^ 0f c 0a Fishery Bulletin Sr ATES O* + SEp * 1982 \ Vol. 80, No. 1 January 1982 MANOOCH, CHARLES S., Ill, and CHARLES A. BARANS. Distribution, abundance, and age and growth of the tomtate, Haemulon aurolineatum, along the southeastern United States coast 1 MURAWSKI, STEVEN A.. JOHN W. ROPES, and FREDRIC M. SERCHUK. Growth of the ocean quahog, Arctica islandica, in the Middle Atlantic Bight. . . 21 TUCKER, JOHN W., JR. Larval development of Citharichthys cornutus, C. gym- norhinus, C. spilopterus, and Etropus crossotus (Bothidae), with notes on larval occurrence 35 WIEBE, P. H., S. H. BOYD, B. M. DAVIS, and J. L. COX. Avoidance of towed nets by the euphausiid Nematoscetis megalops 75 LAROCHE, JOANNE LYCZKOWSKI, SALLY L. RICHARDSON, and AN- DREW A. ROSENBERG. Age and growth of a pleuronectid, Parophrys vetulus, during the pelagic larval period in Oregon coastal waters 93 HJORT, R. C, and C. B. SCHRECK. Phenotypic differences among stocks of hatchery and wild coho salmon, Oncorhynchus kisutch, in Oregon, Washington, and California 105 BAGLIN, RAYMOND E., JR. Reproductive biology of western Atlantic bluefin tuna 121 IRVINE, A. B., R. S. WELLS, and M. D. SCOTT. An evaluation of techniques for tagging small odontocete cetaceans 135 Notes CONOVER, DAVID 0., and STEVEN A. MURAWSKI. Offshore winter migra- tion of the Atlantic silverside, Menidia menidia 145 ROSENBERG, ANDREW A., and JOANNE LYCZKOWSKI LAROCHE. Growth during metamorphosis of English sole, Parophrys vetulus 150 PRATT, HAROLD L., JR., JOHN G. CASEY, and ROBERT B. CONKLIN. Ob- servations on large white sharks, Carcharodon earcharias, off Long Island, New York 153 GIBSON, D. M. A note on the estimation of trimethylamine in fish muscle 157 CRASS, DENNIS W., and ROBERT H. GRAY. Snout dimorphism in white stur- geon, Acipenser transmontawus, from the Columbia River at Hanford, Washing- ton 158 V. Seattle, Washington U.S. DEPARTMENT OF COMMERCE Malcolm Baldrige, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION John V. Byrne, Administrator NATIONAL MARINE FISHERIES SERVICE William G. Gordon, Assistant Administrator Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, ant economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries ir, 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. 1 103. 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 bulle- tin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodi- cal, issued quarterly. In this form, it is available by subscription from the Superintendentof 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. Carl J. Sindermann Scientific Editor, Fishery Bulletin Northeast Fisheries Center Sandy Hook Laboratory National Marine Fisheries Service, NOAA Highlands, NJ 07732 Editorial Committee Dr. Bruce B. Collette Dr. Donald C. Malins National Marine Fisheries Service National Marine Fisheries Service Dr. Edward D. Houde Dr. Jerome J. Pella Chesapeake Biological Laboratory National Marine Fisheries Service Dr. Merton C. Ingham Dr. Jay C. Quast National Marine Fisheries Service National Marine Fisheries Service Dr. Reuben Lasker Dr. Sally L. Richardson National Marine Fisheries Service Gulf Coast Research Laboratory The Fishery Bulletin (USPS 090-870) is published quarterly by Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE. BIN C15700, Seattle, WA 98115. Second class postage paid to Finance Department. USPS, Washington, DC 20260. Although the contents have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the public business required by law of this Department Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget through 31 March 1982. Fishery Bulletin CONTENTS Vol. 80, No. 1 January 1982 MANOOCH, CHARLES S., Ill, and CHARLES A. BARANS. Distribution, abundance, and age and growth of the tomtate, Haemulon aurolineatum, along the southeastern United States coast 1 MURAWSKI, STEVEN A., JOHN W. ROPES, and FREDRIC M. SERCHUK. Growth of the ocean quahog, Arctica islandica, in the Middle Atlantic Bight. . . 21 TUCKER, JOHN W., JR. Larval development of Citharichthys cornutus, C. gym- norhinus, C. spilopterus, and Etropus crossotus (Bothidae), with notes on larval occurrence 35 WIEBE, P. H., S. H. BOYD, B. M. DAVIS, and J. L. COX. Avoidance of towed nets by the euphausiid Nematoscelis megalops 75 LAROCHE, JOANNE LYCZKOWSKI, SALLY L. RICHARDSON, and AN- DREW A. ROSENBERG. Age and growth of a pleuronectid, Parophrys vehdus, during the pelagic larval period in Oregon coastal waters 93 HJORT, R. C, and C. B. SCHRECK. Phenotypic differences among stocks of hatchery and wild coho salmon, Oncorhynchus kisutch, in Oregon, Washington, and California 105 BAGLIN, RAYMOND E., JR. Reproductive biology of western Atlantic bluefin tuna 121 IRVINE, A. B., R. S. WELLS, and M. D. SCOTT. An evaluation of techniques for tagging small odontocete cetaceans 135 Notes CONOVER, DAVID 0., and STEVEN A. MURAWSKI. Offshore winter migra- tion of the Atlantic silverside, Menidia menidia 145 ROSENBERG, ANDREW A., and JOANNE LYCZKOWSKI LAROCHE. Growth during metamorphosis of English sole, Parophrys vetulus 150 PRATT, HAROLD L., JR., JOHN G. CASEY, and ROBERT B. CONKLIN. Ob- servations on large white sharks, Carcharodon carcharias, off Long Island, New York 153 GIBSON, D. M. A note on the estimation of trimethylamine in fish muscle .... 157 CRASS, DENNIS W., and ROBERT H. GRAY. Snout dimorphism in white stur- geon, Acipenser transmontanus, from the Columbia River at Hanford, Washing- ton 158 Seattle, Washington 1982 For sale by the Superintendent of Documents. U.S. Government Printing Office, Washington. DC 20402— Subscription price per year: $15.00 domestic and $18.50 foreign. Cost per single issue: $4.50 domestic and $5.05 foreign. The National Marine Fisheries Service (NMFS) does not approve, recommend or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication fur- nished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary prod- uct 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 pur- chased because of this NMFS publication. DISTRIBUTION, ABUNDANCE, AND AGE AND GROWTH OF THE TOMTATE, HAEMULON AUROLINEATUM, ALONG THE SOUTHEASTERN UNITED STATES COAST 1 Charles S. Manooch III 2 and Charles A. Barans 3 ABSTRACT Tomtates, Haemulon aurolineatum, were widely distributed over sponge-coral habitats throughout the South Atlantic Bight region in depths of 9 to 55 m, although they were occasionally caught in large numbers over sandy bottom habitats. Fish were most common in offshore areas during winter and were not taken in waters of <10°C south of Cape Fear, N.C. Juveniles (<148 mm TL) were caught in the same geographical areas as adults, but were collected in warmer waters than adults during fall and winter. Spawning occurred during the spring. Individuals collected by hook and line and trawl were aged by scales and otoliths. Back-calculated mean total lengths were from 103.0 mm at age I, to 280.5 at age IX. The von Bertalanffy growth equation is I, =310(1 -exp - 0.22017 (t + 1.28)), where t is age in years, and /, is total length at age. The oldest fish sampled was age IX, 289 mm TL. Annual total mortality based on catch curves from 1,496 fish landed by the recreational fishery from 1972 to 1978 was 59% (instantaneous total annual mortality = 0.89). We found that the tomtate grows faster, does not live as long, and has a higher natural mortality rate than most other reef fishes previously studied in the South Atlantic Bight. The tomtate, Haemulon aurolineatum, is a small grunt (Haemulidae), which occurs from Cape Cod, Mass., to Brazil, including the Caribbean, Gulf of Mexico, and Central American coast. The species, previously referred to as Bathystoma rimator, B. aurolineatum, and Haemulon rima- tor (Courtenay 1961), is known vernacularly as xira in Brazil; cuji in Venezuela; rancho, juez, and chankay in Mexico; and mulita, mula, mari- quita, and maruca in Puerto Rico. The tomtate is taken primarily by hook and line off the southeastern United States and by traps, hook and line, and trawl in the more south- ern areas of its range. Unfortunately, commer- cial landings of tomtates in the United States are reported in the collective term "grunts," which includes many different species of the family and therefore precludes species identifications that are needed for fishery management. A Soviet- Cuban cooperative fisheries research program on the Campeche Banks revealed the tomtate as 'Contribution No. 192, Southeast Fisheries Center Beau- fort Laboratory, National Marine Fisheries Service, NOAA; No. 189, MARMAP Program; No. 127, South Carolina Marine Resources Center, Marine Resources Institute. 2 Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. 3 South Carolina Wildlife and Marine Resources Depart- ment, Marine Resources Research Institute, Charleston, SC 29412. the main demersal species caught by trawl from 1962 to 1972 (Sokolova 1969; Sauskan and Olae- chea 1974). Also, exploratory trawling off South Carolina found large quantities of tomtates (Wenner et al. 1979a). Recreational headboat 4 fishermen fishing from North Carolina to Cape Canaveral, Fla., caught an average of 23.2 t (metric tons) of tom- tates in 1976 and 1977 (Dixon 5 ). This species was the most commonly caught haemuline, although second in weight landed to the white grunt, Haemulon plumieri. In this paper we describe the relative abun- dance, spatial and temporal distributions, spawning, age, growth, and mortality for tom- tates along the southeastern United States. METHODS Distribution and Relative Abundance Eight groundfish survey cruises spanning all four seasons (Table 1 ) were conducted on the con- 4 A boat for hire where anglers are charged on a per person basis. 6 R. L. Dixon, Southeast Fisheries Center Beaufort Labora- tory, NMFS, NOAA, Beaufort, NC 28516, pers. commun. Jan- uary 1978. Manuscript accepted September 1981. FISHERY BULLETIN: VOL. 80, NO. 1, 1982. FISHERY BULLETIN: VOL. 80, NO. 1 Table 1.— Groundfish cruises of the RV Dolphin. No. ot No. of tows No. of Cruise Dates trawls with tomtates tomtates DP-7305 23 Oct. -16 Nov. 1973 86 18 2,075 DP-7402 1 Apr. - 9 May 1974 112 19 442 DP-7403 13 Aug. -19 Sept 1974 87 14 581 DP-7501 16 Jan. -10 Apr. 1975 92 10 1,212 DP-7503 30 Aug. -19 Sept 1975 87 20 1,298 DP-7601 12 Jan. - 7 Feb. 1976 86 15 4,005 DP-7603 28 Aug. -21 Sept 1976 89 15 1,749 DP-7701 17 Jan. - 9 Mar. 1977 93 11 3,260 tinental shelf and upper continental slope be- tween Cape Fear, N.C., and Cape Canaveral, Fla., except in spring 1974 when sampling ex- tended to Cape Hatteras. A preassigned number of stations was selected randomly (Grosslein 1969) with a set number in each of six depth zones (9-18 m; 19-27 m; 28-55 m; 56-110 m; 111- 183 m; 184-366 m). Bottom water temperatures were measured at each station with mechanical or expendable bathythermographs. Thirty-minute trawls were made continuously (day and night) from the RV Dolphin, at 6.5 km/h with a towing wire scope of 2.5-3.0:1. The trawl was a 3/4-scale version of a "Yankee No. 36" with a 16.5 m footrope, 11.9 m headrope, and 1.3 cm stretch mesh cod end liner (Wilk and Silverman 1976). Fork lengths (later converted to total lengths) of all fish collected by trawl were recorded to the nearest centimeter. Frozen fish samples were taken to the laboratory for further investiga- tions. An index of relative abundance (Musick and McEachran 1972) was calculated for each depth zone by the following expression: Index of Relative Abundance 2 1n(x+l) n h where n h = number of trawls in the Mh depth zone, and x = number of individuals for each tow in a given depth zone. Because previous investi- gators have shown that trawl catches are usually distributed as a negative binomial (Elliott 1971; Taylor 1953), a In (x + 1) transformation was made on the relative abundance data to permit statistical tests to determine if the differences among habitats within depth zones were signifi- cant. Estimates of biomass standing stock were cal- culated with both transformed, In (x + 1), and untransformed data for comparison of the result- ing values. The stratified mean catch/tow (Coch- ran 1977) was calculated by the expression: *- = N k iNhVh] where y st = stratified mean catch(kg)/tow, N = total area, Nh = area of Mh depth zone (from plani- meter chart measurements), y h = mean catch/tow in the Mh depth zone, and k = number of zones in the set. The area of live-bottom habitat in each depth zone («44.5%) was estimated from the frequency of occurrence of sponge and coral in catches dur- ing 5 yr of bottom trawling with the stratified random sampling design. The areas of sandy- bottom habitats were obtained by subtraction. The estimated population variance of the mean catch(kg)/tow was also calculated by Clark and Brown (1977): S 2 = N mi [N h y 2 ]- NyJ + ! Sf \(N h - *=i r + (N h -N)(N h -n h ) N n h ] where S = estimated population variance, and S h 2 = variance of the Mh zone. The mean catch/tow (y h ) of the transformed In (x + 1) data was estimated for each depth zone following the methodology of Bliss (1967): E(y h ) = exp (y h + S 2 /2) where E(y h ) = the estimated (retransformed) mean catch(kg)/tow in the Mh depth zone, y h and S 2 , both expressed in logarithmic units, are the zone mean and its variance. The same methodol- ogy was applied to obtain the stratified mean catch/tow from transformed data for the whole study area. Biomass estimates were expanded by the area swept method (Rohr and Gutherz 1977), using S&ot = X (P h ) (A„) h -1 where SStot = total standing stock, MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE P = average population expressed as kilograms per km 2 in the /?th depth zone, and A h = total area of the Mh depth zone. The sweep of the "3/4 Yankee trawl" was 8.748 m (Azarovitz 6 ), and 3.241 km was the distance cov- ered during a standard trawl. It should be noted that all estimates were minimum estimates be- cause the sampling efficiency of our gear with regard to tomtates was unknown. Standing stock values calculated for sandy-bottom areas incor- porated such a large number of zero catches that the transformation did not normalize the data, so the resulting values should be considered sus- pect. Age and Growth Scales, otoliths, fish lengths, and fish weights were collected from 1,496 tomtates from the rec- reational headboat fishery operating from North Carolina to Cape Canaveral from 1972 through 1978 and from approximately 100 juvenile fish collected by research trawling off South Caro- lina. Total fish length was recorded in milli- meters and weight in grams. Scales were removed from beneath the tip of the posteriorly extended pectoral fin, soaked in a one-tenth aqueous solution of phenol, cleaned and mounted dry between two glass slides, and viewed at 40X magnification on a scale projector. Measurements were made and recorded from the scale focus to each annulus and to the scale edge in the anterior field for marginal increment analyses and back-calculating fish length at the time of annulus formation. Otoliths (sagittae) were removed by making a transverse cut in the cranium with a hacksaw midway between the posterior edge of the orbit and the preopercle. The skull was pried open and the otoliths were removed with forceps, washed in water, and stored dry in labeled vials. Rings were counted by placing the otoliths in a black- ened-bottom watch glass and then viewing the structures through a binocular dissecting micro- scope with the aid of reflected light. Some of the otoliths from large (older) fish were sectioned with a Buchler, Isomet, 1 1-1 180 7 low-speed saw to facilitate aging. Measurements were not re- corded from otoliths since these structures were used only as a method of validating age deter- mined by reading scales. Lengths by age for fish from all years com- bined were back-calculated from a scale radius- fish length regression. The regression equation was based on the relationship of magnified (40X) scale length to total fish length. Since a majority of the scale measurements were clustered around a relatively narrow size range, we based our regression on a subsample of scale radius and body length measurements. After grouping the measurements into 25 mm body length intervals, we selected approximately 12 from each interval to ensure that the regression provided good representation. The prediction equation took the form TL = a SR 1 '; where TL = total length, SR = scale radius, a = intercept, and b = slope. We substituted the means of the distances from the focus to each annulus for SR in the above equa- tion, calculated the mean fish length for the time of each annulus formation, and then calculated mean growth increment for each age group. Calculation of a theoretical growth curve is useful in modeling of growth in natural popula- tions of fish. Growth parameters such as theo- retical maximum attainable size (LJ, growth coefficient (K), and theoretical time of the begin- ning of growth (to), may be used in constructing population models. The most popular theoretical growth curve, the von Bertalanffy (l t = L^{\ — exp — K(t — to))) was fitted to back-calculated length at age data (Ricker 1975; Everhart et al. 1975). This particular equation also allows us to make comparisons with results obtained by other researchers. The growth parameter, L x , was first derived by fitting a Walford (1946) line: Z, +1 = L x (1 - k) + kl, to back-calculated data where h = total length at age t, and k = slope of the Walford line. The slope (A - ) is equal to e*, thus our first estimate of K = In k. Preliminary values of L x were ob- tained by solving the equation L x = ^-intercept/ (1 — k), and by regressing annual growth incre- ment (X) against fish length at the beginning of the incremental period ( Y) (Jones 1976). By plot- ting log e (L^ - l t ) against t and by using trial val- ues of L„ ranging from lower than the prelimi- 6 T. Azarovitz. Northeast Fisheries Center Woods Hole Lab- oratory. National Marine Fisheries Service, NOAA, Woods Hole, Mass., pers. commun. January 1978. 7 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. FISHERY BULLETIN: VOL. 80, NO. 1 nary values to much greater, we determined the best L x that resulted in the straightest line. The growth coefficient (K) was the slope of this line and was used to solve for t : to = y - intercept of natural log line - log P L 3 K We checked the U value to see if it was biased toward younger or older fish by using the equa- tion to = t(l/K) In (1 - h/LJ for separate ages I- IX (Jones 1976). Mortality Estimates We calculated annual total mortality estimates by analyzing catch curves (Beverton and Holt 1957) based on fully recruited age fish and older. If the log, of the age frequency in the catch is plotted on age, the slope of the linear descending right limb of the curve is equal to the mean in- stantaneous total mortality (Z). To calculate mortality rates, we first needed to assign ages to the 1,100 or so unaged fish. We grouped fish of known age by 25 mm length intervals, calculated the percentage of fish of each observed age in each group, and used these percentages to esti- mate the number of fish of each age for the un- aged group (Ricker 1975). Length- Weight and Fork Length- Total Length Relationships To calculate length-weight and length conver- sion relationships fish lengths were subsampled to provide a fairly equal distribution throughout the size range of fish examined during this study. The length-weight relationship was expressed exponentially, whereas the fork length-total length equation was expressed as a simple linear regression. Spawning Gonads were examined macroscopically by season to determine the approximate time of spawning. Observations on the development of testes were used collaboratively with measure- ments recorded from ovaries. Ovaries were weighed to calculate a seasonal gonad index, or the percentage of gonad weight to fish weight. RESULTS Distribution and Relative Abundance Tomtates were collected throughout the South Atlantic Bight (Figs. 1-4). Although most of the continental shelf is sandy "open-shelf habitat" (Struhsaker 1969), the greatest catches of tom- tates were directly associated with the irregu- larly distributed sponge-coral ("live bottom") habitats (as defined by Wenner et al. 1979a). In- dices of relative abundance over live-bottom areas were significantly larger (P<0.01) than abundance indices from sandy-bottom catches in all seasons and years, except during the cold winter of 1977 (Table 2). Although tomtates occurred in 30-70% of the collections from the sponge-coral habitat, 79.6% of the total number caught during seven cruises, excluding the cold winter of 1977, were at sponge-coral stations (Table 3). During all seasons, catches of tomtates over sand were infrequent, but occasionally large (Wenner et al. 1979a, b, c, d). Occurrence of tomtates in both sandy-bottom and live-bottom habitats increased the difficulty in biomass esti- mations. Information from catches over the sponge-coral habitat with the 30-min tows was expanded to preliminary estimates of biomass (Tables 4, 5), although the catch represented a mixed habitat collection of unknown propor- tions. Standing crop estimates of tomtates from the region between Cape Fear and Cape Canav- eral ranged from 1,730 t (minimum catch, sum- mer 1974) to 12,878 t (maximum catch, winter 1976). Although biomass estimates were calcu- lated separately for each depth zone and stand- ing crop estimates were calculated separately for catches from live-bottom and sandy-bottom habitats (Table 6), all estimates represent mini- mal values because fish availability and vulner- ability to the trawl were not considered. Tomtates, both juvenile (<137 mm TL) and adult, were more abundant in catches in the northern part of the South Atlantic Bight than in catches in the south. During all seasons sampled, between 1973 and 1977, the catch north of lat. 32°32'N, an arbitrary shelf division, was be- tween 59 and 89% of the total catch. The one ex- ception occurred during the cold winter of 1977, when 98% of the total catch (3,192 fish) was made south of lat. 32°30'N at a single station. MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE FIGURE 1. — Spatial distribution and catch per tow of tomtates between Cape Fear and Cape Canaveral, 1 April- 9 May 1974. FISHERY BULLETIN: VOL. 80, NO. 1 80^ 79* V \ 78* ^> 77* 76* 75* ■J? &* V Cape I Feor ll 3 25 .•'10 TOTAL NUMBER OF FISH SUMMER 1974 * None O I to 6 (3 6 to 51 3 51 to 101 9 101 to 501 • 501 to 10,000 34* 33" 75' 32* 31* 30* 76 29 28 27 77' 28* 81° 80* 27* 79* 78° FIGURE 2.— Spatial distribution and catch per tow of tomtates between Cape Fear and Cape Canaveral, 13 August- 19 September 1974. MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE 75* TOTAL NUMBER OF FISH FALL 1973 x None O I to 6 3 6 to 51 3 51 to 101 * 101 to 501 • 501 to 10,000 \ 34* 33' 73 32 \ 31* w 76* 29* 28* 27* 77* 78° FIGURE 3.— Spatial distribution and catch per tow of tomtates between Cape Fear and Cape Canaveral, 23 October-16 November 1973. FISHERY BULLETIN: VOL. 80, NO. 1 79* 78* 77* 76* 75' V TOTAL NUMBER OF FISH WINTER 1976 x None O I to 6 (5 6 to 51 3 51 to 101 • 101 to 501 • 501 to 10,000 34" 33* 75* 32* 31* 30* '76* 29* 28* 27 77' 28* 81* 80" 27* 79° 78° FIGURE 4.— Spatial distribution and catch per tow of tomtates between Cape Fear and Cape Canaveral, 12 January-7 February 1976. 8 MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE Table 2.— f-test and chi-square results of com- parisons between numbers of tomtates in catches over live-bottom and sandy-bottom habitats. Table 4.— Mean catch/tow (f/J values for trawl-caught tom- tates on untransformed and transformed [In (kg + 1)] data by depth and habitat zone for summer 1974. Bliss' (1967) estima- tion of the mean was applied to the transformed values. f-test of X 2 test y„ biomass yv biomass Area of No. I In (x + 1) «i Ix Depth (kg/tow) (kg/tow) zone of Seasons n df of — n df (m) Habitat untransformed transformed (km 2 ) tows Fall 1973 375" 65 14.79" 9-18 live 5.272 12.981 2.622 2 Spring 1974 4.70" 85 15.69" sand 0.324 0.163 15,461 14 Summer 1974 4 15" 66 24.73" 19-27 live 6.804 6804 2.730 1 Winter 1975 2.83" 68 93.36" sand 0.218 0101 16,100 18 Summer 1975 8 18" 66 69.46" 28-55 live 3991 5.196 3,794 5 Winter 1976 877" 67 350.41" sand 0.000 0.000 22,367 14 Summer 1976 11.04" 67 193.32" 56-110 live 1 285 1 120 692 6 Winter 1977 1.59n.s. 70 40.42" sand 000 0.000 4,083 8 *" = significant at 0.01 level; n.s. = nonsignificant at 0.05 level. Table 3.— Catches of tomtates associated with collec- tions within the "live bottom"-sponge/coral habitats. Live bottom stations Tomtate catch % with Total % from Cruise date N tomtates number live bottom Fall 1973 10 39 2,075 296 Spring 1974 11 42 442 55.7 Summer 1974 14 50 581 76.2 Winter 1975 1 9 30 1,212 78.4 Summer 1975 18 70 1,298 98.2 Winter 1976 11 53 4,005 976 Summer 1976 8 53 1,749 91.7 Winter 1977 2 11 55 3,260 1.5 Average 50 79.6 'Sampling season prolonged into spring. 2 Unusually cold winter, data omitted from average. Table 5.— Mean catch/tow (l2cm FALL 1973 14 l~l 10 8 L0-. r- < I.O—i _l UJ 05- >I3 cm _7_ 4 2| » . i" i. n i • • i • i £ WINTER 1976 f I <'2cm O - Depth(m)9-I8 19-27 28-55 56-110 111-183 184^366 I I i i i i X UJ 0.5- Q Z 1.0-1 I.O-i 0.5- >l3cm SPRING 1974 2 -Z. 2T 20 4 JL ± 28 18 0.5- 1.0- 21 rn r~ 1 ° ° l5cm Depth(m)9-I8 19-27 28-55 56 - lio' III- 183 ' 184-366 I I I I "D Figure 5.— Index of relative abundance for tomtates by depth zone during four seasons O'uveniles above the axis, adults be- low: fraction numerator = number of trawls with tomtates- denominator = total number of trawls in depth zone). higher temperature intervals (Fig. 6). Young tomtates (20-63 mm) have previously been col- lected during December in the Florida Keys at a water temperature of 16.2°C (Springer and Woodburn 1960). During summer (1975) juve- niles were collected only in the coolest thermal zone (24.0°-27.9°C), while during spring (1974) both juveniles and adults were collected in the same thermal interval (16.0°-23.9°C). ^°o mt ^ te t may avoid water temperatures of <10 L. b ish were never caught at <10 3°C dur- ing any season, even at five sponge-coral stations in areas where large numbers were caught at >10°C during the previous winter (Fig. 7). 10 SUMMER 1975 FALL 1973 24-27°C 28-3IOC I6-I9°C r- 5 10 15 20 25 5 WINTER 1976 < soon I2-I5°C X O 100 < U X CO o 20-23°C 24-27°C -r — i 1 1 r 5 10 15 20 25 £ SPRING 1974 I6-I9-C 20-23°C ~l ' i ' I ' i 5 10 15 20 25 FISH LENGTH(cm) Figure 6. -Length-frequency distributions (TL) of tomtates by bottom water temperature interval (4°C). Age and Growth Validity of Rings as Annuli Both scales and otoliths were used to age tom- tates. Approximately 75% (397 of 529) of the scale samples and 85% (177 of 208) of the otoliths were legible. Since tomtates have been aged by read- ing scales (Sokolova 1969), we did not try spe- cifically to validate the methods presented here. Several findings, however, pursuant to the goals of this paper, indicate that rings on tomtate scales and otoliths are true annuli. Close exami- nation of otoliths from young-of-year tomtates, collected by trawl, clearly show the formation of one ring per year, and that the first ring (annu- lus) forms between the fall and spring collection periods. The mean length of fish progressively in- creased as the number of scale or otolith rings increased and otoliths and scales agreed closely (Table 7). For instance, if aged by scales, age-I fish averaged 135.4 mm TL; age-II, 181.9; age- Ill, 203.3; age-IV, 220.0; age-V, 234.5; age- VI 255.7; and age- VII, 265.8. If aged by otoliths,' age-I fish averaged 134.3 mm TL; age-II, 164.7; MANOOCH and BARANS. DISTRIBUTION AND ABUNDANCE OF TOMTATE 76* 75° TOTAL NUMBER OF FISH WINTER 1977 « None O I to 6 <3 6 to 51 3 51 to 101 9 101 to 501 • 501 to 10,000 ^ 34* 33* 75 32 3I C 30° 76* 29* 28° 27 V 27° 79° 78° Figure 7.— Spatial distribution and catch per tow of tomtates between Cape Fear and Cape Canaveral during the cold winter of 1977. 11 FISHERY BULLETIN: VOL. 80, NO. 1 TABLE 7.— Comparison of mean empirical length-age data obtained by reading tomtate scales and otoliths. Scales Otoliths Mean Mean Difference Age TL Range in TL Range in in means group N 22 (mm) 84.7 length (mm) SD N (mm) length (mm) SD (mm) 50-142 28 1 54 89.8 50-142 29.9 5.1 1 9 1354 109-171 25.6 23 134.3 80-157 21.5 1.1 2 45 181.9 153-206 13.3 43 164.7 150-185 9.7 17.2 3 81 2030 180-221 9.6 16 197.1 161-212 14.7 5.9 4 134 220.0 195-238 10.5 19 213.0 193-227 11.1 7.0 5 66 234.5 208-257 11.3 12 232.2 226-242 4.1 2.3 6 28 255.7 245-268 6.3 9 253.1 240-262 6.5 2.6 7 5 265.8 260-272 5.5 1 267.0 — — 1.2 8 4 277.0 270-280 5.0 9 3 2867 282-289 4.0 Total 397 age-Ill, 197.1; age-IV, 213.0; age-V, 232.2; age- VI, 253.1; and age- VII, 267.0 mm. The relative length frequencies of the mea- sured distance from the focus of the scale to each ring progressively increased with the number of rings. Significant features of the plotted curves were the occurrence of one mode for each ring, the consistent location of a specific mode on the X-axis for fish of different ages, the increased overlap for each additional ring, and the pro- gressive decrease in the distance between modes for each successive year, indicating less linear growth each year as the fish ages. Growth There was relatively little difference in the mean annual increments of fish aged by scales and those of fish aged by otoliths (Table 7). An- nual growth increments for fish aged by scales for ages I-V were: HI, 46.5 mm; II-III, 21.1 mm; III-IV, 17.0 mm; and IV-V, 14.5 mm. After age V, growth appears to be more irregular, prob- ably a result of the relatively small sample sizes for ages VI, VII, VIII, and IX (Table 7). Lengths by age for fish from all years were back-calculated from a scale radius-fish length regression. The prediction equation was TL = 1.7489 SR 09512 ; r = 0.93 and AT = 103, where TL = total length, and SR = scale radius. By substituting the means of the distances from the focus to each annulus for SR in the above equation, we were able to calculate the mean fish length at the time of each annulus formation, and the mean annual growth increment for each age (Table 8). The von Bertalanffy equation was used to de- scribe theoretical growth. The growth param- eters L x and K were first calculated by fitting a Walford (1946) line to back-calculated data. The equation was Itn = 90.833 + 0.6747*,, r = 0.982. Our first estimate of K was L 0.6747 or 0.3935. This value was used to obtain L x by solving the equation L x = ^-intercept / (1 — k). The initial value for L x of 289, and the subsequent value of 285.7 obtained by regressing annual growth in- crement {X) against fish length at the beginning of the incremental period (Y) (Jones 1976), Table 8.— Calculated total lengths (millimeters) of 346 tomtates aged by scales. Ob« ierved ige N Mean cal culated total length at end of year ; 1 2 3 4 5 6 7 8 9 1 9 1037 II 45 108 1 173.0 III 75 102 5 171.1 199.1 IV 123 102.6 1684 198.8 214.1 V 56 101 1678 2007 2167 226.7 VI 26 102.2 165.5 198.8 221.1 235 7 245 1 VII 5 99.6 165.3 200.9 224.4 240.8 251 3 2588 VIII 4 105 3 171.7 195.5 215.1 2286 2420 252.1 2608 IX Total 3 346 1025 170.6 200.9 222 1 237.7 253.0 264 5 273.5 280.5 Wei ghted mean 103.0 169.3 1993 2160 230.4 2462 2580 2662 280.5 Increment 103.0 66.3 30.0 16.7 14.4 15.8 11.8 8.2 14.3 No. calculations 346 337 292 217 94 38 12 7 3 12 MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE seemed low. Therefore, we plotted log f (L^ — /,) against t by using trial values of L x ranging from 285 to 310 mm. The straightest line resulted from L x of 310 mm. The slope of the line, —0.22017, was selected as the growth coefficient {K) and was used to obtain U (—1.28). Our best estimate of the equation describing the theoreti- cal growth of tomtates is It = 310 (1 - exp - 0.22017(* + 1.28)). annual mortality estimate for 1972 through 1978 was 59% (Z = 0.887). By year, instantaneous mor- tality rates were 1974, 0.669; 1975, 1.035; 1976, 1.017; 1977, 1.041; and 1978, 0.972. Too few fish were sampled from the fishery in 1972 and 1973 to construct catch curves. The instantaneous mortality rate(s) for tomtates was higher than those previously obtained for white grunt, Z = 0.65 (Manooch 1976), or for red porgy, Z= 0.58 (Manooch and Huntsman 1977). Observed, back-calculated, and theoretical lengths at age are presented in Table 9. Table 9.— Total lengths of tomtates at age (ob- served, back-calculated, and theoretical). Age Length at age (mm) Observed Back-calculated Theoretical 1 135 4 103.0 1224 2 181.9 169.3 159.4 3 203.0 199.3 189.2 4 220.0 216.0 213.1 5 234.5 230.4 232.2 6 255.7 246.2 247.6 7 265.8 258.0 259.9 8 2770 266.2 269.8 9 286.7 280.5 277.8 Length- Weight and Fork Length- Total Length Relationships Fish ranging from 52 to 280 mm TL were used to calculate a length-weight relationship. The equation W = 0.0000086L 30905 , r = 0.996 and N = 70, where W = weight in grams and L = total length in millimeters, describes this relation- ship. The equation TL = -1.8196 + 1.1540 FL, r = 0.99 and TV = 100 was derived to convert lengths. Mortality Estimates By age IV, tomtates are fully recruited to the hook and line fishery, the only important method of harvesting this species off the southeastern United States. Instantaneous mortality (Z) esti- mates were obtained by analyzing catch curves of fish aged IV and older (Fig. 8). The mean total 7.0 6.0 > O 5.0 z Ul P. 4.0 a 3.0 o o 2.0 1.0 N= 1.496 Z=-0.887 r= -0.985 m E m 2n 2m ix AGE Figure 8. — Catch curve for tomtates caught by hooked line off the southeastern United States, 1974-77. Spawning Indirect evidence indicates that tomtates of the South Atlantic Bight spawn primarily in April and May. Running ripe males and partly spent females were caught in April 1979 (28-42 m; 16.4°-19.4°C), while a major decrease in mean ovarian weight and maximum ovary weight of mature females occurred after the spring (April 1974) sampling period (Table 10). Throughout the year, many (>38%) of the females sampled each season were in the maturing and ripe condi- tion. The presence of juveniles (33-90 mm TL; mode 80 mm) in bottom trawl collections during summer, and the progressive increase in modal fish lengths in length-frequency distributions Table 10.— Gonad condition of adult tomtates (>15 cm TL) from the South Atlantic Bight. Mean Maximum Runn nq ovarian wt. Gonad ovarian wt. Season Sex N ripe % (9) index' (g) Summer 1974 F 31 5 0.6 0.6 1.7 Fall 1973 F 48 2 0.5 0.5 1.4 Winter 1976 F 36 9 1.3 1.1 8.6 Spring 1979 2 M 13 77 — — — F 34 4.2 3.4 17.0 'Gonad index = (ovary wt./fish wt.) X 100. 2 Females 77% with hydrated eggs. 13 FISHERY BULLETIN: VOL. 80, NO. 1 through a seasonal cycle (Fig. 9), indicated that these juveniles were spawned in spring. 200-, SUMMER 1975 X o < LU X en UJ CD 800- 700- 600- 500- 400- 300- 200- 100- 200- 100- N=4005 — r WINTER 1976 "1 I T "1 — I — r X 1 1 1 — T" SPRING 1974 N = 442 i — i — i — i — i i — i — i — i — i — i — i — r 2 4 6 8 10 12 14 16 18 20 22 24 26 FISH LENGTH (cm) Figure 9.— Length-frequency distributions (FL) from the total catch of tomtates between Cape Fear and Cape Canaveral during each of four seasons. DISCUSSION Distribution and Abundance Tomtates are considered abundant in several habitats in and to the south of the South Atlantic Bight, and indicate daily movements between habitats. Within the Bight, tomtates were com- mon over both live-bottom and shelf-edge habi- tats during earlier (1959-64) exploratory fishing (Struhsaker 1969). Farther south, tomtates were found over broad sandy areas off southern Flor- ida (Craig 1976), near coral stacks in the Tor- tugas Islands (Longley and Hildebrand 1941), and in grass beds and other open areas in the Bahamas (Bohlke and Chaplin 1968). Tomtates were common from nearshore to the offshore reefs in Florida and were abundant on the shrimp grounds of the Dry Tortugas and the Gulf of Mexico (Courtenay 1961). In the Virgin Islands, changes in distribution with respect to habitat type were associated with feeding be- havior. Tomtates feed as individuals or in small schools at night over open sand (Collette and Tal- bot 1972), and they spend the day on the reef segregated into size groups; juveniles school over the highest part of the reef, while adults hover low between the coral colonies (Smith and Tyler 1972). Juvenile tomtates may occur in several habi- tats, either inshore or offshore, which include "live bottom" and rocky outcrops similar to those occupied by adults. Small tomtates (*«33 mm) were abundant over artificial reefs (Parker et al. 1979) and natural ridges in spring through fall off the Carolinas and have been found in the mouths and stomachs of black sea bass in the same areas (Parker 8 ). Young tomtates also fre- quent grass beds (Randall 1968), subtidal mud flats (Reid 1954), and nearshore areas around wharfs (Jordan and Evermann 1896). The pres- ence of young fish among spines of sea urchins (Johnson 1978) suggests that microhabitats may be important to the survival of some early life stages. Juveniles of French grunts, H. flavolin- eatum, and white grunts, H. plumieri, form large multispecies schools closely associated with par- ticular coral formations (microhabitats) during the day and follow precise routes (>100 m) to and from feeding areas (sea grass beds) at night (Ogden and Ehrlich 1977). Biomass and standing stock calculations for tomtate from groundfish trawling were consid- ered preliminary, minimal estimates. More satisfactory estimates should incorporate infor- mation on 1) abundance/biomass sampling con- ducted completely within a known area of a given habitat type, 2) the correct proportional allocation of a day/night catch factor for each habitat sampled, 3) the vulnerability of tomtate to the sampling gear, and 4) estimates of biomass from untrawlable, rocky outcrop, habitats. Un- fortunately, none of the above information is available at present, so our estimates were based upon continuous day/night sampling imposed on the very random nature of sponge-coral habitat distribution. Discrete, short duration trawling completely within the boundaries of the patchy sponge-coral habitats could be directed by pre- trawl bottom mapping with underwater TV 8 R. 0. Parker, Southeast Fisheries Center Beaufort Labora- tory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516, pers. commun. January 1978. 14 MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE (Powles and Barans 1980). This method would allow more accurate quantification of relative abundance differences between habitats and be- tween day and night sampling. Then, a biomass factor could be developed to proportion fish availability to the trawl during daytime collec- tions in sponge-coral habitats and during night in sand bottom habitats. Also, the vulnerability of tomtate, or any groundfish in the South Atlan- tic Bight, to trawl gear is unknown. Several experiments with a headrope mounted TV sys- tem would do much to fill this data gap. Biomass estimates from rocky habitats may have to be extrapolated from nearshore diver counts or off- shore TV counts, but many problems remain in the interpretation of these data. In general, a composite estimate of tomtate biomass or stand- ing stock in the South Atlantic Bight should include difficult to obtain fish behavior informa- tion. Tomtate are relatively shallow water (<50 m) groundfish with a more pronounced tendency for annual depth migrations in populations south of the South Atlantic Bight. Tomtates in the Cam- peche Bank area were most abundant in waters <30 m during all seasons (Sauskan and Olaechea 1974), while tomtates occurred only at depths of <10 m in the Bahamas (Bohlke and Chaplin 1968). Although tomtates remain inshore during winter in Florida (Courtenay 1961), they are not caught by inshore shrimp trawlers off South Carolina (Keiser 1976) and appear to avoid shal- low waters (<20 m) north of Florida during win- ter. There is the possibility that during ex- tremely cold winters, slight migrations (shifts in distribution) southward occur. In contrast to the results of this study, tomtates of the Campeche Bank move onshore during win- ter and fall and offshore in spring and summer and are recruited to the fishery in shallow waters, a great distance from the deeper area where spawning takes place (Sauskan and Olae- chea 1974). The difference in location of spawn- ing and recruitment and lack of large adult fish over reefs in Florida (Stone et al. 1979) and in commercial trawl catches (Sokolova 1969) sug- gests separation of juvenile and adult popula- tions, especially south of Florida. Age and Growth The fact that scales may be used to accurately determine the age of a warmwater marine fish species is not particularly surprising. Scales have been used to age other reef fishes that occur with tomtates in the South Atlantic Bight. Ma- nooch (1976) found annuli on scales from white grunt collected off the Carolinas; Manooch and Huntsman (1977) aged red porgy, Pagrus pag- rus, using both scales and otoliths; and Grimes (1978) determined the age of vermilion snapper, Rhomboplites aurorubens, by reading scales. The theoretical parameters derived in this study are compared with those for tomtates from the Campeche Banks, and with cooccurring spe- cies in the South Atlantic Bight in Table 11. The Campeche Banks fish did not live as long — 5 or 7 yr compared with 9 in the South Atlantic Bight — and had a slightly smaller maximum size (L x ), 295 mm compared with 310 mm. Con- sequently, the growth coefficient, although very similar, is slightly higher— 0.235 compared with 0.200. With the exception of black sea bass, Cen- tra pristis striata, sympatric species previously studied in the South Atlantic Bight were longer lived and slower growing (Table 11). Table 11.— Growth parameters for six species of demersal fish. Scientific L^ Longevity Common name name Area Author M K (TL. mm) (yr) Tomtate Haemulon aurolineatum N.C., S.C, Ga , east coast Florida Campeche Banks This paper Calculated from 0.22017 310 9 Sokolova (1969) 0235 295 5 Sauskan and Olaechea 7 (1974) White grunt H. plumieri N.C., S.C. Manooch (1976) 0.4 & 0.6 0.108 640 13 Red porgy Pagrus pagrus N.C., S.C. Manooch and Huntsman (1977) 0.2 0096 763 15 Vermilion Rhomboplites N.C., S.C. Grimes (1976) 0.25 198 627 10 snapper aurorubens Gag Mycteroperca microlepis N.C, S.C. Ga., east coast Florida Manooch and Haimovici (1978) 0.20 0.121 1,290 13 Black sea Centropristis N.C, S.C. Mercer 1 0.30 0.220 352 8 bass striata 'Linda Mercer, Virginia Institute of Marine Sciences, Gloucester Point, Va 15 FISHERY BULLETIN: VOL. 80, NO. 1 Spawning Growth rates of juvenile tomtates (>130 mm TL/first year) in the South Atlantic Bight and rates estimated from larvae of similar species support the spawning season indicated by analy- sis of gonads. If growth of very early stages of tomtates approximates the 14 mm SL/30 d for white grunts (Saksena and Richards 1975) and French grunts (Brothers and McFarland In press), tomtates of 30-90 mm TL caught in early September may have been spawned between early April and June. Identification of peak spawning period of tomtate by associated larval abundance was impossible due to difficulties in identifying larval haemulids. Populations of tomtates farther south appear to have a prolonged spawning season. Munro et al. (1973) reported collections of ripe females be- tween January and August in Jamaica, while Cervigon (1966) suggested that tomtates spawn throughout the year in Brazil. Tomtates from Campeche Bank spawn primarily during July- September at depths of >50 m and again during winter at shallower depths (Sauskan and Olae- chea 1974). Management We believe management of the tomtate fishery should be considered for three reasons. First, the species is easily captured by a variety of fishing techniques: hook and line, trap, and unlike most other reef fishes, by trawl. Second, fishing effort applied to this, and other associated, species will probably increase. And third, the tomtate is a member of a rather delicate faunal community and is a major source of food for higher trophic level, piscivorous fishes. Unwise harvest of one species could have both physical and energetic impacts on the community as a whole. While regional catches of tomtates may at times be quite large, for instance by recreational anglers on headboats fishing inshore waters, the species ranks low in terms of poundage landed in the South Atlantic Bight by both recreational and commercial fishermen. Because tomtates are small and not competitive in value with other reef fishes in the commercial market, commer- cial hook and line fishermen usually discard the species or use it as bait for larger predatory fishes, such as groupers and snappers (Wen- ner 9 ). Given the geographical range of H. aurolin- eatum, its abundance as indicated by exploratory trawling, and relatively low harvest by fisher- men, one could label it as an "underutilized spe- cies" in the South Atlantic Bight. However, assigning tomtates this status requires a thor- ough understanding of currently operating fish- eries plus a knowledge about the role of the species in the ecosystem. We do not recommend such a designation at this time. Although the distribution of tomtate is contin- uous from the southeastern coast of the United States to the Campeche Banks, the stock fished in each area should be considered separate for assessment and fishery management. Our study and the studies of Sokolova (1969) and Sauskan and Olaechea (1974) show that tomtates are a relatively short-lived, fast-growing reef fish with a high annual mortality rate when com- pared with other reef fishes of the region. Fish with these biological traits usually are not as readily overfished as those that grow more slowly, those that live longer, and those with a lower annual mortality rate. However, many fast-growing, high-mortality species such as mackerels (Scomberomorus), some tunas (Thun- nus), and menhaden (Brevoortia), which are im- portant to large fisheries, have demonstrated some signs of being overfished. By comparing our study with those on the Campeche Bank, we can look at one stock caught at present primarily by hook and line and the other by a more intensive gear, the trawl. The major difference between the South Atlantic Bight tomtates and those from Campeche Bank is that the Atlantic stock is older and larger. There are several explanations other than bio- logical changes in ecology or genetics for the dif- ferences between these stocks. In our study, the tomtates were caught by recreational fishermen using hook and line while those from Campeche Bank were trawled. Hook and line fishing may be more selective of larger fish and some of the smaller fish may be discarded by the fishermen resulting in larger fish of each age being sam- pled. A more likely explanation is that the large, old Atlantic fish have a generally low exploita- tion rate. The Soviet-Cuban trawlers have fished Campeche Bank since 1964 with catches of grunts averaging over 20,000 tons a year and ex- *C. A. Wenner, South Carolina Wildlife and Marine Re- sources Department, Marine Resources Research Institute, Charleston, SC 29412, pers. commun. January 1978. 16 MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE ceeding 60,000 in 1971 and 1975 for the region according to FAO Yearbooks. If, as we suspect, most of these catches were tomtate, the Cam- peche stock has been much more exploited than the Atlantic for the past 10 yr. Regional harvest of tomtates by hook and line will probably remain low. Recreational anglers will continue to catch small numbers, and com- mercial handliners will continue to regard H. aurolineatum as "trash fish" or bait. Any in- crease in the harvest will probably involve an expansion of a trawl fishery off South Carolina, Georgia, and northeast Florida. Prior to development of any U.S. groundfish trawl fishery for tomtate, the possibility of habi- tat destruction by trawl gear should be investi- gated. Some bottom trawl harvest techniques may have detrimental effects on the substrate community in which tomtate are most abundant. Destruction or removal of the sponge/coral in- vertebrates and crab species, or damage to Ocu- lina coral beds, may indirectly reduce future yields of tomtate and other fish species. Also, during several seasons trawls may catch juveniles of a species important to both commer- cial and sport fisheries prior to their recruitment to harvest by hook and line. Bottom trawling for tomtate in "live bottom" areas would catch large numbers of small, commercially unimportant fish species and invertebrates which would in- crease costs of sorting unless the entire catch was processed as a mixed species product. In our study the greatest relative abundance (catch/tow) of adults was during winter at which time commercial harvesting could take advan- tage of any concentrations of fish resulting from a shift to a more offshore distribution of the popu- lation. Reduction of fishing effort during late winter and early spring would allow the un- fished stock to spawn and juveniles to be re- cruited to the fishery at a larger size, possibly regulated by net mesh size. Even in this case, a drastic reduction in population size could ad- versely affect the recreational headboat fishery. ACKNOWLEDGMENTS We thank the Captain and crew of the RV Dol- phin for cooperation and patient efforts through- out the field work; Victor Burrell, Jr., Director of the South Carolina Marine Resources Research Institute for support and encouragement; Julian Mikell for preliminary summary, especially gonad/maturity data; Karen Swanson and Herb Gordy for graphic arts; Dean Ahrenholz, Gene Huntsman, and John Merriner for reviewing the manuscript; and Beverly Ashby and Beverly Harvey for typing. Parts of this research were sponsored by the National Marine Fisheries Service (MARMAP Program Office) under contract No. 6-35147. LITERATURE CITED Beverton, R. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish populations. Fish. Invest. M inist. Agric. Fish. Food (G.B.) Ser. II, 19, 533 p. Bliss, C. I. 1967. Statistics in biology. Statistical methods for re- search in the natural sciences. Vol. 1. McGraw-Hill, N.Y., 558 p. BOHLKE, J. E., AND C. C. G. CHAPLIN. 1968. Fishes of the Bahamas and adjacent tropical waters. Livingston Publ. Co., Wynnewood, Pa., 771 p. Brothers, E. B., and W. N. McFarland. In press. Correlations between otolith microstructure, growth, and life history transition in newly recruited French grunts [Haemulon flavolineatum (Demarest), Haemulidae]. Rapp. P.-V. Reun. Cons. Int. Explor. Mer. Cervigon M., F. 1966. Los peces marinos de Venezuela, Vol. II. Funda- cion La Salle de Ciencias Naturales, Caracas, p. 449-951. Clark, S. H., and B. E. Brown. 1977. Changes in biomass of finfishes and squids from the Gulf of Maine to Cape Hatteras, 1963-74, as deter- mined from research vessel survey data. Fish. Bull., U.S. 75:1-21. Cochran, W. G. 1977. Sampling techniques. 3ded. Wiley,N.Y.,428p. COLLETTE, B. B., AND F. H. TALBOT. 1972. Activity patterns of coral reef fishes with emphasis on nocturnal-diurnal changeover. In B. B. Collette and S. A. Earle (editors). Results of the Tektite program: Ecology of coral reef fishes, p. 98-124. Nat. Hist. Mus. Los Ang. Cty., Sci. Bull. 14. COURTENAY, W. R., JR. 1961. Western Atlantic fishes of the genus Haemulon (Pomadasyidae): Systematic status and juvenile pig- mentation. Bull. Mar. Sci. Gulf Caribb. 11:66-149. Craig, A. K. 1976. Trapping experiments with snappers in South Florida. In H. R. Bullis, Jr. and A. C. Jones (editors), Proceedings: Colloquium on snapper-grouper fishery resources of the western central Atlantic Ocean, p. 222- 236. Fla. Sea Grant Rep. 17, 333 p. Elliot, J. M. 1971. Some methods for the statistical analysis of sam- ples of benthic invertebrates. Freshwater Biol. Assoc. Sci. Publ. 25, 144 p. Everhart, W. H., A. W. Eipper, and W. D. Youngs. 1975. Principles of fishery science. Cornell Univ. Press, Ithaca, N.Y., 288 p. Grimes, C. B. 1976. Certain aspects of the life history of the vermilion 17 FISHERY BULLETIN: VOL. 80, NO. 1 snapper, Rhomboplites aurorubens (Cuvier), from North and South Carolina waters. Ph.D. Thesis, Univ. North Carolina, Chapel Hill, 240 p. 1978. Age, growth, and length-weight relationship of vermilion snapper, Rhomboplites aurorubens, from North Carolina and South Carolina waters. Trans. Am. Fish. Soc. 107:454-456. Grosslein, M. D. 1969. Groundfish survey program of BCF Woods Hole. Commer. Fish. Rev. 31(8-9):22-30. Johnson, G. D. 1978. Development of fishes of the mid-Atlantic Bight: An atlas of egg, larval and juvenile stages, vol. IV. Ca- rangidae through Ephippidae. U.S. Fish Wildl. Serv., Biol. Serv. Program FWS OBS-78/12, 314 p. Jones, R. 1976. Growth of fishes. In D. H. Cushing and J. J. Walsh (editors), The ecology of the seas, p. 251-279. W. B. Saunders Co., Phila. Jordan, D. S., and B. W. Evermann. 1896. The fishes of North and middle America: A de- scriptive catalogue of the species of fish-like vertebrates found in the waters of North America, north of the isth- mus of Panama. U.S. Natl. Mus. Bull. 47:1-3313. Keiser, R. K., Jr. 1976. Species composition, magnitude and utilization of the incidental catch of the South Carolina shrimp fish- ery. S.C. Mar. Resour. Cent. Tech. Rep. 16, 94 p. LONGLEY, W. H., AND S. F. HlLDEBRAND. 1941. Systematic catalogue of the fishes of Tortugas, Florida, with observations on color, habits, and local dis- tribution. Carnegie Inst. Wash. Publ. 535, Pap. Tor- tugas Lab. 34, 331 p. Manooch, C. S., III. 1976. Age, growth and mortality of the white grunt, Hae- mulon plumieri Lacepede (Pisces: Pomadasyidae), from North Carolina and South Carolina. Proc. Annu. Conf. Southeast. Assoc. Game Fish. Comm. 30:58-70. Manooch, C. S., Ill, and M. Haimovici. 1978. Age and growth of the gag, Mycteroperca microlep- sis, and size-age composition of the recreational catch off the southeastern United States. Trans. Am. Fish. Soc. 107:234-240. Manooch, C. S., Ill, and G. R. Huntsman. 1977. Age, growth and mortality of the red porgy, Pagrus pagrus. Trans. Am. Fish. Soc. 106:26-33. Munro, J. L., V. C. Gaut, R. Thompson, and P. H. Reeson. 1973. The spawning seasons of Caribbean reef fish. J. Fish Biol. 5:69-84. Musick, J. A., and J. D. McEachran. 1972. Autumn and winter occurrence of decapod crusta- ceans in the Chesapeake Bight, U.S.A. Crustaceana 22:190-200. Ogden, J. C, and P. R. Ehrlich. 1977. The behavior of heterotypic resting schools of juve- nile grunts (Pomadasyidae). Mar. Biol. (Berl.) 42273- 280. Parker, R. O., Jr., R. B. Stone, and C. C. Buchanan. 1979. Artificial reefs off Murrells Inlet, South Carolina. Mar. Fish. Rev. 41(9):12-24. Powles, H., and C. A. Barans. 1980. Groundfish monitoring in sponge-coral areas off the southeastern United States. Mar. Fish. Rev 42(5)- 21-35. 18 Randall, J. E. 1968. Caribbean reef fishes. T.F.H. Publications, Nep- tune City, N.J., 318 p. Reid, G. K., Jr. 1954. An ecological study of the Gulf of Mexico fishes, in the vicinity of Cedar Key, Florida. Bull. Mar. Sci. Gulf Caribb. 4:1-94. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Fish. Res. Board Can., Bull. 191, 382 p. Rohr, B. A., AND E. J. Gutherz. 1977. Biology of offshore hake, Merluccius albidus, in the Gulf of Mexico. Fish. Bull., U.S. 75:147-158. Saksena, V. P., and W. J. Richards. 1975. Description of eggs and larvae of laboratory- reared white grunt, Haemulon plumieri (Lacepede) (Pisces, Pomadasyidae). Bull. Mar. Sci. 25:523-536. Sauskan, V. I., and A. Olaechea. 1974. Biology and population abundance of the tomtate (Haemulon aurolineatum) (Cuvier and Valenciennes, 1830, Pomadasyidae, Perciformes) on Campeche Bank. [In Russ.] In A. S. Bogdanov (editor), Soviet-Cuban fishery research, Vol. 4, p. 135-510. Institute de Inves- tigacion Cientifica de la Economia Pesquera Marina y de la Oceanografia de toda la Union (VNIRO), Centro de Investigaciones Pesqueras (CIP), Mosc. Smith, C. L., and J. C. Tyler. 1972. Space resource sharing in a coral reef fish commu- nity. In B. B. Collette and S. A. Earle (editors), Results of the Tektite program: Ecology of coral reef fishes, p. 125-170. Natl. Hist. Mus. Los Ang. Cty., Sci. Bull. 14. Sokolova, L. V. 1969. Distribution and biological characteristics of the main commercial fish of Campeche Bank. In A. S. Bog- danov (editor), Soviet-Cuban fishery research, p. 208- 224. Isr. Program Sci. Transl., Jerusalem. Springer, V. G., and K. D. Woodburn. 1960. An ecological study of the fishes of the Tampa Bay area. Fla. Board Conserv. Mar. Lab. Prof. Pap. Ser. 1, 104 p. Stone, R. B., H. L. Pratt, R. O. Parker, Jr., and G. E. Davis. 1979. A comparison of fish populations on an artificial and natural reef in the Florida Keys. Mar. Fish. Rev. 41(9):1-11. Struhsaker, P. 1969. Demersal fish resources: Composition, distribu- tion, and commercial potential of the continental shelf stocks off southeastern United States. Fish. Ind. Res. 4:261-300. Taylor, C. C. 1953. Nature of variability in trawl catches. U.S. Fish Wildl. Serv., Fish. Bull. 54:145-166. Walford, L. A. 1946. A new graphic method for describing the growth of animals. Biol. Bull. (Woods Hole) 90:141-147. Wenner, C. A., C. A. Barans, B. W. Stender, and F. H. Berry. 1979a. Results of MARMAP otter trawl investigations in the South Atlantic Bight. I. Fall 1973. S.C. Mar. Re- sour. Cent. Tech. Rep. 33, 79 p. 1979b. Resultsof MARMAPotter trawl investigations in the South Atlantic Bight. II. Spring, 1974. S.C. Mar. Resour. Cent. Tech. Rep. 40, 78 p. MANOOCH and BARANS: DISTRIBUTION AND ABUNDANCE OF TOMTATE 1979c. Results of MARMAP otter trawl investigations in Wilk, S. J., AND M. J. Silverman. the South Atlantic Bight. III. Summer, 1974. S.C. 1976. Fish and hydrographic collections made by there- Mar. Resour. Cent. Tech. Rep. 41, 62 p. search vessels Dolphin and Delaware //during 1968-72 1979d. Resultsof MARMAPottertrawl investigations in from New York to Florida. U.S. Dep. Commer., the South Atlantic Bight. IV. Winter-early Spring, NOAA Tech. Rep. NMFS SSRF-697, 159 p. 1975. S.C. Mar. Resour. Cent. Tech. Rep. 44, 59 p. 19 GROWTH OF THE OCEAN QUAHOG, ARCTICA ISLANDICA, IN THE MIDDLE ATLANTIC BIGHT Steven A. Murawski, John W. Ropes, and Fredric M. Serchuk 1 ABSTRACT In situ growth rate of the ocean quahog, Arctica islandica, was investigated at a site 53 m deep off Long Island, New York, during 1970-80. Specimens notched during summer 1978 and recaptured 1 and 2 calendar years later yielded information on shell growth and the periodicity of supposed annual marks. Growth of specimens recaptured after 1 year at liberty (n = 67, 59-104 mm shell length) was described by SL,.\ =2.0811 +0.9802 SL,, where SL is shell length in millimeters at age t. Average shell length of marked specimens recaptured during summer 1980 increased 1.17 mm (w = 200), approximately twice that of ocean quahogs recaptured in 1979 (0.56 mm). Band for- mation on the external surface of small ocean quahogs (less than about 60 mm) was apparently an annual event since small specimens recaptured in 1979 formed one such mark during the interval between release and recapture. Small specimens sampled during summer exhibited relatively wide marginal growth from the last external mark to the shell edge, while winter samples had formed new annuli at the shell margin, thus, external bands were formed during early autumn- early winter. Internal banding in shell cross sections of small ocean quahogs correlated in number and position with external features. An equation representing back-calculated growth, based on external banding patterns of small unmarked specimens (19-60 mm) captured during summer 1978, was: SL = 75.68-81.31 (0.9056)', where tis age in years. Length-frequency samples were avail- able for the vicinity of the marking study from routine dredge surveys of clam resources during 1970-80. Growth rates inferred from progressions of length-frequency modes in 1970 and 1980 sam- ples were similar to those computed from mark-recapture and age-length equations. Ocean quahogs are apparently among the slowest growing and longest lived of the continental shelf pelecypods; annual increases in shell length were 6.3% at age 10, 0.5% at age 50, and 0.2% at an estimated age of 100 years. Research on the population dynamics of the ocean quahog, Arctica islandica, has become in- creasingly important in recent years. An inten- sive fishery for the species developed off New Jersey and the Delmarva Peninsula during the mid-1970's. The resulting increases in U.S. land- ings were dramatic: from 588 1 of shucked meats in 1975 to a record 15,748 1 by 1979. Estimates of the growth rate and longevity of ocean quahogs inhabiting the Middle Atlantic Bight are neces- sary to assess potential impacts of various har- vesting strategies on the resources (Murawski and Serchuk 2 ; Mid-Atlantic Fishery Manage- ment Council 3 ). ■Northeast Fisheries Center Woods Hole Laboratory, Na- tional Marine Fisheries Service, NOAA, Woods Hole, MA 02543. 2 Murawski,S. A., and F. M. Serchuk. 1979. Distribution, size composition, and relative abundance of ocean quahog, Arctica islandica, populations off the Middle Atlantic Coast of the United States. ICES/CM. 1979/K:26, Shellfish Comm., 22 p. 3 Mid-Atlantic Fishery Management Council. 1979. Amendment No. 2 for the surf clam and ocean quahog fishery Several early studies alluded to the age and growth rate of Arctica islandica, yet citations were largely anecdotal and generally did not re- flect critical evaluations of the rate of growth or the validity of aging criteria. Turner (1949) reported an observation by G. Thorson that "European investigators who have studied the chemical composition of the shell found reason to believe that it took six years or more for mahog- any (ocean) quahaugs (quahogs) to reach average size." Loosanoff (1953) stated that ocean quahogs he examined for reproductive studies "were adults, several years old, and averaged 3% to 4 inches (89-102 mm) in length." Jaeckel (1952) noted Cyprina (=Arctica islandica) could per- haps attain ages up to 20 "Sie kann hohes Alter (Vielleicht bis zu 20 Jahven) erreichen." Skula- dottir 4 did not elaborate on aging methodologies Manuscript accepted August 1981. FISHERY BULLETIN: VOL. 80. NO. 1. 1982. management plan and final supplemental environmental im- pact statement. Mid-Atlantic Fishery Management Council, Dover, Del., 114 p. 4 Skuladottir, U. 1967. Kraffadyr og skeldyr (Crustacean and mollusks). Radstefna Isl. Verkfraedinga. 52:13-23. 21 FISHERY BULLETIN: VOL. 80. NO. 1 but claimed "the oldest clams were up to 18 years and about 9 cm long. The bulk was in the 10-14 year group and 7-8.7 cm long." The external color of large ocean quahogs (greater than about 60 mm shell length) is usu- ally solid black; however, the periostracum of small individuals is variable in color, grading from pale yellow to deep brown (Loven 1929; Hiltz 5 ). Concentric dark bands appearing in the shell surface of small specimens have thus been interpreted as annuli by several authors. Although Loven did not present age-size rela- tionships explicitly, he did note the presence of external "annual rings" ("Jahresringe") and pre- sented photographs of a size range of small ocean quahogs, illustrating the relationship between numbers of rings and shell lengths. Chandler 6 measured the maximum diameters of concentric rings and derived growth relationships based on eight specimens (96 total measurements, to milli- meters). The largest number of such rings ap- pearing on an individual ocean quahog was 21; the corresponding shell length was 58.5 mm. Caddy et al. 7 presented growth curves, based on external markings, for small ocean quahogs from the Northumberland Strait and Passama- quoddy Bay. Average length at age was consis- tently greater for the more southern area. Unpublished manuscripts by Chene 8 and Mea- gher and Medcof 9 document efforts to more pre- cisely establish ocean quahog growth rates. Mark and recapture experiments were con- ducted in Brandy Cove, New Brunswick. Notched specimens (n = 14), averaging 57.4 mm (shell length) when recaptured, grew an average of 0.6 mm (shell height) between September 1970 (Proceedings of the conference of Islandic Professional Engi- neers. Fish. Res. Board Can., Biol. Stn., St. Andrews, N.B., Trans. Bur., No. 1206.) 5 Hiltz, L. M. 1977. The ocean clam (Arctica islandica). A literature review. Fish. Mar. Serv. Tech. Branch, Halifax N.S., Tech. Rep. 720, 177 p. 6 Chandler, R. A. 1965. Ocean quahaug resources of Southeastern Northumberland Strait. Fish. Res. Board. Can., Manuscr. Rep. (Biol.) 828, 9 p. 7 Caddy, J. F., R. A. Chandler, and D. G. Wildler. 1974. Biology and commercial potential of several underexploited molluscs and Crustacea on the Atlantic coast of Canada. Pre- sented at Federal-Provincial committee meeting on Utiliza- tion of Atlantic Resources, Montreal, Feb. 5-7 1974. Prepared at Fisheries Research Board of Canada, St. Andrews Biologi- cal Station, N.B. "Chene, P. L. 1970. Growth, PSP accumulation and other features of ocean clams (Arctica islandica). Fish. Res. Board Can., St. Andrews Biol. Stn., Orig. Manuscr. Rep. 1104, 34 p. s Meagher, J. J., and J. C. Medcof. 1972. Shell rings and growth rate of ocean clams (Arctica islandica). Fish. Res. Board Can., St. Andrews Biol. Stn., Orig. Manuscr. Rep. 1105, 26 p. and September 1971. Sequential observations of eight small ocean quahogs (mean length 20.16 mm) was undertaken to assess growth rates and seasonal changes in the color patterns of the peri- ostracum. These individuals were held in cages and grew an average 17% in length from 4 June to 31 August 1971. Periostracum formed during the interval was brown, contrasting with yellow material formed before the study was begun. However, this banding pattern may not have been indicative of a normally occurring annual event since "the caged clams were sensitive to experimental treatments and produced distur- bance rings each time they were air-exposed for observation" (Meagher and Medcof footnote 9). Several recent studies have examined banding patterns present in shell cross sections and have attempted to validate the hypothesis of band for- mation as an annual event. Jones (1980) noted that marginal increments of shell deposition be- yond the last band followed a seasonal progres- sion; bands were formed once per year between September and February. The most rapid pro- duction of shell was from late spring to early summer; annulus formation overlapped the spawning period in mature individuals. Thomp- son et al. (1980) presented size-frequency data of small specimens from the Baltic Sea and inter- preted external and cross-sectional banding in these specimens as supporting evidence for an- nual periodicity of band formation in larger (older) specimens from the Middle Atlantic Bight. Thompson et al. further stated that pre- liminary results from radiochemical analysis of shells corroborated age analysis based on shell banding patterns. We initiated a project during summer 1978 to assess in situ growth rates of ocean quahogs at a deepwater site off Long Island, N.Y. Objectives of the study were to obtain growth increment data directly from mark-recapture, further eval- uate the potential of banding patterns (both ex- ternal and in shell cross section) as indicators of age, and correlate growth measurements with a 10-yr time-series of length frequencies collected in the vicinity of the marking site. Length- weight relationships have been established for the Middle Atlantic, based on a synoptic winter survey (Murawski and Serchuk 1979); however, no data have been published on seasonal varia- tions. An additional objective of the project was to compare winter and summer length-weight relations at the marking site. 22 MURAWSKI ET AI..: GROWTH OF OCEAN QUAHOG. ARCT1CA ISLANDICA FIELD STUDIES Intermittent surveys of offshore clam re- sources of the Middle Atlantic Bight have been conducted since 1965 by the National Marine Fisheries Service, and its predecessor the Bu- reau of Commercial Fisheries (Merrill and Ropes 1969; Murawski and Serchuk footnote 2; Serchuk et al. 10 ). Cruises were designed to yield information on temporal and areal aspects of dis- tribution, size composition, and relative abun- dance of both surf clam, Spisula solidissima, and ocean quahog. Stations were sampled in a grid array prior to 1978; surveys from 1978 to 1980 employed a stratified-random scheme. Commer- cial-type hydraulic clam dredges were modified to retain small individuals and used as survey gear; dredge specifications and vessels varied somewhat among cruises (Serchuk et al. footnote 10; Table 1). We selected an area for intensive field study of ocean quahog growth, based on an evaluation of pre-1978 survey data and knowledge of commer- cial fleet activities. Specific criteria were: 1) suf- ficient clam densities for rapid capture of indi- viduals used in the marking experiment, 2) abundant numbers of clams over a wide size range, 3) clam densities similar to sites fre- quented by fishing vessels, and 4) lack of pre- vious exploitation and low probability of near- future use. These specifications were met at a site 48 km south-southeast of Shinnecock Inlet, Long Island, at lat. 40°25.1'N, long. 72°23.7'W. '"Serchuk, F. M.. S. A. Murawski, E. M. Henderson, and B.E.Brown. 1979. The population dynamics basis for man- agement of offshore surf clam populations in the Middle Atlan- tic. Proceedings of the Northeast Clam Industries - Manage- ment for the Future, Coop. Ext. Serv. Univ. Mass. -MIT Sea Grant, p. 83-101. Water depth was 53 m, and substrata consisted of coarse sand and shell, primarily ocean quahog and sea scallop, Placopecten magellanicus. Live invertebrates present in survey samples in- cluded Lunatia heros, Echinarachnius parma, Venericardia borealis, Aphrodite aculeata, and Astarte spp., in addition to ocean quahog and sea scallop. Water depth at the study site precluded ex- tended periods of bottom time using normal scuba methods, thus we elected to sample ocean quahogs with commercial and research dredging vessels. The probability of recapturing marked ocean quahogs at the site was considered to be relatively low because of water depth, width of sampling gear, difficulties in positioning the ves- sel at a precise location, and the accuracy of the loran-C navigation system. Hence it was decided to mark and redistribute large numbers. Incremental increases in clam shell growth corresponding to known time durations can be measured if a point of reference is initially estab- lished at the margin of the growing shell. Growth is determined directly from recaptured speci- mens and shell length at marking can either be measured or back-calculated. Thus we needed only to indelibly etch the shell edge of live qua- hogs and return them to the sea bed, obviating the laborious and time-consuming process of measuring and number-coding individuals prior to release. Notching techniques have been used success- fully to study growth rate and to validate the periodicity of band formation in a number of bi- valve species including soft shell clam, Mya arenaria (Mead and Barnes 1904); hard shell clam, Mercenaria mercenaria (Belding 1912); American oyster, Crassostrea virginica (Loosa- noff and Nomejko 1949); sea scallop (Stevenson Table 1. — Characteristics of survey gear and length-frequency statistics of ocean quahogs collected near lat. 40°25' N, long. 72°24' W, in the Middle Atlantic Bight, 1970-80. Dates Hydraulic dredge blade width (cm) Spacing between 1 bars or rings (mm) Shell ler igth (mm) Vessel X SD Range n RV Delaware II 13 August 1970 122 30 2 741 20.1 25-105 107 RV Delaware II 24 April 1976 122 30 74.1 16.6 40-115 271 RV Delaware II 27 February 1977 122 30 73.4 14.5 45-104 234 RV Delaware II 1 January-2 February 1978 122 30 74.5 14.3 34-113 211 FV Diane Maria 3 26 July-5 August 1978 254 13 74.5 15.4 31-112 1.262 RV Delaware II 9 January 1979 152 25 71.4 145 33-116 1,317 RV Delaware II* 14-21 August 1979 152 25-51 76.5 15.2 38-111 811 RV Delaware II 8 February 1980 152 51 74.2 13.8 38-117 5.546 RV Delaware II* 9 September 1980 152 51 74.8 13.4 40-108 1.899 'Dimension in the portion of the dredge where catch is accumulated 2 Samples measured to the nearest 5 cm. initiation of marking study 4 Recapture of marked individuals. 23 FISHERY BULLETIN: VOL. 80, NO. 1 and Dickie 1954; Merrill et al. 1966); and surf clam (Ropes and Merrill 1970; Jones etal. 1978). Accordingly, we marked ocean quahogs by cut- ting shallow grooves from the ventral margin up the shell surface using thin carborundum discs mounted on an electric grinder (Ropes and Mer- rill 1970). Two parallel grooves 2 mm apart were cut into each shell to distinguish our marks from shells scratched by natural processes or during dredging (Fig. 1). Marking operations were conducted from 26 July to 5 August 1978 (Table 1). A total of 41,816 ocean quahogs was notched by the previously de- scribed technique. Batches of 3,000-5,000 clams were dredged from within 9 km of the planting site, marked, and redistributed. The method of marking and planting clams was rapid; about 1,600 clams were marked per hour. A grid sys- tem based on loran-C coordinates, was used to indicate the location of each batch. Length-fre- quency samples were obtained during the mark- ing phase (Table 1), and 134 small ocean quahogs (19-60 mm) were retained for maturity studies and analyses of exterior and cross-sectional banding. An intensive effort to recapture marked indi- viduals was undertaken, 1 calendar year after planting, during 14-21 August 1979 (Table 1). Forty-three hydraulic dredge tows, each of about 5-min duration, were completed at the site. A Northstar 6000 11 loran-C set and an Epsco loran- C plotter were used in the systematic search of a 20,000 m 2 area. A total of 14,043 ocean quahogs was examined; 74 (0.5%) had been marked. Re- captured specimens were photographed, mea- sured, and frozen intact at sea. A random sample of 126 unmarked ocean quahogs was frozen for length-weight comparison with marked indi- viduals. Marked individuals were again recaptured, approximately 2 yr after planting, on 9 Septem- ber 1980 (Table 1). Two dredge tows yielded 1,899 ocean quahogs; 249 individuals (13.1%) had been marked. Length-frequency measurements were ob- tained from the site during routine assessment surveys in January 1979 and February 1980. Sampling within 10 km of the site was historic- ally serendipitous; catch data were available from four surveys between 1970 and February 1978 (Table 1). Lengths of ocean quahogs taken near the site exhibited a consistent bimodal fre- quency distribution throughout the time-series. Growth rate information from the mark-recap- ture and shell banding experiments was thus compared with that generated from modal pro- gression in sequential length frequencies. A random sample of 278 ocean quahogs taken from the site during February 1980 was frozen whole for length-weight comparison with the August 1979 sample. Small ocean quahogs (<60 mm) were also frozen intact for analysis of the timing of periodic band formation in the shells. LABORATORY STUDIES M ark- Recapture Recaptured specimens were thawed but kept moist during all phases of analysis to prevent shell cracking and disintegration of the perio- stracum. A total of 67 of the 74 specimens recap- tured in 1979 and 200 of 249 specimens recap- tured in 1980 were suitable for growth analysis; the remaining samples were either shell frag- ments or from quahogs obviously dead when re- covered. Shells were measured to the nearest 0.01 mm, using calipers or dissecting microscope equipped with an ocular micrometer. Perio- stracum obscured the shell edge of most speci- mens and was subsequently removed from the vicinity of the mark prior to measurement. Shell lengths were obtained by pressing the perio- stracum against the valves with calipers. Growth increments of recaptured ocean qua- hogs were determined as the linear increase in shell dimension along an imaginary line passing through the umbo and equidistant between grooves that formed the mark (Fig. 1). The linear distance between the umbo and shell edge at the mark was designed as h'\ shell length at marking was computed for each quahog by: Oi-/f — o/v(*l [ {h'm - h',) ] (1) "Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. where SL, = shell length (longest linear dimen- sion) at marking, SLm = shell length at recapture, h\= linear measurement between umbo and edge of the shell equi- distance between grooves, at marking, h'tn = linear measurement between umbo and edge of the shell 24 MURAWSKI ET AL.: GROWTH OF OCEAN QUAHOG, ARCTICA ISLANDICA Figure 1.— Ocean quahog shells used for growth analyses taken near lat. 40° 25'N. long. 72°24'W, in the Middle At- lantic Bight, (a) Specimen 65 mm. shell length, marked during July-August 1978 and recaptured during August 1979. Arrow indicates external growth band formed during the interval be- tween marking and recapture, (b) Ar- row indicates shell growth of a 68 mm specimen from July-August 1978 to Au- gust 1979 with periostracum removed, (c) Arrows indicate positions of most re- cently formed external growth bands on small individuals from August 1979 (right, 43 mm) and February 1980 (left, 45 mm) samples. 25 FISHERY BULLETIN: VOL. 80, NO. 1 equidistant between grooves, at recapture. Marginal growth in shell length was thus equiv- alent to the bracketed term. Implicit in Equation (1) is the assumption that ratios between the linear parameters SL and h! did not change between marking and recapture (isometric growth). The assumption is supported by comparisons of various standard shell dimen- sions (i.e., shell length, height, and width, Chan- dler footnote 6; Northeast Fisheries Center Woods Hole Laboratory unpubl. data), particu- larly considering the relatively small percent changes in shell size between marking and re- capture (Table 2). 1 40 - • • _T • + • ..♦ LP s 1 05 - 2 . • < •* ^^2 • N = 67 2 Ld UJ or o 70 • • • • .. • • Y = 2.0811 -0.0198x r = - 774 1 I (- o or o 35 • • • • • • • • • • I 60 70 80 90 100 SHELL LENGTH ( MM ) L Table 2.— Growth of ocean quahogs marked during August 1978, and recaptured during August 1979 (n = 67), and Septem- ber 1980 (n = 200), at lat. 40°25' N, long. 72°24' W, in the Middle Atlantic Bight. Parameter Year Mean (mm) SD (mm) Range (mm) Shell length at 1979 77.31 14.67 59.12-104 40 recovery 1980 7901 13.91 57.69-103.66 Calculated growth 1979 0.56 0.38 0.08-1.38 increment in shell 1980 1.17 1.04 0.07-4.32 length Calculated shell 1979 7676 14.97 58.15-104.09 length at marking 1980 77.84 14.75 55.46-103.43 Three methods were used to fit growth equa- tions to mark-recapture data. For ocean quahogs recovered 1 calendar year after marking, length at recapture was related to length at marking using Ford-Walford and linear annual increment plots described by Gulland (1969; Fig. 2). Ad- ditionally, a nonlinear exponential equation was fit to increment data and results compared with those assuming the von Bertalanffy model. The von Bertalanffy parameters L^ and K were also estimated using the BGC4 computer program ( Abramson 1971 ). The program was designed for determining growth parameters when lengths of unaged individuals are known at two points in time, based on the algorithm of Fabens (1965). Equations derived from mark-recapture data can be used to describe relative growth from an arbitrary point in time (i.e., SL M , SL t , 2 , ... SL,.„), but without at least one independently derived age-length observation, absolute growth curves cannot be established. Accordingly, anal- yses of external banding patterns of small ocean quahogs were critical in "fixing" growth curves from mark-recapture. Figure 2.— Relation between calculated increment of growth in shell length (millimeters) and initial length for ocean qua- hogs marked during July- August 1978 and recaptured during August 1979 near lat. 40°25'N, long. 72°24'W, in the Middle Atlantic Bight. Shell Banding Small ocean quahogs retained from the July- August 1978 cruise were analyzed for external and internal shell banding patterns. Sequential growth of individual ocean quahogs was followed by measuring the maximum dimension (shell length) of exterior bands appearing on the perio- stracum, using calipers (Fig. 1). Maximum shell length beyond the last band was also recorded. The opposite valve was sectioned from the umbo to the ventral margin and polished (Saloman and Taylor 1969; Jones et al. 1978). An acetate im- pression of the polished surface was made and mounted between glass slides. Images were en- larged with a microprojector to reveal internal banding patterns. Internal lines present in shell cross sections correlated in number and position with external bands when the latter were distinct. The perio- stracum on some shells was eroded near the umbo, obscuring external bands. In these cases "annuli" nearest the umbo were located on the peels, but measurements of shell size could not be made (Table 3). External marks present near the shell margins on some larger specimens also could not be discerned; internal banding was again used to estimate age. Shell length statistics were computed for each age/annulus subclass, weighted lengths at annuli for all ages and 26 MURAWSKI ET AL.: GROWTH OF OCEAN QUAHOG, ARCTICA ISLANDICA Table 3.— Back-calculated growth (shell length, in millimeters) of small ocean quahogs. Samples taken from lat. 40°25' N, long. 72°24' W, 26-29 July 1978, in the Middle Atlantic Bight. Number of annuli Length at capture Length at annulus 1 2 3 4 5 6 7 8 9 10 11 12 13 2 x SD n 18.00 0.00 1 700 000 1 12.30 000 1 3 x SD n 23.36 3.42 9 4.59 0.78 9 1059 266 9 1801 3.14 9 4 x SD n 2973 200 14 4.39 0.73 14 1004 2.13 14 1699 238 14 2438 1 96 14 5 x SD n 34.58 3.19 26 4.43 0.07 26 8.80 1.50 26 14.45 2.29 26 21.72 3.08 26 29.72 3.41 26 6 x SD n 3849 2.73 27 4.07 059 '25 7.77 1.57 27 13.40 249 27 19.13 258 27 26.09 2.73 27 33.88 2.92 27 7 x SD n 41 66 200 29 4.16 1.10 '27 7.66 1.34 29 12.10 1.72 29 17.42 1.57 29 2387 1.87 29 3081 1 98 29 37.61 2.05 29 8 x SD n 46.24 1.78 10 392 0.98 10 7.59 1.44 10 1229 2.39 10 16.92 2.77 10 23 64 2.38 10 29 95 2.52 10 36.63 2.22 10 42.76 1.99 10 9 x SD n 47.60 000 1 3.10 0.00 1 7.50 000 1 11.00 0.00 1 15 90 000 1 21.30 0.00 1 27.40 000 1 33.50 000 1 39.20 0.00 1 44 90 0.00 1 10 x SD n 4823 0.59 3 3.67 0.29 3 6.47 0.50 3 11.77 1.19 3 15.97 2.48 3 20.80 2.31 3 25.57 235 3 31.17 1.89 3 36.90 2.07 3 40 40 0.36 3 45.30 0.30 3 11 X SD n 54 35 205 2 3.90 0.00 '1 5.70 0.42 2 935 0.78 2 13.80 0.28 2 20.30 368 2 27.60 4.81 2 34.20 283 2 40 20 1.41 2 4445 1.06 2 48.50 0.71 2 51.95 1.20 2 12 x SD n 53.87 395 3 3.73 0.35 3 7.23 1.38 3 10.07 2.30 3 12.97 3.28 3 19.13 4.15 3 27.00 9.37 3 31.60 8.56 3 3567 7.90 3 3950 842 3 43.50 8.23 3 44.75 1.91 2 2 49.55 2.90 2 13 x SD n 53 90 000 1 1 5.20 000 1 9.70 0.00 1 12.80 0.00 1 17.50 0.00 1 22.20 000 1 2800 0.00 1 34.70 0.00 1 38 30 000 1 4370 0.00 1 46 40 0.00 1 50 00 0.00 1 52.00 0.00 1 14 2 x SD n 51.15 5.16 2 3.85 0.50 2 7.30 2.26 2 10.65 2.19 2 15.30 0.42 2 22 40 0.57 2 29.10 1.56 2 33.75 1.34 2 38.75 0.07 2 43.40 1.98 2 48.10 0.00 1 16 2 x SD n 57.93 290 4 4.00 0.00 '2 695 1.11 4 12.05 2.24 4 1850 249 4 24.80 3.95 4 31.53 3.75 4 37.25 2.91 4 4260 260 4 46.57 1.59 3 50.30 1.84 2 55.30 0.00 1 18 2 x SD n 57.10 0.99 2 360 0.00 '1 7.55 2.05 2 10.95 3.89 2 1640 5.80 2 2460 5.37 2 29.85 4.46 2 40.10 0.00 1 43 40 000 1 46 80 000 1 49.00 0.00 1 ALLx SD n Min Max 38.94 8.65 134 18.7 60.4 4.21 0.85 125 2.5 70 8.27 1.95 134 5.1 15.8 13.59 3.03 133 7.8 22.5 19.17 369 124 9.3 267 25.44 3.95 110 14.5 364 31.13 3.75 83 18.6 38.1 36.28 3.47 56 24.5 41.9 40.40 4.01 27 29.3 462 42.82 4.41 16 32.4 488 46.52 4.32 13 36.0 52.3 4918 4.58 6 43.4 55.3 4970 2.07 3 47.5 51.6 52.00 0.00 1 52.0 520 'External mark eroded but mark present in shell cross section "Number of annuli exceeds the number of lengths at annulus because marks could be distinguished in shell cross sections that were too closely spaced to discern on shell surfaces. lengths at capture were also determined (Table 3). Specimens recaptured in 1979 ranged in shell length from 59 to 104 mm, most had a deep brown or black periostracum. Several specimens did, however, exhibit the characteristic external banding pattern (Fig. 1), and were useful in vali- dating the presumed annual periodicity of marks. Marginal shell growth beyond the last exter- nal mark was strikingly different among small ocean quahogs from August 1979 and February 1980 samples. Mean lengths at capture for indi- vidual age classes from summer 1978 (particu- larly ages 1-9) were substantially greater than lengths at the last annulus, and were nearly equivalent to mean lengths at the last annulus for the next age class (Table 3). Ocean quahogs from winter 1980 invariably had formed or were forming an annulus at the shell margin (Fig. 1). A similar pattern was noted in shell cross sections. 27 FISHERY BULLETIN: VOL. 80. NO. 1 Modified exponential and logistic growth equations were fitted to mean back-calculated lengths at age, from the July 1978 samples (Table 3), using the asymptotic regression and nonlinear least squares computer programs BMD06R and~ BMD07R, respectively (Dixon 1977; Fig.. 3). Few aged shells were as large as those recap- tured (Tables 2, 3). Growth functions generated from aging data were thus extrapolated to the size range of recaptured specimens and results compared with annual growth increments pre- dicted from mark-recapture (Figs. 2, 3). An age- size point necessary to initiate the mark-recap- ture growth function was computed from growth equations fitted to age-length data generated in shell banding experiments; the mark-recapture equation was then iterated to encompass most shell lengths present at the marking site (Figs. 4,5). SL„,= 20811 +09802 SL 60 50 40 £ 30 ui X C/5 20 10 SL = 7568-81.31 (0.9056) AGE o OBSERVED ■• PREDICTED 6 8 10 12 AGE (YEARS) 14 16 18 Figure 3.— Observed and predicted shell lengths at age for small ocean quahogs sampled during July 1978 near lat. 40°25' N. long. 72°24'W. in the Middle Atlantic Bight. Length-Weight Shell length-drained meat weight relation- ships were computed for samples taken during August 1979 and February 1980. Laboratory 28 SHELL LENGTH MEAT WEIGHT 10 20 30 40 50 60 AGE I YEARS) 80 90 100 Figure 4.— Predicted shell lengths (millimeters) and drained meat weights (grams) at age for ocean quahogs at lat. 40°25' N, long. 72°24'W, in the Middle Atlantic Bight. Growth in length is described by an equation derived from studies of external banding patterns of small individuals (left of dot), and the Ford-Walford equation from mark-recapture data (right of dot). Weights at age are derived by applying the overall length- weight equation presented in Table 5 to calculated mean lengths at age. and statistical methods are given in Murawski and Serchuk (1979). Equations for recaptured and unmarked specimens from August 1979 were compared by covariance analysis to assess effects of marking (Table 4). Presumably, if physiological processes of the animal were sig- nificantly disrupted by the marking procedures, the adjusted mean of the length-weight equation might be statistically lower than that of controls. Seasonal variability in length-weight was in- vestigated by comparing summer and winter equations (Table 5). RESULTS AND DISCUSSION New shell growth of recaptured individuals was clearly discernible in small specimens (<70 mm) not only at the mark, but all along the Table 4.— Ocean quahog shell length-meat weight regression equations, and analysis of covariance for marked and un- marked individuals sampled at lat. 40°25' N, long. 72°24' W, in the Middle Atlantic Bight, during August 1979. Linear regression parameters Sample Intercept (a) Slope (6) r n Marked Unmarked -9.8373 2.9530 -9.0170 2.7637 0.975 55 953 126 Test of adjusted mean Test of slope Sample Adjusted mean dl F df F Marked Unmarked fi709 2 8714 1 ' 178 ° 001 nS 1,177 2.13 n.s. n.s. = P>0.05. MURAWSKI ET AL.: GROWTH OF OCEAN QUAHOG, ARCTICA ISLANDICA Table 5.— Ocean quahog shell length-meat weight regression equations, and analysis of covariance for August 1979 and Feb- ruary 1980 samples taken near lat. 40°25' N, long. 72°24' W, in the Middle Atlantic Bight. Linear regression parameters Sample Intercept (a) Slope (b) r n August 1979 February 1980 All data -9.2901 2.8274 -8.6865 2.7086 -9.0627 2.7871 0.961 181 0.976 278 0.967 459 Test of adjusted mean Test of slope Sample Adjusted mean df F df F February 1980 August 1979 l™l 1,456 58.86" 1,455 3.22 n.s. 20 40 60 80 100 120 SHELL LENGTH ( MM ) "P<0.01; n.s. = P>0.05. ventral margin when the periostracum was re- moved (Fig. 1). A growth interruption was pro- duced at the previous shell edge of small speci- mens; new material was formed slightly below the earlier shell margin and was shinglelike in appearance (Fig. 1). Growth in larger ocean quahogs was less distinct and thus more diffi- cult to measure. Where clear growth interrup- tions were not present, a faint yellowish band contrasting with white shell material was inter- preted as a marking-induced check and growth was measured from that point. Shell growth was assessed midway between grooves that formed the mark since, in the case of larger specimens, the depth of the grooves was actually greater than the amount of new shell deposited (Figs. 1, 2). A total of 11,658 ocean quahogs was measured directly from dredge catches at the marking site during 1970-80 (Table 1; Figs. 5, 6). Although minimum spacing of bars or rings in the rear portion of dredges varied somewhat (Table 1), size selectivity was apparently not significantly altered. Repeated tows were made at the mark- ing site during August 1979 with 25 X 25 mm and later 51 X 51 mm wire mesh in the after por- tion of the dredge. Size distributions of ocean quahogs were nearly identical before and after the alteration. A possible explanation for the lack of differential selectivity is that shell, sand, and live invertebrates may have clogged the dredge at the beginning of tows, negating fur- ther filtering ability. Two discrete length-frequency modes were ex- hibited in all sets of samples (Figs. 5, 6). Few small ocean quahogs (<50 mm) were encoun- FlGURE 5. — Length-frequency distributions (1 mm intervals) of ocean quahogs sampled near lat. 40°25'N, long. 72°24'W, in the Middle Atlantic Bight, April 1976-February 1980. 29 FISHERY BULLETIN: VOL. 80. NO. 1 60 80 SHELL LENGTH (MM) Figure 6.— Length-frequency distributions (5 mm intervals) of ocean quahogs sampled near lat. 40°25'N, long. 72°24'W, in the Middle Atlantic Bight, August 1970 and February 1980. tered from 1976 to 1980 (Fig. 5) and, considering uniformity of modes over time, recruitment was probably equally poor during 1971-76. Thus, cor- responding modes in the 1970 and 1980 samples were probably composed of the same year classes (Fig. 6). Average size of the small mode in- creased about 13 mm during the 9%-yr interval between August 1970 and February 1980, while the large group shifted about 3 mm (Figs. 5, 6; Table 1). Size progression of modes was minimal during 1976-80; intersample variation may be primarily related to differential sample sizes (Table 1). The effects of a sevenfold increase in sampling intensity can be seen by comparing August 1979 and February 1980 frequencies. Modes are smoothed in the latter sample, yet re- spective peaks are at precisely the same 1 mm in- tervals in both (65 and 90 mm). Average shell sizes ranged from 71 to 77 mm; however, trends in shell length among samples were not apparent (Table 1). The average lengths of recaptured ocean qua- hogs (Table 2) were slightly greater than con- current length-frequency samples (Table 1), although length extremes of the marked indi- viduals were not as great. Recaptured ocean qua- hogs also exhibited the bimodal length-frequency distribution (Fig. 2), indicating recaptured specimens represented a relatively unbiased sample of marked individuals and the ocean quahog population in the immediate vicinity of the study area. Calculated increments of shell growth from ocean quahogs recaptured in 1979 ranged from 0.08 to 1.38 mm, and averaged 0.56 mm (Table 2). Those recaptured in 1980 30 grew an average of 1.17 mm (range 0.07-4.32 mm). Thus, incremental growth approximately doubled between summer 1979 and summer 1980, implying growth rates were similar dur- ing the 2 yr of the experiment and that marking procedures probably did not significantly dis- rupt growth patterns. Growth increments of ocean quahogs at liberty 1 yr generally declined with increasing shell length, although there was substantial variation about a linear fit (Fig. 2). The linear equation for predicting annual incre- ment of growth from initial length is given in Figure 2; the Ford-Walford equation is: SL m = 2.0811 + 0.9802 SL t , where SL is shell length (in millimeters) at age t. An exponential equation fitted to data in Figure 2(7= 14.1216 (exp (— 0.0459X))) explained about 8% more of the re- sidual variance about the predicted line than did the linear equation. However, growth rates im- plied from length-frequency analyses were sub- stantially greater than those from the exponen- tial fit, and were similar to rates computed from the linear (von Bertalanffy) model. Thus, the latter model was considered more valid. Esti- mates of the asymptotic length (LJ and growth coefficient (K) from two fitting methods are: BGCU Annual increment L^ (mm) K 107.06 0.0195 104.95 0.0200 Values of L^ from the two methods are >99.5% (BGC4) and 98.5% (annual increment) of the cumulative 1980 length-frequency distribution at the study site. Estimates of K are relatively low and characteristic of slow-growing, long- lived species (Beverton and Holt 1959). Analyses of shell banding features present in small specimens indicate both external and in- ternal marks are produced once during the bio- logical year in these sizes. Several of the small recaptured ocean quahogs exhibited concentric external rings, and these specimens formed one such band during the interval between marking and recapture (Fig. la). Studies of small un- marked individuals retained from summer and winter sampling demonstrate that external and internal marks generally correspond in number and position. Internal marks were particularly useful in age determination when external marks were eroded near the umbo or closely spaced at the shell margin. Small ocean quahogs captured during the summer exhibited wide MURAWSKI KT AL.: GROWTH OF OCEAN QUAHOG. ARCTICA ISLANDICA marginal increments of shell growth from the last external and internal marks to the shell edge, whereas winter samples had recently formed annuli (Fig. lc; Table 3). Thus, mark for- mation probably occurs during the last half of the calendar year. These observations are consis- tent with data presented by Jones (1980). In a study of the seasonality of incremental shell growth, he noted that internal growth bands in shell cross sections were formed from September to February. The formation of growth bands apparently overlaps the spawning period (Jones 1980); however, both events may be related to other physiological or environmental stimuli since specimens that were reproductively imma- ture formed bands concurrently with mature ocean quahogs. Back-calculated mean lengths at age varied considerably depending on the subset of data analyzed in Table 3. Mean lengths at age for all year classes (bottom rows in Table 3) were gener- ally smaller than mean lengths at the last com- plete annulus (rightmost diagonal vector), and growth of recent age groups (2-8) appeared more rapid than for older ocean quahogs (Lee's phe- nomenon; see Ricker 1969). However, conclu- sions regarding the growth of older age groups (9-18) are tenuous due to the relatively small numbers of these ages sampled (87% of the sam- ples were <8-yr-old). Age analyses were limited to ocean quahogs that exhibited suitable contrast on the shell sur- face to discern external concentric rings. Thus, the oldest aged ocean quahogs (particularly ages 14-18) may represent the smallest, slowest grow- ing individuals of their year classes; faster grow- ing individuals may have reached sizes asso- ciated with color changes of the periostracum. Nevertheless, back-calculated mean lengths at age for 14- to 18-yr-old ocean quahogs did not tend to be progressively smaller than means for ages 9-13, perhaps indicating that size selectivity of older individuals was not a significant bias (Table 3). The objectives of fitting statistical models to age-length data were to describe growth during the juvenile and early adult phases of life, and more importantly, to predict ages associated with the lengths of the smallest recaptured speci- mens (59-65 mm) thereby linking the age-length data and mark-recapture results into a contin- uous growth function. Recognizing the disparate nature of data subsets in Table 3, a series of ex- ponential and logistic growth equations were fitted to: 1) weighted mean back-calculated lengths at age for all quahogs, 2) weighted mean lengths at age for ages 2-8, and 3) mean lengths at the last completed annuli (rightmost diagonal vector) for ages 2-10 and 2-13. For our purposes, the applicability of a particular model fit was judged not only by the total amount of variance between length and age explained by the equa- tion, but by predicted annual growth increments in the 59-65 mm range. An appropriate model would fit as much of the age-sample data as pos- sible and yield calculated annual growth incre- ments consistent with those observed from re- captured specimens. Exponential equations utilizing weighted mean back-calculated lengths for ages 2-8, and lengths at the last complete annulus for ages 2-13 yielded unacceptable fits by our criteria. The former equation was calculated with informa- tion from the linear portion of the growth curve, predicted lengths beyond age 8 were unrealistic- ally high. The latter equation incorporated one negative growth increment (between ages 11 and 12) and thus the calculated asymptote was only 62.8 mm; predicted annual growth near the asymptote was considerably less than observed increments for that size (Fig. 2). The logistic growth equation fitted to weighted mean lengths at age for all ocean quahogs (SL = 52.09/1 + exp(2.4722 - 0.4702(0)) was superior to the respective exponential fit considering the residual sums of squares criterion. The reverse was true for the logistic equation describing mean lengths at the last annulus for ages 2-10 (SL = 43.12/1 + exp(2.9361 - 0.8069 (*))). How- ever, asymptotic lengths were, for both logistic equations, well below the range of shell lengths considered in the mark-recapture experiments. Thus, extrapolation of logistic age-length re- lationships, necessary for initializing the Ford- Walford equation, was not feasible. On the contrary, the two exponential equations yielded reasonable asymptotic lengths and adequately described ocean quahog growth relative to that inferred from modal progressions in 1970 and 1980 length-frequency distributions (Fig. 6) and observed growth increments (Fig. 2). Exponential growth equations computed from weighted mean lengths at age for all ocean qua- hogs and mean lengths at the last annulus for ages 2-10 were: SL = 75.68-81.31 (0.9056)' and SL = 72.70-75.22 (0.8935)', respectively. Mean lengths at age predicted from the two equations generally reflect differences among data sets 31 FISHERY BULLETIN: VOL. 80, NO. 1 over the range of shell sizes used to fit the func- tions: however, estimated lengths at age con- verge near the sizes of the smallest recaptured specimens. Estimated lengths at age 20 were 64.49 and 64.29 mm, respectively. Correspond- ing growth increments from age 20-21 were 1.06 and 0.84 mm, well within the range of ob- served growth for those sizes (Fig. 2). If calcu- lated lengths at age 20 are assumed to be the starting points for the Ford-Walford equation (SL,n = 2.0811 + 0.9802 SL t ), the two acceptable exponential equations yield virtually identical growth curves when the Ford-Walford relation- ship is iterated. Additional growth analyses were conducted using the regression equation fitted to weighted mean back-calculated lengths for all ages because the maximum amount of in- formation was used and the equation's behavior in the vicinity of marking data was consistent with empirical observations. However, further research on the growth patterns of small ocean quahogs is indicated in order to resolve differ- ences between various data subsets in Table 3 and thus to define a more appropriate growth model for these sizes. A composite growth curve incorporating the aged samples and mark-recapture data is given in Figure 4. The Ford-Walford equation was iterated to age 100 and a predicted shell length of 96.91 mm. Although ocean quahogs reach a size of at least 117 mm in the vicinity of the marking site (Table 1), ages substantially in excess of 100 are not necessarily implied because of the statis- tical variability in the marking data used to fit the predictor (Fig. 2). Annual growth in shell length is rapid during the first 20 yr of life, but declines significantly thereafter. Average yearly shell growth is 6.3% at age 10, 0.5% at age 50, and 0.2% at age 100. Estimates of the von Bertalanffy parameter to (age at zero length) were computed as -27.29 yr and -27.62 yr for the BGC4 and annual incre- ment equations respectively, with SL20 = 64.49 mm (Gulland 1969, equation 3.5). Although pre- dicted lengths at ages >20 are similar to those in Figure 4, a relatively poor fit to younger ages re- sults from both von Bertalanffy equations. The validity of using the age-length functions given in Figure 4 to describe ocean quahog growth at the marking site can be assessed by comparing predicted growth to that from modal progressions in length-frequency samples. Fre- quency distributions from 1976 to 1980 exhibit inter-sample variability in the position of major modes but no progressive shifts are discernible (Fig. 5). However, expected growth during the 5- yr period (Fig. 4) was smaller than could prob- ably be identified, given the precision of length- frequency sampling (Table 1; Fig. 5). Length modes can be used to compute growth at the site between August 1970 and February 1980 (Fig. 6). Average growth of the smaller mode (52 mm in 1970) was about 13 mm, and the larger mode (87 mm in 1970) added about 3 mm shell length during the 9Y 2 -yr interval (Figs. 5, 6). Ocean qua- hogs 52 mm in length are about 12-yr-old and average 21-yr-old at 65 mm; the estimated age of 87 mm individuals is 60 yr and 90 mm quahogs average 70-yr-old (Figs. 3, 4). Thus, predicted growth during the period 1970-80 is strikingly similar to that inferred from length mode pro- gressions, implying that age analyses and mark- recapture data adequately describe historical ocean quahog growth at the site. The age-length relationships presented herein have been computed for shell sizes in excess of 95 mm and ages up to 100 yr. However, computed relationships for large sizes (>65 mm) are based on average growth rates from mark-recapture results and not from aging of individual speci- mens. It is likely, based on these analyses, that ocean quahogs do reach 100 yr in age; however, direct age determination of large individuals is contingent upon development and validation of suitable methodologies. Internal banding pat- terns present in shell cross sections were useful in aging small specimens since formation of the bands apparently occurs once annually. Seasonal shell formation patterns (Jones 1980) and age analyses of large individuals based on internal banding (Thompson et al. 1980; Jones 1980) are generally consistent with our data. Analysis of shell cross sections of large recaptured speci- mens may be useful in determining the periodi- city of internal banding and the validity of the aging technique for large ocean quahogs; study of this material continues. The regressions of shell length vs. drained meat weight for marked and unmarked ocean quahogs taken during August 1979 were not sig- nificantly different in slope or adjusted mean (Table 4). If in fact soft-tissue robustness is a valid index of relative condition, then marked in- dividuals apparently suffered no lasting effects from the stress of dredging and handling. This observation is supported by the conclusions that incremental shell growth of marked specimens was similar to that computed from progressive 32 MURAWSKI ET AL.: GROWTH OF OCEAN QUAHOG, ARCTICA ISLANUICA length frequencies of the population as a whole, and growth rates of marked individuals were nearly equal between 1978-79 and 1979-80. Length-weight equations from February 1980 and August 1979 were parallel (Table 5); winter samples were apparently heavier in drained meat weight at a given shell length than summer samples. However, the magnitude of predicted differences in weight at length was small (4-11% for 65-115 mm ocean quahogs). Differences may be related to weight changes associated with sexual development, or merely a statistical artifact. Samples from winter and summer were combined to predict average weight for a given length during the year (Table 5). The resulting length-weight equation was applied to computed lengths at age to derive an age-weight relationship (Fig. 4). Initial weight gains are proportionally greater than concomitant length increases, but growth rates are nearly identical at the oldest predicted ages. Average annual increases in drained meat weight are 18.1% at age 10, 1.6% at age 50, and 0.2% at age 100 (Fig. 4). Growth rates determined from the examina- tion of concentric external banding patterns in- dicate small ocean quahogs may grow faster off Long Island than in the Northumberland Strait and in Passamaquoddy Bay (Caddy et al. foot- note 7). However, data are insufficient to con- clude that a latitudinal cline in ocean quahog growth exists. Factors influencing growth rates in a particular area are speculative; however, density dependence must be considered. Muraw- ski and Serchuk (footnote 2) noted relative popu- lation stability and poor recruitment for ocean quahogs in the Middle Atlantic during 1965-77. Stable population size, poor recruitment, and slow growth are characteristic of populations under density dependent regulation. Investiga- tion of ocean quahog growth rates at various den- sities may help to elucidate their interrelation- ship and indicate the population consequences of cropping high density areas. ACKNOWLEDGMENTS In particular we thank ships' personnel and scientific parties aboard the various research vessels during field sampling phases of the proj- ect. Significant technical contributions were made by Lt. Comdr. Ron Smolowitz, NOAA Corps, and Dea Freid of the Northeast Fisheries Center. The manuscript was critically reviewed by Brad Brown, Mike Sissenwine, Emma Hen- derson, Mike Fogarty, and Rich Lutz. LITERATURE CITED Abramson, N. J. (compiler). 1971. Computer programs for fish stock assessment. FAO Fish. Tech. Pap. 101, 149 p. Belding, D. L. 1912. A report upon the quahaug and oyster fisheries of Massachusetts. Commonw. Mass., Boston, 134 p. Beverton, R. J. H., and S. J. Holt. 1959. A review of the lifespans and mortality rates of fish in nature, and their relation to growth and other physio- logical characteristics. Ciba Found. Colloq. Ageing 5: 142-180. Dixon, W. J. (editor). 1977. BMD Biomedical computer programs. Univ. Calif. Press, Berkeley, 773 p. Fabens, A. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. GULLAND, J. A. 1969. Manual of methods for fish stock assessment. Part I. Fish population analysis. FAO Man. Fish. Sci. 4, 154 P- Jaeckel, S., jun. 1952. Zur oekologie der Molluskenfauna in der West- lichen Ostee. Schr. Naturwiss. Ver. Schleswig-Hol- stein 26:18-50. Jones, D. S. 1980. Annual cycle of shell growth increment formation in two continental shelf bivalves and its paleoecologic significance. Paleobiology 6:331-340. Jones, D. S., I. Thompson, and W. Ambrose. 1978. Age and growth rate determinations for the Atlan- tic surf clam, Spisula solidissima (B\va.\v\a.: Mactracea), based on internal growth lines in shell cross-sections. Mar. Biol. (Berl.) 47:63-70. Loosanoff, V. L. 1953. Reproductive cycle in Cyprina islandica. Biol. Bull. (Woods Hole) 104:146-155. Loosanoff, V. L., and C. A. Nomejko. 1949. Growth of oysters, O. virginica, during different months. Biol. Bull. (Woods Hole) 97:82-94. Loven, P. M. 1929. Bietrage zur Kenntnis der Cyprina islandica L. im Oresund. K. Fysiogr. Sallsk. Lund Handl. N.F. 41: 1-38. (Contribution to the knowledge of Cyprina islandi- ca L. in the Oresund, Transl. Lang. Serv. Div., Off. Int. Fish., U.S. Dep. Commer.. NMFS, Wash.. D.C.) Mead, A. D., and E. W. Barnes. 1904. Observations on the soft-shell clams. Thirty- fourth Annu. Rep. Comm. Inland Fish., R.I., p. 26-28. Merrill, A. S., J. A. Posgay, and F. E. Nichy. 1966. Annual marks on shell and ligament of sea scallop, Placopecten magellanicus. U.S. Fish Wildl. Serv., Fish. Bull. 65:299-311. Merrill, A. S., and J. W. Ropes. 1969. The general distribution of the surf clam and ocean quahog. Proc. Natl. Shellfish. Assoc. 59:40-45. Murawski, S. A., and F. M. Serchuk. 1979. Shell length— meat weight relationships of ocean 33 FISHERY BULLETIN: VOL. 80, NO. 1 quahogs, Arctica islandica, from the Middle Atlantic Shelf. Proc. Natl. Shellfish. Assoc. 69:40-46. RlCKER, W. E. 1969. Effects of size-selective mortality and sampling bias on estimates of growth, mortality, production, and yield. J. Fish Res. Board Can. 26:479-541. Ropes, J. W., and A. S. Merrill. 1970. Marking surf clams. Proc. Natl. Shellfish. Assoc. 60:99-106. Saloman, C. H., and J. L. Taylor. 1969. Age and growth of large southern quahogs from a Florida Estuary. Proc. Natl. Shellfish. Assoc. 59:46- 51. Stevenson, J. A., and L. M. Dickie. 1954. Annual growth rings and rate of growth of giant scallop, Placopecten magellanicus (Gmelin) in the Digby area of the Bay of Fundy. J. Fish. Res. Board Can. 11:660-671. Thompson, I., D. S. Jones, and D. Dreibelbis. 1980. Annual internal growth banding and life history of the ocean quahog, Arctica islandica (Mollusca: Bival- via). Mar. Biol. (Berl.) 57:25-34. Turner, H. J., Jr. 1949. The mahogany quahaug resources of Massachu- setts. In Report on investigations of improving the shellfish resources of Massachusetts, p. 12-16. Com- monw. Mass., Dep. Conserv., Div. Mar. Fish. 34 LARVAL DEVELOPMENT OF CITHARICHTHYS CORNUTUS, C. GYMNORHINUS, C SPILOPTERUS, AND ETROPUS CROSSOTUS (BOTHIDAE), WITH NOTES ON LARVAL OCCURRENCE 1 2 John W. Tucker, Jr. 3 ABSTRACT Developmental series of 4 of the 12 species of Citharickthys and Etropus known from the western North Atlantic and Gulf of Mexico are illustrated and described. The series consist of C. cornutus (preflexion to nearly transformed, 2.2-17.4 mm body length. BL), C. gymnorhinus (preflexion to late transformation, 4.4-12.9 mm BL), C. spilopterus (preflexion to juvenile, 3.7-25.4 mm BL), and E. crossotus (preflexion to nearly transformed, 4.6-10.8 mm BL). Data from this study and that for 2 species previously described permit identification of larvae of 6 of the 12 species. For the species investigated, caudal fin formula (4-5-4-4) is the most reliable indi- cator for the group of genera Citharichthys, Cyclopsetta, Etropus, and Syacium. Number of elongate dorsal rays, degree of cephalic spination, and pigmentation are most useful for determining genus for known forms. Number of elongate dorsal rays, number of caudal vertebrae, pigmentation, mor- phology, and number of gill rakers are most useful for identification of Citharichthys and Etropus larvae that have been described. Citharichthys cornutus larvae have no pectoral melanophore, little notochordal pigmentation, heavy lateral pigmentation, 3 elongate dorsal rays, and develop 6 left pelvic rays and 25-26 caudal vertebrae. Flexion is complete at 9-10 mm SL and transformation at about 18 mm SL. Larvae have been collected during all seasons. Caudal fin development in C. cornutus is typical of the four species described here. Citharichthys gymnorhinus larvae have no pectoral melanophore, little notochordal pigmentation, light lateral pigmentation except for a caudal band, 3 elongate dorsal rays, and de- velop only 5 left pelvic rays and 23-24 caudal vertebrae. Flexion is complete at 7-8 mm SL and trans- formation probably at about 18 mm SL. Larvae have been collected during all seasons. Citharichthys spilopterus larvae have no pectoral melanophore, little notochordal pigmentation, light lateral pig- mentation, a blunt snout, a deep body, 2 elongate dorsal rays, and develop 6 left pelvic rays and 23-24 (rarely 25) caudal vertebrae. Flexion is complete at 7-8 mm SL and transformation at 9-11 mm SL. Larvae have been collected from September through April. Etropus crossotus larvae have a melano- phore at the base of the pectoral fin, heavy notochordal pigmentation, heavy lateral pigmentation, 2 elongate dorsal rays, and develop 6 left pelvic rays and 25-26 (very rarely 24) caudal vertebrae. Flexion is complete at 9-10 mm SL and transformation at 10-12 mm SL. Larvae have been collected in May and August and probably occur from March to August. Twelve species of the flatfish genera Citharich- thys and Etropus (subfamily Paralichthyinae, family Bothidae) are recognized from the west- ern North Atlantic (Table 1). Because of their small size at maturity, these fishes are presently used only by the petfood and fish meal industries (Topp and Hoff 1972). However, the abundance of larvae (Richardson and Joseph 1973; Smith et "Contribution No. 1037, Virginia Institute of Marine Sci- ence, Gloucester Point, VA 23062. 2 Derived from a thesis submitted to North Carolina State University in partial fulfillment of the requirements for the Master of Science degree. 3 School of Marine Science of the College of William and Mary. Virginia Institute of Marine Science, Gloucester Point, VA 23062. Manuscript accepted October 1981. FISHERY BULLETIN: VOL. 80. NO. 1. 1982. al. 1975; Dowd 1978) and adults (Dawson 1969; Topp and Hoff 1972; Christmas and Waller 1973) indicates that some species may represent sig- nificant components of estuarine and marine food webs. Larvae in the Citharichthys-Etropus complex are difficult to distinguish and are often ignored or classified as "unidentified bothids" in species composition analyses (e.g., Fahay 1975). Of the 12 western North Atlantic species, only C. arcti- frcms and E. microstomas have been described in detail (Richardson and Joseph 1973). Citharich- thys cornutus, C. gymnorhinus, C. macrops, and E. rimosus have been briefly described by Dowd (1978). Larvae of the remaining species have not been reported previously. Hsiao (1940) mis- 35 FISHERY BULLETIN: VOL. 80, NO. 1 takenly described Bothus sp. larvae as E. cros- sotus. In this paper I present descriptions of larvae of C. cornutus, C. gymnorhinus, C. spilopterus, and E. crossotus and summarize data useful for iden- tifying Citharichthys and Etropus larvae. MATERIALS AND METHODS Abbreviations The following institutional abbreviations are used: CP&L = Carolina Power and Light Com- pany, Raleigh, N.C.; GCRL = Gulf Coast Re- search Laboratory, Ocean Springs, Miss.; GMBL = Grice Marine Biological Laboratory, College of Charleston, S.C.; LSU = Louisiana State Univer- sity, Baton Rouge; NCSU = North Carolina State University, Raleigh; NMFS = National Marine Fisheries Service, NOAA (four laboratories — Beaufort, Galveston, Panama City, and La Jolla); OSU = Oregon State University, Corvallis; RSMAS = Rosenstiel School of Marine and Atmospheric Science, University of Miami, Fla.; SCMRRI = South Carolina Marine Resources Research Institute, Charleston; Texas A&M = Texas A&M University, College Station; UNC = University of North Carolina, Institute of Ma- rine Sciences, Morehead City; USNM = U.S. National Museum of Natural History, Smith- sonian Institution, Washington, D.C.; VIMS = Virginia Institute of Marine Science, Gloucester Point. Specimens Larval and juvenile specimens used in this study were obtained from several sources. Forty- seven C cornutus specimens from SCMRRI (MARMAP ichthyoplankton survey) collections in the South Atlantic Bight and five specimens from RSMAS collections from the Gulf of Mexico off western Florida were used for morpho- metries, counts, and general development. Seven additional RSMAS specimens were used for counts. Other specimens from NMFS (Beaufort) collections in Onslow Bay, off North Carolina, were used for comparison. Twenty-eight C gym- norhinus specimens from SCMRRI collections and 12 from RSMAS collections were used for morphometries, counts, and general develop- ment. Other specimens from NMFS (Beaufort) collections were used for comparison. Fifty-five C. spilopterus specimens from NCSU and per- 36 sonal collections in the Cape Fear River estuary, one from a CP&L collection in the ocean just off Cape Fear, and three from Texas A&M collec- tions in the Gulf of Mexico off Texas were used for morphometries, counts, and general develop- ment. Other specimens from Texas A&M, NMFS (Beaufort, Galveston, and Panama City), and RSMAS collections were used for compari- son and additional count data. Thirty E. cros- sotus specimens from LSU collections from the Gulf of Mexico off Louisiana and one from a NCSU collection were used for morphometries, counts, and general development. Other speci- mens from Texas A&M collections were used for comparison. Comparative larval material of other species was also examined. Citharichthys sp. A (prob- ably C. abbotti) specimens came from Texas A&M; Citharichthys arctifrons specimens from NMFS (Beaufort), SCMRRI, and VIMS; a Cith- arichthys sp. B (probably C. dinoceros) specimen from RSMAS; and Citharichthys (macrops'!) specimens from GCRL, RSMAS, and VIMS. Larvae of the eastern Pacific species Citharich- thys sordidus, C stigmaeus, and C xanthostigma came from NMFS (La Jolla). Other specimens of Pacific Citharichthys spp. came from OSU; Etropus microstomus specimens from NMFS (Beaufort) and VIMS; Etropus sp. A (probably E. rimosus) specimens from CP&L, NMFS (Panama City), and RSMAS; Cyclopsetta fim- briata specimens from NMFS (Beaufort), RSMAS, SCMRRI, and Texas A&M; and Syaci- um papillosum specimens from RSMAS and Texas A&M. Juvenile and adult specimens were examined to determine permanent characters. Specimens of C. arctifrons, C. macrops, C. spilopterus, E. crossotus, E. intermedius (cf. E. crossotus), E. microstomus, and E. rimosus came from USNM; Citharichthys cornutus and C. gymnorhinus specimens from GMBL; Citharichthys macrops specimens from UNC and a personal collection; and Citharichthys spilopterus and E. crossotus specimens from NCSU. Description of caudal skeleton development was based on study of the entire developmental series of C. cornutus and comparison with the series of the three other species described. Calcified components of the caudal skeletons of nearly all the specimens could be seen following light staining with Alizarin Red S in 1% aqueous potassium hydroxide solution. Twenty cleared and stained (Taylor 1967) specimens were exam- TUCKER: LARVAL DEVELOPMENT OF CITHAR1CHTHYS AND ETROPUS ined: C. arctifrons, (2) 40, 117 mm SL; C. comu- tus, (1) 51.5 mm SL; C. spilopterus, (2)41.6, -100 mm SL; C. macrops, (2) 45.7, ~100 mm SL; E. crossotus, (1) 49.4 mm SL; E. microstomus, (12) ~30-100 mm SL. Radiographs of juveniles and adults also were studied: C. arctifrons, (1) 100 mm SL; C. comutus, (16) 30-67 mm SL; C. gym- norhinus, (3) 23-37 mm SL; C. macrops, (75) 47- 113 mm SL; C. spilopterus, (65) 23-109 mm SL; E. crossotus, (62) 29-92 mm SL; E. intermedius (cf. E. crossotus), (2) 80, 92 mm SL; E. micro- stomus, (1)66 mm Sh;E. rimosus, (1)104 mmSL. Counts All larvae were lightly stained with Alizarin Red S in 1% aqueous potassium hydroxide solu- tion for making counts and observing the sequence of ossification. Most specimens were fairly transparent and internal structures were visible without clearing. The following counts were taken from larvae and juveniles with a stereomicroscope: precaudal neural spines, caudal neural spines, hemal spines, precaudal centra, caudal centra (including urostyle), caudal fin rays supported by each hypural ele- ment, dorsal fin rays, anal fin rays, left and right pelvic fin rays, left and right preopercular spines, left and right frontal-sphenotic spines, and left and right upper (premaxillary) and lower (dentary) larval teeth. Morphometries Measurements of various body parts of repre- sentative specimens were made on the left side with an ocular micrometer in a stereomicro- scope. The only exceptions were standard and total lengths of the six longest C. spilopterus (19.4-25.4 mm SL), which were made with di- viders and a millimeter scale. Measurements are defined as follows: Body length (BL) = snout tip to notochord tip for preflexion and flexion larvae (notochord length, NL); snout tip to posterior margin of hypurals for postflexion larvae and juveniles (SL). Upper jaw length = snout tip to posterior mar- gin of maxillary. Lower jaw length = anterior tip of dentary to posterior margin of articular just above the angular. Snout length = horizontal distance from snout tip to anterior margin of left pigmented eye. Eye diameter = horizontal diameter of left pig- mented eye. Head length (HL) = horizontal distance from snout tip to anterior margin of cleithrum at the body midline. Snout to anus length = horizontal distance from snout tip through midline of body to vertical line through anus. Total length = snout tip to posterior margin of finfold prior to caudal fin ray development, then to posterior tip of longest caudal ray. Head depth = greatest vertical depth of head; in preflexion larvae, this is near or just behind the posterior half of the eye, but with develop- ment the greatest depth is progressively more posterior. Body depth at pelvic fin = vertical distance from dorsal to ventral body margin at base of second pelvic ray. Body depth at loop of gut = vertical distance from dorsal to ventral body margin at the deepest par t of the gut ( C. comutus and C. gym- norhinus only). Body depth at anus = vertical distance from dor- sal to ventral body margin at anus. Body depth at third hemal spine = vertical dis- tance from dorsal to ventral body margin at third hemal spine. Caudal peduncle depth = prior to dorsal and anal fin formation, the vertical distance from dorsal to ventral body margin at the shallowest part of the caudal peduncle; after dorsal and anal fin formation, at the posterior edge of dor- sal and anal fins. Developmental Terminology Body length is a useful basis for linking char- acters of unidentified specimens with those in larval descriptions. However, body length may not be the most appropriate basis for comparing larvae of different species, especially bothids, which undergo notochord flexion and transfor- mation at different sizes, usually within a nar- row range for a single species but over a wide range for the family or even within a genus (e.g., Citharichthys). In this paper, both body length and stage of development are indicated for devel- opmental events. Stage of development is de- fined by degree of notochord flexion or degree of transformation. Terminology is similar to that of 37 FISHERY BULLETIN: VOL. 80, NO. 1 Moser et al. (1977) and Sumida et al. (1979), with slight modification because of the peculiarities of bothid development. Preflexion stage = notochord is straight. Early caudal formation = a substage of preflex- ion in which the notochord is still straight, but the caudal fin has begun to form. Flexion stage = notochord is turning upward. There are three substages: Early flexion = notochord is slightly flexed; midflexion = noto- chord is S-shaped and flexed about 30°-60°; late flexion = notochord is turned up and is no longer S-shaped but is not yet in final position. Postflexion stage = notochord is in final position, but transformation is not complete. Transforming larvae = those in which dorsal mi- gration of the right eye can be detected with low magnification. The period of transforma- tion is divided into thirds, depending on the position of the right eye. Juveniles = those specimens in which the right eye has reached its final position on the left side of the head and in which all fin rays have formed. Reported size ranges at transforma- tion are based on available specimens and might not encompass the full possible size ranges. Environmental stimuli inducing transformation may be encountered at differ- ent sizes. Terminology of components of the caudal skel- eton follows Amaoka (1969), except as noted. The caudal fin formula was described by Gutherz (1971) as the number of caudal rays supported by each caudal element, dorsal to ventral. Gutherz (1971) described certain cranial spines of Cyclopsetta fimbriata larvae as origi- nating from the sphenotic bones. Futch and Hoff (1971) described similar spines of Syacium papillosum larvae as originating from the fron- tal bones. In the Citharichthys and Etropus larvae I have examined, similar spines are at the suture between frontal and sphenotic bones. The origin of these could not be determined with cer- tainty, and therefore they are called "frontal- sphenotic" spines. For the larvae described here, the first elon- gate dorsal ray is actually the second ray of that fin. Larval Identification Four developmental series were assembled, primarily on the basis of similar meristics, shape, and pigmentation. Transforming larvae and juveniles were identified first by the pres- ence of known adult characters. Additional lar- val characters observed in those specimens were then used to aid in identification of the smaller specimens. Because all transformed specimens were sinis- tral and the right eye of all transforming speci- mens was migrating, it was decided that the four larval series belonged to one or more of the flat- fish families Bothidae, Scophthalmidae, or Cy- noglossidae. Morphological characters exhibited in the larval series and shared by larvae of these three families are lateral compression, deep head, deep abdomen, and looped gut, and in early larvae a raised and rounded dorsal profile of the head and slender caudal region. Only one scoph- thalmid species, Scophthalmus aquosus, is known from the western North Atlantic (Gutherz 1967; Hensley 1977). The distinctive rhomboid shape, long-based pelvic fins, and dense pigmentation of S. aquosus larvae were lacking in my series of larvae. The small eyes, small head, and confluent dorsal, caudal, and anal fins of cynoglossids were also lacking. In addition, cynoglossids from this region have fewer caudal (usually 9-14) and pelvic (usually 4 left, right) rays than the specimens in my se- ries. Therefore, Scophthalmidae and Cynoglossi- dae were eliminated from consideration. Gutherz (1971) summarized known characters most useful for identifying bothid larvae. Futch (1977) summarized subfamilial larval charac- ters and tentatively recognized two subfamilies, Paralichthyinae and Bothinae. The following discussion is limited to western North Atlantic species. Four paralichthyine genera.— Citharich- thys, Cyclopsetta, Etropus, and Syacium — have a similar combination of transitory (larval) and permanent characters that distinguish them from other bothid genera. These include: 1) adult caudal fin ray formula of 4-5-4-4; 2) placement of the left pelvic fin on the ventral midline and the right above the ventral midline, both origi- nating behind the cleithra (Gutherz 1971); 3) the same basic larval shape; 4) similar larval pig- mentation—on the gas bladder, in dorsal and anal lines, and in the caudal region; 5) larval pre- opercular spines (at least in Citharichthys cor- nutus, C gymnorhinus, C. spilopterus, Cyclop- setta fimbriata, C. chittendeni, Etropus crossotus, E. microstomus, and Syacium papillosum); 6) larval frontal-sphenotic spines (at least Cith- 38 TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND KTROPUS arichthys arctifrons, C. cornutus, C. gymno- rhinus, C. spiloptems, Cyclopsetta fimbriata, C. ch ittendeni, E. crossotus, E. microstomus, and S. papillosum). Caudal formula, pelvic fin place- ment, shape, and pigmentation of larvae in the four series corresponded to this group. Cyclopsetta spp. have 26-28 caudal vertebrae (Gutherz 1967). Larvae of C. fimbriata, C. chit- tendeni, and S. papillosum have 5-10 elongate dorsal rays and well-developed preopercular and frontal-sphenotic spines (Gutherz 1971; Futch and Hoff 1971; Evseenko 1979). Futch and Hoff (1971) listed Syacium generic larval characters. Other Cyclopsetta and Syacium larvae are prob- ably similar. Larvae in the four developmental series had lower caudal vertebral count ranges than Cyclopsetta spp., only 2-3 elongate dorsal rays, and relatively small preopercular and fron- tal-sphenotic spines. Therefore, these two genera were ruled out, leaving Citharichthys and Etropus. Identification to species is described in the individual species accounts. For aid in determining species of Citharich- thys and Etropus, frequency distributions of cau- dal vertebral, anal ray, and dorsal ray counts were tabulated from the literature, and from radiographs of juveniles and adults from the Atlantic off the southeastern United States (Append. Tables 1-3). Ranges of gill raker counts were tabulated from the literature (Append. Table 4). Number of caudal vertebrae (Append. Table 1) was the count most useful for distin- guishing larvae. Vertebral counts can be made before ossification during early or midflexion, and overlap is not excessive. However, care is necessary to avoid inaccurate counts because of fused centra. Caudal neural spines and hemal spines, both of which number one less than cau- dal vertebrae, will stain with alizarin and some- times can be counted before caudal vertebrae, during early or midflexion. The number of gill rakers on the lower limb of the first arch (Ap- pend. Table 4) can be counted in most specimens during transformation and can be very useful for identification of older larvae. The number of anal rays (Append. Table 2) is next in usefulness, followed by the number of dorsal rays (Append. Table 3); however, the overlaps for these counts are great. Efficiency can be gained by plotting individual anal versus dorsal counts on a graph, so that the counts can be used simultaneously. The adult complements of anal and dorsal rays are present by the end of transformation. After the largest specimens in each series were identified, the identities of successively smaller larvae were verified. The most useful characters for untransformed specimens were lateral, pec- toral, and notochordal pigment; number of elon- gate dorsal rays; number of caudal vertebrae; number and size of left pelvic rays; and head shape. DESCRIPTION OF DEVELOPMENTAL STAGES Citharichthys cornutus (Figs. 1-5) Identification Larvae approaching transformation had com- plete complements of countable characters. Those specimens were identified by comparing the following larval counts with known adult counts. Number of specimens is given in paren- theses. Caudal fin formula = 4-5-4-4 (27) Caudal vertebrae = 25(11)-26(16) Gill rakers (lower limb, first left) = 12 (1) Left pelvic rays = 6 (17) Anal rays = 60-66(11) Dorsal rays = 78-84 (11) Of the potential species listed in Table 1, only C. cornutus has counts that agree with these. In addition, larvae were captured over the outer shelf, slightly farther offshore than C. gymnorhi- nus (Fig. 1). This is consistent with bathymetric distribution of adults. Distinguishing Characters Citharichthys cornutus larvae have no pectoral melanophore, and notochordal pigment is re- stricted to the caudal region. Three elongate dorsal rays are present from preflexion (about 4 mm) through transformation. Caudal vertebrae (25-26) can be counted by early flexion (6 mm). Lateral pigment is relatively heavy. Flexion is complete at 9-10 mm SL. The larval mouth and eye are large. Morphology is similar to that of C. gymnorhinus. However, the left pelvic fin of C. cornutus has a full complement of six rays, and in larvae the first ray is not reduced in size. The left pelvic fin of C. gymnorhinus has only five rays, and in larvae the first ray is much reduced com- pared with that of C. cornutus. Length of C. 39 FISHERY BULLETIN: VOL. 80, NO. 1 Table 1.— Distribution of adults of Cithariehthys and Etropus species known from the western North Atlantic and status of knowledge of their larvae. 1 Species Geographic range of adults Depth range of adults (m) Larval descriptions C. arctifrons E. microstomas C. cornutus C. gymnorhinus C. spilopterus E. crossotus C. macrops E. rimosus C. abbotti C. amblybregmatus C. arenaceus C. dinoceros C. uhleri 2 E. intermedws 3 Georges Bank to Yucatan 22-682 Richardson and Joseph 1973 New York to South Carolina 5-91 Richardson and Joseph 1973 Georgia to Brazil 27-366 This paper Florida to Guyana 37-201 This paper New Jersey to Brazil (rare north of Virginia) 1-73 This paper Chesapeake Bay to French Guiana 1-86 This paper Southern Atlantic and Gulf coasts of the United States 1-91 Brief description in Dowd 1978 North Carolina to Mississippi River 5-190 Brief description in Dowd 1978 Veracruz to Campeche, Mexico 0-2 Unknown Western Caribbean off Nicaragua 139-197 Unknown West Indies to Brazil Shallow Unknown Florida to Nicaragua 183-1829 Unknown Haiti Unknown Trinidad to Rio de Janeiro 27 Unknown ons compiled from Goode and Bean 1896; Gutherz 1967, Dawson 1969; Gutherzand Blackman 1970; Toppand Hoff 1977; Wenner et al. 1979; and original data for C. spilopterus. E. crossotus, and C macrops (i.e., 1 m depths). nar.eus Dawson 1969 'Distribut 1972; Leslie . 2 cf C. arenaceus, Dawson 1969 3 cf. E. crossotus, Gutherz 1967 Figure 1.— Occurrence of Citharichtkys cornutus and C. gym- norhinus larvae off the southeastern United States. Numbers are the sums of bongo and neuston tows made per 1° quadran- gle during four RV Dolphin fall, winter, and spring cruises in 1973 and 1974. Symbols indicate positive tows. cornutus at transformation is about 18 mm. Lar- vae may appear in collections year-round. Pigmentation Pigmentation of C. cornutus larvae is rela- tively heavy. Gas bladder, gut, and lateral tail 40 pigment are the most striking. By 2.2 mm NL and throughout larval development, the dorsal one-third of the left side of the gas bladder is fairly heavily pigmented. This pigment may be diffuse or in the form of stellate or punctate melanophores. With growth, the number of mel- anophores increases. The maximum number in a preflexion specimen was five (4.8 mm). The right side of the gas bladder is usually unpigment- ed. During preflexion (2.2 mm, see Fig. 4A), two or three melanophores are present on both the dorsal and the ventral body margins about half- way between the anus and the notochord tip. Another one or two melanophores are present between these two clusters near the lateral mid- line. Later in development, pigment in this area forms a band. One or two small melanophores may be present on the ventral finfold just poste- rior to the hindgut. Three or four melanophores are on the caudal finfold near the ventral body margin just anterior to the notochord tip. Two melanophores are on the ventral surface of the gut loop; additional melanophores appear there during development. A small melanophore ap- pears along the posterodorsal surface of the mid- gut at about 3 mm. Melanophores begin appear- ing on the ventral body margin anterior to the cleithrum at about 3 mm. At about 4.7 mm, one or two melanophores appear along the posterior margin of the articular. Flexion larvae (see Fig. 4B) usually have four or five melanophores on the dorsal one-third of the left side of the gas bladder. Midlateral caudal pigmentation consists of up to six dashlike mel- anophores. Additional, dashlike clusters of pig- ment appear along the dorsal and ventral body TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND ETROPUS margins between the anus and the caudal fin base. During midflexion (6 mm, see Fig. 4B), inter- nal pigment appears along the dorsal notch be- tween the midbrain and hindbrain, and one or two round melanophores appear below the notch. Visible internal notochordal pigment is re- stricted to the vicinity of the external caudal band. The dorsal surfaces of one to three forming centra are darkened by about 6 mm. Several melanophores are present along the ventral body margin from just above the tip of the urohyal to just behind the cleithrum. Internal pigment appears between the hindgut and anal fin origin by midflexion. By late flexion (8 mm, see Fig. 4C), both sides of the gas bladder are obscured by body muscula- ture, and pigment appears diffuse. Notochordal pigment appears as fine dashes along the dorsal surfaces of three to six centra of caudal vertebrae 15-21. As many as 30 or more melanophores may be present along the ventral surface of the gut loop. Pigment along the posterodorsal surface of the midgut extends to the gas bladder and ap- pears as a black lining over the gut. One or more melanophores appear on or just behind the posterodorsal margin of the preopercle. Melano- phores have developed along the elongate second left pelvic ray and begin to develop along the elongate dorsal rays at 8-9 mm. Some larvae have small melanophores near the distal tips of rays at the middle of both dorsal and anal fins. By 8 mm, a group of melanophores has appeared along the middle of the caudal fin. The posterior margin of the articular is covered with a stellate melanophore. By postflexion (9 mm), myoseptal pigment is present in the caudal band as well as adjacent to dorsal and ventral lines. Internal pigment along the brain surface looks diffuse. Pigment appears on the dorsal fin membrane adjacent to the first dorsal ray at about 11 mm. Body musculature tends to obscure dorsal notochordal pigment in larvae longer than 12 mm. Additional midlateral dashlike melanophores appear near the caudal fin base at 13-14 mm (see Fig. 5A). By about 14 mm, all five dorsal and four ventral pigment lines have formed, and myoseptal pigment is well developed. A small amount of pigment is present along the anteroventral edge of the maxillary by about 14 mm. Late transforming larvae have about three small internal melano- phores near the pectoral fin base and just for- ward of the cleithrum beneath the angle of the last gill arch (barely visible through the opercle); these probably develop by about 14 mm. Ventral pigment from the urohyal to the cleithrum per- sists until late transformation. By late transfor- mation (see Fig. 5B), midlateral dashlike melan- ophores are present anterior to the caudal band. Morphology (Figs. 4, 5; Tables 2, 3) General morphological features include later- al compression, a deep head, a deep abdomen, and a looped gut. In early larvae the dorsal pro- file of the head is raised and rounded and the caudal region is slender. The eye is nearly spheri- cal during early development but becomes ellip- soidal in transforming larvae. A ventral choroid fissure is visible from 3-4 mm NL until about the end of the postflexion stage. The nasal capsule is visible by about 3 mm NL. The gas bladder is prominent just above the foregut until the end of postflexion. It bulges slightly on the left side of the body and is not as obvious on the right. A loop forms in the gut by 2 mm NL. The liver occupies a large portion of the anteroventral region of the abdomen. Adult morphometries given in the fol- lowing discussion were derived from Topp and Hoff (1972). The mouth is relatively large in larvae and adults. Larval upper jaw length/BL increases slightly from 10.3% (preflexion) to 11.0% (flexion) and then decreases to 9.8% (postflexion). Adult upper jaw length/BL is 12.8%, range 11.8-13.7%. Larval upper law length/HL decreases from 37% to 34%. Adult upper jaw length/HL is 45%. Lar- val lower jaw length/BL increases slightly from 13.3% (preflexion) to 13.9% (flexion) and then de- creases slightly to 13.0% (postflexion). Larval lower jaw length/HL decreases slightly from 48% to 46% and is only slightly greater than that of C. gymnorhinus. Larval snout length is moderate. Larval snout length/BL increases slightly from 6.2% (preflex- ion) to 7.1% (flexion) and then decreases slightly to 6.3% (postflexion). Adult snout length/BL is 5.5%, range 4.8-6.2%. Larval snout length/HL is constant at about 22-23%. Adult snout length/HL is 19.5%. The larval eye is large, and the relative size of the adult eye is greater than that of any other western North Atlantic Citharichthys or Etropus species except C. amblybregmatus. Larval eye diameter/BL is constant at 9.8% during preflex- ion and flexion and then decreases to 8.5% (post- flexion). Adult orbit length/BL is 10.0%, range 41 FISHERY BULLETIN: VOL. 80, NO. 1 Table 2.— Measurements (mm) of larvae of Citharichthys cornutus. Pref = preflexion, ECF = early caudal formation, Early = early flexion, Mid = midflexion, Late = late flexion, Post = postflexion. S = symmetrical, 1 = to one-third of the way to the dorsal ridge, 2 = one-third to two-thirds of the way to the dorsal ridge, 3 = two- thirds to all the way to the dorsal ridge, R = on the dorsal ridge. sz O) c a >- •D O m c o .2. m Q. a Ol c a> S CO CD 5 o _l D) C 2> W o c w £ E ra TJ . LU O) c a> n ca CD I CO C ca 0£ .- CT1 =3 C O CD C — w -C a> c JD "cO o 1- sz Q. . O "O o o — m CO c CD Q. CD T> >^ "D O CO CD c a. CO £ E O. CD CD J= ■D n O " CO a o c z> ■o CD Q. it O CD Ol ca c o X CD LL. c o o o. CD >. CD -C Ol rr 22 0.33 0.12 0.27 062 1.1 '0.76 '0.61 '0.61 '0.53 '026 Pref S 32 0.35 0.41 0.21 0.31 1.0 1.7 1.2 '1.1 '10 '0.79 '036 Pref S 37 0.40 0.47 0.25 0.37 1.1 1.9 3.8 1.2 '1.2 '12 '0.88 '0.47 Pref s 4.0 0.56 028 0.44 1.1 1.9 14 '1.4 '15 '1.2 '064 Pref s 4.1 0.59 027 0.43 1.2 2.1 4.2 '1.3 '0.57 Pref s 4.5 0.47 0.57 0.27 043 1.2 18 4.6 1.4 '14 '11 '065 Pref s 45 0.44 0.61 0.23 040 1.1 1.9 4.6 1.5 '1.4 '1.5 '1.2 '0.63 Pref s 4.6 050 0.71 0.29 047 1.3 2.2 4.7 1.8 1 7 1.5 0.83 ECF s 48 0.51 0.66 026 0.48 1.3 2.2 5.0 1.6 '1.6 '1.6 '1.4 '082 ECF s 49 0.28 0.47 1.3 2.1 5.0 '1.6 '18 '1.5 '083 ECF s 5.0 0.58 0.73 0.37 0.53 1.5 2.2 1.7 1.9 ECF s 5.7 0.48 0.63 0.32 0.50 1.5 2.3 5.8 1.9 1.9 1 9 1 1 ECF s 5.7 0.53 0.65 0.35 0.52 1.5 2.5 20 1.9 2.0 1.0 ECF s 58 0.65 0.81 0.33 0.63 1.7 2.7 2.2 2.5 2.6 25 16 57 Mid s 6.0 063 0.78 0.40 0.53 1.7 2.4 2.1 2.2 2.4 2.2 1.4 Early s 6.1 0.69 0.86 0.37 0.63 1.8 3.1 2.4 2.7 3.0 2.6 1.8 061 Mid s 63 66 0.84 1 8 2.6 7.4 2.4 2.5 1 7 0.59 Mid s 6.3 0.75 0.96 0.46 0.65 3.0 2.5 28 3.1 2.7 1 8 0.64 Mid s 64 070 0.92 0.41 063 1.8 2.9 7.5 2.5 2.8 3.0 2.7 1.8 0.64 Mid 1 64 0.70 089 0.44 0.63 1.8 3.0 7.2 24 2.6 28 2.5 1.6 054 Mid s 64 0.77 0.90 0.55 0.63 3.0 2.5 2.8 3.1 27 1.9 0.65 Late 1 69 0.87 1.1 0.53 0.73 2.3 3.2 8.6 28 3.2 3.6 3.3 2.3 088 Late 1 7.2 0.76 0.96 0.46 0.74 2.2 3.2 2.8 3.0 3.4 3.1 2.1 084 Late s 7.2 0.77 1.0 0.47 0.75 2.2 3.2 8.7 2.7 3.0 35 3.3 2.2 080 Late s 7.6 090 1.2 0.60 077 2.5 3.6 3.2 37 4.2 4.0 3.0 1.0 Late s 7.6 0.77 099 0.51 0.73 2.2 3.6 9.2 3.0 3.2 3.0 2.3 082 Late s 7.6 0.83 1.0 050 077 2.3 3.3 28 32 3.6 3.3 2.3 088 Late 2 7.6 084 1.1 0.50 23 37 9.3 3.0 3.6 4.0 3.5 2.6 090 Late s 7.7 090 1.1 0.65 0.73 2.4 34 3.2 3.5 3.7 3.5 2.6 0.94 Late s 7 9 0.81 1.0 0.54 0.74 2.4 3.9 9.5 3.5 3.8 3.6 2.5 090 Late 1 82 091 1.2 64 0.81 2.6 3.6 3.7 4.1 4.0 3.0 1.1 Late 1 8.2 0.87 1.1 0.62 0.76 2.5 3.6 3.1 3.4 3.5 3.3 2.5 0.87 Late 1 83 097 1.3 0.71 0.81 2.7 4 1 3.5 4.0 4 8 4.3 3.0 1.1 Late 1 8.3 096 1 2 051 085 26 105 40 46 4.1 30 1.1 Late s 84 090 1.1 0.61 0.80 2.7 4.2 10.2 3.4 38 4.4 3.9 2.8 0.94 Late 1 86 90 1.1 066 0.81 2.7 4.1 10.7 3.2 3.7 4.4 4.1 3.0 1 Late 2 88 095 1.2 0.61 080 2.6 3.7 3.2 3.8 3.9 3.9 29 10 Late 2 8.9 0.90 1 1 062 080 2.6 3.9 3.4 3.8 4.1 38 2.9 1 Late 2 10.4 1.1 1.4 074 093 32 4.6 12.9 4.0 4.6 5.3 5.1 40 1 4 Post 3 10.6 1.1 1.5 77 1.0 3.2 4.8 13.2 4,1 48 5.8 5.6 4.1 1.5 Post 3 10.6 1.2 1.5 0.73 0.94 3.2 4.6 3.9 4.6 5.3 5.1 4.1 14 Post 1 109 1.2 1.5 0.79 099 3 2 4.3 4.0 4.7 5.7 5.3 4.1 1.4 Post 3 11 5 1.2 1.6 0.74 1.1 3.4 4.9 14.1 4.0 4.6 5 5 5.3 4.1 1.4 Post 3 12.0 1 1 1.6 089 99 3.5 4.6 14.7 43 4.9 5.1 52 4.3 1.4 Post 3 12.1 1.3 16 0.73 1.1 3.6 50 46 5 1 6.1 6.0 4.9 1 6 Post 2 12.8 1.1 1.6 0.64 1 1 3.5 5.0 44 4.9 5.8 5.5 4.6 16 Post 3 129 1 2 1.6 073 1.0 3.5 5.1 4.8 5.2 6 1 59 4.8 1.6 Post 3 130 1.2 16 72 1.0 3.5 4.8 15.9 45 54 5.9 59 5.3 16 Post 3 138 1 2 1.7 0.70 1.2 3.7 5.2 5.4 5.9 6.7 6.5 5.5 18 Post 3 15.4 1 4 1.8 095 1.2 40 5.6 185 5 1 6 1 69 68 6 1 1.9 Post 3 174 1.6 2.2 1.2 1.2 48 5.6 21.2 5.7 6.6 7.5 7.6 7.2 2.2 Post 3 17.4 1.7 2.1 1.0 15 4.8 5.6 6.3 6.5 7.7 6.8 2.0 Post R 'Measurement does not include dorsal or anal pterygiophores 9.2-11.1%. Larval eye diameter/HL decreases from 36% to 30% and is similar to that of C. gym- norhinus. Adult orbit length/HL is 35.5%. The head is relatively large in larvae and mod- erate in adults. Larval head length/BL increases from 28% (preflexion) to 30% (flexion) and then decreases to 28% (postflexion). Postflexion head length/BL is similar to those of C. arctifrons and C. ffymnorhinus. Adult head length/BL is 28%, range 27-30%. Larval head depth/BL increases from 34% (preflexion) to 39% (flexion) and then decreases slightly to 36% (postflexion). Larval snout to anus length is relatively great until postflexion. Snout to anus length/BL is 46% during preflexion and flexion and then decreases greatly to 39% (postflexion). The body is relatively deep in larvae and mod- erate in adults. Larval body depth at pelvic fin/ 42 TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND ETROPUS Table 3.— Body proportions of larvae and juveniles of three species of Citharichthys and one species of Etropus. Except for body length, values are in percentage of body length (BL) or of head length (HL) and are given as: mean ± stan- dard deviation (range). (Values derived from Tables 2, 5, 6, 7.) Measurement C. cornutus C. gymnorhinus C. spilopterus E . crossotus Body length (mm) Preflexion 4.6 (3.2-5.7) 46 (4.4-5.0) 3.7 4.6 Flexion 7.4 (5.8-8.9) 6.7 (5.3-7.7) 64 (5.7-6.8) 6.8 (4.9-9.5) Postflexion 12.9 (10.4-17.4) 104 (7.9-12.9) 9.4 (8.3-10.6) 10.2 (9.3-10.8) Early juvenile 10.0 (8 7-11.6) Midjuvenile 20.5 (14.3-25.4) Upper jaw length/BL Preflexion 10 3±1 0(8.4-1 1.6) 9.5+0.7(8.3-10.3) 99 7.0 Flexion 11.0±0.6(10. 1-12.5) 9.3+0.4(8.3-10.0) 7.2+0.8(6.3-7.9) 7.2+0.7(5.9-8.4) Postflexion 9.8+0.8(8.6-10.8) 9.3+0.7(8.1-11.3) 6.7+0.6(5.6-7.9) 7.1+0.5(6.4-7.8) Early juvenile 7.3+0.4(6.1-8.1) Midjuvenile 9.0+0 4(8 4-9.7) Lower jaw length/BL Preflexion 13.3±1. 3(1 1.0-15.3) 11 .5+1 0(10.2-12.9) 12.1 9.6 Flexion 13.9+0.9(12.4-15.5) 12.7+0.6(11.6-13.6) 9.9+0.7(9.1-10.4) 9.6+1.0(8.5-12.3) Postflexion 13.0+0.8(11.9-14.4) 12.7+0.6(11.6-14.2) 9.1+0.5(8 2-10 2) 9.8+0.5(9.2-10.6) Early juvenile 10.3+0.5(8.9-11.5) Midjuvenile 13.1+0.4(12.5-13 8) Snout length/BL Preflexion 6.2+0.7(5.1-7.4) 5 .2+1. 1(3.7-6.7) 7.5 5.2 Flexion 7 1+0.8(5.7-8.6) 5 8+0.6(4.9-7.0) 7.6+0.8(6.8-8.1) 6.4+0.7(5.1-7.5) Postflexion 6.3±0.8(5.0-7.4) 6.1+0.6(4.8-7.6) 6.4+0.6(5.6-7.4) 6.8+0.9(5.4-7.6) Early juvenile 5.8+0.6(4.4-7.2) Midjuvenile 5 0+0.5(4.3-5.5) Eye diameter/BL Preflexion 9 8+0.7(8.8-11 0) 8 8+0.5(8.1-9.5) 9.7 7.4 Flexion 9.8+0.5(8.8-10 8) 8.9+0.8(6.9-10.0) 7.9+0.2(7.6-8.1) 6 9+0.4(6.1-7.7) Postflexion 8.5+0.6(7.2-9.4) 8 8+0 5(7.9-9.8) 6.5+0.6(5.5-7.6) 6 3±0.3(6 1-6.9) Early juvenile 6.8+0.5(5.9-7.7) Midjuvenile 7.0+0.6(6.5-8.2) Head length/BL Preflexion 27.6±2.2(23.8-31.2) 24.8+1.0(23.8-26.6) 28.0 23.4 Flexion 30.4±1 .5(28.0-33.1) 27.9+1.9(25.0-31.1) 26.4+0.8(25.5-27.0) 26 4+1.3(24.2-28 7) Postflexion 28.4+1.6(25.9-30.5) 28.6+1.7(26 8-33.8) 23.9+1.0(22.4-25.7) 26.4+1.5(24 4-28 8) Early juvenile 25.4+0.8(23.9-27.1) Midjuvenile 25.0+0.7(24.2-26.2) Snout to anus length/BL Preflexion 45.8+4.7(39.4-54.1) 43.0±1. 2(42.0-45.1) 40.0 39.1 Flexion 45.9+2.9(40.1-51 2) 44.2+2.0(40.2-46.6) 39 0+1.4(37.5-40.2) 44.2+2.0(39 8-48.0) Postflexion 39.3±4.1 (32.1-45.8) 39.7+2.9(34.6-46 2) 31.8+1.2(29.7-33.5) 38 8+3.2(33.5-42.2) Early juvenile 31.0+1.3(28.8-34.0) Midjuvenile 31 .6±1. 2(29.8-34.3) Total length/BL Preflexion 102.3±0.7(101 .6-103.5) 102.0±0. 6(101. 5-102. 9) 102.2 101.5 Flexion 121.0+39(112.3-126.5) 1160 + 126(102.0-126.2) 128.0 115 0+8.8(100.7-127.4} Postflexion 123 0±1 .4(120 6-124.9) 121 3+1.0(119.8-122.5) 123.1+1.2(121.4-125.0) 122.3+1.3(119.7-123.4) Early juvenile 123.9+2.0(119.4-129.4) Midjuvenile 125.4+1.6(123.4-128.0) Head depth/BL Preflexion 34.5+2.3(31.4-38.8) 29 0±1. 8(26.7-31 .4) 36.6 28.6 Flexion 38.7+2.0(34.3-42 5) 33.3+2.3(28 3-36 2) 39.4+2 0(38.1-41.7) 33.8+1.6(30.7-36 4) Postflexion 36.2+1.9(32.9-38 8) 33.3±1. 6(31 .0-36.3) 32 8+1.1(31.0-34 9) 33.1+1.6(31.5-35 8) Early juvenile 32.0+1.1(29 8-34.9) Midjuvenile 31 .0±1. 0(29 6-33 0) Depth at pelvic fin/BL Preflexion 33.6±2.0(31 .2-37.6) 29.8+2.6(27.0-33.5) 40.3 26.2 Flexion 43.6+2.9(37.1-49.1) 37.5+3.1(31.9-42.7) 46.7+0.9(45 9-47.7) 39.0+4.0(32.7-49.7) Postflexion 41.3+2.4(37.4-45.8) 39.0+1.7(36.4-43.9) 39 0+1.1(36 9-40 8) 40.2+2.4(36.2-43.6) Early juvenile 37.2±1. 2(34.7-40.3) Midjuvenile 35.0+0.9(33.3-36.4) Depth at loop of gut/BL Preflexion 33.7+1.9(31.0-37.0) 29.2+2.9(25.5-33.9) Flexion 48.4+4.4(39 6-57.4) 39.1+4.3(32.7-45.0) Postflexion 48 0+3.6(42 6-54.7) 43.9+2.2(41.5-50.2) Depth at anus/BL Preflexion 28.4+3.6(23 8-33.7) 24.9+2.5(22.4-29.2) 38.7 21 4 Flexion 44.5+4.0(36.1-52 4) 36.9+4.4(30.3-44.2) 50 6+1 2(49 2-51.4) 37.8+5.6(29.3-45.7) Postflexion 46.6+2.8(43.0-52.9) 42.2+1 7(39.7-47.0) 42 9+1.6(39 9-44.9) 43.1+3.5(36.3-46.2) Early juvenile 39.5+1.4(35.9-43.3) Midjuvenile 38.7+0.7(37.2-39.7) Depth at third hemal spine/BL Preflexion 15.6±2.5(1 1.2-19 .5) 14.2+2.3(11.9-18.2) 22.6 116 Flexion 31.4+3.9(22 8-39 8) 26 6+4.3(19.7-33 2) 40 8+2.9(38.7-44.0) 28.2+7.3(18.2-39 1) Postflexion 38.6±1 .9(35.6-41 .8) 34.8+2.8(28.6-42.0) 38.0+1.4(35.3-39.8) 37.7+1.5(36.2-40 2) Early juvenile 37.0+1.4(34.6-40 0) Midjuvenile 40.2±1. 0(39.0-42 3) 43 FISHERY BULLETIN: VOL. 80, NO. 1 Table 3.— Continual. Measurement C. cornutus C. gymnorhinus C. spilopterus E. crossotus Caudal peduncle depth/BL Flexion Postflexion Early juvenile Midjuvenile Upper jaw length/HL Preflexion Flexion Postflexion Early juvenile Midjuvenile Lower jaw length/HL Preflexion Flexion Postflexion Early juvenile Midjuvenile Snout length/HL Preflexion Flexion Postflexion Early juvenile Midjuvenile Eye diameter/HL Preflexion Flexion Postflexion Early juvenile Midjuvenile 11.4+1.4(8.4-13.8) 12.7±0.6(11.6-13 8) 37.4+2.7(32.6-40.7) 36.0+1.8(33.2-39 1) 34.4+1.6(31 6-36.6) 48.0+4.9(41.0-56.5) 45.5±2.9(39.5-51.4) 45.9+1.1(43.6-47.7) 22.4+1 5(20 2-25.2) 23.0+2.0(19.3-27.2) 22.3+2.3(18.4-25 5) 35.6±2 .1(31.0-39.6) 32.2+2.1(29.6-37.5) 29 9+1.5(26 3-31.7) 11.6+1.6(9.3-14.4) 13.2±0. 7(12.2-14. 7) 38.3+2.0(34.9-40.9) 33.3+2.1(29.2-37.6) 32.6+1.9(29.4-37.5) 46.5+3 2(42.7-51.2) 45.6±2. 2(42. 5-49.6) 44.5+1.8(40 2-46.8) 20 8+4.1(15 4-26.8) 20.9+1.6(19.0-24.6) 21.3±1. 9(17.6-24 0) 35 4+2.2(32 3-38.2) 31.9±2.7(27.8-39.1) 30.6+1 6(27.4-33.9) 14.4±1.6(13.1-16 2) 13.5±0.5(12.7-14.3) 13.6+0.6(12.2-14.6) 11.4±0.6(10.0-12.3) 35.6 27.2+2.4(24.7-29.4) 28.1+2.2(24.0-33.8) 28.7+1.5(25.4-31.3) 36.2+1.4(33.9-38 4) 43.3 37.5±1. 4(35.9-38 6) 38.2+1.7(35.9-41.5) 40 5±2 1(36.9-44.1) 52.6+1.5(50 9-55 9) 269 28 8+3 3(25 0-31.2) 27.0±2.5(23.2-31.4) 23.0±2.3(18.3-27.1) 19.9+1.7(17.5-22 5) 34.6 29 9+0.5(29.4-30.3) 27.1+2.0(23.4-31.1) 26.8±2.0(23. 1-30.0) 27.8±1 6(26.4-31.2) 9.6+3 0(4 7-13.4) 12.6+0.6(11.8-13.4) 299 27.2+2.4(21.6-30 5) 26 9±1. 1(24. 7-27 6) 41.1 36.3±3.2(30 8-46 8) 37.2+1 .5(35.4-39 0) 224 24.3+1.9(21.0-28 0) 25 9±2.7(22.3-29 1) 31.8 26.0+1.8(22.0-30.6) 23 6+0.4(23.3-24 3) BL increases from 34% (preflexion) to 44% (flex- ion) and then decreases slightly to 41% (post- flexion). Larval body depth at loop of gut/BL increases from 34% (preflexion) to 48% (flexion and postflexion). Larval body depth at anus/BL increases greatly from 28% to 47%. Larval body depth at third hemal spine/BL increases greatly from 16% to 39%. Adult body depth/BL is 46%, range 43-50%. Larval caudal peduncle depth/BL increases from 11.4% (flexion) to 12.7% (postflex- ion). Adult caudal peduncle depth/BL is 10.5%, range 9.7-11.4%. Fin and Axial Skeleton Formation Development of the caudal skeleton of C. cor- nutus from larva to juvenile (Fig. 2A-E) is typi- cal of the four species described in this paper. The major difference among them is the rate of development. Flexion is complete at about 7-8 mm in C. gymnorhinus and C. spilopterus and at about 9-10 mm in C. cornutus and E. crosso- tus. During preflexion, before caudal formation (2.2-4.5 mm NL), the notochord is straight and there is no evidence of hypural formation. Dur- ing early caudal formation (4.6-5.7 mm NL, Fig. 2A) the notochord is straight and the outline of incipient hypurals 2+3 and 4+5 are visible, but neither hypurals nor incipient caudal rays are calcified. During early flexion (6.0 mm NL, Fig. 2B) the notochord begins to turn upward. Hypurals 2+3 and 4+5 (sometimes hypural 1) and caudal rays begin to stain with alizarin, and the last neural and hemal spines stain with alizarin. Caudal rays form in about equal numbers dorsally and ventrally during flexion, beginning at the posteroventral corner of hypural 4+5 and the posterodorsal corner of hypural 2+3. The 6.0 mm specimen (Fig. 2B) was the smallest in which calcification of caudal rays had begun. During midflexion (6.1-6.4 mm NL, Fig. 2C) the noto- chord is S-shaped and hypurals 1 and 6 and the epural begin to stain with alizarin. During late flexion (6.4-8.9 mm NL, Fig. 2D) the notochord tip points upward and is nearly flexed but is still in contact with hypural 6 and the epural; all hypurals stain with alizarin and the last neural spine touches hypural 6. All rays are formed by about 7.5 mm NL. When flexion is complete (10.4-17.4 mm SL, Fig. 2E) the urostyle is separate from hypural 6 and the epural, and all caudal rays stain with alizarin. Fusion of the epural with hypural 6, and fusion of hypural 4+5 with the urostyle occur at about the time of transformation. The terminol- ogy of Amaoka (1969) is followed here; however, actual fusion of hypurals 2 with 3 and 4 with 5 was not observed. The adult caudal skeleton of C. cornutus (Fig. 3A) is composed of a urostyle, or terminal half 44 TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND ETROPUS A B D Figure 2.— Development of the caudal skeleton of Citharichthys cornutus: A. Preflexion (early caudal formation), 5.7 mm NL; B. Early flexion, 6.0 mm NL; C. Midflexion,6.4mmNL; D. Late flexion, 8.2 mm NL;E. Postflexion, 13.7 mm SL. NS = neural spine, HS = hemal spine, H Yl = hypural 1, HY2+3 = hypurals 2 and 3, 4+5 = hypurals 4 and 5, H Y6 = hypural 6, EP =epural, PV = penulti- mate centrum, UR = urostyle. Scale = 1 mm. centrum (according to Hensley 1977, in the bothid Engyophrys senta this bone consists of the first and second ural centra and the first preural centrum); a penultimate, or second preural, cen- trum (see Hensley 1977); an enlarged hemal spine from the second preural centrum support- ing hypural 1; autogenous, proximally free, hypural 1 which supports three unbranched and one branched ray (equivalent to "parhypural" of some authors— e.g., Futch 1977; Hensley 1977— 45 FISHERY BULLETIN: VOL. 80, NO. 1 see Sumida et al. 1979); autogenous, fused hy- purals 2 and 3, articulating ventrally with the urostyle and supporting four branched rays; fused hypurals 4 and 5, fused with the tip of the urostyle and supporting five branched rays; an autogenous, proximally free element consisting of hypural 6 fused anteriorly with the single epural, one branched and three unbranched rays supported by hypural 6; no evidence of a uro- neural; an enlarged neural spine from the second preural centrum supporting the epural. The cau- dal skeletons of the four species described here are similar to Amaoka's (1969) type 4, except for the lack of a uroneural. Dendritic splitting of hypurals 2+3 and 4+5 occurs in Etropus crossotus by about 40 mm SL (Fig. 3B). The hypurals of adult specimens of Citharichthys spp. examined were sometimes grooved but never split as in E. crossotus. Hypur- als 2+3 and 4+5 of E. microstomas and E. rimosus were similar to those of Citharichthys spp. except for an apparent tendency to split slightly at the distal margins. In C. comutus larvae all precaudal neural spines stain with alizarin by about 4.8 mm NL. Some caudal neural spines and hemal spines stain with alizarin at 4.8 mm NL and all do by 6.1 mm NL. The urostyle stains with alizarin at 6.3 mm NL. All precaudal and caudal centra stain with alizarin by 7.2 mm NL. The smallest speci- FlGURE 3.— Caudal skeletons of two bothids: A. Citharichthys comutus, 51.5 mm SL; B. Etropus crossotus, 49.4 mm SL. Abbreviations as in Fijrure 2. Scale - 1 mm. 46 TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND ETROPUS men in which caudal centra could be counted was 5.8 mm NL (midflexion). The second, third, and fourth dorsal rays are elongate and widely separated at the bases from preflexion (about 4 mm NL) through transfor- mation (17.4 mm SL). During early caudal formation (5.0 mm NL), rays near the middle of the dorsal fin begin to calcify. Calcification pro- ceeds anteriorly and posteriorly. Adult counts are present from late flexion (6.4 mm NL) on- ward. The first ray and the most posterior rays are calcified just prior to transformation (17.4 mm SL). During early caudal formation (5.0 mm NL), anal rays near the middle of the fin begin to cal- cify. Calcification proceeds anteriorly and pos- teriorly. Adult counts are present from late flexion (about 8 mm NL) onward. The most pos- terior rays are calcified just prior to transforma- tion (17.4 mm SL). Development of the left pelvic fin precedes that of the right fin. The left pelvic fin bud ap- pears during preflexion (3.7 mm NL). Rays develop between preflexion (4.0 mm NL) and late flexion (about 8.9 mm NL). The most ante- rior two rays are the first to appear; the second is elongate and the first slightly elongate. The right pelvic fin bud appears during early caudal for- mation (4.9 mm NL). Rays develop between mid- flexion (6.0 mm NL) and late or postflexion (9-10 mm BL). Each complete fin has six rays. Rayless, fanlike, larval pectoral fins were present on the smallest available specimen (pre- flexion, 2.2 mm NL). Calcification of rays in the left fin occurs between about 13 mm and 17.4 mm SL. Calcification of rays in the right fin had not begun in the largest specimen (17.4 mm SL). Cephalic Spination Preopercular spines (Table 4) were present in the smallest preflexion specimen (2.2 mm NL, Fig. 4A). With development (Fig. 4B, C), addi- tional spines appear until maximum numbers of about 33 on the left (range 26-52) and 39 on the right (range 23-50) are reached during late flex- ion (6.4-8.9 mm NL). Thereafter, spines are lost until none or only a few remain at transforma- tion (17.4 mm SL, Fig. 5B). Frontal-sphenotic spines were evident in the second smallest preflexion specimen (3.2 mm NL) and throughout the larval series, though less conspicuous near transformation (13-17 mm SL). The lowermost spine on the left side is usually just above the center of the eye and on the right side slightly anterior to the center of the eye. (During transformation those on the right side are at the anterior margin of the skull.) The spines are arranged in a slightly posteriorly con- cave arch following the curve of the skull. There are usually six (up to eight) spines per side, in- cluding three stronger spines arising from a small bulge of the skull. Larval Teeth (Table 5) No teeth are present at 2.2 mm NL (Fig. 4A). At 3.2-4.1 mm NL, larvae usually have two upper and two lower teeth on each side. A 5.3 mm NL preflexion specimen had three upper and four lower teeth on each side. The same numbers were present in the largest early caudal formation specimen (5.7 mm NL). During flexion, numbers of teeth increase from about four upper and five lower (about 6 mm NL) to about eight upper and seven lower (8.9 mm NL) on each side. Postflex- ion larvae (10.4-13.8 mm SL) have about nine upper and nine lower teeth on each side. The nearly transformed specimen (17.4 mm SL, Fig. 5B) had fewer upper teeth on the left side (about 11) than on the right side (19) but the same num- ber (about 15) in both lower jaws. Transformation Migration of the right eye may begin as early as midflexion (6.4 mm NL) or as late as postflex- ion (10.6 mm SL). The right eye moves from the right side of the head through a space between the dorsal fin and supraorbital bars (Fig. 5A) as in Cyclopsetta fimbriata (Gutherz 1971). The right eye reaches its final position on the left side of the head by about 18 mm SL. No early juvenile specimen was available, but eye migration in one of the 17.4 mm specimens was nearly complete (Fig. 5B). Occurrence Larvae were collected in the Atlantic during February, March, April, May, October, and No- vember (Powles 4 ). There was no apparent size progression by month, indicating an extended spawning season. Water depth was 46-640 m. 4 H. W. Powles, Assistant Marine Scientist, South Carolina Marine Resources Research Institute, P.O. Box 12559, Charles- ton, SC 29412, pers. commun. July 1976. 47 FISHERY BULLETIN: VOL. 80, NO. 1 a c 2 oJ ' be El .2 a 3"° WW f. OS >> Q. I. L. ce O CO II C o &H 01 - X C g U X £ cp n <~ CO 0) c 31 II H <~ . 0) B t* c Q x QJ 0) o £ & — ii CD CO > o cL Oh TJ . — C a o > '3 c5 « o cv 0> as be — US II -*-> CO 0) >> »3 71 -Q J CO 0) c 'a. X. CO _0 Em 3 O E fc. cu ii D. o — s o. -*-> c _c _o _be X "E o> T3 55 C j>> aS u 4H» c< 0) Qj II «*- o >> t. 1_ 01 SO E W 3 ■z B o 1 X OJ •*s< wJ u j oa < H CO o u CI 00 01 II _J — CD O) n co co m n r- r- CM CM CO "J co cm co cm co in O O CO O CD CD CO O) S O) N Ifi t- y- r- i- ^- CM cm r- o co CM CM C\J r- i > i r-(oon CM t- i- i- CM o cm m i- in o cm co O) t- CM LO CO CM i- CM CM t- f~ O) O CO CM CM CM i- I s - CO h- Tj" ro cnj co cj cd m »- cm o cmitj r- co p r- o o ^r in d »- t- coV i- CM CM CM CM CO t coo co in CM in in ce cb s O T— in cd co r- CO t ^r in cb CO co ^r o in co co ■*t in cb oS o o a) ^r t- co ■*r in 1 cb oS •su.*, £ o a 2 a O (J co — CL HI LU 5 _i CL LU LU 2 CO *1 ■c .c I so 8 " 1 J CO a o O i - o 0- O CT> CO r- CNJ "T o ■» co uo ^r CO CM CO CO T CO CO CO 1 CM CM CM CO r- CO r- r - *- ^ t- r- o> co r- CM O) CM i- CO CO ^ *- *- CM CO CO CO } r- CM CD ^ CO o O CM CO (O O) i- t- CM co m co CO CM -«3- ^r CO <-CD^- (M (M ^ CD CM O r- CMCJ r- N ^f t- S O) 00 •- CM CO CO CD CM CM O r- t- (M CM CO CM / O N O ^ O) CO 1 in in CO CD CO CO N cm r- CM TJ- CD TT ^r in cd o i_ LJ CO ^ CO O IT CL LU LU 5 _l 0. »- i i ." 3 ■C o u r* r^ CM CNJ CO CD 'd O I I I s n n UO CD GO o5o2 lu _j a. i- « Q. o c o X 0j IS ? _>> cd i — i cd V II it a >, — ' L. M c & O aj a> J; IS CL ?! a D. 3 •-^ II •2 3 c« . C 0) F !» ^^ ^ CO 3 i- ct! O o -a >> -^ CO 03 a> ll &H o H c o s.a> D. co II § H-H t. °- c _o cu 'x be_a; rt <** *^ ^ CO CO __ o cO D. ^ II CD *j SCO . o _o ^ o qj • — -a x ^ ^S S ii — OJ a; w > as u J _53 " .2 u X qj cp J3 «S E^ 1 " • 12 w S CQ < S ° > — c F CD CO c en E CO _l >< T2 0J CO CO c en 1 - t CO _) ~- en CO CO co^co-^-TfLom rococo CMCNJCMCOCM^f'4 4 ID (O CO COCOCMCMCOCMCMT-CD^rcJ^m cOr^Lqucocouoin colocdo CNJCNJCMcbcO'^ ; -^ ; cDUOcbcbo6 cOTjco^^ifiincONcocoffi I I I I I I I I I I I I CMCMCMCOCM'J-TtLO-^-CDCDCD COCOCNCMCOCMCOCNJCO"** T- h* cosintf)comsinNmt-o cm cNJCMcbcbTT^LOLOcbr^cb ^ _ >U - J - l-) _ >-) - 5-)-0-) 3-JDJ3JDJDJDJ ■^t co o to r^ 01 ■jmixm' >io'pco ijs-oj LLuoLUcp l Sco2^._j h l CL <3) ■^t tt tio cb cd r^ co in n ^ r-. o> coincoincos^co f-ojojooioiMomT-i- ur> Tf r^. oo cm Tt ninnincocdcosuiNNO) 10 io Mnco m o cococoiiOcocoTj-cn •-r-CMCMOIOi-O^TfT-T- UO 'Cf -^ i- o CO CM couocoLtocbcb^tr^^Tr^r^o) D-IDJDJ3JDJD CO rf O CO CO CO co I 5 I co i o V co "" UJ LUCn^Tt-i^-CLco^ •^ uo cb oS o CO 3 C *■ c a; u c L. t X 0) o CO T TJ" O O CO CO COMO TT i- i- CM CO CO ^f CO T -m CD S CO CO O i- m co m cd co co co in CD CO CD s incbcbcbc6c6T-m 4 4- co ri-WCMlOrT- co in cq cm co ^ ^ co m ^r 3JDJDJDJDJDJDJ co r- o m ll in ^ cb I O , m cm uj r- uj cm ^r ^ O) CO ^» ■gCDOJCO^CO "^ " i o V co '" Tl; CL Tt > ta CD cb in m m co co co co tj- t m T-T-cOi-cDTjinto o cq co co CMCo^m'tTt^'io DJDJDJ r- co co r- LLcoajcb-^o cd O co i O "" UJ in O) CD CO •C co O 3 C C co i; £ 8 o CO <0 X a £ L. ii c o 'x a> T3 £ £ c o 'x 1) s- Oh CO Cfl be cs .3 as D O 49 FISHERY BULLETIN: VOL. 80, NO. 1 Figure 5.—Citharickthys cornutus: A. Transforming larva, 14.2 mm; B. Nearly transformed larva, 17.4 mm; C. Adult, 37.2 mm. Scale = 1 mm. Surface temperature and salinity were 20.4°- 27.3°C and 35.5-36.8"/... Almost no larvae were caught east of the average Gulf Stream axis (Fig. 1). The reported northern limit for adults is Flor- ida (with one exception— an adult male taken off Cape Hatteras (Stewart 5 )). Larval occurrences shown in Figure 1 are evidence of the effective- 5 D. J. Stewart, Graduate Student, Laboratory of Limnology, University of Wisconsin, Madison, WI 53706, pers. commun. June 1978. ness of Gulf Stream transport. The eastward shift of positive tows just north of lat. 32°N cor- responds to the location of a semipermanent meander of the Gulf Stream induced by the Charleston Rise (at about lat. 32°N, long. 79°W (Pietrafesa et al. 1978)). In the eastern Gulf of Mexico, larvae smaller than 4 mm NL were common in January, Febru- ary, May, June, July, August, and November, indicating year-round spawning in that area (Dowd 1978). 50 TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND ETROPUS Citharichthys gymnorhinus (Figs. 1, 6, 7) Identification Larvae approaching transformation had com- plete complements of countable characters. Those specimens were identified by comparing the following larval counts with known adult counts. Number of specimens is given in paren- theses. Caudal fin formula = 4-5-4-4 (15) Caudal vertebrae = 23(3)-24(18) Gill rakers (lower limb, first left) = ~12 (1) Left pelvic rays = 5 (12) Anal rays = 55-59 (11) Dorsal rays = 70-75 (11) Of the potential species listed in Table 1, only C. gymnorhinus has counts that agree with these (it is unique in having only five left pelvic rays). In addition, larvae were captured over the outer shelf, but not as far offshore as C. cornutus (Fig. 1). This is consistent with bathymetric distribu- tion of adults. Distinguishing Characters Citharichthys gymnorhinus larvae have no pectoral melanophore, and notochordal pigment is restricted to the caudal region. Three elongate dorsal rays are present from preflexion (4.6 mm) through postflexion (probably through transfor- mation). Caudal vertebrae (23-24) can be counted by early flexion (6 mm). Lateral pigment is rela- tively sparse except for the caudal band. Flexion is complete at 7-8 mm SL. Morphology is similar to that of C. cornutus. However, the left pelvic fin of C. gymnorhinus has a full complement of only five rays, and in larvae the first ray is much re- duced in size compared with that of C. cornutus. Length of C. gymnorhinus at transformation is probably about 18 mm. Larvae may appear in collections year-round. Pigmentation Pigmentation of C. gymnorhinus larvae is moderate. Gas bladder and caudal band pigment are the most striking. By 4.6 mm and throughout larval develop- ment, the dorsal one-third of the left side of the gas bladder is fairly heavily pigmented, usually with distinct melanophores. With growth, the number of melanophores increases. There are usually more of them than in C. cornutus larvae. The maximum number in a preflexion specimen was about 15 (4.6 mm, Fig. 6A). The right side of the gas bladder is either unpigmented or has only one or two melanophores. By 4.6 mm (Fig. 6A) a caudal band of melano- phores is present on the dorsal and ventral fin- folds and sides and margins of the body about halfway from the anus to the notochord tip. This band is more distinct and regular than in other 51 FISHERY BULLETIN: VOL. 80, NO. 1 Figure 6.-Larval stages of Citharichthys gymnorhin- ws: A. Preflexion, 4.6 mm; B. Late flexion, 6.7 mm. Scale = 1 mm. 52 TUCKER: LARVAL DEVELOPMENT OF C1THARICHTHYS AND ETROPUS Figure 7.— Larval stages of Citharichthys gymnorhinus: A. Transforming, 9.6 mm; B. Transforming, 12.6 mm. Scale = 1 mm. 53 FISHERY BULLETIN: VOL. 80, NO. 1 known larvae of western North Atlantic Cithar- ichthys and Etropus species. In preflexion lar- vae, before pelvic rays form, one or two melano- phores are present on the ventral body margin at the future site of the pelvic fin. By 4.6 mm and throughout development (at least to 13 mm), a few external melanophores are present along the posterior surface of the gut loop. A small melano- phore is found over the posterodorsal surface of the midgut of preflexion larvae. Flexion larvae (Fig. 6B) usually have 15-20 melanophores on the dorsal one-third of the left side of the gas bladder. The caudal band is mostly confined to the body and contains myo- septal pigment. Visible internal notochordal pigment is restricted to the vicinity of the exter- nal caudal band. The dorsal surfaces of one or two forming centra are darkened at about 5 mm. By about 6 mm and throughout development (at least to 13 mm), there may be a few melano- phores along the ventral surface of the gut loop. By late flexion, notochordal pigment appears as fine dashes along four to six centra of caudal ver- tebrae 13-19. By late flexion, pigment along the posterodorsal surface of the midgut extends to the gas bladder and appears as a black lining over the gut. By postflexion (about 8 mm, Fig. 7A), both sides of the gas bladder usually are obscured by body musculature, and pigment in this area ap- pears diffuse. Small melanophores appear on the left pelvic fin membrane along both sides of the elongate second ray. Body musculature tends to obscure notochordal pigment in larvae longer than 12 mm. Morphology (Figs. 6, 7; Tables 3, 6) General morphological features are similar to those of C. comutus, with the qualification that the smallest C. gymnorhinus specimen examined was 4.6 mm NL. Adult morphometries given in the following discussion were derived from Gutherz and Blackman (1970) and Topp and Hoff (1972). The mouth is relatively large in larvae and adults. Larval upper jaw length/BL is fairly con- stant at 9.3-9.5%. Adult upper jaw length/BL is 11.2%, range 9.9-13.0%. Larval upper jaw length/ HL decreases greatly from 38% (preflexion) to 33% (flexion and postflexion). Adult upper jaw length/HL is 41%, range 39-45%. Larval lower jaw length/BL increases from 11.5% to 12.7%. Adult lower jaw length/BL is 13.2%, range 11.6- 14.8%. Larval lower jaw length/HL decreases slightly from 46% to 44% and is only slightly less than that of C. comutus. Adult lower jaw length/ HL is 48%, range 43-53%. The larval snout is pointed but relatively short. Larval snout length/BL increases slightly from 5.2% to 6.1%. Adult snout length/BL is 5.4%, range 4.6-6.6%. Larval snout length/HL is con- stant at 21%. Adult snout length/HL is 20%, range about 18-20%. The eye is relatively large in larvae and adults (only slightly smaller than that of C. comutus). Larval eye diameter/BL is constant at about 8.8%. Adult orbit length/BL is 9.6%, range 8.0- 11.4% (Topp and Hoff 1972); eye diameter/BL is 10.1%, range 9.1-11.0% (Gutherz and Blackman 1970). Larval eye diameter/HL decreases from 35% to 31% and is similar to that of C. comutus. Adult orbit length/HL is 35% (Topp and Hoff 1972); eye diameter/HL is 36.5%, range 33-38% (Gutherz and Blackman 1970). The head is fairly long but shallow in larvae and of moderate length in adults. Larval head length/BL increases greatly from 25% to 29%. Postflexion head length/BL is similar to those of C. arctifrons and C. comutus. Adult head length/ BL is 27%, range 25-29%. Larval head depth/BL increases from 29% to 33% and is similar to that of E. crossotus. Larval snout to anus length is fairly great until postflexion. Snout to anus length/BL increases slightly from 43% (preflexion) to 44% (flexion) and then decreases to 40% (postflexion). This length is similar to that of E. crossotus during flexion and postflexion. With the exception of a relatively deep caudal peduncle, the body is of moderate depth in larvae and adults. Larval body depth at pelvic fin/BL increases from 30% to 39%. Larval body depth at loop of gut/BL increases from 29% to 44%. Larval body depth at anus/BL increases greatly from 25% to 42% and during flexion and postflexion is similar to that of E. crossotus. Larval body depth at third hemal spine/BL increases greatly from 14% to 35%. Adult body depth/BL is 47%, range 39-50%. Larval caudal peduncle depth/BL in- creases from 11.6% (flexion) to 13.2% (postflex- ion). Adult caudal peduncle depth/BL is 11.5%, range 10.5-12.6%. Fin and Axial Skeleton Formation Caudal skeleton development is similar to that of C. comutus. Size ranges of available speci- 54 TUCKER: LARVAL DEVELOPMENT OF CITHAKICHTHYS AND ETROPUS Table 6.— Measurements (mm) of larvae of Citharickthys gymnorhinus. Pref = preflexion, ECF = early caudal formation, Early = early flexion, Mid = midflexion, Late = late flexion, Post = postflexion. S = sym- metrical, 1 = to one-third of the way to the dorsal ridge, 2 = one-third to two-thirds of the way to the dorsal ridge, 3 = two-thirds to all the way to the dorsal ridge. sz Dl c a> >- "D O CD sz Ol c a S ro CD a. a. Z> sz Ol c 0) 5 . LU SZ Ol c 0) ■D CO CD I c/) C CO 0£ — Ol zs c O CD c — CO SZ o> c a> ~m o sz a 01 ■o ■o CO CD I ro ft CO « — So, a.— CD o 13 a. >- o "O o o — CO to C CD ro sz Q. CD ■o >. ■o o CO CD c Q. (/I £ E Q. CD CD-C ■o-o >i n sz o *" CO V c •o CD Q. _ SZ Sa Zl > u cS H CO •a c E be s- o ■a «4-H O e be !- o £ 3 O E l* , 1 eci 0) ii >> u a> cfl +j be X s-, £ o £ c EC o CO -»-> crt hn CU E c (!) C > 3 1-3 o d -C en cS > I* cS ^^ £ E OJ OS «*-c o X) CS 0) ffi §■ a. J 1 u as D O 60 TUCKER: LARVAL DEVELOPMENT OF C1THARICHTHYS AND ETROPUS Table 7.— Measurements (mm) of larvae and juveniles of Citharichthys sptiopterus. ECF = early caudal formation, Late = late flexion, Post = postflexion. S = symmetrical, 1 = to one- third of the way to the dorsal ridge, 2 = one-third to two-thirds of the way to the dorsal ridge, 3 = two-thirds to all the way to the dorsal ridge, R = on the dorsal ridge, T = transformed. O) c V >> T3 O CD O) c 0) 3 .5. o. a D Ol c a> 5 re a> 3 o _i O) c o c CO a; ai E 5 T3 ai >. LU O) c Q) ■a re I C re 0£ — o> => c O I re £ c is ft CO C re re Q. 0) TD >. ■D O m CD c a. en «re £ E a. a; a>-<= ■°-o >*— ■D r. o ~ m o c D CD a It re-° O a> O) re c o X c o '55 o a a> >. CD .c n 3.7 0.37 0.45 028 0.36 1.0 1.5 3.8 1.4 '1.5 '1.4 '084 ECF s 5.7 0.42 0.58 0.46 0.46 1.5 22 2.4 2.7 2.9 2.2 074 Late s 6.7 042 0.61 0.53 0.51 1.7 25 8.5 2.5 3.1 3.3 2.6 093 Late s 68 0.54 0.71 046 0.54 1.8 2.7 2.6 3.2 3.5 3.0 1.1 Late 1 8.3 066 080 0.55 0.53 20 2.7 10.4 2.8 3.4 3.7 32 1.2 Post 3 8.7 0.66 88 0.53 0.54 2.3 2.8 11.3 3.0 3.5 3.8 3.4 1.3 Post T 8.9 0.54 0.83 0.50 0.52 2.0 3.0 11.1 3.0 3.6 4.0 3.5 1.2 Post 1 8.9 0.61 0.82 0.61 0.56 2.1 29 11.0 3.0 3.6 4.0 3.4 12 Post 2 89 065 0.85 0.50 0.63 2.2 2.9 11.1 3.1 3.5 3.8 3.4 1.3 Post R 90 0.73 1.0 0.58 0.58 2.4 2.9 11.7 3.1 3.5 3.6 3.4 1.2 Post T 9.0 0.62 0.81 0.60 0.58 22 2.8 110 3.0 3.6 4.0 3.6 1.2 Post s 9.1 0.58 0.80 0.56 0.60 2.1 2.9 11.3 3.0 3.4 36 3.3 1.2 Post 3 91 0.70 0.90 0.65 060 2.4 3.1 112 29 3.3 3.4 3.1 1.2 Post T 9.1 0.67 0.93 056 0.60 23 3.0 11.2 29 3.3 3.5 3.3 1.3 Post T 9.2 0.70 093 058 0.70 2.4 2.8 11.4 32 3.7 3.9 3.6 1.3 Post R 9.2 0.64 0.97 046 0.64 2.3 2.7 11.4 2.9 3.6 3.8 3.4 1.3 Post T 92 0.67 0.87 0.60 062 2.4 2.8 114 3.0 3.6 3.8 3.5 1.3 Post T 9.2 0.68 089 0.62 060 2.3 2.8 11.2 2.9 34 39 3.2 1.3 Post R 93 0.57 080 056 060 22 2.9 11.4 30 3.6 3.9 34 1.2 Post 1 93 0.67 090 065 063 24 2.9 11.7 3.0 34 3.8 3.7 1.3 Post T 9.4 068 093 0.53 068 2.4 2.7 11.7 3.0 3.5 3.7 3.4 1.3 Post T 9.4 069 1.0 056 068 23 3.0 11.6 3.1 35 3.7 38 1.4 Post T 9.4 0.67 094 059 0.67 24 29 11.7 29 3.4 3.7 3.3 1.3 Post T 95 071 1.0 060 0.68 2.4 3.1 118 3.0 3.6 3.8 3.4 1.3 Post T 96 066 0.91 0.71 0.53 2.3 3.1 12.0 3.3 3.8 4.1 38 1.4 Post R 9.7 069 0.93 0.64 0.68 2.5 2.9 11.8 3.2 3.6 3.9 3.6 1.3 Post R 9.7 0.61 0.80 0.70 0.56 2.2 3.3 11.9 3.2 3.9 4.3 3.7 1.2 Post S 9.8 0.65 089 0.58 0.74 2.4 3.2 11.9 3.1 3.8 4.1 36 1.2 Post 1 9.8 0.67 0.87 0.70 064 2.4 3.2 119 3.1 3.8 44 38 1.2 Post 1 10.0 0.72 1.0 054 0.72 2.5 3.0 12.2 3.2 3.6 3.8 3.6 1.3 Post T 10.0 0.76 1.0 0.67 0.73 2.6 3.2 12.5 3.2 3.8 3 9 37 1.4 Post T 10.0 1.1 0.73 2.5 3.0 12.2 3.2 3.7 4.0 3.6 1.3 Post T 102 068 0.90 0.70 0.67 2.4 3.2 12.4 3.2 39 4.2 3.8 1.3 Post 3 102 0.71 1.0 056 064 2.6 3 1 125 3.1 3.5 3.7 3.6 1.4 Post T 10.2 0.78 1.1 0.58 0.75 2.5 32 12.7 3.4 39 4.2 39 1.4 Post T 10.2 0.73 1.1 0.53 0.76 26 3.2 125 33 38 4.0 3.8 1.4 Post T 10.2 1.0 060 0.70 2.4 3.2 12.2 33 3.9 4.1 39 1.4 Post T 10.3 0.83 1.1 0.64 064 2.6 32 12.7 3.4 3.8 4.1 3.8 1.4 Post T 10.3 0.79 1.1 060 0.74 2.7 33 13.0 3.4 3.9 4.2 3.9 1.4 Post T 10.5 0.76 1.1 0.60 0.71 2.6 3.4 13.0 3.4 3.8 4.2 4.0 1.4 Post T 10.5 0.59 0.92 060 066 2.4 3.2 12.9 3.5 3.9 4.4 4.2 1.4 Post R 10.6 0.73 1.1 0.63 068 2.7 32 13.1 3.4 4.0 4.2 4,0 1.4 Post T 10.6 0.65 1.1 050 0.67 2.6 3.2 13.1 3.5 4.0 4.3 4.1 1.4 Post T 10.6 060 090 064 60 2.5 3.2 13.0 3.4 4.1 4.5 4.1 1.4 Post 1 10.7 0.74 0.95 0.47 0.77 2.6 3.2 13.0 3.2 3.8 4.1 3.9 1.4 Post T 10.8 0.82 1.1 0.65 0.72 2.7 3.3 13.3 3.2 39 4.2 3.8 1.4 Post T 10.9 0.76 1.1 0.55 066 2.7 3.1 13.4 3.4 4.0 4.3 4.0 1.4 Post T 10.9 0.75 1.1 0.62 0.67 2.8 3.2 13.3 3.3 4.0 4.3 4.0 1.4 Post T 11.0 0.81 1.1 0.61 0.65 27 3.2 3.4 3.8 4.0 39 1.3 Post T 11.6 0.84 1.3 0.61 089 3.1 3.8 143 3.7 4.2 4.5 4.4 1.5 Post T 143 1.3 1.9 0.77 1.2 3.8 4.9 18.3 4.6 5.2 5.6 5.6 1.6 Post T 16.6 1.6 2.3 0.91 1.3 4.4 5.3 21 5.5 5.9 6.6 7.0 1.7 Post T 167 1.5 2.2 0.75 1.1 4.0 5.0 5.2 60 64 6.6 1.9 Post T 19.4 1.7 2.5 084 1.3 48 6.0 24.1 5.8 6.7 7.5 7.8 2.2 Post T 220 2.1 3.0 1.2 14 54 6.8 27.9 69 7.6 8.6 90 2 7 Post T 23.1 1.9 29 1 1 1.6 5.6 7.1 285 7.0 7.7 8.6 90 2.7 Post T 23.3 2.1 3.1 1.2 1.6 5.8 72 289 7.2 84 92 94 2.8 Post T 24.0 2.2 3.1 1.1 1.6 60 7.8 30.0 7.4 8.3 92 97 2 6 Post T 254 2.3 3.2 1.3 1.6 6.2 8 1 31.9 7.5 89 9.7 3.0 Post T 'Measurement does not include dorsal or anal pteryglophores. Fin and Axial Skeleton Formation Caudal skeleton development apparently is similar to that of C. cornutus. Size ranges of available larval specimens in each stage are as follows: Early caudal formation, 3.7 mm NL; late flexion, 5.7-6.8 mm NL; postflexion, 9.0-10.6 mm SL. All caudal rays are calcified by 5.7 mm. All precaudal neural spines, the first 13 caudal neural spines, the first 13 hemal spines, and no 61 FISHERY BULLETIN: VOL. 80. NO. 1 precaudal or caudal centra stain with alizarin at 3.7 mm NL. All neural spines and hemal spines, some precaudal centra, and the urostyle stain with alizarin at 5.7 mm NL. All precaudal and caudal centra stain with alizarin at 6.7 mm NL. The smallest specimen in which caudal centra could be counted was 5.7 mm NL (late flexion). The second and third dorsal rays are moder- ately elongate and moderately separated at the bases from preflexion (3.7 mm NL) through transformation (about 10 mm SL). No other dor- sal rays were formed at 3.7 mm, but adult counts were present from 5.7 mm NL onward. All dor- sal rays had calcified by postflexion (8.3 mm SL). No anal rays were formed at 3.7 mm, but adult counts were present from 5.7 mm onward. All anal rays had calcified by 8.3 mm. The second left pelvic ray is formed by 3.7 mm; in larger specimens it is elongate. By 5.7 mm, four left and four right pelvic rays are calcified. All six rays in each fin are calcified by 6.8 mm NL. Rayless, fanlike, larval pectoral fins were present in the smallest specimen (3.7 mm). Calci- fication of rays in the left fin begins during post- flexion (9-10 mm SL), and is complete by the end of transformation (9-11 mm SL). Cephalic Spination Preopercular spines (Table 4) were present from early caudal formation (3.7 mm NL, Fig. 8A) through transformation (10.2 mm SL). Maximum numbers may be reached during or before late flexion (31 left, about 36 right); how- ever, counts from early and midflexion larvae are lacking and those from older ones are highly variable. No preopercular spines are evident in juveniles. The 3.7 mm NL specimen had one frontal- sphenotic spine on each side. Several postflexion (8-10 mm SL) specimens had one or two rela- tively inconspicuous frontal-sphenotic spines on each side. These spines may be more numerous in larvae smaller than 5.7 mm NL. None are evi- dent in juveniles. Larval Teeth (Table 5) The early caudal formation (3.7 mm NL, Fig. 8A) specimen had two upper and three lower teeth on each side. During late flexion and post- flexion (5.7-10.6 mm BL), larvae usually have four upper and five lower teeth on each side. Transforming larvae and early juveniles (9.1- 10.7 mm SL) usually have about five upper left (probably about five upper right) and five or six lower left (probably five to eight lower right) teeth. Transformation Migration of the right eye may begin as early as late flexion (6.8 mm NL) or as late as postflex- ion (10.6 mm SL). The right eye moves from the right side of the head around the dorsal fin origin (Fig. 9 A) as in Citharichthys arctifrons and Etropus microstovnus (Richardson and Joseph 1973). The right eye reaches its final position on the left side of the head at about 9-11 mm SL. Occurrence Larvae were collected from September through December in the Gulf of Mexico off Texas (Daher 6 ) and from October through April in the Cape Fear River estuary, North Carolina (pers. obs.). Temperature and salinity ranges at capture locations in the Cape Fear River were 4.1°-26.6°C and 0.0-31.77... Etropus crossotus (Figs. 10, 11) Identification Larvae approaching transformation had com- plete complements of countable characters. Those specimens were identified by comparing the following larval counts with known adult counts. Number of specimens is given in paren- theses. Caudal fin formula = 4-5-4-4 (15) Caudal vertebrae = 24(1), 25(19), 26(3) Gill rakers (lower limb, first left) = ~7 (1) Left pelvic rays = 6 (11) Anal rays = 60-66 (13) Dorsal rays = 76-84 (13) Of the potential species listed in Table 1, only E. crossotus has counts that agree with these. In addition, most specimens were captured west of the Mississippi River in the Gulf of Mexico, an 6 M. A. Daher, Graduate Student, Department of Wildlife Science, Texas A&M University, College Station, TX 77843. pers. commun. June 1978. 62 TUCKER: LARVAL DEVELOPMENT OF C ITH A RICHTHYS AND ETROPUS area from which other Etropus spp. have not been reported. Distinguishing Characters Etropus crossotus larvae have a dashlike melanophore at the base of each pectoral fin. In- ternal pigment along the dorsal surface of the notochord is extensive. Two elongate dorsal rays are present from preflexion (4.6 mm) through transformation. Caudal vertebrae (25-26, very rarely 24) can be counted by midflexion (5.4 mm). Lateral pigment is relatively heavy. Flex- ion is complete at 9-10 mm SL. The larval mouth and eye are small. The left pelvic fin has a full complement of six rays. Length at transforma- tion is 10-12 mm. Larvae usually appear in col- lections from March through August. Pigmentation Pigmentation of E. crossotus larvae is rela- tively heavy. Pigment on the gas bladder and on the ventral and dorsal surfaces of the body is the most striking. Most useful for identification is in- ternal pigment along the dorsal surface of the notochord and a melanophore at the base of the pectoral fin. By about 4.6 mm (Fig. 10A) and throughout larval development, the dorsal one-third of the Figure 10.— Larval stages of Etropus crossotus: A. Preflexion (early caudal formation), 4.6 mm; B. Midflexion, 6.0 mm. Scale = 1 mm. 63 left side of the gas bladder is fairly heavily pig- mented, usually with three or four distinct mel- anophores. The right side of the gas bladder is similarly pigmented until late flexion. Internal notochordal pigment consists of a series of fine dashes along the dorsal surface and is more ex- tensive then in known Citharichthys larvae. Pre- f lex ion and early flexion larvae have up to about 12 pigment dashes between the gas bladder and caudal centrum 15. From about caudal centra 15 to 18 (range 14-20) there is a distinct series of heavy dashes which usually form a nearly solid line throughout development. An internal mel- anophore that appears to be associated with the notochord is located below the hindbrain near the otic capsule, where the notochord joins the brain. Dashlike clusters of pigment develop along the dorsal and ventral body margins be- tween the pectoral fin and the caudal fin base. These clusters have not completely formed in the preflexion specimen, but three dorsal clusters and ventral pigment are present. During pre- flexion, a melanophore may be present on the ventral edge of the caudal finfold, opposite the midpoint of incipient hypural bones. Throughout larval development, a continuous or broken line of pigment (the length of three to five centra) is on the lateral midline about two- thirds of the way from the anus to the notochord tip. One or two melanophores are on each side of the symphysis of the lower jaw. The posterior margin of the articular is covered with a stellate melanophore. A stellate melanophore is present at the junction of left and right branchiostegal membranes, just forward of the isthmus. About one to three internal melanophores are present near the pectoral fin base and just forward of the cleithrum beneath the angle of the last gill arch (visible through the opercle). Usually, a melano- phore is on the anterodorsal edge of the urohyal. The ventral body margin between the isthmus and pelvic fin is fairly heavily pigmented with a few distinct melanophores or a continuous band of pigment. Several melanophores are present along the ventral and lateral surfaces of the abdomen and sometimes along the hindgut near the anus. The lower edge of both pectoral fin bases is lined with a dashlike melanophore. The second left pelvic ray has melanophores along its distal end. A series of small melanophores is pres- ent along the distal tips of anal pterygiophores. During early flexion (4.9 mm), one, or rarely two, diffuse internal melanophores appear above the hindbrain. 64 FISHERY BULLETIN: VOL. 80, NO. 1 During midflexion (5-6 mm, Fig. 10B), mela- nophores appear along the distal ends of the elon- gate dorsal rays. A group of melanophores may be present at the distal ends of the middle anal rays. Melanophores begin appearing at the bases and along the sides of middle caudal rays. During flexion, internal notochordal pigment increases. Midflexion and late flexion larvae have up to about 5 pigment dashes between the cleithrum and gas bladder and up to about 18 dashes between the gas bladder and caudal cen- trum 15. From midflexion through postflexion, a small amount of pigment usually is on the antero- ventral edge of the maxillary. By late flexion (6 mm), the gas bladder has be- come oriented toward the left side, and greater development of musculature obscures the gas bladder from the right side. By about 8.5 mm, musculature begins to obscure notochordal pig- ment, except for the heavy dashes in the caudal band area. There is no evidence of a melanophore on the opercle; however, one or two small melano- phores occasionally appear on the interopercle during late flexion. By about 8.5 mm, concentra- tions of pigment have formed around the first through third left pelvic rays. Pigment at the dis- tal margin of the right pelvic fin appears at about the same time. During postflexion (10.5 mm, Fig. 11A), a small melanophore appears on the upper lip. Groups of melanophores are present along the margins of dorsal and anal fins of some speci- mens. In the nearly transformed specimen (10.3 mm, Fig. 11B), heavy posterior notochordal pigment is still obvious. Additional internal melano- phores have appeared posterior to the hindbrain. Myoseptal pigment is well developed, mostly adjacent to dorsal and anal pigment clusters. As in Citharichthys larvae, this forms a caudal band. A midlateral cluster of melanophores is present near the caudal fin. Melanophores have formed along the anterior surface of the head from the snout to the dorsal fin. External and in- ternal melanophores are present along the hind- gut. Melanophores have formed along the proxi- mal ends of groups of some dorsal and anal rays. Morphology (Figs. 10, 11; Tables 3, 8) General morphological features are similar to those of Citharichthys cornutus, with the qualifi- cation that the smallest E. crossotus specimen examined was 4.6 mm NL. Adult morphomet- TUCKER: LARVAL DEVELOPMENT OF C1THAR1CHTHYS AND ETROPUS Figure 11.— Larval stages of Etropus crossotus: A. Transforming, 10.5 mm; B. Nearly transformed, 10.3 mm. Scale = 1 mm. rics given in the following discussion are from Gutherz (1967). The larval mouth is relatively small. During flexion and postflexion relative mouth size is similar to that of C. spilopterus. The adult mouth is the smallest of known western North Atlantic Etropus and Citharichthys species. Larval upper jaw length/BL is fairly constant at 7.0-7.2%. Lar- val upper jaw length/HL decreases from 30% to 27% (preflexion to postflexion). Adult upper jaw length/HL is 21-27%. Larval lower jaw length/ BL is fairly constant at 9.6-9.8%. Larval lower 65 FISHERY BULLETIN: VOL. 80, NO. 1 Table 8.— Measurements (mm) of larvae and a juvenile of Etropuscrossotus. Pref = preflexion, ECF = early caudal formation, Early = early flexion, Mid = midflexion, Late = late flexion, Post = postflexion. S= symmetrical, 1 =0 to one-third of the way to the dorsal ridge, 2 = one-third to two-thirds of the way to the dorsal ridge, 3 = two-thirds to all the way to the dorsal ridge, T = nearly transformed. c 0) c O) c CD to c a to CD c CD IS E CO O) c 3 ■D to C CO o.c c Q. CO ■D to CO c CO CO .c a CO x> CO c a. CO £ E Q. <» co-c CO o c ■o CO a _ .c 5°- CO a> CO "to c o c o CO o a to >. CO o Q. Q. 3 o o c CD CO CO O CO c — CO o CO CO B& o ■a .c o -* d <» CO" CO ai CD 3 _i oo LU I CO r- I co CO CO O u. rr 4.6 0.32 044 0.24 0.34 1.1 1.8 4.6 1.3 '1.2 1 098 1 053 ECF S 4.9 0.35 0.43 0.27 0.37 1.2 2.1 1.6 1.6 1.4 089 0.23 Early S 5.4 0.38 0.50 0.32 0.39 1.4 24 5.6 1.7 1.9 1.7 099 0.27 Early S 54 0.43 066 0.32 041 1.4 2.4 6.0 1.9 2.1 1.9 1.3 0.48 Mid S 5.5 0.40 0.53 0.40 0.36 1.4 2.5 5.8 1.9 2.0 1.8 1.3 0.43 Mid S 5.5 0.37 0.52 0.35 040 1.5 2.6 6.0 1.9 2.0 1.8 1.2 0.43 Mid S 5.6 043 0.57 0.33 0.40 1.4 2.4 5.9 1.8 2.1 1.9 1.3 0.37 Mid s 5.7 0.40 0.51 0.33 0.37 1.4 2.4 5.7 1.7 1.9 1.7 1.1 0.31 Mid s 5.7 0.35 0.49 0.29 0.35 1.4 2.4 6.0 1.8 1.9 1.8 1.2 0.35 Mid s 5.8 0.34 0.49 0.35 0.38 1.4 1.8 1.9 1.2 0.34 Mid s 6.0 0.43 0.56 0.40 0.42 1.6 2.7 65 2.0 2.2 2.1 1.4 0.43 Mid s 6.0 0.37 0.51 0.32 041 1.5 2.8 6.6 2.0 3.0 18 1.3 0.52 Mid s 6.0 0.50 0.62 0.42 0.43 1.6 2.8 68 2.1 2.4 2.2 1.6 0.53 Mid s 6.1 0.51 0.72 0.37 0.47 1.7 2.6 2.1 2.3 2.3 1.6 068 Late S 62 0.41 0.57 0.38 0.41 1.6 3.0 7.2 2.2 2.5 2.4 1.8 0.63 Late s 6.9 0.50 0.67 0.47 0.46 1.9 3.2 8.2 2.3 2.8 2.6 2.0 0.73 Late S 7.4 0.54 0.74 0.53 0.49 20 3.3 8.8 2.5 3.1 3.0 2.5 0.83 Late 82 049 0.70 0.56 0.50 2.3 38 10.1 2.8 3.5 3.7 3.0 1.1 Late 8.3 0.53 0.71 0.53 0.54 2.2 3.6 10.1 2.8 3.3 3.5 2.8 1.0 Late s 8.3 063 0.78 0.53 0.54 22 3.6 10.1 2.9 3.5 3.8 3.2 1.1 Late 8.5 0.69 084 0.63 0.58 2.4 3.6 10.8 3.0 3.6 3.7 3.1 1.1 Late 9.1 0.70 092 0.57 0.61 24 3.6 11.4 3.0 3.7 39 3.4 1.2 Late 93 0.73 0.96 0.70 0.67 2.7 3.9 3.4 4.0 4.1 3.5 1.2 Late 9.3 0.64 0.84 0.63 0.65 26 4.0 11.9 3.4 38 4.1 3.6 1.2 Late 9.3 0.60 0.86 0.67 0.57 2.4 3.9 11.2 3.0 3.8 4.0 3.4 1.1 Post 9.5 0.71 090 0.70 0.63 2.6 4.0 11.9 3.3 4.0 4.3 3.7 1.3 Late 9.6 0.72 096 0.73 0.61 26 4.1 11.9 3.4 4.2 4.4 39 1.3 Post 10.3 0.80 1.1 0.75 071 3.0 3.4 127 33 3.7 3.7 3.7 1.2 Post 10.5 0.75 0.99 0.63 0.66 2.7 4.0 12.9 3.5 4.3 4.8 4,0 1.3 Post 2 10.5 0.76 1.1 080 0.64 28 4.1 12.9 3.5 4.2 4.6 4.0 1.3 Post 3 10.8 0.71 1.0 0.59 2.6 4.1 13.2 3.4 4.3 4.7 4,1 1.4 Post 1 'Measurement does not inc ude do rsal or anal pt srvaioD lores. jaw length/HL decreases greatly from 41%to 36- 37%. The larval snout is moderate but exhibits a relatively fast growth rate. Snout length/BL in- creases from 5.2% to 6.8%. Snout length/HL increases from 22% to 26%. The eye is relatively small in larvae and mod- erate in adults. Larval eye diameter/BL de- creases from 7.4% to 6.3%. Larval eye diameter/ HL decreases greatly from 32% to 24%. Adult eye diameter/HL is about 22-28%. The larval head is of moderate length but rela- tively shallow depth. In adults, the head is the shortest of known western North Atlantic Etro- pus and Citharichthys species. Larval head length/BL increases from 23% to 26%. Adult head length/BL is 20-25%. Larval head depth/BL increases from 29% to 33-34% and is similar to that of C. gymnorhinns. Larval snout to anus length is moderate. Snout 66 to anus length/BL increases from 39% (pref lex- ion) to 44% (flexion) and then decreases to 39% (postflexion). This length is similar to that of C. gymnorhinus during flexion and postflexion. Early larvae are relatively shallow, but ab- dominal and tail depths increase quickly, and as adults, this species and E. rimosus are the deep- est bodied of known western North Atlantic Etropus and Citharichthys species. During post- flexion, the dorsal and ventral profiles of E. cros- sotus are relatively convex. Larval body depth at pelvic fin/BL increases greatly from 26% to 40%. Larval body depth at anus/BL increases greatly from 21% to 43% and is similar to that of C. gym- norhinus during flexion and postflexion. Larval body depth at third hemal spine/BL increases greatly from 12% to 38%. Adult body depth/BL is 50-58%. Larval caudal peduncle depth/BL in- creases from 9.6% (flexion) to 12.6% postflex- ion). TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND ETROPUS Fin and Axial Skeleton Formation Caudal skeleton development is similar to that of C. cornutus. Size ranges of available speci- mens in each stage are as follows: Early caudal formation, 4.6 mm NL; early flexion, 4.9-5.4 mm NL; midflexion, 5.4-6.0 mm NL; late flexion, 6. 1- 9.5 mm NL; postflexion, 9.3-10.8 mm SL. Caudal rays become calcified between early flexion (4.9 mm NL) and late flexion (about 6.5 mm NL). All precaudal neural spines stain with alizarin at 4.6 mm NL. Some caudal neural spines and hemal spines stain with alizarin at 4.6 mm NL, and all do by 5.6 mm NL. All precaudal and cau- dal centra stain with alizarin at about 6.0 mm NL. The urostyle stains with alizarin at 6.2 mm NL. The smallest specimen in which caudal centra could be counted was 5.4 mm NL (midflexion). The second and third dorsal rays are elongate and moderately separated at the bases from pre- flexion (4.6 mm NL) through transformation (about 11 mm SL). During early flexion (4.9 mm NL), rays near the middle of the fin begin to cal- cify. Calcification proceeds anteriorly and pos- teriorly. Adult counts are present from late flex- ion (about 8.0 mm NL) onward. The first ray and most posterior rays are calcified prior to trans- formation (by about 9.6 mm SL). During early flexion (4.9 mm NL), anal rays near the middle of the fin begin to calcify. Calci- fication proceeds anteriorly and posteriorly. Adult counts are present from late flexion (about 8.0 mm NL) onward. The most posterior rays are calcified during late flexion (about 9.3 mm NL). Development of the left pelvic fin precedes that of the right. The left pelvic fin bud appears dur- ing preflexion (before 4.6 mm NL). Rays develop between early caudal formation (4.6 mm NL) and late flexion (8.5 mm NL). The second ray is the first to appear; it is elongate. The first ray appears soon after the second; it may be slightly elongate. The right pelvic fin bud appears dur- ing midflexion (5.5 mm NL). Rays develop between midflexion (5.8 mm NL) and late flexion (8.5 mm NL). Each complete fin has six rays. Rayless, fanlike, larval pectoral fins are present on the smallest available specimen (4.6 mm NL). Calcification of rays in the left fin oc- curs during late transformation (10-11 mm SL). Fig. 10A). With development (Fig. 10B), addi- tional spines appear until maximum numbers of about 24 on the left (range 18-29) and 22 on the right (range 16-27) are reached during midflex- ion (5.4-6.0 mm NL). Thereafter, spines are lost until none or only a few remain at transforma- tion (Fig. 11B). Most specimens had three or four relatively inconspicuous frontal-sphenotic spines on each side, including one or two that were noticeably stronger. Larval Teeth (Table 5) The early caudal formation specimen (4.6 mm NL, Fig. 10A) had three upper and five lower teeth on each side. Early flexion larvae (4.9-5.4 mm NL) have three upper and five or six lower teeth on each side. During midflexion (5.4-6.0 mm NL), there are three to five upper and five to seven lower teeth on each side. During late flex- ion (6.1-9.5 mm NL), larvae usually have four upper and seven lower teeth on each side. During postflexion (9.3-10.8 mm SL), there are usually four or five upper and seven lower teeth on each side. The nearly transformed specimen ( 10.3 mm SL, Fig. 11B) had seven upper and more than nine lower teeth on each side. Transformation Migration of the right eye may begin as early as late flexion (7.4 mm NL) or as late as postflex- ion (10.8 mm SL). The right eye moves from the right side of the head around the dorsal fin origin (Fig. 11 A) as in C. arctifrons and E. microstomas (Richardson and Joseph 1973). The right eye reaches its final position on the left side of the head at about 10-12 mm SL. Occurrence Larvae were collected in the Cape Fear River Estuary during May (pers. obs.) and in the Gulf of Mexico off Louisiana west of the Mississippi River Delta during July and August (Walker 7 ). Moe and Martin (1965) suggested a spawning season from March to at least June for the east- ern Gulf of Mexico off Florida (based on ripe Cephalic Spination Preopercular spines (Table 4) were present in the smallest preflexion specimen (4.6 mm NL, 7 H. J. Walker, Research Technician, North Carolina State University, Cape Fear Estuarine Laboratory, P.O. Box 486, Southport, NC 28461, pers. commun. July 1977. 67 FISHERY BULLETIN: VOL. 80, NO. 1 females). Capture of a ripe female from the same area in June was reported by Topp and Hoff (1972). Christmas and Waller (1973) suggested that spawning may be nearly continuous throughout the year. However, that observation was partly based on the occurrence of one juve- nile specimen during January and another dur- ing February that could have been spawned in the late summer or early fall. Therefore, the sea- son may extend beyond August, but the evidence is not yet complete. Comparisons Larval Characters Morphology seems to be influenced by the en- vironment and duration of larval existence. Cith- arichthys cornutus and C. gymnorhinus are found in deeper water and may have longer pel- agic larval stages than C. spilopterus or E. cros- sotus. In some respects, the latter two species are similar to each other and dissimilar to the first two. Citharichthys spilopterus and E. crossotus have only two elongate dorsal rays, as opposed to three in the other two species. They have smaller, less conspicuous frontal-sphenotic spines, and fewer larval teeth. (However, the jaws of C. spil- opterus later grow at a relatively fast rate and acquire correspondingly more adult teeth.) Dur- ing transformation, the origin of the dorsal fin is slightly farther forward relative to the right eye in C. cornutus and C. gymnorhinus than in the other two species. (However, after transforma- tion the dorsal origin is more anterior relative to the right eye in all three Citharichthys species than in E. crossotus.) Citharichthys spilopterus and E. crossotus larvae have smaller eyes and mouths than the other two. They also complete transformation at a smaller size. Known similarities among Citharichthys lar- vae that are not shared with Etropus larvae include the absence of a pectoral melanophore (except possibly in C. macrops), less extensive internal notochordal pigmentation, and, later, more gill rakers. Except for a shallower body and smaller eyes, C. arctifrons larvae are mor- phologically similar to those of C. cornutus and C. gymnorhinus. Etropus microstomus larvae are similar to those of E. crossotus. Larval Occurrence Differences among distributions of larvae (Append. Table 5) can be helpful in identifying them to species. Months of occurrence of larvae reported here are those in which larvae have been collected throughout the ranges of the re- spective species (except that data for C. macrops are from the southern part of its range, and data for C. arctifrons and C. arenaceus are from the northern parts of their ranges). Because most sampling was not continuous throughout the year, presence in other months is not precluded; however, enough is known to delineate approxi- mate spawning seasons for most of the species. Larval occurrence of C. cornutus, C. gymno- rhinus, C. spilopterus, and E. crossotus was dis- cussed in the earlier species' accounts. Throughout their ranges, E. microstomus spawns from March through August and E. rimosus spawns from September to April (Leslie 1977). Leslie suggested that spawning of the two species may be temporally distinct. This con- flicts with spawning of E. microstomus reported from May to December (Richardson and Joseph 1973), and my information is not sufficient to re- solve this conflict. However, Scherer and Bourne (1980) collected E. microstomus eggs in Septem- ber and larvae in October in Block Island Sound, which is north of the reported adult range (Table 1). In the eastern Gulf of Mexico, larvae of E. rimosus smaller than 4 mm NL were common in November, January, February, and May (Dowd 1978). Small juveniles (>13 mm SL) of C. abbotti were caught in the Gulf of Mexico from Veracruz to Campeche, Mexico, in June and September (Dawson 1969), indicating a spawning season approximately from May through August, or longer. Citharichthys macrops larvae smaller than 4 mm NL were caught in the eastern gulf in May and November (Dowd 1978). Topp and Hoff (1972) reported juveniles from the same part of the gulf during the fall and winter. The season probably extends from May through November, and possibly longer. Richardson and Joseph (1973) reported a spawning season approximately from May to December for C. arctifrons in the Chesapeake Bight, with peak spawning from July through October. Dawson ( 1969) reported a 27 mm SL specimen of C. arenaceus caught in the British West Indies in November. This may indicate summer spawn- ing, probably during August or September at least. 68 TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS; AND ETROPUS Citharichthys amblybregmatus and C. dinocer- os are deep water forms. Because of the constancy of their environment, they may have extended spawning seasons, but little is known of their habits. Considering the geographic and bathymetric distributions of adults (Table 1) and probable spawning periods (Append. Table 5), it is un- likely that large numbers of larvae of different species of western North Atlantic Citharichthys and Etropus cooccur in the ichthyoplankton at any given time. Among the six deepwater spe- cies, C. amblybregmatus and C. dinoceros larvae probably occur relatively far from shore. Appar- ently, there is little difference between larval occurrence of C. gymnorhinus and C. cornutus, but spawning centers or peak periods could be distinct. Topp and Hoff (1972) suggested that adults of the two species were bathymetrically separated, C gymnorhinus being found in shal- lower water. Etropus rimosus adults occur in shallow water and do not spawn during the sum- mer. Citharichthys arctifrons has a more north- ern distribution than the preceding three species and its spawning peak is in the summer, prob- ably earlier than that of E. rimosus. Among the three coastal species, the geographic range of C. arenaceus is distinct from those of the other two, and C. macrops and E. microstomas cooccur only off the Carolinas. In this area of overlap, C. macrops probably spawns in the fall and E. microstomus in the spring. Among the three estuarine and coastal species, C. abbotti spawns in the warmer months and may be restricted to very shallow water. Citharichthys spilopterus spawns in the colder months, beginning in late summer in the Gulf of Mexico and in mid to late fall off the Carolinas. Etropus crossotus may spawn from March through the summer, or later, in the gulf, but probably does not begin off the Carolinas until after most spawning activity of C. spilopterus is finished. SUMMARY The caudal fin formula (4-5-4-4) is the most re- liable character for linking larval specimens to the group of paralichthyine genera Citharich- thys, Cyclopsetta, Etropus, and Syacium. The most useful characters for identification to genus are number of elongate dorsal rays, de- gree of cephalic spination, and pigmentation. Known western North Atlantic Syacium and Cyclopsetta larvae have 5-10 elongate dorsal rays and well-developed preopercular and frontal- sphenotic spines. Known western North Atlantic Citharichthys larvae have two or three elongate dorsal rays, small (or no) preopercular spines, small frontal-sphenotic spines, no pectoral mel- anophore (except possibly C macrops), little notochordal pigmentation, usually large eyes and mouths, and (except for C. arctifrons) high gill raker counts. Known western North Atlantic Etropus larvae have no or two elongate dorsal rays, small preopercular and frontal-sphenotic spines, a melanophore at the base of the pectoral fin, extensive notochordal pigmentation, small eyes, and low gill raker counts. Table 9 summarizes the most useful charac- ters for distinguishing larvae of the six species of western North Atlantic Citharichthys and Etro- pus that have been described in detail. The best characters for determining species are number of elongate dorsal rays, number of caudal verte- brae, pectoral and notochordal pigmentation, number of left pelvic rays (C. gymnorhinus), head shape and snout to anus length (C. spilop- terus), number of gill rakers, and length at trans- formation. Citharichthys arctifrons larvae have three elongate dorsal rays, no preopercular spines, many caudal vertebrae, a small eye, large mouth, and few gill rakers. Citharichthys cor- nutus larvae have three elongate dorsal rays, a strong first left pelvic ray, heavy pigmentation, a large eye and mouth, and relatively many gill rakers. Citharichthys gymnorhinus larvae have three elongate dorsal rays, few caudal vertebrae, five left pelvic rays (with the first weak), a dis- tinct caudal pigment band, large eye and mouth, and relatively many gill rakers. Citharichthys spilopterus larvae have two elongate dorsal rays, few caudal vertebrae, little pigmentation, a small eye and mouth, very blunt anterior profile, short snout to anus length, and relatively many gill rakers. Etropus crossotus larvae have two elongate dorsal rays, heavy pigmentation, a small eye and mouth, and many (for the genus) gill rakers. Etropus microstomus larvae have no elongate dorsal rays, a small eye, and few gill rakers. ACKNOWLEDGMENTS I wish to thank the following individuals and institutions for their contributions to this study: for loans and gifts of specimens — E. H. Ahlstrom (NMFS, La Jolla); Charles Bennett, William 69 Table 9.— The most useful FISHERY BULLETIN: VOL. 80, NO. 1 characters, in order of ontogenetic appearance, for distinguishing larvae of four species of Cith- ariehthys and two species of Etropus. Character C. arctHrons* C. cornutus C. gymnorhinus C spilopterus E. crossotus E. microstomas* Pectoral melanophore (before transformation) Absent Absent Absent Absent Present Present Notochordal pigment (before transformation) Caudal only Caudal only Caudal only Caudal only From brain to caudal area From brain to caudal area Elongate dorsal rays (before transformation) 3 3 3 2 2 Caudal vertebrae 26-28 25-26 23-24 2 23-24(25) 2 (24)25-26 2 24-25(26) Lateral pigment (before transformation) Moderate Heavy Moderate Light Heavy Moderate Length at flexion (mm) 9 9-10 7-8 7-8 9-10 7 Left pelvic rays (full complement) 6 6 5 6 6 6 Left preopercular spines (during preflexion- 14-31-22 17-22-31 31-31-16 17-20-11 Several flexion-postflexion) Eye diameter/BL in % (during preflexion- 7 10-10- 8 9- 9- 9 10- 8- 7 7- 7- 6 7 flexion-postflexion) Upper jaw length/BL in % (during preflexion- 10 10-11-10 10- 9- 9 10- 7- 7 7- 7- 7 9 flexion-postflexion) Lower jaw length/BL in % (during preflexion- 13-14-13 12-13-13 12-10- 9 10-10-10 flexion-postflexion) Snout to anus length/BL in % (during preflexion- 42 46-46-39 43-44-40 40-39-32 39-44-39 40 flexion-postflexion) Gill rakers on the lower limb of the first arch 6- 8 10-15 9-14 9-15 6- 9 4- 7 (at transformation) Length at transformation (mm) 13-15 -18 -18 9-11 10-12 10-12 Snout spine (at about transformation) May be present Present (in males?) Present (in males 9 ) Absent Absent Absent Symphyseal spine (after transformation) Absent Present (in males?) Present (in males?) Absent Absent Absent 'Data for C. arctifrons and E. mtcrostomus are mostly from Richardson and Joseph (1973). 'Uncommon counts given in parentheses. Birkhead, Ronald Hodson, Wilson Laney, Ed- ward Pendleton, and others (NCSU); Norman Chamberlain (GMBL); Alan Collins (NMFS, Panama City); Mary Ann Daher and John McEachran (Texas A&M); Lise Dowd and Ed- ward Houde (RSMAS); Kathy Kearns (CP&L); Walter Nelson (NMFS, Beaufort); John Olney (VIMS); Howard Powles and Bruce Stender (SCMRRI); Sally Richardson (GCRL); Frank Schwartz (UNC); Victor Springer (USNM); and Frank Truesdale and H. J. Walker (LSU). Tech- nical assistance was provided by Robin Cuth- bertson, Jay Geaghan, Ronald Hodson, Marsha Shepard, and William Watson (NCSU); Frank McKinney (USNM); and the Beaufort NMFS Laboratory. Data and advice were provided by E. H. Ahlstrom, Lise Dowd, Elmer Gutherz (NMFS, Pascagoula), Drew Leslie (Florida State University), Sally Richardson, and How- ard Powles. Nancy Brown Tucker (VIMS) assisted in preparation of the manuscript. John Miller (NCSU), Allyn Powell (NMFS, Beaufort), E. H. Ahlstrom, Jeff Govoni (VIMS), John Olney (VIMS), William Nicholson (NMFS, Beaufort), John Reintjes (NMFS, Beaufort), William Hass- ler (NCSU), B. J. Copeland (NCSU), Leonard Pietrafesa (NCSU), and two anonymous review- ers offered many helpful suggestions for im- proving the manuscript. Carolina Power and Light Company provided financial support. LITERATURE CITED Amaoka, K. 1969. Studies on the sinistral flounders found in the waters around Japan— Taxonomy, anatomy and phylog- eny. J. Shimoneseki Univ. Fish. 18:65-340. Christmas, J. Y., and R. S. Waller. 1973. Estuarine vertebrates, Mississippi. In J. Y. Christmas (editor), Cooperative Gulf of Mexico estua- rine inventory and study, Mississippi, p. 320-403. Gulf Coast Res. Lab., Ocean Springs, Miss. Dawson, C. E. 1969. Citharichthys abbotti, a new flatfish (Bothidae) from the southwestern Gulf of Mexico. Proc. Biol. Soc. Wash. 82:355-372. 70 TUCKER: LARVAL DEVELOPMENT OF CITHARICHTHYS AND ETROPUS DOWD, C. E. 1978. Abundance and distribution of Bothidae (Pisces, Pleuronectiformes) larvae in the eastern Gulf of Mexico, 1971-72 and 1973. M.S. Thesis, Univ. Miami, Miami, 106 p. EVSEENKO, S. A. 1979. Larvae of the flounder Cydopsetta Gill, 1888 (Bothidae, Pisces) from the northwestern Atlantic. Biol. Morya 1979(2):67-75. Fahay, M. P. 1975. An annotated list of larval and juvenile fishes cap- tured with surface-towed meter net in the South Atlan- tic Bight during four RV Dolphin cruises between May 1967 and February 1968. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-685, 39 p. Futch, C. R. 1977. Larvae of Trichopsetta ventralis (Pisces: Bothidae), with comments on intergeneric relationships within the Bothidae. Bull. Mar. Sci. 27:740-757. Futch, C. R., and F. H. Hoff, Jr. 1971. Larval development of Syacium papillosum (Bothi- dae) with notes on adult morphology. Fla. Dep. Nat. Resour. Mar. Res. Lab., Leafl. Ser., Vol. IV, Pt. 1, No. 20, 22 p. Goode, G. B., and T. H. Bean. 1896. Oceanic ichthyology. U.S. Natl. Mus., Spec. Bull., 553 p. GUTHERZ, E. J. 1967. Field guide to the flatfishes of the family Bothidae in the western North Atlantic. U.S. Fish Wildl. Serv., Circ. 263, 47 p. 1970. Characteristics of some larval bothid flatfish, and development and distribution of larval spotfin flounder, Cydopsetta fimbriata (Bothidae). U.S. Fish Wildl. Serv., Fish. Bull. 68:261-283. GUTHERZ, E. J., AND R. R. BLACKMAN. 1970. Two new species of the flatfish genus Citha richthys (Bothidae) from the western North Atlantic. Copeia 1970:340-348. Hensley, D. A. 1977. Larval development of Engyophrys senta (Bothi- dae), with comments on intermuscular bones in flat- fishes. Bull. Mar. Sci. 27:681-703. Hsiao, S. C. T. 1940. A new record of two flounders, Etropus crossotus Goode and Bean and Ancylopsetta dilecta (Goode and Bean), with notes on postlarval characters. Copeia 1940:195-198. Leslie, A. J., Jr. 1977. The systematics of Etropus microstomus (Gill) and E. rimosus Goode and Bean (Pisces: Bothidae), with eco- logical notes. M.S. Thesis, Florida State Univ., Talla- hassee, 81 p. Moe, M.A., Jr., and G. T. Martin. 1965. Fishes taken in monthly trawl samples offshore of Pinellas County, Florida, with new additions to the fish fauna of the Tampa Bay area. Tulane Stud. Zool. 12: 129-151. Moser, H. G., E. H. Ahlstrom, and E. M. Sandknop. 1977. Guide to the identification of scorpionfish larvae (family Scorpaenidae) in the eastern Pacific with com- parative notes on species of Sebastes and Helicolenus from other oceans. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 402, 71 p. Pietrafesa, L. J., J. O. Blanton, and L. P. Atkinson. 1978. Evidence for deflection of the Gulf Stream at the Charleston Rise. Gulfstream 4(9):3-7. Richardson, S. L., and E. B. Joseph. 1973. Larvae and young of western North Atlantic bothid flatfishes Etropus microstomus and Citharich- thys arctifrons in the Chesapeake Bight. Fish. Bull., U.S. 71:735-767. SCHERER, M. D., AND D. W. BOURNE. 1980. Eggs and early larvae of smallmouth flounder, Etropus microstomus. Fish. Bull., U.S. 77:708-712. Smith, W. G., J. D. Sibunka, and A. Wells. 1975. Seasonal distributions of larval flatfishes (Pleuro- nectiformes) on the continental shelf between Cape Cod, Massachusetts, and Cape Lookout, North Carolina, 1965-66. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-691, 68 p. Sumida, B. Y., E. H. Ahlstrom, and H. G. Moser. 1979. Early development of seven flatfishes of the east- ern North Pacific with heavily pigmented larvae (Pi- sces, Pleuronectiformes). Fish. Bull., U.S. 77:105-145. Taylor, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. U.S. Natl. Mus. 122(3596), 17 p. Topp, R. W., and F. H. Hoff, Jr. 1972. Flatfishes (Pleuronectiformes). Mem. Hourglass Cruises, 4 (Pt. 2), 135 p. Wenner, C. A., C. A. Barans, B. W. Stender, and F. H. Berry. 1979. Results of MARMAP otter trawl investigations in the South Atlantic Bight. I. Fall 1973. S.C. Mar. Re- sour. Cent., Tech. Rep. 33, 79 p. 71 FISHERY BULLETIN: VOL. 80. NO. 1 Appendix Table 1.— Frequency distributions of caudal vertebral counts for western North Atlantic species of Citharichthys and Etropus. i Species 21 22 23 24 25 26 27 28 29 W X C. abbotli 19 96 9 124 21 92 C. arenaceus 3 38 8 49 22 10 C. gymnorhmus 9 27 36 23 75 C. spilopterus 23 109 8 140 23 89 E. rimosus 3 50 53 5 m 24.54 E. microstomus 51 61 2 114 24 57 C. macrops 27 46 73 24.63 E. crossotus 1 69 15 85 25.16 C. cornutus 15 29 44 25 66 C. amblybregmatus 5 16 21 25.76 C. arctifrons 5 34 3 42 26.95 C. dinoceros ( 2 ) ( 2 ) 'Compiled from Gutherz 1967; Dawson 1969; Gutherz and Blackman 1970; Leslie 1977; S L. Richardson, Research Assistant Professor, School of Oceanography, Oregon State Univer- sity, Corvallis, OR 97331, pers. commun December 1976 (unpubl. data for E microstomus and C. arctifrons), and original data for larvae, juveniles, and adults of C gymnorhinus, C spilopterus, C macrops, E, crossotus, and C. cornutus. 2 Extremes of counts. Appendix Table 2.-Frequency distributions of anal fin ray counts for western North Atlantic species of Citharichthys and Etropus. 1 Species 48 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 ~ 68 69 70~ C. arenaceus ( 2 ) 1 11 8 14 9 3 E. microstomus ( 2 ) 1 12 8 14 22 32 31 23 15 6 5 ( 2 ) C. gymnorhinus 2 4339 13 6 10 32 C. abbotti 1 5 13 37 35 22 14 4 1 E. rimosus 1 5 5 17 25 42 57 57 38 41 16 6 1 C. spilopterus ( 2 ) 15 24 30 41 26 11 4 ( 2 ) C. macrops 116 2 5 17 16 13 E. crossotus 1 1 , 6 10 12 C. arctifrons ( 2 j 117 6 C. cornutus ( 2 ) 3 2 7 4 C. amblybregmatus 1 3 2 3 1 q C. dinoceros 12 10 12 13 13 3 ( 2 ) 10 23 16 8 6 6 3 11 76 N * 46 53.fr. 160 57.fl 55 56.(| 132 55.! 311 59: 151 58f 73 61 t 60 63. 83 65.3 27 63.CI 22 67.CI n ( 2 > 'Compiled from Gutherz 1967; Dawson 1969; Gutherz and Blackman 1970; Topp and Hoff 1972' Leslie 1977 9 l Rirharrt^n n OM „^ « . . r> < Oceanography, Oregon State University, Corvallis. OR 97331, pers commun December 1976 (unpubl date for C arcWmn^d^Zt^lf^ Pro essor ' Sch°°' C gymnorhinus. C spilopterus. C. macrops, E. crossotus. and C. cornutus lunpuDi aata tor c. arctifrons), and original data for juveniles and adults of Extremes of counts, not included in totals. Appendix Table 3,-Freguency distributions of dorsal fin ray counts for western North Atlantic species of Citharichthys and Etropus.* s P ecies 67 68 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 90 95 E. microstomus n 2 1 2 7 8 22 21 26 24 23 12 6 4 1 1 ( 2 ) C. arenaceus ( 2 ) 1 2 11 10 9 7 3 2 C. gymnorhinus 7 8 12 9 8 5 3 1 E. rimosus 2 8 16 18 43 57 52 48 40 17 4 4 1 C. abbotti 1 5 13 23 30 23 23 11 2 1 C. spilopterus 1 4 4 14 35 32 24 26 11 5 1 C. cornutus 1 1 1 3 5 8 6 5 1 1 1 C. macrops 1 1 1 1 3 9 10 12 12 11 C. arctifrons 1 1 3 8 8 14 14 17 7 2 E crossotus 0.05). Clearest evidence for differential day/night net avoidance by N. megalops is found in the catch data obtained by the 1 m 2 MOCNESS (Fig. 3). Without exception, for each of the eight day/ night pairs of tows taken on Knorr 65 and Knorr 71, the day estimate of numbers per square t uu o — o 80 8 oo 60 — o X °x * 40 h- x x *x x° )f X XX o • _ • • • 20 • ADOLESCENT ° FEMALE x MALE 1 ' i 10 20 NEMATOSCELIS MEGALOPS TOTAL BODY LENGTH (mm) 40 Figure 2.— Relationship between total body length, carapace length, and wet weight of Nematoscelis megalops. 79 APRIL 1977 KNORR 65 FISHERY BULLETIN: VOL. 80. NO. 1 OCTOBER/NOVEMBER 1977 KNORR 71 RING BOB STATION 1 MOC 1-62 MOC1-63 MOC 10-21 MOC 10-28 (00 10 1 10 100 100 10 1 10 100 5} 200 f- 400 >■ 600 P- 600 is 1000 jg 200 iS r- 400 j- 600 n. ki 800 1000 7m'=6 2 */m 2 =35 1 */m 2 »2.1 RING AL STATION 2 MOC 1-72 MOC 1-71 MOC 10-35 10 1 10 100 10 1 1 ' O l ; /m' = 7 7 7m' = 8 .07 7m 2 =26 4 *m'=<01 SLOPE WATER £ 200 is ^ 400 J 600 ft. ^ 800 1000 STATION 3 MOC 1-73 MOC 10-36 10 1 10 1 I 1_ O * °0 o o o o 7m 2 =0 9 738 O O »/_* /m'=0 6 STATION 4 MOC 1-76 MOC 1-75 MOC 10-38 100 10 100 200 400 600 800 1000 1 10 100 7m' = 14 8 046 7m 2 =22 7 1 DAY | NIGHT A NOT SAMPLED * 10° C ISOTHERM DEPTH 7m 2 = 28 4 RING EMERSON STATION 6 MOC 199 MOC 198 MOC 10-59 MOC 10-58 100 10 1 10 100 100 10 I 10 100 * r~~: 1 * - 7m 2 = 1? 7 7m 2 = 73 7m 2 =5 6 RING FRANKLIN 7m 2 = 14 9 200 400 600 800 1000 STATION 7 M0CH02 MOC 1-103 MOC 10-61 MOC 10-62 10 1 10 100 100 10 1 10 too 7m 2 =06 7m' =17 3 7m* » 46 2 7m 2 -- 31 5 200 400 600 800 WOO STATION 8 MOC 1-109 MOC 1-110 MOC 10-65 MOC 10-66 100 10 1 10 100 7m' =6 3 7m 2 =8 5 */m 2 = 27 3 SLOPE WATER '/m' = 22 9 200 400 600 800 1000 STATION 9 MOC 1-117 MOC 1-116 MOC 10-67 MOC 10-68 7m' = 4 1 7m' = 135 4 */m 2 = 31 1 7m 2 = 914 - 1 MOC 1-96 MOC 1-97 STATION 5 MOC 10-57 1 10 100 7m'=73 2 7m' = 4 31 8 7m 2 =235 2 200 400 600 800 1000 200 400 600 800 1000 Figure 3.— Vertical distribution of Nematoscelis megalops in the Slope Water and in variously aged cold core rings based on collec- tions made with the 1 m 2 and 10 m 2 MOCNESSes on two cruises taken 6 mo apart. Night samples are blacked; day samples are crosshatched. 80 WIEBE ET AL.: AVOIDANCE OF TOWED NETS BY NEMATOSCEUS MEGALOPS meter for the water column is less than the cor- responding night catch. In every case, sampling extended below the maximum depth of occur- rence of the population and there is no evidence that any individuals of the population migrated vertically out of the depth zone sampled during the day. Therefore, it is highly significant that all of the day values were less than the respective night ones (P<0.005). This result gains impor- tance if we also consider 10 other day/night pairs of 1 m 2 MOCNESS tows in which N. megalops was collected on previous cruises (Chain 125, Knorr 53, Knorr 62). For nine of these pairs, moderately to dramatically higher catches in the night tow were obtained (Table 2). The single ex- ception to this pattern was a pair of Slope Water tows taken near the continental shelf in the wake region of a warm core ring (tows 41, 42). But these two tows were displaced in space by several miles, and the night tow was taken nearer the warm core ring where a lower catch might have been expected. Of the 18 day/night pairs of 1 m 2 MOCNESS tows, 17 yield higher density estimates at night. Patchiness in the distribution of N. megalops contributed to variability to these estimates but as an unbiased variance component, it does not affect our expectation that one-half of the day and one-half of the night tows in day/night pairs should be the larger. Thus it is unlikely that patchiness of this species is responsible for the significantly higher night catches that we have observed (P<0.001). We know of no other expla- nation than avoidance to explain this result. There are only five pairs of 10 m 2 MOCNESS observations of the vertical distribution of N. megalops. For two of these, the integrated day catch is larger than the corresponding night catch and, therefore, night catches are not sig- nificantly larger than day catches (P>0.05). This result either means that there is no day/night dif- ferential avoidance of the 10 m 2 net or that in the face of other sources of error such as patchiness, we have too few day/night pairs of observations to observe the avoidance effect. If avoidance were affecting only the smaller net then at least we would expect that the 1 m 2 net day catches per unit volume would be consistently smaller than the corresponding 10 m 2 net day catches. We might also expect that night catches with the 1 m 2 net would be smaller than the 10 m 2 net. Neither comparison yields a significant result (P>0.05; day MOCNESS 1 tows greater than day MOCNESS 10 tows in four out of seven com- parisons; night MOCNESS 1 tows greater than night MOCNESS 10 tows in three out of seven comparisons). Thus within the limits of error, by day or by night both net systems provide com- parable estimates of the number of N. megalops living in the water column at a given station. It is possible that the lack of differences in the catching rates between the two nets is due to the different mesh sizes. Small individuals might have been caught more efficiently by the 1 m 2 net while larger individuals could have avoided this net better and conversely for the 10 m 2 net except that small individuals would have been lost due to escapement through the mesh. The size-fre- quency data in Figures 4 and 5 do not support this possibility. While there is considerable vari- ability between net tow pairs, in terms of abso- lute abundance, neither net system systemati- cally catches large or small individual N. mega- lops in the size range counted better than the other. A similar observation can be made if com- parisons are made on the relative abundances in a given size class (Fig. 6). There is one other potentially significant trend in the data that is important to note. The magni- tude of the day/night avoidance does not appear to be uniform with depth. For the 1 m 2 MOC- NESS, largest differences between paired night and day catches where both are positive occur when the center of distribution of N. megalops is above 300-400 m and minimum differences occur at or below these depths (Table 2). Linear regression of the ratio of night to day catch (N/D) versus depth of the center of the distribution at night (50% of occurrence with depth) is signifi- cant at P = 0.1. There is a similar pattern in the 10 m 2 MOCNESS tows, although as mentioned above, the day/night differences in catching rates are considerably smaller. In summary, there is clear evidence for differ- ential day/night avoidance of the 1 m 2 MOC- NESS. Furthermore, there are no significant differences in the size range of adolescent or adult N. megalops caught by the 1 m 2 or 10 m 2 MOCNESS systems nor in either system's esti- mates of its abundance in the water column at a given station when day or night pairs are com- pared. Although differences between pairs of day/night catches for the 10 m 2 MOCNESS are statistically not significant, the entire data set when considered as a whole strongly suggests that N. megalops is also avoiding the 10 m 2 net, albeit to a lesser extent. 81 FISHERY BULLETIN: VOL. 80. NO. 1 ""1 RING BOB STATION I MOC 10-27 D MOC 1-63 N n-K>9 P 2 MOC 10 28 N APRIL 1977 KNORR 65 RING AL STATION 2 MOC 1-72 D MOC 10-35 SLOPE WATER STATION 3 MOC I 73 a MOC K> 36 D STATION 4 MOC I 75 N "■98 IP n.485 MOC 10-38 N OCTOBER /NOVEMBER KNORR 71 O CO coo RING EMERSON STATION 6 MOC I 99 MOC 1-98 N n.27 n-58 MOC K) 59 STATION 7 MOC t 102 r-|i MOC 10 61 D 'ii.ii i i i 50 100 150 MOC K) 58 N . i i i i i i i i i RING FRANKLIN MOC I 103 N J MOC K) 62 N MOC 1-117 L J ...... LU MOC 10 65 D STATION 9 MOC K> 67 ' MOC H09 WET WEIGHT (mg) SLOPE WATER MOC 1116 N MOC 10 68 N MOC 10 66 N I ' ' I ' ' I 50 100 150 STATION 5 MOC I 97 N 50 K>0 D = DAY N NIGHT FIGURE 4. — Comparison of the composite size-frequency distribution (expressed as No./l,000 m 3 for a given wet weight interval)of Nematoscelis megalops caught by the MOC NESS 1 (shaded) and the MOC NESS 10 (crosshatched) for tows taken on the same day or night, n = the number of individuals used to construct the histogram. Application of Barkley Avoidance Theory Catch Since it is likely that N. megalops avoids both net systems, it must detect the approach of either net at some distance in front of the net, resulting in a response which permits a certain percentage of the population to avoid capture. Determina- tion of the avoidance percentage and reaction distance requires an indirect approach, since no other means are available. The theoretical framework on the process of net avoidance devel- oped by Barkley (1964, 1972) provides a means for estimating these parameters according to a quantitative theoretical model. Barkley (1972) formulated the problem in the following way: (volume sampled) X (no. of organisms unit volume ) X (probability of capture) — (losses) (1) "Losses" refers to individuals which are enclosed by the net but escape through the net meshes. For the size range of individuals which consti- tute our "catch," the "losses" term is essentially zero. Since the volume of water sampled has been rather carefully measured, the "probability of capture" (Pc) is of greatest concern. Pc is related to the mean reaction distance (.r<>), the radius of the net mouth (R), the net's speed (U), and the organism's mean escape speed (u,) by the equa- 82 WIEBE ET AL.: AVOIDANCE OF TOWEL) NETS BY NKMATOSCKL1S MEGALOPS RING BOB 6 \ S 01 - MOC I 62 MOC 1-63 N APRIL 1977 KNORR 65 RING AL SLOPE WATER STATION 2 STATION 3 MOC 1 72 MOC 1-73 D MOC 1-27 D I I I M I I I I I M I I MOC 10 28 N ' MOC 10 35 D RING EMERSON STATION 6 MOC I 99 OCTOBER/NOVEMBER KNORR 71 STATION 9 MOC 1-117 D MOC 10-59 m <| I I I i IJ1 I MOC 1-102 MOC 10-58 N i i i i i i i i MOC 1-103 N RING FRANKLIN MOC 1-109 TT - MOC 10-61 50 100 150 MOC 10-62 N 50 WO 150 MOC 10 65 WET WEIGHT (mg) MOC 10 36 ''''' i i i i I i i j SLOPE WATER MOC 1 116 N TV MOC l-IIO N MOC 10-66 N ' ' ' ' ' ' ' i i i i i i . i i i i 50 100 150 STATION 4 MOC 1 75 N IMF MOC 10 38 N j j i i i i STATION 5 MOC 1-97 N MOC 10-57 N i I i L I M I I I I I I 50 100 150 0- DAY N' NIGHT Figure 5.— Comparison of the difference between paired MOCNESS 1 and MOCNESS 10 catches in No./l,000 m 3 for a given wet weight interval. For shaded columns above the line, the MOCNESS 1 catch is greater than the MOCNESS 10 catch and vice versa for crosshatched columns below the line. tion derived by Barkley (1972, equation 6) wherein: 10 m 2 MOCNESS catch volume sampled Pc = 1 Jo u,. R(lf u e 2 ) Vi (2) This expression assumes that as the net moves forward through the water, an individual senses the oncoming net and at a distance Xo in front of the net begins a swimming response in a direc- tion away from the net which is optimal for avoidance. Thus, this equation provides an esti- mate of the minimum probability of capture. As a first step in applying these equations to our data, we may recall that for both the paired night tows and the paired day tows differences between the two net systems were not signifi- cant, i.e., 1 m 2 MOCNESS catch volume sampled If we assume that the number of organisms per unit volume was a constant during the time each pair of tows was taken, then: 10 m 2 MOCNESS P, = 1 m 2 MOCNESS P r and 1 - XloUe 2\'/ i50(ioo' -u;) X\U f 2\ 4 3 4 3 .01 Figure 7.— Examples of relative es- cape speed of Nematoscelis megalops individuals versus the catch per 1,000 m 3 . Superimposed on this plot are the theoretically derived curves of Xr>/R as a function P c and uj U adjusted to give a "best" fit of the observed points. 85 FISHERY BULLETIN: VOL. 80, NO. 1 if individual escape speeds remained constant. However, further analysis reveals that the dif- ference in Xo for a given day/night catch differ- ential could be a function of the relationship between the observed night catch and the true water column abundance. This is clearly evident if we express the ratio of the day catch per vol- It could be argued that the day/night catch dif- ferential is due to differences in escape speed of the individuals rather than a change in their re- action distance. To explore this we have also solved Equation (3) for the ratio of day escape speed, ud, to night escape speed, u N , after assum- ing x D = x N . + m r- -. •-■ & 2 * x_ 2 Ir _ 2 2 ~x~ R L J _ (5) ume sampled {DC) and night catch per volume sampled {NC) in terms of real abundance (^4) and percent capture as expressed in Equation (2): DC NC If we assume that the daytime escape speed, Ud, is equal to the nighttime speed, w,v, and solve for the ratio of the daytime reaction distance, xd, to the nighttime reaction distance, Xn, we have: Xd/Xs 1 - We have evaluated this equation assuming a true abundance of 100 individuals per volume, night- time catches of 99, 90, 10, 1, and 0.1 individuals per volume, and daytime catches of 50, 10 and 1% of the nighttime catch. The ratios of x D /x Ni plotted as a function of the ratio of NC/A Fig. 8a), shows that only very small differences in re- action distance between day and night are re- quired to explain large day/night catch differen- tials when the night catch is 10% or less of the true water column abundance. The fact that we see no significant difference in day/night reac- tion distances suggests our nighttime catches also could be affected strongly by avoidance, and that even at night we have significantly under- estimated the numbers of N. megalops in the water column. Note that the ratio of day/night escape speeds is a function of xs and R as well as DC, NC, and A. The escape speed and radius of net were not in (3) Equation (4) for the ratio of day/night reaction distances. We have evaluated this ratio using the same values noted above. With these results (Fig. 8b), we reach a conclusion similar to that for re- (4) action distance, namely, if reaction distance re- mains constant between day and night, then small differences in escape speed can explain the day/night catch differential when the night catch is 10% or less of the true abundance. There is, however, an entirely different expla- nation which may account for this outcome in application of Barkley avoidance theory to our data. In fitting these data to Barkley's plots of percent capture versus the ratio of x /R, two assumptions were required: 1 ) that all changes in size frequency are due to avoidance and 2) that swimming speed is a function of body size. The second assumption can be examined if one has day/night pairs of tows taken at the same station location with the same size of net. With swim- ming speed a function of size, Barkley's model 86 WIKBE ET AL.: AVOIDANCE OF TOWEL) NETS BY NEMATOSCELIS MEGALOPS IOOO.Of 100.0 ^ ^ 10.0 1.0 DC = Day Catch NC= Night Catch Xn = Day Reaction Distance Xn = Night Reaction Distance A - True Abundance — o — o 2DC = NC — I0DONC — 100 DC - NC ~ U D =Day Escape Speed Un= Night Escape Speed 100.0 t ^ 10.0 °-° 2 DC = NC — 10DC=NC x— 100DC = NC NC/A Figure 8.— Relationships between the ratio of night catch to true abundance (NC/A) and a) the ratio of day and night reac- tion distances (x,,/x s ), and b) the ratio of day and night escape speeds ( u„/u v ). predicts that the ratio of the number of individ- uals caught per size class at night (NC) to those caught during the day (DC) will increase with in- creasing individual size (the inverse of Equation 3). This relationship is illustrated in Table 4 where u N and Ho are assumed to be equal and 10 body lengths/s, Xd — 175 cm, xn = 150 cm, R = 50 cm, and U = 100 cm/s. This ratio increases dramatic- ally with individual size until at the largest size, the model predicts all individuals avoid capture. No such pattern emerges if we compute the ratio NC/DC for each size class in our paired day/ Table 3.—Nematoscelvi megalops reaction distances (xo) for the 1 m 2 and 10 m* MOCNESS nets derived from the plots like those in Figure 7. Station Cruise Tow Day/Night Xo/fl Xo 1 Knorr 65 M-1-62 D 3.4 1.7 M-10-27 33 5 M-1-63 N 34 17 M-10-28 34 5 3 Knorr 65 M-1-73 D 34 1.7 M- 10-36 (') (') 4 Knorr 65 M-1-75 N (') (') M- 10-38 34 5 5 Knorr 71 M-1-97 N 4.5 2.3 M-10-57 44 66 6 Knorr 71 M-1-99 D 4.5 23 M-10-59 4 4 6.5 M-1-98 N 4 5 23 M-10-58 4 4 6.6 7 Knorr 71 M-1-102 D 4.4 22 M-10-61 (') (') M-1-103 N 44 22 M-10-62 44 66 8 Knorr 71 M-1-109 D 35 1.8 M-10-65 44 66 M-1-110 N 43 22 M-10-66 4 3 64 9 Knorr 71 M-1-117 D 3.5 18 M-10-67 4.4 66 M-1-116 N 4.4 22 M-10-68 4.3 65 'Not sufficient points to derive an estimate, Station 2 omitted for this reason. Table 4.— The ratio of night catch to day catch as a function of individual swimming speed (it*) as predicted by Barkley's avoidance model (inverse of Equation 3). w e is assumed to be a function of body size as described in the text. Body wet weight Ue Night Day (mg) (cm/s) catch' catch Ratio 20 1508 0294 0217 1 35 30 1632 0253 177 1.43 40 17 56 216 0.141 1.53 50 18 80 0.181 0.108 1 66 60 20 04 0.149 0080 1 84 70 21.28 0.120 0056 2 12 80 2252 093 0036 2.57 90 2376 0.070 0020 342 100 25.00 0.050 0.009 547 110 2624 0033 0002 14.57 120 27 48 0022 000 — 'Catch units are proportion of individuals present per unit volume night MOCNESS 1 or MOCNESS 10 tows (Table 5). Thus, the assumption of increasing swimming speed with increasing size does not appear to be valid, i.e., for the size range of indi- viduals used in this study, avoidance swimming speeds are essentially the same. One implication of this finding is that the size-frequency distribu- tions evident in the field data may not be seri- ously biased by the avoidance although the esti- mates of average density clearly are. 87 FISHERY BULLETIN: VOL. 80, NO. 1 Table 5.— Ratios of night to day catches (number per square meter) of Nematoscelis megalops as a func- tion of size for stations where both the MOCNESS 1 and the MOCNESS 10 were taken. <» indicates only the night tow caught individuals in the given size class; indicates the opposite patterns. Body wet weight (mg) MOCNESS 1- -tow no.: MOCNESS 10— tow no .: 117/116 62/63 99/98 102/103 109/110 27/28 59/58 61/62 65/66 67/68 10 OC — — — — — OO 4.4 20 OC _ 12 333 0.3 — 24 1 1 <0 1 40 30 66.6 1.3 17 OO 26 59 26 04 08 2 2 40 45.5 4.0 OC 37 26 48 36 08 48 29 50 79 56 oc 1 9 36 7 7 02 56 59 60 46 1.1 — oc 08 32 — 05 28 59 70 2.1 7.7 — OO 07 26 — 09 OC 1.8 80 0.6 7 7 — — 0.7 29 — 2.5 1.0 OC 90 24 OO — OO — 35 — 125 1.7 63 100 OO OC — — OO 1.0 — — OO — 110 120 130 — OO — — 7.1 — — — — - - — — - — - - - DISCUSSION From this application of the Barkley avoid- ance theory, it appears that estimates of N. mega- lops water column abundance could be substan- tially underestimated by both nets, even at night. Minimum probabilities of capture derived from best fits to model expectations are 0.1 or less for night catches and 0.01 or less for day catches. However, the fact that we cannot demonstrate a dependence of the ratio of night to day catches on the size of individuals caught strongly suggests the size dependent swimming speed assumption required to apply the model is not valid for this species, a result which is apparently supported by Kils's (1979) data for Euphausia superba escape swimming (tail swimming). Being unable to make this assumption means that the field population size-frequency distribution which was observed is probably not materially affected by avoidance. Undeniably some fraction of the N. megalops population is avoiding the net systems, and the problem is serious enough to merit an effort to reduce this bias, i.e., to prevent the avoidance from taking place. The usual strategies suggested to reduce net avoidance, increasing net speed or net size, have serious shortcomings in this case. Our evidence strongly implies that N. megalops' response to increased net size is to increase its reaction dis- tance so that the catch rate remains relatively constant. Barkley (1972) reached the same con- clusion in a comparison of 1 m diameter net and 3 m IKMT (Isaacs-Kidd midwater trawl) catching rates of the northern anchovy, Engraulis mor- dax. It is possible that by going to still larger nets (i.e., >10 m 2 mouth areas), a reduction in the bias could be effected. However, larger nets would be impractical, if not impossible, to handle on most oceanographic vessels. As Barkley (1964) has demonstrated, in- creased net speed is not a feasible strategy for avoidance reduction since increasing the towing speed of a net requires a compensatory reduction in net size. The practical limits to increasing the tow speed are reached at 2 to 3 kn, because of un- avoidably extreme wire angles and inordinate amounts of wire required to fish even at moder- ate depths (to 1,000 m). High speed tows gener- ally result in damaged specimens, reducing their value in studies requiring taxonomic identifica- tion or in physiological and biochemical mea- surements. Finally, as speed of net is increased, the effects of escapement through the meshes is enhanced. Another means of reducing avoidance, that of camouflaging the net to reduce an animal's abil- ity to detect its presence and thereby reducing the avoidance reaction distance, has been dis- cussed briefly by Clutter and Anraku (1968). There is evidence that it may be an effective strategy for species such as N. rae.ga/o/xs'(LeBras- seur and McAllister, unpublished data cited by Clutter and Anraku 1968). To use this approach, one must first know what kind of a signal the ani- mal is using to detect the oncoming net. Camou- flaging the net can be accomplished by reducing the signal until it becomes part of the back- ground (omnidirectional noise). Alternatively, the noise level could be increased until the signal is no longer detectable. Signals emanating from a net and towing cable include deformation of flow, near field (displacement dominant) or far field (pressure dominant) sound, and light (bioluminescence) (Clutter and Anraku 1968). The importance of 88 WIEBE ET AL.: AVOIDANCE OF TOWED NETS BY NEMATOSCELIS MEGALOPS these different signals obviously depends upon the net structure and towing cable configuration and upon the ability of N. megalops to sense the various signals. Although there is no direct ex- perimental information about N. megalops' sen- sory capabilities or about the signals being gen- erated by MOCNESS, it seems clear that the primary avoidance stimulus involves day to night variations in light. Nemaioscelis megalops must use vision to detect the net and can better avoid the net during the day than at night be- cause during the day the net is better illumi- nated. A fundamental link between the amount of light present and the magnitude of the avoid- ance is provided by our observation that as indi- viduals live deeper in the water column under substantially reduced daytime light levels, day/ night differences in catch rates decline. But if we accept the results gained by the application of Barkley's model which indicate substantial avoidance takes place at night in the absence of bright sunlight, then other factors must also be important. We propose that bio- luminescence is the principal signal and that vision remains the principal means of detection. Three linesof evidence support the importance of bioluminescence as an avoidance cue. First, in an experiment conducted in the early 1960's, Boden (1969) equipped an IKMT with light meters so that he could monitor the amount of light produced above, below, in front of, and in- side the trawl as it was towed at night. Biolumi- nescent light above the trawl was less than below the trawl but both were considerably lower than that ahead of or in the net. Light within the net was so bright that it recorded off scale and indi- vidual flashes were often too numerous to be re- corded as such. Light ahead of the net was also exceedingly bright. Boden (1969) speculated that the light ahead of the net was caused by or- ganisms flashing either in response to the light within the net or to pressure or sound waves propagating forward from the net. Second, Neshyba's (1967) experiments with a submarine photometer and strobe light showed that meso- pelagic and epipelagic organisms could be stimulated to produce significant amounts of bio- luminescence (10 4 juW/cm 2 ) for a sustained pe- riod by proper strobe light flashing. In the ab- sence of artificial flashing, he observed a much lower level of irregular flashing (10 8 -10 7 /zW/ cm 2 ) similar to that reported by Kampa and Boden (1956), Clarke and Backus (1964), and Boden et al. (1965). Third, it is known that the eyes of euphausiids and decapod shrimps living at midwater depths during the day (i.e., 200-600 m)are sensitive to light levels (10 7 to possibly 10 9 yuW/cm 2 , Clarke 1970) significantly lower than that produced as a result of bioluminescence. These lines of evidence suggest that the light generated by organisms when they come in di- rect contact with the nets or encounter turbu- lence caused by the net is used by individuals ahead of the net to detect its presence and begin an avoidance response. It seems likely that the light ahead of the net observed by Boden (1969) was caused by the same kind of response mech- anism described by Neshyba(1967), i.e., flashing in response to flashing. The tactic of reducing the visual contrast be- tween a net and the surrounding water was demonstrated by LeBrasseur and McAllister (unpublished data cited by Clutter and Anraku 1968) to reduce the avoidance error for euphau- siids both day and night. However, if biolumi- nescence in and ahead of the net is an important cue as we suspect it to be, then a more active means of camouflaging the net is required. It is known from recent evidence (Warner et al. 1979) that decapod Crustacea living at the same depth as N. megalops are easily "blinded" by even moderate amounts of light. This suggests the possibility of equipping the mouth of a net with a "blinding" light system to be used to periodically illuminate a region ahead of the net with enough light to temporarily blind individuals in the net. With the light out, individuals so affected by the light pulse would be unable to see and, therefore, to respond to the much lower light generated by zooplankton being captured by the net. We pos- tulate that individuals outside the zone of tem- porary blindness may respond by electing a startle response, but, because the volume illumi- nated would be so large, their movement would be random with respect to the volume to be fil- tered by the net. Clearly, considerably more re- search is required before this strategy could be considered feasible. There are two precautionary notes that must be made. First, in spite of avoidance error, verti- cal distribution patterns obtained in sampling this species with MOCNESS at different times under different hydrographic regimes are repli- cable (Fig. 3). That is, although avoidance error is strongly affecting the numerical estimates, the shape of the vertical distributions seem much less affected. Thus, in spite of the avoidance, we believe we are obtaining valuable ecological in- 89 FISHERY BULLETIN: VOL. 80, NO. 1 formation about this species. Second, for most species of euphausiids and many copepods, chae- tognaths, and pteropods in our collections, we have no evidence that differential day/night avoidance is taking place. Therefore, for many ecological studies of oceanic zooplankton, nets still seem the most effective tool to use to quanti- tatively collect them. ACKNOWLEDGMENTS We gratefully acknowledge the assistance given us by R. Backus and J. Craddock in work- ing up the MOCNESS 10 samples. L. Haury, A. Morton, J. Wormuth, A. Hart, and C. Polloni pro- vided valuable assistance in making the MOC- NESS 1 collections at sea, and G. Flierl provided helpful suggestions for interpreting the data. We thank C. Miller and L. Haury for critically read- ing the manuscript and we thank the officers and crew of the RV Knorr and RV Chain for their skillful operation of the vessels. This study was supported by the Office of Naval Research con- tracts N00014-66-CO241 NRO38-004, N00014- 74-C0252 NR083-004, and N00014-79-C-0071 NRO83-004 and the National Science Founda- tion grants DES74-02793 AOl and OCE77- 09132. LITERATURE CITED Barkley, R. A. 1964. The theoretical effectiveness of towed-net sam- plers as related to sampler size and to swimming speed of organisms. J. Cons. 29:146-157. 1972. Selectivity of towed-net samplers. Fish. Bull., U.S. 70:799-820. Boden, B. P. 1969. Observations of bioluminescence on SOND 1965 cruise of R.R.S. 'Discovery'. J. Mar. Biol. Assoc. U.K. 49:669-682. Boden, B. P., E. M. Kampa, and J. M. Snodgrass. 1965. Measurements of spontaneous bioluminescence in the sea. Nature (Lond.) 208:1078-1080. Boyd, S. H., P. H. Wiebe, and J. L. Cox. 1978. Limits of Nematoscelis megalops in the Northwest Atlantic in relation to Gulf Stream cold core rings. II. Physiological and biochemical effects of expatriation. J. Mar. Res. 36:143-159. Brinton, E. 1967. Vertical migration and avoidance capability of eu- phasiids in the California Current. Limnol. Oceanogr. 12:451-483. Clarke, G. L. 1970. Light conditions in the sea in relation to the diurnal vertical migrations of animals. InG. B. Farquhar(edi- tor), Proceedings of an international symposium on bio- logical sound scattering in the ocean, p. 41-48. U.S. Gov. Print. Off., Wash., D.C. Clarke, G. L., and R. H. Backus. 1964. Interrelations between the vertical migration of deep scattering layers, bioluminescence, and changes in daylight in the sea. Bull. Inst. Oceanogr. Monaco 64(1318), 36 p. Clutter, R. I., and M. Anraku. 1968. Avoidance of samplers. In Zooplankton sam- pling, p. 57-76. UNESCO Monogr. Oceanogr. Metho- dol. 2. Fleminger, A., and R. I. Clutter. 1965. Avoidance of towed nets by zooplankton. Limnol. Oceanogr. 10:96-104. Haury, L. R., J. A. McGowan, and P. H. Wiebe. 1978. Patterns and processes in the time-space scales of plankton distributions. In J. H. Steele (editor), Spatial pattern in plankton communities, p. 277-327. Plenum Press, N.Y. Isaacs, J. D. 1965. Larval sardine and anchovy interrelationships. Calif. Coop. Oceanic Fish. Invest. Rep. 10:102-140. Kampa, E. M., and B. P. Boden. 1956. Light generation in a sonic-scattering layer. Deep-Sea Res. 4:73-92. KlLS, U. 1979. Swimming speed and escape capacity of Antarctic krill, Euphausia superba. Meeresforschung 27:264- 266. Lai, D. Y., and P. L. Richardson. 1977. Distribution and movement of Gulf Stream rings. J. Phys. Oceanogr. 7:670-683. Laval, Ph. 1974. Un modele mathematique de l'evitement d'un filet a plancton, son application pratique, et sa verifi- cation indirecte en recourant au parasitisme de l'amphi- pode hyperide Vibilia armata Bovallius. J. Exp. Mar. Biol. Ecol. 14:57-87. Mackas, D. L., and C. M. Boyd. 1979. Spectral analysis of zooplankton spatial hetero- geneity. Science (Wash., D.C.) 204:62-64. Mauchline, J. 1967. Volume and weight characteristics of species of Euphausiacea. Crustaceana 13:241-248. McEwen, G. F., M. W. Johnson, and Th. R. Folsom. 1954. A statistical analysis of the performance of the Fol- som plankton sample splitter, based upon test observa- tions. Arch. Meteorol. Geophys. Bioklimatol., Ser. A Meteorol. Geophys. 7:502-527. McGowan, J. A., and V. J. Fraundorf. 1966. The relationship between size of net used and esti- mates of zooplankton diversity. Limnol. Oceanogr. 11: 456-469. Murphy, G. I., and R. I. Clutter. 1972. Sampling anchovy larvae with a plankton purse seine. Fish. Bull., U.S. 70:789-798. Neshyba, S. 1967. Pulsed light stimulation of marine biolumines- cence in situ. Limnol. Oceanogr. 12:222-235. Ortner, P. B., P. H. Wiebe, L. Haury, and S. Boyd. 1978. Variability in zooplankton biomass distribution in the northern Sargasso Sea: The contribution of Gulf Stream cold core rings. Fish. Bull., U.S. 76:323-334. Richardson, P. L. 1980. Gulf Stream ring trajectories. J. Phys. Oceanogr. 10:90-104. 90 WIEBE ET AL.: AVOIDANCE OE TOWED NETS BY NEMATOSCEUS MEGALOPS RlCKER, W. E. 1973. Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. Semenov, V. N. 1969. Observations of krill behaviour in an aquarium. Tr. VNIRO 66:235-239. Smith, P. E., R. C. Counts, and R. I. Clutter. 1968. Changes in filtering efficiency of plankton nets due to clogging under tow. J. Cons. 32:232-248. Vannucci, M. 1968. Loss of organisms through the meshes. In Zoo- plankton sampling, p. 77-86. UNESCO Monogr. Oceanogr. Methodol. 2. Warner. J. A., M. I. Latz, and J. F. Case. 1979. Cryptic bioluminescence in a midwater shrimp. Science (Wash., D.C.) 203:1109-1110. Wiebe, P. H. 1971. A computer model study of zooplankton patchiness and its effects on sampling error. Limnol. Oceanogr. 16:29-38. Wiebe, P. H., and S. Boyd. 1978. Limits of Nematoscelis megalops in the Northwest- ern Atlantic in relation to Gulf Stream cold core rings. Part I. Horizontal and vertical distributions. J. Mar. Res. 36:119-142. Wiebe, P. H., K. H. Burt, S. H. Boyd, and A. W. Morton. 1976. A multiple opening/closing net and environmental sensing system for sampling zooplankton. J. Mar. Res. 34:313-326. Wiebe, P. H., and W. R. Holland. 1968. Plankton patchiness: Effects on repeated net tows. Limnol. Oceanogr. 13:315-321. Wiebe, P. H., G. D. Grice, and E. Hoagland. 1973. Acid-iron waste as a factor affecting the distribu- tion and abundance of zooplankton in the New York Bight. II. Spatial variations in the field and implications for monitoring studies. Estaurine Coastal Mar. Sci. 1: 51-64. 91 AGE AND GROWTH OF A PLEURONECTID, PAROPHRYS VETULUS, DURING THE PELAGIC LARVAL PERIOD IN OREGON COASTAL WATERS Joanne Lyczkowski Laroche, Sally L. Richardson, 1 and Andrew A. Rosenberg- ABSTRACT The age of 331 field-collected English sole, Parophrys vetulus, larvae, 3.1-20.0 mm SL, was deter- mined using daily otolith growth increments. Age in days from hatching was estimated by adding 5, the number of days prior to first increment formation in the laboratory, to the number of increments counted on sagittae. Number of otolith growth increments among larvae of known age in the labo- ratory ranged widely. Yet daily periodicity of increment formation in P. vetulus was inferred from the observations that even under poor growing conditions some larvae added one increment each day since first formation and that, unlike the remaining laboratory-reared larvae in which no pattern was evident, increment addition among larvae in the sea appeared to follow a stable and uniform pattern. Gompertz and von Bertalanffy growth models fitted the resultant size-at-age data equally well; therefore, only the Gompertz model is presented. Larval growth rate decreased from 0.3 mm per day at 8-9 days of age to <0.1 mm per day between 73 and 74 days. The oldest specimen was 74 days old. but most of the larval and transforming specimens collected in plankton samples were <70 days old . Previous estimates of age at length of larval P. vetulus, based on length-frequency modal progres- sion analysis, overestimated the age of larvae >5.5 mm SL by 2-3 times and, correspondingly, the duration of pelagic life was overestimated, 18-20 weeks compared to 8-10 weeks based on otolith- estimated age. Saccular otoliths grow by addition of layers of material differing in the relative amount of the protein, otolin, and calcium carbonate in the aragonite form (Degens et al. 1969; Pannella 1971). This results in growth units or increments composed of an inner light band and an outer dark band. Once the cycle of formation has been established for a species, otolith growth incre- ments can be used to estimate a fish's age and as a record of its past growth. Daily periodicity of increment formation has been confirmed in numerous species by the number of first-order growth increments within annuli in fish over 1 yr of age (Pannella 1971, 1974), by inspection of otoliths from reared fish of known age (Brothers et al. 1976; Taubert and Coble 1977), or from fish maintained in the laboratory for a known period of time (Struhsaker and Uchiyama 1976). Bands of daily increments are often grouped into fort- nightly and monthly growth patterns (Pannella 'Gulf Coast Research Laboratory, East Beach Drive, Ocean Springs, MS 39564. department of Oceanography, Dalhousie University, Halifax, N.S., Canada. Manuscript accepted August 1981. FISHERY BULLETIN: VOL. 80, NO. 1. 1982. 1974; Rosenberg 1980). Subdaily increments, which appear faint and indistinct, when com- pared to daily increments, have been found in some species (Taubert and Coble 1977; Brothers and McFarland in press). The daily increment method of aging larval and juvenile fishes can be used in fishery re- search to document the timing and duration of spawning, development, and major life history stages and events. The singlemost important application is the accurate determination of growth rates during early life in the sea. This technique has been applied to relatively few species, however, and much remains to be learned about how growth may change during development and under varying environmental conditions. Once specific growth rates are avail- able, age-dependent larval mortality rates can be estimated and used to improve estimates of spawning stock biomass and also, perhaps, pro- vide insight into recruitment success. This paper documents the existence of daily growth increments in laboratory-reared and field-caught larvae of an eastern North Pacific pleuronectid, the English sole, Parophrys vetulus. It provides the first accurate estimates of 93 FISHERY BULLETIN: VOL. 80. NO. 1 age at length for larvae of this species and describes the growth of larvae collected in Oregon coastal waters during the 1977-78 spawning season. It is the first detailed study of larval growth of a pleuronectid throughout the pelagic period, and further, provides a basis for the documentation of growth during trans- formation to the adult form (Rosenberg and Laroche 1982) and of juveniles in nursery grounds off the Oregon coast (Rosenberg 1980). METHODS Spawning and Rearing Procedures Ripe adult P. vetulus were collected during fall and winter 1978 with a 12 m otter trawl off the Oregon coast in the vicinity of Hecata Head, approximately lat. 44°10'N, long. 124°18'W, 68-77 m water depth. Eggs were artificially fer- tilized on shipboard (Bagenal and Braum 1971) and transported back to the laboratory in sea- water-filled plastic bags. In the laboratory, eggs were incubated and larvae reared at 12°-13°C and under a 14-h light, 10-h dark photoperiod in filtered seawater taken from the area where the adults were captured. Eggs held in 4 1 glass jars hatched in 3-3% d. The newly hatched larvae were transferred by pipette to new 4 1 glass jars or 8 and 9 1 plastic tubs in which a bloom of the green flagellate Tetraselmis sp. was maintained throughout the rearing period. Approximately every 2 d, one- fourth to one-third of the water in rearing con- tainers was replaced. On day 4 after hatching, Gymnodinium splendens, a naked dinoflagellate, and Brachionus plicatilis, a rotifer, were intro- duced into the rearing containers. After 1-2 wk, G. splendens was no longer added because larvae did not appear to eat this organism. Prey con- centrations were not measured but B. plicatilis, the primary food item, was maintained at high levels, i.e., rotifers were readily visible through- out rearing containers. Artemia salina nauplii and the harpacticoid copepod, Tisbe sp., pro- vided secondary food sources. One to ten larvae were preserved in ~80% ethanol each day after hatching for the first 35 d; subsequently, older larvae were preserved at irregular intervals. Larvae were reared from two separate spawnings, in early and late fall 1978, but since rearing conditions were identi- cal, age and growth data from the two were com- bined. Field and Laboratory Procedures Parophrys vetulus larvae were collected in the field with 70 cm, 0.505 mm mesh bongo nets in bottom to surface stepped, oblique tows. Samples were taken approximately monthly from November 1977 to June 1978 in Yaquina Bay, Oreg., and 2-7 km offshore (lat. ~44°37'N; long. 124°05'W). Samples were drained and pre- served in ~80% ethanol; within 12-18 h the samples were drained again and fresh preserva- tive was added. With each plankton sample sur- face temperature, surface and bottom salinity were recorded and a bathythermograph cast was made. In the laboratory all fish larvae were removed from plankton samples and stored in ~80% ethanol. Otoliths were removed from P. vetulus larvae within 6 mo of initial preservation be- cause longer storage resulted in erosion or com- plete dissolution of the otoliths. Prior to otolith removal P. vetulus larvae were placed in freshwater for ~l-2 min (somewhat longer for specimens >15 mm) to remove or di- lute ethanol in the tissue. A larva was then placed in a drop of water on either a glass slide or large rectangular cover slip under a dissecting micro- scope fitted with polarizing filter and analyzer. Standard length (SL) was measured with an ocular micrometer to the nearest 0.1 mm and both sagittae were dissected out with fine probes at 25 X or 50 X magnification. The larva was re- moved from the slide or slip and the otoliths were left to dry concave side up. Sagittae were then permanently mounted under a cover slip with Pro-Texx, 3 a clear mounting medium. Rectan- gular cover slip mounts, which were thought to improve the optical properties of the preparation, were taped for support to a thin piece of brass for viewing under the microscope. Otolith growth increments, consisting of an inner light band and a narrower, sharply delineated, continuous outer dark band adjacent to it, were counted using a compound micro- scope with bright field illumination at 800 X or 1,250 X magnification. Faint bands inside the otolith nucleus in reared larvae and "subdaily" or weak rings between well-defined growth incre- ments in some older (>30 d old) field-caught fish were not counted. Counts were made on only one sagitta of the pair and were repeated until a :i Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 94 I^KIK'HK KT AL.: AGE AND GROWTH OF PAROPHRYS VETULUS final, "best" count was reached. Successive counts and verification counts which were made by the original reader at a later time usually did not vary by more than ±2. Age estimates could not be obtained for 10% of the field-caught larvae because increments were faint and indistinct or the otoliths were misshapen. Maximum otolith and nucleus diameters were measured to the nearest micron. Photomicrographs were taken at 500 X or 1,000 X magnification under a light microscope. Shrinkage of larvae preserved in 80% ethanol was compared with shrinkage after preservation in 10% seawater-diluted Formalin, the fixative most commonly used to preserve plankton samples. Thirty 7-day-old reared larvae were measured alive and immediately preserved in either 80% ethanol (15) or 10% Formalin ( 15). The live, mean standard lengths of the two groups of larvae were 4.34 and 4.42 mm. After 4 mo in preservative the mean standard length of the ethanol-preserved group was 4.20 mm and of the Formalin-preserved group, 4.196 mm. Mean percent shrinkage or 100(original SL — pre- served SL/original SL) was 3.2% in the ethanol- preserved group and 5.1% in the Formalin- preserved group. The difference in amount of shrinkage between the two groups was highly significant (ANOVA, P<0.01). Care must be taken, therefore, when comparing estimates of size at age based on measurements of larvae pre- served in different fixatives. From this limited investigation it became apparent that Formalin- preserved P. vetulus larvae appear to be some- what smaller at age than ethanol-preserved fish. Statistical Procedures Gompertz and von Bertalanffy growth models were fitted to larval P. vetulus data because the form of the length-age plot was nonlinear with a distinct upper asymptote. A detailed discussion of the Gompertz function, which is the primary model used in this paper, and methods for ob- taining initial parameter estimates are pre- sented by Zweifel and Lasker (1976). The gen- eralized equation of this model is: L, = Loexp[x(l-e- Q ')j, where L, = length at age f; Lo = length at t = (i.e., where the curve intercepts the y-axis); and K= — > or the specific growth rate at t = divided by the rate of exponential decay. Un- transformed data were used in this model be- cause the standard deviation of larval lengths at age remained relatively constant and did not in- crease with age, indicating variance homogene- ity within the data set. The Gallucci and Quinn (1979) version of the von Bertalanffy equation was employed, utilizing the new parameter, w = kL x , where k is the growth constant, and L x , the asymptotic maximum size, which for P. vetulus larvae is the maximum size attained in the plankton prior to transformation into benthic juveniles. The general form of this equation is: L t = f (l - exp [- kit - *>)] I . where t is the time when Lo = (i.e., where the curve intercepts the x-axis). The SPSS NONLINEAR 4 program employ- ing Marquardt's algorithm was used to fit both models. A measure of goodness of fit was pro- vided by the residual sums of squares (RSS), the standard error of the regression (or standard deviation of the residuals), and approximate 95% confidence limits for each parameter assuming linearity. Linear confidence theory can be ap- plied here because the assumption of linearity at the final (least squares) parameter values is a reasonable one (Conway et al. 1970; Kimura 1980). A comparison of the RSS at the final pa- rameter values to the linear estimate RSS pro- vides a measure of the linearity of the sum of squares (SS) function (SPSS NONLINEAR pro- gram). Absolute growth rate or k-U expressed in millimeters per day and specific growth rate or In La — In L\ k-U X 100 expressed as percent per day of length were calculated (Ricker 1979). RESULTS Increment Formation Parophrys vetulus larvae survived and grew in the laboratory for over 35 d after hatching, with some individuals eventually transforming into juveniles. However, growth after yolk-sac ab- sorption, between days 4 and 5, was retarded and 4 SPSS NONLINEAR. Statistical Package for the Social Sciences, Vogelback Computing Center, Northwestern Uni- versity, Evanston, IL 60201. 95 FISHERY BULLETIN: VOL. 80. NO. 1 not comparable to growth in the field. Despite this, growth increments were visible on the otoliths of over 300 reared larvae. Increments, though extremely narrow and crowded, were even visible on the otoliths of larvae as old as 54 d. In the laboratory, the highest incidence of larvae with one growth increment occurred on days 5 and 6 (Table 1; Fig. la, b). This coincided with the time that larvae first began to swim actively near the surface of rearing containers and search for food. By day 5 larvae had also acquired darkly pigmented, iridescent eyes and functional mouths, and had utilized all or almost all their yolk. Age at first increment formation in the field was ascertained by comparing mean otolith diameter (jum) of field-caught larvae with a single increment, to mean otolith diameter of laboratory-reared larvae of known age (Table 1 ). The otolith diameter, 23.8 nm, of field-caught larvae with only one growth increment (SL = 3.7 mm) fell between the mean values for lab- oratory-reared larvae at 5 d, 23.1 (SL = 4.2 mm), and at 6 d, 24.2 (SL = 3.9 mm). Age of all field-caught larvae with one otolith growth increment was, therefore, taken to be 6 d. Age at first increment formation varied among in- dividuals in the laboratory and may, likewise, vary in the field; however, for the purpose of developing a generalized growth model, a single, best estimate of this event was made. The appar- ent smaller size of field-caught larvae with one increment most likely resulted from shrinkage during capture prior to preservation (Theilacker 1980). Larvae sampled in the laboratory were pipetted alive directly into preservative, thus re- ducing the amount of handling-induced shrink- age. Although laboratory results were somewhat ambiguous, daily periodicity of otolith growth increment formation in P. vetulus was inferred from the following observations: 1) despite less than optimum rearing conditions some 14-, 17-, and 20-d-old larvae had added one increment each day since first formation on day 4 (Table 2); 2) no other periodical pattern in increment formation (i.e., other than daily) was observed among laboratory-reared larvae; 3) increment addition among larvae in the sea appeared to follow a stable and uniform pattern. The wide range in number of otolith increments among reared larvae of known age may have been caused by poor growing conditions which re- sulted in stunted body and otolith growth (Table 2). Reared larvae of northern anchovy also failed a Figure 1.— Photomicrographs of Parophrys vetulus otoliths (X 1,000). a. Sagitta (22 ^m in diameter) prior to first in- crement formation from a 4-d-old, laboratory-reared larva; b. Sagitta (24 /iiti in diameter) with two complete increments (highlighted with black lines) from a6-d-old, laboratory-reared larva; c. Sagitta (22 ^m in diameter) with two complete incre- ments (highlighted with black lines) from a 7-d-old, field- caught larva. 96 .AROC'HK ETAL.: AGE AND GROWTH OF PAROPHRYS VETULUS Table 1. — Comparison of mean otolith diameters (OD) of laboratory-reared and field-collected Pnrophrys vetulus larvae. Age of reared larvae represents days from hatching. Age Mean OD No No. growth increments Mean OD No No growth (days) (pm) larvae 1 2 3 4 (//no larvae Increments 146 10 10 1 16.6 12 12 2 188 10 9 1 3 205 11 10 1 4 21.6 14 13 1 21 3 7 5 23.1 24 10 10 4 23.8 4 1 6 24.2 19 4 4 8 2 1 246 10 2 to consistently form growth increments when maintained on low rations (Methot and Kramer 1979). In P. vetulus, delayed inception of incre- ment formation, up to 8 d after hatching, may also have accounted for some of the apparent irregularity in increment formation in the lab- oratory (Table 2). Another factor contributing to ambiguity of laboratory results was the diffi- culty in counting otolith increments in older larvae. Increments in most laboratory-reared fish after 16-25 d were exceedingly faint and, in some fish, no increments could be discerned (Fig. 2a, b). Growth increments were, in gen- eral, clearer and more distinct on the otoliths of field-caught P. vetulus larvae than on otoliths of laboratory-reared fish (Figs, lc, 2c). The steady increase in number of increments with increas- ing otolith diameter and length of pretransfor- mation larvae in the field is evidence that the irregularity in increment formation observed in the laboratory did not occur under natural feed- ing conditions (Figs. 3, 4). Age and Growth Age of field-caught P. vetulus larvae in days from hatching was estimated by adding 5, the number of days prior to appearance of the first otolith growth increment, to the number of in- crements counted on sagittae. Counts of growth increments were obtained from 338 larval and transforming, pelagic specimens ranging from 2.4 to 20.0 mm SL (Fig. 4). But age could be esti- mated for only 331 larvae because increment formation had not yet begun in seven small speci- mens, 2.4-3.7 mm SL (Fig. 5). The oldest P. vetulus taken in plankton samples during 1977- 78 was 74 d (2.4 mo) old and 17.8 mm SL. The next oldest larvae ranged from 65 to 70 d old and were 19-20 mm SL. The length of pelagic life of P. vetulus can be estimated directly from these data to be 2-2.5 mo. Few P. vetulus larvae >20 mm SL, the size at which larvae transform to benthic juveniles (Ahlstrom and Moser 1975; Rosenberg and Laroche footnote 3), were taken in extensive plankton collections off Oregon during the spring months in 1972-75 (Laroche and Richardson 1979). The largest larva taken in those collections was 22 mm SL. Behavior of reared P. vetulus larvae further supports a pelagic phase of 2+ mo. At approxi- mately 60 d of age, larval P. vetulus maintained in the laboratory first exhibited the tendency to rest on their sides on the bottom and to swim with their bodies at an angle to the vertical (J. L. Laroche unpubl. data). Table 2.— Summary of growth in body length (SL) and otolith diameter (OD), and counts of growth increments on otoliths of laboratory-reared Parophrys vetulus larvae. N= number of larvae from which growth increment counts were taken; (N) - number of larvae used in mean otolith diameter calculation. No growth Age (days) Mean SL Range SL (mm) Mean OD Range OD increments N (mm) (fjm) (fjm) Mea 1 Range 4 14 4.1 3.7-4.4 21.6 20-25 0-2 5 24 4.2 3.8-4.5 23.1 20-25 1 0-2 6 19(17) 39 3.1-4.2 24.2 23-27 2 0-4 9 9 4.0 3.7-4.1 24.7 23-26 3 2-4 10 13 4.2 37-46 25.7 25-27 4 3-6 14 13 49 5.8-4.2 287 27-33 8 5-10 17 13 54 45-63 296 28-33 10 5-13 20 7(6) 5.7 5 .1-6.4 31.8 30-36 10 5-16 21 6 59 5.4-6.6 31.2 30-34 13 10-16 26 18(17) 7.0 5.6-8.6 334 30-37 14 10-20 97 FISHERY BULLETIN: VOL. 80. NO. 1 * y a Figure 2.— Photomicrographs of Parophrys vetulus otoliths ( X 1,000). a. Sagitta (30 ^m in diameter) with 16 complete incre- ments from a 21-d-old, laboratory-reared larva; b. Sagitta (32 jum in diameter) with no discernible increments from a 22-d-old. laboratory-reared larva; c. Sagitta (102 ^m in diameter) with 42 complete increments from a 47-d-old, field-caught larva. 98 LA ROCHE ET AL.: ACE AND OROWTH OF PAROPHRYS VETULUS O 00 •00 O PJ OJ IO oo o co o • • • • .V*. •«« . 4 ID k- O 5 0) •a e j= bo 3 OS o bo C £ «- o If) to CM C C V) a! t O a > t= a <1 00 £ CO CO o O 111 t. o ^ 0) V 0> £ ~; ^ jt TT k 43 ^ -*j --j V^i +j k. o O 3 ca in in l O t_ O) 42 E 3 Z 1 CO w as (M S1N3W3UDNI HIMOdO JO ON 99 FISHERY BULLETIN: VOL. 80, NO. 1 70 50 I I a; 1 5 25 _..f> •a. i.. Mi. . 7"7' \ j i i i_ _i i i — i- 5 10 15 STANDARD LENGTH (mm) 20 FIGURE 4.— Number of otolith growth increments related to standard length of 338 larval and transforming, field-caught Parophrys vetulus. Our description of early growth of P. vetulus in Oregon coastal waters at temperatures ranging from 9° to 11°C is based on the ages and lengths of 331 specimens, 3.1-20.0 mm SL, with otolith growth increments. Gompertz and von Berta- lanffy models yielded good and nearly identical fits to the data and similar estimates of growth rate; therefore, the results of only one model (Gompertz) are presented (Table 3; Fig. 5). RSS and linear estimate RSS of the Gom- pertz growth parameters were very similar; thus, the assumption of linearity in computing 95% confidence limits is reasonable, and the computed limits indicate relatively narrow con- fidence regions around the parameters (Table 3). Previous estimates of age at length of larval P. vetulus were derived from the progression of modes in length-frequency distributions of larvae from a time series of (10% Formalin pre- served) plankton samples (Laroche and Richard- son 1979). A comparison of those results with age at length estimated by the Gompertz equation (Zwiefel and Lasker 1976) indicates that the length-frequency method overestimated the age of larvae >5.5 mm SL by 2-3 times (Table 4). Estimates of specific and absolute rates of growth were calculated from length at age for various ages as predicted by the Gompertz model (Table 5). Specific growth rate steadily de- creased between 8 and 74 d. Absolute growth rate was fairly uniform between 8 and 31 d, slowed somewhat between 31 and 41 d, but was more drastically reduced between 73 and 74 d, at which time larvae undergo transformation, a Table 3.— Gompertz equation and estimated parameters describing the growth of 331 Parophrys vetulus larvae in Oregon waters during the 1977-78 spawning season. RSS = residual sum of squares; SE = standard error of the regres- sion; S 2 = variance; CL = confidence limits. Equation L, = 2.073 exp[2.354 (1- Parameters S 2 _ e -0.045,,] RSS Linear est. RSS RSS SE 520 83 1256 Approximate 95% CL Lo = 2.073 K =2.354 a =0 045 0023 0003 0.00001 564892 535.762 533854 564 892 536.093 533.635 L, = 1.779, Li = 2.367 /_, =2.245, U =2.462 L, = 0.040, L 2 = 0.050 Table 4.— Age of Parophrys vetulus larvae; (A) estimated from modal progression in length-frequency distributions of larvae caught during 1971 in biweekly and weekly Formalin-preserved plankton samples (Laroche and Richardson 1979), (B) estimated by the Gompertz equation based on otolith increment counts from ethanol-preserved larvae caught in 1977-78. Estimated age (weeks) SL (mm) 5.5 7.5 95 11.5 13.5 15.5 17.5 w f • -z. : * • < " • v i • en 50 1 1 i : 1 i i 1 1 \ 2 354(l-e"" 0045 h L t = 2073 e 35U ' i i i 1 i i i i 25 50 ESTIMATED AGE (days) 75 Figure 5.— Gompertz curve and equation fitted to length at age of 331 larval and transforming, field-caught Parophrys vetulus with at least one otolith growth increment. period characterized by reduced growth in length (Rosenberg and Laroche 1982). The plot of otolith diameter on standard length of pelagic larval and transforming P. vetulus re- vealed an allometric relationship (Fig. 6). A dis- tinctive feature of this plot was the apparent con- tinued, even accelerated growth of sagittae as P. vetulus larvae reached the size of transforma- tion, 18-20 mm SL, when rate of growth in body length slows down. Physical evidence of acceler- ated growth in otolith diameter relative to body length can be seen by the increased width of the outermost increments on otoliths of larvae older than 30 d (e.g., outer 9-10 increments on sagitta in Fig. 2c). The otolith diameter to standard length relationship, once a mathematical formu- lation has been computed, can be used to back- calculate individual growth histories of larvae and juveniles (Rosenberg 1980; Methot in press), as has been done for adult fishes (Tesch 1968; Ricker 1969). DISCUSSION As in numerous other temperate and some tropical species of fishes, growth increments on the otoliths of P. vetulus larvae appear to be formed daily after yolk-sac absorption when larvae become capable of exogenous feeding. Counts of these increments provide more pre- cise and accurate estimates of larval age and growth rates throughout the larval period than have previously been available. This informa- tion, when combined with abundance data, allows computation of age-dependent mortality rates resulting in more accurate estimates of larval mortality in the sea. Empirically, both the Gompertz and von Bertalanffy growth models fit the larval P. vetulus data well. Both yielded similar values for length at age and growth rates from which age- dependent mortality estimates can be made. There has been much disagreement, on theoreti- cal grounds, as to the appropriateness of either model for describing growth in fishes, although they are mathematically quite similar (e.g., Zweifel and Lasker 1976; Ricker 1979). Despite numerous attempts to attribute biological signif- icance to mathematical models of growth, the best criterion available for choosing a particular model is still goodness of fit to the data (Ricker 1979). In that respect, both models were appro- priate to this data set. A practical measure of the appropriateness of mathematical models is the relative accuracy and stability of pertinent parameter estimates (Gallucci and Quinn 1979). In the Gompertz 101 FISHERY BULLETIN: VOL. 80, NO. 1 230 208 - 228 . '213 210 193 I 150- 125 i 100- 75 50- 3. ~ • • • • • - • . • • '•2 •• • • .* V. •• 2 2* •—€ mm* • • «• 2* « .. «, • am • - - - -•- mm me • ••• • • • 25 nf •• • • *k *: - * : Figure 6.— Otolith diameter related to standard length of 338 larval and transforming, field-caught Parophrys vetulus. J L J I I L _1_ 10 15 STANDARD LENGTH (mm) 20 102 LARIH'HE ETAL.: A(!K AND CROWTH OF PAROPHRYS VETULUS model, the parameter L or the y-intercept has been used as an estimator of length at hatching (Zweifel and Lasker 1976). However, the value of this parameter, 2.07 mm SL, for the P. vetulus data set was low compared to mean hatching lengths of reared larvae: 2.60 (N = 11) and 2.91 (N = 10) mm SL at 12°-13°C (Laroche unpubl. data); and 2.85 (N = 25) mm TL at 10°-1 1 °C (Orsi 1968). Net-caught larvae on which the growth model is based would appear smaller at age be- cause of increased shrinkage during capture (Theilacker 1980). This may account for some of the difference in predicted and observed hatch- ing lengths. Another probable cause of this dis- crepancy is the lack of data points in the <6 d of age region of the plot, i.e., before growth incre- ment formation begins. The value of Lo is based on extrapolation beyond the actual data and may be, therefore, of questionable use as a measure of the appropriateness of this model. Comparison with larval growth in the field at similar temperatures of another pleuronectid, Pseudopleuronectes americanus, provided evi- dence that growth rates predicted by the Gompertz model for Parophrys vetulus are realistic. Larval Pseudopleuronectes americanus between the ages of 28 and 42 d, growing in large enclosures in Narragansett Bay at 10°- 15°C, had a specific growth rate of 1.9% per day of standard length (Laurence et al. 1979). The predicted specific growth rate of Parophrys vetulus of the same age, growing at9°-ll°C, was 2.2%. Larval Pseudopleuronectes americanus be- tween 28 and 42 d of age grew from 6.6 to 8.6 mm SL, while Parophrys vetulus larvae grew from 11.2 to 15.3 mm SL. Although these two species differ in size at age, both transform at approxi- mately the same age, 8-10 wk, and appear to grow at similar rates between 4 and 6 wk of age. Since length at hatching, ~2-3 mm SL, is similar for both species, higher rates of growth prior to and after 4-6 wk probably accounts for the greater size at age of P. vetulus and greater size at transformation, >18 mm SL in P. vetulus versus <10 mm for Pseudopleuronectes ameri- canus. A comparison of otolith-estimated and length- frequency derived age-at-length data indicated that the latter method overestimated age of Parophrys vetulus larvae >5.5 mm SL by 2-3 times. This resulted in a gross overestimate of duration of the pelagic life of this species, 18-22 wk (Laroche and Richardson 1979) compared to 8-10 wk based on the age data presented here. It is unlikely that these large differences are solely the result of different preservatives. Such a large discrepancy between the two methods demon- strates the serious inaccuracies that could result from attempts to estimate age and growth rates from length-frequency data. Such data predict- ably yield low estimates of growth, especially for species with protracted spawning, because of continual recruitment of small larvae to the pop- ulation. Problems of net avoidance by larger specimens further bias length-frequency dis- tributions. The otolith aging method developed in this study could be used further to investigate growth and survival among different cohorts of P. vetulus larvae. Spawning in this species is highly variable in both frequency and timing (Laroche and Richardson 1979). Peak spawning can be bimodal in some years with a 2-4 mo separation between peaks (Kruse and Tyler 5 ). Larvae pro- duced in those two peaks could develop and grow under very different temperature regimes and feeding conditions, which could result in two distinct groups of larvae differing in rates of growth, mortality, and relative contribution to that year class. ACKNOWLEDGMENTS The following individuals are gratefully ac- knowledged for their significant contributions to this study. Rindy Ostermann and Betsy B. Washington assisted in all phases of collection of ripe adult specimens, larval rearing, otolith removal, and data compilation. Rae Deane Leatham and Percy L. Donaghay instructed the first author in culture techniques and generously provided the equipment and space in their laboratory for parts of this work. Eric Lynn, National Marine Fisheries Service, NOA A, La Jolla, Calif., sent us cultures of larval food organisms. Gary Hettman, Oregon Department of Fish and Wildlife, kept us informed throughout the fall and winter of the likely locations of English sole spawning con- centrations. Waldo W. Wakefield and Marky Bud Willis assisted at sea in collecting spawning fish. Chip Hogue, Paul Montagua, Gene Ruff, and Andrew Carey provided laboratory space 5 Kruse, G. H., and A. V. Tyler. 1980. Influence of physical facotrs on the English sole (Parophrys vetulus) spawning season. Unpubl. manuscr., 25 p. Department of Fisheries and Wildlife, Oregon State University, Corvallis, OR 97331. 103 FISHERY BULLETIN: VOL. 80, NO. 1 and their microscope for otolith observations. Otolith photomicrographs were taken with the help and guidance of Michael D. Richardson, Naval Ocean Research and Development Activi- ties, Bay St. Louis, Miss. This work is a result of research sponsored by the Oregon State University Sea Grant College Program (04-8-M01-144), supported by NOAA Office of Sea Grant, Department of Commerce. LITERATURE CITED Ahlstrom, E. H., and H. G. Moser. 1975. Distributional atlas of fish larvae in the California Current region: flatfishes, 1955 through 1960. Calif. Coop. Oceanic Fish. Invest. Atlas 23, 207 p. Bagenal, T. B.. and E. Braum. 1971. Eggs and early life history. In W. E. Ricker (editor), Methods for assessment of fish production in fresh waters, 2d ed., p. 166-198. IBP (Int. Biol. Programme) Handb. 3. Brothers, E. B., and W. N. McFarland. In press. Correlations between otolith microstructure, growth, and life history transitions in newly recruited French grunts [Haemulon flavolineatum (Desmarest), Haemulidae]. Rapp. P.-V. Reun. Cons. Int. Explor. Mer. 178. Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Conway, G. R., N. R. Glass, and J. C. Wilcox. 1970. Fitting nonlinear models to biological data by Marquardt's algorithm. Ecology 51:503-507. Degens, E. T., W. G. Deuser, and R. L. Haedrich. 1969. Molecular structure and composition of fish otoliths. Mar. Biol. (Berl.) 2:105-113. Gallucci, V. F., and T. J. Quinn, II. 1979. Reparameterizing, fitting, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull., U.S. 77:765-776. Laroche, J. L., and S. L. Richardson. 1979. Winter-spring abundance of larval English sole, Parophrys vetulus, between the Columbia River and Cape Blanco, Oregon during 1972-1975, with notes on occurrences of three other pleuronectids. Estuarine Coastal Mar. Sci. 8:455-476. Laurence, G. C, T. A. Halavik, B. R. Burns, and A. S. Smigielski. 1979. An environmental chamber for monitoring "in situ" growth and survival of larval fishes. Trans. Am. Fish. Soc. 108:197-203. Methot, R. D., Jr. In press. Spatial covariation of daily growth rates of larval nothern anchovy, Engraulis mordax, and northern lanternfish, Stenobrachius leucopsarus. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178. Methot, R. D., Jr., and D. Kramer. 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull., U.S. 77:413-423. Orsi, J. J. 1968. The embryology of the English sole, Parophrys vetulus. Calif. Fish Game 54:133-155. Pannella, G. 1971. Fish otoliths: daily growth layers and periodical patterns. Science (Wash., D.C.) 173:1124-1127. 1974. Otolith growth patterns: an aid in age determina- nation in temperate and tropical fishes. In T. B. Bagenal (editor). The proceedings of an international symposium on the ageing of fish, p. 28-39. Unwin Brothers, Surrey, Engl. Ricker, W. E. 1969. Effects of size-selective mortality and sampling bias on estimates of growth, mortality, production, and yield. J. Fish. Res. Board Can. 26:479-541. 1979. Growth rates and models. In W. S. Hoar, D. J. Randall, and J. R. Brett (editors), Fish physiology, Vol. VIII, p. 677-742. Acad. Press, N.Y. Rosenberg, A. A. 1980. Growth of juvenile English sole, Parophrys vetulus, in estuarine and open coastal nursery grounds. M.S. Thesis, Oregon State Univ., Corvallis, 51 p. Rosenberg, A. A., and J. L. Laroche. 1982. Growth during metamorphosis of English sole, Parophrys vetulus. Fish. Bull., U.S. 80(1):150-153. Struhsaker, P., and J. H. Uchiyama. 1976. Age and growth of the nehu, Stolephorus purpureus (Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily growth increments of sagittae. Fish. Bull., U.S. 74:9-17. Taubert, B. D., and D. W. Coble. 1977. Daily rings in otoliths of three species of Lepotnis and Tilapia mossambica. J. Fish. Res. Board Can. 34: 332-340. TESCH, F. W. 1968. Age and growth. In W. E. Ricker (editor), Methods for assessment of fish production in fresh waters, 2d ed., p. 93-123. IBP (Int. Biol. Programme) Handb. 3. Theilacker, G. H. 1980. Changes in body measurements of larval northern anchovy, Engraulis mordax, and other fishes due to handling and preservation. Fish. Bull., U.S. 78:685- 692. ZWEIFEL, J. R., AND R. LASKER. 1976. Prehatch and posthatch growth of fishes — a general model. Fish. Bull., U.S. 74:609-621. 104 PHENOTYPIC DIFFERENCES AMONG STOCKS OF HATCHERY AND WILD COHO SALMON, ONCORHYNCHUS KISUTCH, IN OREGON, WASHINGTON, AND CALIFORNIA 1 R. C. Hjort and C. B. Schreck ABSTRACT Similarities in phenotypic characters (isozyme gene frequencies, life history, and morphology) among 35 stocks of coho salmon, Oncorhynchus kisutch, from Oregon, Washington, and California were compared by using agglomerative and divisive cluster analyses. Coho salmon stocks from similar environments were phenotypically similar. Five groups of stocks were identified by the agglomerative cluster analysis: 1) wild stocks from the northern Oregon coast, 2) wild stocks from the southern Oregon coast, 3) stocks from hatcheries that used wild coho salmon for an egg and sperm source, 4) stocks from large stream systems, and 5) hatchery stocks from the northern Oregon coast. Three trends were indicated by the clustering patterns: 1) stocks that were geographically close tended to be phenotypically similar, 2) stocks from large stream systems were more similar to each other than to stocks from smaller stream systems, independent of geographic proximity, and 3) hatchery stocks were more similar to each other than to wild stocks, and wild stocks were more similar to each other than to hatchery stocks. These trends may be useful to fishery managers for selecting donor stocks from hatcheries for transplanting to stream systems or transferring to other hatcheries. Individual phenotypic characters were correlated with characters of the stream sys- tems. Results of two agglomerative cluster analyses, one of certai n characters of the stocks and one of certain characters of the stream systems, demonstrated a lack of correspondence between stream types and stock phenotypes. Genetic diversity among stocks of anadromous salmonids (Simon and Larkin 1970) is a biologi- cal characteristic that is more frequently dis- cussed than used in fishery management. The tendency to return to native streams reduces gene flow among salmon populations and en- ables the individual stocks to adapt to the native stream systems. The mixing of stocks highly adapted to their native stream systems with other stocks, or transplanting them to other stream systems, may reduce the rate of return or survival rate of the donor stock (Ritter 1975 3 ; Bams 1976). If the survival rate of a salmon stock is related to its degree of adaptation to its stream system, fishery managers may be able to in- crease survival of hatchery fish by planting them in recipient streams having native stocks geneti- 'Oregon State University Agricultural Experiment Station Technical Paper Number 5477. Oregon Cooperative Fishery Research Unit, Oregon State University, Corvallis, OR 97331. Cooperators are Oregon State University, Oregon Department of Fish and Wildlife, and U.S. Fish and Wildlife Service. 'Ritter, J. A. 1975. Lower ocean survival ratio for hatch- ery reared Atlantic salmon (Salmo salar) stocks released in rivers other than their native streams. Int. Counc. Explor. Sea, Anadromous and Catadromous Fish Comm., C. M. 1975/ M 26, 10 p. cally similar to the planted fish. Higher survival should be especially important during the first several generations, while the transplanted stock is adapting to the recipient environment. An additional advantage of using genetically similar stocks might be a reduction in the intro- gression of divergent hatchery genotypes into wild stocks (Reisenbichler and Mclntyre 1977). Genetic descriptions of salmon stocks could benefit salmon management by assisting fishery managers in selecting hatchery stocks and in protecting wild stocks. Obviously, determination of genetic similarity among stocks is not now pos- sible for the entire genome; however, similarity can be estimated by comparing genetically re- lated characters. Two biochemical characters that vary among stocks of coho salmon, Oncor- hynehus kisutch, are transferrin (Utter et al. 1970) and phosphoglucose isomerase (PGI) (May 1975), the electrophoretic expressions both of which were established by breeding studies to be genetically determined. Life history and mor- phological characters also vary among salmonid stocks. Time of spawning (Roley 1973) and fre- quency of occurrence of jacks in the population (Feldmann 1974) both have a genetic basis in Manuscript accepted August 1981. FISHERY BULLETIN: VOL. 80. NO. 1. 1982. 105 FISHERY BULLETIN: VOL. 80. NO. 1 coho salmon but probably have an environmen- tal component as well. A genetic basis, as shown in rainbow trout, Salmo gairdneri, has also been established for numbers of vertebrae (Winter et al. 1980a), scales in the lateral series (Winter et al. 1980a), scale rows (Neave 1944), gill rakers (Smith 1969), branchiostegals (MacGregor and MacCrimmon 1977), and anal fin rays (Mac- Gregor and MacCrimmon 1977). Ricker (1970) hypothesized that the meristic characters of salmonids probably have both genetic and en- vironmental components. The difficulty of deter- mining the importance of these phenotypic characters to the fitness of the stock does not pre- clude the possibility that they could, through selection or pleiotrophic effects, have a bearing on fitness as suggested by Barlow (1961). The objective of this study was to characterize stocks of coho salmon by using enzyme gene fre- quencies, life history characters, and morpho- logical characters. Secondarily, we hoped this in- formation would help provide a basis for select- ing donor stocks in Oregon hatchery programs. The stocks were selected so that comparisons could be made among geographical areas and stream types and between hatchery and wild stocks. We calculated a measure of a phenotypic similarity and used cluster analysis to display the relationships among stocks. Because cluster analyses are arbitrary (Blackith and Reyment 1971), we used two clustering strategies. Factors affecting genetic similarity were hypothesized by determining environmental characteristics common to the similar stocks. Although our analysis is primarily systematic, we correlated the phenotypic characters with variables characteristic of the stream systems. Although correlations do not prove a functional significance, they are included here because in- ferences and hypotheses can be developed from the correlations for future studies. METHODS Sampling We evaluated 10 characters for 15 hatchery stocks (based on samples of 75-100 juvenile coho salmon of the 1976 brood from 14 hatcheries in Washington, Oregon, and California and 9 hatcheries from Oregon for the 1977 brood year) and 12 wild stocks (based on samples of 30-100 juvenile coho salmon of the 1976 and 1977 broods, collected by electrofishing from 12 Oregon stream systems). (See Figure 1 for locations of hatcheries and stream systems.) Because some of the hatcheries have used nonnative egg sources, and stream systems have been stocked with juvenile and adult coho salmon, few pure native stocks remain. We did not use hatchery stocks or • HATCHERY • WILD 1 N TEN MADR CALIFORNIA Figure 1.— Map indicating sample site locations of wild and hatchery coho salmon stocks. Location codes are as follows with the hatcheries in parentheses: ALSE. Alsea River (Fall Creek Hatchery); BEAV, Beaver Creek; BIGC, BigCreek(BigCreek Hatchery); COLM. Columbia River (Cascade Hatchery in 1976 and Bonneville Hatchery in 1977); COQL, Coquille River; COWL, Cowlitz River (hatchery stock reared at Cascade Hatchery in 1976 and Big Creek Hatchery in 1977); KLAM, Klamath River (Irongate Hatchery); MADR, Mad River (Mad River Hatchery); NEHA, Nehalem River; NEST, Nestucca River; NONE, North Nehalem River (North Nehalem River Hatchery); QUIL, Quilcene River (Quilcene River Hatchery); QUIN, Quinault River Hatchery); ROGU, Rogue River (Cole Rivers Hatchery); SALM, Salmon River Hatchery); SAND, Sandy River (Sandy River Hatchery); SILZ, Siletz River; TENM, Tenmile Lakes; TRAS, Trask River (Trask River Hatchery); TRIN, Trinity River (Trinity River Hatchery); UMPQ, Umpqua River (hatchery stock collected from Smith River and reared at Cole Rivers Hatchery). 106 IIJORT and SCHRECK: PHRNOTYI'IC DIFFFRFNC'KS AMONC COHO SALMON fish from tributaries of streams that had re- ceived a large supplement of a normative hatch- ery stock in the previous 6yr. This was to ensure that characterization of the genotype would re- flect environmental considerations rather than introgression of foreign stocks. Morphological Characters For each sample, 15 carcasses were frozen for later counts. Scales in the lateral series were counted in the second row above the lateral line, starting with the anteriormost scale and ter- minating at the hypural plate. Scales above the lateral line were counted from the anterior in- sertion of the dorsal fin to the lateral line. Anal ray counts did not include the short rudimentary anterior rays, and branched rays were counted as one. The total number of gill rakers on the first gill arch was recorded. Alizarin red was used to highlight rudimentary gill rakers. The total number of branchiostegal rays from both sides was counted. Vertebral counts, made on X-ray plates, included the last three upturned centra. Accuracy of morphological counts was checked by recounting two fish from each sample. If errors were found, additional fish from that sample were recounted to correct for any error. Electrophoresis Blood and white muscle samples were col- lected from the fish that were not used for morphological counts. The caudal peduncle was severed and the blood collected in heparin- ized microhematocrit tubes that were then cen- trifuged and stored at — 10°C. White muscle samples (1 cm 3 ) were removed from the anterior dorsal portion of the frozen carcasses, homo- genized with 2 or 3 drops of water, and then cen- trifuged to clear the supernatant. Only the blood serum and supernatant were used for elec- trophoresis. The methodology for electrophoresis of trans- ferrin and phosphoglucose isomerase followed the basic principles of May (1975) with some modifications by Solazzi. 4 The gel and elctrode buffers were described by Ridgway et al. (1970). Four genotypes of transferrin (AA, AC, CC, and BC) in the serum samples were interpreted ac- 4 Solazzi, M. F. 1977. Methods manual for the electro- phoretic analysis of steelhead trout (Salmo gairdneri). Oreg. Dep. Fish Wildl., Res. Sect., Inf. Rep. Ser. Fish. 77-7, 35 p. cording to Utter et al. (1970). Transferrin was recorded as the frequency of the "A" allele, since the "B" allele was relatively rare. The variant allele for the second locus of phosphoglucose isomerase, first observed in white muscle tissue by May (1975), was recorded as the frequency of this variant allele. Life History The life history characters we used were time of peak spawning and proportion of females in the adult population. We estimated the peak spawning times on the basis of interviews with district fishery biologists and hatchery man- agers. Whenever possible, we verified the esti- mates with spawning ground survey records and hatchery records. We stratified the peak spawn- ing times into five segments of 2 wk each. The proportions of adult females (3 yr olds) were estimated from hatchery records and spawning ground surveys. This character is an indirect measure of the proportion of jacks (males that mature at 2 yr of age) in the popula- tion. Populations with high proportions of jacks in a given year should have relatively higher pro- portions of females returning the next year. A direct measure of the proportion of jacks cannot be used because body size differences between jacks and 3-yr-old adults affect the catch in gill net fisheries, retention in hatchery holding ponds, recovery of carcasses on spawning ground surveys, and catch rate in sport fisheries. Environmental Data Stream characteristics include distance up- stream to spawning grounds, basin area, area and length of the estuary on the stream system, latitude, gradient, spring runoff, the presence or absence of the myxosporidan parasite, Cerata- myxa shasta, and the presence or absence of the following nine species of fish: carp, Cyprinus carpio; Oregon chub, Hybopsis crameri; north- ern squawfish, Ptycholcheilus oregonensis; speckled dace, Rhinichthys osculus; redside shiner, Richardsonius balteatus; largescale sucker, Catostomus macrocheilus; brown bull- head, Ictalurus nebulosus; largemouth bass, Micropterus salmoides; and striped bass, Morone saxatilis. To separate the populations that have short and potentially long swimming distances to the spawning grounds, we measured spawn- ing distances from the mouth of the stream sys- 107 FISHERY BULLETIN: VOL. 80. NO. 1 tem to the upper limit of coho salmon spawning, as estimated from Anadromous Fish Distribu- tion Maps 5 and interviews with district fishery biologists. Inasmuch as, intuitively, latitude should be correlated with the temperature and flow regimes of the stream systems, we deter- mined the latitude at the mouth of each stream system. Gradients were calculated from tide- water to the upper limit of coho spawning as a basis for estimating the difficulty of the spawn- ing migration. Because estuary size and length is an estimate of exposure to vibriosis (Harrel et al. 1976) and potential richness of feeding grounds (Myers 1979), we measured the estuary lengths, stream elevations, and distances on United States Geographical Survey Quadrangle Maps. Inasmuch as high flows could have an effect on both the early life history and the smolting proc- esses of juvenile coho salmon, we determined the presence of a spring runoff from snowmelt by interviewing district biologists. We obtained in- formation on the distributions of other fish species in Oregon stream systems from C. E. Bond, 6 and on the distribution of Ceratamyxa shasta from J. E. Sanders. 7 We obtained temperature data from hatchery records to help interpret the morphological data for the hatchery stocks. The average tempera- ture for the first month of incubation was used, because previous studies have indicated that this time is a period during ontogeny when morpho- logical features may be most sensitive to the effects of temperature (Taning 1952). Statistics We calculated averages for the morpho- logical characters, enzyme gene frequencies, and the proportion of females for each stock, and used multivariate analysis of variance and Rao's (1970) test for additional information to deter- mine whether morphological characters differed significantly among stocks. In Rao's test, the statistical significance of each morphological character is determined, given that the other morphological characters are already in the model. Because environmental data on spawn- 5 Anadromous Fish Distribution Maps. Oregon State Water Resources Board, Salem, Oreg. ^arl E. Bond, Professor of Fisheries, Oregon State Univer- sity, Corvallis. OR 97331, pers. commun. April 1979. 7 James E. Sanders, Assistant Fish Pathologist, Oregon Dep. Fish Wildl., Corvallis, OR 97331, pers. commun. February 1979. ing distance, estuary length, estuary size, basin area, and gradient were skewed, we transformed them to natural logarithms to stabilize the vari- ance and improve normality. We standardized the stock characters (z = 0, S 2 = 1) for the cluster analyses, using the standard normal standardization. This standardization expresses the stock character as standard deviations from the character means, thus giving equal weight to each character. We calculated correlation coeffients (Snedecor and Cochran 1967) between the stock characters and the environmental data for all stocks, and between the morphological characters and tem- perature data for hatchery stocks only. The levels of significance for the correlation coefficients were also calculated as described by Snedecor and Cochran (1967). We used two cluster analysis programs to dis- play similarities among stocks. One, a nonhier- archical divisive cluster analysis, minimized the total sum of squares between observations and the cluster means (Mclntire 1973). In the other, a hierarchical agglomerative cluster analysis, Euclidean distance was used as the dissimilar- ity measure, and the clustering strategy was group average (see Sneath and Sokal [1973] or Clifford and Stephenson [1975] for terminology). Standardized data were used in both programs. We used canonical variate analysis to investi- gate the relation among the clusters from the agglomerative cluster analysis (Clifford and Stephenson 1975). Canonical variate analysis produces canonical variables that project groups of multivariate data onto axes separating the groups as much as possible. We plotted the ca- nonical variables against each other in two- dimensional space to determine the relationships among clusters and the discreteness of the clusters. RESULTS AND DISCUSSION Morphological Characters Significant differences (a = 0.01) for each morphological character (Tables 1-3) as indi- cated by multivariate analysis of variance and Rao's test of additional information existed among the 35 samples which consisted of wild and hatchery stocks from two brood years. When morphological characters for each stock between brood years were compared for each of the hatch- eries that were sampled in both years of the study 108 HJORTand SCHRECK: PHENOTYPIC DIFFERENCES AMONG COHO SALMON TABLE 1.— Means, standard errors (in parentheses), and ranges for the morphological characters of the 1976 brood year hatchery samples of juvenile coho salmon and the hatchery water incubation temperatures for the first month of incubation. Sample sizes were 15. The data are listed in north to south order of the sampling locations. Stock and (in parentheses) incubation water tempera- tures (°C) Scales in lateral series Scales above lateral line Anal rays Gill rakers Branchi- ostegals Vertebrae Washington Quilcene River Hatchery (7.3) 126 93 (97) 116-132 28.13 (38) 25-30 14.07 (15) 13-15 2233 (.19) 21-23 27 87 (26) 26-30 64 40 (13) 64-65 Quinault River Hatchery (7 3) 13267 (48) 130-136 2993 (.33) 28-32 1353 (16) 13-15 22 53 (.24) 21-24 2667 (.21) 25-28 6550 (.13) 65-66 Oregon: Cascade Hatchery (6.9) (Columbia River) 13307 (.56) 128-135 28 20 (40) 26-32 1400 (.20) 12-15 22.20 (.35) 20-25 27.27 (37) 25-29 6680 (.22) 66-68 Big Creek Hatchery (6.4) (Columbia River) 132 67 (.57) 128-136 2880 (.28) 28-31 14.31 (.13) 13-15 2287 (17) 22-24 2607 (30) 24-28 65 80 (15) 65-67 Cowlitz Hatchery stock. Cascade Hatchery (6.9) 133 60 (.63) 131-137 27 87 (.24) 26-29 1380 (.14) 13-15 22.20 (.26) 21-24 28 13 (.24) 27-30 6447 (22) 65-68 Sandy River Hatchery (7.0) (Columbia River) 133.13 (72) 128-137 2827 (.42) 24-30 14.33 (13) 14-15 22.07 (.34) 20-25 28 20 (26) 26-30 6607 (21) 65-67 North Nehalem River Hatchery (7.8) 131 93 (64) 128-138 28 33 (30) 26-36 1400 (14) 13-15 2267 (.27) 21-24 2673 (32) 24-28 65 80 (.17) 65-67 Trask River Hatchery (9 8) 132 13 (48) 128-135 2880 (47) 26-32 13.93 (12) 13-15 22 13 (31) 20-24 26 40 (.32) 24-29 6607 (.18) 65-67 Salmon River Hatchery (6.2) 129 40 (54) 125-132 27 00 (.59) 23-33 13 60 (.19) 13-15 22 13 (.24) 21-24 2540 (.24) 24-27 64 93 (30) 62-66 Fall Creek Hatchery (5 7) (Alsea River) 132 00 (.50) 129-135 2867 (.29) 27-31 1400 (.17) 13-15 23 20 (.20) 22-25 27.13 (34) 25-29 65 80 ( 17) 65-67 Umpqua Hatchery stock (Smith River), Cole Rivers Hatchery (3 5) 131 20 (51) 127-134 26 00 (.34) 24-28 13.47 (.19) 13-15 22 13 (.22) 21-23 25 13 (24) 24-26 65 33 (.23) 64-67 California: Irongate Hatchery (5.3) (Klamath River) 13273 (78) 129-138 2907 (18) 28-30 1380 (.14) 13-15 2233 (25) 21-24 27 00 (28) 25-28 6607 (30) 64-68 Trinity River Hatchery (7.3) (Klamath River) 130 87 (.75) 126-137 2827 (.64) 24-33 1360 (.13) 13-14 22 00 (.31) 19-23 26.00 (59) 20-28 66 00 (.14) 65-67 Mad River Hatchery (8.5) 129.20 (.88) 121-134 25.27 (.37) 22-27 13.40 (19) 12-15 2093 (.33) 19-23 23.47 (.51) 20-27 65.60 (.31) 63-68 (Table 4), the agreement between brood years was not particularly high, especially for scale rows and branchiostegal ray counts. Although meristic counts and water temper- atures during the incubation period of the eggs are usually correlated (Barlow 1961), we found that lateral series scale counts provided the only meristic character significantly (a = 0.05) cor- related with the temperature of the hatchery water during incubation. Under the extant en- vironmental conditions, incubation tempera- tures may have little effect in determining the morphological characters of our stocks. Among all possible statistically significant correlations between morphological characters and the stream characteristics in Table 5, only vertebral number and estuary length, and verte- bral number and spawning distance, had corre- lation coefficients >r = 0.50 (Table 6). All other correlations each accounted for <25% of the variation observed. Possibly we overlooked some important environmental gradients, or possibly the selective forces occur during peri- odic environmental extremes or pulses that were not accounted for in our environmental data. Each of the counts significantly correlated with at least two of the characters of the stream sys- tems, suggesting that, if these characters are the 109 FISHERY BULLETIN: VOL. 80, NO. 1 Table 2.— Means, standard errors (in parentheses), and ranges for morphological characters of the 1977 brood year hatchery samples of juvenile coho salmon and the hatchery water incubation temperatures for the first month of incubation. Sample sizes were 15. The data are listed in north to south order of the sampling location. Stock and (In parentheses) Incubation water tempera- tures (°C) Scales in lateral series Scales above lateral line Anal rays Gill rakers Branchi- ostegals Vertebrae Bonneville Hatchery (5.4) (Columbia River) 133 33 (.61) 129-138 2673 (.27) 25-29 1393 (.12) 13-15 22 53 (29) 21-25 27 00 (-32) 25-29 65 80 ( 15) 65-67 Big Creek Hatchery (7.2) 133 60 (46) 130-136 27.20 (.33) 26-30 13.53 (.13) 13-14 23 13 (.22) 22-25 25 60 (24) 23-27 6607 (.21) 65-67 Cowlitz Hatchery stock (7 2) (Big Creek Hatchery) 132 20 (.40) 129-135 26 60 (.41) 25-30 13.60 ( 13) 13-14 21.80 (.22) 20-23 2600 (.34) 24-28 65 67 (.16) 65-67 North Nehalem Hatchery (7.7) 130.93 (.42) 128-134 27.73 (25) 26-29 13.73 (.15) 13-15 23.07 (.27) 21-24 26.13 (.29) 24-28 6527 (.18) 64-66 Trask River Hatchery (9 9) 130.33 (.42) 128-133 25.53 (.32) 23-27 13.67 (.13) 13-14 2273 (.23) 21-24 25 60 (.32) 24-28 65.40 (24) 63-66 Salmon River Hatchery (7 8) 130.53 (.59) 127-135 26 80 (28) 25-29 1367 (.16) 13-14 2240 (.16) 22-24 2627 (.25) 25-29 65 53 (.19) 64-66 Fall Creek Hatchery (7.4) (Alsea River) 131.53 (-68) 127-136 2620 (.33) 24-28 1380 (.17) 13-15 22 53 (.27) 21-24 2607 (.23) 25-28 66.13 (19) 65-67 Umpqua Hatchery stock (8.6) (Smith River) Cole Rivers Hatchery 129 07 (.37) 126-131 26 40 (-32) 24-29 1340 (.13) 13-14 21.47 (.17) 21-23 2587 (.24) 24-28 65 40 (.13) 65-66 Cole Rivers Hatchery (8 6) (Rogue River) 130 33 (.66) 125-134 26 20 (.33) 24-28 13 80 ( 17) 13-15 22.20 (.30) 20-24 26 20 (.24) 25-28 65 20 (26) 64-67 result of selection, several interacting selective forces were involved. Life History Characters Earlier peak spawning times (Table 7) were strongly associated with the northern stream systems and with stream systems having large estuaries (Table 6). However, the correlation of peak spawning time with size of estuary may be biased by the large number of samples from Columbia River hatcheries: spawning times of stocks from the Columbia River are earlier than those of coastal stocks, and the Columbia River has a large estuary. Selection for earlier spawn- ing times through hatchery practices may be the cause for the differences in spawning times be- tween hatchery and wild stocks in the North Nehalem, Trask, and Alsea Rivers. Selection for earlier spawning times has been observed in a steelhead trout hatchery program (Millenbach 1973). At hatcheries using wild stocks as sources for eggs and sperm, peak spawning times were similar to those of naturally spawning fish in the respective stream system. The proportion of females (Table 7) ap- peared to be higher in the southern stream sys- tems, suggesting that jacks were more common there. The effective sex ratio, including jacks, at the time of spawning should be close to 1:1 (Fisher 1930). If only 3-yr-old males and females are counted, the proportion of females should be X).50, the margin above 0.50 depending on how many jacks returned in the previous year. How- ever, the proportion of males was higher than that of females in stocks from the Quilcene, Quinault, Sandy, North Nehalem, Nehalem, Trask, Salmon, Alsea, Umpqua, and Rogue Rivers. Nikolskii (1969) reviewed several pos- sible causes for sex ratios departing from 1:1; however, the reason for the high proportion of males in these stocks is not known. Isozyme Gene Frequencies Transferrin gene frequencies (Figs. 2, 3), cor- related significantly with six of the stream characters (Table 6). The best model from step- wise multiple regression explained only 68% of the variation in gene frequencies. Analysis of the relationships of the "A" allele frequencies with basin area (Fig. 4) and latitude (Fig. 5) explained the variation more simply than did the stepwise regression model. These correlations showed 110 HJORTand SCHRECK: PHENOTYPIC DIFFERENCES AMONG COIK) SALMON Table 8.— Means, standard errors (in parentheses), and ranges of morphological char- acters for 1977 brood year samples of wild juvenile eoho salmon. Sample size was 12 for all stream systems except Tenmile Lakes and Coquille River (15 each). The data are listed in north to south order of the sampling locations. Stream system North Nehalem River Nehalem River Trask River Nestucca River Salmon River Siletz River Beaver Creek Alsea River Umpqua River Tenmile Lakes Coquille River Rogue River Scales in lateral series Scale rows above lateral core Anal fin rays Gill rakers Branchi- ostegals Verte- brae 13225 27.75 13 58 23 25 26 75 65 50 (.88) (33) ( 15) (37) (.25) (.26) 126-137 26-30 13-14 22-25 26-28 63-66 132.50 2667 1375 2300 27 33 6575 (.77) (35) (.13) (25) (.22) (25) 127-136 25-29 13-14 22-24 26-28 64-67 131 17 26 50 1400 2292 2658 65 83 (.47) (.31) ( 12) (31) (26) ( 11) 128-133 24-28 13-15 21-24 25-28 65-66 132 17 26-83 14.00 2292 27.25 65.58 (68) (.40) (12) (34) (.28) (.19) 128-136 25-29 13-15 21-25 26-29 65-67 131.83 26 92 1367 23 08 2667 65 00 (61) (31) (.14) (.29) (19) (.17) 128-135 24-28 13-14 22-25 26-28 64-66 130.33 27 08 1358 23 00 27.50 65.25 (61) (.29) (.15) (21) (.23) (28) 128-135 26-29 13-14 22-24 26-29 63-67 132.27 27 33 13 27 2327 27 18 65 33 (.49) (38) (.20) (36) (.30) ( 14) 130-135 25-29 12 14 21-25 26-29 65-66 131 25 27 17 1367 23 17 26 83 6525 (37) (30) (.19) (34) (.30) (-18) 129-134 26-29 12-14 21-25 26-28 64-66 131 75 2683 13 25 22 92 27 00 65.83 (.70) (40) (13) (19) (28) (.27) 128-136 25-30 13-14 22-24 26-29 65-68 131.73 26 20 13.47 22.53 2660 65.73 (.58) (28) ( 13) ( 19) (31) (.25) 128-136 25-29 13-14 21-24 25-28 64-67 131 67 26 27 13.27 2240 26.47 65.93 (.43) (-43) (18) ( 19) (.27) (21) 129-134 24-30 13-14 21-24 24-28 65-67 132.75 2658 1400 2250 2692 6542 (59) (26) ( 17) (31) (40) (15) 131-137 25-28 13-15 21-25 24-29 65-66 Table 4.— Hatchery stocks of coho salmon in which dif- ferences in morphological characters occurred between the 1976 and 1977 brood years as determined by a two-sample test. Scales Lateral above Branchi- series lateral Anal Gill ostegal Verte- Hatchery scales line rays rakers rays brae Cascade-Bonneville Cowlitz stock Big Creek Trask River Salmon River Alsea River Umpqua River *P<0.05; "P<0 01 that the stocks from large stream systems and the southernmost stream systems had high fre- quencies of the "A" allele, whereas the fre- quencies in the smaller stream systems and northern stream systems were highly variable. Combining these two relationships helps explain the pattern of transferrin gene frequencies. Fre- quencies of the "A" allele were high in stocks from large stream systems regardless of lati- tude, and in southern stocks regardless of stream size. Stocks from smaller stream systems on the northern Oregon coast and in Washington had higher frequencies of the "C" allele. The factors affecting the patterns of trans- ferrin gene frequencies in coho salmon stocks are not known. However, Utter et al. (1980) sug- gested that the frequencies may be influenced by bacteriostatic properties associated with the dif- ferent transferrin alleles. Genotypes of trans- ferrin had differential mortality when exposed to bacterial kidney disease in studies by Suzumoto et al. (1977) and Winter et al. (1980b), and to vibriosis, cold-water disease, and furun- culosis in a study by Pratschner (1978). Trans- ferrin genotype was also related both to differ- ences in juvenile growth rates and to propensity to return as jacks (Mclntyre and Johnson 1977). Ill FISHERY BULLETIN: VOL. 80, NO. 1 100L 80 UJ _l y 60 _i < 40 UJ S 20 or UJ Q. "- 66 |}62 H o-|976 — 1977 57 69 66 63 ,. 68 68 51 65 68 I I I 1 _75 63 62 ii •65 63 Figure 2. — Transferrin gene fre- quencies of hatchery coho salmon. Samples are arranged from north to south. Vertical lines represent 95% confidence intervals; numbers above the line show sample sizes. Location codes are as in Figure 1. * ~ w> WASH COLUMBIA OREGON CALIF Table 5.— Environmental data for the stream systems sampled in this study. Spawning Estuary Estuary Gradi- Runoff Basin distance Lati- area length ent in area Stream system (km) tude (ha) (km) (m/km) spring (km 2 ) Washington Quilcene River 13 47.75 '512 32 192 yes 2 179 Quinault River 92 47 33 '64 32 2.8 yes 2 1,123 Oregon Columbia River Cascade- Bonneville Hatcheries 235 4625 '37,513 2365 yes 3 51,769 Cowlitz Hatchery stock 193 4625 '37,513 1094 08 yes 2 6.418 Big Creek Hatchery 60 46.25 '37.513 43 4 142 no 2 88 Sandy River Hatchery 270 4625 '37,513 1947 7.1 yes 2 1.299 North Nehalem River 45 4625 "1,128 11.3 86 no '233 Nehalem River 195 4568 "1,128 24.1 24 no 2 2.192 Trask River 72 4552 "3,480 209 9.5 no 5 455 Nestucca River 76 45 16 "400 12.9 7.7 no 2 657 Salmon River 29 4505 "82 64 130 no 6 194 Siletz River 122 44 93 "475 37.0 34 no 2 797 Beaver Creek 21 44.52 '3 3.2 4 3 no '31 Alsea River 93 44.43 "858 193 32 no 6 1,227 Umpqua River 372 43.68 "2,285 45.0 1.9 yes 6 1 1,801 Smith River 7 122 4368 "2,285 24 1 2.5 no 2 898 Tenmile Lakes 24 43.57 '1 1.6 3.2 no 5 254 Coquille River 138 43.11 "308 660 34 no 6 2,738 Rogue River 249 42.44 "251 64 1.9 yes 6 13,199 California Klamath River 293 41.58 '200 3.2 2 2 yes 8 31.314 Trinity River 235 41 58 '200 3.2 24 yes 8 7.383 Mad River 72 40 95 '200 64 6.5 no 8 1.255 'Provided by district biologists. 2 Pacific Northwest River Basins Commission 1966. 1967. 1968, 1969, 1972 River Mile Indices Hydrol Hydraul Comm. Personal estimate of area utilized by coho in the Columbia drainage "Gaumer. T . D Demory, and L Osis 1973 Estuary resources use study Fish Comm Oreg , Div Manage Res 5 Water Resources Board of Oregon 1969 Oregon long range requirements for water Salem, Oreg . 397 p Svilsey and Ham Incorp 1974 Estuarme resources of the Oregon coast A natural resource inventory report to the Oregon Coastal Conservation and Development Commission, Portland. Oreg . 233 p 'Source of Umpqua Hatchery stock "United States Geological Survey 1977 Water resources data for California water year 1977 Water Data Rep CA 77-2. Therefore, diseases, life history characteristics, and other factors may play a role in maintaining the patterns of transferrin gene frequencies. Transferrin gene frequencies were in good agreement between the two year classes of Ore- gon coast wild stocks, despite the small size of some of the samples (Fig. 3). The heterogeneity between year classes was greater for the Oregon 112 H.IORT and SCHRKCK: PHKNOTYPIC DIFFKRKNCKS AMONC (OHO SALMON FIGURE 3.— Transferrin gene frequencies of wild coho salmon stocks for 1976 and 1977 brood years. Stocks are arranged from north to south. Bars represent 95% con- fidence intervals and the sample sizes are above the bars. Location codes are as in Figure 1. 23 54 50 28 UJ Ld < UJ c_> or UJ 0- uu- o-l976 . -1977 IP 50 T < < 80- 17 60- 40- 52 to 73 58 32 i 63 > i i >4 58 109 J64 2 | i 49 f 1 20- o i 87 T 86 I J ■*■ ii i - 1 - n- K 60 1 69 i 1 1 1 1— i i »" m s Table 6.— Statistically significant correlation coefficients between the characteristics of the coho salmon stocks and the environmental characteristics of their respective stream systems, r = 0.28 at a = 0.05 and 0.37 at a = 0.01. Characteristics Stock Environmental Correlation Scales in lateral series Spawning distance 0418 Estuary size 0.341 Estuary length 0430 Gradient -0 368 Scale rows Latitude 0360 Spring runoff 0-315 Anal rays Estuary size 0.414 Latitude 0382 Gill rakers Latitude 0346 Basin area -0353 Spring runoff -0319 Branchiostegals Latitude 0431 Spring runoff 381 Vertebrae Estuary size 0350 Basin area 0.445 Gradient -0.432 Spawning distance 549 Estuary length 0533 Proportion of females Latitude -0426 Time of peak spawning Estuary size -0613 Spring runoff -0.345 Estuary length -0391 Latitude -0702 Phosphoglucose isomerase Spring runoff -0410 Transferrin Estuary length 0326 Latitude -0381 Spawning distance 0.590 Basin area 0588 Spring runoff 0528 Gradient -0 596 coast hatchery stocks (Fig. 2). The gene fre- quencies of hatchery stocks may have been altered by earlier importing of stocks with dif- ferent gene frequencies, or by disease epizootics. If fish with certain transferrin genotypes have different resistances to diseases, and if epizootics are more severe because of the higher densities of ttOr ar>- 60-- < ;< 40- • g *H" 10 BASIN AREA (Ln SO. Ml.) Figure 4.— Transferrin gene frequencies for wild and hatchery coho salmon stocks arranged by basin area, in ln square miles. fish in hatcheries, then the transferrin gene fre- quency of a given year class could be altered without affecting the other two year classes. The phosphoglucose isomerase variant (Table 7) was present only in samples from Oregon stocks — particularly those from the northern Oregon coast. May ( 1975) reported this variant in Washington stocks. Similarity of Stocks The groups of stocks of coho salmon found to be most similar by the agglomerative cluster analysis (Fig. 6) were composed of northern 113 FISHERY BULLETIN: VOL. 80. NO. 1 o £ ■*-» ■*-> c <— at o t- L. :« V > ~ CD 13 In o c t. +j CD - - o o — o o 3 c bl c o *N jr ~ _ -4-» ~ en J=. Z^ a V <— s-. o as 01 a! N ~ m aS — 11 cd ft.c E H m en a> T5 en r 01 — jr +j ^_^ C m CJ -_ > C !- <1) 01 -4-> — c as CD V o C b! CD <1) ~ <— «4H c <— o o CJ c 35 c 1- *J (75 L. " >i ft o o a) a; c a> d) as *- P "- 1 • — 0) c a> be p as en a ° .* as aS CD 8.x ° o cd c E - ._ aS -*-> en en O CD -C as zj S "ft £ <" o 15 q C o u o en a* .2 s-s t« J= - 03 ^ C 2*1 Q. CD E S CO "> o Q. CD E H CO " cfl D O ■*» o 3) o~ in O) >. . o CD C .*- CD U- 3 CT 0) Q. CD E H CO <" *. o '- ^p J) IT) 05 >. i u CD C *- CD "- 3 CT *- Ol °^.E CU CO c £ CD > 1- Q. CO CO CO CD > 5 o a? at- ro Pro tioi fem o o 1 it CD to r C CD S .E O- CO 1- °- CO CO CO CD >- ' ^-'co o« a. r <0 Pro tioi fem E CD CO >~ CO E CO CD (/) I I I p o O O O O Tf • en lo en O O 1 r*-. ~- > > > o o o Z Z Z > O z a > u O (D Z O o Q u u o 03 - r^ 01 CD 05 ai co i- CD CD h- O) O) Ol ooooo odduuuudddud (DCVJCUQJCUaJCVCDQJQJCDCD QQQQQQQQQQQQ CDCOCDNCDCOCOCOnCO «^ a>c7iO)Looia>oiO)Tj- Oc^r* 0)0)t7)O)0)CJ)(J)0)0)tT) cocncococD^Tconromm-^ o ooooooooooo CO re 1 1 flag ■^ 3 3 en _ raO O £ O 5 O u 1 8 ? ^ « ^^ - 2 cd o ra rr CD CD > ra tr -c ^ CD CO "O N i- 1 OC > CD CD CD re CD rf lO ■CD 5 2 5rr - EH > S Se <" 5 s e £ zzhZwcomohO CT en O o B2 > I ro DC CD I CO .c ro >. CO c c (- o c ) CD i- ; e & ; ^ o O 114 H.IORT ami SCHKKCK: I'HKNOTYI'IC DIFFERENCES AMONG ('OHO SALMON WOr 80 20 43* 45* LATITUDE 'N 4T Figure 5.— Transferrin gene frequencies for wild and hatch- ery coho salmon stocks arranged by latitudes. STOCK Figure 6. — Dendrogram of the agglomerative cluster analysis for all stocks of wild and hatchery coho salmon of two brood years, 1976 and 1977. Euclidean distance was the dis- similarity measure and group average was the clustering strategy. Location codes are as in Figure 1. The other codes are as follows: H6, hatchery stock of the 1976 brood year; H7, hatchery stock of the 1977 brood year; and W7, wild stock from the 1977 brood year. Oregon coast wild stocks (cluster 1), southern Oregon coast wild stocks (cluster 2), stocks from hatcheries that used wild stocks for the egg source (cluster 3), stocks from large river sys- tems (cluster 4), hatchery stocks and two wild stocks from the northern Oregon coast (cluster 5), and three individual hatchery stocks from California and Washington (clusters 6-8). Canonical variate analysis on the five larger clusters produced three canonical variables that were significant (a = 0.05). When these three variables were plotted against each other, only clusters 1 and 5 (consisting of wild stocks and hatchery stocks, both from the northern Oregon coast) were not completely separate in three- dimensional space. The other three clusters were discrete, suggesting that intercluster differ- ences were stronger than between clusters 1 and 5. Statistical testing for differences between the clusters would not be valid because the necessary assumption of randomness of data is violated. The results of the canonical variate analy- sis must be interpreted with caution because the variation within each cluster was reduced by our using the averages of the morphological char- acters. This reduction of variation facilitates dis- crimination between clusters by canonical vari- ate analysis, so that quantitative comparisons of cluster discreteness cannot be made. Individual phenotypes undoubtedly overlap between stocks or between clusters; however, the multivariate analysis of variance did indicate that significant differences existed among the stocks for each of the morphological characters. We characterized the stocks by the average phenotypes in order to estimate the phenotypes typical for each stream system, and on that basis the results of the canon- ical variate analysis suggested that there were discrete differences between all clusters except 1 and 5. The results of the agglomerative and divisive cluster analyses were similar. At the 13-cluster level of the divisive analysis (Table 8), all but two clusters were identical with clusters from the agglomerative cluster analysis dendrogram. The results of these analyses should be inter- preted cautiously, because they are based on only 10 characteristics— a small number compared with the total number of genetically related characteristics possible. If other characteristics had been used, the results might have differed. Thus, we did not emphasize the exact order or the levels of dissimilarity at which any two clusters joined together; rather, we observed only general trends in the clustering patterns. Three general trends are apparent in the clustering patterns of the agglomerative cluster analysis dendrogram. First, the stocks from the larger stream systems (Columbia, Rogue, and Klamath Rivers) were more similar to each other than to the stocks from smaller streams. The only exceptions to this trend were wild stock from the Umpqua River and the Umpqua and Rogue hatchery stocks. The Umpqua wild stock was 115 FISHERY BULLETIN: VOL. 80, NO. 1 Table 8.— Coho salmon stocks at the 13 cluster level of the divisive cluster analysis. "Wild" denotes wild stocks. Cluster no Divisive cluster analysis stock Brood year 1 Cascade Hatchery 1976 Cowlitz Hatchery 1976 Sandy Hatchery 1976 2 Salmon River Hatchery 1976, 1977 Rogue River Hatchery 1977 Umpqua River Hatchery 1976, 1977 3 North Nehalem wild 1977 Nestucca River wild 1977 Salmon River wild 1977 Siletz River wild 1977 Beaver Creek wild 1977 Alsea River wild 1977 4 Quilcene Hatchery 1976 5 Nehalem River wild 1977 Trask River wild 1977 6 Mad River Hatchery 1976 7 North Nehalem Hatchery 1976, 1977 Trask Hatchery 1976 Alsea Hatchery 1976 8 Umpqua River wild 1977 Tenmile Lake wild 1977 Coquille River wild 1977 9 Trask Hatchery 1977 Alsea Hatchery 1977 10 Quinault Hatchery 1977 11 Bonneville Hatchery 1977 Cowlitz Hatchery 1977 Rogue wild stock 1977 12 Irongate Hatchery 1976 Trinity Hatchery 1976 13 Big Creek Hatchery 1976, 1977 associated with the other southern Oregon coast wild stocks, and the Rogue and Umpqua hatch- ery stocks were in the cluster with other hatch- eries that used wild stocks as egg sources. The second trend observed in the dendrogram was geographical clustering. Three stocks from Washington and California were dissimilar to the Oregon stocks, and the Oregon wild stocks clustered into two groups, northern and southern coastal stocks. The third trend in the dendrogram was for hatchery and wild stocks to cluster independent- ly. One of the clusters was composed entirely of wild stocks from the northern Oregon coast, and another included all but one of the northern Ore- gon coast hatchery stocks, in addition to two wild stocks from the northern Oregon coast. The hatchery stock excluded from this cluster (no. 5) was from the Salmon River, a stock developed from eggs of wild coho salmon; both brood years of this stock were in the cluster of hatcheries that used wild stocks as an egg source. The rest of the northern Oregon coast hatcheries used return- ing hatchery-reared adults for egg sources. The two wild stocks in this cluster were from the Trask and Nehalem Rivers. They are also simi- lar to the other wild stocks; however, because of the mechanics of the group-average clustering strategy, they both clustered first with the hatch- ery stocks. The average Euclidean distance be- tween the Nehalem wild stock and the other wild stocks was less than that between the Nehalem wild stock and the hatchery stocks of cluster 5. The close relationships of the stocks in clusters 1 and 5 were also apparent in the results of the canonical variate analysis, which showed these two clusters to be continuous. The three trends in the clustering pattern indi- cated that coho salmon stocks from similar en- vironments had similar phenotypes. These trends provide some guidance for the transfer of coho salmon stocks. Geographical clustering in- dicates that the phenotypic or perhaps genetic similarity between stocks probably decreases as the distance between stocks increases. Mc- Intyre 8 showed a strong negative correlation for the distance between stream systems and the genetic similarity of the steelhead trout stocks in those stream systems. If a similar relation be- tween phenotype and distance exists among coho salmon stocks, survival rate would be expected to vary inversely with the distance that the stock is transferred from its native stream. The crucial question from the management standpoint, assuming the relationships we found are real, is how far stocks can be transferred before de- creasing survival rate and the increasing genetic impact on the native stocks reduce the practical- ity of such transfers. Although geographical distance can be an im- portant factor in selecting a donor stock, other considerations must also be taken into account. The difference between stocks from large and small stream systems illustrates a problem in basing stock transfers primarily on geograph- ical distance. Stocks from large stream systems were more similar to stocks in distant large systems than to stocks in small stream systems that were geographically close. Other environ- mental variables may also differ, affecting the phenotypes of geographically close stocks. Characteristics such as time of peak spawning or transferrin genotype may be closely related to flow and temperature regimes or to disease organisms present in the stream systems. These characteristics and others not included in this "Mclntyre. J. D. 1976. The report of interbreedingof arti- ficially propagated and native stocks of steelhead trout. Oreg. Dep. Fish Wildl., Res. Sect, Steelhead Annu. Rep., 22 p. 116 HJORT and SCHRECK: PHKNOTYIMC DIFFERENCES AMONG COHO SALMON study all should play a role in choosing stocks for transfer to other stream systems. The third trend mentioned (that of hatchery and wild stocks diverging toward different phenotypes) presents a problem to managers who must choose the best stock for transfer to other stream systems. The separate clustering of hatchery and wild stocks suggests that hatch- ery stocks have become dissimilar to wild stocks — even those that inhabit the same drain- age. Studies with steelhead trout indicated that hatchery fish survived better in hatchery ponds, whereas wild fish had higher survival in streams (Reisenbichler and Mclntyre 1977). The dis- similarity between hatchery and wild stocks may play a role in reducing the survival of hatchery-reared coho salmon when they are re- leased into a stream system. CONCLUSIONS Individual characters of the stocks examined by us showed a variety of responses to stream characters. Time of peak spawning was strongly correlated with latitude, whereas other charac- ters were significantly correlated with several environmental gradients, suggesting that interactions determining stock phenotypes are complex. The variability of the stock character may also change along environmental gradients, as demonstrated by the transferrin genotype (Figs. 4, 5). The results of the cluster analysis indicate that stocks that are geographically close are similar, that stocks from large stream systems are simi- lar to each other, that stocks from coastal stream systems are similar to each other, and that hatch- Similarity of Stream Systems and Wild Stocks Because coho salmon appear to have similar phenotypes in similar environments, one could possibly relate phenotypes of stocks with de- scriptions of their stream basins (Tables 5, 9). However, comparisons of an agglomerative cluster analysis of wild stocks (Fig. 7) with a cluster analysis of stream characters (Fig. 8) in- dicated that they were less similar than we had anticipated — although we expected some differ- ences because the stream characters were not necessarily related to taxonomic characters used in this study. Table 9.— Fish species and myxosporidan parasite. Cerata- myxa skasta, present in the Oregon stream systems. X = present. Stream systems CO o °>.c CD CD 3 a W II u> — cr 3 5 o CO o a CO "> T3 a it* in CD m Q-ra T3 CO CD 33 on c/> CD 3 O E m oro TO X co co 2-c CO Nehalem River Trask River Nestucca River Salmon River Siletz River Beaver Creek Alsea River Smith River Umpqua River Tenmile Lake Coquille River Rogue River X X X X X X •I in 4.0 3.5 3.0. 2.5 z.o 1.5 1.0. .5 >- l- a. in HI 5.0. 4.5 4.0. 3.51 3.0 2.51 Z.O 1.5 1.0 .5 STOCK STREAM SYSTEM Figure 7.— Dendrogram of the agglomerative cluster analy- sis for wild coho salmon stocks with a Euclidean distance dissimilarity measure and group average clustering strategy. Location codes are as in Figure 1. Figure 8.— Dendrogram of the agglomerative cluster analy- sis for stream systems with a Euclidean distance measure and group average clustering strategy. Location codes are as in Figure 1. 117 FISHERY BULLETIN: VOL. 80. NO. 1 ery and wild stocks are dissimilar. In general, coho salmon stocks from similar environments appear to have similar phenotypes; however, groupings obtained from cluster analyses of coho salmon stocks and corresponding stream systems were dissimilar. This dissimilarity may be a result of our using only a small number of characters for analysis. As additional characters are considered, additional trends may become evident. The characters in this study, in concert with other characters, should be used in future evaluations of genetic similarities between stocks for an eventual characterization of stocks that will ensure effective transplantation. In addition to providing information which may be useful for selecting donor stocks for hatchery programs, the results of this study also suggest a potential weakness in hatchery supple- mentation. Selection through hatchery environ- ment and hatchery practices may be changing the overall phenotype of hatchery stocks, as well as the between-year variability of individual genotypes (as we found for transferrin). If these changes result in reduced performance of the donor stocks in other stream systems, practices designed to increase hatchery production must be weighed against the actual benefits to wild production. We believe that this study demonstrates a re- lationship between phenotypic characters and certain habitat types. The differences in pheno- type that are attributable to hatchery or wild origin, geographic proximity, and small or large stream systems may provide a first basis for judging the advisability of stock transfers. ACKNOWLEDGMENTS We express our appreciation to Carl Bond for his suggestions concerning the morphological characters, to Norbert Hartmann for his advice concerning the analysis of the data, to Fred Utter for taking the time to share his knowledge of electrophoresis, and to Al McGie for providing data on the proportion of females on spawning ground surveys. Funds for this project were pro- vided by the Oregon Department of Fish and Wildlife. LITERATURE CITED Bams, R. A. 1976. Survival and propensity for homing as affected by presence or absence of locally adapted paternal genes in two transplanted populations of pink salmon (Oncorhyn- chusgorbuscha). J. Fish. Res. Board Can. 33:2716-2725. Barlow, G. W. 1961. Causes and significance of morphological variation in fishes. Syst. Zool. 10:105-117. Blackith, R. E., and R. A. Reyment. 1971. Multivariate morphometries. Acad. Press, Lond., 412 p. Clifford, H. T., and W. Stephenson. 1975. An introduction to numerical classification. Acad. Press, N.Y.. 229 p. Feldmann, C. L. 1974. The effect of accelerated growth and early release on the timing, size and number of returns of coho salmon (Oncorhynchus kisutch). M.S. Thesis, Univ. Washing- ton, Seattle, 46 p. Fisher, R. A. 1930. The genetical theory of natural selection. Oxf. Univ. Press. (Clarendon), Lond., 272 p. Harrell, L. W., A. J. Novotny, M. H. Schiewe, and H. 0. Hodgins. 1976. Isolation and description of two vibrios pathogenic to Pacific salmon in Puget Sound, Washington. Fish. Bull, U.S. 74:447-449. MacGregor, R. B., and H. R. MacCrimmon. 1977. Evidence of genetic and environmental influences on meristic variation in the rainbow trout, Salmogaird- neri Richardson. Environ. Biol. Fishes 2:25-33. May, B. 1975. Electrophoretic variation in the genus Oncorhyn- chus: The methodology, genetic basis, and practical applications to fisheries research and management. M.S. Thesis, Univ. Washington, Seattle, 95 p. McIntire, C. D. 1973. Diatom associations in Yaquina Estuary, Oregon: a multivariate analysis. J. Phycol. 9:254-259. McIntyre, J. D., and A. K. Johnson. 1977. Relative yield of two transferrin phenotypes of coho salmon. Prog. Fish-Cult. 39:175-177. MlLLENBACH, C. 1973. Genetic selection of steelhead trout for manage- ment purposes. In M. W. Smith and W. M. Carter (edi- tors), International Atlantic salmon symposium, p. 253- 257. Int. Atl. Salmon Found., N.Y. Myers, K. W. 1979. Comparative analysis of stomach contents of cul- tured and wild juvenile salmonids in Yaquina Bay, Ore- gon. In S. J. Lipovsky and C. A. Simenstad (editors), Fish food habits and studies, Proceedings of the Second Pacific Northwest Technical Workshop, p. 155-162. Wash. Sea Grant Publ., Univ. Washington. Seattle. Neave, F. 1944. Racial characteristics and migratory habits in Salmo gairdneri. J. Fish. Res. Board Can. 6:245- 251. Nikolskii, G. V. 1969. Theory of fish population dynamics as the biologi- cal background for rational exploitation and manage- ment of fishery resources. Oliver and Boyd, Edinb., 323 p. Pratschner, G. A. 1978. The relative resistance of six transferrin pheno- types of coho salmon (Oncorhynchus kisutch) to cytopha- gosis, furunculosis, and vibriosis. M.S. Thesis, Univ. Washington, Seattle, 71 p. 118 HJORTaml SCHKKCK: I'HF.NOTVI'H' DIFFERENCES AMONO COHO SALMON Rao, C. R. 1970. Advanced statistical methods in biometric re- search. Hafner Publ. Co., Darien. Conn., 390 p. REISENBICHLER, R. R., AND J. D. MclNTYRE. 1977. Genetic differences in growth and survival of juve- nile hatchery and wild steelhead trout, Salmo gairdneri. J. Fish. Res. Board Can. 34:123-128. RlCKER, W. E. 1970. Hereditary and environmental factors affecting certain salmonid populations. In R. C. Simon and P. A. Larkin (editors), The stock concept of Pacific salmon, p. 19-160. H. R. MacMillan Lectures in Fisheries, Univ. British Columbia, Vancouver. Ridgway, G. J., S. W. Sherburne, and R. D. Lewis. 1970. Polymorphism in the esterases of Atlantic herring. Trans. Am. Fish. Soc. 99:147-151. ROLEY, D. D. 1973. An evaluation of the selective breeding of steelhead trout (Salmi) gairdneri gairdneri) and coho salmon (Oncorhynchus kisutch) at Eagle Creek National Fish Hatchery, Oregon. M.S. Thesis. Univ. Washington, Seattle. 198 p. Simon, R. C. and P. A. Larkin (editors). 1970. The stock concept of Pacific salmon. H. R. Mac- Millan Lectures in Fisheries, Univ. British Columbia. Vancouver. 231 p. Smith, S. B. 1969. Reproductive isolation in summer and winter races of steelhead trout. In T. G. Northcote (editor), Symposium on salmon and trout in streams, p. 21-38. H. R. MacMillan Lectures in Fisheries, Univ. British Columbia, Vancouver. Sneath, P. H. A., and R. R. SOKOL. 1973. Numerical taxonomy, the principles and practice of numerical classification. W. H. Freeman, San Franc, 573 p. Snedecor, G. W„ and W. G. Cochran. 1967. Statistical methods. 6th ed. Iowa State Univ. Press, Ames, 593 p. SlJZUMOTO, B. K., C. B. SCHRECK, AND J. D. MclNTYRE. 1977. Relative resistance of three transferrin genotypes of coho salmon (Oncorhynchus kisutch) and their hema- tological responses to bacterial kidney disease. J. Fish. Res. Board Can. 34:1-8. Taning, A. V. 1952. Experimental study of meristic characters in fishes. Biol. Rev. (Camb.) 27:169-193. Utter, F. M., W. E. Ames, and H. O. Hodgins. 1970. Transferrin polymorphism in coho salmon {Onco- rhynchus kisutch). J. Fish. Res. Board Can. 27:2371- 2373. Utter, F. M., D. Campton, S. Grant, G. Milner, J. Seeb, and L. Wishard. 1980. Population structures of indigenous salmonid spe- cies of the Pacific Northwest. In W.J. McNeil and D. C. Himsworth (editors), Salmonid ecosystems of the North Pacific, p. 285-304. Oreg. State Univ. Press, Corval- lis. Winter, G. W., C. B. Schreck, and J. D. McIntyre. 1980a. Meristic comparison of four stocks of steelhead trout (Salmo gairdneri). Copeia 1980:160-162. 1980b. Resistance of different stocks and transferrin genotypes of coho salmon, Oncorhynchus kisutch. and steelhead trout, Salmo gairdneri. to bacterial kidney disease and vibriosis. Fish. Bull., U. S. 77:795-802. 119 REPRODUCTIVE BIOLOGY OF WESTERN ATLANTIC BLUEFIN TUNA 1 Raymond E. Baglin, Jr. 2 ABSTRACT Ovaries of bluefin tuna, Thunnus thynnus, were collected from the Gulf of Mexico, Florida Straits, Middle Atlantic Bight of the western North Atlantic, and off the northeast coast of the United States. There was relatively little development towards maturity in age 1 through age 7 fish from the Middle Atlantic Bight as evidenced by low gonosomatic index values and histological examination of ovaries. Well-developed ovaries were present in giant bluefin tuna from the Gulf of Mexico and Florida Straits, with heaviest spawning occurring in May. For bluefin tuna measuring 205-269 cm fork length and 156-324 kg round weight, the average number of eggs measuring 0.33 mm in diameter and larger was estimated at 60.3 million, and the average number of eggs measuring 0.47 mm in diameter and larger was estimated at 34.2 million. Atlantic bluefin tuna, Thunnus thynnus, are seasonally distributed over most of the North Atlantic. They are found from Newfoundland to Brazil and from Norway to the Canary Islands (Gibbs and Collette 1967). In the western Atlantic, a sport fishery for bluefin tuna exists off the east coast of the United States from Maine through North Carolina and along the western Bahamas and the eastern coast of Canada. Also, a substantial commercial bluefin tuna fishery exists in the western Atlantic. There is purse seining along the east coast of the United States from Massachusetts to North Carolina and a handline and harpoon fishery off Massachusetts and Maine. A sub- stantial Japanese longline fishery is present off the east coast of the United States and in the Gulf of Mexico. In the eastern Atlantic, a sport fishery for bluefin tuna exists around the Canary Islands and a substantial commercial fishery occurs off Europe and North Africa. Purse seining is conducted off the Atlantic coast of Norway and Morocco, the Mediterranean coast of France, the Adriatic coast of Italy and Yugoslavia, in the Tyrrhenian Sea off Italy, and occasionally in the North Sea off Denmark. An important hook-and- line bait fishery occurs in the Bay of Biscay off France and Spain, off Morocco, the Azores, the 'Contribution Number 81-34 M, Southeast Fisheries Center Miami Laboratory, National Marine Fisheries Service, NOAA, Miami, FL 33149. 2 National Marine Fisheries Service, NOAA, c/o Alaska Department of Fish and Game, P.O. Box 686, Kodiak, AK 99615. Canary Islands, the Mediterranean coast of Spain, and occasionally off Turkey. Trap fisheries were present off southern Portugal, southern Spain, and the Straits of Gibraltar, as well as along the Mediterranean coast of Morocco, Tunisia, and Sicily. There is a significant Japanese longline bluefin tuna fishery in the Mediterranean Sea, Bay of Biscay, and off western Europe. There has been a substantial reduction in the Atlantic-wide catch of bluefin tuna from 38,500 1 in 1964 to 12,500 t in 1973 with no large reductions in effort (Miyake et al 3 .) A number of studies have been made, and are continuing, to understand the reason for this decline (Parks 1977; Shingu and Hisada 1980; Parrack 1980). Of the various aspects of the dynamics of fish populations, the measure of reproductive potential is of primary importance since it is a basic determinant of productivity. It is used to separate subpopulations, to estimate mortality, and, with ichthyoplankton data, to estimate stock size. Two major bluefin tuna spawning areas are located in the Atlantic approximately 4,000 mi apart: In the Gulf of Mexico (Richards 1976; Montolio and Juarez 1977; Rivas 1978) and the Florida Straits (Rivas 1954; Baglin 1976) during April, May, and June; and in the Mediterranean Sea during May, June, and July (Frade and Manacas 1933; Rodriguez-Roda 1964). Al- 3 Miyake, M. P., A. De Boisset, and S. Manning (compilers). 1974. Int. Comm. Conserv. Atl. Tunas, Stat. Bull. 5. Unnumbered pages. Manuscript accepted May 1981. FISHERY BULLETIN: VOL. 80, NO. 1, 1982. 121 FISHERY BULLETIN: VOL. 80, NO. 1 though these two spawning grounds have been well documented, a question remains whether bluef in tuna spawn elsewhere and at other times. Mather 4 believes that some bluefin tuna prob- ably spawn in late spring near the northern edge of the Gulf Stream off the eastern United States. Berrien et al. (1978) have reported collecting bluefin tuna larvae from this area during April and June 1966. These larvae, however, could have drifted to this area from spawning grounds farther south. In this paper, I describe bluefin tuna ovaries from the Middle Atlantic Bight (the U.S. coastal area between Cape Cod and Cape Hatteras) and examine the possibility that this may be another significant spawning area for bluefin tuna. Also, I make some comparisons of female gonadal development between the known spawning areas. My literature review on bluefin tuna shows there is a need for additional information on the bluefin tuna's reproductive potential. A wide range in fecundity estimates was found. Frade (1950) reported that an eastern Atlantic 160 kg bluefin tuna produced 18.7 million eggs. Williamson (1962) stated that the ovaries of a 272.4 kg western Atlantic bluefin tuna con- tained about 1.0 million eggs. Rodriguez-Roda (1967) estimated that off southern Spain a 54 kg fish could produce 5.5 million eggs and a 235 kg bluefin tuna could produce over 30.0 million eggs. Baglin (1976) estimated that a 188.4 kg western Atlantic bluefin tuna could produce 16.7 million eggs and that a 271.5 kg bluefin tuna could produce 31.4 million eggs. Baglin and Rivas (1977) indicated that a 324 kg western Atlantic tuna could produce 57.6 million eggs. I deter- mined the fecundity of bluefin tuna taken from the United States sport fishery in the Florida Straits and Gulf of Mexico and from the Japanese longline fishery in the Gulf of Mexico and compared my findings with previous estimates. I have also examined monthly sex ratios for western Atlantic bluefin tuna. MATERIALS AND METHODS Bluefin tuna from the Gulf of Mexico, Florida Straits, Middle Atlantic Bight, and off the northeast coast of the United States were sampled from anglers' catches. Bluefin tuna samples from purse seine catches came from the Middle Atlantic Bight and the northeast coast of the United States. Bluefin tuna were also sampled from the Japanese longline fishery in the Gulf of Mexico and from the New England handgear fishery. Throughout this paper the classification system of Rivas (1979) was used. Thus, small bluefin tuna are 50-129 cm fork length (FL) and 3-44 kg round weight, medium bluefin tuna are 130-180 cm FL and 45-130 kg round weight, and giant bluefin tuna are >180 cm FL and >130 kg round weight. Sex data were obtained from 283 small and medium bluefin tuna captured by commercial and sport fishermen off the Middle Atlantic Bight (1974-77). Also, sex data were obtained from 3,429 giant bluefin tuna captured by sport and commercial fishermen in the Gulf of Mexico and along the northeast coast of the United States from North Carolina to Maine, and from fish taken by sport fishermen in the Bahamas (1975-78). Straight fork length (cm) was measured with calipers and round weight was recorded in pounds and later converted to kilograms. In some instances where either weight or length was unknown, a functional regression (Baglin 1980) was used for estimating the missing measurement. Small and medium fish were assigned an age based on a length-weight-age conversion table presented by Coan (1976). No ages were assigned to giant fish because of the difficulty experienced in aging them accurately. Ovaries were examined from 81 small and medium bluefin tuna caught from 1974 through 1977 and from 403 giant bluefin tuna collected during 1965 through 1968 and 1974 through 1978. Ovaries were stored in 10% Formalin 5 and later blotted dry and weighed in grams. The gonosomatic index (GSI) (ovary weight as a percentage of total body weight) was used as a gross indicator of maturity. Only fork length was taken from Japanese longline samples from the Gulf of Mexico for which the GSI was calculated using an estimated body weight from the length- weight relationship of Baglin (1980). A detailed examination of ovaries from 292 "Mather. F. J., III. 1973. The bluefin tuna situation. Proc. 16th Annu. Int. Game Fish. Res. Conf., p. 93-120. 5 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 122 BAGLIN. JR.: REPRODUCTIVE BIOLOGY OF WESTERN ATLANTIC BLUEFIN TUNA fish was conducted using histological tech- niques. Ovaries were sectioned at 8 yu and stained either with haematoxylin and eosin or trichrome stain. The oocytes were grouped into stages according to the classification system of Kraft and Peters (1963) and Smith (1965). From each prepared slide, a measurement was taken of the largest egg diameter. The egg diameters were measured with an ocular micrometer at 30 X magnification, and the orientation of egg diameters was assumed to be random. Stage of maturity was thus based on the histological examinations. A test was made for heterogeneity of egg size within the ovary. Thin cross sections were taken from the anterior, middle, and posterior parts of one ovary of a mature fish and each section was subdivided into three subsamples, representing the center, midregion, and periphery of the ovary (Otsu and Uchida 1959). Egg diameters from each area were then measured and compared statistically. Fecundity, defined as the potential number of mature eggs (yolked ova) that could be spawned during one reproductive season, was estimated by using a dry weight method. This consisted of taking samples from the anterior, middle, and posterior parts of each ovary. The eggs from each of these sections and from the remainder of the ovaries were then separated from the ovarian tissue by straining them over a wire screen under running water. The egg samples from each section, in an aqueous solution, were stirred and a subsample from each was pipetted into a beaker. These eggs were stirred and approxi- mately 1-2 g wet weight were taken to be used for the fecundity estimate. The yolked eggs, which were counted and fecundity estimated, were divided into two size categories. Eggs 0.46 mm and larger that were counted were well developed and fully yolked. A second size category for which eggs were counted included smaller eggs (0.32 mm in diameter) that were not in quite as advanced stage of development, but that could possibly undergo further development and be spawned during one reproductive season. The subsample was then weighed to the nearest 0. 1 mg, and the weight of the remaining eggs was recorded in grams. Fecundity estimates, rounded to the nearest 0.1 million eggs, were calculated from the relationship C - (AD/B) + A, where A is the number of mature ova in the subsample, B is the weight of the ova in the subsample, C is the number of mature ova, and D is the weight of ova from both ovaries minus the weight of the subsample. RESULTS AND DISCUSSION Sex Composition From 1974 through 1977, sex was determined for 283 small and medium bluefin tuna from the Middle Atlantic Bight during June, July, and August (Table 1). No significant difference from an expected 1:1 sex ratio was found. Sampling for the remaining months was inadequate. From 1975 through 1978, sex was determined for 3,429 giant bluefin tuna from the Gulf of Mexico, March through June; Bahamas, April through June; and from the northeast coast of the United States, July through October (Table 1). The deviation from an expected 1 : 1 sex ratio was significant for April, May, July, and August. Females were more prevalent than males in spawning aggregations during April and May. Males were more prevalent in feeding schools during July and August. No significant dif- ference from an expected 1:1 sex ratio was found for March, June, September, and October. Sampling during the remaining months was in- adequate. These findings suggest that some giant bluefin tuna segregate into distinct areal groups according to the predominating sex and that sex ratios may change with season. Table 1.— Monthly sex ratios for small and medium (1974-77) and giant (1975-78) western Atlantic bluefin tuna. Sex ratio Size category Month Number males/females Small and medium June 204 1.02 July 35 0.84 August 44 1.10 Giant March 66 1.00 April 292 '0.75 May 356 '0.63 June 106 093 July 800 '1 46 August 1,049 '1.74 September 694 093 October 66 1.13 'Significant departure from null hypothesis at 0.05 level (chi-square). Gonosomatic Index, Gross Morphology, and Size of Ova The external appearance alone of tuna ovaries is inadequate for gross classification of maturity stage (Bunag 1956). The GSI (also called gonad index, maturity index, gonadosomic or gonadal- 123 FISHERY BULLETIN: VOL. 80, NO. 1 somatic index), along with egg diameter measurements, has been a successfully used criterion for selecting specimens for fecundity studies and for determining the spawning peri- ods for various species of fish (Vladykov 1956; Peterson 1961; Erdman, 1968; Mathur and Ramsey 1974; Baglin 1979). On the basis of the GSI and the gross morphology and size of ova from the preserved ovaries, western Atlantic female bluefin tuna may be assigned to one of the following developmental stages: I. Immature — Ovaries are thin, hollow tubes; nearly spherical, transparent oocytes range from 0.03 to 0.13 mm in diameter (these eggs were stained with acetocar- mine to facilitate measuring). These oocytes were also present during all other developmental stages, and there was no sign of previous spawning. GSI ranges from 0.1 to 0.3. II. Maturing — Ovaries are flaccid, opaque ova up to about 0.63 mm in diameter. GSI ranges from 0.4 to 1.9. III. Mature — Ovaries are firm and full of eggs, with many yellow-orange ova up to 0.85 mm in diameter. GSI ranges from 2.0 to 5.3. IV. Ripe — None of the fish studied were found to be in this developmental stage. However, a few transparent ripe eggs were found in an individual classified as Stage III. The largest of these eggs was 1.16 mm in diameter with an oil droplet 0.30 mm in diameter. This egg corresponds to Stage V of Rodriguez-Roda (1967). A fish would be classified as being ripe only when a substantial number of eggs of this size are present. V. Spent— Ovaries are flaccid; completely spent fish had a few degenerating eggs up to 0.63 mm in diameter. Fish in this stage taken in the summer and early fall months had ovaries with a large amount of fatty tissue. GSI >0.2 but <2.0. Size and reproductive data by estimated age and month are presented in Table 2 for 81 small and medium female bluefin tuna from the Middle Atlantic Bight of the western Atlantic and for 15 small and medium eastern Atlantic female bluefin tuna collected by Rodriguez- Roda (1967) and by Cort et al. (1976). Relatively Table 2.— Length, weight, and gonadal data for 81 small and medium female western Atlantic bluefin tuna collected during 1974- 77 and for 15 small and medium female eastern Atlantic bluefin collected during 1963 by Rodriguez-Roda (1967) and during 1976 by Cortet al, (1976). Month Fork length (cm) X SE Range Round weic Iht (kg) Ovary we ight (g) Gonosomatic X SE index (%) Range Age X SE Range X SE Range Number Western Atlantic 3 June 100 0.59 97-102 19.6 0.57 16.3-23.2 25 39 5-52 1 0.02 0.03-0 28 12 July 96 95-100 18.7 16.8-21 8 16 10-23 0.1 0.06-0.14 6 August 101 98-105 20.7 18.8-22.2 18 8-51 01 0.03-0.23 6 4 June 108 107-109 25.4 24.5-26.3 15 10-20 01 0.04-0 08 2 July 121 118-122 34.2 26 8-38.1 76 68-80 02 0.21-0 26 3 August 115 106-124 29.5 19 1-409 49 17-94 0.2 009-0.26 4 5 June 133 1.42 126-138 44 1 1.18 36.3-499 97 95 44-148 02 11-0 33 12 July 133 126-140 47.2 43.6-508 190 149-231 0.4 029-0 53 2 August 129 126-131 41.0 37.9-454 69 40-119 02 11-0 26 4 6 June August 150 150 1 44 142-157 59.1 58.1 2.34 45.4-75.4 204 126 18.5 102-222 04 02 0.03 17-0.51 12 1 7 June 163 090 157-169 80.2 2.53 65.4-91.7 434 97.9 175-1,605 0.5 10 17-1 75 14 July 160 158-165 60.3 53.1-68.1 142 62-211 02 0.13-0.31 3 Eastern Atlantic 4 July 111 '26.2 150 0.6 1 5 May June July August 130 125 126 134 54.0 '37.0 '37.8 48 740 500 450 460 1.4 14 1.2 1.0 1 1 1 1 6 June July August 149 148 152 '61.7 '61.0 63.5 900 860 470 1.5 1.4 07 1 3 2 7 June 171 110.0 1,920 1.9 1 8 May August 180 185 113.0 106.0 2,500 660 2.2 6 1 2 'Round weight estimated using functional regression of Baglin (1980) 124 BAGLIN, JR.: REPRODUCTIVE BIOLOGY OF WESTERN ATLANTIC BLUEFIN TUNA little development towards maturity was evident in the age 1 through age 7 western Atlantic fish with the exception of one age 7 fish collected in June with a GSI of 1.75. The eastern Atlantic small and medium fish, on the average, have a larger mean GSI than the western Atlantic small and medium fish. The largest GSI (2.2) calcu- lated from the eastern Atlantic small- and medium-sized bluefin tuna was found for a fish captured during May at an estimated age of 8 yr. Body and ovary size, by month, for 403 western and 75 eastern Atlantic female giant bluefin tuna are presented in Table 3, with the calculated monthly GSI presented in Figures 1 and 2. The average body and ovary size of the female giant fish from the western Atlantic was larger than that of the eastern Atlantic fish for each month. Well-developed ovaries were present in western Atlantic giant bluefin tuna during April and May. These fish were collected from the Gulf of Mexico longline and sport fishery and from the Florida Straits sport fishery. Ovarian development was minimal during October, as indicated by a low mean GSI, 3.6 3.2 Table 3.— Length, weight, and gonadal data for 403 female giant western Atlantic bluefin tuna collected during 1965-68, 1974-78, and for 75 female giant eastern Atlantic bluefin tuna collected during 1963 by Rodriguez -Rod a (1967). Fork len gth (cm) Round wei X SE ght (kg) Range O vary we ght (g) Num- ber Month X SE Range X SE Range Western Atlantic March 237 3.5 199-256 245 102 143-307 2.849 448 331-6.810 25 April 242 27 190-264 262 8.1 1 1 7-335 8,380 564 409-13,166 41 May 240 1.4 205-269 244 48 156-324 7,708 475 900-14.960 73 June 246 1.7 213-270 262 58 159-351 5.023 341 950-13.600 68 July 254 10 218-272 295 33 205-375 1.927 89 600-7,000 96 August 257 13 224-282 320 44 213-424 2.857 158 600-6.250 79 Sept 256 31 235-269 330 7.5 297-374 2,770 533 750-6.500 11 Oct 242 38 212-257 308 16.8 Easterr 186-370 Atlantic 1,137 71 700-1,400 10 May 211 1.8 200-225 190 4.2 155-219 2.758 318 1.540-6,820 17 June 213 26 204-230 189 5.9 167-235 3,148 403 1,780-6,280 12 July 214 2.3 189-244 169 58 130-263 1,493 117 700-3.580 30 August 207 22 197-232 149 56 125-226 1,112 85 840-1.980 16 MAR. APR. MAY JUNE JULY AUC. SEPT. OCT. N = 2S N = «l N = 73 N = 6S N = 96 N=79 N=ll N=10 3.2-| 3.0 2.8 2.6 2.4 2.2- 2.0 1.8 1.6 1.4 1.2 1.0 .8 .6 .4' .2- O MAY N = 17 1UNE N=12 JULY N=30 AUG. Nil. Figure 1. — Seasonal variation in gonosomatic index of 403 female giant western Atlantic bluefin tuna collected from 1965 through 1968 and 1974 through 1978. The number, mean (horizontal line), range (vertical line), 1 SD on each side of the mean (open box), and 2 SE on each side of the mean (shaded box) are shown. Figure 2.— Seasonal variation in gonosomatic index of 75 fe- male giant eastern Atlantic bluefin tuna collected during 1963 by Rodriguez-Roda (1967). The number, mean (horizontal line), range (vertical line), 1 SD on each side of the mean (open box), and 2 SE on each of the mean (shaded box) are shown. 125 FISHERY BULLETIN: VOL. 80, NO. 1 and reached a peak during April and May as indicated by the highest GSI. Sampling was inadequate for November through February. Well-developed ovaries were present mainly during May and June in eastern Atlantic giant bluefin tuna from the traps at Barbate (Rodriguez-Roda 1967). In May, the mean GSI for the western Atlantic was greater than that for the eastern Atlantic. No great difference was found between the mean GSI for the western and eastern Atlantic giant bluefin tuna for June, July, and August for which data were available for comparison. The plots of the GSI indicate that giant bluefin tuna spawning occurs earlier in the western Atlantic (the drop in GSI occurring in June) than in the eastern Atlantic (the drop in GSI occurring in July). Heterogeneity of Egg Diameters A significant difference in egg diameters was found for the center, midregion, and periphery of the anterior section of an ovary from a mature fish (F=6.1;df = 2, 631; P<0.005). No significant difference in egg diameters was found for the center, midregion, and periphery of the middle or posterior sections of the ovary. A significant difference in egg diameters was also found among the anterior, middle, and posterior sec- tions (F= 11.6; df = 2, 1,843; P<0.001). Because some heterogeneity occurred, estimates of fecundity were based on eggs from each section of both ovaries. Heterogeneity of egg size within an ovary has also been shown for albacore, Thunnus alalunga, (Otsu and Uchida 1959); swordfish, Xiphias gladius, (Uchiyama and Shomura 1974); and white marlin, Tetrapturus albidus, (Baglin 1979). Histology of the Ovaries Microscopic examinations were made of ovarian tissues from 119 small and medium bluefin tuna and 173 giant bluefin tuna. Diameters measured from these prepared slides were considerably smaller than those measured from whole eggs fixed in 10% Formalin (see footnote 5). The oocytes were grouped into the following stages of oogenesis using the system of Kraft and Peters (1963), Smith (1965), and Moe (1969). Stage 1 — Thin layer of cytoplasm surrounding a 126 relatively large nucleus with a single nucle- olus; oocytes are <0.03 mm in diameter. Stage 2 — Dark cytoplasm and many peripheral nucleoli are in the nucleus oocytes are 0.03 mm through 0.13 mm in diameter (resting stage). Stage 3— Yolk vesicles appear in cytoplasm; the membrane called the zona radiata, also referred to as zona pellucida (Hoar 1969) and vitelline membrane (Bodola 1966), appears at the end of this stage; oocytes are 0.17 mm through 0.30 mm in diameter (early vitello- genic stage). Stage 4 — Thick zona radiata, yolk vesicles, and yolk globules are present; oocytes are 0.33 mm through 0.63 mm in diameter (late vitellogenic stage). Stage 5— This final stage was seldom observed during histological analysis. It evidently takes place during a short period of time immediate- ly before ovulation. Eggs in this stage have a lightly staining granular yolk mass with few yolk vesicles and yolk globules, a thin zona radiata, and an irregular shape caused by sectioning. Histological examination of female bluefin ovary sections from the Middle Atlantic Bight revealed the following: Age 1 — Very little sexual differentiation was present in age 1 bluefin tuna (N= 17) collected during May, June, July, and August. Some oogonia were observed within the lamellae. Age 2 — The first appearance of oocyte develop- ment occurred in age 2 bluefin tuna collected during July (N - 4). Both stage 1 and stage 2 oocytes were present, although stage 2 oocytes were most numerous. Age 3— Many stage 2 oocytes and a few stage 1 oocytes were found in age 3 bluefin tuna collected during January and June (N - 13). Only stage 2 oocytes were found in age 3 bluefin tuna collected during July and August (N= 10). Age 4 — Stage 2 oocytes were present in all age 4 females collected during June (Af = 36) (Fig. 3). Also in 11% of these fish, some vitellogenic stage 3 oocytes undergoing absorption were present. Only stage 2 oocytes were present in age 4 bluefin tuna collected during July and August (N = 10). Age 5— Mostly stage 2 oocytes were present in age 5 fish collected during June (N = 16), BAGLIN, JR.: REPRODUCTIVE BIOLOGY OF WESTERN ATLANTIC BLUEFIN TUNA Figure 3.— Ovarian tissue from an age 4 bluefin tuna (119 cm, 33.6 kg) collected off the Middle Atlantic Bight during June 1977. Stage 2 oocytes are present, as indicated by arrow. although a few stage 1 oocytes were also observed. Also, in 44% of these fish some stage 3 oocytes were present, many undergoing absorption. One individual also had some stage 4 oocytes present. These oocytes were in the process of degeneration. Mostly stage 2 oocytes were present in age 5 fish collected during July (N= 2). Both of these fish also had some stage 3 oocytes present, which were undergoing absorption. Stage 2 oocytes were present in all fish collected during August (N = 4). Only one of these fish had stage 3 oocytes present. These stage 3 oocytes were also undergoing absorp- tion. Age 6 — The majority of oocytes observed in age 6 bluefin tuna collected during June (N = 12) were in stage 2 of development and only a few stage 1 oocytes were observed. Many of these fish (83%) had some stage 3 oocytes present, most undergoing the process of degeneration (Fig. 4). One individual had some stage 4 oocytes present, which were also degenerat- ing. Only stage 2 oocytes were found in an age 6 bluefin tuna collected during August. Age 7 — Mostly stage 2 oocytes were found in age 7 fish collected during June (N - 15). Also, some stage 1 oocytes were present in most of these fish and 47% had stage 3 oocytes present, many of which were undergoing absorption. Only stage 2 oocytes were found in an age 7 fish collected during July and another age 7 fish taken during October. Some gonadal development, therefore, occurs in these medium female bluefin tuna. However, the simulation of gonadal maturation by young fish that probably do not spawn has been re- ported for king mackerel, Scomberomorus cavalle, (Beaumariage 1973) and Atlantic sail- fish, Istiophorus platypterus, (Jolley 1977). These authors based their determination on the size of the stage 4 oocytes, on their compactness within the lamellae, and on the appearance of many de- generating oocytes. My observations of medium bluefin tuna seem to correspond with the findings of the above authors, although the most 127 FISHERY BULLETIN: VOL. 80, NO. 1 I mm _ ^*A< Figure 4.— Ovarian tissue from an age 6 bluefin tuna (153 cm, 59.5 kg) collected off the Middle Atlantic Bight during June 1976. Resting stage 2 oocytes are present (upper arrow), as are stage 3 oocytes undergoing the process of degeneration (lower arrow). developed oocytes in the majority of fish that I examined were in stage 3 of development, and the average were in stage 2 based on the mean size of the oocytes measured (Fig. 5). I believe, therefore, that the Middle Atlantic Bight is not a significant spawning area during the summer and probably not at any other time of the year, but samples are not available for other seasons. Also, recently there has been speculation by Rivas (1978) that some of these medium bluefin tuna may migrate during May and June across the Atlantic to the Mediter- ranean Sea to spawn. According to Sella (1929) eastern Atlantic bluefin tuna begin to reproduce in their third year, when they attain a weight of about 15 kg. Rodrlguez-Roda (1967) found that 50% of eastern Atlantic bluefin tuna females are mature at 97.5 cm or 3 yr of age. However, the smallest bluefin tuna that he estimated fecun- dity for was 130.5 cm and 54 kg, which corre- sponds to an age 5 fish according to Coan (1976). Frade and Manacas (1933) found very little development in age 3 females from the eastern Atlantic. Cort et al. (1976) found developing oocytes in eastern Atlantic bluefin tuna measur- ing 148 and 152 cm from the Gulf of Gascony. These fish would be age 6 according to Coan (1976). The only published record I have found of age of maturity of western Atlantic bluefin tuna is that of Westman and Neville (1942). Observing the gross morphology of ovaries, they indicated that western Atlantic bluefin tuna 5 yr of age appeared to be mature, although the gonads gave no indication of the presence of eggs. I have also examined the unpublished cruise report of the MV Delaware, June 1957. Using the conversion table of Coan (1976), all age 3 female bluefin tuna were judged immature, and most age 4 (2 out of 3) were judged immature. No description of the ovaries on an age 5 fish was given, and 67% of age 6 females (N = 6) had well-developed eggs. My analysis of western Atlantic bluefin tuna ovaries indicates that age 6 would probably be 128 BAGLIN, JR.: REPRODUCTIVE BIOLOGY OF WESTERN ATLANTIC BLUEFIN TUNA .55- .50- .45 ^ .40' z "~ .35 QE }HJ .30 U | ... Q .20 (9 (9 W .15 Z !3 io .05 8 l_ _l_ MARCH APRIL MAY JUNE JULY AUG. SEPT. OCT. N=25 N = 65 N = 27 N = 17 N=10 N = 10 N = 16 N = 3 JUNE N = 91 JULY N = 12 AUG. N-16 Figure 5.— Mean egg diameter of largest ovarian egg as determined from histological sections from monthly samples (N) of A) female giant bluef in tuna from the Gulf of Mexico and the Florida Straits (March through June 1955, 65-67. 76-78) and from off New England (July through October 1974-75, 77), and B) female bluefin tuna ages 3-7 from off the Middle Atlantic Bight (June through August 1974-77). the earliest age at which a majority of females could possibly reach maturity. However, a majority of vitellogenic oocytes in these age 6 fish were being absorbed and most likely would not have been spawned during the years when these fish were taken. As previously noted, I observed no vitellogenic oocytes in age 3 fish, but if Sella (1929) and Rodriguez-Roda (1967) are correct, eastern Atlantic bluefin tuna may reach maturity at an earlier age than their western Atlantic counterparts. Vitellogenic oocytes were not found in the two giant bluefin tuna (205 and 207 cm) taken during the March 1966 MV Delaware Cruise 66-2 (lat. 37°24'N, long. 67°32'W). Vitellogenic oocytes were found in one of three giant fish ( 190-213 cm) taken during April 1965 on the MV Delaware Cruise 65-3 (lat. 35°54'N, long. 72°51'W). Evidently this area of the western North Atlantic was not an important spawning area for bluefin tuna during March-April. Vitellogenic oocytes (stages 3 or 4) were present in all giant bluefin tuna taken from the Gulf of Mexico and Florida Straits during 1955, 1967, 1976, 1977, and 1978 during March {N = 24), April (N = 61), and May (N= 54) (Figs. 6, 7). In one of two fish taken in June 1967 and 1977, stage 3 and stage 4 oocytes were present; in the other fish all the vitellogenic oocytes had been spawned or absorbed. On the average, the larg- est oocytes in all of the fish for July (N - 10), August (N - 10), September (N= 16), and October (N = 12) from off New England during 1974, 1975, and 1977 were in stage 2 (Fig. 8). Vitellogenic oocytes were generally absent from these fish. I found degeneration and absorption of ad- vanced unovulated eggs more common as the season progressed. This agrees with the findings of Frade and Manacas (1933) for eastern Atlantic bluefin tuna. Distinctive atretic bodies, formed from the remnants of oocytes that were not shed, were present in the female giant bluefin tuna collected during March through October. Topp and Hoff (1971) also reported atretic body formation in the ovaries of a single giant female collected from the west Florida coast during May. As previously described by Smith (1965), these atretic bodies form a characteristic brownish mass, the corpus atreticum, and are made up of amorphous brownish granules, phagocytes, and clear yellow pigment globules (Fig. 9). I found that empty follicles left behind after the ripe oocytes are released degenerate rapidly. This was also observed for eastern Atlantic bluefin tuna by 129 FISHERY BULLETIN: VOL. 80, NO. 1 130 BAGLIN. JR.: REPRODUCTIVE BIOLOGY OF WESTERN ATLANTIC BLUEEIN TUNA Figure 6.— Ovarian tissue from a giant bluefin tuna (250 cm) collected off the Gulf of Mexico during March 1978. Early stage 4 oocytes (upper arrow), stage 2 oocytes (middle arrow), and stage 3 oocytes (lower arrow) are present. Figure 7.— Ovarian tissue from a giant bluefin tuna (205 cm) collected off the Gulf of Mexico during May 1978. Late stage 4 oocytes (upper arrow), stage 2 oocytes (middle arrow), and stage 3 oocytes (lower arrow) are present. Frade and Manacas (1933), who speculated that the rapid degeneration of the follicles could be caused by pressure exerted by neighboring oocytes. My findings also confirm that western Atlantic bluefin tuna released eggs intermit- tently during April, May, and June, the majority during May. Fecundity Estimates Fecundity estimates were obtained for 28 western Atlantic bluefin tuna, which were collected from the Gulf of Mexico and Florida Straits during April, May, and June of 1967, 1968, 1974, 1975, 1976, and 1978 (Table 4). The reliability of the dry gravimetric method was tested by estimating the fecundity of an individual fish by counting and weighing eggs from four subsamples. Based on these four estimates, the average number of eggs >0.46 mm in diameter was 41.6 million with a range from 40.0 to 43.0 million and a standard error of the mean (SE) of 0.76. The average number of eggs >0.32 mm in diameter was 76.0 million with a range from 72.5 to 82.5 million and SE = 2.2. I am presenting two estimates of potential fecundity, and the estimate based on eggs >0.32 mm in diameter essentially coincides with the size of 0.33 mm used by Rodriguez-Roda (1967) for eastern Atlantic bluefin tuna. This would be the potential number of eggs that could be spawned, assuming there was no degeneration or absorption of advanced unovulated eggs. My histological examinations of bluefin ovaries, Figure 8.— Ovarian tissue from a giant bluefin tuna (256 cm, 268 kg) collected off the northeast coast of the United States during August 1975. Stage 2 oocytes are present, as indicated by arrow. 131 FISHERY BULLETIN: VOL. 80. NO. 1 FIGURE 9.— Ovarian tissue from a giant bluefin tuna (264 cm, 297 kg) collected off the northeast coast of the United States during July 1975. Brown bodies are present, as indicated by arrows. however, revealed the presence of atretic bodies. It was impossible to quantify the number of eggs constituting these atretic bodies, although I assume that absorption would occur principally with the smaller vitellogenic ova. Therefore, I have also estimated the number of eggs >0.46 mm in the most advanced size mode that could potentially be spawned during one spawning season. For bluefin tuna 205-269 cm FL and 156- 324 kg round weight, the average number of eggs 0.33 mm in diameter and larger was estimated as 60.3 million (SE = 4.04) and the average number of eggs 0.47 mm in diameter and larger was estimated as 34.2 million (SE = 2.15). No apparent relationship was found for fecundity as a function of length or estimated weight for the size range of bluefin tuna I studied. MacGregor (1968) said that the relationships of length and weight to egg production are masked in many species of fish, because egg production occurs over a relatively short range in size, and because variation in number of eggs produced at each length is great. Bailey (1964) found no obvious relationship between fecundity and fish size for American smelt, Osmerus mordax. Schenck and Whiteside (1977) and Loesch and Lund (1977) found a great amount of variability in fecundity for a given fish size for the fountain darter, Etheostoma fonticola, and blueback herring, Alosa aestivalis. Since histological examinations of ovaries and estimated GSI showed that the bluefin tuna are multiple spawners, it is possible that some of the fish selected for fecundity estimates had previously shed eggs. Also the rate of absorption of vitellogenic ova could have varied for the fish selected for fecundity estimates. I found, however, that the dry weight of eggs could be used for estimating fecundity. The following relationships were found: F= 5.2895 + 0.0167 W (F? = 0.64), where F is the number of ova >0.46 mm in diameter and W is the dry weight of all eggs separated from the ovarian connective tissue. 132 BAGLIN. JR.: REPRODUCTIVE BIOLOGY OF WESTERN ATLANTIC BLUEFIN TUNA Table 4.— Length, weight, and gonadal data for 28 female western Atlantic bluef in tuna from the Gulf of Mexico and the Florida Straits collected during April, May, and June 1967, 1968, 1974, 1975, 1976, and 1978. The mean and standard error of the mean are given at the bottom of the columns. Body length (cm) Estimated body weight (kg) Dry weight of eggs (9) Gono- somatic index (%) Estimated no. of eggs >0.46 mm >0.32 mm diameter diameter (millions) (millions) 205 156 1.260 5.3 13.6 32.7 222 '188 696 2 1 167 22.7 229 217 1.202 3 1 24.2 46.4 229 217 2.177 5.0 555 960 229 217 1.481 28 339 64.4 231 224 1,330 4.4 268 41 1 236 '197 1.329 32 284 40.3 236 '189 1.404 3.2 296 44.7 238 246 1,788 4.1 390 639 238 246 1,703 4.2 40.1 64.5 238 246 1.796 3.8 34.4 61.7 241 254 1.483 3.4 29.8 62.4 241 254 2.436 48 480 849 241 '247 1,560 3.9 330 44.0 244 263 1.750 36 25.2 56.7 244 263 1.452 3.4 232 42.1 244 263 2,121 5.0 493 93.3 252 289 2.770 4 7 396 946 254 298 1,942 29 41.6 76.0 — 307 1,681 2.5 32.0 59.2 256 307 1.950 2 6 422 79.5 257 '232 1.200 29 243 338 257 309 750 1.9 162 26.2 259 316 2.750 4 2 326 769 259 316 2.500 44 488 806 261 '272 1,488 2.6 31.4 42.3 262 '324 2,593 4.5 57.6 81.6 269 '284 1.950 4.6 40.6 74.8 X243 255 1,734 3.7 342 60.3 SE 2.78 8.37 10279 0.18 2 15 4.04 'Actual weight determined F = -0.9057 + 0.0353 WO? = 0.81), where F is the number of ova >0.32 mm in diameter and W is the dry weight of all eggs separated from the ovarian connective tissue. A reduction in fecundity in older fish has been reported by Bodola (1966) for gizzard shad, Dorosoma cepedianum, and by Loesch and Lund (1977) for blueback herring. No such decline in number of eggs was found for western Atlantic bluefin tuna. My fecundity estimates for western Atlantic bluefin tuna are considerably greater than the estimate given by Williamson (1962). He, however, did not describe how he arrived at his estimate for a western Atlantic bluefin tuna. My estimates more closely agree with estimates presented by Frade (1950) and Rodriguez-Roda (1967) for eastern Atlantic bluefin tuna. Although my estimates were based generally on larger fish, it appears that western Atlantic bluefin tuna are considerably more fecund than eastern Atlantic bluefin tuna. ACKNOWLEDGMENTS I thank L. R. Rivas and W. J. Richards of the Southeast Fisheries Center Miami Laboratory, National Marine Fisheries Service, NOAA and C. L. Smith of the American Museum of Natural History for their many helpful comments on the manuscript. I also thank the two anonymous reviewers who read the manuscript and offered helpful suggestions. LITERATURE CITED Baglin, R. E., Jr. 1976. A preliminary study of the gonadal development and fecundity of the western Atlantic bluefin tuna. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 5 (SCRS-1975):279-289. 1979. Sex composition, length-weight relationship, and reproduction of the white marlin, Tetrapturus albidus, in the western North Atlantic Ocean. Fish. Bull., U.S. 76:919-926. 1980. Length-weight relationships of western Atlantic bluefin tuna, Thunnus thynnus. Fish. Bull., U.S. 77:995-1,000. Baglin, R. E., Jr., and L. R. Rivas. 1977. Population fecundity of western and eastern North Atlantic bluefin tuna, Thunnus thynnus. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 6 (SCRS- 1976):361-365. Bailey, M. M. 1964. Age, growth, maturity, and sex composition of the American smelt, Osmerus mordax (Mitchill), of Western Lake Superior. Trans. Am. Fish. Soc. 93:382-395. Beaumariage, D. S. 1973. Age, growth, and reproduction of king mackerel, Scomberomorus cavalla, in Florida. Fla. Mar. Res. Publ. 1, 45 p. Berrien, P. L., M. P. Fahay, A. W. Kendall, Jr., and W. G. Smith. 1978. Ichthyoplankton from the RV DOLPHIN survey of Continental Shelf waters between Martha's Vineyard, Massachusetts and Cape Lookout, North Carolina 1965- 66. Sandy Hook Lab. Northeast Fish. Cent., Natl. Mar. Fish. Serv. Tech. Ser. Rep. 15, 152 p. Bodola, A. 1966. Life history of the gizzard shad, Dorosoma cepedianum (Lesueur), in Western Lake Erie. U.S. Fish Wildl. Serv., Fish. Bull. 65:391-425. Bunag, D. M. 1956. Spawning habits of some Philippine tuna based on diameter measurements of the ovarian ova. Philippine J. Fish. 4:145-175. Coan, A. 1976. Length, weight, and age conversion tables for Atlantic tunas. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 5 (SCRS-1975):64-66. Cort, J. L., C. A. Fernandez Pato, and E. De Cardenas. 1976. Observations sur la maturation sexuelle du thon rouge, Thunnus thynnus (L.), du golfe de Gascogne. Cons. Int. Explor. Mer, 6 p. Erdman, D. S. 1968. Spawning cycle, sex ratio, and weights of blue 133 FISHERY BULLETIN: VOL. 80, NO. 1 marlin off Puerto Rico and the Virgin Islands. Trans. Am. Fish. Soc. 97:131-137. Frade, F. 1950. Estudos de pescarias do ultramar Portugues os atuns. Minist. Colon. Junta Invest. Colon., Lisbon, 30 P- Frade, F., and S. Manacas. 1933. Sur l'etat de maturite des gonades chez le thon rouge genetique (note Preliminaire). C. R. Assoc. Anat., 15 p. Gibbs, R. H., Jr., and B. B. Collette. 1967. Comparative anatomy and systematics of the tunas, genus Thunnus. U.S. Fish Wildl. Serv., Fish. Bull. 66:65-130. Hoar, W. S. 1969. Reproduction. In W. S. Hoar and D. J. Randall (editors), Fish physiology, Vol. Ill, p. 1-72. Acad. Press, N.Y. Jolley, J. W. , Jr. 1977. The biology and fishery of Atlantic sailfish, Istiophorus platypterus, from southeast Florida. Fla. Mar. Res. Publ. 28, 31 p. Kraft, A. N., and H. M. Peters. 1963. Vergleichende Studien uber die Oogenese in der Guttung Tilapia (Cichlidae, Teleostei). Z. Zellforsch. Mikrost. Anat. 61:434-485. Loesch, J. G., and W. A. Lund, Jr. 1977. A contribution to the life history of the blueback herring, Alosa aestivalis. Trans. Am. Fish. Soc. 106:583-589. MacGregor, J. S. 1968. Fecundity of the northern anchovy, Engraulis mordax Girard. Calif. Dep. Fish Game 54:281-288. Mathur, D., and J. S. Ramsey. 1974. Reproductive biology of the rough shiner, Notopis baileyi, in Halawakee Creek, Alabama. Trans. Am. Fish. Soc. 103:88-93. Moe, M. A., Jr. 1969. Biology of the red grouper, Epinepheius morio (Valenciennes), from the eastern Gulf of Mexico. Fla. Dep. Nat. Resour. Mar. Res. Lab., Prof. Pap. Ser. 10, 91 P- Montolio, M., and M. Juarez. 1977. El desove de Thunnus tkynnus thynnus en el Golfo de Mexico - Estimado preliminar de la magnitud de la poblacion en desove a partir de la abundancia de larvas. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 6 (SCRS-1976):337-344. Otsu, T., and R. N. Uchida. 1959. Sexual maturity and spawning of albacore in the Pacific Ocean. U.S. Fish Wildl. Serv., Fish. Bull. 59:287-305. Parks, W. W. 1977. Cohort and equilibrium yield-per-recruit analyses for the Atlantic bluefin tuna fisheries system accounting two system configurations and two natural mortality models. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 6 (SCRS-1976):385-4. Parrack, M. L. 1980. Trends on the abundance and age structure of Atlantic bluefin tuna. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 9 (SCRS-1979):563-580. Peterson, C. L. 1961. Fecundity of the anchoveta (Cetengraulis mysti- cetus) in the Gulf of Panama. Inter-Am. Trop. Tuna Comm. Bull. 6:55-68. Richards, W. J. 1976. Spawning of bluefin tuna. (Thunnus thynnus) inthe Atlantic Ocean and adjacent seas. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 5 (SCRS-1975):267- 278. Rivas, L. R. 1954. A preliminary report on the spawning of the western north Atlantic bluefin tuna (Thunnus thynnus) in the Straits of Florida. Bull. Mar. Sci. Gulf Caribb. 4:302-322. 1978. Preliminary models of annual life history cycles of the North Atlantic bluefin tuna. In G. D. Sharp and A. E. Dizon (editors), The physiological ecology of tunas, p. 369-393. Acad. Press, San Franc. 1979. Proposed terminology for size groups of the North Atlantic bluefin tuna (Thunnus thynnus). Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 8 (SCRS- 1978):441-446. RODRiGUEZ-RODA, J. 1964. Biologia del Atiin, Thunnus thynnus (L.), de la costa sudatlantica de Espana. Invest. Pesq. 25:33- 146. 1967. Fecundidad del atiin, Thunnus thynnus (L), de la costa sudatlantica de Espana. Invest. Pesq. 31:33- 52. Schenck, J. R.. and B. G. Whiteside. 1977. Reproduction, fecundity, sexual dimorphism and sex ratio of Etheostoma fonticola (Osteichthyes:Per- cidae). Am. Midi. Nat. 98:365-375. Sella, M. 1929. Migrations and habitat of the tuna (Thunnus thynnus L.), studied by the method of the hooks, with observations on growth, on the operation of the fisheries, etc. Translated by W. G. Van Campen, U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 76, 20 p. Shingu, C, and K. Hisada. 1980. Analysis on the Atlantic bluefin tuna stock. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 9 (SCRS-1979):595-600. Smith, C. L. 1965. The patterns of sexuality and the classification of serranid fishes. Am. Mus. Novit. 2207, 20 p. Topp, R. W., and F. H. Hoff. 1971. An adult bluefin tuna, Thunnus thynnus, from a Florida west coast urban waterway. Fish. Bull., U.S. 69:251-252. UCHIYAMA, J. H., AND R. S. SHOMLIRA. 1974. Maturation and fecundity of swordfish, Xiphias gladius, from Hawaiian waters. In R. S. Shomura and F. Williams (editors). Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review and contributed papers, p. 142- 148. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. Vladykov, V. D. 1956. Fecundity of wild speckled trout Salvelinus fontinalis in Quebec lakes. J. Fish. Res. Board Can. 13:799-841. Westman, J. R.. and W. C. Neville. 1942. The tuna fishery of Long Island, New York. N.Y. Board Superv., Nassau Cty., 30 p. Williamson, G. R. 1962. The bluefin tuna in Newfoundland waters. New- foundland Tourist Dev. Off., Saint John's, Newfound- land, Can., 23 p. 134 AN EVALUATION OF TECHNIQUES FOR TAGGING SMALL ODONTOCETE CETACEANS A. B. Irvine, 1 R. S. Wells, and M. D. Scott' ABSTRACT Ninety tags — various combinations of radio tags, spaghetti tags, Roto tags, freeze brands, and tags bolted to the dorsal fin — were placed on 47 Atlantic bottlenose dolphins, Tursiops truncatus, captured near Sarasota, Florida, between January 1975 and July 197(i. In 18 months of field obser- vation, 910 tagged dolphins were sighted; 781 were identifiable, and 129 were not. Twelve naturally marked dolphins were also observed. Radio tagged animals were tracked for as long as 22 days. Repeated observations of tagged animals permitted evaluation of effect on animals and relative merits of the various tags. Freeze brands were most readable from a distance(<30 m ). and most long lived (4.8 years). Other tags were too short lived (bolt tags) or too small to be identified from a dis- tance) Roto tags and spaghetti tags), and all caused tissue destruction. Radio tags caused unexpected dorsal fin damage and were frequently lost prematurely. Taken together, the results suggest that freeze brands are least harmful, and that static tags should be tested on each species to be studied prior to attachment in the field. Cetaceans are difficult to study in the field. Most individuals move almost constantly, rise to the surface only briefly to breathe, and are difficult to differentiate from conspecifics. To facilitate individual recognition, researchers have devel- oped several tagging techniques and tested them on small odontocete cetaceans. Nishiwaki et al. (1966) placed streamer tags on captive rough- toothed dolphins, Steno bredanensis, and concluded that none were effective. On the other hand Perrin et al. (1979) recovered spaghetti tags, another type of streamer, from free-ranging dolphins, Stenella spp., in the eastern tropical Pacific up to 1,478 d after attachment. Roto tags were placed on the spotted dolphin, S. attenuata, and one marked individual was repeatedly identified from a semisubmersible over a period of 3^2 y r (Norris and Pryor 1970). Evans et al. (1972) successfully used radio transmitters, large plastic "button" tags, spaghetti tags, and freeze brands on a total of five species in the Pacific Ocean and Gulf of Mexico. Leatherwood and Evans (1979) have recently reviewed devel- opments and uses of radio tags on cetaceans. Irvine and Wells (1972) reported that an 'Gainesville Field Station, Denver Wildlife Research Center, 412 NE. 16th Ave., Gainesville, FL 32601. department of Zoology, University of Florida, Gainesville. Fla.; present address: Center for Coastal Marine Studies. Uni- versity of California, Santa Cruz, Santa Cruz, CA 95064. department of Zoology, University of Florida, Gainesville, Fla.; present address: Inter-American Tropical Tuna Commis- sion, Scripps Institute of Oceanography, La Jolla, CA 92037. improved button tag was sighted 3 mo after attachment to a bottlenose dolphin, Turxiops truncatus, near Sarasota, Fla. Despite all these improvements in tagging technology, however, little information has been available about long- term effectiveness or affect on the wearers of any type of tag. The tagging program of Irvine and Wells (1972) was reinitiated in the same area in January 1975, after a 4-yr lapse. Using radio transmitters, visual tags, and natural marks we studied the movements and activities of bottle- nose dolphins. Between 29 January 1975 and 25 July 1976, 47 dolphins were captured, tagged, and released a total of 90 times. A summary of the tagging program and an evaluation of the tagging methods used are included below. Detailed analysis of the tagging program results is presented by Irvine et al. (1979, 4 1981). METHODS The study was conducted along 40 km of coast south from Tampa Bay, Fla. The study area included shallow channels and bays bounded by a chain of barrier islands (NOS Chart No. Irvine, A. B., M.D. Scott. R. S. Wells. J. H. Kaufmann, and W. E. Evans. 1979. A study of the movements and activities of the Atlantic bottlenose dolphin, Tursiops truncatus, includ- ing an evaluation of tagging techniques. Available National Technical Information Service, 5285 Port Roval Road, Spring- field, VA 22151 as PB-298042, 54 p. Manuscript accepted July 1981. FISHERY BULLETIN: VOL. 80. No. 1, 1982. 135 FISHERY BULLETIN: VOL. 80, NO. 1 11425). Dolphins were captured by encircling one to nine animals with a 455 m X 4.5 m net dropped from a fast moving boat in shallow water. An inner circle enclosure method (Asper 1975) was used to minimize escapes. The inner circle was partitioned so that individual dolphins could be isolated and entangled without collapsing the entire net on remaining animals. Tangled dolphins were removed from the net and placed for tagging in a stretcher, usually held in the water alongside a boat. All animals were sexed, measured, and photographed before tagging. Previously tagged dolphins were examined and retagged as necessary before being released. The study area was surveyed as thoroughly as possible at least biweekly in a 7.3 m Wellcraft Fisherman 5 boat equipped with a 3 m tuna tower. All dolphin sightings were recorded during 228 surveys; photographs were taken to facilitate identification of tags and distinctive dorsal fins. Radio Tags An improvement (Model PT 219) of the radio tag developed for small pelagic cetaceans by Ocean Applied Research Corporation (Martin et al. 1971) had not been tested on T. truneatus. In our first efforts, the transmitter was attached with plastic straps to a foam-lined fiber glass saddle and secured to the dorsal fin with a stainless steel bolt through the fin. Because saddles provided by the manufacturer were too small for most T. truneatus, the transmitters were attached to fiber glass saddles molded by the authors (Fig. 1 A, C). The saddles were lined with open cell foam to protect the animal from abrasion and to allow water circulation for thermoregulation. Transmitter saddles were attached using techniques developed by other investigators (see review by Leatherwood and Evans 1979). The first seven saddles were attached with single bolts through the dorsal fin. The last three saddles were attached with bolts fore and aft to provide greater stability against water flow (Fig. 1C). Spring-loaded bolts with dissolving nuts were designed to release the saddle and transmitter from the dolphin sometime after the 1-2 mo life of the lithium batteries. Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA, or the U.S. Fish and Wildlife Service. Ten radio tags (designated RT-1 through RT- 10) were attached to dolphins between 29 January 1975 and 9 June 1976. The RT-1 transmitter consisted of a single 35 cm long X 3.7 cm diameter tube with a 63 cm high spring steel antenna on the forward end. Transmissions from RT-1 gradually failed within 2 h, ap- parently due to saltwater leakage into the battery case. Cause of failure could not be confirmed because the transmitter was missing from the saddle when it was sighted 2 d after attachment. Transmitters on subsequent radio tags were attached to the saddle with bolted aluminum plates (Fig. 1A, C) instead of plastic straps. The transmitter antenna on RT-2 was ob- served to be broken off at the base 5 d after attachment. Consequently, transmitter pack- ages on RT-3 through RT-10 were modified to two tubes, 19.2 cm long X 3.8 cm diameter, connected by copper tubing at the forward end. A flexible 42.5 cm high whip antenna extended vertically from the rear of the starboard tube. The tubes, with transmitter assembly in one and batteries in the other, were bolted to either side of the saddle, and the connecting tubing was solidly encased in fiber glass (Fig. 1C). Visual Tags The button tags described by Evans et al. (1972) had proven not to be durable on T. truneatus (Irvine and Wells 1972). Therefore, we elected to try rectangular fiber glass "visual tags" (Fig. 2). These tags were 10 cm X 7.5 cm and made of 0.4 cm thick yellow laminated fiber glass with 5.1 cm high black tape numerals epoxied to the surface. Each tag was held in place by Teflon bolts with stainless steel washers and cotter pins. The bolts were placed near the anterior edge of the tag to produce a streaming effect as the dolphin moved through the water. The bolt hole was bored through the fin and cauterized with a heated rod, and sheathed with Plexiglas tubing in the same manner as for radio tags. Double bolt tags, also yellow rectangles with black numerals, were cut from 0.2 cm thick fiber glass and varied in size from 9.0 cm X 12.9 cm to 10 cm X 15 cm, depending on the size of the dorsal fin to be tagged. The bolts were located near the anterior and posterior edges of the tag. Numerals were 7.7 cm high. Because cotter pins had sheared some of the Teflon bolts on single 136 IRVINE ET AL: AN EVALUATION OF TAGGING CETACEANS FIGURE 1.— A. Single tube transmitter with spring antenna forward (on dolphin RT-2). B. Dorsal fin 8 mo after transmitter in A was attached. C. Twin tube transmitter assembly with whip antenna aft. Dissolving nuts are top center and below the forward portion of the tube. D. Dorsal fin from C 22 d after the transmitter's installation. Note discolored, apparently necrotic, area around forward hole and apparent migration path of top bolt. bolt tags, double bolt tags were attached with 0.64 cm stainless steel bolts and nuts. Freeze Brands When first captured, all dolphins were freeze branded with 5 cm high numerals on both sides of the dorsal fin and on the body below the fin (Figs. ID, 2C, D). Recaptured animals were rebranded as necessary to improve visibility of existing brands. Application times of 15 s with irons cooled in a mixture of Dry Ice and alcohol were used to brand the dolphins captured before August 1975. Thereafter, liquid nitrogen was used as the coolant. The application time remained 15 s. When possible, the skin was rubbed with an alcohol swab to lower the skin temperature by evaporative cooling prior to branding. Before April 1976, the branding irons were applied to the skin with a gentle rocking motion to assure even contact. After that time the irons were held firmly against the skin without motion, and brand visibility was greatly improved. In some cases, however, parts of the brand did not show because of uneven contact (Figs. ID, 2C, D). Roto Tags Numbered Roto tags (NASCO Inc., Ft. Atkinson, Wis., Jumbo size) were attached to the trailing edge of the dorsal fin of all dolphins handled after January 1976. Red tags were attached to females and yellow tags to males. The 137 FISHERY BULLETIN: VOL. 80. NO. 1 Figure 2.— A. Single bolt visual tag held by Teflon bolt and cotter pin. Note tag bolt scar from 1970-71 study. B. Double bolt tag, Roto tag (at top rear of fin), and spaghetti tag (lower right). C. Double bolt tag on free-swimming dolphin. Note freeze brand with incomplete left digit on body below fin. D. Algae-covered tag 2 mo after initial installation. Note indented area of skin where water flow against tag on opposite side of fin caused pressure on near side. Note also discolored tissue around forward bolt hole. TABLE 1.— Comparison of tagging techniques. No. tags installed Tag longevity' No. sight- ings/tag mean total No. identifi- cations/tag mean total % identifiable sightings- other observers No. tags of known fate Tags of known fate lost, broken, or removed Tags of known fate obscured by fouling % total no. Tags fate bee tissu % of known removed ;ause of e damage Tag no % total no. total no Visual tags 16 <5 min to 4 88 78 200 32 6 14 86 12 14 2 14 2 (single bolt) Visual tags 19 >2 mo <2 wk to 10.00 190 984 187 16 16 63 10 31 5 25 4 (double bolt) Roto tags 53 >2 mo <1 d to >5.5 mo 6.45 342 0.53 28 48 40 19 10 5 4 2 Spaghetti tags 17 <1 mo to >13 mo 2.53 43 12 50 6 25 3 Freeze brands 2 47 >4.8 yr 6.57 309 589 277 2 39 — — — — — — Natural marks 3 12 >6yr 7.25 87 7.25 87 1 12 — — — — — - 'Length of time tag was attached and identifiable. 2 Many were redone or "touched up " 'Recognizable dorsal fins. numbers on the tags were too small to be read from the observation boat, but the color codes 138 were useful for recognition of sex, and the positions of a tag often indicated identity. IRVINE ET AL: AN EVALUATION OF TAGGING CETACEANS Spaghetti Tags Spaghetti tags (Floy Tag and Manufacturing, Inc., Seattle, Wash., Model FH 69A) were tested on some dolphins captured from April through June 1976. The attachment technique was similar to that described by Evans et al. (1972), except that the tags were applied to animals in a stretcher. Natural Marks Some dolphins had disfigured or uniquely shaped dorsal fins. A photo catalog of these recognizable untagged animals was compiled as a reference for field identification. RESULTS Nine hundred ten tagged dolphins were sighted; 781 were identifiable, and 129 others were not. When field identification was uncer- tain, photographs of combinations and locations of tags or tag remnants were often examined to verify individual identities. A compilation of tagging and sighting results is presented in Table 1. Radio Tags Ten dolphins were radio tagged and tracked for up to 22 d (Table 2). Eight of these were later recaptured and examined. In five instances, the saddle was lost, apparently because the bolt ripped through the fin (for example, see Figure IB). Fin damage was apparent 3 to 6 wk after tagging by which time saddles no longer fit snugly over the leading edge of the fin. When loosened, the saddles titled backwards creating an obvious drag; this shifting caused the bolts to migrate dorsoposteriorly. When RT-8 was re- captured after wearing a transmitter for 22 d, the two bolt holes had not healed nor appeared infected. The forward bolt had migrated dor- soposteriorly about 1.5 cm (Fig. ID), and the saddle was fouled with algal growth and mono- filament line. When RT-9 was recaptured after 46 d, the saddle and rear bolt were missing, but the front bolt was still present but bent, with part of the dissolving nut attached. The partially healed wounds appeared discolored and necrotic, but showed no obvious signs of infection. Only RT-6 showed no fin damage from the radio tag, but the tag (with malfunctioning transmitter) was removed <8 h after attachment. Two dolphins, RT-9 and RT-10, developed aberrant swimming behavior after 10 and 17 d, respectively. Both animals were observed to respire without bringing the dorsal fin to the surface in a typical cetacean rolling motion, although each could move rapidly under water. Evaluation of the problem was impossible because RT-9 evaded recapture attempts during this period, and RT-10 was not sighted during capture operations. One animal, RT-7, died 17 d after tagging, apparently of causes unrelated to the radio tag. Necropsy results implicated pulmonary damage from parasitism as a cause of death. It could not be determined if the capture-tagging process contributed to the cause of death. Tissue autolysis precluded histopathological examina- tion, and no parasites were found. Table 2.— Radio tagging results. Tag Dolphin Dolphin Date Duration of Probable reason for no. Tag description sex length (cm) attached transmission cessation of transmission RT-1 Single cylinder; forward spring antenna Male 251 29 Jan. 1975 2h Water leak (?) RT-2 Single cylinder; forward spring antenna Male 210 28 Apr 1975 5d Broken antenna RT-3 Twin cylinder; aft spring antenna Male 249 15 Jun. 1975 20 h' Seawater switch failure (?) RT-4 Twin cylinder; aft flexible antenna Female 252 1 Aug. 1975 6d 2 Unknown RT-5 Twin cylinder; aft flexible antenna Female 257 2 Oct 1975 7d 3 Seawater switch failure (?) RT-6 Twin cylinder; aft flexible antenna Male 226 15 Dec 1975 7h Seawater switch malfunction; transmitter removed RT-7 Twin cylinder; aft flexible antenna Male 239 14 Feb 1976 17d Functioning transmitter removed from dead dolphin after 21 d RT-8 Twin cylinder; aft flexible antenna Male 221 15 Apr. 1976 22 d Functioning transmitter removed because of fin damage RT-9 Twin cylinder; aft flexible antenna Female 256 8 May 1976 lOd Unknown. Dolphin did not bring fin above the surface RT-10 Twin cylinder; aft flexible antenna Female 250 9 June 1976 17d Unknown. Dolphin did not bring fin above the surface 'inconsistent signals during the last 6 h direction finder malfunction after 6 d inconsistent signals during the last 3 d 139 FISHERY BULLETIN: VOL. 80. NO. 1 Visual Tags Sixteen single bolt tags were attached between January and December 1975. One was lost within seconds, and three others were lost within 24 h. Two tags had twisted after 2 mo, damaging the fin and requiring removal of the tag. Another tag was believed to have ripped through the fin of a third animal. Two recaptured dolphins had bolt migration scars, and the tags were lost. Of 32 single bolt tags identified in the field, only 3 were sighted more than 2 wk after tagging. From December 1975 through May 1976, 19 dolphins were tagged with double bolt tags. Tags were identified on free-ranging dolphins 187 times through July of 1976, and one tag was sighted 2 mo after attachment. Broken tags were observed eight times, and nine sightings were unidentified due to algae and barnacle fouling (Fig. 2D). Several tags were observed to have only the upper anterior edges broken, implying that breakage was from physical contact. During recaptures, four intact tags were re- moved because barnacles on the inner surface of the tag caused skin abrasions. Six broken tags were removed. Bolt migration was not as common as with single bolt tags, probably because of the stability offered by the rear bolt. Although none of the bolt wounds appeared fully healed, none appeared infected when the animals were recaptured and examined. Visual tags were often discernible up to 200 m away. The numerals were rarely readable at distances >50 m, but even broken tags, tag bolts, and tag scars were useful for identification of some dolphins. Freeze Brands Freeze brands were recognizable on marked animals at distances of <30 m, although photographic analysis was often necessary to confirm identification. Some brands were difficult to identify because they were incom- plete or because of the relatively poor color con- trast of the brand against the skin (Figs. ID, 2C). One of the dolphins captured by Irvine and Wells (1972) in March 1971 and freeze branded (on both sides of the dorsal fin) was captured again in December 1975. The animal had a readable freeze brand on only one side of the fin. On another dolphin branded in the same manner in March 1971 and additionally recognizable because of a deformed lower jaw, the brand was readable in May 1971 (Evans et al. 1972), butthe brand was no longer visible upon recapture in February 1976. Roto Tags From February 1976 through July 1976, 53 Roto tags were placed on 38 dolphins, including 3 animals released with 2 tags. Roto tags were known to be lost from 17 animals and were replaced on 10 of them. A healed indented notch on the trailing edge of the fin was the only evidence of tag loss. Two Roto tags were replaced due to barnacle fouling on the inner surfaces. Brown algae and/or barnacles obscured the tag numbers on most recaptured dolphins, but the tags were still readable on close examination. Roto tag color could be observed from up to 70 m in calm seas. When examined photographi- cally, position of the tag on the fin or placement in relation to other tags or marks helped verify identity. No dolphins were identified exclusively with Roto tags. Spaghetti Tags Seventeen spaghetti tags were attached to 13 dolphins, including 4 dolphins initially released with two tags. None of the animals reacted noticeably to the attachment process. Six tags were missing from four animals re- captured 10 wk after tagging. Three tags were removed from three other dolphins because the entry wounds appeared to be festering. Animals that had lost their tags bore healed but discolored scars that were similar in size to the festering entry wounds described above. No scratches or other evidence that the dolphins may have attempted to remove the tags by rubbing were noted. The wounds, up to 1.9 cm in diameter, apparently were created by movement of the base of the tag streamer on the skin. One spaghetti tag was observed in May 1977, 345 d after attachment. Several orange colored spaghetti tags became faded within 4 wk, an observation not reported by other investiga- tors. Natural Marks Twelve untagged dolphins with recognizable natural marks were identified a total of 87 times. Photographs of an individual taken first in 1970- 71, then during this study in 1976, and by Wells 140 IRVINE ET AL: AN EVALUATION OF TAGGING CETACEANS et al. 6 in 1980 suggest that natural marks are relatively permanent. DISCUSSION The most obvious shortcoming of tags attached to the dorsal fin was the short longevity. Water drag, tissue rejection, and attempts by dolphins to shed tags may have contributed to tag loss and fin damage. We had hoped that tissue would grow tightly around the bolt sheaths and insulate the wound from bolt-induced tissue irritation; however, healing apparently never occurred while bolts were in place. Since tag wounds did not heal, different attachment methods or new designs are needed. Transmitter packages on two killer whales, Orcinus orca, were held for 6 mo by pins implanted diagonally to the plane of the leading edge of the fin (Erickson 7 ), and may offer an alternative method of attachment. The relatively larger fin of a killer whale (vs. a dolphin) may, however, have increased chances of success. Carbon bolts attach human prosthetic devices, 8 and are another attachment method yet to be tested on marine mammals. Radio tags have proved useful to study the ecology of small odontocetes (Evans 1971, 1974; Evans et al. 1972; Gaskin et al. 1975; Wursig 1976), but the configuration used in this study is not recommended for use on T. trwneatus. The fin damage, premature transmitter loss, and unusual swimming behavior which we ob- served, may influence study results. These factors have not been previously documented. Radio tags caused no obvious behavioral effects during captive tests on Delphinus delphis (Martin et al. 1971). In field studies, however, the radio tagged animals have been infrequently sighted and never recaptured, so possible long- term effects of the tags on the animals are unknown. 6 Wells, R. S.. M.D. Scott, A. B. Irvine, and P. T. Page. 1981. Observations during 1980 of bottlenose dolphins, Tursiops truncatus, marked during 1970-1976, on the west coast of Florida. Report to National Marine Fisheries Service, Con- tract No. N A80-GA-A-195, 21 p. Available Center for Coastal Marine Studies. University of California, Santa Cruz, CA 95064. 7 Erickson, A. W. 1977. Population studies of killer whales {Orcinus orca) in the Pacific Northwest: a radio-marking and tracking study of killer whales. Available National Techni- cal Information Service, 5285 Port Royal Road, Springfield, VA 22151 as PB-285615, 34 p. "Anonymous. 1977. The application of high purity carbon technology for Rehabilitation Engineering Center at Rancho Los Amigos Hospital. John F. Kennedy Space Center (NASA) Report SED-77-100, 146 p. Kennedy Space Center. Cape Kennedy, FL 32899. Freeze branding proved the most durable marking method. The variability of marks on the animals captured 5 yr after branding indicates that tissue response to the branding process is inconsistent. Freeze brands have remained readable after several years in captivity, but optimal coolants, application times, and pres- sures have yet to be determined (Cornell et al. 9 ). Our resighting, after almost 5 yr, is the longest yet reported. Twenty-one of 26 of the dolphins originally tagged in this study were observed during 1980 and had freeze brands that were either completely readable in photographs or were legible enough to confirm identifications indicated by other characteristics (Wells et al. footnote 6). Maximum longevity of freeze brands is still unknown, however. The comparatively high incidence of spaghetti tag loss reported here is noteworthy because this tagging method has been previously used with no reports of rejection or abscess (Sergeant and Brodie 1969; Evans et al. 1972; Perrin et al. 1979). Recent tests on captive dolphins have shown, however, that tag loss may be related to tissue rejection, attachment impact, or to the angle of dart entry (Jennings 10 ). Recognition of natural marks provided useful supplementary information in our study, and has been used to study bottlenose dolphins in Texas (Gruber 1981; Shane and Schmidly 11 ) and Argentina (Wursig and Wursig 1977). Close approach to the animals is usually required for field recognition, however, and we felt that photoidentification was necessary to verify most of our sightings. This tagging study has demonstrated that repeated sightings of tagged dolphins are possible and can provide substantial amounts of information about the behavioral ecology of small cetaceans (Wells et al. 1980; Irvine et al. 1981). Selection of the tags to be used should, however, involve consideration of tagging and resighting effort, tag visibility and durability, and potential harm to the tagged animal. Visual Cornell, L. H., E. D. Asper. K. Osborn, and M. J. White. 1979. Investigations on cryogenic marking procedures for marine mammals. Available National Technical Informa- tion Service, 5285 Port Royal Road. Springfield, VA 22151 as PB-291570, 24 p. 10 J. G. Jennings, Fishery Biologist, Southwest Fisheries Center, National Marine Fisheries Service. NOAA, P.O. Box 271, La Jolla, CA 92038, pers. commun. October 1978. "Shane, S. H., and D. J. Schmidly. 1978. Population biology of Atlantic bottlenose dolphin. Tursiops truncatus, in Aransas Pass, Texas. Available National Technical Informa- tion Service, 5285 Port Royal Road, Springfield, VA 22151 as PB-283393, 130 p. 141 FISHERY BULLETIN: VOL. 80. NO. 1 tags are most detectable, but are not durable and may damage the dorsal fin tissues. Freeze brands are durable, but not highly visible. Roto tags are of limited use for field identification except in unusual close range situations (e.g., Norris and Pryor 1970), although a combination color and location of the tag can identify an individual. For free-ranging dolphins, spaghetti tags are the only current tagging option, but identification of these tags usually requires collection of the animal. If animals are to be captured initially, combinations of tag types and use of natural marks can provide effective field identification. Although radio tagging and tag or mark identifications are valuable tools for ecological studies of cetaceans, more development and testing of tags and attachment techniques are needed. Investigators should realize that tagging methods which are successful on one species may not work well on another species. Prior to field studies, tags should be tested on the species to be studied. We also recommend intensive follow-up sighting surveys to maxi- mize data return and to determine the effect of tags and marks on free-ranging animals. ACKNOWLEDGMENTS This project was supported by Marine Mammal commission Contract MM4AC004 to J. H. Kaufmann, W. E. Evans, D. K. Caldwell, and A. B. Irvine; and Contract MM5AC0018 to Kaufmann, Irvine, and Evans. We are indebted to Clyde Jones and Howard Campbell of the Denver Wildlife Research Center and to Robert Hofman of the Marine Mammal Commission for support and encouragement. We also thank Mike Bogan, Larry Hobbs, Steve Leatherwood, and Galen Rathbun for their constructive comments on versions of the manuscript. Field work and data analysis were greatly assisted by volunteers from New College (University of South Florida) and the University of Florida, to whom we are indebted. We thank G. Marlow and M. Haslette and staff from the St. Petersburg Aquatarium for dolphin collections through October 1975, and Snake Eubanks and Joe Mora for their fine work thereafter. We also thank John Morrill (New College, Environmental Studies Program) for providing office space, Mary Moore and Carol Blanton for furnishing dock space, Fred Worl for supplying us with liquid nitrogen, and especially Fran and Jack Wells for providing floor space and much patience to the dolphin trackers, who regularly invaded their home. Estella Duell and Joan Randell typed the manuscript. LITERATURE CITED Asper, E. D. 1975. Techniques of live capture of smaller cetacea. J. Fish. Res. Board Can. 32:1191-1196. Evans, W. E. 1971. Orientation behavior of delphinids: radio tele- metric studies. Ann. N.Y. Acad. Sci. 188:142-160. 1974. Radio- telemetric studies of two species of small odontocete cetaceans. In W. E. Schevill (editor). The whale problem: A status report, p. 385-394. Harv. Univ. Press, Camb., Mass. Evans, W. E., J. D. Hall, A. B. Irvine, and J. S. Leatherwood. 1972. Methods for tagging small cetaceans. Fish. Bull., U.S. 70:61-65. Gaskin, D. E., G. J. D. Smith, and A. P. Watson. 1975. Preliminary study of movements of harbor por- poises (Phocoena phocoena) in the Bay of Fundy using radiotelemetry. Can. J. Zool. 53:1466-1471. Gruber, J. A. 1981. Ecology of the Atlantic bottlenose dolphin (Tursiops truncal us) in the Pass Cavallo area of Matagorda Bay, Texas. M.S. Thesis, Texas A&M Univ., College Station, 182 p. Irvine, A. B., M. D. Scott, R. S. Wells, and J. H. Kaufmann. 1981. Movements and activities of the Atlantic bottle- nose dolphin. Tursiops truncatus, near Sarasota, Florida. Fish. Bull. U.S. 79(4):67 1-688. Irvine, A. B., and R. S. Wells. 1972. Results of attempts to tag Atlantic bottlenosed dolphins (Tursiops truncatus). Cetology 13:1-5. Leatherwood, S., and W. E. Evans. 1979. Some recent uses and potentials of radio-telemetry in field studies of cetaceans. In H. E. Winn and B. L Olla (editors), Behavior of marine animals, Vol. 3, p. 1-31. Plenum Press, N.Y. Martin, H., W. E. Evans, and C. A. Bowers. 1971. Methods for radio tracking marine mammals in the open sea. IEEE 1971 Conf. Eng. Ocean Environ., p. 44-49. Nishiwaki, M., M. Nakajima, and T. Tobayama. 1966. Preliminary experiments for dolphin marking. Sci. Res. Whales Inst. Tokyo 20:101-107. Norris, K. S., and K. W. Pryor. 1970. A tagging method for small cetaceans. J. Mammal. 51:609-610. Perrin, W. F., W. E. Evans, and D. B. Holts. 1979. Movements of pelagic dolphins (Stenella spp.) in the eastern tropical Pacific as indicated by results of tagging, with summary of tagging operations, 1969-76. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF- 737, 14 p. Sergeant, D. E., and P. F. Brodie. 1969. Tagging white whales in the Canadian arctic. J. Fish. Res. Board Can. 26:2201-2205. Wells, R. S., A. B. Irvine, and M. D. Scott. 1980. The social ecology of inshore odontocetes. In L. M. Herman (editor), Cetacean behavior, p. 263-317. Wiley Interscience, N.Y. 142 IRVINE ET AL: AN EVALUATION OF TAGGING CETACEANS WURSIG, B. WiJRSIG, B.. AND M. Wi'lRSK;. 197fi. Radio tracking of dusky porpoises (Lagen- 1977. The photographic determination of group size, orhynchus obscurus) in the South Atlantic: preliminary composition, and stability of coastal porpoises ( Tu /slops analysis. ACMRR Scientific Consultation on Marine truncatus). Science (Wash., D.C.) 198:755-756. Mammals. Bergen, Norway, 21 p. 143 NOTES OFFSHORE WINTER MIGRATION OF THE ATLANTIC SILVERSIDE, MENIDIA MENIDIA ' The Atlantic silverside, Menidia menidia, is an abundant fish in coastal waters of the western Atlantic ranging from Florida to Nova Scotia. During spring, summer, and fall, the habitat of M. menidia includes intertidal creeks, marshes, and the shore zone of estuaries and embayments (Hildebrand and Schroeder 1928; Bigelow and Schroeder 1953). In such areas, ichthyofaunal surveys often cite M. menidia as the most numer- ous species encountered (Mulkana 1966; Rich- ards and Castagna 1970; Chestmore et al. 1973; Briggs 1975; Anderson etal. 1977; Hillmanetal. 1977). The entire life cycle of M. menidia is com- pleted in 1 yr. Reproduction occurs in the spring, juveniles grow rapidly during the summer and reach full adult size by late fall. However, con- siderable uncertainty exists concerning winter ecology and habitat. In populations from Chesa- peake Bay northward, Atlantic silversides are rare or absent from the shallow waters of the shore zone in midwinter ( Warfel and Merriman 1944; Bayliff 1950; Hoff and Ibara 1977; Conover and Ross in press). Hildebrand and Schroeder (1928) and Richards and Castagna (1970) re- ported that M. menidia were captured in mid- winter with bottom trawls in deepwater areas of Chesapeake Bay and deep estuarine channels in eastern Virginia. Catches of M. menidia have also been occasionally reported up to 15 km off- shore (Clark et al. 1969; Fahay 1975). However, Needier (1940) noted that Atlantic silversides could be taken through the ice in Malpeque Bay, P.E.I, (although he presented no data concerning relative seasonal abundance), and investigations in South Carolina found an abundance of M. menidia in intertidal marsh creeks during win- ter (Cain and Dean 1976; Shenker and Dean 1979). Because the Atlantic silverside is an important forage fish (Merriman 1941; Bayliff 1950; Bige- low and Schroeder 1953) and reaches a high level of biomass in the shore zone of marshes and estu- aries (7.8 g/m 2 wet weight) (Conover and Ross in 'Contribution No. 73 of the Massachusetts Cooperative Fish- ery Research Unit. press), the winter movement patterns of this an- nual species could represent a significant path- way of energy flow from and/or within estuarine systems. This paper demonstrates that Atlantic silversides migrate offshore in winter, and we discuss aspects of their winter ecology and distri- bution by examining catch records of the bottom trawl survey program of the Northeast Fisheries Center (NEFC) of the National Marine Fisheries Service (NMFS). Methods A modern series of standardized bottom trawl surveys was begun in 1963 by the Bureau of Com- mercial Fisheries (BCF) Woods Hole Laboratory (Grosslein 1969). Initially, fall surveys encom- passed the general range of offshore groundfish stocks of primary interest (i.e., gadoids) and thus was confined to the area between Hudson Can- yon and Nova Scotia and depths from 27 to 366 m. Later, as the goals and emphasis of the survey program expanded to include a wider variety of species, both fall and spring surveys were con- ducted and the sampling area was extended southward to Cape Hatteras (1967). The offshore survey region was stratified into geographic zones based on depth contours and area (Gross- lein 1969). A stratified random sampling design was employed to locate trawl stations within depth strata and the number of stations was allo- cated in proportion to stratum area. A standard No. 36 Yankee bottom trawl with a 1.25 cm stretched mesh cod end liner was towed at each station for 30 min at an average of 3.5 kn; how- ever, spring offshore surveys since 1973 have used the larger No. 41 Yankee trawl. Stations were sampled continuously 24 h/d during cruises. Synoptic bottom trawl surveys in the near- shore environment were begun in 1972 by the NMFS Sandy Hook Laboratory. Early surveys in the inshore region (defined as depth strata of 5-27 m) assessed the technical and geographic feasibility of using offshore sampling gear in waters as shallow as 5 m. Since autumn 1972, inshore surveys have been conducted each fall and spring with summer cruises added in 1977 and a winter cruise in 1978. Of 18 inshore cruises through 1978, most (17) included the region from FISHERY BULLETIN: VOL. 80. NO. 1. 1982 145 Cape Cod to Cape Hatteras, 4 included the Gulf of Maine, and 7 included the region from Cape Hat- teras to Cape Fear. During 1972-75, all inshore surveys used either a % modified Yankee trawl or the No. 36 Yankee trawl. Since 1976, a No. 41 Yankee trawl has been used. Towing procedures were the same as described for offshore surveys. The seasonal and geographic variation in the ex- tent of inshore surveys reflects their evolution as a monitoring tool. Capture data employed in this study included date, location, time, depth, surface and bottom temperatures, and number collected. Catch loca- tions from all surveys were plotted to the nearest 10' of latitude and longitude on depth contour maps by season. Surface and bottom tempera- tures and depth frequencies were plotted for each occasion that M. menidia were captured. Results Standard bottom trawl tows at 2,057 stations from inshore surveys collected 979 M. menidia at 107 sites (5.2% occurrence), while offshore tows at 10,209 stations captured 464 M. menidia at 72 sites (0.7% occurrence). Because sampling effort by season was not uniform with respect to in- shore and offshore surveys or geographic zones, analysis of catch per effort data (catch fre- quency) was compiled by month for inshore and offshore surveys in three geographic regions (i.e., Gulf of Maine-Georges Bank, Cape Cod- Cape Hatteras, Cape Hatteras-Cape Fear). In the inshore surveys, effort was primarily concen- trated in the Cape Cod-Cape Hatteras region, where the percent frequency of capture of M. menidia was negligible in summer, increased in November (4.9%), peaked in January (34.3%), and declined through the spring (Table 1). Num- ber of stations sampled in the inshore surveys of the Gulf of Maine-Georges Bank and Cape Hat- teras-Cape Fear regions was inadequate for monthly or regional comparisons. In offshore surveys, the monthly pattern of occurrence of M. menidia was similar to that of inshore surveys; catch frequency was zero in summer and autumn, peaked in January (3.8%) in the Gulf of Maine-Georges Bank and in February (11.2%) in the Cape Cod-Cape Hatteras regions, and de- clined thereafter (Table 2). These data support the hypothesis of an offshore winter migration. The geographic distribution of catches by sea- son (Fig. 1) indicates that most collections are confined to a zone within roughly 50 km of the shoreline and within the 40 m depth contour. One collection occurred 170 km from the mainland. Although most catches appear to occur between Cape Cod and Cape Hatteras and especially in the New York Bight, sampling effort among in- shore surveys was much greater in this region as previously noted (Table 1). Although only four collections of M. menidia were observed south of Cape Hatteras (two off Cape Fear, S.C., and two off Cape Romain, S.C.; not appearing in Figure 1), no offshore or inshore surveys were conducted south of Cape Hatteras in winter when catches might be expected. Surface temperatures recorded at 141 of the inshore and offshore stations where Atlantic silversides were captured ranged from l°to22°C, but 86% of these were within a range of 2°-6°C (x = 4.9°C; Fig. 2A). Bottom temperatures re- corded at 135 collecting sites revealed a similar Table 1.— Percent frequency of occurrence of Men idia men id- ia at stations sampled in the inshore survey region (depth strata of 5-27 m) of the NMFS bottom trawl survey program over the continental shelf of eastern North America. Catch sta- tistics are from cruises conducted from 1972 to 1979 and are pooled by month and area of capture. The number in paren- theses is the total number of stations sampled. Gulf of M aine and Cape Cod- Cape Hatteras- Month Georges Bank Cape Hatteras Cape Fear Jan. — (0) 34.3 (70) 0.0 (2) Feb. — (0) — (0) — (0) Mar — (0) 21.4 (206) 0.0 (18) Apr 0.0 (7) 9.6 (240) 00 (25) May — (0) 0.7 (141) — (0) June — (0) 00 (33) — (0) July 0.0 (3) 0.0 (41) 00 (47) Aug 0.0 (80) 0.5 (216) 00 (31) Sept — (0) 0.0 (150) 0.0 (82) Oct. — (0) 0.2 (398) 00 (40) Nov. 0.0 (10) 4.9 (183) 9 1 (22) Dec. — (0) 0.0 (6) 33.3 (6) Table 2.— Percent frequency of occurrence of Men idia men id- ia at stations sampled in the offshore survey region (depth strata 27-366 m) of the NMFS bottom trawl survey program over the continental shelf of eastern North America. Catch sta- tistics are from cruises conducted from 1963 to 1979 and are pooled by month and area of capture. The number in paren- theses is the total number of stations sampled. Month Gulf of Maine and Georges Bank Cape Cod- Cape Hatteras Cape Hatteras- Cape Fear Jan. 38 (159) — (0) — (0) Feb. 04 (221) 11.2 (98) — (0) Mar 00 (386) 43 (925) 0.0 (18) Apr. 02 (1.270) 1.5 (518) — (0) May 0.0 (456) 0.0 (2) — (0) June — (0) — (0) — (0) July 0.0 (310) 0.0 (114) 0.0 (41) Aug. 00 (522) 0.0 (336) — (0) Sept — (0) 0.0 (344) 00 (9) Oct. 0.0 (1,265) 0.0 (1.219) — (0) Nov 0.0 (1.628) 00 (154) — (0) Dec 00 (155) 0.0 (55) — (0) 146 LONG ISLAND CHESAPEAKE BAY GEORGES BANK O 9 . FALL * WINTER • SPRING O 50 IOO Km \CAPE V1ATTERAS Figure 1. — Location of Atlantic silverside catches by season during inshore and offshore bottom trawl surveys of the National Marine Fisheries Service, Cape Hatteras to Nova Scotia, 1963-79 (fall = Sept.-Nov.; winter = Dec.-Feb.; spring = Mar.-May). Seven catch locations do not appear: Two off the northern coast of Maine, one off the outer coast of southern Nova Scotia, two off Cape Fear, S.C., and two off Cape Romain, S.C. pattern: the majority (86%) of all Atlantic silver- side collections occurred within a range of 2°-6° C (x = 5.1°C; Fig. 2B). These data indicate that M. menidia occur over the continental shelf pri- marily under winter temperature conditions after fall overturn when temperatures are iso- thermal. The distribution of Atlantic silversides with respect to depth was examined by comparing catch frequency to depth of capture in 5 m inter- vals. The majority of catches occurred in waters <50 m deep (86%), and 42% of all catches were in depths of 10-20 m (Fig. 3). Maximum depth of capture was 126 m. Some aspects of the winter ecology of Atlantic silversides while at sea can be revealed by exam- ining their vertical distribution in the water column. Vertical distribution was inferred from diel variations in capture times partitioned in- to six 4-h intervals. Chi-square analysis com- paring catch frequency in each time interval to all others combined showed that catch frequen- cies during night intervals (2000-0359 h) were significantly less than expected (P<0.01; Table 3), while catch frequencies during midday inter- vals (0800-1559 h) were significantly greater than expected (P<0.01). Apparently, M. menidia occurred nearer the bottom during daylight hours and hence were more susceptible to bot- tom trawl tows conducted during the day. These observations indicate that while at sea, Atlan- tic silversides are vertical migrators like other planktivores such as Atlantic herring, Ciupea harengus, (Blaxter 1975) and American shad, Alosa sapidissima, (Neves and Depres 1979). 147 o c 40 r 30 20 10 < - -- 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22 SURFACE TEMPERATURE °C ° 40 r = 30 a. < 20 >- U 5 lO o .. I-... 10 20 30 40 50 60 70 80 90 100 120 DEPTH ( m ) 140 Figure 3.— Water depths at which Atlantic silversides were captured during inshore and offshore bottom trawl surveys of the National Marine Fisheries Service over the eastern North American continental shelf, 1963-79. >■ U o 50 r 40 30 20 10 B _u_u_ 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22 BOTTOM TEMPERATURE °C Figure 2.— Water temperatures at stations where Atlantic silversides were captured during inshore and offshore bottom trawl surveys conducted by the National Marine Fisheries Service during 1963-79 over the eastern North American con- tinental shelf. A. Surface temperatures (n = 141). B. Bottom temperatures (w = 135). Discussion The results of this study demonstrate that populations of M. menidia north of Cape Hat- teras undergo an offshore winter migration from inland to inner continental shelf waters. Atlantic silverside winter habitat probably also includes Table 3.— Chi-square analysis of diel variations in catch fre- quencies of Atlantic silversides in combined inshore (5-27 m) and offshore (27-366 m) trawl surveys conducted by NMFS, 1963-79, over the continental shelf of eastern North America. Time of capture (est.) Tows Menidia capt jring (no.) Observed Expected 2 X 0000-0359 11 298 14.3"* 0400-0759 28 298 0.1 0800-1159 46 298 10.5*" 1200-1559 43 298 7.0" 1600-1959 35 298 1.1 2000-2359 16 298 7.7" Totals 179 179 " P<0 01 "*P<0 005 deep inland waters not sampled by NMFS sur- veys, as Hildebrand and Schroeder (1928) and Richards and Castagna (1970) have noted. Since the lower lethal temperature for M. menidia in short-term experiments was 1°-2°C (Hoff and Westman 1966; Conover unpubl. data), the off- shore migration may be promoted by potentially stressful or lethal low water temperatures in shallow inland waters during midwinter. Con- over and Ross (in press) and Warfel and Merri- man (1944) found that Atlantic silversides leave the New England shore zone in November as water temperatures drop to about 6°-8°C. The timing of Atlantic silverside disappearance from shallow inland waters corresponds closely with their appearance in deeper offshore waters. If the offshore migration of Atlantic silver- sides is primarily motivated by low temperature stress, than offshore movements in warmer waters, such as south of Cape Hatteras, would not be expected. Even though our data cannot 148 address this question directly, evidence from ich- thyofaunal surveys in South Carolina indicate that M. menidia abundance remains high in intertidal creeks (Cain and Dean 1976; Shenker and Dean 1979) and in the surf zone of barrier beaches (Anderson et al. 1977) throughout win- ter. The relative abundance of Atlantic silversides over the continental shelf is difficult to judge from this study, since bottom trawling is a rela- tively ineffective method for catching small pelagic fish such as M. men idia (see Conover and Ross in press). In addition, the low overall catch frequency for M. menidia reported herein is pri- marily due to the relatively small number of sta- tions sampled in midwinter when maximum catches might be expected. Neves and Depres (1979) used similar NMFS offshore survey data on a larger pelagic species, the American shad, and reported catches at 527 of the 10,435 stations sampled (5.05%). Considering the methods used, the percent occurrences of M. menidia in the in- shore and offshore surveys of the mid-Atlantic during midwinter (34 and 11%, respectively) may indicate considerable abundance. In a previous study, Conover and Ross (in press) showed that Atlantic silversides reach a high level of biomass during late fall in marsh areas and also suffer a high rate of winter mor- tality (90-99%). Their hypothesis that winter movement and mortality patterns of M. menidia represent a one-way export of biomass from the shore zone of bays, marshes, and estuaries to off- shore communities is strengthened by this study. The causes of high winter mortality experienced by Atlantic silversides at northern latitudes are unknown but conceivably could include preda- tion and perhaps physiological stress imposed by the migration itself and prolonged exposure to cold temperatures. Atlantic silversides could be an important forage fish over the inner continen- tal shelf, but it will require an analysis of the food habits of offshore fishes in midwinter to address this question. Acknowledgments The authors wish to thank the staff of the Re- source Surveys Investigation Section and other members of the staff at NMFS, Woods Hole, who have participated in the cruises, and B. E. Brown and M. R. Ross for reviewing the manuscript. The senior author also received support from the Graduate School of the University of Massa- chusetts and the Massachusetts Cooperative Fishery Research Unit, which is jointly spon- sored by the Massachusetts Division of Marine Fisheries, the Massachusetts Division of Fish and Wildlife, the University of Massachusetts, and the U.S. Fish and Wildlife Service. Literature Cited Anderson, W. D., Jr., J. K. Dias, R. K. Dias, D. M. Cupka, and N. A. Chamberlain. 1977. The macrofauna of the surf zone off Folly Beach, South Carolina. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-704, 23 p. Bayliff, W. H., Jr. 1950. The life history of the silverside Menidia menidia (Linnaeus). Chesapeake Biol. Lab. Publ. 90, 27 p. Sol- omons Island, Md. BlGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Blaxter, J. H. S. 1975. The role of light in the vertical migration of fish— a review. In G. C. Evans. R. Bainbridge, and O. Rack- ham (editors), Light as an ecological factor: II, p. 189- 210. Blackwell Sci. Publ., Oxf. Briggs, P. T. 1975. Shore-zone fishes of the vicinity of Fire Island In- let, Great South Bay, New York. N. Y. Fish Game J. 22: 1-12. Cain, R. L., and J. M. Dean. 1976. Annual occurrence, abundance and diversity of fish in a South Carolina intertidal creek. Mar. Biol. (Berl.) 36:369-379. Chesmore, A. P., D. J. Brown, and R. D. Anderson. 1973. A study of the marine resources of Essex Bay. Mass. Div. Mar. Fish. Monogr. Ser. 13, 38 p. Clark, J., W. G. Smith, A. W. Kendall, Jr., and M. P. Fahay. 1969. Studies of estuarine dependence of Atlantic coastal fishes. U.S. Bur. Sport Fish. Wildl. Tech. Pap. 28, 132 P- Conover, D. O., and M. R. Ross. In press. Patterns in seasonal abundance, growth and biomass of the Atlantic silverside, Menidia menidia, in a New England estuary. Estuaries. Fahay, M. P. 1975. An annotated list of larval and juvenile fishes cap- tured with surface-towed meter net in the South Atlan- tic Bight during four RV Dolphin cruises between May 1967 and February 1968. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-685, 39 p. Grosslein, M. D. 1969. Groundfish survey program of BCF Woods Hole. Commer. Fish. Rev. 31(8-9):22-30. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43(1), 366 p. Hillman, R. E., N. W. Davis, and J. Wennemer. 1977. Abundance, diversity, and stability in shore-zone fish communities in an area of Long Island Sound affected by the thermal discharge of a nuclear power station. Estuarine Coastal Mar. Sci. 5:355-381. 149 Hoff, J. G., AND R. M. Ibara. 1977. Factors affecting the seasonal abundance, com- position and diversity of fishes in a southeastern New England estuary. Estuarine Coastal Mar. Sci. 5:665- 678. Hoff, J. G., and J. R. Westman. 1966. The temperature tolerances of three species of ma- rine fishes. J. Mar. Res. 24:131-140. Merriman, D. 1941. Studies on the striped bass (Roccus saxatilis) of the Atlantic coast. U.S. Fish Wildl. Serv., Fish. Bull. 50 (35), 77 p. MULKANA, M. S. 1966. The growth and feeding habits of juvenile fishes in two Rhode Island estuaries. Gulf Res. Rep. 2:97-168. Needler, A. W. H. 1940. A preliminary list of the fishes of Malpeque Bay. Proc. Nova Scotia Inst. Sci. 20:33-41. Neves, R. J., and L. Depres. 1979. The oceanic migration of American shad, Alosa sapidissima, along the Atlantic coast. Fish. Bull., U.S. 77:199-212. Richards, C. E., and M. Castagna. 1970. Marine fishes of Virginia's eastern shore (inlet and marsh, seaside waters). Chesapeake Sci. 11:235-248. Shenker, J. M., and J. M. Dean. 1979. The utilization of an intertidal salt marsh creek by larval and juvenile fishes: Abundance, diversity and temporal variation. Estuaries 3:154-163. Warfel, H. E., and D. Merriman. 1944. Studies on the marine resources of southern New England. I. An analysis of the fish population of the shore zone. Bull. Bingham Oceanogr. Collect., Yale Univ. 9(2):1-91. David O. Conover Massachusetts Cooperative Fishery Research Unit Department of Forestry and Wildlife Management University of Massachusetts, Amherst, MA 01003 Present address: Marine Sciences Research Center State University of New York at Stony Brook Stony Brook, NY 11 79 J, Steven A. Murawski Northeast Fisheries Center Woods Hole Laboratory National Marine Fisheries Service, NOAA Woods Hole, MA 025b3 GROWTH DURING METAMORPHOSIS OF ENGLISH SOLE, PAROPHRYS VETULUS Among fishes, the period of transformation from the larval to adult form is marked not only by changes in morphology, behavior and in some species, habitat (Jakobczyk 1965; Sale 1969; Hoar 1976; Marliave 1977), but in growth rate as well. Ontogenetic changes in growth have not been well documented principally because a method for determining age of larvae and juve- niles has not, until recently, been available. The discovery of daily growth rings on otoliths has made possible the precise determination of age, in days, of larval and juvenile fishes (Brothers et al. 1976). Changes in growth rates during differ- ent life history stages which could be correlated with behavioral and habitat changes were ob- served in the French grunt, Haemulonflavoline- atum (Brothers and MacFarland in press). Struhsaker and Uchiyama (1976) observed an inflection point in the age-length plot of larval and juvenile nehu, Stolephorus purpureus, indi- cating a change in growth rate. This inflection point corresponded with the size when body depth began to increase in proportion to the length of the fish, but not with changes in diet or habitat that occur over the course of develop- ment. Age estimates based on counts of otolith growth increments have now allowed us to de- termine growth during metamorphosis of the pleuronectid Parophrys vetulus Girard. Methods The results of this study are based on the stan- dard length (SL) in millimeters and age in days of 127 pelagic larvae and transforming individ- uals of P. vetulus ranging 10-20 mm SL, and 106 benthic 0-age individuals from 18 to 35 mm SL. Pelagic specimens were collected off Newport, Oreg. (approximately lat. 44°37'N, long. 124°06' W), from November 1977 through June 1978 with a 70 cm bongo net with 0.505 mm Nitex 1 mesh (see Laroche et al. 1982 for sampling de- tails). Benthic P. vetulus were collected off Moolach Beach, Oreg., 10 km north of Newport, during September 1978 through September 1979 with a 1.5 m wide beam trawl (7 mm stretch mesh). The removal and mounting of saccular otoliths from larvae followed the methods outlined in Methot and Kramer (1979) except that otoliths were mounted on rectangular glass cover slips to improve the optical properties of the prepara- tion. Otolith growth increments were counted at 800 or 1250 X under bright-field illumination. A complete description of the counting technique and validation of the daily periodicity of the rings can be found in Laroche et al. (1982). ■Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 150 FISHERY BULLETIN: VOL. 80. NO. 1, 1982 Otoliths from the benthic individuals were re- moved, mounted on glass microscope slides and ground to a sagittal thin section through the nu- cleus using 600 grit carborundum paper. Incre- ment counts on these ground sections were made at 250-400 X using either bright-field or polar- ized illumination. The age of each fish was defined as the number of daily otolith growth increments plus five, the age at first increment formation for this species (Laroche et al. 1982). In order to characterize changes in body form during metamorphosis, body depth, snout to anus length, lower jaw length, and the distance of migration of the left eye, of 65 larvae and 0-age benthic specimens were measured to the nearest 0.1 mm. Results and Discussion A plot of the length-at-age of P. vetulus larvae and juveniles taken in both pelagic and benthic collections exhibits a prominent plateau between 60 and 120 d of age and between 18 and 22 mm SL (Fig. 1). This plateau shows that there is a period of reduced growth in body length when these fish are undergoing metamorphosis. Plots of body depth and snout to anus length versus standard length both have a well-defined inflec- tion point between 18 and 22 mm SL (Figs. 2, 3). Other morphometric measurement plots (lower jaw length and distance of migration of the left eye) (not shown) also contain an inflection between 18 and 22 mm SL, but less clearly. Changes in body morphology during the growth plateau are illustrated by examining a develop- mental series just prior to metamorphosis (Fig. 4A) and comparing individuals of similar sizes within the plateau (Fig. 4B). Changes in body depth are most pronounced, but eye migration and changes in head morphology are also evi- dent. The definition of two distinct growth stanzas separated by a plateau conforms to several of the criteria outlined by Ricker (1979). However, as he points out, the timing of the inflection points on the size-at-age plot depends on whether length or weight is measured. We have not measured weight in this study. Due to the shape of the length-at-age plot we might expect the length-weight relationship for this species to be complex in form over the inter- val considered here. The age of 18-20 mm SL P. vetulus, taken in 1 ! 1 1 i i 1 1 1 1 O »o O E 30 o O O O £ OOO o o »do o QDOD 0% - O (DO O O o> O O CO O i. O 20 mm ESL, dial calipers were used. The reason for using ENL and ESL rather than standard length (SL) was that in most specimens the sword (bill) was damaged and standard length measurement would have been inaccurate. A series of 220 Xiphias gladius from 3.7 mm ENL to 668 mm ESL captured with plankton nets, or by night light and dip netting, or taken from dolphin fish, Coryphaena hippurus, stomachs were cleared and stained for cartilage and bone by a combined method after Taylor (1967) and Dingerkus and Uhler (1977). Mea- surements of the specimens were taken after clearing and staining, because almost all Xiphias were twisted before clearing but were easily straightened after the clearing. Although we had many smaller sized Xiphias larvae, we could have used more juveniles for our study (Fig. 1). Most of our specimens were col- lected in the Gulf of Mexico but a few were caught in the Caribbean Sea and Atlantic Ocean (Fig. 2). All specimens were examined in 100% glycerin and under 100X to 150X magnification with a high-quality binocular dissecting microscope. Cartilage was viewed with the help of alcian blue stain, but cartilaginous structures that some- times stained weakly or not at all were viewed by manipulating light intensity and the angle of the substage mirror. Onset of ossification was deter- mined by light (pink) alizarin uptake, usually around the margin of a structure. Illustrations were drawn with the help of a camera lucida. The osteological terms used in this study follow those used by Gosline (1961a, b), Nybelin (1963), Gibbs and Collette (1967), Monod (1968), and Potthoff (1975, 1980). Counts of pterygiophores and fin rays include very small vestigial structures. PECTORAL FIN The pectoral fin rays in Xiphias were the first of all fin rays to begin development. The first rays were present at 4.8-5.6 mm ENL (Tables 1, 2). Development of the rays started on the dorsal border of the larval fin blade and proceeded in a Figure 1.— Length-frequency distribution of cleared and stained Xiphias gladim used for this study. 131 188 225 668 4 5 5.5 6.5 7.5 8.5 9.5 10 5 115 12 5 13 5 14 5 15 5 16 5 17 5 18 5 19 5 20 5 25 5 35 5 45 5 65 5 LENGTH.mmENLor ESL 162 POTTHOFF and KELLEY: OSTKOLOCICAL DKVKLOPMKNT IN SWORDFISH Figure 2.— Capture localities (black dots) of larval and juve- nile Xiphias glad i us used in this study. A locality may repre- sent more than one capture. rays were still developing, 2 specimens differed by two rays (1.3%) between sides, 63 differed by one ray (40.9%), and 80 Xiphias (57.8%) had the same count on both pectoral fins. Of 20 specimens 19.6-668 mm ESL, which had adult counts, 10 differed by one ray between sides and 10 had the same count on both sides. The position of the pectoral fin in Xiph ias is on the side of larvae but changes during growth to ventrad in adults near the spot where the pelvic fin is located in most Perciformes. Xiphias lacks a pelvic fin and no vestiges of it were found during development (Gregory and Conrad 1937; Leim and Scott 1966; Ovchinnikov 1970; Yasuda et al. 1978). PECTORAL FIN SUPPORTS Table 1. — Summary of fin development sequence in cleared and stained larvae of Xiph tan gladius. PRC = principal caudal rays, SCR = secondary caudal rays. Length ENL or ESL (mm) Fin First appearance of rays All specimens have rays Full complement of rays Number of rays in fully developed fin Caudal PCR SCR Dorsal fin Anal fin Pectoral fin 5.4 5.4 7.8 5.5 5.3 4.8 6.1 6.1 11.6 6.1 6.1 5.6 26.7 8.8-11.0 26.7 8 1-13.9 78-10.6 14.2-196 34-38 17 8-10 dorsal 8-11 ventral 44-49 16-19 16-19 ventral direction. Adult counts of 16-19 rays were first obtained at 13.3 mm ESL and all specimens >19.5 mm ESL had the adult count (N = 20, X= 17.6, SD = 0.89) (Table 2). Pectoral fin ray counts differed for individual specimens between sides. Of 154 specimens 4.6 mm ENL-19.5 mm ESL, in which the pectoral The pectoral rays were directly and indirectly supported by the bones of the pectoral girdle and its suspensorium. In fully developed juveniles the girdle consisted on each side of a scapula and a distal scapular radial (which supported the dorsalmost ray directly and which orginated from scapular cartilage), four large radials (which supported the remainder of the rays directly), a coracoid, and a cleithrum (Figs. 3-5). The scapula was connected to the coracoid by cartilage (Figs. 4, 5). The pectoral suspensorium consisted of a posttemporal, a supracleithrum, and a single postcleithrum. The posttemporal and supracleithrum were connected from the rear of the skull to the lateral side of the posterior process of the cleithrum. The single post- cleithrum extended over the abdominal area and articulated on the medial side of the posterior process of the cleithrum (Figs. 3-5). The pectoral Table 2.— Development of left pectoral fin rays for Xiphias gladius (3.7 mm ENL-225, 668 mm ESL). X = mean, SD = standard deviation. Length, mm ENL or ESL Number of rays 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 SD 3.6-4.5 4.6-5.5 5.6-6.5 6.6-7.5 7 6-8.5 8.6-9.5 9.6-10.5 10.6-11.5 11.6-12 5 12.6-13.5 13.6-14.5 14.6-15.5 15.6-16.5 16.6-17 .5 17.6-18.5 18.6-195 196-668 7 20 2 2 6 2 2 1 — — 1 1 10 1.6 64 89 10.7 11.3 12.6 13.1 14.1 14.5 14.8 14.8 16.4 165 16.3 17.6 2.03 2 19 1.12 1.37 1.08 1.33 1.20 0.60 1.80 089 1.57 045 1.20 208 089 163 FISHERY BULLETIN: VOL. 80, NO. 2 Fnfld Figure 3.— Left lateral external view of the pectoral girdle and suspensorium from Xiphias gladius, showing the ontogeny. Starting from left the specimens' lengths in millimeters are: top, 5.1 ENL; 7.6 ESL; bottom, 21.4 ESL. A, anterior process of the coraco-scapular carti- lage; Bl, larval pectoral fin blade; CI, cleithrum; Cor, coracoid; Fnfld, larval finfold; P, posterior process of the coraco-scapular cartilage; PCI, posterior process of cleithrum; PstCl, post- cleithrum; Pt, posttemporal; R, radial orignat- ing from larval fin blade; ScF, scapular fora- men; SCI, supracleithrum; ScR, cartilaginous distal radial originating from scapular carti- lage. Cartilage, white; ossifying, stippled. PstCl Figure 4.— Left lateral external view of the pectoral girdle and suspensorium from Xiphias gladius, showing the ontogeny. The specimens' lengths in millimeters ESL are: left, 33.0; right, 64.6. Sc, scapula; for other abbreviations, see Figure 3. Cartilage, white; ossifying, stippled. 164 POTTHOFF and KELLEY: OSTEOLOGKAL DEVELOPMENT IN SWORDFISH Table 3.— Development of the pectoral girdle and sus- pensorium for 190 Xiphias gladius (3.7 mm ENL-64.6 mm ESL). length ranges (mm, ENL, or ESL) are from "first ob- servance" to "first observance in all specimens." PstCl Figure 5.— Left lateral external view of the pectoral girdle and suspensorium from a 187 mm ESL Xiphias gladius. For abbreviations, see Figures 3 and 4. Cartilage, white; ossify- ing, stippled. girdle is only briefly mentioned in Gregory and Conrad (1937) and no detailed description is given. Our smallest 3.7 mm ENL specimen already had rudiments of a pectoral girdle, consisting of a rod-shaped bony cleithrum, an inverted Y- shaped coraco-scapular cartilage without scapular foramen, and a larval fin blade (similar to the 5.1 mm ENL specimen in Fig. 3) (Table 3). The cleithrum later developed a shelflike dorsal posterior process (Figs. 3-5). The coraco- scapular cartilage at first had long dorsal and long posterior processes and a short anterior process. It developed a foramen on the dorsal process, and the anterior process grew relatively larger and ossified into part of the coracoid, while the posterior process atrophied. Ossifica- tion of the scapula started around the scapular foramen and spread over the dorsal process forming the scapula (Figs. 3, 4; Table 3). The larval fin consisted of two parts: a flat cartilagin- ous semicircular blade surrounded on the cir- Appearance in Part cartilage Ossification Posttemporal — 5.3 Supracleithrum — 53 Postcleithrum — 53 Cleithrum — <3.7 Posterior process of c eithrum — 6.2-6.9 Coraco-scapular cartll age <3.7 — Scapular foramen 4.6-5.1 — Scapula — 6.6-8.1 Coracoid — 5.4-6.5 Scapular radial 55-9.3 10.6-15.0 Radial No. 1 52-5.6 8.8 Radial No 2 5.2-5.9 90-10.0 Radial No 3 5.4-9.1 9 1-12.0 Radial No. 4 6.8-9.1 133-147 cumference by a finfold containing larval actinopterygia (Fig. 3). The semicircular carti- laginous pectoral fin blade developed into the four large radials by first forming elongate holes in the blade. These holes then gradually enlarged to the border of the semicircular cartilage blade, forming separate cartilaginous radials, which later ossified (Figs. 3-5; Table 3). The pectoral suspensorium, consisting of the posttemporal, supracleithrum, and postclei- thrum, was of dermal origin (did not form from cartilage) and was first seen ossifying at 5.3 mm ENL (Table 3). The posttemporal was at first a flat rectangular bone with spines. The spines were lost and a dorsal and ventral process developed, giving the posttemporal the charac- teristic inverted C shape (Figs. 3-5; Table 3). The supracleithrum was short at first and had spines. It also lost its spines and developed a long pos- terior process which articulated laterally with the posterior process of the cleithrum (Figs. 3-5; Table 3). Lengthening of the supracleithrum accommodates the migration of the pectoral fin from a lateral position in the larvae to a more ventral position in the adults (Ovchinnikov 1970). The postcleithrum was an elongate rod- shaped bone without spines from the start and articulated medially with the posterior process of the cleithrum (Figs. 3-5; Table 3). DORSAL FIN Dorsal fin rays first appeared almost at the same sizes as the anal and caudal rays (Tables 1, 4). The dorsal fin rays developed in the dorsal finfold first at the middle of the body above the 10th-14th myomere in specimens 5.5-6.1 mm ENL. With growth, addition of dorsal fin rays 165 Table 4.— Summary of dorsal fin ray development for 208 Xiphias gladius (3.7 mm ENL-225, 668 mm ESL). Length. Range, number Mean, number mm ENL of dorsal dorsal fin or ESL N fin rays rays SD 3.6-45 7 — 4.6-5.5 47 0-32 1.3 617 5.6-6.5 31 0-38 23.0 1392 6.6-75 14 27-42 35.7 4.49 7.6-8.5 21 36-45 41.0 2.75 8.6-9.5 13 40-44 42.2 1.44 9.6-10 5 11 40-45 422 1.66 10.6-11.5 8 42-47 438 1 70 11.6-12.5 9 40-48 449 2.40 12.6-135 4 42-46 43.8 1 80 13.6-145 5 43-48 44.8 201 14.6-668.0 38 44-49 464 1.23 Figure 6.— Schematic representation of dorsal and anal fin and pterygiophore development in Xiphias gladius in relation to the vertebral column and head. Pterygio- phores are represented white when cartilagi- nous and black when ossifying. Scales represent interneural and interhaemal space numbers and points on scales align with tips of neural and haemal spines. FISHERY BULLETIN: VOL. 80. NO. 2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1 1 — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i-~i u^mssssm"^ 5-0 mm ENL ft ummmms 5.3mm ENL /////////////// f/L wmvm w\\\v 5.6 mm ESL ^m^^WWW^MWM¥ j 1 2 3 4 S 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ^///////////////////////WM/// ^mw////////////////////////// //////? was in an anterior and posterior direction. The posterior part of the dorsal fin was complete at a smaller size before the anterior part. Adult dorsal fin counts of 44-49 rays (14.6-668 mm ESL, N = 38, X = 46.4, SD = 1.23) were first ob- served at 8.1 mm ESL, and all specimens longer than 13.8 mm ESL had the adult count (Fig. 6; Table 4). Our counts are in agreement with Arata (1954). Some of Arata's specimens, how- ever, did not have adult counts. The sequence of dorsal fin ray development is similar in Xiphias to that of Curyphaena reported by Potthoff (1980). DORSAL FIN PTERYGIOPHORES In juvenile and adult specimens of Xiphias 14.1-668 mm ESL, the pterygiophores consisted of a jointed proximal and distal radial support- ing a fin ray. The distal radial was located between the bifurcate base of the fin ray. Each proximal and distal radial and fin ray were forming a series, hence a serial association. Each fin ray also closely approximated the following posterior pterygiophore in a secondary associa- tion. Distal radials were present for all fin rays in 14 out of 37 juvenile specimens. Of the remaining 23 specimens 19 had one anteriormost ray and 4 had two anteriormost rays without distal radials (Table 5). Exceptions to the serial and secondary fin ray associations were found at the beginning and end of the fins. The anteriormost pterygio- phore supported from one to three rays, most often two (Fig. 7). This pterygiophore consisted of one piece of cartilage, or of a Y-shaped piece, or of two fused pieces (Figs. 7, 8). In 1 of 38 speci- Table 5.— Percent and number of anterior dorsal and anal fin rays without distal radials for 37 Xiphiasgladius( 14.7-668 mm ESL). Percent and number under are specimens in which all fin ravs had distal radials. Number of anter lor dorsal and anal fin rays Item 1 2 Percent without dorsal distal radial(N) Percent without anal distal radial(W) 37,8(14) 86.5(32) 51 4(19) 13.5(5) 108(4) 1000(37) 166 POTTHOFF and KELLEY: OSTEOLOGICAL DEVELOPMENT IN SWORDFISH Art, most Number, dorsal in rays associated dorsal pterrgiophore shape with in tenormos t pterfgiophore 1 2 3 c? 4 23 2 f 6 (7 2 (ntfrior moit Nm er, anal fin rajs associated ■III ptptuiophoif shape with anteriormost pterygiophore 1 2 3 \ 1 21 K 10 4 . 1 Figure 7.— Threee possible shapes of anteriormost dorsal and anal pterygiophores for 37 Xiphias gladius 14.7-668 mm ESL and the number of fin rays associated with each pterygiophore shape. 025mm mens, no rays were associated with the anterior most pterygiophore. The posteriormost dorsal fin ray was double and was serially associated with the posteriormost pterygiophore (Figs. 9, 10). The double ray lacked a secondary associa- tion, but a stay was present under the double ray (Figs. 9, 10). Middle radials were absent in Xiphias. Total dorsal pterygiophore count was either equal to or one to two less than the dorsal fin ray count, depending on the number of rays associated with the anteriormost pterygiophore. In larvae, juveniles, and small adults of Xiphias the dorsal proximal radials inserted in the interneural spaces. In 39 juveniles and small adults with fully formed fins, the first inter- neural space (bounded by head and first neural spine) lacked inserting pterygiophores or pre- dorsal bones. The second interneural space (bounded by first and second neural spines) had four to seven {X- 5.2), the third space had three tofive (X - 4.2), the fourth space had two to three (X = 2.9), the fifth space had two to three (X= 2.4), and the remainder of the interneural spaces had one to three pterygiophores, but usually two Figure 8.— Left lateral view of the two anteriormost dorsal fin pterygiophores with their associated rays in the second interneural space for various sizes of Xiphias gladius. Starting from left the specimens' lengths in millimeters ESL are: top row, 15.9, 20.4; middle row, 26.7, 33.6; bottom row, 52.4, 225. D, distal radial; NPr. neural prezygapophysis; Ns, neural spine; P, proximal radial; R, fin ray. Carti- lage, white; ossifying, stippled. (Fig. 11). Usually the posteriormost dorsal pterygiophore inserted in the 22d interneural space and occasionally in the 21st (Fig. 11; Tables 6, 7). In Xiphias, dorsal fin pterygiophores first appeared in cartilage before the fin rays at 4.8 mm ENL, but not until 6.0 mm ENL did all specimens have cartilaginous pterygiophores. Two Xiphias, 5.1 and 5.6 mm ENL, lacked dorsal pterygiophores but had some cartilagi- nous anal pterygiophores. Dorsal pterygiophores were first seen at the center of the body between the 11th and 18th interneural spaces (Fig. 6; 167 FISHERY BULLETIN: VOL. 80, NO. 2 D Figure 10.— Posterior-most dorsal pterygiophore and its stay from a 668 mm ESL Xiphias gladius. Top, left lateral view of proximal and distal radial, double ray and stay; bottom, dorsal view of stay, enlarged. For abbreviations, see Figures 8 and 9. Cartilage, white; bone, stippled. 2.0 mm Figure 9.— Left lateral view of the posterior- most dorsal pterygiophore from Xiphias gla- dius, showing the ontogeny. Starting from the top and going to the bottom the specimens' lengths in millimeters ESL are: 15.9, 20.4, 26.7, 33.6, 52.4, 225, length unknown for last on bot- tom, weight 61 lb. St, stay; for other abbrevia- tions, see Figure 8. Cartilage, white; ossifying, stippled. Table 6.— Adult and juvenile position of posterior- most dorsal and anal fin pterygiophores in their interneural and interhaemal spaces for 116 Xiphias gladius (7.1-668 mm ESL). Interneural space numbers DORSAL FIN i 18 CO 1 1 CO IO CO ro I co i co i ro i i ro i co i to 1 ro ro 1 ro ro 1 ro 1 ro ro 1 CO 1 CO O 1 CO F 29 36 23 36 37 39 33 38 31 29 35 34 33 32 32 29 26 33 30 30 E 6 i I \ 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 5 i 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 7 18 19 20 21 22 B 5 1 2 3 4 5 6 7 8 £ 10 11 12 13 1415 16 1 71 819J2021 22232425 16|17 1 8 19 20 21 B 10 2 2 2 1 C 11 2 2 2 1 D 28 34 29 37 26 E 03 1 1 CO — i 1 CO 1 ro O 1 ro F ANALFIft Interhaemal space numbers Item 21 20 21 21 22 21 22 22 Number of specimens 29 7 79 1 Percent of specimens 25.0 6.0 68 1 9 Table 7). Addition of cartilaginous pterygio- phores was in both anterior and posterior direc- tions. The posteriormost interneural space number 21 or 22 was filled first (Fig. 6; Table 7). Addition of pterygiophores was then in an anterior direction until the anterior interneural Figure 11. — Schematic presentation of common arrangement of pterygiophores and fin rays in relation to neural and haemal spines and vertebrae in 39 Xiphias gladius (14.7-668 mm ESL). Method of presentation modified after Matsui (1967). A, skull and vertebrae numbers; B, interneural and interhaemal space numbers; C, number of pterygiophores with highest frequency of occurrence found in the respective ("B") interneural or interhaemal space; D, number of fin rays associated with pterygiophores for indicated interneural or interhaemal space; E, highest frequency of occurrence in 39 Xiphias for the number of pterygiophores indicated in "C"; F, range of number of pterygiophores found in the respective ("B") interneural and interhaemal spaces. space number 2 was occupied (Fig. 6; Table 7). Fin rays followed pterygiophore appearance at the center of the body. Addition of rays followed addition of pterygiophores, with some cartila- ginous pterygiophores present anterior and posterior to the developing rays (Fig. 6). Ossification of dorsal pterygiophores first started at 6. 1 mm E N L in the same area and pro- ceeded in the same direction as the cartilage development. Every specimen >8.0 mm ESL had some ossifying pterygiophores, and between 18.2 and 26.7 mm ESL all pterygiophores were ossi- fying. The last pterygiophore to ossify was the anteriormost in the second interneural space. 168 I'OTTHOFF and KELLEY: OSTEOLOGICAL DEVELOPMENT IN SWORUFISH Table 7.— Development of dorsal and anal fin pterygiophores in the interneural and interhaemal spaces for 205 Xiphias gladius. X - mean. With pterygiophores With ossifying pterygiophores Length, Anteriormost space no. (X) Posterlormost space no. (X) Anteriormost space no. (X) Posteriormost space no (X) or ESL Interneural Interhaemal Interneural Interhaemal Interneural Interhaemal Interneural Interhaemal 3.6-4.5 (') (') (') (') ( 2 ) ( 2 ) ( 2 ) < 2 > 4 6-55 '3-11(5 0) '16-18(17 .1) '17-22(20.0) '19-21(20 1) < 2 ) ( 2 ) ( 2 ) ( 2 ) 56-65 '2-6 (3.2) 16-18(165) '20-22(21 4) 20-22(20.6) 2 7-9 (8 0) 2 16-17(16.8) 2 16-18(17.0) 2 17-19(180) 6 6-7 5 2-4 (2.8) 16-17(16.4) 21-22(21 6) 20-21(20 5) 2 4-14(92) 2 16-17(16.5) 2 14-19(173) 2 17-20(189) 7 6-8 5 2-3 (2.1) 16-17(16.5) 21-22(21 9) 20-21(20 9) 2 3-1 1(5.4) 2 16-18(16.4) 2 13-22(19 1) 2 17-21(190) 8.6-9 5 2-3 (2.1) 16-17(16.2) 21-22(21 8) 20-21(21 1) 3-12(4.7) 16-17(16.2) 14-22(194) 16-21(192) 9.6-105 2-3 (2.1) 16-17(163) 21-22(21 7) 20-21(20 6) 2-5 (3.5) 16-17(16.3) 19-22(20.5) 19-21(198) 10.6-11 5 2 16-17(164) 21-22(21 9) 20-21(20.9) 2-5 (3.8) 16-17(162) 18-22(207) 19-21(20 2) 11.6-12.5 2 16-17(16.4) 21-22(21 7) 20-21(20 6) 2-4 (2.8) 16-17(16.4) 18-22(208) 19-21(20 4) 126-135 2 16-17(16.8) 21-22(21 5) 20-21(20 8) 2-3 (2.5) 16-17(16.8) 21-22(21 3) 19-21(20.0) 13.6-145 2 16-17(16.4) 21-22(21.4) 20-21(20 6) 2-3 (2.2) 16-17(16.4) 20-22(21.0) 20-21(20.4) 14.6-155 2 16-17(166) 21-22(21.8) 20-21(20 8) 2 16-17(16.8) 21-22(21 8) 20-21(20 8) 15.6-165 2 16-17(162) 21-22(21.4) 20-21(20 4) 2-3 (2.4) 16 21-22(21 4) 20-21(20 4) 166-668 2 16-17(16.4) 21-22(21.5) 20-22(207) 2 16-17(164) 21-22(21.5) 19-21(20.6) 'No pterygiophores developed in all or some specimens; these were not used for calculation of means 2 No pterygiophores ossified in all or some specimens; these were not used for calculation of means Pterygiophores under the middle of the dorsal fin completed development first. Proximal and distal radials first appeared as one piece of carti- lage. Then the distal radial cartilage separated from the proximal radial. Ossification of the proximal radial cartilage started at the middle and spread outwards proximally and distally toward the ends. The ends remained carti- laginous in adults, and small sagittal keels developed ventrad during ossification (Fig. 12). Extensive lateral keels were observed on the pterygiophores in^he largest 668 mm ESL speci- men. The posteriormost pterygiophores ossified later, but in the same sequence as those in the middle area. The last pterygiophores supported a double ray in series and a stay was present (Figs. 9, 10). The posteriormost pterygiophore and the stay ossified from the same piece of carti- lage (Figs. 9, 10). The anteriormost pterygiophores were the last to ossify. The first anteriormost pterygiophore developed a large anterior sagittal keel (Fig. 8). Distal radials developed from a piece of carti- lage that separated during development from the distal portion of the cartilaginous pterygio- phores and was situated between the bifurcate bases of the serial fin rays (Figs. 8, 12, 13). Ossi- fication of all distal radials occurred after cartilage separation. At first the left and right sides of the distal radial cartilage ossified to form two pieces of bone. Ossification continued until the two bones were joined (Figs. 14, 15). All dorsal fin rays associated with the distal radials had bifurcated bases (Figs. 14, 15). 1.0mm Figure 12.— Left lateral view of a dorsal pterygio- phore from the 11th interneural space of Xiphias gladius, showing the ontogeny. Starting from the top and going to the bottom the specimens' lengths in millimeters ESL are: 15.9, 20.4, 26.7, 33.6, 52.4, 225. For abbreviations, see Figure 8. Cartilage, white; ossifying, stippled. ANAL FIN Anal fin rays first appeared at about the same sizes as the dorsal and caudal rays (Tables 1, 8). The anal rays developed in the anal finfold first at the middle of the fin below myomeres 18-20 in specimens 5.3-6.1 mm ENL. Anal rays were 169 FISHERY BULLETIN: VOL. 80, NO. 2 0.5mm Figure 13.— Anteriormost three vertebrae and pterygiophores with fin rays from a 35.9 mm ESL Xiphias gladius. C, centrum; D, distal radial; F, neural foramen; HPo, haemal post- zygapophysis; NPo, neural postzygapophysis; NPr, neural prezygapophysis; Ns, neural spine; P, proximal radial; Pa, parapophysis; R, ray. 0.5mm 0.1mm 0-5 mm 1 mm Figure 14.— Anterior view of the 12th dorsal ray and its distal radial from Xiph ias gladias, showing the ontogeny. Starting from left the specimens' lengths in millimeters ESL are: top, 64.6, 187; bottom, 225, 668. D, distal radial; R, fin ray. Car- tilage, white; ossifying, stippled. 0.25 mm Figure 15. — Anterior view of two fin rays and their distal radials from juvenile Xiphias gladius. The specimens' lengths in millimeters ESL are: left, 225, first anteriormost dorsal ray; right, 668, next to last posteriormost dorsal ray. D, distal radial; R, fin ray. Cartilage, white; bone, stippled. added in an anterior and posterior direction (Fig. 6). Adult anaUounts of 16-19 rays (10.6-668 mm ESL, TV = 66, X = 17.1, SD = 0.81) were first observed at 7.8 mm ESL and all specimens longer than 10.6 mm ESL had the adult counts (Fig. 6; Table 8). Our counts generally agree with those of Arata (1954), except we had two speci- mens with 19 anal rays; Arata had none. Table 8.— Development of anal fin rays for 213 Xiphias gladius (3.7 mm ENL-225, 668 mm ESL). X = mean, SD = standard deviation. Length, mm ENL Number of anal f n rays or ESL 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 X SD 3.6-4.5 7 — — 4.6-5.5 43 2 — — — — 1 — 1 0.6 2.06 5.6-65 6 — — 2 — 1 3 — 8 2 4 2 4 1 95 5 17 6.6-75 1 — 5 1 4 2 1 14.2 1.50 7.6-85 1 8 5 5 2 160 0.92 8.6-95 2 8 3 16.1 0.72 9.6-10.5 1 6 3 1 1 16.6 1.04 10 6-668 13 37 15 1 17.1 0.81 170 POTTHOFF and KELLEY: OSTEOEOUK'AL DEVELOPMENT IN SWOKDFISH ANAL FIN PTERYGIOPHORES The description of the dorsal fin pterygio- phores in the previous section may be applied to anal fin pterygiophores because of the similari- ties between the two. Anal pterygiophores were inserted in the interhaemal spaces. These spaces were numbered the same as the opposing inter- neural spaces. Anteriormost (first) interhaemal space number 16 or 17 was bound anteriorly by the stomach, intestine, and anus and posteriorly by the first haemal spine. The first haemal spine was positioned on the 16th or 17th centrum. If it occurred on the 16th centrum, it was of variable length and often did not reach the pterygio- phores. If the first haemal spine was on the 17th centrum, it always reached past the pterygio- phores. The count for the 16th and 17th interhaemal space was summed because we were not always able to determine a division between the two spaces (Fig. 11). Total number of anal pterygiophores in 31 of 37 specimens with full counts was one less than the anal fin ray count. In 2 of 37 specimens, it was the same and in 4 of 37 it was two less. The ante- riormost anal pterygiophore supported from one to three rays, most often two (Fig. 7). This pterygiophore consisted of one piece of carti- lage, normal in shape (Fig. 16), or of a vestige (Fig. 7). The vestigial piece may fuse to the next posterior pterygiophore to form an inverted Y shape (Fig. 16), or the inverted Y shape may originate from one piece of cartilage (Figs. 7, 16). An anterior sagittal keel developed on the ante- riormost anal pterygiophore (Fig. 16), but this keel was not as large as on the first dorsal pterygiophore (Fig. 8). The posteriormost anal pterygiophore had the same structure as its dorsal counterpart and in- serted most often into the 20th or 21st inter- haemal space, which was usually one space ante- rior to the posteriormost dorsal insertion (Fig. 11; Table 6). In juveniles and small adults of Xiphias with fully formed fins the anteriormost interhaemal spaces 16 and 17 had 8-11 (X = 9.9, N = 40) pterygiophores. The remaining three or four interhaemal spaces had one to two or one to three pterygiophores each (Fig. 11). The pos- teriormost 21st interhaemal space had none or one to two pterygiophores. Only 1 specimen out of 116 had a pterygiophore in the 22d inter- haemal space (Table 6). Development and structure of the anal fin Figure 16.— Left lateral view of two or three anteriormost anal fin pterygiophores from Xiphias gladius, showing the ontogeny. Starting from left the specimens' lengths in milli- meters ESL are: top row, 15.9, 20.4; bottom row, 33.0, 64.6, 225. D, distal radial; P, proximal radial; R, fin ray. Cartilage, white; ossifying, stippled. pterygiophores was the same as in the dorsal supports. Cartilaginous anal pterygiophores first appeared before anal fin rays and most of the time concurrently with dorsal pterygiophores below myomeres 18-20 (which approximately corresponds to interhaemal spaces 18-20) (Fig. 6; Table 7). Addition of cartilaginous pterygio- phores was in an anterior and posterior direction. The posteriormost interhaemal spaces 20 or 21 were filled first. Last to develop was the anterior- most anal pterygiophore (Fig. 6). Fin rays fol- lowed pterygiophore appearance as in the dorsal fin (Fig. 6). Ossification of anal fin pterygiophores first started between 6.0 and 8.0 mm ENL or ESL in the same area of first appearance in cartilage and proceeded in the same directions as cartilage development (Fig. 6; Table 7). All anal pterygio- phores were ossifying between 12.0 and 25.1 mm ESL. Development and ossification of individual anal pterygiophores is similar to the dorsal pterygiophores (Fig. 16). The posteriormost anal pterygiophore develops a stay and supports a double ray serially as does its dorsal counterpart. Distal radials developed in the anal fin as in the dorsal fin (Fig. 14). Almost all rays had a distal radial between their bifurcate base. Only 5 out of 171 FISHERY BULLETIN: VOL. 80, NO. 2 37 specimens did not have a distal radial for the anteriormost ray (Table 5). CAUDAL FIN Caudal fin rays first appeared at about the same sizes as the dorsal and anal rays (Table 1). The caudal fin rays developed in the caudal fin- fold ventrad in preflexion larvae first on hypurals 2 and 3 and were added in an anterior and pos- terior direction. After complete notochord flexion between 6.3 and 8.0 mm ESL, the secondary caudal rays developed dorsad and ventrad in an anterior direction. Caudal rays were first seen in a 5.4 mm ENL specimen and all larvae longer than 6.1 mm ENL had some caudal rays developing (Table 9). The full complement of 9+8 principal rays developed between 8.8 and 11.0 mm ESL. All Xiphias longer than 26.6 mm ESL had the adult_count of (8-10)+9+8+(9- ll)=34-38 (N = 15, X = 35.9, SD = 1.55) rays (Tables 9, 10). The upper and lower caudal lobe had equal numbers of rays or they differed by one ray (Table 10). A procurrent spur (Johnson 1975) was not oberved in Xiphias. CAUDAL FIN SUPPORTS The caudal fin rays were supported by some of the bones of the hypural complex and only two posteriormost centra (PU2 and urostyle) were in- volved in the support (Fig. 17). The bones which supported the fin rays directly or indirectly in larvae and juveniles of Xiphias were two centra (PU2 and urostyle), one specialized neural arch, three epurals, one paired uroneural, five auto- genous hypural bones, one autogenous par- hypural, and one autogenous haemal spine. One of 164 specimens examined had the unusual count of 16+11=27 vertebrae and had two autogenous haemal spines on preural centra 2 and 3. We were able to see all these supporting bones during development (Figs. 18-23; Table 11), but in the adults some parts were ontogeneti- cally fused. Between 3.7 and 6.2 mm ENL, Xiphias had a straight notochord in the caudal area. Notochord flexion was between 6.3 and 8.0 mm ENL. Before notochord flexion hypurals 1-4, the parhypural (Ph), and the haemal spine and arch (Hs) of the future preural centrum 2 were developing ventrad in cartilage (Fig. 18; Table 11). Dorsad the neural arch (Ns) of the future preural cen- trum 3, the specialized neural arch ("Na") of the Table 9.— Caudal fin ray development for 200 Xiphias gla- dius (3.7 mm ENL-225, 668 mm ESL). SCR, secondary caudal rays. PCR, principal caudal rays. X= mean, SE = standard error of the mean. Specimens are undergoing notochord flexion between dashed lines at 6.3-8.0 mm ENL. Length, Upper Lower Total fin ray count mm ENL or ESL SCR PCR PCR SCR Range X SE N 3.6-4.5 — — 7 4.6-5.5 0-3 0-3 0-6 0.2 1.36 46 56-65 0-6 0-8 0-14 49 3.83 30 66-7.5 2-7 2-8 4-15 9.9 2 99 14 76-8.5 4-8 4-8 0-1 8-17 13.7 262 19 86-9.5 0-1 5-9 6-8 0-2 11-19 15.5 277 12 9.6-10.5 7-9 8 0-2 16-20 17.6 1.33 11 10.6-11.5 0-1 7-9 8 0-2 16-20 18.1 1.50 9 11.6-12.5 0-2 9 8 2 19-21 19.6 085 8 126-13.5 0-3 9 8 1-3 18-23 21 3 2.20 4 13.6-155 0-3 9 8 2-3 19-23 21.5 1.41 8 15.6-17.5 3-5 9 8 3-5 23-27 243 1.71 6 17.6-265 4-7 9 8 4-8 25-32 26.0 1.99 11 266-668 8-10 9 8 9-11 34-38 35.9 1.55 15 Table 10.— Adult caudal fin ray counts for 15 Xiphias gladius (26.7- 225, 668 mm ESL). USCR = upper secondary caudal rays, PCR = prin- cipal caudal rays, LSCR = lower secondary caudal rays. Total fin ray count USCR + PCR + LSCR N 34 8 + 17 + 9 4 35 9 + 17 + 9 2 36 9 + 17 + 10 3 37 10 + 17 + 10 3 38 10 + 17 + 11 3 future preural centrum 2, and the three epurals (Ep) were developing from cartilage (Fig. 19). Appearance of the cartilaginous parts was from anterior to posterior. After notochord flexion a cartilaginous hypural 5 (Hy) and a bony uroneu- ral (Un) developed between 9.8 and 12.5 mm ESL (Figs. 20-21; Table 11). The parhypural and hypurals 1-5 developed from separate pieces of cartilage. This is shown for the parhypural and hypurals 1-2 in Figure 18. Joining of the proximal portions of the par- hypural and hypurals 1-2 by cartilage starts with the parhypural and hypural 1 between 5.4 and 5.6 mm ENL and extends to hypural 2 at 5.7 mm ENL. All specimens have the parhypural and hypurals 1-2 joined proximally with carti- lage at 6.9 mm ENL or ESL as shown in Figures 19 and 20. Hypurals 3-5 are never joined by carti- lage during development (Figs. 19-21). The car- tilaginous proximal joint is lost during develop- ment when the hypurals are fully ossified between 27 and 34 mm ESL (Fig. 22). Ossification of the cartilage bone in the caudal complex of Xiphias started with the preural 172 POTTHOFF anil KELLEY: OSTEOLOGICAL DEVELOPMENT IN SWORDFISH Figure 17.— Left lateral view of the adult caudal complex from Xiphias gladius of un- known length, 48 lb, showing fin ray articula- tion in relation to the caudal parts. Ep, epural; Hs, autogenous haemal spine; PCR, principal caudal rays; Ph, parhypural; Pu, preural centrum; SCR, secondary caudal rays; Un, uroneural; Ur, urostyle. Caudal complex bones, white; caudal rays, stippled. 10mm PCR Figure 18.— Left lateral view of the caudal complex of a 5.1 mm ENL Xiphias gladius. Ha, haemal arch; Hy, hypural; Nc, notochord; Na, neural arch; Ph, parhypural. Cartilage, stip- pled. i 1 0.25 mm Figure 19.— Left lateral view of the caudal complex of a 8.8 mm ESL Xiphias gladius. Hs, haemal spine; "Na", specialized neural arch; Ns, neural spine; for other abbrevia- tions, see Figures 17 and 18. Cartilage, white; bone, stippled. PCR SCR Figure 20.— Left lateral view of the caudal complex of a 12.6 mm ESL Xiphias gladius. HPr, haemal prezygapophysis; NPr, neural prezygapophysis; for other abbrevia- tions, see Figures 17-19. Cartilage, white; ossifying, stippled. Table 1 1.— Length ranges at which parts of the caudal complex appear in cartilage and ossify in 173 Xiphias gladius (5.4 mm ENL-225 mm ESL). Pu = preural centrum. Brackets denote fusion of separate structures during development. Length range Length range (mm. ENL or ESL) (mm. ENL or ESL) of first appearance of first evidence First evidence of in cartilage of ossification fusion (mm. ESL) Pu 2 centrum — 6.2- 90 Specialized neural arch 5.4-6.5 7 1-12.3 Epural anterior 5.7-6.8 10.3-13.7 middle 54-6.8 10.3-13.7 posterior 5.4-7.1 162-176 Uroneural — 98-12.3 Hypural 5 9.8-12.5 16.0-17.7 Hypural 4 5.7-7.9 94-13.7 Hypural 3 53-6 1 7 1-10 7 Urostyle — 6.2-9.1 Hypural 2 5 1-56 7.1-9.7 Hypural 1 5.0-5.5 7.1-9.2 Parhypural 5.0-5.5 7.1-9.2 Pu2 haemal spine 5.1-6.1 7.1-10.9 17.2-267 131 -? 17.2-226 173 Figure 21.— The caudal complex of a 21.4 mm ESL Xiphias gladius. A, left lateral view of the complex; B, left lateral view of normal uroneural, enlarged. HPo, haemal post- zygapophysis; NPo, neural postzygapophysis; for other abbre- viations, see Figures 17-20. Cartilage, white; ossifying, stippled. FISHERY BULLETIN: VOL. 80. NO. 2 SCR PCR 0.25 mm PCR Figure 22.— The caudal complex of a 52.4 mm ESL Xiphias gladius. A, left lateral view of the complex; B, left lat- eral view of the anomalous uroneural, enlarged. A, anomalous secondary haemal spine; F, neural foramen; for other abbreviations, see Figures 19-21. Cartilage, white; ossifying, stippled. 0.25 mm Un Hy 5 Figure 23.— The bones of the caudal complex from an adult Xiphias gladius length un- known, 61 lb. A, left lateral view of the caudal bones; B, left lateral view of the normal uro- neural, enlarged. For abbreviations, see Figures 19-21. Cartilage, white; bone, stippled. NPr NPo 10mm 2mm 174 POTTHOKK ami KELLEY: OSTEOLOGICAL DEVELOI'MKNT IN SWORDFISH centrum 2 and the urostyle at 6.2 mm ENL-9.1 mm ESL. Ossification then proceeded from the haemal spine of the preural centrum 2 dorsad to the hypurals. Last to ossify between 16.0 and 17.7 mm ESL was hypural 5 (Table 11). The special- ized neural arch of preural centrum 2 began ossification at 7.1-12.3 mm ESL followed by the three epurals. The posteriormostepural was last to ossify between 16.2 and 17.6 mm ESL (Table 11). The paired uroneural was not a cartilage bone and it was first present between 9.8 and 12.3 mm ESL before epural ossification (Table 11). In a few specimens the uroneural had an anomalous shape as if it had fused from two parts (Fig. 22). During development of the hypural complex, a parhypurapophysis and a hypurapophysis (Lundberg and Baskin 1969; Nursall 1963) were observed on the parhypural and hypural 1. From a dorsal view the parhypural and hypural 1 are bifurcated as shown in Figure 24. This bifurca- tion can be observed in the adults on the autogenous parhypural but is absent on hypural 1, which then is fused to the hypural plate. A tunnellike foramen develops between the tips and rear of the parhypural prezygapophyses for the haemal canal on the proximal surface of the parhypural. This tunnel was not yet developed in a 44.1 mm ESL specimen (Fig. 24) but was fully formed in our 668 mm ESL specimen. In adults of Xiphias, hypurals 1-4 fuse with each other and the urostyle, forming a single hypural plate with a notch posteriorly at the center. Grooves present on the plate formed be- cause of articulating rays (Gregory and Conrad 1937) (Fig. 23). The epurals, the uroneural, hypural 5, the parhypural, and the haemal spine of preural centrum 2 remained autogenous in the adults. Fusion between hypurals 4 and 3 and 1 and 2 started distad from the articular cartilage in an anterior direction at 17.2-26.7 mm ESL (Figs. 21, 22; Table 11). Fusion of the two hypural plates, however, was in a posterior direction starting proximally. We could not determine the size at which the dorsal and ventral hypural plates fused with each other and with the urostyle because of insufficient samples (Fig. 1; Table 11). The parhypural and hypurals 1-5 supported the principal caudal rays. Only on one occasion did the haemal spine of preural centrum 2 support a principal caudal ray, but this is not shown in Table 12. The distribution of principal rays on the hypural bones can only be seen in Haemal Canal 0.5 mm FIGURE 24.— The parhypural and hypural 1 from a 44.1 mm ESL Xiphias gladius. A, dorsal view, enlarged; B, left lateral view. Hyp, hypurapophysis; Phyp, parhypurapophysis. Carti- lage, white; bone, stippled. larvae and small juveniles (Figs. 19-22; Table 12). Table 12.— Distribution of principal caudal rays on the hypurals in 66 Xiphias gladius (8.8- 64.6 mm ESL). Number of principal caudal rays Part 1 2 3 4 5 6 Parhypural 2 61 3 Hypural 1 1 18 47 Hypural 2 49 17 Hypural 3 15 48 3 Hypural 4 1 18 43 4 Hypural 5 38 28 VERTEBRAL COLUMN Of 164 Xiphias 5.3 mm ENL-668 mm ESL, 1 (0.6%) had 15+10=25 vertebrae, 95 (57.9%) had 15+11=26, 65 (39.7%) had 16+10=26, and 3 (1.8%) had 16+11=27 (Nakamura et al. 1968; Ovchin- nikov 1970). All centra except the first anteriormost, the urostyle, and preural centrum 2 had neural pre- and postzygapophyses, and neural arches and spines (Figs. 25-27). The first anteriormost centrum lacked a neural prezygapophysis(Figs. 13, 27), preural centrum 2 had a neural prezyg- apophysis, a specialized (open) neural arch, and a neural postzygapophysis (Figs. 22, 23). The urostyle had only a neural prezygapophysis (Figs. 21-23). All precaudal vertebrae except the anteriormost had parapophyses (Figs. 13, 25, 175 FISHERY BULLETIN: VOL. 80, NO. 2 1mm Figure 25.— Left lateral view of the second anteriormost vertebra from Xiphias gladius, showing the ontogeny. Start- ing from left the specimens' lengths in millimeters are: top, 5.1 ENL, 7.8 ESL.12.6 ESL; center, 21.4 ESL, 52.4 ESL; bottom, 225 ESL. F, neural foramen; Nc, notochord; NPo, neural postzygapophysis; NPr, neural prezygapophysis; Ns, neural spine; Pa, parapophysis. Cartilage, white (except in 5.1 mm ENL specimen in top row left where entire stippling signifies cartilage); ossifying, stippled. 26). Haemal postzygapophyses were present on precaudal vertebrae numbers 3 to 15, sometimes on 2 to 15 (Figs. 13, 26). All caudal vertebrae had nonautogenous haemal spines, except preural centrum 2 and the urostyle. Preural centrum 2 had an autogenous haemal spine. The urostyle had an autogenous parhypural with a tunnellike foramen for the haemal canal. The parhypural is homologous to the autogenous haemal spine of preural centrum 2 (Figs. 20-24). The 16th centrum sometimes lacked a haemal spine, sometimes had a vestigial haemal spine, or it had a normal haemal spine. Haemal pre- and postzygapophyses were present on all caudal centra except on preural centrum 2 and the urostyle. Neural foramina were present on most precaudal and caudal centra on larger specimens (Figs. 13, 22, 23, 25-28). Five out of eight Xiphias with all ribs devel- oped had six paired ventral ribs, which loosely articulated with the parapophyses on centra 1-4, 14, and 15 (Figs. 25-27). Two specimens had seven pairs of ribs on centra 1-5, 14, and 15 and on centra 1-4 and 13-15. One Xiphias had nine pairs on centra 1-6 and 14-16. The neural arches fuse distally during ossifica- tion to form neural spines. The fusion and spine formation is over a size range and proceeds from posterior in an anterior direction (Fig. 27; Table 13). Our largest four specimens of Xiphias, 131- 668 mm ESL, had three to six anterior neural arches and spines split. These arches and spines remain split in adults (Bruce B. Collette 3 ). Development of the centra starts with the appearance of distally opened cartilaginous neural arches. One arch was seen behind the head on top of the notochord in our smallest 3.7 mm ENL specimen (Fig. 29). As length in Xiphias increased, more arches were added in a posterior direction (Fig. 29; Table 14). All specimens >6.5 mm ENL had the complete count of 25 neural arches. Two cartilaginous split haemal arches were first observed at 5.0 mm ENL when 16 neural arches were present. The two haemal arches were opposite the 16th and future 17th neural arch. Additional haemal arches and spines were added in a posterior direction (Fig. 29; Table 15). 3 Bruce B. Collette, Systematic Zoologist, National Marine Fisheries Service, NOAA, Systematics Laboratory, Washing- ton, DC 20560, pers. commun. July 1981. Table 13.— Number of split neural arches and spines counted from anterior to posterior for various size ranges in 159 Xiphias gladius 5.5 mm ENL-668 mm ESL. N = number of specimens, X= mean. Length, mm ENL or ESL Centrum number with split neural arches and spines 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 N X 5.5-6.9 15 2 5 3 3 5 113 1 1 2 5 38 17.1 7.0-133 5 5 12 18 18 10 3 2 1 — 2 — — — — 1 77 107 13 6-64.6 2 6 8 17 6 — 1 40 86 131-668 1 — 2 1 4 48 176 POTTHOFF and KELLEY: OSTEOLOGICAL DEVELOPMENT IN SWORDFISH 0.25 mm 1mm Figure 26.— Left lateral view of the 15th vertebra from Xiphias gladius, showing the ontogeny. Starting from top left the specimens in millimeters ENL or ESL are as in Figure 25. FBr, foraminal bridge; HPo, haemal postzygapophysis; for other abbreviations, see Fig- ure 25. Cartilage, white (except in 5.1 mm ENL specimen in top row left, where entire stippling signifies cartilage); ossifying, stippled. 0-25mm Figure 27.— First and second anteriormost ver- tebrae from a 12.8 mm ESL Xiphias gladius. Top, left lateral view; bot- tom, dorsal view. For abbreviations, see Fig- gures 25 and 26. All specimens >6.0 mm ENL had the complete count of eight or nine haemal arches and spines. Ossification of the vertebral column started at 4.4 mm ENL anteriorly at the bases of the neural arches. All specimens longer than 5.0 mm ENL had some anterior vertebral column ossification. The ossification was in a posterior direction as length increased until all centra including the urostyle were ossifying in some specimens between 6.1 mm ENL and 8.1 mm ESL(Fig.29). In specimens >8.1 mm ESL all entra had some ossification. The development of the neural and haemal pre- and postzygapophyses is shown in Figures 20-23 and 25-28. Neural prezygapophyses devel- oped on all centra except the anteriormost cen- trum (Figs. 13, 27) and neural postzygapophyses developed on all centra except the urostyle (Figs. 21-23, 25-28). Haemal prezygapophyses devel- oped on all haemal spines and shifted dorsad and anteriorly onto the centrum during ontogeny (Figs. 20-23, 28); the haemal prezygapophyses on preural centrum 2 and on the parhypural re- mained on the autogenous haemal spine and the autogenous parhypural (Figs. 21-23). A neural foramen developed on each centrum except on the urostyle by first developing a neural postzygapophysis (Figs. 25-28). Then an anteriorly directed process developed on the anterodorsal side of the postzygapophysis, which joined the neural spine forming a neural foraminal bridge (Figs. 27, 28). The neural prezygapophysis of the second anterior centrum developed an entirely different shape than all other prezygapophyses and could be taken for a neural spine on small juvenile or 177 FISHERY BULLETIN: VOL. 80, NO. 2 Figure 28.— Left lateral view of the 17th vertebra from Xiphias gladius, showing the ontogeny. Starting from top left the speci- mens in millimeters ENL or ESL are as in Figure 25. Hs, haemal spine; HPr, haemal prezygapophysis; for other abbreviations, see Figures 25 and 26. Cartilage, white (except in 5.1 mm ENL specimen in top row left, where entire stippling signifies cartilage); ossifying, stippled. Table 14. — Development of the neural spines on the anterior to posterior numbered centra for 97 Xiphias gladius 3.7-7.0 mm ENL or ESL. N = number of specimens, X = mean. Length, mm ENL or ESL Centra with neural spines 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 N X 3.6-4.0 1 1 — 4.1-4.5 3 1 — 1 — — — 1 6 28 4.6-5.0 1 2 8 5 2 — 1 — — 1 — — — — 1 21 5.3 5.1-5.5 1 1 1 1 1 — - 2 — — 1 1 — — — — — 1 — 6 5 5 26 18.7 5.6-6.0 1 2 — 3 6 5 17 23.5 6.1-6.5 3 3 11 17 24.5 6.6-7.0 9 9 25.0 178 POTTHOFF anrl KELLEY: OSTEOLOGICAI, DEVELOPMENT IN SWORDFISH 2 4 6 8 10 12 14 16 18 20 22 24 26 I I I I I I I I I I I I I I I I I I I I I I I 1 3.7mm JUL^- 4.4mm 53 mm 6-2 mm 7.2 mm Figure 29.— Schematic presentation of the vertebral column development in Xiphias gladius. Ticks on scale denote centra number and are aligned with the middle of the centrum. Indicated millimeter mea- surements are ENL or ESL. Carti- lage, white; ossifying, stippled. Table 15. — Development of the haemal spines on the anterior to posterior numbered centra for 53 Xiphias gladius 5.0-6.5 mm ENL. N - number of specimens, A' = mean. Length, mm ENL or ESL Centra with hae mal spines 17 18 19 20 21 22 23 24 25 N X 5.0 1 1 17.0 5.1-5.5 1 1 — 1 1 4 3 8 19 23.4 5.6-60 1 — 2 1 2 1 9 16 23.6 6.1-6.5 17 17 25.0 larger specimens (Figs. 8, 13, 25, 27). This prezygapophysis is considerably longer than the neural spine except in large juveniles and adults (Fig. 25). Ribs developed from a short piece of proximal cartilage. The cartilage later ossified and bone cells were added distally directly in the length- ening process of the rib during development. One pair of ribs was first seen on the anteriormost centrum at 8.0 mm ESL in some specimens. All Xiphias >12.2 mm ESL had at least one pair of ribs developing. Development was in a posterior direction on the first four centra and in an anterior direction on centra 15 and 14. When centra 1-3 had developing ribs, usually a pair also was present on centrum 15. Ribs developed over a wide size range. The smallest specimen with a full set of ribs on centra 1-4, 14, and 15 measured 25.1 mm ESL and all specimens larger than 55.1 mm ESL had the full rib complement. Xiphias usually developed ribs on centra 1-4, 14, and 15, but a few specimens also had ribs on centra 5, 6, 13, and 16. BRANCHIOSTEGAL RAYS Branchiostegal rays were first seen in a 4.2 mm ENL specimen and all Xiphias >4.2 mm ENL had some rays. The 4.2 mm ENL Xiphias had four rays on each side but a 4.5 mm ENL specimen had only two (Table 16). Branchioste- gals were added from posterior to anterior di- rection, specimens with developing branchios- tegals had either the same count on both sides or differed by one ray between sides. Adult counts of seven or eight rays were first observed at 5.0 mm ENL and all Xiphias >6.6 mm ENL had 179 FISHERY BULLETIN: VOL. 80, NO. 2 Table 16.— Development of the branchiostegal rays on the left and right sides for 211 Xiphias gladius (3.7 mm ENL-225, 668 mm ESL). A^number of specimens, X=mean, SD = standard deviation. Length, Number branchiostegal rays, left N Number branchiostegal rays , right or ESL 12 3 4 5 6 7 8 X SD 12 3 4 5 6 7 8 X SD 3.6-4.5 1 4.6-55 5.6-6.5 6.6-668 — 1 - 3 7 1 8 1 11 19 2 24 71 3.6 2.12 5.9 1.34 6 7 1 0.57 56 7.4 0.45 7 45 32 127 1 — 1 1 2 8 1 6 1 13 3 16 21 78 2 8 49 3.4 2.12 6 1.34 7.1 057 7.4 0.45 180 POTTHOFF and KELLEY: OSTEOLOGICAL DEVELOPMENT IN SWORDFISH adult counts (Table 16). Of 127 Xiphias (6.6 mm ENL-668 mm ESL), 59 (46.4%) had seven bran- chiostegals on both sides, 37 (29.2%) had eight on both sides, and 31 (24.4%) had seven rays on one side and eight on the other. SQUAMATION Larvae of Xiphias developed four rows of scales on each side with smaller "scatter" scales between the rows (Fig. 30). First to appear be- tween 5.3 and 6.1 mm ENL were some ventral "row" scales on the stomach. These scales were added during growth anterior to the pectoral symphysis and posteriorly to the ventral hypurals. Dorsal row scales were first seen be- tween 5.7 and 6.9 mm ENL, approximately be- tween the 3d and 15th centrum. The addition of dorsal row scales during growth was in an ante- rior direction to the top of the head and in a pos- terior direction to the dorsal hypurals. The two lateral scale rows were first seen in some specimens between 6.5 mm ENL and 8.6 mm ESL, extending from the posterior border of the pectoral fin to about the 16th centrum. Scales were added anteriorly only to the dorsal lateral row to about the operculum and posteriorly to the urostyle. Scatter scales, between the dorsal, ventral, and lateral scale rows first developed between 6.2 and 7.1 mm ENL on the stomach just posterior to the pectoral fin and dorsad to the ventral scale row (Fig. 30). Scatter scales, which were smaller than row scales, spread from the stomach dorsad during growth until the left and right sides in an area from the 4th centrum to the 18th centrum were covered (Fig. 30). Further addition of scatter scales was then in an anterior and posterior direction covering the whole body, the sword, and the caudal fin rays at 61.5 mm ESL, but not the pectoral, dorsal, and anal fins. In our 187 mm ESL specimen the dorsal, anal, and pectoral fin rays were covered with scatter scales. In the literature, Arata (1954); Leim and Scott (1966); Nakamura et al. (1968), and Palko et al. (1981) stated that adult Xiph ias lack scales. Figure 30.— Larval and juvenile Xiphias gladius, depicting the ontogeny of squamation. The size of scales was exaggerated in proportion to the body. Starting from the top and going to the bottom the specimens' lengths in millimeters are: 5.3 ENL, 6.2 ENL, 7.6 ESL, 11.5 ESL, 35.4 ESL, 188 ESL. Our largest 668 mm ESL specimen had scales (Fig. 31), seen through the dissecting microscope on a cleared and stained piece of skin. In this specimen the row scales could no longer be dis- tinguished from the scatter scales. Development of individual scales is similar for the row and scatter scales, except scatter scales start out smaller than row scales but increase in size to equal the row scales during development. Each scale starts as an oval-shaped structure with one posteriorly recurved spine. During development more posteriorly recurved spines are acquired in a row at the center of the scales and the scale margins become progressively crenated (Fig. 32). Finally, in specimens >200 mm ESL the marginal scale crenations become fewer and the recurved spines develop into blunt stubs (Fig. 32). Individual row scales have approximately the same number of spines in a developing specimen, but this does not apply for the scatter scales. Our largest 668 mm ESL Xiphias had developed variable scales which had from one to seven blunt stubby spines; row scales were not distin- guishable from scatter scales in this specimen (Figs. 31, 32). Arata's (1954) work on scale devel- opment agrees with our findings. 0.25mm Figure 31.— Enlarged view of the skin from a 668 mm ESL Xiphias gladius, showing scales with two to six posteriorly re- curved spines. White spaces between scales are skin. Anterior is to the left. 181 FISHERY BULLETIN: VOL. 80, NO. 2 Figure 32.— Scales from Xiphias gladius, showing ontogeny. Starting from left the specimens' lengths in millimeters are: top, 5.4 ENL, 6.2 ENL, 25.1 ESL; bottom, 61.5 ESL, 225 ESL, 668 ESL. Each size in top and bottom rows has an external view (top) and a lateral view (bottom). 0.01mm 1.0mm 0.10 mm 1.0 mm 1.0 mm DISCUSSION Xiphias gladius is a highly modified perci- form fish which, in our opinion, should not be placed as the monotypic family Xiphiidae in the suborder Scrombroidei, as was done by Green- wood et al. (1966). We agree with Gosline (1968) and Fierstine (1974), who placed the monotypic family Xiphiidae under the separate suborder Xiphiioidei. However, Gregory and Conrad (1937) compared Xiphias bones with those of Istiophorus and concluded that xiphiids and istiophorids are separate but parallel families of common scombroid stock. G. David Johnson, who examined the branchial arches of Xiph ias, istio- phorids, and scombrids (unpubl. data), has evi- dence that Xiph ias belongs with the scombroids. We will discuss the modifications and variations that we noted in Xiph ias and compare these with other fish families. The pectoral fin position in Xiphias larvae is lateral, but during growth to adults the fin moves ventrad to an almost pelvic position. Xiphias probably lost its pelvic fin during phylogeny. Remnants of a basipterygium were not found by us or other workers during develop- ment of the larvae (Yasuda et al. 1978). Pectoral fin ray counts of the left and right sides were equal or differed by one ray in juvenile Xiphias. Similar results were obtained for Archosargus (Houde and Potthoff 1976), Coryphaena (Potthoff 1980), and Scombrolabrax (Potthoff et al. 1980). In tunas, larger differences in pectoral fin ray counts between sides were found (Potthoff 1974). With the publication of Dingerkus and Uhler's (1977) cartilage staining technique, Fritzsche and Johnson (1980) reported the development of pectoral radials from a sheet of cartilage in Morone. Swinnerton (1905) reported the same for Salmo salar by using the "reconstruction in wax from serial sections" technique; he called the cartilaginous blade "fin-plate." We saw the same happening in Xiph ias and labeled the sheet of cartilage "blade" (Bl) in Figure 3. It is likely that pectoral radials develop from a cartilagi- nous blade in all Perciformes, and perhaps all lower fishes. Starks (1930) reported a cartilagi- nous blade (radial plate) in adult Dallia pecto- ral is and Roberts (1981) in the salmoniform Sun- dasalangidae; we believe this to be an example of a neotenic structure. The pectoral girdle in Xiphias is reduced as compared with a basic perciform pectoral girdle such as that found in Coryphaena (Potthoff 1980) and in at least some scombrids, e.g., Sardini (Collette and Chao 1975), Acanthoeybium (Conrad 1938), and Thu>nius(de Sylva 1955). In Xiphias, the supratemporal and intertemporal bones are absent and there is only one post- cleithrum. Adult Xiphias have two dorsal and two anal 182 1'OTTIIOFF an.] KKI.I.KV: OSTKOLOOICAI, I)K VKI.OI'M KNT IN SWOKDFISH fins (Leim and Scott 1966; Ovchinnikov 1970), but larvae and juveniles have one continuous dorsal and anal fin (Nakamuraetal. 1951; Yabe et al. 1959). During development the fin rays in the center of the fins stop growing and the rays become subcutaneous. The subcutaneous rays and their pterygiophores are present in the adults and were dissected in our largest 668 mm SL specimen. In three scombrid genera, Scomber, RastreUiger, and Auxis, we find a first dorsal and second dorsal fin separation similar to that in adult Xiphias, except that in these scombrids the two fins are separate initially even though the first and second dorsal fin pterygio- phores are continuous (Kramer 1960; Potthoff pers. obs. on Auxis). There is only one anal fin in these three scombrid genera, whereas adult Xiphias have two anal fins. All dorsal rays in Xiphias are bifurcated at their bases (Figs. 14, 15) as in Coryphaena (Potthoff 1980). This probably is not the case in most perciforms where the spinous rays of the first dorsal fin have a closed base with a foramen and the distal radials are situated outside the bases of the first dorsal fin spinous rays (Kramer 1960; Potthoff 1974, 1975; Potthoff et al. 1980). The anteriormost dorsal pterygiophores in Xiphias insert in the second interneural space (Figs. 11, 13), asinthegempylidsandtrichiurids (Potthoff et al. 1980), but not as in the serranids, sparids, apogonids, scombrolabracids, and scrombrids where the anteriormost pterygio- phores insert in the third interneural space (Matsui 1967; Fraser 1972; Potthoff 1974, 1975; Houde and Potthoff 1976; Fritzsche and Johnson 1980; Potthoff et al. 1980), and not as in the coryphaenids in which they insert in the first space (Potthoff 1980). No predorsal bones were present in Xiphias. All scombrids and most scombroids also lack predorsal bones, however some gempylids, e.g., Ruvettus (Potthoff et al. 1980), have one predorsal. Most other perci- formes have predorsals in the first and second interneural spaces. The first dorsal pterygiophore in Xiphias is variable in development (Figs. 7, 8) and originates either from one or two pieces of cartilage. In scombrids (Potthoff 1974, 1975), a two-part development of the first dorsal pterygio- phore was not evidenced, but in Morone it was (Fritzsche and Johnson 1980). The last (posteriormost) pterygiophore of Xiphias has a serially associated double ray and a stay (Figs. 9, 10). In Xiphias, as probably in all Perciformes, the stay develops from the prox- imal radial cartilage. The stay is not posteriorly bifurcated as in most scombrids (Potthoff pers. obs.), nor does it ossify into two parts as in most gempylids and some trichiurids (Potthoff et al. 1980). Xiphias lacks middle radials as does Cory- phaena (Potthoff 1980), whereas many Per- ciformes probably have middle radials at least for some of the posteriormost dorsal and anal pterygiophores (Kramer 1960; Berry 1969; Potthoff 1974, 1975; Houde and Potthoff 1976; Potthoff et al. 1980; Fritzsche and Johnson 1980). In Xiphias the caudal rays are supported by only two centra (urostyle and preural centrum 2) (Figs. 17, 21, 22). This is unusual, because in most perciforms three centra support the caudal rays (Berry 1969; Houde and Potthoff 1976; Potthoff 1980; Potthoff et al. 1980; Fritzsche and Johnson 1980), and in most scombrids four or five centra support the caudal rays (Collette and Chao 1975; Potthoff 1975; Collette and Russo 1978), except in Scomber and RastreUiger where three centra support caudal rays (Potthoff pers. obs.). Xiphias lacks a second uroneural in the caudal complex which is present in the basic perciform caudal such as in Archosargus (Houde and Potthoff 1976), Elagatis (Berry 1969), Scom- brolabrax (Potthoff et al. 1980), Morone (Fritzsche and Johnson 1980), and Coryphaena (Potthoff 1980), but is absent in the scombrids (Potthoff 1975). The single uroneural of Xiphias does not fuse to the urostyle in adults as in Thunnini and Sardini (Collette and Chao 1975; Potthoff 1975; Collette and Russo 1978), but in several specimens anomalous shapes of the uroneural were observed (Fig. 22). We believe that Xiph ias has lost preural cen- trum 3, because a centrum having an autogenous haemal spine and a neural spine with articular cartilage is lacking (Figs. 20-23). However, 1 specimen out of 164 examined with the unusual vertebral count of 16+11=27 (typical counts 15+ 11 or 16+10=26) had two autogenous haemal spines on preural centra 2 and 3. To our knowl- edge, a perciform caudal with only one autogen- ous haemal spine as in Xiphias has not been reported previously. We cannot totally rely on Monod (1968) or any other osteological descrip- tive work dealing only with adult fish because Potthoff (1975) showed that some autogenous hypural parts fuse during development and can- not be recognized in adults. There is considerable fusion of caudal complex 183 FISHERY BULLETIN: VOL. 80, NO. 2 bones in Xiphias. Hypurals 1-4 and the urostyle fuse to one posteriorly notched hypural plate during development (Fig. 23); the three epurals, the uroneural pair, hypural 5, and the par- hypural remain autogenous, whereas in Thun- nini and Sardini only one epural remains autogenous and the paired uroneural fuses to the urostyle (Collette and Chao 1975; Potthoff 1975; Collette and Russo 1978). In Xiphias, hypurals 1- 4 develop initially from distinctly separate pieces of cartilage and fusion of the hypurals into the notched hypural plate occurs. In Scombridae a similar yet different development takes place, because in Thunnini hypurals 1 and 2 originate from one distinctly larger piece of cartilage, whereas in Scomber (Pneumatophorus), hypurals 1 and 2 originate from separate pieces of cartilage as in Xiphias (Kramer 1960). The caudal rays in adult scombrids, except Scombrini, cover the whole hypural plate (Collette and Chao 1975; Collette and Russo 1978), whereas in Xiphias a smaller area is covered by the rays (Figs. 17, 22, 23). When the rays are disarticulated from the hypural plate in adult Xiphias, long vertical depressions caused by the rays can be observed on the hypural plate (Fig. 23). Xiphias has a greater number of precaudal than caudal vertebrae (Fig. 6) (Leim and Scott 1966; Ovchinnikov 1970). The same tencency was observed in the gempylids (Matsubara and Iwai 1958; Potthoff et al. 1980) and the opposite tendency in the scombrids (Conrad 1938; de Sylva 1955; Mago Leccia 1958; Kramer 1960; Gibbs and Collette 1967; Matsui 1967; Potthoff and Richards 1970; Collette and Chao 1975). Generally, the tendency in the perciform fishes is to have a higher caudal vertebral count; the most typical count being 10+14=24 vertebrae (John- son 1981). The neural and haemal arches in Xiphias first develop distally opened (split) (Fig. 27). Dur- ing development the neural and haemal arches fuse forming spines. Fusion of the neural and haemal spines proceeds from posterior in an ante- rior direction (Table 13). In other perciforms studied by Potthoff, split arches were sometimes observed on small larvae on the anteriormost first and second centra only, but these two arches fused to spines during development. Adult Xiphias retain three to six anteriormost split neural arches (Bruce B. Collette footnote 3). Rib development and position is unique in Xiphias. Commonly, perciforms have pairs of dorsal (epipleural) ribs on the precaudal verte- brae starting on the first centrum and pleural ribs starting on the third centrum (Houde and Potthoff 1976; Potthoff et al. 1980). These ribs develop from anterior in a posterior direction. Xiphias, however, has lost many of its ribs. Generally, there are only one pair of ribs on each of the first four centra, which develop from anterior in a posterior direction and one pair on the last two precaudal vertebrae which develop from posterior in an anterior direction. We do not know if the ribs in Xiphias were originally epipleural, pleural, or a combination of epi- pleural and pleural. We were able to determine, however, the cartilage origin of ribs in Xiphias. Tibbo et al. (1961) stated that ribs in adult Xiphias are short and poorly developed, but no details on rib position were given. An account of rib development in lower and higher fishes is given by Emelianov (1935). He found that some bony fish develop ribs from cartilage, in others rib development from cartilage is bypassed and ribs develop directly from bone cells, and still in others, parts of the ribs develop from cartilage and other parts of the same rib develop directly from bone. In Xiphias the proximal portions of each rib originate from cartilage, the distal portions develop directly as bone. The branchiostegal ray count in Xiphias may vary by one ray from specimen to specimen or it may vary between left and right sides in a speci- men. Usually, branchiostegal ray counts are conservative and characterize fish families and sometimes genera (Kishinouye 1923; McAllister 1968; Fraser 1972; Ahlstrom etal. 1976; Kendall 1979; Matsuura 1979), however variability has been reported in some groups such as Carangidae (McAllister 1968). We cannot make firm conclusions about the phylogenetic status of Xiphias. From our study we conclude that Xiph ias is a perciform fish that differs from other perciforms to warrant the separate suborder Xiphiioidei. We were unable to determine relationship with the scombroids (gempylids, scombrids). A comparison with istiophorids remains to be done, and we believe we furnished sufficient material to facilitate such a comparison. ACKNOWLEDGMENTS We thank Bruce B. Collette, Edward D. Houde, G. David Johnson, Izumi Nakamura, 184 P0TTH0FF and KELLFY: OSTEOLOGICAL DEVELOPMENT IN SWORDFISH William J. Richards, and Joseph L. Russo for their critical comments on the manuscript. We thank Joaquin Javech for the fine illustrations. Our thanks also go to the persons who supplied us with specimens: Steven Berkeley, Edward D. Houde, G. David Johnson, Mark M. Leiby, William J. Richards, and Donald P. de Sylva. We thank Steven Loher for taking osteological ob- servations and Kelly Clark for clearing and staining. We thank Phyllis Fisher for typing the manuscript. LITERATURE CITED Ahlstrom, E. H., J. L. Butler, and B. Y. Sumida. 1976. Pelagic stromateoid fishes (Pisces, Perciformes) of the eastern Pacific: kinds, distributions, and early life histories and observations on five of these from the northwest Atlantic. Bull. Mar. Sci. 26:285-402. Arata, G. F., Jr. 1954. A contribution to the life history of the swordfish, Xiphias gladius Linnaeus, from the South Atlantic coast of the United States and the Gulf of Mexico. Bull. Mar. Sci. Gulf Caribb. 4:183-243. Berry, F. H. 1969. Elagatis bipinnulata (Pisces: Carangidae): Mor- phology of the fins and other characters. Copeia 1969: 454-463. COLLETTE, B. B., AND L. N. CHAO. 1975. Systematics and morphology of the bonitos {Sa rda) and their relatives (Scombridae, Sardini). Fish. Bull., U.S. 73:516-625. COLLETTE, B. B., AND J. L. RUSSO. 1978. An introduction to the Spanish mackerels, genus Scorn beromorus. In Proceedings of the Mackerel Col- loquium, p. 3-16. Gulf States Mar. Fish. Comm., Brownsville, Tex. Conrad, G. M. 1938. The osteology and relationships of the wahoo (Acanthocybium solandri), a scombroid fish. Am. Mus. Novit. 1000, 32 p. de Sylva, D. P. 1955. The osteology and phylogenetic relationships of the blackfin tuna, Thunnus atlanticus (Lesson). Bull. Mar. Sci. Gulf Caribb. 5:1-41. DlNGERKUS, G., AND L. D. UHLER. 1977. Enzyme clearing of alcian blue stained whole small vertebrates for demonstraton of cartilage. Stain Technol. 52:229-232. Ehrenbaum, E. 1905. Fam. Xiphiidae, Xiphias gladius L. In E. Ehrenbaum, Nordisches Plankton, Zoologischer Teil, Vol. 1, p. 35. Von Lipsius & Tischer, Kiel and Leipzig. New ed.— A. Asher & Co., Amsterdam, 1964. Emelianov, S. W. 1935. Die Morphologie der Fischrippen. Zool. Jahrb., Abt. Anat. Ontog. Tiere 60:133-262. Fierstine, H. L. 1974. The paleontology of billfish— the state of the art. In R. S. Shomuraand F.Williams (editors). Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review and con- tributed papers, p. 34-44. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. Fraser, T. H. 1972. Comparative osteology of the shallow water cardi- nal fishes (Perciformes: Apogonidae) with reference to the systematics and evolution of the family. Rhodes Univ., J. L. B. Smith Inst. Ichthyol. Spec. Publ. 34, 105 p. Fritzsche, R. A., and G. D. Johnson. 1980. Early osteological development of white perch and striped bass with emphasis on identification of their larvae. Trans. Am. Fish. Soc. 109:387-406. Gibbs. R. H., Jr., and B. B. Collette. 1967. Comparative anatomy and systematics of the tunas, genus Thunnus. U.S. Fish Wildl. Serv., Fish. Bull. 66:65-130. Gosline, W. A. 1961a. The perciform caudal skeleton. Copeia 1961: 265-270. 1961b. Some osteological features of modern lower teleostean fishes. Smithson. Misc. Collect. 142(3), 42 p. 1968. The suborders of perciform fishes. Proc. U.S. Natl. Mus. 124(3647), 78 p. Greenwood, P. H., D. E. Rosen, S. H. Weitzman, and G. S. Myers. 1966. Phyletic studies of teleostean fishes, with a provi- sional classification of living forms. Bull. Am. Mus. Nat. Hist. 131:341-355. Gregory, W. K., and G. M. Conrad. 1937. The comparative osteology of the swordfish (Xiphias) and the sailfish (Istiophorus). Am. Mus. Novit. 952, 25 p. Houde, E. D., and T. Potthoff. 1976. Egg and larval development of the sea bream Archosargus rhomboidalis (Linnaeus): Pisces, Sparidae. Bull. Mar. Sci. 26:506-529. Johnson, G. D. 1975. The procurrent spur: an undescribed perciform caudal character and its phylogenetic implications. Occas. Pap. Calif. Acad. Sci. 121, 23 p. 1981. The limits and relationships of the Lutjanidae and associated families. Bull. Scripps Inst. Oceanogr., Univ. Calif. Vol. 24, 114 p. Jones, S. 1958. Notes on eggs, larvae, and juveniles of fishes from Indian waters. I. Xiphias gladius Linnaeus. Indian J. Fish. 5:357-361. 1962. Notes on eggs, larvae and juveniles of fishes from Indian waters. XIV. Further notes on Xiphias gladius Linnaeus. Indian J. Fish. 9(Sect. A):525-529. Kendall, A. W., Jr. 1979. Morphological comparisons of North American sea bass larvae (Pisces: Serranidae). U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 428, 50 p. Kishinouye, K. 1923. Contributions to the comparative study of the so- called scombroid fishes. J. Coll. Agric, Imp. Univ. Tokyo 8:293-475. Kramer, D. 1960. Development of eggs and larvae of Pacific macker- el and distribution and abundance of larvae 1952- 56. U.S. Fish Wildl. Serv., Fish. Bull. 60:393-438. Leim, A. H., and W. B. Scott. 1966. Fishes of the Atlantic Coast of Canada. Fish. Res. Board Can., Bull. 155, 485 p. Lundberg, J. G., and J. N. Baskin. 1969. The caudal skeleton of the catfishes, order Silu- riformes. Am. Mus. Novit. 2398. 49 p. 185 FISHERY BULLETIN: VOL. 80. NO. 2 Mago Leccia, F. 1958. The comparative osteology of the scombroid fishes of the genus Scomberomorus from Florida. Bull. Mar. Sci. Gulf Caribb. 8:299-341. Markle, G. E. 1974. Distribution of larval swordfish in the Northwest Atlantic Ocean. In R. S. Shomura and F. Williams (editors), Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review and contributed papers, p. 252- 260. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. Matsubara, K., and T. Iwai. 1958. Anatomy and relationships of the Japanese fishes of the family Gempylidae. Mem. Coll. Agric. Kyoto Univ., Fish. Ser. Spec. No., p. 23-54. Matsui, T. 1967. Review of the mackerel genera Scomber and Rastrelliger with description of a new species of Rastrelliger. Copeia 1967:71-83. Matsuura, K. 1979. Phylogeny of the superfamily Balistoidea (Pisces: Tetraodontiformes). Mem. Fac. Fish., Hokkaido Univ. 26:49-169. McAllister, D. E. 1968. The evolution of branchiostegals and associated opercular, gular, and hyoid bones and the classification of teleostome fishes, living and fossil. Natl. Mus. Can., Bull. 221, 239 p. Monod, T. 1968. Le complexe urophore des poissons teleosteens. Mem. Inst. Fond. Afr. Noire 81, 705 p. Nakamura, I., T. Iwai, and K. Matsubara. 1968. A review of the sailfish, spearfish, marlin, and swordfish of the world. [InJpn.] Kyoto Univ., Misaki Mar. Biol. Inst, Spec. Rep. 4, 95 p. (Engl, transl. by National Marine Fisheries Service.) Nakamura, H., T. Kamimura, Y. Yabuta, A. Suda, S. Ueyanagi, S. Kikawa, M. Honma, M. Yukinawa, and S. MORIKAWA. 1951. Notes of the life-history of the sword-fish, Xiphias gladius Linnaeus. Jpn. J. Ichthyol. 1:264-271. Nursall, J. R. 1963. The hypurapophysis, an important element of the caudal skeleton. Copeia 1963:458-459. Nybelin, 0. 1963. Zur Morphologie und Terminologie des Schwanz- skelettes der Actinopterygier. Ark. Zool., Ser. 2, 15:485-516. OVCHINNIKOV, V. V. 1970. Swordfishes and billfishes in the Atlantic Ocean; ecology and functional morphology, fin Russ.] Atl. Sci. Res. Inst. Fish. Oceanogr. (Engl, transl. by Isr. Program Sci. Transl., 77 p.; available U.S. Dep. Commer., Natl. Tech. Inf. Serv., Springfield, Va., as TT71-50011.) Palko, B. J., G. L. Beardsley, and W. J. Richards. 1981. Synopsis of the biology of the swordfish, Xiphias gladius Linnaeus. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 441, 21 p. Potthoff, T. 1974. Osteological development and variation in young tunas, genus Thunnus (Pisces, Scombridae), from the Atlantic Ocean. Fish. Bull., U.S. 72:563-588. 1975. Development and structure of the caudal complex, the vertebral column, and the pterygiophores in the blackfin tuna (Thunnus atlanticus, Pisces, Scombridae). Bull. Mar. Sci. 25:205-231. 1980. Development and structure of fins and fin supports in dolphin fishes Coryphaena hi ppurus and Coryphaena eguiselis (Coryphaenidae). Fish. Bull., U.S. 78:277- 312. Potthoff, T., and W. J. Richards. 1970. Juvenile bluefin tuna, Thunnus th ynnus (Linnaeus), and other scombrids taken by terns in the Dry Tortugas, Florida. Bull. Mar. Sci. 20:389-413. Potthoff, T., W. J. Richards, and S. Ueyanagi. 1980. Development of Scombrolabrax heterolepis (Pisces, Scombrolabracidae) and comments on familial relation- ships. Bull. Mar. Sci. 30:329-357. Regan, C. T. 1924. A young swordfish (Xiph ias gladius), with a note on Clupeolabrus. Ann. Mag. Nat. Hist., Ser. 9, 13:224-225. Richards, W. J. 1974. Evaluation of identification methods for young billfishes. In R. S. Shomura and F. Williams (editors), Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972, Part 2. Review and contributed papers, p. 62-72. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675. Roberts, T. R. 1981. Sundasalangidae, a new family of minute freshwater salmoniform fishes from Southeast Asia. Proc. Calif. Acad. Sci., Ser. 4, 42:295-302. Sanzo, L. 1910. Uovo e larva di pesce-soada (Xiphias gladius L.). Riv. Mens. Pesca Idrobiol. 12:206-209. 1922. Uovae larve di Xiph ias gladiush. Mem. R. Com. Talassogr. Ital., ano 79, 17 p. Starks, E. C. 1930. The primary shoulder girdle of the bony fishes. Stanford Univ. Publ., Biol. Sci. 6:147-240. Swinnerton, H. H. 1905. A contribution to the morphology and development of the pectoral skeleton of teleosteans. Q. J. Microsc. Sci., New Ser., 49:363-382. TAning, A. V. 1955. On the breeding areas of the swordfish (Xiphias). Deep-Sea Res., Pap. Mar. Biol. Oceanogr., suppl. to vol. 3:438-450. Taylor, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. U.S. Natl. Mus. 122(3596), 17 p. Tibbo, S. N., L. R. Day, and W. F. Doucet. 1961. The swordfish (Xiphias gladius L.), its life-history and economic importance in the northwest Atlantic. Fish. Res. Board Can., Bull. 130. 47 p. Yabe, H. 1951 . Larva of the swordfish, Xiph ias gladius. [In Jpn., Engl, summ.] Jpn. J. Ichthyol. 1:260-263. Yabe, H., S. Ueyanagi, S. Kikawa, and H. Watanabe. 1959. Study on the life-history of the sword-fish, Xiphias gladius Linnaeus. [In Jpn., Engl, summ.] Rep. Nankai Reg. Fish. Res. Lab. 10:107-150. (Transl. by Musam Fujiya available.) Yasuda, F., H. Kohno, A. Yatsu, H. Ida, P. Arena, F. Li Greci, and Y. Taki. 1978. Embryonic and early larval stages of the sword- fish, Xiphias gladius, from the Mediterranean. J. Tokyo Univ. Fish. 65:91-97. 186 AGE AND GROWTH OF LARVAL ATLANTIC HERRING, CLUPEA HARENGUS L., IN THE GULF OF MAINE-GEORGES BANK REGION BASED ON OTOLITH GROWTH INCREMENTS R. Gregory Lough, Michael Pennington, George R. Bolz, and Andrew A. Rosenberg 1 ABSTRACT An estimate of the age and growth of herring larvae over their first 6 months of life is made by examining presumed daily growth increments in their otoliths. A Gompertz growth curve fitted to 311 autumn-spawned specimens collected in the Gulf of Maine-Georges Bank region describes the mean length at age (based on a range of 7-160 otolith increments (from an initial hatching size of 5. 7 mm SL to a mean length of 30.9 mm at 175 days. A larva with 7 growth increments isestimated to be on average 25 days old with a mean length of 12.7 mm. Larvae reared in the laboratory at 10°C began initial increment deposition on average 4.5 days from hatching at the time of yolk-sac absorption, and the second increment was deposited an average of 12 days from hatching. The rearing experi- ments were terminated before an increment-day relation could be established, but the third incre- ment was estimated to be formed on average 22 days from hatching. Support for the assumption that increment deposition becomes daily at least after the third increment is made by two independent methods. Based on the fitted Gompertz curve, average growth rates for herring larvae increased from 0.25 mm/day at hatch to 0.30 mm/day at 20 days and declined to <0.15 mm/day after 75 days of age during the winter period. This agrees closely with estimated field rates. Atlantic herring, Clupea harengus L., spawn de- mersal eggs during late summer-autumn on the shoaler «40 m bottom depth) regions of Georges Bank and around the perimeter of the Gulf of Maine ( Bigelow and Schroeder 1953; Boyer et al. 1973; Lough and Bolz 1979 2 ). Hatching occurs after 8-9 d at 10°C (Cooper et al. 3 ), and shortly thereafter the larvae are dispersed throughout the water column by the vigorous tidal stirring characteristic of this region (Bumpus 1976). By following length-frequency means or modes be- tween successive surveys, average larval growth rates have been estimated on field populations of larvae in the Gulf of Maine-Georges Bank region by Tibbo et al. (1958), Tibbo and Legare (1960), Das (1968, 1972), Graham et al. (1972), Sameoto (1972), Boyar et al. (1973), Lough et al. (1979), 4 'Northeast Fisheries Center Woods Hole Laboratory, Na- tional Marine Fisheries Service, NOAA, Woods Hole, MA 02543. -'Lough, R. G., and G. R. Bolz. 1979. A description of the sampling methods, and larval herring (Clupea harengus L.) data for surveys conducted from 1968-1978 in the Georges Bank and Gulf of Maine areas. Northeast Fisheries Center, Natl. Mar. Fish. Serv., NOAA, Woods Hole Lab. Ref. 79-06, 230 p. Cooper, R. A., J. R. Uzmann, R. A. Clifford, and K. J. Pecci. 1975. Direct observations of herring ( Clupea harengus haren- gus L.) egg beds on Jeffreys Ledge, Gulf of Maine in 1974. ICNAF Res. Doc. 75/93, 6 p. 4 Lough, R. G., G. R. Bolz, M. D. Grosslein, and D. C. Potter. 1979. Abundance and survival of sea herring (Clu pea haren- and others. Larval herring grow at an overall average rate of about 5 mm/mo (0.2 mm/d) from hatch (6 mm SL) to metamorphosis in the spring. Metamorphosis is a gradual transition to adult characteristics generally achieved by the time the fish are 50-55 mm, but some studies report metamorphosis occurring at much smaller lengths of 30-35 mm (Blaxter and Staines 1971; Boyar et al. 1973; Ehrlich et al. 1976; Doyle 1977). Knowledge of larval herring growth is an im- portant component in the estimation of age- specific mortality rates, which can be used to study variations in larval survival in relation to size of succeeding year classes. However, field estimates of larval growth only provide average rates of growth so that their use in comparative studies is limited by the sometimes polymodal length frequencies and subjective nature of con- necting corresponding length modes. With the development of accurate growth models, popula- tions can be compared by region and season with various environmental factors which may be affecting growth and hence survival of larvae. Techniques are now available for the accurate aging of larval and juvenile fishes based on Manuscript accepted November 1981. FISHERY BULLETIN: VOL. 80, NO. 2. 1982. gus L.) larvae in relation to environmental factors, spawning stock size, and recruitment for the Georges Bank area, 1968- 1977 seasons. ICNAF Res. Doc. 79/VI/112, 47 p. 187 FISHERY BULLETIN: VOL. 80, NO. 2 growth increments or lamellae in their otoliths, thus providing a detailed chronological record of events in the growth history of an individual fish (Pannella 1971, 1974; Scott 1973; Brothers et al. 1976; Struhsaker and Uchiyama 1976; Ralston 1976; Taubert and Coble 1977; Barkman 1978; Methot and Kramer 1979; Radtke 1980 5 ; Radtke and Waiwood 1980; Steffensen 1980; Wilson and Larkin 1980; Uchiyama and Struhsaker 1981; Barkman et al. 1981; Brothers 1981; Brothers and McFarland 1981; Methot 1981). Evidence for the presence of apparent daily otolith growth increments in larval herring collected along western Gulf of Maine in October 1976 was given in a preliminary report by Rosenberg and Lough. 6 Further work by Lough et al. 7 led to the development of a growth model extending through the autumn-winter period. Townsend and Graham (1981) recently used otolith aging techniques to examine the age structure of larval herring entering the Sheepscot River, Maine, estuary during autumn-winter 1978-79. The objective of this study is to summarize our findings on the age and growth of larval herring otoliths during the first 6 mo of life, from hatch- ing to a length of ca. 31 mm, based on larvae reared in the laboratory and collected in the Gulf of Maine-Georges Bank region. Also, aGompertz growth curve is fitted to the length-at-age data based on "daily" growth increment in their oto- liths to describe the shape of the average larval herring growth curve in this region from Octo- ber through March 1976-77. The present study was initiated by the International Commission for the Northwest Atlantic Fisheries (ICNAF) (Lough et al. 1981) and the U.S. participation was conducted concurrently as part of the MAR- MAP (Marine Resources Monitoring, Assess- ment, and Prediction) program of the Northeast Fisheries Center, which measures long-term changes in the variability of fish stock abun- dance off the northeast coast of the United States (Sherman 1980). METHODS Larval herring for otolith studies were col- lected at selected stations within a standard grid of sampling stations covering the western Gulf of Maine, Georges Bank, and Nantucket Shoals areas on five ICNAF larval herring surveys con- ducted from October 1976 through March 1977 (Table 1, Fig. 1). Larvae normally were collected at stations where high densities were encoun- tered. Standard ICNAF double-oblique contin- uous hauls (61 cm bongo net, 0.505 and 0.333 mm mesh nets) were made at each station to a maxi- mum depth of 100 m, or to within 5 m of the bot- tom in shoaler areas, while the vessel was under- way at 3.5 kn. A standard haul ranges in duration from 5 to 25 min; each bongo net filtering be- tween 100 and 1,000 m 3 of water depending on the duration (maximum depth) of the haul. Fur- ther details of the sampling gear and protocols can be found in Lough and Bolz (footnote 2). Immediately after the nets were brought aboard the vessel, larvae were sorted from the untreated 0.505 mm mesh plankton sample and frozen in dishes. Extra hauls occasionally were made to collect sufficient numbers of larvae. Tempera- 1 1 I I fi t/6 J55 ) t //? \L~ r A -! ! ~s\y * ' —" /&'/ )^^X(P ,5*' / VI — 44° * a' efts If l/f I * V 'J * f'* \}\ 1 L %f *--» r < rJ •' "44 \ \ > ^S # 38 s V / ; J . 105 ~> / / * 122 ( — 7 '• )( GULF OF MAINE "-< '^— - / s"\ZS'P I CST'< 102 , — ^ i t \ ~ r ~\ ( \ *v — 42° V "^^2 • no ! jSfcn L-/~' — 7 ~~ N 72 ) '** \ * /' 35 GEORGES f f — NANTUCKET BANK i SHOALS / •20 • ,„ '*' a 33 100 M^ _- • 19 /' / ^--. --- " / ~— -~ s ~40' 1 70° I 68" 1 66- I 5 Radtke, R. L. 1980. The formation and growth of otoliths from redfish (Sebastes spp.) larvae from the Flemish Cap (Di- vision 3M). NAFO SCR Doc. 80/1 X/ 153, 6 p. 6 Rosenberg, A. S., and R. G. Lough. 1977. A preliminary report on the age and growth of larval herring (Clupea haren- giis) from daily growth increments in otoliths. ICES CM. 1977/L:22. 15 p. Figure 1.— Station locations in the Gulf of Maine-Georges Bank region where larval herring were collected for otolith aging over the 1976 spawning season. 7 Lough, R. G., M. R. Pennington, G. R. Bolz, and A. S. Rosen- berg. 1980. A growth model for larval sea herring (Clupea harenyus L.) in the Georges Bank-Gulf of Maine area based on otolith growth increments. ICES CM. 1980/H:65, 22 p. 188 LOlHill KT AI..: ACK AND GROWTH OK LARVAL ATLANTIC HERRING 5 o 3 ca _c o> 3 .2" o o ~ 5 CD TO c aa P CO be CD c be O s_ - * ° i ■s ° c O * a- o c- -a cd 8 >> t« "^ C efl CD a> E *. .S a) o *j cu c &•- en £ be j S be l- 3 t. O cd j- J= -C ca > t- ca i—* Z s c 3 O 3 ca ca .o c o ca (72 W J CO < o c » ra - E CD O CD CO ~ ~ T3 Ol E <= C F m cd cz O - S Z™ 2 EO o c5° C\J E^ g^E o m CD i m 55 „.2>S Is' PI Q O 2 O to s Eg o p O CO CO to ^ oS — . o — • a> a) co *— — o —CD co o CO O CO CM i- CO N CD tJ; cDtbcsjco cnj r^ ^ cm ■. M I - I V H -* ^ V* I — -» » -*• t-i-C\JCM I- CM ^t CM CNJ tJ O CO *~ CN CNJ r- r- co cm r^. CM ^ CM CO i- •^ CO CD CO O CM O *- i- c\i cm - in ^ co t- o w o oir^p'- cm r^ -(J- N r- CO CO r- CJ) CO CO CO o i- i- t-CMi-^-CM CM CM CM CM CO lO O) O) S CO CO CO CO t- O) p O h-- CM CO CO C\i o cp ^r in tt CO CO s s cb cd o m p O O CO CO ZQQZ Z ZZZZZ ZZQO Q o m m o o cm to o o in t-cooco o Ot-i-co m ocsjcoom ocmcmco co CO^incO CD O ID O »- O) SOCOCD CM O'-'-O O OOt-i-O O *- *~ *- cm 1^ CD OOOU O CDCDCUCDCD 0) D ■3 n c OO o C CJ 0) CD CD a) O zo? O Z S w CNJ CJ in r^ o n r\j n o \r> O O u-> o CNJ in m Cl PI in t CJ co T— CM CO O i— co o o o O o CD CO i^ r- s CD CO n CD co < 1 c ) O C5 CD UD UD UD UJ CO UJ UJ r-. CD CJJ IV h- f- r^ CD COCO^t 1 ^- i- »-»-C\ICO'- oococo a ■c o 5 5 CO ■D c CO Q) to 3 «~ 2 en 5 > CO co CD CO 0) ^2 C « CD CD CO g- a E E 2 H CD .*: o 3 ™ E >■ ZO ture data at each station were obtained from ex- pendable bathythermograph traces or surface bucket readings. Larvae were reared from fertilized eggs in the laboratory in order to determine the age at which increment deposition first begins in larval her- ring otoliths. A batch of herring eggs, stripped from several ripe and running adults collected along the western Gulf of Maine near Jeffreys Ledge, was fertilized on 17 October 1978 and reared at the NMFS Narragansett Laboratory at 10°C by G. Laurence for use in various feeding experiments. Larvae were maintained in special rearing aquaria described by Beyer and Laur- ence (1981) with a photoperiod of 12 h light and 12 h dark and fed wild plankton at high den- sities (>3 plankters/ml). Approximately 15 lar- vae were removed from the rearing aquaria daily from hatching on 28 October through 15 November and preserved in 75% ethyl alcohol. Prior to removing the otoliths, larvae were staged according to Doyle (1977) and measured for standard length (snout to caudal peduncle) and head length (snout to sagitta in normal posi- tion) to the nearest 0.1 mm. The largest otoliths (sagittae) were removed from both sides of the head when possible and mounted in Canada bal- sam or Permount. 8 The otoliths were whole mounted and little difficulty was found in read- ing them intact so that further preparation was unnecessary. The 2-sagittae and 2-astericae ob- tained per individual from the laboratory-reared larvae were virtually impossible to distinguish at this early stage; however, the number of growth increments was identical for both sets of otoliths from the same individual. The otoliths were viewed by transmitted light and growth increments were counted using a Zeiss compound microscope-video system with a magnification range of 630X for the largest oto- liths and 1000X or 2000X for the smallest. Differ- ential interference microscopy was particularly helpful in distinguishing increments of the smallest otoliths. The resolving power of our microscope is in the range of 0.2-0.5 /nm. A mini- mum of three counts was made on all otoliths or counts were repeated until a mean value was reached with a maximum acceptable range of 5% variability. Routine otolith measurements made to the nearest micron as illustrated in Figure 2 included the following: 1) anterior-posterior di- 8 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 189 FISHERY BULLETIN: VOL. 80, NO. 2 %. - ;. f < ' BHAi D Figure 2.— Sagittae of herring larvae, Clupea harengus. Bar on photographs represents 10 fim; pr - primordium, a = anterior, p = posterior, nc = nuclear check. A. Otolith from laboratory-reared larva, 8.4 mm SL, showing 2 growth increments (1000X). Additional increments are optical artifacts. B. Otolith with 23 increments showing band of thin, poorly defined 4-5 increments around nucleus; 18.6 mm SL; Annandale 76-01, Stn. 38. C. Otolith with 51 increments (630X). Note the first 3-4 thin, poorly defined increments immediately surrounding nuclear check; 19.9 mm SL; Researcher 76-01, Stn. 105. D. Otolith with 54 incre- ments (630X), posterior view. Note pattern of increment thickness from initial thin, poorly defined 7-9 increments encircling a heavy nuclear check increasing to maximum thickness at 10th-35th increments and then decreasing thickness towards the edge; 21.6 mm SL; Researcher 76-01, Stn. 35. ameter (otolith length); 2) lateral diameter: a line perpendicular to anterior-posterior axis; 3) nu- cleus diameter: whole otolith at hatching without increments or to inner edge of first increment; 4) anterior radius: nucleus center (primordium) to anterior edge; and 5) posterior radius: nucleus center (primordium) to posterior edge. Selected otoliths were photographed, enlargements made, and increment thicknesses were mea- sured across a posterior radius from the nucleus using a Zeiss MOP Digital Image Analyzer Sys- tem. All field-collected larvae used in this study for otolith aging were frozen, whereas the labora- tory-reared larvae were preserved in 75% ethyl alcohol, and larvae referred to in other corrobo- rative field studies were preserved in Formalin. Theilacker (1980) reported that the amount of shrinkage of northern anchovy larvae, Engraulis mordax, varies with fish size and duration of time larvae are retained within the net. Larvae smaller than 11 mm SL net-treated for 20 min could shrink as much as 19% of their live length prior to preservation. We estimate that nearly all herring larvae collected on ICNAF surveys have been dead at least 20 min prior to preservation. An additional 3% shrinkage due to 5% Formalin preservation was recommended by Theilacker 190 LOUGH ET AL.: ACE AND GROWTH OF LARVAL ATLANTIC HERRING for all body parts after net-treatment, whereas preservation in ethyl alcohol (80%) did not cause any additional shrinkage in standard length. Townsend and Graham (1981) indicated that frozen herring larvae (27-45 mm TL) may shrink 3-4% more than Formalin-preserved larvae. From our experience we find that length mea- surements of frozen larvae can be more variable than those of Formalin-preserved larvae; how- ever, a thorough study has not been made. No correction factor was applied to our field-col- lected frozen larvae because of the uncertainty of time prior to preservation and the effect of freezing on shrinkage. We do not feel that the Gompertz population growth curve fit to the un- corrected field-collected larvae would be signifi- cantly altered with respect to shape compared with corrected data. When a direct comparison is made in this paper between laboratory-reared and field-estimated larval lengths, a shrinkage correction factor applied to the lab data will be specified based on Theilacker's ( 1980) work which probably is adequate for all clupeidlike larvae. RESULTS Otoliths from 311 herring larvae caught in plankton hauls were processed in this study cov- ering their first 6 mo of life from October through March 1977 (Table 1, Fig. 1). Approxi- mately 58% of the larvae were collected along the western Gulf of Maine, 23% from Nantucket Shoals, and 19% from Georges Bank. ICNAF surveys have never been conducted during the month of January so that there is a gap in time in our collection of larval otolith data from mid- December 1976 to mid-February 1977. The field- collected larvae ranged in length from 11 to 35 mm with most of the western Gulf of Maine lar- vae falling into the 11-31 mm size range; the Georges Bank larvae, 19-25 mm; and the Nan- tucket Shoals larvae, 26-35 mm. The number of otolith increments counted from the field-col- lected larvae ranged from 7 to 160. Since we were not able to collect any recently hatched lar- vae <11 mm length for otoliths on these surveys, laboratory-reared larvae had to suffice for the smallest size. Laboratory-Reared Larvae Hatching of the laboratory-reared larvae occurred over a 5-d period with 50% hatch esti- mated on 28 October 1978 for a mean incubation of 11 d. Yolk-sac resorption was estimated to be 50% complete 4-5 d after hatching, and 99% com- plete 6 d after hatching. The larvae began actively feeding at yolk-sac resorption and ap- peared to be healthy without any abnormalities throughout the more than 3 wk of rearing. Mor- tality over the first 13 d from hatching averaged 12%/d which is considered low. The age of larvae from hatching midpoint with 0-3 increments is given in Table 2 and Figure 3. The first incre- ment appeared on larval otoliths that ranged in age from to 9 d from hatch with a middate of 4.5 d which indicates that the first increment is de- posited near the end of yolk-sac resorption. Lar- vae staged according to Doyle (1977) showed a progression of the three substages la-lc over the first 3 d from hatch so that after the third day only remnants of yolk sac remained. The second growth increment occurred in lar- vae 6-18+ d old with a middate of 12 d from hatch or 7.5 d from the middate of the first increment formation. The third increment was observed for the first time on a larva 16 d from hatch, but un- fortunately sampling was terminated before the Table 2.— Distribution of otolith growth increments, age in days from hatching midpoint, and mean stan- dard length of herring larvae reared in the laboratory at 10°C. Age (d) Mean standard length (mm) 1 No Increments Midpoint Range larvae 3 0-6 8.0 8 1 4.5 0-9 8.1 25 2 12 6-18 + 9 1 21 3 (22 est.) 16- 9.5 3 'Measurements made on larvae preserved in 75% ethyl alcohol Unpreserved mean length of 55 larvae at hatch was 7.66 mm, standard deviation of 58 (Beyer and Laurence 1981). UJ a o 2 DAYS FROM HATCH Figure 3. — Otolith increment deposition for herring larvae sampled from 50% hatch through 18 d of rearing in the labora- tory at 10°C. Encircled points represent yolk-sac larvae. Num- bers above points denote numbers of larvae >1. 191 FISHERY BULLETIN: VOL. 80. NO. 2 complete age distribution of 3-increment larvae could be determined. If the range of ages of 3- increment larvae is similar to the 2-increment larvae, then the estimated age of 3-increment larvae would range from 16 to 28 d with a mid- date of 22 d from hatch. Otolith Growth The growth and morphology of young herring otoliths has been described previously by Hempel (1959) for specimens ranging in length (total) from 25 to 130 mm collected in the German Bight, by Watson (1964) for Maine herring of 85- 285 mm TL, and by Messieh (1975) for Bay of Fundy herring of 32-118 mm TL. Here we de- scribe the growth and morphology of herring otoliths (sagitta) in relation to head length for Gulf of Maine-Georges Bank larvae ranging in size from 5.7 mm (hatching) to 35 mm SL (prior to metamorphosis). The shape of the larval herring otolith at hatching is essentially spherical having a slight convex distal side and a flat proximal side with three or four furrows radiating from a distinc- tive central core called a primordium (see Fig. 2). A slight protuberance is apparent on the anterior edge of the otolith from larvae starting at about 20 mm SL which develops into the adult rostrum. With further elongation along the anterior-posterior axis, the otoliths become gen- erally pear-shaped at metamorphosis (45 mm TL) and attain the typical shape of adult herring otoliths by 75 mm (Messieh 1975). The mean diameter of the nucleus, defined here as the size of the otolith at hatching prior to increment deposition, is 22.5 ^m (1.1 /xm SD) based on the laboratory-reared larvae. The otolith increases exponentially in length (anterior-posterior axis) to a mean size of 456 /um at 35 mm SL based on the composite field data. Successive dark and light layers are deposited around the nucleus as the otolith grows. A single growth increment comprised of a dark plus light band is generally presumed to represent 1 d. The otolith nucleus of the field-caught larvae is readily discernible as its margin is usually darkened to form a nuclear check (see Fig. 2). In some otoliths an additional increment was seen inside of the check. Messieh and Moore 9 also have observed 1 or 2, and some- times up to 5, faint increments inside the nuclear check of otoliths from herring larvae collected in the Gulf of St. Lawrence. However, no nuclear check was evident in otoliths of the laboratory- reared larvae and no increments were observed within the defined nucleus. Nuclear diameters of the field-caught larvae all fell within the 95% confidence limit of the mean nuclear diameter determined from the laboratory-reared larvae. Immediately surrounding the nucleus of the field-caught larvae, the first 3-9 growth incre- ments appear to be less well defined than suc- ceeding increments, i.e., lower optical density and thinner in width. Distinctive, darker than normal growth layers were noted across an oto- lith transect but they did not suggest any pattern or complex periodicity as observed by Pannella (1971), nor was there any evidence of subdaily rings as observed for some species by Taubert and Coble (1977), Brothers (1981), and Brothers and McFarland (1981). Scanning electron microscopy techniques will be necessary to re- solve the presence or absence of faint incre- ments. The thickness of successive growth increments was measured on otoliths from three field- caught larvae along a posterior radius starting from the nucleus edge (Fig. 4). Measurements were made only through the penultimate incre- ment in each case as the marginal increment was still in the process of formation and could not always be read clearly. Increment thickness ranged from about 0.6 to 2.4 ^m along the radii. All three otoliths show the same general pattern of increment thickness up to about 23 increments where the first 7-10 increments are relatively thin (0.8-1.5 jum) and increase to near maximum thickness (2.3 yum) by about 75 increments. The thickness of the first 3 increments from the lab- oratory-reared larvae was consistently 0.8-1.0 /xm, which compares closely with the initial in- crement thickness of the field-caught larvae. Otolith B tends to suggest a prolonged period of relatively thick increments before thinner ones start to be formed, whereas otolith C appears to form thin increments immediately after maxi- mum increment thickness at around 15 incre- ments. The form of the otolith growth curves can be seen more readily in Figure 5 where otolith radii are plotted against the number of incre- ments for the three larvae. These curves suggest that otolith growth is initially slow, increases 9 S. N. Messieh and D. S. Moore, Marine Fish Division, Fish- eries and Oceans, Bedford Institute of Oceanography, Dart- mouth, N.S., Canada, B2Y 4A2, pers. commun. August-Sep- tember 1981. 192 LOUCH KT AL.: A(JK AND GROWTH OK LARVAL ATLANTIC HKRRING FIGURE 4. — Change in increment thickness for three field-collected her- ring larvae. Measurements were made along a posterior radius from nucleus edge through the penultimate incre- ment. A. Total of 23 increments (see Fig. 2B); 18.6 mm SL larva; Annan- dale 76-01, Stn. 38. B. Total of 54 increments (see Fig. 2D); 21.6 mm SL larvae; Researcher 76-01, Stn. 35. C. Total of 150 increments, only initial 80 measured; 31.0 mm SL larva; Anton Dohm 77-01, Stn. 33. 2.5 o o CO CO LLl 0.5 \ 10 hi v v A -~ . n,1 / A ' V I I 'fUl 20 30 40 50 INCREMENT 60 70 80 110.0 100.0 900 800 o 70 cc / ,-'• 10 30 40 50 INCREMENT 70 80 Figure 5.— Otolith radii vs. number of increments for the same three larvae in Figure 4. rapidly, and then levels off at some point. This same general pattern of otolith microstructure was observed by Brothers and McFarland (1981) for French grunts. Various allometric relations were examined between otolith size and growth of the field- caught larvae, and a few are presented here to show the homogeneity of the measurements from the three spawning populations sampled: west- ern Gulf of Maine, Georges Bank, and Nantucket Shoals. A plot of the otolith anterior vs. posterior radii in Figure 6 shows a linear relationship. The posterior radius becomes increasingly longer than the anterior radius with increment deposi- tion. Otolith length plotted against head length 200 -5: cc 100 o Y = 5.33 + 0.77 X r = 0.94 W. GULF OF MAINE • GEORGES BANK O NANTUCKET SHOALS O 50 100 150 200 POSTERIOR RADIUS ( MICRONS ) 250 Figure 6.— Otolith anterior-posterior relation for herring lar- vae collected from the three areas: western Gulf of Maine, Georges Bank, and Nantucket Shoals, with composite regres- sion line and correlation coefficient (r). 193 FISHERY BULLETIN: VOL. 80, NO. 2 in Figure 7 also can be expressed by a simple lin- ear relationship. The long axis of the otolith shows a positive allometry with respect to head length for recently hatched larvae to a size of about 35 mm SL. Hempel (1959) reported nearly isometric growth between head and otolith length after metamorphosis for German Bight herring. 500 450 - 400 - o °y O y° Ul 350 O o ojj 1 as ? 300 - Y = -79.1 + 86.09 X orf> u s Q I 250 r = 0.90 o o UJ 200 <%$ $P B *%& oVo J ■,<*•" 1 - O 150 O $0* • W. GULF OF MAINE GEORGES BANK • O 100 50 - NANTUCKET SHOALS 1 2 3 4 5 6 HEAD LENGTH ( MM ) Figure 7. — Otolith length (anterior + posterior radii)-head length relation for herring larvae collected from the three areas: western Gulf of Maine, Georges Bank, and Nantucket Shoals, with composite regression line and correlation coeffi- cient (r). function of age where r, the number of incre- ments, represents age plus some unknown con- stant (see Pennington 1979 for details of the model fit). The fitted equation was found to be L = 12.70 exp[0.89(l - exp[-0.03(r - 7)])] for r>7, (1) where 12.70 = Li, the mean length of a 7-incre- ment larvae. Equation (1) may be rewritten as L = 30.90 exp[-1.07 exp(-0.03 r)], r>l, (2) where 30.90 = L x , the asymptotic limit of mean growth during the October-March period. As- suming: 1) for at least r>7, increments are de- posited daily and 2) a curve in the form of Equa- tion (2) approximates growth from hatch, then denoting age by x, x = r + c, r>7, from 1), where c is an unknown constant, or r = x — c, x>c + 7. Thus Larval Growth A composite plot of larval length versus num- ber of otolith increments is presented in Figure 8. A Gompertz growth curve was fitted to the field data to produce a description of the mean growth of larval herring based on the 311 speci- mens with otolith growth increments ranging from 7 to 160. The Gompertz-type curve (Laird 1969) has been used to describe growth of a wide variety of organisms that often grow exponen- tially at a rate which is decaying exponentially. Previous use of the Gompertz model to more accurately describe the growth of young fish has been made by Kramer and Zweifel (1970), Saka- gawa and Kimura (1976), Zweifel and Lasker (1976), and Methot and Kramer (1979). Using the field data as a starting point, it was assumed that increments were deposited daily at least after the 7th increment so that the equation L = L 7 exp[fc(l - exp[-«(r - 7)])], r>7, was taken to represent mean larval length as a 194 L = 30.90 exp(-1.07 exp[-0.03(x - c)]), x>c + 7, (3) ' ■■■.•"' - v—-""-"" : T^~- • - :£■■}. ■' ""C 30.90e" . „ -0.03. 1.70e ' - ■■<*■■?■:■ & -. - 1 i 1 fio 7 70 90 OTOLITH INCREMENTS 80 100 120 ESTIMATED AGE ( DAYS I Figure 8.— Composite standard length-otolith increment plot for field-collected herring larvae in the Gulf of Maine-Georges Bank region, October 1976-March 1977. A Gompertz curve is fitted to larval length at estimated age, x, over their first 6 mo of life from a mean hatch length of 5.7 mm to an asymptotic limit of mean growth of 30.9 mm. A larva with 7 otolith incre- ments is estimated to be on average 24.8 d old. See text for de- tails. LOUGH ET AL: AGE AND GROWTH OF LARVAL ATLANTIC HERRING which, if assumption 2) is reasonable, Equation (3) holds for ,r>0. Letting L denote mean length at hatch (x = 0), then solving Equation (3) for c yields, Ln(3.431 - Ln U) - 0.065 0.026 Table 3 gives an estimate of the age of larvae with 7 increments (24.8 d) derived from the mean length of recently hatched larvae collected on the Jeffreys Ledge spawning beds (Cooper et al. foot- note 3). When the mean hatching size (L ) = 5.7 mm, c — 17.8 d, and from Equation (3), length as a function of age is given by L = 30.90 exp[-1.70 exp(-0.03 x)\ x>0. (4) From Equation (4) the mean length at age along with 95% confidence limits, and growth rate (millimeters/day) are estimated from the time of hatch through 175 d in Table 4. Also, the fitted growth curve is shown in Figure 8 with the estimated larval age referenced to the lower scale. The growth curve is based on data with more than 6 increments and a mean length of 5.7 mm at hatch. Obviously, if the functional form changes between age and the age correspond- ing to 7 increments, then the predicted age of fish with 7 or more increments is biased. This growth curve is based on larvae that sur- vived to the age when caught. Therefore, the back-casted curve represents the mean length of larvae for a given age which survive and hence, may be higher than the mean length of the total population. The mean lengths at age of laboratory-reared larvae having 1 and 2 increments from Table 2 fall reasonably close to the extrapolated curve near the origin. The mean length of the labora- tory-reared larvae at hatch was reported by Beyer and Laurence (1981) to be 7.66 mm (SD = 0.58 mm). After correcting for a 20-min net- treatment and Formalin preservation shrinkage factor to compare with the field data, their re- ported mean hatching size is estimated to be 6.4 mm, which is not significantly different from the Jeffreys Ledge diver-collected, Formalin-pre- served yolk-sac lar vae of 5. 7 mm mean SL. An estimate of \/var(.r |r), the standard devia- tion of age for a fixed number of increments, was made from the field data by Pennington (1979), and its value of 2.9 d compares closely with the Table 'i.— Age of larval herring with 7 otolith growth incre- ments estimated from an initial mean hatching size of 5.7 mm (0.54 mm SD) and 95% confidence intervals of the mean. Stan- dard lengths of 100 newly hatched yolk-sac larvae (Formalin' preserved) were measured from egg bed samples collected by divers 2 on the Jeffreys Ledge study site (38 m depth), 8 October 1974. Hatch 95% confidence Estimated age of larva 95% confidence length (L ) . lntervals with 7 increments mlervals (mm) lower upper (d) lower upper 5.7 5.6 5.8 248 244 25.2 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 2 Northeast Fisheries Center's Manned Undersea Research and Tech- nology (MURT) Dive Team. Table 4.— Mean standard length at age, 95% confidence limits, and growth rate (mm/d) of larval herring from hatch through 175 d estimated from the Gompertz growth model fit. Mean 95% confide nee limits Growth rate Age(d) length (mm) lower upper (mm/d) 5.7 5.4 6.0 025 1 59 56 6.2 0.26 2 62 5.9 6.5 0.26 3 64 6.2 6.7 0.26 4 67 6.4 7.0 0.27 5 70 6.7 7.3 0.27 6 7.2 69 7.5 0.27 7 7.5 7.2 7.8 0.28 8 78 75 8.1 028 9 8 1 7.8 8.4 0.28 10 84 8 1 8.7 029 20 113 11.0 11.6 030 30 142 14.0 14.5 0.29 40 170 16.8 17.2 0.27 50 19.5 193 197 0.23 75 243 24.0 24.6 0.15 100 27.3 268 277 09 125 29.0 28.4 29.5 0.05 150 299 29.3 30.5 0.03 175 30.4 298 31.0 0.01 rough estimate of 3.1 d obtained from the labora- tory data (Lough et al. footnote 7). The first larva with 3 increments observed during the laboratory-rearing occurred on day 16 after estimated hatch. The mean age of fish with 3 increments cannot be estimated directly because sampling stopped after 18 d. But assum- ing a range of ages of 12 d (4 standard devia- tions), the mean age of a 3-increment larva would be approximately 22 d. Assuming daily incre- ment deposition for the population after the third increment, a 7-increment larva would have an average age of 26 d, which compares well with the field estimate of 25 d. Messieh and Moore (footnote 9), working with autumn-spawned herring larvae in the Gulf of St. Lawrence, recently estimated the age of larvae at the time of the nuclear check completion to be 15-17 d from hatching on average. 195 FISHERY BULLETIN: VOL. 80, NO. 2 Growth Curve Compared with Other Field Studies Direct observations of herring egg beds by div- ers were made on Jeffreys Ledge, Gulf of Maine, in 1974 by Cooper et al. (footnote 3). Spawning occurred between 29 September and 3 October 1974 at about 35-50 m depth when the bottom water temperature was 9.6°C. Larval hatching began on this site on 6 October and was com- pleted by 11 October, a 5-d period. Careful visual examination of the egg bed by the divers sug- gested that major hatching began on 7-8 Octo- ber. Newly hatched larvae collected on the egg bed have already been reported in Table 3 to have a mean Formalin-preserved length of 5.7 mm (0.5 mm SD). A special 24-h vertical series of plankton hauls was made slightly downstream of the egg bed 11-12 October (Delaware 7/74-12). The mean Formalin-preserved length of all lar- vae collected by day and night hauls was 6.7 mm (0.6 mm SD) (Lough and Cohen 10 ). Approxi- mately 4 d transpired between the middates of maximum hatching and their collection by the 24-h vertical study yielding an average growth rate of 0.25 mm/d. According to the fitted Gom- pertz growth curve (Table 4), 4-d-old larvae are estimated to have reached a mean length of 6.7 mm at a mean growth rate of 0.26 mm/d (range: 0.25-0.27 mm/d) which are essentially the same as the field estimates. Graham and Chenoweth (1973) made direct observations of larval herring over egg beds on northeastern Georges Bank during autumn 1973. Submersible observations indicated that hatching occurred between 25 September and 5 October, a 10-d period. Larvae hatched in sea- water from eggs brought on shipboard 27 Sep- tember varied in length from 5 to 7 mm with over 90% at 6 mm. On 1 October, larvae collected within the vicinity of the egg beds varied from 5 to 9 mm in length but the mean was 7.1 mm about 4 d from hatching (27 September-1 October). Growth rate of these recently hatched larvae over the 4 d was estimated to be 0.28 mm/d, which is slightly higher but still comparable with the fitted growth curve. Growth of larval herring based on the Gom- I0 Lough, R. G., and R. E. Cohen. 1982. Vertical distribu- tion of recently-hatched herring larvae and associated zoo- plankton on Jeffreys Ledge and Georges Bank, October 1974. Lab. Ref. 82-10. Northeast Fisheries Center Woods Hole Laboratory, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. pertz curve was 0.25 mm/d at hatch, increased to 0.30 mm/d at 20 d, and declined thereafter to <0.15 mm/d after 75 d. The average growth rate over 150 d from hatch was 0.20 mm/d which is similar to average seasonal estimates found in most other studies of herring larvae. By follow- ing length-frequency modes for Georges Bank- Nantucket Shoals herring larvae collected on the 1971-78 ICNAF surveys, Lough etal. (footnote 4) found an average rate of 0.195 mm/d as the best compromise to describe average growth over the 7-30 mm size classes (163 d). Boyar et al. (1973) estimated larval herring growth in the Georges Bank-Gulf of Maine region, September-June, to average 0.17 mm/d with a range of 0.14-0.25 mm/d. The form of the growth curve appears to be universal for herring larvae with a cessation in growth most noticeable during mid-larval life before increasing rapidly again at the time of metamorphosis. When Sette (1943) replotted the Clyde Sea, spring-spawned larval herring data of Marshall et al. (1937), he concluded that two logarithmic curves provided a better description of growth with a decrease in slope at a length of 19.5 mm. Graham et al. (1972) also showed a de- crease in growth after about 20 mm for autumn- spawned herring larvae along the coastal west- ern Gulf of Maine. Townsend and Graham (1981) followed two groups of larvae that entered the Sheepscot River estuary of Maine that grew about 0.2-0.3 mm/d from October to early Janu- ary and from late February to early March, but experienced similar cessation of growth from late January to early February. Das (1968, 1972) followed length modes of Bay of Fundy-Gulf of Maine area herring larvae from hatching in Sep- tember and estimated growth rates to be 0.29 mm/d in the autumn, gradually declining to <0.14 mm/d during late autumn and winter months, and then increasing geometrically to >0.36 mm/d in the spring and early summer. Messieh and Moore (footnote 9) also reported a rapid increase in growth at metamorphosis for herring larvae collected in the Gulf of St. Law- rence. DISCUSSION The available data indicate that the age and growth of herring larvae in the Gulf of Maine- Georges Bank region can be accurately esti- mated from otolith microstructure, although we have no direct evidence of the increment-day relation. A Gompertz growth curve fitted to the 196 LOUGH KT AL.: AGE AND GROWTH OF LARVAL ATLANTIC HERRING field-caught larvae, which describes the length at age from an initial mean hatching size of 5.7 mm to an upper asymptotic mean length of 30.9 mm, agrees well with average growth rate esti- mates from other studies. Our field data begin with a 7-increment larva of 12.6 mm SL, which also is nearly identical to the mean length at increment age estimated by the growth curve. From the growth model a 7-increment larva is estimated to be on average 25 d from hatch (5.7 mm) having grown at an average rate of 0.28 mm/d. This implies that increment deposition does not occur daily over these 25 d or that varia- tion in the timing of first increment deposition is high. If one assumes daily increment deposition from yolk-sac resorption (4.5 d), a 7-increment larva would be 11.5 d old, inferring the larva has grown at an average rate of 0.60 mm/d, which is rather high based on field and laboratory esti- mates. Herring larvae <15 mm have estimated growth rates typically in the range of 0.25-0.30 mm/d with an upper limit of about 0.35 mm/d. The apparent delay in increment formation observed in the laboratory-reared herring larvae after the first increment at yolk-sac resorption may be due to rearing conditions, although we have no reason to suspect they were less than optimal. Other studies have shown that the for- mation of daily growth increments can be af- fected by variations in food ration, temperature, light-dark cycle, age of fish, and stressful condi- tions in general (see references in first section of paper). Increment formation appears to be spe- cies-specific and, for clupeoid species like En- graulis mordax (Brothers et al. 1976) and Clupea harengus (this study) with relatively small eggs and short incubation period, the initial incre- ments begin at the time of yolk-sac resorption (Radtke and Waiwood 1980). A dark band or check observed around the nucleus of most of the larval otoliths collected in the field, but not ap- parent in the laboratory-reared larvae, may cor- respond to the time of yolk-sac resorption as Radtke and Waiwood (1980) found for larval cod otoliths. The nuclear check may be the result of i several thin increments grouped together. Uchi- yama and Struhsaker (1981), working with Pacific tunas, found that countable growth in- crements were formed only when the fishes were fed to satiation throughout the day. The nuclear check and the succeeding 10 or so thin incre- ments observed for the field-caught herring lar- vae may be related to the inability of a first- feeding larva to meet its maximum daily ration during the transition from its yolk supply to exogenous feeding. Initial feeding efficiency is low for herring larvae, <5% success at yolk-sac resorption, but increases to about 40% 2 wk after hatching and 70% after 5 wk (Blaxter and Staines 1971). Karris (1959) observed a rapid leveling off of growth after hatch in four species of fish and Zweifel and Lasker (1976), after fitting a two- stage Laird-Gompertz growth curve to a number of larval fish species, one from hatching to yolk- sac resorption and another to more rapid growth at the onset of feeding, suggested that this phe- nomenon was almost universal in larval growth. It is conceivable that during this period of re- duced growth, increment deposition also may be delayed or diminished until the larva learns to capture sufficient numbers of prey and begins growing rapidly again. Although larval herring appear to be very re- sistant to the range of temperatures normally encountered (Blaxter 1960), the effect of tem- perature on increment formation is not known. Water temperatures observed in herring spawn- ing areas in the Gulf of Maine-Georges Bank region are typically as high as 12°-14°C in early autumn and decline to near 0°C in winter (Table 1), approaching their lower lethal limit (Graham and Davis 1971; Chenoweth 1970). Yolk-sac utili- zation in herring larvae is directly related to water temperature (Blaxter 1956; Blaxter and Hempel 1963, 1966; Blaxter and Ehrlich 1974) and variations in water temperature at hatch can reduce or extend the time to first feedi ng and consequently, otolith increment formation. Yolk- sac resorption is completed at 4-5 d at 10°C and 6 d at 8°C. Feeding of larvae is believed to com- mence at or prior to the end of yolk-sac resorption when the maximum body weight (excluding yolk sac) is reached after about 3 d at 8°C and 2 d at 12°C. Larvae reared at 10°C would initiate feed- ing 2-3 d after hatch. There is some evidence to indicate that early larval herring growth is better at higher temperatures (Blaxter 1962), although food availability is considered the more important factor in controlling growth processes and survival of larval fish in general (May 1974). Increment formation of the green sunfish, Le- pomis cyanellus, could be stopped when growth was slowed sufficiently by simulated winter con- ditions (Taubert and Coble 1977). The slowing of growth during the winter period observed for larval herring in the Gulf of Maine-Georges Bank region also may affect their increment for- mation but further research will be required to 197 FISHERY BULLETIN: VOL. 80, NO. 2 determine the effect of environmental variables on the relationship between otolith and larval growth. ACKNOWLEDGMENTS We thank P. Hamer and R. Cohen for their help with the laboratory and data processing. G. Laurence, NMFS Narragansett Laboratory, R.I., graciously provided the laboratory-reared larvae and unpublished experimental data used in this study. This report is M ARM AP Contribu- tion MED/NEFC 81-6. LITERATURE CITED Barkman, R. C. 1978. The use of otolith growth rings to age young Atlan- tic silversides, Menidia menidia. Trans. Am. Fish. Soc. 107:790-792. Barkman, R. C, D. A. Bengston, and A. D. Beck. 1981. Daily growth of the juvenile fish (Menidia menidia) in the natural habitat compared with that of juveniles reared in the laboratory, In R. Lasker and K. Sherman (editors), The early life history of fish: recent studies. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178: 324-326. Beyer, J. E., and G. C. Laurence. 1981. Aspects of stochasticity in modelling growth and survival of clupeoid fish larvae. In R. Lasker and K. Sherman (editors), The early life history of fish: recent studies. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178: 17-23. Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Blaxter, J. H. S. 1956. Herring rearing— II. The effect of temperature and other factors on development. Mar. Res. Ser. Scott. Home Dep. 5, 19 p. 1960. The effect of extremes of temperature on herring larvae. J. Mar. Biol. Assoc. U.K. 39:605-608. 1962. Herring rearing— IV. Rearing beyond the yolk-sac stage. Mar. Res. Dep. Agric. Fish. Scotl. 1, 18 p. Blaxter, J. H. S., and K. F. Ehrlich. 1974. Changes in behavior during starvation of herring and plaice larvae. In J. H. S. Blaxter (editor), The early life of fish, p. 575-588. Springer- Verlag. Berl. Blaxter, J. H. S., and G. Hempel. 1963. The influence of egg size on herring larvae (Clupea harengus L.). J. Cons. Cons. Int. Explor. Mer 28:211- 240. 1966. Utilization of yolk by herring larvae. J. Mar. Biol. Assoc. U.K. 46:219-234. Blaxter, J. H. S., and M. E. Staines. 1971. Food searching potential in marine fish larvae. In D. J. Crisp (editor), Fourth European Marine Biology Symposium, p. 467-485. Cambridge Univ. Press, Cam- bridge. Boyar, H. C, R. R. Marak, F. E. Perkins, and R. A. Clifford. 1973. Seasonal distribution and growth of larval herring (Clupea harengus L.) in the Georges Bank-Gulf of Maine area from 1962 to 1970. J. Cons. Cons. Int. Explor. Mer 35:36-51. Brothers. E. B. 1981. What can otolith microstructure tell us about daily and subdaily events in the early life history of fish? In R. Lasker and K. Sherman (editors), The early life his- tory of fish: recent studies. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:393-394. Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Brothers, E. B., and W. N. McFarland. 1981. Correlations between otolith microstructure, growth, and life history transitions in newly re- cruited French grunts [Haemulon flavolineatum (Desmarest), Haemulidae]. In R. Lasker and K. Sherman (editors), The early life history of fish: recent studies. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:369-374. Bumpus, D. F. 1976. Review of the physical oceanography of Georges Bank. Int. Comm. Northwest Atl. Fish. Res. Bull. 12: 119-134. Chenoweth, S. B. 1970. Seasonal variations in condition of larval herring in Boothbay area of the Maine Coast. J. Fish. Res. Board Can. 27:1875-1879. Das, N. 1968. Spawning, distribution, survival, and growth of larval herring (Clupea harengus L.) in relation to hydro- graphic conditions in the Bay of Fundy. Fish. Res. Board Can., Tech. Rep. 88, 129 p. 1972. Growth of larval herring (Clupea harengus) in the Bay of Fundy and Gulf of Maine area. J. Fish. Res. Board Can. 29:573-575. Doyle, M. J. 1977. A morphological staging system for the larval de- velopment of the herring, Clupea harengus L. J. Mar. Biol. Assoc. U.K. 57:859-867. Ehrlich, K. F., J. H. S. Blaxter, and R. Pemberton. 1976. Morphological and histological changes during the growth and starvation of herring and plaice larvae. Mar. Biol. (Berl.) 35:105-118. Farris, D. A. 1959. A change in the early growth rates of four larval marine fishes. Limnol. Oceanogr. 4:29-36. Graham, J. J., and S. B. Chenoweth. 1973. Distribution and abundance of larval herring, Clupea harengus harengus Linnaeus, over egg beds on Georges Bank. Int. Comm. Northwest Atl. Fish. Res. Bull. 10:141-149. Graham, J. J.. S. B. Chenoweth, and C. W. Davis. 1972. Abundance, distribution, movements, and lengths of larval herring along the western coast of the Gulf of Maine. Fish. Bull., U.S. 70:307-321. Graham, J. J., and C. W. Davis. 1971. Estimates of mortality and year-class strength of larval herring in western Maine, 1964-67. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 160:147-152. Hempel, G. 1959. Zum Wachstum der Otolithen bei Jungheringen. Helgol. Wiss. Meeresunters. 6:241-259. Kramer, D., and J. R. Zweifel. 1970. Growth of anchovy larvae (Engraulis mordax Gir- ard) in the laboratory as influenced by temperature. Calif. Coop. Oceanic Fish. Invest. Rep. 14:84-87. 198 LOUGH KT AL.: AGE AND GROWTH OF LARVAL ATLANTIC HERRING Laird, A. K. 1969. The dynamics of growth. Res./Dev. 20(8):28-31. Lough, R. G., (!. R. Bolz. M. D. Grosslein, and D. C. Potter. 1981. Abundance and survival of sea herring (Clupea harengus L.) larvae in relation to environmental factors, spawning stock size, and recruitment for the Georges Bank area. 1968-77. In R. Lasker and K. Sherman (editors). The early life history of fish: recent studies. Kapp. P.-V. Reun. Cons. Int. Explor. Mer 178:220-222. Marshall, S. M., A. G. Nicholls, and A. P. Orr. 1937. On the growth and feeding of the larval and post- larval stages of the Clyde herring. J. Mar. Biol. Assoc. U.K. 22:245-268. May, R. C. 1974. Larval mortality in marine fishes and the critical period concept. In J. H. S. Blaxter (editor), The early life history of fish, p. 3-19. Springer- Verlag, Berl. Messieh. S. N. 1975. Growth of the otoliths of young herring in the Bay of Fundy. Trans. Am. Fish. Soc. 104:770-772. Methot, R. D. 1981. Spatial covariation of daily growth rates of lar- val northern anchovy, Engraulis mordax, and northern lampfish, Stenobrach ius leucopsarus. In R. Lasker and K. Sherman (editors), The early life history of fish: recent studies. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:424-431. Methot, R. D., Jr., and D. Kramer. 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull, U.S. 77:413-423. Pannella, G. 1971. Fish otoliths: daily growth layers and periodical patterns. Science (Wash., D.C.) 173:1124-1127. 1974. Otolith growth patterns: an aid in age determina- tion in temperate and tropical fishes. In T. B. Bagenal (editor), Proceedings of an international symposium on the ageing of fish, p. 28-39. Unwin Brothers, Surrey, Engl. Pennington. M. R. 1979. Fitting a growth curve to field data. In J. K. Ord, G. P. Patil, and C. Taillie (editors), Statistical distribu- tions in ecological work, p. 419-428. Int. Coop. Publ. House, Fairland, Md. Radtke, R. L., and K. G. Waiwood. 1980. Otolith formation and body shrinkage due to fixa- tion in larval cod (Gadus morhua). Can. Tech. Rep. Fish. Aquat. Sci. 929, 10 p. Ralston, S. 1976. Age determination of a tropical reef butterflyfish utilizing daily growth rings of otoliths. Fish. Bull., U.S. 74:990-994. Sakawaga, G. T., and M. Kimura. 1976. Growth of laboratory-reared northern anchovy, Engraulis mordax, from southern California. Fish. Bull., U.S. 74:271-279. Sameoto, D. D. 1972. Distribution of herring (Clupea harengus) larvae along the southern coast of Nova Scotia with observa- tions on their growth and condition factor. J. Fish. Res. Board Can. 29:507-515. Scott, J. S. 1973. Otolith structure and growth in northern sand lance, Ammodi/tes duhius, from the Scotian Shelf. Int. Comm. Northwest Atl. Fish. Res. Bull. 10:107-115. Sette, O. E. 1943. Biology of the Atlantic mackerel (Scomber scont- brus) of North America. Part I: Early life history, in- cluding growth, drift, and mortality of the egg and larval population. U.S. Fish Wildl. Serv., Fish. Bull. 50:149- 237. Sherman, K. 1980. MARMAP, a fisheries ecosystem study in the Northwest Atlantic: Fluctuations in ichthyoplankton- zooplankton components and their potential for impact on the system. In F. P. Diemer, F. J. Vernberg, and D. Z. Mirkes (editors), Advanced concepts in ocean mea- surements for marine biology, p. 9-37. Belle W. Baruch Institute for Marine Biology and Coastal Research. Univ. S.C. Press, Columbia. Steffensen, E. 1980. Daily growth increments observed in otoliths from juvenile East Baltic cod. Dana 1:29-37. Struhsaker, P., and J. H. Uchiyama. 1976. Age and growth of the nehu, Stolephorus purpur- eas (Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily growth increments of sagittae. Fish. Bull., U.S. 74:9-17. Taubert, B. D., and D. W. Coble. 1977. Daily rings in otoliths of three species of Lepomis and Tilapia mossambica. J. Fish. Res. Board Can. 34:332-340. Theilacker, G. H. 1980. Changes in body measurements of larval northern anchovy. Engraulis mordax, and other fishes due to handling and preservation. Fish. Bull., U.S. 78:685- 692. Tibbo, S. N., and J. E. Henri Legare. 1960. Further study of larval herring (Clupea harengus L.) in the Bay of Fundy and Gulf of Maine. J. Fish. Res. Board Can. 17:933-942. Tibbo, S. N., J. E. Henri Legare, L W. Scattergood, and R. F. Temple. 1958. On the occurrence and distribution of larval her- ring (Clupea harengus L.) in the Bay of Fundy and the Gulf of Maine. J. Fish. Res. Board Can. 15:1451-1469. Townsend, D. W., and J. J. Graham. 1981. Growth and age structure of larval Atlantic her- ring, Clupea harengus harengus, in the Sheepscot River estuary, Maine, as determined by daily growth incre- ments in otoliths. Fish. Bull., U.S. 79:123-130. Uchiyama, J. H., and P. Struhsaker. 1981. Age and growth of skipjack tuna, Katsmconus pel- amis, and yellowfin tuna, Thunnus albacares, as indi- cated by daily growth increments of sagittae. Fish. Bull., U.S. 79:151-162. Watson, J. E. 1964. Determining the age of young herring from their otoliths. Trans. Am. Fish. Soc. 93:11-20. Wilson, K. H., and P. A. Larkin. 1980. Daily growth rings in the otoliths of juvenile sock- eye salmon (Qncorhynchus nerka). Can. J. Fish. Aquat. Sci. 37:1495-1498. Zweifel. J. R., and R. Lasker. 1976. Prehatch and posthatch growth of fishes— a gen- eral model. Fish. Bull.. U.S. 74:609-621. 199 INCREMENT FORMATION IN THE OTOLITHS OF EMBRYOS, LARVAE, AND JUVENILES OF THE MUMMICHOG, FUNDULUS HETEROCHTUS 1 R. L. Radtke 2 and J. M. Dean 3 ABSTRACT The formation of otoliths and the effect of light cycles on increment formation were studied in embryos, larvae, and juvenile mummichogs, Fundulus heteroditus. We found that increments in the sagitta of mummichogs were a reliable indicator of the daily age of the fish. Calcification of the sagitta was initiated in the core, after matrix formation, at stage 24 of embryological development. The sagitta was the first calcified tissue to develop and there were two or three increments formed before hatching. Daily increment formation in the sagitta was initiated by light and controlled by a 24-hour photoperiod. When embryos were subjected to a 24-hour dark or <24-hour (6L:6D) photo- period, daily increment formation was disrupted. Laboratory experiments at24°C and 30°C con- firmed that there was one increment formed each day, which was independent of growth rate and which validated the age of fish in field collections. Wild populations reproduce in the intertidal zone, a physically stressed environment and, judging by the age, which was estimated from incremental data, reproduction is synchronized with tidal cycles. Interpretation of increments in the hard tissues of fish has long been utilized as a method to esti- mate age composition of adult populations. Most of the interpretive emphasis has been placed on otoliths and scales. However, the process of age determination isnotasimpleone(Bagenal 1974). Otoliths are especially useful for determining the age of fishes, such as larval forms, which lack scales or have very small ones. The otoliths of teleosts consist of deposits of cal- cium carbonate in the form of aragonite (Irie 1955; Degens et al. 1969). The morphology of these structures is so specific it can be used as a taxonomic character (Messieh 1972; Hecht 1978). Three structures (the sagitta, lapillus, and the asteriscus) are found in the membranous labyrinth of inner ear on each side of the brain cavity (Lowenstein 1971; Popper and Coombs 1980). The sagitta is often the largest and is most often used for age determinations and, unless otherwise stated, was the otolith used in the present study. ■This is contribution 390 of the Belle W. Baruch Institute for Marine Biology and Coastal Research, University of South Carolina, Columbia, SC 29208. 2 Belle W. Baruch Institute for Marine Biology and Coastal Research and the Department of Biology, University of South Carolina, Columbia, SC; present address: The Pacific Game- fish Foundation, P.O. Box 25115, Honolulu, HI 96825. 3 Belle W. Baruch Institute for Marine Biology and Coastal Research and the Department of Biology, University of South Carolina, Columbia, SC 96825. Manuscript accepted November 1981. FISHERY BULLETIN: VOL. 80, NO. 2, 1982. Pannella(1971, 1974) postulated that daily in- crements are found in otoliths of adult fishes, and Brothers et al. (1976) showed that such incre- ments can indeed be found in otoliths of young fishes and be used for age estimation. Struhsaker and Uchiyama (1976) postulated that back calcu- lation of daily increment data from otoliths could be used to age the nehu, a tropical marine fish, and Ralston (1976) obtained similar results with a tropical butterfly fish. Taubert and Coble (1977) did direct age observations of otoliths in juvenile freshwater fish and Barkman (1978) was equally successful with the young of a tem- perate estuarine species Menidia menidia. A more accurate daily journal is available in the otoliths of most young fishes than can be found in their scales, since scales are often absent in the early stages of development (Bagenal 1974), and scale metabolism is dynamic (Yamada and Watabe 1979). The discovery of daily increments in otoliths increases the resolution and precision of age de- termination and promises to provide fishery biologists with new levels of information. The deposition of the increments in a rhythmic fash- ion could be a mark of a daily event, and possibly a measure of growth, but the full extent of the influence of external and internal factors on the formation of otolith increments has not been de- termined. There is need of knowledge about the age com- 201 FISHERY BULLETIN: VOL. 80. NO. 2 position of larval fish populations, since this in- formation can provide estimates of growth, mor- tality, and rates of survival (Gulland 1977). The highest mortality of fishes is during the growth period from larvae to juveniles (Hjort 1914; Tanaka 1972) and consequently, the survival and growth of larval fishes has a pronounced effect upon recruitment (Larkin 1978). It should be possible, by using otoliths for estimation of the age, to determine the growth rates and the age structure of larval fish populations. Daily increments have been correlated with natural temperature cycles, light and food for freshwater species by Brothers (1978, 1980 4 ). Taubert and Coble (1977) postulated that daily increments in otoliths of freshwater sunfish re- sulted from a 24-h diurnal light cycle that en- trained an internal clock. To utilize daily depositional increments of the otoliths in the analysis of fish population dynam- ics, it is important to understand the physiologi- cal mechanisms involved in the formation and growth of increments and otoliths. Age estima- tion requires knowledge of 1) age when incre- ment formation begins; 2) factors which control the deposition of daily increments in the otoliths; and 3) length of time daily increments are formed without growth interruption. Informa- tion in these areas will make it possible to better understand age and growth in wild populations of fish. An important area for research in the field of age and growth is the experimental study of the factors which influence the deposition of incre- ments in otoliths. Brothers et al. (1976) showed that daily increments began to form at different ages in different species. Some species hatch with increments already formed, while others apparently do not form increments until later. Thus, it is necessary to study the formation of in- crements in each species and correlate incre- ment formation with external factors before accurate age determinations can be made. The mummichog, Fundulus heteroclitus, is an abundant estuarine fish and an important com- ponent of the estuarine ecosystem (Cain and Dean 1976; Valiela et al. 1977; Kneib and Stiven 1978; Merideth and Lotrich 1979). The biology of Fundulus is well-known and its embryology is well-defined (Armstrong and Child 1965). The objectives of this study were to 1) delineate 4 E. B. Brothers, Section of Ecology and Systematics, Cornell University, Ithaca, NY 14850, pers. commun. October 1980. the structure and formation of otoliths in the embryological and early larval stages of the mummichog, 2) determine the effect of photo- period on incrementdeposition in embryonic and postlarval mummichog otoliths, 3) measure the effects of temperature on body growth and the deposition of increments in otoliths, and 4) test whether growth and age data can be obtained in wild populations of mummichogs by counting the increments in otoliths. METHODS Adult F. heteroclitus used as spawning stock were collected from North Inlet Estuary (lat. 32°20'N, long. 79°10'W) and North Edisto Estu- ary (lat. 32°26'N, long. 80°12'W), near George- town, S.C. Fertilized eggs were collected as pre- viously described by Middaugh and Dean (1977). Only embryos which developed according to the criteria of Armstrong and Child (1965) were uti- lized in the embryological studies, and only lar- vae which hatched within 6 h of hatch induction were used in the growth studies. The embryo is the stage from fertilization to hatching; from hatching to yolk-sac absorption is the larval stage and the mummichog was considered a juvenile after yolk-sac absorption (Hubbs 1943). The terms used to describe growth increments in otoliths are confused, as the increments in lar- vae are variously referred to as lamellae, rings, or layers. The term increment in this study re- fers to a unit formed by an unbroken incremental zone and a discontinuous zone after core forma- tion (Fig. 1), Wild and Foreman (1979). Newly hatched larvae were kept at 24°C and 30°C±1°C (Radtke and Dean 1979) and were fed brine shrimp, Artemia nauplii, ad libitum and maintained in L12:D12 with a daily change of water (30%„) to determine the effect of the rate of growth on otolith size and increment number. A daily sample of 10 larvae was collected for laboratory experiments from each group for the first 10 d, and every 5 d thereafter for 30 d. Stan- dard lengths (SL) were measured on each larva and its otoliths were removed. Photomicro- graphs were made of each otolith for increment counts. Juvenile mummichogs were collected from We Creek in North Edisto Estuary on 9 June 1977 (28°C, 297..). Each fish was weighed, measured for standard length (SL), and its otoliths extract- ed for increment counts from photographs. Sta- tistical analyses of the data were done with 202 RADTKE and DEAN: INCREMENT FORMATION OF OTOLITHS OF MUMMR'HOO Figure 1.— SEM of the sagitta from a 12-d-old Fundulus heteroclitus. I is the unbroken incremental zone, D is the discontinuous zone, and I + D = 1 increment. Bar = 1 p. 203 FISHERY BULLETIN: VOL. 80. NO. 2 standard tests and models as described in Sokal and Rohlf (1969). Removal, Preparation, and Inspection of Otoliths Otoliths were removed from embryos, larvae, and juveniles with fine insect needles mounted on wood rods. The larvae are transparent and the otoliths are birefringent under polarized light, so it is possible to view the sagitta during the dissection. The sagittae were washed with dis- tilled water, dried, and mounted on glass slides with Euparol 5 mounting medium, and viewed with a compound light microscope. Photomicrographs were made of each otolith for counts of increments and measurement of otolith diameters. (Thelof the outside edge of the sagitta was considered as a portion of the last in- crement.) To make increment counts, the back of each photograph was marked and the photo- graphs were shuffled. The counting process was performed three times, which gave three un- biased readings for each otolith. If two of the counts were identical, that value was accepted as the increment count for a particular otolith. In cases where all counts differed, the middle count was chosen unless all counts varied more than two increments from each other, in which case that otolith was disqualified and not used in the final tabulation. Sagitta were measured at the widest diameter on the photographs using a cali- per calibrated on a photographed micrometer. Sagittae viewed with light microscopy showed fine lines in the I that were concentric with the D; these fine lines have been referred to as "sub- units." In the otoliths of young mummichogs the so-called subunits could not be observed in decal- cified sections with light microscopy or SEM (Fig. 1). The D and I compose an increment and are readily differentiated with light microscopy in Fundulus sagittae (Fig. 2). Whole sagittae used for SEM studies were attached to viewing stubs in 5-min epoxy resin. The sagittae were ground to the core in the trans- verse plane on graded grinding stones, polished with diamond-polishing compound, and cleaned with 95% ethanol. The polished surface was de- calcified with 7% EDTA (pH 7.4)(disodium ethy- lenediaminetertacetate) for 1 to 5 min. The speci- mens were coated with gold ( 150A) and observed with a SEM. 5 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Embryological Formation of Otoliths Fertilized eggs were kept in light 12 h and in the dark 12 h (L12:D12) at24°C in hatching jars with recirculating seawater (30 ..). A sample of 10 eggs was collected each day until hatching and viewed under polarized light (120X) to de- termine when calcification was initiated. The embryos were classified according to Armstrong and Child (1965) with the number of embryos with calcified sagitta noted in each stage. Calcified sagittae were removed from the em- bryos and mounted for examination with light microscopy to determine the time of increment formation. Ten- and 14-d embryological sagittae were viewed using SEM to confirm the light microscope observations. Effect of Light on Increment Formation in Embryos and Larvae To determine the influence of light on incre- ment formation, developing embryos and larvae were subject to the following conditions: EMBRYOS: Group ED24— Embryo-dark-24 h, fertilized in the dark, kept in constant darkness until sam- pled 3 d after hatching. Group EL24— Embryo-light-24 h, fertilized in the light, kept in constant light until sampled 3 d after hatching. Group ED24+L— Embryo-dark-24+L, fertilized in the dark, kept in constant darkness except for 1 min of light exposure 10 d after fertiliza- tion. Sampled 3 d after hatching. Group EL12:D12— Embryo-light-12 h:dark-12 h, fertilized, placed in L12:D12 and sampled daily. All groups were maintained at 24°C and the water (30 '/..) was changed daily. The water was changed in the ED24 group and ED24+L group by pouring the eggs onto a 505 jjl mesh net mounted on the end of 10 cm plastic tubing. The eggs were then washed off the netting with a wash bottle and the entire exercise was per- formed in total darkness. Hatching in the ED24 group and ED24+L group was determined by touch, because embryos are hard and easily dis- tinguished when they have hatched. A daily 204 RADTKE and DEAN: INCREMENT FORMATION OF OTOLITHS OF MUMMK'HOO Figure 2.— Light micrograph of a Fundulus heteroclitus sagitta on the day of hatching. C is the core and II and 12 are increments formed after core formations but prior to hatching. Bar = 0.05 mm. sample of three eggs was taken after day 14 to determine the events in otolith development. Sagitta were removed from 10 larvae of each group according to the above schedule and photo- micrographed. LARVAE: Developing embryos were maintained at 24°C in L12:D12 in hatching jars with running sea- water (30'/..). Upon hatching, the larvae were di- 205 FISHERY BULLETIN: VOL. 80, NO. 2 vided and subjected to the following conditions: Group LaD24— Larvae-dark-24 h, constant darkness. Group LaL24— Larvae-light-24 h, constant light. Group LaL6:D6— Larvae-light-6 h:dark-6 h. Group LaL12:D12— Larvae-light-12 h:dark- 12 h. All groups were fed newly hatched brine shrimp ad libitum and kept at 24°C with daily changes of water at 30V... Samples of 10 larvae from each group were taken at days 0,6,9, and 16 except the L12:D12 group, which was sampled daily. Sagitta were removed from each sample and photomicrographed. Scanning electron mi- crographs were made of samples for comparison with the light micrographs. RESULTS Formation of Otoliths in Embryos The sagittae were the first tissues to calcify and were discernible on days 3 and 4 at embry- onic developmental stages 24-28 (Armstrong and Child 1965). An amorphous mass was discernible in the labyrinth region of the larva before calcifi- cation was initiated. This mass, the core organic matrix, had a gellike consistency and could be dissected. Calcification was initiated in the core of the sagitta of 30% of the embryos on day 3 and 100% of the cores showed calcification by day 4 (Fig. 3). Increment formation began on day 12, and 20% had one increment. On day 13, 80% had one increment and 20% had two increments. On day 14, the day of hatch, 20% had one increment, 70% had two increments (Fig. 2), and 10% had three increments. Calcification began with formation of crystals which extended to the edge of the core matrix. Histochemical analyses have shown that calcifi- cation begins in the core at the same time that the core becomes birefringent (J. Yamada 6 ). Multi- ple spherules (Fig. 4a, b) are common in the calci- fied core but their origin and sequence of devel- opment is unknown. The newly formed sagittae had a mean diameter of 0.024±0.004 mm. Calci- fication continued and additional crystals ex- tended beyond the original boundary in an in- terlocking fashion until the diameter reached 0.048+0.008 mm at day 9 and developmental stage 36. At this time only the core region could be observed, with no increments (Fig. 3). Two days later (on day 11 postfertilization), incre- ment formation was initiated around the core, and the mean sagitta diameter had reached 0.074±0.008 mm. When viewed with trans- mitted light, the concentric increments consisted of alternate narrow, dark discontinuous zones (D) and wider, lighter, incremental zones (I) (Fig. 3). The D intersected the I at right angles and were concentric with the core and outer sur- face of the otolith. Upon hatching at day 14, post- fertilization, two or three increments were readily discernible as daily increments started forming 2-3 d before hatching. Otoliths examined with the SEM confirmed the increment counts determined under trans- mitted light and showed the orientation of the crystals (Fig. 1). Effect of Light on Increment Formation in Embryos and Larvae The light cycle to which an embryo or larva was exposed had an effect on increment forma- tion and hatching time. Embryos in the L12:D12 cycle had two or three increments prior to hatch- ing and one increment per day after hatching (Table 1, Fig. 3). Embryos kept in L12:D12 hatched at 14 d while those exposed to other light cycles had longer incubation times and a differ- ence in increment formation during incubation and after hatching was apparent in the other groups (Table 2). Embryos incubated inconstant dark (ED24)had a delayed hatch, suppressed in- crement formation (Fig. 5), and a smaller otolith Table 1. — Sagitta were from Fundulusheteroclitus embryos and larvaeincubatedonaL12:D12eycleat 24°C (N = 10/d). 6 Faculty of Fisheries, Hokkaido University, Hakodate. Hokkaido, Japan, pers. commun. Age (days Otolith diameter after hatching) Increment count (mm) 2.75±0.5 0.128*0011 1 400 ±0.0 150+0015 2 450+0.5 0.158 ±0.015 3 5.50±0.5 180+0 140 4 620 + 1.1 187+0011 5 7.20±1.0 0203+0003 6 8.50 + 1.1 0.220 ±0.1 11 7 950+06 240+0.110 8 1060+0.8 0255+0.042 9 12.00±1.0 0.268+0 058 10 1260+0.9 0.280+0.018 15 17 50±1.2 0350+0.120 206 RADTKE and DKAN: INCREMENT FORMATION OF OTOLITHS OF MUMMICHOG m .02 mm Figure 3.— Light micrograph of the core of the sagitta of Fundulus heteroclitus taken on day 10 of embryo formation. No increments have yet formed. Bar = 0.02 mm. diameter (Table 2A, Fig. 5). The embryos incu- bated in constant light (EL24) hatched at 15 d postfertilization and showed 6.0+0.67 incre- ments when sampled 3 d after hatching (Table 2B). Constant light conditions did not signifi- cantly alter increment formation; the constant light group (EL24) showed the same number of increments as in the EL12:D12 group at 3 d of age. Thus, the effect of light on embryonic incre- ment formation and otolith diameters was the same for the EL24 and EL12:D12 groups. A 1-min light stimulus on day 10 of ED24+L 207 FISHERY BULLETIN: VOL. 80, NO. 2 ** W FIGURE 4.— a) SEM of the core of the sagitta of Fundulus heteroclitus showing the core (C) and the multiple primordia (P) sur- rounds the spherules. Bar = 1 n. b) SEM showing a spherule (S) in the multiple primordia (P) of the core (C) or the sagitta. B = 0.5 M . 208 RADTKK and DKAN: INCRKMKNT FORMATION OF OTOLITHS OF MUMMICHOO resulted in increment formation (Table 2C) in embryos otherwise maintained in constantdark- ness. Increment counts for ED24+L were less TABLE 2.— Effect of photoperiod and light stimuli on incre- ment formation in sagittal otoliths of Fundulus heteroclitus embryos. A) B) C) A B C Embryos were incubated in total darkness and sagitta were removed 3 d after hatching. (ED24) Embryos were incubated in constant light and sagitta were removed 3 d after hatching. (EL24) Embryos were incubated in constant darkness with 1 mm of light at day 10 of development. Sagitta were removed 3 d after hatching. N = 10 in all groups. (ED24+L) Sagitta diameter (mm) (X+SD) Increment numbers (X±SD) 0.071 ±0.008 0.177+0.013 0.14610012 0.610.77 6.0±0.67 4.7±0.48 than those of the EL24 group and the EL12:D12 group at 3 d after hatching, but were very close to the increment counts found at day 2 of the EL12: D12 group. The effect of light on larvae which were main- tained under EL12:D12 during embryonic de- velopment and then transferred to constant darkness after hatching was notasevidentasthe effect of light was in the embryos maintained in constant darkness. Larvae raised in constant darkness (LaD24) showed a rapid addition of in- crements between day and day 6 after hatch- ing, but few increments formed after day 6 (Table 3). When the LaD24 data were compared with the data from the larvae hatched and raised % ,^•4" Figure 5.— Light micrograph of the sagitta from a newly hatched larvae incubated for the total embryonic period in total darkness. Core formation is present but no increments have formed. Bar = 0.02 mm. 209 FISHERY BULLETIN: VOL. 80, NO. 2 Otolith diameter Increment count (mm) 275+05 0.128+0.011 13.89±1.69 0.181+0.010 1522+083 0.190+0.010 15.60±2.01 0.204+0.015 Table 3.— Otoliths are from Fundulus heteroclitus larvae. Embryos were incubated on a L12:D12 cycle and the larvae transferred to constant darkness at 24 °C immediately after hatching. Age 6 9 16 in EL12:D12 (Table 1), sagitta of LaD24 had re- duced increment numbers after day 6 and sagitta diameters in experimental fish were smaller than those found in the control (LaL12:D12). Some groups that had increment formation (LaD24 and LaL6:D6) during the first 6 d had in- crements formed after day 6 that were unclear and it was difficult to differentiate the D and I in the outer areas. However, the LaL12:D12 group showed distinct increments beyond day 6. The ED24 group larvae were sluggish upon hatching as were the ED24+L group. The larvae appeared to be normal in every other fashion except that the yolk sacs were notably smaller than the 12L:12D group. Effect of Temperature and Body Growth on Otolith Formation in Larvae An increase in temperature caused an increase in the growth rate in the larvae (Fig. 6a). The 30°C group grew significantly faster than the 24°C group (P<0.05). The 30°C larvae, also formed otoliths (Fig. 6b) which were significantly larger (P<0.05) in di- ameters than those in fish held at24°C. However, the difference in growth rates had no effect on the increment counts from either group (Fig. 6c). Both showed daily increment formation in their otoliths but the faster growing otoliths had wider daily increments, which accounted for the in- creased diameter measurements. When the oto- lith diameter data were pooled and compared with length data, the relationship was highly correlated (r = 0.95; Fig. 7). Estimation of Age of Wild Fish It is difficult to gain any insight into the age structure of the wild population from the length- frequency histograms, e.g., larvae collected 9 June 1977 had a standard length-frequency x t- o -z. UJ 35r 30 25 20 15 10 5 30°C Y=6.I7 + .83X '' r=.99 ,'/ 24°C Y = 6.26 +.66X r=.99 J L J I 0.7- 0.6 /-\ ^ ^ 0.5 v^ rr LlI 0.4 h- LJ 2 0.3 < Q 0.2 0.1 35 30 w 25 i- lu 20 UJ IK en 15 o ^ 10 5 0. B 30°C Y = .I25 +.0I9X ,' r=.98 / 24°C Y = .131 + .0I5X r = .98 J i 30 U C Y=2.82 + IX hr = .99 24°C Y=2.58 +IX r=.99 -L -L J 10 15 20 25 30 AGE (DAYS) Figure 6.— A regression plot of A) standard length (SL), B)the diameter of the sagitta, and C) the numbers of increments of the sagitta of Fundvlus heteroclitus reared at 24°C and 30°C plotted against age of the fish. 210 RADTKK and DKAN: INCREMENT FORMATION OF OTOLITHS OF MUMMICHOO 15 20 LENGTH ( MM I Figure 7.— A regression plot of the diameters of all of the 24° and 30° Fundulus heteroclitus sagitta plotted against the stan- dard length of the fish. mode of 23 mm (Fig. 8a). However, otolith incre- ment-frequency histograms of the sample en- abled us to differentiate cohorts (Fig. 8b). A statistical analysis of the data from the field population showed that the relationship between the length of the fish and otolith diameter was linear (y = 0.01 + 0.027a, r = 0.90). The relation- ship of increment number and otolith diameter was curvilinear (y = 0.7601- 0.02 17.r + 0.000rx 2 , r = 0.92). Thus, the diameter of the otolith in- creased as the fish grew; the width of the incre- ment was wider in younger, smaller fish than in older, larger fish; and the number of increments increased as the length of the fish increased. When the time of hatching was estimated, using increment counts (Fig. 9a), groups were found that correlated with the occurrence of new and full moons. We observed that the increments tended to be more distinct in larvae collected from the field than in laboratory-reared larvae. When ages were adjusted for the two or three prehatching increments, the relationship was even more obvious (Fig. 9b). Incremental data indicated that the fish collected hatched at the new and full moon spring tides. DISCUSSION Embryological Formation of Otoliths Otoliths (sagittae) are the first calcified tissues to form in developing F. heteroclitus embryos, and although they are prominent and easily observed features that have been presented in numerous developmental studies, their forma- tion is not discussed. Long and Ballard (1976) clearly showed otoliths that formed at stage 20 in LENGTH { MM jll^ INCREMENTS Figure 8.— a) A length-frequency histogram of all fish col- lected in the sample, b) A histogram showing the frequency of the increment number of the sagitta of same fish as in 8a. • s o • o a *i .U* IS iIj 30l 1 a .si JUNE DATE (1977) JUNE DATE (19771 Figure 9.— Fundulus heteroclitus larvae collected on 9 June 1977. a) Estimated hatching dates are determined from num- bers of increments in the sagitta; b) estimated hatching times are adjusted for the two increments formed prior to hatching. Also represented are diurnal high tides (upper lines) and low tides (lower lines) and lunar phase (open circles = full moon, closed circles = new moon). embryos of the white sucker, and Armstrong and Child (1965) showed otoliths in mummichog em- bryos at stage 23 with calcification at stage 24, 211 FISHERY BULLETIN: VOL. 80. NO. 2 which agreed with this study, but their ontogeny is not well known. The importance and function- al nature of the early otolith calcification has not yet been determined. Two or three increments were easily visible in the mummichog otolith at the time of hatching. Accurate age determination of field samples could be affected until the number of increments formed at the time of hatching is considered. Brothers et al. (1976) studied increment forma- tion in several fish species and found that the California grunion, Leuresthes tenuis, had two increments at hatching. Some species, such as the northern anchovy, Engraulis mordax, had no increment formation until the time of yolk-sac absorption, 6 d after hatching (Methot and Kramer 1979). Taubert and Coble (1977) found that three species of Lepomis began increment formation at swim up. Scott (1973) studied the otolith structure in larvae of the northern sand lance, Ammodytes dubius, and suggested that otoliths first formed in the postlarvae at a mean total length of 2.4 cm. However, his interpreta- tion was a result of back calculations, not direct observations of otol iths from known age or larval stages of the fish. We have found that multiple spherules in the core of the sagitta, followed by numerous incre- ments, are formed prior to hatching in the Asiatic salmon or masou, Oncorhynchus masou; chum salmon, 0. keta; pink salmon, 0. gorbuscha; Arctic char, Salvelinus alpinus; brook trout, S. fontinalis; rainbow trout, Salmo gairdneri; and the sculpin, Cottus nozawa. The juveniles of the live bearing guppy, Lebistes reticulatus, and mosquitofish, Gambusia affinis, form a large number of increments prior to being spawned (Radtke and Dean unpubl. data). Mummichogs, California grunion, and the Atlantic silverside, Menidia menidia, have tidally correlated incu- bation periods of about 10 to 14 d and the salmo- nids incubation period can exceed 50 d. In con- trast, the northern anchovy and spot, Leiostomus xardhurus, have short incubation periods of <2 d. This indicates that embryos which have longer incubation periods and large yolk sacs may form several increments before hatching, while em- bryos that have short incubation periods might not start increment formation until hatching or after yolk-sac absorption (Brothers et al. 1976; Methot and Kramer 1979). Much work remains to be done on a range of species before we can attempt to interpret the functional significance of increment formation in embryos. The Effect of Light on Increment Formation in Embryos and Larvae The increments observed in otoliths in this and other studies (Pannella 1971; Brothers et al. 1976; Struhsaker and Uchiyama 1976; Taubert and Coble 1977; Barkman 1978; Methot and Kramer 1979) appear to be indicators of daily biological events. Rhythmic physiological activi- ties, such as the occurrence of rhythmic mineral deposition in coral (Wells 1963), crayfish gastro- liths (Scudamore 1947), and marine bivalves (Clark 1968; Pannella and MacClintock 1968), are controlled to a large extent by environmental changes synchronized to the diurnal astronomi- cal cycle. The only examination of the effect of endogen- ous daily biological rhythms on fish otoliths was by Taubert and Coble (1977), who studied the effect of environmental factors on daily incre- ment formation of Tilapia mossambica larvae hatched in constant light. Their different experi- mental groups all showed increment formation but it was not always daily. They found normal increment formation in all experimental groups with a 24-h periodicity and any other cycle other than 24-h period disrupted increment formation. Since daily cycles are known to occur in blood chemistry of fish (Garcia and Meier 1973), those daily chemical changes could be reflected in the daily increments of the otoliths. Mugiya (1966) found monthly changes in total and diffusible calcium in the endolymph of the semicircular canals of the rainbow trout and the flatfish, Kareius bicolaratus, and he related his finding to the formation of the opaque and translucent zones found in adult otoliths. Daily changes in the calcium metabolism of the fish also occur (Mugiya et al. 1980) which are reflected in the formation of the I and D. Daily increments were formed in F. heterocli- tus larvae kept in a L12:D12 cycle, but were absent when the developing embryos were kept in constant darkness (Fig. 5). Light had a defi- nite effect on increment formation, as embryos kept in constant light showed increment forma- tion and otolith diameters that were comparable with the L12:D12 group. An insight into this dis- crepancy was gained in the analysis of the group which initiated increment formation after a light stimulus on day 10 after fertilization. The possibility that light is a synchronizing stimulus at the cellular level was demonstrated by Pitten- 212 RADTKK and DEAN: INCRKMKNT FORMATION OF OTOLITHS OF MUMMIOIKx; drigh and Bruce (1957), who showed that a light stimulus synchronized emergence in fruit flies. More study is necessary to determine the timing of light needed for increment formation as well as the quantity and quality of light necessary. Whether the control of increment formation is an endogenous or exogenous rhythm (Harker 1957) is beyond the scope of these experiments. But the experiment on increment initiation in the dark group with 1 min of light exposure on day 10 in- dicated that light can act as a synchronizing stimulus, similar to that observed by Pittendrigh and Bruce (1957). Mugiya et al. (1980) found that D formation was initiated when light inter- rupted a photo period of 12L:12D or longer light period, but they did not determine the minimum dark period necessary for formation of the Dor the free running period for the D and I. When F. heteroclitus larvae were hatched in L12:D12 and then placed in light regimes other than a 24-h photoperiod, the increment forma- tion became aphasic in each group and incre- ment formation occurred at a slower rate. The "biological clock" of this group seemed to be out of phase under photoperiods other than those with a 24-h periodicity. A great deal of very exciting work is necessary to resolve these funda- mental questions on increment control. Effects of Temperature and Body Growth on Otolith Formation in Larvae Under the various experimental conditions employed in this study, daily otolith increments formed regardless of body growth or otolith growth rate (Fig. 6a, b, c), so it was possible to determine age and daily growth rates of individ- ual larvae which lived under different environ- mental conditions. Although F. heteroclitus lar- vae grew faster at 30°C than at24°C, the number of increments was still directly related to chrono- logical age. This documents the reliability of oto- lith increments for the age estimation of mum- michog larvae. It has been demonstrated that daily increments exist in several other species of fish (Pannella 1971, 1974; Brothers et al. 1976; Struhsaker and Uchiyama 1976; Ralston 1976; Taubert and Coble 1977) and the relationship be- tween increment counts and fish and otolith size was shown for the Atlantic silversides(Barkman 1978). In this study, otolith diameter increased with increased body length and increments formed on a daily basis with wider increments found in younger fish than older fish. This is con- sistent with the fact that younger fish are grow- ing faster, and although the relationship is non- linear, it is predictable and these results are consistent with those of Methot and Kramer (1979). Estimation of Age of Wild Fish Daily increments observed in field samples were easier to interpret than increments found in laboratory-reared larvae. We were not able to make age estimations of field collections of mum- michogs from length-frequency histograms, but it was possible to determine the age and growth rate of individual larvae from increment counts. Ralston (1976) and Struhsaker and Uchiyama (1976) determined growth rates of the millet- seed butterfly fish, Chaetodon miliaris, and the nehu, Stolephorus purpureus, respectively, and found that the growth, as represented in incre- mental units in the otolith, was nearly linear. Similar results were obtained by Barkman (1978) for Atlantic silversides and Methot and Kramer (1979). Our results are consistent with theirs: that increment formation is independent of growth rate but is age dependent; thus growth rates can be estimated for individual larval fish. Analysis of the age structure of samples of wild larval mummichogs showed that larvae hatched on or near the time of full and new moons. This is corroborated by observations on the reproduc- tive biology of F. heteroclitus by Taylor et al. (1977, 1979) and DiMichele and Taylor (1978), New Zealand white bait, Galaxias maculatus, by McDowell (1968), and Atlantic silversides by Middaugh (1981). Eggs of the California grun- ion, an intertidal spawner, have been found to hatch during spring tides (Clark 1925) and have otolith increments at hatching (Brothers et al. 1976). An analysis of age structure of wild populations of mummichog larvae, as deter- mined from their otoliths showed that South Carolina mummichogs spawn from March to mid-August and have a lunar spawning perio- dicity during that season. Analysis of otolith in- crements enabled us to differentiate individual fish in the wild population of the same size but of different ages. Photoperiod is a critical factor in increment formation, but other factors such as diurnal migratory behavior, rhythmic feeding, tempera- ture, respiration, and tidal rhythms might also 213 FISHERY BULLETIN: VOL. 80. NO. 2 play significant roles. Even though the control and/or mechanism of daily increment formation in larval fish is not fully understood, the incre- ments are a powerful tool for analysis of individ- ual growth and age determination of very young fish. ACKNOWLEDGMENTS We wish to thank R. Feller and D. Middaugh for their constructive criticism and D. Dunkel- berger for his assistance with the SEM work. This manuscript was based on a dissertation sub- mitted as partial fulfillment of the Ph.D. degree, University of South Carolina. LITERATURE CITED Armstrong, P. B., and J. S. Child. 1965. Stages in the normal development of Fundulus heteroclitus. Biol. Bull. (Woods Hole) 128:143-168. Bagenal, T. B. (editor). 1974. The proceedings of an international symposium on the ageing of fish. Unwin Brothers, Surrey, Engl., 234 p. Barkman, R. C. 1978. The use of otolith growth rings to age young Atlan- tic silversides, Menidia menidia. Trans. Am. Fish. Soc. 107:790-792. Brothers, E. B. 1978. Exogenous factors and the formation of daily and subdaily growth increments in fish otoliths. (Abstr.) Am. Zool. 18:631. Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Cain, R. L.. and J. M. Dean. 1976. Annual occurrence, abundance and diversity of fish in a South Carolina intertidal creek. Mar. Biol. (Berl.) 36:369-379. Clark, F. N. 1925. The life history of Leuresthes tenuis, an atherine fish with tide controlled spawning habits. Calif. Dep. Fish Game, Fish Bull. 10, 51 p. Clark, G. R., II. 1968. Mollusk shell: Daily growth lines. Science (Wash., D.C.) 161:800-802. Degens, E. T., W. G. Deuser, and R. L. Haedrich. 1969. Molecular structure and composition of fish oto- liths. Mar. Biol. (Berl.) 2:105-113. DlMlCHELE, L., AND M. H. TAYLOR. 1978. Environmental influences on the hatching of Fun- dulus heteroclitus. (Abstr.) Physiologist 21(4):29. Garcia, L. E., and A. H. Meier. 1973. Daily rhythms in concentration of plasma Cortisol in male and female gulf killifish, Fundulus grandis. Biol. Bull. (Woods Hole) 144:471-479. Gulland, J. A. (editor). 1977. Fish population dynamics. Wiley, Lond.. 372 p. Hecht, T. 1978. A descriptive systematic study of the otoliths of the neopterygean marine fishes of South Africa. Part I. In- troduction. Trans. R. Soc. S. Afr. 43:191-197. Hjort, J. 1914. Fluctuations in the great fisheries of northern Eur- ope viewed in the light of biological research. Rapp. P.- P.-V. Reun. Cons. Int. Explor. Mer 20:5-38. Hubbs, C. L. 1943. Terminology of early stages of fishes. Copeia 1943:260. Irie, T. 1955. The crystal texture of the otolith of a marine tele- ost, Pseudosciaena. J. Fac. Fish. Anim. Hus. Hiro- shima Univ. 1:1-13. Knieb, R. T., and A. E. Stiven. 1978. Growth, reproduction and feeding of Fundulus heteroclitus on a North Carolina salt marsh. J. Exp. Mar. Biol. Ecol. 31:121-140. Larkin, P. A. 1978. Fisheries management — an essay for ecologists. Annu. Rev. Ecol. Syst. 9:57-73. Long, W. L., and W. W. Ballard. 1976. Normal embryonic stages of the white sucker, Catostomus commersoni. Copeia 1976:342-351. Lowenstein, O. 1971. The labyrinth. In W. S. Hoar and D. J. Randall (editors). Fish physiology. Vol. 5, p. 207-240. Acad. Press, N.Y. McDowell, R. M. 1968. Galaxias maculatvs (Jenyns), the New Zealand whitebait. N.Z. Mar. Dep. Fish Res. Div. Bull, New Ser. 2, p. 1-84. Merideth, W. H., and V. A. Lotrich. 1979. Production dynamics of a tidal creek population of Fundulus heteroclitus (Linnaeus). Estuarine Coastal Mar. Sci. 8:99-118. Messieh, S. N. 1972. Use of otoliths in identifying herring stocks in the southern Gulf of St. Lawrence and adjacent waters. J. Fish. Res. Board Can. 29:1113-1118. Methot, R. D., Jr., and D. Kramer. 1979. Growth of northern anchovy, Engraulis mordajc, larvae in the sea. Fish. Bull.. U.S. 77:413-423. Middaugh, D. P. 1981. Reproductive ecology and spawning periodicity of the Atlantic silverside, Menidia menidia. Copeia 1981:766-775. Middaugh, D. P., and J. M. Dean. 1977. Comparative sensitivity of eggs, larvae, and adults of the estuarine teleosts, Fundulus heteroclitus and Me- nidia menidia to cadmium. Bull. Environ. Contam. Toxicol. 17:645-652. Mugiya, Y. 1966. Calcification of fish and shell-fish— VI. Seasonal change in calcium and magnesium concentration of the otolith fluid in some fish, with special reference to the zone formation of their otolith. Bull. Jpn. Soc. Sci. Fish. 32:549-557. Mugiya, Y., N. Watabe, J. Yamada, J. M. Dean, D. G. Dun- kelberger, and M. Shimizu. 1981. Diurnal rhythm in otolith formation in the gold- fish, Carassius auratus. Comp. Biochem. Physiol. 68A:659-662. Pannella, G. 1971. Fish otoliths: Daily growth layers and periodical patterns. Science (Wash., D.C.) 173:1124-1127. 1974. Otolith growth patterns: an aid in age determina- tion in temperate and tropical fishes. In T. B. Bagenal 214 RADTKE and DEAN: INCREMENT FORMATION OF OTOLITHS OK MUMMK'lloo (editor). The proceedings of an international symposium on the ageing of fish, p. 28-39. Unvvin Brothers. Sur- rey, Engl. Pannella. G., and C. MacClintock. 1968. Biological and environmental rhythms reflected in molluscan shell growth. Paleontol. Soc. Mem. 2:64- 80. PlTTENDRIGH, C. S., AND V. G. BRUCE. 1957. An oscillator model for biological clocks, In D. Rudnich (editor), Rhythmic and synthetic processes in growth, p. 75-109. Princeton Univ. Press, Princeton. Popper. A. N.. and S. Coombs. 1980. Auditor mechanisms in teleost fishes. Am. Sci. 68:429-440. Radtke, R. L., and J. M. Dean. 1979. Feeding, conversion efficiencies and growth of lar- val mummiehogs, Fundulus heteroclitus. Mar. Biol. (Berl.) 55:231-237. Ralston, S. 1976. Age determination of a tropical reef butterflyfish utilizing daily growth rings of otoliths. Fish. Bull., U.S. 74:990-994. Scott, J. S. 1973. Otolith structure and growth in northern sand lance, Ammodytes ditbius. from the Scotian shelf. Int. Comm. Northwest Atl. Fish. Res. Bull. 10:107-116. Scudamore. H. H. 1947. The influence of the sinus glands upon molting and associated changes in the crayfish. Physiol. Zool. 20:187-208. Sokal, R., and F. Rohlf. 1969. Biometry: The principles and practice of statistics in biological research. Freeman, San Franc, 776 p. Strihsaker, P., and J. H. Uchiyama. 1976. Age and growth of the nehu, Stole/thorns purpur- etts(Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily growth increments of sagittae. Fish. Bull., U.S. 74:9-17. Tanaka, M. 1972. Studies on the structure and function of the diges- tive system in teleost larvae— V. Epithelial changes in the posterior-gut and protein ingestion. [In Jpn., Engl, summ.] Jpn. J. Ichthyol. 19:172-180. Taubert, B. D., and D. W. Coble. 1977. Daily rings in otoliths of three species of Lepomis and Tilapia mossambica. J. Fish. Res. Board Can. 34:332-340. Taylor, M. H., L. DiMichele, and G. J. Leach. 1977. Egg stranding in the life cycle of the mummichog. Fundulus !nt, rod it us. Copeia 1977:397-399. Taylor, M. H., G. J. Leach, L. DiMichele, W. M. Levitan, and W. F. Jacob. 1979. Lunar spawning cycle in the mummichog, Fun dulus heteroclitus (Pisces: Cyprinodontidae). Copeia 1979:291-297. Valiela, I., J. E. Wright, J. M. Teal, and S. B. Volkmann. 1977. Growth, production and energy transformations in the salt-marsh killifish Fundulus heteroclitus. Mar. Biol. (Berl.) 40:135-144. Wells. J. W. 1963. Coral growth and geochronometry. Nature (Lond.) 197:948-950. Wild, A., and T. J. Foreman. 1980. The relationship between otolith increments and time for yellowfin and skipjack tuna marked with tetra- cycline. [In Engl, and Span.] Inter-Am. Trop. Tuna Comm. Bull. 17:509-560. Yamada. J., and N. Watabe. 1979. Studies on fish scale formation and resorption. I. Fine structure and calcification of the scales in Fundulus heteroclitus (Athereniformes: Cyprinodontidae). J. Morphol. 159:49-66. 215 THE LARVAL DEVELOPMENT OF SERGESTES SIMILIS HANSEN (CRUSTACEA, DECAPODA, SERGESTIDAE) REARED IN THE LABORATORY Margaret Knight 1 and Makoto Omori 2 ABSTRACT The larval development of Sergestes similis Hansen reared in the laboratory includes the following stages: nauplius I-IV, protozoea I-III, and zoea I-II. These forms together with the first two postlarval stages are described and illustrated. Sergestes sim His and S. arcticus, closely related species which comprise the arcticus species group, are very similar in larval as well as adult morphology especially in the ornate armature of protozoeal carapace apparently specific to the group. In contrast, the two species of the atlanticus group, S. atlanticus and S. comutus, differ distinctly from each other in carapace armature of the protozeal stages. The difference between these two species groups in variation within each group indicates that larval morphology may be of value in the study of interspecific relationships within Sergestes. Sergestes similis and Sergia lucens, species of closely related genera, differ in number of naupliar stages, in armature of body in protozoeal and zoeal phases, and in development of some appendages. The pelagic shrimp Sergestes similis is abundant in the North Pacific Drift ranging from Japan to North America between 40° and 50°N, and is a prominent constituent of the plankton in the cooler waters of the California Current. Within the genus Sergestes (Omori 1974), S. similis is located in the arcticus species group, as defined by Yaldwyn (1957), which includes only the two species 5. arcticus Kroyer and S. similis Hansen. Sergestes arcticus is widely distributed, occurring in the North Atlantic, the Mediter- ranean, and all sectors of the Southern Ocean, while S. similis is restricted to the subarctic and transitional zones of the North Pacific; available data indicates that the species are geographi- cally isolated from one another (Judkins 1972). The life history and distribution of S. similis and its importance in oceanic ecosystems of the Pacific have been discussed by Pearcy and Forss (1969), Omori et al. (1972), and Omori and Gluck (1979). The purpose of this paper is to describe and illustrate the larval development of S. similis and to compare the larvae with those of the closely related species S. arcticus described by Wasserloos (1908), Hansen (1922), and Gurney •Scripps Institution of Oceanography, University of Cali- fornia, La Jolla, CA 92093. 2 Research Laboratory of fisheries resources. Tokyo University of Fisheries, Konan, Minato-ku, Tokyo 108, Japan. Manuscript accedpted October 1981. FISHERY BULLETIN: VOL. 80, NO. 2, 1982. and Lebour (1940). The larvae of S. similis are also compared with the early stages of S. atlanticus and S. comutus (Gurney and Lebour 1940), which comprise the atlanticus group, to note the difference in variation within species groups in protozoeal morphology, and with the larvae of Sergia lucens (Omori 1969) to note the differences between species of closely related genera. The description of Sergestes similis is based on both individuals reared in the laboratory by Omori (1979) during his study of the growth, feeding, and mortality of larval and postlarval stages of the species off southern California, and on specimens from preserved plankton samples. Gurney and Lebour (1940), in the major work on larvae of the genus, remarked that "perhaps the most interesting feature of the development of Sergestes is the striking difference which exists between the larvae of the different species, while the adults are often separable with diffi- culty," and suggested that knowledge of the larvae, when complete, may give a better indica- tion of the relationships of species than adult morphology. METHODS Omori (1979) described the procedures used for rearing the larvae of S. similis in the lab- oratory. Larvae from the population of the 217 FISHERY BULLETIN: VOL. 80, NO. 2 species off the coast of southern California were obtained for study from preserved plankton samples taken on Scripps Institution of Ocean- ography Expedition X and CalCOFI Cruises 6904 and 6905 during April and May of 1964 and 1965. At least five individuals of each developmental stage were dissected in glycerine for study of appendages. Some specimens of each stage were prepared for study and dissection by digesting away all soft tissue in heated aqueous KOH and then staining with Chlorazol Black E. Drawings were prepared with the drawing attachment of a Wild M20 3 microscope. Measurements of reared and planktonic larvae of S. similis were compared by Omori (1979, table 6); the mean body lengths (with standard deviation in parentheses) of larval stages obtained at 14°C are repeated here by stage for convenience. The larvae were mea- sured along the midline from anterior margin of forehead to posterior margin of telson. The postnaupliar developmental phases have been named protozoea, zoea, and postlarva following Omori (1979), and the terminology of Gurney and Lebour (1940) has been followed in describing the armature of carapace. In the protozoeal phase the outgrowths of the carapace are referred to as processes with secondary spines and spinules, while in the zoeal and postlarval stages the outgrowths are called spines with secondary spinules. Segmentation of two of the appendages proved difficult to determine. The basal segmentation of the exopod of the second antenna in protozoeal stages I-III was not clear. In S. similis it appeared that there were incomplete sutures within segments 1 and 3, giving 12 outer margin and 10 inner margin sutures; we have numbered the segments along the inner margin. The articu- lation of coxa, basis, and endopod of the second maxilla in protozoeal and zoeal phases also proved confusing. We have followed Gurney (1942) in referring to the medial lobes as bifid endites of coxa and basis, and have assumed from the morphology of the postlarval appendage that the endopod consists of 5 segments, although the articulation of segment 1 and basis was not clear. In the description of larval stages, only changes in structure and armature of body and appendages are discussed; if an appendage is not 3 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. mentioned, it may be assumed that there has been no change from the preceding stage except increase in size. In order to compare the basic pattern of development between Sergestes similis and Sergia lucens, we reexamined a number of larvae and postlarvae of S. lucens from the rearing experiment in July 1965. RESULTS The larval development of Sergestes similis includes the following stages: nauplius I-IV, protozoea I-III, and zoea I-II. The first two postlarval stages are also described. Nauplius I (Fig. la, e) Body length: 0.34 mm (0.01). Body ovoid with two posterior spines which curve posterodorsally and are slightly swollen basally. Antennule (Fig. 2a) unsegmented with 4 smooth setae, 2 terminal and 2 subterminal, and small terminal spine. Antenna (Fig. 3a) unsegmented; exopod with 5 setae; endopod with 3 setae, 2 terminal and 1 sub- terminal; all setae smooth. Mandible (Fig. 4a) biramous and unseg- mented, each ramus with 3 smooth setae. Nauplius II (Fig. lb) Body length: 0.38 mm (0.01). Body slightly longer and narrower posteriorly than in stage I, with 2 pairs of spines on posterior margin, outer pair very short, tiny rudiments of third inner pair sometimes visible. Antennule (Fig. 2b) with 1 subterminal medio- ventral seta and 3 terminal processes including 2 setae with setules and 1 small aesthetasc. Antenna (Fig. 3b) unsegmented; exopod with 6 setae and sometimes with small distal spine, distolateral seta smooth and others plumose; endopod with 2 plumose setae and 1 small spine terminally. Mandible (Fig. 4b) with 3 plumose setae on each ramus. Nauplius III (Fig. lc, f) Body length: 0.42 mm (0.02). Body with posterior portion tapering, pos- terior margin slightly indented medially with 4 218 KNIGHT and OMORI: LARVAL DEVELOPMENT OF SERGESTES SIMILIS Figure 1. — Sergestes similis. Nauplius I-IV, a-d, dorsal view; nauplius I, III-IV, e-g. lateral view without appendages. pairs of spines, outer pair tiny, relatively long third pair armed with spinules and articulated basally. Antennule (Fig. 2c) sometimes with a second seta on inner margin, incipient segmentation, and few rows of tiny spinules. Antenna (Fig. 3c) with incipient segmentation of protopod and exopod sometimes visible; exopod with 7 setae and small distal spine, distal 2 setae with small setules, other setae plumose; endopod with 3 terminal setae and 1 seta on inner margin. 219 FISHERY BULLETIN: VOL. 80, NO. 2 1 J Figure 2.— Sergestes similis. Antennules: a-d, nauplius I-IV; e-g, protozoea I-III; h-i, zoea I-II; j, postlarva I; setules omitted on i and j. 220 KNICHT and OMORI: LARVAL DKVKLOl'MKNT OK Sh'RCh'STKS SIMILIS f h Figure S.—Sergestes similis. Antenna: a-d, nauplius I-IV; e, protozoea I;f-g,zoeaI-II; h, postlarval.tipof scale; setules omitted on g. 221 FISHERY BULLETIN: VOL. 80, NO. 2 FIGURE 4.— Sergestes similis. Mandibles: a-c, nauplius MI, IV; d-f, protozoea I-III; g-h, zoea MI; i, postlarva I. Labrum: j, protozoea I; k, zoea I. 222 KNICHT and OMORL LARVAL DKVKLOI'MKNT OF SERGESTES SIMILIS Mandible unchanged. Anlagen of maxillules, maxillae, and first and second maxillipeds visible. Nauplius IV (Fig. Id, g) Body length: 0.49 mm (0.03). Body with abdomen forming, posterior margin with distinct medial indentation and 4 pairs spines, third pair still relatively long, rudi- ments of fifth inner pair sometimes visible, spinules present on spines 2-4, and sometimes on 1; third pair articulated, other spines fused with telson. Antennule (Fig. 2d) with 2 inner setae, terminal setation unchanged; proximal two- thirds with indistinct segmentation most clearly visible along inner margin; about 17 rows of tiny spinules encircle antennule associated, in segmented section, with distal margin of seg- ment. Antenna (Fig. 3d) with protopod of 2 indistinct segments; exopod with approximately 8 seg- ments (basal segmentation unclear, specimens cleared and stained have indication of 10 seg- ments on outer margin and about 8 on inner margin), with 8 or 9 setae and sometimes a small distal spine, distal 3 setae with small setules, others plumose; endopod at least 2-segmented, small distinct distal segment with 4 terminal setae, proximal segment with 2 setae on outer margin and sometimes with incomplete basal segmentation; both rami encircled with rows of tiny spinules. Mandible (Fig. 4c) with basal portion swelling with development of gnathal lobe, tissue with- drawing from rami. Rudiments of maxillules, maxillae, and 2 pairs of maxillipeds present posterior to mandibles. Protozoea I (Fig. 5a, b) Body length: 0.82 mm (0.02). Carapace with following processes: 1 pair anterolateral, each branching to 3 large spines and occasionally 1-3 small spines (5 of 20 reared larvae with small spines on one or both processes, 20 larvae from the plankton with 3 large spines only); 1 pair lateral with 1-3 large basal spinules; 1 posterodorsal with few large basal spinules, usually 2; all processes with small spinules to tip. Anterior margin of forehead with pair of small papillae. Prominent, round dorsal organ present in protozoeal phase. Thorax with evidence of segmentation, abdomen unsegmented. Telson forked, each fork with 2 small smooth ventral spines and 4 long curving processes armed with spinules. Antennule (Fig. 2e) of 3 segments, proximal segment subdivided into 5 small segments; proximal and middle segments with 1 and 2 setae, distal segment with 8 processes including 5 setae, 3 terminal and 2 proximal, and 3 aesthetascs. Gurney (1942) noted that distal seg- ment with aesthetascs is homologous with outer flagellum of later stages and that peduncle is therefore of 2 segments. Antenna (Fig. 3e) with exopod of 10 segments, terminal segment with 3 setae, segments 2-9 with 1 distal seta on inner margin, segments 3 and 5 with 1 distal seta on outer margin as well; endopod 2-segmented, distal segment with 5 terminal setae, long proximal segment with 5 setae on inner margin— 3 distal and 2 proximal on slight protuberance; basis with 2 setae on inner margin; structure unchanged in proto- zoeal phase. Mandibles (Fig. 4d) without palp, gnathal lobe of each mandible with 1 strong serrated spine on cutting edge between incisor teeth and molar area. Labrum (Fig. 4j) with long anteroventral spine in protozoeal phase. Maxillule (Fig. 6a) with exopod a small round lobe bearing 4 plumose setae; endopod 3- segmented with 3-2-5 setae progressing distally; basal and coxal endites with 4 and 5 setae, re- spectively. Maxilla (Fig. 7a) with segmentation indistinct; exopod small and oblong with 5 long plumose setae; endopod 5-segmented with setation of 4-2- 2-2-3, segment 1 rarely with 3 setae; basal and coxal endites bifid, the 4 median lobes with 8-4-4- 4 setae. First maxilliped (Fig. 8a) with exopod of 1 seg- ment bearing 7 long plumose marginal setae; endopod 4-segmented with 3-2-2-5 setae; basis with 12 setae in groups of 3 along medial margin; coxa with 5 setae; inner margins with fine setules as well. Second maxilliped (Fig. 9a) with exopod of 1 segment bearing 6 marginal plumose setae; endopod 4-segmented with 2-1-2-5 setae; basis with 5 and coxa with 2 setae. Third maxilliped a small bud. Protozoea II (Fig. 10a, b) Body length: 1.21 mm (0.10). 223 FISHERY BULLETIN: VOL. 80. NO. 2 Figure 5. — Sergestes similis. Protozoea I: a, dorsal view; b, lateral view. Carapace with rostrum but without pair of anterolateral processes; all processes with relatively large spines which branch distally into several small spinules, the processes themselves do not branch distally but bear small spinules to tip; rostral process with 3 pairs of spines, each lateral process usually with 7, rarely 6 or 8, spines, and posterior process with 2-4, usually 3 or 4, pairs of spines. Eyes stalked and moveable with papilla on anterior margin of stalk. Thorax 224 KNICHT and OMORI: LARVAL DKVKLOl'MKNT OF SERGh'STKS SIMIUS Figure 6.— Sergestes similis. Maxillule: a, protozoea I; b, protozoea II, coxal and basal endites; c, protozoea III, basal endite; d, zoea I; e, postlarva I. with segments delineated; abdomen and telson as in preceding stage. Antennule (Fig. 2f) with 4 subdivisions of proximal segment and distal segment with 9 processes, including 5 setae and 4 aesthetascs. Mandibles (Fig. 4e) with median armature asymmetrical, right mandible with 2 and left mandible with 5 strong spines, the spine nearest molar area is strongest on each mandible, 2 long spines on right mandible separated by short tooth. Maxillule (Fig. 6b) with 6 setae on basal endite and 6 or 7, usually 7, setae on coxal endite. Maxilla (Fig. 7b) with setation of 8-4-5-5 on medial lobes. First maxilliped with 7 or 8, usually 8, setae on coxa. Second maxilliped with endopod setation of 2- 2-2-5; basis with 5 or 6, rarely 6, setae. Third maxilliped a small rudiment, some- times slightly bifid at tip. Anlagen of thoracic legs 1-5 may be visible. Trotozoea III (Fig. 11a, b) Body length: 1.90 mm (0.18). Carapace with 1 pair supraorbital processes which curve dorsolateral^ in addition to arma- ture of preceding stage; all processes but rostrum armed with strong spines which branch 225 FISHERY BULLETIN: VOL. 80. NO. 2 Figure T.—Sergestes similis. Maxilla: a, protozoea I; b, protozoea II, basal endite; c, zoea I; d, postlarva I. distally into spinules, all processes terminate in single spine and bear spinules to tip; supra- orbital processes each with 9-14, usually 10-12, spines; each lateral process with 5-8, usually 7, spines; and posterodorsal process with 7-13, 226 usually 10-12, spines. Eyestalks longer than in stage II. Abdomen with 5 segments articulated, segment 6 still fused with telson; segments 1-5 with 1 pair lateral spines, segment 6 with bira- mous, unsegmented, nonsetose uropods and 1 KNKJHT and OMORI: LARVAL DEVELOPMENT OF SEHCKSTKS SIMILIS d c Figure 8. — Sergestes simiiis. First maxilliped: a-b, protozoea I, III; c, zoea I; d, postlarva I; setules omitted on b and c. pair small ventolateral spines proximal to uropods; smooth ventral spines of telson rela- tively larger than in stage I. Antennule (Fig. 2g) with proximal of 3 seg- ments without subdivisions and segment 2 with 5 setae, otherwise setation unchanged. Mandibles (Fig. 4f) usually with 3 and 6 spines on right and left cutting edges, 1 of 10 larvae 227 FISHERY BULLETIN: VOL. 80, NO. 2 Figure 9.—Sergestes similis. Second maxilliped: a-b, protozoea I, III; c, zoea I; d, postlarva I; setules omitted on b. 228 KNICHT and OMORI: LARVAL DEVELOPMENT OF SERGESTES SIMILIS Figure 10.— Sergestes similis. Protozoea II: a, dorsal view; b, lateral view. 229 FISHERY BULLETIN: VOL. 80, NO. 2 Figure ll.—Sergestes sim ilia. Protozoea III: a, dorsal view; b, lateral view; c, third maxilliped and thoracic legs of late stage larva. 230 KNICHT and OMORI: LARVAL DF.VF.LOI'MKNT OF SKRCESTES SIMMS with armature of stage II, long spine nearest incisor teeth on right mandible separated from other long spines by several small teeth. Maxillule (Fig. 6c) with 7 setae on both basal and coxal endites; as in earlier stages distal stout seta on basal endite with long basal spinules, other stout setae with short spinules. Maxilla with setation of 9-5-6-5 on medial lobes, rarely with 8 setae on proximal lobe of coxal endite. First maxilliped (Fig. 8b) with 9 setae on exopod. Second maxilliped (Fig. 9b) with endopod setation of 3-2-2-5 and exopod with 8 setae; coxa with 1 or 2 setae. Third maxilliped and thoracic legs 1-5 (Figs, lie, 12a, d) biramous, unsegmented, and nonsetose with exopod slightly longer than endopod. Zoea I (Figs. 13, 14a) Body length: 3.25 mm (0.13). Carapace altered with change in phase, now with 10 spines including rostrum, 1 pair supra- orbital, 1 pair hepatic, 2 pairs lateral, and 1 posterodorsal, all spines armed only with spinules except rostrum which bears a strong basal dorsal spine with spinules; dorsal organ present in zoeal phase but smaller than in pre- ceding stages. Eyes with long slender stalk bearing single ventral papilla in both stages of phase. Abdomen of 6 segments with following arma- ture: segments 1-5 with 1 pair lateral spines which decrease in length posteriorly, segments 1-6 with 1 posterodorsal spine longest on segments 3-5, segment 6 with 1 pair small ventrolateral spines and segment 1 with 1 pair triangular dorsolateral processes; posterodorsal and lateral spines armed with spinules, lateral spines of segments 1 and 2 with relatively long spinules proximally on posterior margin, seg- ments with dorsal and lateral setae as figured. Telson (Fig. 15e) slender with 1 pair lateral spines on rounded margin and produced distally into 2 long slender spines which bear 4 spinules near one-third their length and tiny spinules distally. Antennule (Fig. 2h) with peduncle unseg- mented and with following armature: basal lateral spine; 13-16, usually 15, long plumose setae along inner, outer, and distal ventral margins; smaller setae distributed near basal spine and along dorsal surface of peduncle in clusters of 2-2-3-1-3. Flagella unsegmented; outer flagellum with 3 small spines and 1 seta distally, and dorsal tier of 6 aesthetascs near two- thirds the length of flagellum; inner ramus very small with 2 terminal spines. Antenna (Fig. 3f) with scale (exopod) slender bearing 1 small subterminal ventral seta, 1 subterminal seta on outer margin, and 10 or 11, usually 11, setae on distal and inner margins, all setae with setules, terminal setae relatively stout and graded in size from short lateral to long medial seta; flagellum (endopod) with about 8 segments, proximal segment about the length of scale, distal segment with 4 terminal spines, 1 seta projecting laterally from each side, and 1 seta directed anteriorly; a strong spine with basal spinule appears on inner margin of flagellum before midpoint of segment 1 and dis- tally on segment 6. Mandibles (Fig. 4g) with 4 and 7 relatively long spines on right and left blades between incisor teeth and molar surfaces, bud of palp present. Labrum (Fig. 4k) with anteroventral spine present but shorter than in preceding phase. Maxillule (Fig. 6d) with 11 setae on basal endite. Maxilla (Fig. 7c) with exopod modified bear- ing 1 long plumose seta on proximal lobe and 4 small processes approximately in position of plumose setae of preceding stage; endopod un- changed; medial lobes with setation of 9-5-6-6. First maxilliped (Fig. 8c) with form as in protozoeal phase; exopod with 13 marginal setae; endopod with setation of 4-3-2-5; basis with 13 and coxa with 8 setae. Second maxilliped (Fig. 9c) somewhat modi- fied, long flexible exopod with 7 or 8, rarely 8, setae and resembling exopod of thoracic leg rather than form of preceding phase; endopod 4- segmented with 3-0-2-5 setae; basis with 9 and coxa with 2-4 setae. Third maxilliped (Fig. 12b) functional and pediform; exopod with 19-23, usually 21, setae; endopod 4-segmented, usually with setation of 3- 4-4-5, rarely with 5 setae on segment 2; basis with 4 setae, coxa nonsetose. Legs 1-5 functional; legs 1-3 (Fig. 12e) similar, shorter than third maxilliped; exopods with 20- 22, frequently 21, setae; endopods 4-segmented with 3-4-4-4 setae and bases with 3 or 4 setae. Legs 4 and 5 (Fig. 16c, d) smaller than first three pairs; exopods with 17-19 setae; endopods 231 FISHERY BULLETIN: VOL. 80, NO. 2 O.I mm I 1 Figure 12.— Sergestes similis. Third maxilliped: a, protozoea III; b, zoea I; c, postlarva I. Leg 1: d, protozoea III; e, zoea I; f, postlarva I. 232 KNIOHT and OMORI: LARVAL DKVKLOl'M KNT OF SKRdh'STKS SIM I LIS Figure 13.— Sergestes similis. Zoea I, dorsal view. 233 FISHERY BULLETIN: VOL. 80, NO. 2 Figure 14.— Sergestes similis. Abdomen, lateral view: a, zoea I; b, zoea II. 3-segmented with 3-3-4 setae on leg 4, rarely 4 setae on segment 2, and 2-3-4 setae on leg 5; bases with 1 seta, coxae nonsetose. Pleopods (Fig. 15a) present on abdominal seg- ments 1-5 and variable in size within stage; exopods nonsetose decreasing in length from pleopod 1 to 5; pleopod 5 with nonsetose endopod about two-thirds length of exopod, pleopod 4 with small bud of endopod, pleopods 2 and 3 sometimes with some swelling in position of endopod. Uropods with rami articulated; protopod with lateral spine and posterior projection (Fig. 14a); exopod and endopod long, slender, and fringed with plumose setae except proximal to smooth spine on lateral margin of exopod. Zoea II (Figs. 14b, 17) Body length: 4.42 mm (0.20). Carapace, abdomen, and telson (Fig. 15f) with armature as in preceding stage; spines shorter relative to size of larva and lateral spines of abdominal segments 1 and 2 without long pos- terior spinules. Antennule (Fig. 2i) with peduncle bearing 16- 19, usually 17 or 18, marginal plumose setae and small setae in clusters of 3-4-3-1-3; outer flagellum with 1 distal seta and usually unseg- mented, sometimes constricted at two points distal to tier of 6 aesthetascs, rarely with weak sutures; inner ramus without spines. Antenna (Fig. 3g) with scale armed with long subterminal spine on outer margin bearing spinules, a small subterminal ventral seta, and 14 or 15 marginal plumose setae, terminal setae no longer stout; flagellum long, with 19-25 seg- ments in three reared larvae, terminal segment with 2 spines and 3 setae, 1 seta projects laterally from each side. Mandibles (Fig. 4h) with armature un- changed; palp larger than in zoea I, unseg- mented and nonsetose. Labrum with short remnant of anteroventral spine. 234 KNIGHT and OMOKI: LARVAL PKVKLOI'MKNT OF SKKUKSTKS SI.MIIJS e f 9 Figure 15.— Sergestes similis. Pleopods: a-b, zoea I-II; c-d, postlarva I-II. Telson: e-f, zoea I-II; g-h, postlarva I-II. Maxillule with 12 or 13, usually 12, setae on basal endite and 8 or 9 setae on coxal endite. Maxilla unchanged except that exopod relatively larger with small processes now tiny. First maxilliped with 13 or 14, usually 13, setae on exopod. Second maxilliped with endopod setation of 4- 2-3-5, rarely 5 and 4 setae on segments 1 and 3; exopod with 7 setae; basis with 9 or 10 and coxa with 3 or 4 setae. Third maxilliped with 4 or 5, rarely 4, setae on distal segment of endopod and 3 setae on basis. Legs 1-3 with endopod slightly longer than exopod, legs 2 and 3 with distal margin of 235 FISHERY BULLETIN: VOL. 80. NO. 2 endopod segment 3 swelling in formation of chela (Fig. 16a); exopods with 20-24, usually 22, setae; endopods usually with setation of 3-4-5-4, rarely 5 and 4 setae on segments 2 and 3; bases with 3 or 4 setae. Leg 4 exopod with 18- 21 setae, endopod usually with setation of 3-3-4, rarely with 2 and 4 setae on segments 1 and 2. Leg 5 exopod with 16-19 setae, endopod usually with setation of 2-3-4, rarely with 3 setae on segment 1. Pleopods (Fig. 15b) nonsetose but longer than in preceding stage, exopods again decreasing in length from anterior to posterior pairs; pleopods 2-5 with endopod which increases in size posteriorly with variation in size within stage with age. 0. 1 mm I 1 c-d,f Figure 16.— Sergestes similis. Leg 2: a, zoea II; b, postlarva I. Leg 4: c, zoea I. Leg 5: d, zoea I. Legs 4 and 5: e-f, postlarva HI. 236 KNIGHT and OMORI: LARVAL DEVELOPMENT OF Sk'RHhSTh'S SIMIUS Figure n.—Sergestes similis. Zoea II, dorsal view. 237 FISHERY BULLETIN: VOL. 80. NO. 2 Postlarva I (Fig. 18a) Body length: 5.07 mm (0.26). Carapace with armature reduced, 2 pairs of lateral spines of preceding phase missing or only tiny remnants; rostrum, supraorbital, and hepatic spines, and posterodorsal spine shorter relative to length of carapace. Abdomen with lateral spines of segments 1-5 and posterodorsal spines of segments 1 and 2 small and without spinules. Telson (Fig. 15g) with posterior fork spines much shorter in relation to body of telson than in zoeal phase, with spinules reduced and with pair of plumose setae on inner margin near base of fork; relative length of terminal setae and fork spines vary within stage. Antennule (Fig. 2j) with peduncle 3-seg- mented and fringed with marginal plumose setae, basal segment with statocyst and lateral spine; rows of small setae now situated on distal margins of segments; outer flagellum with 10 segments, proximal segments 1-3 with 1-2-6 medioventral aesthetascs, 6 aesthetascs of seg- ment 3 set proximally on small protuberance; inner flagellum with 2 segments. Antenna (Fig. 3h) with subterminal lateral spine of scale smaller than in zoea II, scale with 20-22 marginal setae; flagellum very long, about 2.6 times body length in one reared larva. Mandibles (Fig. 4i) with cutting edges smooth between simplified incisor and molar processes, left mandible with notch opposing incisor tooth of right mandible; palp with 5-7 setae and some- times indistinctly 2-segmented. Labrum without spine. Maxillule (Fig. 6e) with endopod reduced to small nonsetose rudiment and with tiny vestige of exopod, basal and coxal endites with increased numbers of setae. Maxilla (Fig. 7d) with endopod reduced to small nonsetose rudiment; scaphognathite (exopod) large with 17 or 18 marginal setae, 1 posterior seta relatively long; coxal and basal endites bifid, medial lobes with 2-2-4-4 to 6 setae. First maxilliped (Fig. 8d) with small non- setose exopod and endopod, coxa with medial setae and small epipodite, basis with broad flat medial lobe armed with setae along inner margin. Second and third maxillipeds and legs 1-3 with small nonsetose remnant of exopod. Second maxilliped (Fig. 9d) with endopod long, 5-segmented, recurved at articulation of merus and carpus, and armed with strong setae, articulation of ischium and basis indistinct if visible; coxa with bud of epipodite. Third maxilliped (Fig. 12c) with endopod long, 5-segmented, and with strong marginal setae. Leg 1 (Fig. 12f) with clusters of strong barbed setae at articulation of propodus and carpus, legs 2 (Fig. 16b) and 3 with small setose chela, leg 2 with small spine on lateral margin of ischium. Leg 3 slightly longer than first maxilliped. Legs 4 and 5 (Fig. 16e) reduced to irregular, nonsetose bifid rudiments. Pleopods (Fig. 15c) with setose exopods; endopod of pleopod 5 setose, rarely endopod 4 with 1 or 2 terminal setae, as before endopods increase and exopods decrease in length from anterior to posterior pairs; protopod with 1 distal seta on inner margin of pleopods 1-3; endopods vary in size within stage. Postlarva II (Fig. 18b) Body length: 5.80 mm (0.20). Carapace and abdomen with armature reduced in size, small posterodorsal spine of carapace may be missing and dorsal spines of abdomen segments 1 and 2 very small. Telson (Fig. 15h) with posterolateral spines reduced in length and usually with 3 pairs plumose lateral setae in addition to terminal pair. Antennule with outer flagellum, in exuvia of one reared larva, with about 17 segments and in- creased number of aesthetascs on proximal seg- ments; inner flagellum with 2 or 3 segments. Antenna with subterminal lateral spine of scale smooth or with few spinules and reduced in length, scale with 24-28 marginal setae. Mandibles with palp 2-segmented bearing 11- 14, usually 11 or 12, setae. Maxillule with endopod more distinctly shaped; vestige of exopod not present. Maxilla with 25-27 setae on scaphognathite; endopod larger than in preceding stage with outer basal seta and sometimes inner seta; medial lobes with setation of 2-3-5 to 8-8 or 9. First maxilliped with endopod, exopod, and epipodite larger than in postlarva I, rarely endopod slightly longer than exopod with some indication of segmentation. Second and third maxillipeds and legs 1-3 without vestige of exopod; legs 1 and 2 with small ischial spine; legs 4 and 5 (Fig. 16f) more distinctly formed, nonsetose, and with leg 4 longer than leg 5. Pleopods 3-5 (Fig. 15d) with setose endopods, 238 KNIGHT and OMORI: LARVAL DKVKLOPMKNT OK SlCUdKSTKS SIMIUS Figure 18.— Sergestes similis. a, postlarva I; b, postlarva II. 239 FISHERY BULLETIN: VOL. 80, NO. 2 rarely endopod of pleopod 2 with 1 or 2 small setae; protopods of pleopods 1-3 with 2 setae; those of pleopods 4 and 5 with or without 1 seta on protopod. DISCUSSION Yaldwyn (1957) defined two subgenera, Sergestes s.s. and Sergia, within what he termed the rather unwieldy genus Segestes s.l. Recently, the subgenera were raised to full genera by Omori (1974). The species of Sergestes have specialized luminescent modifications of the gastrohepatic gland called organs of Pesta and are without cuticular pigmentation and dermal photophores, while species of Sergia are without organs of Pesta and, with some exceptions, have uniform cuticular pigmentation and often dermal photophores. The two genera are themselves divided into species groups, six in Sergestes and three in Sergia, which appear to be easily distinguished and are considered to be natural phyletic units (Judkins 1978). The arcticus group includes only two species, Sergestes arcticus and S. similis, and is characterized by the morphology of third maxilliped, fifth pereiopod, antennular pedun- cle, and petasma (Yaldwyn 1957). Sergestes similis differs from S. arcticus in having a more slender and fragile body and antennular peduncle, in a longer and more upwardly directed rostrum, in proportions of posterior arthrobranchs above the third and fourth pereiopods, and in some proportions and armature of petasma and thelycum (Milne 1968). The close relationship of S. similis and S. arcticus which has been inferred from adult morphology may also be seen in their larval morphology, especially in the shape of eye, in the ornate armature of protozoeal carapace, and in the armature of carapace, abdomen, and telson in the zoeal phase. Gurney and Lebour (1940) described larvae now known to be representative of all of the species groups within Sergestes s.l. and noted that the protozoea II and III of S. arcticus were very distinct in form of eye and peculiarly branched spines. They described some features of protozoea II and III, zoea I and II, and postlarva I of S. arcticus and gave figures of the second protozoea and zoea, with telson of postlarva I. They stated that the "brushlike endings" of the long spines on rostral, lateral, and posterior processes of protozoea II and on supraorbital, lateral, and posterior processes of protozoea III were most characteristic of the species. The protozoea II and III of S. similis, identified in this study, have the same distinctive armature of carapace spines. The larval stages of S. arcticus discussed by Gurney and Lebour (1940) resemble the compar- able stages of S. similis in the details they described and figured. Gurney and Lebour, however, do not deal with the structure of mouth- parts and thoracic appendages; rather, they note that these appendages seem to be uniform throughout the genus and refer the reader to the earlier descriptions of S. arcticus by Wasserloos (1908) and Hansen (1922). Gurney, in a later work (1942), does figure the appendages of protozoea III of S. cornutus, an atlanticus group species, and they appear very similar to those of the same stage of S. similis. The protozoeal stages of S. arcticus described by Wasserloos (1908), on the other hand, differ from those of S. similis in setation and/or segmentation of antennule, antenna, and mouthparts, but appendages are not figured; the armature of carapace differs in protozoea II and III on lateral and supraorbital processes, re- spectively. The species appear similar in described and figured features of the zoeal phase. Hansen (1922) offered a brief summary of Wasserloos' description of the protozoeal phase and added both generic and specific comments, with figures, on the zoea and postlarva of S. arcticus from his own observations. He noted that the mouthparts of the protozoea are like those of the zoeal stages which he described in some detail but which do not always agree with details of the protozoeal phase described by Wasserloos. Hansen also noted that the rostrum in protozoea III is little modified from stage II, yet conspicuous secondary spines are lost in this molt. In the zoeal phase, S. similis larvae differ from those of S. arcticus, as described by Hansen, in segmentation of maxillule and first maxil- liped. Unfortunately, because the descriptions of S. arcticus by Wasserloos (1908) and Hansen (1922) were found to be inconsistent with each other and with that of Gurney and Lebour (1940), and they could not be interpreted with confidence, a detailed comparison with S. similis was not possible. A reexamination of the larval stages of S. arcticus is needed to detect specific differences that may exist between the apparently very similar arcticus group species. 240 KNIGHT and OMORI: LARVAL DEVELOPMENT OF Sh'RCKSTh'S SIMIL1S Gurney and Lebour (1940) believed the elaborate protozoeal phase of Sergestes s.l. to be of particular importance, as it might "point to a satisfactory subgeneric grouping of the adults." They separated the second and third protozoeae of thirteen species of Sergestes s.l., representative of all of the nine species groups later defined by Yaldwyn (1957), into three types: dohrni, ortmanni, and hispida. The carapace has the same number of processes in all three types but the armature of the processes differs as follows: dohrn i type with numerous long lateral spines on supraorbital, lateral, and posterior processes; ortmanni type with long spines on supraorbital processes but with long spines only at the bases of lateral and posterior processes on carapace; h ispida type without long spines on supraorbital, lateral, or posterior processes, although there may be basal spines on lateral and posterior processes. Gurney and Lebour observed that the ortmanni armature seems to be derived from the dohrni type in that it retains long lateral spines on supraorbital processes. These larval types do correspond with three divisions of species within Sergestes s.l. Of the species described by Gurney and Lebour (1940), the hispida type larvae all belong to the genus Sergia, while the dohrni and ortmanni types belong to Sergestes; S. corniculum is of the ortmanni type, but all other species of Sergestes identified are of the dohrni type. The zoeal stages could not be separated into groups which corresponded to the protozoeal types. Gurney and Lebour (1940) noted that the dohrni type carapace was found in a number of species which were not supposed to be particu- larly closely related and which could not be grouped further by structure of the protozoeal phase. The identification of S. similis larvae has proved this untrue with respect to the arcticus group species, but apparently it does apply to species of the atlanticus group, the only other species group within Sergestes, or Sergia, all of whose protozoeal stages are identified. Gurney and Lebour described the larvae of Sergestes atlanticus and S. cornutus, the two species which comprise the atlanticus group, and observed that larval morphology did not corroborate the close relationship implied by the morphology of adult petasma. The carapace armature in protozoea II and III of the arcticus and atlanticus groups is compared in Table 1 to show the range of variation within each group; the species groups themselves are not considered to be closely re- lated within the genus (Judkins 1972). Sergestes arcticus and S. similis may have the same armature in both protozoeal stages, while S. atlanticus and S. cornutus differ in each stage; all of the lateral spines of the atlanticus group have smooth tips rather than the brushlike endings characteristic of the arcticus group. The difference in larval morphology within the atlanticus group is in accordance with the signi- ficant difference described by Foxton (1972) between S. atlanticus and S. cornutus in morphology of the organs of Pesta. This dis- crepancy was one of two exceptions noted by Foxton to a generalization that species of Sergestes that are the most similar in other adult diagnostic characters usually have identical or closely similar organs of Pesta; he does not note any difference between the arcticus group Table 1.— Comparison of the number of long lateral spines which arm carapace processes in protozoea II and III of two species groups of Sergestes; the lateral spines have smooth tips in the atlanticus group and branching tips ("brushlike endings") in the arcticus group (descriptions of the atlanticus group and S. arcticus are taken from Gurney and Lebour (1940). Carapace articus group atlanticus g roup processes S. similis S arcticus S. atlanticus S cornutus Protozoea II Rostrum 6 in 3 pairs as similis 7 rather irregularly arranged 8 in 4 pairs + 2 ventral Lateral, each 6-8, usually 7 '7 9 8 Posterior 4-8, with 3 large pairs 6 in 3 pairs 6, process swollen basally 4 in 2 pairs Protozoea III: Rostrum with spinules only as similis 3 ventral 7 ventral Supraorbital, each 9-14. usually 10- 9. orientation as ca 15; processes curve 12-19; processes direct- 12; processes in similis inward to meet and ed anterolateral^ curve postero- overlap lateral^ Lateral, each 5-8, usually 7 7 17-20 12-14 Posterior 7-13. usually 10- 12 in 5-6 pairs 10 in 5 pairs 16 arranged in circle on large basal swelling 8 in 4 pairs 'Gurney and Lebour (1940) report eight long spines on the lateral carapace process, but their figure shows seven with brushlike endings and the simple spmulose tip of process, the common armature in S. similis 241 species in morphology of this feature. The cor- respondence between variation in morphology of protozoeal stages and organs of Pesta within the two species groups indicates that, with identifi- cation of additional species, larval morphology may prove useful in the study of interspecific re- lationships within Sergestes, as predicted by Gurney and Lebour (1940). The larvae of S. similis were also compared with those of hispida type Sergia lucens (Omori 1969), one of the seven species comprising the challengeri group. They were found to differ in body armature, as expected from difference in protozeal type, in form of telson, and in develop- ment of some appendages, as shown in Table 2. They also differ in number of naupliar stages. Four distinct stages were observed in the naupliar phase of Sergestes similis, while in Sergia lucens nauplius I and II were found and the latter developed gradually to molt into protozoea I. When this finding is coupled with the observations by Nakazawa (1916, 1932), they suggest that there are zero to two molts during the naupliar phase of S. lucens. An assessment of the significance of these observations requires additional knowledge of larval development within the two closely related genera and their species groups. FISHERY BULLETIN: VOL. 80, NO. 2 ACKNOWLEDGMENTS This work was supported by the Marine Life Research Program, the Scripps Institution of Oceanography's component of the California Cooperative Oceanic Fisheries Investigations. We are very grateful to Kuni Hulsemann and her colleagues for a translation of Wasserloos (1908), to A. Fleminger for the method and materials to treat larvae with KOH and Chlorazol Black E, and to E. Brinton for criticism of the manuscript. LITERATURE CITED FOXTON. P. 1972. Further evidence of the taxonomic importance of the organs of Pesta in the genus Sergestes (Natantia, Penaeidea). Crustaceana 22:181-189. Gurney, R. 1942. Larvae of decapod Crustacea. RaySoc. Publ. 129, 306 p. Ray Soc, Lond. Gurney, R., and M. V. Lebour. 1940. Larvae of decapod Crustacea. Part VI; The genus Sergestes. Discovery Rep. 20:1-68. Hansen, H. J. 1922. Crustaces decapodes (Sergestides) provenant des campagnes des yachts "Hirondelle" et "Princess Alice" (1885-1915). Resultats des Campagnes Scien- tifiques Accomplis sur son Yacht, par Albert I" 64:1- 232. Table 2.— Some differences between larvae of Sergestes similis and Sergia lucens. Feature Sergestes similis Sergia lucens Carapace armature: Protozoea I Protozoea II Protozoea III Zoea l-ll Postlarva I Abdomen armature; Zoea l-ll Telson: Zoea l-ll Postlarva I Antennule: Zoea l-ll Antenna: Zoea I Mandible: First maxilliped: Zoea l-ll anterolateral process branches to 3 spines posterodorsal process a single spine with basal spinules all processes with long spines which branch to spin- ules distally rostrum with small spinules, armature ot other pro- cesses as in II with 2 pairs lateral spines lateral spines remnants only, other spines present lateral spines decrease in length posteriorly, spines 1 and 2 with relatively long spinules in I fork with 2 outer and 2 inner spinules, invagination does not reach lateral spines fork relatively wide with tiny spinules outer flagellum unsegmented in rarely 2- or 3-seg- mented in II and shorter than peduncle endopod 8-segmented and longer than rostrum palp appears in zoea I exopod with 13 or 14 setae anterolateral process branches to 4 spines posterodorsal process branches to 3 spines all processes with small spinules only as in II with 3 pairs lateral spines rostrum and small posterodorsal spine present; supra- orbital spine and basal spine of rostrum sometimes present, lateral and hepatic spines missing lateral spines increase in length posteriorly, without long spinules in I fork with 1 outer and 5/6 inner spinules in I, with 2 outer spinules in II, invagination about as deep as lateral spines fork narrow, with 2 large inner setae outer flagellum 2-segmented in I, ca. 8-segmented in II, and longer than peduncle endopod 2-segmented and shorter than rostrum palp appears in zoea II, rarely in zoea I exopod with 12 setae 242 KNIGHT and OMORI: LARVAL DEVKLOl'MKNT OF SKRdKSTh'S SIMILIS JUDKINS, D. C. 1972. A revision of the decapod crustacean genus Sergestes (Natantia, Penaeidea) sensu lain, with emphasis on the systematics and geographical distribu- tion of Neosergestes, new genus. Ph.D. Thesis, Univ. California, San Diego, 274 p. 1978. Pelagic shrimps of the Sergestes edwardsii species group (Crustacea: Decapoda: Sergestidae). Smithson. Contrib. Zool. 256, 34 p. Milne, D. S. 1968. Sergestes similis Hansen and S. consobrinus n. sp. (Decapoda) from the northeastern Pacific. Crusta- ceana 14:21-34. Nakazawa, K. 1916. [On the development of Sakura-ebi.] [In Jpn.] Zool. Mag. Tokyo 28:485-494. 1932. [On the metamorphosis of Sakura-ebi.] [In Jpn.] Zool. Mag. Tokyo 44:21-23. Omorl M. 1969. The biology of a sergestid shrimp Sergestes lueens Hansen. Bull. Ocean Res. Inst. Univ. Tokyo 4, 83 p. 1974. The biology of pelagic shrimps in the ocean. Adv. Mar. Biol. 12:233-324. 1979. Growth, feeding, and mortality of larval and early postlarval stages of the oceanic shrimp Sergestes sim His Hansen. Limnol. Oceanogr. 24:273-288. Omorl M., and D. Gluck. 1979. Life history and vertical migration of the pelagic shrimp Sergestes similis off the southern California coast. Fish. Bull., U.S. 77:183-198. Omorl M., A., Kawamura, and Y. Aizawa. 1972. Sergestes similis Hansen, its distribution and im- portance as food of fin and sei whales in the North Pacific Ocean. In A. Y. Takenouti (chief editor), Biological oceanography of the northern Pacific Ocean, p. 373-391. Idemitsu Shoten, Tokyo. Pearcy, W. G., and C. A. Forss. 1969. The oceanic shrimp Sergestes similis off the Oregon coast. Limnol. Oceanogr. 14:755-765. Wasserloos, E. 1908. Zur Kenntnis der Metamorphose von Sergestes arc- ticus Kr. Zool. Anz. 33:303-331. Yaldwyn, J. C. 1957. Deep-water Crustacea of the genus Sergestes (Decapoda, Natantia) from Cook Strait, New Zealand. Zool. Publ. Victoria Univ., Wellington 22, 27 p. 243 GROWTH OF JUVENILE ENGLISH SOLE, PAROPHRYS VETULUS, IN ESTUARINE AND OPEN COASTAL NURSERY GROUNDS Andrew A. Rosenberg 1 ABSTRACT The growth of English sole juveniles, during 1978-79, from estuarine and open coastal nursery grounds on the Oregon coast is described in detail. Counts of fortnightly growth rings on otoliths were used to determine size-at-age. Mean growth rates were similar for the two areas, but variabil- ity in size-at-age was much greater among fish captured in the estuary. Back calculation of individual growth, using radial measurements on the otoliths, showed that growth proceeds linearly during the first year of life. Differences in average growth among indi- vidual fish account for the high variability in size-at-age among fish found in the estuary. Fish from the estuary grew slightly faster, on average, in 1979 compared with 1978. The settlement date of English sole larvae to the benthic habitat, determined from age data, occurred over the winter and spring in the open coastal nursery area. In the estuary, settlement was concentrated in the early winter. The life cycle of many marine fishes contains a stage in which the juveniles of the species are concentrated in a specific area or nursery ground where the adults are uncommon. On both the east and west coasts of North America, es- tuarine areas are extensively used as nursery grounds for a large number of species (Gunter 1961; Pearcy 1962; McHugh 1967; Haedrich in press). Many east coast fishes are considered to be dependent on estuarines during early life. On the west coast, estuarine dependence has not been clearly demonstrated (McHugh 1967; Pearcy and Myers 1974). The high productivity of estuarine areas, pro- viding improved growth conditions for juvenile fish, the apparent lack of large predators, and re- duction of competition among age groups of a species are frequently invoked explanations for estuarine dependence (Haedrich in press; Kuipers 1977). Unfortunately, it is difficult to test these hypotheses for most species of fish, be- cause it is uncommon to find a species which uses both estuarine and nonestuarine nursery envi- ronments in a small geographic area. The commercially important pleuronectid Parophrys vetulus Girard, found off the Oregon coast, uses both estuarine and nonestuarine habi- tats as nursery areas during the first year of life (Laroche and Holton 1979). This study examines the growth of the English sole, Parophrys vetulus, •School of Oceanography, Oregon State University, Cor- vallis, Oreg.; present address: Department of Biology, Dal- housie University, Halifax, Nova Scotia, B3H 4J1 Canada. juveniles from two nursery grounds: the Ya- quina Bay estuary (Pearcy and Myers 1974) and the open coastal area off Moolach Beach, Oreg. (Laroche and Holton 1979). Size-at-age data, obtained from daily and fort- nightly growth ring counts on otoliths, are used to detail growth during the first year. Daily growth rings on otoliths have been documented in many species of fish (Pannella 1971; Brothers et al. 1976; Struhsaker and Uchiyama 1976; Tau- bert and Coble 1977). Pannella (1974) reported fortnightly banding patterns in several species as well. Laroche et al. (1982) have provided lab- oratory evidence for the daily periodicity of growth rings on P. vetulus otoliths. METHODS Sampling was conducted from September 1978 through September 1979 at Moolach Beach and Yaquina Bay. The sampling stations are shown in Figure 1. A tow was made at each station with a 1.5 m wide beam trawl lined with 7 mm stretch mesh. Tows were for 5 min in Yaquina Bay and for 10 min at Moolach Beach. The beam trawl was equipped with a 1.0 m circumference odom- eter wheel to measure distance travelled on the bottom. Measurements of bottom temperature and salinity were made at each station. All fish captured were preserved in a strongly buffered 10% solution of Formalin 2 in seawater. Manuscript accepted November 1981. FISHERY BULLETIN: VOL. 80, NO. 2, 1982. 2 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 245 FISHERY BULLETIN: VOL. 80, NO. 2 In the laboratory, all fish were identified and measured for standard length (SL). Both saccu- lar otoliths were removed from each English sole. In cases where large numbers of P. vetulus were captured, individuals were selected to cover the size range of the sample. The otoliths were mounted on microscope slides in the syn- thetic mounting medium Protexx. One otolith from each fish was ground on 600 grit carborundum paper to a thin section along a sagittal plane through the nucleus. The sections were examined under 250X magnification, using either bright-field or polarized illumination. Counts of fortnightly rings were made on each otolith. No fortnightly rings could be detected in the central area of the otoliths, which apparently represents the time the larvae are in the plank- ton. Therefore, daily rings were counted from the nucleus out to the first fortnightly ring. The actual age of each fish was calculated by sum- FlGURE 1.— The study area. Sam- pling stations are indicated by the letters A through G. ming the number of daily rings in the nuclear area, the number of fortnightly rings times 14, and the mean age of first ring formation, which was taken to be 5 d for this species (Larocheetal. 1982). The count of rings on each otolith was repeated until the same count was obtained three times. As a further check on the accuracy of the counts, a set of 42 otoliths was recounted several months later and a mean error computed. Counts of the number of daily rings between fortnightly rings on 40 otoliths and the number of fortnightly rings between consecutive annual rings on 15 otoliths from older specimens were made as tests of fortnightly periodicity. Individual growth curves of 25 fish were back calculated by making radial measurements to every other fortnightly ring along the same axis from the nucleus to the anterior edge of the oto- lith. From these measurements and the linear relationship between otolith radius and standard length of the fish, 3 lengths-at-age for the various points in the life of an individual were calculated. RESULTS Counts of daily rings between fortnightly rings yielded a mean of 13.95 with a standard devia- tion of 0.68. The mean number of fortnightly rings between consecutive annual rings was 26 with a standard deviation of 1.13. The mean dif- ference between repeated counts of fortnightly rings made a substantial period of time apart was 1.45 rings. Figure 2 shows the daily and fort- 3 A regression of standard length on anterior otolith radius was performed on 60 data points. The resulting equation was: F = 0.86x + 4.5, where Y is standard length in mm and jc is the distance from the nucleus to the anterior edge of the otolith in arbitrary units, r 2 for this regression is 0.98. 246 ROSENBERG: GROWTH OF JUVENILE ENGLISH SOLE ■98 mm Figure 2.— An otolith from a 110 mm SL Parophrys vetulus captured in Yaquina Bay. Arrows indicate clear fortnightly rings. There are 21 fortnightly rings on this otolith. The actual age was calculated to be 363 d (see text). nightly patterns of a P. vetulus otolith. The first fortnightly ring is formed consistently when the fish is 60 to 75 d old, i.e., the beginning of the metamorphic period (Rosenberg and Laroche 1982). Basic growth data for the two nursery areas were obtained from size-at-age information. The data for 218 fish captured at Moolach Beach (Fig. 3) show that there are two linear portions of the data, with different slopes, separated by an 247 FISHERY BULLETIN: VOL. 80, NO. 2 Figure 3.— Size-at-age data for Pa- rophrys vetulus captured in the Moo- lach Beach nursery area. 20 40 60 80 100 120 140 160 160 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 Age (days) inflection point. There is no evidence of an upper asymptote in the data, so the use of growth models such as the Gompertz or von Bertalanffy equations is inappropriate. A least squares mul- tiple regression on these data was performed using the following model: Y = Bo + BiX + B 2 Ai + B 3 A 2 + E (1) where Y is the standard length in millimeters, X is the age in days, A\ is a dummy variable whose value is zero to the left of the inflection point and one to the right of the inflection point, and A 2 is equal to X times A\, i.e., the interaction term. The B terms are the regression coefficients and E indicates the error terms. The point of inflec- tion which produced the smallest residual sum of squares was found to be 140 d for the Moolach Beach data. The fitted equation is: Y = 16.87 + 0.051X - 32.92A + 0.2SA 2 . An analysis of variance for the regression (Table 1A) shows that a good fit was obtained with this model, and the data set has a relatively low esti- mated variance. The slopes of the regression be- low and above 140 d were computed as 0.051 and 0.279, respectively. These slopes are estimates of the mean growth rate per day for juvenile P. vetulus utilizing the Moolach Beach nursery area. The lower portion of the data, below 140 d of age, shows a plateau in growth attributed to the metamorphic period (Rosenberg and Laroche 1982). Regression of the size-at-age data for Yaquina Bay juveniles (Fig. 4: 186 data points) yields the fitted equation: Y = 13.01 + 0.083X - 33.45,4, + 0.201^ 2 . 248 The analysis of variance for this model (Table IB) once again shows that a good fit was obtained, but the estimated variance is much higher than for the Moolach Beach data. The inflection point with the smallest residual sum of squares was also 140 d of age for the Yaquina Bay data. The slopes below and above the inflection are 0.083 and 0.284, respectively. The first step in comparing the regression lines of growth for English sole from the two nursery grounds was to test for statistical equal- ity of variances. This was done by examination of the ratio of the mean square errors of the fitted regressions, 19.88 for the Moolach Beach data and 95.01 for the Yaquina Bay data. The ratio is distributed as F(184:216) and the variances are significantly different at the P = 0.001 level. Since the variances are unequal, statistical tests for equality of slopes or intercepts are not strictly valid (Scheffe 1959). However, the slopes are similar, 0.279 and 0.284. Back-calculated growth for individuals from both areas are in good agreement with growth Table 1.— Analysis of variance for the least squares multiple regression analysis of size-at-age data. A Moolach Beach regression Y = 16 87 + 0.01 5X - 32.92 A, + 023 A 2 Multiple R = 0.975 R 2 = 0.950 Source DF Sum of squares Mean square Regressior 3 80815.2 269384 Residual 214 42537 19.9 F value = 13550 B Yaquina Bay regression: Y = 1301 + 0083X - 33.45 A, + 0.201 A 2 Multiple R = 943 R 2 = 0.890 Source DF Sum of squares Mean square Regressior 3 1395756 46525.2 Residual 182 17308.1 95.1 F value = 489 3 ROSENBERG: GROWTH OF JUVENILE ENGLISH SOLE Figure 4.— Size-at-age data for Pa- rophrys vetulus captured in the Ya- quina Bay nursery area. I50T 140 130 120- 110- 100 90 80 70 60 50 40 30 20 10 YAQUINA BAY tzi- ' ' . %; 20 40 60 80 100 120 140 160 180 200 220 240 260 280 3O0 320 340 360 380 400 420 440 460 480 500 Age ( days ) estimates from the size-at-age data (Figs. 5, 6). The plots are, in general, linear. Slight changes in slope do occur in all the lines. This may indi- cate small variations in individual growth through the juvenile period, changes in the lin- ear nature of the relationship between otolith growth and overall fish growth, or measurement error. By inspection, these variations do not occur at coincident times among individuals. For the Moolach Beach data (Fig. 5), the average slope of the eight lines ranged from 0.20 to 0.28. Average growth was not significantly different in the 2 yr (nonparametric rank sum test). Individual growth back calculated from the otoliths of 16 fish captured in Yaquina Bay range in average slope from 0.19 to 0.32 (Fig. 6). The growth rates of fish collected in 1978 versus 1979 were significantly different (P — 0.05, nonpara- metric rank sum test). The range in average slope for the 1978 group is 0.19 to 0.25 and for the 1979 group, 0.21 to 0.32. Since the sample size was small, this test is inconclusive, but examina- tion of the size-at-age data by year (Fig. 7) tends to support the results of the back calculations. The influx of fish to the Moolach Beach nur- sery ground, determined by back calculating the date of recruitment to the sampling gear for each fish, was distributed over the winter and spring (Fig. 8). During the summer, recruitment de- clined and was zero by July 1978 and by Septem- ber 1979. For juveniles captured in the estuary, recruit- ment appeared to be concentrated over a few winter months (Fig. 9). A sharp peak is evident o ■D O to 120 110- 100 90 80 70 a> 60 50- 40" 30" 20 10 MOOLACH BEACH JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1978 JAN FEB MAR APR MAY JUN JUL AUG 1979 SEP Figure 5.— Back-calculated growth of eight individual Parophrys vetulus from Moolach Beach during 1978-79. 249 FISHERY BULLETIN: VOL. 80, NO. 2 o (75 JAN FEB MAR ' APR MAY JUN JUL AUG SEP 1978 I 1979 Figure 6.— Back-calculated growth of 16 individual Parophrys vetulus from Yaquina Bay during 1978-79. 120 YAQUINA BAY & 1978 • 1979 41 a " * .'A m • V' ••• • .. »:- . • ••• • • • 160 240 320 Age (days ) Figure 7.— Size-at-age data plotted by year of capture of Pa- rophrys vetulus. in November, December, and January. As in the Moolach Beach data, recruitment goes to zero in the summer, but reappears in the fall among Yaquina Bay fish. DISCUSSION Several previous studies have attempted to estimate growth rates for English sole juveniles (Table 2). For the purposes of comparison with the data reported here, the total length measure- ments used in other studies were converted to MOOLACH BEACH _c OC T NOV DEC 1977 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC ' JAN FEB MAR APR MA, JUN JUL AUG SEP 1978 '. 1979 Monlh ot Recruitment Figure 8.— Distribution of Parophrys vetulus recruitment to the sampled population at Moolach Beach during 1978-79. Full recruitment to the sampling gear was estimated to occur at 120 d of age. standard length using the relationship given by Laroche and Holton (1979). The recalculated daily growth estimates from all of these other studies are similar, but are substantially higher than my estimated daily growth rates. Smith and Nitsos (1969) and Van Cleve and El-Sayed (1969) determined growth during the first year of life by back calculating the size of the fish when the first detectable annulus on the inter- opercular bone was formed. This occurs during the fish's first slow growth season, which may be at various ages due to the protracted spawning period of this species. Growth back calculations of individual fish (Figs. 5, 6) do not show a clear slow growth period during the first year. 250 ROSENBERG: GROWTH OF JUVENILE ENGLISH SOLE YAQUINA BAY rfUm OCT NOV OEC ; JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC'JAN FEB MAR APR MAT JUN JUL AUG SEP 1977 ! 1978 '• 1979 Month of Recruitment Figure 9.— Distribution of Parophrys vetulus recruitment to the sampled population in the Yaquina Bay estuary during 1978-79. Full recruitment to the sampling gear was estimated to occur at 120 d of age. The other two studies (Westrheim 1955; Ken- dall 1966) utilize the technique of following modal progressions through time in length- frequency distributions. These estimates are strongly influenced by the efficiency of the sam- pling gear. If the smaller fish are sampled less efficiently than the larger, growth will be over- estimated. Emigration of small individuals, immigration of larger fish, and differential mor- tality of small fish would all produce overesti- mates of growth using this method. Also, length- frequency modal progression may give variable results dependent on the method of choosing the modes. Variability in the size-at-age data was much higher for fish sampled in the estuary compared with those sampled in the open coastal area, but the mean growth rates for fish from the two areas were similar. Physical factors may affect growth variability. Yaquina Bay has highly variable temperature and salinity. Frey 4 found differ- ences of up to 57.. salinity and 2°C between high and low tides in the lower bay. Bottom tempera- ture in the estuary ranges between 5° and 15°C through the year, and salinity from virtually to 347... At Moolach Beach in contrast, a more con- stant environment may be expected. The open coastal region does not have a large source of freshwater to influence salinity and tempera- ture. Huyer (1977) and Huyer and Smith (1978) reported that bottom water salinity off the Ore- gon coast fluctuates about 17.. from winter to summer. Temperature varies from 6.5°C in sum- mer, due to seasonal upwelling, to about 10°C in winter. There are two ways in which growth variabil- ity can be reduced. Either outlying individuals have their growth rates altered towards the mean or they are removed from the population. Particularly good or bad growth conditions in an area would affect the growth of all individuals, and alter the mean. Emigration and mortality are the two possible removal processes. The size- at-age plot for Moolach Beach (Fig. 3) and other data (Laroche and Holton 1979) indicate that most P. vetulus juveniles move out of the near- shore area at between 70 and 80 mm SL. Emi- gration from the estuary appears to be at a larger size, approximately 100 mm SL (Westrheim 1955; Olson and Pratt 1973). Predation in the estuary is probably low com- pared with the open coast. Few large fishes are regularly found in the bay, although birds may be significant predators. Kuipers (1977), in a study of an estuarine nursery for plaice in the Wadden Sea, reported predation mortality to be low in contrast to a coastal nursery area studied by Steele and Edwards (1970). Finally, intraspecific competition may affect growth. The estimated densities of juvenile Eng- lish sole in the estuary are a consistent order of magnitude greater than at Moolach Beach (Kry- gier and Pearcy 5 ). Competition may potentially 4 B. Frey, School of Oceanography, Oregon State University, Corvallis, OR 97331, pers. commun. March 1980. 5 E. E. Krygier and W. G. Pearcy, School of Oceanography, Oregon State University, Corvallis, OR 97331, pers. commun. March 1980. Table 2.— Summary of growth estimates from previous studies: the data has been recal- culated so that direct comparisons can be made (see text). Location Size at 1 yr of age (mm SL) Yaquina Bay, Oreg Monterey Bay. Calif Puget Sound, Wash Puget Sound. Wash Daily growth rate (mm/d) 117 0.40 108-126 0.36-0 43 128 44 — winter 48 summer 0.73 Source Westrheim 1956 Smith and Nttsos 1969 Van Cleve and El-Sayed 1969 Kendall 1966 251 FISHERY BULLETIN: VOL. 80, NO. 2 emphasize differences among individuals and increase observed variability. The most plausible mechanism for explaining low growth variability at Moolach Beach com- bines limitation and removal processes. If the population is food limited in the open ocean and selective predation on smaller, slower growing individuals is occurring, the observed variability in size-at-age will be small. Using the otolith aging technique this hypothesis is testable. It re- quires a comparison of the size-at-age distribu- tion of fish found in the stomachs of predators with the distribution shown in Figure 3. A hypothesis arising from this study is that survival, not growth, is enhanced in the estua- rine nursery ground compared with the open coast. Testing of this hypothesis will be an impor- tant step in understanding the role that estuaries play in the life history of many fishes. ACKNOWLEDGMENTS I would like to thank W. G. Pearcy, C. B. Miller, and A. V. Tyler for their guidance and assist- ance; B. Frey, E. Krygier, and C. Creech for pro- viding accessory data; and J. L. Laroche and W. Wakefield for their invaluable assistance throughout the course of this study. Funding was provided by Oregon State Uni- versity Sea Grant Project No. A/OPF-1. LITERATURE CITED Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Gunter, G. 1961. Some relations of estuarine organisms to salinity. Limnol. Oceanogr. 6:182-190. Haedrich, R. L. In press. Estuarine fishes. In B. H. Ketehum (editor), Ecosystems of the world: Vol. 22. Estuaries and enclosed seas. Elsevier Press, Amsterdam. Huyer, A. 1977. Seasonal variation in temperature, salinity and density over the continental shelf off Oregon. Limnol. Oceanogr. 22:442-453. Huyer, A., and R. L. Smith. 1978. Physical characteristics of Pacific northwestern coastal waters. In R. W. Krauss (editor), The marine plant biomass of the Pacific Northwest coast, p. 37-55. Oregon State Univ. Press, Corvallis. Kendall, A. W., Jr. 1966. Sampling juvenile fishes on some sandy beaches of Puget Sound, Washington. M.S. Thesis, Univ. Wash- ington, Seattle, 77 p. Kuipers, B. R. 1977. On the ecology of juvenile plaice on a tidal flat in the Wadden Sea. Neth. J. Sea Res. 11:56-91. Laroche, J. L., S. L. Richardson, and A. A. Rosenberg. 1982. Age and growth of a pleuronectid, Parophrys vetu- lus, during the pelagic larval period in Oregon coastal waters. Fish. Bull., U.S. 80:93-104. Laroche, W. A., and R. L. Holton. 1979. Occurrence of 0-age English sole, Parophrys vetu- lus, along the Oregon coast: an open coast nursery area? Northwest Sci. 53:94-96. McHugh, J. L. 1967. Estuarine nekton. In G. H. Lauff (editor). Estu- aries, p. 581-620. Am. Assoc. Adv. Sci. Publ. 83. Misitano, D. A. 1976. Size and stage of development of larval English sole, Parophrys vetulus, at time of entry into Humboldt Bay. Calif. Fish Game 62:93-98. Olson, R. E., and I. Pratt. 1973. Parasites as indicators of English sole (Parophrys vetulus) nursery grounds. Trans. Am. Fish. Soc. 102: 405-411. Pannella, G. 1971. Fish otoliths: daily growth layers and periodical patterns. Science (Wash., D.C.) 173:1124-1127. 1974. Otolith growth patterns: an aid in age determina- tion in temperate and tropical fishes. In T. B. Bagenal (editor), The proceedings of an international symposium on the ageing of fish, p. 28-39. Unwin Brothers, Sur- rey, Engl. Pearcy, W. G. 1962. Ecology of an estuarine population of winter floun- der, Pseudopleuronectes americanus(Wa.\ba.um). Bull. Bingham Oceanogr. Collect., Yale Univ. 18:1-78. Pearcy, W. G., and S. S. Myers. 1974. Larval fishes of Yaquina Bay, Oregon: A nursery ground for marine fishes? Fish. Bull., U.S. 72:201-213. Rosenberg, A. A., and J. L. Laroche. 1982. Growth during metamorphosis of English sole, Parophrys vetulus. Fish. Bull., U.S. 80:152-155. Scheffe, H. 1959. The analysis of variance. Wiley, N.Y., 477 p. Smith, J. G., and R. J. Nitsos. 1969. Age and growth studies of English sole, Parophrys vetulus, in Monterey Bay, California. Pac. Mar. Fish. Comm. Bull. 7:73-79. Steele, J. H., and R. R. C. Edwards. 1970. The ecology of 0-group plaice and common dabs in Loch Ewe. IV. Dynamics of the plaice and dab popula- tions. J. Exp. Mar. Biol. Ecol. 4:174-187. Struhsaker, P., and J. H. Uchiyama. 1976. Age and growth of the nehu, Stolephorus purpur- eus (Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily growth increments of sagittae. Fish. Bull., U.S. 74:9-17. Taubert, B. D., and D. W. Coble. 1977. Daily rings in otoliths of three species of Lepomis and Tilapia mossambica. J. Fish. Res. Board Can. 34:332-340. Van Cleve, R., and S. Z. El-Sayed. 1969. Age, growth, and productivity of an English sole (Parophrys vetulus) population in Puget Sound, Wash- ington. Pac. Mar. Fish. Comm. Bull. 7:51-71. Westrheim, S. J. 1955. Size composition, growth and seasonal abundance of juvenile English sole (Parophrys vetulus) in Yaquina Bay. Fish. Comm. Oreg. Res. Briefs 6:4-9. 252 POPULATION FLUCTUATIONS OF CALIFORNIA SEA LIONS AND THE PACIFIC WHITING FISHERY OFF CENTRAL CALIFORNIA 1 David G. Ainley, Harriet R. Huber, 2 and Kevin M. Bailey 3 ABSTRACT Seasonal fluctuations in the number, age ratios, and diet of California sea lions, Zalopkus califor- nianus, were studied at the Farallon Islands, central California, from 1971 to 1980. During these years, average monthly numbers increased geometrically, except for April and May. Before 1977, the annual peak in population occurred during April and May, almost no animals were present late June to early July, and a slight peak occurred during fall; adult males predominated. Beginning in 1977, fall numbers equaled or exceeded those in spring, large numbers remained throughout summer, and subadults predominated. We hypothesize that seasonal fluctuations in sea lion num- bers were related to the availability of their principal prey, Pacific whiting, Merluccius productus, and that the changes that began in 1977 were related to termination of the whiting fishery off central California beginning that year. The California sea lion, Zalophus californianus, ranges along the North American west coast from the Gulf of California to British Columbia. Bartholomew (1967) hypothesized that most adult males migrate to the north from breeding sites in Baja California and southern California beginning in midsummer and remain there until the early spring when they return south, and that females and young animals remain in the vicin- ity of breeding areas or move somewhat south- ward during the nonbreeding season. This has become the accepted explanation to account for the seasonal movements in the population (e.g., Mate 1975). Preliminary analysis of census and diet information collected at the Farallon Islands during 1971-80 led to a related hypoth- esis that the movements of male sea lions toward the north could be a response to the seasonal oc- currence and availability of an important prey species, the Pacific whiting, Merluccius produc- tus (Huber et al. 4 ). This information was later quoted by Fiscus (1979). Additional analysis, presented here, provides more insight into the ecological relationship between the two species. The Pacific whiting is an abundant midwater fish of the continental slope and shelf off Cali- 'Contribution No. 232 of the Point Reyes Bird Observatory. 2 Point Reyes Bird Observatory, Stinson Beach, CA 94970. 3 College of Fisheries, University of Washington, Seattle, WA 98195 l o o o o CO CO ro uj o o o o en ai o o CO CD o o 10 o TIME OF CENSUSES Figure 1.— The number of California sea lions hauled out during hourly periods at Shubrick Point, Southeast Farallon Island; the mean and ± standard deviation are shown based on 12 all-day watches during April and May 1974. FISHERY BULLETIN: VOL. 80, NO. 2 RESULTS AND DISCUSSION California Sea Lion Biology Aside from the one pup born at Southeast Farallon, every year since 1974 except 1978 (plus its mother and at least one bull) (Pierotti et al. 1977; Huber et al. [footnote 4]; Point Reyes Bird Observatory unpubl. data), the California sea lion population was comprised of nonbreeding males. Major breeding sites are located in the southern California islands (Bartholomew 1967; LeBoeuf and Bonnell 1980). From 1971 to 1976 a large peak in numbers was reached each year at the Farallones in late April or early May, when animals migrating south toward southern breed- ing sites hauled out for short periods (Fig. 2). A majority of animals departed (temporarily?) each evening to feed (Hobson 1966); about an hour after dawn they began to return and by early afternoon maximum numbers were hauled out. Numbers present each day rapidly declined in late May, and by late June only a few Zalophus hauled out. Population size increased again in late July, reached a peak in August or Septem- ber that was much smaller than in spring, and than declined to a level maintained through the <" 500- z o 1977-80 EC LU [TJ 3 Z 500- Figure 2.— The mean ( ± standard deviation) number of Cali- fornia sea lions hauled out at Southeast Farallon Island each month during two periods: 1971/73-76 and 1977-80; below each curve are the number of censuses each month and above are the proportion of adults present. 254 AINLEY ET AL.: POPULATION FLUCTUATIONS OF SEA LIONS winter. Average monthly population size in- creased slightly from one year to another (Fig. 3). The proportion of adults present each month ranged between 73 and 95%. Since 1977, population fluctuations of the Cali- fornia sea lions have been markedly different in several ways. First, except for April and May, average monthly population size began to in- crease rapidly from one year to the next (Fig. 3). This was especially evident for the summer and fall and thus, secondly, by 1978 the timing of the annual maximum population had shifted and fall counts were exceeding those of the spring peak (Fig. 2). In fact, for each month except April and May, average monthly numbers in- creased geometrically from 1971-73 to 1980 (least squares; r ranged 0.7745 to 0.9537, P<0.01). Finally, the percentage of adults during 1977-80 was reduced to a range between 15 and 35%. These differed significantly from percentages of adults in the period 1971/73-76 (P<0.01; per- centage test, Sokal and Rohlf 1969:608). Young animals were thus migrating north rather than remaining in southern California and Baja Cali- fornia waters as Bartholomew (1967) had noted in earlier years. Seasonal population fluctuations and age ratios at the Farallones from 1971 to 1976 were largely similar to those at coastal sites, as mea- sured at Ano Nuevo Island (80 km away, Orr and Poulter 1965; Lance and Peterson 1968), and 1971/73 74 Figure 3. — The average number of California sea lions hauled out annually at Southeast Farallon Island. Dots above each year are monthly averages; the curve is described by the geo- metric equation: y = a\ x , where a = 9.5 X 10" 8 and A = e 02979 ; r = 0.6557, P<0.01. sites farther north (Mate 1975). Exceptional at the Farallones was the fact that there was almost no fall peak, whereas at coastal sites it greatly exceeded the peak in spring. When the fall peak increased in 1977 the Farallon pattern became similar to coastal sites. However, it is possible that the age composition, for which few compar- ative data are available, and the size of the spring peak were changing then at the Farallons. At coastal sites there is a small spring peak and a large fall peak, but at the Farallones the two peaks became equal in magnitude. The diet of California sea lions at the Faral- lones, as revealed by regurgitated items, has been comprised of at least 20 species of prey (Table 1) (some otoliths could actually have come from the stomachs of sea lion prey). Outstanding were the predominance of Pacific whiting, par- ticularly from April to August, and the diversifi- cation in diet from September to March. The whiting eaten averaged 25 to 36 cm in length and were 2 to 3 yr of age (Bailey and Ainley in press). Except for the short period during summer when they were away at breeding sites, Califor- nia sea lions were most abundant when whiting predominated in their diet. At coastal sites of central California, the market squid, Loligo opalescens, along with whiting and northern anchovies, Engraulis mordax, are dominant prey of this pinniped (Morejohn et al. 1978). California Sea Lions and the Pacific Whiting Fishery From 1967 to 1972 most Pacific whiting were caught off the coasts of British Columbia, Wash- ington, and Oregon (Fig. 4). After 1972, catches increased off the California coast, and especially high catches of around 100,000 t occurred from 1974 to 1976. This southward shift of fishing is believed to be due to a depletion of large adults in the Pacific Northwest. Fishing off central Cali- fornia targeted juvenile whiting. 5 After the FCMA restriction on fishing south of lat. 39°N, the total whiting catch dropped significantly (Fig. 4). Whiting prevalence in the diet of Farallon sea lions was directly correlated to the average monthly number of trawlers fishing for whiting in the Farallon area (Table 2; r = 0.747, t = 3.55, 5 Anonymous. 1976. Summary of National Marine Fish- eries Service views on the status of the Pacific hake resource. Unpubl. rep., 4 p. Northwest and Alaska Fisheries Center. NMFS. NOAA, 2725 Montlake Blvd. E., Seattle. WA 98115. 255 FISHERY BULLETIN: VOL. 80, NO. 2 Table 1.— Percent composition of California sea lion diet as determined by otoliths and beaks regurgitated at haul out sites, Southeast Farallon Island, 1974-78. Months: J F M A M J J A S O N D Cephalopods Octopus rubescens 1 2 Berryteuthis (?) sp. 1 Gonatus sp. 1 Loligo opalescens 3 Fishes Merluccius productus 54 36 28 87 94 98 96 84 38 43 28 30 Sebastes spp 45 27 61 11 5 <1 14 30 16 69 30 Porichthys notatus 1 6 1 1 <1 <1 1 31 1 Engraulis mordax 20 <1 1 Glyptocephalus zachirus 1 <1 10 1 1 40 Chilara taylori 8 <1 9 Parophrys vetulus 2 <1 3 Genyonemus lineatus <1 <1 2 2 Citharichthys sordidus 1 2 3 Microgadus proximus 1 2 <1 <1 Atherinopsis californiensis <1 <1 Leptocottus armatus <1 2 Zalembius rosaceus <1 2 Microstomas pacificus <1 Trachurus symmetricus 4 <1 <1 Clupea pallasi <1 Lyopsetta exilis <1 Total prey (no.) 11 147 55 550 1,077 291 45 535 267 102 140 10 IOO z cr o < < I- o o IOO" o o I967 69 71 Figure 4.— The total catch of whiting in the Pacific coast fish- ery and the proportion of that catch taken off California, 1967- 79. df = 10, P<0.01, Spearman rank correlation). Considering the whole coast of California, trawl- ers concentrated near the Farallones, at least from 1974 to 1976, when fishery surveillance rec- ords were available to us. If we assume that the number of trawlers and the prevalence of whit- ing in sea lion diets, in conjunction with sea lion population size, reflect whiting availability, we conclude that both sea lions and humans were attracted to continental slope waters at the same time in order to catch whiting. The only differ- ence was that the sea lions departed at the peak of Table 2.— Number of stern trawlers fishing for Pacific whit- ing over the California continental slope between lat. 39° and 37°N from January through December 1974-77; data summar- ized from NMFS monthly surveillance reports. Year M M O N 1974 13 43 55 60 57 55 11 1975 3 8 60 64 90 64 2 ? 1976 10 35 55 50 38 13 1977 x 1974/76 1 2 28 47 67 58 32 34 3 harvest in order to return to traditional breeding sites. Associated with the unavailability of whiting, both fishing activity and the preponderance of whiting in the sea lion diet dropped off from Sep- tember to March. During the winter months adult whiting migrate off the continental shelf to spawn in deeper waters of the continental slope (Bailey 1980), and juveniles probably show the same behavior. In addition, during the spawning months they do not diurnally migrate but remain deep (Nelson and Larkins 1970). They are thus unavailable to both the fishery and the sea lions. We offer the following hypothesis to explain the patterns observed in the sea lions' behavior. First, they are attracted to continental slope waters of central California by whiting which, due to their own migrations, become available there during spring and summer. The trawler fishery, also attracted by greater fish availabil- ity, was perhaps depleting whiting stocks sea- sonally to such an extent during the early to mid- 1970's that by late summer when sea lions were 256 AINLEY FT AL.: POPULATION FLUCTUATIONS OF SEA LIONS returning north from breeding sites, offshore waters near the Farallones were no longer as attractive to the pinnipeds as during the spring. The sea lions thus remained along the coast to feed on other prey. Then in 1977, when trawlers no longer fished for whiting off central Califor- nia, the sea lions responded in three ways, all possibly due to increased food supply during summer and fall: 1) Young animals moved farther north or farther off the coast than previ- ously, 2) more adults remained during summer instead of migrating south, and 3) adults return- ing from southern breeding sites moved offshore in larger numbers than they had in previous falls. The size of the sea lion population peak during spring was not affected by termination of the fishery, because fishing was only just getting under way each year at that time. Adding coincidental support to the hypothesis that the 1966/76 whiting fishery off central Cali- fornia was indirectly depressing the numbers of California sea lions in the vicinity are data from other localities. Populations of California sea lions at breeding sites on southern California islands have been increasing geometrically for the past several decades (Bartholomew 1967; LeBoeuf and Bonnell 1980; LeBoeuf 6 ). At the crease in numbers at the Farallon Islands is likely a reflection of this. Successive counts at coastal Ano Nuevo Island during the early 1960's also reflected this increase, but beginning some- time between 1963 and 1967 numbers began a decline there that lasted through 1975; since then, however, they have begun to increase again (LeBoeuf and Bonnell 1980; LeBoeuf). At the Monterey breakwater, about 80 km farther south, D. J. Miller 7 has noted that numbers of subadult California sea lions since about 1978 have been much higher than in previous years. Changes in the occurrence of another pinni- ped, the northern fur seal, Callorhinus ursinus, at the Farallones, provide additional support to the hypothesis. Also an important whiting pred- ator (Fiscus 1979), this species breeds at San Miguel Island in southern California and in the Bering Sea, and during the nonbreeding season frequents waters of the California continental slope. From 1970 to 1976 we observed individual fur seals at the farallones on only 3 single days, 6 B. J. LeBoeuf, Division of Natural Sciences. University of California at Santa Cruz, Santa Cruz, CA 95064, pers. com- mun. June 1981. 7 D. J. Miller, California Department of Fish and Game. Mon- terey, CA 93940. pers. commun. June 1981. each 2 yr apart. Since then, however, their occur- rence has changed dramatically: the species has occurred annually during the summer and fall, and at least 10 different individuals(determined by tags or peculiar scars) have hauled out, some repeatedly, for periods of variable length. Two that hauled out were tagged at San Miguel; another has hauled out for 5 yr in succession. The fur seal breeding population on San Miguel Island has been increasing geometrically from the early 1960's to the present(LeBoeuf and Bon- nell 1980) and the increasing occurrence of this species on the Farallones is likely a reflection of this trend. The dramatic jump in numbers at the Farallones beginning after 1976, however, is out of line with the continuous increase in breeding numbers. Cessation of the whiting fishery off central California in 1976 may account for the change at the Farallones, just as this may be re- sponsible for the change in population dynamics of California sea lions in central California. ACKNOWLEDGMENTS Field work at the Farallon Islands was funded by the Point Reyes Bird Observatory, U.S. Fish and Wildlife Service, Marine Mammal Commis- sion, and National Marine Fisheries Service (Marine Mammal Laboratory and Southwest Fisheries Center). Logistic support was provided by the U.S. Coast Guard and the Oceanic Society, San Francisco Bay Chapter. The Farallones com- prise a national wildlife refuge, and we thank the personnel of the San Francisco Bay National Wildlife Refuge for their help. Marine mammal food items were collected under NMFS permit No. 146; fish otoliths were identified by J. E. Fitch and cephalopod beaks were identified by D. G. Ainley; G. Galbraith, Division of Enforce- ment and Surveillance, NMFS, provided data on the occurrence of whiting trawlers. C. S. Strong, T. J. Lewis, R. J. Boekelheide, R. P. Henderson, R. R. Le Valley, S. H. Morrell, J. W. Higbee, B. Bainbridge, K. Darling, and W. Clow assisted with counts. O'B. Young helped to prepare the manuscript. Valuable comments were offered by R. L. DeLong, D. P. DeMaster, C. H. Fiscus, B.J. LeBoeuf, and P. F. Major. LITERATURE CITED Bailey, K. M. 1980. Recent changes in the distribution of hake larvae: causes and consequences. Calif. Coop. Oceanic Fish. Invest. Rep. 21:167-171. 257 FISHERY BULLETIN: VOL. 80, NO. 2 Bailey, K. M., and D. G. Ainley. In press. The dynamics of California sea lion predation on Pacific whiting. Fish. Res. Bartholomew, G. A. 1967. Seal and sea lion populations of the California Islands. In R. N. Philbrick (editor), Proceedings of the symposium on the Biology of the California Islands, p. 229-244. Santa Barbara Botanic Garden, Santa Barbara. Dark, T. A. 1975. Age and growth of Pacific hake, Merluccius pro- duces. Fish. Bull., U.S. 73:336-355. Dark, T. A., M. 0. Nelson, J. J. Traynor, and E. P. Nunna- LEE. 1980. The distribution, abundance and biological char- acteristics of Pacific whiting, Merluccius productus, in the California-British Columbia region during July- September 1977. Mar. Fish. Rev. 42(3-4):17-33. Fiscus, C. H. 1979. Interactions of marine mammals and Pacific hake. Mar. Fish. Rev. 41(10):l-9. Hobson, E. F. 1966. Visual orientation and feeding in seals and sea lions. Nature (Lond.) 210:326-327. Lance, C. C, and R. S. Peterson. 1968. Seasonal fluctuations in populations of California sea lions. Ano Nuevo Rep. 2:30-36. (Univ. Calif., Santa Cruz.) LeBoeuf, B. J., and M. L. Bonnell. 1980. Pinnipeds of the California Islands: abundance and distribution. In D. M. Power (editor), The California Islands: Proceedings of a Multidisciplinary Symposium, p. 475-493. Santa Barbara Museum of Natural His- tory. Mate, B. R. 1975. Annual migrations of the sea lions Eumetopias jubatus and Zalophus californianus along the Oregon coast. Rapp. -P.-V. Reun. Cons. Int. Explor. Mer 169: 455-461. Morejohn, G. V., J. T. Harvey, and L. T. Krasnow. 1978. The importance of Loligo opalescens in the food web of marine vertebrates in Monterey Bay, California. In C. W. Recksiek and H. W. Frey (editors), Biological, oceanographic, and acoustic aspects of the market squid, Loligo opalescens Berry, p. 67-98. Calif. Dep. Fish Game, Fish. Bull. 169. Nelson, M. O., and H. A. Larkins. 1970. Distribution and biology of Pacific hake: A synop- sis. In Pacific hake, p. 23-33. U.S. Fish Wildl. Serv. Circ. 332. Orr, R. T., and T. C. Poulter. 1965. The pinniped population of Ano Nuevo Island, California. Proc. Calif. Acad. Sci., Ser. 4, 32:377- 404. Pierotti, R. J., D. G. Ainley, T. J. Lewis, and M. C. Coulter. 1977. Birth of a California sea lion on Southeast Farallon Island. Calif. Fish Game 63:64-66. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. W. H. Freeman, San Franc, 776 p. 258 FEEDING BEHAVIOR OF THE HUMPBACK WHALE, MEGAPTERA NOVAEANGLIAE, IN THE WESTERN NORTH ATLANTIC James H. W. Hain, 1 Gary R. Carter, 1 Scott D. Kraus, 2 Charles A. Mayo, 3 and Howard E. Winn 1 ABSTRACT Observations on the feeding behavior of the humpback whale, Megaptera novaeangliae, were made from aerial and surface platforms from 1977 to 1980 in the continental shelf waters of the north- eastern United States. The resulting catalog of behaviors includes two principal categories: Swim- ming/lunging behaviors and bubbling behaviors. A behavior from a given category may be used independently or in association with others, and by individual or groups of humpbacks. The first category includes surface lunging, circular swimming/thrashing, and the "inside loop" behavior. In the second category, a wide variety of feeding-associated bubbling behaviors are described, some for the first time. The structures formed by underwater exhalations are of two major types: 1) bubble cloud— a single, relatively large (4-7 m diameter), dome-shaped cloud formed of small, uniformly sized bubbles; and 2) bubble column— a smaller (1-1.5 m diameter) structure composed of larger, randomly sized bubbles, used in series or multiples. Both basic structures are employed in a variety of ways. Many of these behaviors are believed to be utilized to maintain naturally occurring concentrations of prey, which have been identified as the American sand lance, Ammodytes americanus, and occasionally as herring, Clupea harengus. This paper reports on the feeding behavior of the humpback whale, Megaptera novaeangliae, in the continental shelf waters of the northeastern United States. We describe several feeding be- haviors reported for the first time, as well as a number of behaviors known from other areas but not previously reported for these waters. Our col- lective observations provide the beginning of a more complete catalog than has previously been available. Early observations of humpback feeding be- havior were made by Ingebrigtsen (1929) from the Norwegian Sea near Bear Island: "It [the humpback] employed two methods of capturing 'krill' when the latter was on the sur- face of the water. One was to lie on its side on the surface and swim round in a circle at great speed, while it lashed the sea into a foam with flukes and tail and so formed a ring of foam. The frightened 'krill' gathered together in the circle. This done the humpback dived under the foam-ring and a moment later came up in the center to fill its open mouth with 'krill' and ■Graduate School of Oceanography, University of Rhode Island, Kingston, RI 02881. 2 Present address: New England Aquarium, Central Wharf Boston, MA 02110. 3 Provincetown Center for Coastal Studies, P.O. Box 826, Provincetown, MA 02657. ^n . « water, after which it lay on its side, closed its mouth, and the catch was completed. "The other method was to go a short distance below the surface of the water, swimming in a ring while at the same time it blew off. The air rose to the surface like a thick wall of air bub- bles and these formed the 'net'. The 'krill' saw this well of air bubbles, were frightened into the centre, and then the manoeuvre of the first method was repeated." Some 45 yr later, "bubblenetting" was reported from Alaskan humpbacks by Jurasz and Jurasz (1978), and later described in detail (Jurasz and Jurasz 1979). With the exception of the work of Watkins and Schevill (1979), accounts of feeding behavior of this species in the waters of the west- ern North Atlantic are few and largely anecdot- al. MATERIALS AND METHODS Observations were made from dedicated air- craft (a Cessna 337 Skymaster and a Beechcraft AT-11 4 ), from dedicated surface vessels (the 27.5 m Dolphin III and the 21.3 m Tioga), from plat- forms-of-opportunity, and from shore stations. Manuscript accepted November 1981. FISHERY BULLETIN: VOL. 80, NO. 2, 1982. "Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 259 FISHERY BULLETIN: VOL. 80, NO. 2 All data were collected by experienced observ- ers. Photographs taken in both 35 mm and 70 mm format documented most observations, and were supplemented by written and occasionally tape-recorded field notes. From the aircraft, ob- servers estimated the critical dimensions of feed- ing-associated structures with respect to refer- ences such as the whale's body or flipper length. From shipboard, more precise measurements were obtained through reference to known dimensions on the vessel, or to a 25 cm diameter fiberboard disk which had been deployed in the immediate vicinity of the whale. RESULTS Feeding behaviors were observed on more than 150 occasions in the period April 1977 to May 1980. Observations were made in the area of West Quoddy Head, Mt. Desert Rock, Stellwagen Bank, the waters east and southeast of Cape Cod, and southeast of Block Island (Fig. 1). Feeding, or apparent feeding, was reported for individ- uals and for groups of up to 20 whales. Behaviors Circular Swimming/Thrashing On 2 December 1978, a single humpback whale was observed and photographed swim- ming in a broad (23 m) circle, roiling the surface as it swam. Tail slashing (a rapid sideways sweeping of the flukes) may have accompanied this behavior. Dense flocks of birds were present over the whale, and dolphins were present by the head and body. The presence of both of these feeding-associated elements, as well as the re- semblance to observations by Ingebrigtsen (1929), suggested that feeding was taking place. This initial observation was substantiated in May 1980 when a number of shipboard observa- tions confirmed the behavior as feeding asso- ciated. An initial thrust of the flukes was fol- lowed by the whale's swimming in a broad circle, roiling the surface with flippers and flukes. This was followed in many, but not all, cases by a feed- ing rush through the circle. This behavior was repeated many times by a single animal over a period of several hours. The circular swimming/thrashing behavior, observed on two occasions, each time involving a single whale, is considered relatively uncom- mon. FIGURE 1.— Study area where observations of feeding behav- iors were made. Place names on chart are those referred to in text. Lunge Feeding Lunge feeding is defined as an upward rush at the water surface with the longitudinal axis of the body intersecting the plane of the surface at an angle of 30°-90°. As the whale breaks the surface, the mouth is agape, and quite often a greatly distended throat region is seen. Up to one-third of the body length clears the surface before the whale falls or settles back into the water. Observations and photographs of prey at the surface, in the mouths of the whales, and picked up by closely associated birds leave no doubt that this is a capture mode of feeding be- havior. This common behavior has been recorded in 21% of our feeding observations, from single animals as well as from groups. When several animals fed together, the lunges often were simultaneous and in close proximity (3 m). In several cases, two or more animals came in con- tact, bumping each other as they lunged. Bouts of lunge feeding may contain on the order of 20 lunges (3 animals in one case) in 25 min. The speed at which the lunge takes place is highly variable. At times, the whale bursts through the surface in a vigorous upward rush. At other times, the rise to the surface and the subsequent extension of the rostrum and dis- 260 HAIN ET AL.: FEEDING BEHAVIOR OF HUMPBACK WHALE tended lower jaw above the surface are quite gradual. In several instances, humpbacks were observed feeding in this manner (the slow, grad- ual rise) in formation. Five or six whales ar- ranged side by side and slightly staggered of one another acted in unison. This behavior has been similarly described from Alaskan waters and termed "echeloned" lunge feeding (Jurasz and Jurasz 1979). Inside Loop Behavior On 23 May 1980, a single humpback was ob- served feeding for over 1 h. The whale repeatedly displayed a behavior we have termed an "inside loop." As the whale begins a shallow dive, it sharply strikes the water's surface with its flukes. This action creates an area of turbulence in the water estimated to have an average diame- ter of 9 m. This area of foam and bubbles is seen clearly as the whale swims away at a shallow dive angle with the pectoral fins held horizon- tally. The whale, swimming rapidly, then rolls 180°, so that the white ventral surface of the flukes can be seen just below the surface. An in- side loop (a sharp U-turn in the vertical plane) follows immediately, so that the whale is now swimming toward the area of turbulence. Fi- nally, the whale is seen rising vertically in a slow lunge, with mouth widely agape, through the center of the turbulence created by the fluke slap. The horizontal distance covered by this "out-and-back" motion was on the order of l%-2 body lengths of the whale. The behavioral se- quence is illustrated diagrammatically in Figure 2. Several variations on the basic behavior were observed. The humpback did not feed through the area of turbulence in every instance. Occa- sionally, the whale would surface to the side of the disturbance, not always feeding. On other occasions, a second whale would enter the gen- eral area and subsequently be seen lunge feeding through the disturbance created by the flukes of the first whale, either alone or in unison with the original whale. The inside loop behavior, observed on a single occasion, involving a single whale later joined by a second, is at present considered relatively un- common. Bubbling Behaviors Underwater exhalations or bubbling behav- iors were seen in association with feeding, or apparent feeding, in 52% of our feeding observa- tions. These exhalations appear to be of two ma- jor types, forming what we have termed "bubble columns" and "bubble clouds." In general, bubble columns and bubble clouds have been observed with about equal frequency. BUBBLE COLUMNS.-Bubble columns are formed by the underwater exhalations of a whale swimming from 3 to 5 m (estimated) below the surface. As the bubble bursts are released, they rise vertically to the surface in the form of a somewhat ragged column. The columns are 1-2 m in diameter and are composed of random-sized bubbles estimated to be generally >2 cm. Series of from 4 to 15 bubble columns are used to form rows, semicircles, and complete circles or bubble nets (Figs. 3, 4). JV> B Figure 2.— "Inside loop" type of feeding behavior. A. Upon making a shallow dive, humpback whale strikes the surface sharply with flukes. B. Fluke slap creates an area of turbu- lence (foam/bubbles) as whale swims away in a shallow dive, flippers held horizontally. C. Whale executes a 180° roll and now does a sharp inside loop, or U-turn in the vertical plane. D. Whale lunge feeds through the area of disturbance created by original fluke slap. 261 FISHERY BULLETIN: VOL. 80, NO. 2 B )^ Figure 3.— The seven types of bubbling behaviors associated with feeding in humpbacks. A through D are structures using bubble columns, which are 1- 1'/ 2 m in diameter and composed of nonuniform-sized bubbles (estimated at >2 cm). E through G are bubble cloud structures, 4-7 m in diameter, and composed of uniform-sized bubbles (estimated at<2cm). A. Bubble row. B. Bubble row with "crook," whale feeding location shown. C.V or semicircle shaped bubble curtain. Whale feeds in and through open side of the semicircle. D. Complete cir- cular formation, or bubble net. E. Single bubble cloud. In this example, one of several variations, whale lunge feeds through center. F. Triangular formation of multiple bubble clouds. G. Linear formation of multiple bubble clouds. In the simplest configuration, bubble rows, the whale creates a line of columns (generally 4-6). When this has been completed, the whale turns sharply and feeds, open-mouthed, either at or below the surface, at an acute angle to the screen formed by the row of bubble columns. In some cases, the whale continues to release bubble bursts during its turn, so that the line of bubble 262 columns has a "crook" in the end where the whale feeds. The behavior associated with a semicircle of bubble columns is similar, in that once a semi- circle (or "V") has been constructed, the whale appears and feeds toward the concave portion of the screen. Complete circles of bubble columns, termed bubble nets (Jurasz and Jurasz 1979), have been HAIN ET AL.: FEEDING BEHAVIOR OF HUMPBACK WHALE I A • \ i/ i B c r ' ***"***. N. V D y - . -■■% -^ > \ ^ - • V > * E F Figure 4.— Aerial views of bubble net construction by a humpback whale. A through E are 5 frames from a 29-frame sequence; F is from a sequence immediately following. Underwater exhalations are used to form a bubble net approximately 15 m in diameter, composed of some 15 individual bubble columns. Arrows in B indicate undersides of left pectoral fin and flukes. In A through C, whale is rotated on its longitudinal axis so that the blowhole and dorsal surface are toward the center of the circle. In D, whale turns sharply about on the right pectoral fin and prepares to pass through the center of the net. A stream of turbulence is seen trailing from the dorsal fin area, which is being sharply thrust to the whale's left. In E, the whale is seen in feeding posture, mouth agape, under- water in the center of the net. In F, the whale surfaces and blows weakly before exiting the area of the net. Photographs by S. Kraus. 263 FISHERY BULLETIN: VOL. 80, NO. 2 seen on a relatively few occasions, approximately 8% of our observations. Our clearest observations have been from aircraft, particularly on 23 April 1979 when several sequences of bubble net for- mation were photographed (Fig. 4). The whale, maintaining its longitudinal body axis on a nearly horizontal plane, swims some 3-5 m (esti- mated) below the surface in a circular pattern. The dorsal surface (and blowhole) of the whale is rotated toward the center of the circle so that the flippers are oriented nearly in the vertical plane. As the whale swims in this manner, approxi- mately 15 bubble bursts are released, which rise to the surface as columns and appear to form an effective corral. As the circle or net nears com- pletion, the whale appears to pivot on the axis of the flippers. The flukes are thrust to the outside, and a stream of underwater turbulence is seen trailing from the region of the dorsal fin. The whale then banks to the inside and turns sharply into and through the center of the net — all below the surface of the water. The aerial photographs show apparent feeding, i.e., the mouth is agape and the lower jaw region is greatly distended. Only after this stage does the whale rise to the surface, pause, and blow one or more times be- fore exiting the area of the bubble net. Measure- ments show the circle to be approximately equal in diameter to the whale's length— about 13-15 m. While bubble nets constructed in both the clockwise and counterclockwise directions have been observed, the clockwise direction appears to be more common. There are several variations to the behavior described above. Shipboard observations in May 1980 showed that bubble nets are not restricted to 360° circles, but instead may include from V/ 4 - 2 complete revolutions as the whale swims in a spiral of decreasing radius. Often, smaller bursts of smaller bubbles made up the greater portion of the outer ring, with the bursts and bubbles both increasing in size within the inner ring. Additionally, a line of bubbles 10-30 m in length would often directly precede the forma- tion of the circular portion of the bubble net. This gave the overall structure the shape of a "6" or a "9." Finally, surface lunge feeding (gradual rise type), rather than underwater feeding, was re- ported from this series of shipboard observa- tions. BUBBLE CLOUDS.-Bubble clouds form the second major category of bubbling behaviors associated with feeding. There are several marked differences to the bubble columns de- scribed above. In this case, a single underwater exhalation forms a single, relatively large (4-7 m diameter), dome-shaped "cloud" made up of small (estimated to be <2 cm), uniformly sized individual bubbles (Fig. 5). In a few observations where we were able to see the early stages of bub- ble cloud formation, the cloud appeared quite narrow initially, about 2-3 m in diameter, but ex- panded as it rose toward the surface. In many observations, schools of American sand lance, Ammodytes americanus, were visible over wide areas in patches at the surface in the general area of feeding activity, but prior to the onset of any bubbling behavior in their immediate vicin- ity. In all observations, the whale dove out of sight to produce the bubble cloud which rose gradually toward the surface. The prey, appear- ing as a disturbance at the surface, would at times leap vigorously into the air when the bub- ble cloud surfaced into the school. The subsequent appearance of the whale rela- tive to the bubble cloud displayed a good deal of variation. Observations to date suggest five pos- sible variations, as illustrated in Figure 6. When lunge feeding through the cloud's center was seen (Fig. 6A), the speed of the lunge was slower than lunge feeding observed in the absence of clouds. In the second type of behavioral sequence (Fig. 6B, the slow, horizontal appearance of the whale in the surfaced cloud), over 70 bubble cloud observations recorded from shipboard in 1978-79 suggest a repetitive, rigidly patterned activity composed of the following: 1) The whale sounds, usually with flukes in the air. 2) A cloud of bubbles appears beneath the sea surface up to 2 1 / 4-3% min after sounding. 3) The whale, not obviously swimming, rises slowly to the surface. Its back first appears in the center of the spent cloud of bubbles 5-9 s after the first bubbles in the cloud reach the surface. 4) Three to ten blows and slow, shallow diving precede the sounding dive which begins the next sequence. In this common activity, the actual feeding prob- ably takes place in the cloud and below the sur- face, with the whale's appearance marking the conclusion of the episode. Although no feeding is visible at the surface, the presence of a number of important elements (prey abundant in bubble clouds, similarity of structure to those in known 264 HAIN ET AL.: FEEDING BEHAVIOR OF HUMPBACK WHALE feeding events, repeated occurrence in known feeding areas, and the presence of feeding birds) is strongly suggestive of a feeding-associated be- havior. Bubble clouds were also observed being used in series or multiples. These clouds possess the characteristics described above but are used in groups, generally three, by one or more hump- backs. Two varieties have been seen (Fig. 3F, G): 1) individuals or groups of humpbacks blow clouds in either triangular or random patterns, and feed in the midst of the clouds or within a particular cloud— observed on a number of occa- sions and considered relatively common; and 2) an individual whale was seen to blow three lin- early connected clouds, and then swim on the surface very slowly through the formation- observed on a single occasion and considered un- common. A final variation, which may or may not be directly associated with feeding, is poorly under- stood. At times, a lunge-feeding whale will ex- hale underwater, lunge feed to the surface, and be followed shortly by one to three bubble clouds appearing at the surface, closely adjacent to the whale but arriving at the surface after the whale instead of before, as described above. Behavioral Strategies Figure 5.— Aerial views of bubble cloud formation and asso- ciated feeding. A. Dome-shaped bubble cloud, formed by underwater exhalation, seen rising toward surface. B. Bub- ble cloud after intercepting plane of surface— upper portion of structure is flattened. C. Lunge-feeding whale appears through center of bubble cloud. Photographs by A. Frothing- ham. It has been our experience that a given hump- back whale will generally repeat a fairly rigid feeding pattern over a period of time. However, several individual humpbacks or groups of humpbacks feeding in the same area may or may not display the same feeding strategy. Several examples illustrate this observation. In two instances on Stellwagen Bank in 1978, all humpback whales (five and seven individuals) within a 20 km 2 area displayed bubble cloud feeding (slow rise type) for an entire 1-h period of observation. Every whale in sight appeared to be using the same strategy. During two of the three observation periods on one day in 1979, bubble clouds were formed by one individual in the vicinity of extensive schools of American sand lance, while three other whales were lunge feeding (no bubbling asso- ciated) several hundred meters away. On a third occasion, a single humpback on the northern side of a school of American sand lance was observed forming bubble clouds (with ap- parent subsurface feeding), while three other animals, working the same school of American 265 FISHERY BULLETIN: VOL. 80, NO. 2 B Figure 6.— The five feeding variations associated with bubble clouds. A. Whale lunge feeds vertically through the center of the cloud, as in Figure 5. B. Whale apparently feeds under- water and upon completion rises slowly through the center of the spent bubble cloud; the whale's body is on a horizontal plane and the mouth is not agape. C. Whale lunge feeds to one side of cloud. D. Whale surfaces alongside cloud, emits a weak blow, dives, and reappears lunge feed- ing through the center of the cloud. E. Whale swims vertically up alongside the rising cloud, and then passes horizontally, mouth agape, between the still-rising cloud and the water's surface. sand lance, were generating bubbles in rows, as well as randomly, and lunge feeding. Prey Species Shipboard observations, primarily on Stell- wagen Bank, provide direct visual and photo- graphic evidence that concentrated schools of American sand lance are a frequent prey species in the area. American sand lance was identified in 50% of feeding events from the Dolphin III on Stellwagen Bank in 1978 and in 75% of observa- tions in 1979. Photographs show American sand lance in the corners of the whale's mouth, being picked up by closely associated birds, and in con- centrated surface schools in which the whale is feeding. At least one other species is a target for hump- back feeding. It appeared that humpbacks in the West Quoddy Head area took herring, Clupea harengus, close inshore and in coves, using the bubble cloud and lunge feeding techniques on a number of occasions. 5 DISCUSSION Humpback whales in the North Atlantic feed on a wide variety of prey species, with krill and schooling fishes the most important (Tomilin 1967). In Canadian waters, humpbacks feed heavily on capelin, with krill second in impor- 5 S. K. Katona and P. V. Turnbull, College of the Atlantic, Bar Harbor, ME 04609, pers. commun. October 1980. 266 HAIN KT AL.: FKKDINC BEHAVIOR OF HUMPBACK WHAI.K tance, although the data also suggest haddock, mackerel, whitefish, and sand lance (Mitchell 1973; Sergeant 1975 6 ). The American sand lance has been suggested as a prey species in the Cape Cod area by Overholtz and Nicolas (1979). Our direct evidence confirms their observations and demonstrates the importance of this prey species in these waters. The sand lance is similar in size, summer habitat, and schooling behavior to the more northern capelin, Mallotus villosus (Over- holtz and Nicolas 1979), and therefore may oc- cupy a similar role in the diet of humpbacks in more temperate latitudes. Interestingly, Meyer et al. (1979) reported a significant increase in the relative abundance of sand lance since 1975 on Stellwagen Bank, a trend which was typical of the northwestern Atlantic from Cape Hatterasto the Gulf of Maine. Indirect evidence suggests herring as a prey species in the northern Gulf of Maine. Watkins and Schevill (1979) also tentatively identified herring, along with pollock, Pollachius virens, from Cape Cod waters. These observations will require confirmation as additional knowledge on prey species in New England waters is gained. With regard to the capture mode of feeding be- havior, our observations on lunge feeding closely corroborate those of Watkins and Schevill (1979) and Jurasz and Jurasz (1979). The observations on underwater feeding by humpbacks were almost always in association with bubble struc- tures, although Watkins and Schevill (1979) described several instances of underwater feed- ing in the absence of such structures. "Apparent circling behavior" during feeding was reported by Watkins and Schevill (1979). Our description of what we term circular swim- ming/thrashing behavior expands somewhat on their observations. We speculate that the use of anatomical structures and swimming motion in the manner described bears some generic resem- blance to the "flick feeding" reported from Alas- kan waters by Jurasz and Jurasz (1979). This would seem to be particularly true for the inside loop behavior we have described. These behav- iors may be placed together into a major subdivi- sion of feeding behaviors, the various bubbling behaviors being the other major subdivision. The effect of the whale's feeding behavior on the prey species, and the advantage conferred to the whale, remains a subject for conjecture, since few data are available. The bubbling behaviors are perhaps the most intriguing. Based on ex- periments with artificial bubble curtains, it is known that under certain circumstances, cur- tains of bubbles form an effective barrier to schooling fish (Brett and Alderdice 1958; Smith 1961; Bates and VanDerwalker 1964). Whatever the precise mechanism, it seems reasonable to conclude that humpback whale bubble nets can, and do, effectively corral schools of prey. Whether bubble nets concentrate the prey 7 or merely enclose and maintain naturally occurring concentrations of prey (as hypothesized here) can only be resolved by further study. The humpback appears well suited to these be- haviors; Edel and Winn (1978) have described in some detail the locomotion, maneuverability, and flipper movement required to execute the behaviors described here. It has been suggested (Howell 1970; Brodie 1977) that flashes from the long, white flippers are used to concentrate or herd the prey. This may play a role in the circling behavior, the bubble-netting, and perhaps other types of feeding. In the case of bubble-netting, in addition to their hydrodynamic function, the ver- tical orientation of the two extended flippers may act in unison with the bubble screen to help form the "curtain" which herds and/or entraps the prey. While bubbling behavior appears to be com- monly associated with feeding (52% of our feed- ing observations), some caution is in order. Underwater bubbling, even in the presence of feeding activity, may not always be directly re- lated to feeding (see also Watkins and Schevill 1979). Underwater exhalations from humpbacks in nonfeeding situations have also been observed. On occasion, underwater exhalation by hump- backs when approached by ships has been re- corded. From field observations and study of photographs, the possibility that some swim- ming and bubbling behavior may be "play" be- havior, particularly when displayed in the pres- ence of closely associated dolphins, is recognized. In the Pacific, Hubbs (1965) described under- water exhalations with no clearly apparent func- tion, and Forestell and Herman 8 described the 6 Sergeant, D. E. 1975. An additional food supply for humpback (Megaptera noraeangiiae) and minke whales (Ba- laenoptera acutorostrata). Int. Counc. Explor. Sea, Mar. Mamm. Comm., CM. 1975/No. 13:1-7. 7 Earle, S. A. 1979. Quantitative sampling of krill (Eu- phausia pacifica) related to feeding strategies of humpback whales (Megaptera novaeangtiae) in Glacier Bay, Alaska. Paper presented at The Third Biennial Conference of the Biol- ogy of Marine Mammals, 7-11 Oct. 1979, Seattle, Wash. "Forestell, P. H., and L. M. Herman. 1979. Behavior of 267 FISHERY BULLETIN: VOL. 80. NO. 2 apparent use of bubble screens as camouflage by an escort whale in order to protect a calf or mother-calf pair. It is likely that some functions of bubbling still remain to be discovered. At times, bubbling may be purely adventitious. The humpback possesses a diverse repertoire of feeding behaviors. Whether environmental factors influence the choice of feeding method is presently unknown. Perhaps, as suggested by others (Jurasz and Jurasz 1979; Watkins and Schevill 1979), various prey species or densities elicit different feeding strategies and behaviors. For less mobile prey or high prey densities, rela- tively simple devices may be sufficient. For more mobile and evasive species, or for more efficient feeding in lower densities, more sophisticated methods may be advantageous. ACKNOWLEDGMENTS This study was supported by the Bureau of Land Management, U.S. Department of the In- terior, under contract number AA551-CT8-48 to the Cetacean and Turtle Assessment Program, University of Rhode Island. For their skill and assistance, we thank Captain A. Avellar and the crew of the Dolphin HI; Captain W. Simmons and the crew of the Tioga; and survey pilots T. Flynn, J. McMicken, and J. Rutledge. We grate- fully acknowledge the critical review of the manuscript by R. Edel, S. Katona, J. Roanowicz, and W. Watkins. LITERATURE CITED Bates, D. W., and J. G. VanDerwalker. 1964. Exploratory experiments on the deflection of juve- nile salmon by means of water and air jets. Fish-Pas- sage Res. Prog. Rev. Prog., U.S. Bur. Commer. Fish., Seattle 3:1-14. escort accompanying mother-calf pairs of humpback whales. Paper presented at The Third Biennial Conference of Marine Mammals, 7-11 Oct. 1979, Seattle, Wash. Brett, J. R., and D. F. Alderdice. 1958. Research on guiding young salmon at two British Columbia field stations. Fish. Res. Board Can., Bull. 117, 75 p. Brodie, P. F. 1977. Form, function and energetics of Cetacea: A dis- cussion. In R. J. Harrison (editor), Functional anatomy of marine mammals, Vol. 3, p. 45-58. Acad. Press, N.Y. Edel, R. K., and H. E. Winn. 1978. Observations on underwater locomotion and flip- per movement of the humpback whale Megaptera novae- angliae. Mar. Biol. (Berl.) 48:279-287. Howell, A. B. 1970. Aquatic mammals; their adaptation to life in the water. Dover Publ., N.Y. Hubbs, C. L. 1965. Data on speed and underwater exhalation of a humpback whale accompanying ships. Hvalr. Skr. 48: 42-44. INGEBRIGTSEN, A. 1929. Whales caught in the North Atlantic and other seas. Int. Counc. Explor. Sea, Rapp. P.-V. Reun. 56:1-26. Jurasz, C, and V. Jurasz. 1978. Humpback whales in southeastern Alaska. Alaska Geogr. 5(4):116-127. • 1979. Feeding modes of the humpback whale, Megaptera novaeangliae, in Southeast Alaska. Sci. Rep. Whales Res. Inst., No. 31:69-83. Meyer, T. L., R. A. Cooper, and R. W. Langton. 1979. Relative abundance, behavior, and food habits of the American sand lance, Ammodytes americanus, from the Gulf of Maine. Fish. Bull., U.S. 77:243-253. Mitchell, E. D. 1973. Draft report on humpback whales taken under spe- cial scientific permit by eastern Canadian land stations, 1969-1971. Int. Comm. Whaling, 23d Rep. Comm., Lond., p. 138-154. OVERHOLTZ, W. J., AND J. R. NICOLAS. 1979. Apparent feeding by the fin whale, Balaenoptera physalus, and humpback whale, Megaptera novaeangli- ae, on the American sand lance, Ammodytes americanus, in the Northwest Atlantic. Fish. Bull., U.S. 77:285- 287. Smith, K. A. 1961. Air-curtain fishing for Maine sardines. Commer. Fish. Rev. 23(3):1-14. Tomilin, A. D. 1967. Mammals of the USSR and adjacent countries. Cetacea 9:1-717. Isr. Prog. Sci. Transl. Jerusalem. Watkins, W. A., and W. E. Schevill. 1979. Aerial observation of feeding behavior in four ba- leen whales: Eubalaena glacialis, Balaenoptera bore- alis, Megaptera novaeangliae. and Balaenoptera physa- lus. J. Mammal. 60:155-163. 268 THE INTERRELATION OF WATER QUALITY, GILL PARASITES, AND GILL PATHOLOGY OF SOME FISHES FROM SOUTH BISCAYNE BAY, FLORIDA Renate H. Skinner 1 ABSTRACT This study investigated monogenetic trematode infestation of the gills and gill pathology of yellow- fin mojarra, Gerres cinereus (Gerreidae); gray snapper, Lutjanus griseus (Lutjanidae); and timucu (needlefish), Strongylura timucu (Belonidae) in relation to water quality in south Biscayne Bay, Florida. Two habitats of the three species in the bay, one in the southeast and the other in the south- west, differed in water quality whereas physical and environmental parameters were similar. The water in southwest Biscayne Bay contained.high amounts of ammonia, trace metals, and pesticides which were not present in the southeast bay. The gills of hosts from the habitat with inferior water quality were heavily infested with the Monogenea (Platyhelminthes) Neodiplectanwm wenningeri (on G. cinereus), Ancyrocephalus sp. (on L. griseus), and Ancyrocephalus parvus (on S. timucu) and suffered from excessive mucus secretion, epithelial hyperplasia, fusion of gill lamellae, clubbing and fusion of filaments, and aneurisms. Only light infestations and little or no abnormal tissue changes were noted in fish from the area of good water quality. The findings led to the conclusion that the pollutants in the water acted as an irritant, stressing the fisji, and producing physical and physiological changes which reduced resistance to infestation by Monogenea. Manmade pollution of coastal waters of south- east Florida has reached a critical level in the most populated areas, causing substantial envi- ronmental degradation (Carter 1974) and the loss of valuable fishing grounds, and making some areas unsuitable for recreation. In recent years, the pollution of Biscayne Bay, Fla. (Fig. 1) has become a major issue. The shore of north Bis- cayne Bay is bordered by Miami and Miami Beach, and lined by bulkheads. It receives a large amount of runoff water from the metropol- itan areas (Waite 1976). Although the south- western part of the bay still retains much of its natural shoreline and mangrove forests, it is broken by drainage canals intended to lower the water level in neighboring agricultural and ur- ban areas. These canals therefore carry agricul- tural, industrial, and urban wastes into that part of the bay (Waite 1976). The southeastern shore- line of Biscayne Bay is formed by a chain of islands which is part of Biscayne National Park with no major direct sources of water pollution. The purpose of this study was to investigate if differences existed in the ectoparasite fauna and possible gill pathology in the same three species of fish living in southwest Biscayne Bay in the '3834 El Prado Boulevard, Miami, FL 33133. Manuscript accepted November 1981. FISHERY BULLETIN: VOL. 80. NO. 2. 1982. entrances of three drainage canals on one hand and the relatively clean waters of the southeast bay in the National Park on the other. The effect of water quality on the incidence and intensity of infestation by ectoparasites was investigated along with the frequency and kind of abnormal tissue changes of the gills. Included were those ectoparasites that came close to 100% incidence on their hosts and had a direct life cycle. Three species of Monogenea of the suborder Monopis- thocotylea fell into this category. Monogenea (Platyhelminthes) of the gills are common in fish. Since parasites affect the health of fish, they can be the cause of or a contributing factor to host mortality and epizootics (Iversen et al. 1971). Disease and mass mortality in aquacul- ture, often occurring under crowded conditions, are known to have been caused by the genera Gyrodactylus, Dactylogyrus, and Tetraonchus (Wobeser et al. 1976). Since exchange of gases in the gills takes place through a single thin epithe- lial layer separating the blood from the external environment (Anderson and Mitchum 1974), parasites may cause extensive damage to host gill tissue. Although many adverse circumstances weak- en fish and make them more susceptible to dis- eases, presently available literature is mainly concerned with bacterial diseases (Pippy and 269 FISHERY BULLETIN: VOL. 80, NO. 2 Black Creek Canal Moody Canal Mowry Canal ATLANTIC OCEAN N + 10 km FIGURE 1.— Collection stations in south Biscayne Bay, Fla. Hare 1969; Bullock et al. 1971; Burrows 1972; Snieszko 1974). Information concerning parasit- ic diseases in relation to water quality has been obtained in artificial situations such as aquacul- ture facilities rather than the natural environ- ment. According to Hoffman (1976) eutrophica- tion and pollution probably affect helminth parasites as well as the hosts, but no precise studies have been made. Deleterious effects on various marine biota due to manmade pollution have been investigated, among them disease of fishes and Crustacea (O'Connor 1976; Overstreet and Howse 1977; Sindermann 1979). Overstreet and Howse (1977) suggested that poor environ- mental conditions may favor parasitic infesta- tion by stressing the host, causing disease and lowering resistance. In pioneering literature on fish diseases, gill damage other than parasitic was described as due to exposure (Osburn 1911), industrial pollu- tants (Plehn 1924), and fertilizers (Schaperclaus 1954). More recent literature implicates phenols (Reichenbach-Klinke 1965), ammonia (Reichen- bach-Klinke 1966; Smith and Piper 1975), pesti- cides (Lowe 1964; Walsh and Ribelin 1975), and environmental stress, defined as a change from the normal which reduces the chances for survi- val (Snieszko 1974). Damage to the gills in re- sponse to various toxins in the water was reported by Herbert and Shurben (1964), who suggested 270 SKINNER: INTERRELATION OF WATER QUALITY. GILL PARASITES, AND GILL PATHOLOGY that "sublethal effects of each poison can sum within the individual fish and kill it." Minimal risk, hazard and lethal levels, and median lethal concentrations (LC50, the concentration that kills 50% of the test organisms in 96 h) of certain pollu- tants in the marine environment are published by the National Academy of Sciences Environ- mental Studies Board (1972). Although both the National Academy of Sciences and conservation organizations have emphasized the need for eco- logical information on long-term effects of pesti- cides on wildlife at sublethal doses, most field studies are done after the animals have been found dead. Mitrovic (1972) asked for studies of subtle damage resulting from long-term expo- sure at subacute levels. Local studies are needed, since environmental conditions vary with loca- tion, and temperature, salinity, and pH play a part in the toxicity of poisons (Trussel 1972). Subtle indications of damage, according to John- son (1968), may be a change in behavior caused by lowered efficiency of the organism. He sug- gested that the illustration of physiological and ecological effects of sublethal quantities of envi- ronmental pollutants will lead to a more realistic view in establishing tolerance levels for all toxic pollutants. This requires year-round monitoring to take into consideration seasonal variation, variation in drainage as a result of precipitation, runoff, and irrigation, as well as fluctuating physical or chemical factors. MATERIALS AND METHODS The study period extended from May 1975 to August 1976. The three host species were yellow- fin mojarra, Gerres cinereus (Walbaum), a bot- tom feeder; gray snapper, Lutjanus griseus (Lin- naeus), a predator; and timucu (needlefish), Strongylura timucu (Walbaum), a surface feeder, since they were available on both sides of south Biscayne Bay and remained in one locality for extended periods (Cervigon 1966; Randall 1968; Cressey and Collette 1971; Starck 1971). Collection stations for the fish were the mouths of Black Creek, Moody, and Mowry Canals; the Arsenicker Keys; Elliott Key; and a canal in Sands Key (Fig. 1). South Florida Water Man- agement District maintains salinity control structures a short distance inland from the south- western shoreline of Biscayne Bay at Black Creek, Moody, and Mowry Canals. Directly up- stream from the gates the water is brackish. The flood gates open automatically according to the difference in the water level on both sides when the water exceeds a certain height on the upland side. For many months during the study period the gates of the salinity structures remained closed because of the low freshwater table inland and the danger of saltwater intrusion into inland wells. Fish were collected downstream from the salinity control structures near the entrances of Moody and Mowry Canals into the bay, at the confluence of Black Creek and Goulds Canals where they enter the bay, in mangrove creeks and close to shore at Arsenicker and Elliott Keys, and in a manmade canal and lagoon in the inte- rior of Sands Key. Collections were made between April 1975 and August 1976 during three to five trips per week, depending upon weather conditions. The total number of fish collected was 356, of which 186 were from the southwest locations and 170 from southeast Biscayne Bay (Table 1). Only one spe- cies was collected on a given day from one area to prevent exchange of parasites from one host spe- cies to the other. The yellowfin mojarra were caught by gill net, 75 mm mesh size (stretch), and occasionally on hook and line; the gray snapper on hook and line; and the timucu (needlefish) on hook and line, and by beach seine. Collection trips to the stations were alternated regularly, depending on weather conditions and need. Be- cause of the gear used, size ranges of fish were the same in both localities. Sex ratios were simi- lar, with an average of 52% males and 48% fe- males. Fish were collected at depths between 0.5 m and 2.5 m. Water samples for the salinity read- ings were obtained from depths of 0.3 m, 1.0 m, and 3.0 m. An average of 2.5 salinity measure- ments per month were made at each station and each depth. To avoid contamination a closed, weighted plastic bottle was lowered to the desired depth where it opened and filled with water. Additional salinity data for the years 1975 and 1976 from Black Creek, Mowry, and Moody Canals downstream from salinity structures were made available by the U.S. Geological Sur- vey (USGS) and South Florida Water Manage- Table 1.— Numbers of fish hosts collected in the southeast and southwest locations in Biscayne Bay, Fla., between May 1975 and August 1976. S E Biscayne Bay S W Biscayne Bay Gerres cinereus Lut/anus griseus Strongylura timucu Total 69 57 44 170 52 80 54 186 271 FISHERY BULLETIN: VOL. 80, NO. 2 ment District (SFWMD) (unpubl. data). Tem- perature and dissolved oxygen measurements were taken at 0.3 m and 1.0 m depths. The aver- age of two or three readings represent one mea- surement. The average number of measurements was two to three per month at each station. Data of hydrogen ion concentration expressed as pH were obtained from the Dade County Depart- ment of Environmental Resources Management (DERM) and USGS (unpubl. data). DERM and USGS obtained routine monthly water quality data for Black Creek, Mowry, and Moody Canals downstream from salinity struc- tures and made them available for this study. The DERM laboratory also made water quality analyses of eight southeast Biscayne Bay water samples collected at intervals of 2 mo. The sam- ples, taken from slick-free water (see Discussion section), were kept in plastic bottles which con- tained a few milliliters of hydrosulfuric acid for preservation, and were refrigerated until arrival in the laboratory. Chemical analyses were made for total ammonia nitrogen, nitrite, nitrate, phos- phate, and total organic carbon. DERM and USGS furnished data on heavy metals in Black Creek and Mowry Canals and pesticides in Black Creek Canal. Fish were kept alive in an aerated plastic con- tainer until arrival and dissection at the labora- tory. Body surface, fins, gills, gill covers, and mouth were searched for parasites; the gill arches and single parasites fixed and preserved; and the parasites identified and counted. When parasites were too numerous for total counts, estimations of numbers per gill arch were made from counts per gill filament. Formalin, 2 AFA (alcohol-formol-acetic acid) fixative, and Bouin's solution were used to fix whole gill arches and trematodes. They were preserved in 70%ethanol. For the purpose of identification whole mounts were made of Trematoda using Harris hema- toxylin and Permount. Whenever possible, origi- nal descriptions of parasites were used for identi- fication together with Yamaguti's (1963, 1971) keys for identification of trematode genera. His- tological sections of 12 entire gill arches from 12 fish were examined. Arches were decalcified prior to embedding and cut at 8 /xm. Sections were mounted, stained with hematoxylin and eosin (H&E) and Periodic Acid Schiff (PAS). Histological techniques were after the method described by Humason (1972). Statistical evalua- tion of all station salinity data consisted of calcu- lations of standard deviation (Snedecor and Cochran 1967). RESULTS Water Quality Variations in salinity occur in Biscayne Bay from year to year because of climatic conditions. The salinity readings of all stations were similar during the dry season of 1976, mainly January to June (Table 2). Maximum salinities in both the bay and canal entrances were 40-417.. at this time. More freshwater discharge into the canals and Biscayne Bay during the rainy season in the fall accounted for a slight drop in salinity and some fluctuation mainly in the canals at that time. The lowest salinity reading from surface water samples from the entrance of Moody Canal in September indicated that freshwater dis- charge was more noticeable in this narrow canal than in the others. Some salinity measurements were taken immediately after freshwater dis- charge (see Table 2, footnotes). A typical reading showed that the fresh water flowed as a shallow surface layer about 30 cm deep out of the canals. During freshwater discharge, salinity at the sur- face varied from 57.. to 157..; at a depth of 30 cm it rose by 15-207.., and at a depth of 1 m it was close to the reading before the discharge, indicating that there was little vertical mixing. The statistical analysis of monthly averages of salinity data of all stations of 1.0 m depth showed that two-thirds of the values fell within one stan- dard deviation of 1.85 on each side of the mean of 36.87... Temperatures reflected seasonal changes at all stations and were similar, with most values between 20° and 30°C (Table 2). Differences may reflect the time of day when readings were taken. Dissolved oxygen concentration fluctuated mainly at canal entrances. Values ranged from 4 ppm to above 8 ppm (Table 2). Values below 6.8 ppm did not occur in southeast Biscayne Bay. Phosphates, ammonia, nitrites, and nitrates present at the collection sites from May 1975 to August 1976 are listed in Table 3. The southeast Biscayne Bay water quality data were similar to those of de Sylva 3 and Bader and Roessler 4 . In general, southeast bay values were low in all 2 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 3 de Sylva, D. P. 1970. Ecology and distribution of post- larval fishes in southern Biscayne Bay, Florida. Univ. Miami 272 SKINNKR: INTERRELATION OF WATFR QUALITY. GILL PARASITES, AND OILL I'ATIIOLOOY — M ■*-> co >> co CQ o> c & o en s en o> •S j= 3 O CO XI C aS & 03 4) C oa en jj CO £ © 3 i — i o en x C -j = 01 o. +J T3 a c eS _rt fc ■/ t- ■5 -*j c c Hi u co c o o >> — p O P a ■— ' — c c 0> US bi >. >> X •~ o Z ~ CD j> § "3 en .ii Cfl 0> 0> -z s~ tu — ' !_ 3 -*-> S3 •~ OJ a o 0) be ot) t- 0) > DC] _>, c o CM w J oa < I Q. „ E CL d d) > O < D d> > < o c o 5 a. E 01 1- X re 5 6 z eji > < _ • c '— 5 >~ E 03 CO X re 5 6 z ~~- C .2 CD CO « o n o _i I Q. E o. Q. > O < D O) > < O c o 5 a. E tg X re 5 o Z D) > < C *-" 5 >. E re CO X re 5 o Z -~. c o a, re re O T3 o - ' SSSNNCONSCOCONS cgrroiOiDCDOJOJO C0COCOCOC\IC\jC\JC\JCNiCvJC\JC\l oO'-oooin'-'-coifitoo) COCOCOCOC\IC\JCNC\JC\JC\JC\)C\J COCOCOCOCOCSJC\JC>iO0C\JC\JC\J COCOCO'^C^CNJ'^-^tCOCO'— C\J mmif)^"coa>coooa)o>ooLn CNicvjcompCNjpcqo c\j cq cnj ^t -«t '^t -^ (b s cb cd tb ui tj tj o i a)--oos(OOwm(D(Oo C\J I WCOCOCOC\IC\JC\JC\JC\iC\jC\JCO .*-C\JTtCMCM | i- CO CNJ C\J »- CNJ S^>i?fCi"iPcDcococccofS o 13 o o 5 CO CO CO CO CO CO CD r>-OtQ.>CJri}t:'-2 > §33ll2<2- m m£ir>£'£cni £ > <0<0 jc\jcsjc\icNjc>jcocococo CVJCOCOCOCOt-t-COCNJOJCOCOCOCOCVJ TrcocMcocoiooiLnoicoqw't inu^i^r^c\ji-CM^(Ococ\jcocp ^^^^^to^iricb^ririifico a>o^O'-'-cj>r^cNiojcoiO(or s - C>JC>JCOCOCOCVlCNiC\JC>JCMC>JC>JCVJ oiaioO'-oscNi-coinmin CNC\JC0C0C0C\ICNIC\JC\JC\JCS1CNJC\J a>CT>o*-*-r-c\jcvj(omr--co C\JC\JCOCOCOCNJC\JC\ICNJ rjia^ > 6 3 3 3 ec um<0)OZQ CO § re o.re ? u. S < S 5 -^ TtLf)uoco^rcDtcocor^cocococJ> cw cococococococococococococo 6 y TfNO)COCO(OCOCOCOCOOO>0 ™ COCOCOCOCOCOCOCOCOCO^CO'J m n'-CNTtCM^f'-COCMCO'-COCVJOvJ CD eg cjo^^ro^^^-o^^^oioio^o^oi m ^c >, cJ)Q.--'>cj c fjii;'-S' ?5^ 5 o 5 E 5 o o 5 c OJ Q. o 5 c o to c OJ QJ CO C a eg o re a CO *" o LO a 01 CO c 3 CO Cr re cr CD c "O cn co ID) ci) c OJ c c "D re > a -a re c eg OJ > a >> >. CO c c c re in ai re LO en LO to in co m 0) eg r^ r^ a O) o> E (^ QJ > >^ 3 U 3 _j u> d> —3 ~3 3 < G r^ O o CO ir> rg n « 273 FISHERY BULLETIN: VOL. 80. NO. 2 Table 3.— Monthly nutrient 1 values (mg/1) in southeast and southwest Biseayne Bay, Fla., locations (at depth 0.3-1.0 m) (Dade County Department of Environmental Resources Management unpubl. data). Location/ Location/ date PC TOC TAN N0 2 N0 3 date PO4 TOC TAN N0 2 NO3 SE. Biseayne Bay at Elliot Key S.W Biseayne Bay at Mowry Canal July 1975 0.00 7.0 0.00 0.00 0.00 May 1975 0.02 — 0.48 0.01 0.00 Sept. 1975 0.00 1 0.00 0.00 001 June 1975 0.00 6.0 0.44 0.01 0.20 Dec. 1975 0.02 9.0 00 0.00 0.07 July 1975 000 1.0 030 002 0.12 Feb. 1976 2 0.05 120 0.00 0.003 0002 Sept. 1975 0.00 1.0 0.10 0.01 0.02 Apr. 1976 0.00 0.00 0.00 0.002 Oct. 1975 0.01 3.0 0.05 0.02 0.68 June 1976 0.017 0.00 002 0025 Dec. 1975 0.01 1.0 0.04 001 0.50 July 1976 3 0.338 0.224 0.012 000 265 Jan. 1976 0.11 3.0 0.04 0.03 0.59 Aug. 1976 3 0.380 0336 0.012 000 0.168 Feb. 1976 2 0.40 30 0.05 0.01 0.48 Mar. 1976 0.07 — — — — S.W. Biseayne Bay at Black Creek Canal Apr. 1976 0.01 2.0 0.03 0.01 0.17 May 1975 0.32 7.0 0.10 0.01 0.01 May 1976 0.05 3.0 0.04 002 0.70 June 1975 0.64 7.0 0.48 0.02 0.30 Aug. 1976 3 0.01 6.0 0.03 0.01 0.32 July 1975 0.12 4.0 0.72 0.21 0.48 Aug. 1975 0.48 — — — S.W Biseayne Bay at Moody Canal Sept 1975 0.40 60 0.59 0.10 056 May 1975 000 — 0.06 0.00 0.00 Oct. 1975 0.25 4.0 0.45 0.06 0.43 June 1975 002 10 0.04 0.00 0.01 Nov. 1975 0.09 — — — July 1975 0.00 — — — — Dec. 1975 0.25 40 0.44 0.24 1.5 Sept. 1975 0.01 — — — — Jan. 1976 0.00 40 1.0 030 2.1 Oct. 1975 0.02 1.0 0.06 0.02 0.32 Feb. 1976 2 0.50 60 0.51 11 1.8 Dec. 1975 0.01 3.0 0.04 0.02 0.48 Mar. 1976 0.18 — — — Jan. 1976 0.00 3.0 0.04 003 0.50 Apr. 1976 — 6.0 0.03 0.16 1.2 Mar. 1976 0.14 3.0 003 0.01 0.30 May 1976 1.0 0.32 0.03 0.36 Aug. 1976 3 0.04 11.0 0.31 0.01 008 'TOC = Total organic carbon, TAN and in suspension. 2 Arsenicker Keys. 3 Sands Key. Total ammonia nitrogen. The term "total" refers to the amount present both in solution nutrients. Nutrient concentrations were consid- erably higher in the southwest, especially am- monia values. The water sample taken in April 1976 near Arsenicker Keys contained high total organic carbon compared with the other samples because of the proximity of an extensive man- grove coastline and the presence of mangrove detritus in the water. The two July 1976 water samples from Sands Key were taken in a canal and small lagoon inside the Key surrounded by mangroves and connected to the bay. At low tide, about two-thirds of the bottom muds of the lagoon Sch. Mar. Atmos. Sci., Prog. Rep. Fed. Water Qual. Admin., 198 p. 4 Bader, R. G., and M. A. Roessler. 1971. An ecological study of south Biseayne Bay and Card Sound, Florida. Rosen- stiel Sch. Mar. Atmos. Sci., Univ. Miami. were exposed, the canal was rich in fish, and wading birds fed in the flats at low tide. The somewhat higher content of ammonia, nitrates, and phosphates was due to decaying vegetation, the exposed mud flats, animal concentrations, and little flushing. Trace metals were present in water samples from the southwest locations only (Table 4). None were detected with standard methods in water samples from the southeast bay. Those pesticides either present or not de- tected in the water in Black Creek during the time of this study are shown in Table 5. None were detected with standard methods in the southeast bay. As in all the other southeast bay samples, no pesticides or heavy metals were de- tected in Sands Key samples. The junction of Black Creek and Goulds Canals and the south- Table 4.— Potentially harmful trace metals (m g/1, total ) in southwest Biseayne Bay locations of Black Creek and Mowry Canals, Fla., from May 1975 to May 1976 (USGS unpubl. data). Black Creek Canal Mowry Canal Hazard Minimal risk May Oct Jan Apr. May Oct Jan. Apr marine marine 1975 1975 1976 1976 1975 1975 1976 1976 biota 1 biota' Source As 2 2 — 2 — 1 1 1 50 10 Paints; pesti- cides; industry Pb 4 7 19 20 ~ 7 9 38 50 10 Gasoline fuel; industry Mn 4 20 — 100 20 Industry, paints Hg 02 02 02 0.5 0.1 02 06 0.1 Pesticides; paint plastics and paper industry Zn 2 2 20 — — — — 100 20 Plating industry 'Natl Acad Sci., Natl. Acad Eng . Environ Stud Board (1972) 274 SKINNER: INTERRELATION OF WATER QUALITY, GILL PARASITES. AND GILL PATHOLOGY Table 5.— Pesticides (/ug/1) in southwest Biscayne Bay location of Black Creek Canal, Fla., from July 1975 to August 1976 (USGS unpubl. data). 1 Date Diazinon 2 2,4-D 3 Silvex 4 Parathion 5 July 1975 Dec 1975 Aug. 1976 0.02 0.06 0.00 0.00 0.00 0.27 002 0.00 10 0.02 0.00 'Pesticides not found present in Black Creek Canal water samples were: Aldrin. Chlordane. DDD, DDE. DDT. Dieldrin. Endrin, Ethion, Heptachlor, Heptachlorepoxide. Lindane. Malathion, Methyl-parathion, Methy Itrithi- on, PCB, Toxaphene. Trithion. and 2,4,5-T. 2 0,0-Diethyl 0-(2-isopropyl-6-methyl-4-pynmidinyl) phosphorothioate 3 2,4-Dichlorophenoxy (acetic acid). 4 2-(2.4.5-Trichlorophenoxy) propionic acid. 5 0,0-Diethyl-0-p-nitrophenyl phosphorothioate west bay was found to be the highest in nutrients and trace metals. Direct sources of pollution may have been waste discharge from boats, marinas, agriculture, suburban developments, and the nearby county dump. Slightly lesser amounts were found at the Moody and Mowry stations which were located some distance from inhabited areas. Pesticide data were not available from the Moody and Mowry stations, although chemical pest and weed control conducted at the time along the banks and in the vicinity of the canals would have been a direct source of pesticides in the water. Parasites The parasite fauna in both the southeast and southwest Biscayne Bay habitats was similar in kind for all three hosts, consisting mainly of pre- viously reported ectoparasites of marine fishes of the same and related species or those sharing similar habitats. The three monogenetic gill parasites — Neodiplectanum wenningeri, Ancyro- cephalus sp., and A. parvus— showed close to 100% incidence and were therefore suitable for this study. Incidence of infestation was as fol- lows: N. wenningeri on G. cinereus, 97% in south- east Biscayne Bay, 100% in southwest Biscayne Bay locations; Ancyrocephalus sp. on L. griseus, 100% in southeast Biscayne Bay, 100% in south- west Biscayne Bay; A. parvus on S. timucu, 100% in southeast Biscayne Bay, 100% in southwest Biscayne Bay. The difference in intensity of infestation of hosts by these parasites was striking, with few parasites on host gills from the southeast loca- tions and extremely large counts on hosts from the southwest locations (Table 6). Pathological Changes in Host Gills Neodiplectanum wenningeri created compara- tively little histological disturbance of the gills when infestation was light. Damage was often mechanical and gill lamellae were deflected. In severe cases of infestation, however, the lamellae were covered with N. wenningeri, and an in- crease in mucus production was noticed along with clubbing of filaments where parasites were attached. Similarly, when numerous, Ancyro- cephalus sp. and A. parvus caused pathological changes at the site of attachment. Localized host reaction to the parasites' hooks included epithe- lial hyperplasia and heavy mucus production (Fig. 2), and the respiratory epithelium was lost in some instances. Often the side of the filament opposite the worm attachment was also affected Table 6.— Averages of some nutrients, trace metals, and pesticides in water samples, and Monogenea and gill pathology of the three host species in southeast Biscayne Bay and southwest Biscayne Bay at Black Creek Canal from May 1975 to August 1976. Southeast E iiscayne Bay Sol ithwest Biscayne Bay at Black Creek Canal No. of No of Component Min. Avg. Max. samples Mm. Avg. Max. samples Total ammonia nitrogen mg/l 000 0.00 0012 6 003 045 1.0 11 Arsenic fjg/\ 00 000 0.00 6 2.0 20 2.0 3 Lead jug/1 0.00 0.00 000 6 4.0 12.5 20.0 4 Manganese /jg/l 000 0.00 0.00 6 4.0 60 20.0 4 Mercury fjg/\ 0.00 0.00 0.00 6 0.2 0.28 0.5 4 Diazinon fjg/\ 0.00 000 0.00 6 0.02 0026 006 3 2,4-D fjg/\ 0.00 0.00 0.00 6 000 0.09 0.27 3 Silvex /jg/l 0.00 0.00 0.00 6 000 0.04 0.10 3 Parathion /L/g/l 0.00 0.00 0.00 6 0.00 0.01 0.02 2 Neodiplectanum wenningeri no./gill arch 0.00 0625 5 69 25 725 >100 52 Ancyrocephalus sp no./gill arch 0.1 1.4 8 57 69 124.75 >500 80 Ancyrocephalus parvus no./gill arch 0.3 2.25 4.5 44 61 8925 >200 54 Pathological changes' None None Slight 170 Moderate Severe Severe 186 Slight = mucus production above normal: moderate of lamellae, loss of structure. heavy mucus production and epithelial hyperplasia; severe = fusion 275 FISHERY BULLETIN: VOL. 80. NO. 2 Figure 2.— Photomicrograph of Ancyrocephalus sp. on the gills of Lutjanus griseus, 22.0 cm SL, from southwest Biscayne Bay, Fla. (PAS, 75X) showing hyperplasia, loss of respiratory epithelium, excessive mucus, lamellar fusion, and aneurisms, a) para- site; b) mucus. in a similar manner (Fig. 3). In addition to injury caused by the hooks of the parasite, the lamellae were deflected and adhered to each other, thus reducing the gill surface effective for gas ex- change. In severe cases when a number of worms were attached to the tips of filaments, clubbing of filaments was almost always present, as was obliteration of normal filament structure. The affected filaments appeared white in fresh prep- arations and the gills were congested with mu- cus. Histological changes of the gills not associated with parasites were found in hosts from south- west Biscayne Bay stations. Few southeast Bis- cayne Bay fish showed above-normal production of mucus in the gills. Increased mucus produc- tion was evident in all fish from the southwest locations, and pathological changes ranged from moderate to severe (Table 6). Abnormal color changes were frequent in southwest Biscayne Bay fish and were usually associated with over- production of mucus which congested the gills. Histological sections of gills from these fish showed that whole filaments were lined with mucus and that it filled the spaces between the filaments. Additionally, mucus-producing cells were concentrated, sometimes in several layers, at the tips of gill filaments which had lost their normal structure. Fusion of gill lamellae along entire filaments, epithelial hyperplasia, club- bing of lamellae or obliteration of lamellar struc- ture, aneurisms, and clubbing of filaments occurred frequently, along with proliferation of cells at the bases of lamellae. DISCUSSION According to Grundmann et al. (1976), hel- minth populations in a natural environment are well regulated to a point of host comfort. Although the results from the southeastern habi- tat in this study agreed with this statement, those from the southwest bay locations did not. Disease caused by parasites often requires exogenous as well as endogenous factors (Sindermann 1979). Exogenous factors, as defined by Cameron 276 SKINNER: INTERRELATION OF WATER QUALITY, GILL PARASITES. AND GILL PATHOLOGY (If (UE "■*)&*#> V a A. i3? .* ^r£ b : , x } * Figure 3.— Photomicrograph of two Aneyrocephalus partus on the gills of S. timncu, 24.3 cm SL, from southwest Biscayne Bay, Fla. (PAS, 300X). Lamellae are deflected and obstructed, a) and b) parasites. (1958), are alterations in the ecology of the para- sites or hosts by some abnormal or unnatural event, most often manmade. The most outstanding difference between the southeast Biscayne Bay and southwest locations was the difference in chemical water quality. According to Klontz (1972), fish are so intimately associated with their aqueous environment that physical or chemical changes in this environ- ment are often rapidly reflected as measurable physiological changes in the fish. In general, re- actions of fish gills to an irritant include inflam- mation, hyperplasia, lamellar fusion, excessive mucus production, clubbing of filaments or la- mellae, and formation of aneurisms. Aneurisms may be a specific tissue reaction due to injury or toxic substances, especially am- monia or herbicides in the water or food (Eller 1975). Ammonia frequently has been reported to cause extensive gill damage. Although much of the data on the degree of toxicity of ammonia is not satisfactory (National Academy of Sciences Environmental Studies Board 1972), it has been shown that the more toxic component of ammonia solutions is the unionized ammonia (NH3). An in- crease in pH from the normal level increases the toxicity, because along with temperature it con- trols the degree of dissociation (Trussel 1972). A decrease in dissolved oxygen concentration in- creases the toxicity of unionized ammonia (Na- tional Academy of Sciences Environmental Studies Board 1972). Even low concentrations may cause pathological changes in marine and freshwater organisms (Doudoroff and Katz 1950; Flis 1968; Larmoyeux and Piper 1973). In addition to exhibiting gill damage, after expo- sure to NH 3 freshwater fish were susceptible to ectoparasites, according to Reichenbach-Klinke (1966). Prolonged exposure to nonlethal dosage of ammonia in salmon led to hyperplasia of gill epithelium and epizootic bacterial gill disease in a study by Burrows (1964). Pollutants such as metals and pesticides show similar effects on fish gills (Gardner 1975). The LC50 and sublethal effects of pesticides are pres- ently under scrutiny. According to Anderson (1971), for pollutants not to influence the physiol- ogy and behavior of fish, "safe" concentrations 277 FISHERY BULLETIN: VOL. 80, NO. 2 should be 0.01-0.05 of the lethal concentrations. Since a small rise in temperature or salinity can shift the LC50 by one order of magnitude (Eisler 1972), most pesticides may be more harmful than previously assumed. A synergistic effect of sev- eral sublethal concentrations of pollutants is pos- sible. They may exist in such low concentrations that conventional analysis or collection methods will not detect them, especially herbicidal con- taminants. However, Seba and Corcoran (1969) found that surface slicks formed by a film of or- ganic matter concentrated pesticides in south- west Biscayne Bay to detectable levels, up to 137 times as much as slicks in the Florida Current. Although the reaction of gill tissue to toxic chemicals appears to be nonspecific in regard to the particular chemicals present, and it is there- fore difficult to indict any one particular com- ponent or group of components in nature, the overall result of gill damage is impairment of function. Regardless of cause, pathological changes reduce the useful respiratory surface and make gas exchange difficult, which stresses the fish and eventually weakens it. Disease has been known to change behavior in fish (National Academy of Sciences 1973) and in- fluence their chance for survival. Impaired func- tion of an organ and reduced efficiency require expenditure of energy which cannot be used for other life processes such as feeding, reproduc- tion, and predator avoidance. In case of gill dam- age, metabolic activity must be reduced to a minimum in order to reduce oxygen demand (Wedemeyer et al. 1976), and the fish become weakened and stressed. Selye's (1950) definition of stress was used in reference to fish by Wede- meyer (1970): "the sum of all the physiological responses by which an animal tries to maintain or reestablish a normal metabolism in the face of a physical or chemical force." Unfortunately, some of the metabolic changes may also contrib- ute to increased susceptibility to disease (Wede- meyer et al. 1976). When fish are weakened by environmental fac- tors, chemicals, or poor nutrition, their resist- ance to infestation and infection by Monogenea, Trichodina, and bacteria is reduced (Schaper- claus 1954; Wedemeyer et al. 1976). These facts are well known to the aquaculture and aquarium industries. Most research on immune reactions is done in human and veterinary medicine, but parallels can be drawn since fishes' immune sys- tems, although less advanced, resemble those of other vertebrates (Sindermann 1970). Mucus antibody may be active against some external in- festations (Anderson 1974); thus, a parasite must be able to avoid the immune reaction of the host (Williams 1970). Stress-provoked physiological changes may cause a disturbance of the host's im- mune system, and damaged or irritated gills can then become heavily infested with parasites. Snieszko (1974) shared the belief of other scien- tists that the aggravating effect of stress from various types of pollution caused a high inci- dence of infectious disease in fishes, and men- tioned that this belief, unfortunately, was not yet adequately documented. Sindermann (1979) summarized some of the recent supporting evi- dence that toxins have a deleterious effect on the immune response of fishes. This study of Bis- cayne Bay fishes suggests that, in the presence of sublethal quantities of pollutants in a natural marine environment, fish suffered from gill damage which produced stress, physiological and physical compensation, leading to weaken- ing, reduced immunity, and heavy parasitic in- festation. ACKNOWLEDGMENTS I thank Edwin S. Iversen, Eugene F. Corcoran, and Donald P. de Sylva of the Rosenstiel School of Marine and Atmospheric Science, University of Miami; and George T. Hensley and Lanny R. Udey of the School of Medicine, University of Miami, for their advice and assistance during this study. Special thanks go to James T. Tilmant, Bis- cayne National Monument; James F. Redford, Jr., Dade County Commissioner; Henry J. Schmitz and Edward Gancher, Dade County De- partment of Environmental Resources Manage- ment; Robert L. Taylor, South Florida Water Management District; and Keith Dekle, Florida Power and Light Company, for help with field work and obtaining water quality data. I am grateful to Fay Mucha, Rosenstiel School of Marine and Atmospheric Science, University of Miami; and Elaine Kraus, Medical School, University of Miami, for assistance with histo- logical work and photomicrography. Funds were provided by the Rosenstiel Fund, University of Miami, RSMAS. The Richard G. Bader Memorial Student Fund supplied a gill net and photographic material. This is a contri- bution from the Rosenstiel School of Marine and Atmospheric Science, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149. 278 SKINNER: INTERRELATION OF WATER QUALITY. GILL PARASITES. AND GILL PATHOLOGY LITERATURE CITED Anderson, B. G., and D. L. Mitchum. 1974. Atlas of trout histology. Wyo. Game Fish Dep., Cheyenne, 110 p. Anderson, D. P. 1974. Diseases of fishes. Book 4; Fish immunology. T.F.H. Publ., Neptune City, N.J., 239 p. Anderson, J. M. 1971. II. Sublethal effects and changes in ecosystems. Assessment of the effects of pollutants on physiology and behaviour. Proc. R. Soc. Lond. B. Biol. Sci. 177:307- 320. Bullock, G. L., D. A. Conroy, and S. F. Snieszko. 1971. Bacterial diseases of fishes. T.F.H. Publ.. Nep- tune City, N.J., 151 p. Burrows, R. E. 1964. Effects of accumulated excretory products on hatchery-reared salmonids. U.S. Fish Wildl. Serv., Res. Rep. 66, 12 p. 1972. Salmonid husbandry techniques. In J. E. Halver (editor), Fish nutrition, p. 375-402. Acad. Press, N.Y. Cameron, T. W. M. 1958. Parasites and parasitism. Methuen & Co., Ltd., Lond., 322 p. Carter, L. J. 1974. The Florida experience: land and water policy in a growth state. Johns Hopkins Univ. Press, Baltimore, 355 p. Cervigon, F. 1966. Los peces marinos de Venezuela. Vol. I and II. Fund. La Salle de Ciencias Naturales, Caracas, 951 p. Cressey, R. F., and B. B. Collette. 1971. Copepodsand needlefishes: a study in host-parasite relationships. Fish. Bull., U.S. 68:347-432. DOUDOROFF, P.. AND M. KATZ. 1950. Critical review of literature on the toxicity of in- dustrial wastes and their components to fish. I. Alkalies, acids, and inorganic gases. Sewage Ind. Wastes 22:1432-1458. Eisler, R. 1972. Pesticide-induced stress profiles. In M. Ruivo (editor), Marine pollution and sea life, p. 229-233. FAO Fish. News (Books) Ltd., Lond. Eller, L. L. 1975. Gill lesions in freshwater teleosts. //; W. E. Ribe- lin and G. Migaki (editors), The pathology of fishes, p. 305-330. Univ. Wis. Press, Madison. Flis, J. 1968. Anatomicohistopathological changes induced in carp (Cyprinus carpio L.) by ammonia water. Pt. I. Effects of toxic concentrations. Pt. II. Effects of subtoxic concentrations. Acta Hydrobiol. 10:205-238. Gardner, G. R. 1975. Chemically induced lesions in estuarine or marine teleosts. In W. E. Ribelin and G. Migaki (editors), The pathology of fishes, p. 657-693. Univ. Wis. Press, Madi- son. Grundmann, A. W., R. G. Warnock, and D. L. Wassom. 1976. Some mechanisms of natural regulation of para- sitic helminth populations. Am. Midi. Nat. 95(2):347- 360. Herbert, D. W. M., and D. S. Shurben. 1964. The toxicity to fish of mixtures of poisons. I. Salts of ammonia and zinc. Ann. Appl. Biol. 53:33-41. Hoffman, G. L. 1976. Fish diseases and parasites in relation to the envi- ronment. Fish Pathol. 10:123-128. Humason, G. L. 1972. Animal tissue techniques. 3d ed. W. H. Free- man and Co., San Franc, 641 p. Iversen, E. S., N. Chitty, and N. Van Meter. 1971. Some Myxosporida from marine fishes in south Florida. J. Protozool. 18:82-86. Johnson, D. W. 1968. Pesticides and fishes — a review of selected litera- ture. Trans. Am. Fish. Soc. 97:398-424. Klontz, G. W. 1972. Haematological techniques and the immune re- sponse in rainbow trout. Symp. Zool. Soc. Lond. 30:89- 99. Larmoyeux, J. D., and R. G. Piper. 1973. Effects of water reuse on rainbow trout in hatcher- ies. Prog. Fish-Cult. 35:2-8. Lowe, J. I. 1964. Chronic exposure of spot, Leiostomus xanthurus, to sublethal concentrations of toxaphene in seawater. Trans. Am. Fish. Soc. 93:396-399. Mitrovic, V. V. 1972. Sublethal effects of pollutants on fish. In M. Ruivo (editor), Marine pollution and sea life, p. 252-257. FAO Fish. News (Books) Ltd., Lond. National Academy of Sciences. 1973. Aquatic animal health. Natl. Acad. Sci., Wash., D.C., 46 p. National Academy of Sciences, Environmental Studies Board. 1972. Water quality criteria 1972. Environ. Prot. Agency, Wash., D.C., 594 p. O'Connor, J. S. 1976. Contaminant effects on biota of the New York Bight. Proc. 28th Annu. Gulf Caribb. Fish. Inst. p. 50- 63. Osburn, R. C. 1911. The effects of exposure on the gill filaments of fishes. Trans. Am. Fish. Soc. 40:371-376. OVERSTREET, R. M., AND H. D. HOWSE. 1977. Some parasites and diseases of estuarine fishes in polluted habitats of Mississippi. Ann. N.Y. Acad. Sci. 298:427-462. Pippy, J. H. C, and G. M. Hare. 1969. Relationship of river pollution to bacterial infec- tion of salmon (Salmo salar) and suckers {Catostomus commersoni). Trans. Am. Fish. Soc. 98:685-690. Plehn, M. 1924. Praktikum der Fischkrankheiten. E. Schweiz. Velagsbuchhandlung, Stuttgart, Germany. 479 p. Randall, J. E. 1968. Caribbean reef fishes. T.F.H. Publ., Neptune City, N.J., 318 p. Reichenbach-Klinke, H.-H. 1965. Der Phenolgehalt des Wassers in seiner Auswirk- ung auf den Fischorganismus. Arch. Fischereiwiss. 16:1-16. 1966. Untersuchiingen uber die Einwirkung des Am- moniakgehalts auf den Fischorganismus. Arch. Fisch- ereiwiss. 17(2):122-132. Schaperclaus, W. 1954. Fischkrankheiten. Akad.-Verlag, Berlin, 708 p. Seba, D. B., and E. F. Corcoran. 1969. Surface slicks as concentrators of pesticides in the 279 FISHERY BULLETIN: VOL. 80. NO. 2 marine environment. Pestic. Monit. J. 3:190-193. SlNDERMANN, C. J. 1970. Principal diseases of marine fish and shellfish. Acad. Press, N.Y., 369 p. 1979. Pollution-associated diseases and abnormalities of fish and shellfish: A review. Fish. Bull., U.S. 76:717- 749. Smith, C. E., and R. G. Piper. 1975. Lesions associated with chronic exposure to ammo- nia. In W. E. Ribelin and G. Migaki (editors). The pathology of fishes, p. 497-514. Univ. Wis. Press, Madi- son. Snedecor, G. W., and W. G. Cochran. 1967. Statistical methods. 6th ed. Iowa State Univ. Press, Ames, 593 p. Snieszko, S. F. 1974. The effects of environmental stress on outbreaks of infectious diseases of fishes. J. Fish Biol. 6(2):197- 208. Starck, W. A., II. 1971. Biology of the gray snapper, Lutjanusgriseus (Lin- neaus), in the Florida Keys. Stud. Trop. Oceanogr. 10: 11-23. Univ. Miami Press. Trussel, R. P. 1972. The percent un-ionized ammonia in aqueous am- monia solutions at different pH levels and temperatures. J. Fish. Res. Board Can. 29:1505-1507. Waite, T. D. 1976. Man's impact on the chemistry of Biscayne Bay. In A. Thorhaug and A. Volker (editors), Biscayne Bay: Past/Present/Future, p. 279-285. Univ. Miami Sea Grant Spec. Rep. 5. Walsh, A. H., and W. E. Ribelin. 1975. The pathology of pesticide poisoning. In W. E. Ribelin and G. Migaki (editors), The pathology of fishes, p. 515-541. Univ. Wis. Press, Madison. Wedemeyer, G. 1970. The role of stress in the disease resistance of fishes. In S. F. Snieszko (editor), A symposium on diseases of fishes and shellfishes, p. 30-35. Am. Fish. Soc, Spec. Publ. 5. Wedemeyer, G. A., F. P. Meyer, and L. Smith. 1976. Diseases of fishes. Book 5: Environmental stress and fish diseases. T.F.H. Publ., Neptune City, N.J., 192 p. Williams, H. 1970. Host-specificity of fish parasites. J. Parasitol. 56(2):482-483. Wobeser, G., L. F. Kratt, R. F. J. Smith, andG. Acompanado. 1976. Proliferative branchiitis due to Tetraonchus rau- chi (Trematoda: Monogenea) in captive arctic grayling (Thymallus arcticus). J. Fish. Res. Board Can. 33: 1817-1821. Yamaguti, S. 1963. Systema helminthum. Vol. IV. Monogenea and Aspidocotylea. Intersci. Publ., N.Y., 699 p. 1971. Synopsis of digenetic trematodes of vertebrates, Vol. I, 1074 p. Keigaku Publ. Co., Tokyo. 280 THE EFFECT OF PROTEASE INHIBITORS ON PROTEOLYSIS IN PARASITIZED PACIFIC WHITING, MERLUCCIUS PRODUCTUS, MUSCLE Ruth Miller and John Spinelli 1 ABSTRACT Since the enactment of the Fishery Conservation and Management Act of 1976, the U.S. fishing industry has intensified its interest in Pacific whiting, Merluccius products, as an additional food resource. In some fishing areas, Pacific whiting is infected with a protozoan parasite, Myxosporidia kudoa, which produces a proteolytic enzyme that degrades the textural quality of muscle as it is processed or cooked. Several enzyme inhibitors were evaluated for their potential to inactivate the enzyme, thereby preserving the texture of the fish during processing. It was found that protease inhibitors such as those found in egg white, potato, and soy and lima beans were ineffective as inhibitors. Compounds that react with sulfhydryl groups, on the other hand, were found to be active inhibitors. These compounds include hydrogen peroxide (free and alkaline), potassium bromate, iodoacetate, and N- ethylmaleimide. The most promising results were obtained with potassium bromate or combinations of dibasic phosphate peroxide and potassium bromate. These reagents mixed into ground parasitized pacific whiting muscle inhibited proteolysis sufficiently during frozen storage and later cooking to maintain texture comparable with nonparasitized fish. The Fishery Conservation and Management Act of 1976 has intensified the interest of the fishing industry in Pacific whiting, Merluccius produc- tus, as an additional food resource. Although Pacific whiting has been extensively fished by the Russian and Polish fishing fleets, it has attracted only slight commercial interest in the United States, primarily because its texture and color are somewhat less desirable than that of other gadoid species such as cod and haddock. In 1970, Dassow et al. observed that the textural change in cooked Pacific whiting was due to the presence of a protozoan parasite, Myxosporidia kudoa. This parasite produces a proteolytic enzyme capable of breaking the chemical bonds of the muscle fibers which are responsible for the characteristic texture of fresh fish. The activity of the enzyme increases as the temperature increases. Thus, during conventional processes such as baking, broiling, or pan frying, the gradual increase in heat enhances proteolysis until the product reaches the temperature of inactivation of the enzyme. One method of handling the problem of the parasitic enzyme is rapid cooking (deep-fat frying of sticks and por- tions) where the temperature of inactivation is achieved before proteolysis destroys the texture 'Utilization Research Division Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 2725 Montlake Boulevard East, Seattle, WA 98112. of the fish (Patashnik et al. 2 ). Another possibility would be to inactivate the enzyme with an inhibitor. In the work presented here, several enzymic inhibitors were evaluated to determine their effectiveness in inhibiting proteolysis in Pacific whiting muscle. The concentration of enzyme inhibitor sufficient to prevent organoleptic textural alteration was also determined. METHODS Pacific whiting were caught off the coast of Astoria, Oreg., by commercial trawlers, filleted and frozen within 24 h, and stored at — 20°C. The presence of the parasite was determined directly by visual evidence of black and white spores, by microscopic identification of the spores, or, indirectly, by baking a segment of muscle in a covered container for 20 min at 162°C. Soft or mushy muscle indicated the presence of the parasitic enzyme. To ascertain the effects of enzyme inhibitors under uniform conditions, tests for proteolytic activity were carried out on diluted blends of fish Manuscript accepted November 1981. FISHERY BULLETIN: VOL. 80. NO. 2. 1982. 2 Patashnik, M.. H.S. Groninger. H. Barnett, G. Kudo, and B. Koury. 1981. Pacific coast whiting (Merlucciuit productus). I. Abnormal muscle texture caused by myxosporidian-induced proteolvsis. In prep., 34 p. Northwest and Alaska Fisheries Center* Natl. Mar. Fish. Serv., NOAA, 2725 Montlake Blvd. E., Seattle. WA 98112. 281 FISHERY BULLETIN: VOL. 80, NO. 2 muscle and on ground (minced) muscle. Condi- tions for testing were kept close to those under which we knew the parasitic enzyme functioned. The pH was maintained at that of the fish (6.8), the substrate was the fish muscle, and the tem- perature was moderate (45°C). Blended Fish Blended fish muscle was prepared by blending two parts 0.1 M NaCl with one part ground fish in a Lourdes Blender 3 in a quantity large enough to serve for several tests. The pH (6.8) of the solutions of the various potential inhibitors was maintained by the addition of dilute NaOH or HC1. In a 50 ml polycarbonate tube, 2 ml of the blended fish was mixed with 1 ml 0.1 M NaCl, as a control, or with 1 ml of the potential inhibitor. The tubes, covered with parafilm, were incu- bated for 90 min at 45°C. Duplicate samples of the control and test material were kept at 0°C in order to know the soluble protein level before in- cubation. This figure was subtracted from the quantity of soluble protein that was the result of increased proteolysis in the incubated sample. The reaction was stopped by the addition of 3 ml of 10% trichloroacetic acid. After 30 min at room temperature, the tubes were centrifuged at 9,750 g for 10 min. Protein determinations by the Lowry method (Lowry et al. 1951) were done on 1 ml of the supernatant. The effectiveness of the inhibitor was gauged by comparison of the proteolysis of the control (0.1 M NaCl) with that of the potential inhibitor. Since over a period of time the amount of proteolysis was bound to vary, a control was run with each experiment. In order to calculate the amount of inhibition, an arbi- trary figure of 100% was assigned to the control and the effectiveness of the inhibitor was expres- sed as percent inhibition by the following formula: g protein/ml of test X 100 = % proteolysis g protein/ml of control 100 — % proteolysis = % inhibition. Ground Fish Ground fish was prepared by putting partially frozen fillets through a 4mm die. Ten parts of ground fish were thoroughly mixed with 1 part 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. of 0.1 M NaCl or the inhibitor solution. Three grams of this material was incubated in a 50 ml covered polycarbonate tube for 30 min at 45°C. The reaction was stopped by the addition of 3 ml of 10% trichloroacetic acid. The remaining treatment was the same as with the blended fish. Preparation of Ground Fish Blocks for Storage A quantity (about 200 g) of the ground para- sitized Pacific whiting was mixed with 0.1 M NaCl (approximately the ionic strength of muscle) as a control or an inhibitor in the ratio of 10 parts fish to 1 part solution. Before the blocks were placed in storage, aliquots were taken to test for inhibition and inhibitor residues. The blocks (3" X 1" X 8") were stored at -20°C for 1 mo. At the end of the month, aliquots were retested for inhibition and inhibitor residues. Effect of Proteolytic Inhibition on Texture The blocks of parasitized whiting made for the storage study and a similar block made from nonparasitized Pacific whiting were used to test the effectiveness of maintaining texture by inhibiting proteolysis. Duplicate portions (3" X 1" X Y 2 ") were cut from each block and baked in a covered dish (3 1 / 2 " X 2" X l 1 //'). The baked por- tions were randomly mixed before presenting them to an experienced panel for texture and organoleptic evaluation. In order to express the results in numerical values, numbers were assigned to the texture categories: firm (1); soft (2); mushy (3). Aliquots were taken at the same time to test for percent inhibition. Oxidative Effect on Amino Acids Amino acid analyses, using the Beckman 118 CL Amino Acid Analyzer (Spackman et al. 1958), were done on acid hydrolysates of non- parasitized fish, parasitized fish with no treatment, and parasitized fish which had been treated with either 0.5% disodium phosphate peroxide plus 0.025% potassium bromate or 0.5% dipotassium phosphate peroxide plus 0.025% potassium bromate. Enzyme Inhibitors All chemicals were of reagent grade. Trypsin 282 MILLER and SPINELLI: EFFECT OF PROTEASE INHIBITORS ON PROTEOLYSIS inhibitors were purchased from Sigma Com- pany. Dibasic phosphate peroxides were pre- pared in our laboratory according to the method of Nakatani and Katagiri (1970). The potato extract was prepared in our laboratory accord- ing to the method of Melville and Ryan (1972). Test for the Presence of Peroxides or Bromates The following method of measuring peroxides and bromates was adapted from two methods, that of Price and Lee (1970) and that of the Asso- ciation of Official Analytical Chemists handbook (1975): 4 ml H 2 1 ml of oxidant standard or 1 g fish 1 ml saturated KI 1 ml 0.001 M ammonium molybdate in 1 N H2SO4. Shake for 1 min, titrate to a light yellow with 0.1 N sodium thiosulfate, and add a few drops of 1% starch; continue titrating to the end point. Both hydrogen peroxide and potassium bromate liberate iodine by oxidation; therefore, this method can be used to indicate the presence of either one. Quantification was determined by comparison with a known standard expressed in milliequivalents. RESULTS AND DISCUSSION Tests with Blended Fish Blended fish was used to test a variety of potential inhibitors which are listed with concentrations and results in Table 1. The enzyme inhibitors tested included trypsin inhib- itors from four sources: soybeans, lima beans, turkey egg white, and chicken egg white. We also tested crude potato extract which has been shown to contain several protease inhibitors (Melville and Ryan 1972; Ryan et al. 1974; Bryant et al. 1976; Hass et al. 1976). None of the tested enzyme inhibitors caused significant in- hibition in concentrations that would be suitable for use in food systems. From the remaining potential inhibitors which included metal chelators, oxidizers, and sulfhydryl binding compounds, we found hydro- gen peroxide, potassium bromate, dibasic phos- phate peroxides, iodoacetate, and N-ethylmaleim- TABLE 1.— Protease inhibitors. Inhibitor Concentration Active site Effect' EDTA 0.3X10'' M 0.3X10 3 M 0.3X10 ' 5 M Chelates, Metals ± + + Sodium pyrophosphate 5 5 5 XXX CO CO CO 000 Mg, Mn, Zn, other metals ± ± Sodium oxalate 0.3X10 ■' M 0.3X10 ' 3 M Ca. Mg ± Cysteine 0.3X10 ' M 0.3X10 3 M 0.3X10 5 M Fe, Cu, other metals + ± ± o-Phenanthroline 0.3X10" 2 M 0.3X10" 4 M Fe. Co, Zn, other metals ± Sodium fluoride 0.3X1 0" 1 M 0.3X1 0' 3 M Mg, Ca. other metals ± Iodoacetate 0.3X10"' M 0.3X10' 2 M 2.0X10 2 M Sulfhydryls, imid- azoles, thio ethers - N-ethylmaleimide 0.3X1 0' 2 M 1.5X10 2 M 0.75X10 2 M Sulfhydryls - Hydrogen peroxide 1.0% 0.1% 0.5% Oxidizes - Disodium phosphate peroxide 0.3% 0.5% Oxidizes _ Dipotassium phosphate peroxide 0.3% 0.5% Oxidizes — Potassium bromate 0.05% 0.025% 0.001% Oxidizes - Soybean 1 mg/ml Trypsin ± Lima bean 5 mg/ml Trypsin ± Chicken egg white 5 mg/ml Trypsin ± Turkey egg white 5 mg/ml Trypsin ± Potato extract 2.5 mg/ml 5.0 mg/ml 10.0 mg/ml Chymotrypsin ± Carboxypeptidase ± Serine endopeptidase ± Metallocarboxy pep- tidase ± 'increased proteolysis change ±. +, decreased proteolysis — , no signficant ide to warrant further investigation. The reaction with iodoacetate and N-ethylmaleimide indicated that we were dealing with a thiol enzyme. Tests with Ground Fish Both hydrogen peroxide (H2O2) and potas- sium bromate (KBrOa) are currently being used in the U.S. food industry to impart desired func- tional and organoleptic properties to the foods to which they are added. For example, KBr0 3 is used in breadmaking to improve the physical properties of the dough (Tsen 1968). H2O2 has been used as a preservative in dairy products (Cuq et al. 1973) and as a bleaching agent in some fish products (Sims et al. 1975; James and McCrudden 1976). The dibasic phosphate perox- ides have been used as a stablilizer for H2O2 in various food products such as soy products, meat, fish, and cereals (Pintauro 1974). 283 FISHERY BULLETIN: VOL. 80. NO. 2 After testing for inhibition effects in the (model) blended system, tests were conducted on ground (minced) parasitized Pacific whiting to test those which demonstrated inhibitory potential and could be used in food systems. Hydrogen Peroxide In the ground parasitized Pacific whiting, hydrogen peroxide was significantly less effec- tive in inactivating the proteolytic enzyme than it had been with the blended fish. This was explained by the fact that catalase is known to be present in muscle to destroy hydrogen peroxide formed in aerobic muscle fiber (Deisseroth and Dounce 1970). There was a difference between the blended and ground muscle both in protein concentration and distribution of the catalase. In order to demonstrate the difference more specif- ically, we compared the protein concentration and the catalase activity in the two systems. Pro- teins were determined by the macro-Kjeldahl, percent protein N method. Catalase activity was determined by measuring the disappearance of peroxide residues after 0.3% H2O2 (0.146 meq) was mixed with 1 g of blended or ground fish. The results in Table 2 show 40% less protein, which includes catalase, in blended fish than in ground fish. When 0.3% H1O2 was added to the blended fish, hydrogen peroxide was more slowly de- graded and thereby had longer contact time with the enzyme of the parasite. The location of the catalase was shown by washing out all intercel- lular catalase from ground muscle, then ^in- stituting the catalase activity by crushing or manipulating the washed muscle fibers. A con- centration of 3% H 2 2 was needed to counteract all catalase activity, but a concentration of this magnitude also destroyed the tissue structure. It was obvious that hydrogen peroxide alone would be impractical to use as a protease inhibitor. Potassium Bromate Because of the difference in protein concentra- tion in ground fish, it was necessary to increase the concentration of potassium bromate from Table 2.— Comparison of protein concentration and peroxide residues in blended or ground parasitized Pacific whiting. 0.01% to 0.05% in order to achieve a 63-66% in- hibition of proteolytic activity. This was shown to be sufficient to maintain the texture of para- sitized Pacific whiting. Tsen (1968) suggested that there was a syn- ergistic effect between potassium bromate, a slow oxidizer, and faster oxidizers such as iodates, acetone peroxide, or azodecarbonamide; therefore, potassium bromate was tested with hydrogen peroxide in varying concentrations. The results were not synergistic but 0.025% KBr0 3 with 0.5% H2O2 was as effective as 0.05% KBr0 3 (Table 3). Table 3.— Effect of hydrogen peroxide and potassium bromate on proteolysis in ground parasitized Pacific whiting. Oxidant % inhibiton Control— no treatment 0.5% H 2 2 43 0.05% KBr0 3 63 0.025% KBrOs 47 0.01% KBr0 3 35 0.05% KBr0 3 in 0.5% H 2 2 66 0.025% KBr0 3 in 0.5% H 2 2 64 Treatment Percent protein N % peroxide residues remaining of fish time 5 mm Blended fish Ground fish 988 16.51 100 (0.146 meq) 28 (0.041 meq) 76 (0.1 10 meq) 9 (0.010 meq) Dibasic Phosphate Peroxides The adduct of hydrogen peroxide with dibasic phosphates has been found to facilitate the use of hydrogen peroxide by stabilizing it in food sys- tems (Pintauro 1974). It seemed possible that these compounds might protect hydrogen perox- ide from catalase long enough for it to be effective in inhibiting proteolysis. We tested 0.3% and 0.5% of both disodium phosphate peroxide (Na2HPCv H2O2) and dipotassium phosphate peroxide (K2HPO4H2O2) with ground parasitized Pacif- ic whiting. When these compounds were compared in terms of milliequivalents of perox- ides with equivalent concentrations of hydrogen peroxide alone, disodium phosphate peroxide had 23% milliequivalents of peroxide and dipo- tassium phosphate peroxides 15%. The dipotas- sium phosphate peroxide seemed less stable than disodium phosphate peroxide judging from its effervescence. Both dibasic phosphate peroxides were tested alone and with potassium bromate (Table 4). As found earlier in combination with hydrogen peroxide, 0.025% KBr03 enhanced the proteolytic inhibition of both concentrations of dibasic phosphate peroxides which meant effec- tive inhibition could be achieved with lower con- centrations of each of the oxidants. The results of testing these inhibitors estab- lished concentrations and combinations which 284 MILLER and SPINKLLI: EFFECT OF PROTEASE INHIBITORS ON PROTEOLYSIS Table 4.— Effect of dibasic phosphate peroxide on proteolysis in ground parasitized Pacific whiting- Oxidant % inhibition Control— no treatment 0.3% Na 2 HP0 4 -H 2 2 35 3% K 2 HP0 4 H 2 2 9 0.3% Na 2 HP0 4 H 2 2 + 0.025% KBr0 3 64 3% K 2 HPO„-H 2 2 + 0.025% KBr0 3 62 5% Na 2 HP0 4 H 2 2 45 0.5% K 2 HP0 4 -H 2 2 24 5% Na 2 HP0 4 H 2 2 + 025% KBrOa 73 5% K 2 HP0 4 -H 2 2 + 0.025% KBr0 3 67 were effective in inactivating the parasitic enzyme in parasitized Pacific whiting. We then determined whether 1) the inactivation would be maintained during a freeze-thaw cycle after 1 mo of storage at — 20°C, 2) inactivation was suf- ficient to maintain a desirable texture, and 3) the treatment with oxidizing agents would adverse- ly affect the amino acids, thereby decreasing the nutritional quality of the protein. Effect of Frozen Storage The prolonged effect of frozen storage on in- hibition was determined on samples of ground parasitized Pacific whiting treated with various inhibitors. Aliquots of these samples were tested at the time of preparation for percent inhibition and the presence of oxidant residues. All samples were stored at — 20°C for 1 mo at which time these tests were repeated, and as the results show in Table 5 there was no decrease in the inhibition of proteolysis. The ground fish treated with 0.5% H2O2 had no detectable residues even imme- diately after treatment, but maintained the in- activity of the enzyme. The residual bromate was dependent on concentration. The samples con- taining 0.025% and 0.05% KBr0 3 still had slight amounts of bromate. Bushuk and Hlynka (1960) reported that 80 ppm of bromate in bread dough disappeared completely after baking for 20 min. We baked portions of ground fish, treated with 0.05% KBr0 3 , for 20 min at 162°C. There were no detectable residues indicating there would not be significant residues in normally cooked fish. Effect of Inhibition on Texture Results of the organoleptic evaluation for tex- ture are shown in Table 6. These results demon- strate that there is a correlation between the per- centage of inhibition and the maintenance of firm texture. Samples which had the highest inhibition were judged to have texture compar- able with nonparasitized fish. Oxidative Effect on Amino Acids Some amino acids are susceptible to oxidation, particularly methionine which is readily oxidized to methionine sulfoxide and, under severe conditions, to methionine sulfone. We were using relatively mild conditions compared with other investigators, but we lacked informa- tion on the effect of potassium bromate or the combination of potassium bromate and hydrogen peroxide. We therefore compared the amino acid profiles of acid hydrolysates of nonparasitized Pacific whiting, parasitized with no treatment, and two samples of parasitized ground fish, one of which was treated with 0.5% Na 2 HP0 4 H 2 02 + 0.025% KBr0 3 , the other with 0.5% K 2 HP0 4 - H2O2 + 0.025% KBr0 3 . We compared the profiles for differences that might suggest significant de- struction of any of the amino acids. Acid hydrol- ysis converts methionine sulfoxide to methionine so a difference would only show if methionine were converted to methionine sulfone. No signif- icant differences were found in any of the amino acids (Table 7). Table 5.— Storage study of oxidants in ground parasitized Pacific whiting. % Oxidant inhibition resi due Oxidant Otime 1 mo time 1 mo Control— no treatment 5% Na 2 HP0 4 H 2 2 + 0.025% KBrOa 73 81 + 2 + 0.5% K 2 HPO«-H 2 2 + 0.025% KBr0 3 67 75 + + 0.5% Na 2 HP0 4 H 2 2 + 0.01% KBr0 3 62 59 + N.D. 3 0.5% K 2 HPO„-H 2 + 0.01% KBr0 3 37 46 + N.D. 0.5% H 2 2 49 52 N.D. N.D. 05% KBr0 3 66 66 + + 0.5% H 2 2 + 0.025% KBr0 3 63 69 + + 0.5% Na 2 HP0 4 H 2 2 + 0.5% H 2 2 34 47 + N.D. 'Storage at -20°C 2 + = presence of residue oxidant. 3 N.D. = not detectable. Table 6.— Texture evaluation of treated parasitized Pacific whiting. Sample and treatment Texture evaluation Nonparasitized Pacific whiting Parasitized Pacific whiting— no treatment Parasitized Pacific whiting treated with 0.5% H 2 2 Parasitized Pacific whiting treated with 05% KBr0 3 Parasitized Pacific whiting treated with 5% Na 2 HP0 4 -H 2 2 + 025% KBr0 3 Parasitized Pacific whiting treated with 0.5% K 2 HP0 4 H 2 2 + 0.025% KBr0 3 Parasitized Pacific whiting treated with 5% K 2 HP0 4 -H 2 2 % inhibition '1.1 2.6 2.6 13 1.1 69 14 63 18 64 2.8 28 'Categories: 1 = firm, 2 = soft. 3 = mushy 285 FISHERY BULLETIN: VOL. 80, NO. 2 Treatment of Fillets Since a large portion of any food fish such as Pacific whiting is sold in the form of fillets, it would be preferable to treat the fillets as well as the minced fish. Recently Spinelli 4 reported on the use of adding aqueous additives into fillets by high pressure injection. The work showed that it is possible to disperse precisely given amounts of aqueous additives into fillets taken from several species of fish. SUMMARY The proteolytic activity in minced parasitized Pacific whiting can be effectively inhibited by the addition of hydrogen peroxide, potassium bromate, dibasic phosphate peroxides, iodo- acetate, and N-ethylmaleimide. In human food systems, the only acceptable compounds of those mentioned to achieve this inhibition are hydro- gen peroxide, potassium bromate, or the dibasic phosphate peroxides. The most effective inhib- itors at low concentrations were 0.05% KBr03 and either 0.5% Na 2 HP0 4 H 2 2 + 0.025% KBr0 3 or 0.5% K 2 HP0 4 H 2 2 + 0.025% KBr0 3 . These inhibitors retained their inhibitory effect during 1 mo of storage at — 20°C. The inhibition was suf- ficient to maintain a firm texture when portions of the treated ground parasitized Pacific whiting were cooked. Catalase in whiting muscle rapidly degraded added hydrogen peroxide, but did not destroy potassium bromate; however, potassium bromate was reduced to undetectable levels when the material was cooked. LITERATURE CITED Association of Official Analytical Chemists. 1975. Official methods of analysis, 12th ed. Assoc. Off. Anal. Chem., Wash., D.C., 1094 p. Bryant. J., T. R. Green, T. Gurusaddaiah, and C. A. Ryan. 1976. Proteinase inhibitor II from potatoes: Isolation and characterization of its protomer components. Bio- chemistry 15:3418-3424. BUSHUK, W., AND I. HLYNKA. 1960. Disappearance of bromate during baking of bread. Cereal Chem. 37:573-576. Cuq, J. L., M. Provansal, F. Guilleux, and C. Cheftel. 1973. Oxidation of methionine residues of casein by hydrogen peroxide. J. Food Sci. 38:11-13. Dassow, J. A., M. Patashnik, and B. J. Koury. 1970. Characteristics of Pacific hake, Merluccius Table 7.— Percent of amino acid in hydrolysate of ground Pacific whiting muscle. Amino acid Non- 0.5% Na 2 HPGv 0.5% K 2 HPO«- para- Nontreated H 2 2 + 0.025% H a 2 f 0.025% sitized parasitized KBr0 3 treated KBr0 3 treated Aspartic acid 94 9.5 9.6 9.6 Threonine 4.5 49 46 46 Serine 48 5.0 5.0 50 Glutamic acid 13.6 13.6 13.8 13.8 Proline 3.3 34 3.4 36 Glycine 7.2 70 7.0 7.0 Alanine 8.8 84 8.7 8.5 Valine 5.6 5.5 5.6 55 Methionine 26 2.7 26 2.7 Isoleucine 4 2 4.2 4.2 4.2 Leucine 7.6 7.6 7.7 76 Tyrosine 2 2 2.2 1.9 2.2 Phenylalanine 2 8 2.8 2.8 2.8 Histidine 1.6 1.6 16 1.6 Lysine 7.7 7.8 78 7.8 Arginine 4.0 3.9 4.0 4.0 4 Spinelli, J. 1980. Injection of aqueous additives into fish by high-pressure injection. Paper presented at Pacific Fish- eries Technologists meeting, Astoria, Oreg., 3/16-19/80. Northwest and Alaska Fisheries Center, Natl. Mar. Fish. Serv., NOAA, 2725 Montlake Blvd. E., Seattle, WA 98112. productus, that affect its suitability for food. In Pacific hake, p. 127-136. U.S. Fish Wildl. Serv., Circ. 332. Deisseroth, A., and A. L. Dounce. 1970. Catalase: Physical and chemical properties, mech- anism of catalysis, and physiological role. Physiol. Rev. 50:319-375. Hass, G. M., R. Venkatakrishnan, and C. A. Ryan. 1976. Homologous inhibitors from potato tubers of serine endopeptidases and metallocarboxypeptidases. Proc. Natl. Acad. Sci., U.S. 73:1941-1944. James, A. L., and J. E. McCrudden. 1976. Whitening of fish with hydrogen peroxide. Pro- ceedings: The production and utilization of mechanical- ly recovered fish flesh (minced fish) 7/8 April 1976, p. 54- 55. Interox Chemicals (U.K.) Ltd., Luton, Bedford- shire, Minist. Agric, Fish., Food, Torry Res. Stn. Lowry, O. H., N. J. Rosenbrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. Melville. J. C, and C. A. Ryan. 1972. Chymotrypsin inhibitor I from potatoes. Large scale preparation and characterization of its subunit components. J. Biol. Chem. 247:3445-3453. Nakatani, H., and K. Katagiri. 1979. Foodstuffs with phosphate peroxide additive. U.S. Patent Office. Patent No. 3,545,892. Pintauro, N. D. 1974. Food additives to extend shelf life. Noyes Data Corp.. Park Ridge, N.J., 400 p. Price, R. J., and J. S. Lee. 1970. Inhibition of Pseudonwnas species by hydrogen peroxide producing Lactobacilli. J. Milk Food Technol. 33:13-18. Ryan, C. A., G. M. Hass, and R. W. Kuhn. 1974. Purification and properties of a carboxypeptidase inhibitor from potatoes. J. Biol. Chem. 249:5495-5499. Sims, G. G., C. E. Coshan, and W. E. Anderson. 1975. Hydrogen peroxide bleaching of marinated her- ring. J. Food Technol. 10:497-505. Spackman, D. H., W. H. Stein, and S. Moore. 1958. Automatic recording apparatus for use in the chro- matography of amino acids. Anal. Chem. 30: 1 190-1206. Tsen, C. C. 1968. Oxidation of sulfhydryl groups of flour by bromate under various conditions and during the breadmaking process. Cereal Chem. 45:531-538. 286 FEEDING HABITS OF STOMIATOID FISHES FROM HAWAIIAN WATERS Thomas A. Clarke 1 ABSTRACT Stomachs were examined from over 2,800 specimens of stomiatoids collected near Hawaii. Small Vinciguerria nimba ria ate mostly small copepods and ostracods, while large fish appeared to switch to large amphipods and small euphausiids. The remaining planktivorous species, sternoptychids and small gonostomatids, fed primarily on large calanoid copepods and small euphausiids. All of these appeared to feed by active, visual searching, and preferred prey were probably more visible than other zooplankton in appropriate size ranges. Diets and preferences of the planktivorous stomiatoids were similar to or identical with those of one or more species of myctophids which share the same habitat. The large gonostomatids ate micronekton but appeared to feed in the same manner as the small individuals and species. The species from six other families, which appear to be morphologically adapted to ingest rela- tively large prey, did in fact feed mostly on prey 20% of their body length or longer. Only two species ate zooplankton as well. Most species with chin barbels were nearly or exclusively piscivorous, and those without barbels ate few or no fish. The barbel and analogous structures appear to be used pri- marily to attract and aid in the capture of relatively large fish. Apparent preferences for certain types of prey by the piscivorous species indicate that interspecific differences in barbel features are related to dietary specialization. Based on feeding incidence and estimates of stomach evacuation time, the piscivorous stomiatoids appear to consume a large fraction of the standing crop of plank- tivorous fishes each year. Stomiatoid fishes are important components of the micronekton in most tropical and temperate oceanic areas (e.g., Maynard et al. 1975). Most species occur in the upper 1,000 m and undertake diel vertical migrations (Clarke 1974 and others cited therein). They include both small, plank- tivorous species and generally larger forms with certain morphological features apparently re- lated to capture of relatively large prey. Little is known of the feeding habits of these fishes and, consequently, of their role and impor- tance in the pelagic food web. Diets of a few planktivorous species have been reported, but usually from few specimens and without identifi- cation of prey beyond major taxa. Clarke (1978) showed that some planktivores feed while at depth during the day. Knowledge of the prey of nekton-eating species has consisted mainly of in- cidental reports scattered throughout the litera- ture rather than systematic investigations of large numbers of specimens. This paper presents results of examination of stomach contents of over 70 species of stomia- toids from an extensive series of collections near ■Department of Oceanography and Hawaii Institute of Ma- rine Biology, P.O. Box 1346, Kaneohe, HI 96744. Manuscript accepted November 1981. FISHERY BULLETIN: VOL. 80, NO. 2. 1982. Hawaii in the north central Pacific Ocean. Almost all the species are vertical migrators; the abundant, nonmigrating species of Cyclothone, Sternoptyx, and Argyropelecus (which are the subjects of separate studies by other investiga- tors) are not included. Diets of the planktivorous species are compared with estimates of prey abundance in appropriate depth ranges in order to determine whether composition and apparent preference are similar to those of cooccurring, nonstomiatoid planktivores which feed in the upper layers at night (Clarke 1980). Data from the nekton-eating stomiatoids allows considera- tion of preference, feeding methods, and the impact of these predators on the planktivorous micronekton in the community. METHODS Specimens for this study were collected ca. 20 km west of the island of Oahu, Hawaii (ca. lat. 21°20-30'N, long. 158°20-30'W) in waters 2,000- 4,000 m deep. Previous studies in this area have considered the vertical distribution and certain other aspects of the ecology of stomiatoids (Clarke 1974) and the feeding chronology of five species (Clarke 1978). Other investigations in the 287 FISHERY BULLETIN: VOL. 80, NO. 2 area have been summarized by Maynard et al. (1975). Over 2,800 specimens of nine families were examined. Based upon preliminary results, mor- phology, and the literature, the species were separated into two groups, each of which was treated differently. Members of the Photichthyi- dae, Sternoptychidae, and Gonostomatidae were considered planktivores; and those of the Astro- nesthidae, Chauliodontidae, Idiacanthidae, Mel- anostomiatidae, Stomiatidae, and Malacostei- dae as nekton-eating species. All specimens of planktivorous species were taken with a 3 m Isaacs-Kidd midwater trawl which terminated in a 1 m diameter cone of ca. 3 mm mesh netting with a ca. 2 1 nonfiltering cod end bucket. Towing procedures were the same as described in Clarke (1980). The trawl was low- ered to a given depth as rapidly as possible, towed for 2-3 h at ca. 2 m/s, and retrieved as rap- idly as possible. A time-depth recorder of the appropriate range was attached to the trawls; depth records were accurate to 2-4% of the depth fished. In addition to night tows at 70-170 m described in Clarke (1980), specimens were also taken from day tows at 400-800 m and night tows at 225-250 m made in September 1973 and November 1974 (Table 1). During some of the deeper tows, the trawl changed depth by as much as 50-100 m during the "horizontal" portion of the tow. Since the most abundant planktivores were known to feed during the day (Clarke 1978), zoo- plankton were sampled at 400-500 m during the day (Table 1) with opening-closing 70 cm diame- ter bongo nets (505 yum mesh). The nets were low- ered closed, opened for 0.5-1 h atca. 1 m/s ship's speed, then closed, and retrieved. Time-depth re- corders attached to the nets indicated vertical movement of up to 100 m during the open por- tions of the tows. Volume sampled by each net was estimated from the mouth area, duration of the open portion of the tow, and an estimated speed of 1 m/s. All material was preserved immediately after capture and held in ca. 4% formaldehyde/sea- water. Except for certain trawl samples where large numbers of Vinciguerria nimbaria were caught and only individuals with obviously full stomachs were selected, all specimens of each species considered were measured (standard length, SL, to the nearest millimeter) and stom- achs examined. Intact prey items were identi- fied, counted, and measured to the nearest 0.1 mm (prosome length for copepods, total length without telson for malacostracans, and total length or maximum dimension for other prey). Identifiable remains among partially digested material were recorded. Any remains of chae- tognaths (the only gelatinous prey found) were counted as intact since they are probably de- graded much more rapidly than other prey types. Items in the mouth or esophagus were not counted; their limbs and bodies were not com- pressed, indicating that they had been taken after capture. Otherwise, there was no evidence of postcapture ingestion by the fishes. Most prey types found intact in the stomach were also re- corded as digested remains that had almost cer- tainly been eaten well before capture, and, conversely, several types of abundant zooplank- ton were rarely or never found in the stomachs, as would be expected if the fish were feeding in- discriminately in the net. There was no evidence that food was regurgitated during or after cap- ture; I found no everted stomachs and no digested remains in the esophagus. Zooplankton from the bongo net samples were counted from either the entire sample (euphausi- ids and other relatively large types) or 1/16-1/32 aliquots taken with a Folsom plankton splitter. For both plankton and intact prey items, eu- phausiids and most copepods (with the exception of unidentifiable copepodites, which were fairly common in all the plankton samples) were identi- fied to species. Ostracods (mostly Conchoecia TABLE 1.— Dates, local (Hawaiian Standard) times, and depths of tows with 3 m Isaacs-Kidd midwater trawl and 70 cm bongo plankton nets off Oahu, Hawaii. Times for trawls are for the period at depth; total times including descent and ascent are in parentheses. Times for bongos are for the open period only. Depth figures are the ranges during "horizontal" portions of tow or modal depth if the range was <20 m. Date Time Depth (m) Trawls: 25 Sept 1973 1523-1723 (1500-1750) 400 9 Nov. 1974 1540-1740 (1535-1802) 400-450 25 Sept. 1973 0748-0954 (0721-1028) 450-500 25 Sept. 1973 1148-1348 (1115-1435) 525 11 Nov. 1974 0818-1118 (0736-1154) 550-600 9 Nov. 1974 0808-1108 (0730-1132) 550-650 26 Sept 1973 0820-1020 (0742-1120) 600 9 Nov. 1974 1230-1430 (1155-1500) 600-650 12 Nov. 1974 0756-1000 (0722-1100) 650 (briefly to 800) 26 Sept. 1973 1227-1427 (1142-1550) 700-800 10-11 Nov. 1974 2303-0100 (2250-0115) 250 11 Nov. 1974 0155-0505 (0145-0515) 225 Bongos: 14 Sept. 1973 0930-1033 400-425 14 Nov. 1974 0808-0908 400-425 14 Sept. 1973 1103-1135 425-525 14 Sept 1973 1241-1311 550 288 CLARKE: FEEDING HABITS OF STOMIATOID FISHES spp.) and amphipods were not further identified, and other prey types were identified only to major taxa. Prey types of the same genus or of similar size, pigmentation, etc. were often lumped for convenience of presentation of re- sults. Densities of zooplankton (Table 2) were calculated from the counts (corrected for any subsampling) and estimated volumes filtered; however, since these are based on so few samples, they can be considered as only rough estimates of prey abundance at the depths where the fishes were caught. Furthermore, the densities of types under 1.0 mm long are underestimated due to mesh escapement; for most of these, densities are probably about 4-5 times higher than estimated from the samples (Clarke 1980). The nekton-eating species were much less abundant than the planktivores, and their feed- ing incidence and number of prey per fish were lower; consequently, in order to gather as much data as possible I examined specimens from a wide variety of trawl samples taken between 1969 and 1978. These included both horizontal and oblique samples in the upper 1,200 m— mostly either above 350 m at night or deeper dur- ing the day. Almost all were taken with a 3 m Isaacs-Kidd trawl towed at ca. 2 m/s. The termi- nal section was of fine (333 yum) plankton mesh for about two-thirds of the samples. For a few rare species I also took material from collections with a 5 m Isaacs-Kidd, a 3 m Tucker, or a 2/3 Cobb pelagic trawl. Data from the more abun- dant fishes were grouped by arbitrary size classes or by time of capture. For the latter, "day" included all tows started and completed between sunrise and sunset plus a few dusk tows which were completed after sunset but fished at or near the day depths of the fishes. Similarly, "night" included tows taken wholly between sun- set and sunrise plus a few dawn tows that fished at or near night depths of the fishes. Specimens were identified, measured, and classified into one of four categories: Damaged — the stomach ruptured or lost during capture, Empty — stomach completely empty or with only a trace of unidentifiable remains, Remains — prey completely disintegrated but identifiable to major taxon, Intact — prey in one piece or a few large pieces. Sizes (standard length of fishes, length without telson of crustaceans, and mantle length of squids) of all intact prey items were re- corded. Depending upon size of the item and de- gree of digestion, the accuracy of these measure- ments was an estimated ±1-5 mm. Relative length of the prey items, as percentage of stan- dard length of the predators, was used for pre- sentation. Intact crustaceans and many of their remains could be identified to genus or species, but only a fraction of the intact fishes could be unquestionably identified. Where a fish prey could not be identified positively, a probable identification could be often given based on a process of elimination. Because of their photo- phores, myctophids and some stomiatoids could be identified as such at more advanced stages of digestion than other fishes; in most cases where an item was clearly not a myctophid, it was in good enough condition to be more precisely iden- tified. There was little evidence of postcapture inges- tion of large items. A few very fresh items, i.e., those without a coating of stomach mucus or with the limbs not flattened against the body, were not counted. Most of these items were still partly in the esophagus and were usually types not found as digested remains in the same predator spe- cies, e.g., a euphausiid in an otherwise piscivo- rous species. As with the planktivores, there was no evidence of postcapture regurgitation of prey by nekton-eating species. Most of the nekton-eating species proved to eat only small nekton (prey >10 mm long). Zooplank- ton (usually copepods) were very rarely found in their stomachs— always in near-perfect condi- tion and never as digested remains. Certain spe- cies of these fishes, however, appeared to eat both small and large prey, and zooplankton were rou- tinely found in their stomachs. In spite of the fact that many specimens of these species were taken by trawls with a fine mesh terminal section, there was little evidence that the data were biased by postcapture ingestion. As with the strictly planktivorous species, the types of prey found intact included only a narrow range of the types collected in the cod end of the trawl rather than a mixture as would be expected from indis- criminate ingestion in the net, and digested re- mains of the same types of prey were also re- corded for these species. Finally, as will be shown below, the incidence of small items in the stomachs decreased with size of the fish; this would not be expected if these species were for some reason prone to ingestion after capture. (At towing speeds of less than ca. 1.5 m/s, post- capture ingestion of both large and small items appears to be a serious problem. During the course of this study, I examined specimens from several tows taken at 1.0-1.5 m/s. Zooplankton— 289 FISHERY BULLETIN: VOL. 80, NO. 2 co o> 1-2 u — "S s .fa X 5 W c cc a! Q LO a V CO. CO 2 2 c CO >- a, cd O o >> c .g 'E - 3 4- -a s «g E < 3 © • -a <=> £ "5 © _Q 'S f si » "& H 'O £2 C o C a> ct) I flee - E ■2 £ cd 3 co to CO 4> «fe CD CO -i-i a> o fl „. - . *> o fet- a ft si o ° ^* 1|| ^ o S *0 — ■> 5 .g : cd O- C c A C CL> CO '-£ "3 a W * S '. o e CM 6 cd to E < o ^ ■^ t^ -^ CM CM en I »- o o o o o d d *-" d d d I I a> r» r- in i- CO CM CM i- ° 1 1 P + O O r- o o o + o o o d d d d d d d I + I I + I OOOU")OOOC\IOt- oooooooooo I I O i- CM O d d d d o d CM CM i- CO ^ ° P + I + + o o o o d + a VI 5 CL o a O CO CO CD CD CD CD CO O O C C O O co Q. CO . o o o 9- CM CM CO ^1 © oi to co -c co ^ r a ac O CD O O -u V Q. co ^ ~0 cr ° E E en E T E *T P T-' C\i Al O ™ 9- 3 °- „ d cr a. m g- a <" cl c co d Q. m co to o > D. to CD CO ^ CO ^ ^ "O "O o — "• O O co "o Q. CL>- : = C C 2 3 ~ O CD CO W i; o "D CO CD CO 3 -C O 3 J5 S W <0 fO TO to vj yj ^ w w ^ 3 cm CO to r- ^£,■0 5 ? o o - y cu o — m i25OD.O0.il o O 1 CM 1 p SXOCB O "- CM CO d d o dddd cvji-mcsjcNJcnoojoocDcD OOOOi-OOOCSJOO cm m CO CM O O CM r- I o d> d d i- i 5 " u"> r^ ^ CD co co d d d -r^ d d O , »-CM(J)CMOCO T-Ni-T-COCMlfi^T-COOJ •*t | t-Oi-COi-CM LOt-CMCOCOOOO^-OO d dddcMod ddddddddddd co r- co o d d oo)0)omio | CD O CO CD LO t- d d d d d o co in in o 1 LO CD | CO CT) lO 1 P 1 OJ 0> CD *- LO CM | o O) o o o o d d ' d ^ o ri-r^NOd CO -t *- ~^ I «>. o o o LOT-T-cocor^coo OOOT^i-IOt-CO dddddord en cm o o I I S O -n* CO CO CO OOt- WOO h- CM CM O o o o o o c Q. t, to ;, 3 . . C to (0 " TO ttgtop « to C £ CO O Joe CD CD CO O < < O S c o <0 CO o o O ~ to T3 | ogg LTiSoo? a. a. to t, E ■2 E Bo ~ CM SV -~ CD co co I' CD — Ct3 QJ o< qo.« CM CO "6 Al | a a E ~ E E to E v P P V r " o o c c en j- c co to c A CO CO Q. Q.§ 9-T3 VI A CO D) CO * z. /i, «. [^ s t; CD CD CO CO ■c x: o o 3 3 LU Uj CO ^1 CO co CO <0 CO S E S SEE CD CD CO E E E o o o L*. h» W 3 3 3 CD CD CD a S a 3; S; S; E E g, EE5 CO CO Q. CL a 'VI A CD Cl co CL to to "> . 3 3 to cl "O "O CD to CD Q. -Q -Q ^ (^ CO to 2 2 X 3 ^ to 2 2 S Q.CL-S CD CD CD CO 3 *C 290 CLARKE: FEEDINC HABITS OF STOMIATOII) FISHES up to 10-12 assorted copepods and ostracods, all apparently recently ingested — were found in several stomachs of fishes that otherwise had eaten only relatively large items. I also found several apparently freshly ingested euphausiids and sergestid shrimps in stomachs of fishes that otherwise appeared to be strictly piscivorous. Specimens from these slow tows were not in- cluded in the data presented here.) Estimates of biomass and relative abundance of vertically migrating fishes in the study area and of feeding incidence of the nekton-eating stomiatoids were made from catches of a series of oblique 3 m Isaacs-Kidd trawl tows taken at approximately monthly intervals between Au- gust 1977 and November 1978. A time-depth re- corder and a flowmeter were mounted on the trawl for all tows. All fishes were identified, spe- cies were grouped by taxa and known or prob- able feeding habits, and wet weights of each group determined for each sample. All nekton- eating stomiatoids from the series were exam- ined and are included in the results below. The 58 night tows in this series fished between the surface and ca. 350 m and covered the night- time depth range of all vertically migrating spe- cies. The relative abundances of the different taxonomic and trophic groups were calculated based upon total numbers from all the night tows. Biomass (wet weight) per unit area of each group was calculated as in Maynard et al. (1975) for each sample. The overall mean of all samples and all seasons was used as the estimated aver- age biomass. The 28 day tows covered the day depth ranges of the vertically migrating species (ca. 350-1,000 m), but for various reasons it was not possible to reliably estimate volume filtered (and therefore biomass per unit area) from these tows. The numbers of nekton-eating stomiatoids and of prey species in the catch and the numbers of prey found in the stomachs of the stomiatoids from both day and night tows were used to esti- mate feeding incidence relative to the numerical standing crop of prey species. RESULTS Photichthyidae Vinciguerria nimbaria (Table 3) from three samples within its day depth range were divided into two size groups (16-25 mm and over 25 mm SL). Catches of three of the six size-depth groups were high, and only fish with visually apparent full stomachs were selected for examination (Table 3, columns 3, 5, 6). All fish of the other groups were examined, but total numbers of in- tact prey were still quite low for these. Overall the most frequent items in the stom- achs were small copepods and ostracods. Oncaea spp. were the dominant prey in most size-depth groups, and in all samples Oncaea— especially the small forms — were more frequent in the diets of the smaller fish than in those of the large. Be- yond this, however, the diet composition varied between groups without much apparent relation to size or depth, e.g., Clausocalanus spp. and Pleuromamma gracilis were important fractions of the prey of the small fish from 400 m and both size groups from 450 to 500 m; candaciids and Scolecithrix danae were taken by most groups, but decidedly more frequently (as percentage of prey items) by the large fish from 400 m; the fre- quency of ostracods varied among the groups from 3% to 42% of the total items. Some of this variability was undoubtedly a consequence of small sample sizes from three groups, but part resulted from large numbers of certain prey types occurring in only one or a few of the fish from a given size-depth group. Exam- ples include (see appropriate column of Table 3): All 7 P. gracilis from 1 of 6 fish with prey (col- umn 1); all 5 amphipods from 1 of 3 fish (column 2); 4 of 5 Undinula spp. from 1, 4 of 5 Sapph irina spp. from another, and all 11 pelecypod larvae from another out of 18 fish (column 3); 13 of 41 Clausocalanus spp. in 1 and 34 of 73 P. gracilis in 3 others out of 20 fish (column 5); 14 of 15 Scole- cithrix danae in 1 and 24 of 34 P. gracilis in 2 others out of 9 fish (column 6). The presence or absence of only one or two fish such as these had an important effect on percentages of certain items in the estimated diet of a size-depth group. Vinciguerria nimbaria over 30 mm SL had eaten considerably larger items more frequently than smaller fish. The only large day-caught specimen (37 mm) contained remains of another fish, a 3.0 mm amphipod and two Nematobrach- ion spp., each ca. 15 mm long. Ten specimens over 30 mm were taken in a night tow at 70 m. Most items in the stomachs of these fish were on the borderline between "intact" and "remains" and difficult to count similarly to those from the day specimens, but it was clear that euphausi- ids— mostly Stylocheiron spp.— were the most frequent items and that small copepods were much less important than in the smaller fish. Re- mains of six to nine Stylocheiron each were found 291 FISHERY BULLETIN: VOL. 80, NO. 2 Table 3.— Numbers of intact items of different prey types from stomachs of Vinciguerria nimbaric i and V. poweriae from several depth s and times. Remains of types not found intact are denoted by ' 'r"; in column seven (V. nimbaria. Night, 70 m), numbers of nearly intact remains (see text) are given in parentheses. V. nimbaria V. poweriae Day Night 70 Day 400-300 23-29 Night Depth (m) 400 400-450 450-500 225-250 Size (SL, mm) 17-25 26-30 16-24 26-30 17-25 26-30 31-39 20-34 No examined 14 8 18 13 20 9 10 11 16 No. w/intact prey 6 3 18 4 20 9 8 7 No. of intact prey 54 33 270 31 437 192 57 17 No of prey type: Neocalanus spp. — — — — 1 1 r(4) — — Nannocalanus minor — — — — — 1 — — — Undinula spp. — — 5 — 1 1 — — — Clausocalanus spp. 3 — 1 — 41 13 — — — Euchaeta spp — 1 — — — 4 r(2) — — Aetideidae — — — — — 2 r(1) — — Scolecithnx danae — 5 3 1 2 15 — — — Scolecithricidae <1 mm 1 — — 1 2 2 — — — Scolecithricidae >1.0 mm — 1 — — 5 1 — 2 — Pleuromamma abdominalis — — — r 3 1 — 1 — Pleuromamma spp CIV, CV — — 1 — 8 2 — — — Pleuromamma gracilis 7 r 5 1 73 34 — — r Lucicutia spp. <1.3 mm 3 r 3 — 20 9 — — 1 Helerorhabdus spp — — 1 — 2 1 — — — Augaptilidae — — — — — 1 — — — Candacia spp 1 1 2 — 4 6 r(1) 3 r Paracandacia spp — 7 2 — 1 3 r(13) 1 2 Unident calanoid — 1 — — 6 2 r(5) — — Corycaeus spp — 1 10 — 11 1 r(1) — — Oncaea mediterranea 9 2 66 6 82 13 r(2) 6 5 Oncaea conilera 10 2 43 4 55 17 — 7 — Oncaea venusta 1 — 19 1 7 4 — 3 3 Oncaea spp <0.6 mm 13 2 29 — 67 24 — 3 1 Sapphirina spp — — 5 1 — 1 rd) — — Aegisthes spp — — — — — 1 — — — Euphausia spp — r r — 6 2 — 2 — Stylocheiron spp. — — — — — r r(28) r r Nematobrachion spp. — — — — — — — 1 r Euphausiid larva — — 1 — 3 4 r(1) — — Caridean larva — 1 — — — 1 — — — Amphipod <2.0 mm — 1 4 — 2 — — 1 — >2.0 mm — 4 3 1 4 4 r(46) — — Ostracod <1.0 mm 2 — 27 4 16 9 — 5 2 >1.0 mm 4 1 20 9 14 8 r(9) 16 3 Gastropod larva — — 8 — 1 1 — 6 — Pelecypod larva — — 11 — — — — — — Heteropod {Atlanta spp ) — 1 1 — — — — — — Chaetognath — 1 — — — 1 r(1) — — Fish larva — 1 r 2 — 2 r(2) r r in four of the large V. nimbaria. Amphipods were also apparently important items in the diet of these large fish, but similar to the above exam- ples, about 38 of the approximately 46 amphi- pods recorded were eaten by only 2 of the 10 fish. Vinciguerria poweriae was taken in small numbers in the same day tows as V. nimbaria and at night at 225-250 m (Table 3). Both the inci- dence of fish with intact prey and the number of prey per fish were lower at night, indicating that, like V. nimbaria (Clarke 1978), V. poweriae feeds during the day. The items and remains found in stomachs of both groups indicate that V. poweriae's diet is generally similar to that of V. nimbaria of the same sizes. The lower per- centages of Oncaea spp., higher percentages of ostracods, and less diversity may have been an artifact of small sample size. Sternoptychidae Valenciennellus tripunctulatus and Danaphos oculatus (Table 4) were taken in day tows with and slightly deeper than the Vinciguerria spp. The small Valenciennellus tripunctulatus — mostly from the shallower tows— had eaten some Oncaea spp., ostracods, and small (1.0-1.5 mm) calanoids, but most of their prey and all of those of the larger fish were medium to large cala- noids. Few prey were found in D. oculatus, but with the exception of a small scolecithricid and remains of an ostracod, all were large calanoids. Gonostomatidae The diet of Gonostoma atlanticum (Table 5) from day tows consisted of essentially the same 292 CLARKE: FKKDINC HABITS OF STOMIATOII) FISIIKS TABLE 4.— Numbers of intact prey from stomachs of ValrtiricnnrllNs tripiuictii- latu8 from day samples at four different depths and of Danaphos oculatus com- bined from four different day samples. Remains of types not found intact are denoted by "r." Danaphos Valenciennellus tripunctulati JS oculatus Depth (m) 400 400-450 450-500 525 400-600 Size (SL, mm) 22-30 23-29 28-31 29-34 28-40 No. examined 11 10 3 4 21 No. w/intact prey 11 10 3 4 10 No. of intact prey 84 44 7 34 21 No. of prey type: Neocalanus spp. — 1 — — — Eucalanus spp 3 — — 1 — Clausocalanus spp. 4 1 — 1 — Aetideidae <2.0 mm 4 — — 2 3 >2.0 mm 10 3 1 — 8 Euchaeta media 12 4 2 6 5 Scolecithricidae <1.0 mm 4 2 — — 1 >1.0 mm 3 1 1 9 — Pleuromamma xiphias 11 11 1 12 3 Pleuromamma xiphias CV 9 1 — — — Pleuromamma abdominalis 3 2 2 1 — Pleuromamma abdominalis CV 1 — — — — Pleuromamma gracilis 6 — — — — Heterorhabdus papilliger 4 1 — — — Heterorhabdus spp. 2 — — — — Candacia longimana — — — 1 1 Oncaea conifera 2 5 — — — Oncaea spp <0 6 mm — 1 — — — Corycaeus spp — 1 — — — Ostracod <1.0 mm — 2 — — — 1.1-1 9 mm 1 3 — — r Unident. calanoid 5 4 — 1 — Chaetognath — 1 — — — Table 5.— Numbers of intact prey from stomachs of four species of gonostomatid fishes taken day and night and combined from two or more samples within given depth ranges. In this table, a few stage V copepodites of Pleuromamma xiphias and P. abdominalis are included with adults. Prey types not found as intact items are denoted by "r." Additional remains from fish of column seven (225-250 m depth) included a penaeidean shrimp and a large Metridia sp. Data for Gonostoma elongatum and G. ebelingi over 120 mm SL are in Table 6. Gonostoma atlanticum Gonostoma elongatum Gonostoma Day ebelingi Diplopho Day + s taenia * Day N ight Day 400-800 34-78 N ight night Depth (m) 400-525 170-250 110-170 29-88 225-250 93-120 400-500 34-77 525-650 94-117 400-650 + 90 Size (SL. mm) 22-45 i 46-65 25-44 46-54 53-93 103-171 No examined 34 29 26 9 24 15 17 9 21 9 14 No. w/intact prey 25 21 7 4 6 9 2 7 4 8 9 No. of intact prey 74 46 10 5 22 18 3 20 9 13 19 No. of prey type: Neocalanus spp. — — — — 1 — — — — 1 — Undinula sp. — — — — — — — — — 1 — Eucalanus spp. 1 — — — — — — — — — — Aetideidae 2 2 — 1 4 — — 2 3 — — Euchaeta media r 1 — 1 — — — 1 — — — Scottocalanus spp r 2 1 2 r — — — 4 — — Amallothrix spp. 2 1 — — — — — — — — — Pleuromamma xiphias 33 10 1 — 11 15 2 5 r 2 2 Pleuromamma abdominalis 8 3 2 — — 1 — — — 4 — Pleuromamma gracilis — 1 — — — — — — — — — Lucicutia spp 1 — 1 — — — — — — — — Candacia longimana 6 3 1 r 2 — r r — — — Unident calanoid 1 — 2 1 — — — — — — — Oncaea spp 3 — — — 1 — — 6 — — — Euphausia spp. 13 14 2 — — 1 r 1 — 5 4 Stylocheiron spp. 2 — — r r r r r — — — Nematoscelis spp — 6 — — — — — — 1 — — Nematobrachion sp. — — — — — — 1 — — — — Thysanopoda aequalis — 3 — r 1 — — 1 — — 2 Thysanopoda spp — — — — — — r — — r 1 Euphausnd larva — — — — — — — — — — 1 Ostracod 1 — — — — 1 — 4 1 r — Amphipod — — — — 2 — r — — — 7 Fish 1 — — — r — r r _ 2 293 FISHERY BULLETIN: VOL. 80. NO. 2 types of copepods eaten by the sternoptychids plus small (8-12 mm) species of euphausiids. The euphausiids were over twice as frequent and, among the copepods, P. xiphias and P. abdomi- nalis much less important in the diet of the larger of the two size groups of fish. Gonostoma atlanticum appears to feed by day (Clarke 1978); as expected, the remains and few intact prey items found in night-caught specimens were similar to those from day-caught fish. Gonostoma elongatum were divided into three size groups. Specimens <90 mm SL from both day and night tows (Table 5) contained mostly large copepods, the majority of which were P. xiphias. Euphausiids or their remains were found in several specimens; only one, a Thysano- poda aequalis, was over 10% of the fish's length. Intermediate-sized G. elongatum (93-120 mm SL) were taken only at night, and most stomachs contained only digested remains. The frequency of euphausiids in the diet appeared higher than in the small fish, and one plus the remains of two others were over 10% of the fish's length. Gonos- toma elongatum over 120 mm (Table 6) had eaten large prey in all but two cases. Relative sizes of most measurable items were about 10%, but val- ues ranged from 3.8% to 27% (excluding two cope- pods and a somewhat suspicious pyrosome). Penaeidean shrimps and euphausiids were the most frequent items and remains, but fish were taken by several and squid by two of the large specimens. Limited data for G. ebelingi and Diplophos taenia indicated that both diet and differences between size groups were similar to those of G. elongatum, but there were some differences in important prey types. Data for G. ebelingi came exclusively from day tows. Small fish (Table 5) had eaten small zooplankton— Oncaea spp. and ostracods — as well as the larger P. xiphias and euphausiids; the intermediate-sized individuals had eaten only large zooplankton. The largest fish (Table 6) had eaten only fish and crustaceans over 10 mm long; the relative sizes of intact items were 11-24%. Diplophos taenia (Table 5) were mostly from day tows. Small fish had eaten med- ium to large copepods and Euphausia spp. The large fish contained few copepods or their re- mains; most prey were small euphausiids or the large (5-6 mm) amphipod Vibilia spp. The two largest fish examined had eaten myctophids. One of the myctophids (Lampanyctus sp.) and a T. tricuspidata were relatively large (29 and 22%, respectively), but all other items were <10%. Astronesthidae Astronesthes indicus under 60 mm SL fed mostly on copepods and ostracods (Table 7). Small prey types, especially Oncaea spp., were more frequent in diets of fish under 30 mm SL. Of the two species of scolecithricid copepods eaten, the smaller Scolecithrix danae(ca.. 1.5 mm prosome length) was more frequent in the diet of the fish under 30 mm than in the 31-60 mm fish, but the larger Scottocalanus spp. (over 3 mm PL) were more frequent in the larger fish. Euphausi- ids were only slightly more frequent in the diet of the 31-60 mm fish than in that of the smaller ones; remains of euphausiids, including five in one fish, were found only in the 31-60 mm group. The few individuals over 60 mm SL (Table 6) were mostly empty; only a myctophid and fish re- mains were found. The smallest individual of A. "cyaneus" (15 mm SL) had eaten small zooplankton, but those 20-47 mm SL (Table 6) had eaten only Euphausia spp. — some up to almost one-half their own length. Fish remains were found in two of the three larger fish examined. The small and inter- mediate-sized A. splendidus had eaten a few copepods and a small euphausiid, but all other prey of all sizes were relatively large— an aver- age of 41% of SL— and all but two were fish (Table 6). Small A. "similis" (Table 6) contained only fish remains; the large individuals contained fish and a single euphausiid whose relative length was considerably less than those of the fishes eaten. (See Clarke 1974, regarding differ- ences between the two provisionally identified species and A. cyaneus and A. similis.) The items found in Heterophotus ophistoma (Table 6) were unique in several respects, but the significance of these cannot be assessed from the insufficient data here. One of the small speci- mens contained squid remains — otherwise found in only two specimens of G. elongatum. The four large specimens contained two sergestids, a Ster- noptyx sp. — the only nonmigrating fish found in any stomiatoid, and remains of a Parapandalus sp. — the only adult caridean shrimp found. All of these items were relatively smaller than prey of most other nekton-eating species. Chauliodontidae Chauliodus sloani (Table 6) had eaten mostly fish; only those <120 mm had taken crusta- ceans—mostly euphausiids— frequently. The 294 CLARKE: FEEDING HABITS OF STOMIATOII) FISHES Table 6.— Summary of stomach analyses for nekton-eating stomiatoids. See text for definition of categories. Under "Time" (first column): D = day, N = night, B = both combined. Under "Remains recorded" (last column): e = euphausiid, s = sergestid, c = un- identifiable crustacean, m = myctophid, f = unidentifiable fish, sq = squid. See text for explanation of groups of unidentified Eustom ias spp. Relative I engths of prey No % of undamaged specimens in % of predator SL Time SL (mm) specimens (damaged) Empty Remains only Intact items (No. of items) Remains Family/species Fish Crustaceans recorded Gonostomatidae: Gonostoma elongatum D 138-207 11(0) 64 9 27 13-20(2) 10-13(3) e.sq' N 126-210 10(0) 10 40 60 13-17(2) 6-27(6) e.c.f.sq 2 Gonosloma ebelingi D 121-143 19(0) 58 32 11 — 11-24(3) e.c.m.f Astronesthidae: Astronesthes indicus B 64-152 20(2) 83 11 6 29(1) — f Astronesthes "cyaneus" B 15-47 30(0) 60 23 17 — 24-48(6) e.c 3 B 114-164 3(0) 33 67 — — m,f Astronesthes splendidus B 22-39 31(1) 53 23 23 32-63(5) 31-41(2) e.m.f B 41-58 17(0) 76 6 18 41-44(2) — e.m.f 5 B 66-95 14(0) 57 7 36 21-64(5) — m Astronesthes "similis" B 23-68 27(1) 73 27 — — — m.f B 98-122 5(0) 40 20 40 25-41(2) 8(1) f Heterophotus ophistoma B 35-70 8(1) 86 14 — — — sq B 141-320 4(0) 25 25 50 7(1) 6-11(2) c Chauliodontidae: Chauliodus sloani D 20-60 43(4) 62 18 20 33-45(7) 20(1) e.m.f N 20-60 57(9) 58 23 19 31-63(6) 10-20(2) e.m.f D 61-120 12(2) 50 20 30 21(1) 11-16(3) e.f N 61-120 33(9) 54 25 21 22-42(6) — m.f D 121-255 24(6) 44 22 33 14-33(5) 13(1) f N 121-232 23(11) 50 25 25 14-19(3) — c.f Idiacanthidae: Idiacanthus fasciola D 50-100 55(25) 90 10 — — — m.f N 50-100 73(24) 88 4 8 16-22(4) — f D 101-200 38(6) 72 19 9 17-20(3) — m.f N 101-200 57(8) 78 14 8 9-20(4) — f D 201-375 37(1) 75 14 11 13-23(4) — f N 201-372 102(7) 84 6 9 10-23(8) 4-8(2) f Melanostomiatidae: Thysanactis dentex D 121-167 29(0) 86 — 14 30-48(4) — — N 121-174 51(5) 67 13 20 21-42(5) 14-29(3) f 6 Eustomias bifilis D 50-90 40(5) 91 — 9 19-20(3) — — N 50-90 95(2) 85 4 11 17-47(10) — m,f D 91-165 36(5) 74 13 13 8-21(4) — m,f N 91-170 60(2) 91 5 3 15-33(3) — f Eustomias enbarbatus B 56-219 26(3) 78 13 9 16-41(2) — f Eustomias spp. (3,low) B 50-160 46(4) 79 7 14 17-32(6) — f Eustomias longibarba B 66-152 50(3) 79 9 11 24-42(5) 25(1) f Eustomias gibbsi B 61-141 35(3) 91 3 6 34-37(2) — f Eustomias spp. (3. hi) B 55-161 134(7) 83 9 8 17-34(10) — f Eustomias "silvescens" B 60-180 32(1) 68 6 26 23-48(8) — f Eustomias spp. (2) B 60-161 152(0) 80 8 12 17-76(18) 14(1) f Bathophilus kingi B 24-140 3(2) 76 17 7 23-40(3) — f Bathophiius spp. B 26-90 27(3) 67 21 12 45-67(3) — f Photonectes spp B 22-78 14(2) 42 25 33 34-72(4) — f B 132-240 10(0) 90 — 10 26(1) 16(1) — Leptostomias spp B 35-290 31(2) 83 7 10 13-29(3) — f Melanostomias spp B 62-165 8(0) 75 12 12 33(1) — f Stomiatidae: Stomias danae B 42-183 12(0) 75 8 17 24-33(2) — f Malacosteidae: Aristostomias spp. B 33-140 25(8) 71 24 6 35(1) — m.f Photostomias spp B 29-51 37(6) 68 13 19 — 9-30(4) s.c Photostomias sp. 1 B 52-102 73(4) 74 16 10 — 15-28(8) s,c Photostomias sp. 2 B 51-90 54(2) 67 21 12 — 29-42(6) s.c Photostomias sp. 2 B 91-140 38(1) 89 8 3 — 30(1) s.c Small intact items also recorded: 'Euchirella sp. 2 Pleuromamma xiphias, candean larva, pyrosome. 3 Oncaea spp , ostracod. 4 P. xiphias, P. abdommalis, euphausiid larva. 5 P. xiphias, Euchirella sp. 6 lsopod. fishes eaten by the smallest size group were rela- tively larger (30-63%) than the fishes from the larger C. sloani (14-29% with one exception) or any of the crustaceans (10-20%). Of the 28 fish eaten, 18 were myctophids of at least 5 different genera (Ceratoscopelus, Hygophum, Notolych- wus, Lampcuiyetus, and T ri phot u run); five others were definitely not myctophids and included one and probably a second Vinciguerria nimbaria and what was most likely a Bregmaceros sp. 295 FISHERY BULLETIN: VOL. 80, NO. 2 Table 7. — Summary of stomach analyses of planktivorous sizes of Astronesthes indicus and Thysanactis dentex. For large prey types, the range of relative lengths in percentage of predator length is given in parentheses after the count. Data for larger fishes of both species are in Table 6. Astronesthes indicus Thysanactis dentex Size (SL, mm) 15-30 31 ■60 43-90 91- 120 Time Day Night Day Night Day Night Day Night No. examined (No. damaged) 23(0) 37(1) 43(1) 55(2) 78(1) 104(3) 37(0) 79(5) Percent undamaged: Empty 70 67 86 77 57 44 73 68 Remains only — 3 2 4 10 19 8 8 Intact items 30 31 12 19 32 38 19 24 Intact large items 4 8 2 9 16 21 16 22 No. of prey type: Eucalanus sp. — 1 — — — — — — Aetideidae 3 — — — 8 8 — 2 Scolecithrix danae 3 6 1 2 — — — — Scottocalanus spp. — 1 1 4 1 — — — Pleuromamma xiphias 1 — — — 24 21 2 13 Pleuromamma abdominalis — — — — 5 3 — 3 Other calanoid 6 3 — 5 5 3 — — Oncaea spp 16 9 1 4 — — — — Aegisthes sp. — 1 — — — — — — Euphausia spp 1(21) 2(19-35) 1(14) 4(12-30) 2(7-13) 3(10-13) 1(10) 3(10-11) Nematoscelis sp. — — — — — 1(11) — — Thysanopoda aequalis — 1(21) — — 8(12-19) 18(11-19) 1(10) 4(10-13) Thysanopoda spp — — — 1(30) 1(24) 3(23) 1(22) 3(21-28) Euphausiid larva 3 6 1 1 2 1 — — Decapod larva — 1 — — — 1 — — Sergestes spp — — — — — 2(18-20) — 1(11) Ostracod 21 15 14 8 — 1 — — Fish — — — — 2(23-24) 3(15-36) 3(16-43) 7(18-53) Idiacanthidae All sizes of Idiacanthus fasciola had eaten fish nearly exclusively (Table 6). Of the 23 fish, 15 were myctophids of at least 5 genera (Bolinich- thys, Ceratoscopelus, Diaphus, Lamp any ctus, and Triphoturus). Only one of the others, possibly a stomiatoid, was definitely not a myctophid. The largest prey of all sizes of /. fasciola were about 20% of the predator's length, but the minimum and average relative size of prey were somewhat higher in the small /. fasciola. The only two crus- taceans found were intact, but neither appeared to have been very recently ingested. No crusta- cean remains were found, and the two intact crustaceans were smaller than all (substantially smaller than most) of the fishes eaten. Two other, smaller items — a pyroosome and a copepod — found in /. fasciola were not counted because they showed no sign of digestion or compression. Thus /. fasciola must have occasionally fed in the net and may have ingested the crustaceans there. Whatever the case, crustaceans are certainly a very minor part of the diet. Melanostomiatidae Thysanactis dentex under 120 mm SL had eaten zooplankton as well as large prey (Table 7). The 43-90 mm size group had eaten small eu- phausiids — mostly Thysanopoda aequalis 11- 19% of their length— and large, pigmented cope- pods — mostly Pleuromamma xiphias; however, several relatively larger (15-36% of SL) fishes, Thysanopoda spp., and sergestids were also found. Fish 91-120 mm had eaten copepods and small euphausiids much less frequently; the bulk of the diet was relatively large fish and crusta- ceans. With the exception of a single isopod, the items and remains from fish >120 mm (Table 6) included only relatively large prey: other fishes, a large Thysanopoda spp., and two sergestids. Of the 24 intact fishes from all sizes of Thysanactis dentex, 9 were definitely myctophids of at least 5 different genera (Bolinichthys, Diaphus, Diogen- ichthys, Lampadena, and Triphoturus), and 11 were definitely of other families, including 6 Bregmaceros spp. and 2 Melamphaes spp. Most of the high values of relative size were for the slen- der Bregmaceros spp. The three B. japonicus from the 91-120 mm SL Thysanactis dentex were 45-53% compared with 11-28% for the remaining fishes and large crustaceans. Among the prey from T dentex, over 120 mm SL, the three Breg- maceros sp. (c.f. B. macclellandi) ranged from 39 to 42%, while with the exception of an unidenti- fied fish at 48%, the remaining fish prey were 14- 32%. There were approximately 30 species of Eu- stomias in the collections, many of them either 296 CLARKE: FEEDING HABITS OF STOMIATOID FISHES undescribed or of uncertain status. Eustomias bifilis was the only one for which large numbers were available, and only 4 others were repre- sented by more than 25 specimens (Table 6). The remaining identifiable species were pooled according to pectoral ray and photophore counts along with specimens whose barbels had been damaged and could not be identified to species. Those designated "3, low" were all damaged spec- imens with 3 pectoral rays and 15 or fewer VAL and VAV photophores. Eustomias bifilis and E. enbarbatus were the only other species from the area with the same counts. Those designated "3, hi" included 69 specimens of at least 6 unde- scribed species and 65 damaged specimens, all with 3 pectoral rays and over 15 VAL and VAV photophores — the same counts as for E. longi- barba and E. gibbsi. Those designated "2" in- cluded 6 damaged specimens and 146 others of about 20 species, which, like E. "silvescens," had only two pectoral rays. The 2-rayed species have shorter and generally more ornate barbels than any of the 3-rayed species (cf. illustrations in Morrow and Gibbs 1964). All prey items and remains from the 3-rayed species with low counts were fish. Of 20 intact items from E. bifilis, 11 were the myctophid Bolinichthys loyigipes and 6 were myctophids of at least 3 other genera {Benthosema, Diogenich- thys, and Hygophum). One of the three unidenti- fied items was definitely not a myctophid and was probably a Howella sp. The range of relative size of prey (15-34% of SL) was large, but there was no trend with the size of the predator. One and probably both of the intact fish found in E. enbarbatus were Howella sp. The six intact items from the damaged specimens (most of which were probably the abundant E. bifilis) included three Bolinichthys longipes, a Benthosema, an un- identified myctophid, and an unidentified fish. The prey of E. longibarba, E. gibbsi, and the other species with three pectoral rays and high photophore counts were, with one exception, fish. Of the 17 intact fish, 15 were myctophids includ- ing 7 and probably 8 Bolinichthys longipes and at least 2 other genera (Benthosema and Cerato- scopelus). The median relative size of fish prey for these Eustomias spp. (25%) was significantly higher (P<0.05, Mann-Whitney test, one-tailed probability) than that for the Eustomias spp. with three rays and low photophore counts (20.5%). One specimen of E. longibarba had eaten a large euphausiid, Thysanopoda pectinata. One of the Eustomias spp. with two pectoral rays had eaten asergestid shrimp, but all other prey of this group were fish. These Eustomias spp. appeared to eat fewer and different mycto- phids than did any of the 3-rayed species. Eu- stomas "silvescens" (cf. fig. 106A in Morrow and Gibbs 1964), the most commonly taken species of this group, had eaten three Scopelosaurus spp., three myctophids (two Bolinichthys longipes and a Diaphus), and two unidentified fish. Stomachs of the remaining species contained a total of 18 intact fish: 12 myctophids, 2 Howella sp., and 4 Scopelosaurus spp. (plus 2 more of the latter that were too digested to measure). Five and probably six of the myctophids were Diaphus spp., and only three and probably four were B. longipes. In the 3-rayed species of Eustomias, Diaphus was found only once, and B. longipes was the most common prey. Although data are too few to be certain, some of the 2-rayed species appeared to have diets that were restricted or included high proportions of relatively rare fishes. For one un- described form, all four items were Diaphus spp.; for another, two out of four were Howella sp.; and for a third and fourth, two out of two items and two out of four remains, respectively, were Scopelosaurus spp. The median relative size of prey of the 2-rayed species (27%) was sig- nificantly (P = 0.01) higher than that for the 3- rayed species with low counts, but did not differ from that for the 3-rayed species with high counts. The two crustaceans recorded from Eustomias spp. appear suspicious and indicative of postcap- ture ingestion, especially since no digested crus- tacean remains were found in any of the stom- achs. The two items showed no obvious signs of having been eaten after capture, but neither were they much digested. The only indirect evi- dence that these were actual prey items and not eaten in the net is that I have found both crusta- ceans and their remains in the stomachs of several E. bulbornatus, a species which does not occur in the study area. Since at least one species of the genus appears to eat crustaceans, it is pos- sible that others may do so occasionally. Based upon a limited amount of data (Table 6), the remaining melanostomiatid genera, as well as Stomias danae (Stomiatidae) and the Aristo- stomias spp. (Malacosteidae), are piscivorous. All the identifiable fish eaten by these species were myctophids. All three items from Lepto- stomias spp. were Notolychnus valdiviae. The relative size of prey of the small Photonectes spp. and several of the Bathophilus spp. was high— 297 FISHERY BULLETIN: VOL. 80. NO. 2 over 50% in several cases. The only crustacean found was a partially digested penaeidean shrimp, together with a myctophid, in a large Photonectes sp. Malacosteidae Only 24 items were found in the 100 Malacos- teusniger examined (Table 8). The most frequent items were copepods; these included some small harpacticoids, but most were large aetideids or scolecithridids. Similar-sized Pleuromamma xiphias were conspicuously absent. Two fish (86 and 93 mm SL) had eaten somewhat larger prey, and remains of relatively large prey were found only in the three largest fish examined. The inci- dence of intact prey was much lower in the larger of the two size groups. As indicated in Clarke (1974), two species of Photostomias occur near Hawaii; neither is iden- tical with P. guernei, the only presently recog- nized species. The form designated species 1 here matures at about 60 mm SL and grows to ca. 100 mm SL, while species 2 matures at about 120 mm SL and grows to >150 mm SL. Individuals less than ca. 50 mm SL cannot be reliably separated. The data given in Table 6 are limited to speci- mens that were analyzed after I had learned to separate the species as well as possible; the text below, however, also includes prey identifica- tions and relative sizes from 54 other specimens from earlier collections. These 54 specimens were no longer conveniently available to me after Table 8.— Summary of stomach analyses of Malacosteus niger with list of all items and remains found. % undamaged specimens Size (SL, mm) No. examined (damaged) Empty Remains only Intact items 24-90 44(3) 71 2 27 91-188 56(0) 88 4 9 SL — items or remains: 30 Undeuchaeta plumosa 37 Candacia longimana, Chirundina streets/, aetideid CV, cope- pod remains 61 Undeuchaeta major 70 Oncaea sp 70 remains of 3-4 copepods 71 2 C. streets/, 2 U. major, Euchirella curticauda 80 2 C. streetsi, U. plumosa, Lophothrix sp. 81 aetideid CV 84 Oncaea sp. 85 Oncaea sp 86 Lophothrix sp., Euphausia hemigibba, myctophid (10 mm SL) 87 Sapphirina sp. 93 Nematoscelis tenella 96 Amallothrix sp. 97 Euchirella sp.. remains Scaphocalanus sp. 101 Corycaeus sp 110 Thysanopoda sp. remains 111 Gaetanus kruppi, fish remains 188 fish remains I had learned to separate the species, and could not be identified with certainty from notes taken at the time of examination. Both species ate crustaceans exclusively, and with few exceptions the prey and identifiable re- mains were sergestid shrimps, mostly small Ser- gestes spp. Two large individuals of species 2 had eaten Gennadas spp., and an unidentified small specimen had eaten a Nematobrachion, the only euphausiid found. Aside from the Gennadas occurring only in species 2, there was no evidence of difference in diet between the two species. Ex- cept for a juvenile shrimp eaten by a small fish, relative length of prey was 15-42% of SL with a median of 28.5%. DISCUSSION Vinciguerria nimbaria, V. poweriae, Valenci- ennellus tripunctulatus, Danaphos oculatus, and Gonostoma atlanticum and small G. elongatum, G. ebelingi, and Diplophus taenia were planktiv- orous, i.e., almost all prey were <5-10 mm long. Clarke (1978) showed that four of these species feed primarily by day, and the limited data here indicated that Vinciguerria poweriae does also. The majority of the diets of V. n imbaria and V. poweriae <30 mm SL consisted of small cope- pods and ostracods. Vinciguerria nimbaria >30 mm SL appeared to feed mostly on substantially larger prey — amphipods and small euphausiids, but large calanoid copepods were not important at any size. In the western Pacific, V. nimbaria, apparently smaller than the smallest size group covered here, were also reported to feed mostly on small copepods and ostracods (Ozawa et al. 1977). Certain prey types found in stomachs of V. nimbaria, e.g., Scolecithrix danae, Paracandacia spp., Oncaea venusta, Stylocheiron spp., were either absent or very rare in the daytime plank- ton samples, but most were present at moder- ately high densities within the nighttime depth range of V. nimbaria (Clarke 1980). Based on diel changes in state of digestion of prey, Ozawa et al. (1977) concluded that V. nimbaria fed at sunset and at sunrise; their evidence for feeding at sunrise is indirect and equivocal. Clarke's (1978) data do not preclude feeding during the upward migration at sunset, but give no indica- tion of feeding at night or sunrise. Thus, while some of the prey types not present in the plankton by day may have been taken at sunset, it seems unlikely that any would remain intact until late 298 CLARKE: FEEDINC. HABITS OF STOMIATOID FISHES afternoon (when two of the day trawls were made) the next day. Vinciguerria nimbaria could conceivably undertake short, irregular excursions to shallower water during the day, or alternatively, may have a strong preference for rare, but perhaps vulnerable "stragglers" from populations with shallower centers of abun- dance. For several prey types, most of the items re- corded were found together in one or a few of the fish examined. This indicates that V. nimbaria often feeds on patches or aggregations of certain prey types. My earlier observation (Clarke 1978) that V. nimbaria stomachs tend to be either quite full or nearly empty throughout the day is also indicative of encounters with patches of prey. Since patchiness would increase the variability of encounter rates by both individual fish and the plankton nets, this might explain why some prey types were poorly represented by the few plank- ton samples as well as the large apparent differ- ences in diet between small samples of fish. Wherever and however V. nimbaria feeds, it clearly showed preference for certain prey types. Some types which were abundant in the zoo- plankton samples, e.g., Oncaea spp., Clausocal- anus spp., small ostracods, were eaten frequently by fish <30 mm SL; but many other types, e.g., Eucalanus spp., scolecithricids (except Scoleci- thrix danae), Metridia spp., large Pleuromamma spp., and chaetognaths, also abundant were either absent or poorly represented in the diet. The types poorly represented in the diet were mostly either larger, less pigmented, or more translucent than those frequently eaten, regard- less of whether the latter were rare or abundant in the plankton. The diet and apparent prefer- ences of small V. nimbaria are most similar to but not identical with myctophids such as Ben- thosema suborbital and Bolinichthys longipes which feed on small zooplankton (Clarke 1980). Vinciguerria nimbaria >30 mm SL showed apparent preference for Stylocheiron spp. and amphipods, both of which were rather uncom- mon within the day depth range. In contrast to both the remaining planktivorous stomiatoids and several myctophids which also feed on large zooplankton (see below), V. nimbaria ignored the large calanoids which were fairly abundant at the deeper end of its depth range (Table 2). The diets of the remaining planktivorous sto- miatoids were nearly restricted to large cala- noids and small euphausiids. The cope pods eaten were fairly abundant within the day depth ranges of the fishes (Table 2), but were appar- ently preferred over similar-sized Eucalanus spp., augaptilids, and chaetognaths which were also fairly abundant. The latter types are very translucent compared with the types eaten and probably less detectable visually. The Gonosto- ma spp. and D. taenia have relatively smaller eyes than V. nimbaria (data given in Grey 1964). Thus, the apparent preferences of these gono- stomatids may result from their being poorly equipped to detect small, translucent, or other- wise less visible prey. (The sternoptychid species both have relatively large eyes, but they are tubular and directed upward, and