.<**: °7% *^TCS 0« * Fishery Bulletin in 3 National Oceanic ar d Atmospheric Administration • National Marine Fisheries Service APR 'i S r Woods Hole, .Mass. -\ Vol. 75, No. 1 January 1977 CLARK, STEPHEN H., and BRADFORD E. BROWN. Changes in biomass of finfishes and squids from the Gulf of Maine to Cape Hatteras, 1963-74, as determined from research vessel survey data 1* NELSON, WALTER R., MERTON C. INGHAM, and WILLIAM E. SCHAAF. Larval transport and year-class strength of Atlantic menhaden, Brevoortia tyrannus . . 23 STRUHSAKER, JEANNETTE W. Effects of benzene (a toxic component of petro- leum) on spawning Pacific herring, Clupea harengus pallasi 43 HOSIE, MICHAEL J., and HOWARD F. HORTON. Biology of the rex sole, Glypto- cephalus zachirus, in waters off Oregon 51 HOUDE, EDWARD D. Abundance and potential yield of the round herring, Etru- meus teres, and aspects of its early life history in the eastern Gulf of Mexico ... 61 HAEFNER, PAUL A., JR. Reproductive biology of the female deep-sea red crab, Geryon quinquedens, from the Chesapeake Bight 91 PRIST AS, PAUL J., and LEE TRENT. Comparisons of catches of fishes in gill nets in relation to webbing material, time of day, and water depth in St. Andrew Bay, Florida 103 WHITE, MICHAEL L., and MARK E. CHITTENDEN, JR. Age determination, repro- duction, and population dynamics of the Atlantic croaker, Micropogonias undulatus 109 RICHARDSON, SALLY L., and WILLIAM G. PEARCY. Coastal and oceanic fish larvae in an area of upwelling off Yaquina Bay, Oregon 125 ROHR, BENNIE A., and ELMER J. GUTHERZ. Biology of offshore hake, Merluccius albidus, in the Gulf of Mexico 147 NORRIS, KENNETH S., ROBERT M. GOODMAN, BERNARDO VILLA-RAMIREZ, and LARRY HOBBS. Behavior of California gray whale, Eschrichtius robustus, in southern Baja California, Mexico 159 PEARCY, WILLIAM G., MICHAEL J. HOSIE, and SALLY L. RICHARDSON. Dis- tribution and duration of pelagic life of larvae of Dover sole, Microstomas pacificus; rex sole, Glyptocephalus zachirus; and petrale sole, Eopsetta jordani, in waters off Oregon 173 TRENT, LEE, and PAUL J. PRISTAS. Selectivity of gill nets on estuarine and coastal fishes from St. Andrew Bay, Florida 185 MacINNES, J. R., F. P. THURBERG, R. A. GREIG, and E. GOULD. Long-term cadmium stress in the cunner, Tautogolabrus adspersus 199 LEONG, RODERICK. Maturation and induced spawning of captive Pacific mackerel, Scomber japonicus 205 (Continued on back cover) \t Seattle, Washington U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Administrator NATIONAL MARINE FISHERIES SERVICE Robert W. Schoning, Director Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. EDITOR Dr. Bruce B. Collette Scientific Editor, Fishery Bulletin National Marine Fisheries Service Systematics Laboratory National Museum of Natural History Washington, DC 20560 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. William H. Bayliff Inter-American Tropical Tuna Commission Dr. Roger F. Cressey, Jr. U.S. National Museum Mr. John E. Fitch California Department of Fish and Game Dr. William W. Fox, Jr. National Marine Fisheries Service Dr. Marvin D. Grosslein National Marine Fisheries Service Dr. Edward D. Houde University of Miami Dr. Merton C. Ingham National Marine Fisheries Service Dr. Reuben Lasker National Marine Fisheries Service Dr. 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 through 31 May 1977. Fishery Bulletin CONTENTS Vol. 75, No. 1 January 1977 CLARK, STEPHEN H., and BRADFORD E. BROWN. Changes in biomass of finfishes and squids from the Gulf of Maine to Cape Hatteras, 1963-74, as determined from research vessel survey data 1 NELSON, WALTER R., MERTON C. INGHAM, and WILLIAM E. SCHAAF. Larval transport and year-class strength of Atlantic menhaden, Brevoortia tyrannus . . 23 STRUHSAKER, JEANNETTE W. Effects of benzene (a toxic component of petro- leum) on spawning Pacific herring, Clupea harengus pallasi 43 HOSIE, MICHAEL J., and HOWARD F. HORTON. Biology of the rex sole, Glypto- cephalus zachirus, in waters off Oregon 51 HOUDE, EDWARD D. Abundance and potential yield of the round herring, Etru- meus teres, and aspects of its early life history in the eastern Gulf of Mexico ... 61 HAEFNER, PAUL A., JR. Reproductive biology of the female deep-sea red crab, Geryon quinquedens, from the Chesapeake Bight 91 PRIST AS, PAUL J., and LEE TRENT. Comparisons of catches of fishes in gill nets in relation to webbing material, time of day, and water depth in St. Andrew Bay, Florida 103 WHITE, MICHAEL L., and MARK E. CHITTENDEN, JR. Age determination, repro- duction, and population dynamics of the Atlantic croaker, Micropogonias undulatus ' 109 RICHARDSON, SALLY L., and WILLIAM G. PEARCY. Coastal and oceanic fish larvae in an area of upwelling off Yaquina Bay, Oregon 125 ROHR, BENNIE A., and ELMER J. GUTHERZ. Biology of offshore hake, Merluccius albidus, in the Gulf of Mexico 147 NORRIS, KENNETH S., ROBERT M. GOODMAN, BERNARDO VILLA-RAMIREZ, and LARRY HOBBS. Behavior of California gray whale, Eschrichtius robustus, in southern Baja California, Mexico 159 PEARCY, WILLIAM G., MICHAEL J. HOSIE, and SALLY L. RICHARDSON. Dis- tribution and duration of pelagic life of larvae of Dover sole, Microstomus pacificus; rex sole, Glyptocephalus zachirus; and petrale sole, Eopsetta jordani, in waters off Oregon 173 TRENT, LEE, and PAUL J. PRISTAS. Selectivity of gill nets on estuarine and coastal fishes from St. Andrew Bay, Florida 185 MacINNES, J. R., F. P. THURBERG, R. A. GREIG, and E. GOULD. Long-term cadmium stress in the cunner, Tautogolabrus adspersus 199 LEONG, RODERICK. Maturation and induced spawning of captive Pacific mackerel, Scomber japonicus 205 (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 Notes AUSTIN, C. BRUCE. Incorporating soak time into measurement of fishing effort in trap fisheries 213 MISITANO, DAVID A. Species composition and relative abundance of larval and post-larval fishes in the Columbia River estuary, 1973 218 GUNN, JOHN T., and MERTON C. INGHAM. A note on: "Velocity and transport of the Antilles Current northeast of the Bahama Islands" 222 CREASER, EDWIN P., JR., and DAVID A. CLIFFORD. Salinity acclimation in the soft-shell clam, Mya arenaria 225 GRAVES, JOHN. Photographic method for measuring spacing and density within pelagic fish schools at sea 230 MORROW, JAMES E., ELDOR W. SCHALLOCK, and GLENN E. BERGTOLD. Feeding by Alaska whitefish, Coregonus nelsoni, during the spawning run 234 FISHER, WILLIAM S., and DANIEL E. WICKHAM. Egg mortalities in wild pop- ulations of the Dungeness crab in central and northern California 235 Vol. 74, No. 4 was published on 18 February 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. CHANGES IN BIOMASS OF FINFISHES AND SQUIDS FROM THE GULF OF MAINE TO CAPE HATTERAS, 1963-74, AS DETERMINED FROM RESEARCH VESSEL SURVEY DATA Stephen H. Clark and Bradford E. Brown 1 ABSTRACT Trends in finfish and squid biomass for the 1963-74 period in the International Commission for the Northwest Atlantic Fisheries (ICNAF) Subarea 5 and Statistical Area 6, as evidenced by autumn bottom trawl survey data, were reviewed. Commercial statistics reported to ICNAF reveal that landings for groundfish species of major commercial importance peaked in 1965 and subsequently declined with shifts in directed effort to major pelagic species (for which landings peaked in 1971). Trends in landings for species of lesser commercial importance primarily reflect increasing effort throughout this period. Relative abundance indices (stratified mean catch in kilograms per tow) from the autumn bottom trawl survey revealed drastic declines in abundance of haddock, Melanogrammus aeglefinus; silver hake, Merluccius bilinearis; red hake, Urophycis chuss; and herring, Clupea harengus, during this period although decreases were observed for nearly all finfish species of commercial importance. Possible evidence of changes in species composition were also observed, in that white hake, Urophycis tenuis; Atlantic mackerel, Scomber scombrus; and squids, Loligo pealei and lllex illecebrosus , have shown pronounced increases in relative abundance in recent years coincident with declines in other species occupying similar ecological niches. Analysis for four strata sets (Middle Atlantic, southern New England, Georges Bank, and Gulf of Maine areas) reveal unadjusted declines in biomass ranging from 37% on Georges Bank to 74% in the Middle Atlantic area; by combining data for all strata, a decline of 32% was obtained for the 1967-74 period (including the Middle Atlantic section, added in 1967), while for all remaining strata (1963-74) the corresponding figure is 43%. By adjusting biomass components according to catchability and computing stock size estimates for the entire biomass, a 65% decline was obtained for all strata (including the Middle Atlantic) using untransformed abundance indices, and a 66% decline was computed from retransformed abundance indices. For the remaining strata (Middle Atlantic strata excluded) declines of 47% and 46% were obtained, respectively. By combining these data sets, the corresponding figures were 51% and 47%. Stock size estimates for 1975 approximated 2.0 x 10 e tons, one-fourth of the estimated virgin biomass level and one-half of the level corresponding to maximum sustainable yield. The continental shelf waters of the northwest Atlantic adjacent to the U.S. coast support a valuable and productive fishery resource. Prior to 1960, this area was exploited almost exclusively by a coastal fleet of U.S. vessels of under 300 gross registered tons. Landings averaged less than 500 x 10 3 tons 2 annually (International Commission for the Northwest Atlantic Fisheries 1953-1961), a level substantially lower than the estimated maximum sustainable yield (MSY) of approx- imately 950 x 10 3 tons obtained for this area by various investigators (Au 3 ; Brown et al. 4 ; Brown •Northeast Fisheries Center, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. 2 Landings and estimated stock levels in this paper are given in terms of metric tons. 3 Au, D. W. K. 1973. Total sustainable finfish yield from Subareas 5 and 6 based on yield per recruit and primary pro- duction consideration. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1973, Res. Doc. No. 10, Serial No. 2912 (mimeo.), 7 p. 4 Brown, B. E., J. A. Brennan, E. G. Heyerdahl, M. D. Gross- et al. in press). In the early 1960's, however, distant-water fleets of the U.S.S.R., Poland, and other nations entered the fishery and as that dec- ade progressed these fleets underwent continual modernization and expansion. As a result, fishing effort and landings have increased greatly in this area in recent years. Brown et al. (in press) es- timated that during the 1961-72 period stan- dardized effort increased sixfold, while landings more than tripled. Assessments now indicate that all major stocks in this area are fully exploited and some, notably haddock, Melanogrammus aeglefinus, and herring, Clupea harengus, on Georges Bank and yellowtail flounder, Limanda ferruginea, off southern New England have been lein, and R. C. Hennemuth. 1973. An evaluation of the effect of fishing on the total finfish biomass in ICNAF Subarea 5 and Statistical Area 6. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1973, Res. Doc. No. 8, Serial No. 2910 (mimeo.), 30 p. Manuscript accepted September 1976. FISHERY BULLETIN: VOL. 75, NO. 1, 1977. FISHERY BULLETIN: VOL. demonstrably overfished (Hennemuth 5 ; Brown and Hennemuth 6 ; Schumaker and Anthony 7 ). In addition, the June 1975 report of the ICNAF Standing Committee on Research and Statistics (STACRES) indicates that finfish landings for the 1971-74 period have substantially exceeded the 5 Hennemuth, R. C. 1969. Status of the Georges Bank haddock fishery. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1969, Res. Doc. No. 90, Serial No. 2256 (mimeo.), 21 p. s Brown, B. E., and R. C. Hennemuth. 1971. Assessment of the yellowtail flounder fishery in Subarea 5. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1971, Res. Doc. No. 14, Serial No. 2599 (mimeo.), 57 p. 7 Schumaker, A., and V. C. Anthony. 1972. Georges Bank (ICNAF Division 5Z and Subarea 6) herring assessment. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1972, Res. Doc. No. 24, Serial No. 2715 (mimeo.), 36 p. MSY point (International Commission for the Northwest Atlantic Fisheries 1975c). This expansion in fishing activity in recent years has stimulated considerable interest in its possible effects on biomass levels and productiv- ity. Edwards (1968) developed biomass estimates for the area extending from Hudson Canyon to the Nova Scotia shelf (strata 1-40, Figure 1) by ad- justing 1963-66 U.S. research vessel survey catches to compensate for availability and vul- nerability to the survey gear by species and es- timated that the annual harvest from this area (1.2 x 10 6 tons) approximated one-fourth of the fishable biomass during that period. He also re- ported a rapid decrease in fishable biomass during B FIGURE 1. — Northwest Atlantic area from Nova Scotia to Cape Hatteras, (a) delineated into strata by depth, and (b) delineated into major units for analytical purposes, with ICNAF division boundaries superimposed. CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS the early and mid-1960's and noted that while the decrease had obviously been greater in the case of species for which there were directed fisheries, declines had nevertheless been general. Gross- lein 8 examined autumn research vessel survey data (stratified mean catch per tow, pounds) for the 1963-71 period for southern New England and Georges Bank (strata 1-12, 13-23, and 25, Figure 1) and observed reductions in abundance of over 90% for haddock and ocean pout, Macrozoarces americanus, and more moderate reductions in other components of the groundfish community. Overall, Grosslein's data indicated declines in finfish biomass of 62% and 74% for southern New England and Georges Bank strata, respectively. Brown et al. (see footnote 4) presented additional analyses of Grosslein's data and documented pronounced declines for nearly all groundfish species or species groups, skates (Raja spp.), and sea herring; the decline for all species combined (with individual species weighted by cumulative landings for the 1962-71 period) was 64%. Brown et al. (in press) updated these analyses by includ- ing 1972 data and found an overall decline of 56%. Since 1950, fishery management in the northwest Atlantic region has been conducted under the auspices of ICNAF, an international body currently consisting of 18 member nations pledged to cooperate in research and management of marine fishery resources in the northwest Atlantic area. This Commission, after considering the advice of various standing committees and subcommittees, formulates regulations, estab- lishes quotas or "total allowable catches" (TAC's), and handles other matters necessary for the conservation of fish stocks in the seven regions composing the ICNAF Convention Area. The present study is concerned with the southernmost regions within this area adjoining the U.S. coast (ICNAF Subarea 5 and Statistical Area 6, Figure 1, hereafter referred to as SA 5 and 6). In response to accumulating evidence indicat- ing biomass declines in SA 5 and 6, STACRES in 1973 recommended an overall TAC for this area for 1974 (International Commission for the Northwest Atlantic Fisheries 1974d). Accord- 8 Grosslein, M. D. 1972. A preliminary investigation of the effects of fishing on the total fish biomass, and first approxi- mations of maximum sustainable yield for finfishes in ICNAF Division 5Z and Subarea 6. Part I. Changes in the relative biomass of groundfish in Division 5Z as indicated by research vessel surveys, and probable maximum yield of the total groundfish resource. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1972, Res. Doc. No. 119, Serial No. 2835 (mimeo.), 20 p. ingly, a TAC of 923.9 x 10 3 tons was adopted by the Commission for 1974 to stabilize biomass levels (International Commission for the Northwest Atlantic Fisheries 1974a); for 1975, this figure was reduced to 850 x 10 3 tons (In- ternational Commission for the Northwest At- lantic Fisheries 1974b). In addition, STACRES further recommended that biomass levels, as measured by bottom trawl surveys, be used to monitor the effect of this regulation (International Commission for the Northwest Atlantic Fisheries 1974d). The validity of such an approach is well documented. Grosslein (1971) has presented evidence that abundance indices derived from bottom trawl surveys are of sufficient accuracy to monitor major changes in stock size; for selected groundfish species, current levels of sampling appear adequate to detect changes on the order of 50%. Similarly, Schumaker and Anthony (see footnote 7) and Anderson 9 have found that trends in bottom trawl survey data accurately reflect major changes in stock abundance for pelagic species (herring and Atlantic mackerel, Scomber scombrus, respectively). The objective of the present study was to further investigate changes in biomass of finfishes and squids in SA 5 and 6 as evidenced by trends in U.S. research vessel survey data. In this study, we have expanded on previous analyses of untransformed data (Grosslein see footnote 8; Brown et al. see footnote 4; Brown et al. in press) so as to include all available data from SA 5 and 6 for the 1963-74 period. In addition, we have attempted to com- pensate for anomalies in survey catch data and bias resulting from catchability differences by transforming and weighting data by species and summarizing resulting values to provide com- bined biomass estimates by year. We believe that the resulting trends obtained are more realistic than those derived from unadjusted survey data. In this paper, we define biomass as consisting of weight of all species of finfishes and squids re- ported to ICNAF, excluding other invertebrates and large pelagic species such as swordfish, Xiphias gladius; sharks other than dogfish (Squalus acanthias and Mustelus canis); and tunas, Thunnus spp. We have also chosen to exclude inshore species such as American eel, 9 Anderson, E. D. 1973. Assessment of Atlantic mackerel in ICNAF Subarea 5 and Statistical Area 6. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1973, Res. Doc. No. 14, Serial No. 2916 (mimeo.), 37 p. FISHERY BULLETIN: VOL. 75, NO. 1 Anguilla rostrata; white perch, Morone ameri- cana; and Atlantic menhaden, Brevoortia tyrannus. The latter species is an important component of the biomass, but is taken primarily inshore in the southern portion of SA 6 and is, therefore, not of direct interest in the present study. The term species, for convenience, refers to both species and species groups. Terms such as other pelagics, other fish, and groundfish refer to species so designated in ICNAF statistical bulletins (International Commission for the Northwest Atlantic Fisheries 1965-1973, 1974c, 1975a). BOTTOM TRAWL SURVEY PROCEDURES Autumn bottom trawl survey data have been collected by the U.S. National Marine Fisheries Service RV ALBATROSS IV since 1963; the RV DELAWARE II has also participated infre- quently. In all of these surveys, both vessels have used the standard "36 Yankee" trawl with a 1.25- cm stretched mesh cod end liner. This trawl measures 10-12 m along the footrope and 2 m in height at the center of the headrope, and is equipped with rollers to make it suitable for use on rough bottom (Edwards 1968). The area sampled extends from Nova Scotia to Cape Hatteras. A stratified random sampling design has been used in this survey (Cochran 1953); thus, the survey area has been stratified into geographical zones (Figure 1) primarily on the basis of depth (Grosslein 1969). During 1963- 66, only strata from the New Jersey coast northward (1-42, Figure 1) were sampled; addi- tional strata (61-76, Figure 1) were added in autumn 1967 to cover the mid- Atlantic region (Grosslein 10 ). An additional section covering part of the Scotian Shelf was also added in 1968 but is not considered in this study. In each cruise, sampling stations were allocated to strata roughly in proportion to the area of each stratum and were assigned to specific locations within strata at random. A 30-min tow was taken at each station at an average speed of 3.5 knots. After each tow, weight and numbers captured, fork length, and other pertinent data were re- corded for each species. Data were summarized, '"Grosslein, M. D. 1968. Results of the joint USA-USSR groundfish studies. Part II. Groundfish survey from Cape Hatteras to Cape Cod. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1968, Res. Doc. No. 87, Serial No. 2075 (mimeo.), 28 p. audited, and transferred to magnetic tape follow- ing the completion of each survey. The reader is referred to Grosslein (1969, footnote 11) for further details concerning survey procedures. Following procedures given by Cochran (1953:66) we calculated stratified mean catch per tow values in terms of weight by y, = VN 2 [au] (1) h = V where y st = stratified mean catch per tow, N h = area of the hth stratum, N = total area of all strata in the set, ft, — mean catch per tow in the hth stratum, and k = number of strata in the strata set. We calculated the estimated population variance as S 2 = 1/A7 k I h = l N h y h '- Ny st 2 +1^ /! = ! (N h 1) + (N h - N) (N h - n h ) N m (2) where S 2 = estimated population variance, n h = number of tows in the hth stratum, s/, 2 = variance within the hth stratum, and y st , N, N h , y h , and k are defined as before. We used stratified mean weight per tow (kilograms) in preference to numbers as an index of biomass change due to its convenience when working with different species groups and the high degree of variability in numbers associated with fluctuations in recruitment. Obviously, numbers would also tend to overemphasize the importance of small organisms in the community under study, as pointed out by Odum and Smalley (1959). RECENT TRENDS IN LANDINGS Commercial landings as reported to ICNAF (International Commission for the Northwest Atlantic Fisheries 1965-1973, 1974c, 1975a, "Grosslein, M. D. 1969. Groundfish survey methods. NMFS, Woods Hole, Mass., Lab. Ref. No. 69-2, 34 p. CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHKS AND SQUIDS footnote 12) for the major species groups consid- ered in this paper (principal groundfish, princi- pal pelagics, flounders, other groundfish, other pelagics and other fish, and squid, Table 1) are given in Figures 2 and 3. Effort was concentrated on principal groundfish during the mid-1960's; landings peaked at approximately 643 x 10 3 tons in 1965, declined to approximately 575 x 10 3 tons in 1966, and dropped off sharply thereafter (Fig- ure 2). Statistical data for individual species (International Commission for the Northwest Atlantic Fisheries 1965-73, 1974c, 1975a, see footnote 12) reveal that this pattern resulted primarily from great increases in landings of cod, haddock, and silver and red hake in the mid- 1960's, followed by subsequent declines. Landings of redfish and pollock have increased somewhat in more recent years, but not enough to offset de- clines in the remaining species. Landings for principal pelagics during this period (herring and mackerel) declined initially followed by a subsequent upswing. This can be attributed primarily to a diversion of USSR effort from herring to haddock and hake in 1965 and 1966 (Schumaker and Anthony see footnote 7). In 1967, however, the USSR redirected much of its effort back to the Georges Bank herring stock and also initiated an intensive mackerel fishery (Anderson see footnote 9) and other distant water fleets also began to exploit these species at about this time. This increase in effort produced in- creased landings of herring and mackerel to a total TABLE 1. — Scientific and common names of species considered 1 in this study, grouped as in ICNAF statistical bulletins. "International Commission for the Northwest Atlantic Fisheries. 1975. Provisional nominal catches in the Northwest Atlantic, 1974 (Subareas 1 to 5 and Statistical Areas and 6). Int. Comm. Northwest Atl. Fish. Annu. Meet. 1975, Summ. Doc. No. 32, Serial No. 3590 (mimeo.), 61 p. ^— ^— Principal groundfish Principal pelagics / \ ' 1 \ 1 / \ / \ ' ^-' / \ ' ^ / \ 1 / \ ' / \ ' / V / AS. / N~ — -~">v / \ N • S/ / \ \ \ \ 63 64 65 66 67 68 69 70 71 72 73 74 YEAR FIGURE 2. — Landings of principal groundfish and principal pelagics in ICNAF Subarea 5 and Statistical Area 6, 1963-74. Common name Scientific name Principal groundfish (except flounders): Cod Haddock Redfish Silver hake Red hake Pollock (saithe) Flounders: American plaice Witch Yellowtail Winter flounder Summer flounder Other groundfish: Angler Cusk Ocean pout Sculpins Scup Searobins White hake Principal pelagics: Herring Mackerel Other pelagics and other fish: Butterfish Spiny dogfish Skates and rays Squid: Short-finned squid Long-finned squid Gadus morhua Melanogrammus aegletinus Sebastes marinus Merluccius bilinearis Urophycis chuss Pollachius virens Hippoglossoides platessoides Glyptocephalus cynoglossus LJmanda ferruginea Pseudopleuronectes americanus Paralichthys dentatus Lophius americanus Brosme brosme Macrozoarces americanus Myoxocephalus spp. Stenotomus chrysops Prionotus spp. Urophycis tenuis Clupea harengus Scomber scombrus Poronotus triacanthus Squalus acanthias Raja spp. ///ex illecebrosus Loligo pealei 1 Note that for all groupings except principal groundfish, principal pelagics, and squid, other species were considered but are not mentioned specifically. g 70 S — Fkujnders Other groundfish — Other pelages and other fish Squid 63 64 65 66 67 68 69 70 71 72 73 74 FIGURE 3. — Landings of flounders, other groundfish, other pelagics and other fish, and squid in ICNAF Subarea 5 and Statistical Area 6, 1963-74. FISHERY BULLETIN: VOL. 75, NO. 1 of approximately 667 x 10 3 tons in 1971 (Figure 2). Landings of herring and mackerel peaked in 1968 and 1972, respectively (International Commission for the Northwest Atlantic Fisheries 1965-1973, 1974c, 1975a, see footnote 12). Landings for the remaining species groups (Figure 3) generally reflect decreasing abundance in response to increasing effort. Landings of flounders were relatively constant but did in- crease until 1969 followed by a gradual decline. The somewhat anomalous 1969 value resulted primarily from sharply increased catch of yellow- tail by distant water fleets (Brown and Henne- muth see foonote 6). Steadily declining landings of other groundfish throughout the period of study can be attributed in part to declining abundance, while other pelagics and other fish show a general increase which would appear to be associated with increased effort as shown later. Squid landings also increased sharply since 1970. As TAC's have been imposed for certain stocks since 1970, their possible influence should be considered. It is not believed, however, that quota management affected these trends appreciably. Species subject to quota management in 1970 and 1971 (i.e., haddock and yellowtail) had already been seriously depleted, while in 1972 and 1973 TAC's did not appear to be limiting with the ex- ception of those imposed for haddock, yellowtail, and herring, and for the latter two species TAC's were in fact exceeded (International Commission for the Northwest Atlantic Fisheries 1975c). It appears likely that TAC's imposed for 1974 had a greater effect, particularly in the case of herring and mackerel; also, the overall TAC of 923.9 x 10 3 tons (referred to above) undoubtedly limited total catches by nation to some degree although it was exceeded by approximately 75 x 10 3 tons (In- ternational Commission for the Northwest At- lantic Fisheries 1975c). In summary, however, it would appear that the influence of quota management on the overall trends depicted in Figures 2 and 3 was relatively minor for the level of effort being exerted which, as noted previously, increased by a factor of six during the period 1962-72. It is not possible to speculate whether or not significant additional effort would have been added in 1973 and 1974 (say from new entrants to the area), had there not been regulations. The possible influence of bias upon reported landings remains to be mentioned. In ICNAF statistical bulletins, some landings have been recorded as "not specified," e.g., "groundfish (not specified)," "other pelagics (not specified)," etc. Insofar as possible, we have combined these landings with landings data reported by species within each species group. In recent years, however, an improvement has occurred in re- porting accuracy which appears to have affected the relative amounts of "not specified" landings (and thus annual totals as depicted in Figures 2 and 3). For instance, examination of data in ICNAF statistical bulletins (International Commission for the Northwest Atlantic Fisheries 1965-1973, 1974c, 1975a) reveals a decrease in the relative percentage of "not specified" groundfish of from 15 to 20% of the other groundfish category in the mid-1960's to approximately 10% in 1970-73, while for "other fish" a complete reversal of this trend occurred. The "not specified" proportion of the total "other fish" category increased from approximately 10% in the mid-1960's to 25-30% during 1970-73. This implies that landings for principal groundfish and other species may have been erroneously included under other groundfish to a greater extent in former years, thus biasing the observed trend for other groundfish down- ward, while the trend for other pelagics and other fish may have been biased upward due to inclusion of previously omitted landings data in more recent years. The actual extent to which trends depicted in Figures 2 and 3 were distorted by this factor is problematical, but it should be noted that for principal groundfish, principal pelagics, floun- ders, and squid, more important (and/or more readily identified) species were involved which probably were not affected by reporting inac- curacies to the same degree. Consequently, it is our judgement that trends for the remaining species groups were probably not appreciably biased. CHANGES IN BIOMASS Unweighted Analyses Summaries of survey data by species and area permit preliminary evaluation of the magnitude and direction of change in selected biomass components in recent years and of the degree of year-to-year variability that may be encountered. Accordingly, we examined trends for different species and strata sets and for data summed over all strata before attempting transformation or weighting procedures. Individual strata can be grouped for analysis on CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS the basis of stock structure, ecological factors, exploitation patterns, and availability of survey data. In the present paper, we have selected four major strata sets in SA 5 and 6 based on the above factors (Figure 1) which we considered separately prior to examination of data for the area as a whole. These are as follows: 1. Middle Atlantic area (strata 61-76, cor- responding approximately to ICNAF Di- visions 6B and C); 2. Southern New England area (strata 1-12, corresponding approximately to ICNAF Divisions 6A and Subdivision 5Zw); 3. Georges Bank (strata 13-25, corresponding approximately to ICNAF Subdivision 5Ze), and 4. Gulf of Maine (strata 26-30 and 36-40, corresponding approximately to ICNAF Division 5Y). The rationale for this arrangement is based on differences in faunal assemblages although exploitation patterns and data availability were also considered. A number of stock identification studies support such an arrangement (Wise 1962; Grosslein 1962; Anthony and Boyar 1968; Ridg- way et al. 13 ; Anderson 14 ; and others). In addition, Grosslein's 15 study indicated a relatively high diversity of species in the southern New England-Middle Atlantic areas in contrast to the Gulf of Maine, with Georges Bank being a rather transitional area. Exploitation patterns and reporting of commercial fishery statistics also dictate some form of division between Subdivision 5Ze and the Subdivision 5Zw-Statistical Area 6 region and other areas to the north or south (Fig- ure 1). Finally, the fact that survey data are nonexistent for Middle Atlantic strata prior to 1967 required a division between this area and the remainder of SA 5 and 6 for analytical purposes. Trends in relative abundance from 1963 to 1974 (stratified mean catch per tow [kilograms], U.S. autumn bottom trawl survey data) are given by area for selected species in Tables 2-5 and for major ICNAF categories in Figures 4-9. Pro- nounced declines of principal groundfish are evident both on Georges Bank and in the Gulf of Maine, with lesser declines in the remaining areas (Figure 4). The trends observed resulted primarily from declining relative abundance of haddock and silver and red hake (Tables 2-5). Haddock, in particular, appears to have greatly decreased on 13 Ridgway, G. J., R. D. Lewis, and S. Sherburne. 1969. Serological and biochemical studies of herring populations in the Gulf of Maine. Cons. Perm. Int. Explor. Mer, Memo No. 24, 6 p. 14 Anderson, E. D. 1974. Comments on the delineation of red and silver hake stocks in ICNAF Subarea 5 and Statistical Area 6. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1974, Res. Doc. No. 100, Serial No. 3336 (mimeo.), 8 p. 15 Grosslein, M. D. 1973. Mixture of species in Subareas 5 and 6. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1973, Res. Doc. No. 9, Serial No. 2911 (mimeo.), 20 p. TABLE 2. — Stratified mean catch per tow (kilograms) for selected species of finfish and squid, Albatross IV autumn bottom trawl survey data, 1967-74, Middle Atlantic area (strata 61-76). Species 1967 1968 1969 1970 1971 1972 1973 1974 Principal groundfish: Silver hake 0.9 0.9 0.1 0.2 0.3 0.5 0.4 '0.0 Red hake 0.1 0.8 0.5 0.2 0.4 0.2 0.1 0.0 Flounders: Yellowtail 3.4 5.5 3.6 '0.0 0.3 0.1 1 0.0 0.0 Winter flounder 1.7 1.3 0.6 1 0.0 0.2 0.1 0.1 0.0 Summer flounder 2.0 1.5 0.8 '0.0 0.4 0.1 0.3 0.8 Other 0.7 2.0 0.6 4 0.8 10 1.6 0.5 Other groundfish: Angler 0.7 0.6 0.3 '0.0 0.1 1.4 0.9 '0.0 Scup 2.6 0.8 8.4 0.1 0.3 3.2 0.2 0.7 Searobins 130.1 13.8 5.4 6.9 3.1 1.7 1.9 1.9 Other 05 0.3 0.3 '0.0 '0.0 '0.0 '0.0 0.0 Principal pelagics: Herring 0.0 0.0 0.0 0.0 0.0 0.0 '0.0 0.0 Mackerel '0.0 0.1 0.0 00 '0.0 0.0 0.0 0.0 Other pelagics and other fish: Butterfish 3.6 18.1 3.9 5.4 5.0 4.2 11.0 3.7 Spiny dogfish 47.8 3.1 4.9 0.0 0.0 0.0 '0.0 Skates and rays 4.0 8 4 29.5 7 12.8 6.6 10.4 5.4 Other 2 9.8 7.0 4.5 59 9.6 3.1 94 3.3 Squid: Short-finned squid 0.3 0.2 0.1 0.4 0.2 0.3 '0.0 0.1 Long-finned squid 10.6 9.3 9.2 48 2.5 12.6 11.2 11.1 Total finfish and squid 218.8 73.7 72.7 31.3 36.0 35.1 47.5 27.5 'Less than 0.05. 2 Does not include data for tunas, sharks, swordfish, American eel, or white perch. FISHERY BULLETIN: VOL. 75, NO. 1 TABLE 3. — Stratified mean catch per tow ( kilograms) for selected species of finfish and squid, Albatross IV autumn bottom trawl survey data, 1963-74, southern New England area (strata 1-12). Species 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 Principal groundfish: Cod 3.0 0.5 18 0.7 2.9 08 1.5 0.6 0.1 2.1 '0.0 04 Haddock 2.7 7.1 1.2 0.1 0.5 '0.0 0.1 0.5 0.1 0.0 '0.0 0.0 Silver hake 5.2 5.7 7.6 36 4.4 4.8 2.3 2.6 4.6 4.0 3.2 1.3 Red hake 8.1 4.4 5.6 2.9 2.7 4.4 48 3.9 3.4 6.6 3.0 05 Flounders: Yellowtail 12.0 11.8 8.7 79 11.9 11.1 12.3 137 7.6 26.8 2.6 1.2 Winter flounder 2.4 3.1 3.1 2.1 1.5 1.0 1.3 2.4 1.0 3.0 0.5 0.4 Other 4.8 3.8 2.7 45 1.9 29 1.7 1.9 1.3 2.9 2.4 2.9 Other groundfish: Angler 4.4 7.0 49 6.7 1.9 1.2 2.5 28 1.5 9.8 2.9 1.0 Ocean pout 0.7 0.4 0.3 1.1 0.6 0.5 0.3 0.3 0.1 0.1 0.2 0.0 Sculpins 0.3 1.0 1.7 2.5 1.6 1.0 1.4 1.1 0.3 2.2 0.1 0.1 Scup 1.3 2.5 0.7 0.5 0.6 0.4 1.6 0.4 0.2 1.9 1.6 1.4 Searobins 1.0 0.8 0.5 0.7 08 0.3 0.5 0.2 0.3 4.7 0.3 0.1 White hake 12 04 0.6 1.2 1.3 1.4 0.6 0.5 0.4 0.4 01 1 Other 0.1 0.1 0.1 '0.0 0.3 '0.0 0.1 0.1 0.3 '0.0 '0.0 0.0 Principal pelagics: Herring 0.2 '0.0 0.5 1.8 05 0.1 '0.0 '0.0 '0.0 '0.0 0.0 00 Mackerel '0.0 "0.0 '0.0 '0.0 1.0 0.2 3.9 '0.0 0.1 '0.0 '0.0 '0.0 Other pelagics and other fish: Butlerfish 26 6.0 4.5 1.5 22 4.0 6.5 1.1 58 2.4 63 6.1 Spiny dogfish 71.2 194 4 93.0 924 969 585 216.5 676 13.2 327 46.1 18.6 Skates and rays 15.8 10.4 11.3 13.6 3.7 1.2 2.3 2.9 6.6 9.1 3.0 32 Other 2 01 1.9 2.0 0.7 1.7 1.3 4.1 5.1 4,1 3.1 5.3 52 Squid Short-finned squid ( 3 ) 4 0.1 "0.1 «0.1 05 07 1 0.3 03 0.6 1 02 Long-finned squid ( 3 ) «1.2 "16 "22 2.0 122 181 3.6 5.4 67 167 12.1 Total finfish and squid 137.1 262 6 1525 1468 141.4 108 282 5 1116 567 119.1 944 548 'Less than 0.05 2 Does not include data for tunas, sharks, swordfish. American eel. or white perch 3 Data not recorded 4 Squid catches for 1964-66 prorated by species according to relative percentages caught in later years TABLE 4. — Stratified mean catch per tow (kilograms) for selected species of finfish and squid, Albatross IV autumn bottom trawl survey data, 1963-74, Georges Bank area (strata 13-25). Species 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 Principal groundfish: Cod 11.0 7.1 7.2 5.0 8.4 5.3 4.9 7.8 6.1 14.2 19.1 5.1 Haddock 51.2 75.2 56.1 21.4 20.5 9.3 5.8 10.6 3.6 5.1 7.2 2.8 Redfish 0.9 4.0 1.1 2.0 2.6 3.5 6.5 4.6 1.9 3.9 2.6 1.9 Silver hake 5.4 1.7 1.6 2.1 1.0 2.2 1.6 2.3 1.2 2.4 2.4 1.5 Red hake 7.4 2.2 1.8 1.2 0.8 1.1 1.5 0.9 1.9 1.2 2.8 1.4 Pollock 2.3 2.1 1.7 2.9 1.1 1.0 1.4 0.4 2.2 1.0 1.6 0.4 Flounders: American plaice 5.5 2.0 1.2 3.3 1.7 1.3 1.1 1.5 0.9 0.9 0.9 0.4 Witch 1.0 0.5 0.5 1.5 0.6 0.9 0.5 1.5 0.5 1.0 1.5 0.4 Yellowtail 8.2 8.4 5.6 2.5 4.5 6.7 5.4 3.0 3.7 4.0 3.8 2.2 Winter flounder 1.8 2.1 2.0 3.6 1.3 1.5 1.7 4.7 1.0 1.5 1.6 1.5 Other 1.0 0.7 0.6 1.1 1.1 1.2 1.3 0.4 0.6 1.3 3.5 1.8 Other groundfish: Angler 3.5 2.6 5.0 5.8 0.5 1.9 1.1 0.7 0.6 1.6 2.2 1.1 Ocean pout 1.7 1.0 0.9 0.9 0.2 0.1 '0.0 0.1 '0.0 0.4 0.2 '0.0 Sculpins 3.4 1.8 3.3 3.3 2.0 3.8 3.1 4.9 3.1 2.8 3.6 2.0 White hake 1.4 0.5 0.8 '0.0 1.6 1.0 1.8 2.4 2.2 2.2 3.5 2.0 Other 0.5 0.5 0.6 1.0 0.7 1.0 0.2 0.5 0.1 0.4 0.7 0.3 Principal pelagics: Herring 1.0 0.2 0.9 1.5 0.6 0.2 0.2 '0.0 0.3 0.1 '0.0 '0.0 Mackerel '0.0 0.0 0.1 0.1 0.2 0.2 0.4 0.1 '0.0 0.4 '0.0 0.3 Other pelagics and other fish: Butlerfish 0.7 1.3 0.3 0.1 0.6 1.0 0.3 0.2 1.1 1.2 0.4 1.0 Spiny dogfish 2.9 3.0 3.5 1.8 2.5 5.6 2.4 3.5 3.3 9.7 36.2 2.2 Skates and rays 31.3 15.0 21.7 17.7 15.2 12.3 8.7 15.7 8.9 15.4 28.9 15.4 Other 2 0.5 0.4 0.5 0.5 0.5 0.4 0.4 0.2 0.6 0.9 1.0 2.8 Squid: Short-finned squid ( 3 ) "0.2 «0.5 "0.3 0.1 0.3 '0.0 0.2 0.4 0.2 5.0 0.1 Long-finned squid ( 3 ) 4 0.2 "0.5 "0.4 0.4 0.4 1.5 1.1 1.0 1.1 0.1 2.2 Total finfish and squid 142.6 132.7 118.0 80.0 68.7 62.2 51.8 67.3 45.2 72.9 128.8 48.8 'Less than 0.05. 2 Does not include data for tunas, sharks, swordfish, American eel, or white perch. 3 Data not recorded. 4 Squid catches for 1964-66 prorated by species according to relative percentages caught in later years. CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS TABLE 5. — Stratified mean catch per tow (kilograms) for selected species of finfish and squid, Albatross IV autumn bottom trawl survey data, 1963-74, Gulf of Maine area (strata 26-30 and 36-40). Species 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 Principal groundfish: Cod 10.9 14.1 7.4 8.0 5.7 12.0 9.5 10.2 10.2 8.0 5.4 5.5 Haddock 39.1 14.2 12.8 10.1 9.8 11.9 7.8 4.3 5 1 3.2 5.3 2.2 Redfish 269 59.1 14.0 31.8 25.7 432 21.3 33.8 25.4 250 17.3 264 Silver hake 28.3 4.8 8.7 4.2 26 2.0 26 2.4 3.0 63 4.0 3.9 Red hake 4.9 0.7 1.0 0.8 0.3 0.1 0.3 0.1 1.0 2.0 05 0.5 Pollock 8.6 7.8 3.6 2.4 2.9 5.4 13.1 3.6 5.5 84 5.9 62 Flounders: American plaice 6.2 3.6 6.0 6.3 3.5 4.3 3.5 2.5 2.9 2.2 2.9 2.3 Witch 36 23 2.5 4.5 2.0 3.7 5.1 3.4 3.2 2.3 1.3 1.6 Other 1.1 0.4 1.0 0.1 '0.0 0.1 1.2 0.3 0.1 0.7 0.2 0.6 Other groundfish: Angler 3.7 1.6 1.9 3.6 1.7 2.0 4 5 3.1 4.0 1.5 3.6 2.3 Cusk 2.2 1.2 1.3 3.8 1.1 1.8 1.7 2.0 1.8 3.0 1.3 0.5 White hake 7.8 52 7.9 9.5 4.2 5.8 17.7 16.3 15.3 16.9 15.9 14.0 Other 0.3 0.4 0.6 1.0 0.2 0.5 0.1 0.6 0.3 0.8 0.4 0.3 Principal pelagics: Herring 16 0.1 0.2 0.3 0.1 '0.0 "0.0 0.1 0.6 '0.0 '0.0 '0.0 Mackerel '0.0 0.0 0.0 '0.0 00 '0.0 '0.0 '0.0 '0.0 '0.0 '00 '00 Other pelagics and other fish: Spiny dogfish 58.2 10.6 11.8 4.0 78 22.8 98 18.3 119 17.3 7.2 8.7 Skates and rays 15.1 9.4 111 17.4 4.9 10.0 14.4 16.2 12.1 7.9 7.6 4.4 Other 2 2.5 0.1 0.2 0.3 0.4 0.2 0.1 0.3 0.2 0.3 0.2 02 Squid: Short-finned squid ( 3 ) '•"0.0 "0.2 "0.4 0.1 0.1 0.1 0.3 0.5 0.2 06 1.2 Long-finned squid ( 3 ) "0.0 '• 4 0.0 "0.1 '0.0 '0.0 '0.0 '0.0 '0.0 '0.0 '0.0 '00 Total finfish and squid 221.0 135.6 92.2 108.6 73.0 125.9 112.8 117.8 103.2 106.0 79.6 80.8 'Less than 0.05 2 Does not include data for tunas, sharks, swordfish, American eel, or white perch 3 Data not recorded. 4 Squid catches for 1964-66 prorated by species according to relative percentages caught in later years. Middle Altanlic So New England Georges Bank Gulf ol Marne 72 73 74 FIGURE 4— Catch of principal groundfish in U.S. autumn bot- tom trawl surveys for the Middle Atlantic (strata 61-76), 1967- 74, and for southern New England (strata 1-12), Georges Bank (strata 13-25), and the Gulf of Maine (strata 26-30 and 36-40), 1963-74. Georges Bank and in the Gulf of Maine and to be almost nonexistent in southern New England waters. Relative abundance indices for redfish and pollock, however, appear to have remained rela- tively stable (Tables 4, 5). Cod declined somewhat in the Gulf of Maine but remained relatively sta- ble in other areas (Tables 3-5). Catches of flounders indicate substantial de- clines in relative abundance for all areas (Figure 5) and nearly all species (Tables 2-5) with yellow- tail declining very sharply in recent years. Unusually high catches of yellowtail were taken in southern New England waters in 1972 (Figure 5, Table 3); factors involved are unclear but appear to reflect changes in availability, as actual in- creases in abundance do not appear to have oc- curred (Parrack 16 ). Data for other groundfish (Figure 6) suggest a decline in biomass for Middle Atlantic strata, an increase for Gulf of Maine strata, and relatively stable levels elsewhere. The observed trend for Middle Atlantic strata is strongly influenced by large catches of searobins in 1967 (Table 2) which 16 Parrack, M. L. 1973. Current status of the yellowtail floun- der fishery in ICNAF Subarea 5. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1973, Res. Doc. No. 104, Serial No. 3067 (mimeo.), 5 p. FISHERY BULLETIN: VOL. 75, NO. 1 g 25 £ 20 3 15 Middle Atlantic So Mew England Georges Bank Gulf of Mome FIGURE 5.— Catch of flounders in U.S. autumn bottom trawl surveys for the Middle Atlantic (strata 61-76), 1967-74, and for southern New England (strata 1-12), Georges Bank (strata 13- 25), and the Gulf of Maine (strata 26-30 and 36-40), 1963-74. Middle Atlantic So New England Geofges Bonk Gulf of Mome 66 69 70 YEAR FIGURE 6. — Catch of other groundfish in U.S. autumn bottom trawl surveys for the Middle Atlantic (strata 61-76), 1967-74, and for southern New England (strata 1-12), Georges Bank (strata 13-25), and the Gulf of Maine (strata 26-30 and 36-40), 1963-74. continued to decline in succeeding years. Ocean pout also appear to have declined sharply during the period of study in southern New England and Georges Bank strata (Tables 3, 4). Abundance of white hake, however, appears to have increased in the Gulf of Maine in recent years (Table 5), leading to an increase in other groundfish biomass for these strata. Principal pelagics appear to have declined in relative abundance although considerable fluctuation is evident (Figure 7). Most of this variation is, however, associated with the pres- ence of outstanding year-classes of herring in the early and mid-1960's (Schumaker and Anthony see footnote 7) and the appearance of an out- standing year-class of mackerel in 1967 (Anderson see footnote 9). Considerable fluctuation is also evident in catches of other pelagics and other fish (Figure 8, Tables 2-5) although the trend is generally downward (anomalous peaks relate primarily to high catches of spiny dogfish in cer- tain years). Data for squid (Figure 9) indicate increased abundance although catches of long- finned squid appear to be lower in 1970 and 1971 in Middle Atlantic strata and from 1970 to 1972 in southern New England strata than in the years immediately preceding and following (Tables 2, 3). The actual degree of change throughout the period of study is uncertain, however, in that complete records of catches for squid were not kept prior to 1967. A summary of trends in relative abundance by area is given in Tables 6 and 7 and Figure 10. We computed percentage changes from mean catch values (averaged over 1967-68 and 1973-74 for Middle Atlantic strata and 1963-65 and 1972-74 for all other strata sets). We obtained declines of FIGURE 7.— Catch of principal pelagic species in U.S. autumn bottom trawl surveys for the Middle Atlantic (strata 61-76), 1967-74, and for southern New England (strata 1-12), Georges Bank (strata 13-25), and the Gulf of Maine (strata 26-30 and 36-40), 1963-74. 10 CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS M-ddle AltantK — — So New England Georges Bonk Gulf of Mome FIGURE 8. — Catch of other pelagics and other fish in U.S. au- tumn bottom trawl surveys for the Middle Atlantic (strata GI- TS), 1967-74, and for southern New England (strata 1-12), Georges Bank (strata 13-25), and the Gulf of Maine (strata 26-30 and 36-40), 1963-74. over 90% for certain species, while for all data combined we obtained declines of 74%, 52%, 37%, and 41% for the Middle Atlantic, southern New England, Georges Bank, and Gulf of Maine areas, respectively. Omission of catches of searobins for the Middle Atlantic area, however, reduces that value to 52%. Further omitting data for squid for all strata sets (as squid catches were inadequately recorded during the early years of the survey) provides corresponding declines of 62% , 58% , 38% , and 41%. Consequently, even greater declines may be more realistic than those initially com- puted. After examining data for the above strata sets, we evaluated trends for the entire region by combining data over all strata (Tables 8, 9) and compared between means of initial and final periods (1967-68/1973-74 data for all strata; 1963-65/1972-74 data, Middle Atlantic strata excluded). For 1967-74, all strata (Table 8), we observed a decline of 32%, while for 1963-74, FIGURE 9. — Catch of squid in U.S. autumn bottom trawl surveys for the Middle Atlantic (strata 61-76), 1967-74, and for southern New England (strata 1-12), Georges Bank (strata 13-25), and the Gulf of Maine (strata 26-30 and 36-40), 1963-74. 58 69 YEAR FIGURE 10.— Catch of total finfish and squid in U.S. autumn bottom trawl surveys for the Middle Atlantic (strata 61-76), 1967-74, and for southern New England (strata 1-12), Georges Bank (strata 13-25), and the Gulf of Maine (strata 26-30 and 36-40), 1963-74. Middle Atlantic strata excluded (Table 9), the decline is 43% . The corresponding figures are 37% and 46%, respectively, with squid omitted. The above data demonstrate that significant changes in biomass levels occurred in SA 5 and 6 after the early 1960's. It will be noted, however, that the summaries presented above are biased by "catchability" differences among species and do 11 FISHERY BULLETIN: VOL. 75, NO. 1 TABLE 6. — Stratified mean catch per tow (kilograms) for selected species, Albatross IV fall survey data, Middle Atlantic (1967-68 and 1973-74) and southern New England (1963-65 and 1972-74) areas. 1 Mean catch per tow values represent simple averages of values given in Tables 2 and 3 for these areas and years. Middle Atlantic Southern New England Species 1 967-68 mean 1973-74 mean % change 1963-65 mean 1972-74 mean % change Principal groundfish: Cod 0.0 0.0 1.7 0.8 -53 Haddock 0.0 0.0 3.7 2 0.0 -99 Silver hake 0.9 0.2 -78 62 2.8 -55 Red hake 0.5 0.1 -80 6.0 3.4 -43 Flounders: Yellowtail 4.5 2 0.0 -99 10.8 10.2 - 6 Summer flounder 1.8 0.5 -72 0.5 0.9 +80 Winter flounder 1.5 0.1 -93 2.9 1.3 -55 Other 1.2 1.1 -8 3.3 1.8 -45 Other groundfish: Angler 0.7 0.5 -29 5.4 4.5 -17 Ocean pout 2 0.0 2 0.0 - 0.5 0.1 -80 Sculpins 0.1 0.0 -100 1.0 0.8 -20 Scup 1.7 0.5 -71 1.5 1.6 +7 Searobins 71.9 1.9 -97 0.7 1.7 + 143 White hake 0.1 0.0 -100 0.8 0.2 -75 Other 0.3 2 0.0 -99 0.1 2 0.0 -99 Principal pelagics: Herring 0.0 2 0.0 +0 0.2 2 0.0 -99 Mackerel 0.1 0.0 -100 2 0.0 2 0.0 +0 Other pelagics and other fish: Butterfish 10.9 7.4 -32 4.4 4.9 + 11 Spiny dogfish 25.5 2 0.0 -100 119.4 32.5 -73 Skates and rays 6.2 7.9 + 27 12.5 5.1 -59 Other 8.4 6.4 -24 1.3 4.5 +246 Squid: Short-finned squid 0.3 0.1 -67 0.1 0.3 +200 Long-finned squid 9.9 11.1 + 12 1.4 11.8 + 743 Total finfish and squid 146.5 37.8 -74 184.4 89.2 -52 'Middle Atlantic and southern New England areas represented by strata sets 61-76 and 1-12, respectively 2 Less than 0.05. TABLE 7. — Stratified mean catch per tow (kilograms) for selected species, Albatross IV fall survey data, Georges Bank and Gulf of Maine areas, 1 1963-65 and 1972-74. Mean catch per tow values represent simple averages of values given in Tables 4 and 5 for these areas and years. Georges Bank Gult of Maine Species 1963-65 mean 1 972-74 mean % change 1963-65 mean 1 972-74 mean % change Principal groundfish: Cod 8.4 12.8 +52 10.8 6.3 -42 Haddock 60.8 5.0 -92 22.0 3.5 -84 Redfish 2.0 2.8 +40 33.3 22.9 -31 Silver hake 2.9 2.1 -28 13.9 4.7 -66 Red hake 3.8 1.8 -53 2.2 1.0 -55 Pollock 2.0 1.0 -50 6.7 6.8 + 1 Flounders: American plaice 2.9 0.7 -76 5.3 2.4 -55 Yellowtail 7.4 3.4 -54 0.4 0.2 -50 Winter flounder 2.0 1.5 -25 0.4 0.3 -25 Witch 0.7 1.0 +43 2.8 1.7 -39 Other 0.8 2.2 + 175 0.1 2 0.0 -99 Other groundfish: Angler 3.7 1.6 -57 2.4 2.5 +4 Cusk 0.3 0.2 -33 1.6 1.6 Ocean pout 1.2 0.2 -83 2 0.0 0.1 +474 Sculpins 2.8 2.7 -4 0.2 0.2 White hake 0.9 2.6 + 189 6.9 15.6 + 126 Other 0.2 0.3 +50 0.3 0.1 -66 Principal pelagics: Herring 0.7 2 0.0 -99 0.6 2 0.0 -99 Mackerel 2 0.0 0.2 + 300 2 0.0 2 0.0 Other pelagics and other fish: Spiny dogfish 3.1 16.0 +416 26.9 11.1 -59 Skates and rays 22.7 19.9 -12 11.9 6.6 -45 Other 1.2 2.4 + 100 0.8 0.2 -75 Squid: Short-finned squid 0.4 1.8 +350 0.1 0.7 +600 Long-finned squid 0.4 1.1 + 175 2 0.0 2 0.0 Total finfish and squid 131.3 83.3 -37 149.6 88.5 -41 'Georges Bank and Gulf of Maine areas represented by strata sets 13-25 and 26-30 and 36-40, respectively. 2 Less than 0.05. 12 CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS TABLE 8. — Stratified mean catch per tow (kilograms) for selected species of finfish and squid, Albatross IV autumn bottom trawl survey data, 1967-74, Middle Atlantic, southern New England, Georges Bank, and Gulf of Maine (strata 61-76, 1-30, and 36-40). Species 1967 1968 1969 1970 1971 1972 1973 6.4 19 74 Cod 4.5 50 4.4 5.1 46 6.4 2.9 Haddock 8.1 5.8 38 4.0 2.4 2.2 3.3 1.3 Redfish 8.2 136 7.9 11.1 7.9 8.3 5.7 7.7 Silver hake 2.3 2.5 1.8 2.0 2.4 3.6 2.6 1.9 Red hake 1.0 1.6 1.8 1.3 1.7 2.6 1.6 0.7 Pollock 1.2 1.9 4.2 1.2 2.2 2.7 2 1 18 Yellowtail 4.8 5.6 5.2 42 29 79 16 1.0 Other flounder 4.6 5.4 5.1 5.1 3.5 4.3 4.2 3.5 Herring 0.3 0.1 0.1 '0.0 0.3 0.1 '0.0 '0.0 Mackerel 03 02 1.1 '0.0 '0.0 0.1 '0.0 '0.0 Other finfish 2 809 47.1 892 493 33.6 43.3 54.5 27.4 Short-finned squid 0.2 0.3 0.1 0.3 0.4 03 0.3 0.4 Long-finned squid 2.8 5.1 6.8 22 2.1 4.6 76 5.8 Total finfish and squid 119.2 94.2 131.5 85.8 64 86.4 89.9 54.4 'Less than 0.05. 2 Does not Include data for tunas, sharks, swordfish, American eel. or white perch. TABLE 9. — Stratified mean catch per tow (kilograms) for selected species of finfish and squid, Albatross IV autumn bottom trawl survey data, 1963-74, southern New England, Georges Bank, and Gulf of Maine areas (strata 1-30 and 36-40). Species 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 Cod 8.5 7.6 5.6 4.8 5.7 6.4 5.6 6.4 5.7 8.1 80 3.6 Haddock 31.6 31.2 229 10.5 10.2 7.4 4.7 5.1 3.1 2.8 42 1.6 Redfish 10.3 23.1 5.8 12.4 10.4 17.1 10.0 14.0 10.0 10.5 7.2 100 Silver hake 13.8 4.1 6.1 3.3 2.7 2.9 22 2.4 29 4.4 3.2 2.4 Red hake 6.7 2.3 2.7 1.6 1.2 1.8 2.1 1.6 2.0 3.2 2.0 0.8 Pollock 4.1 3.6 1.9 1.8 1.5 2.3 5.3 1.5 2.7 3.4 2.7 2.2 Yellowtail 6.6 6.5 4.5 3.2 5.2 5.6 5.6 5.3 3.6 98 2.1 12 Other flounder 88 6.1 6.7 92 4.6 5.6 5.8 63 4.0 5.1 4.8 4.0 Herring 1.0 0.1 0.5 1.2 0.4 0.1 0.1 0.1 03 0.1 '0.0 0.1 Mackerel '0.0 '0.0 0.1 0.1 0.4 0.2 1.4 0.1 0.1 0.2 0.1 0.1 Other finfish 2 75.6 89.4 61.8 629 49.8 45.8 97.5 55.5 33.7 49.4 599 306 Short-finned squid ( 3 ) '"0.0 "0.1 "0.1 0.2 0.4 0.1 0.3 0.4 0.3 0.4 05 Long-finned squid ( 3 ) "0.5 "0.8 "1.0 0.8 4.0 6.2 1.5 2.0 2.5 6.7 4.5 Total finfish and squid 167.0 174.5 119.5 112.1 93.1 99.6 146.6 100.1 70.5 99.8 101.3 61.6 'Less than 0.05. 2 Does not include data for tunas, sharks, swordfish. American eel, or white perch 3 Data not recorded. "Squid catches for 1964-66 prorated by species according to relative percentages caught in later years. not reflect the relative magnitude of various species within the biomass as a whole. For in- stance, herring and mackerel together appear to have constituted over 50% of the biomass present during this study (Edwards 1968; International Commission for the Northwest Atlantic Fisheries 1974e, footnote 17) yet account for less than 1% of the weight taken in autumn bottom trawl surveys. Furthermore, the aggregated distribution of finfishes and squid in nature, and the behavior of the gear employed, insure that catch data for individual species will seldom be normally dis- tributed but rather will tend to conform to the negative binomial or some other contagious form (Taylor 1953). In the following sections, we utilize selected transformation and weighting procedures in attempts to correct for these factors. ^International Commission for the Northwest Atlantic Fisheries. 1975. Report of the herring working group, April 1975. ICNAF Annu. Meet. 1975, Summ Doc. No. 19, Serial No. 3499 (mimeo.), 31 p. Weighted Analyses Catchability differences among species imply that trends in biomass as defined in this study will be primarily determined by trends for species most vulnerable to the survey gear unless adjustments in terms of catchability are made. Accordingly, we developed catchability coefficients by year for the species and species groups in Tables 8 and 9 for use in computing weighting factors by relating stratified mean catch per tow by stock to available estimates of stock size, all computations being in terms of weight. Annual estimates of stock size (weight at the beginning of year i) were required for this purpose for each individual stock for which TAC's have been established (International Commission for the Northwest Atlantic Fisheries 1975c); thus, separate estimates were required for cod in 5Y 18 and 5Z, haddock in 5Ze, silver hake in 18 Alphanumeric designations refer to divisions and sub- divisions of SA 5 and 6 given in Figure 1. 13 FISHERY BULLETIN: VOL. 75, NO. 1 5Y, 5Ze, and 5Zw-SA 6, red hake in 5Ze and 5Zw-SA 6, yellowtail in 5Ze, 5Zw, and SA 6, and herring in 5Y and 5Z-SA 6. (We considered the remaining species and species groups indicated as stocks for the purpose of this analysis.) Silver hake, herring, and mackerel stock sizes were available from virtual population analyses in previous assessments (International Commission for the Northwest Atlantic Fisheries 1974e, see footnote 17; Anderson 19,20 ), while annual esti- mates for haddock and red hake had also been computed earlier (Hennemuth see footnote 5; Anderson 21 ; Clark 22 ) using average weight or mean weight at age data and the relationship: calculated stock size for each year using Equation (3); 1964-66 stock sizes were then assumed to be similar to the 1967-68 average as commercial abundance indices were stable through this period. We then obtained values for succeeding years by adjusting the 1967-68 average by stock abundance indices based on pre-recruit survey catches (Brown and Hennemuth see footnote 6; Parrack 23 ), i.e., Stock size in year i = Mean stock size for 1967-68 Abundance index for year i Mean abundance index for 1967-68 (4) C t =N,F l /Z l (l - expt-ZJ) (3) where C t = landings (number) in year i, N t = stock size ( number) at the beginning of year i, F- = instantaneous fishing mortality rate in year i, and Z = instantaneous total mortality rate in year i ( =F t + M, the instantaneous natural mortality rate). Approximations of stock size for both long- finned and short-finned squids are also available for recent years ( International Commission for the Northwest Atlantic Fisheries 1975c). We used these approximations for all years in view of uncertainty regarding stock size and historical trends in abundance for these species (Interna- tional Commission for the Northwest Atlantic Fisheries 1975c). Stock size estimates for the remaining species and species groups are currently unavailable, and we computed estimates by a variety of procedures. For yellowtail, we assumed an F value of 1.0 for the southern New England (5Zw) stock in 1967-68 (M = 0.2 in all cases) based on earlier assessment work (Brown and Hennemuth see footnote 6), and 19 Anderson, E. D. 1975. Assessment of the ICNAF Division 5Y silver hake stock. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1975, Res. Doc. No. 62, Serial No. 3544 (mimeo.), 13 p. 20 Anderson, E. D. 1975. Assessment of the ICNAF Subdivision 5Ze and Subdivision 5Zw-Statistical Area 6 silver hake stocks. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1975, Res. Doc. No. 94, Serial No. 3574 (mimeo.), 17 p. "Anderson, E. D. 1974. Assessment of red hake in ICNAF Subarea 5 and Statistical Area 6. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1974, Res. Doc. No. 19, Serial No. 3165 (mimeo.), 27 p. 22 Clark, S. 1975. Current status of the Georges Bank (5Ze) haddock stock. Int. Comm. Northwest Atl Fish. Annu. Meet. 1975, Res. Doc. No. 48, Serial No. 3527 (mimeo.), 9 p. For an estimate of SA 6 stock size, we obtained values for the 1963-66 period by multiplying the computed average stock size value for southern New England by the ratio between mean survey abundance indices between the SA 6 and southern New England stock areas and the ratio between the actual bottom areas considered; we obtained the remaining values using stock abundance indices (Parrack see footnote 23) as above. For the Georges Bank (5Ze) stock, we assumed an F value of 0.8 in 1964 and 1965 (Brown and Hennemuth see footnote 6), calculated stock sizes by Equation (3), and averaged these values to obtain an initial estimate; we then adjusted this value by means of commercial abundance indices (Brown and Hennemuth see footnote 6; Parrack see footnote 23) according to Equation (4) to obtain estimates for later years. The Cape Cod yellowtail stock was considered to have been relatively stable in recent years; we computed an estimate for 1969 by Equation (3) assuming an F value of 0.8 and added the resulting value to each Georges Bank stock size estimate to obtain combined estimates for the Georges Bank area. We obtained stock size estimates for the re- maining stocks from Equation (3) using available estimates of F and M and historical catch data (International Commission for the Northwest Atlantic Fisheries 1965-1973, 1974c, 1975a, see footnote 12). We computed an average stock size for the entire 1965-75 period for 5Y cod using mortality rates reported by Penttila and Gifford 24 , 23 Parrack, M. L. 1974. Status review of ICNAF Subarea 5 and Statistical Area 6 yellowtail flounder stocks. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1974, Res. Doc. No. 99, Serial No. 3335 (mimeo.), 17 p. 24 Penttila, J. A., and V. M. Gifford. 1975. Growth and mortal- ity rates for cod from the Georges Bank and Gulf of Maine areas. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1975, Res. Doc. No. 46, Serial No. 3525 (mimeo.), 13 p. 14 CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS while for 5Z cod we computed an average figure for the 1970-75 period using mortality rates from the above paper and obtained values for the remain- ing years by adjusting this average by commercial abundance indices reported by Brown and Heyerdahl. 25 We followed an analogous procedure in the case of "other finfish" by computing a value for 1967 (chosen to be in the middle of the period) assuming an F value of 0.4 and M = 0.2; we then calculated commercial abundance indices from historical catch data and total effort estimates for SA 5 and 6 (Brown et al. in press) and obtained stock size estimates for the remaining years by adjusting the 1967 value by means of these abundance indices according to Equation (4), as above. For redfish, other flounders, and pollock, we computed average values from Equation (3) using available sustainable yield estimates and as- sumed values of F, as follows (M = 0.2 in all cases): Sustainable yield estimate Species Period (tons x 10' 3 ) F Redfish 1964-75 16 (Mayo 26 ) 0.4 Other flounders 1964-69 25 0.7 Other flounders 1970-75 20 0.9 Pollock 1964-75 27 16 0.4 Turning to survey abundance indices, an in- herent problem in any analysis of trawl data lies in the fact that the computed means and variances are seldom, if ever, independent. The present data are no exception; Grosslein (1971) has found that in the present survey individual stratum var- iances are approximately proportional to the squares of the stratum means, indicating that a logarithmic transformation is appropriate (Steel and Torrie 1960). Under these conditions, use of a logarithmic scale transformation tends to nor- malize the data and render means and variances independent, thereby permitting use of paramet- ric statistical methods (obviously, anomalous fluctuations in observed trends are also reduced 25 Brown, B. E., and E. G. Heyerdahl. 1972. An assessment of the Georges Bank cod stock (Div. 5Z). Int. Comm. Northwest Atl. Fish. Annu. Meet. 1972, Res. Doc. No. 117, Serial No. 2831 (mimeo.), 24 p. 26 Mayo, R. K. 1975. A preliminary assessment of the redfish fishery in ICNAF Subarea 5. Int. Comm. Northwest Atl. Fish. Annu. Meet. 1975, Res. Doc. No. 59, Serial No. 3541 (mimeo.), 31 p. "Pollock in ICNAF Divisions 4VWX, Subarea 5, and Statis- tical Area 6 are currently considered as a unit stock. Accord- ingly, this figure represents the SA 5 and 6 proportion of the estimated sustainable yield for this stock as determined from historical catch data. considerably). Accordingly, we computed strati- fied mean catch per tow values for all stocks using In (kilograms + 1) values for each tow; strata sets used are given by species and stock in Table 10. We then computed estimates of stratified mean catch per tow in original units by retransforming as suggested by Bliss (1967:128) according to the relation: E(y st ) = exp (y st + S 2 /2) (5) where E(y st ) represents the estimated (re- transformed) stratified mean catch per tow andy s , and S 2 represent the stratified mean and the estimated population variance, respectively, in logarithmic units, computed as in Equations (1) and (2) above. We also calculated untransformed (y st ) values for the stocks and strata sets in Table 10 for comparative purposes. After obtaining stock size estimates and abundance indices as described above, we com- puted catchability coefficients for all years by dividing both untransformed and retransformed stratified mean catch per tow for year i by the appropriate stock size value at the beginning of year i + 1 (or by the computed average stock size). Deviations from the arithmetic mean were then plotted by year; where trends were apparent, TABLE 10. — Strata sets used in computing stratified mean catch per tow values by stock. Strata sets Middle Atlantic Southern New Species and stock north' England north 2 Cod 5Y 3 26-30, 36-40 26-30, 36-40 5Z 5-30, 36-40 5-30, 36-40 Haddock 5Ze 13-25 13-25 Redfish 18, 22, 26-30, 36-40 1-30,36-40 Silver hake 5Y 26-30, 36-40 26-30, 36-40 5Ze 13-25 13-25 5Zw-6 61-76, 1-12 1-12 Red hake 5Ze 13-25 13-25 5Zw-6 61-76, 1-12 1-12 Pollock 61-76, 1-30, 36-40 1-30, 36-40 Yellowtail 5Ze 13-25 13-25 5Zw 5-12 5-12 6 69-76, 1-4 1-4 Other flounders 61-76, 1-30, 36-40 1-30. 36-40 Herring 5Y 26-30, 36-40 26-30, 36-40 5Z-6 63-76, 1-25 1-25 Mackerel 61-76, 1-30, 36-40 1-30, 36-40 Other finfish 61-76, 1-30, 36-40 1-30, 36-40 Short-finned squid 61-76, 1-30, 36-40 1-30, 36-40 Long-finned squid 61-76, 1-30, 36-40 1-30, 36-40 'Strata for the Middle Atlantic area 2 Since 1963 (strata 1-40). Alphanumeric designations refer to shown in Figure 1 . (61-76) added in 1967. divisions and subdivisions of SA 5 and 6 15 FISHERY BULLETIN: VOL. 75, NO. 1 linear regressions were fitted to the data to evaluate the degree of relationship. A significant (P<0.01) negative trend was obtained for haddock for both untransformed and retransformed data (Figure 11). This could have resulted from over- estimates of stock size in later years or actual differences in catchability associated with changing availability as stock size decreased. A plot of numbers captured per tow by year during the period of study suggested that actual dif- ferences in catchability may have occurred (Fig- ure 11); accordingly, we divided the period of study into two units (1963-68 and 1969-74) for the purpose of calculating weighting coefficients for the species. The dividing line was taken as the point in which the percentage of tows containing five haddock or less reached 90%. In the case of species for which more than one stock had been defined, some question existed as to whether coefficients should be computed for the entire species or on a stock basis. As no consistent trends had been found for these species over time, one-way analysis of variance was used to test for differences between stocks, using years as repli- cate observations. These tests revealed significant differences (P<0.05) between individual stocks for all species except yellowtail (i.e., cod, silver and red hake, and herring). We therefore retained individual stocks as discrete units in computing biomass declines (i.e., no attempt was made to combine stocks on a species basis). After obtaining the desired sets of catchability coefficients for all stocks, we obtained weighting coefficients by calculating arithmetic means of untransformed and retransformed sets (Tables 11, 12), using the entire set except in the case of haddock as explained above. We then computed biomass estimates by year, viz. TABLE ll. — Weighting coefficients calculated by stock from untransformed and retrans- formed survey data, 1967-74, Middle Atlantic, southern New England, Georges Bank, and Gulf of Maine area (strata 61-76, 1-30, and 36-40). Calculated from Species Untransformed data Retransformed data' and Weighting Coefficient Weighting Coefficient stock 2 coefficient 3 of variation 4 coefficient 3 of variation 4 Cod: 5Y 39.954 0.31 44.545 0.44 5Z 5.160 0.52 3.433 0.50 Haddock 5 : 5Ze 14.146, 10.193 0.25, 0.46 15.591, 7.461 0.71, 0.56 Redfish 40.063 0.29 49.188 0.32 Silver hake: 5Y 8.714 0.80 8.348 0.94 5Ze 0.727 0.30 0.650 0.31 5Zw-6 1.325 0.33 1.101 0.40 Red hake: 5Ze 6.565 0.65 5.384 0.74 5Zw-6 2.341 0.74 1.422 0.71 Pollock 4.069 0.45 1.442 0.37 Yellowtail: 5Ze 17.391 0.24 15.106 0.31 5Zw 45.722 0.79 42.229 0.70 6 67.795 0.95 39.969 076 Other flounders 10.897 0.18 11.134 0.17 Herring: 5Y 0.125 >1.0 0.039 0.97 5Z-6 0.010 >1.0 0.002 0.75 Mackerel 0.015 >1.0 0.005 0.57 Other finfish 12.809 0.31 14.553 0.14 Short-finned squid 0.302 0.37 0.206 0.34 Long-finned squid 5.240 046 4.302 0.65 'Estimated stratified mean catch per tow values computed from transformed data according to the relation, E(yst) = oxp(y s t +S 2 /2), where y st and S 2 represent the mean and estimated population variance, respectively, on the transformed scale. 2 Weighting coefficients calculated by individual stock for cod, haddock, silver hake, red hake, yellowtail, and herring: stock areas are given in Figure 1 . Stock areas for the remaining species are equivalent to all strata in SA 5 and 6 covered during 1967-74. /TiM+i] 3 Weighting coefficients calculated as n where C, = stratified mean catch per tow (tons) in year/ and S / + 1 = stock size at the beginning of the following year. All values x 10 8 . Coefficient of variation calculated over all years. 5 Weighting coefficients computed separately for 1967-68 and 1969-74 data due to apparent changes in catchability. 16 CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS 46 613 X \ LINEAR . RETRANSFORMED 1963 64 73 1974 V^ 1963 n - 57 _lm_ „b«hJ 1964 n =63 1965 n 66 1966 n - 67 1967 n • 67 )- 1968 )-■ n>69 3 ^2 g g ? Si e 88 ^ii i ii i 7 i 8 ^ ~- (\j O V J1 10 ^ - 1969 II. n = T3 1970 ll. n =70 1971 Il— n .73 1972 - n =73 - 1973 n =73 - 1974 1.. n .74 o m o o o g 1 — (SJ (O ^ i. i i I I O 1.00 121.231 >1.00 Other flounders 13.016 0.22 14.293 0.25 Herring: 5Y 0.178 >1.00 0095 >1.00 5Z-6 0.027 >1.00 005 0.94 Mackerel 015 >1.00 006 056 Other finfish 12.569 0.31 13.648 0.18 Short-finned squid 254 0.70 0.177 0.63 Long-finned squid 3.124 0.80 2.099 >1.00 1 Estimated mean catch per tow values computed from transformed data according to the relation, E(y st ) = ex P(ysf + S 2 /2), where y st and S 2 represent the mean and estimated population variance, respectively, on the transformed scale. 2 Weightmg coefficients calculated by individual stock for cod, haddock, silver hake, red hake, yellowtail, and herring, stock areas are given in Figure 1 Stock areas for the remaining species are equivalent to all strata in SA 5 and 6 covered during 1967-74 /=1 [ C A + l] 3 Weightmg coefficients calculated as ^ where C, = stratified mean catch per tow (tons) in year/ and S, + 1 = stock size at the beginning of the following year All values « 10 8 . Coefficient of variation calculated over all years 5 Weighting coefficients computed separately for 1967-68 and 1969-74 data due to apparent changes in catchability k 1 Cv/Wj .7 = 1 L - for all i (6) where C y refers to stratified mean catch per tow for the 7th stock in the itb. year and Wj refers to the weighting coefficient for thejth stock (Tables 13, 14), summation being over k stocks. For the purposes of this paper, we consider each computed estimate as representing stock size at the begin- ning of the year following collection of the survey data (i + 1), as catchability coefficients were calculated by relating catch per tow values in autumn of year i to stock size at the beginning of year i + 1 (above). Note that with the exception of 1970 figures for "all data" (Tables 13, 14), values computed from retransformed data agree reasonably well with those computed from un- transformed values; consequently the general appropriateness of assuming a lognormal dis- tribution for these data is confirmed. The average stock size estimate for 1964-66 obtained for all species of 5.0 x 10 6 tons (Table 14) is almost identical to that obtained by Edwards (1968) for the same area and period (5.1 x 10 6 TABLE 13.— Stock size estimates (tons x 10 3 ) for ICNAF Sub- area 5 and Statistical Area 6, 1967-74, Middle Atlantic, southern New England, Georges Bank, and Gulf of Maine, inclusive (strata 61-76, 1-30, and 36-40). Calculated with Untransformed data Retransformed data All Data for principal All Data for principal Year data pelagics excluded data pelagics excluded 1968 7,481 1,783 8,012 1,806 1969 3,826 1,795 5,209 1,880 1970 9,555 1,859 5,158 1,750 1971 2,097 1,567 2,964 1,736 1972 3,156 1,331 3,062 1,418 1973 3,136 1.870 3,661 1,825 1974 2,098 1,841 2,541 1,760 1975 1,828 1.107 1,934 1.119 18 CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS TABLE 14.— Stock size estimates (tons x 10 3 ) for ICNAF Sub- area 5 and Statistical Area 6, 1963-74, southern New England, Georges Bank and Gulf of Maine, inclusive (strata 1-30 and 36-40). Calculated with Untransformed data Retransformed data All Data for principal All Data for principal Year data pelagics excluded data pelagics excluded 1964 6.616 3,317 7,357 3,640 1965 2.780 2,373 2.677 2.151 1966 5,079 2,088 5,382 2.184 1967 8.331 1.610 7.770 1,605 1968 6,056 1,478 6.431 1,493 1969 3.400 1.787 4,238 1.763 1970 1 1 .490 2.012 5,158 1.867 1971 2,174 1.642 2.828 1.759 1972 2.644 1,411 2.751 1,501 1973 3.231 1.964 3.622 1.937 1974 2.371 2.009 2,717 1.931 1975 2,036 1.217 1.981 1.165 tons). Edwards obtained biomass estimates by adjusting minimum biomass figures for each species by a factor accounting for differences in availability and vulnerability, and although estimates obtained for individual species by these methods differed in certain cases it can be seen that, on the average, results are quite comparable. The data of Tables 13 and 14 again reveal pronounced declines. In Table 13 (1968-75, all strata) comparisons of averages for "all data" between 1968-69 and 1974-75 reveal a 65% decline for untransformed data and a 66% decline in the case of retransformed values; with principal pelagics excluded, the corresponding figures are 18 and 22%, respectively. In Table 14 (1964-75, Middle Atlantic strata excluded) comparisons between averages for "all data" for 1964-66 and 1973-75 reveal declines of 47% and 46% for un- transformed and retransformed values, respec- tively, while with principal pelagics excluded the corresponding figures were 33% and 37%. The greater decrease for the 1968-75 period for "all data" might appear somewhat anomalous but actually results primarily from appearance of the outstanding 1967 mackerel year class. As the estimates in Tables 13 and 14 purport to measure declines in biomass in SA 5 and 6, it might logically be argued that they could be combined in some way (use of the 1968-75 data would be preferable in that survey coverage extended further to the south). Paired £-tests indicated no differences between corresponding stock size estimates in Tables 13 and 14 for the 1968-75 period. Therefore, we combined the 1968-75 estimates in Table 13 with the 1964-67 estimates in Table 14 (Figures 12, 13) and computed percentage changes between the means of the 1964-66 and 1973-75 periods, as before. For "all data," we obtained declines of 51% and 47% with untransformed and retransformed values; with herring and mackerel excluded, the cor- responding figures were 38% and 41%. Analysis of both untransformed and re- transformed data yield essentially similar results. The data of Figures 12 and 13 illustrate the ef- fectiveness of the transformation in reducing anomalies caused by variability in the data. For untransformed estimates (Figure 12) it will be £ 4.000 FIGURE 12 — Estimates of fishable biomass by year for ICNAF Subarea 5 and Statistical Area 6, 1964-75, calculated with un- transformed survey data. Curves were plotted by combining 1968-75 estimates from Table 13 with 1964-67 estimates from Table 14. 1 1 _ 9.000 ^^— Alldolo Dolo »0f pr-incipoi pdog-cs eicluded 7,000 - - 6.000 - 5 DOC - 4,000 - 3.000 \ * - 2,000 V *"--. ____ ... ^,- .. - 1,000 i , 1 1 I 1 i _J FIGURE 13.— Estimates of fishable biomass by year for ICNAF Subarea 5 and Statistical Area 6, 1964-75, calculated with retransformed survey data. Curves were plotted by combining 1968-75 estimates from Table 13 with 1964-67 estimates from Table 14. 19 FISHERY BULLETIN: VOL. 75, NO. 1 noted that an anomalous peak occurs in 1970, which examination of biomass estimates on a per- species basis revealed to have been caused by anomalously high mackerel catches in certain tows during the 1969 survey. The influence of this factor appears to have been compensated for by use of the logarithmic transformation (Figure 13). On the other hand, the anomalously low data point for 1965 (Figures 12, 13) appears to have been caused by anomalously low catches of herring in that year, a circumstance in which the trans- formation was ineffective. It does appear, how- ever, that by and large the transformation was of definite value in following trends through time, although estimates for most of the years consid- ered proved to be similar. The above analyses clearly indicate that biomass levels have decreased significantly in SA 5 and 6 in recent years; the trend observed cor- relates well with increases in fishing effort ob- served by Brown et al. (in press). In addition, we have also found evidence indicating that major changes in species composition have occurred as well. The apparent increase in white hake abundance in the Gulf of Maine in recent years (Table 5) could have resulted from population increases in response to reductions in other groundfish species. Similarly, increased mackerel abundance coincident with declining abundance of herring (Tables 3, 4) may indicate some form of species interaction coincident with exploitation, while apparent increases in abundance of squid (Tables 2-7, Figure 9) may have occurred in re- sponse to declining abundance of finfish species. The relationships involved are unclear at present and further study is obviously necessary. Comparisons of annual landings data since 197 1 (over 1.0 x 10 6 tons) with biomass estimates in Tables 13 and 14 indicate that the fraction of the biomass harvested annually has increased sig- nificantly in recent years (i.e., from less than one- fifth of the total in the early and mid-1960's to between one-third and one-half of the total at present). Furthermore, landings since 1971 have exceeded the composite MSY figure of 950 x 10 3 tons calculated by Brown et al. (in press) based on the Schaeffer yield model. This information, together with declines in stock size approximating 50% as indicated in this paper, imply that a significant degree of overfishing has occurred and that stock size has been reduced below the level corresponding to MSY. Back-calculations for all species in Tables 13 and 14 provide an average stock size estimate of approximately 7.0 x 10 6 tons prior to 1964, from which (allowing for the U.S. coastal fishery in previous years) it may be in- ferred that the actual virgin biomass for this fishery probably approximated 8.0-9.0 x 10 6 tons. Since the Schaeffer yield model postulates that MSY will be taken at a stock level corresponding to one-half the maximum (Schaeffer 1954), we may in turn assume that a stock level of ap- proximately 4.0-4.5 x 10 6 tons should be main- tained for SA 5 and 6 if MSY from this resource is to be achieved. In contrast, estimates for fishable biomass in the present paper approximate 2.0 x 10 6 tons at the start of 1975, implying that a lengthy period of reduced exploitation is necessary if stocks are to be rebuilt to the MSY level. In April 1975, the Assessments Subcommittee (STACRES) reviewed evidence relating to de- clines in biomass in SA 5 and 6 in recent years and concluded that substantial reductions in catch would be necessary if stocks are to recover (In- ternational Commission for the Northwest At- lantic Fisheries 1975c). Accordingly, a TAC of 650 x 10 3 tons was recommended to ICNAF and approved at the Seventh Special Commission Meeting (International Commission for the Northwest Atlantic Fisheries 1975b) in Sep- tember. Even with a reduction of this magnitude, STACRES estimated that a minimum of 7 yr would be required for this resource to recover to the MSY point. ACKNOWLEDGMENTS We thank Judith Brennan for her helpful comments and suggestions on data analysis, Kathryn Paine for her assistance with computer programming, and Elizabeth Bevacqua and Maureen Romaszko for numerous tabulations of the data. Richard C. Hennemuth reviewed the manuscript and made suggestions for im- provement. The work of the numerous biologists and technicians who have participated in Alba- tross IV autumn bottom trawl surveys and the processing of the sample data since the beginning of the program is also sincerely appreciated. LITERATURE CITED ANTHONY, V. C, AND H. C. BOYAR. 1968. Comparison of meristic characters of adult Atlantic herring from the Gulf of Maine and adjacent waters. Int. Comm. Northwest Atl. Fish. Res. Bull. 5:91-98. 20 CLARK and BROWN: CHANGES IN BIOMASS OF FINFISHES AND SQUIDS Bliss, C. I. 1967. Statistics in biology; statistical methods for research in the natural sciences, Vol. I, 558 p. McGraw-Hill, N.Y. BROWN, B. E., J. A. BRENNAN, E. G. HEYERDAHL, M. D. GROSSLEIN, AND R. C. HENNEMUTH. In press. The effect of fishing on the marine finfish biomass in the Northwest Atlantic from the eastern edge of the Gulf of Maine to Cape Hatteras. Int. Comm. Northwest Atl. Fish. Res. Bull. 12. Cochran, w. G. 1953. Sampling techniques. John Wiley & Sons, Inc., N.Y., 330 p. EDWARDS, R. L. 1968. Fishery resources of the North Atlantic area. In D. Gilbert (editor), The future of the fishing industry of the United States, p. 52-60. Univ. Wash. Publ. Fish., New Ser., 4. GROSSLEIN, M. D. 1962. Haddock stocks in the ICNAF convention area. Int. Comm. Northwest Atl. Fish. Redbook 1962, Part III, p. 124-131. 1969. Groundfish survey program of BCF Woods Hole. Commer. Fish. Rev. 31(8-9):22-35. 1971. Some observations on accuracy of abundance indices derived from research vessel surveys. Int. Comm. Northwest Atl. Fish. Redbook 1971, Part III, p. 249- 266. INTERNATIONAL COMMISSION FOR THE NORTHWEST ATLAN- TIC FISHERIES. 1953-1973. Statistical Bulletin 1-21. 1974a. Proceedings, Third Special Commission Meeting, October 1973. ICNAF Proceedings 1974, p. 4-34. 1974b. Proceedings, 24th Annual Meeting, June 1974. ICNAF Proceedings 1974, p. 107-256. 1974c. Statistical Bulletin 22, 239 p. 1974d. Report of the Standing Committee on Research and Statistics, October 1973. ICNAF Redbook 1974, p. 5-8. 1974e. Report of the Standing Committee on Research and Statistics, May-June 1974. ICNAF Redbook 1974, p. 63- 142. 1975a. Statistical Bulletin 23, 277 p. 1975b. Proceedings, Seventh Special Commission Meeting, September 1975. 1975c. Report of the Standing Committee on Research and Statistics (STACRES), Annual Meeting-May-June 1975. ICNAF Redbook 1975, p. 11-111. ODUM, E. P., AND A. E. SMALLEY. 1959. Comparison of population energy flow of a herbivorous and a deposit-feeding invertebrate in a salt marsh ecosystem. Proc. Natl. Acad. Sci. 45:617-622. SCHAEFFER, M. B. 1954. Some aspects of the dynamics of populations impor- tant to the management of the commercial marine fisheries. Bull. Inter- Am. Trop. Tuna Comm. 1:27-56. Steel, R. G. D., and J. H. Torrie. 1960. Principles and procedures of statistics with special reference to the biological sciences. McGraw-Hill, N.Y., 481 p. Taylor, C. C. 1953. Nature of variability in trawl catches. U.S. Fish Wildl. Serv., Fish. Bull. 54:145-166. WISE, J. P. 1962. Cod groups in the New England area. U.S. Fish Wildl. Serv., Fish. Bull. 63:189-203. 21 / LARVAL TRANSPORT AND YEAR-CLASS STRENGTH OF ATLANTIC MENHADEN, BREVOORTIA TYRANNUS 1 Walter R. Nelson, 2 Merton C. Ingham, 3 and William E. Schaaf 2 ABSTRACT A Ricker spawner-recruit model was developed for Atlantic menhaden, Brevoortia tyrannus, from data on the 1955-70 year classes. The number of eggs produced by the spawning stock was calculated as the independent variable to account for changes in fecundity due to changes in population size and age structure. A survival index was developed from deviations around the Ricker curve and was regressed on several environmental parameters to determine their density-independent effects. The recruit- environment model accounted for over 84% of the variation in the survival index. Zonal Ekman transport, which acts as a mechanism to transport larval menhaden from offshore spawning areas to inshore nursery grounds, was the most significant parameter tested. Ricker functions for good and poor environmental years were developed, indicating the wide range of recruitment that can be expected at different stock sizes. Comparisons of spawner-recruit relations for Pacific sardine and Atlantic menhaden indicated striking similarities. Surplus yield for the Atlantic menhaden fishery was cal- culated from observed and predicted survival, and compared with the actual performance of the fishery. One of the more intriguing and important prob- lems in fishery science, that of the relative influence of spawning stock size and environ- mental variation on year-class strength, has resulted in a long-standing controversy among fishery biologists. The two principal reasons for investigating the effects of stock size and en- vironmental change on year-class strength are, of course, to understand what has happened and to predict what will happen. Since environmental conditions will produce varying recruitment at a given stock size, one must determine both the reproductive potential under average en- vironmental conditions, i.e., the density- dependent spawner-recruit curve, and the effect of varying environmental conditions, or the density-independent function. The difficulty comes, as Clark and Marr (1955) point out, in separating the relative influences of the two functions. A prerequisite for such an attempt is a reliable long-term series of data, adequate to estimate the size of the spawning stocks, the number of recruits, the age structure of the populations, the patterns of environmental var- iation, and the rate at which the resource is being harvested. "MARMAP Contribution No. 88. 2 Atlantic Estuarine Fisheries Center, National Marine Fisheries Center, NOAA, Beaufort, NC 28516. 3 Atlantic Environmental Group, National Marine Fisheries Service, NOAA, Narragansett, RI 02882. Biologists are in general agreement that the most critical survival period for many marine fishes is during the time of egg and larval drift. Major factors affecting survival during this period are food (Cushing 1969), cannibalism by filter- feeding parents (Radovich 1962; Murphy 1967), and ocean currents (Sette 1943). The first two of these factors are density dependent and tend to control population growth. Transport by ocean currents to or from areas favorable to survival is density independent and has been used to explain successful year classes of Atlantic mackerel by Sette (1943) and Atlantic haddock by Walford (1938). A relationship between winds and year- class success for the East Anglian herring fishery was reported by Carruthers (1938). Cushing (1969) pointed out that ". . . correlations between recruitment and winds were often successful for a period of years, after which they failed catas- trophically." Other density-independent factors, such as temperature, particularly in the sense of long- term climatic change, have been related to changes in spawning success and location. For example, a change in the environment of the Pa- cific sardine over a period of time which resulted in a change in normal distribution patterns and a series of poor year classes was postulated by Radovich (1962). Sissenwine (1974) documented a significant relationship between atmospheric temperature and the recruitment and equilibrium catch of yellowtail flounder, but did not explain Manuscript accepted June 1976. FISHERY BULLETIN: VOL. 75, NO. 1, 1977. 23 FISHERY BULLETIN: VOL. 75, NO. 1 the mechanism by which temperature anomalies influence the fishery. Cushing (1969) listed three sources of variation which might affect recruitment: year-to-year environmental changes, larger scale climatic changes, and differences due to stock density. The year-to-year effects were considered by Cushing to be randomly distributed around the stock and recruitment curve and not of major consequence in the long-term regulation of fisheries. Over a number of years, variations around a stock and recruitment curve may tend to cancel one another and the fishery may provide a relatively stable yield. However, when a fishery is overexploited and subjected to poor survival as a result of en- vironmental conditions, stock size may be reduced to a small fraction of that necessary to maintain a maximum sustainable yield (MSY). Further, with overcapitalization, fishing effort may remain high, preventing a resurgence of the stocks by maintaining a spawning stock too small to pro- duce a large year class under favorable en- vironmental conditions. From this standpoint, a predictive capability, based on knowledge of density-dependent and density-independent recruitment could be vital to the maintenance of adequate stock size through a reduction in effort, or to the harvesting of surplus population beyond that necessary to maintain the MSY. Fisheries, in the generic sense, operate over long periods of time. Fishermen, fish processors, and consumers operate on a much shorter time scale and large, unexpected, year-to-year fluctuations in stock size have significant economic and social impact. The Atlantic menhaden, Brevoortia tyrannus, is a species that has supported a significant fishery since the middle of the 19th century (Reintjes 1969). Landings from the fishery have been sampled extensively since 1955 and the major characteristics of the stocks and the fishery have been determined. Information for a variety of stock sizes and from a range of environmentally different years is available, and the stocks have been subjected to heavy fishing pressure (Schaaf and Huntsman 1972). A study of forecasting methods and the de- velopment of a forecast for the Atlantic menhaden fishery was carried out by the National Marine Fisheries Service (Schaaf et al. 4 ). The manuscript points out that knowledge of the biology of re- cruitment of the Atlantic menhaden is needed to take advantage of strong year classes through the development of short-term fishing strategies. Knowledge of poor year classes would also be beneficial from a standpoint of avoiding excessive fishing pressure on the stocks. A single year class is harvested by industry over a 4- to 5-yr period, and its failure could be masked to some extent by overfishing of other year classes taken concurrently, resulting in serious stock depletion. Conversely, a large year class may lead to a large increase in fishing effort which con- tinues after the year class has been harvested, leading to overcapitalization and overfishing in subsequent years of reduced stock size. A large year class, followed by several poor year classes is potentially disastrous to the fishing industry and to the stocks. Knowledge of the recruitment pro- cess and the ability to predict year-class strength is necessary if the fishery is to operate at the MSY level. Detailed information on the composition of Atlantic menhaden stocks obtained yearly since 1955 shows a range in numbers recruited into the fishery of from 11.5 billion in 1958 to 0.9 billion in 1967. Although some of the variation in re- cruitment can be attributed to fluctuations in the size of the spawning stock (Schaaf and Huntsman 1972), the wide range of fluctuations between years with similar spawning stock sizes suggests that environmental factors are influencing the survival of prerecruits. This study attempts to identify those factors, determine their relative influences, and develop a predictive model to account for the variations between actual and expected recruitment into the Atlantic menhaden fishery. SPAWNING AND LARVAL DISTRIBUTION Gravid or running-ripe Atlantic menhaden are rarely caught and spawning has not been ob- served. Without conclusive information, the time and place of spawning has been inferred by the relative ripeness of maturing ova, the occurrence of partially spent ovaries, and the distribution and occurrence of eggs and small larvae. Higham and Nicholson (1964:262) reported that "Schaaf, W. E., J. E. Sykes, and R. B. Chapoton. 1973. Forecast of 1973 Atlantic and Gulf menhaden catches based on the histor- ical relation of catch and fishing effort. Unpubl. manuscr., 22 p. Atlantic Estuarine Fisheries Center, National Marine Fisheries Service, NO A A, Beaufort, NC 28516. 24 NELSON ETAL.: LARVALTRANSPORTOFB/?£VOO/?77A TYRANNHS ". . . (only 11 specimens containing numerous ripe ova were encountered in the routine field examination of several hundred thousand fish during 4 years of sampling), . . . ." Based on a sample of approximately 37,000 female menhaden from all Atlantic coast fishing areas, they con- cluded, p. 270, "Spawning apparently occurred in the North Atlantic Area [north of Long Island] from May to September; in the Middle Atlantic [south to Cape Hatteras], from March through May and again in September and October; and in the South Atlantic [south of Cape Hatteras] , from October through March." Based on the percent- ages of sexually active (ripening but not ripe) females in their samples, it appears that a major- ity of spawning activities take place in the South Atlantic Bight. The spawning cycle appears to be one of limited spawning during a spring north- ward migration, limited early and late summer spawning as far north as Cape Cod and occasion- ally into the Gulf of Maine, increased spawning activity during a southward fall migration, and intensive (90-100% sexually active) winter spawning in the South Atlantic Bight. Spawning activities through the winter are difficult to determine because the stocks move offshore and there is no fishery for menhaden during that period. This is the only time during the year that menhaden schools are not available in coastal waters, and that fact leads to specula- tion about an offshore spawning migration. Available information about the distribution of menhaden eggs and larvae has been reviewed by Kendall and Reintjes (1975) and Chapoton. 5 In- ferences regarding spawning activities have been drawn from various surveys of restricted time and coverage which have been conducted on the east coast since 1937 (Permutter 1939), primarily in sounds, bays, and creeks. Only two egg and larval research efforts have provided large-scale sys- tematic coverage of major menhaden spawning areas on the Atlantic coast. Those are the cruises of the MV Theodore N. Gill (Reintjes 1961) and the RV Dolphin (Kendall and Reintjes 1975). The distribution of larvae collected by the Dolphin cruises is in general agreement with the spawning cycle documented by Higham and Nicholson (1964). RV Dolphin cruises covered the entire continental shelf from Cape Lookout, N.C., to 5 Chapoton, R. B. 1972. On the distribution of Atlantic menha- den eggs, larvae, and adults. Unpubl. manuscr., 69 p. Atlantic Estuarine Fisheries Center, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. Martha's Vineyard, Mass., in 14 transects from December 1965 to May 1966. The southern part of the menhaden spawning range was covered by cruises of the Theodore N. Gill in 1953 and 1954 (Reintjes 1961). The absence of menhaden larvae during all but the winter cruises led Reintjes to conclude that menhaden spawn along the south Atlantic coast generally from December to February. The southern limit of the spawning range of the Atlantic menhaden is undetermined because a southerly species, the yellowfin menhaden, Brevoortia smithi, has an overlapping spawning range. Those larvae col- lected by the Theodore N. Gill off southern Florida were probably B. smithi and those collected off Cape Lookout, the other area of larval concentra- tion located by the Theodore N. Gill, were un- doubtedly B. tyrannus. Based on the distribution of juveniles and adults, it seems safe to assume that Atlantic menhaden spawn as far south as northern Florida, but at a low intensity in the ex- treme southern part of their range. Reintjes (1969) hypothesized that much of the spawning takes place south of Cape Hatteras. Atlantic menhaden appear to spawn over most of the continental shelf. The general timing se- quence and location of spawning during migra- tions indicates that eggs and larvae are subjected to an open ocean environment for a sufficient length of time to be affected by oceanic conditions. Both the Dolphin and Theodore N. Gill cruises resulted in catches of small larvae from nearshore to the edge of the shelf. Dolphin records show a general increase in average size of larvae from offshore to inshore stations as well as increased distance offshore from north to south. Major sum- mer spawning in the New York-New England area appears to occur well inshore, and large numbers of eggs and larvae have been taken in bays and sounds from Long Island north. Matth- iessen (1974) reported concentrations of eggs that exceeded 20,000/100 m 3 in June 1972 in Nar- ragansett Bay, R.I., and computed the total pro- duction of eggs in the Bay during the summer of 1973 as being in excess of 4.64xlO n . Concentrations of eggs and small larvae are found progressively nearer the offshore edge of the shelf during the fall and winter southward migra- tion. Massmann et al. (1962) found larvae as small as 7 mm 79 km off Chesapeake Bay, and concluded that spawning and hatching occurred more than that distance offshore. Reintjes (1968) reported an extensive patch of menhaden eggs in Onslow Bay, 25 FISHERY BULLETIN: VOL. 75, NO. 1 N.C., in December 1966, 40 km from shore and estimated their age at 8 to 55 h. Theodore N. Gill cruises resulted in the location of larval menhaden up to 220 km off Cape Fear, N.C., in February 1954, although most larvae taken during the Gill cruises were over the shelf. Cruises of the RV Undaunted during the winter of 1970-71 also yielded larvae 170-175 km off Cape Fear. PHYSICAL OCEANOGRAPHY OF THE SPAWNING REGION An excellent summary of the oceanography of the coastal waters of the U.S. east coast was re- cently prepared by Bumpus ( 1973) and the reader is referred to that for detailed information. Bum- pus identified three distinct subdivisions as the Gulf of Maine, Middle Atlantic Bight (Cape Cod to Cape Hatteras), and South Atlantic Bight (Cape Hatteras to Cape Canaveral). Although menha- den are periodically taken north of Cape Cod, Mass., migratory intrusions do not occur there routinely and the area is not one of significant menhaden spawning activity. A brief summary of oceanographic conditions in the other two regimes of significant menhaden spawning activities follows. In the Middle Atlantic Bight the Gulf Stream diverges abruptly toward the northeast, passing Cape Hatteras, and the space between the Shelf Water masses and the Gulf Stream left by this divergence is occupied by the Slope Water mass. Flow in the Shelf Water and Slope Water is generally slow and southward, more or less parallel to the isobaths except for portions of the Slope Water mass near the Gulf Stream which have a northward to northeastward motion im- parted by transfer of momentum from the Gulf Stream. At Cape Hatteras the southward flowing waters generally turn to flow northward and an unknown fraction of these waters becomes en- trained within the Gulf Stream. The southward drift of Shelf Water is partly driven by the pres- sure field developed around river effluent plumes, and in times of low runoff and southeasterly winds the flow may be reversed. Menhaden spawning takes place throughout the Middle Atlantic Bight and oceanographic conditions there should have a major influence on the distribution and survival of eggs and larvae. In the South Atlantic Bight the Gulf Stream current forms the seaward boundary of the region of intensive Atlantic menhaden spawning. The current's mean position is parallel to and a short distance (37-74 km in Carolina coastal waters) from the edge of the continental shelf (180-m isobath). A mass of Shelf Wa^er which has lower salinity and lower temperature, except in sum- mer, than the Gulf Stream water is found shoreward of the Gulf Stream. Motion of the Shelf Water mass is generally slow and variable, re- sponding to local winds, but not customarily flowing southward, unlike the pattern of flow of the Shelf Water in the Middle Atlantic Bight. Occasionally southward flows have been identified near the coast, and the cuspate formations of Raleigh Bay, Onslow Bay, and Long Bay suggest southward flow nearshore as part of a large counterclockwise eddy in each bay. The existence of these eddies, although suspected, never has been conclusively demonstrated. Stefansson et al. (1971) found, based on geopotential topography from six cruises in 1966-67, that there was always an indication of a counterclockwise eddy in Onslow Bay. The pattern found in Raleigh Bay was less permanent and influenced by the influx of Virginian Coastal Water from the north. LARVAL TRANSPORT Menhaden larvae, spawned offshore, move into estuaries before metamorphosing to juveniles, after traversing long, open ocean distances. The larvae are 18-22 mm in length when they enter estuaries after an oceanic phase of IV2 to 2 mo. Very few small larvae (<12 mm) have been taken in estuaries along the central and southern U.S. Atlantic coast, even though eggs and young larvae have occasionally been taken near shore. The timimg of larval entrance is apparently controlled to some extent by the larvae and is somewhat independent of water movement. During earlier larval stages, however, there is a passive drift period in which larval movement is the result of ocean currents. Based on the rate of fin de- velopment, the completely passive phase probably ends when a length of 10-12 mm is reached. Depending on water temperature, menhaden reach that length in 30-45 days (William F. Het- tler pers. commun., Atlantic Estuarine Fisheries Center). Currents with an onshore component, par- ticularly during the passive larval phase, would seem to be important for transportation of the larvae from offshore spawning areas to estuarine nursery grounds. There are no documented 26 NELSON KT AL.: LARVAL TRANSPORT OFBREVOORT1A TYRANNUS physiological requirements for estuarine de- pendence, but metamorphosing larvae are rarely taken in the ocean, indicating that apparent requirements (food, shelter, etc.) provided by estuaries are essential in the life cycle of menhaden. Transport to the vicinity of estuaries should increase the opportunity for entering nursery grounds, resulting in good year classes from years of strong onshore transport. Weak onshore transport or water movement offshore would increase the distance that must be actively traversed, reduce chances of survival, and result in a poor year class. If variation in survival is due to variation in the efficiency of transport of larval menhaden from offshore areas to estuaries, then knowledge of the transport mechanisms would be useful for understanding and predicting variation in year-class strength. Menhaden larvae have been found to be more abundant in the upper 15 m of the water column than in the underlying 18-33 m in extensive surveys of our Atlantic shelf waters (Kendall and Reintjes 1975; Chapoton see footnote 5). It is assumed, therefore, that they remain in the upper mixed layer and are transported along with it. Horizontal transport in the surface layer is principally the result of extensive quasi-steady- state currents and local, variable currents, which are strongly influenced by wind and run-off. Steady state currents, by definition, cannot be responsible for year-to-year variation in larval transport and recruitment, so attention was first turned to the local, variable currents which are superimposed on the quasi-steady-state circula- tion of the surface layer. In the search for a westward transport mechanism which varies seasonally and from year-to-year, wind drift data computed from mean monthly atmospheric pressure distributions for the period 1946 to the present were considered first. In particular, plots of zonal (eastward or westward) wind-driven (Ekman) transport produced by the Pacific Environmental Group, NMFS, NOAA were studied (for method see Bakun 1973). A grid point (lat. 35°N, long. 75°W) located about 56 km southeast of Cape Hatteras was selected as being representative of the wind field in the area of interest. The seasonal variation of Ekman transport at lat. 35°N, long. 75°W generally includes relatively strong WSW-SW- SSW transport during the first quarter of each year. Because of the SW-NE trend of the coastline south of Cape Hatteras, Ekman transports sig- nificantly west of southwestward (those with a stronger westward component) would be most effective in transporting eggs and larvae toward estuarine nursery areas. Plots of the monthly zonal transport at this point revealed conditions of eastward or weak westward transport during most of the year, shifting to moderate or strong west- ward transport during January-March; a periodicity which matched that of spawning of menhaden south of Cape Hatteras (Figure 1). In coastal waters of the Middle Atlantic Bight between Virginia and Long Island, N.Y., com- putations of monthly zonal Ekman transport exhibited a pattern similar to that found south of Cape Hatteras. Monthly zonal Ekman transport values computed for this area show that stronger westward transport generally occurs in the November-February period of menhaden spawn- ing activities, possibly providing a mechanism for transporting menhaden larvae into the vicinity of estuarine environments. A model of the circulation of the shelf waters off the Chesapeake Bight was developed and cited for its application to menhaden year-class strength by Harrison et al. (1967). The model was used in an attempt to explain the difference in "production of young menhaden" in Chesapeake Bay from the 1958 year class, an unusually productive one, and the 1964 year class, which was well below average. The model yielded inappropriate surface current regimes to explain strong shoreward larval transport in 1957-58, and Harrison et al. chose near-bottom currents, which appeared more favorable, as an explanation. As cited earlier, data collected in comparative net tows indicate that menhaden larvae are more abundant in the upper layer than the near-bottom layer, a condition which weakens the premise on which the argu- ment is based. Application of the Ekman drift data to the problem of explaining the large difference in menhaden production in Chesapeake Bay in 1958 and 1964 leads to a more satisfactory biological conclusion than the bottom-layer-transport model used by Harrison et al. (1967). The average monthly westward Ekman transports for the November-March period at two points in the Middle Atlantic Bight for 1957-58 (Table 1) were about twice as large as those for 1963-64, qual- itatively implying that variation in wind-driven surface layer transport of larvae may be at least partly responsible for the amount of variation in menhaden year-class strength. 27 FISHERY BULLETIN: VOL. 75. NO. 1 5 U LU O X «/» z o ►- z < EASTWARD TRANSPORT EASTWARD TRANSPORT WESTWARD TRANSPORT 1969 , 1970 1971 1972 1973 . I ' " "' .... .1. ... . I., i , 1 1 .1, i J M I RN.IMMJSN J M M .1 " JMMJSNJMMJSNJMMJSNJMMJSNJMMJSNJMMJSNJMMJSNJMMJSNJMMJSNJMMJSN FIGURE 1. — Monthly average zonal Ekman transport at lat. 35°N. long. 75°W, 1955-73. January-March spawning period is shaded. TABLE 1. — Average westward wind-driven Ekman transport computed for November-March 1957-58 and 1963-64 at lat. 39°N, long. 72°W and lat. 39°N, long. 75°W. Transports expressed in metric tons per second per kilometer of front. Year Lat. 39°N, long. 72W Lat. 39°N, long. 75°W 1957-58 1963-64 480 250 520 260 SPAWNER-RECRUIT RELATION Over the 16 yr from 1955 to 1970, there was a sharp decline in the size of the Atlantic menhaden spawning stock and the size of resultant year classes. From 1964 to 1970, the annual catch of spawning age fish averaged only 14% of the previous 9 yr. Resultant per-year recruitment from 1964-70 averaged 42% of that for the previ- ous 9 yr (Schaaf 1972). A description of the aver- age relationship between spawning stock size and recruitment is useful for examining this coinci- dent reduction and for predicting the expected fate of the fishery under different exploitation regimes. A stock-recruit function is also the necessary starting point for developing an index of survival (observed recruitment to that expected from number of spawners) against which one may assess the impact of density-independent en- vironmental effects of recruitment. The approach in this study has been to determine if selected density-independent environmental factors could explain deviations from a postulated spawner- recruit model. Ricker's (1954) comprehensive study of stock- recruitment formulated a dome-shaped model, with strong compensation, resulting in decreased recruitment at stock sizes beyond some maximum value. It is described by the equation: R = Se (S r~ S),Sm where R = recruitment S = spawning stock e = base of natural logarithm S r = maximum equilibrium stock S m = stock size yielding maximum absolute recruitment. Ricker's model states that some stock size (S m ) produces maximum recruitment, and that, be- cause of density-dependent mortality and growth, 28 NELSON ET AL.: LARVAL TRANSPORTOFBfi£VOO/?77A TYRANNUS stocks greater than S m produce progressively fewer recruits. There is a size-dependent fecundity relationship for Atlantic menhaden (Higham and Nicholson 1964), and growth rates are slower for large year classes (Gene R. Huntsman, pers. com- mun., Atlantic Estuarine Fisheries Center). Also, adult menhaden are indiscriminate filter feeders and are known to ingest their own eggs. Calcula- tion of a density-dependent index for Atlantic menhaden (i.e., the slope of a regression of InR on InS ) yields a value of 0.238. This index falls within the category described by Cushing (1971) as hav- ing a slightly convex spawner-recruit curve. The average fecundity of Atlantic menhaden (113,000 eggs per female) calculated from data used in this study, also places the species in groups which Cushing describes as having a dome-shaped spawner-recruit curve. Accordingly, the Ricker model has been used in this analysis, instead of models proposed by Beverton and Holt ( 1957), and others. Schaaf and Huntsman (1972) presented a Ricker spawner-recruit curve for Atlantic menhaden. The same catch data and basically the same methodology were used in this study, with one important modification. Instead of using the estimated total number of spawning age fish as the independent variable to estimate recruitment, the potential number of eggs that could be produced from the spawning stock was used. This annual potential is influenced by the age distribution of the spawners and their average size. The potential number of eggs produced each year and at each age (Table 2) was calculated from the estimated number of age 3 and older females (1955-70), their back-calculated length, and the following fecun- dity relation from data presented by Higham and Nicholson (1964): ME) = 0.3149+0.0176(/) where E = thousands of eggs produced per female at length, and / = back-calculated length at age of an- nulus formation for age-3 and older fish. Another deviation from the data used on the original Ricker spawner-recruit curve by Schaaf and Huntsman (1972) is the calculated number of recruits in the 1955-70 year classes. The numbers differ between the two studies because: 1) some adult menhaden were reaged following the initial study which brought about slight changes in estimates of year-class size, 2) the maximum instantaneous fishing mortality rates were av- eraged for age-specific exploitation rates for age 2-5 fish and were not weighted for numbers at age as was done in the earlier study, and 3) the exploitation rate of age-1 fish was estimated each year based on the exploitation rate of age 2-5 fish instead of an estimated exploitation rate of two- thirds that of older fish as was done in the previous study. This was necessary because shifts in fishing area and effort in recent years have increased the vulnerability of age-1 fish. The parameters of the Ricker model were es- timated from a linear regression of ln(i?/S) on S. Fitting the model (Figure 2) yielded an estimate of S m equal to 60 x 10 12 eggs. This is equivalent to 531 million spawning females spread over ages 3-6, and would produce an average recruitment of 3.68 billion fish at age 1. TABLE 2. — Estimated number of eggs produced by spawning stock of Atlantic menhaden for each year class by age, 1955-70. Age Year 3 4 5 6 7 8 + Total eggs 4.3 x W" - 1955 36.2 72.1 12.6 0.9 0.3 126.4 1956 45.7 11.1 52.8 12.5 3.4 1.1 126.6 1957 15.5 15.1 12.2 13.8 1.8 0.6 59.0 1958 11.4 6.3 6.8 4.9 3.0 0.3 32.7 1959 49.0 10.8 5.0 6.0 2.5 1.1 74.4 1960 18.1 368 12.6 4.7 1.7 0.5 74.4 1961 146.2 5.5 12.0 1.4 0.6 0.2 165.9 1962 23.9 56.7 7.2 6.4 0.9 0.2 95.3 1963 15.4 8.8 12.2 3.3 1.1 0.2 41.0 1964 8.5 3.8 1.9 2.1 0.5 0.1 16.9 1965 7.8 1.7 0.3 0.4 0.2 + 10.4 1966 3.9 0.9 0.1 + 0.1 + 5.0 1967 9.7 1.0 0.1 + 10.8 1968 6.7 2.0 0.2 + 8.9 1969 9.4 1.4 0.1 + 10.9 1970 7.7 2.9 0.2 10.8 + = less than 0.05 x 10 12 . 20 40 60 80 100 120 140 160 ISO SIZE OF SPAWNING STOCK {NO OF EGGS * 10' J ) FIGURE 2. — Ricker spawner-recruit relationship for Atlantic menhaden, 1955-70. 29 FISHERY BULLETIN: VOL. 75, NO. 1 Because the regression of \n(R/S) onS, as is done for the Ricker equation, will automatically give a significant correlation coefficient, a nonlinear fitting procedure was also applied to the data (Marquardt 1963). A comparison of the residual mean squares of the two procedures yielded anF of 1.02, indicating no significant difference in the fit of the Ricker curve to the spawner-recruit data between the standard technique and the nonlinear estimation. Few published stock-recruitment curves appear to fit the observed data well, and the one for At- lantic menhaden is no exception. Application of a power function of the form R = aSh to the data resulted in a fit that was not significantly better from that of the Ricker function. The purpose of the study, however, is to examine and explain the deviations from the curve caused by density- independent factors, to see if they can be predicted, and consequently to improve upon a management plan based solely on a long-term, average MSY concept. The survival index (Table 3) represents the ratio of observed recruits (the number of age l's in the population as estimated from the catch of age l's and estimated exploitation rates) to the number calculated from the Ricker spawner- recruit model. This ratio is an index of survival, independent of density, and should reflect those environmental effects which influence survival of menhaden from the time of spawning until the time of recruitment to the fishery at age 1. INFLUENCE OF EKMAN TRANSPORT AND OTHER FACTORS The influence of transport processes in the southern part of the spawning range is indicated in Figure 3 which depicts the Ekman transport index for the January-March spawning period for 1955-70 and the estimated number of menhaden recruits at age 1 from the year class. The re- sponsiveness of survival to transport shows up well in the Figure where years of strong westward transport correspond with large year classes, and weak transport years with smaller year-class size. Also, increases and decreases in recruitment from one year to the next generally coincide with an increase or decrease in westward transport in the year in which the year class was produced. The correspondence is weaker in the 1968-70 year classes, although it follows the general pattern. Intense fishing pressure over a number of years changed the age structure of the spawning TABLE 3.— Estimated number of eggs, observed and expected number of recruits at age 1, and density-independent survival index for Atlantic menhaden, 1955-70. No. No. of observed No. of expected Survival Year of eggs recruits (fi ) recruits (ft. ) index class x 10 12 x 10 6 x 10 6 ^o Rc 1955 126.4 5,019 2,569 1.95 1956 126 6 4.984 2.568 1.94 1957 56.0 2.538 3,688 069 1958 32.7 11,540 3,166 3.64 1959 74.4 2,007 3.599 056 1960 74.4 2,568 3,598 0.71 1961 165.9 1,553 1,751 089 1962 953 1,740 3.253 0.54 1963 41.0 1.378 3,457 0.40 1964 16.9 1,408 2.134 066 1965 10.4 1,406 1,472 0.96 1966 5.0 1.579 773 2.04 1967 10.8 922 1,505 0.61 1968 8.9 1,324 1,282 1.03 1969 10.9 2,763 1,521 1.82 1970 10.8 1,415 1,499 0.94 stocks to a considerable extent. For example, approximately 40% of the estimated spawning stock in 1958 were 4 yr or older. The number of age 4 and older fish in the 1969 spawning population was only about 9%, and the average number of eggs per spawning female was about 50,000 less than in 1958. Thus, fishing pressure brought about an even greater reduction in spawning potential than is apparent when considering the number of spawners alone, because of a reduction in the average age. This reduction in real spawn- ing potential reduced the opportunity for a large- scale response to favorable transport in the 1968- 70 year classes. Comparison of the density-independent survi- val index with Ekman transport yields a sur- prisingly consistent relationship (Figure 4). A _^ s z o 12 sj *^ CD o - 10 X o No. of R«cruitt 7 _ 130 O < H U O at 3 Qr- 6 A Ekman Tramporl— — ^ ^ / \ ' 90 at o I in A '*"' "*\ 1 1 / 1 \ 60 a. Z < lu \ at i / 1 1 i-*ti ' O 2 - 1 v / 30 at CO t i n r^i < * 4> 3 \l h — * \ i i/t z n * 70 Y EAR CIA ss FIGURE 3.— Observed number of Atlantic menhaden recruits at age 1 and sum of average monthly zonal Ekman transport at lat. 35°N, long. 75°W for January-March of spawning years, 1955-70. 30 NELSON ET AL : LARVAL TRANSPORT OF BREVOORTIA TYR ANNUS o z < > > at 3 lo 46 J6 40 - 50 60 70 80 WESTWARD TRANSPORT (METRIC TONS X 10/SEC/KM) FIGURE 4. — Linear regression of calculated survival index (observed recruits/calculated recruits) for Atlan- tic menhaden on sum of January-March zonal Ekman transport at lat. 35°N, long. 75°W, 1955-70. linear regression of survival indices against transport values for the January-March spawning periods at lat. 35°00'N and long. 75°00'W results in an r of 0.789 significant at the 0.001 level with 14 df (Figure 4). This accounts for approximately 629c (r 2 = 0.622) of the variation between observed and expected recruitment. Since the transport is indicative of conditions over only a portion of the total spawning range of Atlantic menhaden, and since r 2 accounts for such a large share of the total variation in overall recruitment, the actual effect of transport processes in the southern spawning area must be of overriding significance for the survival of spawn south of Cape Hatteras. With the exception of 1966, the index of survival was greater than 1.0 only when the Ekman transport index indicated a strong westward transport for the January- March period of menhaden spawning activities south of Cape Hatteras. The transport data fall conveniently into groups of 0-200, 200-500, and 500-1,000 metric tons/s- km of ocean front. Five years of strong westward transport (>500) were found, and in all of these years the survival index was greater then 1.0. The observed recruitment exceeded the expected by an average of 108%, with the 1958 year class showing the largest value. In 6 yr of low westward trans- port (0-200), the survival index was never greater than 1.0. In 5 yr of moderate or "average" west- ward transport, (200-500) high survival occurred in 1 yr, and poor or moderate survival in the other 4 yr, indicating the influence of additional factors over the spawning range that are operating to produce variations in year-class strength. The high index for 1966 may partially result from the fact that the estimated spawning stock production of 5 x 10 12 eggs was, by far, the lowest of any year on record (Table 2). Under such low stock size, density-dependent survival may have exceeded that indicated by the Ricker curve, creating an artificially high index of survival. A slight un- derestimation in the computation of the number of spawners would also create a very high survival index, since the slope of the Ricker curve is ex- tremely steep as spawning stock size approaches zero (Figure 2). Transport values at lat. 33°N, long. 78°W, approximately 200 nautical miles southwest of lat. 35°N, long. 75°W were also considered. The data are from a point offshore of Long Bay, S.C., the southernmost of the cuspate Carolina bays, and serves as an indicator of Ekman transport in the extreme southern part of the Atlantic menhaden spawning range. A significant corre- lation existed between transport for the January-March period and the survival index (Table 4). Due to the correlation between the two transport values south of Cape Hatteras, however, little additional variation is accounted for by the southernmost transport value (Table 5). Since transport is a function of wind stress and Coriolis force, movements of air masses through the southeastern United States would give parallel transport values at the two locations, with inten- sity of transport dependent on variations within the air mass. The large amount of variation ac- counted for by the two transport indices south of Cape Hatteras is sufficient to account for the rela- tive success or failure of a year class, and supports the observation that a significant portion of menhaden spawning takes place south of Cape Hatteras. 31 FISHERY BULLETIN: VOL. 75, NO. 1 TABLE 4. — Stepwise regression of survival index of Atlantic menhaden on environmental factors. Correlation Individual Error Cumulative Time of with level of Cumulative mean percent of Factor No. year survival index significance correlation square variance Zonal Ekman transport *i Jan. -Mar 0.789 0.001 0.789 0.298 62.2 lat. 35°N, long. 75°W Chesapeake Bay *6 July-Sept. -0.216 — 0.825 0.271 68.0 discharge Zonal Ekman transport *3 Nov -Feb. 0.352 — 0.840 0270 70.6 lat. 39°N, long. 72°W Zonal Ekman transport ** Nov -Feb 0.519 0.05 0896 0.198 80.3 lat. 39°N, long. 75°W Minimum temp * 5 Jan. -Feb. -0 177 — 0.914 0.181 83.6 Delaware Bay entrance Zonal Ekman transport x 2 Jan-Mar. 0.720 0.005 0.919 0.190 84.5 lat. 33°N, long. 78°W TABLE 5. — Regression coefficients between independent en- vironmental variables used in the recruit-environment predic- tive equation for Atlantic menhaden. See description of X's in Table 4. x 2 *3 *4 *5 *6 *, 0.789 0.645 0.644 -0.333 0.032 X, 0.589 0.701 -0.580 -0.068 x, 0.868 -0.403 0.174 ** -0.510 0.213 * 5 0.023 Wind-driven transport off Delaware Bay was studied as being representative of menhaden spawning areas in the Middle Atlantic Bight. Because the transport values are produced in a 3° grid by the Pacific Environmental Group, there were no available data for a point located centrally on the continental shelf. Two locations were chosen: one at lat. 39°N, long. 75°W, near the mouth of Delaware Bay, the other at lat. 39°N, long. 72°W, near the outer edge of the continental shelf. The two locations are approximately 260 km apart in an east-west direction, and are felt to be representative of Ekman transport over the broad shelf area near the east-west axis of the Middle Atlantic Bight. The entrance of larvae into estuaries of the Middle Atlantic Bight occurs variably from September to June, with peak immigration oc- curring in the winter. Reintjes and Pacheco ( 1966) reported on 6 yr of larval collection at Indian River, Del., and showed high rates of influx from December through March. The peak month varied from year to year, but stayed within the December-March period. Correlation coefficients between summed transport values for November-February (the peak period of larval drift) and the survival index (Table 4) are not as large as those from south of Cape Hatteras, but the effect of transport on survival at the inshore point (lat. 39°N, long. 75°W) is significant at the 0.05 level. The transport values from the inshore and offshore points account for approximately 27% and 12%, respectively, of the total variance in the survival index for Atlantic menhaden. When combined with the transports south of Cape Hatteras, these values for the Middle Atlantic Bight account for an additional 12+% of the re- sidual variance. Correlation coefficients are lower than those found for the South Atlantic Bight, and may be indicative of: 1) major nearshore spawning activities, reducing the need for a suitable transport mechanism; 2) a lower level of spawning in the area; or 3) a lower level of recruits per spawner due to mortalities from other en- vironmental factors in the area. The model of circulation off Chesapeake Bay developed by Harrison et al. (1967) and discussed in the Larval Transport section would be ap- propriate if larval menhaden were demersal in nature. However, since larvae are more abundant in the upper water column, we would expect a negative relationship between discharge and survival in the Middle Atlantic Bight because high surface discharge would impede larval entrance into estuaries. Chesapeake Bay was chosen to test that hypothesis because of its im- portance as a major nursery area. Average monthly discharge rates from the Susquehanna, Potomac, and James rivers were used in the test because they constitute over 90% of the total inflow into Chesapeake Bay. Discharge during the third quarter (July-September) of the year pre- ceding the year-class year was chosen because there is a lag time of up to 90 days between stream flow and bay discharge (Harrison et al. 1967). The influence from run-off would be felt at the mouth of the Bay in the October-December period when larvae begin entering in increasing abundance. A correlation between the survival index and discharge rate did not result in a significant 32 NELSON ETAL.: LARVAL TRANSPORT OFBREVOORT1A TYRANNUS coefficient (Table 4). When combined with the other factors considered above, Chesapeake Bay discharge accounts for an additional 6% of the residual variance in density-independent year- class strength. A fairer test of the effects of dis- charge on larval transport would require that we isolate that portion of the total larval production that would enter Chesapeake Bay under varying conditions. Our knowledge of Atlantic menhaden spawning activities is not sufficient to do this with reasonable precision. An absence or reduction in the number of larvae in estuaries during periods of extreme cold has been noted by June and Chamberlin (1959) and Reintjes and Pacheco ( 1966). Kendall and Reintjes (1975) hypothesized that severe winters, par- ticularly in the northern segment of the spawning range, result in heavy kills of overwintering lar- vae in the estuaries. In addition, laboratory ac- climation studies have shown high mortality rates when menhaden larvae were held for several days at temperatures below 3°C (Lewis 1965). A time series of minimum mean monthly sea surface temperatures was located for the mouth of Dela- ware Bay from National Ocean Survey Tide Sta- tion Observer Records (U.S. Department of Commerce 1973). These data were considered representative of mid-to-northern coastal areas in the Middle Atlantic Bight. Correlation of the survival index for the entire population and the minimum temperature yielded a low correlation coefficient (Table 4). The correlation is somewhat of an artifact, however, and probably is biased by the positive correlation between Ekman transport and year-class strength. Westward Ekman transport is generated by winds from the north. Years of high westward transport in winter months are years of sustained north winds, which are associated with cold air masses. Under such conditions, we would expect cooler sea-surface temperatures in those years, particularly in or near shallow estuarine areas. There may be a posi- tive correlation between temperature and survi- val, but the relationship probably is masked by the overriding effects of wind-generated Ekman transport (Table 5). The low correlation coefficient could also indicate that only a small portion of the population would overwinter in northern waters where temperature stress might be a significant factor. If low temperature reduces survival, a transport mechanism to carry fall-spawned larvae south- ward along the Middle Atlantic Bight into the vicinity of estuaries that have milder winter temperatures would be a positive survival factor. Therefore, the meridional (north-south) compo- nent of Ekman transport in the Middle Atlantic Bight at lat. 39°N, long. 72°W near the edge of the shelf off Delaware Bay was considered. A corre- lation between the survival index and the southward transport for the October-December spawning period resulted in a coefficient of 0.336, which accounts for about 10% of the total variance in density-independent recruitment. However, the contribution to reduction in residual variance was minimal, because all of the variation due to southward transport was accounted for by linearly related east-west zonal Ekman components al- ready considered. A relatively steady state southward transport mechanism exists in the Middle Atlantic Bight in the form of a southward flowing current over the shelf (Bumpus 1973). Because this current is quasi-permanent, vari- ations in southward Ekman transport may be of little significance and may only create minor fluctuations in strength of an existing transport mechanism. RECRUIT-ENVIRONMENTAL MODEL The logic used in the selection of environmental parameters for inclusion in a model of en- vironmental effects is depicted schematically in Figure 5. The heavy line represents an intuitive weight of density-dependent and density- independent factors in the survival of menhaden larvae from the time of spawning through their oceanic phase. In the upper Middle Atlantic Bight, for example, spawning takes place close to shore or in major bays and sounds, reducing or eliminating the time spent by larvae in the open ocean. This would reduce dependence on favorable currents for transport. Under such conditions, environ- mental factors influencing mortality may be rela- tively stable, with variation in the number offish spawning in the area being the probable cause of most of the variation in the number of recruits produced. In the South Atlantic Bight, however, spawning takes place offshore, and dependence on favorable ocean currents would seem to have greater weight than spawning stock size on survival. Large annual variations in transport would produce large variations in survival in the South Atlantic Bight at a given stock size. The lower Middle Atlantic Bight seems to be an in- tergrade between the two extremes, with sig- 33 FISHERY BULLETIN: VOL. 75, NO. 1 STOCK SIZE (Survival Density Dependent) ENVIRONMENTAL CONDITIONS (Survival Density Independent) YEAR-CLASS SUCCESS FACTORS Actual Survival Index (Ro/Rc) Predicted Survival Index FIGURE 5. — Schematic representation of logic used in the de- velopment of the survival index predictive model. Location of environmental parameters used in the model is indicated byXn, description of parameters in Table 4. nificant spawning taking place farther offshore as adults migrate southward in the fall. This should result in increased significance of oceanic trans- port factors from north to south in the determi- nation of year-class strength. The hypothesis of increasing importance of transport as spawning activities move progressively farther offshore is supported by the highly significant correlations between the survival index and transport values south of Cape Hatteras and similar correlations which have a lower level of significance off Dela- ware Bay. The selection of locations and time periods for Ekman transport data was based on the availa- bility of data for specific coordinates, desire for representation of broad spawning areas, and estimates of larval drift time and direction (Figure 5). Of the many possible environmental factors which could influence survival during the oceanic phase, three (transport, temperature, and river discharge) were chosen because they appeared to be factors of major importance and data series were available for the same period in which vital statistics of the Atlantic menhaden populations have been taken. x UJ Q Z -j < > > 19SS 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 FIGURE 6. — Multiple regression of the survival index for Atlan- tic menhaden on environmental factors, 1955-70. Predictive equation and listing of environmental factors presented in text. Correlation coefficients and model data presented in Tables 4 and 6. TABLE 6. — Data used in recruit-environment predictive model for Atlantic menhaden. Location of factors identified in Figure 5, individual factors identified in Table 4. Year class S.I. Environmental factors X, X 2 *3 X 4 *5 x 6 1955 1.95 70 74 272 152 2.4 9 1956 1.94 64 107 307 271 1.0 30 1957 0.69 12 13 83 34 3.4 29 1958 3.64 94 124 141 169 0.6 9 1959 0.56 7 13 126 82 1.5 24 1960 0.71 40 42 121 78 1.6 11 1961 0.89 30 33 155 129 1.6 22 1962 0.54 27 63 149 79 1.4 17 1963 0.40 3 70 206 158 -0.3 9 1964 0.66 14 43 120 127 2.0 16 1965 0.96 11 32 96 35 1.8 7 1966 2.04 26 55 125 104 1.3 7 1967 0.61 4 21 98 63 1.8 8 1968 1.03 60 96 161 97 0.0 23 1969 1.82 92 76 317 212 0.4 16 1970 0.94 39 47 185 156 0.2 28 The multiple-regression model developed to relate recruitment to environmental variables yields a correlation coefficient of 0.919, significant at 0.003 with 9 df (Figure 6). Model data are given in Table 6. The model accounts for over 84% of the variance in the actual survival indices (Table 4). Translated into recruits, the model indicates that over 84% of the variation between actual re- cruitment into the fishery and expected re- cruitment during the 1955-70 period is accounted for by environmental fluctuation. The model is described by the equation: S.I. = 0.4148 + 0.0205XJ +0.00530Z 2 - 0.00807X 3 + 0.00950X 4 + 0.23967X 5 - 0.02679X 6 ± e where S.I. = survival index computed by dividing observed recruits by expected re- cruits 34 NELSON ET AL.: LARVAL TRANSPORT OFBREVOORTIA TYRANNUS X x = sum of monthly average zonal (westward) Ekman transport rates for January-March of the year-class year at lat. 35°N long. 75°W X 2 = sum of monthly average zonal (westward) Ekman transport rates for January-March of the year-class year at lat. 33°N, long. 78°W X 3 = sum of monthly average zonal (westward) Ekman transport rates for November-December of the year prior to the year class and January- February of the year-class year at lat. 39°N, long. 72°W X 4 = sum of monthly average zonal (westward) Ekman transport rates for November-December of the year prior to the year class and January- February of the year-class year at lat. 39°N, long. 75°W X 5 = minimum mean sea surface temper- ature at the mouth of Delaware Bay in the year-class year X 6 = sum of monthly average discharge rates from Susquehanna, Potomac, and James rivers in July-September of the year preceding the year-class year e= error term. The predicted number of recruits for each year is given by: R p - R CI x S.I. where R p = Rr, = predicted number of recruits number of recruits calculated from the Ricker curve at spawning stock size in the ith year. A correlation between the observed number of recruits (R ) and the predicted recruits (R p ) for each year yields a coefficient of 0.943 and a slope of 0.914 with no systematic bias around the regres- sion line. Further evidence of the validity of the model is the failure of adjustments to increase the percent of variance accounted for by the en- vironmental factors. The initial model, based on judgments of the proper time and location of environmental parameters, yielded a higher correlation coefficient than any subsequent mod- els in which any of the parameters or time-spans were varied away from those which were consid- ered the most significant from a biological stand- point. The parameters were not selected by a screening process from a large number of vari- ables, but were selected because of their probable impact on survival. The four largest year classes ( 1955, 1956, 1958, and 1969) during the 16-yr period are accurately described by the model. The average error of prediction for these years is 4.3% and the maximum error is 6.3%. Smaller year classes are not described with the same degree of accuracy, although the mean error for the 16-yr period is reduced from 1.5 billion fish using only the Ricker curve to 610 million individuals per year by the model, and the standard error of the mean is re- duced from 501 to 155 million fish. The multiple-regression model has a high correlation coefficient and therefore describes the data well. Its value for prediction is somewhat more tenuous and requires testing on a sub- sequent set of data to determine its accuracy. The model was not broken into separate time-series units for testing because of the brevity of the 16-yr data base. The model is a first-cut approximation for the evaluation of transport and other factors. The number of variables included tends to increase the R 2 value, even though some parameters do not show individual significance levels when corre- lated with the survival index. However, only the Chesapeake Bay discharge has a /3 value of which ±2 standard errors encompasses 0, indicating that the factor is probably not significant. The other parameters are associated with the same major air mass movements, and are therefore interrelated. A more sophisticated model should be based on either principal components regression or Ridge regression techniques to correct for the inter- dependence of some of the parameters and to improve the predictive capability. A reduction in the number of variables used is desirable from a statistical standpoint because of the short time span of the data base. Regression of the survival indices on the three transport values off of Cape Hatteras (lat. 35°N, long. 75°W) and Delaware Bay (lat. 39°N, long. 72°W; lat. 39°N, long. 75°W) yields an R 2 of 0.741 (12 df, P<0.001). The ab- breviated model accounts for a significant portion of the variance around the spawner-recruit curve. It describes the data for high and low survival years nearly as well as the full model and probably has a similar predictive capability. Determination of the actual influence of the other factors (dis- 35 FISHERY BULLETIN: VOL. 75, NO. 1 charge and temperature) which were included because of their potential biological importance will require a greater knowledge of spawning intensities and a longer term data base. Overall, the model implies a predictive capabil- ity for large year classes and for extremely poor year classes. The model provides a satisfactory indication of the general magnitude of a year class prior to entering the fishery in 14 of the 16 yr. For initial model purposes, the survival index was not computed beyond 1970 because the 1971 year class is still being harvested by the fishery, and the total catch from that year class necessary for verification of the number of recruits is not known. Forecasting in real time can be ac- complished by inserting the routinely available environmental data into the survival index equation. The expected number of recruits for a given year class is obtained by determining age structure and abundance of 2-yr-old and older fish from fishery landings the previous fishing season, estimating an exploitation and survival rate to determine the number that will survive to spawn the next year class, calculating the expected number of eggs produced, and estimating the expected number of recruits from the Ricker function. Multiplying the expected number of recruits by the predicted survival index gives the predicted number of recruits. Estimates of the number of recruits can be made as early as April of the year-class year, and can be revised when ac- tual exploitation rates are determined to allow better estimates of the size of the spawning stock which produces the year class. Thus, an initial prediction of the number of recruits can be made approximately 1 yr before they become available to the fishery the following spring. DISCUSSION Refinement of the predictive capability of the recruit-environment model is dependent on in- creased knowledge of the biology of Atlantic menhaden and on better understanding of the effects of the many factors that influence dis- tribution, abundance, and survival. The model is concerned only with variation introduced into year-class size during the relatively short life phase in which larvae are oceanic and before metamorphosis takes place. The model concen- trates on those factors which influence larval distribution and act as a mechanism to transport larvae into the vicinity of estuarine nursery grounds, thereby increasing survival. Major sources of variation such as food availability and predation have not been directly considered. However, since these factors are, to some extent, influenced by the number of larvae produced by the spawning stock, variations induced by them should be partially accounted for by the density- dependent Ricker function. The actual fluctuation in availability of food could only be determined by broad-scale surveys over the entire menhaden spawning range and would require a continuous time series for a number of years. Likewise, the determination of predation and cannibalistic influences would require extensive field surveys and controlled laboratory experiments. Problems in determining the influence of pertinent environmental factors are compounded by the large geographic range of menhaden spawning activities. The influence of any one particular factor at a specific location could only be determined if the amount of spawning at that location was known. Comparison of environmen- tal factors against a survival index for the entire stock, as has been done in this study, requires the selection of broad-scale factors having major influence over large portions of the spawning range, or the selection of representative data which provide a generalized environmental index for a selected factor. Localized variations may be highly significant, but masked by overall survival success or failure without knowledge of localized spawning intensity. Cushing (1969, 1974) cited failures in attempts by other authors to correlate year-class strength and winds (or pressure gradients), and suggested that variation in wind direction may be a greater source of variation than the strength of winds from a single direction. The U.S. east coast is composed of an almost continuous series of bays and sounds, which extend both north and south of the major spawning region for Atlantic menhaden. Under these circumstances, variations in wind direction would probably influence the route of larval drift. However, unless northward or southward larval movement was extreme, larvae would not be transported away from suitable nursery areas as long as there was a significant onshore component of wind-driven circulation. Thus wind direction would be a significant factor only if that direction reduced the westward component of Ekman transport or if the normal seasonal wind pattern reversed, generating eastward (offshore) trans- port. 36 NELSON ET AL.: LARVAL TRANSPORTOFfifl£VOOft77A TYRANNUS Comparison with Pacific Sardine Computed survival indices allow comparisons between the Pacific sardine and Atlantic menhaden, in addition to those detailed by McHugh (1969). Radovich (1962) presented data for Pacific sardine showing the effect of good, average, and poor environmental conditions on the spawner-recruit relationship. He used maximum and minimum parabolas based on highest and lowest recruitment years and iden- tified the area between the curves as indicative of the effects of the environment as well as spawning stock size on recruitment. A similar approach, modified by using the right-hand skewed Ricker curve yields similar results (Figure 7). Year clas- ses used in the computation of the maximum and minimum recruitment curves for Atlantic menhaden were not selected for high and low recruitment as was done by Radovich, but were selected because they represented extremes in the variation of transport factors. The maximum recruitment curve was developed from year-class size during the 3 yr of highest (3^700 metric tons/ skm) southern onshore transport (1955, 1958, 1969). Similarly, the minimum recruitment curve was computed from year-class size during the 3 yr of lowest (<100 metric tons/s-km) onshore transport (1959, 1963, 1967). The two curves represent a wide range of environmentally in- duced fluctuation around the stock and re- cruitment curve calculated from the 1955-70 data base. No statistical significance can be attached to the upper and lower curves because each is based on three data points. However, the figure indicates the range of variance that masks the density- dependent function if pertinent environmental factors are not identified and weighted for effect at various stock sizes. The greater slope of the maximum curve is of particular interest, indicat- ing a significant loss of potential recruits in good environmental years if adequate stock size is not maintained. Additional parallels can be drawn between Pacific sardine and Atlantic menhaden spawner- recruit relationships during periods of overfishing and low survival. A comparison of spawning stock size and year-class size for the two species linked in chronological order shows striking similarities (Figure 8). In each case, there was a period of several years at high stock size in which the size appeared to be near or past the maximum needed to produce large numbers of recruits. A series of |333-S]/55 SIZE OF SPAWNING STOCK (NO OF EGGS x 10") FIGURE 7. — Ricker spawner-recruit relationships calculated for years of good and poor environmental conditions. The upper curve is calculated from observed recruitment during the three greater years of Ekman transport, the middle curve is calculated from the 16-yr data set, and the lower curve is calculated from observed recruitment during the three lesser years of Ekman transport. good year classes ( 1937-39 for sardine; 1955, 1956, and 1958 for menhaden) was followed by a series of poor survival years (1940-45 for sardine, 1959-64 for menhaden). These reductions in recruitment, combined with excessive fishing pressure, reduced spawning stock size drastically, leading to a re- stabilization of stock and recruitment around small stock levels. In the case of menhaden, the 5-yr period of decline reduced the spawning stock size by an order of magnitude. By 1966, spawning potential had dropped to a low of 5 x 10 12 eggs from the 1961 high of 165 x 10 12 . The parallel between the two sets of data is a cause for concern, because the decline and apparent restabilization of Pacific sardine stocks was followed by a com- plete collapse of the fishery. Henry (1971:23) in his analysis of the decline of the Atlantic menhaden fishery stated, "I am concerned that the stocks of Atlantic menhaden may have been reduced to a level that is having an adverse effect on recruit- ment." Clark (1974:14), in a study of the effects of schooling on population dynamics on small school- ing species (as in the case with Atlantic menha- den), concluded that, "A commercial fishery based on such a species might be expected to experience a rather spectacular population collapse, which could be brought on either as a direct result of an increased fishing effort which suddenly trans- forms the system into an unstable mode, or as an indirect result of fishing which reduces resiliency and renders the population vulnerable to the ef- fects of random environmental fluctuations." The possibility of a complete collapse in the Atlantic 37 FISHERY BULLETIN: VOL. 75, NO. 1 o z O 7 PACIFIC SARDINE 2 9 39 ! i 3? 40 v\ ' 33> .--'43 \35 34 u: ATLANTIC MENHADEN 55 % -'56 FIGURE 8. — Year-class size related to spawning stock size and linked in chronological order for Atlantic menhaden and Pacific sardine. Pacific sardine figure after Radovich (1962:134). SPAWNING STOCK SIZE (BILLIONS OF FISH) SIZE OF SPAWNING STOCK |NO OF EGGS * 10"l menhaden fishery, given high fishing effort and additional years of poor survival, cannot be dis- counted. Fortunately, there are significant differences in the environment, biology, and fishery of Pacific sardine and Atlantic menhaden. One of the more important differences is the estuarine depen- dence of menhaden. In every year, at least some estuarine systems on the east coast should provide favorable environments, insuring good survival of larvae which reach those nursery grounds. Also, spawning activities spread over the entire coast should include at least some areas conducive to survival, reducing the chance of almost no survival over the entire range. Climatic change which shifts the distribution of menhaden spawning activities would not likely shift the spawning region far enough away from suitable nursery areas to cause the type of massive failure that occurred in the sardine fishery. Another significant factor in the collapse of the sardine stocks was an increase in the stock size of compet- ing species, filling the niche in the ecosystem as the sardine population decreased. Although there is no fishery for species which are potentially competitive with Atlantic menhaden and adequate stock data on such species are not av- ailable, there are no indications of large increases in abundance of any coastal pelagic species, and the niche available to menhaden appears to be open. However, John Radovich (pers. commun., California Department of Fish and Game) points out that "the value of not having identified an increase in competitors for the menhaden may be of little significance because: 1) The sardine collapse and failure to recover may have happened without a 'competing' species such as the anchovy. 2) Available forage and habitat may be utilized through slight increases in the abundance of several species, and hence go unnoticed. 3) The capacity within a trophic level may vary considerably so that actual changes in the abundance of competing species may be masked by changes in available forage and habitat." The menhaden fishery is somewhat self- regulating, in that low stock levels have brought about economic conditions which forced a reduc- tion in effort and closure of processing plants. The closure of plants in the northeast United States during the late 1960's reduced fishing effort on older age-groups, halting the drastic decline in spawning stock size (Schaaf in press). This action, plus good survival in 1966 which produced the spawning stock for the high transport, large year- class year of 1969, is probably responsible for the brief resurgence of the fishery in the early 1970's. Implications for the Fishery Implications for the fishery are rather straightforward: in years of poor environmental conditions recruitment is low regardless of stock size; extremely low spawning stock sizes in years of poor environmental conditions result in re- cruitment below the level needed to maintain the fishery; favorable environmental years will 38 NELSON ETAL.: LARVAL TRANSPORT OF BREVOORTIA TYRANNUS produce exceptional year classes and a propor- tionally greater harvestable surplus at stock sizes near the spawning optimum; and a series of poor environmental years (1959-64), coupled with excessive fishing pressure, will reduce stock size to a level which produces little harvestable surplus. During the 16 yr covered by this study ex- tremely large year classes were produced in 3 yr (1955, 1956, and 1958). Favorable conditions in 1969 resulted in a high survival rate, but only produced 2.7 billion recruits because of small spawning stock size. In one other year (1966) survival occurred that was greater than expected, but at extremely low stock size. In the other 11 yr recruitment was either near, or well below the expected level, compounding the stock depletion caused by excessive fishing pressure. The drastic reduction in stock size resulted in a restabilization of the stock-recruitment relationship around a low stock level. This is evidenced by the steady decline in catches from 1956 to a low of 162,000 metric tons in 1969, followed by slightly higher catches in succeeding years (Table 7). Extremely large catches in the late 1950's are the result of the unusual coincidence of 3 high survival years within a 4-yr span. Average survival over the 16-yr period was much lower, and average year- class size would be considerably smaller, even at optimum spawning stock size. Schaaf and Huntsman (1972) gave MSY es- timates for Atlantic menhaden of 600,000 metric tons based on an equilibrium catch-effort curve from historic data and 380,000 metric tons from a population-prediction model. The population- prediction model dampens the effects of large year classes and probably comes closer to representing long-term MSY than the higher estimates. The maintenance of optimum spawning stock size and several year classes in the spawning stock is vital to insure adequate response to favorable environmental conditions. Based on the estimated survival rates over the 16-yr period, and the optimum spawning stock size from the Ricker function, surplus yield was calculated under conditions which would maintain four spawning groups (ages 3-6) in the populations. The calcu- lation of surplus yield is based on an instantane- ous natural mortality of 0.42 and fishing mortality of 0.36 spread over 6 yr within a year class (ages 1-6) and assuming that one-half of the age-1 re- cruits are vulnerable to the fishery. A full complement of years 1-6, from year-class data available after 1954, was not obtainable until 1961, when 6-yr-old fish were harvested from the 1955 year class. Under the conditions imposed on the harvest, the allowable catch, computed for 1961-71, averaged 419,000 metric tons/yr (Table 7). Extremes in the allowable catch would have ranged fron 227,000 to 633,000 metric tons, depending on the size of year classes which con- stituted stock size in a particular year. This catch is similar to the MSY estimates of Schaaf and Huntsman (1972), and was computed for a period in which most of the year classes had less-than- expected survival. The survival index was well below 1.0 from 1959 to 1964, a period of six con- tinuous years, and is reflected by the decline in surplus stock during that period. Actual catches made by the fishery from 1955 to 1971 (Table 7) averaged approximately the same as MSY, but TABLE 7. — Catch of Atlantic menhaden at MSY for actual survival rates, 1955-70 year classes, fishery landings, 1955-71, and predicted surplus from recruit-environment model. Potential catch at Sm Actual catch by fishery Predicted catch Year of No. in Wt (thousand Wt/fish No. in Wt (thousand Wt/fish Wt (thousand harvest billions metric tons) (9) billions metric tons) (9) metric tons) 1955 3.12 641.4 206 1956 3.56 721.1 203 1957 3.51 602.8 172 1958 2.72 510.0 188 1959 5.35 659.1 123 1960 2.78 529 8 191 1961 1.68 632.9 377 2.60 575.9 222 510.9 1962 1.38 488.1 354 2.01 537.7 268 466.7 1963 1.10 4100 373 1.76 346.9 197 412.5 1964 0.88 339.0 385 1.73 269.2 156 392.5 1965 0.76 226.6 298 1.50 273.4 182 295.5 1966 099 254.9 257 1.34 219.6 164 374.2 1967 1.72 367.4 214 0.98 193.5 197 371.5 1968 1.62 472.0 291 1.14 234.8 206 405.1 1969 1.40 426.0 304 0.87 161.6 185 387.0 1970 1.81 464.7 257 1.40 259.3 185 471.5 1971 1.78 525.6 295 0.97 250.3 258 521.6 Mean 1.37 418.8 306 2.20 410.4 178 419.0 39 were taken by extensive overfishing in the late 1950's and early 1960's, with a resultant decrease in spawning stock size and age structure. The average catch from 1955 to 1963 was 596,000 metric tons, well above the MSY level. The fishery also took greater numbers of fish of smaller size than was compatible with management to insure adequate numbers of spawners. Thus overfishing, which reduced stock size, was compounded by a series of poor environmental years, further re- ducing the spawning stock to a level below that necessary to provide large surplus yields from the higher survival years of 1966 and 1969. Had optimum spawning stock size been maintained, the fishery should have been able to increase its yield during the 1967-71 fishing seasons by an average of 231,000 metric tons/yr. The value of a predictive model lies in its usefulness for developing strategies to take advantage of exceptional year classes or to avoid overexploitation of poor year classes. Catches based on the number of recruits calculated from the survival index model are similar to MSY and to those averaged by the fishery (Table 7). However, the absolute mean error from the al- lowable surplus is approximately 134,000 metric tons/yr for the actual fishery landings (1961-71) and 48,000 metric tons/yr if harvest had been limited to the predicted surplus. Some overfishing would have occurred because of errors in pre- diction, but it would have been significantly less than that imposed by the fishery during earlier years. Fishing at a level necessary to harvest the predicted surplus would have provided reasonably stable catches, maintained several age-classes in the fishery, maintained adequate spawning stock, and prevented excessive exploitation of the stocks, all desirable factors in the management of fishery resources. ACKNOWLEDGMENTS The authors acknowledge a debt to the late Robert L. Dryfoos who was instrumental in the initiation of this work. We also express our ap- preciation to David R. Colby for assistance in computer analyses, to Herbert R. Gordy for the illustrations, and to Valerie N. Ward for assis- tance with the manuscript. FISHERY BULLETIN: VOL. 75, NO. 1 LITERATURE CITED BAKUN, a. 1973. Coastal upwelling indices west coast of North Ameri- ca, 1946-71. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-671, 103 p. BEVERTON, R. J. H.. AND S. J. HOLT. 1957. On the dynamics of exploited fish populations. Fish. Invest. Minist. Agric, Fish. Food (G.B.), Ser. II, 19, 533 p. BUMPUS, D. F. 1973. A description of the circulation on the Continental Shelf of the east coast of the United States. Prog. Oceanogr. 6:111-157. CARRUTHERS, J. N. 1938. Fluctuations in the herrings of the East Anglian autumn fishery, the yield of the Ostend spent herring fishery, and the haddock of the North Sea — in the light of relevant wind conditions. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 107(3): 10-15. Clark, c. W. 1974. Possible effects of schooling on the dynamics of exploited fish populations. J. Cons. 36:7-14. Clark, F. n., and J. C. Marr. 1955. Population dynamics of the Pacific sardine. Calif. Coop. Oceanic Fish. Invest. Prog. Rep. July 1953-March 1955, p. 11-48. CUSHING, D. H. 1969. The fluctuation of year-classes and the regulation of fisheries. FiskeriDir. Skr. Ser. HavUnders. 15:368-379. 1971. The dependence of recruitment on parent stock in different groups of fishes. J. Cons. 33:340-362. 1974. The natural regulation of fish populations. In F. R. Harden Jones (editor), Sea fisheries research, p. 399-412. John Wiley and Sons, N.Y. harrison, w., j. j. norcross, n. a. pore, and e. m. Stanley. 1967. Circulation of shelf waters off Chesapeake Bight. Surface and bottom drift of Continental Shelf waters between Cape Henlopen, Delaware, and Cape Hatteras, North Carolina June 1963-December 1964. U.S. Dep. Commer., ESSA Prof. Pap. 3, 82 p. henry, k. a. 1971. Atlantic menhaden (Brevoortia tyrannus) resource and fishery — analysis of decline. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-642, 32 p. HIGHAM, J. H., AND W. R. NICHOLSON. 1964. Sexual maturation and spawning of Atlantic menhaden. U.S. Fish. Wildl. Serv., Fish. Bull. 63:255-271. June, F. C, and J. L. Chamberlin. 1959. The role of the estuary in the life history and biology of Atlantic menhaden. Proc. Gulf Caribb. Fish. Inst., 11th Annu. Sess., p. 41-45. KENDALL, A. W., JR., AND J. W. REINTJES. 1975. Geographic and hydrographic distribution of Atlantic menhaden eggs and larvae along the middle Atlantic coast from RV Dolphin cruises, 1965-66. Fish. Bull., U.S. 73:317-335. LEWIS, R. M. 1965. The effect of minimum temperature on the survival of larval Atlantic menhaden, Brevoortia tyrannus. Trans. Am. Fish. Soc. 94:409-412. 40 NELSON ETAL.: LARVAL TRANSPORT OF BREVOORTIATYRANNUS MARQUARDT, D. W. 1963. An algorithm for least-squares estimation of non- linear parameters. J. Soc. Ind. Appl. Math. 11:431-441. MASSMANN, W. H., J. J. NORCROSS, AND E. B. JOSEPH. 1962. Atlantic menhaden larvae in Virginia coastal waters. Chesapeake Sci. 3:42-45. MATTHIESSEN, G. C. 1974. Rome Point Investigations, Quarterly Progress Rept. Sept. - Nov., 1973. Marine Research Inc., East Wareham, Mass. Mimeo. Data Rep., lip. McHUGH, J. L. 1969. Comparison of Pacific sardine and Atlantic menhaden fisheries. FiskeriDir. Skr. Ser. HavUnders 15:356-367. MURPHY, G. I. 1967. Vital statistics of the Pacific sardine (Sardinops caerulea) and the population consequences. Ecology 48:731-736. PERMUTTER, A. 1939. Section I. An ecological survey of young fish and eggs identified from tow-net collections. In A biological survey of the salt waters of Long Island, 1938, Part II, p. 11-71, N.Y. Conserv. Dep., Suppl. 28th Annu Rep., 1938, Salt- water Surv. 15. RADOVICH, J. 1962. Effects of sardine spawning stock size and environ- ment on year-class production. Calif. Fish Game 48:123- 140. REINTJES, J. W. 1961. Menhaden eggs and larvae from MV Theodore N. Gill cruises, South Atlantic coast of the United States, 1953- 54. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 393, 7 p. 1968. Development and oceanic distribution of larval menhaden. In Report of the Bureau of Commercial Fisheries Biological Laboratory, Beaufort, N.C., p. 9-11. U.S. Fish Wildl. Serv., Circ. 287. 1969. Synopsis of biological data on the Atlantic menhaden, Brevoortia tyrannus. U.S. Fish Wildl. Serv., Circ. 320, 30 P- REINTJES, J. W., AND A. L. PACHECO. 1966. The relationship of menhaden to estuaries. In R. F. Smith, A. H. Swartz, and W. H. Massmann (editors), A symposium on estuarine fisheries, p. 50-58. Am. Fish. Soc. Spec. Publ. 3. RICKER, W. E. 1954. Stock and recruitment. J. Fish. Res. Board Can. 11:559-623. SCHAAF, W. E. 1972. Dynamics of Atlantic menhaden. Brevoortia tyrannus, population inferred from statistics of the purse-seine fishery: 1955-1969. Ph.D. Thesis, Univ. Michigan, Ann Arbor, 42 p. (Diss. Abstr. Int. 33:5153B.) In press. Fish population models: Potential and actual links to ecological models. Proceedings of a symposium "Ecological modeling in a resource management framework." Resour. of the Future, Inc., Wash., D.C. SCHAAF, W. E., AND G. R. HUNTSMAN. 1972. Effects of fishing on Atlantic menhaden stock: 1955- 1969. Trans. Am. Fish. Soc. 101:290-297. SETTE, O. E. 1943. Biology of the Atlantic mackerel (Scomber scombrus) of North America. Part I: Early life history, including the growth, drift, and mortality of the egg and larval popu- lations. U.S. Fish Wildl. Serv., Fish Bull. 50:149-237. SISSENWINE, M. P. 1974. Variability in recruitment and equilibrium catch of the southern New England yellowtail flounder fishery. J. Cons. 36:15-26. STEFANSSON, U., L. P. ATKINSON, AND D. F. BUMPUS. 1971. Hydrographic properties and circulation of the North Carolina shelf and slope waters. Deep-Sea Res. 18:383- 420. U.S. Department of Commerce. 1973. Surface water temperature and density- Atlantic coast North and South America. U.S. Dep. Commer.. NOAA, Natl. Ocean. Surv. Publ. 31-1, 109 p. WALFORD, L. A. 1938. Effects of currents on distribution and survival of the eggs and larvae of the haddock (Melanogrammus aeglefinus) on Georges Bank. U.S. Bur. Fish. Bull. 49:1- 73. 41 EFFECTS OF BENZENE (A TOXIC COMPONENT OF PETROLEUM) ON SPAWNING PACIFIC HERRING, CLUPEA HARENGUS PALLASI Jeannette W. Struhsaker 1 ABSTRACT When female Pacific herring were exposed to low (parts per billion) levels of benzene for 48 h just prior to their spawning, a significant reduction occurred in survival of ovarian eggs and resultant embryos and larvae through yolk absorption. The reduction in survival of ovarian eggs was approximately 10-25%, for embryos from fertilization to hatching, 26%, and for embryos and larvae through yolk absorption, 43%. Exposure to benzene also induced premature spawning and resulted in aberrant swimming behavior and disequilibrium in adults of both sexes. The maximum accumulation of 14 C-labeled benzene and/or metabolites in ovarian eggs (14 times initial concentration in water in 24-48 h; 1.4 /il/g from 0.1 /id/liter) was greater than in later egg and larval stages as measured in other experiments. Conservatively estimating the total reduction in survival in these experiments to be approximately 50% through yolk absorption, I surmise that the effect of exposing spawning herring to only one toxic component of petroleum could have a significant effect on the population. The fish in these experiments were exposed to relatively high parts per billion levels, but they were exposed for a relatively short period (48 h); it is probable that in the estuary, if chronically exposed over a longer period of time to low parts per billion levels of aromatic components, the populations could be seriously affected. When the spawning female herring is compared with other life history stages, we find that the spawning stage is clearly the most sensitive of those tested. If fishes prove generally to be most sensitive to petroleum components during their spawning seasons, fishery management decisions should take this factor into consideration in protecting the resources. In studies of pollutant effects on marine or- ganisms, emphasis should be placed on critical or sensitive life history stages. With this in view, research on petroleum effects on fish has been directed more recently toward egg, embryo, and larval stages (Kiihnhold 1969, 1972; Evans and Rice 1974; Struhsaker et al. 1974). Results in many studies revealed that fish egg and larval stages were surprisingly resistant to crude oil and water-soluble and aromatic fractions of crude oil. Some of this resistance in fish is probably at- tributable to the presence of enzymes for metabolic detoxification of components with ensuing rapid depuration and physiological homeostasis (Lee et al. 1972; Neff 1975; Korn, Hirsch, and Struhsaker 1976, footnote 2). I have observed, as expected, that the effects of exposure of monoaromatics such as benzene are more severe at all life history stages if fishes are 'Southwest Fisheries Center Tiburon Laboratory, National Marine Fisheries Service, NOAA, 3 150 Paradise Drive, Tiburon, CA 94920. 2 Korn, S., N. Hirsch, and J. W. Struhsaker. 1976. The uptake, distribution, and depuration of 14 C-benzene and 14 C-toluene in Pacific herring (Clupea pallasi). Unpubl. manuscr. otherwise stressed by environmental extremes or are in poor "condition" from inadequate nutrition. On this basis it is suggested that the female at time of spawning may be the most sensitive stage to toxic oil components. In herring, for example, the fish often feed poorly for some time prior to spawning and have low fat and energy reserves associated with the production of eggs (Blaxter and Holliday 1963). Anadromous fishes or fishes such as herring which migrate into estuaries for spawning may also be exposed to environmental extremes, particularly to changes in salinity, which produce additional stress. Further, since aromatics are highly lipid-soluble, it might be expected that benzene would accumulate to high levels in ovarian eggs. These factors could lead to significant reductions in fecundity and serious consequences for populations over long chronic exposures. The purpose of this experiment was to examine the effect of benzene on female Pacific herring, Clupea harengus pallasi Valenciennes, just prior to spawning. We have also studied benzene effects on other life history stages of the herring (Struhsaker et al. 1974; Korn et al. see footnote 2; Manuscript accepted September 1976. FISHERY BULLETIN: VOL. 75, NO. 1, 1977. 43 FISHERY BULLETIN: VOL. 75, NO. 1 Eldridge et al. 3 ). So far as we know, there is no similar study, exposing fish just prior to spawning, for any oil component. Benzene was selected for most of our studies on herring because of its relatively high proportion in the water-soluble fraction of crude oil and refined products (Anderson et al. 1974), high solubility in water and relative toxicity (Benville and Korn 1974, footnote 4; Korn, Struhsaker, and Benville 1976). Monoaromatics were tested individually rather than exposing fishes to the total oil or total water-soluble fraction in order to more specifically delineate physiological responses to a known toxic component. Initial research on Pacific herring adults, eggs, and larvae was conducted with high (ppm level) concentrations of benzene (Struhsaker et al. 1974; Korn, Struhsaker, and Benville 1976). Because of the high volatility of benzene, such concentrations would probably occur only briefly after cata- strophic incidents, such as tanker accidents and well blowouts. Subsequently, we tested levels in the low ppb (parts per billion) range as being more representative of chronic exposures and poten- tially more damaging over a long period to marine populations. In this study, ripe male and female herring were exposed just prior to spawning to 100 nl/liter (ppb) and 800 nl/liter (ppb) benzene for 48 h. The re- labeled benzene and its metabolites were mea- sured in the ovaries to determine uptake, ac- cumulation, and depuration. Exposure effects on behavior, the mortality of eggs in the gonads of females, and rate of delayed mortality in embryos at hatching and larvae through yolk absorption were also recorded. METHODS Pacific herring were captured 4 December 1974 during the spawning season in San Francisco Bay by a local bait dealer. The fish were captured with a lampara net and wet-brailed from the net into the vessel bait wells. The fish were transported immediately in the bait vessel to the Tiburon Laboratory dock and then transferred to 1,900- liter tanks in the laboratory. Fish were "running ripe" when captured. Because the purpose of these 3 Eldridge, M. B., T. Echeverria, and J. W. Struhsaker. Manuscr. in prep. The effect of benzene on the energetics of Pacific herring (Clupea harengus pallasi) embryos and larvae. 4 Benville, P., Jr., and S. Korn. Manuscr. in prep. The acute toxicity of six mono-cyclic aroma tics to striped bass (Morone saxatilis) and bay shrimp (Crago sp.). experiments was to expose fish prior to spawning, an acclimation period of only 24 h was allowed. Previous experience with ripe herring has shown that they usually spawn shortly after capture. Fish were initially placed in circular tanks with double sand-filtered, open flow seawater at ambient conditions in the bay at the time. Initial handling mortality was negligible. During the experiment, conditions were as follows: salinity, 23.0-24.0%o; temperature, 10.0°-11.5°C; oxygen, 6.0-10.5 ppm. An ambient benzene concentration was undetectable at the ppb level. Since herring generally feed poorly when spawning, neither exposed nor control fish were fed during the ex- periment. The exposure treatments were as follows: Control: nl/liter (ppb) benzene; open flow system, no benzene exposure; approximately 100 fish (50 males, 50 females). Exposed: 800 nl/liter (ppb) benzene, open flow system, constant exposure for 48 h; ap- proximately 100 fish (50 males, 50 females). Exposed: 100 nl/liter (ppb) 14 C-labeled benzene; static system, declining exposure, 48 h; 25 females only; linear decrease in benzene concentration to approximately 10% of initial concentration remaining at end of 48 h. All benzene exposures were terminated and open flow reestablished in the 100 ppb static exposure tank at the end of 48 h. The static ex- posure of 14 C-labeled benzene was to determine the uptake, accumulation, and depuration of benzene in the gonads of females. The open flow constant exposure and control were primarily to establish morphological and mortality effects on the ovarian eggs and delayed effects on sub- sequent larval development and mortality. The behavior of fish was observed before sampling. Subsamples of females were taken daily for 6 days — 2 days during exposure and 4 days after. Fish were removed randomly until 10 females were obtained from the control and 800 ppb exposure conditions. Five females were removed daily from the static 100 ppb exposure. Concentrations of benzene in the water of all tanks were also measured daily. Each female sampled was measured (standard length), weighed (wet weight), and the ovaries dissected out. The ovaries were also measured (total length) and weighed (wet weight); the left ovaries were examined microscopically, the right 44 STRUHSAKER: EFFECTS OF BENZENE ON SPAWNING HERRING ovaries prepared for radiometric or gas chromatograph analyses. Methods of preparation for radiometric and chromatograph measure- ments are described elsewhere (Benville and Korn 1974; Korn, Hirsch, and Struhsaker 1976, see footnote 2). It should be emphasized that the radiometric technique measures total radioactiv- ity and concentrations calculated may include metabolites of benzene as well as benzene itself. Ovaries were examined under a dissecting microscope for developmental stage [Hjort's stage (Bowers and Holliday 1961)] and the presence of opaque dead or dying eggs, and the gross ap- pearance (color and degree of deliquescence) was ranked. The maximum diameters of 10 eggs from the ovary of each female were measured and the eggs examined for abnormal development. On day 3, after cessation of exposure, pieces of clean plastic screening were placed around the standpipe in the center of the 800 ppb and 100 ppb exposure and control tanks to provide substrate for spawned eggs. Males were placed with females in the 100 ppb tank. After spawning occurred, the screens were removed and eggs examined for developmental stage and mortality. Pieces of screen with 75 eggs on each (most in 4-cell stage) were cut apart. Pieces of screen were selected with a single layer of relatively separated eggs because previous experience showed reduced survival in dense egg clusters. Two pieces of screen with 75 eggs each were placed in each 8-liter rearing container (total of 150 eggs). There were five rep- licate containers for each treatment and control (total of 15 containers). Temperature during development was 11.0°-12.0°C, and salinity, 22.0%o. Other rearing conditions were as pre- viously described (Struhsaker et al. 1974). Hatching occurred 10 days after fertilization, and percent survival at hatching was determined from three replicate counts of swimming larvae in each container and by counting the number of dead and abnormal embryos left on the screen. The screens were removed and surviving larvae fed the rotifer, Brachionus plicatilis, through the remainder of the experiment (past yolk absorption to larval day 7). Surviving larvae were counted and the percent survival through yolk absorption determined from the original egg number. Data were analyzed, depending upon variables, with the methods of analysis of variance and covariance using University of California Biomedical programs, BMD 01V, 02V, and 03V (Dixon 1970). RESULTS No adult mortalities occurred during the 6 days of the experiment. Stress behavior was noted in exposed fish, particularly at the constant 800 ppb exposure. Definite distress was observed by the end of the first day, although oxygen levels were above saturation. Milling was disrupted, fish were gaping at the surface, and many exhibited dis- equilibrium. Even after cessation of exposure, stress behavior continued for the duration of the experiment. Control fish may also have been stressed by the capture conditions and the short acclimation period, but they exhibited none of the stress symptoms of exposed fish and milled normally. Although behavior was abnormal in exposed fish, spawning occurred in the tanks. In fact, the stress from benzene exposure appeared to pre- maturely induce spawning. This is illustrated in Table 1 by the percentage of exposed fish which were spent (Stage VII) compared with control fish. At the end of the 6-day experimental period, 73% (100 ppb) and 70% (800 ppb) of the exposed fish were spent, compared with only 25% of the con- trols. The higher percentage of spent females in the 100 ppb static treatment than in the 800 ppb open flow treatment during the first 4 days may be a result of additional stress imposed by static conditions. At all treatments, most unspent ovaries were ripe (Stage VI); only 7-10% were immature (Stages III-V) (Table 1). No changes in growth (as indicated by wet weight and length) were expected in females over the short experimental period. However, these measurements were taken to determine the similarity of fish between the treatments and to adjust effect of size on the differences in weights of ovaries between the treatments. Ovary length and weight and egg diameters were measured to determine if benzene uptake affected the growth or resorption of ovaries or eggs and to determine the ripeness or proximity to spawning. Data are summarized in Table 2. Egg diameter did not correlate with any other measurement variable. Analysis of variance revealed no significant difference (P>0.25) in egg diameter between and 800 ppb benzene treatments. Since the size range of females varied somewhat between the two treatments (Table 2), analysis of covariance was used to compare the weights of females and ovaries between concentrations and days after adjustment for the effect of lengths (Table 3). No 45 FISHERY BULLETIN: VOL 75, NO. I Table l- -Effects of benzene exposure on ovaries and eggs of Pacific herring. Benzene concentration (nl/l; ppb) No. of ovaries examined Percent o f eggs in stage' No. of ripe ovaries examined Stages dead < lll-VI Hours lll-V Immature VI Ripe VII Spent >ggs (Days) No. % 24 10 10 80 10 9 d) 100 5 40 60 2 800 9 40 49 11 8 48 10 90 10 9 (2) 100 5 20 80 1 1 100 800 10 10 60 30 7 1 14 72 10 20 70 10 9 (3) 100 5 20 40 40 2 2 100 800 9 56 44 6 6 100 96 10 70 30 7 1 14 (4) 100 5 20 20 60 1 1 100 800 10 10 57 33 6 6 100 120 10 10 40 50 5 (5) 100 5 100 0-AII spent — — 800 9 100 0-AII spent — — 148 10 60 40 6 (6) 100 5 100 0-AII spent — — 800 10 100 0-AII spent — — Totals 60 7 68 25 36 1 3 (6 days) 100 30 7 20 73 6 4 67 800 57 10 20 70 24 13 54 'Hjort's stage; Bowers and Holliday (1961). TABLE 2. — Mean and range of female standard length, wet weight; ovary length and wet weight; and maximum egg diameter for Pacific herring. Linear equation describes the regression of wet weights on lengths for both whole female fish and left ovaries. Sample size = 59 females; 59 ovaries (spent females excluded). Female Standard length (X) Benzene concentration Range Mean (ppb) (cm) (cm) Ovary (Stages lll-VI) Wet weight (Y) Range (g) Total length (X) Wet weight (/) Max egg diameter Mean (g) Range (cm) Mean (cm) Range (g) Mean (g) Range (mm) Mean (mm) 16.8-22.4 19.3 76.8-239.6 136.8 7.7-11.5 10.4 6.7-30.8 18.2 1.20-1.50 1.30 800 16.4-21.5 18.7 75.3-189.6 120.3 7.5-14.3 9.3 6.3-26.5 13.6 1 20-1.56 1.30 Total 16.4-22.4 19.0 75.3-239.6 126 2 7.5-14.3 9.9 6.3-30.8 15.9 1.20-1.56 1.30 Regressions' Y = -339 96 ^24 98X Y = -19.26 + 3.56X 800 Y = -267.50 • 20 89X Y = -12.84+2.90X 'Tests of significance between slopes (to) and elevations (a) of regressions showed no significant difference (0.1000250 P>0.250 P>0.250 NS 2 NS NS Analysis of dependent variable (wet wt ovary) after adjustment for covanate (wet wt female) Source of variation Between concentrations (C) (0 vs. 800 ppb) Between days (D) Interaction (CD) Within cells df SS MS F ratio Probability 1 3 3 31 6940 2.5351 19.4057 165.5181 06940 8450 64686 53393 0.13 0.16 1.21 P>0250 P>0.250 P~>0.250 NS NS NS Analysis of dependent variable (wet wt ovary) after adjustment for covanate (total length ovary) S ource of va riation df SS MS F ratio Probability Between concentrations (C) (0 vs 800 ppb) Between days (D) Interaction (CD) Within cells 'F 0.05=4.16, df = 1,31;F 0.05=2.91, df=3,31. 2 NS = not significant 1 04585 04585 004 P>0 250 NS 3 27.2532 9.0844 0.71 P>0.250 NS 3 8.0860 2.6953 021 P>0 250 NS 31 3982616 12.8471 46 STRl'HSAKKR: EFFECTS OF BENZENE ON SPAWNING HERRING significant difference (P>0.25) between con- centrations or days or interaction was found. Tests between slopes (b) and elevations (a) of the re- gression lines of weights on lengths of females and weights on lengths of ovaries (Snedecor and Cochran 1967:432-436) showed no significant differences (P>0.10) between and 800 ppb concentrations (Table 2). Microscopic examination of the ovaries, however, revealed the presence of dead eggs in ovaries of exposed fish by the second day of expo- sure (Table 1). No dead eggs were found in control fish until day 4, and then only a few (15-20 eggs) in one female, the rest of the ovary appearing nor- mal. Ovaries of exposed fish contained sig- nificantly larger numbers of opaque dead eggs (more than 10%) and were generally paler yellow and deliquescent. By the end of 6 days, 67% (100 ppb) and 54% (800 ppb) of exposed females were found with ovaries containing dead or dying eggs. The uptake and depuration of benzene in ovaries of females exposed to a static initial concentration of 100 nl/liter (ppb) 14 C-labeled benzene is shown in Figure 1, together with data determined from other larval studies for later stages (Eldridge, Struhsaker, and Echeverria 5 ). Uptake was rapid, so that a maximum accumu- lation (1.4 /u.l/g; ppm) was reached in 24 h. This level was maintained through the 48-h exposure period. After open flow was reestablished and exposure ended, benzene and/or metabolites were depurated until they reached an undetectable level in 96 h. The figure shows that levels ac- cumulated in ovarian eggs were higher and sustained longer than in later egg and larval stages from other experiments with comparable exposure conditions. Results of rearing experiments with eggs from females exposed to and 800 ppb unlabeled benzene are summarized in Tables 4 and 5. Survival was also reduced in eggs and larvae from females exposed to an initial concentration of 100 ppb labeled benzene. However, results were obscured by an additional variable. Eggs taken from the static exposure tank were covered by filamentous bacterial growth early in develop- ment and many eggs died as a result. In the other treatment with open flow and in controls, eggs did not undergo this mortality due to epifloral growth. i.o 09 Q- 8 Cl ^ 7 3. ~ 06 oj 6 5 05 o 0) to 3 04 Cl a. O 03 O E o f 0.2 Q. 3 c (V CD Eggs In Ovary 6 12 18 24 48 Time (h) 72 96 s Eldridge, M. B., J. W. Struhsaker, and T. Echeverria. Manuscr. in prep. The uptake, accumulation and depuration of 14 C-labeled benzene in embryos and larvae of Pacific herring (Clupea harengus pallasi). FIGURE 1. — Accumulation of 14 C-labeled benzene in different early Pacific herring developmental stages exposed to an initial concentration of 100 nl/liter (ppb) in a static system. Concentra- tions shown on y-axis were calculated from total radioactivity and may include metabolites derived from benzene as well as benzene. Spawned eggs were in a stage just prior to blastopore closure; post yolk-sac larvae were fed the rotifer, Brachionus plicatilis, containing high accumulated levels of labeled ben- zene. ND = not detectable. The 100 ppb treatment, therefore, was not in- cluded in the analysis. Analysis of variance showed survival at hatching and survival of lar- vae through yolk absorption were significantly less in exposed eggs (800 ppb) than in control eggs (P<0.1; Table 5). Exposure to ppb benzene levels for only 48 h reduced survival by about 43% through yolk absorption to larval day 7 (Table 4). DISCUSSION When female herring were briefly exposed to low levels of benzene for 48 h just prior to spawning, a significant reduction occurred in survival of eggs and resultant larvae from the ovary through yolk absorption. It is probable that further mortality would have occurred in later larval stages if the experiments were continued. When this result is compared with that from exposing other life history stages after spawning (Struhsaker et al. 1974; Eldridge et al. see footnote 5) where survival is not affected except at ppm levels, it appears that the spawning female and ovarian eggs are the most sensitive stages. 47 FISHERY BULLETIN: VOL. 75, NO. 1 TABLE 4. — Mean percent survival through hatching and yolk absorption of Pacific herring larvae from eggs of benzene-exposed and control females. Stage Benzene concentration (nl/l; ppb) Total no. of eggs Mean survival (%) 95% confidence Interval (%) Mean reduction survival 1 (%) Embryos to hatching Hatched larvae through yolk absorption 800 800 750 750 750 750 92.9 666 76.7 34.4 91.5-94.3 64.1-69.1 74.5-78.9 32.0-36.8 -26.3 -43.3 1 See Table 5 for test of significance. TABLE 5. — One-way analysis of variance in survival of Pacific herring embryos to hatching and larvae through yolk absorption (larval day 7). Ripe females exposed prior to spawning. Five replicate containers per treatment; 150 eggs/container. ( Arcsin transformation applied to percent survival data.) Percent survival to hatching Source of variation df SS MS F ratio Probability Between concentrations vs. 800 ppb Within groups Total 12 14 1 .3442 0.0843 1 .4285 06721 0070 95.6098- P<0.01 Percent survival through yolk absorption Source of variation df SS MS F ratio Probability Between concentrations vs. 800 ppb Within groups Total 2 12 14 0.8053 0.1599 0.9652 0.4026 0.0133 30.2147* P<0.05 Although male herring were not studied in de- tail here, their behavior was severely disrupted, as in the females. Testes of mature, spawning her- ring have been found to contain higher levels of cholesterol (a lipid) during spawning than at other times in their adult life (Blaxter and Holliday 1963), and it is possible the lipid-soluble benzene may accumulate to high levels in testes of ripe males. Effects on males and their spermatozoa, as well as effects on females, may have contributed to reduction in survival of fertilized eggs through yolk absorption in these experiments. Reference to Figure 1 shows that the maximum accumulation of labeled benzene in ovarian eggs was greater than in later egg and larval stages as measured in other experiments. Accumulation in ovarian eggs of exposed females was approxi- mately twice that in eggs exposed just after spawning and prior to blastopore closure and about six times that in embryos exposed just after yolk-sac absorption. Accumulation for the first 48 h of water column exposure in these stages ap- pears to correlate with the yolk volume of the eggs and larvae, decreasing as yolk is utilized, as would be expected with lipid-soluble benzene. However, the decreased accumulation may also relate to the development of enzymes enabling later stages to metabolize benzene and subsequently depurate more rapidly. After being fed Brachionus plicatilis, which accumulate high levels of benzene (Echeverria 6 ), the fish larvae rapidly accumulated benzene from their food (Figure 1). Other studies of accumulation in tissues of adult herring (Korn et al. see footnote 2) show that only one site, the gall bladder with bile, accumulates higher con- centrations than ovarian eggs (30 times and 14 times initial concentration, respectively). I have noted previously (Struhsaker et al. 1974) that the percentage survival of eggs through hatching is significantly less (approximately 25% less;P<0.01) in Pacific herring eggs collected from San Francisco Bay than in those from Tomales Bay. Although other environmental differences may be involved, this reduction in hatching suc- cess may well relate to the effects of accumulated pollutants in the gonads of spawning fish in the relatively more polluted San Francisco Bay wa- ters and warrants further study. Estimating that the reduction in survival of eggs through yolk absorption of spawning exposed females is at least 43%, the effect on Pacific her- ring populations exposed to only one toxic component of petroleum could be significant. Considering that the total water-soluble fraction contains many other toxic aromatics, it is possible "Echeverria, T. Manuscr. in prep. Uptake and depuration of 14 C benzene in the rotifer, Brachionus plicatilus. 48 STRUHSAKER: EFFECTS OF BENZENE ON SPAWNING HERRING that long-term chronic exposures to low levels may be decreasing population survival in polluted areas. In addition, chlorinated hydrocarbons in pesticides may also be accumulating in the gonadal lipids and interacting with petroleum hydrocarbons producing even more deleterious effects. More studies of the effects of these components on spawning fish are clearly needed. If fishes prove generally to be the most sensitive to accumulated oil components during their spawn- ing season, fisheries management decisions should take into consideration their protection from damaging levels, particularly at spawning time. ACKNOWLEDGMENTS I thank the staff of the Physiology Program, SWFC Tiburon Laboratory, for assisting me in these experiments. I am grateful to Norman Abramson and Vance E. McClure for reviewing the manuscript and for making suggestions. Dale Straughan, Institute of Marine and Coastal Studies, University of Southern California, also reviewed the manuscript and made several improvements. LITERATURE CITED ANDERSON, J. W., J. M. NEFF, B. A. COX, H. E. TATEM, AND G. M. HIGHTOWER. 1974. Characteristics of dispersions and water-soluble extracts of crude and refined oils and their toxicity to estuarine crustaceans and fish. Mar. Biol. (Berl.) 27:75-88. BENVILLE, P. E., JR.. AND S. KORN. 1974. A simple apparatus for metering volatile liquids into water. J. Fish. Res. Board Can. 31:367-368. BLAXTER, J. H. S., AND F. G. T. HOLLIDAY. 1963. The behavior and physiology of herring and other clupeids. Adv. Mar. Biol. 1:261-393. BOWERS, A. B., AND F. G. T. HOLLIDAY. 1961. Histological changes in the gonad associated with the reproductive cycle of the herring (Clupea harengus L.). Dep. Agric. Fish. Scotl., Mar. Res. 1961(5), 16 p. DIXON, W. J. (editorl 1970. Biomedical computer programs. Univ. Calif. Press, Berkeley, 600 p. EVANS, D. R., AND S. D. RICE. 1974. Effects of oil on marine ecosystems: A review for administrators and policy makers. Fish. Bull., U.S. 72:625-638. KORN, S., N. HIRSCH, AND J. W. STRUHSAKER. 1976. Uptake, distribution, and depuration of 14 C-benzene in northern anchovy, Engraulis mordax, and striped bass, Morone saxatilis. Fish. Bull., U.S. 74:545-551. KORN, S., J. W. STRUHSAKER, AND P. BENVILLE, JR. 1976. Effects of benzene on growth, fat content, and caloric content of striped bass, Morone saxatilis. Fish. Bull., U.S. 74:694-698. KUHNHOLD, W. W. 1969. Der Einfluss wasserloslicher Bestandteile von Roholen und Rohblfraktionen auf die Entwicklung von Heringsbrut. [Engl, abstr.] Ber. Dtsch. Wiss. Komm. Meeresforsch., Neue Folge 20:165-171. 1972. The influence of crude oils on fish fry. In M. Ruivo (editor), Marine pollution and sea life, p. 315-318. Fishing News (Books) Ltd., Surrey, Engl. LEE, R. F., R. SAUERHEBER, AND G. H. DOBBS. 1972. Uptake, metabolism and discharge of polycyclic aromatic hydrocarbons by marine fish. Mar. Biol. (Berl.) 17:201-208. NEFF, J. M. 1975. Accumulation and release of petroleum-derived aromatic hydrocarbons by marine animals. Symposium on chemistry, occurrence, and measurement of polynuc- lear aromatic hydrocarbons, p. 839-849. Div. Pet. Chem., Inc., Am. Chem. Soc. Chicago Meeting, 1975. SNEDECOR, G. W., AND W. G. COCHRAN. 1967. Statistical methods. 6th ed. Iowa State Univ. Press, Ames, 593 p. STRUHSAKER, J. W., M. B. ELDRIDGE, AND T. ECHEVERRIA. 1974. Effects of benzene (a water-soluble component of crude oil) on eggs and larvae of Pacific herring and north- ern anchovy. In F. J. Vernberg and W. B. Vernberg (editors), Pollution and physiology of marine organisms, p. 253-284. Academic Press Inc., N.Y. 49 BIOLOGY OF THE REX SOLE, GLYPTOCEPHALUS ZACHIRUS, IN WATERS OFF OREGON Michael J. Hosie 1 and Howard F. Horton 2 ABSTRACT Data are presented on the life history and population dynamics of rex sole, Glyptocephalus zaehirus Lockington, collected from Oregon waters between September 1969 and October 1973. Length-weight relationships vary little between sexes or with time of year. Otolith annuli form primarily from January through May and were used for age determination. Age and length are highly correlated (r = 0.9945 for males and 0.9864 for females), with females growing faster and living longer than males. Estimates of total instantaneous mortality rate (Z) appear less variable when calculated by the catch-curve method (mean Z of 0.64 for males and 0.51 for females), than by the Jackson method. Age at 50% maturity occurs at 1 6 cm ( about 3 yr ) for males and at 24 cm (about 5 yr ) for females. Spawning off northern Oregon occurs from January through June, with a peak in March-April. Fecundity is correlated (r = 0.9620) with length offish. There were 15 recaptures (0.59% ) from 2,537 fish tagged off northern Oregon during March and June 1970. Maximum movement of recaptured fish was only 53.9 km, but the low recovery precludes definite conclusions. Twenty loci were detected by starch-gel electrophoretic analysis using rex sole muscle tissue. Of these, three loci were polymorphic, but showed no discernible variation between collections from northern, central, and southern Oregon in April 1973. Investigation into the life history of rex sole, Glyptocephalus zaehirus Lockington, by the Ore- gon Department of Fish and Wildlife provided new information on this species. The broad objective was to develop knowledge of the biology and population dynamics of rex sole found off the Oregon coast which would enhance management of this species. Specific objectives were to: 1) determine the length-weight and age-length relationships; 2) estimate the total instantaneous mortality rate by two independent methods; 3) determine rela- tionships of maturity and fecundity with length and age, and with the spawning season; and 4) determine if rex sole off Oregon are composed of separate stocks 3 which undergo predictable movements. The rex sole is a slender, thin flatfish belonging to the family Pleuronectidae (Starks 1918; Nor- man 1934), the right-eyed flounders. Of the three species of Glyptocephalus , rex sole is the only one reported in the eastern Pacific Ocean (Pertseva- Ostroumova 1961). Geographically distributed 'Oregon Department of Fish and Wildlife, Marine Field Laboratory, P.O. Box 5430, Charleston, OR 97420. 2 Department of Fisheries and Wildlife, Oregon State Uni- versity, Corvallis, OR 97331. 3 The rex sole spawning in a particular marine location (or portion of it) at a particular season, and which do not interbreed to a substantial degree with any group spawning in a different place, or in the same place at a different season (modified from Ricker 1972). from southern California to the Bering Sea (Miller and Lea 1972), it is found bathymetrically to 730 m (Alverson et al. 1964). Rex sole is important in the commercial trawl fishery from California northward through British Columbia. In 1972, rex sole was the fifth most important flatfish in weight (1.54 million kg [3.4 million pounds]) in the domestic northeastern Pacific trawl food fishery. Glyptocephalus zaehirus is also important in the domestic trawl fishery for animal food (Best 1961; Niska 1969), although this fishery has declined in recent years. On the continental shelf off the northern three-fourths of the Oregon coast, rex sole was third in biomass 4 and first in numbers of all flatfish caught with an 89-mm (3.5-inch) mesh trawl. There is little published information on the biology of rex sole. Villadolid (1927) and Frey (1971) reported briefly on the time of spawning, size and age at maturity, and food habits for specimens captured off California. Hart (1973) summarized the life history of rex sole off Canada and suggested that the lack of information re- sulted in doubtful deductions. An aging study was conducted on rex sole by Villadolid (1927) who used scales. Domenowske (1966) used otoliths, Manuscript accepted August 1976. FISHERY BULLETIN: VOL. 75. NO. 1. 1977. 4 Demory, R. L., and J. G. Robinson. 1973. Resource surveys on the continental shelf of Oregon. Fish Comm. Oreg. t Commer. Fish. Res. Dev. Act Prog. Rep., July 1, 1972 to June 30, 1973, 19 p. (Unpubl. manuscr.). 51 FISHERY BULLETIN: VOL. 75, NO. 1 scales, and interopercles for aging rex sole; by comparing the age-length relationships, he concluded otoliths were the most readable structure. Vanderploeg (1973) conducted food habit studies on rex sole collected off Oregon. Porter (1964) described the larvae of rex sole, and Waldron (1972) and Richardson (1973) reported on distribution and abundance of rex sole larvae. Tsuyuki et al. (1965) conducted a general starch- gel electrophoresis study on the muscle proteins and hemoglobin of 50 species of North Pacific fish and found that rex sole differed from 10 other pleuronectids tested. Benthic distribution of rex sole was investigated by numerous workers 4 (Alverson et al. 1964; Day and Pearcy 1968; Demory 1971; Alton 1972). Limited tagging studies (Manzer 1952; Harry 1956) were con- ducted to determine movements of rex sole, but no tagged fish were recaptured. METHODS Rex sole were collected by otter trawl off Oregon from the Columbia River south to Cape Blanco at depths of 18-200 m during September 1969-73. Most data were obtained from rex sole captured incidentally to a study of pink shrimp, Pandalus jordani, distribution during 1969-70. 5 Rex sole were also obtained from commercial trawl land- ings at Astoria, Oreg., in 1970 and 1973; at Charleston and Brookings, Oreg., in 1973; and from research vessel catches during the 1971-73 Fish Commission of Oregon (FCO) groundfish surveys. 4 6 All specimens were frozen until time of examination. Rex sole were sexed by examination of gonads, measured for total length (TL) to the nearest centimeter, and weighed to the nearest gram. The left otolith was removed for aging studies, stored in a 50:50 solution of glycerin and water, and read using reflected light on a dark background ( Powles and Kennedy 1967). The length-weight relationship, by calendar quarters, of rex sole collected off central and northern Oregon in 1969-72 was determined by the least squares method using the logarithmic 5 Lukas, G., and M. J. Hosie. 1973. Investigation of the abundance and benthic distribution of pink shrimp, Pandalus jordani, off the northern Oregon coast. Fish Comm. Oreg., Commer. Fish. Res. Dev. Act, Final Rep., July 1, 1969 to June 30, 1970, 45 p. (Unpubl. manuscr.). 6 Demory, R.L. 1974. Resource surveys on the continental shelf of Oregon. Fish Comm. Oreg., Commer. Fish. Res. Dev. Act Prog. Rep., July 1, 1973 to June 30, 1974, 6 p. (Unpubl. manuscr). form of the equation W =aL b , where W is weight in grams, L is length in centimeters, and a and b are constants. Estimates of the lineal growth of rex sole were obtained from the age-length relationship of fish collected off northern Oregon in September- October 1969 and September 1971. A mean total length (TL) at each age was determined from these data and expressed mathematically in terms of the von Bertalanffy growth equation (Ricker 1958; Ketchen and Forrester 1966). To obtain the calculated growth parameters, we used ages 1.5-10.5 yr for males and 1.5-12.5 yr for females. Estimates of the instantaneous total mortality rate (Z) were made using age group data obtained from FCO groundfish cruises off northern Oregon in 1971 and 1973 and off central Oregon in 1972. Two methods, a catch curve (Ricker 1958) and the Jackson technique (Jackson 1939), were used for the analyses. To determine maturity stages, gonads were examined according to the procedures described by Hagerman (1952), Scott (1954), and Powles (1965). Definitions used for maturity stages are listed in Table 1. Fecundity was determined from 13 fish collected in February 1970 and measured to the nearest millimeter (TL). Both ovaries were removed from TABLE 1. — Description of reproductive phases of rex sole gonads used in this study. Sex Maturity stage Description Females Immature (A): Ovaries very small (<40 mm TL), whitish in color, semitransparent, and gelatinous. No eggs dis- cernible to the naked eye. Mature (B): Ripening. Ovaries enlarging, becoming reddish- orange colored and granular in consistency, full of developing eggs that can be recognized by direct observation. (C): Ripe. Ovaries full of mostly reddish-orange colored granular eggs, although a few transparent ova are present. Ova can be extruded from the fish by using considerable pressure. (D): Spawning Ovaries full of entirely translucent eggs which will run with slight pressure. (E): Spent. Ovaries flaccid, usually empty although occasionally a few eggs will remain. Ovarian membrane very transparent and saclike. (F): Recovering. Ovaries filling with fluid, and reddish- orange in color. No ova detectable to the naked eye. Males Immature (A): Testes very small (<3 mm TL), translucent in color and not extending into the abdominal cavity. Mature (B): Ripening Testes enlarged, extending posteriorly into abdominal cavity, light brown to cream colored, but retain sperm under pressure. (C): Ripe and/or spawning. Testes full and cream colored. Sperm will run under no or only slight pressure. (D): Spent-recovering. Testes shrunken and trans- parent or dark brown in color. 52 HOSIE and HORTON: BIOLOGY OF REX SOLE each fish and stored in 10% Formalin. 7 Estimates of fecundity were obtained gravimetrically, following the method described by Harry (1959). To obtain fish for tagging, short tows of about 15 min were made in March and June 1970 off northern Oregon near the mouth of the Columbia River. Any rex sole caught were held for 15-60 min in a tank containing running seawater. Fish in good condition were tagged and released. Petersen disc (vinyl) tags, 16 mm in diameter, were at- tached by a stainless steel pin inserted through the musculature about Vz inch below the midbase of the dorsal fin. Fishermen were advised of the tagging program, and a $0.75 reward was offered by the FCO for each tagged rex sole returned. Electrophoresis was used to investigate stock identification of rex sole. A preliminary electro- phoretic examination was conducted using muscle tissue of 145 rex sole collected in April 1973 in three nearly equal samples taken off northern, central, and southern Oregon. Tissue extraction and starch gel electrophoresis procedures followed the methods of Johnson et al. (1972). Tests were conducted for polymorphisms in muscle protein and the five enzyme systems: aspartate aminotransferase (AAT) A-I and A-II; lactic dehydrogenase (LDH); peptidase A-I and A-II; phosphoglucomutase (PGM); and tetrazolium oxidase (TO). RESULTS AND DISCUSSION Length-Weight Relationships Length and weight were closely correlated, with 7 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. the derived coefficient of determination (r 2 ) vary- ing from 0.9902 to 0.9988 for males and from 0.9872 to 0.9966 for females (Table 2). These coefficients of determination varied little by season, possibly because of the extended spawning period (Villadolid 1927) in the first half of the year. Based on data in Table 2, we calculated mean weights by season at representative lengths. For both sexes weight increase was greatest in the third quarter, average in the second quarter, and slowest in the first and fourth quarters (Table 3). Among mature fish, about 30 cm TL and larger, females generally were slightly heavier than males of the same length (Figure 1). A total of 950 males and 1,121 females were included in the length-weight data analyzed. Age and Growth Validity of the Aging Technique Opaque or hyaline zones occur on the margin of rex sole otoliths. These zones mark the respective periods of rapid or slow growth. Examination of 265 otoliths from rex sole <27 cm TL collected off northern Oregon from September 1969 through July 1970 revealed that hyaline edges were first observed in September (Figure 2). No hyaline edges were present the previous June or July. In the fall the percentage of otoliths with a hyaline zone on their edge began to increase. By January the majority of otoliths had a hyaline zone on their edge. The percentage rapidly increased and peaked in March when 92.3% had hyaline zone margins. Conversely, opaque zones on edges were at their lowest in March, gradually increasing until June or July when all otoliths had opaque edges. TABLE 2. — Length-weight relationship (log 10 W = log 10 a + b logL) by quarterly period for male and female rex sole collected off central and northern Oregon, 1969-72. ' Period and sex Number of fish Constant log a Constant b Standard deviation Correlation coefficient Coefficient of determination January-March: Male 119 -3.1447 3.5551 0.1437 0.9972 0.9944 Female 68 -3.0978 3.5095 0.1587 0.9936 0.9872 Both 187 -3.1248 3.5258 0.1539 0.9932 0.9864 April-June: Male 386 -2.8398 3.3557 0.1501 0.9994 0.9988 Female 356 -2.9398 3.4345 0.1488 0.9980 0.9960 Both 742 -2.8903 3.3914 0.1567 0.9984 0.9968 July-September: Male 350 -3.0884 3.5598 0.1461 0.9982 09964 Female 621 -2.9886 3.5112 0.1661 0.9983 0.9966 Both 971 -3.0631 3.5553 0.1788 0.9988 0.9976 October-December: Male 95 -2.9823 3.4423 0.1269 0.9951 0.9902 Female 76 -2.9795 3.4423 0.1599 0.9972 0.9944 Both 171 -2.9500 3.4252 0.1562 0.9973 0.9946 'Regression analysis conducted on 1 1- to 36-cm males and 1 1- to 51 -cm females. 53 FISHERY BULLETIN: VOL 75, NO. 1 TABLE 3. — Computed mean weight per quarter at selected lengths of male and female rex sole, using regression formulas from Table 2. Sex Total length (cm) Computed mean weight (g) per quarter 1 I II III IV Male Female 15 25 35 15 25 35 45 11 67 221 11 64 210 506 13 71 220 13 73 231 547 13 77 256 14 83 271 655 12 68 215 12 68 231 514 'I = Jan-Mar.; II = Apr-June; III = July-Sept.; IV = Oct -Dec HOOr KX>0 900 o»800 £ 700 ^ 600 o 500 O m z 400 < Id 2 300 • MALES (N=950) O FEMALES (N=II2I) ^oo •cP 200 I00 d? .CP d?" Jp _l_ 10 15 20 25 30 35 TOTAL LENGTH (Cm) 40 45 50 FIGURE 1. — Length-weight relationship for male and female rex sole collected off central and northern Oregon, 1969-72. Body weights obtained from an average of quarterly mean values. From these observations, we concluded that the hyaline margin is deposited on otoliths during each winter and spring for all sizes of rex sole. Thus, these hyaline zones are interpreted as an- nuli with a year's growth occurring between successive hyaline margins. These results are similar to those of Villadolid (1927) who found northern California rex sole formed a scale an- nulus in March through May. Age-Length Relationship After 3.5 yr of age, females were consistently longer than males at a given age. Females also attained an older age and longer length. Statistics for both males and females followed the von Bertalanffy growth curve, as a good fit was ob- tained for most age groups (Figure 3, Table 4). I00r 90 80- ~ 70 w 60 >- o z 50 LU uj 40 or (13) (20) (18) 30 20 10- (93) (28) (32) S N D 1969 (12) (13) (20) (16) MONTH M A M J J A 1970 FIGURE 2. — Percent frequency of hyaline edges found on otoliths of 265 rex sole (<27 cm TL) collected off northern Oregon, September 1969-July 1970. Numbers in parentheses represent sample size. 40 r 30 20 I 10 "-* I H z UJ _l 50 _1 < 40 30 20 10 MALES I, =33.43 ( N = 257 ) [h -0.1778 (t-0.8551) 1749 (t -0 5667)1 6 8 10 12 age (yr) 14 FIGURE 3. — Age-length relationship for male and female rex sole collected off northern Oregon, September-October 1969 and September 1971. The calculated length at infinity (Loo) of 33.43 cm for males was close to the computed mean value of 29.33 cm (Table 4). For females theL^ of 54 HOSIE and HORTON: BIOLOGY OF REX SOLE TABLE 4. — Computed mean length at age and mean length at age estimated by von Bertalanffy growth equation for 45 unsexed, 189 male, and 212 female rex sole collected off northern Oregon in September-October 1969 and September 1971. Age' (yr) No. 1.5 3 45 2.5 13 3.5 36 4.5 29 5.5 15 6.5 17 7.5 23 8.5 23 9.5 16 10.5 10 11.5 6 12.5 1 13.5 14.5 15.5 16.5 17.5 18.5 Male Computed mean Estimated mean length (cm) length 2 (cm) 9.20 12.61 17 00 19.52 21 66 24.55 2539 2582 27.37 28.90 29.33 27.00 9.44 13.36 16.65 19.39 21 69 23.62 25.22 26.57 27.69 28 63 2942 30.07 No 3 45 7 33 11 19 14 9 17 24 28 20 14 4 2 6 1 3 Female Computed mean Estimated mean length (cm) length 2 (cm) 920 12.71 16.64 20.45 24.95 25.64 26.33 28.05 3037 31.03 33.35 3245 33.75 33.50 37.00 47.00 0.00 47.30 891 13.44 17.25 2045 23.14 25.39 27.29 28 88 30.21 31 34 32.28 33.07 33 73 34 29 34.76 'These fall-caught fish were assumed to be about one-half way through the growing season, based upon otolith readings. 2 Von Bertalanffy growth equations were based on 1- to 10-yr-old males (L a = 33.43 cm, K -0 8551 yr), and 1- to 12-yr-old females (L x = 37.21 cm, K = 0.1747, ( = -0.5667 yr) 3 Sexes were not separated for age 1 fish (45 specimens) = 0.1778, t n 37.21 cm fit observed data through age 15.5, but was far below the maximum computed mean TL of 47.30 cm. The apparent discrepancy does not in- validate the data because Knight (1968) noted that L x is not the maximum obtainable length, but rather a mathematical tool needed in compu- tations for the von Bertalanffy growth equation. This is exemplified by our collection of a 23-yr-old ( ± 1 yr), 59-cm female rex sole off northern Oregon in February 1970, which we consider as about the maximum length and age of rex sole. Hart (1973) reported a 24-yr-old rex sole was collected off British Columbia, but no length was given. Mortality Rate Estimates of the total instantaneous mortality rate (Z) derived from data in Table 5 and using the catch curve method varied from 0.53 to 0.70 for males and from 0.44 to 0.55 for females (Table 6). In this analysis the natural logarithm of the numbers of males and females caught at each age was plotted against the respective age class (Figures 4, 5). The total mortality rate was the best fitted slope on the right side of the catch curve, determined by linear regression using ages rang- ing maximally from 6 to 16 yr (Table 5). Estimates of Z using the Jackson method ranged from 0.43 to 0.61 for males and from 0.20 to 0.52 for females (Table 6). In this method annual survival rate (S) is: TABLE 5. — Numbers of rex sole per age group caught during groundfish surveys off northern Oregon in 1971 and 1973 and central Oregon in 1972. Age Number males Number females (yr) 1971 1972 1973 1971 1972 1973 2 7 14 11 19 26 3 50 68 75 59 70 116 4 67 142 45 102 124 56 5 270 290 337 353 207 514 6 244 663 387 329 732 613 7 375 278 881 418 501 1,217 8 380 412 432 400 560 570 9 215 274 382 366 465 596 10 320 45 106 582 108 201 11 67 123 42 138 283 94 12 76 24 72 247 32 219 13 5 14 11 69 57 30 14 10 2 50 10 26 15 5 7 20 10 16 2 2 7 3 9 18 9 3 21 4 Total 2,093 2,358 2,781 3,149 3,184 4,291 TABLE 6. — Estimates of the total instantaneous mortality rate (Z) of rex sole collected off northern Oregon in September 1971 and 1973 and off central Oregon in September 1972. Age of Catch curve Jackson method Year maximum Ages estimates of estimates of and sex numbers utilized Z Z 1971: Male 8 8-16 0.70 0.43 Female 10 7-16 0.44 0.20 1972: Male 6 6-13 0.53 0.44 Female 6 6-16 0.55 0.31 1973: Male 7 7-13 0.68 0.61 Female 7 7-14 0.54 0.52 Mean: 1 Male 0.64 0.49 Female 0.51 0.34 'Based on simple average of Z's for the 3 yr. 55 FISHERY BULLETIN: VOL. 75, NO. 1 8 6 4 2 c - 8 »- x < o cr UJ CD o 3 8- 6- 4 2r N = I833 r =0.9215 1971 1972 _l I I I I I I I I l_ 1973 • N = I926 r =0.9558 _i i i i i i i i i 8 12 AGE 16 20 FIGURE 4. — Catch curves of male rex sole collected off Oregon in September 1971, 1972, and 1973. I < o in a. UJ m 1971 N = 2297 r =0 9135 j i i i i i i i ' 1972 1973 FIGURE 5. — Catch curves of female rex sole collected off Oregon in September 1971, 1972, and 1973. S = Nt + Ns + ... + Nr Ne + Nl + ... + Nr-l where N is the number of fish of age group r caught. Annual mortality rate is 1 - S and the corresponding instantaneous rate of total mortal- ity is obtained from the expressions = e z , where e and Z are derived from Ricker (1958). The catch curve method probably gives more reliable estimates of Z than those obtained using the Jackson method. In the Jackson method the larger samples of younger fish strongly affect the estimates, with the older age groups weighted less. Thus, the Jackson method substantially underestimates the mean Z for the entire right limb of the catch curve. Reproduction Size at Maturity Some males were mature at 13 cm while no females reached maturity until 19 cm (Figure 6). 30 20 10 I u_ fe 50 cr iu 40 CD 5 | 30 20 10 — MATURE -o IMMATURE A LENGTH AT 50% MATURITY Q LENGTH AT 100% MATURITY VUV^. 10 14 18 22 26 30 34 38 42 46 50 54 58 62 TOTAL LENGTH (Cm) FIGURE 6. — Size composition of immature and mature rex sole, by sex, collected off northern Oregon, September 1969-July 1970. About 50% of the males were mature at 16 cm, and all were mature at 21 cm. For females, 50% were mature at 24 cm and 100% were mature at 30 cm. Lengths at 50% and 100% maturity correspond to 56 HOME and IIORTON: BIOLOGY OF REX SOLE about ages 3 and 5 for males and 5 and 9 for females (Table 4). The only maturity data on rex sole available from other areas is that of Villadolid (1927). He found that both males and females off San Francisco, Calif., were fully mature at age 4, which corresponded to about 21.8 cm for males and 22.8 cm for females. Possibly rex sole mature earlier in the southern portion of their range. Spawning Duration of the spawning period was from January through June, with a peak in March- April (Figure 7). Although samples were not obtained during August and December, the percentage offish in each reproductive phase gives a good indication of the spawning time. The 6-mo spawning period we found is longer than the January through April spawning re- ported by Villadolid ( 1927) for rex sole collected off central California in 1925 and 1926. Paul Reed (FCO, pers. commun.) found a prolonged spawning from January through August for 3,189 rex sole collected off northern California in 1949-54 and 100 50 (20) (77) (16) (64) (60) (37) (84) (33) (55) (50) RIPENING _□_ 2 ioo r UJ =3 o UJ a. u. 50 o UJ RIPE AND SPAWNING n n XL I00r (—1 r^ 50 SPENT AND RECOVERY nil SONDJFMAMJJA 1969 1970 MONTH FIGURE 7. — Annual cycle of reproduction in 496 rex sole (274 males and 222 females) collected off northern Oregon, Sep- tember 1969-July 1970. The number in each monthly sample is shown in parentheses. 1962-63. This suggests the duration of rex sole spawning varies by area and year. Fecundity Examination of 13 mature females ranging from 240 to 590 mm TL yielded fecundity esti- mates of 3,900 and 238,100 ova, respectively. The numbers of ova generally increased with size of the female. In 11 of 13 fish, the right ovary con- tained more ova than the left ( 100 to 12,700 more). A linear regression fitted to the fecundity-length data gave a correlation coefficient of 0.9620 (Fig- ure 8). The formula for the regression line was F = 5.3797 x 10" 7 L 422667 , where F is fecundity in number of ova and L is fish TL in millimeters. 300 200 300 400 500 TOTAL LENGTH (mm) 600 FIGURE 8. — Fecundity-length relationship for 13 rex sole col- lected off northern Oregon, February 1970. Stock Identification Tagging Experiment A total of 2,537 rex sole were tagged and re- leased off the northern Oregon coast in April (200) and June 1970 (2,337). There were 15 recaptures (0.59% recovery) by July 1974, all from the June 1970 tagging (Table 7). Maximum movement was 53.9 km, and 788 days was the longest time at liberty. There was little change in the depth range occupied by recaptured fish, which were released in 42-154 m and recovered by trawls in 51-101 m. These results suggest only limited movement by rex sole. However, tag returns were too few to justify definite conclusions. This low recovery is similar to reports of rex sole tagged off British Columbia (Manzer 1952 [90 tagged]) and Oregon (Harry 1956 [19 tagged]) from which no fish were recovered. 57 FISHERY BULLETIN: VOL 75, NO. 1 TABLE 7. — Release and recovery data on 2,537 rex sole tagged off northern Oregon, April and June 1970. Date Number Number Percent tagged recovered recovery Distance traveled (km) Days at liberty April 1970 June 1970 200 2,337 15 000 0.64 0.0 1.5 17.1 0.0 3.7 23.0 14.1 2.2 8.0 14.3 0.9 38.9 539 unknown 39 523 0.0 4 4 5 18 40 189 240 278 279 294 346 364 374 450 788 Total 2.537 15 0.59 The low returns possibly were caused by rex sole not surviving the tagging process. Manzer (1952) reported rex sole reacted badly to capture and tagging. Most tagged rex sole released at the ocean surface did not immediately descend. Instead, unlike most other flatfish species, they curled into a semicircle and moved across the water surface in a skipping motion. This peculiar reaction might have resulted in a high initial tagging mortality from predation. It may also indicate a stress condition from which fish did not recover. Starch-Gel Electrophoretic Analysis There were 20 loci detected in the muscle tissue of 145 rex sole. Of these loci 13 were enzymes and 7 were muscle proteins (Table 8). Only three of the loci (15%) were polymorphic. The polymorphism was found in only three of the eight systems studied or examined. AAT staining occurred in two anodal regions (A-I and A-II). Zone II was the only polymorphic region, having A, B, C, and D alleles (Figure 9, Table 9). The enzyme peptidase also had two anodal re- ORIGIN O OBSERVED ■i NOT OBSERVED □ CD CD CD CD - CD a CD o □ CD CD CD CD CD CD CD □ 1 1 1 1 1 ZONE I (POLYMORPHIC) ZONE n (MONOMORPHIC) AA AB BB BC CC CD DD AD AC BD AAT PHENOTYPES FIGURE 9. — Diagrammatic representation of aspartate aminotransferase (AAT) phenotypes in starch gel from 145 rex sole collected off Oregon, April 1973. TABLE 9. — Frequencies of aspartate aminotransferase (AAT) phenotypes in 145 rex sole collected off Astoria, Charleston, and Brookings, Oreg., in April 1973. Item Astoria Charleston Brookings Sample size 52 43 50 Date 5, 9 April 30 April 8 April AAT phenotypes: AA 3 8 6 AB 18 3 10 BB 9 10 11 BC 12 12 9 CC 3 2 3 CD 1 DD AD 1 1 AC 4 6 9 BD 1 1 2 Frequency of alleles: A 0.28 0.30 0.31 B 0.47 0.42 0.43 C 0.23 0.26 0.24 D 0.02 0.02 0.02 gions. Only zone II was polymorphic, with A and B alleles (Figure 10, Table 10). A third enzyme, PGM, was polymorphic, having only one locus which had A 1 , A, and B alleles (Figure 11, Table 11). No discernible variation in the frequency or kinds of phenotypes found was observed between rex sole collections from off Astoria (northern), TABLE 8. — Results of electrophoretic tests of muscle tissue samples from 145 rex sole collected off Oregon, April 1973. No. of Proposed Proposed no. Type of bands in no. of of alleles per alleles Phenotypic Protein 1 starch gel loci locus found variation AAT A-I 1 1 — Monomorphic AAT A-II 4 4 A,B,C,D Polymorphic LDH 1 1 — Monomorphic Peptidase A-I 1 1 — Monomorphic Peptidase A-II 2 2 A.B Polymorphic PGM 3 3 A'.A.B Polymorphic TO 1 1 — Monomorphic Muscle proteins 2 7 7 1 — Monomorphic 1 AAT (aspartate aminotransferase); LDH (lactate dehydrogenase); PGM (phospho- glucomutase); TO (tetrazolium oxidase). 2 Analysis of muscle proteins was nonspecific, with 6 anodal ( + ) bands and 1 cathodal ( -) band found. 58 HOSIF and HORTON: BIOLOGY OF REX SOLE FIGURE 10. — Diagrammatic representation of peptidase phenotypes in starch gel from 137 rex sole collected off Oregon, April 1973. TABLE 10. — Frequencies of peptidase anodal zone II phenotypes in 137 rex sole collected off Astoria, Charleston, and Brookings, Oreg., in April 1973. Item Astoria Charleston Brookings Sample size 1 50 43 44 Date 5, 9 April 30 April 8 April Peptidase phenotypes: AA 10 10 13 AB 30 17 22 BB 10 16 9 Frequency of alleles: A 0.50 0.43 0.55 B 0.50 0.57 0.45 'An additional two rex sole from the Astoria sample and six fish from the Brookings sample did not develop distinct patterns and hence are not included. Charleston (central), or Brookings (southern) Oregon (Tables 9-11). These data are insufficient to warrant extended speculation. However, they suggest that geographic selection or variation in rex sole off Oregon, if any, may not revolve around the genetic system included in the eight systems tested. Other alternatives, such as testing ad- ditional genetic systems or possible use of hel- minth parasites as biological tags, should be investigated to provide a more extensive evalua- tion of the population structure of rex sole off Oregon as a possible adjunct to effective man- agement decisions. ACKNOWLEDGMENTS Financial support was provided by the Fish Commission of Oregon (now Oregon Department of Fish and Wildlife [ODFW]). James Meehan, Gerald Lukas, Bill Barss, Edwin Niska, Jack Robinson, Robert Demory, and Brent Forsberg (all ODFW) helped collect and tag rex sole. Paul Reed UJ + UJ + -1 1 i _l UJ _l _l < uj a _i _l < li- Cl u. z 1 7DNF T o o (MONOMORPHIC) z F- o o F- 0. O UJ -, O. > A t- < n CD ZONE n (POLYMORPHIC) UJ > A A ui B □ □ F- B or < _l UJ ORIGIN i 1 1 ac AA AB BB PEPTIDASE PHENOTYPES ORIGIN □ - CD CD CD 1 I C3 1 CD l A'A AA AB PGM PHENOTYPES BB FIGURE ll. — Diagrammatic representation of phospho- glucomutase (PGM) phenotypes in starch gel from 145 rex sole collected off Oregon, April 1973. TABLE 11. — Frequencies of phosphoglucomutase (PGM) phenotypes in 145 rex sole collected off Astoria, Charleston, and Brookings, Oreg., in April 1973. Item Astoria Charleston Brookings Sample size 52 43 50 Date 5. 9 April 30 April 8 April PGM phenotypes: A'A 1 AA 51 42 49 AB 1 BB 1 Frequency of alleles: A 1 000 0.00 0.01 A 0.98 099 0.99 B 0.02 0.01 0.00 (ODFW) provided spawning data on northern California rex sole. Allyn Johnson (National Marine Fisheries Service) conducted the elec- trophoretic analysis. The assistance of Rudy Lovvold of the MV Sunrise, and Thomas Oswald and Olaf Rockness of the RV Commando is ap- preciated. W. G. Pearcy (Oregon State Universi- ty), S. J. Westrheim (Canada Department of the Environment), and Robert Loeffel (ODFW) criticized the manuscript. LITERATURE CITED Alton, M. S. 1972. Characteristics of the demersal fish fauna inhabiting the outer continental shelf and slope off the northern Oregon coast. In A. T. Pruter and D. L. Alverson (editors), The Columbia River estuary and adjacent ocean waters, p. 583-634. Univ. Wash. Press, Seattle. ALVERSON, D. L., A. T. PRUTER, AND L. L. RONHOLT. 1964. A study of demersal fishes and fisheries of the 59 FISHERY BULLETIN: VOL. 75, NO. 1 northeastern Pacific Ocean. H. R. MacMillan Lectures in Fisheries, Univ. B.C., 190 p. BEST, E. A. 1961. The California animal food fishery 1958-1960. Pac. Mar. Fish. Comm., Bull. 5:5-15. DAY, D. S., AND W. G. PEARCY. 1968. Species associations and benthic fishes on the con- tinental shelf and slope off Oregon. J. Fish. Res. Board Can. 25:2665-2675. DEMORY, R. L. 1971. Depth distribution of some small flatfishes off the northern Oregon-southern Washington coast. Fish Comm. Oreg., Res. Rep. 3:44-48. DOMENOWSKE, R. S. 1966. A comparison of age estimation techniques applied to rex sole, Glyptocephalus zachirus. M.S. Thesis, Univ. Washington, Seattle, 102 p. FREY, H. W. 1971. California's living marine resources and their utili- zation. Calif. Dep. Fish Game, Sacramento, 148 p. HAGERMAN, F. B. 1952. The biology of the Dover sole, Microstomas pacificus (Lockington). Calif. Dep. Fish Game, Fish Bull. 85, 48 p. HARRY, G. Y. III. 1956. Analysis and history of the Oregon otter-trawl fishery. Ph.D. Thesis, Univ. Washington, Seattle, 328 p. 1959. Time of spawning, length at maturity, and fecundity of the English, petrale, and Dover soles (Parophrys vet- ulus, Eopsetta jordani, and Microstomas pacificus, re- spectively). Fish Comm. Oreg., Res. Briefs 7:5-13. Hart, j. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can., Bull. 180, 740 p. Jackson, c. h. n. 1939. The analysis of an animal population. J. Anim. Ecol. 8:238-246. Johnson, A. G., F. M. Utter, and H. O. Hodgins. 1972. Electrophoretic investigation of the family Scor- paenidae. Fish. Bull., U.S. 70:403-414. Ketchen, k. S., and C. R. Forrester. 1966. Population dynamics of the petrale sole, Eopsetta jordani, in waters off western Canada. Fish. Res. Board Can., Bull. 153, 195 p. Knight, w. 1968. Asymtotic growth: an example of nonsense dis- guised as mathematics. J. Fish. Res. Board Can. 25:1303-1307. MANZER, J. I. 1952. Notes on dispersion and growth of some British Co- lumbia bottom fishes. J. Fish. Res. Board Can. 8:374-377. Miller, D. J., and R. N. Lea. 1972. Guide to the coastal marine fishes of California. Calif. Dep. Fish Game, Fish Bull. 157, 235 p. NlSKA, E. L. 1969. The Oregon trawl fishery for mink food, 1958- 65. Pac. Mar. Fish. Comm., Bull. 7:90-101. Norman, J. R. 1934. A systematic monograph of the flatfishes (Heterosomata). Vol. 1. Psettodidae, Bothidae, Pleuronec- tidae. Br. Mus. (Nat. Hist.), Lond., 459 p. PERTSEVA-OSTROUMOVA, T. A. 1961. The reproduction and development of far-eastern flounders. Akad. Nauk. USSR, Inst. Okeanol., 484 p. (Transl. Fish. Res. Board Can. Transl. 856.) Porter, P. 1964. Notes on fecundity, spawning and early life history of petrale sole (Eopsetta jordani) with descriptions of flatfish larvae collected in the Pacific Ocean off Humboldt Bay. California. M.S. Thesis, Humboldt State Coll., Ar- eata, Calif, 98 p. POWLES, P. M. 1965. Life history and ecology of American plaice (Hip- poglossoides platessoides F.) in the Magdalen Shallows. J. Fish. Res. Board Can. 22:565-598. POWLES, P. M., AND V. S. KENNEDY. 1967. Age determination of Nova Scotian greysole, Glyp- tocephalus cynoglossus L., from otoliths. Int. Comm. Northwest Atl. Fish., Res. Bull. 4:91-100. RICHARDSON, S. L. 1973. Abundance and distribution of larval fishes in waters off Oregon, May-October 1969, with special emphasis on the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 71:697-711. RICKER, W. E. 1958. Handbook of computations for biological statistics of fish populations. Fish. Res. Board Can., Bull. 119, 300 p. 1972. Hereditary and environmental factors affecting cer- tain salmonid populations. In R. C. Simon and P. A. Lar- kin (editors), The stock concept in Pacific salmon, p. 19- 160. H. R. MacMillan Lectures in Fisheries, Univ. B.C. SCOTT, D. M. 1954. A comparative study of the yellowtail flounder from three Atlantic fishing areas. J. Fish Res. Board Can. 11:171-197. STARKS, E. C. 1918. The flatfishes of California. Calif. Fish Game 4:161- 179. TSUYUKI, H, E. ROBERTS, AND W. E. VANSTONE. 1965. Comparative zone electropherograms of muscle myogens and blood hemoglobins of marine and freshwater vertebrates and their application to biochemical sys- tematics. J. Fish. Res. Board Can. 22:203-213. VANDERPLOEG, H. A. 1973. The dynamics of 65 Zn in benthic fishes and their prey off Oregon. Ph.D. Thesis, Oregon State Univ., Corvallis, 104 p. VILLADOLID, D. V. 1927. The flatfishes ( Heterosomata) of the Pacific coast of the United States. Ph.D. Thesis, Stanford Univ., Palo Alto, 332 p. WALDRON, K. D. 1972. Fish larvae collected from the northeastern Pacific Ocean and Puget Sound during April and May 1967. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-663, 16 p. 60 ABUNDANCE AND POTENTIAL YIELD OF THE ROUND HERRING, ETRUMEUS TERES, AND ASPECTS OF ITS EARLY LIFE HISTORY IN THE EASTERN GULF OF MEXICO 1 Edward D. Houde 2 ABSTRACT Eggs and larvae of the round herring, Etrumeus teres, were surveyed from plankton collections made in the eastern Gulf of Mexico from 1971 to 1974 to determine adult stock size, spawning areas, and spawning seasons and to study aspects of its early life history. Spawning occurred from mid-October through May where depths ranged from 30 to 200 m, surface temperatures from 18.4° to 26.9°C, and surface salinities from 34.5 to 36.5°/oo. A major spawning area was present 150 km from Tampa Bay between lat. 27°00' and 28°00'N and long. 083°30' and 084°30'W. Mean relative fecundity of 8 adult females was 296.5 ova per gram and the sex ratio of 71 adults was 1:1. The development time of eggs from spawning to hatching was approximately 2.0 days at 22°C. Three methods were used to determine adult biomass. The most probable annual estimates of biomass were approximately 700,000 metric tons in 1971-72 and 130,000 metric tons in 1972-73. The best estimates of the range of potential annual yields to a fishery were from 50,000 to 250,000 tons. Abundance and mortality rates of larvae were estimated in each year. It is probable that more than 99.4% mortality occurred between spawning and the 15.5-mm larval stage during 31 days in 1971-72 and more than 98.3% mortality occurred for the same period in 1972-73. Round herring, Etrumeus teres (DeKay), is one of several clupeid fishes that are abundant in conti- nental shelf waters of the eastern Gulf of Mexico. Distribution and abundance of this species was determined, based on egg and larvae surveys, as part of a program to investigate abundance and fishery potential for sardinelike fishes in the east- ern Gulf. It is generally believed that several species of underexploited clupeid fishes from this area could provide significant catches (Bullis and Thompson 1967; Bullis and Carpenter 1968; Wise 1972) that would supplement yields of the heavily exploited Gulf menhaden, Breuoortia patronus. The egg and larvae surveys were carried out in 17 cruises from 1971 to 1974. Preliminary reports on clupeid abundance, based on these surveys, have been published (Houde 1973a, 1974) and overall results of the surveys were recently summarized (Houde 1976; Houde et al. 1976; Houde and Chitty 1976). There are eight apparently discrete populations of Etrumeus in the world oceans. Whitehead (1963) has placed all of the forms in the single species E. teres. Recorded populations occur in the Contribution from Rosenstiel School of Marine and Atmo- spheric Science, University of Miami, Miami, Fla. 2 Division of Biology and Living Resources, Rosenstiel School of Marine and Atmospheric Science, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149. western Atlantic from Cape Cod into the Gulf of Mexico, in the eastern North Pacific from the Gulf of California to north of Los Angeles, in the central North Pacific near Hawaii, in the Indo-Pacific off the south and west coasts of Australia, in the western North Pacific off the coasts of Japan, in the western Indian Ocean off the east coast of South Africa, in the Red Sea, and near the Gala- pagos Islands in the Eastern Pacific. Eggs and larvae of E. teres have been described from some areas where they occur (Blackburn 1941; Uchida et al. 1958; Mito 1961; Houde and Fore 1973; O'Toole and King 1974; Watson and Leis 1974). Ito (1968) examined fecundity and maturity of round herring from the Sea of Japan. Spawning by Hawaiian round herring recently was discussed by Watson and Leis (1974). Dis- tribution and abundance of round herring eggs and larvae were reported in the Gulf of California (Moser et al. 1974; De la Campa de Guzman and Ortiz Jimenez 1975) and in the northern Gulf of Mexico by Fore (1971). Khromov (1969) found Etrumeus larvae to be common in plankton catches during a winter survey of the eastern Gulf of Mexico. Round herring are fished commercially off Japan and South Africa. A catch of approximately 26,000 metric tons was made by South Africa in 1973 (Food and Agriculture Organization 1974; Manuscript accepted August 1976. FISHERY BULLETIN: VOL.75, NO. 1, 1977. 61 O'Toole and King 1974), and the Japanese catch was 40,400 metric tons in that year (Food and Agriculture Organization 1974). The species is not fished at present in the Gulf of Mexico. Salnikov (1969) reported that round herring was abundant in the northeastern Gulf of Mexico, and Harvey Bullis (pers. commun.) stated that it was plentiful in the eastern Gulf, based on acoustic traces and trawl catches made by National Marine Fisheries Service research vessels. Our initial surveys of eggs and larvae indicated that it might be abun- dant in the eastern Gulf (Houde 1973a), and Fore (1971) reported round herring eggs and larvae to be abundant in the northern Gulf of Mexico. In the absence of a commercial fishery, catch and effort statistics, and other data on abundance, I have estimated the adult biomass in the eastern Gulf from the abundance of eggs that were spawned annually. This fishery-independent technique of biomass estimation can provide preliminary knowledge of fishery potential (Ahlstrom 1968) and is considered to be a useful biomass estimat- ing procedure (Saville 1964; Smith and Richardson in press). METHODS Survey Area and Times Seventeen plankton surveys were made in the eastern Gulf of Mexico between lat. 24°45' and 30°00'N (Figure 1) in 1971-74 (Table 1). Most sampling stations were located on the broad conti- nental shelf, where depths ranged from 10 to 200 m, but a few stations were over the continental slope where depths were greater. Potential sam- pling stations were on transects running parallel to lines of latitude; transects were spaced at 15- nautical-mile (27.8-km) intervals. Stations were located at 15-mile (27.8-km) intervals on each transect, except for those stations beyond the 200-m depth contour, which were placed at 30- mile (55.6-km) intervals (Figure 1). Not all sta- tions were sampled on each cruise (Table 1). Other details of survey planning and design have been reported elsewhere (Rinkel 1974; Houde et al. 1976; Houde and Chitty 1976). Beginning with cruise IS 7205 (Table 1), sam- pling was restricted to stations on alternate tran- sects. The three stations nearest to shore (at 27.8-km intervals) were sampled on each of the designated transects but only stations at 30-mile (55.6-km) intervals were sampled offshore. A few FISHERY BULLETIN: VOL. 75, NO. 1 T" FIGURE 1.— Area emcompassed by the 1971-74 eastern Gulf of Mexico ichthyoplankton surveys. Plus symbols ( + ) represent stations that were sampled during the survey. The 10-, 30-, 50-, and 200-m depth contours are indicated. additional stations were added on 1974 cruises in areas where depth was less than 10 m; no round herring eggs or larvae occurred at these stations and they were not important with regard to spawning by this species, but they were important in determining spawning and distribution of other Gulf clupeids. Plankton Sampling A paired 61-cm Bongo net plankton sampler was used on all cruises except cruise GE 7101, in which a 1-m ICITA [International Cooperative Investi- gations of the Tropical Atlantic (Navy)] plankton net with 505-^m mesh was towed. Meshes on the Bongo sampler were 505 /xm and 333 fxm. Ichthyoplankton was sorted from the 505-^tm mesh net and plankton volumes were determined from the 333-/u,m mesh net catch (Houde and Chitty 1976). Net tows were double oblique from within 5 m of bottom to surface or from 200-m depth to surface at deep stations. Nets were towed at approximately 3.0 knots (1.5 m/s) in 1971, but towing speed was reduced on later cruises and averaged 2.3 knots (1.2 m/s) (Table 2). Stations were sampled whenever the ship occupied them; thus, tows were made during either daylight or darkness, depending on the time of arrival at a station. Prior to cruise GE 7208, all tows consisted of 62 HOUDE: ABUNDANCE AND POTENTIAL YIELD OE ROUND HERRING TABLE 1 . — Summarized data on cruises to the eastern Gulf of Mexico, 197 1 -74, to estimate abundance of round herring eggs and larvae. (GE = RV Gerda, 8C = RV Dan Braman, TI = RV Tursiops, 8B = RV Bellows, IS = RV Columbus Iselin, CL = RV Calcnus.) Number of stations Positive stations Positive stations Mean egg abundance under 10 m 2 Mean larvae abundance under 10 m 2 Cruise Dates for eggs' for larvae 2 All stations Positive stations All stations Positive stations GE 7101 3 1-8 Feb. 1971 20 4 9 39.37 196.88 7.34 16.30 8C 7113 TI 7114 7-18 May 1971 123 2 24 0.21 12 88 300 1580 GE 7117 26 June-4 July 1971 27 0.00 — 000 — 8C 7120 TI 7121 7-25 Aug. 1971 146 0.00 — 0.00 — TI 7131 8B 7132 GE 7127 7-16 Nov. 1971 66 15 20 41.41 187.73 4.18 14.20 8B 7201 GE 7202 1-11 Feb. 1972 30 8 13 151.20 604.81 20.29 49.97 GE 7208 1-10 May 1972 30 2 2 1.38 22.11 0.28 4.44 GE 7210 12-18 June 1972 13 0.00 — 0.00 — IS 7205 9-17 Sept. 1972 34 0.00 — 0.00 — IS 7209 8-16 Nov. 1972 50 5 2 0.83 8.30 1.61 40.28 IS 7303 19-27 Jan. 1973 51 12 20 23.77 101.04 19.12 48.76 IS 7308 9-17 May 1973 49 2 3 2.48 6072 229 37.41 IS 7311 27 June-6 July 1973 51 0.00 — 0.00 — IS 7313 3-13 Aug. 1973 50 0.00 — 0.00 — IS 7320 6-14 Nov. 1973 51 8 5 4.11 26 22 111 1 1 32 CL 7405" 28 Feb.-9 Mar. 1974 36 0.00 — 000 — CL7412 1-9 May 1974 44 1 1 0.49 21.50 3.98 175.07 'Positive station is a station at which round herring eggs were collected. 2 Positive station is a station at which round herring larvae were collected. 3 An ICITA, 1-m plankton net was used on this cruise. On all other cruises a 61 -cm Bongo net was used. 4 No stations in offshore areas were sampled, accounting for the failure to collect round herring eggs or larvae on this i TABLE 2. — Summary of plankton tow characteristics for 17 ichthyoplankton cruises to the eastern Gulf of Mexico. The 61-cm Bongo net sampler was used on all cruises except GE 7101 in which a 1-m ICITA net was used. Standard error Mean Standard error Mean volume Standard error of Number Mean volume of towing of filtered per volume filtered of filtered volume filtered speed towing speed unit depth per unit depth Cruises stations (m 3 ) (m 3 ) (m/s) (m/s) (m 3 /m) (m 3 /m) GE7101 8C7113& TI 7114 GE 7117 8C 7120 & TI 7121 8B7132&TI 7131 GE 7202 & 8B 7201 GE 7208 GE 7210 IS 7205 IS 7209 IS 7303 IS 7308 IS 731 1 IS 7313 IS 7320 CL 7405 CL7412 20 358 335 <55 m deep 124 >55 m deep 675.25 160.17 104.39 231 .93 30 29 7.27 0.92 11.80 1.44 1.17 1.18 0.03 0.01 0.01 49.69 3.60 11.04 2.37 11.58 0.11 0.57 0.07 wire release at 50 m/min to desired depth and retrieval at 20 m/min. In later cruises, two types of tow were used, a shallow-water tow at stations less than 55 m deep and the usual 50 m/min release-20 m/min retrieval tow at deeper stations (Table 2). The shallow-water tow was of 5-min duration; it consisted of 1 min for wire release and 4 min for wire retrieval. The objective at shallow stations was to filter 100 m 3 of water. This objective was met, but the volume of water filtered per unit of depth fished by the net was increased significantly at the shallow stations relative to deeper stations (Table 2). This discrepancy in type of tow was considered to be more desirable than the alterna- tive situation, which existed in 1971, when as lit- tle as 25 m 3 of water were filtered at some of the shallowest stations. Tows at stations deeper than 55 m filtered between 100 and 400 m 3 . A stopwatch was used to monitor each tow and the wire angle was measured at the end of each minute of a tow. A time-depth recorder gave a record of tow characteristics. Volume filtered was determined from a flowmeter in the mouth of the 505-yu.m mesh net. 63 FISHERY BULLETIN: VOL. 75, NO. 1 Plankton Samples All samples were preserved immediately in 10% seawater Formalin 3 buffered with marble chips. Samples were transferred to 5% buffered Forma- lin after they had been stored in the laboratory for 1 mo. Houde and Chitty (1976) have discussed methods used to determine plankton volumes. All fish eggs and larvae were sorted from each 505-/xm mesh net plankton sample under a dissecting mi- croscope for later identification and enumeration. Eggs and larvae of round herring are distinctive and easily identified (Houde and Fore 1973). Round herring eggs from each station were enu- merated; larvae were enumerated and measured with an ocular micrometer in a dissecting micro- scope. Temperatures and Salinities Temperature and salinity profiles of the water column at each station were obtained on all cruises. 4 Usually a mechanical bathythermo- graph cast was made to describe the vertical tem- perature profile. This was followed by a hydrocast consisting of from two to seven 1.7-liter Niskin bottles with reversing thermometers. Samples for salinity were brought to Rosenstiel School of Marine and Atmospheric Science for analysis. On cruises IS 7308 and IS 7320 a salinity- temperature depth unit was used in place of the Niskin bottles to obtain temperature and salinity data. Round herring egg and larva data were examined in relation to temperatures and salinities at stations where they were collected. Determining Egg and Larvae Abundance Catches of round herring eggs and larvae at each station were standardized to give abundance in numbers under 10 m 2 of sea surface: n. c j 2 j 10 (1) where n, = the number of individuals (eggs or lar- vae) at station j under 10 m 2 of sea surface 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ••Temperature and salinity data for these cruises can be re- trieved from the MAFLA file at the National Oceanographic Data Center, Washington, D.C. c = the catch of eggs or larvae at station^' Zj = the depth of tow (in meters) at station/ Vj = the volume filtered by the net (in cubic meters) at station j. Both total larval abundance under 10 m 2 and lar- val abundance in each 1.0-mm length class under 10 m 2 were determined. Numbers of eggs or larvae also were estimated in the area represented by each station. These areas were determined by the polygons described by the perpendicular bisectors of lines from the station in question to adjacent stations (Sette and Ahlstrom 1948): Pj CjZj Aj (2) where p • = the estimated total number of eggs or larvae in the area represented by sta- tion j Cj, Zj, and Vj are defined in Equation (1) Aj = the area (in square meters) rep- resented by station j . Total larvae and larvae by 1.0-mm length classes were estimated for each station area. Most sta- tions represented areas ranging from 0.75 to 3.15 x 10 9 m 2 . The estimated total number of eggs and larvae, as well as larvae by 1.0-mm length classes, was estimated for the entire area represented by each cruise: P, = I (3) 7 = 1 where P, = the cruise estimate (i.e., the total number of eggs or larvae estimated in the area represented by cruise i) k = the number of stations sampled dur- ing cruise i Pj is defined by Equation (2). Variance estimates on the abundance of eggs were obtained for each cruise using a combination of methods outlined by Cushing (1957) and Taft (1960). Only stations at which round herring eggs had been collected at least once during the 1971-74 survey period were included in obtaining these estimates. Other stations were considered to be outside the area of spawning, because round her- ring eggs were never collected there. These usu- ally were the three stations on each transect that 64 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING were located closest to the coast (Figure 1). An estimate of the variance in egg abundance under a square meter of sea surface (s y ) was obtained from the log 10 ((CjZjVvj) + 0.1 egg catch at each station during a cruise (Cushing 1957). The log 10 variance estimate so obtained was backtransformed to ob- tain the untransformed estimate of variance. The variance estimate for a cruise was calculated using the estimator given by Taft (1960) that as- sumes random sampling. It is: 7 = 1 A 2 50 m were calculated from pooled data of all cruises that had round herring eggs. The «50 ra logio mean was 1.6351 (n = 25, Sj = 0.1609); the >50 m log 10 mean was 1.5585 (n = 32, S; = 0.1209). These means did not differ significantly (f-test;P>0.50). However, the area between the 30- and 50-m depth contours was less than that included between the 50- and 200-m contours. The total area between the 30- and 200-m depth contours was considered to be the spawning area; 40.1% of the area is in the 30- to 50-m zone while 59.9% is between 50 and 200 m. Thus, the total abundance of eggs in the area where depths exceeded 50 m probably was greater than abundance in shallower areas. The 50-m depth contour divides the shelf area in the eastern Gulf into approximate halves. For eight cruises in which sampling effort was distributed nearly equally to include potential spawning area in water =£50 m and >50 m (cruises 8C 71 13-TI 7114, 8B 7132-TI 7131-GE7127, 8B 7201-GE 7202, GE 7208, IS 7209, IS 7303, IS 7308, and IS 7320), the summed totals of egg abundance from the areas represented by stations on these cruises were compared with respect to the 50-m depth contour. A total abundance of 11.92 x 10 11 eggs was esti- mated for stations =£50 m; total abundance was 16.73 x 10 11 at deeper stations. If these egg abun- dance estimates reflect relative adult abundance, then 41.6% of the adult population was located in depths =£50 m and 58.4% was distributed at depths >50 m. The total abundance of eggs, and appar- ently of adult round herring, is directly propor- tional to the surface area of the two depth zones. Some small fraction of the spawning population inhabited depths greater than those sampled in 70 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING GE 7101 Ltrkus teres egos February 1971 (£7101 I ; 6 TERES LARVAE FEBRUARY 1971 1 i \ 50m- \ 4 1 • • i* + + ' 4 + 4 . 4- Number under IOm 2 4 • IOO0 1 86° 84' 8C 7113 S Tl 71114 etrumeus teres eggs May 1971 30° i r — r— t 4 4 4 + + V \ 50m-. + * + * * V-j \ *-»++++ + + V \ +\ +44 + 41 ««. + '"# 4 4 4 + + / W) 4*44*44/ Y 28 e + + 4*.+ 4 4 + >J7 \ + t V, + * n^ \ 4 4 4 V + 4 + \ i + + + \+ + + \r? r © xooo ' ' 1 84* 6E 7127, TI 7131 & 8B 7132 Etrumeus teres eggs November 1971 30° 1 1 - T \ » + *■ * + + V. 50m-* + + r \^/f • ^ + * + + + ^l •gV + + 4- + + I + *■ •• 4- + + t / 28° «■•«©• * 4 SJ/ * • Ml • + 4 •,{ V - * 4 • + \ • 4 f \ t + t i t * * ^t • ° 1 i.'b° Number under 10m 2 l , \ / • I000 30" 50m- u^ \ 28* M o\ i:6- ?4° 1 Number under IOm 2 • 1 • " • I-IO • 11-100 • 101 - IOOO © XOOO 1 • l_ .£: 4 4 • + 4 + 4 4 4 4 + z~-"^ 84' 8C 7113 S TI 7111 f.truheus teres larvae Nay 1971 1 M^,^ i i 30° 50m-. f + + + 4 V + 4 4++ \- *-.+ +4-4+4 V 4\ 4+44+ I • fc 4 + + 4 + / • •+444+/ 28° 4 \ J(V^ 4 • *« • 4 4 Sj^ • ••""* + 4 4 <,£ •••\++»\ +>•+•* 4 4 4 \ _ •••4 44+^M 4- 4- +'.+ + + + f • 4 4\4 +4 4- + 4 +',• 4 + + + •+*«++ + + 4- o\ Lb" 1 Number under IOm 2 4 • I-IO • 11-100 • IOI-IOOO + + +i 4 4- +4 + + ,4 -»- 4 4 4 +J • + 4 + + + <*--'' © >I000 I * 84° GE 7127, TI 7131 & 8B 7132 Etrumeus teres larvae November 1971 1 r— 30° \ 4 4 + * + + V 50m« * +"*■"*■■*■ V-« • #.+ + * + 1 4 +#•+++- / 28° v - • • • y tt » yj o\ 1 Number under IOm 2 : ?■ • l-IO \ T. [• • 11-100 • 101 - IOOO ' ' * + j*!*^ © >I000 1 FIGURE 3. — Distribution and abundance of round herring eggs and larvae. Catches are standardized to numbers under 10 m 2 of sea surface. A, B: Cruise GE 7101, February 1971. C, D: Cruise 8C 7113- TI 7114, May 1971. E, F: Cruise GE 7127-TI 7131-8B 7132, November 1971. 71 FISHERY BULLETIN: VOL. 75, NO. 1 8B 7201 ft GE 7202 Etrumeus teres eggs February 1972 SB 7201 & GE 7202 Etrumeus teres larvae February 1972 — l — 1 M, _^-^ -1 T 30" 50m« *■ * • + + + \^- *- \ + ■*■ + J 28" • \ - o\ 26" Number under 10m 2 • I000 GE 7208 Etrumeus teres eggs Hay 1972 30° 50m-. ^ \ + / 26* \ f^ V - + + + + 1 + * \ K» o\ + + + + 'i + *- «^ + \ + + 2fo° Number under 10m 2 + • I000 | 30° \ 50m« • • • • '• * * 28" + • \ + o\ 26° Number under 10m 2 •\ - >> fl • 1 - 10 + • w + + • 11-100 • 101-1000 * * •*.. +- .*•»--' © >I000 84" GE7233 Etrumeus teres larvae May 1972 30° 30m--, i r \ + 28" (?\A V - + + X . ,/ * \ * *\tf • <) ) f + + 1 -t- + 2fo° Number under 10m 2 + • <\ • ! + * * \ y • 1 - 10 • 11-100 • 101 - 1000 • ; + * .>--" © >I000 IS 7209 Etrumeus teres eggs November 1972 30° 30m- .^ ' ' f ***** \ \ *■ N « + * * + / Vy 28° • \ * * * -/^ \ ♦- + • * \ + * st€ *\ *> + * V. J" 2b" Number under 10m 2 t f +■ ! + * *" * V • 1 - 10 • 11-100 • 101 - 1000 © >I000 IS 7209 Etrumeus teres larvae November 1972 • I - 10 • 11-100 • 101 - 1000 © >I000 FIGURE 4.— Distribution and abundance of round herring eggs and larvae. Catches are standardized to numbers under 10 m 2 of sea surface. A, B: Cruise 8B 7201-GE 7202, February 1972. C, D: Cruise GE 7208, May 1972. E, F: Cruise IS 7209, November 1972. 72 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING IS 7303 Etrumeus teres eggs January 1973 IS 75)3 Etrumeus teres larvae January 1973 JO' 26* u*-^ 50m-. • \ * ' ' ' 1 \ • \ * • • \j- I000 1 80° 50m-.. \_ • • • • V • \ * * * / • \ • • • <-/*y- \ - • »'•-., • • - l/ • • \ + + \j- Tc. Number under 10m- • * <-\ * + • • ': • * * * \ y • 1-10 • 11-100 • 101 - 1000 • •: • .-<"'-'' © >I000 i i . _i_ IS 7308 Etrumeus teres eggs Hay 1973 64" IS 7308 Etrumeus teres larvae May 1973 30° 50m-., -t- \ 28* + o\ *■ * *-\ +■ + + 2b" Number under 10m 2 + * * ; +■ +- + \ y • 1 - 10 • n-ioo • 101 - 1000 *j * * ...*■*--' © >I000 30° 1 1 r- 50m '-. \~» • * + * + * / \ 26° • \ + - • ♦[ft. V * * \ + * *\l o\ * +■ \ *■ + * »x 26" Number under 10m 2 * . .,.+ . . ': . ^9 / • 1 - 10 • 11-100 • IOI-IOOO *l * * .-«--"' © >I000 1 ' « IS 7320 Etruheus teres eggs November 1973 84" IS 7320 Etruheus teres larvae November 1973 1 1 — V&A" \ + t + + \ 50m.. • + + + V + + \ • *■ + / \ + \ • + + * 0") * + + 1 + + + + Number under 10m 2 • ! + + -+■ * \ y • 1-10 • 11-100 • IOI-IOOO + tj + .'•"'-'' © >I000 1 Number under 10m 2 + • I000 86" FIGURE 5. — Distribution and abundance of round herring eggs and larvae. Catches are standardized to numbers under 10 m 2 of sea surface. A, B: Cruise IS 7303, January 1973. C, D: Cruise IS 7308, May 1973. E, F: Cruise IS 7320, November 1973. 73 FISHERY BULLETIN: VOL. 75, NO. 1 a 7112 Etrikeus teres eggs a 7112 Etrueus teres LARVAE May 1971 \ — — r- r \ 50m- X ' * • * *~\ - \ - \ * * *\|T . I000 k^-^ ><~^ \ * u^^'tX \ 50m-. \ . . :N\ \ \ ••••) 1) - V ' * w \ " • '"■■•. •♦*•'/ \ V*\r ' ) Y.\*\ J Number under 10m 2 • ' *^ _J • 1 - 10 • II-IOO • 101-1000 ! . ..'»■-'' © >I000 FIGURE 6. — Distribution and abundance of round herring eggs and larvae on cruise CL 7412, May 1974. Catches are standardized to numbers under 10 m 2 of sea surface. our survey and the relative abundance of adults in water >50 m deep may be higher than the esti- mated 58.4%. Because the intensity of spawning was the same in depths, =£50 and >50 m, adults apparently are not more abundant per unit of sea surface in deeper water but their greater abun- dance reflects the larger area of habitat suitable for round herring where shelf waters are >50 m deep. Temperature and Salinity Relationships Round herring eggs were collected when surface temperatures ranged from 18.4° to 26.9°C. They occurred at surface salinities of 34. 50-36. 50°/oo. Because no vertically stratified tows of the Bongo sampler were made, the percentage of eggs or lar- vae that occurred in surface waters is unknown. Surface temperatures from November to May were 0°-3°C higher than those at 50 m when verti- cal sections along transects at three latitudes were examined for each cruise in which round herring eggs or larvae were collected. Surface salinities differed by less than 0.5°/oo from those at 50-m depth, except on cruise IS 7320 when surface salinities ranged from 0.6 to 1.0%o less than those at 50 m. It is reasonable to believe that surface temperatures and salinities are representative of conditions where pelagic eggs were incubated and where larvae were found. Salinity may not be an important factor affecting spawning since the range of surface salinities at which eggs were col- lected nearly encompasses the entire range of salinities found in offshore waters of the eastern Gulf. Larvae =£5.0 mm SL are from to about 6 days old. They occurred where surface tempera- tures ranged from 20.5° to 26.9°C and surface salinities from 34.10 to 36.80%o. The percentage cumulative frequency distri- butions (Figure 7) of stations where eggs or =s5.0-mm larvae occurred in relation to tempera- ture and salinity were examined. For eggs, 82.5% of the occurrences were between 21° and 26°C sur- face temperature, while 87.2% of the =s5.0-mm larvae occurrences were in that temperature range. Only 10.5% of the egg occurrences were at stations where surface temperatures exceeded 26°C and only 6.4% of the =£5.0-mm larvae occur- rences were at such stations. The distribution of egg occurrences in relation to temperature was similar in the 1971-72 and 1972-73 spawning sea- sons. In 1971-72, 78.3% of the eggs occurred at stations where surface temperatures were less than 25°C; in 1972-73, 79.0% of the occurrences were at temperatures below 25°C. Comparable data were not available for the 1973-74 spawning season. More than 50% of round herring eggs and =s5.0-mm larvae were collected at stations where surface salinity exceeded 36.00%o (Figure 7). For eggs, considering all years' data, 79.7% of the oc- currences were at surface salinities from 35.50 to 36.50%o; for «5.0-mm larvae, 80.0% of the occur- rences were in that salinity range. In 1971-72, 88.0% of the egg occurrences were at stations with surface salinities from 35.50 to 36.50%o; in 1972- 74 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING TEMPERATURE SALINITY FIGURE 7. — Percent cumulative frequency dis- tribution of 1971-74 stations where round her- ring eggs occurred in relation to surface tem- peratures (A) and to surface salinity (C), and =£5.0-mm SL larvae occurred in relation to surface temperature (B) and surface salinity (D). 100 90 80 70- 60- 50- >- 40 U " 50 ^ 100 X 3 9 ° * 80 £ 70 Q. 60 - 50- 40- 30 20 10 Etrumeus teres eggs Etrumeus teres larvae -5mm -T 1 r- Etrumeus teres larvae '5mm 100 '-•') 80 70 t'i 50 40 SO 20 10 IB I 20.1*- 22 r- 24 1*- 26 r- 3401- 21 0' 23 0* 25 - 27 0' 34 25 TEMPERATURE CLASS CO 34 51- 35 01- 35 51- 3601- 34 75 35 25 35 75 36 25 SALINITY CLASS (%.) 36 51- 36 75 73, 94.7% of the egg occurrences were in that salin- ity range. There were seven egg occurrences at less than 35.50%o surface salinity on cruise IS 7320 (November 1973). This cruise influenced the cumulative frequency distribution of egg occur- rences in relation to salinity (Figure 7) over all years. Data for the entire 1973-74 spawning sea- son were not available to compare occurrence of eggs in relation to salinity with 1971-72 and 1972-73 data; but, the frequency distribution ap- parently would have been shifted to lower salinities in that year, reflecting low surface salinities that prevailed in the eastern Gulf in fall 1973. Egg and Larvae Abundance in Relation to Zooplankton There was no apparent relationship between zooplankton volumes and round herring egg or larvae abundance. Zooplankton volumes (cubic centimeters/1,000 m 3 strained) were determined at each station for cruises in 1972 through 1974. Round herring egg abundance and larvae abun- dance were examined in relation to zooplankton volume for stations included in those cruises but the correlations were not significant. Fecundity and Maturity A total of 71 adult round herring was examined, of which 39 were males and 32 were females. Based on this sample, the sex ratio did not differ significantly from 1:1 (x 2 = 0.69; 0.2520.0 mm were rarely collected during the survey. Frequencies for each length class in Figure 9 are given as esti- mated abundance during each cruise (Equation (3)). No area adjustments have been made in Fig- ure 9 for the two cruises that did not cover the entire spawning area. Round herring larvae <4.0 mm SL usually were in poor condition, with curved or deformed bodies, and their measure- ments are underestimates of true length. O'Toole and King (1974) hatched eggs that they had col- lected and reported that preserved, newly hatched round herring larvae were 3.75-4.00 mm long. The 4.1- to 5.0-mm SL length class was the most abun- dant class in my survey (Figure 9). I assumed that this length class was fully vulnerable to the sam- pling gear, although some escapement may have occurred for larvae of this size. The ratios of night-caught to day-caught larvae 80 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING 48 44 40 36 32 28 24 20 16 12 8 4 48 44 40 36 32 28 24 20 16 12 8 4 28 24 20 16 12 6 4 28 24 20 16 12 8 4 .rfTTh-i-^ 8C7II3 - TI 7114 On- ~L-TK GE7I27 - TI 7131 -8B7I32 887201-GE 7202 m . ^ . n _Q_ GE 7208 TlD IS 7209 D IS7303 "h-TT-T-i-n h , , rh r~i , rnx^ IS 7320 CL 7412 tLu 2.1- 4 1- 6.1- 8.1- 10.1- 12.1- 14.1- 161- 181- 30 50 70 90 110 13 15 17.0 19 STANDARD LENGTH CLASSES (mm) FIGURE 9. — Length-frequency distributions of round herring larvae for 1971-74 cruises to the eastern Gulf of Mexico. Fre- quencies are expressed as estimated abundance of larvae in each length class within the area represented by the cruise. by length classes were examined over all cruises and they indicated that considerable net avoid- ance was occurring in the day relative to that occurring at night. The data were plotted by 2-mm length classes (Figure 10), and functions were fitted to allow estimation of the night-caught to day-caught ratio for larvae in any length class. The ratio increased rapidly for larvae of 4.0-13.0 mm, but then decreased from a factor of more than 3.0 to about 1.0 when larvae had grown to 18.0 mm. Two power functions were fitted: for larvae 2.1- 14.0 mm SL the function was R = 0.3041 X° ' 9115 , where R is the ratio of night-caught to day- caught larvae andX is standard length of larvae; for 12.1- to 20.0-mm SL larvae the function was R = 44,521.54 X" 37298 . Larva catches made at daytime stations were adjusted by R (Equation (ID). Exponential functions or a single poly- nomial could have been used in place of the power functions to describe the relationship, but the power functions provided reasonably good fits to the data and were acceptable for correction pur- poses. No adjustments were made for larvae <4.0 mm or > 18.0 mm because there was no observable difference in night or day catches for larvae of those lengths. The round herring larvae night to day catch ratios are unusual with respect to the observed : 3 50 <2 50 <0 50- Y- 0.3046X 09 " 5 Y- 44521. 54X 50 7.0 9.0 110 130 150 MIOPOINT OF LENGTH CLASS (mm) 17 19 FIGURE 10. — Night to day ratios of sums of catches, standardized to numbers under 10 m 2 of sea surface, for round herring larvae collected in 1971-73 in the eastern Gulf of Mexico. The ratios were calculated for larvae within each 2-mm length class from 2.1 to 20.0 mm SL. Fitted power functions describe the relation- ships for larvae from 2.1 to 13.0 mm SL and for larvae from 13.1 to 20.0 mm SL. Larval abundance estimates for each length class at stations occupied during daylight were corrected by the appropriate ratio factor for each length class to account for daytime avoidance. 81 FISHERY BULLETIN: VOL. 75, NO. 1 decrease in the ratio for larvae >13.0 mm. The ratio increased in other studies on clupeoid larvae throughout the size range of larvae that were col- lected (Ahlstrom 1954, 1959b; Lenarz 1973; Mat- suura in press), and this is true for other species of clupeid larvae that I have studied in the Gulf of Mexico. The return of the ratio toward unity after round herring larvae reached 13.0 mm must indi- cate that larvae 13.0-18.0 mm became as good at avoiding the gear at night as during the day. The alternative explanation, which seems unlikely, is that larger larvae lost the potential to avoid the gear during daylight. Daylight is only one factor that could allow larvae to avoid the gear and ad- justment of catches to account for it can only par- tially correct for avoidance losses. The correction was made, however, in an attempt to get the best estimate possible for round herring lar- val mortality during the 1971-72 and 1972-73 seasons. Larval abundance estimates, corrected for day- time avoidance, were determined by 1-mm length classes for the 1971-72 and 1972-73 seasons (Fig- ure 11) (Equation (10)). Except for larvae in the 4.1- to 5.0-mm length class, which were twice as abundant in 1972-73, total abundance of larvae was similar in the two seasons. The greater abun- dance of 4.1- to 5.0-mm larvae in 1972-73 could have reflected the reduction in towing speed from the previous season. Escapement of small larvae through the meshes may have been more impor- tant in 1971-72 when towing speed averaged about 0.7 knot faster. Abundance of round herring larvae decreased exponentially as lengths increased during each season (Figure 11). Fitted exponential functions for 5.1- to 16.0-mm larvae in 1971-72 and 4.1- to 16.0-mm larvae in 1972-73 provided estimates of the instantaneous mortality coefficients per mil- limeter increase in length (Figure 11). The coefficients were Z = 0.2269 in 1971-72 andZ = 0.3647 in 1972-73. These correspond to percentage losses per millimeter increase in length of 20.3% in 1971-72 and 30.5% in 1972-73. Confidence in- tervals at the 0.95 probability level were Z = 0.2269 ± 0.0930 in 1971-72 and Z = 0.3647 ± 0.1179 in 1972-73. The null hypothesis of no difference in mortality coefficients between years was accepted at the a = 0.05 probability level U-test; 0.05T> 2 ,. »|. 4|. si- 6i- 7|- Bh 91- HI- 13 1- 151- 17 1- 19 1- 21. t- 23 1- 23 I- 271- 29 1- 3 40 30 60 70 8090(0 120 14 160 18 20 22.0 24 26 28 30 LENGTH CLASS (mm) FIGURE ll. — Length-frequency distributions of annual larval abundance estimates of round herring larvae collected in the eastern Gulf of Mexico. Frequencies in each 1-mm length class are expressed as estimated annual abundance and have been corrected for daytime avoidance. Fitted exponential functions provide estimates of the instantaneous coefficient of decline in abundance by length, 1971-72 and 1972-73. similar to those reported by Lenarz (1973) from several years of data on Pacific sardine and north- ern anchovy, Engraulis mordax. He reported a range of instantaneous coefficients of 0.15-0.33, averaging 0.22 for Pacific sardine, that correspond to a 20% loss per millimeter of growth. For an- chovy his instantaneous coefficients ranged from 0.32 to 0.46, averaging 0.39, a mean decrease of 32% per millimeter of growth. Matsuura (in press) has measured the rate of decline in catches of Brazilian sardine, Sardinella brasiliensis, obtain- ing an instantaneous coefficient of 0.4962, corre- sponding to a 39% decrease in catch per millimeter of growth. Most of the decline in catch of larger round herring larvae presumably was due to lar- val mortality but gear avoidance also must be important. For this reason mortality curves were fitted only for larvae 16.0 mm or less in length. Catches of larger larvae were sporadic and possi- bly greatly influenced by gear avoidance. Larval mortality is best expressed as a function of age. If it is assumed that growth of round her- ring larvae is exponential from the post yolk-sac 82 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING stage to 20.0 mm SL, then the instantaneous mor- tality coefficients, based on mean estimated ages of larvae, can be calculated using Equations (12)- (16). No information on growth rates of round her- ring larvae was available, but mean daily growth increments of other Gulf of Mexico clupeid species have been determined in laboratory rear- ing experiments and they range from 0.3 to 1.0 mm (Richards and Palko 1969; Saksena et al. 1972; Houde 1973b; Houde and Swanson 1975). Growth rates in those experiments exceeded 0.7 mm/day only when temperature was above 26°C. Mean daily growth of round herring larvae proba- bly is between 0.3 and 0.7 mm. Duration of the egg stage from spawning until hatching is about 2.0 days. The duration of nonfully vulnerable length classes was estimated from a knowledge of growth rate and development times of other clupeid species that have been reared in the laboratory. Larvae of yellowfin menhaden, Brevoortia smithi, did not begin to grow in length until nearly 4 days after hatching at 26°C (Houde and Swanson 1975) when they were about 4.5 mm SL; larvae of Harengula jaguana did not grow significantly until they were nearly 3 days old and 4.5 mm SL at 26°-28°C (Houde et al. 1974). The exponential growth phase was assumed to begin in the 4.1- to 5.0-mm length class for round herring. The non- fully vulnerable length classes of 2.1-5.0 mm in 1971-72 were assigned durations that varied from 4.0 to 7.0 days; the nonfully vulnerable 2.1- to 4.0-mm length classes in 1972-73 were assigned durations of 1.5-3.0 days. Various combinations of mean daily growth increments and durations of nonfully vulnerable length classes were entered into the program to estimate mortality in relation to age of larvae (Equations (12)-(16)) for 1971-72 and 1972-73. Examples, for one combination of values of the variables in 1971-72 and one combi- nation in 1972-73, are provided in Table 10 and Figure 12. Given mean daily growth increments of 0.3-0.7 mm (corresponding to instantaneous growth coefficients of 0.0299-0.0698) and the most proba- ble durations of nonfully vulnerable length clas- ses, the probable range of instantaneous mortality coefficients was 0.0866-0.1739 in 1971-72 and 0.0835-0.1719 in 1972-73 (Table 11). In terms of daily mortality the 1971-72 probable estimates ranged from 8.3 to 16.0%; in 1972-73 they ranged from 8.0 to 15.8% . Although the estimated range is great, it is nearly the same for the two seasons. Varying duration of the nonfully vulnerable length classes had only minor effects on mortality rate estimation (Table 11), but varying the growth rate had important effects. The values ofiV , they-axis intercepts, provide yet another series of estimates of annual spawn- ing, because they estimate the numbers of eggs present at time zero. The intercept values are gen- erally lower than spawning estimates by the other methods and are not considered to be good esti- mates of spawning. It seems that the exponential model of loss fits the decrease in larval abundances reasonable well, but that a greater than expected mortality occurs between egg and fully vulnerable larval length classes. Figure 12 illustrates this possibility. If only larval mortality had been con- sidered, rather than total mortality from egg to 16.0-mm larvae, the instantaneous coefficients TABLE 10. — Two examples of data treated to obtain class durations and mean ages of round herring larvae from the eastern Gulf of Mexico. Abundance estimates are then corrected for duration, and the duration-corrected abundances were subsequently regressed on mean ages to obtain mortality rates (Table 11). Data are from 1971-72 and 1972-73 egg and larvae abundance estimates that were pre- viously corrected for daytime avoidance. In these examples the mean daily growth increment (b) was set at 0.50. The nonfully vulner- able length classes were 2.1-5.0 mm in 1971-72 with duration of 6 days, and 2.1-4.0 mm in 1972-73 with duration of 2.5 days. Calculat- ing procedures are given in Equations (12)-(16). The regressions for these data are given in Figure 12. 1971-72 1972-73 Mean Duration-corrected Mean Duration-corrected Abundance Duration age abundance Abundance Duration age abundance Class (no. x 10") (days) (days) (no. x 10 11 ) Class (no. x 10") (days) (days) (no. x 10") Eggs 2,128.39 2.00 1.00 1,064.20 Eggs 388.94 2.00 1.00 194.47 2.1-5.0 72.90 6.00 5.00 12.15 2.1-4.0 43.89 2.50 3.25 17.56 5.1-6.0 61.96 3.26 9.52 19.00 4.1-5.0 117.78 3.98 6.37 29.58 6.1-7.0 38.96 2.76 12.87 14.11 5.1-6.0 55.29 3.26 10.39 16.95 7.1-8.0 31.70 2.39 15.74 13.24 6.1-7.0 69.81 2.76 13.75 25 28 8.1-9.0 35.92 2.11 18.25 16.99 7.1-8.0 35.42 2.39 16.62 14.79 9.1-10.0 46.88 1.89 20.48 24.77 8.1-9.0 34.55 2.11 19.13 16.34 10.1-11.0 22.29 1.71 22.49 13.02 9.1-10.0 17.08 1.89 21.36 9.02 11 1-12.0 11.60 1.56 24.32 7.41 10.1-11.0 7.44 1.71 23.37 4.34 12.1-13.0 26.81 1.44 25.99 18.63 11.1-12.0 22.99 1.56 25.20 14.70 13.1-14.0 12.25 1.33 27.53 9.19 12.1-13.0 6.67 1.44 26.87 4.63 14.1-15.0 989 1.24 28.97 7.97 13.1-14.0 4.79 1.33 28.41 3.59 15.1-16 3.31 1.16 30.31 2.85 14.1-15.0 0.74 1.24 29.85 0.59 15.1-16.0 4.36 1.16 31.19 3.76 83 FISHERY BULLETIN: VOL. 75, NO. 1 1000- 500 - r-100 b X (jj 50 o z < Q Z m < a UJ >" 10 UJ x < i _L 2 4 6 6 10 12 14 16 16 20 24 28 ESTIMATED MEAN AGE (DAYS) 32 FIGURE 12. — Estimated abundance of egg and larval stages of round herring in the eastern Gulf of Mexico in 1971-72 and 1972-73. Abundance is expressed as a function of estimated age. Fitted exponential functions give estimates of the instantaneous rates of decline in abundance for eggs and larvae up to 31 days of age. The two symbols enclosed in circles represent nonfully vulnerable length classes and were not included in the re- gression estimates of instantaneous decline. would have been lower. In 1971-72, Z = 0.0563 for fully vulnerable larval stages and Z = 0.1123 for those stages in 1972-73. The results suggest that egg and nonfully vulnerable larvae mortality were higher in 1971-72 than in 1972-73. Mortality of vulnerable larval stages appears to have been higher in 1972-73 when the population declined by 10. 6% /day as opposed to 1971-72 when it declined only 5.5%/day. The higher mortality rate of larvae in 1972-73 also was apparent in the mor- tality estimates based on larval lengths (Fig- ure 11). High mortality of eggs or newly hatched larvae may be characteristic of many clupeids, including round herring. Smith (1973) recently reported that Pacific sardine eggs experience high mortal- ity, the instantaneous rate being Z = 0.31 during that stage. Pilchard, Sardina pilchardus, eggs undergo high mortality during early embryonic stages (Southward and Demir 1974) and embryos ofClupeonella delicatula suffered high mortality, especially under unfavorable temperature re- gimes (Pinus 1974). The best probable estimates of mortality from the egg to 16.0-mm larval size are near the middle of the ranges given in Table 11, at instantaneous growth rates of 0.0498. In 1971-72, Z = 0.1317 is the most probable estimate while Z = 0.1286 seems most probable in 1972-73. These estimates correspond to average daily losses of 12.3% in 1971-72 and 12.1% in 1972-73. Estimates of the instantaneous mortality coefficients based on the two examples given in Table 10 and Figure 12 coincide with what I believe may be the best esti- mates of mortality. Confidence limits, at the 0.95 probability level, were placed on the instantane- ous mortality coefficients derived from these examples. They were wide, ranging from Z = 0.0635-0.1999 in 1971-72 andZ = 0.0823-0.1749 in 1972-73. The coefficients Z = 0.1317 in 1971-72 and Z = 0.1286 in 1972-73 did not differ sig- nificantly between years U-test; P>0.50). The estimates of mortality rates could be too high if avoidance by larvae was increasing sig- nificantly as they grew, reducing their probability of capture. If growth was not exponential, but linear, during the larval phase, then the mortality estimates may be too low, because duration- corrected abundances gave relatively high values to older larvae that presumably were growing through length classes at an increasing rate. Because of the difficulty in ageing eggs or larvae of marine fishes, few estimates of mortality rates in relation to age have been reported. Ahlstrom (1954) reported that about one Pacific sardine larva survived to 21.25 mm/100,000 eggs spawned during the first 40-45 days of life, which corre- sponds to an instantaneous daily loss rate of 0.16- 0.17. Japanese sardine was investigated by Nakai and Hattori (1962). They reported survival from egg to the 15.0 mm stage as 0.10% in 54 days, corresponding to an instantaneous rate of Z = 0.1279. This rate is nearly identical to that which 84 HOUDE; ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING TABLE 11. — Summary of mortality estimates for round herring larvae from the eastern Gulf of Mexico, 1971-72 and 1972-73. Estimates were obtained from the exponential regression of egg and larvae abundances on mean age. Instantaneous growth and mortality coefficients were calculated for various possible combinations of mean daily growth increment and duration of the nonfully vulnerable larval stages. Egg stage duration was assumed to be 2.0 days. Nonfully vulnerable larval stages were 2.1-5.0 mm SL in 1971-72 and 2.1-4.0 mm SL in 1972-73. Explanation of the estimating method is given in Equations (12)-(16). Season Mean daily growth increment (mm) Instantaneous growth coefficient (g) Nonfully vulnerable larvae duration (days) Instantaneous mortality coefficient (2) /-axis intercept, N a (no. x 10") Daily mortality rate, 1 - exp(-Z) 1971-72 1972-73 0.3 0.0299 4.0 0.0866 103.25 0.0830 0.3 0.0299 5.0 0.0866 112.07 0.0830 0.3 0.0299 6.0 0.0866 121.40 0.0830 0.3 0.0299 7.0 0.0866 131.21 0.0829 0.5 0.0498 4.0 0.1331 186.35 0.1246 0.5 0.0498 5.0 0.1325 208.29 0.1241 0.5 0.0498 6.0 0.1317 231.46 0.1234 0.5 0.0498 7.0 0.1307 255.74 0.1225 0.7 0.0698 4.0 0.1739 285.65 0.1596 0.7 0.0698 5.0 0.1718 324.45 0.1579 0.7 0.0698 6.0 0.1693 364.72 0.1558 0.7 0.0698 7.0 0.1665 406.00 0.1534 0.3 0.0299 1.5 0.0842 71.56 0.0808 0.3 0.0299 2.0 0.0840 73.89 0.0805 0.3 0.0299 2.5 0.0837 76.26 0.0803 0.3 0.0299 3.0 0.0835 78.68 0.0801 0.5 0.0498 1.5 0.1303 114.55 0.1222 0.5 0.0498 2.0 0.1295 119.80 0.1214 0.5 0.0498 2.5 0.1286 125.12 0.1207 0.5 0.0498 3.0 0.1278 130.52 0.1200 0.7 0.0698 1.5 0.1719 160.03 0.1580 0.7 0.0698 2.0 0.1702 168.78 0.1565 0.7 0.0698 2.5 0.1683 177.58 0.1549 0.7 0.0698 3.0 0.1665 186.39 0.1533 is most probable for round herring larvae. Hard- ing and Talbot (1973) and Bannister et al. (1974) reviewed the results of several years' investiga- tions on plaice, Pleuronectes platessa. They found that instantaneous mortality coefficients varied from only 0.0209 to 0.0685 from egg stage 1 to larval stage 4 during the long larval life of more than 150 days. Mortality of haddock eggs and lar- vae was reported by Saville (1956), who gave a series of estimates that ranged from 4 to 16%/day (Z = 0.04-0.17) during a 4-yr survey of egg and larvae abundance at Faroe. Jack mackerel, Trachurus symmetricus, larvae have a high rate of mortality (Lenarz 1973), losses ranging from 57 to 67% per millimeter of growth. Farris (1961) re- ported mortality of jack mackerel larvae in rela- tion to age. The instantaneous mortality rate, cal- culated from his data, was 0.23 during the first 30 days of life. Mortality of Japanese mackerel, Scomber japonicus, larvae was very high (Watanabe 1970), 99.95% mortality having occur- red between the egg and 15-mm larval stage in about 23 days. This corresponds to an instantane- ous rate of Z — 0.3295. Round herring larval mor- tality rates apparently are similar to those of other clupeoids from temperate or subtropical marine waters (Ahlstrom 1954; Nakai and Hattori 1962; Lenarz 1973). On average they are slightly higher than those reported for haddock (Saville 1956). Round herring larvae have mortality rates that are much higher than those reported for North Sea plaice larvae and lower than those reported for jack mackerel or Japanese mackerel larvae. If any period can be considered critical in the early life of round herring, it must occur between the time that eggs are spawned and when larvae reach 5.5 mm long. Greatest losses occurred at that time in 1971-72 and 1972-73 (Figure 12). Abundance estimates declined by more than 92% between the egg and 5.5-mm larvae in 1971-72. A decline of more than 78% in abundance was esti- mated between egg and 5.5-mm larvae in 1972-73 (Table 12, Figure 12). For larvae longer than 5.5 mm mortality decreased, the decrease in rate being especially great in 1971-72. The number of survivors and percentage survi- val of round herring larvae at various stages were estimated (Table 12) from the number of spawned eggs obtained by Method I and the information on growth and mortality that is summarized in Table 1 1 . The Method I spawning estimate was assumed to be a better estimate of initial number of eggs than they- intercept estimates in Table 11. There was an apparent high mortality between spawn- ing and hatching which exceeded 75% in 1971-72 (Table 12). The larval populations were reduced by 85 FISHERY BULLETIN: VOL. 75. NO. 1 TABLE 12. — Estimated numbers and percentages of survivors of round herring larvae at hatching, 5.5 mm SL and 15.5 mm SL in 1971-72 and 1972-73. Estimates are made for three possible growth rates (see Table 11). Duration of the nonfully vulnerable larval stages was set at 6.0 days for 2.1-5.0 mm larvae in 1971-72 and 2.5 days for 2.1-4.0 mm larvae in 1972-73. The number of spawned eggs in each year was based on estimates by Method I (Table 5). Predicted numbers at hatching, 5.5 mm and 15.5 mm are calculated from exponential functions based on Table 11 data. Season Instantaneous growth coefficient (a) Number of spawned eggs (x 10 11 ) Instantaneous mortality coefficient (Z) Number hatching (x 10") % mortality to hatching' Number of 5.5-mm larvae (x 10") % mortality to 5.5 mm N 15.5 ( umber of -mm larvae x 10") % mortality to 15.5 mm 1971-72 1972-73 0.0299 0.0498 0.0698 0.0299 0.0498 0.1683 1 ,064.20 1,064.20 1 ,064.20 194.47 194.47 194.47 0.0866 0.1317 1693 0.0837 0.1286 0.1683 102.09 177.86 259.96 64.51 96.74 126.83 90.3 83.3 75.6 66.8 50.3 34.8 48.77 66.06 78.40 23.00 32.89 41.00 95.4 93.8 92.6 88.2 83.1 78.9 2.43 4.27 6.35 1.26 2.27 3.37 99.8 99.6 99.4 99.3 98.8 983 'Hatching assumed to occur at 2.0 days. more than 99.4% at 15.5 mm in 1971-72 and by more than 98.3% in 1972-73. The 15.5-mm stage would be attained at about 31 days if the instan- taneous growth coefficient was 0.0498 (equal 0.5-mm mean daily growth increment). At that growth rate, approximately 4 larvae/1,000 eggs spawned in 1971-72 and 12 larvae/1,000 eggs spawned in 1972-73 would have survived to 15.5 mm and 1 mo of age. SUMMARY 1) Surveys of eggs and larvae were used to inves- tigate spawning, to determine adult stock size, and to study aspects of the early life history of round herring in the eastern Gulf of Mexico during 1971-74. 2) Spawning takes place from mid-October to the end of May between the 30- and 200-m depth contours. About 60% of the total spawning occurred at depths greater than 50 m. Most spawn- ing apparently occurred during January and February. 3) Eggs occurred when surface temperatures ranged from 18.4° to 26.9°C, and surface salinities from 34.5 to 36.5%o. Larvae =s5.0 mm SL were collected when surface temperatures were from 20.5° to 26.9°C, and surface salinities from 34.1 to 36.8%o. Of the eggs 82.5% and of the ^5.0-mm larvae 87.5% were collected when surface temper- atures were from 21° to 26°C. More than 50% of the eggs and =£5.0-mm larvae were collected where surface salinity exceeded 36.0%o. 4) There is a major spawning area between lat. 27°00' and 28°00'N and long. 083°30' and 084°30'W. The center of the area is located about 150 km west by southwest of Tampa Bay in depths of 50-200 m. 5) The fecundity of eight round herring females 130-165 mm SL ranged from 7,446 to 19,699. Mean relative fecundity was 296.5 ova/g (S~ = 33.7). Gonads of round herring collected from Au- gust to November were ripening or near ripe. Those collected in June were spent. The sex ratio of 71 round herring adults did not differ sig- nificantly from 1:1. 6) The time from spawning to hatching, based on observations of development stages in planktonic eggs, was about 2.0 days at 22°C. 7) Adult biomass was determined by three methods from data on estimated annual spawn- ing. The Sette and Ahlstrom's (1948) and Simpson's (1959) techniques gave estimates that ranged from 130,000 to 715,000 metric tons in 1971-72 and 1972-73. The geometric mean of eight individual estimates by Saville's (1956) method was 181,200 metric tons, the arithmetic mean being 415,175 metric tons. But, the best estimates by Saville's method were from two individual cruises in midwinter. These were 673,481 metric tons in 1971-72 and 136,330 metric tons in 1972- 73. Those estimates were nearly the same as esti- mates obtained by the other two methods. Spawn- ing biomass apparently was higher in 1971-72 than in 1972-73. 8) The estimated concentration of biomass be- tween the 30- and 200-m depth contours, based on the stock size estimates, was from 67.6 to 120.0 kg/hectare in 1971-72 and from 5.9 to 28.3 kg/hec- tare in 1972-73. 9) The annual potential yield of round herring to a fishery, if instantaneous natural mortality coefficients lie in the range 0.5-1.0, ranged from 32,750 to 420,700 metric tons. The most probable mean annual potential yield estimates are in the range 50,000 to 250,000 metric tons. This is equiv- alent to 6.5-32.5 kg/hectare in the 30- to 200-m depth zone. 10) Total abundance of larvae was estimated in 1971-72 and 1972-73. The 4.1- to 5.0-mm length 86 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING class was nearly twice as abundant in 1972-73 as in 1971-72. Other length classes were somewhat more abundant in 1971-72 catches. 11) Mortality rates of larvae were estimated by length and for estimated ages. For lengths, the instantaneous coefficients of decline in catches wereZ = 0.2269 in 1971-72 and Z = 0.3647 in 1972-73, corresponding to 20.3 and 30.5% losses per millimeter of growth. For ages, a range of estimates of daily mortality, based on varying growth rates and nonfully vulnerable larva stage durations, was obtained. The most probable daily mortality estimates were Z = 0.1317 in 1971-72 and Z = 0.1286 in 1972-73, corresponding to per- centage losses of 12.3 and 12.1% on a daily basis. 12) It is probable that more than 99.4% mortal- ity from eggs to 15.5-mm larvae occurred in 1971- 72, and that more than 98.3% mortality occurred during that period in 1972-73. About 4 larvae/ 1,000 eggs spawned survived to 31 days and 15.5 mm in 1971-72, while about 12 larvae/1,000 eggs survived to that stage in 1972-73. ACKNOWLEDGMENTS This project was initiated as part of cooperative efforts to investigate biological and physical pro- cesses in the eastern Gulf of Mexico. Assistance was provided by many people and agencies. Par- ticular thanks go to Murice Rinkel of the State University System of Florida, Institute of Oceanography, for his help in coordinating EGMEX and Western Florida Continental Shelf cruises, as well as reduction of physical oceano- graphic data. The 1971 plankton surveys were coordinated with the National Marine Fisheries Service MARMAP program in the eastern Gulf of Mexico and special acknowledgments go to the following personnel: Ed Hyman, Larry Ogren, William J. Richards, Charles Roithmayr, and Stuart Smith. My students and technical person- nel deserve thanks for long hours spent at sea and tedious hours sorting and enumerating; among these are Steven Berkeley, Alfred Cardet, Reuben Charles, Ann and Nicholas Chitty, Lise Dowd, John Klinovsky, Walter Stepien, A. Keith Taniguchi, and Gregg Waugh. Harvey Bullis and Paul E. Smith of the National Marine Fisheries Service criticized earlier drafts of this paper, and I thank them for the suggestions and ideas that they provided. This research was sponsored by NOAA Office of Sea Grant, Department of Commerce, under Grant 04-3-158-27 to the University of Miami. LITERATURE CITED AHLSTROM, E. h. 1954. Distribution and abundance of egg and larval popu- lations of the Pacific sardine. U.S. Fish Wildl. Serv., Fish. Bull. 56:83-140. 1959a. Distribution and abundance of eggs of the Pacific sardine, 1952-1956. U.S. Fish Wildl. Serv., Fish. Bull. 60:185-213. 1959b. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. U.S. Fish Wildl. Serv., Fish. Bull. 60:107-146. 1968. An evaluation of the fishery resources available to California fishermen. In The future of the fishing indus- try of the United States, p. 65-80. Univ. Wash., Publ. Fish., New Ser. 4. ALVERSON, D. L. 1971. Manual of methods for fisheries resource survey and appraisal. Part I. Survey and charting of fisheries re- sources. FAO, Fish. Tech. Pap. 102, 80 p. ALVERSON, D. L., AND W. T. PEREYRA. 1969. Demersal fish explorations in the northeastern Pacific Ocean — an evaluation of exploratory fishing methods and analytical approaches to stock size and yield forecasts. J. Fish. Res. Board Can. 26:1985-2001. Anonymous. 1958. Gulf exploratory fishery program. Commer. Fish. Rev. 20(7):29-32. BANNISTER, R. C. A., D. HARDING, AND S. J. LOCKWOOD. 1974. Larval mortality and subsequent year-class strength in the plaice (Pleuronectes platessa L.l. In J. H. S. Blaxter (editor), The early life history of fish, p. 21- 37. Springer- Verlag, N.Y. BEVERTON, R. J. H. 1963. Maturation, growth and mortality of clupeid and engraulid stocks in relation to fishing. Rapp. P.- V. Reun. Cons. Perm. Int. Explor. Mer 154:44-67. BLACKBURN, M. 1941. The economic biology of some Australian clupeoid fish. Aust. Counc. Sci. Ind. Res. Bull. 138, 135 p. BULLIS, H. R., JR., AND J. S. CARPENTER. 1968. Latent fishery resources of the central West Atlantic region. In The future of the fishing industry of the Unit- ed States, p. 61-64. Univ. Wash., Publ. Fish., New Ser. 4. BULLIS, H. R., JR., AND J. R. THOMPSON. 1967. Progress in exploratory fishing and gear research in region 2 fiscal year 1966. U.S. Fish Wildl. Serv., Circ. 265, 14 p. CUSHING, D. H. 1957. The number of pilchards in the Channel. Fish. In- vest. Minist. Agric. Fish Food (G.B.), Ser. II, 21(5), 27 p. De la Campa de Guzman, S., and J. M. Ortiz Jiminez. 1975. Distribucion y abundancia de larvas de peces en el Golfo de California durante abril-mayo de 1973, con espe- cial referenda a sardina monterrey y japonesa. Inst. Nac. Pesca. INP/SC:11, 25 p. Dryfoos, R. L., R. P. Cheek, and R. L. Kroger. 1973. Preliminary analyses of Atlantic menhaden, Bre- voortia tyrannus, migrations, population structure, 87 FISHERY BULLETIN: VOL. 75, NO. 1 survival and exploitation rates, and availability as indi- cated from tag returns. Fish. Bull., U.S. 71:719-734. FARRIS, D. A. 1961. Abundance and distribution of eggs and larvae and survival of larvae of jack mackerel (Trachurus symmet- ricus). U.S. Fish Wildl. Serv., Fish. Bull. 61:247-279. FOOD AND AGRICULTURE ORGANIZATION. 1974. Catches and landings, 1973. FAO, Yearb. Fish. Stat. 36, 590 p. Fore, p. l. 1971. The distribution of the eggs and larvae of the round herring, Etrumeus teres, in the northern Gulf of Mexico. (Abstr.) Assoc. Southeast. Biol. Bull. 18:34. GULLAND, J. A. 1971. The fish resources of the ocean. Fishing News (Books) Ltd., Surrey, Engl., 255 p. 1972. The scientific input to fishery management deci- sions. In Progress in fishery and food science, p. 23-28. Univ. Wash., Publ. Fish., New Ser. 5. Harding, D., and J. W. Talbot. 1973. Recent studies on the eggs and larvae of the plaice (Pleuronectes platessa L.) in the Southern Bight. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 164:261-269. HOLDEN, M. J., AND D. F. S. RAITT. 1974. Manual of fisheries science. Part 2. Methods of re- source investigation and their application. FAO, Fish. Tech. Pap. 115, 214 p. HOUDE, E. D. 1973a. Estimating abundance of sardine-like fishes from egg and larval surveys, eastern Gulf of Mexico: prelimi- nary report. Proc. Gulf Caribb. Fish. Inst., 25th Annu. Sess., p. 68-78. 1973b. Some recent advances and unsolved problems in the culture of marine fish larvae. Proc. World Maricult. Soc. 3:83-112. 1974. Research on eggs and larvae of fishes in the eastern Gulf of Mexico. In R. E. Smith (editor), Proceedings of marine environmental implications of offshore drilling in the eastern Gulf of Mexico, p. 187-204. State Univ. Syst. Fla., Inst. Oceanogr., St. Petersburg. 1976. Abundance and potential for fisheries development of some sardine-like fishes in the eastern Gulf of Mexico. Proc. Gulf Caribb. Fish. Inst., 28th Annu. Sess., p. 73-82. HOUDE, E. D., S. A. BERKELEY, J. J. KLINOVSKY, AND C. E. DO WD. 1976. Ichthyoplankton survey data report. Summary of egg and larvae data used to determine abundance of clupeid fishes in the eastern Gulf of Mexico. Univ. Miami Sea Grant Tech. Bull. 32, 193 p. HOUDE, E. D., AND N. CHITTY. 1976. Seasonal abundance and distribution of zoo- plankton, fish eggs and fish larvae in the eastern Gulf of Mexico, 1972-74. U.S. Dep. Commer., NOAA Tech. Rep. NMFSSSRF-701, 18 p. HOUDE, E. D., AND P. L. FORE. 1973. Guide to identity of eggs and larvae of some Gulf of Mexico clupeid fishes. Fla. Dep. Nat. Resour., Mar. Res. Lab., Leafl. Ser. 4(23), 14 p. HOUDE, E. D., W. J. RICHARDS, AND V. P. SAKSENA. 1974. Description of eggs and larvae of scaled sardine, Harengula jaguana. Fish. Bull., U.S. 72:1106-1122. HOUDE, E. D., AND L. J. SW ANSON, JR. 1975. Description of eggs and larvae of yellowfin menha- den, Brevoortia smithi. Fish. Bull., U.S. 73:660-673. ITO, S. 1968. Observations on the ovarian ova of the round her- ring, Etrumeus micropus (Temminck et Schlegel). Bull. Jap. Sea Reg. Fish. Res. Lab. 19:11-17. KHROMOV, N. S. 1969. Distribution of plankton in the Gulf of Mexico and some aspects of its seasonal dynamics. In A. S. Bogdanov (editor), Soviet-Cuban fishery research, p. 36-56. VNIRO, TsRI, 1965. (Translated from Russ. by Isr. Program Sci. Transl., available U.S. Dep. Commer., Clgh. Fed. Sci. Tech. Inf., as TT 69-59016.) LENARZ, W. H. 1973. Dependence of catch rates on size of fish larvae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 164:270-275. Martinez, S., and E. D. houde. 1975. Fecundity, sexual maturation, and spawning of scaled sardine (Harengula jaguana Poey). Bull. Mar. Sci. 25:35-45. Matsuura, Y. In press. A study of the life history of Brazilian sardine, Sardinella brasiliensis . IV. Distribution and abundance of sardine larvae. Bol. Inst. Oceanogr. (Sao Paulo). MITO, S. 1961. Pelagic fish eggs from Japanese waters — I. Clupeina, Chanina, Stomiatina, Myctophida, Anguillida, Belonida and Syngnathida. [In Jap., Engl, summ.] Sci. Bull. Fac. Agric, Kyushu Univ. 18:285-310. MOSER, H. G., E. H. AHLSTROM, D. KRAMER, AND E. G. STE- VENS. 1974. Distribution and abundance of fish eggs and larvae in the Gulf of California. Calif. Coop. Oceanic Fish. In- vest. Rep. 17:112-128. NAKAI, Z., AND S. HATTORI. 1962. Quantitative distribution of eggs and larvae of the Japanese sardine by year, 1949 through 1951. Bull. Tokai Reg. Fish. Res. Lab. 9:23-60. O'TOOLE, M. J., AND D. P. F. KING. 1974. Early development of the round herring, Etrumeus teres (de Kay) from the South East Atlantic. Vie Milieu, Ser. A, 24:443-452. PINUS, G. N. 1974. Some factors influencing early survival and abun- dance of Clupeonella in the Sea of Azov. In J. H. S. Blaxter (editor), The early life history of fish, p. 81-86. Springer- Verlag, N.Y. RICHARDS, W. J., AND B. J. PALKO. 1969. Methods used to rear the thread herring, Opis- thonema oglinum, from fertilized eggs. Trans. Am. Fish. Soc. 98:527-529. RINKEL, M. O. 1974. Western Florida continental shelf program. In R. E. Smith (editor), Proceedings of marine environmental implications of offshore drilling in the eastern Gulf of Mexico, p. 97-126. State Univ. Syst. Fla., Inst. Oceanogr., St. Petersburg. SAKSENA, V. P., C. STEINMETZ, JR., AND E. D. HOUDE. 1972. Effects of temperature on growth and survival of laboratory-reared larvae of the scaled sardine, Harengula pensacolae Goode and Bean. Trans. Am. Fish. Soc. 101:691-695. SALNIKOV, N. E. 1969. Fishery research in the Gulf of Mexico and the Caribbean Sea. In A. S. Bogdanov (editor), Soviet- Cuban fishery research, p. 78-171. VNIRO TsRI, 1965. (Translated from Russ. by Isr. Program Sci. Transl., 88 HOUDE: ABUNDANCE AND POTENTIAL YIELD OF ROUND HERRING available U.S. Dep. Commer., Clgh. Fed. Sci. Tech. Inf., as TT 69-59016.) Saville, A. 1956. Eggs and larvae of haddock {Gadus aeglefinus L.) at Faroe. Scott. Home Dep. Mar. Res. 1956(4), 27 p. 1964. Estimation of the abundance of a fish stock from egg and larval surveys. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 155:165-170. SCHAAF, W. E., AND G. R. HUNTSMAN. 1972. Effects of fishing on the Atlantic menhaden stock: 1955-1969. Trans. Am. Fish. Soc. 101:290-297. SETTE, O. E., AND E. H. AHLSTROM. 1948. Estimations of abundance of the eggs of the Pacific pilchard (Sardinops caerulea) off southern California dur- ing 1940 and 1941. J. Mar. Res. 7:511-542. SIMPSON, A. C. 1959. The spawning of the plaice (Pleuronectes platessa) in the North Sea. Fish. Invest. Minist. Agric. Fish. Food (G.B.), Ser. II, 22(7), 111 p. Smith, p. e. 1973. The mortality and dispersal of sardine eggs and lar- vae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 164:282- 292. Smith, p. e., and S. Richardson (editors). In press. Manual of methods for fisheries resource survey and appraisal. Part 4. Standard techniques for pelagic fish egg and larvae survey. FAO, Rome. Southward, a. j., and n. demir. 1974. Seasonal changes in dimensions and viability of the developing eggs of the Cornish pilchard (Sardina pilchar- dus Walbaum) off Plymouth. In J. H. S. Blaxter (editor), The early life history of fish, p. 53-68. Springer- Verlag, N.Y. TAFT, B. A. 1960. A statistical study of the estimation of abundance of sardine (Sardinops caerulea) eggs. Limnol. Oceanogr. 5:245-264. TANAKA, S. 1960. Studies on the dynamics and the management offish populations. Bull. Tokai Reg. Fish. Res. Lab. 28:1- 200. UCHIDA, K., S. IMAI, S. MITO, S. FUJITA, M. UENO, Y. SHOJIMA, T. SENTA, M. TAHUKU, AND U. DOTU. 1958. Studies on the eggs, larvae and juvenile of Japanese fishes. Series I. [In Jap.] 2d Lab. Fish. Biol., Fish. Dep., Fac. Agric, Kyushu Univ., 89 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. 62:1-283. Watson, W., and J. M. leis. 1974. Ichthyoplankton of Kaneohe Bay, Hawaii. Univ. Hawaii, UNIHI-Sea Grant Publ. TR-75-01, 178 p. WHITEHEAD, P. J. P. 1963. A revision of the recent round herrings (Pisces: Dus- sumieriidae). Bull. Br. Mus. (Nat. Hist.) Zool. 10:305- 380. WISE, J. P. 1972. U.S. fisheries: A view of their status & poten- tial. Mar. Fish. Rev. 34(7-8):9-19. 89 REPRODUCTIVE BIOLOGY OF THE FEMALE DEEP-SEA RED CRAB, GERYON QUINQUEDENS, FROM THE CHESAPEAKE BIGHT 1 2 Paul A. Haefner, Jr. 3 ABSTRACT Collections of the deep-sea red crab, Geryon quinquedens, were made at depths from 270 to 1,300 m in the vicinity of Norfolk Canyon in the northwest Atlantic Ocean in November 1974, September 1975, and January 1976. The gross morphology and histology of ovary development are described. The size range in which relative growth of the abdomen changes is associated with maturation of the vulvae, copulation and insemination, gonad development, and egg extrusion. Females become sexually mature within the intermolt size range 65-75 mm carapace length (80-91 mm carapace width). Most intermolt females s*76 mm carapace length show signs of copulation and insemination, and their ovaries are in intermediate to advanced stages of development. Few females <75 mm are ovigerous. Historically the red crab, Geryon quinquedens Smith, has been seldom utilized commercially (Schroeder 1959; McRae 1961). Explorations have established that red crabs can readily be captured by pot or trap fishing in many regions along the eastern United States. The commercial potential of this crab has spurred investigations of the general biology and distribution (Le Loeuff et al. 1974; Haefner and Musick 1974; Wigley et al. 1975; Gray 4 ; Dias and Machado 5 ; Ganz and Herrmann 6 ) as well as technological and economic aspects of harvesting and processing (Meade and Gray 1973; Holmsen and McAllister 1974). The present study was prompted by recognition that biological data on sexual maturity are re- quired for proper management of red crab stocks. This paper presents data on collections from Chesapeake Bight and deals with various aspects of reproductive biology of the female crab: ovary development, size composition of catch, size of 'Research cruises supported by National Science Foundation Grant GA-37561, J. A. Musick, principal investigator, and by the University of Virginia Institutional Grant Program for P. A. H. participation. Contribution No. 777, Virginia Institute of Marine Science, Gloucester Point, VA 23062. 'Virginia Institute of Marine Science, Gloucester Point, VA 23062. "Gray, G. W., Jr. 1969. Investigation of the basic life history of the red crab (Geryon quinquedens). R.I. Div. Conserv. P.L. 88- 309, Proj. 3-46-R Completion Rep., 36 p. 5 Dias, C. A., and J. S. Machado. 1974. Preliminary report on the distribution and relative abundance of deep-sea red crab (Geryon sp.) off Angola. Sci. Pap. No. 26, 12 p. In Scientific papers presented to the second session of the International Commission for the Southeast Atlantic Fisheries (Madrid, December 1973). Publ. Mimeogr. M. E. Bioceanol. Pescas, Angola 12, 75 p. 6 Ganz, A. R., and J. F. Herrmann. 1975. Investigations into the southern New England red crab fishery. R.I. Dep. Nat. Resour. Div. Fish. Wildl. Mar. Fish. Sect., 78 p. ovigerous individuals, abdomen width-carapace length relationship, development of vulvae, and evidence of copulation and insemination. METHODS Red crabs were collected at depths from 270 to 1,300 m in Norfolk Canyon and vicinity (lat. 36°32'-37°10'N; long. 74°10'-74°46'W) in Novem- ber 1974 (RV James M. Gilliss 74-04), September 1975 (RV James M. Gilliss 75-08), and January 1976 (RV James M. Gilliss 76-01). Based on the recommendations of Gray (see footnote 4), all female crabs were measured for short carapace length (CL, distance from the diastema between the rostral teeth to the posterior edge of the carapace, along the midline); width of the fifth abdominal segment was recorded for 190 crabs. Carapace length may be converted into carapace width (CW) by using the equation CW = 11.04 + 1.06CL, r = 0.98, based on measurements of 268 female crabs. Pleopods and vulvae were examined to deter- mine if mating and egg extrusion had occurred. Eggs or egg remnants or their absence on pleopods, variations in the size, shape and physi- cal condition of vulvae, and the relative size of seminal receptacles were noted. Selected samples of the spermathecal fluid were withdrawn directly from incisions in the receptacle and examined mi- croscopically for presence of sperm or spermato- phores. Ovaries were initially classified to relative size following the scheme used for the rock crab, Cancer irroratus (Haefner 1976). The scheme for Manuscript accepted June 1976. FISHERY BULLETIN: VOL. 75, NO. 1, 1977. 91 FISHERY BULLETIN: VOL. 75, NO. 1 red crabs was quantified by measuring ovary volume and deriving gonad indices (Giese and Pearse 1974) for the various stages. Certain ovar- ian samples were selected on the basis of relative size and color and treated in the following manner. Displacement of ovaries was measured by placing the entire, excised ovary in volumetrically graduated tubes containing a known quantity of seawater. Ovary volume (V in milliliters) was used to compute a gonad index: G, = (Ovary weight)/* Total body weight) x 100, where weights in grams were calculated as follows Ovary weight = 1.025 V , assuming ovarian specific gravity equals that of seawater. Total body weight was derived from the following relationship based on measurements of 142 females: log body weight = -3.134 + 2.8833 log length, r = 0.968. Portions of the ovaries were then preserved in Davidson's fixative for histological processing and in Gilson's fluid (Bagenal and Braum 1971) for measurement of ova size. Histological sections were stained in haematoxylin and eosin and mounted in Per- mount. 7 Descriptions of developmental stages were made from the resultant slides. Samples in Gilson's fluid were shaken to release ova which were then observed with a dissecting 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. microscope. The diameters of 20 spherical ova from each sample were measured with a cali- brated ocular micrometer. Misshapen ova were not considered. Similarly, 20 extruded eggs from 11 ovigerous crabs were removed and measured (length and width). A mean diameter was com- puted for each crab. RESULTS AND DISCUSSION The Ovary The following account of the gross morphology and histology of the red crab ovary is based on examination of the gross anatomy of 255 crabs and on histological preparations from 34 crabs. The ovary is an H-shaped organ located dorsally just beneath the carapace (Figure 1). Two horns extend anterolateral^ from either side of the gastric mill and lie dorsal to the hepatopancreas. At the posterolateral borders of the gastric mill, near the origin of the posterior mandibular muscle bundles, the anterior horns are joined by a commissure. Two posterior horns, which lie ven- tral to the heart, extend posteriorly on either side of the intestine. The seminal receptacles arise from the midlateral border of the posterior horns and open externally through gonopores (vulvae) on thoracic sternite VI, immediately adjacent to sternite V. FIGURE 1. — Dorsal dissection of female Geryon quinquedens. Heart and medial portion of branchial chamber removed. Anterior (aov), posterior (pov) and commissure (cov) of ovary, gastric mill (g), gill (br), intestine (i), hepatopancreas (hp), seminal receptacle (sr), midgut caeca (mc). 92 HAEFNER: REPRODUCTIVE BIOLOGY OFGERYONQUINQUEDENS Very Early Development In very early development (Table 1 ), the ovary is small ( <0.2 ml in volume; horn width 0.5 mm) and colorless. A central lumen is not apparent from gross morphological examination, although the precursor of one is indicated in Figure 2. Lobation is not obvious in this stage. The bulk of the organ consists of fibrous connective tissue, apparently stratified, and blood sinuses (Figure 2A). The outer connective tissue wall of the ovary is not readily distinguishable from the inner connective tissue. Various cell types are present. Most cells contain one oval nucleus (7.2 fim long) while other larger, less numerous cells have a large round nucleus (7.2 /xm in diameter). Ova diameters are small (40-172 /xm) and confined to germinative areas or strands. In some instances, it is difficult to free the ova from the surrounding tissue even after treatment in Gilson's fluid. The germinal zone consists of columnar cells with (12 /xm) elongate nuclei (Figure 2B). throughout the ovary. Cells in an early stage of oogenesis, recognizable by vacuolate nuclei (Fig- ure 3B), are small (14-53 /urn) compared with the more advanced ova (74-278 /xm) characterized by more compact nuclei and the presence of cytoplasmic yolk granules (Figure 3C, D). They are surrounded by a single layer of follicular cells (Figure 3D) which are spindle shaped with an elongate nucleus (72 /xm). Intermediate Stage As the ovary progresses to the intermediate stage of development, accumulating yolk, it gradually occupies more space (G, = 1.4-2.7) in the visceral cavity and changes color (Table 1). The ovarian architecture is little changed from that of earlier stages; connective tissue is confined to the margin of the ovary and to the interstices between the now obvious lobes. Germinative zones are present. Ova are larger (1 12-537 /xm) than those in earlier stages. Early Development White, ivory, light gray, or light yellow ovaries which are small (0.2-2.0 ml volume, 2-6 mm horn width) may exhibit histological development in advance of the previous stage. Most of the organ is filled with ova in various early stages of de- velopment (Figure 3A). Connective tissue is still prevalent around the margin, penetrating the ovary in numerous locations to form small lobes which are not readily visible from a gross mor- phological aspect. The germinal zone is well defined and branches Mature Stages A fully mature ovary nearly obscures the hepatopancreas in dorsal view. Only a small por- tion of the hepatopancreas and the slightly coiled midgut caecae are visible between the ovary and branchial chamber (Figure 1). The high gonad indices (>2. 7) attest to the large volume (8-32 ml) of the organ at these stages of development. The color remains variable but is generally darker than that of earlier stages as reddish and brownish hues become evident (Table 1). The predominant histological feature in a TABLE 1. — Descriptive stages of Geryon quinquedens ovary: color variation, horn size, volume, gonad index, and ova diameter. Stage of ovary Color of ovary Horn width range (mm) n Ant. Post. Ovary volume (ml) n X Range Gonad index Ova diamet n X er (nm) development n X Range Range Very early Colorless, white, ivory 7 0.5-2.2 0.5-1.3 8 <0.2 0.1-0.2 8 0.29 0.09-0.88 3 102 49-172 Early White, ivory, light gray, light yellow 12 2-6 2-6 15 1.1 0.2-2 15 0.75 0.19-1.75 10 168 74-278 Intermediate Ivory, white, light yellow, yellow, yellowish orange, light brownish orange 7 8-15 6-10 12 5.2 4.5-7 12 205 1.45-2.73 10 289 112-537 Advanced Yellow, yellowish orange, brownish orange, reddish brown, brownish purple 4 16-23 6-12 6 13.4 8-12 6 4.24 2.74-6.02 6 508 298-666 Mature Yellowish orange, orange, brownish orange, brownish purple 12 20-32 10-18 10 28.9 21-32 11 8.22 6.00-11.85 9 611 484-788 Redeveloping Ivory, yellowish orange, light brownish orange, reddish brown, reddish orange, brownish purple 6 8-20 5-7 14 9.0 2.5-21 14 2.67 1.04-7.25 16 347 148-671 93 FISHERY BULLETIN: VOL. 75, NO. 1 f ! v life -■ M «* ' IS © #»y* *. €> .# * I 4. $5 ^=75 mm were in intermediate to advanced stages of ovarian development. Early developmental stages can occur in large crabs, particularly after recent ovulation. This is evident from the distribution of ovigerous crabs and those with egg remnants on the pleopods. Such ovaries, in redevelopment stages, can recede to early developmental stages. Size at Sexual Maturity Hartnoll (1969) regarded a crab as mature "when it enters the intermolt during which it is first able to copulate successfully." It is generally accepted that in brachyurans maturity in some females cannot be determined from the condition of the gonads because development and ovulation often occur a considerable time after mating. In the case of red crabs, several criteria were examined in an effort to define the size (age) at which females mature. These included the size distribution of ovigerous and nonovigerous fe- males, the incidence of physical indicators of copu- lation, and changes in the features of the vulvae and abdomen. Ovigerous Females The size-frequency distribution of 755 females captured in November 1974, September 1975, and January 1976 reveals the incidence of ovigerous individuals and those with egg remnants on the pleopods (Figure 6). In November and September, 27.3% and 15.7%, respectively, of females 3=71 mm CL (97 mm CW) were ovigerous; 9.0% of females 2*71 mm in September carried egg remnants. In January, 25.5% of females 2=71 mm CL were berried; two of these showed some evidence of egg hatching. Most (94%) of the ovigerous individuals and those with egg remnants were between 71 and 1 13 mm CL (97-131 mm CW); only four crabs were smaller. Physical Evidence of Copulation In numerous species of crabs, recent copulation by the female is indicated by the presence of a hardened mass of spermatozoa and associated secretions protruding from the vulvae (Hartnoll 1969). This so-called sperm plug does not occur in Geryon quinquedens . The exoskeletons of red crabs that have not recently molted are blackened or discolored in abraded or damaged areas and are usually in- fested with lepadid barnacles Trilasmis sp. The association of lepadids and discoloration serves as an indicator of a time lapse since the last molt, although the exact length of time cannot presently be determined. It was reasoned that abrasion and damage of vulval margins due to copulation would result in similar discoloration. This was verified 96 HAEFNER: REPRODUCTIVE BIOLOGY OFGERYON QUINQUEDENS MM •4k 3» - FIGURE 4. — Redeveloping ovary of Geryon quinquedens from ovigerous crab. A. 25 x . Germinative zone (g) and developing ova are evident. B. Higher magnification (125x) showing prevalence of fibrous connective tissue (f) among various sizes of developing ova. 97 FIGURE 5.— Distribution of female Geryon quinquedens according to size (carapace length) and stage of ovarian development. November 1974 and September 1975 samples pooled. Black areas indicate ovigerous crabs and those with egg remnants on pleopods. VERY EARLY n^l nfJl t*\ hr^lrnn D ,-D , n r— i FISHERY BULLETIN: VOL. 75, NO. 1 EARLY ^ r, nBp U 30 40 q n ,- jliui^ INTERMEDIATE , "1 , , 5-1 N = 50 -0— P^r-^r^ jlfrh 5 , N=22 MATURE SHORT CARAPACE LENGTH (mm) 10 OVIGEROUS NOVEMBER 1974 n^Hrn h^V^ N=208 p n , n , < > Q Z o cr LU CD 15 10 SEPTEMBER 1975 tL 20 15 10 5 OVIGEROUS EGG REMNANTS H 1 1 - 1 — I 1 ' T n , n , n JANUARY 1976 OVIGEROUS Un. u ^M fl H J ~ L | S N^332 N--2I5 30 40 50 60 70 80 90 100 110 120 SHORT CARAPACE LENGTH (mm) FIGURE 6.— Size-frequency distribution of female Greyon quinquedens captured in November 1974 (a), September 1975 (b), and January 1976 (c). Ovigerous individuals are indicated in black; those with egg remnants on pleopods by horizontal stripes. by examining the spermathecal contents of 67 crabs with discolored vulvae (14 with extruded eggs, egg remnants, or damaged pleopods and 53 with clean, intact pleopods). Eleven (79%) of the recently ovulated females (78-103 mm CD and 47 (89%) females with clean pleopods (45-105 mm CD contained sperm (Figure 7). Twenty-one crabs (50-75 mm CD with immature vulvae were similarly examined; none had sperm in the spermathecae. Another 17 crabs (50-72 mm CL) with immature vulvae were not examined for the presence of sperm because the spermathecae were undeveloped; only the tubular vagina was present between the ovary and gonopore. Blackened vulval margins may be used as a criterion to indicate that copulation of the female crab has occurred, if other obvious signs (eggs or remnants) are absent. The 89% incidence among nonovigerous females supports this contention. The 79% incidence among ovulated females is low, 98 HAEFNER: REPRODUCTIVE BIOLOGY OFGERYON QUINQUEDENS FIGURE 7. — Isolated sperm from spermatheca of 83-mm CL Geryon quinquedens. Nonmobile processes extend from nuclear region surrounding a central, refringent structure, most likely the acrosome (Brown 1966). Interference microscopy, 1300x. but expected. None of these crabs had swollen or turgid spermathecae of the type shown in Figure 1. In most cases, only residual quantities of semi- nal secretions were present in the receptacles, indicating that most of the deposit had been used in past ovulation(s) or absorbed. The presence of discolored vulval margins among large crabs suggested that they may pro- vide a physical criterion for copulation, similar to those demonstrated for other brachyurans ( Veillet 1945; Butler 1960; Hartnoll 1969). Vulval mar- gins of 93.5% of the females 2=70 mm CL examined (n = 328) were blackened (Figure 8). All females <70 mm CL had vulvae with intact margins. Not included in Figure 8 are an unusually small inseminated female (47 mm CL) and the ovigerous 64-mm CL specimen included in Figure 6b. One crab (47 mm CL) with small (1.2 mm long), but open, mature-type vulva was sperm positive. This unusually small crab had obviously mated but the vulval margins were not blackened. It is physically possible for a female this small to mate with a male of similar size. I have observed morphologically functional pleopods, with penis inserted in the first pair, on male crabs as small as 38 mm CL. The size at which males become physiologically mature is not known, but it must be relatively small. Change in Vulvae Although variable in form, vulvae of G. quinquedens undergo a recognizable growth and development pattern which parallels growth in body size and ovarian development. Six types are recognized (Figure 9). The first form vulvae (a) are slitlike and tightly closed. The observed size range appears to be related to crab length (Table 2). Form (b) vulvae are recurved, closed, and slightly larger than the longest form (a) vulvae. Forms (c) and (d), irregularly shaped and partially open, range from a size comparable to the largest vulvae of type (a) to that of type (e). Unusually large (d) vulvae (2.6 mm) were observed in a 78-mm CL crab. Form (e) vulvae are oval, gaping, and appear to immediately precede the mature vulva. Form (f) is the enlarged (2.4-3.9 mm), gaping, and usu- ally blackened vulvae of the larger, mated crabs. TABLE 2. — Incidence of vulval type and size range in relation to carapace length of female Geryon quinquedens. Type Carapace length (mm) Vulval length range n (mm) a 4 20-33 4 0.2-0.3 10 57-66 10 0.6-0.9 b 5 56-60 no data 15 61-74 8 0.7-1.2 c 9 50-60 5 0.5-0.8 17 61-74 13 0.8-1.5 d 8 61-72 6 0.7-1.3 1 78 1 2.6 e 3 47-60 3 0.6-1.2 9 61-72 7 0.8-1.2 f 1 45 1 3.0 51 70-103 12 2.4-3.9 Change in Abdomen Width The abdomen width (Y) to carapace length (X) relationship is allometric and is transformed to a straight line by the equation: log Y = -0.875 +1.321 logX, n = 251; r = 0.990 The relationship changes in the 60- to 75-mm CL range (Figure 10) so linear regressions were calculated separately for crabs with mature (f) vulvae: FIGURE 8.— Size-frequency dis- tribution of female Geryon quin- quedens with immature gonopores (white) and with discolored gonopore margins (black). November 1974, September 1975, and January 1976 collections pooled. BLACKENED VULVAL MARGINS O IMMATURE VULVAE 50 60 70 80 SHORT CARAPACE LENGTH (mm) 00 MO 120 99 FISHERY BULLETIN: VOL. 75, NO. 1 Wm I 3. *&S--2^ .^€ ... / Si**** • v -. ~ns^ *'**>, VI- "*--. O.I6mm 0.65mm 'V -^ 0.65 mm ^§& *!» /v* •sjlt ■-W 5 "-**** 0.65mm i* 5 *****.-**.; % m < *} > ■"• " . w -*««<*>*"'' 065mm r x-f f ' '" ^ji.T. ..■■■■ Jkh^'-' ■•■'"■ "A *'«~., •^-j- ->**■' .-— " 1.33mm FIGURE 9. — Structural variation in vulvae of female Geryon quinquedens. Portions of thoracic sternites V, VI, VII illustrated, a. First form, slitlike, from 20-mm CL crab. b. Recurved, closed, 66 mm CL. c and d. Irregular shape, partially open, 74-mm and 71-mm CL crabs, respectively, e. Oval, gaping, 68 mm CL. f. Oval, enlarged, with blackened margins, 90 mm CL. 100 HAEFNER: REPRODUCTIVE BIOLOGY OFGERYON QUINQUEDENS Y = -8.286 + 0.662X, n = 160; r - 0.943 and those with immature vulvae: Y = -8.512 + 0.64LY, n = 91; r = 0.971. The size range in which relative growth of the fifth abdominal segment changes is clearly as- sociated with the maturation of the vulvae, copulation and insemination, gonad development, and extrusion of eggs. Females become sexually mature within the intermolt size range 65-75 mm CL (80-91 mm CW). Most intermolt females 3=76 mm CL show signs of copulation and insemina- tion, and their ovaries are in intermediate to advanced stages of development. Few females <75 mm CL are ovigerous. ACKNOWLEDGMENTS I am indebted to the following personnel at Virginia Institute of Marine Science who con- tributed their expertise to the project: F. A. Perkins, photomicrography; Patsy Berry, micro- technique and photography; Peggy Peoples and Kay Stubblefield, art work; W. A. Van Engel, manuscript review; and those associated with the canyon cruises. LITERATURE CITED BAGENAL, T. B., and E. braum. 1971. Eggs and early life history. In W. E. Ricker (editor), Methods for assessment offish production in fresh water, p. 166-198. IBP (Int. Biol. Programme) Handb. 3. BROWN, G. G. 1966. Ultrastructural studies of sperm morphology and sperm-egg interaction in the decapod Callinectes sapidus. J. Ultrastruct. Res. 14:425-440. BUTLER, T. H. 1960. Maturity and breeding of the Pacific edible crab, Cancer magister Dana. J. Fish. Res. Board Can. 17:641- 646. CRONIN, L. E. 1942. A histological study of the development of the ovary and accessory reproductive organs of the blue crab, Callinectes sapidus Rathbun. M.S. Thesis, Univ. Maryland, College Park, 37 p. Q I- 2 s O UJ (0 £ C CD IE o .c V) S m c CD T3 c ra Is 11 CD > c «§ c — ro ro t/t CJ - IX Q. o Q >- x Q X co X "D C CO CO o CO c\i co *r r-. T-' ^ C\i in co cd cm d cm iioq ^ iod^t I C\J CO T CM CM i- C\i CO CO CM CM CD O CD CM t- t- t- CO C\i CO r^ CNJ CO d m- in cm co d d m o in T^ cvi T^ in in p 6 CD CO i- CNJ o m in c\i d -r- CD cm co cvi r^ ■*' r^ CM CD CM O CM O O I I ^ "-coon o co co in d cm cm d o ■» ocn .- cm in i-' 1 " o 1 5 1 1 1 cb ^ 1 in in ,_ »- *3- o CD i- ,— ,— T CM o o in CM r- o CO i- "~ o O d t- ■* CD o <=> h- ■* co o C\J o 1 1 o o o 1 1 b d d d 1 1 ■^ in co *- co in cm ^t o co d d d W d COCOOON t-^ d r-' d d o t- o m I o in ^ d d d ' d r^ m cd o in | o cm cm W ^ d I cm d I I I I I I I I I I »- cm in *- | »- en r-. | |inco o in cm o 0*- o in co o t^ d d d I ° Oi-n d d cm r- d d O ^3" I I *t 00 W cm I I i-" d o o ^r d N CM CD f^ CO 00 o 'tci^dci N CO CD CO N ^ CD O CD CM f a. CD T3 IX if OCOO) •- d CM *-' CM d co r^ cm -^r co CD ^ CM d CM w i- co 't o ^in t-' in ■* d co i: d T *ct oo co oo in ■- oj d d ^ oj r^ r-' S t- CD CO CO N CO CM CM CM "^ co cd d CD T3 CO C SI CD C T3 CD 2 -C E CO c ■;= ^ Cu E 8-!2 1 Cu *= col^ U J CD CO O w m > "cd ^ cr C CO CD $ 5 s "S c»«' u c?E E axOicoo. O 0) > > a) CD ? fiS in ro co c c m m u U rt !t c r U) o en en CO T3 O c CO o CO cu -C CO CO a 'b >> CO T3 S CO •s o ci-i CO CD 2 CO u C co CD c CD CD £ CD XI CO C o CO -1-t CO u J OQ < IX J>- x >- IX IX o o d d I in P P I P I I I CM O O O O CO CO ? in CDCOCMCO^t I CM CM i- i- i- I CO CD COO r- CMOOCMCO o o o o o o o o o o T- P o CO CD p *- o r^ o o d ^t C\i CM in >* CM CM C\J CO - T— b b b to b b in 1 b 1 >°. d d m C\J d d i- 1 o 1 m co m en CO ,_ CM en CM o ^ 1 ^ *" T_ T— T ~ ,— ^_ ^_ CM T _ ,_ CO f. o ,_ o 1 P d d O o d d o O o d ^ T3 CO CD CO S fi ja a. CD CD 2- t> X >- X •^■T-Tt'^-'^- COOOCOO N indcM->ci : - m di d <5 NSTtCOO Tt c\i ■•- c\i in CM o m o I ciriN ' ' i- CD coscDino CM i- >- i- CM CM t O CM d d d d d i- in s coco cm oo o m co i- in co i- in cm p p CM d^-dddwd O CO CO CO CM O CO oi Wi-doi do o o o o 1 I in O 1 1 in o d d o o 1 1 CM o 1 1 h~ d CM CD CM i- CM CO •" in CO o CM 1 CO CD CM O i- CM o o o co CD I o h- in d d O o d d o O 1 o d d 0) r~ .,_ CO p o ,_ CO 1 f~ o co CM d CO *T CM d '" ■* 1 CO d T-' CD CD t ^ •q; co CD CM o o o ■* d d N in d <6 c\i CD CD o d CM CM i- CM CO i- in CO o in in CO in CO CD CD in en CD CD t- CO T— T_ C0 CO CM o ,_ CO ,_ *■ ,_ y- co in d d o o d d '- o o o d d 00 o CO ,- CM CD CD CD CO CO CO CM d ro CM CD •* d 1^ CM CM <» 1- I- i_l -* *jw — a* ^r *— CO CD CO CD > _J < 106 PRISTAS and TRENT: CATCHES OF FISHES IN GILL NETS TABLE 4. — Statistical comparisons between catches from low- (S), mid- (M), and deep- (D) water depth zones. shal- Species Depth, mean catch, and significance lines' Error df Gulf menhaden Sea catfish Bluefish Yellowfin menhaden Little tunny Atlantic sharpnose shark Spanish mackerel Atlantic croaker Garftopsail catfish Hybrid menhaden Striped mullet Pinfish s M D 1.6 6.0 s 13.5 M D 1.0 1.5 2.3 D M s 0.3 1.5 2.2 M s D 2.6 2.7 5.5 S M D 0.2 2.5 3.3 S M D 0.9 2.3 3.1 S M D 0.5 1.1 1.4 D M S 0.3 0.5 3.5 S M D 0.0 0.6 M 2.4 s D 1.9 2.2 2.9 D M S 0.0 0.1 4.3 D M S 0.1 09 1.0 213 246 195 69 123 111 198 123 168 36 54 102 1 Any two means not underscored by the same line were significantly different at the 5% level. menhaden, little tunny, Atlantic sharpnose shark, Spanish mackerel, and gafftopsail catfish were caught in greater numbers as depth increased, and sea catfish were caught in greatest numbers in the deep zone. Conversely, catches decreased with increasing depth for bluefish, Atlantic croaker, striped mullet, and pinfish. Net Damage Monofilament nets were damaged less than multifilament nets in each depth zone fished. In terms of the amount of surface area damaged, shallow nets received the least and deep nets the greatest (Table 5). When corrected to percent of total webbing damage in nets at each zone, shal- TABLE 5. — Average daily net damage in square meters and percent of total net area in relation to depth of net and to webbing material. Depth of net Monofilament Multifilament (m) m 2 Percent m 2 Percent 1.5 0.11 0.21 0.16 0.33 3.0 0.16 0.16 0.23 0.23 6.1 0.31 0.15 0.44 0.22 Average of three nets 0.25 0.16 0.34 0.24 low nets received the greatest proportion of damage. Blue crab, Callinectes sapidus, caused damage to both webbing types. Multifilament webbing was damaged the most, possibly because 87% of all blue crabs taken were caught in multi- filament webbing. SUMMARY AND DISCUSSION In this study, catch per net was higher with monofilament than with multifilament gill nets; over 58% of the 12 most abundant species and over 71% of the 4 most abundant food and recreational fishes (bluefish, Spanish mackerel, Atlantic croaker, and striped mullet) were caught in mono- filament nets. Catch per net was much greater at night than during the day; about 93% of the 12 most abundant species and about 82% of the 4 most abundant food fishes were taken at night. Total catches of the 12 most abundant species were 816 (22%), 1,063 (28%), and 1,859 (50%) fish in the shallow, mid, and deep zones, respectively. For evaluation where the amount of webbing could be an important cost factor, total catches in each depth zone were converted to catches per unit surface area of webbing by dividing total catches for the shallow, mid, and deep zones by one, two, and four, respectively. Catches per unit area of webbing for the 12 species combined were 816 (45%), 531 (29%), and 465 (26%) fish for the shallow, mid, and deep zones. For the four most abundant species of food fishes unadjusted catches per unit area of net were 407 (56%), 196 (27%), and 126 (17%), and adjusted catches per unit area of net were 407 (76%), 98 (18%), and 32 (6%) fish for the shallow, mid, and deep zones. Thus, on either basis, fishing in the shallow zone was the most productive. Other factors of importance in this study in terms of overall efficiency included net damage, ease of fishing, cost, and storage of webbing. Daily average net damage was 0.16% for monofilament and 0.24% for multifilament webbing. Fish could be removed faster and fewer crabs were caught in monofilament nets. Monofilament nets tangled less and were set and retrieved faster than multi- filament nets. Disadvantages of monofilament compared to multifilament nets were: greater cost per pound (almost double); more storage room required; and greater difficulty of repairing the webbing owing to the requirement of double knots to prevent slippage. 107 ACKNOWLEDGMENTS We express sincere appreciation to John Ham- ley of the University of Toronto and Edwin A. Joyce, Jr. and his staff of the Florida Department of Natural Resources for their time in reviewing this manuscript and for their beneficial comments. We are deeply grateful to Dennis Anderson and Maxwell Miller for their assistance in the field during this study. LITERATURE CITED HOPKINS, T. L. 1966. The plankton of the St. Andrew Bay system, Flori- da. Publ. Inst. Mar. Sci., Univ. Tex. 11:12-64. ICHIYE, T, AND M. L. JONES. 1961. On the hydrography of the St. Andrew Bay system, Florida. Limnol. Oceanogr. 6:302-311. MAY, N., L. TRENT, AND P. J. PRISTAS. 1976. Relation offish catches in gill nets to frontal periods. Fish. Bull., U.S. 74:449-453. MCNULTY, J. K., W. N. LINDALL, JR., AND J. E. SYKES. 1972. Cooperative Gulf of Mexico estuarine inventory and study, Florida: Phase I, area description. U.S. Dep. Commer., NOAA Tech. Rep. NMFS CIRC-368, 126 p. FISHERY BULLETIN: VOL. 75, NO. 1 MIHARA, T., A. BRITO, J. RAMIREZ, AND J. V. SALAZAR. 1971. La pesca experimental con filete de ahorque en el Golfo de Paria. Proyecto Invest. Desarrollo Pesq. Venez., Inf. Tec. 23, 15 p. NATIONAL MARINE FISHERIES SERVICE. 1975. Fishery statistics of the United States 1971. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv. Stat. Dig. 65, 424 p. NATIONAL OCEAN SURVEY. 1971. Tide tables, high and low water predictions 1972, east coast of North and South America including Green- land. U.S. Dep. Commer., Natl. Ocean Surv., 290 p. REINTJES, J. W. 1969. Synopsis of biological data on the Atlantic menha- den, Brevoortia tyrannus. U.S. Fish Wildl. Serv., Circ. 320, 30 p. SIEBENALER, J. B. 1955. Commercial fishing gear and fishing methods in Florida. Fla. State Board Conserv., Univ. Miami Mar. Lab., Tech. Ser. 13, 45 p. STEEL, R. G. D., AND J. H. TORRIE. 1960. Principles and procedures of statistics with special reference to the biological sciences. McGraw-Hill, N.Y., 481 p. Waller, R. a. 1961. Ostracods of the St. Andrew Bay system. M.S. Thesis. Florida State Univ., Tallah., 46 p. 108 AGE DETERMINATION, REPRODUCTION, AND POPULATION DYNAMICS OF THE ATLANTIC CROAKER, MICROPOGONIAS UNDULATUS 12 Michael L. White and Mark E. Chittenden, Jr. 3 ABSTRACT A validated scale method of age determination is described for the Atlantic croaker, Micropogonias undulatus. Two age-classes were usually observed, but only one was abundant. Mean total lengths were 155-165 mm at age I and 270-280 mm at age II based on three methods of growth estimation. Fish matured near the end of their first year of life when they were about 140-170 mm total length. Spawning occurred from at least September through March but there was a distinct peak about October. Somatic weight-length relationships varied monthly, and changes appeared to be associated with maturation and spawning. Somatic weight reached a maximum in June, and the minimum was observed in March. Maximum somatic weight loss (24%) occurred in March, but no data were obtained from December through February. In estuaries, age croaker apparently occupied soft-substrate habitat and older fish occurred near oyster reefs. Life spans were only 1 or 2 yr, and the total annual mortality rate was 96%. The above life history pattern appears similar for croaker found throughout the Carolinian Province. Contrasts are presented to illustrate differences in the life histories and population dynamics of croaker found north and south of Cape Hatteras, N.C. A parallel is drawn with apparently similar changes in the American shad, A/osa sapidissima, and the suggestion is made that changes in the population dynamics of species that traverse the Cape Hatteras area may represent a general phenomenon. The Atlantic croaker, Micropogonias undulatus (Linnaeus), ranges in the western Atlantic from the Gulf of Maine to Argentina (Chao 1976). It is potentially a very important protein source be- cause it is one of the most abundant inshore fishes of the northern Gulf of Mexico (Gunter 1938, 1945; Moore et al. 1970; Franks et al. 1972) and the Atlantic Ocean off the southeastern United States (Haven 1957; Bearden 1964; Anderson 1968). Much work has been done on this species. However, many aspects of its life history and population dynamics are not clear; because no reliable method of age determination exists, and reproduction has not been studied intensively. A few early workers, including Welsh and Breder (1924) and Wallace (1940), attempted to age croaker using scales; but criteria for marks were not described and methods were not validated. More recent workers, in general, have not at- tempted to use hard parts to determine croaker age and growth. The scale method is difficult to apply to croaker (Joseph 1972), and this may be related to its migratory habits and extended 'Based on a thesis submitted by the senior author in partial fulfillment of the requirements for the MS degree, Texas A&M University. technical article TA 12419 from the Texas Agricultural Experiment Station. department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843. spawning season (Suttkus 1955). Only Wallace (1940) studied reproduction using a large series of gonads. However, he worked north of Cape Hatteras, N.C. The life history of croaker found north of Cape Hatteras seems quite different from that of individuals in the Carolinian Province. Studies of reproduction in croaker found south of Cape Hatteras have been based on few fish (Gunter 1945; Bearden 1964) or fish less than 200 mm long (Hansen 1969). This paper describes a validated method of age determination for croaker, their weight-length and girth-length relationships, habitat segrega- tion between age-groups, spawning seasonality, somatic weight variation, growth, maximum size, life span, and total annual mortality rates. Final- ly, it contrasts the life histories of croaker found north and south of Cape Hatteras. Geographically, statements made herein apply to the Carolinian Province and/or more northerly waters. With modifications, particularly ones due to calendar differences in seasons, our findings may also apply in the southern hemisphere; but further work is needed there. MATERIALS AND METHODS Collections were made from commercial shrimp trawlers during 1974 in the Gulf of Mexico off Manuscript accepted June 1976. FISHERY BULLETIN: VOL. 75, NO. 1, 1977. 109 FISHERY BULLETIN: VOL. 75, NO. 1 Freeport-Galveston and Port Aransas, Tex., and Cameron, La. Fish were also collected by trawling in Palacios, Galveston, and Matagorda bays, Tex., and Calcasieu Lake, La. Additional fish, herein- after termed reef fish, were captured by angling with dead shrimp bait (about 25 mm long) near an oyster bar in Galveston Bay. Collection months are indicated in Figure 1. A sample was taken from each trawl catch by shoveling into a 25-liter container small portions of the catch from various areas of the deck. Unusually large fish were arbitrarily selected to obtain older fish to develop an ageing technique. Total length was measured on each croaker. Total and gonad weights and girth at the origin of the dorsal fin were determined for fish over a broad size range during each sampling period. Scales below the lateral line posterior to the pectoral fin were removed from 1,123 fish, were pressed on plastic slides, and were examined using a scale projector. Scales were examined from small numbers of croaker collected off Mississippi and Fort Pierce, Fla., and in Chesapeake Bay, Va., to judge whether or not the method of age deter- mination proposed herein is valid throughout their range in the Carolinian Province and more northerly waters. The size and appearance of the gonads of more than 1,700 fish were examined, and ovaries were classified following Nikolsky (1963) as summarized by Bagenal and Braum (1971) except that the immature and resting stages were combined. The regressions of somatic, gonad, and total weights, and girth on total length were computed to express the best linear or quadratic fit using the Statistical Analysis System (Service 1972). Sex data were pooled to compute total weight-length, somatic weight-length, and girth-length re- gressions, because F tests (Ostle 1963:204) indicated that pooled regression lines were appropriate. each sex began by late August, increased greatly during September, reached a peak in October, declined greatly by November, and was at the latter level in March. Similarly, the coefficients of determination (r 2 ) of the regression lines (Table 1) show that gonad weight variation in each sex was increasingly associated with length until October and then greatly declined. Therefore, it appears that peak spawning occurred in October. Fish captured in the Gulf and by the reef were in all stages of development during September, as were trawl-caught bay fish in October (Figure 3). Therefore, spawning apparently began at least by late September, and some individuals finished or had nearly finished spawning then. Most spawn- ing occurred during October in agreement with the gonad weight-length analyses, because most fish captured in the Gulf were still immature in September. Most fish captured near the reef and in the Gulf were ripe or spent during October and November. Specimens captured in the Gulf during late March were in a resting stage or nearly spent, so that spawning is apparently completed by late March except by a few individuals. Croaker started to mature at about 140-170 mm total length. Extrapolated x -intercepts or inflec- tion points of the regressions of gonad weight on total length occur in that size range for each sex (Figure 2). Developing fish as small as 136 mm were observed. Many aspects of croaker spawning appear similar throughout the Carolinian Province. The prolonged spawning period suggested by our data is consistent with frequently reported collections offish about 25-40 mm long from October to June (many references including Suttkus 1955; Bear- den 1964; Hansen 1969; Parker 1971; Swingle 1971; Christmas and Waller 1973; Hoese 1973). The apparent peak of spawning after September agrees with Pearson ( 1929), Hildebrand and Cable SPAWNING Spawning occurred over a protracted period extending at least from September to late March, but there was a distinct peak about October. The regressions of gonad weight on length were not significant during May, June, or July for either sex. The mean gonad weight in this period was 0.10 g, and its 95% confidence limits were 0.09- 0.11 g. The regressions of gonad weight on length (Figure 2) indicate that gonad development in TABLE 1. — Analyses for the regressions of gonad weight (Y) in grams on total length (X ) in millimeters for each sex and month. All regressions were significant at a = 0.0001. Sample Sex Month size r 2 Equation Males August 67 0.46 Y = 389 • 004X September 108 0.68 Y = -4.737 + 0.033X October 64 0.73 Y = -8.804 + 0.055X November 46 0.32 Y = -2.782 + 0.01 8X March 35 0.43 Y = -3.785 + 0.021 X Females August 92 0.47 Y = -0.426 + 0.004X September 286 0.63 Y = -11.920 + 0.080X October 154 0.67 Y = -27.135 + 0.177X November 69 0.28 Y = -15.570 + 0.097X March 41 0.32 Y = -13.359 + 0.077X 110 WHITE and CHITTENDEN: AGE DETERMINATION OK ATLANTIC CROAKER I S 5- < 20 - BAY GULF REEF BAY GULF REEF _^£&- -cc r\ Ar^ 20 - BAY 30 - GULF 5 - REEF BAY GULF > o z HI D O uj K 0C CO U. D a 13 5 - REEF — L. 10 - BAY GULF 5 - REEF ^ ^ ^ ft LU 00 BAY 10- a. LU 5 — CO GULF REEF f- r\ r\ /V\ ^n /~\ 10 - O 5 ^ BAY <— ^ «-> ^- o GULF O 5- Q r-> ^^^—^^ REEF W y- 1 ^ i s. s\ f\ /\ /S e- cc UJ CO 5- BAY f\ <-\ r\ *~~i GULF > ^-^ ^ Q i^. n REEF ^_ T — . 60 1^ 100 T 150 200 250 TOTAL LENGTH (MM) 300 350 FIGURE 1. — Length frequencies of Atlantic croaker in each area each month. Frequencies are moving averages of three. Ill FISHERY BULLETIN: VOL. 75, NO. 1 September GULF BAY REEF September TOTAL LENGTH (MM) FIGURE 2. — Gonad weight-length regressions for Atlantic croaker by sex and month. The length of each line shows the observed size range. (1930), Suttkus (1955), and Bearden (1964); and size at maturity agrees with Pearson (1929), Bearden (1964), Hansen (1969), and Hoese (1973). The general similarity of croaker reproduction suggests that 15 October, which approximates the time of peak spawning, would be appropriate as a defined hatching date in warm-temperate waters. 100 I " 50 > > 3 O -> z O a 3 < n = 60 Lj < 50 100 -rr l -f=t- n = 120 I 1 I I 100 -r^ ^ n = 448 50 100 1 50 n = 378 100 50 O 50- z t^JS 12 3 4 5 50 n=150 L 100 n = 187 iu 100 50 - -rf^l 3=^^, n = 137 100 ^ n 50- | v^V ^|^ 12 3 4 5 GONAD CONDITION 100-1 50 N n = 55 100- 50- 100 50- 12 3 4 5 FIGURE 3. — Gonad condition of Atlantic croaker by months and areas. The ordinate represents percent of the sample. Gonad conditions on the abscissa are: (1) immature or resting, (2) maturation, (3) maturity, (4) reproduction, and (5) spent. SOMATIC WEIGHT VARIATION Somatic weight-length relationships varied monthly, and these changes appeared to be as- sociated with maturation and spawning. Peak somatic weight occurred during June except in fish smaller than about 140 mm. Somatic weights predicted by the regression equations for other months (Table 2) were compared with predicted weights in June (Figure 4). The somatic weight of individuals smaller than about 140 mm increased from May to at least September. Fish about 140- 160 mm showed progressive somatic weight loss from June to September-October. The smallest fish greater than 160 mm, in general, showed the greatest somatic weight loss (or smallest gain); TABLE 2. — Analyses for the regressions of somatic weight (7) in grams on total length (X) in millimeters for each month. All regressions were significant at a = 0.0001. Month Sample size Equation May 120 099 June 686 0.99 August 299 0.99 September 501 0.97 October 265 0.98 November 162 0.91 March 93 0.99 Y = 39.5303 - 0.8538X + 0.0057X 2 Y = 71.1692 - 1.3371X + 0.0076X 2 Y = 120.4035 - 1.9159X + 0.0092X 2 Y = 158.951 1 - 2.3706X + 0.01 03X 2 Y = 148.7089 - 2.201 6X + 0.0097X 2 Y = 73.4739 - 1 2980X + 0072X 2 y = 132.7087 -1.8537X + 0.0080X 2 and somatic weight loss, in general, seemed to progressively increase from June to September- October. Somatic weight loss during the fall in fish larger than 140 mm was greatest in September- 112 WHITE and CHITTENDEN: AGE DETERMINATION OF ATLANTIC CROAKER 30 25 - 20 - 15 - 10 5 I < 5 - 10 - z uu a. 15 UJ Q. 20 25 H + 30 35 - 40 - 45- 50 May March Novemb September August August September 50 200 TOTAL LENGTH (MM I —I 300 FIGURE 4. — Monthly somatic weight changes in Atlantic croaker. Percentage changes are in comparison to somatic weights in June. The lengths of the curves represent observed size ranges. October just prior to the time of peak spawning. However, greatest somatic weight loss was ob- served in March when individuals of 170-250 mm had lost 20-24% of the June weight. The ob- served somatic weight-length relationships and apparent weight changes in November may be anomalous. Absolute somatic weight decreased in fish smaller than 140 mm, but the percentage weight loss in fish greater than 160 mm was about 5%. Croaker mature at about 140-160 mm, and most fish were small and immature in November. These smaller fish may have just begun to mature for spawning, and their inclusion in the data may have biased the observed pattern in November. This interpretation is supported by the regression coefficients of X and X 2 which were markedly smaller during November than during other months in the August-March period (Table 2). Somatic weight changes have not been reported for croaker. Additional data, especially from the post-peak spawning period December to February, are needed to fully understand their annual cycle of somatic weight change. Possibly, the percen- tage of somatic weight loss may be greater in late fall and winter than we observed in March. AGE DETERMINATION AND GROWTH General Basis for the Method of Age Determination Scale marks similar to annuli were distin- guished by standard criteria, especially cutting over and differential spacing of circuli. Croaker appear to form two marks on their scales each year except that no mark is formed during their first winter. Some fish form no mark during their first year if 1 5 October is defined as the hatching date of croaker. Even-numbered marks (cold-period marks) form from about December to March, and odd-numbered marks (warm-period marks) form from about May to November. Fish that do not form a mark in their first year would not have mark numbering that corresponds to the typical odd and even system. Cold-period marks were most distinct and were used as "year" marks, although they represent 1-1 V2 yr of growth. Recognition of the first cold-period mark is the basis for this method. Subsequent marks, espe- cially cold-period marks, seem to be easily identified. Age determination was validated by: 1) es- tablishing the time of year when each mark forms, 2) establishing age through analysis of length frequencies, and 3) showing that modes of back- calculated and observed lengths at each age agree with age determination by length frequencies. Repeated reading suggests this method of age determination is consistent. We found 91% agreement between the first reading of scales from 200 fish (112 age and 88 age I) and a second reading 3 mo later. We have suggested 15 October as a defined hatching date for croaker. Definition of a hatching date is essential in age and growth studies, so that year classes and age groups can be referenced. In the northern hemisphere 1 January is a standard defined hatching date. That date is convenient and has biological reality, especially for species that spawn in the spring and summer of one year. In more northerly waters, furthermore, growth seasons tend to be short; and spawning tends to be restricted in time and often occurs about when the annulus forms. Croaker of the Carolinian Pro- vince, in contrast, have a long, possibly year- round, growing season; and their spawning "season" is so long that it takes place over much of two calendar years. Therefore, it seems more convenient and biologically sound to select their 113 FISHERY BULLETIN: VOL. 75, NO. 1 peak spawning period as a denned hatching date upon which year class and age group terminology is based. Using an October hatching date, the year class would pertain to the fall calendar year and would include any fish of that spawning cycle hatched in the following winter and spring. A virtual annulus would be designated as of October. Characteristics of Scale Markings Used to Determine Age The first mark is typically a more or less in- distinct mark formed in warm periods. It is characterized by cutting over in the lateral field, but it has little or no differential spacing of circuli before and after the mark (Figure 5a). This mark is often difficult to distinguish after the heavier second mark is formed. The typical second mark is formed in cold periods. It is the most diagnostic feature for age determination in croaker, and its recognition is the basis for our method. This mark is characterized by heavy cutting over of circuli and differential spacing of circuli in the lateral field (Figure 5b). Generally, circuli are closely spaced before the second mark and more widely spaced after it. When the first mark is absent or difficult to see, the typical second mark is readily distinguished. The third mark is typically formed in warm periods and is similar to the first mark (Figure 5c). We examined only six fish whose scales had the fourth mark, and its criteria may need modification. However, the fourth mark apparently forms in cold periods and apparently, resembles the second mark in having heavy cut- ting over and differential spacing of circuli (Figure 5c). Croaker from a broad geographical range seemingly can be aged by the method proposed, although further work is needed to establish this. Scales offish from Mississippi, Fort Pierce (Figure 6a), and Chesapeake Bay (Figure 6b, c) showed markings similar to those on scales from Texas fish. Croaker scales from Florida generally had more or less indistinct cutting over and seemed FIGURE 5. — Top. Scale from a 190-mm croaker showing mark 1. This fish was approaching age I when it was captured off Texas in September. The axis depicted shows how measurements were made to determine when each mark formed. Middle. Scale from a 255-mm croaker showing marks 1 and 2. This fish was ap- proaching age II when it was captured off Texas in August. Bottom. Scale from a 310-mm croaker showing marks 2,3, and 4. This was an age 11+ fish captured off Texas in March. 114 WHITE and CHITTENDEN: AGE DETERMINATION OF ATLANTIC CROAKER difficult to read, possibly because the fish were collected in tropical waters of southern Florida where temperature changes are not as extreme as further north. Only six fish from Texas had scales with four marks. In contrast, scales from some Chesapeake Bay fish had six marks (Figure 6c). Croaker that live in the Carolinian Province south of Cape Hatteras live only 1 or 2 yr (see General Discussion) and, therefore, tend to have comparatively few marks on their scales. These fish might be easier to age than croaker that live north of Cape Hatteras. The latter fish apparently survive longer and, therefore, probably tend to have more marks on their scales. Times of Mark Formation The time when each annuluslike mark formed was determined by plotting for each month the distance from the scale margin to the last mark. Distance was measured across the lateral field of the scale (Figure 5a). Croaker generally form two marks per year except during their first year. Scales with no marks had the smallest distance between the scale margin and focus in May (Fig- ure 7). The radius increased from May to October as scales grew during that period. Therefore, apparently no mark is formed during the first winter; and some croakers form no mark during the first year of life if 15 October is defined as their hatching date. Scales with one mark had the mark closest to the scale edge in warmer months. In March the mark was far removed from the scale margin, suggesting that the first mark normally forms in warm months. Apparently this mark formed on some fish throughout the period May to at least October. The increment between the scale margin and the first (or third) mark did not in- crease with time, but the reason for this is not clear. Scales with two marks showed the second mark closest to the scale margin in March. The increment between this mark and the scale edge increased until June and then remained nearly constant through November. Therefore, the sec- ond mark apparently forms during the colder FIGURE 6. — Top. Scale from a 305-mm croaker showing marks 1, 2, 3, and 4. This was an age 11+ fish when it was captured off Florida in March. Middle. Scale from a 293-mm croaker showing marks 1 and 2. This fish was approaching age II when it was captured in Chesapeake Bay in July. Bottom. Scale from a 508-mm croaker showing marks 1, 2, 3, 4, 5, and 6. This fish was approaching age IV when it was captured in Chesapeake Bay during July. 115 NO MARKS FISHERY BULLETIN: VOL. 75, NO. 1 MARK 1 MARK 2 MARK 3 MARK 4 I o cc < 15n 5- J ^T- i /\ i — i—n 20 40 15-, $ 5 -l 15^ Z 5- ■h — r— i — r ^i a i i i i — r~i — i T-i — i — i -r~T — i — i A^ r — i — i — i > o z LJJ o LU CL I5n 5- 15n t — i — r t — I r-i — r— i ' i ' i — i — i T- 1 -! 1 1 Vm £ 15". CD 2 5^ o o cc LU cd I5n § 5 -> t — i — n — r~-i T"""! 1 1 -\ — r i r i in l n i T— I 1 1 t-^i — i — i — i — i r I 1 1 ! I ^-T-. 20 40 60 20 40 20 40 20 40 DISTANCE (MM X 42) FIGURE 7. — Distance from scale margin to the last mark or to the focus if no marks were present. months. Scales with three marks showed the third mark being formed throughout the warm months, the only period when scales with only three marks were available. Scales with four marks were observed only during March. The increment on these scales suggests that the fourth mark was formed during winter or spring. However, further data are needed to establish this. Our findings on times of mark formation agree with Haven's (1954) suggestion that croaker form one fall and one winter mark each year in Chesapeake Bay and with Richards' (1973) computer-simulated findings that the related black drum, Pogonias cromis, forms one mark a year until maturity and two marks a year thereafter. 116 WHITE and CHITTENDEN: AGE DETERMINATION OF ATLANTIC CROAKER Age Determination and Growth by the Length-Frequency Method Our length-frequency distributions suggest two croaker year classes occurred off Texas. One age group greatly predominated in the length fre- quencies of trawl-caught fish from the bay and Gulf during June (Figure 1). The size range of that age group was primarily about 100-150 mm in the bay and about 120-160 mm in the Gulf. Young-of- the-year first appear in Texas bays about November and increase in size from about 10-50 mm during January to 30-85 mm in March, 40-100 mm during May, and 70-130 mm in June (Gunter 1945; Parker 1971; Gallaway and Strawn 1974). Therefore, the fish we captured by trawling during June must be young-of-the-year. These young-of- the-year fish grew to about 1 10-170 mm in August, 120-175 mm in September, and 140-180 mm in October when they reached age I. Similar sizes in October have been recorded by Gunter (1945), Parker (1971), and Gallaway and Strawn (1974). The fish that became age I in October were about 130-190 mm in November, and fish captured in March were about 165-220 mm. The large fish caught in June by angling near the oyster reef were about 190-270 mm and apparently were survivors of the year class that became age I on the preceding 15 October. These age 1+ fish were about 200-310 mm in September when they approached age II. This agrees with Gunter's (1945) size estimates for age II croakers off Texas. With minor differences, length frequencies reported throughout the Carolinian Province by many workers, including Hildebrand and Cable (1930), Gunter (1945), Suttkus (1955), Bearden (1964), Hansen (1969), Christmas and Waller (1973), Hoese (1973), and Gallaway and Strawn (1974), show growth and age composition similar to our findings. Growth north of Cape Hatteras seems similar to that in the Carolinian Province. Haven (1957) presented monthly length fre- quencies of fish he considered young-of-the-year. His fish ranged from about 150 to 220 mm in September, but the mode was about 175-180 mm. Agreement of Observed and Back-Calculated Lengths with Length-Frequencies Observed sizes at ages 0, I, and II agree closely with ages determined by length frequencies (Figure 8). Only age fish were captured in May and age I fish in July, so that graphs are not presented for these months. The frequencies show overlap in size between the various ages each month. This is to be expected, especially in a species having a prolonged spawning season, and makes it impossible to use the length-frequency method to assign age confidently where sizes at age overlap. The observed lengths of age fish in September were primarily 130-170 mm (mean = 151 mm), but they ranged from about 110 to 220 mm. This age group was about 140-220 mm (mean = 158 mm) during October when they became age I and about 130-220 mm (mean = 172 mm) during November. The observed lengths of age I fish in September were about 200-340 mm with the mean being 253 mm. This age group was about 190-360 mm (mean = 274 mm) in October when they became age II. Lengths back-calculated to cold-period marks reasonably agree with the sizes at age I estimated by length frequencies in October (Figure 9). However, cold-period marks apparently begin to form generally after October; so that the back- calculated lengths should be larger than the observed lengths in October. The similarity suggests Lee's phenomenon, possibly due to selective mortality favoring survival of smaller croaker. Back-calculated lengths were somewhat smaller than the sizes at age 1+ in March, as would be expected. Back-calculated lengths from age 1+ fish were primarily 110-210 mm at age I with a mean length of 165 mm. In agreement, back-calculated lengths from six age 11+ fish had a mean of 181 mm at age I and 270 mm at age II. The body-scale regression equation used to back- calculate length was: Y = 2.6000 + 4.6389Z - 0.0122X 2 where Y represents total length in millimeters, andX represents the scale radius (millimeters x 42). The sample size was 1,123, and the total length range was 90-360 mm. About 88% of the variation in total length was associated with variation in scale radius. Growth estimates based upon the length- frequency method and from observed and back- calculated estimates using the scale method show very close agreement. Mean lengths in October were about 155-165 mm at age I and 270-280 mm at age II depending upon how age was determined. The wide back-calculated and observed size ranges found at age may be due to the long 117 FISHERY BULLETIN: VOL. 75, NO. 1 3—1 5 I o < 10- 5 — 5 80 40 — > l- LU u D D O < 2 — 15 — 5 — LU CO 5 — I- 15 — CL LU OO 5 — cc 2 — LU CO o 2 — o o LU CO > O 2 10 — 5 — 2 — 10 — Age II+ Observed ■"^ *- --^- Age I + Observed - ^ „S Sss. r-. *. Aged By L/F ri-, , I + II + ? ^-~x Q^ ^-^ ,-> ^^-. Age I Observed n » «■ n n Age Observed Age I Observed Age Observed m am r-irs /-vr\ \r^,-i r-^ Aged By L/F •«--* Age I Observed _£ a. Q £\_ Age Observed ~V\ Aged By L/F fl /— ^s. r-, A. Age II Observed — - Age I Observed s^> rW. m Age II Observed Age I Observed a ^^ ^^ _ca_ oa ^ Aged By L/F II 70 100 T 150 fca IR-i Back Calculated Length at Age I. o -c Age I Fish Captured in October, Age Determined by L/F Method. 5- t — i — i — r 70 100 — i — r 160 130 TOTAL LENGTH (MM l — l — I — T 190 220 — I — I — I — i 250 280 FIGURE 9.— Back-calculated length frequencies at age I and length frequencies (L/F) of age I fish in October and age 1+ fish in March. Frequencies are moving averages of three. spawning season and/or prolonged time span when the cold-period mark may form. HABITAT SEGREGATION BETWEEN AGE GROUPS A portion of all croaker age groups apparently utilized bays as feeding grounds during the warmer months, but age I and older fish seemed to occupy different habitat than young-of-the-year. Croaker captured by angling near the oyster reef from June to August were about 200-270 mm in length (Figure 1) and seemed common there. In contrast, trawl-caught bay fish were generally much smaller than 200 mm. Reef and trawl- caught bay individuals were then about age 1 + and age 0, respectively. Many other workers, including Reid (1955), Perret (1966), Nelson (1969), Hansen (1969), Parker (1971), Hoese (1973), and Gallaway and Strawn (1974), have also captured few individuals greater than 200 mm by trawling in bays, but they captured many small specimens like we did. Therefore, although capture by angling may have selected larger fish near the reef, the two age-groups seem to segregate by habitat: young-of-the-year occupy soft substrates, and age I and older fish occur near oyster reefs (and similar hard substrates?). This agrees with Harden Jones' (1968) generalization that the feeding grounds of adult fishes are sepa- rate from their spawning grounds and nurseries. Age I and older fish seemed to remain near oys- ter reefs until they migrated to sea to spawn. Fish caught near oyster reefs were much larger than those caught by trawling in the Gulf or bays until September-October (Figure 1). Specimens larger than 191 mm were not collected in the Gulf until September, which is about when spawning begins in the northern Gulf (Gunter 1945; Suttkus 1955; present study). Simmons and Hoese (1959) captured fish less than 175 mm long throughout the summer as they migrated to the Gulf, but these workers captured fish similar in size to our reef fish only during September. The larger young-of-the-year began moving to sea by late spring or early summer. Trawl-caught fish in the bay were smaller than those in the Gulf during June (Figure 1) when modal length for young-of-the-year was about 120 mm in the bay and about 140 mm in the Gulf. The difference in size between young-of-the-year in the bay and Gulf agrees with Gunter (1945), Haven (1957), and Reid and Hoese (1958) who found a size gradient in estuaries, the smallest young-of-the- year being farthest up the estuary. Haven (1957) and Hoese et al. (1968) suggested that the gradient was due to gradual seaward dispersal of the largestyoung, and Parker (1971) and Franks etal. (1972) suggested that young-of-the-year began moving to sea at about 85-100 mm long. Evidently the Gulf becomes a very important nursery by midspring or early summer, because young croaker compose about 24-29% by number of the fishes found on the white shrimp grounds of the Gulf then (Miller 1965, table 3; Chittenden and McEachran 1976). MAXIMUM SIZE AND AGE, LIFE SPAN, AND MORTALITY RATE Croaker in the Carolinian Province are typi- cally small and have a short life span and high mortality rate. Most fish we collected were less than 200 mm long and the largest was 357 mm. The largest croaker observed in warm-temperate waters generally have been less than 300 mm (many workers including Hildebrand and Cable 1930; Reid 1955; Bearden 1964; Miller 1965; Nel- son 1969; Hansen 1969; Parker 1971; Hoese 1973), although some workers captured fish as large as 330-380 mm (Pearson 1929; Gunter 1945; Suttkus 1955; Franks et al. 1972; Christmas and Waller 1973). Rivas and Roithmayr (1970) found a 668 mm specimen, but this is exceptional. 119 FISHERY BULLETIN: VOL. 75, NO. 1 Our length frequencies suggest that two year classes occurred, but only one was abundant. This agrees with other reported length frequencies from warm-temperate waters (see references cited in section on Age Determination and Growth by the Length-Frequency Method). Therefore, the typical croaker life span in warm-temperate water appears to be only 1 or 2 yr. Age 11+ fish captured in March were the oldest fish we examined in agreement with other estimated maximum ages from the Carolinian Province (Gunter 1945; Suttkus 1955; Bearden 1964; Hoese 1973). Fish associated with oyster reefs are larger and a year older than trawl-caught bay or Gulf fish during the summer. However, the abundance of these age I croaker must be small compared with the abundance of age croaker, because the geograph- ical area occupied by oyster reefs is comparatively small. Croaker have a high total annual mortality rate as their short life span requires. We found only six age 11+ fish in 1,123 aged. Greatest mixing of age-groups probably coincides with fall spawning in the Gulf. We observed 1 1 age I + and 250 age 0+ fish in random samples from trawl catches made 25-27 September 1974, so that the observed total annual mortality rate was about 96% assuming negative exponential survivorship. This must approximate the total annual mortality rate throughout the Carolinian Province because maximum sizes and ages, length frequencies, and life spans appear similar throughout this area. The observed total annual mortality rate agrees closely with the theoretical total annual mortality rate. Following the reasoning of Royce (1972:238) the negative exponential survivorship relation S = N t /N = e~ Zt can be solved for an approximate instantaneous total mortality rate over the entire life span which can be used to estimate average annual total mortality rates. A species with a life span of 1 or 2 yr would have a theoretical approximate total annual mortality rate of 90- 100%. TOTAL WEIGHT-LENGTH AND GIRTH-LENGTH RELATIONSHIPS The regression of total weight in grams (Y) on total length in millimeters (X) was expressed by the equation: log 10 Y = -5.26 + 3.15 log 10 X. This relationship was based on a sample size of 2,081 fish in the length range 90-360 mm. About 98% of the variation in log 10 total weight was associated with variation in log 10 total length. The arithmetic mean log 10 X was 2.21056, and arithmetic mean log 10 Y was 1.71546. The regression of girth in millimeters (Y) on total length (X) in millimeters was expressed by the linear equation: Y = -11.84 + 0.71X. This relationship was based on a sample size of 2,081 fish in the length range 90-360 mm. The arithmetic mean girth was 108.07 mm. About 94% of the variation in girth was associated with variation in total length. GENERAL DISCUSSION Many aspects of the life history of Atlantic croaker in the Carolinian Province appear dif- ferent than those of fish found in cold-temperate waters north of Cape Hatteras except that the growth rates appear similar. In general, our data and the literature agree that in warm-temperate waters: 1) peak spawning occurs about October but the spawning season is long and lasts from about September to at least March, 2) maturity is reached at about 140-180 mm long as the fish approach age I, 3) maximum size is about 300-350 mm and most fish are so small (about 200 mm or less in length) that they do not support commercial food fisheries, 4) the life span is about 1-2 yr and maximum age is typically about 2 yr, 5) most fish live only to about age I, and 6) total annual mor- tality rate is about 95%. In contrast, fish living north of Cape Hatteras generally: 1) Have a spawning season (July or August- December?) that starts earlier and may end earlier (Welsh and Breder 1924; Hildebrand and Schroeder 1928; Wallace 1940; Pearson 1941; Massmann and Pacheco 1960). However, the time when spawning ends is not certain. Haven (1957) captured many young 20-30 mm TL from February to April, but their significance is not clear; they could represent late-winter spawning or, perhaps, fall spawning with little or no overwinter growth. Peak spawning seemingly occurs no later than midfall, because all the adult fish that Wallace (1940) examined had spent or 120 WHITE and CHITTENDEN: AGE DETERMINATION OF ATLANTIC CROAKER recovering gonads in late November and thereafter. 2) Reach maturity when greater than 200 mm long as they approach at least age II (Welsh and Breder 1924; Wallace 1940; Haven 1954). 3) Have a maximum size of about 500 mm (Hildebrand and Schroeder 1928; Gunter 1950) and large average size so that they have supported important commercial food fisheries (Gunter 1950; Haven 1957; Joseph 1972). Maturity is reached about 1 yr later in cold- temperate waters and typical sizes are much larger, although growth rates appear similar. Therefore, the typical maximum age is probably about 2-4 yr north of Cape Hatteras. If so, the total annual mortality rate must be lower north of Cape Hatteras. Assuming negative exponential survivorship, the theoretical approximate total annual mortality rates would be 90, 78, and 68% for life spans of 2, 3, and 4 yr, respectively. The existence of an abrupt change at Cape Hatteras in the life histories and population dynamics of species whose ranges traverse this area has apparently not been recognized, par- ticularly as a possible general phenomenon; although Cape Hatteras has long been recognized as a significant zoogeographic boundary [see Briggs' (1974) review]. Gunter (1950) noted dif- ferences in the sizes and some aspects of the life histories of certain fishes of the Gulf of Mexico and mid-Atlantic coast of the United States. However, he gave no consideration to the possibility that an abrupt change might occur near Cape Hatteras. Although the Cape Hatteras connection has not been recognized, the pelagic, anadromous American shad, Alosa sapidissima, also shows changes in life history there that are similar to those herein documented for croaker. Runs of shad native to streams north of Cape Hatteras consist primarily of somewhat older fish (ages IV- VII and older) and include many repeater spawners in contrast to the younger fish (ages IV- VI) and the complete or virtual absence of repeat spawners south of Cape Hatteras (for pertinent literature see Walburg and Nichols 1967; Chittenden 1975). La Pointe (1958) reported similar growth rates in shad native to streams throughout their range. Therefore, the geographic differences in age compositions should result in differences in life spans, ages at maturity, maximum ages, maximum and average sizes, and mortality rates as in croaker. The life histories and population dynamics of two species with different life styles but primarily coastal habit have been shown to change abruptly at Cape Hatteras. This may represent a general phenomenon as Gunter (1950) apparently ob- served. However, similar comparisons are necessary in other species, especially noncoastal forms, to see how far the inference extends. The reason for the geographical differences in population dynamics is not clear. However, shad exhibit great somatic weight loss (about 25-55% depending upon sex and size) associated with migration and spawning (Leggett 1972; Chitten- den 1976). Leggett (1972) suggested that the low frequency of repeat spawning shad in southern streams might be due to increased use of body reserves during spawning migrations that occur at higher average temperatures. Croaker also show somatic weight loss associated with mat- uration and spawning, although we did not ob- serve weight loss comparable to that in shad. However, we had no data for the post-peak spawning period December-February when weight loss may have been greater. It is pertinent here that Chittenden has observed many emaciated spot, Leiostomus xanthurus, in the Gulf of Mexico during January, which is about when this species spawns. The observed differences in population dynamics north and south of Cape Hatteras may be largely the result of different temperature regimes that affect age at mat- uration, spawning-associated somatic weight loss, and the magnitude of a subsequent post-spawning mortality. ACKNOWLEDGMENTS For assistance with field collections we are indebted to R. Clindaniel, C. H. Stephens, G. Graham, J. Surovik, M. Carlisle, and to Captains R. Foreman, R. Foreman, Jr., J. Torres, H. For- rester, and M. Forrester. C. E. Bryan and W. Cody of the Texas Parks and Wildlife Department made collections offish from the Gulf in November. S. M. Lidell directed us to large croakers near the reef. J. Merriner and J. Musick of the Virginia Institute of Marine Science loaned scales from Chesapeake Bay. J. McEachran, W. Neill, R. Noble, L. Ringer, R. Stickney, K. Strawn, and M. VanDenAvyle of Texas A&M University reviewed the manuscript and L. Ringer programmed certain statistical 121 FISHERY BULLETIN: VOL. 75, NO. 1 analyses. Financial support was provided, in part, by the Texas Agricultural Experiment Station and the Office of Sea Grant, NO A A. LITERATURE CITED Anderson, W. W. 1968. Fishes taken during shrimp trawling along the south Atlantic coast of the United States, 1931-35. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 570, 60 p. BAGENAL, T. B., AND E. BRAUM. 1971. Eggs and early life history. In W. E. Ricker (editor), Methods of assessment offish production in fresh waters, p. 166-198. IBP (Int. Biol. Programme) Handb. 3. Blackwell Sci. Publ., Oxf. BEARDEN, C. M. 1964. Distribution and abundance of Atlantic croaker, Micropogon undulatus, in South Carolina. Contrib. Bears Bluff Lab. 40, 23 p. BRIGGS, J. C. 1974. Marine zoogeography. McGraw-Hill, N.Y., 475 p. CHAO, L. N. 1976. Aspects of the systematics, morphology, life history and feeding of western Atlantic Sciaenidae (Pisces: Perciformes). Ph.D. Thesis, College of William and Mary, Williamsburg, 342 p. CHITTENDEN, M. E., JR. 1975. Dynamics of American shad, Alosa sapidissima, runs in the Delaware River. Fish. Bull., U.S. 73:487-494. 1976. Weight loss, mortality, feeding, and duration of res- idence of adult American shad, Alosa sapidissima, in fresh water. Fish. Bull., U.S. 74:151-157. CHITTENDEN, M. E., JR., AND J. D. MCEACHRAN. 1976. Composition, ecology, and dynamics of demersal fish communities on the northwestern Gulf of Mexico con- tinental shelf, with a similar synopsis for the entire Gulf. Sea Grant Publ. No. TAMU-SG-76-208, 104 p. Christmas, J. Y., and R. S. Waller. 1973. Estuarine vertebrates, Mississippi. In J. Y. Christmas (editor), Cooperative Gulf of Mexico estuarine inventory and study, Mississippi, p. 320-434. Gulf Coast Res. Lab. Franks, J. S., J. Y. Christmas, W. L. Siler, R. Combs, R. Waller, and C. Burns. 1972. A study of the nektonic and benthic faunas of the shallow Gulf of Mexico off the state of Mississippi as re- lated to some physical, chemical and geologic factors. Gulf Res. Rep. 4:1-148. GALLAWAY, B. J., AND K. STRAWN. 1974. Seasonal abundance and distribution of marine fishes at a hot-water discharge in Galveston Bay, Texas. Con- trib. Mar. Sci. Univ. Tex. 18:71-137. GUNTER, G. 1938. Seasonal variations in abundance of certain estuarine and marine fishes in Louisiana, with particular reference to life histories. Ecol. Monogr. 8:313-346. 1945. Studies on marine fishes of Texas. Publ. Inst. Mar. Sci. Univ. Tex. 1:1-190. 1950. Correlation between temperature of water and size of marine fishes on the Atlantic and Gulf coasts of the United States. Copeia 1950:298-304. HANSEN, D. J. 1969. Food, growth, migration, reproduction, and abun- dance of pinfish, Lagodon rhomboides, and Atlantic croaker, Micropogon undulatus, near Pensacola, Florida, 1963-65. U.S. Fish Wildl. Serv., Fish. Bull. 68:135-146. HARDEN JONES, F. R. 1968. Fish migration. Edward Arnold Publ., Lond., 325 p. Haven, D. S. 1954. Croakers. Va. Comm. Fish. 54th and 55th Annu. Rep. 1952-1953, p. 49-53. 1957. Distribution, growth, and availability of juvenile croaker, Micropogon undulatus, in Virginia. Ecology 38:88-97. HlLDEBRAND, S. F., AND L. E. CABLE. 1930. Development and life history of fourteen teleostean fishes at Beaufort, N.C. Bull. U.S. Bur. Fish. 46:384-488. HlLDEBRAND, S. F., AND W. C. SCHROEDER. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43:1-366. HOESE, H. D. 1973. A trawl study of nearshore fishes and invertebrates of the Georgia coast. Contrib. Mar. Sci. Univ. Tex. 17:63-98. HOESE, H. D., B. J. COPELAND, F. N. MOSELY, AND E. D. LANE. 1968. Fauna of the Aransas Pass Inlet, Texas. III. Diel and seasonal variations in trawlable organisms of the adja- cent area. Tex. J. Sci. 20:33-60. Joseph, E. B. 1972. The status of the sciaenid stocks of the middle Atlantic coast. Chesapeake Sci. 13:87-100. LAPOINTE, D. F. 1958. Age and growth of the American shad, from three Atlantic coast rivers. Trans. Am. Fish. Soc. 87:139-150. LEGGETT, W. C. 1972. Weight loss in American shad (Alosa sapidissima, Wilson) during the freshwater migration. Trans. Am. Fish. Soc. 101:549-552. MASSMAN, w. h., and a. l. pacheco. 1960. Disappearance of young Atlantic croakers from the York River, Virginia. Trans. Am. Fish. Soc. 89:154-159. MILLER, J. M. 1965. A trawl study of the shallow Gulf fishes near Port Aransas, Texas. Publ. Inst. Mar. Sci. Univ. Tex. 10:80-107. Moore, d., H. a. Brusher, and L. Trent. 1970. Relative abundance, seasonal distribution, and species composition of demersal fishes off Louisiana and Texas, 1962-1964. Contrib. Mar. Sci. Univ. Tex. 15:45-70. NELSON, W. R. 1969. Studies on the croaker, Micropogon undulatus Linnaeus, and the spot, Leiostomus xanthurus Lacepede, in Mobile Bay, Alabama. M.S. Thesis, Univ. Alabama, University, 85 p. NIKOLSKY, G. V. 1963. The ecology of fishes. Academic Press, N.Y., 352 p. OSTLE, B. 1963. Statistics in research. 2d ed. Iowa State Univ. Press, Ames, 585 p. PARKER, J. C. 1971. The biology of the spot, Leiostomus xanthurus Lacepede, and Atlantic croaker, Micropogon undulatus (Linnaeus), in two Gulf of Mexico nursery areas. Sea Grant Publ. TAMU-SG. 71-210, 182 p. Pearson, j. C. 1929. Natural history and conservation of the redfish and other commercial Sciaenids on the Texas coast. Bull. U.S. Bur. Fish. 44:129-214. 1941. The young of some marine fishes taken in lower 122 WHITE and CHITTENDEN: AGE DETERMINATION OF ATLANTIC CROAKER Chesapeake Bay, Virginia, with special reference to the grey sea trout, Cynoscion regalis (Bloch). U.S. Fish Wildl. Serv., Fish. Bull. 50:79-102. PERRET, W. S. 1966. Occurrence, abundance, and size distribution of fishes and crustaceans collected with otter trawl in Vermilion Bay, Louisiana. M.S. Thesis, Univ. Southwest. La., Lafayette, 64 p. REID, G. K., JR. 1955. A summer study of the biology and ecology of East Bay, Texas. Part I. Introduction, description of area, methods, some aspects of the fish community, the in- vertebrate fauna. Tex. J. Sci. 7:316-343. REID, G. K., AND H. D. HOESE. 1958. Size distribution of fishes in a Texas estuary. Copeia 1958:225-231. RICHARDS, C. E. 1973. Age, growth and distribution of the black drum (Pogonias cromis) in Virginia. Trans. Am. Fish. Soc. 102:584-590. RIVAS, L. R., AND C. M. ROITHMAYR. 1970. An unusually large Atlantic croaker, Micropogon undulatus, from the northern Gulf of Mexico. Copeia 1970:771-772. ROYCE, W. F. 1972. Introduction to the fishery sciences. Academic Press, N.Y., 351 p. SERVICE, J. 1972. A user's guide to the statistical analysis system. N.C. State Univ., Raleigh, 260 p. Simmons, E. G., and H. D. Hoese. 1959. Studies on the hydrography and fish migrations of Cedar Bayou, a natural tidal inlet on the central Texas coast. Publ. Inst. Mar. Sci. Univ. Tex. 6:56-80. SUTTKUS, R. D. 1955. Seasonal movements and growth of the Atlantic croaker (Micropogon undulatus ) along the east Louisiana coast. Proc. Gulf Caribb. Fish. Inst., Annu. Sess. 7:151-158. Swingle, H. a. 1971. Biology of Alabama estuarine areas — cooperative Gulf of Mexico estuarine inventory. Ala. Mar. Res. Bull. 5, 123 p. Walburg, C. H., and P. R. Nichols. 1967. Biology and management of the American shad and status of the fisheries, Atlantic coast of the United States, 1960. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 550, 105 p. WALLACE, D. H. 1940. Sexual development of the croaker, Micropogon undulatus, and distribution of the early stages in Chesapeake Bay. Trans. Am. Fish. Soc. 70:475-482. WELSH, W. W., AND C. M. BREDER, JR. 1924. Contributions to life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141-201. 123 COASTAL AND OCEANIC FISH LARVAE IN AN AREA OF UPWELLING OFF YAQUINA BAY, OREGON Sally L. Richardson and William G. Pearcy 1 ABSTRACT A 1%-yr survey of planktonic fish larvae collected from 2 to 111 km off the mid-Oregon coast in 1971-72 yielded 287 samples which contained 23,578 individuals in 90 taxonomic groups, 78 identified at the species level. Two distinct faunal assemblages were found: a "coastal" assemblage 2 to 28 km offshore and an "offshore" assemblage 37 to 111 km from shore. The coastal group was dominated by Osmeridae, Parophrys vetulus, Isopsetta isolepis, and Microgadus proximus. The offshore group was dominated by Sebastes spp., Stenobrachius leucopsarus, Tarletonbeania crenularis, Lyopsetta exilis, and Engraulis mordax. Peak abundance in both assemblages occurred between February and July when >9(Wc of all larvae were taken. Larval distribution patterns in each assemblage were similar in 1971 and 1972, but larval abundance was greater in 1971 than 1972. Ninety-nine percent of the larvae in 53 taxa designated as coastal and 96% of the larvae in 31 taxa designated as offshore were taken 2 to 28 km or 37 to 111 km offshore respectively. This separation of coastal and offshore larvae may be explained, in part, by adult spawning locations and current circulation patterns. The species of larvae present in the coastal assemblage were similar to those in Yaquina Bay, but dominant species were quite different. The coastal zone is an important spawning area for P. vetulus, which utilizes Yaquina Bay estuary as a nursery during part of its early life. In this paper, distribution patterns, seasonality, species composition, dominance, and relative abundance of larval fishes in an upwelling area off Yaquina Bay, Oreg., are described. Included are the most comprehensive time series of data yet available on larval fishes in the northeast Pacific Ocean north of California, data on the greatest number of distinct larval taxa yet reported for this area, and the first quantitative information on coastal and offshore assemblages of larval fishes off the northwest coast of the United States. Larval fish distributions are discussed in rela- tion to current circulation patterns and spawning location of adults. Results are compared with Pearcy and Myers' (1974) study of larval fishes of Yaquina Bay. The data on fish larvae are com- pared with data on zooplankton (Peterson and Miller 1975, footnote 2), shrimp larvae (Rothlis- berg 1975), and crab larvae (Lough 1975) collected at the same time and location. Distribution patterns of larval fishes off the mid-Oregon coast 'School of Oceanography, Oregon State University, Corvallis, OR 97331. 2 Peterson, W. T., and C. B. Miller. 1976. Zooplankton along the continental shelf off Newport, Oreg., 1969-72: distribution, abundance, seasonal cycle and year to year variations. Oreg. State Univ. Sea Grant Coll. Prog. Publ. ORESU-T-76-002, 111 p. are discussed in relation to a broader geographic area in the northeast Pacific. PREVIOUS STUDIES IN THE NORTHEAST PACIFIC This review includes only studies of a general survey nature conducted in ocean waters from northern California to the Gulf of Alaska, excluding the Aleutian Chain and Bering Sea. Studies in sounds, bays, and estuaries are not considered. Prior to 1972, data on ichthyoplankton in the northeast Pacific were sparse and essentially nonquantitative because of the gear used — Isaacs-Kidd Midwater Trawls and Northern Pa- cific area (NORPAC) nets (Motoda et al. 1957). Surveys were designed primarily for biomass estimates of pelagic invertebrates and fishes. The ancillary data on fish larvae, often not identified to species, were usually presented in the form of appendix tables [Aron 3 for northern Washington Manuscript accepted September 1976. FISHERY BULLETIN: VOL. 75, NO. 1, 1977. 3 Aron, W. 1958. Preliminary report of midwater trawling studies in the Pacific Ocean. Univ. Wash. Dep. Oceanogr. Tech. Rep. 58, 64 p. 125 to southwest Alaska; Aron 4 for southern Califor- nia to southwest Alaska; Pearcy 5 for Oregon; Porter (1964) for northern California (flatfish only); LeBrasseur 6,7 for the northeast Pacific; Day (1971) for Washington to British Columbia]. Two additional reports (Aron 1959; LeBrasseur 8 ) briefly mentioned larval fishes in the text. More recent reports have been based on surveys designed specifically to sample ichthyoplankton using meter nets and bongo nets [Waldron (1972) off Oregon, Washington, and British Columbia in April-May 1967; Richardson (1973) off Oregon from May to October 1969; Naplin et al. 9 off Washington and British Columbia in October- November 1971; Dunn and Naplin 10 off Alaska in April-May 1972; Pearcy and Myers (1974) off Yaquina Bay from June 1969 to June 1970]. Results were quantitative and more refined species lists were provided. However most of these studies were restricted in seasonal coverage to periods of less than 1 yr. Pearcy and Myers (1974) presented a year-long data set but listed only yearly mean abundances. Discussion of larval distribution patterns in all these papers was limited. Waldron (1972) arbitrarily divided his data into two groups located inshore or offshore of the 914-m contour and discussed larval abun- dances in each region. Pearcy and Myers (1974) discussed horizontal variations in larval dis- tributions with respect to larvae that occurred offshore and those that occurred in Yaquina Bay. Vertical distribution and day-night differences have not been discussed, although Richardson (1973) compared deep (to 200 m) and shallow (upper 20 m) tows. 4 Aron, W. 1960. The distribution of animals in the eastern north Pacific and its relationship to physical and chemical conditions. Univ. Wash. Dep. Oceanogr. Tech. Rep. 63, Ref. 60-55, 65 p. + 156 append. 5 Pearcy, W. G. 1962. Species composition and distribution of marine nekton in the Pacific Ocean off Oregon. Oreg. State Univ., Dep. Oceanogr., A.E.C. Prog. Rep. 1, Ref. 62-8, 14 p. 6 LeBrasseur, R. J. 1964. Data record: a preliminary checklist of some marine plankton from the northeastern Pacific Ocean. Fish. Res. Board Can., Manuscr. Rep. Ser. (Oceanogr. Limnol.) 174, 14 p. 7 LeBrasseur, R. 1970. Larval fish species collected in zoo- plankton samples from the northeastern Pacific Ocean 1956- 1959. Fish. Res. Board Can. Tech. Rep. 175, 47 p. 8 LeBrasseur, R. J. 1965. Seasonal and annual variations of net zooplankton at Ocean Station P, 1956-1964. Fish. Res. Board Can., Manuscr. Rep. Ser. (Oceanogr. Limnol.) 202, 162 p. 9 Naplin, N.A., J. R. Dunn, and K. Niggol. 1973. Fish eggs, larvae and juveniles collected from the northeast Pacific Ocean, October-November 1971. NOAA-NMFS Northwest Fish. Cent., MARMAP Surv. I, Rep. 10, 39 p. + 121 tables. 10 Dunn, J. R., and N. A. Naplin. 1974. Fish eggs and larvae collected from waters adjacent to Kodiak Island, Alaska, during April and May 1972. NOAA-NMFS, Northwest Fish. Cent., MARMAP Surv. I, Rep. 12, 61 p. FISHERY BULLETIN: VOL. 75, NO. 1 MATERIALS AND METHODS Most data came from samples taken at 12 stations, located 2 to 111 km offshore along an east-west transect (lat. 44°39.1'N) off Newport, Oreg., just north of Yaquina Bay (Figure 1). The transect extended over the continental shelf and slope; depths ranged from 20 to 2,850 m. Samples were taken every month from January 1971 to August 1972 except in January and February 1972, although not every station was sampled 46' 45' 44' 43« 42' WASH 93 74 56 37 18 6 / mfxa/PORT III 65 46 28 9 2rS BROOKINGS 1 CALIF. I26 c 125* I24 e 123° FIGURE 1. — Location of the major bongo net sampling stations (circles) along an east-west transect (lat. 44°39.1'N) off Yaquina Bay, Oreg., and a 24-h station (square) occupied in May 1972. Numbers are kilometers from the coast. 126 RICHARDSON and I'EARCY: COASTAL ANDOCEANIC FISH LARVAE every month (Table 1). Of the 287 station oc- cupancies, 219 were made during daylight, 50 at night, and 18 at dusk or dawn. In addition, a series of replicate tows was made on 28-30 June 1971, which included two daytime and two nighttime hauls at stations 2, 6, and 9 and one daytime and one nighttime haul at stations 46, 56, 65, and 74. Samples were collected with a 70-cm (mouth diameter) bongo net without a closing mechanism. The bongos had two cylindrical-conical nets of 0.571-mm mesh Nitex 11 which were 4.6 m long and had a filtering area to mouth area ratio of about 10:1. Tsurumi-Seiki Kosakusho (TSK) flowmeters were positioned off center in the mouth of each net. A 40-kg multiplane kite-otter depres- sor (Colton 1959) was attached to the cable be- neath the bongos which produced a 2:1 wire out to depth fished ratio. A time-depth recorder (bathykymograph) was attached to the cable above the bongos to record depth and path of tow. The net was towed along depth contours parallel to the coast at a vessel speed of 2-3 knots. Tows were made obliquely through the water column in equal stepped intervals from the bottom or 150 m to the surface. Tow times ranged from 8 to 39 min and were usually between 10 and 30 min. Volume of water filtered ranged from 283 to 1,411 m 3 and was usually between 500 and 1,000 m 3 . At each station a bathythermograph (BT) cast was made to the bottom or 140 m, a surface bucket "Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. temperature was recorded, and surface and deep (bottom or 140 m) salinity samples were taken. Plankton samples were preserved at sea in 10% buffered Formalin. One sample from each bongo pair ( 287 samples) was sorted for fish larvae except for the replicate series where both samples of each pair (7 of the 287 samples plus 33 additional samples) were sorted. All fish larvae were re- moved from each sample and were stored in 5% buffered Formalin. Larvae were identified to the lowest possible taxonomic group, enumerated, and measured (standard length). Numbers of larvae from each sample were standardized to number under 10 m 2 of sea surface. This standardized number was used in all analyses unless indicated otherwise. In addition to the above samples, a 24-h station was occupied 18 km offshore at a location 46 km north of the Newport transect at lat. 45°04.0'N (Figure 1 ) on 30-31 May 1972. Water depth ranged from 158 to 164 m. Four depth strata (0-10, 11-50, 51-100, and 101-150 m) were sampled. Tows were designed to filter approximately the same volume of water in each stratum (x = 912 m 3 ± 142). The nonclosing bongo gear was lowered rapidly to the maximum depth of the zone to be sampled, towed obliquely through the depth zone in equally spaced steps, and then retrieved quickly to minimize contamination. Two tows were made in each depth stratum in daylight and again at night, which yielded 32 ( 16 pairs) samples. All fish larvae were sorted, identified, and enumerated. Numbers TABLE 1. — Summary of 287 station occupancies made on an east-west transect (lat. 44°39.1'N) off Yaquina Bay, Oreg., 1971-72. 2 6 9 18 Station (km from coast) 28 37 46 56 65 74 93 111 Month 20 46 59 85 95 Bottom depth (m) 142 330 220 340 1.060 1,300 2,850 1971: Jan. 2 2 2 2 Feb 2 2 2 2 — 2 — 2 — 1 1 1 Mar. 2 2 2 2 1 2 — 2 1 2 2 2 Apr. — — 1 1 1 1 1 1 — 1 1 — May 3 3 3 3 3 3 3 3 2 2 2 2 June 2 2 2 2 2 2 2 2 2 2 2 2 July 2 2 2 2 1 1 1 1 1 1 1 1 Aug. 2 2 2 2 2 2 2 2 2 2 2 1 Sept 1 1 1 1 1 1 1 1 1 1 1 1 Oct. 1 1 1 1 1 1 1 1 1 1 1 1 Nov. 1 1 1 1 1 1 1 1 1 1 1 1 Dec. 1 1 1 1 1 1 1 1 1 1 — — 1972: Jan. Feb. Mar. 3 2 3 3 3 3 3 3 2 2 2 2 Apr. 2 1 2 2 1 1 1 1 — — — — May 1 1 1 1 1 1 1 1 1 1 1 1 June 2 2 2 2 2 2 2 2 2 2 2 2 July 1 1 1 1 1 1 1 1 1 1 1 1 Aug. 1 1 1 1 1 127 FISHERY BULLETIN: VOL. 75, NO. 1 of larvae from each of these samples were stan- dardized to numbers per 1,000 m 3 of water filtered. TAXONOMIC PROBLEMS The 287 samples yielded 23,578 fish larvae in 27 families and 1 order (Table 2). To date 90 taxonomic groups have been identified, 78 at the species level, although 17 of these, primarily in the Cottidae and Stichaeidae, are still only numbered "larval types" 12 which are considered to be identified at the level of distinct species. These larval types have not yet been named because large specimens needed for positive identification were absent from the collections. This is the greatest number of species recorded from a larval fish study in the northeast Pacific which reflects, in part, refinements in larval fish identification as well as the intensity of the sampling effort which yielded many complete developmental series. Many of these larvae, particularly the coastal forms, have not yet been described in detail in the literature. While identification of many of the abundant larvae, particularly the pleuronectids and myctophids, has been accomplished with cer- tainty, a few major taxonomic problems remain, most notably with the osmerids and the scor- paenids, primarily Sebastes spp. We have not yet been able to identify the larval osmerids ( <30 mm) to species, of which there are five possibilities: Allosmerus elongatus, Hypomesus pretiosus, Spirinchus starksi, Spirinchus thaleichthys, and Thaleichthys pacificus. Available descriptions (Morris 13 ; Yap-Chiongco 1941; DeLacy and Batts 14 ; Dryfoos 1965; Moulton 1970) are in- adequate to distinguish all five species. We have not even established "larval types" below the family level. No attempt was made to separate Sebastes spp., another problem group, into "larval types" (species or species groups) although a few distinct kinds appeared to be present. Samples from Ore- gon waters may contain some 35 species and 12 The term larval type used in this paper refers to a particular kind of larva which may be distinguished from other larvae on the basis of larval characters but which has not yet been named. The term does not necessarily denote identification to the species level and is not intended to have any taxonomic implications. 13 Morris, R. Some notes on the early life history of the night surf smelt, Spirinchus starski (Fisk) 1913. Unpubl. manuscr., 37 p. 14 DeLacy, A. C, and B. S. Batts. 1963. A search for racial characteristics in the Columbia River smelt. Res. Fish., Fish. Res. Inst. Univ. Wash. Contrib. 147:30-32. identification of the larvae is difficult (Moser 1967, 1972; Moser et al. in press). One other problem group is the Cyclopteridae. Based on its broad distribution pattern, our Cyclopteridae spp. 1 probably represents a multispecies group, perhaps Liparis spp., but we have not yet been able to subdivide it on the basis of larval characters. These identification problems impose limita- tions on analysis of ichthyoplankton data. Caution must be exercised in interpretation of results when multispecies groups constitute a major proportion of larvae taken, such as Sebastes spp. and osmerids off Oregon. SAMPLING VARIABILITY A series of replicate oblique tows (four day and four night samples at stations 2, 6, 9; two day and two night samples at stations 46, 56, 65, 74) made in June 1971 was examined to assess sampling variability. Species composition of day and night tows at a station was similar, based on common larvae collected and their relative rank abun- dance. Total larvae in night catches exceeded those in day catches at all stations except 65 and 74 (Figure 2). Large day-night differences oc- curred at stations 6 and 9. This was primarily due to increased catches of large (>23 mm) osmerid larvae at night (Figure 3), which presumably avoided the net by day or were deeper, although 76 to 87% of the water column was sampled in daytime. Even so, osmerids were the most abundant larvae captured in all samples from these two stations. At station 2, the increased night catches were due to an increase in the numbers of large larvae (including osmerids), as well as an increase in the number of species captured (7-10 in daytime vs. 13-14 at night). Both Isopsetta isolepsis (most >16.5 mm) and Micro- gadus proximus (most >29 mm), species com- mon at stations 6 and 9 during day and night, were collected only at night at station 2. At sta- tions 46 and 56, night catches yielded increased numbers of Engraulis mordax (4-10 mm) and Stenobrachius leucopsarus (4-15 mm) while night catches of Sebastes spp. (3-9 mm) were half the daytime numbers (3-12 mm). At station 65, E. mordax (6-10 mm) was again more abundant in night tows while Stenobrachius leucopsarus was much less abundant at night, composing only 10 and 34% of the numbers of larvae in the two night- time tows (6-13 mm) but 61 and 54% in the two 128 RICHARDSON and PEARCY: COASTAL AND OCEANIC FISH LARVAE TABLE 2. — Species composition 1 and abundance 2 offish larvae taken 2 to 111 km off of Yaquina Bay, Oreg., from January 1971 to August 1972. Taxa Total standardized abundance 2 Coastal Offshore Taxa Total standardized abundance 2 Coastal Offshore 1.17 1.55 1.75 13.27 1.48 0.87 28.43 1.16 3.14 15.85 34.09 79.70 4.45 27.04 6.75 32.47 3.35 0.32 0.70 33.81 37 80 1.03 1.37 1.12 0.77 5.53 1.04 6.56 1.09 0.70 71.17 15.60 258.50 2.31 60.30 1.80 2.59 7.53 57.19 4.80 0.64 7.09 1.57 18.27 113.81 2.70 259 1,157.90 12.53 1.31 96.54 475.23 8.24 81.74 1,479.59 37.62 187.40 1.72 308.12 1.13 16.84 17.22 47.71 49.09 11,474.46 10,868.04 Clupeidae: + Clupea harengus pallasi (c) Engraulidae: + - Engraulis mordax (o) Osmeridae: + - Undetermined spp. (c) Bathylagidae: - Bathylagus milleri (o) - Bathylagus ochotensis (o) - Bathylagus pacificus (o) Melanostomiatidae: - Tactostoma macropus (o) Chauliodontidae: - Chaulidous macouni (o) Paralepididae: - Lestidiops ringens (o) Myctophidae: + - Lampanyctus regalis (o) - ?Loweina rara 3 (o) - Protomyctophum crockeri (o) + - Protomyctophum thompsoni (o) + - Stenobrachlus leucopsarus (o) + - Tarletonbeania crenulahs (o) - Undetermined spp. (o) Gadidae: + - Microgadus proximus (c) Ophidiidae: - Brosmophycis marginata (o) - Ophidiidae sp. 1 (o) Scorpaenidae: + - Sebastes spp. (o) + - Sebastolobus spp. (o) Hexagrammidae: + - Hexagrammos spp. (o) + - Ophlodon elongatus (c) Anoplopomatidae: + - Anoplopoma fimbria (o) Cottidae: Artedius sp. 1 (c) Artedius sp. 2 (c) Chitonotus pugetensis (c) Cottus asper (c) Enophrys bison (c) Hemilepidotus hemilepidotus (c-o) Hemilepidotus spinosus (c-o) Icelinus sp. 1 (c) Leptocottus armatus (c-o) • Nautichthys oculofasciatus (c) Oligocottus sp. 1 (c) Paricelinus hopliticus (c) Psychrolutes-hke sp. 1 (o) Radulinus asprellus (c) Rhamphocottus richardsoni (c) Scorpaenichthys marmoratus (c) Cottidae sp. 1C (c) Cottidae sp. 12 (c) Cottidae sp. 19 (c) Cottidae sp. 20 (c) Undetermined spp. (c) Agonidae: + + + + - + - + - + - + - + + + 64.19 + Agonopsis emmelane (c) + - Bathyagonus spp. (c-o) 13.39 1,000.70 + Occella verrucosa (c) + Odontopyxis trispinosa (c) 5.749.53 13.65 + Pallasina barbata (c) + Stellerina xyosterna (c) 2.90 + Zeneretmus latifrons (c) 131.46 + Agonidae sp. 6 (c) 34.18 Cyclopteridae: + Lipans pulchellus (c) 2.05 + - Cyclopteridae spp. 1 (c-o) + Cyclopteridae sp. 3 (c) 29.47 + - Undetermined spp. (c) Bathymasteridae: 5.78 + - Ronquilus jordani (c) Blennioids: 0.82 37.04 + Undetermined spp. (c) 1.15 Clinidae: 34.03 + Gibbonsia Imontereyensis (c) 9.97 173.77 Stichaeidae: 45.30 3,648.00 + Anoplarchus sp. 1 (c) 2.29 635.20 + Chirolophis sp. 1 (c) 7.24 + Lyconectes aleutensis (c) + Lumpenus sagitta (c) 580.28 5.44 + Plectobranchus evides (c) + Stichaeidae sp. 1 (c) 2.86 + - Stichaeidae sp. 2 (c) 1.32 + Stichaeidae sp. 4 (c) Ptilichthyidae: 180.66 3,967.82 + Ptilichthys goodei (c) 0.60 19.21 Pholidae: + Apodichthys flavidus (c) 0.44 2.94 + Pholis spp. (c) 53.44 1.24 Icosteidae: - Icosteus aenigmaticus (o) 0.93 7.34 Ammodytidae: + Ammodytes hexapterus (c) 189.26 7.94 Gobiidae: 139.96 + Clevelandia ios (c) 7.55 Centrolophidae: 145.43 - Ichichthys lockingtoni (o) 60.65 6.63 Bothidae: 13.13 6.44 - Citharichthys sordidus (o) 69.04 29.78 + Citharichthys stigmaeus (c) 54.46 1.94 + - Citharichthys spp. 4 (o) 18.60 5.50 Pleuronectidae: 0.77 - Atheresthes stomias (o) 3.15 + - Embassichthys bathybius (o) 0.79 - Eopsetta jordani (o) 2.21 + - Glyptocephalus zachirus (o) 58.45 9.19 + - Hippoglossoides elassodon (c-o) 0.77 + - Isopsetta isolepis (c) 21.84 + - Lepidopsetta bilineata (c) 5.94 + - Lyopsetta exilis (o) 42.70 + - Microstomus pacificus (o) 0.33 + - Parophrys vetulus (c) 1.12 + - Platichthys stellatus (c) 21.55 + - Psettichthys melanostictus (c) Unidentified larvae Fragments 1 General distribution patterns are given for each taxon: + = taken 2 to 28 km offshore - = taken 37 to 1 1 1 km offshore c = coastal type ( >80% of all larvae taken 2 to 28 km from coast) o = offshore type ( >80% of all larvae taken 37 to 1 1 1 km from coast) c-o = neither c or o type (<80% of all larvae taken in either coastal or offshore area). 2 The sum of the standardized numbers (number under 10 m 2 sea surface) of larvae from each sample in the coastal (2-28 km) and offshore (37-111) km assemblages (139 and 148 samples, respectively), identification based on one partly mutilated specimen. 4 Specimens too small to identify to species. 129 FISHERY BULLETIN: VOL. 75, NO. 1 lO.OOOr- E o < > rr < Ll_ o cr UJ QQ _l < 1000 DAY • NIGHT O DAY 100 t o § o (§> <8 10 8 2 6 9 46 56 STATIONS 65 74 FIGURE 2. — Day and night catches offish larvae on transect off Yaquina Bay, Oreg., June 1971. daytime tows (4-16 mm). Decreased larval abun- dances at night at station 74 were due mainly to reduced numbers of S. leucopsarus (5-13 mm at night, 5-16 mm in day). Thus avoidance of the net by large larvae in daytime seemed to account for much of the day-night variation at the coastal sta- tions 2, 6, and 9. Differences at the offshore stations may have been due to patchiness of small larvae. Variability among repeated samples was examined at the three inshore stations where four day and four night replicate samples were taken at each station. Coefficients of dispersion were calculated for total larvae, osmerids, and total larvae minus osmerids (Table 3). Values were close to 1.0 for total larvae minus osmerids at 200 r- 100 900 1- % < _i 800 700 Ll. o 600 o rr 00 500 o o 4 00 • 1 300- 200- 100- 20 30 STANDARD LENGTH (mm) FIGURE 3. — Day and night length frequencies of osmerid larvae collected at 6 and 9 km off Yaquina Bay, Oreg., June 1971. Numbers of larvae were combined for both nets from four day and four night hauls. stations 6 and 9 and for total larvae at station 2 where osmerids were not abundant suggesting that larvae were randomly distributed. Coeffi- cients were large, however, for total larvae and for osmerids at 6 and 9 where smelt larvae were abundant, except at night at station 9. These large coefficients of dispersion indicate high con- tagion, possibly caused by schooling behavior of large osmerid larvae. TABLE 3. — Coefficients of dispersion (s 2 /x) for total larvae, osmerids, and total larvae minus osmerids in replicate tow series made in June 1971 on the transect (lat. 44°39.1'N) off Yaquina Bay, Oreg. Station 2 Day Night Station 6 Station 9 Item Day Night Day Night Total larvae Osmendae Total larvae minus Osmendae 0.49 0.97 12.44 16.40 0.81 11.96 12.81 3.18 11.56 0.57 14.49 0.82 0.81 1.23 VERTICAL DISTRIBUTION One attempt was made to study the vertical distribution patterns of larvae in the coastal zone 18 km offshore north of the Newport transect (Figure 1). Thirty-two samples were taken within four depth strata (0-10, 11-50, 51-100, and 101-150 130 RICHARDSON andPEARCY: COASTAL AND OCEANIC FISH LARVAE m) during a 24-h period in May 1972. Essentially, the entire water column was sampled. The volume of water filtered by each type of tow was about the same and the number of day and night tows in each stratum was equal. Because the nets had no opening-closing device, samples from all but the 0- to 10-m stratum were contaminated with catches from overlying waters. However, the maximum tow time spent outside the desired stratum was 20% for the deepest tows and was usually <10% for the intermediate depths. Therefore, no cor- rection factor was applied to the data. The greatest number of larvae and taxa was taken near the surface both day and night (Table 4). The 51- to 100-m stratum yielded the fewest larvae and taxa while the 11- to 50- and 101- to 150-m strata were intermediate. More larvae were taken at night, primarily in the 0- to 10-m stratum where avoidance during the day would be expected to be greatest. Mean larval length in this stratum was much greater at night which also indicated daytime avoidance by large larvae in surface waters. Mean larval length was also high in the 101- to 150-m stratum day and night, primarily because of the abundance of large osmerids there. Of the 22 taxa taken, those represented by more than 10 larvae were examined for trends in dis- tribution (Table 4). Clupea harengus pallasi (25-31 mm, x 28), Ammodytes hexapterus (17-37 mm,x33), and Ronquilus jordani (6-21 mm,f 13) were concentrated in the upper 10 m at night and were completely absent in daytime collections from all depths. They exhibited strong daytime avoidance, indicated by night/day ratios. Large Sebastes spp. larvae (9-11 mm, x 10) were only taken at night and perhaps avoided by day, whereas small larvae (3-4 mm, x 4) were taken both day and night in the upper two strata. Stenobrachius leucopsarus (5-11 mm, x 8) and Isopsetta isolepis (14-23 mm, x 20) occurred predominantly in the upper two strata but showed no evidence of daytime avoidance. Mean larval lengths were about the same by day and night. Of the remaining taxa, Radulinus asprellus (9-15 mm, x 12) appeared to occur throughout the water column in similar numbers and lengths during both day and night. Cyclopteridae spp. 1 (4-8 mm,* 5) occurred mainly near the surface in daytime but only in the 51- to 100-m stratum at night, possibly a result of patchiness or con- tamination of the deeper hauled net in the surface stratum. Only osmerids occurred primarily near the bottom (101-150 m), by day and night. Some were taken near the surface at night which may indicate vertical migration by some individuals or avoidance by day. Preliminary examination of specimens did not reveal the surface- and bottom- occurring osrrierid larvae to be different species. Mean lengths for deep- and surface-caught os- merids were about the same, 21 and 23 mm. ASSEMBLAGES Two separate assemblages of fish larvae were distinguished, using a similarity coefficient ma- trix based on Sander's (1960) dominance-affinity index (J lowest percent of all larvae in common between two stations). In 1971 a coastal as- semblage occurred at stations 2 to 28 km offshore, which was distinct from another assemblage occurring at stations farther offshore (Figure 4). A similar pattern was found in 1972 during the 6 mo for which data were available. In 1971, the mean affinity value among stations 2, 6, 9, 18, and 28 was 65.81 and among stations 46, 56, 65, 74, 93, and 111 it was 60.61. In 1972, the mean affinity values for these same sets of stations were 43.21 and 56.61, respectively. Sebastes spp. were TABLE 4. — Number/l,000m 3 , number of taxa, and mean length offish larvae by day, night, and depth strata taken during a 24-h period 18 km off the mid-Oregon coast (lat. 45°04.0'N) in May 1972. N/D = night to day ratio. Each number is the sum of four replicate samples. CO O) c s CO CO -C CD Q. Q. CO CO IB CO T3 to <0 3 Depth strata co CD CD CD = a.™ 3 Q. CO T3 *i_ CD E CO O CO 11 CO 2 CD CO 3 C $ CD « 70 00 15000-6999 [33000-4999 D<30 00 FIGURE 4. — Station to station similarity-coefficient matrices for 1971 and 1972 data on larval fishes based on Sander's (1960) dominance affinity index. All taxa except Sebastes spp. were included in the analysis. excluded from the analysis to minimize masking effects that might have arisen because of the multispecies nature of the group. Since osmerids were known to be essentially coastal forms, they were not excluded. Peaks in larval abundances were associated with the location of these two assemblages with an apparent transitional zone of low larval abun- dance between them (Figure 5). In both 1971 and 1972 abundance was relatively high inshore, dropped to a low at 28 km, and then increased seaward. Larval taxa were determined to be associated with the coastal or offshore zone on the basis of whether 80% or more of all larvae were taken at stations 2 to 28 (coastal = C) or stations 37 to 111 (offshore = O). Using these criteria, 84 of the 90 taxa (93%) could be designated as coastal or offshore (Table 2). Fifty-three taxa in 16 families and 1 order were coastal. Of these, 49 were identified to species, 3 to family, and 1 to order. Ninety-nine percent of all larvae in these 53 taxa were taken in the coastal zone 2 to 28 km offshore. Thirty-one taxa in 15 families were offshore. Of these, 26 were identified to species, 4 to genus, and 1 to family. Ninety-six percent of all larvae in these 31 taxa were taken 37 to 111 km offshore. Only six taxa could not be designated as coastal or offshore. This was probably due in part to rarity, e.g., Hippoglossoides elassodon (total standard- ized number = 5.29; 51% were C and 49% were 0),Bathyagonus spp. (3.30; 47% C and 53% O), and to multispecies groups, e.g., Cyclopteridae spp. 1 (30% C and 70% O) and Bathyagonus spp. In- terestingly, 96% of all Sebastes spp. larvae were taken in the offshore area. Leptocottus armatus was primarily coastal since 77% of all larvae were taken there. Only one sample outside the coastal area (Station 37, in February 1971) contained L. armatus larvae, but they were present in moder- ate numbers. Hemilepidotus hemilepidotus (67% C and 33% O) and//, spinosus {IWc C and 30% O) distributions are more difficult to explain. Hemilepidotus spinosus larvae in the coastal area were smaller (4-9 mm, x 5.3) than those farther offshore (6-12 mm, x 8.9) as were//, hemilepidotus (4-6 mm,x 5.2 in the coastal area and 8-11 mm, x 9.3 offshore). Hemilepidotus spinosus larvae are sometimes abundant (>600 larvae/15 min tow) in the neuston (upper 15 cm of the water column), particularly at night (Richardson unpubl. data). These data suggest that larvae which are as- sociated with surface waters may undergo some kind of offshore transport which does not affect nonneustonic species. Modes of reproduction differ considerably between those species designated as coastal and those designated as offshore. Of the 53 coastal taxa 132 RI< II I ARDSON and PE ARCY: COASTAL AND OCEANIC FISH LARVAE 440f 400 360 320 280- 240- O 200- 160 120- > K 80 < O 4 0h -z. 1971 mm FEB -MAR -APR I 1 MAY -JUN-JUL C—l AUG-SEP-OCT-NOV-DEC tl i ti fa . 2 6 9 28 37 46 56 65 FIGURE 5.— Mean standardized abundance offish larvae by station in 1971 and 1972. 200 160 120- 1 — TT 74 93 1972 MAR -APR CZ) MAY -JUN-JUL \lA 2 6 9 (Table 2), 87% presumably come from demersal eggs (Breder and Rosen 1966) including all the osmerids, cottids, agonids, cyclopterids, and blennioids as well as Clupea harengus pallasi, Ophiodon elongatus, Ronquilus jordani, Am- modytes hexapterus, andClevelandia ios. The eggs of Microgadus proximus are unknown but may also be demersal, as are those of M. tomcod in the Atlantic. Those not derived from demersal eggs, i.e., the six coastal flatfishes, come from small (~1 mm or less in diameter) planktonic eggs. Of the 31 offshore taxa, 81% presumably come from plank- tonic eggs. Eggs of the bathylagids, myctophids, bothids, and Engraulis mordax are probably all relatively small (~1 mm or less) whereas those of Chauliodus macouni, Anoplopoma fimbria, Icos- teus aenigmaticus, Atheresthes stomias, Embas- sichthys bathybius, Glyptocephalus zachirus, and Microstomas paciftcus are large, usually >2 mm. Eggs of Tactostoma macropus, Icichthys locking- is 28 37 46 56 65 STATIONS 74 93 III toni, Eopsetta jordani, and Lyopsetta exilis are in- termediate in size. Eggs of Sebastolobus spp., also of intermediate size, occur in floating masses rather than individually (Pearcy 1962). Larvae of the live-bearers Brosmophycis marginata, Sebastes spp., and possibly Ophidiidae sp. 1 are extruded. Of the offshore taxa, only Hexagrammos spp. and perhaps Psychrolutes-like sp. 1 come from demersal eggs. Coastal Assemblage One hundred thirty-nine samples were taken in the coastal assemblage, five at night, four at dusk or dawn, and the rest during daylight. All but four samples contained larvae, yielding 16,197 specimens or a standardized total [^ (number of larvae under 10 m 2 sea surface in each sample)] of 11,474. 133 FISHERY BULLETIN: VOL. 75, NO. 1 Species Composition and Dominance Seventy-three taxa assigned to 19 families and 1 order were taken in the coastal samples (Table 2). Of these, 62 were identified to species including unnamed numbered larval types considered to be distinct species, 7 to genus, 3 to family, and 1 to order. Margalef's (1958) formula for diversity (D = S — 1/ln N, where S = number of species, N = total number of individuals), which provides a measure of species richness, yielded a value of 7.43 for the coastal assemblage, which was higher than that for the offshore assemblage. Dominant taxa within the coastal assemblage were determined by a ranking method (Biological Index = BI) modified from Fager (1957), which takes into account both abundance and frequency of occurrence. By this method, the most abundant species in each sample is given five points, the next four, etc. Scores for each taxon are summed for all positive samples and divided by the total number of samples taken. The top 13 coastal dominants 15 (Table 5) accounted for 91.8% of the total larvae captured within 28 km of the coast over the entire sampling period. These same 13 taxa were also the 13 most abundant, although not always in the same order as dominance. Osmerids were overwhelmingly the most dominant taxonomic group making up 50% of the total larval catch. They were the most abundant and most frequently taken larvae in the coastal assemblage. Parophrys vetulus and Isopsetta isolepis were also important in terms of abun- dance. These three taxa, together with fourth ranked Microgadus proximus, composed 78% of all larvae taken. Seasonality Obvious trends in seasonality were apparent from the 1971 data, which included samples from every month (Figure 6). Ninety-three percent of all larvae were taken during the 6-mo period from February through July. Two abundance peaks occurred within that period, one in February- March (24% of all larvae) before upwelling, and one in May-July (68% of all larvae) during the upwelling season. Larval abundance decreased greatly in August and remained low through December. Mean number of larvae under 10 m 2 was 142 in February-March, 202 in May-July, and e o < > q: < o 1) in the February-March period were P. vetulus (BI = 4.09), Ammodytes hexapterus (BI = 1.76), /. isolepis (BI = 1.73), and Osmeridae (BI = 1.51). Together they made up 70% of the total larvae. Parophrys vetulus alone accounted for 44%. Dominant taxa from May to July 1971 were Osmeridae (BI = 4.12), /. isolepis (BI = 2.21), M. proximus (BI = 2.03), and Lyopsetta exilis (BI = 1.07). Together they made up 90% of the total number of larvae in those months. Osmerids accou