..<' °?°^. ^ 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 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 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 0,865 936 970 SH BD;ABD-II ABDE ABD-II A2 0,311 0.316 338 360 ^~^-, 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 — — — — — — 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 — — — — — — 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 (%) (%) (%) 0.5 413 248 223 065 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 21.5 062 03 0.17 14 2.5 352 27 6 24 1 089 049 005 390 226 205 0058 03 0.17 13 3.5 284 28 1 246 084 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 3 068 03 0.17 12 6.5 298 302 261 086 043 004 374 22 20.3 0068 03 016 0.13 7.5 280 29 9 269 092 042 004 383 21 9 200 0.074 03 015 10 8.5 289 29 9 256 098 043 004 398 220 202 0074 03 16 0.11 9.5 291 29.8 25.7 096 0.46 004 396 21.8 20.0 0.075 0.2 16 0.08 10.5 275 30 5 266 093 044 004 387 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 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 /-''' •''■ ^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 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 ■•■ 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 10 20 30 10 .>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 Pelecypoda; Saxidomus gigantea^ Protothaca staminea^ Mya truncala'' 481 81 Macoma inquinata'' Macoma incongrua^ Mytilus edulis. musseP 2 0.3 Pododesmus macroschisma Clinocardium nuttalli ' Cephalopoda; Octopus sp 4 0.6 Echinodermata: Asteroidea; Evasterias troschelii 5 0.8 Echlnoidea; Strongylocentrolus drobachiensis 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 48 ='10.9 52 11.8 43 57 "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 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 2.0-5.0 33 2.0 Squid — 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 22 0.3 0.5-2.0 89 9.3 Stomatopod larvae — 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 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 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 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 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 3-4 50 1.8 11 Cladocerons 0.7-1.0 33 0.9 0.4 — 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 — 20-25 17 0.2 0.1 Cumaceans — 2 17 0.1 <0.1 Tanaids 2.0 17 0.6 0.3 — Isopods — 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 — Chaetognaths 2.0-15.0 100 75.3 36.9 3-55 100 440,0 272,8 Larvaceans 1.0-3 100 25.3 12.4 2-4 100 87,4 54,2 Salps — (?) 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 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 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 FIGURE 4 158 F M ' A m' J ' J ' A SO N D I J F M A M J J ' A ' S ' ' 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 / '^^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 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 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 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 Q 8 21 1 22 5 0.2 125 76 201 42 23 20 26 350 15 391 84 4.4 153 95 15 263 54 3.0 48 1 38 7 4 50 11 0.6 14 2 3 19 4 0.2 Total 27 409 23 4 463 100 292 173 18 483 100 A 1000 Surf J 8 2 6 8 1 0,1 48 12 60 11 0.7 20 84 447 8 539 76 6.1 25 315 22 1 363 67 4.1 48 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 K 8 12 1 13 3 0.1 15 8 23 4 03 20 25 411 16 452 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 L 1 1 <1 <0.1 8 12 2 14 4 02 1 16 4 21 6 0.2 20 15 285 24 1 325 85 3.7 29 173 31 3 236 67 2.7 48 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 M 8 13 13 1 0.1 1 62 19 82 13 0.9 20 22 864 18 904 96 10.2 24 444 60 2 530 83 60 48 14 9 6 29 3 03 14 6 8 28 4 0.3 Total 22 891 27 6 946 100 25 520 85 10 640 100 All day tows: Surf 1 1 <1 <0.1 8 2 64 4 70 2 0.2 2 266 119 387 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 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 5.4 374 146 24 544 30 6.1 8 10 268 29 2 309 27 3.5 886 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 40 4 0.5 1 22 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 5 551 126 682 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 12 227 239 17 2.7 48 7 27 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 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 e 1 13 5 19 3 02 1 159 82 242 15 09 20 11 271 218 71 571 93 64 37 774 328 86 1.225 79 46 48 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 8 17 32 49 13 0,6 2 71 44 117 7 04 20 6 151 123 15 295 79 33 115 913 153 16 1,197 74 4.5 48 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 8 2 2 1 <0,1 27 11 38 3 0.1 20 47 150 9 206 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 1 1 <1 <0.1 8 1 5 6 3 0,1 2 33 6 41 4 0.2 20 25 106 3 134 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 1 1 <1 <0.1 1 1 <1 <0.1 8 30 7 37 6 0-4 1 105 26 132 6 0.5 20 14 467 46 527 89 59 60 1,775 124 2 1,961 90 74 48 11 6 10 27 5 0-3 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 1 1 <1 <01 2 2 <1 <01 8 2 65 46 113 5 03 6 395 169 570 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 86 6 1.0 89 648 76 18 831 14 3.1 48 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 43 245 10 28 9 900 315 44 1,268 23 48 8 1.488 274 16 1.778 70 20.0 10 2,642 365 36 3.053 56 11.5 20 32 412 3 447 17 50 110 858 24 6 998 18 3.7 48 7 57 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 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 10 291 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 19 824 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 10 20 30 40 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 % Bntisti Columbia COMMERC lAL 1 25% Washington SPORT 1 20% COMMERCIAL 1 24 % Oregon SPORT ] 3% COMMERCIAL ] 2% California SPORT % COMMERCIAL % Columbia River SPORT % 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 — — — 773 18.9 Marine commercial 4 2.031 269 5 — — — 2.309 56,4 Columbia River sport 14 16 — — — 30 0.7 Columbia River gillnet 11 388 633 17 — — — 1.049 25.6 Columbia River Indian' 11 147 81 — — — 239 58 Total 182 3.068 1.128 22 — — — 4.400 107 4 1962: Marine sport — 34 142 28 — — 204 64 Marine commercial — 234 135 14 — — 383 12,0 Columbia River sport — — — 0,0 Columbia River gillnet — 10 242 88 — — 340 106 Columbia River Indian' — 40 — — 40 1,3 Total — 44 658 251 14 — — 967 303 1963: Marine sport — — 120 368 133 — 621 255 Manne commercial — — 23 966 282 9 — 1.280 52,6 Columbia River sport — — — 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 Columbia River gillnet — — — 15 260 351 19 645 242 Columbia River Indian' — — — 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 618 683 106 — — — 1,407 36.7 Columbia River sport — — — 0.0 Columbia River gillnet 38 402 111 — — — 551 14.4 Columbia River Indian' 4 — — — 4 0.1 Total 23 734 1.192 226 — — — 2.175 56.8 1962: Marine sport — 84 11 8 — — 103 3.5 Marine commercial — 240 194 23 — — 457 15.5 Columbia River sport — 16 — — 16 0.5 Columbia River gillnet — 6 21 46 10 — — 83 2.8 Columbia River Indian' — — — 0.0 Total — 6 345 267 41 — — 659 22.3 1963: Marine sport — — 140 167 66 12 — 385 17.0 Marine commercial — — 366 320 53 — 739 32.7 Columbia River sport — — — 0.0 Columbia River gillnet — — 7 32 44 50 — 133 5.9 Columbia River Indian' — — — 0.0 Total' — — 147 565 430 115 — 1.257 55.6 1964: Marine sport — — — 38 61 40 139 5.2 Marine commercial — — — 132 533 69 734 27.6 Columbia River sport — — — 17 17 0.6 Columbia River gillnet — — — 3 41 68 112 4.2 Columbia River Indian' — — — 3 3 0.1 Total — — — 41 196 631 137 1,005 37,7 'Setnet and dip net fisheries. Southeostern Alosko COMMERCIAL British Columbio COMMERCIAL Woshin q ton 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' 25 109 6 2 31 23 30 9 56 141 36 2 05 1,2 0.0 0.3 0.0 Total 142 84 9 235 2.0 OxBow Hatchery Marine sport Manne commercial Columbia River sport Columbia River gillnet Columbia River Indian' 18 3 78 107 27 6 41 17 3 16 6 14 118 154 17 44 3 3.0 39 0.4 11 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 Washin g ton SPORT COMMERCIAL Ore gon 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' 3 89 29 50 35 5 4 139 71 5 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' 3 19 66 6 28 38 24 4 3 47 107 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 Washin g ton SPORT COM ME RClflL Oregon SPORT COMMERCIAL Colifornia SPORT COM MERCl AL Columbto River SPORT GILLNET INDIAN Southeastern Alasko COMMERCIAL British Columbia COMME RClAL Woshin g ton 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 81 388 8.9 Marine commercial 3 633 617 32 1.285 29.4 Columbia River sport 0.0 Columbia River commercial 72 45 117 2.7 Columbia River Indian' 47 21 68 1.5 Total 164 898 764 32 1,858 425 Big Creek Hatchery: Marine sport 70 209 73 352 5.0 Marine commercial 240 144 7 391 5.5 Columbia River sport 0.0 Columbia River commercial 93 78 171 2.4 Columbia River Indian' 0.0 Total 70 542 295 7 914 12.9 'Setnet and dip net fisheries. Southeastern Alosko COMMERC tAL British Columbio COMMERCIAL Woshin g ton SPORT COMMERCIAL Oregon SPORT COMMERCIAL California SPORT COMMERC lAL Columbia River SPORT GILLNET INDIAN Southeostern Alosko COMMERC tAL British Columbia COMME RC I AL Washin g ton 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 Columbia River commercial 17 Columbia River Indian' 4 Total 165 300 284 Little White Salmon Hatchery: t^arlne sport 40 37 Marine commercial 4 84 125 Columbia River sport Columbia River commerical 5 8 Columbia River Indian' Total 4 129 170 Total catch Potential contribution (in thousands) 264 9.4 8 472 16.8 0.0 5 22 0.8 4 0.1 13 762 27.1 77 28 213 7,7 0,0 13 0.5 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 Washin g ton SPORT COMMERCIAL Ore gon 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 21 — — — 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 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 — — 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 — — — 00 Columbia River gillnet — — 88 1,194 2,168 315 — 3,765 101 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 — — — 0.1 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 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 3.595 319.412 011 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 5.347 579.967 009 Bonneville 1964 762 27 28 2 1 2,711 957.110 003 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 5,446.439 008 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 4,638,237 007 '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 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 23 43 4 1 30 1962 18 41 2 39 1963 34 41 3 {') 21 1964 24 45 7 (') 23 Kalama River 1961 2 48 23 1 e) 26 1962 1 58 24 2 15 1963 2 42 43 3 11 1964 60 23 3 1 13 Elokomin 1961 21 47 13 3 16 OxBow 1961 13 51 15 2 19 Grays River 1962 12 74 5 / 2 Cascade 1962 43 36 2 1 19 Klickitat 1963 39 32 15 4 10 Big Creek 1963 16 57 8 1 19 Bonneville 1964 29 46 17 4 4 Little White 1964 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 477.462 275,965 Speelyai 456.550 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 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 0.3 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 Washin g ton SPORT COMMERCIAL Ore gon SPORT COMMERCIAL California SPORT COMMERC lAL Co umbia River SPORT 6ILLNET INDIAN 3% 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 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 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 1-m^ Petersen grab. 1.0-mm mesh sieve; dredges 2 0, 1 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 5-6 2 5-8 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 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 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 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 5.3 4.6 5.9 3.7 2.9 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 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 4.6 2.5 8.7 2.0 3.7 Isopoda 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 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 Polychaeta 5.3 9.1 23.5 25.9 30.9 47.1 44.1 62.9 Pelecypoda 4.0 45.5 29.4 25.9 21.0 21.7 23.5 22.2 Gastropoda 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 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^ > 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 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 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 . mm 0. 5 mm Figure 6. — Stage VI (megalopa) of Pandalus goniurus: A, ros- trum; B, telson. 1 . 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.£ 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 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 .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!iM i „titii ,,i,, mjMmj ,, m ,,i nMM i, ,mMLMrf ^^ MiiiM i „mM i Mm mMiMiMtiiiM i MUniMinm„i..fcu„i..m„ i ,; t^ O S. brasiliensis • S. maculatus * S. sierra ^2j^>.cz>- \ Bloef.elcJs* ,^-' --^'^-. "mr An^fso^^ \. r»' I|ii llllijn | l i | i llll| l i|ii ^a 25' I ' T'lillM ' T ' lM'T 'imy Hu !i i''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 — — A. nasuta 656 .06 .67 1 .01 .11 A. cubana 248 .02 .44 1 .01 .11 Harengula jaguana 2,725 .26 .67 — — Opisthonema oglinum 521 .05 .33 — — Sardinella anchovia 3 .00 .11 — — 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 — — 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 — — 3 .03 .22 Bathygobius soporator 6 .00 .22 — — Gobiosoma robustum 632 06 .44 336 4.15 .89 Gobionellus boleosoma — — 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 97 27 46 72 11 92 92 10 93 87 4 88 67 3 61 55 7 26 33 53 1 50 4 6 43 31 17 21 25 29 17 43 1 39 2 40 39 15 24 18 20 33 1 31 2 3 30 3 27 4 22 25 21 2 19 3 22 19 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 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).