O CM I I o p o I I odd ^ o o .- O O I I CM o r^ I I o >- d o o^ f^ O "^ 00 I I o m o I I — d -- I I I I 881 1 1 5 in 5 d d d d II I II 8 II I d I II I I II II I II M II O O O O O d I II I 8 I I I O O o o d d I I I 857 FISHERY BULLETIN: VOL. 76, NO. 4 u o 0) a. 0)^ o- OS £ c Og 0) p is E Q. < X S z u a. a. < o o o d in CO m in o o u ^ c!¥ n < > CM o ra o CVJ o - d o O) CO "^ o <- (D 1- CD O O O ^ ■- O CO o o o »- o CO CO 88 CM 8 f^ o o o o ..- o in >- o in ID o o o o T- o o O 1- Tt CM C\J CO O O '^ O O o o o o o o o o o o ■^ CO C\i CO CM o o o o o o o o p o o o o o o CO O) 1^ CD O O I I IT I C^ o o ! I o I T- o d> o cri CD o '- o -^ a> lo CO o o ■^ a^ en -^ CM O ^ ■^ P CM CM o ^ o o o o o o o '~ r^ 1^ CM h- o ^ r^ 00 ^ CD Tj o ^ \n y- -- <:> ■•- ^' CD CD CD O O O ■a- Ol CO T- 1 o o 1 1 o o 5 o o o o I I I 8 I in o -^ '- o o CM CT) r: r: I o o T o o I o o d o d co.^mr.-incoCM»- OOCMf^-OOO"! oooooooo doododdo o o C75 CD *- I O .- O I O CM O in en I S 'D o o o d o o o (D CO I O - - I o o d '- ■^ r^ o in rvi o o o o T- o o o o ■- o CO CD 11188 o o I I CM ■- I O O O O ^cOf.-cor^ooin oo^ocoocdcm opoooo--p dddoO'-dd CO CM O Tt I I o o O CO o o o o I I I Tj- ^ t- CO o O^ ■^ '- O I I OJ O IT) I o p p I loop I O O O O ■^ CD o o o CO in o o o in o o CD in o CO 00 o in o o o o o o o o o o o o in CO o s o o o CD o o in CO o CO in o o o o o o o o o cDr^cDcoi^o-^cMo coo o in'-';j-Lnincoor-~o i -^co i »- OOOOCOt-OOO I "-p I "- ddoddodO'- (d c> d COCU'^COOOO CDCD >- ^cooino.* I ■•- o I 2 O'-oop^ Ipp IP dodo*-o c> ci o .^ CO r^ o CO o o ■^ o o CM t^ I o >- o o -^ o I d d d <- d d ■- d o p p I T- o o o d d d d I I I ocoa)'^cocO'-CM-<^i- COO 0«3'COir)C\J(DC7)01^0 I O'- OCOOOO'-O'^COP Ipp '-^OCDcbOCDOOoO (DO 0) ro -^s •p -O DC V 3 © ^ SZ (0 CJl O CJ Q. .^ ".' ^- ~ ~ o o'-^ 6 -^ ^ C ^ -« C3. £ 3 .S B E ^ o .c = c _ D o •= 0 M 0)01 — T 5 ■* ^ -2 Q. w ^-CJZ ro^-!r= C3)U 5g-«I^IS5l£KS^^I^Bl££5S££& <(f>(/iincr.O<<00^00<<00(n 858 BROWN ET AL.: LINEAR PROGRAMMING SIMULATIONS Appendix Table 2. — Linear programming simulation by country in ICNAF Subarea 5 and Statistical Area 6, 1975 catches to maximize total catch (1,000 t). Simulated using 1973 bycatch ratios. Actual directed and total catches are included also. Total Total allowable catch Simulated Actual allowable catch Simu ated Actua Directed Total Directed Total Directed Total Directed Total Species sought constraint catch catch catch catch Species sought constraint catch catch catch catch BULGARIA POLAND Atlantic cod 0.07 — 0.03 — — Atlantic cod 0.49 — 0.37 — 0.48 Redfish 0.50 — 0.03 — — Redfish 0.40 — — — <0.01 Sliver hake 200 — 0.92 1.02 1.92 Silver hake 5.30 — 0 13 024 0.38 Red hake 5,41 — 0.23 — 0.03 Red hake 2.20 2.12 2.20 — — Yellowtail flounder 0.14 — 0.06 — <0.01 Pollock 0.35 0.28 0.35 — 0.02 Other groundfish 0.65 — 0.13 — 0.34 Other groundfish 1.40 — 1.40 — 1.11 Atlantic herring 1 20 0.47 1.20 — 0.42 Atlantic herring 38.40 32.14 38 40 33.05 38.46 Atlantic mackerel 18.75 1864 1875 18.47 18.75 Atlantic mackerel 90.00 81.45 90.00 68.45 74.28 Other pelagic 075 — 0.15 — 0.39 Other pelagic 220 0.15 220 0.17 3.77 Other fish 2.60 — 0.48 — 2.63 Other fish 6.40 0.34 6.40 — 1.71 Squids 1.70 — 024 — 0.21 Squids 6.80 0.45 3.42 3.25 6.84 Total 34 40 2222 24.70 Total 153.94 144.87 127.05 CANADA ROMANIA Atlantic cod 4 82 0.55 1.31 1.10 1.93 Haddock 0.01 — <0.01 — — Haddock 1.20 — 0.60 0.44 1.44 Redfish 0.34 — <0.01 — 0.01 Redfish 0.50 — 0.02 0.01 0.06 Silver hake 0.50 — <0.01 — 0.12 Pollock 2.46 — 2.46 4.13 4.74 Yellowtail flounder 0.01 — <0.01 — — American plaice <0.01 — <0.01 — 0.02 Other groundfish 0,15 — 0.01 — <0.01 Witch flounder <0 01 — <0.01 — 0.01 Atlantic herring 0.20 0.20 0.20 1.54 1.54 Yellowlail flounder 0.02 ~- <0.01 — 0.01 Atlantic mackerel 3.75 0.05 0.10 — 0.07 Other flounder 0.03 — 0.02 — 0.05 Other pelagic 0.13 — 0.13 — — Other groundfish 0.78 0.70 076 030 0.66 Other fish 0.02 — 0.02 — — Atlantic herring 9.00 9.00 9.00 5.08 5.08 Squids 0.60 — <0.01 — 0.05 Atlantic mackerel 7.50 — 0.06 — <0.01 Total 5.71 0.46 1.79 Other pelagic 0.01 001 0.01 — — SPAIN Total 26.32 14.24 14.00 Atlantic cod 7.09 1 49 1.49 407 4.07 FRANCE Haddock 0.30 — 0.10 — 0.07 Other groundfish 0.02 — 002 — — Red hake 0.07 — <0.01 — 0.01 Atlantic herring 1.87 1.87 1.87 3.34 3.34 Pollock 0.42 — 0.42 — 0.10 Squids 3.40 087 087 — — Other groundfish 0.10 — 0.05 — 0.42 Total 5.29 2.76 3.34 Squids Total 13.00 20.98 13.00 13.00 15.06 9.90 9.90 14.57 FEDERAL REPUBLIC OF GERI^ANY Atlantic cod 009 — 0.01 — 0.02 USSR. Silver hake 0.50 — 0.04 — 0.04 Atlantic cod 2.50 — 0.24 — 2.43 Pollock 1.60 1.60 1 60 0.10 0.15 Haddock 0.05 — 0.05 — 0.01 Other groundfish 0.90 0,48 090 — 0.02 Redfish 1.44 — 1.44 — 1.37 Atlantic herring 24.50 24 50 24.50 22 99 2301 Silver hake 113.30 40.20 41.22 71.38 88.88 Atlantic mackerel 1.40 0.99 1.40 0.08 0.47 Red hake 44.40 — 11.18 4.50 26.12 Other pelagic 0.51 — 0.35 — 1.46 Pollock 1.26 — 0.20 — 0.19 Other fish 0.39 — 025 — — American plaice 0.20 — 0.05 — 0.18 Squids 1.00 0.68 1.00 — 003 Witch flounder 0.20 — 0.05 — 0.20 Total 3089 30.05 25.20 Yellowtail flounder 0.84 — — — 0.08 Other flounder 0.60 — 0.20 — 0.56 GERtVIAN DEfVlOCRATIC REPUBLIC Other groundfish 16.70 — 2.79 — 2.86 Atlantic cod 1.30 — 0.03 — 0.03 Atlantic herring 42.10 1.91 5.28 37.08 40.95 Redfish 0.63 — 0.02 — 0.01 Atlantic mackerel 101.25 1.96 14.80 99.91 106.31 Silver hake 3.10 — 0.06 — 0.04 Other pelagic 4.40 4.15 4.40 — 0.68 Pollock 3.50 3.49 350 <0.01 0.10 Other fish 28.90 — 8.20 5.99 34.08 Other groundfish <0.01 — — — 0.07 Squids 8.50 — 3.00 3.53 8.94 Atlantic herring 31.90 13.00 13.75 27.00 30.90 Total 366.64 93.10 313.84 Atlantic mackerel 5625 20.00 20.14 47.95 48.34 Other pelagic 0.06 0.06 0.06 UNITED STATES Other fish 294 — 2.90 0.12 2.18 Atlantic cod 28.00 14.35 28.00 12 46 23.41 Squids 1.30 — 0.06 — 0.90 Haddock 4.50 — 4.50 086 5.09 Total 100.98 40.52 82.63 Redfish 20.62 18.24 20.62 7.07 896 Silver hake 43.00 34.49 43.00 17.79 2059 JAPAN Red hake 12.90 9.71 12.90 0.11 2.43 Atlantic cod 0.05 — — — — Pollock 11.50 4.19 11.50 3.80 806 Redfish 0.50 — 0.12 — 0.02 American plaice 2.50 — 1.10 0.26 2.19 Silver hake 7.30 — 0.35 — <0.01 Witch flounder 4.10 — 1 65 0.36 2.03 Red hake 0.03 — — — <0.01 Yellowtail flounder 15.00 11.02 15.00 14.99 19.32 Pollock 0.25 — 0.25 — — Other flounder 17.30 — 6 28 11.81 19.39 Other flounder 0.06 — 0.04 — — Other groundfish 44.88 26.20 34 80 10.34 19.11 Other groundfish 0 10 — 0.10 0.33 1.13 Atlantic herring 2465 23.20 24.65 3576 36.09 Atlantic herring 1.16 1 09 1 16 1.88 1.88 Atlantic mackerel 4.70 4.11 4.70 0.54 1.65 Atlantic mackerel 0.80 0.31 065 008 0.20 Other pelagic 9.52 5.95 9.52 1961 23.40 Other pelagic 9.30 6,71 9.30 265 3.62 Other fish 13.60 8.62 13.60 17 02 27.65 Other fish 1.50 0,37 1.50 — — Squids 5.60 1.04 5.60 0.21 1.67 Squids 24.30 9.89 12.58 13.25 1399 Total 262.37 237.42 221.04 Total 45.35 26.05 20.84 859 ~3iZRY BULLETrS: VOL. 76. NO 4 APfEKTHX Table 3. — Linear TjroGT= T in ICNAF St- =tic6. Actual (Lr .Area 6 of catches to i rs are included also. scares Z -e-r?: --• = : ?: e: --■ = . . .- = ' :s-~ :i"r' s:- :=::' 0.49 ^.^ 0-46 ; -: — I Cv — <0-01 : ;: — : :-9 --2- 0.38 : rf — _ i^ — 1-13 .-: -. 2t :' 35 -iC' 33-05 38.46 --_ ^ . = - ^: 92^ 68.62 78-05 ' z 1'. ' -: 13.20 3.25 8-55 zl ^ 144.37 127.05 0.01 _ <0JJ1 _ _ 0-34 — — — 0.01 Z5Z — 0.01 — 0.12 — C 31 — — <0 C" — <0.01 . _ . — Z-I-i V54 1.54 3 t'z ;.'4 C-'5 — 0-07 0-62 — 0-06 — 0-05 5.71 0^ 1.79 7.09 1.71 1 T1 4.07 4-07 ^ -3^ — 030 — 0.07 : -.1 — 0 09 0-53 '':'.'. 13.00 ■3 :c 5-9C 9-90 22 9>c 15.10 1457 2 iC' _ * ";: _ 2.43 0-05 — 0.05 — 0.01 1 4.4 — 1 44 — 1-37 - - -,- ' — - 71 38 88.88 ' i- — ■ :- — 1-02 D^JC C.",4 5-14 4j0 29-17 45 10 34,46 42.10 37.08 40.95 39-85 47-33 99.91 106.99 : " -; — 10.01 9.52 43.02 3cc 5^ 127 02 313.84 2?:-: 15.54 1246 23.41 - : : — 4.50 0-86 509 21 t- 16.59 20.62 7.07 8.96 43-00 32-77 43.00 17.79 20-59 35 5: "32 34.20 27.42 4293 " r :: 1 : "r c- 69^ 14.25 29-60 iL- 11 • .11 24-65 35.76 36.09 14-22 12-56 14.22 20.15 25.05 19.20 9-86 19^ 17.23 29.32 262-37 245^1 221.04 Tec* 0.70 0-01 0-50 2Xi 0.14 ace -2C 19o0 43C 34.41 --2C OjC 0-06 9.00 751 25-32 2-50 245C 1-91 18-; -2C — 0.42 =-50 ia47 19.15 4-30 — 2-84 5 ~i. 24.:'0 : OC 5 06 14.00 054 O-O8 '.33 130 — 053 — 3.10 — 350 0-54 31.90 3C-53 5631 2'.Zi 424 — 54-17 3 08 £2-63 050 — " " : 7-X — 0 'A — '_- — 0 35 — : V: : .-. - - - 1.16 •-15 1-16 --88 " 55 10-10 4.49 450 273 3-32 25-30 — -I SI 13.24 -3 yz. 860 NOTES EFFECT OF SWIMMING SPEED ON THE EXCESS TEMPERATURES AND ACTIVITIES OF HEART AND RED AND W HITE MUSCLES IN THE MACKEREL, SCOMBER JAPOSICUS Body temperatures of most fish t^-pically are about the same as the water in which they swim for much of the heat generated by muscular activity is ducted away via the circulating blood and lost b\' convection at the gills and body surface. Some scombrids and lamnid sharks consei'\'e muscle heat using countercurrent vascular heat exchangers tretia mirabiliai so that temperatures are maintained significantly above ambient in the brain, eyes, red and white swimming muscles, and viscera (Carey et al. 1971; Stevens and Frj' 1971; Linthicum and Carey 1972; Graham 1973). In other fishes lacking these heat conserving devices, only small temperature excesses above ambient have been recorded, but rarely more than VC (Stevens and Fry 1974). Since heat production must depend primarilj' on work output by the locomotor musculature, we have examined effects of swimming speed on the magnitude of the small temperature excesses in a "cool" scombrid not equipped with the retia exchangers, the mackerel. Scomber Japojjicus (locally the Pacific mackerel = chub mackerel). Another important question concerning scom- brid locomotion is how contractions of red and white muscle fibers are staged as swimming speed increases. It is generally thought that red muscle provides power for cruise swimming and that white muscle functions in "burst" swimming (Rayner and Keenan 1967). Red muscle is pre- dominately aerobic and utilizes fatty acids as the major energy source whereas white muscle (which uses glycogen) usually functions anerobically (Gordon 1968; Bilinski 1974). The second objective of our study was to determine how heart rate and red and white muscle activity of S.japonicus are affected by swimming speed. For this purpose, electrodes were implanted into the pericardial space and in swimming muscles of fish so that simultaneous records of electrocardiograms (ECG's) and red and white electromyographs (EMG's) could be obtained. The genus Scomber is a primitive member of the familv Scombridae (Kishinouve 1923). It has a fusiform shape, is less heavilj' bodied than the skipjack tuna. Katsuwonus pelamis. and other tunas, but shares several characteristics with warm-bodied species; they swim continuously (swim bladders are reduced or absent), have high rates of oxygen consimiption (Baldwin 1923: Hall 1930). and have high blood hemoglobin levels (Greer-Walker and Pull 1975 1. They are also ob- ligatorily dependent upon ram gill ventilation as adults (Roberts 1975) and have large gill siirface areas with a high diffusion efficiency (Hughes 1966; Steen and Berg 1966). Materials and Methods Surgical Procedures and Swimming Experiments The general procedure was to implant either thermocouples or cardiac (ECG) and muscle (EMG) electrodes into mackerel which were then placed in a Blazka-Fry tunnel respirometer ( 12 cm i.d.) to swim at controlled velocities. Fifteen specimens (35-40 cm fork length (FL); 0.38-0.62 kg) were obtained from regularly replenished and maintained mackerel stocks at the Southwest Fisheries Center La Jolla Laboratory. National Maiine Fisheries Service. NOAA. After netting, each fish was anesthetized in a large basin of oxygenated seawater containing 0.2 g 1 of tricaine methanesulfonate (Crescent Research Chemical, Inc.)^ and placed on an operating table where its gills were perfused continuousl.v with a fast flow of oxygenated seawater containing a small amount of the same anesthetic (0.08 gl). Thermocouples (0.127 mm in diameter copper constantan, polyvin3i chloride insulation) or electrode pairs (hooked. 0.07 mm in diameter stainless-steel, epoxy insulated) were implanted within the pericardial cavity just posterior to the ventricle, and in red and white muscles just under the lead- ing edge of the second dorsal fin. The white muscle thermocouple tip was placed midway between the vertebral column and the lateral edge of the body at the level of the horizon- tal midline. Preliminary dissections confirmed that red muscle in S. japonicus occurs in bands 'Reference to trade names does not imply endorsement by the National Marine Fisheries Ser\ice. NOAA. FISHERY BULLETIN; VOL. 76. NO. 4. 1979. 861 that are concentrated below the skin along the lateral midline and become thicker posteriorly (see also Kishinouye 1923, fig. 16; Braekkan 1959, fig. 1). To ensure that the tip of the red muscle thermocouple would remain in place, the wire was passed from near the second dorsal fin obliquely through white muscle and then into the thin red muscle band. Once inserted, its position was easily verified by gentle fingertip probing. To facilitate positioning of the two muscle ther- mocouples, 3-4 cm deep holes were tapped with a 20-gage hypodermic needle. The heart ther- mocouple was passed into the pericardial cavity through a 17-gage needle that was subsequently withdrawn. All wires were anchored in place by skin sutures. Wire leads ( 1 m long) to the recorder were lap wound together, passed posteriorly, and sutured to the dorsal midline near the finlets to prevent tangling around the tail. Implanting re- quired about 15 min after which the fish was transferred to the respirometer swimming tube where aerated water was circulated over the gills by the driving impeller at a slow speed. Two hours recovery from anesthesia and a brief period of swim training was required before a fish could maintain station in the tube and regulate swimming speed in response to water flow. This time delay also allowed stabilization of tissue temperature at ambient conditions following surgery. Adaptation to the swimming chamber was car- ried out at a basal swimming speed which is 1.5 BL/s (body lengths per second) for S. Japan icus (Magnuson 1973). This speed is alsojustabove the velocity required for sustained ram gill ventila- tion(Roberts 1975). Flow rates in the respirometer were calibrated with a ducted flowmeter (Marine Advisors, Inc. model B-7C ) and controlled by alter- ing the applied armature voltage to the impeller pump motor. Eight fish were used for excess tem- perature measurements and seven were used to monitor EMG (4) and EGG (3) patterns. Calibration Procedures Thermocouples were made by soldering to- gether the twisted bared tips of the copper and constantan wires and sealing them with epoxy cement. The three tissue thermocouples and a ref- erence thermocouple (for respirometer water temperature) were each connected in series (con- stantan leads) to an ice-bath reference couple (0°G) and to an RS Beckman Dynograph (copper leads) through a high-quality, shorting rotary-switch. This arrangement permitted rapid switching be- tween thermocouples without opening the recorder circuit. Thermocouples were standardized in a water bath at 20°±0.05°G before and after each trial. Paired electrodes for recording ECG's and EMG's were prepared and implanted (in the same sites used for thermocouples) as described by Roberts (1975). The EGG and EMG signals were preamplified using high impedence, probe am- plifiers (Grass, P511DR) to improve the frequency response of the RS Dynograph. Seawater was kept continuously flowing through the respirometer tube and ambient tem- perature was maintained within 2.0°G in each ex- periment by mixing warm and cold seawater at the outlet taps of the laboratory seawater system. Over the 2-mo course of experiments, respirome- ter temperatures ranged from 16° to 22°G. Results Changes in excess tissue temperatures that ac- company increased swimming speed in the mack- erel are best seen in a particularly successful trial with fish number 6 (Figure 1). Similar, but some- what variable records of heart, and red and white muscle temperatures were obtained for all fish (Table 1). While cruising at low speeds, excess tempera- tures reached a maximum of about 0.3°G in the red and white muscles, but doubled within 3 min swimming at enforced higher speeds (3.2-4.5 BL/s). Excess temperatures recorded in the heart averaged about one-half of the excess developed in muscles at all swimming velocities. When swim- ming speeds were reduced once again to slow cruising, excess temperatures returned to pre- burst levels within 8-15 min. During bouts of prolonged high-speed swim- ming (5-6 min), water in the swimming tunnel was warmed about 1°C due to frictional heating even though a continuous exchange of seawater was maintained from the supply tap (about 15 1/min). This thermal error was minimized by rapidly ac- celerating the fish from slow cruising to its pre- determined, burst-swimming velocity. In Figure 1 for example, the fish was accelerated from 1.4 to 3.9 BL/s in about 5 s followed by sustained swim- ming for 3 min, and then rapidly decelerated to 1.4 BL/s. Equilibration of tissue thermal excess (i.e., generation minus dissipation) occurred in most 862 white muscle FK'iURE 1. — Temperature excess in the heart and in red and white muscles re- corded from Scomber japonicus no. 6(35 cm FL, 0.45 kg) swimming at speeds from 1.4 to 3.9 BL/s. Arrows indicate timing and direction of speed changes. Ambient temperature, 19.5°-19.6°C. r*! I I ' I I 400 600 1000 Table l. — Temperature excesses as AT (°C) recorded for seven Scomber japonicus swimming at basal and moderately fast speeds in body lengths per second (BL/s).' Fish number Item 1 2 3 4 6 7 8 Mean Fork length (cm) 35.6 39.2 389 38 1 350 36 3 34,3 368 Weight (kg) 0,54 0,62 059 0.55 045 0-58 039 0 53 Highest AT at basal speed (1 3-1,9 BUS) in: Red muscle 0.2 0.1 02 02 02 0,3 05 0,24 White muscle 0.2 0.3 03 02 04 03 0.3 0.29 Heart 0.0 0.3 0.1 0.1 0,2 02 {') 0.15 Highest AT and swimming speeds (BLs) in: Red muscle 0.9 0.75 0,55 0,85 03 0,5 0,8 0.66 (4.2) (3.2) (3 7) (42) (39) (38) (43) (3,9) White muscle 0.75 0.8 0,65 06 065 0,3 0,7 064 (4.2) (3.2) (3,7) (3,9) (39) (3,8) (4.3) (3,9) Heart 0.45 {') {') 0,4 0,25 0,25 {') 0.34 (4.2) (4,5) (3,9) (38) (4.1) Maximum trial speed (BL/s) 4.2 3,2 3,7 45 39 38 43 39 Water temperature, range (°C)^ 16.1-17,0 16.5-17.0 16-8-17,8 17,1-17.5 19.5-196 20,5-21,0 21 1-21,8 'Fish no 5 omitted because it would not swim in the respirometer tube, ^Thermocouple malfunction ^Starting temperature is that of the seawater supply from mid-June to mid-July. cases within the 3-min swimming bouts. Although the thermal excess was greater in white muscle of fish number 6 ( Figure 1 ) , mean maximum temper- ature excesses recorded in red and white muscles of the seven mackerel were about the same (Table 1). Variability observed in excess temperature measurements seems attributable to different per- formances of individual fish. Some specimens had more body fat than others and did not swim stead- ily. Others were affected by the trailing ther- mocouple cable as evidenced by their tail-beat pat- terns. The cable also added drag which reduced speed but probably increased total heat production at a specific speed. None of the fish trailing ther- mocouple cables could swim steadily above 5 BL/s, whereas fish trailing the thinner EGG and EMG cables could maintain a speed of 6 BL/s. Some of the variability in recorded thermal excesses may have also been due to the slightly differing loca- tions of thermocouples in each fish. In addition, trauma due to thermocouple insertion, which probably interrupts normal blood flow locally may have been a factor influencing thermal convection. In a few cases, thermocouple signals changed abruptly possibly because of insulation failure at the tip due to rapid body flexing of fishes at higher swimming speeds. A wide range was found in heart rates of mack- erel cruising at 1-1.5 BL/s (mean, 106; range, 80-140 beats/min). With acceleration to 4-5 BL/s, the mean heart rate increased by 549f , (mean, 130; range, 112-150), but rapidly returned to the rest- ing rate within a few minutes of deceleration. 863 2iiri:^sr=L ~wi Ji^ EI H SL ~ Yisutb 2 .. A: sji^wsr -csunf-rnog f^vfii. OBiPsr Ti?^^i^ ^)eerL been T^ fmr -w^~g- TTtygf'^ IPSTE: a".. V— rii.-Tig' — •ar'r "rpn — Tv^ar ly- T-eC Tr~-=.r":T- ■ 2 5£iIT5r^=iC ttt.-t-i r r-^t rr^ yrr.-j- 2d£ 4 3L = . "Whis iinsii& EM^'5 nf :iie fiTs: inaii- irr in; xt xrtr "'^■grt^r skst bi "^ra^i m=- "^tst- smi'C aSDuen: r^ji-nr^ toidiSE fnuL lirse icijsr^s: ^in merr Tries. T£rj£3tQii= :x ±IM3 PTrrp-TTTirio;: gf smEt KjTLH T5S1 ?rrir «*t. i/f irnBC±B^ HI £_ -Wt :TTir--T<£ S»S5E£ ■ ^T^ :i lie: 1^ rsEor^ iar iij£ "n^r -r J^^nr^ Jl ir ~rm= ^ase- ceari sna^ oirrss or :r=jc irnsr;^„ iJin; scirnnxiHiJi nf ijESTciEST^ *rtr;'..-^Tr"'t^ ^t-SUS IE rsi- y^ir "Brrir Tfc^ snaSj" 3ii?w5 t^MT Scr»"n;.r>e' ."! u 15 i ..c 'WHS ziD'. - _- ~~~.fr7r- r-^' - T ^ _-^. ^.- T_&_ sorrresijvenr or ijt cdt 3£n X 1 - TS23aiir "RXJ ~r~ nj TsacKE iuDCfd.. J~rsi- : z ' - . - lOISliSE "ITDrLLli be £r>DJ=d 2£ it •n-iT^-qs v " " ■ : rnr^^ if^^ -^^-^^ -''ill is Aj^ii^ ,}^J}}L U^J^JL :jJaV..V^ 0.x ~, iMA"'*** *M ' : 1 t HI - • ft 1/ . . . i t4in i;,,,,, .:im Ht}'^- ' 1 '^/^if I BL% 2 0 2 5 3J 4>4 50 ^0 5^ 187 5tt vBiin* EJft'S^ S5ft»!t Anttrjeatv 5S 200 $&$ Wis- graerqjpgr rj^r me- ?tr~S IZ r^?fi ' _»-<" «Z- ij^ "^O'JJ"— iTl- jlf^ rr. 5v : : -.-;i^tL his -iix ice^c ^^: ^ itctn -5 i±Ls ii_ ^^ »rea id c^rr dtses -re: Wt -:cxuT r^kAAA?^ ,t»i ••^-*^ ^ - ~- in tttos s£Uki> c^ " ~ - - - -"-ted wiuii tail o^3.is • ery Sio* speed* . -:t 2 BL. s- - rni? " :.cl^ -^— rr.-"—! ^T^r . - . s- rl^^iSt SvCatr litkr .f^r--Tvr ^SK^ K i «u nsc .::.se rsT::*s? ,c 5^ r^ r* r-r tfoe 5»55 For scombrids, which swim continuously and rely upon forward motion to ventilate their gills, the existence of a relatively high speed for the division of labor between red and white muscles, has been assumed primarily on the basis of work done by Rayner and Keenan (1967). These inves- tigators concluded that in the skipjack tuna, red muscle alone powered cruise swimming and white muscle only became active at burst velocities. The initial objective of Rayner and Keenan's study was to demonstrate contractile properties of red mus- cle, and to this end they blocked white muscle activity (pentobarbital) and worked exclusively with tranquilized (propriopromazine) or sedated fish. Moreover, their specimens were restrained in a fixed position and artificially ventilated by per- fusion tubes in the mouth. Thus the movements by these skipjack tunas that were identified as "low frequency swimming," were in fact only casually related to the swimming requirements for gill ventilation and hydrostatic equilibrium; both are controlling factors in normal swimming (Magnu- son 1973; Roberts 1975). Our results with S. japonicus contrast in that they show both red and white muscles function in low-speed swimming. Also, Dizon and Brill (A. E. Dizon, Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, Honolulu, HI 96812. Pers. commun., Sep- tember 1977) recorded red and white EMG's from yellowfin tuna, Thunnus albacares, and found that white muscle activity begins at swimming velocities of <3 BL/s — a speed only slightly above the minimum for hydrostatic equilibrium and well below maximal burst capabilities (Magnuson 1973). These observations indicate that in fast- swimming scombrids, patterned staging of red and white muscle activity may differ in that activity begins in white fibers at very low speeds, and that both red and white muscle remain active through- out a wide range of sustainable speeds as well as at burst velocities. Implicit in this idea is the pre- sence of a high scope for aerobic activity in scom- brid white muscle which has been recently dem- onstrated for the skipjack tuna (Guppy et al. in press). Also required by the hypothesis are specializations in red muscle for high-speed con- traction which is supported by the findings of Johnston and Tota (1974) that high levels of myofibrillar ATPase occur in the red muscle of bluefin tuna, T. thynnus. What physiological advantage might be gained by a 1°C thermal excess during fast swimming? Assuming a Qjo of 2 then a 10% increase in metabolism would afford about a 2-3% rise in swimming speed, but an insignificant change in overall swimming efficiency (Webb 1971). An in- teresting speculation is that the extensive heat- exchanging vascular network used for en- dothermy in the scombrids may have initially evolved to meet the high oxygen requirements of red and white myotomal muscle. More metabolic heat is produced during aerobic respiration and natural selection may have proceeded toward a vascular design that maximized oxygen delivery, yet augmented muscle function by conserving heat and insulating the swimming musculature from ambient conditions. Acknowledgments This work was conducted at the Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service (NMFS), NOAA where John L. Roberts was supported by a NOAA Senior Research Associateship. We thank J. R. Hunter, R. Lasker, and G. D. Sharp for advice and many stimulating discussions. H. T. Hammel of the Scripps Institution of Oceanography advised us on the preparation and use of thermocouples. Q. Bone and J. R. Hunter critically read drafts of this paper and made many suggestions. Useful technical ad- vice and assistance were provided by J. Brown and R. Leong of NMFS. Literature Cited Baldwin, F. M. 1923. Comparative rates of oxygen consumption in marine forms. Proc. Iowa Acad. Sci. 30:173-180. BlLINSKl, E. 1974. Biochemical aspects offish swimming. /nD.C.Ma- lins and J. R. Sargent ( editors), Biochemical and biophysi- cal perspectives in marine biology. Vol. 1, p. 239-288. Academic Press, N.Y. BONE, Q. 1966. On the function of the two types of myotomal muscle fibre in elasmobranch fish. J. Mar. Biol. Assoc. U.K. 46:321-349. 1975. Muscular and energetic aspects of fish swim- ming. In T. Y.-T. Wu, C. J. Brokaw, and C. Brennen (editors), Swimming and flying in nature. Vol. 2, p. 493- 528. Plenum Press, N.Y. BRAEKKAN, O. R. 1959. A comparative study of vitamins in the trunk mus- cles of fishes. Fiskeridir. Skr. Ser. Teknol. Unders. 3(8): 1-42. Carey, F. G., J. M. Teal, J. W. Kanwisher, K. D. Lawson, AND J, S, Beckett, 1971. Warm-bodied fish. Am. Zool. 11:137-145. 866 George, J. C. 1962. A histophysiological study of the red and white mus- cles of the mackerel. Am. Midi. Nat. 68:487-494. GORDON, M. S. 1968. Oxygen consumption of red and white muscles from tuna fishes. Science (Wash., D.C.) 159:87-90. GRAHAM, J. B. 1973. Heat exchange in the black skipjack, and the blood- gas relationship of warm-bodied fishes. Proc. Natl. Acad. Sci. U.S.A. 70:1964-1967. 1975. Heat exchange in the yellowfin tuna, Thunnus alba- cares, and skipjack tuna, Katsuwonus pelamis. and the adaptive significance of elevated body temperatures in scombrid fishes. Fish. Bull., U.S. 73:219-229. Greer Walker, M. 1971. Effect of starvation and exercise on the skeletal muscle fibres of the cod (Gadus morhua L.) and the coal fish iGadus virens L.) respectively. J. Cons. 33:421-427. Greer- Walker, M., and G. A. Pull. 1975. A survey of red and white muscle in marine fish. J. Fish Biol. 7:295-300. Guppy, M., W. C. Hulbert, and p. W. Hochachka. In press. The tuna power plant and furnace. In G. D. Sharp and A. E. Dizon (editors). The physiological ecology of tunas. Academic Press, N.Y. Hall, F. G. 1930. The ability of the common mackerel and certain other marine fishes to remove dissolved oxygen from sea water. Am. J. Physiol. 93:417-421. HUDSON, R. C. L. 1973. On the function of the white muscles in teleosts at intermediate swimming speeds. J. Exp. Biol. 58:509- 522. HUGHES, G. M. 1966. The dimensions offish gills in relation to their func- tion. J. Exp. Biol. 45:177-195. JOHNSTON, I. A., AND B. TOTA. 1974. Myofibrillar ATPase in the various red and white trunk muscles of the tunny iThunnus thynnus L.) and the tub gurnard iTrigla lucerna L.). Comp. Biochem. Physiol. 49B:367-373. JOHNSTON, I. A., W. Davison, and G. Goldspink. 1977. Energy metabolism of carp swimming muscles. J. Comp. Physiol. 114:203-216. KISHINOUYE, K. 1923. Contributions to the comparative study of the so- called scombroid fishes. J. Coll. Agric, Imp. Univ. Tokyo 8:293-475. LINDSEY, C. C. 1968. Temperatures of red and white muscle in recently caught marlin and other large tropical fish. J. Fish. Res. Board Can. 25:1987-1992. LINTHICUM, D. S., AND F. G. CAREY. 1972. Regulation of brain and eye temperatures by the bluefin tuna. Comp. Biochem. Physiol. 43A:425-433. MAGNUSON, J. J. 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. Roberts, J. L. 1975. Active branchial and ram gill ventilation in fishes. Biol. Bull. (Woods Hole) 148:85-105. STEEN, J. B., AND T. BERG. 1966. The gills of two species of haemoglobin-free fishes compared to those of other teleosts — with a note on severe anaemia in an eel. Comp. Biochem. Physiol. 18:517-526. 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. 1974. Heat transfer and body temperatures in non- thermoregulatory teleosts. Can. J. Zool. 52:1137-1143. STEVENS, E. D., AND A. M. SUTTERLIN. 1976. Heat transfer between fish and ambient water. J. Exp. Biol. 65:131-145. WEBB, P. W. 1971. The swimming energetics of trout. II. Oxygen con- sumption and swimming efficiency. J. Exp. Biol. 55:521-540. JOHN L. ROBERTS Department of Zoology University of Massachusetts Amherst, MA 01003 Department of Zoology San Diego State University San Diego, CA 92182 JEFFREY B. Graham THERMAL BEHAVIORAL RESPONSES OF THE SPECKLED SANDDAB, CITHARICHTHYS STIGMAEUS: LABORATORY AND FIELD INVESTIGATIONS The speckled sanddab, Citharichthys stigmaeus, is a small bothid flatfish that is common in southern CaliforniaiFord 1965; Stephens etal. 1974). These authors and Helly' have suggested that tempera- ture may have a significant effect on localized population abundances and distributions of speck- led sanddabs. No studies to date, however, have examined in detail the relationship between temperature and fish behavior and distribution. We designed this work to study the speckled sanddab population in King Harbor, Redondo Beach, Calif. This harbor (Figure 1), which re- ceives the thermal effluent from an electricity generating station as well as cold upwelled water from the adjacent Redondo Submarine Canyon, contains a highly diversified thermal environ- ment (Stephens 1972). 'Helly, J. J., Jr. 1974. The effects of temperature and tem- perature selection on the seasonality of the bothid flatfish, Citharichthys stigmaeus. Honors Thesis, Occidental Coll., Los Ang., 34 p. FKSHERY BULLETIN: VOL. 76. NO 4. 1979. 867 Figure l. — Location of field sampling stations for speckled sanddabs in King Harbor, Redondo Beach, Calif PACIFIC OCEAN 300 m Methods We collected adult speckled sanddabs between August 1975 and January 1976 with a 3-m otter trawl. The fish were transported to the laboratory in aerated seawater and acclimated to a range of normally occurring temperatures (10.0°-19.7°C) according to the methods of Ehrlich et al. (1979). Prior to acclimating the speckled sanddabs, we removed the gill isopod Liuonica vulgaris indi- vidually with forceps. During holding and accli- mation, we fed the fish to satiation daily with live and frozen Artemia salina. The behavioral re- sponses of the fish to temperature were studied using a 3.6-m long horizontal gradient and employing the techniques of Ehrlich et al. (1978). Each experiment lasted for 7-8 h with observa- tions every 15 min. We shifted isotherm positions during each experiment to separate selection of temperature from preference for a given position within the experimental chamber. Speckled sanddab abundance and distribution were studied using timed diver transects at six stations (Figure 1). Two divers swimming side by side for 5 min traversed each 6-m wide transect. They recorded the species and number of indi- viduals observed in the same area. The transects by each pair of divers were run in duplicate on a monthly basis at each station from September 1974 through February 1976 and quarterly there- after. In the analyses, we used the largest number of individual fish counted by either diver, but the average count of the two independent observa- tions was used for estimates of large groups of fishes. The divers recorded the temperature at least twice during each transect, with thermome- ters readable to 0.5°C. Results and Discussion We examined the effects of acclimation temper- ature, size, and sex of speckled sanddabs on their temperature selection during 11 experiments (Ta- ble 1). The presence of some skewed temperature- specific frequency distributions (Table 1) pre- cluded comparison of the results with parametric statistics. We tested these distributions for homo- geneity using a Kruskal-Wallis test (Steel and Tor- TABLE 1. — Temperatures selected by speckled sanddabs in laboratory experiments. No. test No fish Standard lengtti (mm) Acclimation temperature Selected temperature (°C) Coefficent of skGwness Coefficient of kurlosis Date animals observations Mean SD Sex Mean SD Mode (5,) (92) 21 Aug. 1975 9 265 963 06 not noted 14.0 10.5 3.4 10 0.493- 2.788 15 Dec 8 230 91.5 1.3 not noted 10.0 123 4.5 9 0.543- 2.824 18 Dec. 6 218 88.5 3.2 M 19.7 10.5 4.1 8 0673- 2.928 22 Dec. 9 220 903 3.5 F 18.9 10 1 26 9-10 0427 3.438 5 Jan. 1976 9 210 77.1 4,0 M 15.2 109 2.6 11 0,220 3.512 8 Jan. 9 251 82.0 6.7 F 15.2 10.4 3.2 8-9 0,465 3.036 9 Jan. 9 250 82.0 67 F 15.2 11.6 3.6 8 0.573- 2,673 26 Jan. 6 162 763 5,6 M 12.0 9.9 4.3 8 0400 0.592- 27 Jan. 6 167 71.5 27 M 12.0 10.5 2,5 9-10 0.291 2.372 29 Jan. 6 157 73.2 2.6 M 120 11,1 30 9 0-724- 2.538 30 Jan. 6 160 75.5 5.1 M 12,0 98 2.5 8 0.422 2.078 P<0.05. 868 rie 1960) and detected no significant differences (X^odf = 15.99, P>0.05). The overall mean selected temperature from pooled data was 10.8°C (SD = 3.1°C), and the mode was 9°C; 709^ of all occurrences were in the range of 8°-13°C (Figure 2). The frequency distribution, however, was sig- nificantly skewed towards warmer temperatures (Figure 2). Brett (1971) showed that the preferred tempera- ture coincided with the optimal temperature for growth of sockeye salmon, Oncorhynchus nerka. Crawshaw (1977) found that physiological re- sponses are often optimized in a zone of efficient operation rather than at a peak, and temperature preference reflects this. We are not aware of any work on the effects of temperature on the physiol- ogy of speckled sanddabs. Considering the work of Brett (1971) and Craw- shaw ( 1977), it is not unreasonable to suspect that speckled sanddabs may have a range of tempera- tures (approximately 8°-13°C) for efficient growth. The skewness may partially be due to activity increasing with temperature that could have re- sulted in occasional excursions into a greater number of compartments (temperatures) than at colder temperatures. Ehrlich et al. (1978) also suggested that skewness of a temperature-specific frequency distribution could result from preferred temperatures approaching lethal limits. These limits are not known for speckled sanddabs. De- Witt (1967) suggested that skewness of distribu- tion could result from the regulation of body tem- 15-1 (ij o z UJ (T lo- ir O o o o 5- ■f-f T — I — I — I T T T T T T — r — I — 7~~l — I — I — I — I 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 202! TEMPERATURE CO Figure 2. — Temperature-specific occurrences of speckled sanddabs, based on pooled data from 2,290 fish observations. The frequency distribution was significantly skewed toward warmer temperatures (^, = 0.571, fgsdf = 2.33, 0.0180% of the total, a finding consistent with other mercury values reported (Kamps et al. 1972; Westoo 1973). In muscle and liver tissues of blue marlin, Makaira nigricans Lacepede, how- ever, only a small portion of the total mercury was found to be organic mercury. Additional studies on marlin landed during fishing tournaments in 1972 (Schultz et al. 1976) and 1973 (Schultz and Crear 1976) revealed low levels of organic mercury in six other tissues. These studies also showed that the difference between total and organic mercury was indeed inorganic mercury. G. Westoo (National Swedish Food Administration, Stockholm. Pers. commun., 1972) had previously identified the or- ganic fraction as methyl mercury. An assessment of mercury is complicated by the presence of selenium. Selenium has been shown to reduce the toxicity of mercuric chloride and methyl mercury in laboratory animals when given as selenite, selenomethionine, or as selenium pre- sent in tuna (Pah'zek et al. 1971; Ganther and Sunde 1974). The presence of selenium in tuna, a principal food item of marlin (NaughtonM, indi- cates that it should also be present in marlin. For this report, nine tissues from blue marlin were analyzed for selenium, total mercury, and organic mercury. Materials and Methods Samples of muscle, liver, kidney, spleen, pyloric caecum, stomach, gill, gonad, and blood were col- lected from 46 marlin landed during a fishing tournament in Kailua-Kona, Hawaii, during Au- gust 1974. The tissues were ground with Dry Ice^ in a blender and stored in acid-washed plastic vials. The organic extraction was carried out as de- scribed by Rivers et al. (1972), i.e., a benzene ex- traction of the methyl mercury was reextracted with cysteine, oxidized with permanganate, and reduced to elemental mercury with stannous ion prior to being volatilized into the flameless atomic absorption apparatus. Total mercury digestions were performed (Rivers et al. 1972) but with 10 ml of concentrated nitric acid instead of 30 ml. All analyses were made with a Perkin-Elmer 303 atomic absorption spectrophotometer equipped with a vapor chamber (Manning 1970). Selenium was determined by a fluorometric technique (Watkinson 1966), as modified by S. Nishigake (Tokyo Metropolitan Research Labo- ratory of Public Health, Tokyo, Japan. Pers. com- mun., 1975), i.e., following sample digestion with nitric and perchloric acids, the selenium was com- plexed with 2,3-diaminonaphthalene and this fluorescent compound then extracted into cyclo- hexane. All analyses were made using a Turner Model 110 fluorometer equipped with a primary filter at 369 nm and a secondary filter at 522 nm. 'Naughton, J. J. 1973. To all billfishermen. (Summary report of 15th Hawaiian International Billfish Tournament, 27-31 Au- gust 1973), 9 p. Southwest Fisheries Center Honolulu Laborato- ry, National Marine Fisheries Service, NOAA, Honolulu, HI 96812. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 872 FISHERY BULLETIN: VOL. 76, NO. 4, 1979. Results A summary of mercury and selenium concentra- tions in the marlin is given in Table 1. Average total mercury and selenium values were greatest in kidney (26.33 mg/kg Hg, 23.42 mg/kg Se) and least in blood (0.18 mg/kg Hg, 1.29 mg/kg Se) and gill (0.32 mg/kg Hg, 1.29 mg/kg Se). Average methyl mercury was highest in muscle (0.40) mg/kg) and lowest in blood (0.04 mg/kg) and gill (0.06 mg/kg). The percentage of organic to total mercury ranged from 19f in kidney to 279f in gonad. The molar ratio of mercury to selenium ranged from 0.06 in blood to 0.62 in muscle. (Molar ratio is computed here using sample average statistics of combined male and female data.) An analysis of variance revealed significant dif- ferences (P<0.05) in total mercury between males and females for all tissues except gill and blood. A similar pattern was found for selenium. The or- ganic mercury levels were not statistically differ- ent (P>0.05) between the sexes. In earlier studies on marlin caught from the same area during fishing tournaments in 1971, 1972, and 1973, mer- cury concentrations in both sexes were found to be similar. Table 2 presents correlation coefficients for body weight and tissues. In most cases, the relation- ships are positive and highly significant. Figures 1-6 illustrate the dependence of mercury and selenium on weight and on each other. Discussion The data clearly demonstrate that methyl mer- cury concentrations are low relative to total mer- cury in blue marlin. Westoo (pers. commun.) and Nishigaki (pers. commun.) confirmed this low percentage in our samples based on subsamples sent to them. Nishigaki has also confirmed our selenium results. In a study of 37 Pacific blue marlin (50-238 kg, average 109 kg) from Japanese waters, Nishigaki (pers. commun.) found total mercury levels rang- ing from 0.02 to 13.0 mg/kg (average 2.83 mg/kg) and methyl mercury ranging from 0.02 to 1.28 mg/kg (average 0.57 mg/kg). Selenium values for 11 marlin ranged from 0.52 to 1.99 mg/kg, averag- ing 0.97 mg/kg. These values are similar to our findings for mercury and selenium in marlin from Hawaiian waters. Table l. — Summary of mercury and selenium data in 46 blue marlin from the Hawaiian Islands by sex.' Total Hg (mg kg) Organic Hg (mg, kg) o„ ^.^a^.r Hg Se (mg/kg) Hg/Se Mean Range Mean Range of total Hg Mean Range ratio Tissue M&F M F M&F M&F M F M&F M&F M&F M F M&F M&F Muscle Liver Kidney Spleen Stomach F>ylonc caecum Gill Gonad Blood 3.12 11.58 26 33 693 1.27 1 83 0.32 0.40 0.18 2.83 12.53 2625 6.34 1 33 1.74 0.26 030 0.18 3 76 9.46 26.53 821 1,16 2.03 0.44 0.68 0.18 0.09-10 00 0.13-39 20 0.18-77.00 0 08-17 60 0.06-3.00 0.17-4.50 008-096 0.03-2.15 0.02-0.53 040 026 0 22 0 18 0.12 030 006 0.11 0.04 032 0 21 0 16 0 12 0.08 022 003 006 004 0.58 038 0.35 030 021 0.49 0.11 025 0.05 0.02-1.02 009-0.76 004-086 002-0.66 003-0.47 008-0.98 0.01-0.34 0.03-0.65 ;0.01-0.11 13 2 1 3 9 16 19 27 22 198 1.88 2.21 0.63-5.32 17 47 20 36 1106 2 50-61 12 23.42 2533 1906 2 63-56 25 9.31 912 973 0 63-24 25 291 316 239 135-4 03 4 92 1.29 2.14 1.29 5.10 1.31 1.97 1.33 4 52 1.26 2.60 1.19 2.28-10.10 0.71-2.21 1 .24-3.80 0.72-2.30 0.62 026 0.44 0.29 0.17 0.15 0.10 0.07 0.06 'Weights of 32 males ranged from 58 to 112 kg (average 80); 14 females weighed 39. 75, and 115-342 kg (average 166) Average for all samples was 106 kg. Table 2. — Correlation coefficient of mercury, selenium, and weight in 46 blue marlin from the Hawaiian Islands by sex. Total Hg/wt Organic Hg/wt Organic total Hg Sewt Total Hg/Se Organic Hg/Se Tissue M&F M F M&F M&F M&F M F M&F M&F Muscle 069" 090" 083" 0.79" 072" 068" 0.88" 0.85" 0.93" 061" Liver 023 076" 069" 0.76" 0 13 0.00 0.68" 0.38 0.80" -008 Kidney 0.49" 083" 079" 087" 042" 032- 0,79" 0.73" 0 91" 0.30- Spleen 06r- 088" 075" 0,84" 0.51" 0.45" 072" 0.61- 0.87" 0.41" Stomach 0.38- 0.77" 082" 0.81" 0.35- -0.27 0 18 0.15 0.32- -033- Pylonc caecum 0.59" 083" 082" 086" 059" -0.11 032 -0.10 025 -0.14 Gill 086" 070" 090" 0.88" 0.77" 004 0.08 016 021 -0.14 Gonad 085" 0.75" 0.95" 082" 0.76" 015 -0.10 -0.53 010 0.25 Blood 0.28 0.32 057- 0.36- 0.44" -004 -0.17 049 0.44" 009 •P<0.05. ••P<0.01. 873 12 >- q: O II - 10 o«-^rs"oi'--i'»!r"' 150 200 WEIGHT (kg) 350 3 50 100 150 200 WEIGHT (kg) 250 350 874 12 Fk;L'RK 1— The relationship of total and organic mercury in muscle tissue of 46 blue marlm from the Hawaiian Is- lands to fish weight. 10 FIGURE 3— Relationship be- tween total mercury and selenium in muscle tissue of 46 blue marlin from the Hawaiian Islands. a: 3 O ir UJ g a FEMALES X MALES 2 3 4 SELENIUM mg/kg (ppm) FIGURE 2.— Relationship be- tween selenium in muscle tis- sue and weight of 46 blue marlin from the Hawaiian Is- lands. FIGURE 4.— Relationship be- tween organic mercury and selenium in muscle tissue of 46 blue marlin from the Hawaiian Islands. 1.2 II 10 i 8 ir O (T UJ S (J z < l PCB. (Incompletely washed glassware and new batches of reagents are com- mon sources of high blanks.) Recovery Carry out the complete procedure with solutions of standards of known concentration in the range anticipated for the samples to assure quantitative recovery. (Some loss of chlorinated hydrocarbons always occurs in the absence of proteins and lipids, which act as keepers. A minimum recovery of 80^f is essential. Doping samples originally containing very low levels of chlorinated hydrocarbons gives results which better reflect the accuracy of the method: 859'f or higher recovery of DDE, TDE, DDT, and PCB, and 80-85'7r recovery of dieldrin and endrin.) Extraction 1) Weigh approximately 10 g of the material (see Procedure Variations for exceptions), to be analyzed into a Virtis flask and record exact weight to the desired degree of accuracy. 2) Add 20 ml of a 1:1 IPA/benzene mixture to the flask. 3) Homogenize at about 23,000 rpm for 5 min. 4 ) Rinse homogenizer blade and Teflon cap with hexane so that the hexane drips into the Virtis flask. Fill the flask with hexane nearly to the bot- tom of the flask neck. 5) Place the Virtis flask in a hot-water bath (ca. 85°C) or sand bath. (An electric fry pan with a layer of sand covered with water provides an economical heating bath.) Boil moderately for at least 45 min, adding hexane whenever the level falls to about one-third of the flask capacity. (If the solution boils too rapidly, material will be lost in the spray. The rate of boiling must be adequate to distill all the H^O, IPA, and benzene from the flask, because they interfere with the cleanup.) When adding hexane, do it so as to rinse down the sides of the flask as well. Keep the water level in the bath and the hexane level in the flask adjusted so that the flask does not become buoyant and tip over. After 45 min of boildown, reduce the vol- ume of the solution to about 20 ml (ca. 1 cm from the bottom of the flask). Cool. If a layer of water separates, add Na.^ SO^ and allow to stand 1-3 min. 6) Filter through a funnel plugged with glass wool into a 50-ml graduated centrifuge tube. Rinse the flask with three or four 5-ml portions of hexane and pour the rinse solutions through the filter into the centrifuge tube. 7) Concentrate the extract in a hot- water bath to desired volume (20-25 ml), record volume, and pour most of extract into a 23-ml borosilicate screw-cap bottle (with Teflon-lined cap) contain- ing 1 g Na^SOj. (The extract can be stored in this condition for extended periods.) (If the extract, prior to concentration, still contains traces of H2O or IPA, as indicated by cloudiness, add hexane while concentrating in order to remove the H2O or IPA.) Oil Determination 8) Pipette exactly 1 ml of the extract into a tared aluminum weighing dish and allow it to 882 evaporate 4-6 h at room temperature to minimum weight. (Since marine oils oxidize, the weight of'oil begins to rise again after a few hours.) Weigh the residue, which contains the oil in 1 ml of extract. (,"lcaiuip 9) Prepare.' a Flurisil (.olumn by filling a 9 mm i.d. X 150 mm glass tube, plugged with glass wool, with ca. 5 cm of Florisil. Wash the column with at least 15 ml of hexane added 1 ml at a time. Allow the hexane level to drop to 1-2 mm, but not to dryness. Pipette 1 ml of the extract onto the col- umn. For samples with very high oil content (refer to step 8 for the amount of oil in 1 ml of the ex- tract), adjust the volume placed on the Florisil so that no more than 0.1 g, and preferably no more than 0.08 g, of oil is placed on the Florisil. Elute with 1-ml portions of hexane; collect the first 12-13 ml of the eluate in a 13-ml graduated centrifuge tube. I Note: Once the Florisil has been wetted, it must always have solvent above it.) 10) If DDT and PCB are not going to be sepa- rated, concentrate (in a tube heater) the eluate to an appropriate volume for gas-liquid chromato- graphic analysis."* If separating DDT and PCB. evaporate the eluate to slightly less than 1 ml. Separation of DDE, TDE, and DDT from PCB Quantitation of TDE and DDT is often difficult and quantitation of the PCB is usually impossible unless the DDT family is separated from the PCB. Separation is achieved by chromatography on silica gel. The behavior of DDT and PCB during solid-liquid chromatography is very similar, and obtaining optimal separation requires careful con- trol of all the parameters of the procedure. Even so, DDE does not separate entirely from PCB. Therefore, the DDE in the PCB fraction is quanti- tated'' and included with that in the DDT fraction. Evaluate the degree of separation of DDE, TDE, and DDT from PCB by chromatographing stan- dard solutions of these compounds according to the procedure described below. Adjust the time and "•For a detailed de.scription of gas chromatography of chlori- nated hydrocarbon pollutants, see the Pesticide Analytical Manual il977), available from Management Methods Branch, DMS, ACA, HFA-250, 5600 Fishers Lane, Rockville. MD 20857. or National Technical Information Service iNTIS), Springfield. Va. The manual provides extensive background on residue analysis. •'^Although DDE elutes from the gas-liquid chromatograph at the same time as one of the PCB peaks, measurement of the other five intense PCB peaks provides accurate quantitation of PCB. temperature of activation, the degree of rehydra- tion, the amount of sil ica gel, and the volume of the pentane fraction to obtain the optimum separation of DDT and TDE from PCB: that is, to maximize the amount of PCB in the pentane fraction and the amount of DDT and TDE in the benzene fraction. 1 1 ) Activate the silica gel by heating at 215°C for 16 h. (The time and the temperature are ad- justed to obtain an arbitrarily, but consistently activated product with suitable separating characteristics, since complete dehydration occurs over a long period of time.) Cool to room tempera- ture in a desiccator. Rehydrate by placing 98 g silica gel in a glass-stoppered bottle and adding 2 g distilled water. Stopper the bottle and shake and tumble until the water is evenly distributed. Allow the silica gel to equilibrate for 2-4 h before use. 12) Place a portion of this prepared silica gel in a beaker and cover with pentane. Let stand 5-10 min to return to room temperature. (Because the chromatographic columns do not contain stop- cocks, a special technique is required for packing.) Quickly transfer silica gel to a glass-wool- stoppered column, 9 mm i.d. x 250 mm long, wet with pentane. (A disposable transfer pipette with the narrow part of the tip removed works quite well for transferring the silica gel slurry. ) Tap the column gently to facilitate packing. Make sure that there is always enough pentane above the column to allow the silica gel to settle slowly in order to eliminate air bubbles and prevent the top of the column from running dry. Pack the column to a height of ca. 8 cm. (Throughout the whole separation procedure the silica gel must always have .solvent above it and must be free of bubbles and cracks, which interfere with the desired sep- aration. If the column runs dry or cracks, discard it.) Rinse the column with 15-20 ml of pentane. 13) Allow the pentane level to descend to 1-2 mm (not dry) and place the Florisil eluate (ca. 1 ml) on the column with a Pasteur capillary pipette. Rinse the Florisil eluate tube with three or four 1-ml portions of pentane, and transfer each rinse successively to the column. After the sample and rinses have been adsorbed, fill the tube with Use the amount of PCB in those peaks to determine the size of the peak overlapping the DDE and correct the apparent total DDE I actually DDE plus PCB) to obtain the tme DDE concentration. The electron-capture detector is so much more sensitive to DDE than PCB, that the correction affects the accuracy of DDE de- temiination only to a small extent. Consequently the variability in amount of DDE in the pentane fraction does not markedly affect the accuracy of DDE analysis. 883 pentane. Collect 42 ml of pentaneeluateina 50-ml graduated centrifuge tube. (This fraction contains PCB and some DDE.) 14) After the appropriate volume of pentane eluate has been collected, place a second 50-ml graduated centrifuge tube under the silica gel col- umn. Then fill the tube with benzene. Collect 35 ml of benzene eluate. (This fraction contains most of the DDT complex.) (DDE elutes very rapidly with benzene. If benzene is added to the column before the second centrifuge tube is in place, the DDT complex will often be found in the pentane fraction.) 15) Concentrate each fraction in a boiling water bath to less than the desired final volume and quantitatively transfer with hexane to a vol- umetric flask of the desired final volume. (The 50-ml centrifuge tubes are not very accurate vol- umetiic containers.) Proceed with gas-lic|uid chi-omatographic analysis (see footnote 4). Notes on DDT/PCB Separation Proccdiirc 1) Elution with hexane instead of pentane dur- ing the silica gel chromatography fails to provide the necessary separation of DDT from PCB. Hexane is reported to contain variable amounts of benzene, which would obviously affect an already delicate separation. Use of UV-quality pentane or hexane has been recommended by others, and might allow use of hexane in warm weather. 2) For high residue level samples, evaporation of the Florisil eluate to ca. 1 ml is not necessary; instead an appropriate aliquot is used. However, no more than 1 ml of the eluate should be placed on the silica gel column because the hexane may con- tain benzene. Procediirt' Variations During the 45-min boildown, scrape the mate- rial on the bottom of the flask. Pile up the solids to leave areas of the flask bottom in direct contact with the solvent to improve boiling action and prevent bumping. Cool. If water separates, add Na.SO,. Filter through glass wool and proceed as usual (step 6 under Extraction). In oi'der to compensate for the low residue level usually found in plankton, place 2 ml of the extract on the Floi'isil column for cleanup. I'islimcals or Dr\ leeds If the standard extraction procedure is used for meals and animal feeds, the finely ground meal forms a layer on the bottom of the Virtis flask which causes bumping and loss of solvent during the 45-min boildown. Extraction with hexane pro- vides as good recovery as IPA/benzene. This sub- stitution allows omission of the boildown. 1) Homogenize sample with 20 ml of hexane. Wash down Virtis blades and Teflon top with min- imal amount of hexane. Add 10 g Na^SO^. 2) Immediately filter through glass wool tightly wadded to remove as much of the solids as possible. Wash flask and funnel with a minimal amount of hexane so that the volume of the cen- trifuge tube is not exceeded (35-40 ml). 3) Stir to mix and centrifuge at 1,500 rpm for 45-60 min at 10°C. (There should be about 35 ml of clear solution with less than 1 ml of solids.) 4) Record volume, subtracting 50% of the vol- ume occupied by the solids. (Although this in- volves an approximation, the error involved should be no more than 2^/c.) 5) Decant the supernatant liquid into a storage bottle containing 1 g Na.,S04. Proceed with Florisil cleanup (step 9 under Cleanup). Plankton and Other High-Moistiire Samples Plankton do not homogenize well using the standard procedure. They also contain water in excess of the amount that 20 ml of 1: 1 IPA/benzene can absorb. Addition of Na.2S04 before homogeni- zation overcomes both difficulties. Na2S04 not only absorbs water, but also serves as a grinding aid. To the weighed sample of plankton add 25 ml of 1 : 1 IPA benzene; then add 25 ml hexane and 10-15 g Na^SOj. Homogenize 5 min and proceed as usual. Fish Oil Homogenization and extraction of oil are un- necessary. 1) On an analytical balance weigh accurately about 2 g of oil into a 50-ml graduated centrifuge tube. 2) Dilute to about 20 ml with hexane and swirl to dissolve the oil completely. 3) Record the volume. 4) Place in a storage bottle, containing 1 g Na., SO4 , and proceed with the usual cleanup ( step 9 under Cleanup). 884 NOTE: For greater accuracy in oil analysis, weigh accurately about 2.0-2.5 g oil into a 25-ml vol- umetric flask. Dilute to volume with hexane. Shake thoroughly. Place sample in storage bottle and proceed at step 9 under Cleanup. Paper This procedure is included because occasionally fishery samples come in contact with contami- nated packaging and labelling materials, such as carbonless carbon paper and cardboard. Although the procedure has not been validated by collabora- tive studies, it provides guidelines for an analysis relevant to fishery studies. Cut the paper (or cardboard* into small pieces, approximately 1 cm on a side, with a scissors or office paper cutter, which has been cleaned thoroughly with iso-octane or hexane. Mix the paper thoroughly and weigh ca. 7 g into a Virtis flask. Add 15 ml distilled water and mix thoroughly with the paper. Allow the mixture to stand 2-5 min, stir again, and dilute with portions of 1:1 IPA/benzene to a total of 70 ml. Homogenize the mixture briefly at low speed. Push the paper down with the Virtis blades, then homogenize briefly. Repeat the process until the paper is com- pletely homogeneous, approximately 10 min total homogenization time. Follow the usual boildown and Florisil procedures. Proccdiiri.' For Dieldrin and Kndrin Saponification'' and Extraction 1) Weigh 10 g of material to be analyzed into a 250-ml Erlenmeyer flask. For oil, use only 2 g. 2) Dissolve 10 g of KOH in 6 ml distilled water. Slowly dilute with 34 ml of ethyl alcohol (95 or 100%). Swirl until clear. 3) Pour the alcoholic KOH over the sample and heat in a water bath without boiling for 20 min; the exact temperature is not critical. 4) Allow the mixture to cool. Pour the liquid portion into a 250-ml separatory funnel and rinse out the Erlenmeyer flask with 50 ml of water, divided into 4 or 5 portions. Avoid pouring any solids into the separatory funnel. For finely pow- dered samples like meal, filter the sample through ^Saponification in the pre.sence of some proteinaceous mate- rials has been reported to cause degradation of dieldrin. As stated above, recovery studies are always important. a wad of glass wool and rinse the glass wool care- fully with each rinse. 5) Add 15 ml hexane to the separatory funnel and shake for 2 min. Open the stopcock several times during shaking to relieve the pressure buildup. Allow the layers to separate completely, usually about 30 min. 6) Drain off the aqueous layer into the Erlen- meyer flask from which it came. Pour the hexane layer into a 30-ml beaker. Do not let any water escape into the beaker. Cover the beaker tightly with aluminum foil. 7) Pour the aqueous layer back into the separatory funnel and repeat step 5. 8) Drain off the aqueous layer and discard it. 9) Return the hexane extract in the beaker to the separatory funnel. Rinse the sides of the beaker with 1 ml hexane. Add the hexane wash to the extract in the separatory funnel. 10) Wash the hexane extract with 10 ml water by rotating the separatory funnel gently to avoid emulsion formation. Do not shake. Allow the layers to separate and discard the aqueous layer. 11) Pour the hexane layer into a 50-ml graduated cylinder. Do not transfer any water to the cylinder. Record the volume. Pour the extract into a 23-ml borosilicate screw-cap bottle (with Teflon-lined cap) containing 1 g NagSO^. Cleanup 12) Prepare a Florisil column by filling a 9 mm i.d. X 150 mm glass tube, plugged with glass wool, with ca. 4 cm of Florisil. Or use a 7 mm i.d. x 150 mm tube containing ca. 5 cm of Florisil. (The longer column of adsorbent may give slightly bet- ter cleanup.) Wet the Florisil with benzene and wash it with 4 to 5 ml benzene added 1 ml at a time. Wash it next with 10 ml hexane (or more) added 1 ml at a time. Pipette 2 ml of the hexane extract onto the column. Put a 12-ml graduated centrifuge tube under the column. Flute with hexane added 1 ml at a time. Collect the first 12 ml of hexane eluate, which contains DDMU (the dehydro- chlorination product of TDE), DDE, and PCB. (They may be quantitated if desired.) Place a second 12-ml centrifuge tube under the column. Change eluant to benzene; add it 1 ml at a time. Collect the first 10 ml of eluate. Place a modified micro Snyder column on the centrifuge tube to prevent loss of residues during evapora- tion. Concentrate the eluate containing dieldrin and endrin to an appropriate volume (1-3 ml) for 885 gas-liquid chromatographic analysis (see footnote 4). Ac know Icclgments We thank the following laboratories for helping to establish the validity of these methods by par- ticipating in verification studies: Seattle, Detroit, and New Orleans District Laboratories of the U.S. Food and Drug Administration; the Gulf Breeze, Fla., Field Station of the Environmental Protec- tion Agency; the Wisconsin Alumni Research Fund Institute, Madison, Wis.; and the Iowa De- partment of Agriculture, State Chemical Labora- tory, Des Moines. Robert Reinert willingly shared his methods with us before they were published. Daniel B. Menzel, formerly of the Institute of Marine Re- sources, University of California, Berkeley, Calif., generously shared his knowledge of the intricacies of chlorinated hydrocarbon analysis. Laura G. Lewis provided technical support in developing this procedure. Literature Cited Horowitz, W. (editor). 1970. Fat-containing foods. /n Official methods of analysis of the Association of Official Analytical Chemists. 11th ed., p. 480. Assoc. OfT. Anal. Chem., Wash., D.C. Pksticide Analytical Manual. 1977. Gas-liquid chromatography. In Pesticide analytical manual. Vol. I, Chapter 3. U.S. Dep. Health, Educ, and Welfare, Food and Drug Admin., Wash., D.C. PORTKR, M. L., S. J. V. Young, and J. A. Burke 1970. A method for the analysis offish, animal, and poul- try tissue for chlorinated pesticide residues. J. Assoc. Off. Anal. Chem. 53:1300-1303. Reinert. R. e. 1970. Pesticide concentrations in Great Lakes fish. Pestic. Monit. J. 3:233-240. Snyder. D., and R. reinert 1971. Rapid separation of polychlorinated biphenyls from DDT and its analogues on silica gel. Bull. Environ. Con- tam. Toxicol. 6:385-390. Virginia F. Stout Northwest and Alaska Fisheries Center National Marine Fisheries Service, NOAA 2725 Mont lake Boulevard East Seattle. WA 98112 F. Lee Beezhold Northwest and Alaska Fisheries Center Present address: Food Chemical and Research Laboratories. Inc. 4900-9th N.W., Seattle, WA 98107 GROWTH OF JUVENILE SPOT PRAWN, PANDALUS PLATYCEROS, IN THE LABORATORY AND IN NET PENS USING DIFFERENT DIETS Floating net pens have been used to culture Pacific salmon, genus Oncorhynchu.'^. in the marine wa- ters of the West Coast since 1969 ( Mahnken 1975). Although it has been a monoculture effort to date, use of a companion crop species such as the spot prawn, Pandalus platyceros Brandt, could diver- sify and enhance this industry. In 1975 the National Marine Fisheries Service selected the spot prawn to examine as a potential companion species to net pen-reared salmon. The spot prawn was selected as a candidate for several reasons; 1 ) it has a rapid growth rate and large size compared with other pandalids (Butler 1964); 2) it can be successfully cultured to maturity in captiv- ity (Prentice 1975); 3) it will reproduce in captiv- ity, often for two consecutive years (Rensel and Prentice 1977); 4) it is gregarious and is normally not cannibalistic; 5) it adapts to vertical or hori- zontal substrates; and 6) it scavenges for, and ac- cepts, a wide variety of foods (Wickins 1972). Coincident to investigating the prawn as a com- panion crop to salmon, several prawn diets were evaluated with prawns held in tanks and net pens at the NMFS Aquaculture Experiment Station on Puget Sound near Manchester, Wash. These ex- periments wer^ conducted using diets made up of underutilized marine species or fishery byprod- ucts that are available to most salmon farmers. Materials and Methods The spot prawns used in the experiments were laboratory-reared progeny of females captured in Hood Canal, Wash. Three concurrent experiments with juvenile prawns ( <1 yr of age) began 10 July 1975 (Table 1). Experiment A was conducted in the laboratory where prawns were held in flowing seawater tanks at 110 animals/m^ of immersed substrate. Four diets were evaluated: 1) steamed mussel, Mytilus ediilis. meat; 2) chopped salmon that had died in nearby net pens; 3) feces and pseudofeces from the Pacific oyster, Ci-assostrea iii^as (eight oysters per replicate having a mean weight (total) of 153 g); and 4) no food (control). Diets 1 and 2 were fed every other day while diet 3 was always present in varying amounts. A sample of 10 prawns for each of four replicates was measured during each of 886 fishery BULLETIN: VOL 76, NO, 4, 1979. Table 1. — Growth and survival of juvenile Pandal us platyceros during three tests. Location Diet No of replicates Start of experiment No of prawns Mean in eactn weight replicate (g) 60 days after start of experiment Mean Mean survival weight (°°) (g) c nd of experiment Experiment No of days Mean survival (°o) Mean weight (g) A (prawns alone) Laboratory tanks Mussel Salmon Oyster wastes No food 4 4 4 4 25 25 25 25 0.72 0.62 069 0.64 82 83 64 26 252 2.27 1.06 1 07 90 90 (') (') 74 71 (') (') 3.14 2.61 (') (') B (prawns alone) Net pens Mussel Salmon^ 2 2 200 200 0.64 063 98 98 3,14 279 365 365 69 64 10.30 10.61 C (prawns and salmon) Net pens Variety of feeds^ 1 100 064 93 3,42 206 93 860 'Terminated at 60 days ^Includes net-fouling organisms. ^Includes salmon mortalities, uneaten fish feed (Oregon moist pellets), salmon feces, and net-fouling organisms three 30-day sampling periods. Carapace lengths' and individual wet weights (nonblottedi were measured to the nearest 0.5 mm and 0.01 g, re- spectively. In all experiments growth data were analyzed by one-way analysis of variance and sur- vival by chi-square tests. In Experiment B, prawns were held in two net pens measuring 1.2 x 1.8 x 1.8 m and constructed of knotless nylon web (6.8-mm stretched measure mesh I. Each pen was vertically divided into three equal compartments (6.5 m-^ of substrate each) with only the outer two being stocked with prawns. The prawns were stocked at a density of 30.8/m^ of immersed substrate. The net pens were covered with black plastic to prevent bird preda- tion and to reduce light intensity. Two dietary treatments were evaluated, mussel meat and salm- on. Each treatment consisted of two replicates and was fed exclusively on one of the two diets. Feed was to excess every other day. A sample of 50 prawns/replicate (100/ treatment) was measured for length and weighed during each of eight sampling periods, except dur- ing the last three periods in which all survivors were measured and weighed. The prawns were sampled at the beginning of the experiment and 32, 60, 88, 146, 205, 292, and 365 days later. In Experiment B, in addition to evaluating the diets, we studied the net cleaning ability of the prawns and the value of the organisms on the net (i.e., the net fouling organisms) as a supplemen- tary food source. Samples of net fouling organisms were taken from inside and outside of a compart- ment containing prawns (test) and inside and out- side of one not containing prawns (control). The nets were selected at random, and each sample ' Carapace length is defined as the distance from the base of the eyestalk to the posterior middorsal edge of the carapace. consisted of all the organisms on the net within the area of two 20-cm-diametei" circles; one circle was at 0.25-m depth, and the other was at 1 .0-m depth. The material collected was identified, enumer- ated, and measui'ed volumetrically. The nets were sampled during November 1975 and March 1976. In Experiment C, juvenile prawns were stocked in a net pen with coho salmon, Oncorhynchus kisutch, (age-group 0) that averaged 20 g each. A single net pen (without dividers) having a sub- strate area of 10.8 m^ was used. Prawn density was 9.3/m^ of substrate, and salmon density was 82'm'' of water. The salmon were fed Oregon moist pel- lets at 37f body weight/day; however, no feed was provided for the prawns other than what they could scavenge. All the prawns were measured at each of seven periods: at the beginning of the ex- periment and 15, 33, 60, 89, 146, and 206 days later. Care was taken to standardize culture condi- tions such as lighting, substrate type, and water temperature within each experiment. This was not practical between experiments because of in- herent differences between laboratory and net pen work. Stocking density differed between experiments, but its impact on growth and survival (agnostic behavior and feeding dominance) was minimized by distributing an excess of food throughout the rearing enclosure. In several years of behavioral observations we have rarely seen overtly agres- sive or cannibalistic behavior in spot prawns. Results and Discussion Mussel-fed juvenile prawns had the best survi- val and growth rate of all the prawns raised in the laboratory tanks (Experiment A); 74% of the 887 prawns survived, and they had a final mean weight of 3. 14 g (Table 1 ). The prawns fed salmon were significantly smaller (P<0.01) than the mussel-fed group. However, the growth of prawns in both the mussel and salmon diet groups equaled or exceeded the growth reported for a natural population in British Columbia (Butler 1964). Growth was similar between oyster waste and no food diet groups (Table 1), but survival was significantly different (P<0.01), 64'/f for oyster waste and 269^ for no food supplement. Prawns fed oyster wastes or receiving no food grew at a slower rate than the prawns in the other two diet groups; this portion of the study was terminated after 60 days. The poor growth of prawns fed oyster wastes contrasts with the good growth of lobster //o/??a/7/.s cifucricanus fed algae and oyster wastes in a sew- age enriched raceway system (Mitchell 1975). In that study intermediate organisms that fed on the solid wastes of oysters were also available as a food source for juvenile lobsters. In our study, raw unfiltered seawater was used, but cleaning the tanks twice weekly prohibited the establishment of intermediate organisms. Prawns in the net pens (Experiment B) grew significantly faster than those in the laboratory (Experiment A) for both mussel- (P<0.01) and salmon-fed (P>0.01) treatments (Table 1). Further, there was no significant difference in the growth (P>0.10) or survival (P>0.25) of prawns fed mussel or salmon in the pens as there had been in the laboratory tanks. In the net pens, the presence of net fouling or- ganisms as an additional food source for the prawns could explain the improved growth over that seen with the same basic diets used in the laboratory. Net fouling organisms could have pro- vided nutritional requirements that were deficient in the basic diets as provided in the laboratory. Prawns in both diet groups were observed re- moving organisms that were on both the inside and outside surfaces of the net pens using their second periopods. The amount of net fouling or- ganisms was reduced by the prawns in these en- FlGURE 1. — Webbing of a three-chambered net pen showed reduced fouhng in the right and left chambers (spot prawns present) compared with the center chamber without prawns. 888 closures ( Figure 1 , Table 2). The largest amount of organisms occurred at the shallower depths of the nets as in Moring and Moring's (19751 study of salmon net pens at the same site. Mussels, asci- cliaiis. and t uliicolous polychactt's, Spirarhis sp., c(iiiti'il)utc'(l lln' most fouling iti the control cham- bers (without prawns). Except for a few Spirorbis sp.. the net pen chambers with spot prawns were completely clean. Little algal growth was present due to the reduction of light by the black plastic covers. Rearing prawns and salmon in the same net pen (Experiment C) proved encouraging. After 6'j mo of culture the growth of these prawns (Figure 2) exceeded that of the monoculture Experiment B (P<0.01t and that reported for a natural popula- tion (Butler 1964). Survival was 93' f and not sig- nificantly different (P >0.95) from that of Experi- ment B. There was no evidence of adverse salmon/prawn interaction. The types of food available to the prawns when reared with salmon included: dead fish, uneaten fishfood pellets, fish feces, and net- fouling organisms. The relative contribution of each was unknown. A limiting factor to stocking juvenile prawns in commercial salmon net pens is the requirement that the prawns must be large enough to prevent them from going through meshes of the net. Smaller "nursery" nets of reduced mesh size could be hung inside the main salmon nets until the prawns reach a suitable size (about 4 g, or 3 mo of age). Several advantages might accrue from using a scavenger, such as the spot prawn, as a companion crop in salmon culture. In Experiment B a reduc- tion in net fouling was seen in net pens with prawns (Table 2. Figure 1). This reduction will aid salmon culture because it would allow greater water circulation within the enclosure, thus in- creasing dissolved oxygen levels and flushing of CO 5 10 < cr Q_ o 8 S^ 6 UJ Ijj < UJ > < DIET A Experiment C (voriety of foods) • Experiment B (mussel) ■ Experiment B (salmon) British Columbia natural population (Butler 1964) X _L 3 6 9 NUMBER OF MONTHS 12 FlClKK 2. -Growth of juvenile Pandali/s plalyrcros in net pens compared with a natural population. metabolic wastes from the pens. A reduction in net maintenance cost might also be realized. Experi- inents B and C demonstrated that fish in the pres- ence of net-fouling organisms was an acceptable food for the prawn (Figure 2); the utilization of dead salmon by a scavenger would be a valuable conversion of an otherwise unused protein and would reduce the labor needed to remove salmon mortalities from the system. Further experiments are needed to determine proper stocking densities of prawns and salmon to maximize growth rates, survival, and to make op- timum use of the net cleaning activities of the prawns. Further, while our studies showed that a single diet fed to prawns in the laboratory was not adequate for rapid growth, other studies (Kelly et al. 1976) have shown that combination raw diets could produce adequate growth. These combina- tion diets need to be evaluated in the net pen system. T.AHl.K 2. — Displacement volumes i milliliters) of fouling or- ganisms on the inside and outside surfaces of net pens with and without Pandcihis platycerog from July 1975 to March 1976. All pen chambers were clean at the start of the experiment. Sample area was 314.2 cm^ of vertical mesh. Location of sample Depth of sample (m) Net pen compartment with prawns Net pen compartment without prawns Nov. 1975 Mar 1976 Nov, 1975 IVIar, 1976 Inside 0 25 0 40 0 00 2640 41 55 of net 1 00 0 00 0 00 1050 12 03 Outside 025 0 25 0 50 9 00 17 00 of net 1.00 0,10 0,40 600 11,53 Literature Cited Butler, T. H. 1964. Growth, reproduction, and distribution of pandalid shrimps in British Columbia. J. Fish. Res. Board Can. 21:1403-1452. KKLLY, R. O. a., a. W. H.\SKI.TINK. .-\Nn E. E. EliKRT. 1976. Mariculture potential of the spot prawn, Pandalus platyceros Brandt. Aquaculture 10( 1 1:1-16. Mahnken, C. V. W. 1975. Status of commercial net pen farming of Pacific salmon in Puget Sound. In J. W. Avault, Jr.. and R. Miller 889 (editors), Proc. 6th Annu. Meet. World Mariculture Soc, p. 285-298. MITCIIKI.I,. J. R. 1975. A polyculture system for commercially important marine species with special reference to the lobster. Honuiriisaiueritaruis. In J. W Avault, Jr., and R. Miller (editors). Proc. 6th Annu. Meet. World Mariculture Soc. p. 249-259. MOKIN'C;, J. R., .AM) K. A. MOKINC. 1975. Succession of net biofouling material and its role in the diet of pen-cultured chinook salmon. Prog. Fish- Cult. 37(l):27-30. PRE.NTRK, E. F. 1975. Spot prawn culture: status and potential. /« C. W. Nyegaard i editor). Proceedings of a seminar on shellfish farming in Puget Sound, Oct. 7, 1975, Poulsbo, Wash., p. 1-11. Wash. State Univ., Coll. Agric, Coop. Ext. Serv., Pullman. RKN.SKI., J. E., .^\\^ E. F. PKKNPKK. 1977. First record of a second mating and spawning of the spot prawn, Pandalus platycerox, in captivity. Fish. Bull., U.S. 75:648-649. WlCKl.WS, J, F. 1972. Experiments on the culture of the spot prawn Pan- dalus platyccros Brandt and the giant freshwater prawn Macrobrachium roscnhcrgti ide Man). Fish. Invest., Minist. Agric, Fish. Food (G.B.)., Ser. II, 27(5), 23 p. John E. Rk.wskl Squaxin Island Tribe Route 1 , Box 25 7 She! ton. WA 98584 E.AKl, F. PKKNTKK Northwest and Alaska Fisheries Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, W A 98112 LARVAL LENGTH-WEIGHT RELATIONS FOR SEVEN SPECIES OF NORTHVC EST ATLANTIC FISHES REARED IN THE LABORATORY Growth is an important connecting link in the functional influence of biotic and abiotic factors on the dynamics of fish populations. Length-weight relations are used by fishery scientists to describe the growth characteristics of species or popula- tions and as a basis of evaluating the consequences of environmental influences on growth. Length- weight relations are also used in assessing produc- tion when combined with age and growth informa- tion and in determining length or weight in a situation where either one or the other is unknown due to sampling procedures. Studies of the early life of fishes are receiving increasing emphasis, particularly with regard to growth and survival in the larval stage. Survival during this period is thought to be minimal and potentially variable from year to year. Small changes of tenths of a percent in mortality have the potential to produce orders of magnitude dif- ferences in eventual adult populations. Larval growth can be influenced due to food limitations and varying abiotic factors (Houde 1974; Lasker 1975; Laurence 1977). Because of these facts, fishery scientists are particularly concerned with two aspects: 1 ) quantifying variable larval growth and survival, relating it to subsequent year-class recruitment, and applying it to traditional stock- recruitment relationships where recruitment has often been considered constant; and 2) the poten- tial use of this type of information in evaluating the increasing effects of pollution or other en- vironmental perturbations because of the fragility and sensitivity of larvae to changing or altered environmental variables. Solutions to these problem areas require quan- titative knowledge of growth parameters of larval fishes, and length-weight relations can be helpful in providing information or establishing relation- ships between pertinent sets of data. It is gener- ally thought that weight' is a better measure of absolute growth offish larvae than length as well as the prime determinant of condition when com- bined with length. Many species exhibit allo- melric or dispi'oportionate length-weight growth. This is especially true during the period of metamorphosis when some species display vary- ing or unusual body proportions with age (Blaxter 1969) and length does not increase in proportion to increasing weight. Additionally, recent attempts to construct models of larval survival, as influenced by environmental variables and den- sity dependent feeding relationships, require weight detei'minations for estimates of biomass and caloric tui-nover between larval and prey trophic levels. There is an extensive data base to asses larval fish growth and survival based on ichthyoplank- ton collected on survey cruises during the last 75 yr by marine laboratories throughout the world. Unfortunately, almost all of these data are in standard or total length measurements as they are much more easily and rapidly taken than dry weights. The difficulty involved in obtaining dry 'Weight for species in this research refers to dry weight. Dry weight is the most accurate for fish larvae because accurate wet weights are difficult to obtain and yield variable results on or- ganisms as small as fish larvae. 890 FISHKKV Bl'I.I.KTlN Vol. 7fi. NO 4. 1979. weights of young stages has been overcome to a certain degree with the advent of experimental laboratory programs at a few research facilities during the last 10 yr. The experimental larval fish program at the Northeast Fisheries Center Narragansett Laboratory, National Mai'ine Fisheries Service, NOAA, has been studying grow th, metabolic, and trophodynamic factors for a number of important commercial and sport species, and it is the object of this report to present larval length-weight rela- tions for seven species including Atlantic cod, Gadus ruorhua: haddock, Melanogrammus aeglefinun: scup, Stcnotonuis chrysops: Atlantic herring, Cliipea harengiis; winter flounder, Pseudoplci/ronectes omericanus\ summer floun- der, Paralicbthys dentatus\ and yellowtail floun- der, Limanda ferruginea . The larval length- weight relations presented here are previously unreported in the literature for six of the seven species with the exception of the Atlantic herring, which is included because it represents the only data available for western North Atlantic stocks. Materials and Methods All larvae were obtained from experimental spawning of adults in the laboratory and reared by techniques reported by Smigielski ( 1975a, b) and Laurence ( 1975). The length-weight data were col- lected coincident with a variety of experimental studies on larval growth, survival, metabolism, and feeding reported by Laurence (1974, 1977, 1978, and as yet unpublished). In all cases the data were collected from larvae reared at prey concentrations in the range of 0.5 to 3.0 organisms ml. Concentrations of 0.5 and above have been shown to be adequate for normal grow th in the studies cited above. Rearing temperatures were optimum for growth and survival or within a 3°C nonlethal range about the optimum depend- ing upon the experiment from which the data were taken. Optimum temperatures determined in laboratory studies for rearing the seven species were 7°C for cod and haddock, 8°C for winter flounder, 10°C for herring and yellowtail flounder, 16°C for summer flounder, and 18°C for scup. Length measurements were taken from the tip of the snout to the end of the notochord in the preflexion stage. During flexion of the notochord measurements were taken to a line vertically per- pendicular to the tip of the notochord until the hypural bones became prominent or exceeded the line vertically perpendicular to the notochord tip. At this time, a standard length measurement to the posterior end of the hypural plate was re- corded. Since the original experiments were not designed for developmental anatomy purposes, the different flexion stages were not recorded coin- cident with the length and weights. Lengths were recorded to the nearest 0.1 mm with a filar ocular micrometer. Dry weights were determined after rinsing larvae in distilled water, pipeting onto a glass Petri dish, and drying to a constant weight at 60°-90°C for 24 h. Individual dry weights were recorded to the nearest 0.1 /xg on a gram electrobalance. All measurements were made on post yolk-sac larvae that were freshly sacrificed and unpre- served. The data points for each species represent lengths and weights for individual larvae except for winter flounder and haddock. The data for these two species are the means of lengths and weights for samples of 10-25 larvae collected on a weekly basis during different experiments. The experimental procedures precluded the matching of individual lengths with weights for these two species. Regression equations and associated parame- ters were calculated as geometric mean, func- tional regressions using log base 10 transformed data according to the methods of Ricker (1973) rather than using the previously standard predic- tive, regression techniques. Ricker demonstrated the advantages of using functional rather than predictive regression calculations to reduce bias in length-weight conversions where the populations of measurements are typically open ended, where only a portion of the length and weight distribu- tions are represented, and where the variability may be more inherent in the biological material itself rather than the means of measuring length and weight. Results and Discussion The exponential relation between length and weight for all seven species are presented in linearized form by logarithmic transformation in Figures 1-7. The larvae studied in this research are from different taxonomic families (Clupeidae, Gadidae, Sparidae, Bothidae, and Pleuronec- tidae), represent different adult life styles (pelagic and demersal), develop in a range of different temperatures, and demonstrate different patterns of metamorphosis from larval to juvenile stages. 891 •4. 0 3.0 ID a. 2. 0 1. 0 STRNDHRD LENGTH (MM) 10. 0 SUMMER FLOUNDER L06r--0. 763+3. 780LO6X 100.0 10000. 0 1000. 0 100. 0 10. 0 0. 1 1.1 LOG STflNDHRO LENGTH (MM; FiGl'RE 1. — Standard length-dry weight relationship of larval summer flounder. STRNDRRD LENGTH CMM) 10. 0 4.0 3.0 - ^ CD o 3. 3. (- I— I U3 |J3 UJ Ul rz 3 V V q: or Q a o o J 2.0 - 1. 0 0.1 1.1 LOG STRNDRRD LENGTH (MM.) 100. 0 r 10000.0 1000. 0 (SI a. I 01 Q 100. 0 10. 0 Figure 3. — Standard length-dry weight relationship of larval cod. 4. 0 3.0 2. 0 I.O STRNDHRD LENGTH (MM) 10.0 100. 0 10000. 0 1000. 0 a. 100. 0 10. 0 0. 1 1.1 LOC S^RNDflRD LENGTH (MM) Figure 2.— Standard length-dry weight relationship of larval haddock. Points represent means for length and weight of sam- ples of 10-25 larvae. 892 4. 0 - a. ::. 3.0 - >- tr a o J 2. 0 1. 0 STRNDRRD LENGTH (MM) 10.0 100.0 rELLOWTRiL FLOUNDER L06r = -1. 017+3. 909LO6X - 10000.0 ID a. 1000.0 7r. 01 Q 100. 0 10. 0 0. 1 1.1 LOG STRNDRRD LENGTH (MM) Figure 4. — Standard length-dry weight relationship of larval yellowtail flounder. STRNDRRD LENGTH (MM) 10. 0 4.0 3.0 a. X 2.0 kl INTER FLOUNDER L0GY = -1. 347t-4. 769LOGX 10000. 0 1000. 0 1.0 Q 100.0 10. 0 0. i LOG STRNDRRD LtNGTH fMM) Figure 5. — Standard length-dry weight relationship of larval winter flounder. Points represent means for length and weight of samples of 10-25 larvae. In spite of these differences, a visual examination of the length-weight regression equation coefficients and associated parameters for all species reveals no obvious correlations with the differences (Table 1). It would not be prudent to statistically test for differences or associations be- tween the species because data for haddock and winter flounder were averaged. Ricker ( 1973) cau- tions that averaging changes the variances as- sociated with the variables, particularly the inde- pendent variable, so that a comparison between STRNDRRD LENGTH (MM) 10.0 100. 0 4.0 3.0 3. o J 2. 0 - 1.0 10000. 0 1000.0 I C3 a 100.0 10. 0 0. 1 1.1 LOG STRNDRRD LENGTH (MM) Figure 6. — Standard length-dry weight relationship of larval Atlantic herring. STRNDRRD LENGTH (MM) 10.0 100.0 10000.0 CD 3. 1000.0 5 a 100. 0 1. 0 10.0 0.1 1.1 LOG STRNDRRD LENGTH (MM) FIGURE 7.— Standard length-dry weight relationship of larval scup. 893 Table l. — Regression parameters for length-weight relations of seven species of laboratory-reared larval north- west Atlantic fishes. Coefficient Standard error 95°o CI about Number Correlation of Regression of regression regression Larval species sampled coefficient determination coefficient coefficient coefficient Summer flounder 57 0997 0 994 3,780 0 039 3702-3858 Yellowtail flounder 80 0.995 0 990 3.909 0 044 3821-3953 Herring 98 0,997 0 993 4.295 0037 4,221-4 369 Soup 100 0-997 0 993 3.756 0028 3692-3820 Cod 104 0 997 0 995 4.081 0029 4 023-4 104 Haddock' 23 0997 0 995 4476 0,071 4 328-4,624 Winter flounder' 36 0991 0 982 4 769 0 110 4 545-4 993 IDala represent means for length and weigfit of samples of 10-25 larvae averaged and unaveraged data is not valid: al- though he does not discredit the use of averaged data by itself. Also, seven species, some of which are closely related taxonomically, probably do not constitute enough cases for drawing conclusions about functional differences. Consequently, these length-weight relations should properly be con- sidered individually as empirically derived rela- tions for each particular opecies. The length-weight relation of fishes usually ap- proximates the cube law relationship in which the weight is proportional to the cube of the length (Beckman 1948; Rounsefe'l and Everhart 1953). This is usually true for adult fishes; however, re- sults of this research imply that it is not necessar- ily .so for larvae. All the length exponents for the species investigated in these studies were >3.6 with a mean value of 4.152. It would seem then that the dry weight of larval fishes may be more closely proportional to length to the fourth power rather than cubed. Length-weight relations for fish larvae are scarce in the literature. Examina- tion of the data available (log,,, formulation) seems to substantiate that the length exponent is always greater than three and more closely ap- proximates four. Marshall et al.( 1937) presented a total length-dry weight equation for larval her- ring, the only species with data available to com- pare with this study, equivalent to log W = -5.6990 + 4.52 log L. The length exponent is >4 and similar to the value of 4.295 for herring in this research. Ehrlich et al. ( 1976) also presented a simi- lar standard length-dry weight relation for Firth ofClyde herring larvae ( log W = -5.7052 -(-4.5710 logL) as well as a relationship of log W = -4.3043 -i- 3.9155 log L for larvae of plaice, Pleuronectcs plati'ssa. Stepien (1976) reported a standard length-dry weight relationship for larval sea bream, Anhosargus rhonihoidcilis, of log W = -0.5144 + 4.2816 log L, and Lasker et al. ( 1970) reported a standard length-dry weight relation- ship for northern anchovy larvae, Engrau lis nior- dax. of log H' = -3.8205 + 3.3237 log L. It is acknowledged that variables such as tem- perature and feeding conditions can influence growth and complicate length-weight relations. These factors may have contributed to some var- iability in the present study. However, it is felt that these influences were minimized by the experi- mental feeding levels and temperatures which were within ranges for adequate growth and sur- vival, and any changes in length or weight were most likely mitigated together causing little effect on the form of the length-weight relation. This is supported in studies of haddock larvae (Laurence 1974) where condition factors were similar and randomly associated with prey concentrations >0.5 organisms/ml. The use of larval length-weight relations for extrapolation may result in some underestimation or overestimation at the smallest and/or largest sizes due to changes in growth rates for yolk-sac or metamorphosing larvae. Farris (1959) suggested that growth rates of larval marine fishes could be separated into three different phases; the first two prior to yolk absorption and the third following. Zweifel and Lasker ( 1976) presented a mathemat- ical interpretation of larval growth with age defined by the Laird-Gompertz growth function. They noticed two growth cycles; one extending from hatching to yolk absorption and the other following yolk absorption. This variability in the small sizes is probably not inherent in the data of this study because larvae were not included until yolk was absorbed and active feeding had com- menced. Some variability may be present in the upper range of sizes in these length-weight rela- tions. In some cases data for larger larvae are not as extensive as for smaller larvae. Also, the major- ity of the largest individuals for each species were either undergoing or had completed metamor- phosis where changes in growth rates of length or 894 weight might cause allometry. Zvveifel and Lasker 11976) briefly considered the length-weight rela- tion in terms of a modified Gompertz-type relation and noted overestimation problems in extrapola- tion at the largest sizes. Length-weight relations have merit, but their usefulness is greatly enhanced when combined with other studies, particularly those on age. Length- weight by itself does not necessarily imply rate of change because of the potential influence the environment may have on changing growth with time. However, when correlated with age and compensated for change in rate due to biotic and abiotic influences, length-weight studies can be an important component in estimating growth, sur- vival, and population production. Acknowledgments I wish to thank J. B. Colton, Jr. and K . Sherman for their critique of the manuscript, and B. R. Burns, T. A. Halavik, and A. S. Smigielski for their assistance in rearing the larvae and collect- mg measurements. Literature Cited BECKMAN. W. C. 1948. The length-weight relationship, factors for conver- sions between standard and total lengths, and coefficients of condition for seven Michigan fishes. Trans. Am. Fish. Soc. 75:237-256. BL.A.XTER. J. H. S. 1969. Development: eggs and larvae. In W. S..Hoar and D. J. Randall (editors). Fish physiolog>'. Vol. 3. p. 177-252. Academic Press, N.Y. EHRLICH. K. F., J. H. S. BLAXTER. AND R. PEMBERTON. 1976. Morphological and histological changes during the growth and starvation of herrmg and plaice larvae. Mar. Biol. (Berl.i 35:105-118. FARRIS. D. A. 1959. A change in the early growth rates of four larval marine fishes. Limnol. Oceanogr. 4:29-36. HOUDE. E. D. 1974. Effects of temperature and delayed feeding on growth and survival of larvae of three species of subtropi- cal marine fishes. Mar. Biol. (Berl.) 26:271-285. Lasker. R. 1975. Field criteria for survival of anchovy larvae: The relation between inshore chlorophyll maximum layers and successful first feeding. Fish. Bull., U.S. 73:453-462. Lasker, R., H. M. Feder. G. H. Theilacker, and R. C. May. 1970. Feeding, growth, and survival ofEngraulis mordax larvae reared in the laboratory. Mar. Biol. (Berl.) 5:345- 353. Laurence. G. C. 1974. Growth and survival of haddock Melanogrammus aeglefinus larvae in relation to planktonic prey concentra- tion. J. Fish. Res. Board Can. 31:1415-1419. 1975. Laborator>' growth and metabolism of the winter flounder Pseudopleuronectes americanus from hatching through metamorphosis at three temperatures. Mar. Biol. (Berl.) 32:223-229. 1977. A bioenergetic model for the analysis of feeding and survival potential of winter flounder, Pseudopleuronectes americanus, larvae during the period from hatching to metamorphosis. Fish. Bull., U.S. 75:529-546. In press. Comparative growth, respiration and delayed feeding abilities of larval cod [Gadus morhua) and had- dock (Melanogrammus aeglefinus) as influenced by tem- perature during laboratory studies. Mar. Biol. (Berl.) Marshall. S. M., A. G. Nicholls. and A. P. Orr. 1937. On the growth and feeding of the larval and post- larval stages of the Clyde herring. J. Mar. Biol. Assoc. U.K. 22:245-268. RlCKER, W. E. 1973. Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. Rounsefell, G. a., and W. H. EVERHART 1953. Fishery science: its methods and applications. John Wiley and Sons Inc., N.Y., 444 p. Smigielskl a. S. 1975a. Hormone-induced spawnings of the summer floun- der and rearing of the larvae in the laboratory. Prog. Fish-Cult. 37:3-8. 1975b. Hormonal-induced ovulation of the winter floun- der, Pseudopleuronectes americanus. Fish. Bull., U.S. 73:431-438. Stepien, W. p., Jr 1976. Feeding oflaboratory-reared larvae ofthe sea bream Archosargus rhomboidalis (Sparidae). Mar. Biol. (Berl.) 38:1-16. Zweifel. J. R., AND R. Lasker 1976. Prehatch and posthatch growth of fishes— a general model. Fish. Bull., U.S. 74:609-621. Geoffrey C. Laurence Northeast Fisheries Center Narragansett Laboratory National Marine Fisheries Service. NOAA R.R. 7 A. Box 522 A Narragansett, RI 02882 EFFECT OF THERMAL INCREASES OF SHORT DURATION ON SURVIVAL OF EUPHAUSIA PACIFICA Euphausiids are an important source of food for many valuable species of fish including herring, cod, pollock, and salmon. Cooney (1971) reported that Euphausia pacifica was the most abundant species associated with the diffuse scattering layer at all locations in Puget Sound, Wash. He found that during the day euphausiids are most abun- dant between depths of 50 and 100 m and that at night most of the population migrates into the upper 50 m. Cooney's findings indicate that great numbers of euphausiids could be drawn through FISHERY BULLETIN: VOL. 76. NO. 4, 1979. 895 the condenser cooling systems of thermal nuclear power plants where they would encounter a siz- able thermal shock. Zooplankton entrained in a power plant cooling system located in a saltwater environment could be subjected to an average temperature increase ranging from 12° to 16°C (Coutant 1970). In some plants, the increases are as high as 19°C. Maximum temperatures would be reached in <1 min in the condenser and would be maintained for at least 9 min in a diffuser discharge system and at almost maximum temperature, for possibly up to 21 min, in a discharge canal system. Other factors that could cause damage to euphausiids in a cool- ing system include pressure changes, abrasion, and toxic substances. I simulated the thermal conditions encountered in a cooling system to determine the temperature increases that E. pacifica could resist for short periods (15 and 30 min). This information can be applied to the design and operation of cooling sys- tems to protect zooplankton. These studies were conducted at the National Marine Fisheries Service's Mukilteo Field Sta- tion, Washington, during 1971-74. Methods Euphausiids for these experiments were cap- tured during daylight hours in Port Gardner of northern Puget Sound, Washington, between Mukilteo and Gedney Island. A 10-m net with 333- /Lim aperture Nytex^ netting was towed at a depth of about 60 m at a rate of 4.6 km/h. Tows were usually of a 5-min duration. A 946-ml glass bottle was used as a collection receptacle to protect the animals. As soon as the net was retrieved, the catch (con- sisting mostly of euphausiids) was divided be- tween two or three 18.9-1 Nalgene carboys filled with fresh seawater and covered with black polyethylene sheeting to exclude light. The catch was taken immediately to the laboratory, usually <2 km away, where the euphausiids were sepa- rated from other organisms in the catch and placed in 5-1 battery jars (23 x 14 x 17 cm) of fresh seawater. They were then placed in a dark, low- temperature incubator set at their previous am- bient seawater temperature where they were held before and after testing. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. The test apparatus consisted of a series of 5-1 battery jars filled with seawater that were main- tained at specific temperatures by immersion heaters activated by temperature controllers (Craddock 1976). The jars were in a primary bath of running seawater at ambient temperature and air was continuously bubbled into the jars to eliminate stratification. Test containers for holding the euphausiids were polyvinyl chloride boxes of 5 cm^ with two opposing sides having 4-cm diameter cutouts cov- ered with 333-/xm aperture Nytex netting to allow free water circulation. Styrofoam glued to the boxes provided flotation (Figure 1). The temperature-time regime to which the euphausiids were subjected was designed to simu- late their passage through a condenser cooling system. Coutant (1970) depicted a hypothetical temperature-time course for organisms entrained in condenser cooling water and discharged by dif- fuser or by discharge canal. An animal could be subjected to maximum temperature increases for up to about 10 min in a diffuser and up to about 20 min in a discharge canal system. Relative to his study, I chose 15- and 30-min exposure tests to represent the longest exposure that might be en- countered. To simulate these conditions, test euphausiids were subjected to a given tempera- ture ranging from 14° to 29°C for 15 or 30 min, starting from temperatures of 11° or 9°C. Euphausiids used as controls were always kept at the prevailing ambient temperature (approxi- mately the same as the subsurface temperature of Puget Sound). Five 15-min tests were conducted during June- July 1971, four 30-min tests were run during June-August 1971, and two 15-min tests were made during March-April 1974. The euphausiids were held 18 h or longer before testing to eliminate handling mortality and were then counted into test containers in seawater while the secondary baths were being raised to the test temperatures. Either 5 or 10 euphausiids were tested in each container, depending upon the numbers available for that particular test. When all secondary baths became equilibrated at the test temperatures, the boxes containing the euphausiids along with a small amount of water were placed in the test baths. Water in the test containers was within 0.5°C of the test tempera- ture in an average of 28 s after introduction. At the end of the exposure period, the test boxes contain- ing the euphausiids were removed from seawater at the test temperature and placed in fresh seawa- 896 Figure l. — Test chambers and apparatus for testing thermal effects on Euphausia pacifica. ter at the acclimation temperature and main- tained in the low temperature incubator. All lots were checked for mortality at 5, 10, and 15 min after introduction to the test temperature when the duration of the exposure was 15 min and also at 20 and 30 min after introduction for the 30-min exposure. All were again checked at 1, 24, and 48 h after testing. The 48 h survival was taken as diag- nostic for the TL5o's. Temperature effects were evaluated on the basis of mortality during tests and 48 h after testing. Forty-eight hours was assumed to be a reasonable holding period to check for delayed mortalities — yet not long enough to cause mortality due to confinement and lack of food. A euphausiid was considered dead if no movement of the thoracic appendages, pleopods, or antennae could be de- tected using 3 X magnification. I modified the term median tolerance limit (TL50) to indicate the maximum 15- or 30-min exposure temperature survived by at least 50'7f of the experimental ani- mals 48 h after testing. This should be considered the maximum temperature-time combination re- sisted. Lengths of the test animals between the ex- treme tips of the rostrum and telson were taken at the end of each test. The mean-lengths of the euphausiids tested at the various seasons ranged from 12.11 to 18.37 mm (Table 1). The actual range was from 9 to 27 mm. Those tested in the early part of June were the largest; they exceeded those tested later in June by an average of 6.26 Table 1.— Sizes (milHmeters) of Euphausia pacifica tested. Dates Mean Range Dates Mean Range 1971; 1974: June 2-9 18,37 14-26 Mar, 12 14.15 10-20 July 21-30 12.11 9-16 Apr. 12-16 13.85 10-23 Aug 4-11 13.14 10-19 897 mm. Those tested in August, March, and April had an average spread of only 1.01 mm. Results Controls in the different tests suffered no mor- tality during the exposure period except the June-July tests, where the 15-min group lost 7% of controls by the end of 48 h and the 30-min group lost 10% by the end of 48 h. The data were cor- rected to reflect the loss of the controls in the June-July tests, using the method of Tattersfield and Morris ( 1924) as reported by Sprague ( 1969). Acclimation temperature influenced resistance. The TL50 of euphausiids given a 15-min exposure to elevated temperature was 25°C for those accli- mated to 11°C; it was 23°C for those acclimated to 9°C. Exposure to 26°C resulted in survivals of 32% and exposure to 27°C resulted in almost im- mediate death ( <15 min). In the 15 min 9°C accli- mation test (March-April 1974), the TL50 was at 23°C and 47% were still surviving 48 h after expo- sure at 24°C. However, 15 min after exposure to 25°C, only 13% remained alive and all were dead in <15 min at 26°C. Figure 2 depicts the survival after a 15- and 30-min exposure to elevated tem- peratures and after a 48-h holding period. Increasing the duration of exposure to test temperatures from 15 to 30 min when the ambient temperature was 11°C decreased the TL50 by 1° to 24°C. Of those tested at 25°C, only 44% survived 48 h after testing. At 26°C, only 2.5% survived the 30-min test period. None survived the test period at 27°C. The logistic model was fitted to the data from the three different thermal shock tests. The prob- ability of survival was taken to be the form P [survival at temperatures] = l/d+e"*^''^ where e = 2.718. This is the so-called logistic model and a and b are parameters which are estimated using the data. In the 15-min exposure of June-July 1971, a = 0.6544 and h = -16.4138; in the March-April 1974 exposure, a = 0.9568 and b = -22.2860; whereas in the 30-min exposure July- August 1971, d = 0.5173 and b = -12.2572. The estimates of TL50 and an approximate 95% confidence interval for it follow for the three tests: 1) 25.08°C, 24.51°-25.65°C; 2) 23.29°C, 22.76°- 23.82°C; and 3) 23.69°C, 22.95°-24.44°C. There was no obvious difference in the effect of a I5MIN 100 90 80 ^ 70 1 60 JUNE - JULY CO 50 50% S^ 40 - 30 15 MIN EX 20 - o CO I I I I i_l 1 l_i I . I ■ l_i I 1 I V I . I 10 12 14 16 18 20 22 24 26 28 30 TEST TEMP (°C) 30 MIN /^ 48 H ""^ / V JUNE - JULY 50 40 - 30 - 20 - 10 - 0 V/^ 8 \ / 50% ^ 30 MIN EX -1— J — l-j 1—1 1 ] I I I . I I . J 10 12 14 16 18 20 22 24 26 28 30 TEST TEMP t°C) 100 90 80 I 60 a: 50 I5MIN 30 - 20 - 10 - 0 Ly/ 48 H MARCH -APRIL 50% — 15 MIN EX I . I I ■ I ■ I ■ I ■ I ■ I I I ] 8 10 12 14 16 18 20 22 24 26 28 30 TEST TEMP (°C) Figure 2. — Survival (including mean and range) ofEuphausia pacifica after 15- or 30-min exposures to elevated temperatures and subsequent holding at 9° or \\°C ambient temperatures for 48 h. 898 short exposure to increased temperature on the largest or the smallest euphausiids tested. Two out of three groups tested for 15 min in early June (the largest euphausiids) exceeded 50'7r mortality after exposure to 26"C as did the two groups tested in late July (the smallest euphausiids). D ISC us-sion The intake of a condenser cooling system may entrain large quantities of euphausiids — depending to some extent on the depth of the in- take, the season of the year, and even the time of the day. During the summer, fall, and winter, the young euphausiids make diurnal vertical migra- tions from the 50- to 100-m strata, rising daily to the surface during the dark hours. After sexual maturity in the early spring they descend even deeper until they inhabit depths over 200 m dur- ing their second winter. The following spring they rise to the surface for the second time to breed. The young euphausiids thus spend much of their first year at depths above 50 m, and older adults are again near the surface in the spring (Ponomareva 1963). Gilfillan (1972) pointed out that E. pcuifica is widely distributed and is abundant in water hav- ing differing temperature characteristics. His studies showed that E. pcuifica from the Pacific Ocean were more easily stressed by changes in temperature and salinity than those from the west entrance of Strait of Juan de Fuca — which, in turn, were more readily stressed than those from Saanich Inlet. His results indicate that E. pcuifica from inner Puget Sound would be among the most resistant to thermal stress of these different groups. Temperatures encountered by euphausiids in Puget Sound normally vary only slightly from the surface to 100 m and deeper. From October through about May there is usually no change in temperature from the surface to 100 m, whereas in the summer the surface to 10 m or less may be a few degrees warmer (Lincoln and Collias^). Seasonal temperature variations in most of Puget Sound are also small, ranging from a low of 7°or 8°C in February to 11°C in late July, August, and September. Even considering their vertical migrations in summer, euphausiids are normally ^Lincoln, J. H., and E. E. Collias. 1970. Skagit Bay Study Progress Report No. 3. Univ. Wash. Dep. Oceanogr., Seattle, Ref. M70-111, 88 p. subjected to only slight temperature fluctuations and, therefore, the mortalities observed at simu- lated condenser cooling temperatures are not sur- prising. Once entrained in a condenser cooling system, the euphausiids would encounter an abrupt tem- perature increase of 12°-16"C (Coutant 1970), which could increase temperatures above the am- bient temperature of Puget Sound to the critical range for survival. There are periods from July through September when surface temperatures may reach or exceed 15°C in portions of Puget Sound (Lincoln and Collias see footnote 2). Nor- mally, surface temperatures do not exceed 14°C. Cooney ( 1971) noted high surface temperatures in June of 16.7°- 19°C. These temperatures could re- sult in condenser cooling temperatures of 27°C and above, which this study found to be 1007f lethal in a very short time. Data from this study indicate that even a short passage time through a condenser (15 min) at temperatures of 23°-24°C could kill from 11 to 53V. of the euphausiids by thermal causes alone. The added loss due to abrasion, pressure, and toxic substances is unknown. To minimize damage to the euphausiid popula- tions, condenser cooling system intakes should be located deep enough to take advantage of the cold- est cooling water available to minimize tempera- tures in the system. A very deep intake (just below 100 m) would probably minimize the entrainment of euphausiids. A surface intake would be espe- cially harmful because of the higher surface tem- peratures and because of the swarming of euphausiids on the surface. Plant lights at night could cause the surface swarming. Literature Cited Cooney, R. T. 1971. Zooplankton and micronekton associated with a dif- fuse sound-scattering layer in Puget Sound, Washing- ton. Ph.D. Thesis, Univ. Washington, Seattle, 208 p. COL'TANT. C. C. 1970. Entrainment in cooling waters: Steps toward pre- dictability. Proc. 50th Anna. Conf. West. Assoc. State Game Fish Comm,, p. 90-105. CR.'SiDDOCK. D. R. 1976. Effects of increased water temperature on Daphnia putex. Fish. Bull., U.S. 74:403-408. GILFILLAN E. 1972. Reactions of Euphausia pad fica Hansen (Crustacea) from oceanic, mixed oceanic-coastal and coastal waters of British Columbia to experimental changes in tempera- ture and salinity. J. Exp. Mar. Biol. Ecol. 10:29-40. 899 PONOMAREVA, L. A. 1963. Evfauziidy sevemoi poloviny Tikhogo okeana, ikh rasprostranenie i ekologiya massovykh vidov (Euphausi- ids of the North Pacific, their distribution and ecolo- gy). Moscow. Izd. Akad. Nauk SSSR, 142 p. (Translated Isr. Program Sci. Transl., 1966. 154 p.; available U.S. Dep. Commer., Natl. Tech Inf. Serv.. Springfield, Va., as TT 65-50098.1 SpR.'^GUE. J. B. 1969. Review paper: Measurement of pollutant toxicity to fish. I. Bioassay methods for acute toxicity. Water Res. 3:793-821. T.-XTTERSFIELD. F., AND H. M. MORRIS 1924. An apparatus for testing the toxic values of contact insecticides under controlled conditions. Bull. Entomol. Res. 14:585-590. species (Lowell 1971; Morris and Kanayama'^l in- dictates that spawning takes place close to shore. The larvae and juveniles lead a pelagic existence for about 3 mo, juveniles moving to shallow in- shore areas at fork lengths (FL) between about 50 and 100 mm. The fish become sexually mature males at 20-25 cm FL and subsequently undergo a sex reversal, passing through a hermaphroditic stage and becoming functional females between 30 and 40 cm FL. Adults inhabit inshore rocky and sandy areas, frequently in zones of turbulence. Methods Donovan R. Craddock Northtvest and Alaska Fisheries Center- National Marine Fisheries Service. NOAA 2725 Montlake Boulevard East Seattle. WA 98112 LUNAR SPAWNING OF THE THREADFIN, POLYDACrVLL'S SEXFIUS, IN HAWAII' Recent evidence indicates that lunar spawning rhythms are more common in fishes than was once thought. Johannes ( 1978) listed 50 species of tele- ost fishes with lunar spawning rhythms, most of them tropical and all of them marine or catadro- mous. In the course of developing methods for cul- turing the threadfin, Polydactylus sexftlis (Cuvier and Valenciennes), we found that this species dis- played a lunar spawning rhythm (May 1976). The lunar pattern of spawning had been indicated by a previous field study (Lowell 1971) and is consis- tent with fishermen's lore (Hosaka 1944), but proof was lacking and details of the rhythm were unknown. In this paper we present detailed infor- mation on the lunar spawning of P. se.x fills along with observations of spawning behavior, using re- sults from captive fish. Polydactylus sex fills is a much sought-after food fish in Hawaii and supports an important sport fishery as well as a small commercial fishery (Rao^). Information on the life history of this 'Contribution No. 552, Hawaii Institute of Marine Biology. ^Rao, T. R. 1977. Enhancement of natural populations of moi (Polydactylus sexfilis) in Hawaii through the release of hatchery-reared juveniles - a feasibility study of sea ranch- ing. Univ. Hawaii, Hawaii Inst. Mar. Biol., Tech. Rep. 33, 46 p. Hawaii Institute of Marine Biology, P.O. Box 1346, Kaneohe, HI 96744. Juvenile P. sexfilis were captured by seine on reef flats along windward Oahu in September and October 1970 and reared to sexual maturity in tidal ponds at Coconut Island, in Kaneohe Bay, Oahu. The fish were daily fed chopped squid or smelt, commercial trout chow (409r protein), or trout chow supplemented with chopped squid. In May 1973, 30 mature fish (18 females and 12 males) were transferred to a 18-m^ nylon net en- closure suspended from Styrofoam'* floats and an- chored off the leeward (southwest) side of Coconut Island. In June and July 1973, a small number of these fish were removed to laboratory tanks and used in experiments on hormone-induced spawn- ing. During this work, ovarian biopsy samples were examined which contained residual eggs and indicated that the fish had been spawning spon- taneously. In order to monitor any such spawning, an airlift egg collector was installed (May et al.^) in the center of the net in July 1973 and operated continuously (except for a few days when equip- ment malfunctioned) between 14 July 1973 and 31 December 1975. Polydactylus sexfilis produces pelagic eggs, so that the collector obtained a sam- ple of eggs at each spawning. Every morning the entire contents of the collector were harvested and examined under a dissecting microscope, and the number of P. sexfilis eggs was estimated by sub- ^Morris, D. E., and R. Kanayama. 1964-69. Life history study of the moi, Polvdactvlus sexfilis. Job Completion Rep., Projects No. F-5-R-li to F-5-R-17, Div. Fish Game, State of Hawaii. Division of Fish and Game, Honolulu, Hawaii. ■•Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ■■^May, R. C, G. 8. Akiyama, and M. T. Santerre. 1976. A simple method for monitoring the spawning activity offish in net enclosures. International Milkfish Workshop-Conference, May 19-22, 1976, Tigbauan, Iloilo, Philipp. Working Pap. 10, p. 133-138. Southeast Asian Fisheries Development Center, Kalayaan Building, Dela Rosa corner Salcedo Sts., Makati, Metro Manila, Philipp. 900 FISHERY BULLETIN: VOL. 76. NO 4, 1979. sampling. In 1976 and 1977, the collector was op- erated only during the spawning periods, which were predictable on the basis of data collected dur- ing the previous years. Eggs of P. sexfilis were distinguished from occasional eggs of other species by their diameter (800-825 ^im) and by comparing hatched larvae with larvae obtained from hormone-induced spawnings; the identification was further corroborated by rearing larvae to the juvenile stage on several occasions. Since P. sexfilis undergoes a male-female sex reversal, ad- ditional males were added to the population each year to maintain a female to male sex ratio be- tween 1:1 and 1.5:1 during the spawning season. In 1976, additional males were not added until the third spawning' month, so data on spawning were not available for the first two spawning months of that year. At the end of this study in December 1977, the population of spawners numbered 57. In order to determine the exact time of day when spawning occurred, water was sampled from the holding net continuously with a centrifugal pump at 10 1/min and passed through a small collecting basket (500-/i,m nylon mesh) on a barge anchored next to the net. The collecting basket was moni- tored visually, and the time when eggs first ap- peared was noted. The superstructure of the barge provided a barrier between the observer and the holding net, so that activities associated with monitoring the basket did not disturb the spawn- ers in the net. Results Lunar Spawning Rhythm Eggs of P. sexfilis were first observed in the collector from 23 to 25 July 1973. Subsequently the fish were found to spawn at night over a period of 3-7 (in one instance, 10) days once each month, always in proximity to the last quarter phase of the moon (Figure 1). and in a spawning season that extended from May or June to October ( Table 1). Because the lunar month is 29.5 days long and does not coincide exactly with the calendar month, the calendar dates of spawning (Table 1) were generally 1-3 days earlier in each succeeding month. The first spawning of each monthly series was usually preceded by 1 or 2 days on which the fish fed less actively than normal. Counts of P. sexfilis eggs from the collector were made in Au- gust 1973, and thereafter. Judging from the sam- ples obtained by the collector, relatively few eggs 10 5 r 1 1 ■ ' ' 1 ' ' 1 ■ ' I 1 - 1973 1 1 T — 0 ,ll 1., 1 1" 2 20 3 X 15- 1974 , — Ul 5 b. O 0 ill 1 1 ■ a: UJ m 25 Z Z 20 1 MO) 15 1975 - 10 - 5 0 1 1 1 1 1 1 ^ 1 i M r r , 1 1 1 , - 1 . . , . 1 "i . . . . /i , , , /I Figure l. — Numbers of Polydactylus sexfilis eggs obtained by the airlift egg collector between August 1973 and October 1975. In July 1973, eggs were noted on 3 days but were not counted. Carets indicate the time of the last quarter phase of the moon. Table l. — Dates on which eggs were produced by captive Polydactylus sexfilis during study period. For each spawning month, the upper date is the first observed day of spawning, the lower date the last day. Spawning monttn Year III IV VI 1973 (') (') (') 22 Aug 25 Aug 19Sept.2 22 Sept 18 0ct.2 21 Oct. 1974 13 [Way 9 June 9 July2 7 Aug 5 Sept 6 Oct. 15 May 14 June 14 July 13 Aug 10 Sept 9 Oct 1975 1 June2 29 June 29 July 27 Aug 25 Sept 2 25 Oct. 5 June 3 July 1 Aug. 30 Aug 29 Sept 27 Oct. 1976 (') V) 17 July 15 Aug 1 3 Sept 2 1 3 Oct 20 July 18 Aug 16 Sept, 16 Oct 1977 9 May 7 June^ 6 July 4 Aug 3 Sept 1 Oct 14 May 1 1 June 1 1 July 13 Aug 7 Sept 5 Oct 'Data not available ^Collector malfunctioned on previous day: initial spawning day possibly ear- were produced on the first and last days of spawn- ing, with a peak usually on the second or third day (Figure 1). However, because the eggs collected represented only a sample of the total number of eggs produced, it is possible that the peaks reflected some sampling variability. The time of spawning relative to the lunar cycle appeared to change as the spawning season pro- gressed. The first spawning series of the year al- ways began (Figure 2) on the day of the last quar- ter (in one case the egg collector malfunctioned prior to the last quarter, and it is possible that the first day of spawning occurred earlier). In sub- sequent months the series began 1-4 days prior to 901 the last quarter (Figure 2). In April 1974 the col- lector obtained several thousand eggs over a 3-day period about 12 days before the last quarter; al- though the eggs were of the same diameter as P. sexfilis eggs, the larvae were not reared for posi- tive identification. In view of the consistency of subsequent spawning data, and the low ovarian weights which two previous studies found in P. sexfilis in April (Lowell 1971; Morris*'), we believe these were eggs of another species. Some of the eggs found in the collector were transparent, buoyant, and developing normally, while others sank and were opaque and obviously not viable. Among the spawnings in which 1,000 or more eggs were collected, an average of 52*%^ of the eggs were viable (range, 0-989^). The airlift collector apparently damaged the eggs to some extent. For example, on 13 June 1974, when 75^^ of the eggs from the collector were viable, eggs obtained by dip net at the time of spawning showed over 90^'r viability. It is not known how ^Morris, D.E. 1964. Life history study of the moi, Po/vrfac- tylus sexfilis. Job Completion Rep., Project No. F-5-R-11, Div. Fish Game, State of Hawaii, 15 p. Division of Fish and Game, Honolulu, Hawaii. Days Before 4 3 2 1 3 1 2 Days After 3 4 5 6 7 ■T I I ; I 1 ■ ' '. ■ '. 1 n 1 1 r— ' . ■ ■ ■ 1 x: d TTT ^^^^^ 1 o HI \///>//////x c a nz: 1 1 V/////////i 1 ■11 CO ■ 1974 lU 1975 7A 1976 D 197 7 1 1 > 1 1 2: 1 1 \/yyyyyyyyy'\ \ ' 1 Hj^^^^^l BHI 3ZI 1 V/////////^ i 1 1 1 , Figure 2. — Duration of spawning cS. Poly dactyl us sexfilis rela- tive to the time of the last quarter ((J) for the six spawning months in 1974, 1975, and 1977. Data are given only for the last four spawning months of 1976. many of the fish in the net participated in each spawning or whether the same fish spawned each month, although Kanayama'^ believed that indi- vidual P. sexfilis spawned more than once in a season. Time of Spawning The developmental stage of eggs found in the collector in the morning indicated that spawning had occurred shortly after sunset. Visual monitor- ing of water sampled continuously from the net on 34 spawning nights showed that with few excep- tions the fish spawned between 2030 and 2130 h (Figure 3, Table 2). The times recorded were those when eggs were first observed in the collector; it is possible that additional spawnings took place slightly later on the same night, and behavioral observations (see below) indicated that this may have been true on at least some nights. The time of spawning did not vary with the time of sunset (Figure 3) and appeared unrelated to the time of moonrise (which occurred generally between 2300 and 0400 h during the spawning season). ''Kanayama, R. 1967. Life history study of the moi, Polydactylus sexfilis. Job Completion Rep., FVoject No. F-5-R- 15, Div. Fish Game, State of Hawaii, 9 p. Division of Fish and Game, Honolulu, Hawaii. 06 02 22 - § LiJ 2 h- 14 10 06 - _l l_ I II 111 IV V VI SPAWNING MONTHS Figure 3. — Times of .spawning oi^ Polydactylus sexfilis during the six spawning months iroman numerals) of 1974 in relation to the tidal cycle. Dots indicate observed times of spa wning in 1974, and the horizontal lines delineate the time of spawning as indi- cated by data from 1974, 1975, and 1977 (see Table 2). Dotted horizontal lines show the times of sunset in 1974. Vertical lines for each spawning night indicate time between the evening high and low tides, i.e., the duration of the outgoing tide, as measured by a tide gage at Coconut Island in 1974. 902 Table 2. — Times of first spawning in a captive population of Polydactylus sexfilis on nights in 1975 and 1977. Where a single time is given, the egg collector was examined continuously; in other cases, the collector was examined at intervals of 5-15 min. Numbers in parentheses indicate times of peak fish activity, presumed to be the exact time of spawning. Disc ussion Time of first Time of first Dale spawning Date spawning 1975: 26 Sept 2045-2100(2052) 2 June 2115-2118 27 Sept 2054-2100 3 June 2110-2115 28 Sept 2050-2054 (2053) 4 June 2110-2115 29 Sept 2135-2140 (2138) 5 June 2115-2120 26 Oct 2050-2100 (2056) 30 June 2115-2120 27 Oct. 2105-2110 (2107) 2 July 2125-2130(2127) 1977. 3 July 2140-2145 7 June 2105 30 July 2045-2100(2058) 9 July 2147 31 July 2120-2125 10 July 2140 28 Aug. 2115-2118 (2117) 6 Aug 2050 29 Aug. 2129 4 Sept 2103 30 Aug. 2115-2118 (2117) 3 Oct. 2038 Data from a tide gage located at Coconut Island were available for 1974 and showed that spawning nearly always took place on the outgoing tide (Figure 3). Although tide gage data were not available for subsequent years, tides predicted from tide tables showed patterns similar to the 1974 tide gage data. For 1975, 1976, and 1977, the time of spawning (i. e., 2030-2130 h) was compared with predicted tides and again found to occur mostly during the outgoing tide. Spawning occur- red on the outgoing tide in 739r of the spawning nights during 1974-77. Spawning Beha\ ior Observations of spawning behavior were made initially by watching bioluminescence caused by the fish's movement; later, direct observations were made by shining lights on the water at the time of spawning. The level of activity of the fish gradually increased beginning around 2015 h and culminated in the spawning act as determined by the appearance of eggs in the centrifugal pump samples. Occasionally the fish broke the surface of the water during the period of increased activity. During courtship the fish swam rapidly around the net in a circular manner in groups of two or three. They appeared to be chasing one another, and often one fish would contact another from be- hind, either dorsally or ventrally, with snout. Spawning appeared to take place between pairs rather than among larger groups offish. Increased activity usually continued for 20 or 30 min after eggs were first noted. The captive population of P. sc.xftlis was clearly spawning with a well-defined lunar rhythm. Other evidence implies that this is a natural be- havior for this species. Lowell (1971) set gill nets weekly in certain shoal areas of Oahu from April to August 1970 and reported that exceptionally large catches of P. se.xfilis per effort occurred "af- ter the full moon and continuing until the last quarter (1 week duration)," and because of the stage of gonadal development among fish in such catches, he termed them "spawning runs." In July 1970, female fish caught 3 days before the last quarter all had well-developed ovaries, but fish caught 4 days after the last quarter had spent ovaries. Fishermen seem to have been aware of the habits of P. scxfilis; for a long time: Hosaka (1944:117) stated, "Moon light nights are best for moi (= P. sexfilis) fishing, and this is especially true when the moon is in the last quarter phase." Spawning at the time of the last quarter phase of the moon appears to be rare among fishes. Of the 50 lunar spawners which Johannes (1978) listed, only two besides P. sexfilis spawned on the last quarter; both of these are species of Amphiprion, and one spawned on the first as well as the last quarter. Since the various species covered in Johannes' list occurred in different geographic lo- cations, the variations in spawning days, taken together with variations in spawning times, could reflect local adaptations such as would occur if egg or larval survival were related to tides or currents. The coincidence of spawning in P. sexfilis with the outgoing tide indicates that the remarkably precise timing of spawning may act as a mechanism for offshore dispersal of eggs and lar- vae. Lowell (1971) noted that there was a strong, oceanward current at his sampling site during falling tide, when he estimated spawning occur- red, and results of ichthyoplankton surveys indi- cated that P. sexfilis eggs and larvae are not found in inshore waters in Hawaii (Leis and Miller 1976; Miller et al.»; Watson and Leis»). Johannes ( 1978) ^Miller, J. M., W. Watson, and J, M. Leis. 1973. Larval fishes. In S. V. Smith (editor), Atlas of Kaneohe Bay, a reef ecosystem under stress, p. 101-105. Univ. Hawaii Sea Grant Tech. Rep. 72-1. Sea Grant College Program, University of Hawaii, Honolulu, HI 96822. ^Watson, W., and J. M. Leis. 1974. Ichthyoplankton of Kaneohe Bay, Hawaii: a one-year study of the fish eggs and larvae. Univ. Hawaii Sea Grant Tech. Rep. 75-1, 178 p. Sea Grant College Program, Universitv of Hawaii, Honolulu, HI 96822. 903 pointed out that spawning on outgoing tides is a common phenomenon among coastal marine fishes in the tropics, and he believed it evolved as a strategy for ensuring that eggs and larvae are transported away from the heavy concentration of predators in shallow water. Johannes noted that nocturnal spawning is also common in tropical reef fishes and serves to reduce predation both on the eggs and on the spawners. The first P. sexfilis spawning of the year appears to be anomalous in that it occurs relatively late with respect to the last quarter. If offshore trans- port confers an important selective advantage on P. sexfilis, the lateness of the first spawning is maladaptive because it results in release of eggs early relative to the outgoing tide (see Figure 3). The initial phase of the spawning season may thus result in few viable offspring and could represent a gradual initiation of the main spawning season, delayed perhaps by the lower water temperatures which usually prevail during the first spawning month (Batheni"). No observations on the spawning behavior of a polynemid fish have been published previously. In P. sexfilis the sexes apparently pair and spawn after a brief courtship involving rapid following and nosing of one fish by another. The spawning behavior of this species is similar in many respects to that of the Pacific bonito, Sarda chiliensis, in- cluding behaviors described by Magnuson and Prescott (1966) as "circle swimming," "tail nos- ing," and "following." The circling behavior noted among P. sexfilis may have been imposed by the confinement of the net enclosure, but S. chiliensis also showed tight circling behavior at the time of gamete release in a very large tank at Marineland of the Pacific, and circling prior to spawning occurs naturally in mullets ( Helfrich and Allen 1 975) and some (perhaps many) other tropical fishes (R. E. Johannes, Hawaii Institute of Marine Biology, P.O. Box 1346, Kaneohe, HI 96844. Pers. com- mun., December 1977). The circling behavior dur- ing spawning observed in captive P. sexfilis thus may not be abnormal for this species but may, as Magnuson and Prescott (1966) theorized for S. chiliensis, serve to enhance the probability of fer- tilization. '-^ Acknowledgments We wish to thank R. E. Johannes for reading and commenting on the manuscript, and Lloyd Watarai, Steven Shimoda, Michael Muranaka, and Michael Matsukawa for their help in main- taining fish and collecting data. This work was supported by the University of Hawaii Sea Grant College Program, NOAA, Office of Sea Grant, U.S. Department of Commerce, under Grant No. 04-3- 158-29, 04-5-158-17, 04-6-158-44026, and 04-6- 158-441 14, and by the State of Hawaii through the Hawaii Insitute of Marine Biology, the Office of the Marine Affairs Coordinator, and the Depart- ment of Planning and Economic Development. Literature Cited Helfrich, P., and P. M. Allen. 1975. Observations on the spawning of mullet, Creni.*7iu^j/ crenilabis (Forskal), at Enewetak, Marshall Islands. Mi- cronesica 11:219-225. HOSAKA, E. Y. 1944. Sport fishing in Hawaii. Bond's, Honolulu, 198 p. Johannes, R. E. 1978. Reproductive strategies of coastal marine fishes in the tropics. Environ. Biol. Fishes 3:65-84. Leis, J. M., AND J. M. Miller. 1976. Offshore distributional patterns of Hawaiian fish larvae. Mar. Biol. (Berl.) 36:359-367. Lowell, N. E. 1971. Some aspects of the life history and spawning of the moi {Polydactylus sexfilis). M.S. Thesis, Univ. Hawaii, 45 p. MAGNU.SON, J. J., AND J. H. PRESCOTT. 1966. Courtship, locomotion, feeding and miscellaneous behaviour of Pacific bonito iSarda chiliensis). Anim. Behav. 14:54-67. May, R. C. 1976. Studies on the culture of the threadfin, Polydactylus sexfilis, in Hawaii, FAO Technical Conference on Aquaculture, Kyoto, Jpn., 26 May-2 June 1976. FIR:AQ/ Conf/76/E.5. FAO, Rome, 5 p. ROBERT C. May Hawaii Institute of Marine Biology Kaneohe, Hawaii Present address: Asian Development Bank P.O. Box 789, Manila, Philippines Gerald S. Akiyam.a Michael T. Santerre Hawaii Institute of Marine Biology P.O. Box 1346, Kaneohe, HI 96744 '"Bathen, K. 1968. A descriptive study of the physical oceanography of Kaneohe Bay, Oahu, Hawaii. Univ. Hawaii, Hawaii Inst. Mar. Biol., Tech. Rep. 14, 353 p. Hawaii Institute of Marine Biology, P.O. Box 1346, Kaneohe, HI 96744. 904 THE ROLES OF PRIOR RESIDENCE AND RELATIVE SIZE IN ( OMPEII LION FOR SHELTER BY LHE MALA^ SLAN I'RAW N, MAC ROliltKLlULM KOSISBHRGIl^ Behavorial dominance, tenitoiiality. and their re- lationship to survival and population density have been the subject of extensive research (reviewed by Brown and Orians 1970; Ito 1970; Brown 1975). Generally dominance (behavioral) hierarchies imply some form of ranked order (reviewed by Marler and Hamilton 1966; Eibl-Eibesfeldt 1970; Ito 1970) whereby the alpha animal(s) has prefer- red access to food, shelter, or mates. Dominance may develop within a short time after an initial encounter (Dingle and Caldwell 1969), is partially controlled by differences in relative size (Marler and Hamilton 1966). and in some species is mod- ified by relative location in space (Brown 1963). This latter modification is related to Noble's ( 1939) original concept of territory. Noble referred to territory as "any defended area." This area could serve as a "retreat" ( in contrast to a sexual or nesting area) that "is occupied because it is famil- iar and defended because any newcomer is irritat- ing to the resident." Such space-related aggressive behavior has been reported in numerous animals (Brown and Orians 1970). Territorial behavior can be related to "defense" of 1 ) a breeding area ( Buechner 1 96 1 ; Watson 1964); 2) a renewable resource such as food (Stimson 1970); or 3 ) a physical shelter I Crane 1958; Reese 1964; Fielder 1965; Hughes 1966; Dingle and Caldwell 1969). Often the out- come of such a defensive action is exclusion of the intruder by the resident. Since this area is "famil- iar" to the resident and unfamiliar to the "new- comer," it follows that the resident has some type of advantage. This "prior resident effect" has been observed in a number of species (Braddock 1945, 1949; Miller 1958; Hughes 1966; Baird 1968; Dingle and Caldwell 1969; Selander 1970). Thus in many animals, spacing behavior is a powerful mechanism that can regulate resource utilization and influence distribution patterns. Many of the above-mentioned studies and re- views dealt with animal populations in natural open systems subject to both immigration and emigration. In contrast, aquaculture systems are 'Contribution No. 544 from the Hawaii Institute of Marine Biology, University of Hawaii, Kaneohe, Hawaii. closed and deal with confined high-density popula- tions. In the case of Macrobrachi ion rosenbergii, ponds are stocked with postlarvae, and harvesting of adults begins 9-12 mo later. The same space- related behavioral mechanisms observed in open systems may be operating in these high-density ponds. Circumstantial evidence indicates that this is occurring in ponds containing M. rosenbergii. Animals of the same age exhibit large variation in size at the end of several months of growth (Fujimura nad Okamoto 1970). Malecha (1977) reported that small M. rosenbergii can greatly in- crease their size when larger animals are absent. This has been called the "Bull Effect" by Fujimura and Okamoto (1970). Similar observations have been reported for carp (Nakamura and Kasahara 1955, 1961; Wohlfarth and Moav 1972), trout (Brown 1946), and salmon (Symons 1971). One hypothesis advanced by Nakamura and Kasahara ( 1961) is that the larger animals are outcompeting the smaller subordinates for food. Maerobraehiiini rosenbergii is a large freshwa- ter prawn. Its native distribution ranges from Pakistan to Papua, New Guinea, and Palau (Johnson 1960; McVey 1975). Usually it is found in fresh and brackish streams and pools. The eggs hatch near ocean waters, and the adults are found up to 200 km from the coast (Ling 1969). Generally males are thought to stay in upstream waters while the females undergo a seasonal migration, moving downstream and into brackish waters (Raman 1967). Relatively little is known of M. rosenhergii's behavioral ecology but Raman ( 1964) reported juveniles "hiding in crevices or among submerged plants along river banks." In order to understand how social behavior affects resource utilization by M. rosenbergii. three experiments were conducted in which shelter was the limiting resource, and relative size and prior residence were measured as variables. Methods The three experiments consisted of: a prior resi- dent experiment, a simultaneous introduction ex- periment, and a control experiment. The prior res- ident experiment was used to test for the role of prior residence and relative size in competition for shelter. The simultaneous introduction experi- ment tested for the role of relative size on competi- tion in the absence of a "prior resident effect." The control experiment tested for the effect of handling and capture. FISHERY BULLETIN: VOL. 76. NO. 4. 1979. 905 Water conditions were maintained via an air lift filter. The animals were fed a dry pellet diet ap- proximately every other day (see diet #5, Balazs et al. 1973). A 12-12 photoperiod with one-half hour twilight lighting at "sunrise" and "sunset" was employed. Prior RLsiticnt 1 xpcrmicni Eai'lier experiments revealed a foi'm of shelter preference or selection operating in M. rosenhcrgii (Peebles 1977). The shelters used in this experi- ment were identical to those most frequently selected by animals in the earlier experiments. One shelter was placed in each experimental tank. A shelter consisted of six concrete bricks arranged into a double open ended square tunnel ( 19.3 x 19 X 11.4 cm tall). Refuge othei" than the shelter was eliminated by the use of oblong experimental tanks ( 137 x 75 x 92 cm deep) and the suspension of the air lift filters just below the water surface ( the usual position for these filters was on the bottom i. Water depth was 34 cm. Adults from commercial ponds were placed in two separate holding tanks, where they were kept for no longer that 1 wk. Two animals were re- moved, one each from the separate holding tanks. Three body characteristics were measured: stan- dard length (tip of telson to orbit of eye), and lengths of left and right chelae. The animals were tagged by means of a small plastic "bread bag twist-tie" that was color coded and tied around the tail. It took about 15 s to attach. Following tagging the two animals were placed separately in ex- perimental tanks. Three observations were made before the introduction of the "immigrant" and four observations were made after the introduc- tion. The preintroduction observations were made on the second, third, and seventh days after the animals were placed in their separate experimen- tal tanks. There were three observations per ani- mal, each lasting 3 min. After the preintroduction week a coin was flipped to determine which animal would be the immigrant. The immigrant was des- ignated as the introduced specimen and was moved via a dip net from its tank to the resident's tank. The resident was the animal that was not moved from one experimental tank to another. The postintroduction observations were made on the day of introduction, and the second, third, and seventh days after introduction. The observation performed on the day of introduction was 15 min and designed to monitor agonistic interactions as- sociated with the initial encounters of the paired animals. The remaining three postintroduction observations were 3 min each and designed to re- cord the animal's position within the tank. All observations were made between 1000 and 1530. Since these animals are nocturnal, movement and behavioral interactions were minimal during the daytime. A total of 36 animals (18 immigrants, 18 resi- dents) were used. Paired animals were of the same sex. This controlled for the possible confounding effect heterosexual courtship behavior might have on competition foi- shelter occupancy. Simultaneous Introdut tion Fxpcrinicnt The treatment of the simultaneous introduction experiment differed from the prior resident exper- iment in four ways: 1 ) only males were used; 2) the animals were simultaneously introduced into the oblong tanks; 3) two additional body characteris- tics were measured (body weight and carapace length); and 4) the animals were not separately observed prior to introduction. Fifteen trials were run employing a total of 15 pairs or 30 animals. Observations were made on the day of simultaneous introduction, and the sec- ond, third, and seventh days after introduction. The observation performed on the day of simul- taneous introduction was 15 min and designed to monitor agonistic interactions associated with ini- tial encounters of the paired animals. The remain- ing three postintroduction observations were 3 min each and designed to record the animal's posi- tion within the tank. Control Experiment Eleven controls were run to test the effect of handling. Animals were selected, measured, tag- ged, and placed individually in experimental tanks. One week later the control was netted, held in the air, and reintroduced into the same experi- mental tank. Observations were made for the week before and the week after netting (mock im- migration). Operation.il Definitions Successful: an animal that was in a shelter at the end of the 7-day period following immigration. Unsuccessful: an animal that was not in a shel- 906 ter at the end of a 7-day period following immigra- tion. Push: an aggressive act where one animal pushes one of its chelae against the body of another animal. Nip: an aggressive act where one animal closes down the tips of its chela on the body part of another animal. Tete-a-tete: a type of aggressive act charac- terized by a head to head confrontation with at least one nip or one push. The tete-a-tete appeared to be difficult enough in orientation from the push and the nip to be placed in a separate category. Further observation and analysis might not sup- port this separation. Shove: an aggressive act where one animal holds both chela forward and parallel while charg- ing into the flanks of another animal. Bout: an agonistic exchange between two ani- mals where at least one aggressive act occurred. A bout was considered terminated when aggressive acts stopped or one animal moved away and was not chased. Bouts were measured in units of ag- gressive acts. Bout length: the number of aggressive acts that occurred during a bout. Body characteristics: standard length (cen- timeters*, right and left chelae length (centime- ters), weight (grams), and carapace length (cen- timeters). Body size index: the number of body characteris- tics in which an animal was larger. It was derived as follows: animal A larger than animal B in stan- dard length and right chela length, then A's body size index is two. In the Prior Resident Experi- ment three body characteristics were measured, thus the maximum body size index in this experi- ment was three. In the Simultaneous Introduction Experiment five body characteristics were mea- sured, thus in this experiment the maximum body damage index was five. Results Control Experiment Ten out of 11 animals were in the shelter on every observation period before mock immigra- tion. The remaining animal was in the shelter on one of the three observation periods. The same 10 were in the shelters on all observations following mock immigration, while the same remaining one was never observed in a shelter after immigration. It was concluded that the act of netting had no effect on shelter use. Prior Resident Experiment Shelters were occupied on every observation by every animal during the preimmigration week. Following immigration all shelters were occupied on every observation period. On several occasions more than one animal was in a shelter during the first two observation periods following immigra- tion. However, by the end of the week, observation period 4, one animal was in a shelter while the other was usually at the opposite end of the tank. When the data were examined by immigrant ver- sus resident for shelter use over the 7-day period, an interesting change became apparent (Figure 1). On the day of immigration, residents were oc- cupying shelter significantly more often than im- migrants (Binomial Test,P = 0.044, Siegel 1956). By the second observation period and for the re- maining two observations there were no sig- nificant differences between residents and immi- gi-ants in frequency of shelter use (Binomial Test: day 2 after immigration,P = 1.0, day 3,P = 0.814; day 7,P = 0.814). Examining the data for the effect of size (Figure 2) revealed that successful animals were sig- nificantly larger than their unsuccessful paired 12 3 4 OBSERVATION PERIOD Figure l. — Shelter usage by observation period for 18 pairs of Macrobrachium rosenhergii. The data from prior resident exper- iments are summed for the 18 pairs. During observation period 1, 18 residents (circles) and 7 immigrants (dots) were inside shel- ters. On observation period 1 there were seven cases of double occupancy; observation period 2, two cases; and observation period 3, none. 907 => < -7 u < u Z) -> Ll_ 1/5 o ^ □£ U z) z 1 Z 3 ' BODY SIZE I NDEX Figure 2. — Frequency of relative body size for successful (open bar) and unsuccessful (solid bar) prawns in the prior resident experiment. A body size index of one indicates one Macro- hrachium rosenbergii was larger than the other in one body trait but smaller in the other two body traits. partners (Kolmogorov-Smirnov Two Sample Test: D max = 11,/? = 18,P = 0.01). SiniiiltaneoLi,>> Introduction Experiment A similar effect of size on shelter use was ob- served in the simultaneous introduction experi- ment (Figure 3; r, = 0.579, P<0. 001). Once again larger animals used the shelters more often than their smaller partners. Aggressive behavior was observed only on the day of introduction. The nip and push occurred more often than the shove or tete-a-tete ( Figure 4). Generally aggressive interactions were limited to a few (one to three) bouts per 15 min (Figure 5), and these bouts were usually one or two aggres- sive acts long (Figure 6). Prior Resident Experiment by Simultaneous Introduction Experiment A Kolmogorov-Smirnov chi-square approxima- tion (Goodman 1954; Siegel 1956) revealed that animals of the simultaneous introduction experi- ment were more aggressive on the day of introduc- tion than were animals in the prior resident exper- iment on the day of immigration (x^ = 15.54, P<0.002 for number of bouts/animal per 15-min period; x^ - 13.877, P<0.002 for number of ag- 40n 3 5- N = NIP P = PUSH T=TETE-A-TETE >30- S= SHOVE Z25- LJ 320- O ^15- ^10- 5- 1 N T 49 to Qi m ^!^3 o LU CO Z) O >2 LU LU -^ CO 1-20- U O Li. 5- 1 2 3 4 5 6 7 8 BOUT LENGTH Figure 6. — Frequency of bout length ( number of aggressive acts per bout per animal) for the simultaneous introduction of 14 pairs of male Macrobrachium rosenbergii. gressive acts/animal per 15-min period). Simul- taneous introduction animals exhibited a total of 68 aggressive acts occurring in 43 bouts in =28 animals), while the prawns from the prior resident experiment exhibited only five aggressive acts in five bouts in = 32 animals). Discussion The results indicate that when M. rosenbergii compete for shelter at least three factors, i-elative size, prior residence, and length of time contes- tants are paired, play important roles in determin- ing who occupies a shelter. It has long been recog- nized that in crustaceans relative size plays a large role in determining dominance (Allee and Douglas 1945; Bovbjerg 1953, 1956, 1960; Lowe 1956). More recent observations have confirmed the size dominance relationship (Hughes 1966; Crane 1967; Griffin 1968; Hazlett 1968; Dingle and Caldwell 1969; Warner 1970; Rubenstein and Hazlett 1974; Jachowski 1974; Molenock 1976; Sinclair 1977). However, relative size does not ap- pear equally important in all species (Hazlett and Estabrook 1974). In prawns, relative size strongly influences the outcome of competition. When two prawns en- counter one another in an area new to both, the larger animal usually has the advantage. Often these encounters are characterized by a limited series of pushes with one or the other chela. The function of the pushing might be threefold: 1) to test their opponent's weight (rest inertia), 2) to determine the opponent's molt state, and 3) to see if the opponent is capable of pushing back (has chelae). Other crustaceans appear to measure their opponent's physical strength by means of physical interactions involving the chelae (Griffin 1968; Schone 1968). In Cambarellus shufeldtii, claw removal causes dominant animals to drop in rank (Lowe 1956). InM. rosenbergii deaths related to agonistic behavior usually occurred near ec- dysis and often the first appendages lost during an agonistic encounter were the chelae (Peebles 1977). Smaller animals have been observed success- fully defending shelters from attempted occupa- tion by larger congeners (Bovbjerg 1953; Griffin 1968; Sinclair 1977). This is related to the prior resident phenomenon and it is central to Nobel's (1939) definition of territory. Resident M. rosen- bergii, regardless of their relative size, success- fully retained their shelters. The mechanism the residents employed apparently was not limited to direct physical interaction. Immigrants and resi- dents seldom fought. Generally immigrants were inactive upon placement into a tank housing a resident. The immigrant's aggressive behavior was well below its counterpart in the simultane- ous introduction group. Only occasionally (Figure 1) did the immigrant seek out the shelter. This latter behavior is in direct contrast to the control gi'oup. A control group animal was usually back in its shelter within 1 min after reintroduction. Pos- sibly an exocrine was an agent of communication between resident and immigrant prawns, since a novel environment did not inhibit exploration in animals of the simultaneous introduction experi- ment; and animals from the control experiment reintroduced into tanks contaminated with their own exocrines, rapidly entered their shelter. The advantage conferred upon resident M. rosenbergii appears to disappear within a short period of time. The smaller resident can defend its shelter against intrusion for no longer than a few days (Figure 1). Apparently relative size can over- come the prior resident effect if resident and im- migrant continue to encounter one another. Simi- lar observations were reported by Lowe ( 1956). In the case of the C. shufeldtii, a dominance hierar- chy was established before shelters were intro- duced. Dominant C. shufeldtii displaced subordi- nates from occupied shelters. In my experiments, M. rosenbergii first exhibited territoriality as de- termined by the presence of the prior resident ef- fect. Territoriality then broke down, due to con- tinued encounters, into simple dominance. 909 The important point addressed in this paper is not who wins or loses the encounter but which animal gains access to the I'esource. Investigators whose observations w^ere limited to the first en- counter might suggest that residents almost al- ways outcompete intruders foi- shelter. However. 1 have shown that in a closed system the prior resi- dent effect breaks down into simple size-related dominance. These results offer a behavioral ex- planation for the known and i-ecognized bull effect in praw n aquaculture ponds. Larger animals have preferential access to food and shelter, two impor- tant resources which are often dispersed in a clumped or patchy fashion. Acknow Icdgnicnts I wish to thank E. Reese, A. Kinzie, S. Malecha, R. May, T. Smith, and the editors and reviewers of Fishery Bulletin for the advice they gave during the development of this paper. Most of the work was carried out at the Hawaii Institute of Marine Biology with funds awarded by University of Hawaii Sea Grant College Program (through In- stitutional Grant No. 04-3-158-29 from NOAA Office of Sea Grant) and administered through the Department of Genetics at the University of Hawaii. Literature Cited allee, w. C, and M. B. DOUGLIS. 1945. A dominance order in the hermit crab, Pagurus lon- gicarpus Say. Ecology' 26:411-412. B.AL.AZS, G. H., E. Ross, .AND C. C. BROOKS. 1973. Preliminary studies on the preparation and feeding of crustacean diets. Aquaculture 2:369-377. BAIRD, R. C. 1968. Aggressive behavior and social organization in Mol- lienesia latipinna Le Sueur. Texas J. Sci. 20:157-176. BOVBJERG. R. V. 1953. Dominance order in the crayfish Orconectes virilis (Hagen). Physiol. Zool. 26:173-178. 1956. Some factors affecting aggressive behavior in crayfish. Physiol. Zool. 29:127-136. 1960. Behavioral ecology of the crab, Pachygrapsus cras- sipes. Ecology 41:668-672. BRADDOCK, J. C. 1945. Some aspects of the dominance-subordination rela- tionship in the fish Platypoecilus maculatus. Physiol. Zool. 18:176-195. 1949. The effect of prior residence upon dominance in the fish Platypoecilus 7Tiaculatus. Physiol. Zool. 22:161-169. Brown, J. L. 1963. Aggressiveness, dominance and social organization in the Stellar Jay. Condor 65:460-484. 1975. The evolution of behavior. W. W. Norton & Co.. N.Y., 761 p. BKOWN, J. L., AND G. H. ORIANS. 1970. Spacing patterns in mobile animals. Annu. Rev. Ecol. Syst. 1:239-262. Brown, M. E. 1946. The growth of brown trout tSalmo Truita Linn.) I. Factors influencing the growth of trout fry. J. Exp. Biol. 22:118-129. BUECHNKR. H. K. 1961. Territorial behavior in Uganda Kob. Science (Wash., DC.) 133:698-699. Crank, J. 1958. Aspects of social behavior in fiddler crabs, with spe- cial reference to Uca maracoani (Latreille). Zoologica (N.Y.) 43:113-130. 1967. Combat and its ritualization in fiddler crabs • Ocypodidae) with special reference to Uca rapax (Smith). Zoologica (N.Y.) 52:49-76. Dingle, H., and R. L. Caldwell. 1969. The aggressive and territorial behaviour of the man- tis shrimp Gonodactylus bredini Manning (Crustacea: Stomatopoda). Behaviour 33:115-136. ElBL-ElBESFELDT, L 1970. Ethology, the biology of behavior. Holt, Rinehart and Winston, N.Y., 530 p. FIELDER, D. R. 1965. A dominance order for shelter in the spiny lobster Jasus lalandei (H. Milne-Edwards). Behaviour 24:236- 245. Fujlml'ra, T., and H. Okamoto. 1970. Notes on progress made in developing a mass cultur- ing technique for Macrobrachium rosenbergii in Hawaii. Indo- Pacific Fisheries Council (FAO), Thai- land. G<:)ODMAN, L. A. 1954. Kolmogorov-Smirnov tests for psychological re- search. Psychol. Bull. 51:160-168. Griffin, D. j. G. 1968. Social and maintenance behaviour in two Austra- lian ocypodid crabs (Crustacea: Brachyura). J. Zool. (Lond.) 156:291-305. Hazlett, B. a. 1968. Size relationship and aggressive behavior in the hermit crab Clibanarius vittafus. Z. Tierpsychol. 25:608-614. Hazlett, B. A., and G. F. Eastbrook. 1974. Examination of agonistic behavior by character analysis. I. The spider crab Microphrys bicornutus. Be- havior 48:131-144. HI'(;HES, D. a. 1966. Behavioural and ecological investigations of the crab Ocypode ceratophthalmus (Crustacea: Ocyp- odidae). J. Zool. (Lond.) 150:129-143. ITO, Y. 1970. Groups and family bonds in animals in relation to their habitat. In L. R. Aronson, E. Tobach, D. S. Lehrman, and J. S. Rosenblatt (editors). Development and evolution of behavior, p. 389-415. W. H. Freeman, San Franc. JACHOWSKI, R. L. 1974. Agonistic behavior of the blue crab, Callinectus sapidus Rathbun. Behaviour 50:232-253. Johnson, D. S. I960. Some aspects of the distribution of freshwater or- ganisms in the Indo-Pacific area, and their relevance to the validity of the concept of an Oriental region in 910 zoogeography. Proc. of the Centenary and Bicentenary Congress of Biology, Singapore, 1958, p. 170-181. Ll.\(;, S. \V. 1969. The general biology and development of Macru- hrachium rosenhi'rgii ( de Mani. FAO Fish. Rep. 57:589-606. Lowe, M. E. 1956. Dominance-subordinance relationships in the crawfish Camharellus shufeldtit. Tulane Stud. Zool. 4:139-170. M.M.KCH.A, S. R. 1977. Genetics and selective breeding of M. rusenber- gii. In J. A. Hanson and H. L. Goodwin (editors), Shrimp and prawn farming in the Western Hemisphere, p. 328- 355. Dowden, Hutchinson & Ross, Inc., Stroudsburg, Pa. M.ARLKR, P., .AND W. J. H.AMILTON III. 1966. Mechanisms of animal behavior. John Wiley & Sons, N.Y., 771 p. McVey, J. p. 1975. New record of Minn)brachium rosenbergii ( De Mani in the Palau Islands (Decapoda, Palaemonidaei. Crus- taceana 29:31-32. MILLER, R. B. 1958. The role of competition in the mortality of hatchery trout. J. Fish. Res. Board Can. 15:27-45. MOLEN'OC'K, J. 1976. Agonistic interactions of the crab Pe/ro/;.s7/!es (Crus- tacea, Anomura). Z. Tierpsychol. 41:277-294. N.AK.A.ML'R.A, N., .AN'D S. K..\.S,AH.-\R.\. 1955. A study on the phenomenon of the tobi-koi or shoot carp — I. On the earliest stage at which the shoot cai-p appears. Bull. Jpn. Soc. Sci. Fish. 21:73-76. 1961. A study on the phenomenon of the tobi-koi or shoot carp — IV. Effects of adding a small number of larger indi- viduals to the experimental batches of carp fry and of culture density upon the occurrence of shoot carp. Bull. Jpn. Soc. Sci. Fish. 27:958-962. Noble, G. K. 1939. The role of dominance in the social life of birds. Auk 56:263-273. Peebles, J. B. 1977. Movement, dispersion and mortality in Macro- brachium rosenbergii. Ph.D. Thesis. University of Hawaii, Honolulu, 220 p. Raman, K. 1964. On the location of a nursery ground of the giant prawn Macrobrachium rosenbergii (de Man). Curr. Sci. (India) 33:27-28. 1967. Observations on the fishery and biology of the giant freshwater prawn Macrobrachium rosenbergii de Man. In Symposium on Crustacea, Pai-t I, p. 649-669 Mar. Biol. Assoc. India, Ernakulum. Reese, E. S. 1964. Ethology and marine zoology. Oceanogr. Mar. Biol. Annu. Rev. 2:455-488. RLBENSTEIN, D. I., AND B. A. HAZLETT. 1974. Examination of the agonistic behaviour of the crayfish Orconectes virilis by character analysis. Be- haviour 50:193-216. SCHONE, H. 1968. Agonistic and sexual display in aquatic and semi- terrestrial brachyuran crabs. Am. Zool. 8:641-654. SELANDER, R. K. 1970. Behavior and genetic variation in natural popula- tions. Am. Zool. 10:53-66. SlEGEL, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill Book Co., N.Y., 312 p. SLNCLAIR, M. E. 1977. Agonistic behaviour of the stone crab, Menippe mer- cenaria (Say). Anim. Behav. 25:193-207. Stlmson, J. 1970. Territorial behavior of the owl limpet, Lottia gigantea. Ecology 51:113-118. symons, p. E. K. 1971. Behavioural adjustment of population density to available food by juvenile Atlantic salmon. J. Anim. Ecol. 40:569-587. WARNER, G. F. 1970. Behaviour of two species of grapsid crab during in- traspecific encounters. Behaviour 36:9-19. W.ATSON, A. 1964. Aggression and population regulation in red grouse. Nature (Lond.) 202:506-507. WOHI.FARTH, G. W., AND R. MOAV. 1972. The regression of weight gain on initial weight in carp. I. Methods and results. Aquaculture 1:7-28. John B. Peebles Hawaii Institute of Marine Biology University of Hawaii Kaneohe. HI 96744 PRINCIPAL SPAWNING AREAS AND TIMES OF MARINE FISHES, CAPE SABLE TO CAPE HATTERAS The purpose of this compendium is to summarize spawning areas and seasons of the more abundant marine fishes of the continental shelf between Cape Sable, N.S., and Cape Hatteras, N.C., as an aid to the identification offish eggs and larvae and planning and scheduling ichthyoplankton sur- veys. We have used the term "marine" to encom- pass fishes which spawn at sea (in contrast to es- tuarine spawners), although some of the species included spawn in both environments contingent on geographic location ( e.g., winter flounder which spawn exclusively in estuaries in the Middle At- lantic Bight and offshore in the Gulf of Maine and Atlantic menhaden which spawn in estuaries along southern New England and in the New York Bight and offshore in the lower Middle Atlantic Bight and in the South Atlantic Bight). The Gulf of Maine is defined as the oceanic bight bounded by Nantucket Shoals and Cape Cod on the west (long. 70°W) and Cape Sable on the east (long. 65°W) including Georges and Browns Banks and waters out to the 200-m contour (Col- ton 1964). The Middle Atlantic Bight is the area FISHERY BULLETIN VOL. 76, NO. 4. 1979. 911 inshore of the continental slope bounded by Cape Cod and Nantucket Shoals to the east (long. 70°W ) and Cape Hatteras to the south (lat. 35°N). The New York Bight, as defined in the MESA New York Bight Atlas Monograph Series (Bowman and Wunderlich 1977), is the offshore waterarea in the bend of the Atlantic coastline from Long Island (long. 7r30'W) to New Jersey (lat. 38°30'N). A chart of the Gulf of Maine and Middle Atlantic Bight and the names of places and areas referred to in the spawning summary are given in Fig- ure 1. 76* 74* 72* 70' 68' 66' 46* 44' 42< 40* 38* 36* 76** 74** 72** 70" 68" FIGURE 1.— Orientation chart of the Gulf of Maine and Middle Atlantic Bight. 46* 44' 42* 40* 38* 36* 6 6* 912 In this summary (Table 1) we have treated the Gulf of Maine and the Middle Atlantic Bight sepa- rately for there is an abrupt general division be- tween the biological and physical properties of water east and west of Cape Cod. The boreal wa- ters over most of the Gulf of Maine are well mixed by strong tidal currents, while the circulation of the warmer shelf waters west of Cape Cod is more sluggish, and its chemical and physical properties are less complex (Colton 1964). The offing of Cape Cod also appears to be a definite transition zone (probably thermal) for some northern and south- ern species of fishes and invertebrates, both pelagic and benthic (Colton 1964). The species composition and abundance of fishes vary marked- ly between the two regions, with boreal, non- migratory species dominating the Gulf of Maine and warmwater, migratory species prevailing in the Middle Atlantic Bight. The bulk or total spawning of many species of fishes is restricted to areas east (e.g., haddock, pollock, redfish) or west (e.g., bluefish, menhaden, anchovies) of Nantuck- Table 1. — Principal spawning areas and times of marine fishes. Cape Sable to Cape Hatteras. Family Species Comon Name Gulf of Maine Sub Area J .lantic Right Sub Area J F M A M J J A S 0 N D F M A M J J A b 0 N u Clupeidae Brevoortia tyrannus Atlantic menhaden t * * * Clupea harenqus harenqus Atlantic herring Georges Bank * * N. of Delaware i j * 1 Western Nova Scotia ! * 1 1 i i i 1 1 \ 1 1 Jeffries Ledge & ; i ' ' Stellwagen Bank i i , K- ! 1 i 1 1 1 Engraulidae Anchoa hepsetus striped anchovy Nantucket Shoals ' ! 1 j i 1 1 * • 1 1 ' j • Engraulis eurystole silver anchovy 1 1 , Gadidae Brosme brosme Enchelyopus cimbrius cusk fourbeard rockling * * * * 1 Gadus morhua ! Melanograiimus aeqlefinus Atlantic cod haddock Georges Bank Browns Bank Nantucket Shoals Georges Bank * 1 — k- -- * * ! 1 I 1 1 ' i * i * * i : . . , , Browns Bank ' * ■ * 1 . ' ; ! ,.,... I South Channel * , I ^ ' 1 1 I Merlucdus albidus offshorp hake 1 i i ! "TT^ -i ■ 1 Merluccius bilinearis silver heke NE Georges & Cent. Gulf ; * * Nant. Shoal s- |Virginia j - * 1 ! 1 1 Southern Georges ■ * , * i i • ' ; ■ ! 1 1 ' i 1 1 Pollachius virens 1 pollock Mass. Bay _^ Stellwagen South Channel * . • • : i i 1 ' Urophycis 1 Chester! long finned hake 1 . ■ i : 1 ■ NY Bight ,,, [ '. . 1 ! * i : i j ' i ' i 'l 1 1 1 1 1 1 1 " j Urophycis chuss red hake S. Georges Nant. Shoals ; ; * * 1* { ,' * '* 1 ■ Urophycis reqius spotted hake i NY Bight- , 1 C. Hatteras * i Urophycis tenuis Pomatomus saltatrix white hake bluefish I Cont. Slope * ' * ___ * J i ■ . ! 1 i * * 1 1 Scianidae Leiostomus xanthurus spot j , i Ches. Bay- 1 1 Cape Hatteras . i : : * Micropogon undulatus Atlantic croaker 1 Ches. Bay- Cape Hatteras * ' * 1 Cynoscion 1 regal is weakfish ; Ches. Bay- I 1 , 1 1 ; ! 1 Montauk, LI 1 1 ' 1 1 ill 1 1 913 Table L— Continued. Family Species Common Name Gulf of Maine Middle Atlantic Biqht Sub Area J F M A H J J A 5 0 N D Sub Area J F M A M J J A S 0 N D Labridae Tautoqa onitis tautog cunner Mass. Bay S. Georges Nant. Shoals * -- * \ Tautogolabrus ■k . adspersus i Scombri dae Scomber scombrus Atlantic mackerel W. Gulf Cape Cod Bay 1 i * Cape Cod- Chesapeak Bay ' * i Scorpaenidae Sebastes marinus redfish Scotian Shelf & * * ] 1 i ! Cent. Gulf i ' Triglidae Prionotus carolinus northern searobin i ' 1 , ' 1 Block Island- 1 j 1 Cape Hatteras Cottidae Hyoxocephalus octodecem- spinosus longhorn sculpin * ' 1 i : 1 1 i * 1 i i ; Amnodytidae Stromateidae Ammodytes sp. Peprilus triacanthus sand lance * * J ' ' u ^ * butterfish SW Georges Nant. Shoals : * * 1 * * j 1 1 1 1 Bothidae Citharichth^s arctitrons Gulf Stream flounder SW Georges ' Nant. Shoals . .„-ui-4 1 :■ 1- 1 1 i*i 1 i Hippoglossina oblonqa fourspot flounder Nant. Shoals- ' ' ' 1 : * 1 South ' ; j ■ I 1 ; I 1 Paralichthys dentatus summer flounder Nant. Shoals- ' ' _^_ South ' 1 ¥ 1 1 1* : i 1 ' Scophthalinus aquosus windowpane Georges Bank | Nant. Shoals- ! South ! -- -- " -- ' i ! i i ' ' * Pleuronec- tidae Glyptocephalus cynoglossus witch flounder 1 * J ! : Cape Cod- Delaware Bay * * 1 ! 1 1 Hippoqlossoides platessoides American plaice 1 i * * South of Martha's Vine- yard - -- 1 1 1 Limanda ferruqinea yellowtail flounder Browns Bank | j ! 1 . ' : 1 ' * 1 ; * * 1 1 f^" t - 1 Georges Bank Nant. Shoals- South 1 * * 1 ' 1 1 1 1 1 1 1 ! ; 1 I Pseudopleuro- nectes americanus winter flounder Georges Bank 1 1 1 i j 1 ' r 1 1 Known spawning season. Uncertain spawning season. *Peak spawning. et Shoals, although there are exceptions to this general rule (notably, yellowtail flounder and silver hake). The spawning summaries are based primarily on published data collected on Bureau of Commer- cial Fisheries (now National Marine Fisheries Service) ichthyoplankton surveys of the Gulf of Maine and Middle Atlantic Bight made in the 1950's and 1960's and listed in the References. Published data from earlier studies (e.g., Fish 1929; Walford 1938; Pearson 1941; Sette 1943; Bigelow and Schroeder 1953) and some unpub- lished information from more recent National Marine Fisheries Service ichthyoplankton sur- veys have also been utilized. We have not at- tempted to make the bibliography encyclopedic. However, the papers cited include references to all pertinent spawning summaries. Spawning areas and seasons were determined on a basis of the occurrence of eggs and/or early stage (yolk-sac) larvae. The families are arranged in phyletic sequence (Greenwood et al. 1966) and the species are listed in alphabetical order. Common names follows those recommended by Bailey et al. ( 1970). References Bailey, R. M., J. E. Fitch, E. S. Herald, E. A. Lachner, C. C. LINDSEY, C. R. ROBINS, AND W. B. SCOTT. 1970. A list of common and scientific names of fishes from the United States and Canada. 3d ed. Am. Fish. Soc. Spec. Publ. 6, 150 p. 914 BIGELOVV, H. B., AND W. C. SfHROKDKK. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Bowman, M. J., and L. D. Windkki.ich. 1977. Hydrographic properties. MESA N.Y. Bight Atlas Monogr. 1, 78 p. COLTON, J. B., Jr. 1964. History of oceanogi-aphy in the offshore waters of the Gulf of Maine. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 496, 18 p. CoLTON, J. B., Jr., and J. M. St. Onck. 1974. Distribution of fish eggs and larvae in continental shelf waters. Nova Scotia to Long Island. Ser. Atlas Mar. Environ., Am. Geogr. Soc. Folio 23. Fahay, M. p. 1974. Occurrence of silver hake, Merluccius hilinearis, eggs and larvae along the Middle Atlantic continental shelf during 1966. Fish. Bull., U.S. 72:813-834. FISH, C. J. 1929. Production and distribution of cod eggs in Mas- sachusetts Bay in 1924 and 1925. U.S. Bur. Fish., Bull. 43(21:253-296. Greenwood, P. H.. D. E. Ro.sen, S. H. Weitzman, and G. S. Myers. 1966. Phyletic studies of teleostean fishes, with a provi- sional classification of living forms. Bull. Am. Mus. Nat. Hist. 131:339-455. HiGH.\M, J. R., AND W. R. Nicholson. 1964. Sexual maturation and spawning of Atlantic menhaden. U.S. Fish Wildl. Serv., Fish. Bull. 63:255- 271. Kendall, A. W., Jr., and J. W. Reintjes. 1975. Geographic and hydrographic distribution of Atlan- tic menhaden eggs and larvae along the Middle Atlantic coast from RV Dolphin cruises, 1965-66. Fish. Bull., U.S. 73:317-335. Marak, R. R., and J. B. Colton, Jr. 1961. Distribution offish eggs and larvae, temperature, and salinity in the Georges Bank-Gulf of Maine area, 1953. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 398, 61 p. Marak, R. R., J. B. Colton, Jr., and D. B. Foster. 1962. Distribution of fish eggs and larvae, temperature, and salinity in the Georges Bank-Gulf of Maine area, 1955. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 411, 66 p. Marak, R. r., J. B. Colton, Jr., D. B. Foster, and D. Miller. 1962. Distribution of fish eggs and larvae, temperature, and salinity in the Georges Bank-Gulf of Maine area, 1956. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 412, 95 p. NoRCROss, J. J., S. L. Richardson, W. H. Massman, and E. B. Joseph. 1974. Development of young bluefish iPomatomus salta- trix) and distribution of eggs and young in Virginian coastal waters. Trans. Am. Fish. Soc. 103:477-497. Pearson, J. C. 1941. The young of some marine fishes taken in lower Chesapeake Bay, Virginia, with special reference to the gray sea trout, Cynoscion regalis (Bloch). U.S. Fish Wildl. Serv., Fish. Bull. 50:79-102. Richards, S. W., and a. W. Kendall, Jr. 1973. Distribution of sand lance, A/?!woc/y/t's sp., larvaeon the continental shelf from Cape Cod to Cape Hatteras from RV Dolphin surveys in 1966. Fish. Bull., U.S. 71:371-386. Richardson, S. L., and E. B. Joseph. 1973. Larvae and young of western north Atlantic bothid flatfishes Etropus microstomus and Citharichthys arcti- frun^ in the Chesapeake Bight. Fish. Bull., U.S. 71:735-767. SE'H'E, O. E. 1943. Biology of the Atlantic mackerel wScomber scom- brus) of North America. Part 1: Early life history, includ- ing growth, drift, and mortality of the egg and larval populations. U.S. Fish Wildl. Serv., Fish. Bull. 50:149- 237. S.MITH, W. G. 1973. The distribution of summer flounder, Paralichthys dentatus, eggs and larvae on the continental shelf be- tween Cape Cod and Cape Lookout, 1965-66. Fish. Bull., U.S. 71:527-548. S.MITH, W. G., J. D. SIBUNKA, AND A. WELLS. 1975. Seasonal distributions of larval flatfishes (Pleuro- nectiformes) on the continental shelf between Cape Cod, Massachusetts, and Cape Lookout, North Carolina, 1965-66. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-691, 68 p. JOHN B. Colton, Jr. Northeast Fisheries Center Narragansett Laboratory National Marine Fisheries Service. NOAA Narragansett. RI 02882 Wallace G. Smith Arthur W. Kendall, Jr. Peter L. Berrien Michael P. Fahay Northeast Fisheries Center Sandy Hook Laboratory National Marine Fisheries Service. NOAA Highlands. NJ 07732 RECENT SIGHTINGS OF THE BLUE WHALE, BALENOPTERA MUSCULUS, IN THE NORTHEASTERN TROPICAL PACIFIC The blue whale, Balenoptera rnusculus, in the North Pacific, migi'ates to the Gulf of Alaska and Aleutians in the summer for feeding (Nishiwaki 1966). It is believed to migrate to tropical waters in winter for calving, but sightings of blue whales in lower latitudes are rare (Tomilin 1957). In mid-July 1928, Cruikshank reported seeing ". . . several blue whales . . ." at lat. 11°32'N and long. 91°58'W (Kellogg 1929). A Peruvian fishery reported taking 247 blue whales between De- cember 1925 and March 1926 (Ingebrigtsen 1929). Potentially these were from a North Pacific stock, since the Southern Hemisphere blue whale is most numerous in the Antarctic at this time. Volkov and Moroz (1977) noted an abundance of baleen fishery BULLETIN: VOL. 76, NO. 4, 1979. 915 whales between lat. 7° and 10°N. Although indi- vidual species of baleen whales were not enumer- ated by Volkov and Moroz, two sightings of blue whales were made on 29 March 1975 and are pre- sented here (Table 1, Vni/shitelnyi cruise). Typically blue whales are seen along the Baja California coast in October while migrating southward, and subsequently reappear off Baja California in large numbers in March-June on their northward migration (Rice 1974). The whereabouts of the North Pacific blue whales dur- ing the winter months is completely unknown, but this is probably due to the lack of sighting effort. For instance, Japanese whale scouting has been carried out systematically since 1965, but their effort has been restricted to the Pacific waters north of lat. 20°N (Wada 1977). Two theories have evolved regarding the win- tering grounds of the blue whale. Wheeler ( 1946), suggested that blue whales winter within a lim- ited area of the subtropics. He maintained that whales congregate in large groups in areas not frequented by vessels. A second theory maintains that wintering blue whales disperse between the feeding grounds and the tropics (Harmer 1931; Mackintosh 1942). Presented in this note is a 3-yr record of blue whales sighted in the northeastern tropical Pacific. Reference is made to migration, whale groupings, behavior, and to the oceano- graphic features of the sighting area. These recent sightings were made by trained observers aboard vessels involved with the National Marine Fisheries Service Tuna/Porpoise Research Pro- gram. Sighting information from other experi- enced observers has also been contributed. Shipboard identification of rorquals is difficult and this problem was compounded by the fact that most of the sightings mentioned in this paper were incidental to ship's activities. However, the blue whale is easily discerned from other rorquals by the recognition of the following combination of characteristics: 1. Mottled blue-grey coloration. All other ror- quals are uniformly steel grey on the dorsal surface. 2. A small dorsal fin of varying shapes located in the posterior third of the body. The dorsal fin of the sei, fin, and brydes whales is larger, falcate shape, and placed farther anterior than the blue whale dorsal fin. 3. A U-shaped rostrum. The rostrum shape of other balenopterids is more pointed. 4. Tall, dense, disperse blows. Generally, the blow of the sei and brydes whales is low and dissi- pated, while the fin whale has a tall conical- shaped blow. A total of 1 1 cruises are discussed in this report, covering the period from January to May for 1971, 1 975, 1 976, and 1 977. 1 The area of effort and sight- ings of blue whales are reported in Figure 1. The 'No blue whale sightings were made in 1977. although two cruises have been included in Figure 1 to complement survey effort. Table l. — Annotated listofblue whale sightings by National Marine Fisheries Service observers in the northeastern tropical Pacific, January-May 1971-76. Date Lat., Long. No. of whales Observations Cruise/Observer 8 Jan, 1971 07 54N, 095 52W 1 Waufz/us Leatherwood 23 Jan, 1975 01 30N.083 04 W 1 Dove for 7 mm Pan Pacific Friedrichsen 3 Feb 1975 07=29 N. 093 48 W cow and calf 3 dives, 5 57 min/dive: lengtfi cow 27 m, calf 8-10 m Aquanus.Wade 4 Feb. 1975 07 48'N, 097 40W 1 Surfaced in middle of tuna scfiool F/nesferre Walker 7 Feb. 1975 07 45'N, 098 2rW 1 5 dives, 10.1 1 mm dive AquariusiWade 7 Feb. 1975 07 47'N, 098 24'W 1 2 dives, 8 19 mm dive Aquarius Wade 7 Feb, 1975 07 52'N. 098 47 W 2 Small unidentified wfiale, 10 m, witfi visible blow, swimming with a large blue whale AquariusiVJade 7 Feb 1975 07 47'N. 09900W 1 3 dives. 1 1 39 mm dive, exposed tail fluke on all dives Aquarius \Nade 9 Feb 1975 08 50'N, 096 04 W 4-6 Paired groups, length estimate: 27 m Pan Pacrf/c/Friedrichsen 10 Feb. 1975 0833N, 096 47'W 1 Pan Pac///c Friedrichsen 15 Feb 1975 0836'N, 096 29W 1 Pan Pac/l/c/Friedrlchsen 17 Feb 1975 08°58'N, 096°54W 8-10 Mostly pairs dispersed over several square miles; all headed northeast Pan Pac/ftc/Frledrichsen 17 Feb 1975 08 53N, 096 36W 1 Pan Pac/ftc, Friedrichsen 29 Mar 1975 08 55N, 093 34'W 10-13 Vnushitelnyi Rice 29 Mar 1975 09 07N. 093 55' W 6-7 Vnushitelnyi Rice 13 Feb 1976 09 44'N, 092 25'W 2 Cromwell Friedrichsen et al. 13 Feb 1976 09'44'N. 092 31 'W 1 Length estimate: 23 m Cromwell Friedrichsen et al. 13 Feb. 1976 09 44'N, 092"35'W 4-5 Whales dispersed over 4-5 mi^ Cromwell Friedrichsen et al. 13 Feb. 1976 09 44'N. 092 52 'W 1 Exposed tail flukes prior to sounding Cromwell Friedrichsen et al. 28 May 1 976 10 31'N,092 45'W 1 Length estimate: 20 m Martinac Friedrichsen 916 (PO'w no**.' I- -T 1 1 1 — h 1 1 n~l 1 1 1 1 1 1 — I 1 1 1 — t _,OOV -1 — — 1 — — , dO-w — 1 — 1 1 — 1 I 80-~ I — 1 — 1 i~ 3 3 1 S^^V i H— -T- 2 1Z^T\ .•si 5^ -r- 1 1 M Ik \ \ 1 1 1 1 1 1 216 Yk\ '^ \ z 1 II iM z 1) H; >x / — . _i_. - - -■ - -- -■ - -- __!_- - -- --- --- - - -- -- _^ 2 1 II Tl 3 4VJ^ \ MEXICO i 11 II 1 1 1 4 4 ixSi \ I '2 II 1 1 1 2 Z 3Z 2 n \ X I2 1 1 1 1 1 12 2 2 3 k ) V -f^ 1 — 1 1 zll 1 1 1 1_ 1 112 13 1 1 N \ 1 L 1 ' 12 < 2 .3 1 1 1 2 \ V. t "11' 1 1 1- 2 Z 2 2 2 1 3 \] \ ==%iH-v 4 11 1 IIZ ; Z1M2 1 3 2 1 V i 2 1 '' 1 Z 1 \ ' 1 1 1 3 1 1 1 2 I 2 z 21 _^ , / 1 1 1 111 I'lll Z123I 1 1 1 1 "r I I \ -- -- --- - ■- -■ • -- T""^"! "1 1 1 2 122224 1 1 Z 2 1 3 1 1 zTz 1 \ K i 1 2'l 1 ■ 13 2 13 1 4 4 1 1 2 \ 1 1 1 1 3 2: ^ p-^ ^ £\ NU'ARACtUA 111^ 11 ' Z 1 Z Z 1 1 Z 3 2 1 1 l|2 3 3 3 \ 321 M2l 211 ' IZZ\33 1 2 3 2|Z z ilz M 4 1 1 1 \ 2 3 Z 1 ' 1 1 1 1 i 1 1 3.a-P^..TDM 2 3 3 1 1 1 5 i 5 1 1 4|2 ,^.B 2 •^ 2 \ J Ws V 2I11M2 1 ll 226feZ32 3 2 3 3 2 2 Z 2 5 5 6 2{5 ^ 4 2 2 z \ 1 rt"* SU ^ u! \ '\ 1 J_;_L 11 M 7 M 2 1 1 3 1 2 2 2 3 6H 2M^3 1 1 2^ \ 1 2| v. NN/ U 1 1 ll 11 ' 2 3 12 2 1 3 2 5 4 3 5 ! VlZRiHi 2 3 z 2 3| M ~\ N ^ 111:1 1 Z 2 1 2 1 2 2 1 1 il 2 1 1 4 ^ M 2 z 1 1 2 1 1 : 1 111' 2 1 [ 2 2 1 1 12 1 1 1 .^.^i4 zlz 1 -■ "' 1" ■ ^ 1 1 1 1 ^ "2 1 1 1 |2 1 1 1 4 1 zll 1 2 ll||2l2l 11.1111131 1 1 1 1 1 1 1 1 1 3 1 1 2 1 111 2' II 1 1 1 3 Zll 1 2111 1 1 1 1 1 1 1 2 1 1 2 1 z 1 3 z /^ 1222! |2|i\ 11 1 1 2 1 Z M z Z D z i2IZl'lllMl 1 ' II 2433 3 2 1 1 1 1 ' 1 1 1 1 2 2 Z 1 1 1 1 Z 1 1 1 M 1 1 1 1 1 1 2 1 1 1 1 1 1 2 1 1 2 1 1 1 I OAl^p-VtO^ r^'^^o. 2 1 1 111 1 1 / ^ ; '■ 1 1 1 11 2 1 1 1 1 1 1 z: i 1 1 1 1 r • 1 1 1 1 1 z ] 1 2 2| 1 1 1 1 1 1 1 1 c I 1 Iz 2 I 1 1 I 2 2 1 z < 1 ' 1 1 1 1 111 Z ' 1 1 c J 1 1 1 1 1 1 i ' 1 1 1 2 1 1 1 1 1 1 1 1 1 ( 1 1 l( 1 1 1 2 1 1 1 2 2 1 ll 1 1 1 > 1 1 1 Z 1 1 1 1 1 1 1 ' i ' 1 Z 1 1 1 ll 1 1 1 \ 1 ' 1 1 1 1 2I 1 1 1 ^ : ! Ill 1 I 1 1 z 1 1 ll 1 1 1 1 i 1 I 1 ;l 1 1 ]l \ 11 1 ITT 1 1 1 +-3^ - O -icr Figure 1 . — Survey effort for blue whales in the northeastern tropical Pacific over 1 1 cruises from January to June in 197 1 , 1975, 1976, and 1977. Each numeral represents the total number of times that vessels entered a 1° block. Darkened blocks indicate blue whale sightings. fact that blue whales have been sighted in the same general area for three winters indicates that North Pacific blue whales may have a distinct wintering ground to which they migrate each year. The location of the suggested wintering grounds indicated by the sighting data are lat. 7°29'- 10°31'N and long. 95°25'-99°00'W.2 Only 1 of the 20 blue whale sightings (lat TSO'N and long. 83°04'W, 23 January 1975, a solitary individual) was outside of these bounds. The equatorial sight- ing location of this whale may indicate that either it was not from the North Pacific population, or that North Pacific blue whales do not restrict mi- gration to the hypothetical wintering grounds. ^Cruikshank's observations, in 1928, were also in this area. Whale Groupings and Behavior Blue whales are believed to be found singly or in pairs (Leatherwood et al. 1976). In fact, Nemoto ( 1964) reported that blue whales observed on the summer feeding grounds were solitary. However, five of the sightings reported here were aggrega- tions of whales dispersed over several square miles. Many of the whales were paired. The multi- ple sightings on 7 February 1975 and 13 February 1976 appear to be mostly of solitary animals (Ta- ble 1). However, on both days, no less than 40 n. mi. separated the first and last whales sighted. Also, four out of the five whales observed on 7 February 1975 were headed in a northeasterly direction. This information may indicate that these appar- ently solitary whales were part of a large dis- persed group. 917 At least one cow and calf were observed and possibly a second pair (Table 1). This is the first actual record of a blue whale calf in the tropics, although historically it has been believed that blue whales have their calves in the warm tropical waters (Mackintosh 1966:126).3 Steve Leatherwood, W. A. Walker, and D. W. Rice, who contributed blue whale sighting data and Karen J. Rice for technical assistance. Last, we thank Robert Schoning of the National Marine Fisheries Service for releasing this data for publi- cation. Oceanoyraphic Features Literature Cited There are several unique oceanographic fea- tures which relate to the sighting location of the blue whales. Cromwell (1958) and Wyrtki (1964) discussed the Costa Rican Dome which is located at approximately lat. 9°N, long. 89°W. The dome is apparently a permanent topographic feature (150 km X 300 km) and is formed by the convergence of several major current systems. These currents typically create an area of nutrient transport or upwelling. High standing stocks of zooplankton in the area near the Costa Rican Dome (lat. 7°25'N- 10°N) has been reported by several authors (Reid 1962; Blackburn et al. 1970; Holmes"*). Volkov and Moroz (1977) suggested that the high stable food base of the area creates a habitat suitable for nonmigratory populations of baleen whales. North Pacific blue whales may also use this area for their winter feeding grounds. In conclusion, the recent sightings of blue whales in the tropics indicates that North Pacific blue whales have a wintering area to which they return each year. Since most of the cruises have occurred largely during the winter months, more information must be collected to determine if whales are found in this area the year round. The high standing stock of zooplankton in this area may indicate that this is a winter feeding area, as well as a calving ground. AckiiDw ledgments We are particularly indebted to E. D. Mitchell, Fisheries Research Board of Canada, for his criti- cal evaluation of the manuscript. D. W. Rice, Na- tional Marine Fisheries Service; K. S. Norris, University of California at Santa Cruz, and Ron Garrett, Wilderness Research Institute, also made helpful suggestions on the draft. We also thank ^The average water temperature for 1 1 sightings was 26.5°C. ■'Holmes, R. W. 1970. A contribution to the physical, chem- ical, and biological oceanography of the northeastern tropical Pacific. (Unpubl. manuscr.) Institute of Marine Resources. Scripps Institute of Oceanography, Univ. Calif La Jolla, Calif AEC-UCSD-.34P99-4. BLACKBURN, M., R. M. LAURS, R. W. OWEN, AND B. ZEITZSCHEL 1970. Seasonal and areal changes in standing stocks of phytoplankton, zooplankton, and micronecton in the east- ern Tropical Pacific. Mar. Biol. (Berl.) 7:14-31. Cro.mwell. T. 1958. Thermocline topography, horizontal currents and "ridging" in the Eastern Tropical Pacific. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm. Bull. 3:133-164. HAR.MER. S. F. 1931. Southern whaling. Proc. Linn. Soc. Lond. 142:85- 163. INGEBRIGTSEN, A. 1929. Whales caught in the North Atlantic and other seas. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 56:1-26. Kellogg, R. 1929. What is known of the migrations of some of the whalebone whales. Smithson. Inst., Annu. Rep. 1928, p. 467-494. LEATHERWOOD, S., D. K. CALDWELL, AND H. E. WiNN 1976. Whales, dolphins, and porpoises of the western North Atlantic. A guide to their identification. U.S. Dep. Commer., NOAA Tech. Rep. NMFS CIRC-396, 176 p. Mackintosh, N. A. 1942. The southern stocks of whalebone whales. Discov- ery Rep. 22:197-300. 1966. Distributionof southern blue and fin whales. InK. S. Norris (editor), Whales, dolphins, and porpoises, p. 125-144. Univ. Calif Press, Berkeley. Nemoto, T. 1964. School of baleen whales in the feeding areas. Sci. Rep. Whales Res. Inst. Tokyo 18:89-110. NlSHlWAKL M. 1966. Distribution and migration of the larger cetaceans in the North Pacific as shown by Japanese whaling re- sults. In K. S. Norris (editor). Whales, dolphins, and por- poises, p. 170-191. Univ. Calif. Press, Berkeley. Reid, J. L. 1962. On circulation, phosphate-phosphorus content, and zooplankton volumes in the upper part of the Pacific Ocean. Limnol. Oceanogr. 7:287-306. RICE, D. W. 1974. Whales and whale research in the eastern North Pacific. In W. E. Schevill (editor). The whale problem: a status report, p. 170-195. Harvard Univ. Press, Cambr., Mass. TOMILIN, A. G. 1957, Zveri SSSR i prilezhashchikh stran Vol. IX Kitoob- raznye (Mammals of the U.S,S,R, and adjacent countries Vol. IX Cetacea). Akad, Nauk SSSR, Mosk,, 756 p, (Trans- lated by Isr. Program Sci. Transl., Jerusalem, 1967, 717 p.) 918 VOLKOV, A. F.. AND I. F. MOROZ. 1977. Oceanological conditions of the distribution of cetacea in the Eastern Tropical part of the Pacific Ocean. Int. Whaling Comm. Rep. 27:186-188. W.-\n,A. S. 1977. Indices of abundance of large-sized whales in North Pacific in the 1975 whaling season. Int. Whaling Comm. Rep. 27:189-192. WHEELER. J. F. G. 1946. Observations on whales in the South Atlantic Ocean in 1943. Proc. Zool. Soc. Lond. 116:221-224. WYRTKI. K. 1964. Upwelling in the Costa Rica Dome. U.S. Fish Wildl. Serv.. Fish. Bull. 63:355-372. Lawrence S. Wade Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service. NOAA La Jolla. Calif. Present address: P.O. Bo.x 4455 Areata. CA 95521 Gary L. Friedrichsen Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NOAA La Jolla. Calif. Present address: P.O. Box 890 Areata. CA 95521 A substantial sport fishery exists for white mar- lin in the Atlantic off North and South America. In the United States, the major sport fisheries occur along the Middle Atlantic States, from New Jersey to North Carolina, off southeast Florida, and along the Gulf Coast States. Important sport fisheries also occur in the Bahamas, off Havana, Cuba, and along the coast of Venezuela ( Mather et al. 1972). Another important sport fishery re- cently developed off eastern Brazil (Anonymous 1976). The white marl in is also an incidental catch of commercial longline vessels fishing for tuna in the Atlantic and Gulf of Mexico (Mather et al. 1975). The marlin is highly prized as a food item in some countries (Kume and Joseph 1969). My review of the literature on white marlin shows that there is a need for additional informa- tion on sex composition and length-weight rela- tionships. Until recently, no information was available regarding its reproductive potential (Baglin^). In this paper I update reproductive and sex ratio data presented by Bagliij (see footnote 2) and include length-weight relationships. Materials and Methods SEX COMPOSITION, LENGTH-WEIGHT RELATIONSHIP, AND REPRODUCTION OF THE WHITE MARLIN, TETRAPTURUS ALBIDUS, IN THE WESTERN NORTH ATLANTIC OCEAN' In the Atlantic, white marlin, Tetrapturus al- bidus. range from lat. 35°S to 45°N with concen- trations in the western Atlantic, including the Gulf of Mexico, and the Caribbean Sea ( Mather et al. 1975). Tag returns show that some white mar- lin migrate seasonally from the U.S. Middle At- lantic Bight (the coastal area between Cape Cod and Cape Hatteras) in the summer to the south- eastern Caribbean Sea in the winter ( Mather et al. 1972). Commercial catches by Japanese longline vessels support the tagging results, but the catches also indicate that a second group of white marlin moves from a wintering area in the south- eastern Caribbean to summer grounds in the Gulf of Mexico (Ueyanagi et al. 1970; Mather et al. 1972; Wise and Davis 1973). White marlin from the northern Gulf of Mexico (hereafter referred to as the gulf), the Florida Straits, the western Bahamas, and the Middle At- lantic Bight of the western North Atlantic (hereaf- ter referred to as the Atlantic) were sampled from anglers' catches at sport fishing tournaments and at Pflueger Marine Taxidermy, Inc., Hallandale, Fla. One marlin was collected by longline in the Windward Passage between Cuba and Hispaniola during RV Oregon Cruise 66. Sex data were obtained from 1,128 white marlin captured by anglers in the gulf ( 1971-77) and from 720 white marlin caught by anglers from the At- lantic (1972-77). Lengths and weights were obtained from 904 white marlin captured in the gulf (1971-76) and from 489 white marlin captured in the Atlantic ( 1972-76). Body lengths ( straight distance from tip of lower jaw to tips of midcaudal rays) were mea- sured in centimeters (Rivas 1956); weights were recorded to the nearest pound and converted to kilograms. 'Contribution No. 78-44M, Southeast Fisheries Center Miami Laboratory-, National Marine Fisheries Service, NOAA, Miami, Fla. ^Baglin, R. E., Jr. 1977. Maturity, fecundity and sex composi- tion of white marlin I Tetrapturus albidusl . Collective Volume of Scientific Papers 6(79):408-416. International Commission for the Conservation of Atlantic Tunas, Madrid, Spain. FISHERY BULLETIN: VOL 76. NO. 4, 1979. 919 Ovaries from 186 females caught from 1972 through 1976 were examined. Fresh ovaries were either blotted dry and weighed in grams or stored in 107f Formalin'^ and weighed later. No sig- nificant difference was found between the mean weight of fresh and preserved ovaries (F = 0.0001; df = 1, 16;P>0.75). The gonosomatic index (GSI), ovary weight as a percentage of total body weight, was used as an indicator of maturity. Preserved eggs 0.56 mm in diameter and larger were counted when estimating fecundity. Eggs <0.56 mm were less spherical in shape and in an earlier stage of development. The 0.56-mm size was determined by measuring the diameters of 3,912 eggs from mature, partly spent, and spent fish. Small transparent ova were stained with aceto-carmine to facilitate measuring. Egg diameters were measured with an ocular microm- eter at 30 X magnification and the orientation of egg diameters was assumed to be random. Thin cross sections were taken from the anterior, mid- dle, and posterior parts of one ovary of a mature fish and subdivided into three subsamples, repre- senting the center, midregion, and periphery of the ovary (Otsu and Uchida 1959). Fecundity was defined as the potential number of mature eggs (yolked ova in the most advanced size mode) that could be spawned during one re- productive season and was estimated using a dry weight method. For six fish in which entire ovaries were saved for fecundity analysis, subsamples consisted of a thin cross section taken from the anterior, middle, and posterior parts of each ovary. The eggs in these subsamples were separated from the ovarian tissue, enumerated, dried, and weighed according to the procedure described by Baglin (see footnote 2). For six other fish, only the ovary weight and a single cross section from the middle of the ovary were taken; these cross sec- tions comprised the subsamples. The eggs were separated, counted, dried, and weighed. A dry/wet weight regression was used to estimate the total dry weight of the eggs in these ovaries, which were not saved. Before the eggs in the subsample were counted, 25 eggs 0.30 mm and larger from the two most advanced modes were randomly selected and measured. Eggs in this second most advanced mode were included to give an indication of the percentage of eggs in both modes because future histological studies may indicate that these smaller eggs undergo further development and are also spawned. Fecundity estimates, rounded to the nearest 0.1 million eggs, were calculated from the relationship: C = iAD/B) + A, where A is the number of mature ova in the subsample, B is the weight of the ova in the subsample, C is the number of mature ova, and D is the weight of ova from both ovaries. Results and Discussion Sex Q)mposition From 1971 through 1977, sex was determined from 1,128 white marlin from the gulf (Table 1). The deviation from an expected 1:1 sex ratio was significant from May through October. Sampling was inadequate for the remaining months. Females were more prevalent than males for each month studied. From 1972 through 1977, sex was determined for 720 white marlin from the Atlantic (Table 1). There were 323 sex determinations from the Florida Straits (March through May) and 397 from the Middle Atlantic Bight (June through Sep- tember). Sampling was inadequate from October through February. No significant difference from an expected 1:1 sex ratio was found for March, May, July, August, and September, but a sig- nificant difference was found for April and June. For the months in which the sex ratio was sig- nificantly different from the expected 1:1 ratio, females were more prevalent. deSylva and Davis (1963) found a significant difference from an expected 1:1 sex ratio (60% females) when they combined their data from the Middle Atlantic Bight for the summers of 1959 and 1960. They presented monthly sex composi- TABLE 1. — Monthly sex ratios for white marlin from the north- ern Gulf of Mexico! 197 1-77), Florida Straits and Middle Atlantic Bight (1972-77). ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Number of Sex ratio Location f^ontfi white marlin (females/males) Gulf of (Vlexico May 21 4.25' June 85 4.00* July 374 3.16* August 444 1.63* September 150 1.50* October 54 1.84* Florida Straits March 103 0.87 April 172 1.96* May 48 1.40 Ivliddle Atlantic June 55 3.23* Bigtit July 56 1.67 August 219 0.80 September 67 0.97 'Significant departure from null fiypothesis at 0.05 level (chi-square) 920 tion data for 1960 only. My analysis of their data shows a significant difference for June and July but no significant difference for August and Sep- tember. Their findings for June, August, and Sep- tember agree with those in this study. The ex- treme difference in se.x ratios found in the present study for the gulf (May through October) and for the Florida Straits in April has not been reported previously. The above findings suggest that some white marlin segregate into distinct areal groups according to the predominating sex and that sex ratios may change with season. A similar occur- rence has been noted for the blue marlin, Makuira nigricans (Kume and Joseph 1969). Length- Weight Relationship The average length of females is greater than that of males from both the gulf and the Atlantic (Figure 1). It is also apparent (Figures 2, 3) that the average length of females is greater than that of males from each area for each month studied. This difference may be due to faster growth of females or higher mortality of males and should be considered in future growth studies of the white marlin. The length- weight relationship by sex was de- termined for white marlin taken in the gulf from 1971 through 1976 (Figure 4) and in the Atlantic from 1972 through 1976 (Figure 5). Analysis of covariance (Table 2) indicated that length-weight regression coefficients were significantly different between gulf females and males (F = 16.0; df = 1, 900;P<0.001), gulf males and Atlantic males (F = 19.2; df= 1, 514; P<0.001), and gulf females and Atlantic females (F = 10.8; df = 1, 871;P<0.001). The adjusted means were also significantly differ- ent between Atlantic females and males (F = 13.4; df = 1, 486; P<0.001). These findings agree with those of Lenarz and Nakamura (1974), who found a significant difference between sexes in the rela- tionship between weight and eye-fork length for white marlin from the gulf during 1971. Analysis of covariance was conducted for the length-weight relationship, on a monthly basis, for which sufficient samples were available: gulf females versus Atlantic females in May, June, August, and September, and gulf males versus Atlantic males for June, August, and September. A significant difference in the regression coefficients was found only for the August males (F = 13.7; df= 1,211;P<0.001). A significant differ- ence in adjusted means was found for females dur- 215- 210- 205- 200- E 195- u 190- o 185- 180- 5 175- o it. I 170- < 165- I 155J 150- 145- 140- 135- ■r ? N=277 31% N=248 51% 0*' N = 627 69% N=241 49% GULF ATLANTIC Figure 1. — Comparison of length between female and male white marlin collected in the northern Gulf of Mexico (1971 through 1976) and in the Atlantic (1972 through 1976). The number, percent, mean (horizontal line), range (vertical line), 1 SD on each side of the mean ( open box), and 2 SE on each side of the mean (shaded box) are shown. ing June (F = 9.3; df = 1, 74;P<0.005) and August (F = 12.0 df = 1, 286;P<0.001), and for males during September (P = 7.6; df = 1, 73; P<0.01). Differences between length-weight relation- ships of white marlin from the gulf and the Atlan- tic suggest the possibility of separate groups in- habiting the two areas. Tag returns, however, showed there is at least some migratory move- ment from the Middle Atlantic Bight to the gulf. To date, tag return data have not shown white marlin migrations in the reverse direction, al- though one fish tagged in the gulf was recaptured off Cuba, giving some support to the likelihood that they do migrate in the opposite direction (Chester C. Buchanan, Southeast Fisheries Center, National Marine Fisheries Service, NOAA, 75 Virginia Beach Drive, Miami, FL 33149, pers. commun.) 921 215- 210- 205- 200- J 190- o Z 185- lii _l 180- ae u 175- I 170- < 165- 160- O 155- 150- 145- 140- 135 N = 17 N=75 ? N = 123 36% ? N = 231 75% N^222 64% N = 74 61% II N=30 62% N=18 38% N = 47 39% MAY JUNE JULY AUGUST SEPTEMBER OCTOBER Figure 2. — Monthly comparisons of length between female and male white marlin collected in the northern Gulf of Mexico during 1971 through 1976. The number, percent, mean (horizontal line), range (vertical line), 1 SD on each side of the mean (open box), and 2 SE on each side of the mean (shaded box) are shown. 205 200 195- 190- 185- Ul K u. ' 160- 5 1551 Uj I50H O 145- 140- 135- MARCH APRIL MAY JUNE JULY AUGUST SEPTEMBER Figure 3. — Monthly comparisons of length between female and male white marlin collected in the Atlantic during 1972 through 1976. The number, percent, mean (horizontal line), range (vertical line), 1 SD on each side of the mean (open box), and 2 SE on each side of the mean (shaded box) are shown. 922 H I 4S X u J|40- o GULF OF MEXICO L0G,oW=-4 49978 •2 65546 LOG,oL N:627 /— L0CnoW=-3 10422* 2 01104 LOC.qL — r- 2 15 2 20 2 25 2 30 LOG,o LOWER JAW- FORK LENGTH (cm) Figure 4. — Length-weight relationship (log transformation) for female and male white marlin from the northern Gulf of Mexico. w I 40' 5 0 u O I 35- 1 30' ATLANTIC LOC,nW= -5 52016 'S 13550 LOCoL N=248 LOG,„W=-4 96728 •2 87607 LOG,^ L —I 1 1 1 1 2 15 2 20 2 25 2 30 2 35 LOG,o LOWER JAW- FORK LENGTH (Cm) Figure 5. — Length-weight relationship (log transformation) for female and male white marlin from the Atlantic. Reproduction A significant difference in egg diameter was found among the anterior, middle, and posterior sections of an ovary from a mature fish {F — 1.1; df = 2, 2,676; P<0.001). There was no significant Table 2. — Regression equations, number, sum of squares of jc, and mean square calculated for the length-weight relationship (logio transformations) of white marlin from the northern Gulf of Mexico and the Atlantic. S.v^-blx-j y + b(X - X) N Xx' N-2 Gulf and Atlantic females: 1.41704 + 288186(X - 2.22302) Gulf and Atlantic males: 1 .34355 + 2.37655(X - 2.20228) Gulf females: 1.39996 + 2.65546(X - 2.22174) Gulf males: 1.32735 + 2 01104(X - 2.20363) Atlantic females: 1.46021 + 3.13550(X - 2.22624) Atlantic males: 1.36217 + 2.87607(X - 2.20073) May gulf females: 1 48714 + 2.60014(X - 2.24572) May Atlantic females: 1.44276 + 2.96929(X - 2.21669) June gulf females: 1,44221 + 3.22066(X - 2.23317) June Atlantic females: 1.40502 + 2.60180(X - 2.23216) June gulf males: 1 35686 + 3.55091(X - 2.21054) June Atlantic males: 1.24868 + 3.01101(X ~ 2.17815) August gulf females: 1.38976 + 2.72170(X - 2.21831) August Atlantic females: 1.41118 + 2.81615(X - 2.21783) August gulf males: 1.32575 * 1.93866(X - 2.20207) August Atlantic males: 1.35384 + 2.91546(X - 2.20279) September gulf females: 1.40689 + 3.01922(X - 2.22399) September Atlantic females: 1 42732 ^ 3,10221(X - 2.22137) September gulf males: 1.32709 + 1.39787(X - 2.20530) September Atlantic males: 1.34390 + 2.01746(X - 2.19767) 875 0.604832 0.00336706 518 0.254956 0.00299720 627 0.396458 0.00252756 277 0.128593 0.00250737 248 0.204773 0.00377638 241 0.125280 0.00244554 17 00164014 0.00365206 16 0.0085322 0.00226977 51 0.0210122 0.00162473 26 0.0163375 0.00323226 11 0.00624798 0.00085387 10 0.00707686 0.00165193 222 0 121132 0.00226799 67 0.0391569 0.00206003 123 0.524543 0.00179221 92 0.0484607 0.00169370 74 0.0440863 0.00317564 17 0.0150308 0.00160237 47 0.0199974 0.00220416 29 0.0105484 0.00144724 difference in mean diameter among the center, midregion, and periphery within each of the three sections. Because some heterogeneity occurred, estimates of fecundity were based, when possible, on eggs from each section of both ovaries. Heterogeneity of egg size within an ovary has also been shown for albacore, Thunnus alalunga (Otsu and Uchida 1959), and swovdfish., Xiphias gladius (Uchiyama and Shomura 1974). The left ovary (X = 25.0 cm, S^ = 0.732) was significantly longer (F = 35^7; df = 1, 196; P<0.001) than the right ovary (X = 19.4cm,S.v = 0.561). Eldridge and Wares ( 1974) reported differ- ential growth in the size of ovaries for striped marlin, Tetrapturus audax, and for sailfish, Is- tiophorus platypterus. Both were similar to the white marlin in having larger left ovaries. Well-developed ovaries were present only in 12 white marlin collected during April and May in the Florida Straits. These fish had a GSI of about 923 &'Jc or greater and were used for estimating fecun- dity. The mean GSI showed that ovarian weights were lowest during October and increased from November through May (Figure 6). The mean GSI of 2.6 for April and May is lower than the 4.5 mean GSI found by Krumholz (1958) for late April. The GSI of 9.3 (Table 3) agrees with the highest GSI of 9.76 found by Krumholz. The high mean GSI val- ues determined by me for April and May, with the sudden decrease in June, indicated that spawning probably occurred during April and May (Figure 6). Therefore, only one spawning season per year was indicated for the Florida Straits. White marlin may also spawn in other areas. One fish captured in April 1976 in the Windward Passage had ripe eggs measuring 1.16 mm. Hayasi et al. (1970) found white marlin with mature gonads during April-June in the northern Carib- bean. Erdman (1956) found well-developed Figure 6. — Seasonal variation of mean gonosomatic index in 186 white marlin collected from 1972 to 1976 (number offish indicated above histograms). Table 3.— Weight, length, and gonadal data for 12 female white marlin from the Florida Straits collected during 1972, 1974, and 1975. The mean and standard error of the mean are given at the bottom of the columns. Estimated number of eggs Body weigtit (kg) Body lengthi (cm) Ovary wet weight (g) Gono- somatic index 0,55 mm in diameter (millions) ■0,29 mm in diameter' (millions) 26.8 160 21.600 6.0 5.4 26.8 169 22,050 7,6 4.8 30.4 168 22,324 7,6 70 30.4 176 21,700 5.6 38 31.3 168 2,908 93 10.4 32.7 166 2,150 66 7 1 32.7 166 2,693 8,2 10 1 33.6 167 2,161 6,4 7.6 35.0 169 22,250 6,4 6.5 35.4 170 22,320 6,6 7.5 36.3 171 2,488 68 105 37.2 179 22,050 5.5 8 1 324 169 2,224 69 7.4 0.98 14 107 032 0.62 10.4 8.0 117 7.3 18.6 11.8 16.8 11.9 10.2 14.4 20.2 14.5 13.0 1 16 'Estimated using actual percent from 0 30 to 0 55 mm in diameter from 25 eggs measured for each fish. 2Entire ovaries available. ovaries in white marlin caught off Puerto Rico in April and found well-formed eggs in a fish taken in June from the same locality. The smallest fish approaching a ripe condition with large ovaries weighed 26.8 kg (Table 3). Ueyanagi et al. ( 1970) reported that white marlin reach sexual maturity at 130 cm eye-fork length. Using the conversion equation of Lenarz and Nakamura (1974), 130 cm eye-fork length would be equal to about 20.3 kg. Frequency distributions of white marlin ovum diameters were made from measurements on 3,912 ova from spent, partly spent, and mature fish (Figure 7). Spent fish caught during May and June contained mostly eggs 0.15 mm in diameter and smaller. Eggs from a partly spent fish caught during June had a frequency mode of about 0.35 mm, with few eggs larger than 0.60 mm. Some of the larger eggs appeared to be undergoing absorp- tion. Jolley ( 1977), in his histological examination of spent sailfish, found degeneration and absorp- tion of advanced unovulated eggs common. Mer- rett (1970), studying several species of billfish from the Indian Ocean, suggested that there also may be at least a partial resorption of resting 5 o S300 Z 250- 150- 100- 50 1 0.05 0 15 0.25 0.35 045 0 55 0 65 0 75 0 85 0 95 105 DIAMETER (mm) 1 1.15 Figure 7. — Frequency distribution of white marlin ovum diameters for: A, four spent fish (827 ova) in May and June; B, one partially spent fish (406 ova) in June; C, one mature fish (2, 679 ova) in April. 924 oocytes. I found frequency modes of about 0.35 mm and 0.65 mm in a mature fish caught in April. Only eggs measuring 0.56 mm and larger were included when estimating fecundity. Because there were two frequency modes present in mature fish, an estimate of the number of eggs 0.30 mm in diameter and larger is also presented (Table 3). Fecundity, based on the number of ova in the most advanced size mode, ranged from 3.8 to 10.5 million eggs iX = 7.4, Sy = 0.62) for white marlin weighing 26.8 to 37.2 kg (Table 3). The number of mature ova per gram of body weight ranged from 125 to 332 (X = 227, %- = 16.76). The average number of eggs measuring 0.30 mm in diameter and larger was estimated as 13 million (S^ = 1.16). Fecundity was based on the number of fully yolked eggs, forming a group distinct from another group of developing eggs. Fecundity would vary depending on whether smaller eggs develop further or are absorbed. If fractional spawning occurs, as reported for sailfish by Jolley ( 1977), the eggs in the next distinct group should be included in seasonal fecundity estimates. Acknowledgments I thank the boat captains and sport fishermen who provided samples, and Pflueger Marine Taxidermy, Inc., Hallandale, Fla., for allowing National Marine Fisheries Service (NMFS) per- sonnel to sample available specimens. I thank Grant Beardsley, Chester Buchanan, Eugene Nakamura, William Richards, Luis Rivas, and James Tyler of the Southeast Fisheries Center (NMFS), and John Jolley of the Florida Depart- ment of Natural Resources for their helpful com- ments on the manuscript. Literature Cited Anonymous. 1976. Brazil: A sleeping giant awakens. Int. Mar. Angler 38(6):6-7. DE Sylv.a, d. p., and W. p. Davis. 1963. White marlin, Tetrapturus albidus, in the Middle Atlantic bight, with observations on the hydrography of the fishing grounds. Copeia 1963:81-99. Eldridge, M. B., and p. G. Wares. 1974. Some biological observations of billfishes taken in the eastern Pacific Ocean, 1967-1970. In R. S. Shomura and F. Williams (editors). Proceedings of the Interna- tional Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review and contributed papers, p. 89-101. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. ERDMAN, D. S. 1956. Recent fish records from Puerto Rico. Bull. Mar. Sci. Gulf Caribb. 6:315-340. Hayasi, S., T. Kato, C. Shinou, S. Kume, and Y. MORITA. 1970. Status of the tuna fisheries resources in the Atlantic Ocean, 1956-1967. [In Engl, and Jpn.] In Resources and fisheries of tunas and related fishes in the Atlantic Ocean. Far Seas Fish. Res. Lab. (Shimizu), S. Ser. 3:1- 72. Jolley, J. W., Jr. 1977. The biology and fishery of Atlantic Sailfish, Is- tiophorus platypterus , from southeast Florida. Fla. Mar. Res. Publ. 28, 31 p. krumholz, l. a. 1958. Relative weights of some viscera in the Atlantic marlins. Bull. Am. Mus. Nat. Hist. 114:402-405. KUME, S., AND J. Joseph. 1969. Size composition and sexual maturity of billfish caught by the Japanese longline fishery in the Pacific Ocean east of 130°W. Bull. Far Seas Fish. Res. Lab. (Shimizu) 2:115-162. LENARZ, W. H., AND E. L. NAKAMURA. 1974. Analysis of length and weight data on three species of billfish from the western Atlantic Ocean. In R. S. Shomura and F. Williams (editors), Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review and contributed papers, p. 121-125. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. M..\THER, F. J., Ill, H. L. Clark, and J. M. Mason, Jr. 1975. Synopsis of the biology of the white marlin, Tetrap- turus albidus Poey ( 1861 ). InR. S. Shomura and F. Wil- liams (editors). Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 3. Species synopses, p. 55-94. U.S. Dep. Commer., NOAA Tech Rep. NMFS SSRF-675. Mather, F. J., Ill, A. C. Jones, and G. L. Beardsley, Jr. 1972. Migration and distribution of white marlin and blue marlin in the Atlantic Ocean. Fish. Bull., U.S. 70:283- 298. MERRETT, N. R. 1970. Gonad development in billfish (Istiophoridae) from the Indian Ocean. J. Zool. (Lond.) 160:355-370. Otsu, T.. and R. N. UCHIDA. 1959. Sexual maturity and spawning of albacore in the Pacific Ocean. U.S. Fish Wildl. Serv., Fish. Bull. 59:287-305. RiVAS, L. R. 1956. Definitions and methods of measuring and counting in the billfishes (Istiophoridae, Xiphiidae). Bull. Mar. Sci. Gulf Caribb. 6:18-27. Uchiyama, J. H., AND R. S. Shomura. 1974. Maturation and fecundity of swordfish, Xiphias gladius, from Hawaiian waters. In R. S. Shomura and F. Williams (editors), Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review and contributed papers, p. 142-148. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. Ueyanagi, S., S. Kikawa, M. Uto, and Y. Nishikawa. 1970. Distribution, spawning, and relative abundance of billfishes in the Atlantic Ocean. [In Jpn., Engl, ab- str.] Bull. Far Seas Fish. Res. Lab. (Shimizu) 3:15-55. 925 WISE, J. P., AND C. W. Davis. 1973. Seasonal distribution of tunas and billfishes in the Atlantic. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-662, 24 p. Raymond E. Baglin, Jr. Southeast Fisheries Center Miami Laboratory National Marine Fisheries Service, NOAA 75 Virginia Beach Drive Miami. FL 33149 RECORDS OF PISCIVORUS LEECHES (HIRUDINEA) FROM THE CENTRAL COLUMBIA RIVER, WASHINGTON STATE No records of leech infestations on fish of the Co- lumbia River exist in the published literature. As a whole, the freshwater hirudinean fauna of the Pacific Northwest remains a relatively unsur- veyed, little known, and neglected biotic group. This is due, in part, to problems in leech identifica- tion as well as in obtaining representative collec- tions. We obtained leeches from the external surface, oral cavity, and gill chambers of fish during a continuing environmental assessment program on the central Columbia River above Richland, Wash. (Benton and Franklin Counties), from 1975 through 1977. This paper identifies four piscivo- rous species, provides new host and distribution records, and reviews some recent taxonomic changes for the species encountered. Ecological observations are included. The leeches recorded herein are Myzobdella lugubris Leidy 1851, Piscicola salmositica Meyer 1946, Placobdella montifera Moore 1906, and Ac- tinobdella inequiannulata Moore 1901. Methods and Site Description Fish were collected at monthly or bimonthly intervals by a variety of gear (gill nets, trammel nets, hoop nets, beach seines, and electroshocker) from January 1975 to December 1977. Over 20,000 fish, representing nearly 40 species, were examined during this period (Gray and Dauble 1977). Leech specimens were preserved in 10% Formalin^ solution, either when captured or after being examined alive in the laboratory. Our leech collections were more qualitative than quantitative because leech-fish associations in nature are normally periodic and facultative despite the nutritional requirement of piscivorous leeches for fish blood. Also, piscivorous leeches can readily detach from fish captured by most types of fishing gear, particularly from fish recovered when moribund or dead. Occurrence of many freshwater leech species can be correlated with characteristic aquatic habitats. Water quality parameters vary season- ally in the central Columbia River, as follows: dissolved oxygen, 8.0-12.0 mg/1; pH, 7.4-8.6; phos- phate (as PO4), 0.03-0.04 mg/1; ammonia- nitrogen, 0.01-0.2 mg/1; hardness (Ca, Mg), 55-75 mg/1; and alkalinity (CaCOg), 50-67 mg/1. Water temperatures range from 1° to 3°C in midwinter to about 21°C in late August and early September. There are no significant quantities of organic and inorganic pollutants (our data). The water carries minimal silt loads. The central Columbia River in the Hanford Reach where our collections were made (river km 550-629) survives as the last free-flowing section of the main channel above Bonneville Dam. Dec- ades of hydroelectric development have trans- formed other sections into a consecutive series of river-run reservoirs. River flows in the study area usually range from about 2,000 m^/s over much of the year to over 12,000 m^/s during the annual spring spate, when surplus runoff is passed down- river over spillways from reservoirs (Nees and Corley^). Additionally, Hanford flows are now regulated at Priest Rapids Dam in response to daily and weekly power demand peaks, causing water levels in the river to fluctuate widely. This periodically exposes and inundates a rocky or muddy shoreline zone, apparently restricting development of a di- verse leech fauna along the river margins. Water levels in Wanapum Reservoir behind Priest Rapids Dam (river km 639) and in Umatilla Res- ervoir behind McNary Dam (river km 470) are relatively stable, although subject to controlled summer drawdowns. Substantial populations of such common omnivorous leeches as Erpobdella punctata (Leidy 1870), Helobdella stagnalis (Lin- naeus 1758), and T/ieromjzon spp. occur along the 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ^Nees, W. L., and J. P. Corley. 1974. Environmental sur- veillance at Hanford for CY-1973. Unpubl. manuscr., 56 p. R&D Rep., BNWL-1881, Battelle, Pacific Northwest Laboratories, Richland, WA 99352. 926 FISHERY BULLETIN: VOL, 76, NO 4, 1979. margins of these mid-Columbia River reservoirs (our observations). Results and Discussion Four leech species were recovered from Colum- bia River fish (Table 1). About 907^ of the speci- mens were Myzobdella lugubris. Two families belonging to the order Rhynchobdella were rep- resented, Glossiphoniidae and Piscicolidae. Mem- bers of this order typically possess a small pore on the anterior sucker for a mouth, from which a muscular pharyngeal proboscis can be protruded, and lack biting jaws or denticles. Relatively few glossiphoniids are piscivorous (Klemm 1975). However, the piscicolids characteristically are ec- toparasites offish and feed on fish blood. The mus- cular proboscis of piscicolids is effective in pene- trating epidermal layers of fish wherever scales are reduced or absent, although gills are favored feeding sites. Myzobdella lugubris Myzobdella lugubris has not been recorded pre- viously as a common ectoparasite of Columbia River fish. Its distribution and host records are included in previous publications that refer to the genxis Illinobdella in which M. lugubris was for- merly placed. Myzobdella lugubris and M. (syn. Illinobdella) moorei (Meyer and Moore 1954) were until re- cently believed to be two distinct species. The former was considered characteristic of brackish and marine waters, while the latter was consi- dered characteristic of freshwater. The distinction was primarily ecological since anatomical fea- TabLE 1. — Piscivorous leeches (Hirudinea) collected in this sur- vey from teleost fishes in the central Columbia River, Washington State. Species Host infected Piscicolidae: Myzobdella Norttiern squawfish, Ptychocheilus lugubris oregonensis Ctiiselmouth, Acrocheilus alutaceus Brown bullhead, Ictalurus nebulosus Largescale sucker, Catostomus macrocheilus Bridgelip sucker, C columbianus Piscicola Chinook salmon, Oncorhynchus tshawytscha. salmositica fry Sucker, Catostomus sp., fingerling Glossiphoniidae: Placobdella montifera Sucker, Catostomus sp , fingerling Actinobdella inequiannulata Largescale sucker, C macrocheilus tures of both species were remarkably similar. However, M. lugubris and M. moorei are now con- sidered to be a single euryhaline species (Sawyer et al. 1975). Further, it now appears that all members of the related piscicolid genus Illinobdella are synony- mous with M. lugubris. Thus species reported in the literature as Illinobdella alba, I. elongata, and /. richardsoni, as well as M. ( =/.) moorei, all prey- ing on fish in North American waters (Meyer 1940, 1946b), are junior synonyms ofM. lugubris, which holds taxonomic priority. Locality and host records of the ubiquitous M. lugubris under these synonyms are given by Hoffman (1967), Klemm ( 1972a, b, 1977), and Sawyer et al. (1975). Studies on M. lugubris infesting the blue crab, Callinectes sapidus, and the white catfish, Ictalurus catus , in a South Carolina tidal river support this synonymy (Daniels and Sawyer 1975). Myzobdella lugubris was recovered from a wide size range of adult chiselmouth, Acrocheilus alutaceus, and northern squawfish, Ptychocheilus oregonensis, in our collections, and less frequently from adult suckers, Catostomus macrocheilus and C columbianus. The associations were apparently facultative. Myzobdella lugubris occurred primar- ily in the oral cavity of chiselmouth (Figure 1 ) and northern squawfish, where they were retained during the struggle of hosts captured in overnight net sets. Leeches were recorded and counted the next day when fish were recovered. Many leeches on the external surfaces of moribund or dead fish may have detached before net recovery. Myzob- della lugubris were never found in the mouth of suckers but only in the axila of pelvic or pectoral fins, on fin rays, or in the gill chambers. Myzobdella lugubris was also a fairly common ectoparasite of brown bullhead,/, nebulosus, col- lected by angling in backwater sloughs of the cen- tral Columbia and lower Snake Rivers during the summer. Infestations on bullheads usually con- sisted of one or two small leeches attached to the pectoral or pelvic fins. The incidence of M. lugubris on adult chisel- mouth and northern squawfish (Table 2) shows infestations only during June, July, and August when C6lumbia River water temperatures ranged from 13° to 21°C. The leeches were primarily sexu- ally mature. Collections from the oral cavity of chiselmouth in October 1975, 1977 and November 1977 contained numerous small, immature leeches that had apparently hatched within the preceding 1 or 2 mo. 927 >-#■ »*^^ 0 mm Figure l. — Four sexually mature Myzobdel I a lugubns in the oral cavity of chiselmouth collected in the central Columbia River. The small subterminal mouth of the host is bordered by a cartilaginous upper and lower lip for grazing on sessile algae. The leeches occupy most of the available space in the oral cavity when the mouth is closed. Table 2. — Incidence of infestation of chiselmouth and northern squawfish by the piscicolid leech, Myzohdel la luguhris, indicated by infestation ratio.' Ctiiselmouth Norttie rn squawfi sh Month 1975 1976 1977 1975 1976 1977 Jan. 0/5 0/4 0/5 0/2 0/1 Feb, 0/1 — — 0/3 0/2 0/1 Mar. — — — 0/1 — — Apr. 0/5 0/3 0/4 0/14 0/9 0/7 May 0/22 0/5 0/10 0/34 0/7 0/14 June 0/32 0/16 6/27 0/33 0/3 0/19 July 1/32 8/19 1/12 3/38 0/16 6/25 Aug. 1/19 0/4 1 9/42 0/13 0/2 0/8 Sept 0/46 0/9 0/13 0/18 0/15 0/4 Oct 27/37 0/10 ^3/13 0/23 0/2 0/1 Nov 0/7 0/2 27/9 0/4 — 0/3 Dec. 0/6 0/7 0/2 0/2 — 0/1 'Infestation ratio = number of fish infested/number of fish examined. ^Numerous small leeches, recently hatched, were attached to some hosts. According to Sawyer etal. ( 1975), M. lugubris is a relatively warm-water species encountered most often at 21°-30°C, occasionally at 16°-20°C, and less often in colder water. They reported that the leech appeared to be injured if the water was sud- denly cooled to 10°-15°C in laboratory experi- ments. Obviously some M. lugubris survive over winter at low temperatures, but it must remain inconspicuous due to dormancy in temperate re- gions of North America. We have never recovered M. lugubris during winter in the central Columbia River, either from fish or from benthic samples 928 designed to quantitatively collect invertebrates. Large, adult M. liigubns from Columbia River fish were characterized by a green background coloration, superficially suggesting that they fed on algal cells ingested by the host. Chiselmouth and suckers commonly feed on sessile, green- colored diatoms from bottom substrates. However, microscopic examination revealed that this col- oration was due entirely to pigments in the adult leech's musculature and not to the presence of algae in their digestive tract. Myzobdella lugubris fed entirely on fish blood cells and plasma. Feeding on fish blood is clearly required by M. lugubris for growth and reproduction. Copulating M. lugubris were noted on fish. But since deposi- tion of a cocoon requires hard substrates, sexually mature leeches must eventually detach, thus free- ing fish of infestations. The breeding, growth, and reproductive cycle of piscivorous leeches may ac- count for the periodic infestations of fish so fre- quently documented in the literature. In the Columbia River, the cycle in M. lugubris is correlated with a seasonal change in water temperatures, with peak activity occurring in late summer and fall. In tidal estuaries on the east coast of the United States, M. lugubris has a life cycle that involves tw^o hosts, a fish and a crustacean. It engorges on fish blood before detaching to deposit cocoons on crabs (Daniels and Sawyer 1975). This indicates possible involvement of a freshwater crustacean in the life cycle of M. lugubris in the Columbia River. The only large crustacean available is Pacifasticus leniusculus, but extensive collections of this crawfish in previous years by the senior author disclosed no attached leeches. Therefore, stones are probably used as cocoon deposition sites in the Columbia River. Piscivorous leeches are often vectors of hemoflagellates (genera Trypanoplasma and Trypanosoma) found in the blood of freshwater and marine fishes iKhaibulaev 1970; Becker 1977). Although we have occasionally detected Trypanoplasma in Columbia River fish, we found no hemoflagellates in the digestive tract of over 20 M. lugubris taken from various hosts. We did not examine histopathology of leech at- tachment and feeding sites in the oral cavity of infested Columbia River fish, although petechiae were evident during the fall on some chiselmouth. Inflammatory conditions and hyperplasia were described previously from a massive infestation of M. lugubris (misidentified as Cystobranchus vir- ginicus) on white catfish in Virginia (Paperna and Zwerner 1974). Piscicola salmositica The salmonid leech, P. salmositica, was de- scribed from specimens taken, in part, from sea- run steelhead trout, Sal mo gairdneri, transferred from the Columbia River to Mason Creek, Chelan County, Wash. (Meyer 1946a). The leeches infest- ing the fish originated either in the Columbia River or from Mason Creek. Thus the salmonid leech has previously been recorded from the upper Columbia River system. This species is usually associated with fall spawning runs of adult salmo- nid fishes in coastal streams, but it occurs elsewhere in the Pacific Northwest (Becker and Katz 1965a). We collected several P. salmositica at various times from chinook salmon, Oncorhynchus tshawytscha , fry and once from a fingerling sucker. Each infestation consisted of a solitary leech attached to the dorsal surface of its host and feeding on blood. All specimens were taken in April and June as water temperatures (10°-14°C) increased and consisted of small leeches that pre- sumably had hatched from cocoons the previous fall or winter. Three specimens contained develop- ing trypanoplasms among their intestinal con- tents, evidence of prior feeding on infected fish. Therefore, P. salmositica is confirmed as a vector transmitting trypanoplasms among various fish in the central Columbia River. The salmonid leech is the only known vector of the piscine hemoflagel- lateTrypanoplasma salmositica (Katz 1951) in the Pacific Northwest (Becker and Katz 1965b). Piscicola salmositica requires meals of fish blood before detaching to deposit cocoons on bottom sub- strates (Becker and Katz 1965a). Thus salmonid leeches presumably occur among and infest popu- lations of anadromous chinook salmon that spawn each fall in the central Columbia River near our fish collection sites. However, we have not de- tected P. salmositica on transient adult fall chinook salmon returning from the sea to spawn or from downstream drifting, spawned out salmon carcasses. Neither have we found salmonid leeches on several adult steelhead trout and spring-run chinook salmon examined during the summer at the Ringold Hatchery (Washington State Department of Game) above Richland. On the basis of our observations, P. salmositica is not an abundant leech in the central Columbia River. 929 Placobdella montifera One immature P. montifera was recovered from the dorsal surface of a fingerling sucker in early October 1976. We also collected one adult speci- men from beneath shoreline rocks at Umatilla Reservoir during June where it was depositing a cocoon. The species is not a common ectoparasite of fish. It probably occurs mostly along reservoir shorelines where water levels remain relatively stable, rather than along the margins of the free- flowing Columbia River above Richland. Placobdella montifera has been reported to at- tack aquatic worms, insect larvae, mussels, frogs, toads, and fish, but the only specific host records are fish (Hoffman 1967; Klemm 1972a, 1975, 1976; Sawyer 1972; and others). This leech, as do most glossiphoniids, broods its cocoon and carries its young. An uncommon but widely distributed species, P. montifera is listed as having been re- ported previously from Washington (Klemm 1972b). Distributional records probably valid in- clude British Columbia, Saskatchewan, Ontario, and the northern states east of the Mississippi River southward to Georgia (Sawyer 1972; Klemm 1977). The host relationship for glossiphoniids is gen- erally considered to be less obligatory than for piscicolids, and most are omnivorous feeders. Ap- parently A. inequiannulata, P. montifera, and P. pediculata Hemingway 1908 are the only three American glossiphoniids consistently reported to parasitize fish. Several authors have reported P. pediculata from the freshwater drum, Aplodinotus grunniens, and Sawyer ( 1972) has indicated a high degree of host specificity; it has not been reported from the Pacific Northwest, nor would it be ex- pected in this region due to its narrow host prefer- ence. Acknowledgments Donald J. Klemm, Research Aquatic Biologist, U.S. Environmental Protection Agency, collabo- rated with us on leech identification and literature survey, and critically reviewed the manuscript. Fish collections were supported largely by en- vironmental assessment programs for Washington Public Power Supply System under Contract 2311201335 between United Engineers and Constructors and Battelle, Pacific Northwest Laboratories. Actinobdella inequiannulata Six A. inequiannulata were collected from the axila of the pelvic and pectoral fins of one adult largescale sucker in mid-August 1975. According to Sawyer (1972), this glossiphoniid is known from Illinois, Minnesota, and Ohio; Klemm (1972a, b, 1977) adds Michigan, Pennsylvania, and New York; and Daniels and Freeman (1976) add On- tario. Actinobdella triannulata Moore 1924, a name common in earlier literature, is now consid- ered a junior synonym of A. inequiannulata (Daniels and Freeman 1976; Klemm 1977). Daniels and Freeman (1976) provide a rede- scription of A. inequiannulata on basis of speci- mens collected from two species of suckers (genus Catostomus) and preserved material from the U.S. National Museum. The species was earlier consid- ered as free-living with no known hosts (Sawyer 1972; Klemm 1972a). Since its synonym A. trian- nulata displayed a predilection for suckers (Hoffman 1967), the host preference of A. in- equiannulata is now partially resolved. Little is known of its ecology and life cycle. We did not examine our specimens for ingested fish blood. Literature Cited Becker, C. D. 1977. Flagellate parasites offish. /« J. P. Kreier (editor), Parasitic protozoa. Vol. 1, Taxonomy, kinetoplastids, and flagellates offish, p. 357-416. Academic Press, N.Y. Becker, C. D., and M. Katz. 1965a. Distribution, ecology and biology of the salmonid leech, Piscicola salmositica (Rhynchobdellae: Pis- cicolidae). J. Fish. Res. Board Can. 22:1175-1195. 1965b. Transmission of the hemoflagellate Cryptobm salmositica Katz, 1951, by a rhynchobdellid vector. J. Parasitol. 51:95-99. Daniels, B., and R. S. Freeman. 1976. A review of the genus Actinobdella Moore, 1901 (Annelida:Hirudinea). Can. J. Zool. 54:2112-2117. Daniels, B. a., and R. T, Sawyer. 1975. The biology of the leech Myzobdella lugubris infest- ing blue crabs and catfish. Biol. Bull. (Woods Hole) 148:193-198. Gray, R. H., and D. D. Dauble. 1977. Checklist and relative abundance of fish species from the Hanford reach of the Columbia River. North- west Sci. 51:208-215. HOFFMAN. G. L. 1967. Parasites of North American freshwater fishes. Univ. Calif Press, Berkeley, 486 p. KHAIBULAEV, K. KH. 1970. The role of leeches in the life cycle of blood parasites of fishes. [In Russ., Engl.^summ.] Parasitologya (Lenningr.) 4:13-17. 930 \ Klemm. D. J. 1972a. The leeches (Annelida: Hirudinea) of Michi- gan. Mich. Acad. 4:405-444 1972b. Freshwater leeches (Annelida: Hirudinea) of North America. U.S. Environ. Prot. Agency, Biota of Freshwa- ter Ecosystems, Ident. Man. 8, 53 p. 1975. Studies on the feeding relationships of leeches (An- nelida: Hirudinea) as natural associates of mollusks. Sterkiana 58:1-50, 59:1-20. 1976. Leeches (Annelida: Hirudinea) found in North American mollusks. Malacol. Rev. 9:63-76. 1977. A review of the leeches ( Annelida-Hirudinea) in the Great Lakes region. Mich. Acad. 9:397-418. MEYER, M. C. 1940. A revision of the leeches (Piscicolidae) living on fresh-water fishes of North America. Trans. Am. Mi- crosc. Soc. 59:354-376. 1946a. A new leech, Piscicola sahnositica n. sp. (Pis- cicolidae), from steelhead trout iSatnio gairdneri gairdneri Richard.son, 1838). J. Parasitol. 32:467-476. 1946b. Further notes on the leeches (Piscicolidae) living on fresh-water fishes of North America. Trans. Am. Mi- crosc. Soc. 65:237-249. Papern.'\. l, and D. E. ZWERNER. 1974. Massive leech infestation on a white catfish (/c- talurus catus): a histopathological consideration. Proc. Helminthol. Soc. Wash. 41:64-67. Sawyer. R. T. 1972. North American freshwater leeches, exclusive of the Piscicolidae, with a key to all species. 111. Biol. Monogr. 46, 154 p. Sawyer. R. T., A. R. Lawi.er. a.vd R. M. Overstreet 1975. Marine leeches of the eastern United States and the Gulf of Mexico with a key to the species. J. Nat. Hist. 9:633-667. C. Dale Becker Dennis D. Dauble Ecosystems Department Battelle. Pacific Northwest Laboratories Richland, WA 99352 INDUCED SPAWNING AND LARVAL REARING OF THE YELLOWTAIL FLOUNDER, LIMANDA FERRUGINEA The yellowtail flounder, Limanda ferruginea (Storer), a commercially important flatfish, occurs in North American continental waters from the north shore of the Gulf of St. Lawrence southward to the lower part of Chesapeake Bay (Bigelow and Schroeder 1953). The yellowtail flounder spawns from March through August where water temper- atures over its range vary from about 5° to 12°C (Colton 1972). The eggs are pelagic and lack an oil globule; diameter of the live eggs (range 0.79-1.01 mm I averages 0.88 mm (Colton and Marak 1969). A program to obtain viable yellowtail eggs through hormone induction, to rear larvae through metamorphosis, and to determine the mechanisms of survival of early life stages under controlled laboratory conditions was undertaken. The suc- cessful induction of yellowtail flounder and sub- sequent rearing of the larvae through metamor- phosis marks the first time the early life history of this flatfish has been completed in the laboratory. Materials and Methods Adult yellowtail flounder were captured by otter trawling in Block Island Sound in the winters of 1974, 1975, and 1976 and transported to the Nar- ragansett Laboratory in a 380-1 live car equipped with an aerator. In the laboratory the fish were held in a 28,000-1 aquarium. A continual supply of filtered seawater was pumped to the aquarium from Narragansett Bay. Individuals presumed to be sexually mature were selected by length. Available length-weight data (Lux 1969) indicated that yellowtail flounder in southern New England waters mature when they attain a length near 35 cm or an age of 3 yr (Lux and Nichy 1969). After acclimating in the laboratory, the fish were segregated by sex, mea- sured and weighed, and tagged with numbered plastic pennants secured through the caudal pe- duncle. Yellowtail flounder were sexed by holding the white underside to the light and looking through the flesh. The outline of the ovary extend- ing posteriorly from the mass of viscera can read- ily be seen even in immature females (Royce et al. 1959). Yellowtail flounder are delicate and excit- able. To minimize injury, the fish were anes- thetized in a solution of tricane methanesulfonate (MS-222') at a concentration of 1:20,000 (Leitritz and Lewis 1976) during each examination. While the fish were held in captivity, a photo- period of 1 1 h of light and 13 h of dark simulated spawning light conditions. Four banks of fluores- cent lights (each bank composed of 16 40-W bulbs) were suspended 4 m from the ceiling and mechani- cally timed. The light banks were sequentially turned on and off in the morning and evening at 15-min intervals to simulate dawn and twilight. Prior to receiving hormones the fish were fed a daily diet of chopped frozen hake, whiting, or squid. Dur- ing the trials the fish were not fed. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. FISHERY BULLETIN: VOL. 76. NO. 4. 1979. 931 The effectiveness of the pituitary preparations was evaluated by monitoring the gonosomatic index (GSI), ovulation, success of egg fertilization, and hatching success. Hormones were prepared on the day of injection, and dosages were established by the weight of each individual fish. A saline solu- tion of isotonic sodium chloride was used as a car- rier. All injections were administered (2-cm'^ syringe, 20 gage 3.85-cm needle) intramuscularly into the back below the dorsal fin. Inserting and withdrawing the needle slowly aided in retaining most of the fluid in the flesh. After injection, the flesh of the fish was massaged to diffuse the fluid into the muscles. Sexually mature fish were hand stripped and the eggs fertilized in a polyethylene pan. Several thousand eggs were collected at each spawning, and the sperm of two males was used to fertilize the eggs from each female. Yellowtail flounder are nonsynchronous spawners (Bigelow and Schroeder 1953), and multiple spawnings occur- red among most induced fish. The fecundity of yellowtail flounder increases with age and body length, and an individual female may yield from 350,000 to more than 4,000,000 eggs during the spawning season (Pitt 1971). The state of ova maturation of the experimental fish was observed at the start and termination of each experiment. Before injecting, a polyethylene cannula was inserted into the oviduct and oocyte samples were orally withdrawn. The oogenesis of oocytes was divided by microscopic observation into three general histological stages: Stage I - the primary oocyte stage, oocytes con- tained cytoplasmic vacuoles and mea- sured between 0.1 and 0.25 mm. Stage II - the yolk globule stage, cytoplasm of oocytes was filled with dense yolk granules and measured up to 0.6 mm. Stage III - ripe stage, hyaline oocytes present and measured 0.75-1.00 mm in size. Fertilized eggs were incubated in static, aer- ated, black-sided aquaria that had been inocu- lated with the green algae Dunaliella sp. A single application of penicillin (25 international units iIU]/ml) and streptomycin (0.02 mg/ml) at the con- centration of 50 mg/1 was effective in controlling bacterial contamination of the aquaria in almost all cases. Three series of experiments were undertaken to determine the effectiveness of the hormone injec- tions (Table 1). The first trial was conducted in winter 1975 to determine if induced spawning would occur at low winter water temperatures. The second and third were conducted in the springs of 1976-77 and coincided with the yellow- tail flounder's natural spawning season. Hormone dosage levels of 2, 5, and 10 mg/kg fish and frequency of injecting were dictated by previ- ous successful results obtained with the summer flounder, Paralichthys dentatus, (Smigielski 1975a) and winter flounder, Pseudopleuronectes americanus, (Smigielski 1975b). After each trial the female fish were killed and reweighed, the ovaries were examined, and gonosomatic indices were recorded. Prior to receiving hormone injec- tions, all the female test fish in the first trial were in Stage I of oocyte oogenesis, and most females prior to the second and third trials were in Stage II. Males were not injected in the second and third trials because they were sexually ripe. Rfsults and Discussion First Trial In the first trial (Table 1), most females in the group receiving 10 daily injections of 2 mg pitui- tary were refractory with low GSI values (7-13%). One fish hydrated but did not ovulate, and a small number of Stage II ova were found in the ovaries. Hydration is an increase in total body weight. The weight gain is due mostly to water intake and is reflected by higher GSI values as most of the water appears to go into the gonads. Excessive hydration is manifested by grossly bloated fish which in some instances can hydrate to the point of death without ovulating. Table l. — Hormone dosages, water temperatures, and number of yellowtail flounder in each trial. Trial 1 - January 1975 Water temperatures. 3 -6 C (Mean 510) Wate Tnal 2 ■ April 1976 >r temperatures. 7 - (Mean 9.2 C) IOC Water Trial 3 • April temperatures. (Mean 10.1 1977 8,5'- C) 12,5'C Dally carp pituitary dosages Number of females Number of males Uninjected controls Number of females Number of males Sham injected controls Number of females Number of males Sham Injected controls 2.0 mg/kg fish 5.0 mg/kg fish 10.0 mg/kg fish 6 6 6 4 4 4 3 3 3 8 8 7 0 0 0 4 4 4 9 0 4 932 In the group of females receiving 10 daily injec- tions of 5 mg, three fish were refractory with low GSI values (10-15'^>f ), and three hydrated but did not ovulate. Two of the latter fish contained a small number of Stage II ova; the other developed a cloacal plug of membranous tissue and Stage I ova. In the group receiving 10 daily injections of 10 mg, all were refractory with low GSI values (9- 1 17( ), except for one fish that hydrated but did not ovulate. A membranous plug developed in the cloaca of this fish, and it was bloated. A very small number of Stage II ova were found in the ovaries. There was no indication of sexual ripening in the uninjected control fish, and their GSI values were low dO-lS'^). Copious semen was obtained from the males injected at all three dosages; however, fertilization was not attempted. It was reasoned from the first trial that low GSI values (7-159r ) of females coupled with low water temperatures (3°-6°C, mean 5.1°C) that were less than optimum inhibited the effectiveness of the hormones, for although some fish hydrated, they did not ovulate. Second Trial • The results obtained from the second trial were variable (Table 2). All but one fish receiving injec- tions of 2 mg hydrated and ovulated. Two fish died during the trial; one, after yielding spawn on two occasions, developed a membranous plug and be- came grossly bloated. In the group receiving injections of 5 mg, two fish ovulated but the eggs obtained were not fer- tile. Three other fish developed plugs and hy- drated to the point of death. Injections were dis- continued at the first sign of abnormal hydration, but the fish continued to imbibe water. In the group that received 10 mg, five fish ex- perienced excessive hydration manifestated by bloating, plug formation, and, in two instances, death. The membranous plugs were identical to those that developed in the test group that re- ceived hormone dosages of 5 mg. The controls had four fish with signs of hydration but no Stage III ova were found in their ovaries. Third Trial The results of the third trial paralleled those of the second trial at a dosage of 2 mg. Seven of the experimental females hydrated normally and ovu- lated (Table 3). Fertilization and hatching of these eggs were satisfactory and the larvae were nor- mal. The remaining two fish died during the trial, and their ovaries had a small number of Stage III ova. The control fish neither hydrated nor ovu- lated; GSI values were fairly high, but Stage III ova were absent. The anomalous hydration with bloating and formation of membranous plugs during hormonal induction is not unique. Clemens and Grant ( 1964) injected female goldfish, Carassius auratus, with carp pituitary and observed that the gonadal water content increased, apparently in association with ovulation. The hormone regulating the hydra- tion process appeared to be a gonadotropin. Shehadeh and Ellis (1970) reported the forma- tion of plugs in the cloaca in striped mullet, Miigil cephaliis, treated with a combination of salmon pituitary and Synahorin. Sinha ( 1971 ) studied the gonadal hydration response of Puutis gonionotus using the second fraction of molecular seived carp pituitary extract and suggested that the second fraction is involved in osmoregulation, since an injection of an additional amount enhances the rate of water transport resulting in maturation. Hirose and Ishida (1974) studied the effects of Cortisol and human chorionic gonadotropin (HCG) in ayu., Plecoglossus altiuelis, and reported that the water content of the ovary from hormone-treated fish increased by 6%. Smigielski (1975b) reported a similar response in winter flounder injected with pregnant mare serum (PMS) and HCG. Hirose ( 1976) demonstrated that gonadotropin-treated ayu imbided a greater quan- tity of water than control ayu. He suggested that gonadotropin may act on the sodium and po- tassium system or permeability of the egg mem- brane. Hydration is a normal and necessary prelude to maturation and ovulation. The cases of abnormal hydration experienced with yellowtail flounder may be attributed to an adverse reaction to hor- mone dosage. Most of the test fish that hydrated abnormally and became bloated had an increase in body weight of more than 10%. The increase in body weight appeared to be a result of the fish imbibing an excess amount of water. An excessive amount of introduced hormone may upset the water transport or sodium potassium systems, re- sulting in more water bing imbibed. In conclusion, it appears that water tempera- tures higher than 6°C and GSI values approaching 209^^ coupled with carp pituitary injections ap- proximating 2 mg/kg offish is an effective combi- 933 Table 2. — Effects of carp pituitary on yellowtail flounder receiving daily injections. All fish were exposed to 11L:13D photoperiod and water temperatures of 8.5°-12.5°C (Mean 10. TC). Symbols: + = did, 0 = did not hydrate or ovulate. Dosage No. of Total Initial injec- length body weight Weight change GSI tions (mm) (g) (°o initial wt) (% final wt) Effect Hydrated Ovulated Date of spawning 1976 Ferti- lization (%) Hatch (%) 2.0 mg/kg fish Controls 5.0 mg/kg fish Controls 10 mg/kg fish Controls 5 6 4 5 10 5 10 5 4 5 5 7 10 10 348 340 353 435 280 350 396 392 421 817 263 381 + 6.94 - 1 84 -330 -0.96 -049 1.12 19,7 5.6 5.1 4.3 7,0 8.4 450 882 + 954 240 387 721 + 0.98 11,2 420 667 t2.12 18.6 392 611 + 2.63 19.6 361 409 + 1.09 10.5 381 562 + 0.44 9.8 401 683 .+2.11 18.6 501 1,489 + 9.62 29.5 307 566 + 11.69 26.1 435 1.131 + 13.17 249 348 762 -019 4.6 351 491 + 8.47 30.8 336 418 ^0.96 11.0 361 521 + 2.92 19.3 392 587 + 2.87 19.5 406 702 + 2.19 15.1 339 367 + 1 87 129 467 1,259 + 11.15 286 346 593 + 13.57 27,6 352 463 + 15.11 29,6 433 1,245 + 10.76 25,3 341 485 + 14.06 29.1 360 501 + 1.16 126 348 437 + 1.84 14.2 343 428 + 1.23 15.1 389 672 + 2.63 17.9 430 891 + 4.78 19.3 369 551 + 1.63 14.3 + + 0 0 +" + 2 0 0 0 + + 0 0 + Apr 6 80 Apr 7 70 + Apr 6 75 Apr 7 70 Apr 8 80 Apr 9 90 Apr 11 70 + Apr 4 75 Apr 5 80 Apr 6 70 Apr 7 70 Apr 9 75 + Apr 5 80 Apr 7 75 Apr 8 80 + Apr 8 80 Apr 9 80 Apr 11 70 +2 Apr 1 80 Apr 3 75 + Apr 5 80 Apr 6 75 0 0 0 0 + 3 0 0 0 0 +3 0 0 0 0 0 0 0 + 3 0 0 0 0 0 0 0 0 70 75 55 60 70 75 60 80 75 60 50 75 75 70 70 60 50 80 75 70 80 65 'Died, -' 10% Stage III ova in ovaries, not fertilized. ^Plug formed, fish became bloated and hydrated to point of death. ^Stage III ova in ovaries, not fertilized. "Plug formed, fish became bloated. nation for inducing spawning of yellowtail floun- der. Larval Rearing Fertilized yellowtail flounder embryos were in- cubated in 64-1, rectangular, black-sided, static, well-aerated aquaria at a density of approxi- mately 80 embryos/1. The incubating and rearing temperature was 10°C and the salinity 32%o. Banks of 40-W timed fluorescent lights suspended 1 m over the aquaria simulated a day and night regimen of 15 h light and 9 h dark ( 15L:9D). The aquaria were inoculated with green algae (Dunaliella sp.) which may have aided in the re- moval of metabolic waste products produced by the larvae. The algae also served to sustain the zoo- plankton introduced as food. A single application of penicillin (25 lU/ml) and streptomycin (0.02 mg/ml) was effective in controlling bacterial con- tamination in almost all cases. At 10°C, hatching occurred 6-7 days after fertili- zation. Yolk absorption occurred 4-5 days after hatching, which coincided with first feeding. The larvae averaged 2.75 mm long upon hatching and possessed a completely formed gut. The eyes were 934 Table 3. — Effects of carp pituitary onyellowtail flounder receiving daily injections. All fish were exposed to llL:13Dphotoperiod and water temperatures of 8.5°-12.5°C (Mean 10.1°C). Symbols: + = did, 0 = did not hydrate or ovxilate. Dosage No of Injec- tions Total length (mm) Initial body weight (g) Weight change ("o initial wt) GSI (°o final wt) Date of Fertili- fcftect spawning zation Hatch Hydrated Ovulated 1977 (%) (%) 2.0 mg/kg fish Controls 5 6 5 3 391 552 453 898 474 1.298 424 878 412 733 364 486 396 598 405 694 437 945 392 601 446 891 409 667 359 472 + 0.56 + 9.13 + 470 + 7.16 + 2.04 + 0.97 + 1.62 + 2.79 + 3.17 + 2.16 + 2.79 + 1,92 *2 19 46 21.2 6.9- 20.8 5.3 10.7 B.8 7.7 93 18.0 18.9 167 18.4 + ' + + + + 0 0 0 0 Apr 12 75 0 + Apr 12 80 Apr 14 80 0 + Apr. 16 50 Apr 17 60 Apr 18 60 + Apr 17 60 Apr 18 70 + Apr. 17 70 Apr 18 70 + Apr 16 85 + Apr. 14 70 0 0 0 0 80 70 75 60 40 45 55 65 60 55 60 80 'Died: Stage III ova in ovaries pigmented at 1 day and the mouth was functional at 1-3 days after hatching. No abnormahties were observed in hormone-induced larvae. Twenty larvae were sampled weekly (Table 4). The specific growth rates for the 63-day period from first feeding averaged 9.9T7(,day for dry weight (micrograms) and 2.759^ standard length (millimeters). The first fish metamorphosed 54 days after hatching at 17.00 mm standard length. By the 63d day after hatching, all the larvae had completed metamorphosis, and average length was 17.40 mm. Wild copepod nauplii were collected daily from a nearby estuary and fed to the larvae after being sieved to obtain the proper particle size. Larvae of the yellowtail flounder required small food or- ganisms (<100/xm in largest dimension) to in- itiate feeding. The most difficult aspect of rearing the larvae was the problem of obtaining enough food organisms in the size range required. Larval mortality was high for the first 2 wk of feeding, possibly caused by starvation. However, yellow- tail flounder larvae are able to survive for consid- T.ABLE 4. — Size of yellowtail flounder larvae reared artificially at 10°C from hac ing to metamorphosis. Average sizes of 20 larvae are followed by standard deviation. Days after l^ean length Mean dry weight first feeding (mm) (MQ) 1 3.08^0.20 .16.2=4.3 7 3.39*0.25 19.8=43 14 5 16i047 560=201 21 5.92-079 899 = 574 27 682±0,51 126.5 = 340 34 668 + 089 161 5 = 664 41 8 95=1 03 608 6 = 340 2 48 10.53±4.38 1.133-3 = 1.267 8 55 14.73=398 5.576.6 = 2.694.1 63 17.40±2.33 8,635.9 = 3,058.4 erable periods of days without exogenous food. Some larvae were maintained at 8°C and fed suc- cessfully and survived after being deprived of food for 10 days after hatching (Smigielski unpubl. data). As the larvae increased in size through metamorphosis, larger food organisms such as adult copepods, the rotifer Branchion us plicatilis, and the brine shrimp, Artemia sp., were offered. Acknow ledgments The author expresses his appreciation to Hugh Poston of the Tunison Laboratory of Fish Nutrition, Cortland, N.Y., and Geoffrey C. Laurence of the Northeast Fisheries Center Narragansett Laboratory, National Marine Fisheries Service, NOAA for their many helpful criticisms of the manuscript, and Kathy Dorsey and Thomas Halavik for their technical assistance. Literature Cited BIGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Clemens, H. P., and F. B. Grant. 1964. Gonadal hydration of carp [Cyprinus carpio) and goldfish iCarassius auratus) after injections of pituitary extracts. Zoologica (N.Y.) 49:193-210. COLTON, J. B., Jr. 1972. Temperature trends and the distribution of groundfish in continental shelf waters. Nova Scotia to Long Island. Fish Bull., U.S. 70:637-657. Colton, J. B., Jr., and R. R. Marak. 1969. Guide for identifying the common planktonic fish eggs and larvae of Continental Shelf waters, Cape Sable to Block Island. U.S. Bur. Commer. Fish., Biol. Lab., Woods Hole, Mass., Lab. Ref. 69-9, 43 p. 935 HIROSE, K. - 1976. Endocrine control of ovulation in medaka iOryzias latipes) and ayu iPlecoglossus altivelis). J. Fish. Res. Board Can. 33:989-994. HiROSE, K., AND R. ISHin.A. 1974. Effects of Cortisol and human chorionic gonadotrophin (HCG) on ovulation in ayuPlecoglossus altivelis (Temminck & Schlegel) with special respect to water and ion bal- ance. J. Fish Biol. 6:557-564. Leitritz, E., and R. C. Lewis. 1976. Trout and salmon culture (hatchery methods). Calif. Dep. Fish Game, Fish Bull. 164, 197 p. LUX, F. E. 1969. Length-weight relationships of six New England flatfishes. Trans. Am. Fish. Soc. 98:617-621. Lux, F. E., AND F. E. NICHY. 1969. Growth of yellowtail flounder, L/wo/!(/o ferruginea (Storer), on three New England fishing grounds. Int. Comm. Northwest Atl. Fish. Res. Bull. 6:5-25. Pitt, T. K. 1971. Fecundity of the yellowtail flounder (Limanda fer- ruginea) from the Grand Bank, Newfoundland. J. Fish. Res. Board Can. 28:456-457. ROYCE, W. F., R. J. BULLER, AND E. D. PREMETZ. 1959. Decline of the yellowtail flounder {Limanda fer- ruginea) off New England. U.S. Fish Wildl. Serv., Fish. Bull. 59:169-267. Shehadeh, Z. H., and J. N. Ellis. 1970. Induced spawning of the striped mullet Mugil cephalus L. J. Fish Biol. 2:355-360. SINHA, V. R. P. 1971. Induced spawning in carp with fractionated fish pituitary extract. J. Fish Biol. 3:263-272. Smigielski, a. S. 1975a. Hormone-induced spawnings of the summer floun- der and rearing of the larvae in the laboratory. Prog.- Fish Cult. 37:3-8. 1975b. Hormonal-induced ovulation of the winter floun- der, Pseudopleuronectes americanus. Fish. Bull., U.S. 73:431-438. Alphonse S. Smigielski Northeast Fisheries Center Narragansett Laboratory National Marine Fisheries Servece, NOAA R.R. 7A, Box 522 A Narragansett. RI 02882 TRACE METAL CONTAMINATION OF THE ROCK SCALLOP, HINNITES GIGANTEUS, NEAR A LARGE SOUTHERN CALIFORNIA MUNICIPAL OUTFALL' Los Angeles County's submarine discharge of municipal wastewater from the Joint Water Pol- lution Control Plant (JWPCP) off Palos Verdes Peninsula is the single largest anthropogenic source of trace metals to the marine ecosystem off southern California. The 1974 annual mass emis- sion rates of chromium, copper, and zinc via this discharge (4.8 x 10" 1/yr, which underwent pri- mary treatment only) were about 400, 300, and 850 t, respectively; these were approximately 10 times the corresponding inputs measured in 1971-72 surface runoff from southern California ( Young et al. 1973). As a result, bottom sediments around this submarine outfall system are highly contaminated by a number of trace metals (Gallo- way 1972; Young et al. 1975). Here we report ab- normal levels of seven metals in three tissues of the filter-feeding rock scallop, Hinnites gigan- teus,^ that was collected in the discharge zone and thus had been exposed to suspended wastewater particulates. (The adductor muscle of this bivalve mollusc is considered to be a delicacy, and scallops near the discharge are sought by sport divers.) Procedures During 1974, divers collected eight scallops within the size range generally consumed ( 10 to 25 cm in diameter) from depths of about 20 m at three stations in the discharge zone between Whites Point and Point Vicente: these stations were <1 km off Palos Verdes Peninsula. Six scallops in the same size range also were taken from control sta- tions at similar depths off Santa Catalina and Santa Barbara Islands (Figure 1). To check our 1974 results, during 1976 eight specimens within this size range were again collected from this re- gion in the discharge zone. However, we were not able to obtain additional island samples; there- fore, five specimens were collected from each of two coastal stations located approximately 50 km to the north and south of Palos Verdes Peninsula. The samples were frozen in plastic bags after col- lection. Later, digestive gland, gonad, and adduc- tor muscle tissues were excised from each speci- men before it was fully thawed, using a new carbon steel scalpel and a cleaned Teflon^ sheet; the tissues were placed in cleaned polyethylene vials. Care was taken to avoid contaminating the gonadal or muscle tissue samples with sediments or juices from the digestive glands. Following dissection, each sample ( 1 to 2 g wet weight) was digested in 10 ml of a 1:1 nitric acid 'Contribution No. 85 of the Southern California Coastal Water Research Project. '^Formerly Hinnites multirugosus (Roth and Coan 1978). ■''Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 936 FISHERY BULLETIN: VOL. 76. NO. 4. 1979. 1 ■ ■ ■ I 1 LOS ANGELES l°N \ PALOS VERDES ) PENINSULA POINT ,* ^r.- — -"""x VICENTE rVwHITES \ 40- . POINT JWPCP OUTFALLS •0 fi — V SANTA ^--^^^ BARBARA i ) \ 20' f ^ N SANTA CATALINA 1 0 15 1 1 30 1 K \1 1 1 1 119°W 40- 20' 118°W FliilRK 1. — Outfall and island control sites ofT Los Angeles, Calif., for collection of rock scallops. solution (ultrahigh-purity reagent grade) until the remaining volume was about 3 ml. This proce- dure was repeated once, and the final residue was filtered through an acid-washed Whatman No. 40 filter. The filtrate was then diluted to an appro- priate volume, and the treated sample was analyzed by atomic absorption spectrometry. Silver, chromium, copper, nickel, and lead were measured by injecting 2.5 /u,l of sample into a graphite furnace; cadmium and zinc levels were determined by aspirating the sample into an air- acetylene flame. Process blanks were analyzed with all samples. Typical blank corrections were