UNITED STATES DEPARTMENT OF THE INTERIOR Walter J. Hickel, Secretary Leslie L. Glasgow, Assistant Secretary for Fish and Wildlife, Parks, and Marine Resources Charles H. Meacham, Commissioner, U.S. FISH AND WILDLIFE SERVICE Philip M. Rocdpl, Director, BUREAU OF COMMERCIAL FISHERIES FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE VOLUME 66 ^2t «IQ ISSUED 1967 AND 1968 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON, D.C. 1970 CONTENTS OF VOLUME 66 «^ Number 1 (Issued January 1967) Page DiSTKiBUTioN OF LARVAL TUNAS i:n MARQUESAS WATERS. By Eugene Ij. Xakauiura and Walter M. Matsumoto 1-12 Association op fishes with flotsam in the offshore waters of Central America. By John R. Hunter and Charles T. Mitchell 13-29 Analog coMpnTEK models of fish populations. By Ralph P. Silliman 31-^6 A serologically detected serum factor associated with maturity in English sole, Parophrys vctuliis, and Pacific halibut, Hippoglossus stcnolcpis. By Fred M. Utter and George J. Ridgway 47-58 Effects of water velocity on passage of salmonids in a transportation channel. By Josejih R. Gauley 50-63 Comp.vrati\-e anatomy and systematics of the tunas, genus Tliunnus. By Robert H. Gibbs. Jr. and Bruce B. Collette 6.1-130 Influence of Rocky Reach Dam and the temperature of the Okanogan River on the upstream migration OF sockeye salmon. By Richard L. Major and James L. Mighell 131-147 Seasonal occurrence and size distribution of postlarval brown and white shrimp near Galveston, Texas, with notes on species identification. By Kenneth N. Baxter and William C. Renfro 149-158 Codiiim ENTERS Maine waters. By Gareth W. Coffin and Alden P. Stickuey 153-161 Laboratory evaluation of red-tide control agents. By Kenneth T. Marvin and Ralphael R. Proctor, Jr 163-164 Number 2 (Issued August 1968) Objective studies of scales of Columbia River chinook salmon. Oucorhi/nchiis txliairiilsclKi ( Walbaum). By Ted S. Y. Koo and Andhi Isarankura ^ 165-lSO Catch and estimates of fishing effort and apparent abundance in the fishery for skipjack tuna (Katsiiwo7iiis pelamis) in Hawaiian waters, 1952-62. By Richard X. Uchida 181-194 Characteristics of the blood of adult pink salmon at three stages of maturity. By Kenneth E. Hutton 195-202 Cross-beactive properties of antisera prepared in rabbits by stimulations with teleost vitellins. By Fred M. Utter and George J. Ridgway 203-207 Occurrence of macrozooplankton in Tampa Bay, Florida, and the adjacent Gulf of Mexico. By John A. Kelly, Jr. and Alexander Dragovich 209-221 The geostrophic circulation and distribution of water properties off the coasts of Vancouver Island and Washington, spring and fall 1963. By W. James Ingraham, Jr 223-2.j0 Systematics and biology of the bonefish, Albiila nemoptcra (Fowleb). By Luis R. Rivas and Stanley M. Warlen 251-258 Response of marine organisms during the Solar Eclipse of July 1963. By Bernard E. Skud 259-271 Shell deformity of mollusks attributable to the hydroid, IIiKtractinia echinata. By Arthur S. Merrill 273-279 Offshore distribution of Hydractinia echinata. By Arthur S. Merrill 281-283 Cyclopoid copepods of the genus Tiicca (Tuccidae), parastic on I>iodontid and Tetradontid fishes. By Ju-Shey Ho 285-298 Morphology and distribution of larval wahoo. Acanthocybium solandri (Cuvieb) in the central Pacific Ocean. By Walter M. Matsumoto 299-322 Seasonal distribution and relati\-e abundance op planktonic-stage shrimp (Penaeus spp.) in the North- western Gulf of Mexico. 1961. By Robert F. Temple and Clarence C. Fischer 323-334 Principal diseases of commercially important marine bivalve mollusca and Crustacea. By Carl J. Siuder- mann and Aaron Rosenfiekl 335-385 Some features of the Gulf Stream off Chesapeake Bay in the spring op 1963. By P. A. Mazeika 387-423 Interaction of food level and exploitation in experimental fish populations. By Ralph P. Silliman 425-^39 III S70-35S— 70 1\' CONTEXTS OF \OLUME 6 Number 3 (Issued September 1968) PMgO MODBI.S OF OCEA.NR- MK.liATIOXS OF PACIFIC SALMON AND COMMENTS ON GUIDANCE MKCIIANISMS. B.V William F. Koyee, Lynwood S. Smith, iiiid Allan C. Hartt 441-IC.:; llYDKOi.or.icAL AND Bifn.odicAi, ciiARACTKKisTics OF FLORIDA'S WEST CoAST TRIBUTABIES. By Alexander Dragovicli. .fi)hn A. Kelly. .Ir., and II. Grant Goo■ 1 i ^ 1 1 1, J 1 ^ JAN. FER MAR APR. MAY JUNE OCT. NOV. DEC. FiGUUE 3. — Seasonal variation in the abundance of larval skipjack. The slender bars represent averages for the diel variability station, the thick bars averages for the offshore surveys. The averages for tlie offshore surveys represent more samples an(J a greater areal and temporal cover- age than those for the diel variability station. Based on the offshore data, skipjack spawning ap- pears to be greater during the southern summer (January to February) and early fall (March to April), but paucity of data during the southern win- ter and spring (June to October) makes this con- clusion tenuous. VERTICAL AND HORIZONTAL DISTRIBUTION Strasburg (19G0) oliserved tliat most larval tunas were captured within the up])cr (iO m. of water, witli 20 to '25 percent of the catch witliin the 70- to 130-m. depth, and practically none below 140 m. Since the possibility of larval tunas in waters below 140 m. had been indicated in the earlier cruises (fig. 2), special plankton hauls were planned for the diel variability station during HMS-45. A closing net of dimen- sions similar to those of the open net was added to the towing cable to permit sampling at depths be- tween 140 and 280 m. Although simultaneous sam- pling was planned for deptli ranges of to 140 and 140 to 280 m., the actual ma.ximum depth sampled by the upper net ranged between 121 and 150 m., while the depths sampled by the lower net ranged between 70 and 262 m. (appendix table A-4). No larval tunas were caught by the lower net. Larvae were caught only in 6 of the 12 hauls by the upper net. Although the results did not conflict with those of Strasburg, the meager catches and the departures from the planned sampling ckptlis caused the results to be inconclusive. A relation lietween larval skipjack distribution and area was not evident. Xo significant aical associa- tion (Kendall coefficient of concordance) was found in the average abundance of larvae for the several legs of the offshore survey track, nor was a relation between larval distribution and proximity to land found in a comparison (Kendall cooflicient of con- cordance) of the average abundance by inner, mid- dle, and outer 75-mile sections of the offshore survey legs (tables 4 and 5). T.VBi.K 4. — Ai'cr(i(jr iiiiinlxr of liirnil akipjuck iindir 10 m.' of ocean surf ace for the tiorlh, soidh, east, and trest tegs of the off- shore suri'cys (fig. 1). [Numbers of samples on which the averages arc based are in parentheses] Cruise North leg South leg East leg West leg CHO-s.";... HMS-13... CHa-38... nMS-45... 2.5 (12) 5. 9 (6) 11.4 (fi) 1. 2 (5) .3.3 (10) 5. 6 (6) 4. 2 (6) 2. 5 (6) 0.9 (12) 4.4 (4) 4. (6) 8. 8 (5) 3.9 (12) 3. 6 (6) 3. 9 (6) 0.9 (6) u S. FISH AN D WILDLIFE SEKVICE Table 5.— Average number of larval skipjack under 10 m.- of ocean surface for the inner, middle, and outer 75-mile sections of the of shore surveys. (Numbers of samples on which the averages are based are in parentheses] Cruise Inner 75 miles Middle 75 miles Outer 75 miles riIO-35 KMS^S CHO-38 HMS-45 2.3 (14) 1.9 (8) 6. 2 (8) 3.2 (5) 2. 7 (16) 6. 1 (6) 5. 7 (8) 4.8 (8) 2. 9 (16) 7. 1 (8) 6. 7 (8) 1. 9 (8) Larval skipjack had been found to be widely dis- tributed in the northeastern part of French Oceania previous to the series of cruises in this report. Fig- ure 4 illustrates the locations around the Marquesas » LARVAL SKIPJACK • OTHER LARVAL TITNAS • r ■*ll\IAR0UES.4S .ISLANDS I ^.TAHITI t * MS" W0« -10' 135° 130° W. Figure 4.— Stations in northeastern French Oceania where larval tunas were collected on cruises earlier (1952-57) than those covered by this study. (Data from Matsumoto, 1958, and Strasburg, 1960.) Islands where larval tunas were collected during cruises of vessels of the Bureau of Commercial Fisheries prior to those listed in table 1 . ABUNDANCE OF LARVAL TUNAS AND INVERTEBRATE PLANKTON Relations between the abundance and distribution of invertebrate plankton and of larval tunas, if any exist, are obscure. If the plankton volumes and abundance of larval tunas are averaged for each of the offshore surveys, an obvious positive correlation CHG-35 H,MS-43 CHG-38 HMS-45 OCT.- NOV JAN.- FEB. MAR-APR. MAY 1957 1958 1958 1958 Figure 5. — Average abundance of zooplankton, larval tunas, and tuna schools for the offshore surveys. can be seen (fig. 5), but in individual samples no significant correlations (Spearman rank correlation) were found. High and low measures of abundance of larval tunas were found in high as well as in low volumes of plankton. Strasburg (I960) reported that high catches of larval tunas came from samples of low and moderate volumes of plankton, while the samples with lowest and highest plankton volumes contained smaller numbers of larval tunas. ABUNDANCE OF LARVAL AND ADULT TUNAS Strasburg (1960) found a tendency for larval tunas to occur in larger abundance where there were more adults of the same species although the correla- tion was statistically nonsignificant. Similar com- parisons by species were not possible with our data. Schools of adult tunas were located by sighting the associated bird flocks. Because of the necessity of covering a certain distance of the offshore surveys LARVAL TUNAS IN MARQUESAX WATERS Table 6. — Average jiumhcr of larval tunas under 10 m.- of ocean surface and number of schools sighted per 10-mile run for the legs of the offshore surveys Cr uise North leg East leg South leg West leg Entire survey Schools Larvae Schools Larvae Schools Larvae Schools Larvae Schools Larvae CHO-35 0.405 1.020 .909 .591 3.2 7.2 14.6 3.9 0.18.1 .884 .390 .275 1.8 5.4 4.4 9.4 0.207 . 569 .591 1.01.1 3.4 6.8 4.3 2.9 0.226 .560 .708 .380 4.3 3.6 3.9 2.9 0.256 .758 .649 .559 3.2 IIMS-43 5.8 CHO-38 6.8 HMS^S 4.7 within an allotted timo, investigation of schools was discontinued if the response to chumming was un- favorable. Consequently, the specific identity of many of the schools was not determined, and our examination of the relation between larval and adult abundance was in terms of the aggregate of all tuna species. In comparing the number of schools sighted per 10-mile run and the average number of larval tunas under 10 m.^ of ocean surface for the legs of the off- shore surveys and for the inner, middle, and outer 7>5-mile sections of these legs, no significant correla- tions (Spearman rank correlation) were found (tables C) and 7). For the entire offshore survey area, the averages for both adult tuna schools and larval tunas were highest during either Hi\IS-43 or CHG-38, slightly lower during HMS-45, and lowest during CHG-35. A similar pattern was found in the aver- age of zooplankton volumes. The variations of all three averages are illustrated in figure 5. Table 7. — Average number of larval tunas under 10 m."- of ocean surface and number of schools sighted per 10-mile run for 75-mile sections of the legs of the offshore surveys Cruise Inner 75 m iles Middle 75 miles Outer 75 miles Schools Larvae Schools Larvae Schools Larvae cno-35 0.452 .867 1.145 .412 2.8 3.0 6.8 3.8 0. 305 .554 .460 .586 3.0 7.4 7.1 6.0 0.0.30 .833 .383 .695 3.7 UMS-43 . . 7.3 CHO-38 UMS-45 6.5 3.9 INFERENCES CONCERNING SKIPJACK SPAWNING Inferences about spawning based on the size of the larva upon hatching have been discussed by Matsu- moto (1958). He hypothesized that skipjack arc 2.5 mm. or less at hatching, that the eggs and larvae are planktonic and therefore subject to dispersion by (■urrcnts, but that their displacement from the spawning site would be relatively insignificant unless the currents were exceptionally strong. Larval skipjack have been taken throughout the area around the Marquesas Islands (fig. 4). Most of the catch consisted of larvae between 3 and 4.5 mm. long, so we may assume that they had hatched re- cently. Since the currents around the Mar(|uesas Islands are suspected to lie weak (Sverdriip, John- son, and Fleming, 1942: p. 702), these larvae could not have drifted very far from the spawning sites. Thus, skipjack spawning appears to occur tiuoiigh- out the sampled area. Matsumoto (1958) has reported larval skipjack catches from long. 180° to 120° W., and on the basis of records of larvae and juveniles taken in the Pliilip- pine Islands (Wade, 1951) and off the coast of Cen- tral America (Schaefer and Marr, 1948; Mead, 1951) and of juveniles caught in the Marshall Islands (Marr, 1948), he has indicated the possibility that skipjack spawn throughout the equatorial waters of the Pacific. Subsequently, Klawe (1903) noted the occurrence of larval skipjack in the eastern tropical Pacific. Matsumoto (unpublished) recently ob- tained larval skipjack from areas west of 180°, par- ticularly around the Marshall Islands and the east- ern part of the Caroline Islands. Capture of larval skipjack in localities still farther west in the Mari- anas and Palau Islands was reported by Yabe, Yabuta, and Uoyanagi (1903). These records con- firm Matsumoto's hypothesis of the transoceanic distribution of larval skipjack in the Pacific. Matsumoto (1958) also has reported the north- south distribution of larval skipjack as extending from lat. 25° N., to 1432° S. in the central Pacific. The southern limit now may be extended to at least lat. 18° S. Table 8 shows the months during which larval skipjack have been taken in northeastern French Oceania on various cruises by vessels of the Bureau of Commercial Fisheries. They were captured in all months except July, in which no sampling was 6 U.S. FISH AND WILDLIFE SERVICE done. Thus, skipjack spawiing can be inferred to occur throughout the year in these waters. Yoshida (1965) hkewise concluded from a study of skipjack ovaries that skipjack spawn year-round in the Marquesas. Yoshida also concluded that spawning is greatest from Novemlier through April. Although the seasonal distribution of larval skipjack (fig. 3) is consistent with Yoshida's results, the data do not permit a comparison for all seasons. Table 8. — Months, years, and cruises during wliich Uirnd skiii- jack hare been captured in northeastern French Oceania. Ital- icized cruises sampled the Marquesan offshore survey area. [Data prior to October 1957 from Matsumoto (1958) and Strasburg (1960)] Month Year 1952 1956 1957 1958 HMS-i3 HMS-38 CHG-^IS HMS-38 CHG-31 HMS-i3 HMS-3.3 CH0-3S CHG-38 HMS-iS H MS-IS HMS-iS July CHG-30 CHG-30 CHG-35 CHGSf CHG-SS November HMS-18 SUMMARY 1. The results of a study of the distribution of larval tunas in IMarquesan waters are presented. Data were collected in 1957 and 1958 on repeated transits of a standardized offshore survey pattern and on repeated visits to a single station where diel variability of larval abundance was studied. 2. Larval tunas were sorted and counted from 113 plankton samples from the offshore surveys and 92 from the diel variability station. Larval abun- dance is expressed as the number of larvae under 10 m.= of ocean surface down to a depth of 140 m. 3. Duplicability of larval catches by oblique tows taken at niglit or day was demonstrated. 4. Greater abundance of larval skipjack during darkness was evident at the diel variability station in December 1957 and March 1958 and of larval yellowfin in January 1958. Greater abundance of larval skipjack at twilight was found in April and .June of 1958. Results of attempts to determine whether larvae were below 140 m. were inconclusive. 5. Data from the offshore surveys indicate greater abundance of larval tunas during the Mar- quesan summer and fall (January to April) than during other months. G. Larval skipjack have been collected through- out the area of northeastern French Oceania bounded by long. 130° W. and 147° W. from the Equator to lat. 18° S. 7. High and low catches of larvae occurred in samples of high as well as of low plankton volume. 8. Average abundances of zooplankton, larval tiuias, and tuna schools for the offshore surveys were lowest during CHG-35 (Oct.-Nov. 1957), highest during either HMS-43 (Jan.-Feb. 1958) or CHG-38 (Mar.-Apr. 1958), and intermediate during HMS-45 (May 1958). 9. According to records of the localities of the capture of larvae and juveniles, skipjack spawn throughout the tropical and subtropical zones of the Pacific Ocean. In northeastern French Oceania, skipjack appear to spawn throughout the year. The data are consistent with the conclusion reached from a study of skipjack ovaries that the spawning of skipjack in northeastern French Oceania is most active from November through April. LITERATURE CITED Austin, Thom.\s S. 1957. Summary, oceanographic and fishery data, Mar- quesas Islands area, August-September, 1956 (EQUA- PAC). U.S. Fish Wild!. Serv., Spec. Sei. Rep. Fish. 217, v-|-186pp. King, Joseph E., Thomas S. Austin, and M.\xwell S. Doty. 1957. Preliminary report on expedition EASTROPIC. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 201, iv-|-155 pp. King, Joseph E., and Joan Demond. 1953. Zooplankton abundance in the central Pacific. U.S. Fish Wildl. Serv., Fish. Bull. 54: 111-144. Klawe, Witold L. 1963. Observations on the spawning of four species of tuna {Neothunnus macropterus, Katsuwonus pelamis, Auxis thazard and Euthynnus lineatus) in the eastern Pacific Ocean, based on the distribution of their larvae and juveniles. Inter-Amer. Trop. Tuna Comm. Bull. 6(9): 449-540. 1\I.\RR, John C. 1948. Observations on the spawning of oceanic skipjack {Katsuwonus pelamis) and yellowfin tuna {Neothunnus macropterus) in the northern Marshall Islands. In Schaefer, Milner B., and John C. Marr, Contributions to the biology of the Pacific tunas, pp. 201-206. U.S. Fi.sh Wildl. Serv., Fish. Bull. 51. Matsumoto, Walter M. 1958. Description and distribution of larvae of four species of tuna in central Pacific waters. U.S. Fish Wildl. Serv., Fish. Bull. 58: 31-72. LARVAL TUNAS IN MARQUESAN WATERS 1959. Descriptions of EuthynTixts and Atixis larvae from the Pacific and Atlantic Oceans and adjacent seas. CarlsbcfK Foundation's Ocesmographical l"'xpedition Round the World 1928-30 and previous "Dana"- Expeditions. Dana Rep. 50, 34 pp. 1962. Identification of larvae of four species of tuna from the Indo-Pacific renion. 1. Carlsberg Founda- tion's Oceanographical Expedition Round the World 1928-30 and previous "Dana"-Expeditions. Dana Rep. 55, 14 pp. Mead, Giles W. 1951. Postlarval Meothunnus macropterus, A uxis (hazard, and Euthijnnus lineatus from the Pacific coast of Cen- tral America. U.S. Fish Wildl. Serv., Fish. Bull. 52: 121-127. SciiAEFER, Mii.NER B., and John C. Makh. 1948. Spawning of yellowfin tuna (Xeothunnus macTop- lerus) and skipjack {Katsuwonus pelamis) in the Pacific Ocean off Central America, with descriptions of juve- niles. In Contributions to the biology of the Pacific tunas, p. 187-196. U.S. Fish Wildl. Serv., Fish. Bull. 51. SiEGEL, Sidney. 1956. Nonparametric statistics for the behavioral sci- ences. McGraw-Hill Book Company, Inc., New York. 312 pp. SxRASBrRG, Donald W. 1960. Estimates of larval tuna abundance in the central Pacific. I'.S. Fish Wildl. Serv., Fish. Bull. 60: 231- 255. SvERDRVP, H. v., Martin W. Johnson, and Richard H. Fleming. 1942. The oceans; their physics, chemistry, and general biology. Prentice-Hall, Inc., New York, 1087 pp. Wade, Charles B. 1951. Larvae of tuna and tuiia-like fishes from Philip- pine waters. I'.S. Fish Wildl. Serv., Fish. Bull. 51: 445-485. WiLso.N, Robert C, Eugene L. Nakamira, and Howard O. Yo.SHIDA. 1958. Marquesas area fishery and environincntal data, October 1957-June 1958. U.S. Fi.sh Wildl. Serv., Spec. Sci. Rep. Fi.sh. 283, vi-|-105 pp. Wilson, Robert C, and Murice O. Rinkel. 1957. Marquesas area oceaiiographic and fishery data, January-March 1957. I'.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 238, v-|-136 pp. Yabe, Hiroshi, Yoichi Yabita, and Sno.ii I'evanagi. 1963. (\)m])arative distribution of eggs, larvae and adults in relation to biotic and abiotic environmental factors. Proceedings of the World Scientific Meeting on the Biology of Tunas and Related Species, 2-14 July 1962, FAG Fish. Rep. (H. Rosa, Jr., editor), no. 6, vol. 3: 979-1009. Yoshida, Howard O. 1960. Marquesa.s area fishery and environmental data, January-March 1959. I'.S. Fi.sh Wildl. Serv., Spec. Sci. Rep. Fish. 348, v-|-37 pp. 1965. Skipjack tuna spawning in the .Maniuesas Islands and Tuamotu .\rchipelago. I'.S. I'ish aiul Wildl. Serv., Fish. Bull. 65(2): 479-488. U.S. FI.SII ASD WILDLIFE SERVICE APPENDIX Table A-1. — Data on planhton hauls and numbers of larval tunas collected at the diel variahiliiy station (9°3I,' S., tSS'SO'W.) Charles H. Gilbert cruise 35 Station Date (1957) Collection time (-H9ZT) Depth of tow Water strained Larvae in sample' SJ YF BE AL EU AU UN Total 18 Oct. 21.... ...do ...do ...do. ...do ...do ...do ...do ...do ...do ...do ...do Oct. 22 ...do ...do ...do Dec. 1 ...do ...do. ...do ...do. ...do ...do ...do ...do ...do ...do ...do Dec. 2 ...do ...do ...do 0739-0811 0814-0845 103.5-1105 1108-1137 1335-1405 1406-1434 1632-1703 1706-1737 193:5-2005 2007-2037 223.V2305 2306-2338 0143-0213 0214-0244 0431-0459 0501-05:52 0608-0638 0639-0711 090.5-0936 0939-1008 1205-12:57 1239-1310 1506-1,537 1539-16 1805-1835 1836-1907 2103-21.33 2134-2006 0005-0035 0037-0106 0304-0334 0336-0405 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-143 0-140 0-140 0-140 0-140 0-140 0-143 0-140 0-140 0-140 0-140 0-142 0-142 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 MJ 2185. 1 2112.5 1927. 1895. 9 1789.0 1860.8 2082.5 2068.2 2331.2 1842.2 1735.6 1979. 1 20.39. 6 2078.5 1733.9 2073. 1 1705.9 1936. 4 1709. 6 1917.3 1681.6 1599.5 176.5.8 1809. 1:561.9 1352 2 1375. 9 1503.4 1481.8 1516.9 1564.6 1649.7 A'o. 5 4 2 13 No. No. No. No. No. No. No. 5 1 3 K?) 6 4(?) 17 1 5 11 5 9 2 2 5 7 17 6 11 19 6 6 6 6 5 7 12 11 14 18 26 8 1 1 .. 4 10 1 10 1 3 1 1 3 119 17 1 I 20 6 2 2 10 1 7 1 2 9 5 128 5 13 23 30 1 14 I 2 2(?) 1(?) IC?) 30 132 1 4 49 56 8 1 ii3 =Kalsuwonus pelamis: YF = Thunnus albacares: BE=Thunnus obesus: At=Thunnus alalunga: EV ^Eulhimnus affinis; AV =.iuxis sp.: UN = unidentified tuna. . . , , , i .,, ^ / i- * j, f ii 2 Sample considered atypical because salps comprised 95 percent (estmiated) of tlie volume and 71 percent (estimated) of the organisms. Table A-2.— Data on plankton hauls and numbers of larval tunas collected at the diel variabititij station {9°3y S., 139°50' W.) on Hugh M. Smith cruise 45 Station Date (1958) Collection time ( +9ZT) Depth of tow Water strained Larvae in sample' SJ YF BE AL EU AU UN Total Jan. 23 ...do ...do ...do Jan. 24.... ...do ...do ...do ...do ...do ...do ...do 1622-1655 1825-1855 2005-2040 2207-2238 0013-0048 0210-0241 0408-0438 0615-0649 0803-08:52 1012-1042 1208-1245 1401-1431 M. 0-145 0-145 0-158 0-140 0-140 0-141 0-148 0^158 0-140 0-149 0-158 0-140 1991.9 2023.7 2104. 9 1711.4 2223.4 1887. 6 1787. 1716.8 1528. 1568.6 1986.8 1559. 6 A'o. 2 3 A'o. 5 2 11 11 13 4 1 A^o. A'o. A'o. A'o. A'o. A'o. 7 4 2 5 4 4 2 4 22-4 4 20 ' 2 1 1 20 22-7 7 1 1 2 2 5 i 8 ^ ^i =Katsuwonus pelamis; YF unidentified tuna. = Thunnus albacares: hE =ThuTtnus obesus: AL=Tfiunnus alalunga: Elj =Euthynnus affinis; AU=,'Iujis sp.; UN = LARVAL TUXAS IX MARQUESAX WATERS Table A-3.- -Dala on plankton hauls anil numbers of larml tunas collccteil at the die! mriahiHty station (0°34' S., 139°50' W.) Chaile.s H. Gilbert cruise 3S Station Date (1958) Collection time ( +9ZT) Depth of tow Water strained Larvae in sample' SJ YF BE AL EU AV UN Total 34-A Mar. 6.... ...do ...do ...do ...do ...do ...do ...do Mar. 7.... ...do ...do ...do Apr. 17... ...do ...do ...do ...do ...do Apr. 18... ...do ...do ...do ...do 0S06-0830 1003-10:i3 120.5-1237 llOS- 14.38 lli(W-I636 1S{)7-1837 2ll(liV-2036 221(1-2240 {l«0IV(l(>37 0205-0235 0403-0434 O604-06:i4 1311-1341 1503-15.33 1702-1732 1903-19.33 2102-2132 2306-2.336 0102-C1.13 0302-0332 0.503-0.5:i3 0702-0732 0903-0934 M. 0-140 0-141 0-140 0-140 0-140 0-142 0-142 0-140 0-142 0-140 0-142 0-140 0-140 0-140 0-140 0-142 0-140 0-140 0-142 0-142 0-140 0-142 0-140 A/. J 1885. 5 1476. 7 1896. 6 1938.8 1486.8 1949. 7 1902. 2 1624.9 1737.6 1907.0 1884. 9 1703.9 1608.2 1 629. 9 1512.7 1847.0 16:!0. 8 1719.8 1697.4 16tiS.4 147S. 7 1478.4 1696. 3 1573.3 No. 3 No. No. No. A'o. No. No. No. 3 34-B 1 34-C 3 1 1 34-1) 34-E 34-F I 34-0.. 4 10 5 8 14 8 1 34-H... 1 34-1 1 34-J ' 34-K... 14 93-A 11 ft3-B 93-C ■ 9.3-D. 4 93-E n 2 2 9.3-F 2 2 1 3 3 2 2 93-0... ft3-H 1 93-1 3 93-J 93-K 2 2 93-L ...do 1102-1132 . ' ^^—^"'""^o'""' Pelamis; YF = Thunnua albacares: hE=Tkunnus obesua: AL^Thunnus alalunga: KV = Eudiimnua affinis- AU = unidentified tuna. ^ Cable meter failed. .4uxi8 sp.; UN = T.\BLE A-4. -Data on plankton hauls and numbers of larral tunas collected at the diet rariabiliti/ station {!)°34 S., 1,S0''.50' W.) on Hugh M. Smith cruise /f'> Station Date (1958) Collection time ( +9ZT) Depth of tow Water strained Larvae in sample' SJ YF BE AL EU AU UN Total 120-1 Junes ...do ...do ...do ...do ...do ...do ...do ...do ...do June 9 ...do ...do ...do ...do ...do ...do ...do ...do ...do ...do ...do ...do ...do 1.522-1552 1522-1.552 1710-1739 1710-1739 1914-1941 1914-1941 210.5-21.33 2105-2133 2313-2347 2313-2347 0110-0144 0110-J? 0308-0335 0.308-0335 0510-05;i9 0510-05:J9 0711-0737 0711-0737 0910-0937 0910-0937 1108-1149 1108-1149 1306-13.33 1306-1333 A/. 0-122 112-2.30 0-125 96-239 0-126 112-2)8 0-126 71-238 0-137 103-240 0-137 70-240 0-121 107-229 0-126 103-2!8 0-125 94-245 0-126 76-2i8 0-1.50 71-262 0-123 116-239 Ar.3 937. 5 606. 7 797.5 445. 845.2 1041. 1 941.4 100.3.0 865.6 394. 852. 6 1482.7 794.0 598.8 817.8 681.1 784.3 447.1 718.5 939. 8 8,55. 4 •71.3 702.0 920.6 A'o. A'O. A'o. A'o. A-o. No. No. A'o. 120-2 121-1 3 121-2 :i I22-I 3 122-2 123-1.. 123-2... 124-1.. 21 124-2 ti 125-1 i 125-2 126-1 126-2 127-1 4 4 128-1. 128-2. 129-1 _ f) 131-2 1 identi^irrumr"'"""" P^'"'""' YF = rftunnua albacares; BE =Tkunmis obeaus; \L=Thun7tu8 alalunga; EU =EHthynnus affinU; AU =Anxis sp.; VS -- ' A 27-tniii. juvenile. ' Messenger time was not recorded. * I-'lowmeter recording was extremely low. 10 U.S. FISH AND WILDLIFE SEUVICE Table A-5. — Data on larval tunas collected during the of shore survey on Charles H. Gilbert cruise S5 station 35.. 36.., 38.. 38A 38B 39.. 41.- 42.. 44.. 45.. 47.. 48.. 50.. 61.. 53.. 54.. 55.. 56.. 58.. 59.. 60.. 61.. 62A 63.. 65.. 66-. 67.. 68.- 70-. 71.. 73.- 74.. 76.. 77.. 79.. 80-. 82.. 83- 84.. 85-. 87.. 88-- 90.. 91.. 93.. 94.. Long. W. Date (19.57) Collection time ( +9ZT) Depth of tow Water strained Larvae in sample' YF AU UN Total 9°16' 9° 16' 9°22' 9°22' 9°14' 9°14' 9»14' 9°14' 9'=17' 9°I7' 9°17' 9°17' n°03' ll'OS' II "22' J1022' 12°23' 12°23' 12°09' 12°09' 9°36' 9°36' 7°20' 7°20' 7°06' 7°06' 6°04' 6''04' 6°26' 6°26' 8°36' 8°36' 8°54' 8°54' 9-13' 9°13' 9°12' 9° 12' 9°15' go,- 9°18' 9° 18' 9° 15' 9"=15' 9°13' 9°13' 137°62' 137°52' 137°30' 137''30' 136°20' 136°20' 136°34' 136''34' 1.39°02' 139°02' 139°I6' 139°16' 1.39°.33' 1.39°33' 139°27' 139°27' 139''36' 139°36' 1.39°.30' 1.39°30' 139°44' I39°44' 139°32' 139°32' 139°30' 139°30' 139'>.50' 139°50' 1.39°52' 139°52' 139°3r 139°3I' 139''38' 139''38' 14r35' 141°35' 142°00' 142°00' 142''46' 142°46' 142°23' 142°23' 140°33' 140°33' HO^IO' 140''10' Oct. 24. ...do... Oct. 26. ...do... ...do... ...do... Oct. 26. ...do... ...do... ..-do... Oct. 27. ...do... ...do... ...do... Oct. 28. ...do... ...do.- ...do... Oct. 29. ...do.. Oct. 30. ...do.. Nov. 1. ...do - Nov. 2- ...do.- ...do.. ...do.- Nov. 3. ...do.. ...do-- ...do-- Nov. 4- ...do-. ...do.. ...do.- Nov. 5. ...do.- ...do.- ...do-- Nov. 6. ...do.. ...do.- ...do-- Nov. 7. ...do.- 2301-2329 2333-0001 02.59-0330 0.332-C40I 22.56-2327 2329-0001 02.58-0329 0330-0401 2269-2329 2.331-0003 0254-0324 0326-0357 2300-2330 2332-0000 0268-0328 0330-0401 2300-2.331 2332-0000 0268-0327 0329-0358 0300-0328 0329-0400 2268-2329 2.331-2359 0257-0327 0332-0400 2267-2327 2329-0000 0302-0333 0349-0419 2305-2334 2335-0005 0300-0330 0334-0404 2268-2328 2331-0002 0256-0326 0328-0366 2259-2329 2332-0004 0258-0329 0334-0403 2256-2326 2328-2358 0268-0328 0332-0402 M. 0-182 0-140 0-140 n-140 0-140 0-140 0-173 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-HO 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 0-140 A/.' 1681. 1546. 1316. 1487. 1760. 1797 175C. 1694. 1486 1582 1558. 1463. 15,30. 1354. 1533. 1666. 2011. 1722. 1627. 145.3. 1134. ia36. 1663. 1414. 3365. 1633. 1788. 1915. 1807. 1765. 1390. 1408. 854. 715. 2218. 2168. 1813. 1786. 1490. 1580. 1496. 1449. 1267. 1192. 1264. 1074. No. No. No. No. No. No. 3 8 8 7 1 5 2 5 8 6 5 6 1 6 2 4 6 8 7 5 17 15 1 1 4 12 1 5 6 1 SJ — Katsuu'onus pelamis: YF = Thunnus albacares; BE unidentified tuna. - A juvenile about 25 mm. long. = Thunnus obesus: .\.Jj=Thunnu$ alalunga; T^V = Euthynnus ajfinis: .W =Auxis sp.: UN = Table A-6. — Data on larval tunas collected during the offshore surrey on Hugh M. Smith cruise 4J station Position Date (1958) Collection time ( +9ZT) Depth of tow Water strained Larvae in sample' Lat. S. Long. W. SJ YF BE AL KU AU UN Total 30 32 34 36 39 41 42 44 47 49 67 69 60.. 62 64 66 70 9°12' 9° 12' 9°10' 9°10' 9°37' lO^H' 12°02' 12°31' 11°16' 10°43' 9°07' 9°08' 9°13' 9° 15' 9°13' 9° 16' 7°35' 6°52' 5°37' 6°21' 8°30' 9°03' 139°I7' 138°48' 136°50' 136''18' 139°40' 139°38' 139'>36' 139°34' 139°42' 139°39' 141 °05' 141°32' 143°02' 142°26' 140°41' 140°06' !39°40' 139°40' 1.39°50' 139-38' 139°40' 139°45' Jan. 27...- Jan. 28.... ...do Jan. 29.... Jan. 30 Jan. 31.... ...do Feb. 1 ...do Feb. 2 Feb. 5 Feb. 6 ...do Feb. 7 ---do Feb. 8 Feb. 9 Feb. 10.... ...do Feb. 11.... ...do Feb. 12.... 2116-2161 0313-0342 2110-2140 031.3-0343 2109-2139 0312-0342 2112-2142 0313-0343 21 10-2141 0313-0343 2118-2145 0319-0344 2113-2143 0315-1.337 2109-21.33 0318-0348 2110-2137 0.330-0359 2113-2142 0319-0347 2120-2146 0312-0341 A/. 0-141 0-160 0-147 0-145 0-148 0-148 0-164 0-185 0-141 0-141 0-141 0-142 0-140 0-145 0-167 0-140 0-140 0-141 0-140 0-141 0-140 0-140 M.' 1755.7 1300. 1750.6 1617.9 1688.4 1516.8 1466. 3 11.58.6 1551. 1 682.7 1264.6 1384. 6 1403. 1 1468.9 1039. 8 1511.9 1164.9 1386. 1 1363. 7 1279.6 957 7 No. A'o. No. A'o. No. A'o. A'o. A'o. 2 13 6 ""'' 2 1 5 14 6 3 10 3 12 3 3 10 3 3 2 17 1 4 13 6 2 13 6 1 13 3 8 1 1 1 14 3 73 76 76 78 1 9 1 1 1199.9 1 6 ' 4 10 ' ?^J =Katsuwonus pelamis unidentified tuna. \F = Thunnus albacares: BK =Thunnus obesus: .\h=Thunnus alalunga: KV =Euthynnus alUnis; AV =Auxis sp.; UX ^ LARVAL TUXAS IX MAKQUESAX WATERS 11 Table A-7. — Data on larval hums collected during the offshore survey on Charle.s H. Gilbert cruise SS Station Position Date (1958) Collection time { +9Z'r) Depth of tow Water strained Larvae in sample' Lat. S. Long. W. SJ YF BE AL EU AU UN Total 46 9-11' 9-09' 9-1 r 9-07' 9-09' 9-08' 7-32' 6-57' .5-44' 6-20' 8-10' 8-38' 9-08' 9-07' 9-0.5' 9-04' 9-09' 9-12' 10°,52' 11-28' 12-32' 12-00' 10-02' 1.38-06' 137-34' 1.36-15' 136-56' 138-53' 1.39-27' 139-46' 139-44' 139-40' 139-39' 139-44' 139-57' 141-14' 141-52' 142°,58' 142-28' 140-28' 139-57' 139-45' 1.39-48' 139-43' 1.39-41' 139-44' 1,39-51' Mar. 26... Mar. 27... ...do Mar. 28... ...do Mar. 29... ...do Mar. 30... ...do Mar. 31... ...do Apr. 1 Apr. 3 Apr. 4 ...do Apr. 5 ...do Apr. 6 ...do Apr. 7 ...do Apr. 8 -..do Apr. 9 2204-2234 0405-0435 2104-2135 0304-0334 2122-2151 0303-0332 2104-2134 0305-0,336 2102-2132 0307-0.337 2103-2133 0306-0.336 2104-2134 0309-0339 2102-2133 0304-0.334 2103-2132 0309-0,339 2107-2137 0304-0335 2103-2133 0305-0,335 2100-2130 0302-0,332 Af. 0-140 0-142 0-142 0-140 0-140 0-140 0-142 0-142 0-140 0-140 0-142 0-142 0-140 0-140 0-142 0-142 0-140 0-140 0-142 0-140 0-140 0-140 0-142 0-142 1416.2 1647.2 1703.6 1776.7 1.394. 2 1.338.6 1275.6 1363. 8 1644. 4 1,505. 8 1475.9 14,56. 4 1.54.3.7 1751.3 1885. 5 1657.8 1212.8 13.33.3 1407.8 No. No. No. No. No. No. No. No. 49 61 5:! 1 5 6 6 16 57 59. 60 62 « 66 69 14 29 2 4 12 13 8 1 6 9 3 2 10 2 4 18 2 5 1 1 16 S 72 74 75 6 // 78 80. 81 S3 84 8 (1 6 1627. 7 1,546.5 1921.8 1406. 7 1536.7 6 5 12 1 6 1 12 6 1 i^J =A'a(siiiranus pelamis: YF = Tliurmua albacares; liF, =Tliunn!ts obesus: .\L = unidentified tuna. 'Thunmis alalunga; EU =Eulhynnixs affinis; AV =Auxi) sp.; UN = Table .\-8.- —Datii on larval liitias collected duritig the offshore survey on Hugh M. Smith cruise J,.5 Station Position Date (1958) Collection time (-F9ZT) Depth of tow Water strained Larvae in sample' Lat. S. Long. W. SJ YF BE AL EU AU UN Total 78 79 82 83 m 87.. 88 9-08' 9-12' 9-10' 9-13' 9-12' 9-12' 9-12' 9-12' 9-14' 9-12' 7-39' 6-54' 5°:!8' 6-24' 8-25' 10°.55' 11-30' 12-41' 11-58' 10-11' 9-25' 1.38-10' 137-22' 136-10' 137-02' 139-07' 141-22' 142-02' 143-08' 142-20' 140-.36' 139-45' 1.39-41' 1.39-40' 139°.38' 1.39-42' 139-40' 139-39' 139°,39' 139-40' 139-34' 139-34' May 15... May 16. . . ...do May 17... ...do May 19... May 20... May 2lV.'' ...do May 23... May 24... ...do May 25... ...do May 27... May 28... May 29." May 30.." 2004-2030 0317-0346 2000-2028 0320-0347 1959-2029 2000-2029 fl3I3-0:)43 19,55-2026 0309-0.338 2002-2031 2001-2030 0314-0343 19.5S-202X 0310-0344- 19.5'.)- 2034 2000-2028 0312-0.339 2001-2040 0306-0'!36 1953-2,125 0315-(351 M. 0-140 0-161 0-140 0-139 0-125 0-140 0-137 0-140 0-140 0-137 0-140 0-140 0-1. 'W 0-140 0-137 0-139 0-144 0-1.33 0-124 0-140 0-140 1392. 8 1611. 1 1432. 8 1321.9 1601.3 1,578. 1 13.57.3 4 34 6 2 3 No. No. No. No. No. No. A-o. 4 34 2 1 4 «1 94 ■97 «8 101 102 105 107 108 110 111... . 113 114 15;!6. 1,592.1 13.59. 1 1421.6 1426.5 2 3 1 5 4 1 1 1 9 Ifl (( Q 1491.0 1 7.59, 2 1926. 3 1301.2 216,5.0 2347. 5 14.59. 1 1928.4 944.9 2 2 2 4 -. 3 1 3 2 4 is' i" ir. ' S.I — hatsuu'onus unidentified tuna. pelamis; YF =Thunnus albacares: BE=Thunnus obesua; AL=Thunnus alalunga: EV =Euthynnu8 affinis: AV =A uxis 8p.; UN = 12 U.S. FISU AND WILDLIFE SERVICE ASSOCIATION OF FISHES WITH FLOTSAM IN THE OFFSHORE WATERS OF CENTRAL AMERICA By John R. Hunter, Fishery Biologist (Research), and Charles T. Mitchell Fishery Aid Bureau of Commercial Fisheries Tuna Resources Laboratory, La Jolla, Calif. 92038 During April, May, June, and October, 1963, a total of 70 purse seine collections were made of the fishes associated with floating objects. Nearly all of these collections were from the offshore waters of Costa Rica. Twelve families of fishes (Lobotidae, Carangidae, Cory- phaenidae, Mullidae, Kyphosidae, Pomacentridae, Scombridae, Blenniidae, Stromateidae, Mugilidae, Polynemidae, and Balistidae) and 32 species were repre- sented in the collections. Most of the species were present during both spring and fall, and nearly all of the fishes were juveniles. Nine of the 32 species, including the 2 most abundant ones, Caranx caballus Giinther and Selar crumenoph- thalmus (Bloch), were carangids. The lengths of two species, Abudefduf saxatilis (Linnaeus) and Seriola sp. were greater the farther an object was located from shore. Some species such as C. caballus. Psenes paci- ficus Meek and Hildebrand, and Canthidermis macu- latus (Bloch) were present in almost a complete series of juvenile stages; others as Chromis atrilobata Gill, ABSTRACT Pseudupeneus grandisquatnis (Gill), and Agonostomus monticola (Bancroft) were represented by only a single juvenile stage. More fishes were collected under large objects than under small objects. The total number of individuals present near moored objects after 5 days did not differ from the numbers present after 20 or more days. The coloration of fishes was related to their as- sociation behavior. Silvery colored fishes did not re- main as close to the object as did the more darkly colored species. Most adult fishes, which did not re- main as near the object as did juveniles, appeared beneath an object only intermittently. Canthidermis maculatus, however, maintained close contact with drifting objects both as adults and juveniles. Observations of the behavior of species are discussed in relation to the mechanisms for the association of fish with flotsam that have been postulated by other authors. None of their hypotheses was supported by our data. Additional mechanisms were postulated. The association of fishes with floating objects has been exploited by a number of fislieries. .Japanese pole-and-line fisheries and American purse seine and Hve-bait fisheries take advantage of the association of yellowfin tuna, Thunnus albacarcs (Bonnaterre), and oceanic skipjack, Katmwonus prlamis (Lm- naeus), with algae, logs, and other flotsam (Uda, 1933; McXeely, 1901). Uda and Tsukushi (1934), and Yabe and Mori (1950) reported that log-asso- ciated schools of tuna provide a consistently higher yield per unit fishing effort than unassociated schools. Moored rafts of bamboo or palm fronds are used to attract dolphin-fish, Coryphaena htppurus (Lin- naeus), in seine fislieries of Japan (Kojima, 1955, 1956, 1960a, 1960b, and 1961). Moored cork-slabs serve the same purpose for Maltese fishermen (Galea, 1961). Two types of palm-frond rafts are used by Indonesian fishermen to attract various clupeids, scombrids, Decapkrus spp., and other carangids (Hardenberg, 1950; Soemarto, 1960). In addition to these commercially important species, many others of lesser or no commercial value are also encountered (Murray and Hjort (1912), Yabe and Mori (1950), Uchida and Shojima (1958), Besednov (1900), Kojima (1960a), Mansueti (1963), and Gooding and Magnuson'). Note; Approved for piiblication Marrti 8, 196fi. ' ReRinald M. Gooding and John J. Magnuson — C)baervations on the ecology and beha\ ior of fishes around a drifting raft near Hawaii during the first 48 hours adrift. Manuscript. Bureau of Commercial Fisheries Bio- logical Laboratory, Honolulu, Hawaii. FISHERY bulletin: VOLUME 66, NO. 1 13 Gooding and Magnuson reviewed the hypotheses that have been advanced to explain this habit: (1) attraction by food (smaller fish, algae, decaying palm fronds, and plankton made more visible by the shade of the object); (2) negative phototaxis in response to the shadow cast by the object; (3) shel- ter from predators; and (4) use of the object as a spawning substrate. They also suggested an addi- tional hypothesis that floating objects are cleaning stations where pelagic fishes go to have their para- sites removed by other fish. This paper provides information on the ecologv and behavior of fishes associated with floating ob- jects in the offshore waters of Central America. Special attention is given to the frequency, abun- dance, and size of the species which compose flotsam- associated aggregations and how these characteristics are related to the location and size of the object. These studies are the framework upon which future behavior investigations will be based. The aim of our program is to determine whether a de\'ice can be designed tliat will be maximally efficient in aggregat- ing tuna and skipjack. The potential value to the tuna fishery of establishing svich devices has been discussed by .\lverson and W'ilimovsky (19G3). PROCEDURES Nearly all of our studies were in the offshore waters of Costa Rica (fig. 1) because yellowfin tuna and skipjack are often associated with the flotsam in this region (logbook records obtained thiough the courtesy of the Inter-American Tropical Tvuia Com- mission). Several collections were near the coast of southern Mexico and 1 near Cocos Islantl. Sam- ples were collected by encircling flotsam and its associated fauna with a small %-inch (11 mm.) stretch-mesh purse seine, 12 feet deep (3.7 m.) and 1 10 feet (33.5 m.) long (.\asted, MS.)=. An average of 60 percent of the fishes observed beneath an object were captured in the seine. Fish larger than 100 mm. standard length may have escaped the net, and fish smaller than 15 mm. occasionally swam through the webbing. When the net was set, fish tended to stay near the flotsam or even swim up- ward. Thus, fish swimming at a depth greater than the maximum depth of the seine also may have been caught. Sampling errors due to fish escaping from or entering the seined cylinder of water were probably small. - Donald C". .\astp; both stages were captured in the seine. To determine if the size of the fishes was related to the distance of an object from shore, the shortest distance to the shore from the location of each col- lection was measured to the nearest nautical mile. The length measurements of species from different collections captured at the same distance from shore were combined, and a mean and range were estab- lished (figs. 2, 3, and 4). The mean and minimum length of Al)udcf(hif saxatilis and Seriola sp. increased with the distance of an object from shore (chi-scjuare test for two inde- pendent samples, p<.01) — figs. 2 and 3. The 16 U.S. FISH AND WILDLIFE SERVICE 33 122 42 83 5326 2 26 I ABUDEFDUF SAXATILIS 3 DISTANCE FROM SHORE NAUTICAL MILES Figure 2. — Range and mean standard length of Abudefduf saxatiUs collected beneath flotsam in the offshore waters of Central America at various distances from shore. Broken lines indicate range for spring collection, solid lines, range for fall collections. Open circles indicate mean for spring, filled circles, mean for fall; circles without bars indicate single fish. Upper numerals are total number of Abudefduf captured in spring of 1963, lower numerals, fall 1963. ~ _ . 2 A - 3 - - 1 - - SPRING ? T5 6 1 19 4 1 10215 - 1 17 6 FALL 160 SERIOLA SP 150- MO 130 120 [ 110 100 4 -| 90 _ L. 80 70 60 ■ • • 50 40 30 1 20 1 t ■• 10 i J 5 10 15 20 25 30 DISTANCE FROM SHORE NAUTICAL MILES Figure 3.— Range and mean standard length of Seriola sp. collected beneath flotsam in the offshore waters of Central America at various distances from shore. Broken lines in- dicate range for spring collections, solid lines range for fall collections. Open circles indicate mean for spring, filled circles, mean for fall; circles without bars indicate single fish. I'pper numerals are total number of Seriola sp. cap- tured in spring of 1963, lower numerals, fall of 1963. maximum size of these two species did not show this change. The mean length of Selar crumenophtJial- mvs also increased with distance, but this change was not as marked as in the other two species (fig. 4). 130 T 120 60 50 40 30 -i 20- 10- 72 32 44617517136 9 95 - SELAR CRUMENOPHTHALMUS 1 1 t 1 ' ] . , 1 - : ] J L. 5 10 16 20 2= 3U DISTANCE FROM SHORE NAUTICAL MILES Figure 4. — Range and mean standard length of Selar cru- menophlhalmus collected beneath flotsam in the ofTshore waters of Central America at various distances from shore. Broken lines indicate range for spring collections, solid lines, range for fall collections. Open circles indicate mean for spring, filled circles, mean for fall. Upper numerals are total number of Selar captured in spring of 1963, lower numerals, fall of 1963. Abudefduf sa.ratilis spawns inshore on rocks or reefs, and the males defend the clutch of eggs (R. Rosen- blatt, personal communication); thus, larvae and juveniles of this species may be more abundant in- shore. It seems possible that individuals captured offshore were originally recruited to the object in- shore and accompanied it as it drifted away from land. The larger size of the offshore specimens could be attributed to growth while the fish were associated with the object. In the remainder of the fishes no obvious relation was evident between the distance from shore and the minimum or mean length; however, the ranges of sizes at which these fishes were associated with flotsam differed widely. Some species were repre- sented by almost a complete series of juvenile stages. Caranx caballus and Psenes pacificus are good exam- ples of this group (fig. 5). Less abundant species in this group were Elagatis bipinnulatus, Kyphosus ele- gans, Sedator ocyurus, and Canthidermis maculatus FLOTSAM IN OFFSHORE WATERS 17 PSENES PACIFICUS N-822 jJBUkiui.^ PSEUOUPENEUS GRANOtSOUAMtS N'339 oi — , — ^Jm^K^L^ , — , JO «Q 60 BO ;-Afi;ANX CABALLUS N'6.ai5 CHROMIS ATRILOSATA H J 160 leO zoo J20 O 20 40 STANDARD LENGTH Milt IMF TFRS associated aggregations of fish: (1) frequency— the total number of collections in which a species oc- curred; (2) abundance — the range and median of the numbers of individuals per collection in which the species was found; and (3) dominance— the iiiinibor of samples in which a particular species or a group including this species comprised 50 percent or more of the total number of indi\'i(luals in a given collection. As the structure of the poinilations in the spring was similar to that in the fall, the two series of collections were combined to cak'ulate tiiese statistics. Fifteen of the 32 species had freciucncies greater than 10. 'i1ie.se were ranked from 1 to 15 on the basis of their fre(iuency, median abundance, and dominance. Tied \alues were given the average of the ranks (table 2 and fig. 6). The remainder of the species was ranked only by frequency (table 3). 1'"igi:re 5. — Length frequency for Psenes pacificus, I'scvilii- peneus grandisquamis, Caranx cabaUus, and Chromis atrilo- bata. Numbor.s arc totals for combined spring ami fall collections. (table 1). The size range of juveniles of other species was extremely restricted. Chromis alrilohata and Pscudiipcncns grandisquamis had this compact type of size distribution (fig. 5). Other species, not figured, which al.so had a limited size range included Agonoslomus monticoJa, Polijdactylus approximans, and Blcnniolus brevipiiuiis. Pscvduprnrus, Chromis, Agonoslomus, and Polydactylus have pelagic juvenile stages but as adults inhabit other areas. The upper size limit of these species in our collections probably was determined by the size at which they end the pelagic phase of their lives. Blcnniolus brevipinnis is a small species; females 19.5 mm. long can be se.xually mature (Krejsa, 19(i0). Adults and juveniles have been found near drifting logs as well as among rocks and coral heads in inshore areas (Krejsa, IDtiO). Apparently for both adults and juveniles of this species, drifting objects are a suitable pelagic substitute for inshore habitats. FREQUENCY, ABUNDANCE, AND DOMINANCE OF FISHES COLLECTED BENEATH FLOTSAM The characters used by Fager and McCiowan (19G3) for the analysis of zooplankton populations were used to describe the structure of the flotsam- FREQUENCY DOMINANCE ABUNDANCE FBEQUENCY DOMINANCE I ABUNDANCE I 1 — -H FnaiiK G. — Ecological characters of the 15 species most fre- (|Uently captured beneath llot.sam in the offshore waters of Central America in 196:5. Ivicli species was ranked sepa- rately hy.fmiiinicij, the total number of collections in which the species occurred; abundance, median of the numbers of individuals per collection in which the species was found; and dominance, the number of collections in which a specie.s was among those making up 50 percent of tlie individuals. Lines indicate the rank held by each sjiecies in the three rankings. Values upon which ranks wer<' based are shown in each column. In the second column, parentheses enclo.se th(! range of the number of individuals pi'r collection of occurrence. Lor clarity, we separated the 15 species into 2 groups: left half of figure, .species who.se ranked abundance was the same as or lower than the ranked frequency; and right half, species whose ranked abundance was higher than the ranked frequency. The total number of collections w.as 70. 18 U.S. FI.SH AND WILDLIFE SERVICE Table 2. — Ecnhgiral characters of the 15 most frequently captured species collected beneath flotsam in the offshore waters of Central America in 1963^ Species Frequency- Rank-i Abundance RankJ Dominance* R.inkJ Range' Median" 5S 19 14 19 29 12 50 27 26 11 41 25 18 43 12 1 9.5 12 9.5 6 13.5 2 6 7 15 4 8 11 3 13.5 (1-822) (1-31) (1-12) (1-99) (1-64) (1-12) (1-340) (1-74) (1-87) (1-161) (1-44) (1-292) (1-10) (1-135) (1-176) 45 2 2 7 :! 3 10 3 4 4 n 8 1 8 3 13.5 13.5 5 10.5 10.5 2 10.5 7.5 7.5 6 4 15 3.5 10.5 38 1 1 1 2 1 12 3 .■! 1 9 7 3 1 12 12 12 9 12 2 7 7 12 4.5 3 15 4.5 7 ' Tlie total number of collections was 70. ~ The total niunber of collections in wliich the species occurred. 3 Rank based on figures in adjacent columns- t ■ j- ■ i i < The number of samples in which a particular species or a Kroui) including this species comprised 50 percent or more of the total number of individuals in a Given collection, = RaUKc of the numbers of individuals per collection in which species was found. ^ Median of the numbers of individuals per collection in wiiich species was found. T.\BLE 3. — Frequency, abundance, and dominance of spenVs occurring in 10 or fewer collections made in the offshore waters of Central America in 1963. Listed in order of frequency Species Agnvonfom tat iiwnficola PoJijiimii/l'is a I'proximavs Kt/phnfin^ ibiiainf Alirqil rlijcma Bali.tfps itttliilepis Kitphnsns tiiinlofiiis Lohotes pnciftcii^ Seriola colhuriii Coriitihaeiid ti'tppurus CoTliptlaeim cquisetis Kiipbosus sp Pofrjdach/lits opercularis.. All. r is ttiazard. Eiilhi/nuits lineatus..', Aliilera motioceros Atiitera scripfa Frequency in 9 8 5 5 3 3 2 2 2 2 Abundance' (I-ll) (1-12) (1-12) (1-2) (1-1) (1) (1) (2-3) (I) (2-3) (1) (1) (1) (1) (I) (1) Dominance 1 Ficures in parentheses show range in number of individuals per collec- tion of occurrence. Nine of the 32 species were carangids, and all but 1 of these, Scriola colhurni, were among the 15 mo.st frequent species. The carangid, C. cabaUus was by far the most frequent, abundant, and dominant species collected. This fish contributed oO percent or more of the individuals in more than half of the collections. Selar crumcnophthalnius, also a caran- gid, ranked second in frequency, abundance, and dominance. Xo other family «as represented as frequently in the collections. The family Kyphosi- dae was represented by four species but only one, Sedator ocyurus, occurred in more than 10 collections. On the basis of their rank by frequency, abun- dance, and dominance, the 15 most frequent species can be divided into three groups: (1) species that occupied about the same rank in all three categories; (2) species that were captured frecjuentlj' but were not abundant in the collections in which they oc- curred ; and (3) species captured less freciuently that were abundant in the collections in which thej' oc- curred. The three highest ranking species, C. caballns, Selar crtmicnophthalmiis, and Pscnes paci- ficus were in the first group. Ahiidcfduf saxatilis and Blenniolus hrevipinnis exemplify the second group, and Chronu's atn'Iobatn and Canthidermis ma- cidatiis the third. The factors responsible for the tlifferences in fre- ciuenc.v and abundance of species are unknown. For some species, evidence suggested that schooling was a significant factor. All of the 15 most fre- quently captured species, except Abudcfduf and Blenniolus, schooled either with their own or other species beneath flotsam. Abudefduf remained near the object and appeare(.l to defend small territories; Blenniolus maintained contact with the surface of the object and were not aggregated. Possibly the solitary or individual habits of these species were responsible for their lower abundance. Juvenile Chromis school at the stage at which they associate with flotsam. This species was dominant in seven of eight collections made in the same area on the same day. The median number of individuals in these seven collections was 199. Chromis was domi- nant only once in the remainder of the collections, and the median number of fish per collection was two. The irregular abundance of Chromis could be ascribed to a tendency toward the recruitment of an entire school. FLOTSAM IN OFFSHORE WATERS 19 Canthidermis macitlalus also was a schooling spe- cies and tended to be abundant in the collections in which it occurred. This species did not show the limited temporal and regional abundance described for juvenile Chroynis, but the distribution of Can- Ikklerrnis appeared to extend farther to sea than other species. The collection farthest from shore (200 nautical miles) contained 87 Canthidermis and 4 Batistes polyUpis. Only Canthidermis was ob- served beneath other drifting material in the same area. The four B. polylrpis were located inside the cavit}^ of a large bamboo stem and probably did not represent a usual component of high-seas aggrega- tions. Had we taken more collections from flotsam drifting 100 or more nautical miles from shore we feel the frequencies for Canthidermis would have increased proportionately. Decapterus sp. ranked fifth in abundance but only once dominated a collection. This species nearly always schooled with Selar crumenophthalmus but was less abundant in the mixed schools. Decapterus was captured wiUiout Selar in only 7 of 19 collec- tions. Thus whenever a large number of Decapterus was taken, the number of Selar was usually larger. Hence, Decapterus rarely dominated a collection. The use of only the numbers of individuals for the determination of dominance instead of numbers and weights obscured some of the relations among species. Had weights as well as numbers been used, Canthidermis and Psenes probably would have tlominated more collections and Chromis, Pscudu- pcneus, and Ahudefduf fewer. Owing to their large size range and abundance, little difference would be e.xpected in the values for C. cahallus and Selar. OBJECT SIZE To determine if the length or the number of fishes was related to the size of the object, we recorded for each collection the volume of the object in cubic centimeters, the total number of fishes captured, and the mean length of all fishes in the collection. Of the two variables only the number of individuals in the collection was obviously related to the volume of the object. Collections made beneath large objects tended to be larger than those taken beneath smaller objects (table 4). Field observations indicated that the frequency of occurrence of larger fishes may be related to the size of the object. The largest object studied, an entire tree, was too large to be encompassed by the purse seine. The tree was 1 m. in diameter at the root section, had a trunk diameter of 0.3 m. and was over 10 m. long. Associated with the tree was the largest aggregation of adult fishes seen during the study. There were large schools of adult Sectator ocyurus, Canthidermis maculatus, Coryphacna hip- purus, and Euihynnus lineatus in addition to nu- merous juvenile fishes. A portion of the school of adult Sectator is shown in figure 7A, and in 7B some of the adult Canthidermis are shown among the branches of the tree. For comparison, two groups of juvenile fishes that were associated with two smaller objects are pictured in 7C and 7D. Yabe Table 4. — Number of collections made of fishes associated with objicts of three size classes and Die number of fislies these collections contained (Collections were made in the offshore waters of Central -America in 1963] Kish in collection Collections from objects of dilTerent volume (cubic centimeters) Total 101-5,000 5,001-100,000 100,001-5.000,000 Number 1-10.... 11-100... lOl-KWO.. .Number Number 14 19 Number 6 19 Number 8 22 38 Total.. 7 36 25 681 • Two collections omitted owing to lack of volume measurements. and Mori (1950) captured, by hook and line, fishes associated with a tree of similar dimensions (1 m. in diameter at the butt and 15 m. long). The lengths of the fish of the species they captured exceeded the lengths of the fish in our purse seine collections by about a factor of 10. In our study, the juvenile fishes that were associated with the tree were of the same species and about the same size as those col- lected beneath smaller objects. Thus the size of the object appeared to be related to the presence or absence of large or adult fishes rather than differ- ences among juveniles. ARTIFICIAL MOORED OBJECTS To study the rate of recruitment of fishes to float- ing materials, eight objects of various types were moored near the Costa Kican coa.st for periods of 15 hours to 4() days. Six objects were not visited from the time they were moored until the day the collec- tion was made. Two balsa logs were moored in the same locality at the same time and were obser\ed daily until collections were made on the fifth day. 20 U.S. FISH .\ND WILDLIFE SERVICE Figure 7 -Fishes associated with drifting objects in the offshore waters of Central America in 1963. A, a school of adult Scclator ocyurus associated with an entire tree; the small fish in background were juvenile Selar cmmenophlhalmus. B an aggregation of adult Canthidermis maculntus in the branches of the same tree; all but three had a dark coloration. C, a group of juvenile Pscnes pacificus. Xaucmks duclor, and other carangids beneath a drifting plank; Psenes are m a dense clump directly below the plank, Xaucrabs can be recognized by the presence of dark vertical bars. D, juvenile Canthi- dermis maculatus, Xaucrates duclor, and various juvenile carangids beneath a drifting log. Divers made daily counts of the number of individu- als of each species beneath each of the two logs. The volumes of the two balsa logs calculated from their measurements were 0.021 m.' (log A) and 0.065 m.' (log B). Counts of the number of indi\nduals beneath logs A and B were 20 and 96 for the second day and 121 and 80 for the third day. By the fourth day it was not possible to make an accurate tabulation because the number of fish under each log was well over 100. On the fifth day 198 individuals were captured under log A and 349 under log B. Prior to being moored log B was encountered 27 miles from shore and 236 fish representing 8 species were captured at that time. Thus more fish were captured after the log was moored 5 days than were collected when the log was drifting 27 miles from shore. Fewer species were represented, however. The larger number of indi- viduals captured beneath log B may reflect the difference in volume of the logs. Although logs A and B were moored only 100 m. apart, their associated fish populations differed in rLOTS.\M IN' OFFSHORE WATERS 21 species composition, dominant species, and tlie time at which each species was fu'st observed. Xo ordered recruitment of species was evident (table 5). The number of fish collected beneath inflated truck inner tubes varied greatly (table 6). All the tubes were identical in size and shape \vith the except- tion of the tube that had 10 manila lines attached. After periods longer than 5 days the number of fish collected beneath the tubes did not increase sub- stantially with time through 20 or more days. The number of fish appears to increase rapidly during the first few days and thereafter to remain at about the same level. Because a drifting object passes through inshore spawning areas, juveniles of species that spawn inshore would be expected to be more abundant beneath a drifting object than beneath an object moored offshore. Abndefduf sa.ratilis and perhaps Seriola sp. spawned inshore. Both species showed a Table 5. — Sj>ecies recruited to two balsa logs, A (mlumc, 0.021 m.' and B {volume, 0.065 /«.') moored 100 m. apart 7 nautical miles from the Costa Rican coast in 196S Log A Log B Species Day species first observed' Fish captured on fifth day Day species first observed' Fish captured on fifth day Pseudupe-neus grandisquamis .Sf/ar crume uojihthalmus 3 3 2 3 5 5 2 3 Number 87 79 17 8 6 1 2 3 2 3 4 5 4 3 4 Number 31 304 129 Chroniis atrilahata Elagntis hipiutiutatus Psene.^ paciltcus . 2 Blenitiolus hreripinnis 2 Polydadylua approrimans. . Aluiera Total _ 198 474 ' No underwater observations were made on tlie dav the log was moored (Day 1). T.\Bi,E 6. — \umber of fish and species recruited to various ob- jects moored near the Costa Rican coast in 1963 Object Time elapsed after establishment Fish Species Distance from shore Season Truck inner tube'. .. Balsa log (.\) Balsa log (li) Truck inner tube 3 feet by 3 feet by X inch plywood... Truck inner tube 'rruck inner tube Truck iiuier tube 15 hours 5 days* 5 days! 6 days 20 days 20 days _ 28 days 46 days Number 203 19S 474 263 2 13 492 118 Number 6 7 6 8 1 4 5 Nautical miles 31 7 8 2 7 8 9 Spring Fall ' Do. Spring Fall Spring Do. Do. ' Attaelied to this inner tube were ten 9^-incij (19 mm.) manila lines 10 m. long. AH tubes had a volume of 0.286 m.' • Established at same time at same locality. high freciucncy in the fall collections, but neither was found under the two balsa logs moored in the fall. With these two exceptions, no difference existed in the species composition or in the size of the individu- als between populations of fishes associated with moored objects and those associated with drifting objects. BEHAVIOR DISTRIBUTION AND FRIGHT BEHAVIOR When disturbed, nearly all species swam toward the drifting object and maintained a position much closer to it than when undisturbed. The fishes showed tliis behavior when a school of porpoise, Sknelln qnif/nani, passed a log, when four ()orpoi.se, Tiu: Unpublished data. Bureau of Conunercial Fisheries. I.a JoUa. California. 26 U.S. FISH AND WILDLIFE SERVICK gests the association has a selective advantage. This behavior was not, however, a mechanism for the association, because fishes remained near the object in the absence of predators. Use of the object as a spawning substrate would apply only to adult fishes. Adults, however, repre- sented only a small portion of the total individuals present. Adults of two species, Euthynnus lineatus and Canthidermis maculatus, were in reproductive condition. Euthynnus does not have attached eggs, however (Calkins and Klawe, 1963), and no eggs were found on any of the floating materials. The cleaning-station hypothesis suggested by Gooding and Magnuson (see footnote 1 ) could apply only to some of the fishes. Canthidermis maculatus alone showed behavior that could possibly be inter- preted as cleaning. This species was taken in only 14 of the 70 collections. Except in the collection made farthest from shore, where Canthidermis and Balistes polylepis were the only species present, no differences in the species composition were evident in the collections that contained Canthidermis. Thus, if Canthidermis regularly consumes parasites of other fishes the activity does not appear to influence the presence of these fishes. Artificial reefs established in sandy locations rapidly attract groups of fishes that would not other- wise inhabit these areas (Carlisle et al., 1964). The artificial reef provides the habitat requirements for certain fishes in an otlierwise unsuitable area. Simi- larly, a drifting object may provide a suitable habitat for inshore fishes that have pelagic juvenile stages or that have become displaced from shore. This explanation seems to be plausible for the presence of Abudefduf saxatilis, Blenniolus breripinnis, Batistes polylepis, and other species. The very restricted size range of Chromts atrilohata and Pseud upcneus grandisquamis suggests that these species are avail- able for association during a limited period. Many of the Pseudupencus were near the size at which metamorphosis takes place. Approaching metamor- phosis was indicated by the slight color changes in some of the individuals and by pronounced changes in coloration after the fish were kept in a shipboard aquarium 34 hours. Possibly large premetamorphic juveniles maj' be attracted to objects because of changes associated with metamorphosis; for these fish the object may represent an inshore or non- pelagic habitat. In summary, we found little evidence to support the mechanisms postulated by other authors. We have suggested two mechanisms: (1) fishes are at- tracted to drifting materials because the object functions as a schooling companion, and (2) for species not adapted to a pelagic life and others under- going a change from a pelagic to other modes of existence, drifting materials may function as a sub- stitute for a reef or other substrate. In both situa- tions the object may have the same function, that is, provide a visual stimulus in an optical void. The occurrence of juvenile fishes beneath flotsam was much more frequent than that of adults. That some species, as Chromis atrilohata, Pseudupeneus grandisquamis, and Agonostomus monticola, are pela- gic only as juveniles can explain the absence of the adults. Of the species that are pelagic as juveniles and adults, the juveniles were in tlie vicinity of an object for longer periods and remained closer to the object than did the adults. Owing to their larger size and faster swimming speed, adults are probably less susceptible to predation. Thus, for adult fishes the selective advantage of maintaining a close associ- ation with a drifting object may be small. It is also possible that development is accompanied by an increase in the specificity of the responses of school- ing fishes to other schooling companions. The val- ence of flotsam as a schooling object would then be lowered and intermittent association with drifting objects might be expected. ACKNOWLEDGMENTS Richard R. Whitney gave editorial and statistical assistance. Frank J. Hester navigated our vessel during the fall cruise, assisted in making the collec- tions, and took the underwater motion pictures. Donald C. Aasted de\eloped the procedure for han- dling the small purse seine and assisted in making the collections. Frederick H. Berry, Bureau of Com- mercial Fisheries Biological Laboratory, Brunswick, Ga., and Richard H. Rcsenblatt, Scripps Institution of Oceanography, La JoUa, Calif, helped identify the fishes. Rosenblatt also read the manuscript and offered suggestions. LITERATURE CITED Alverson, Dayton L., and Norman J. Wilimov.sky. 196.3. Catching fish— 1990 style. Fish. News 19(10): 1, 5, 9, and 12. Atz, Jambs W. 1953. Orientation in schooling fishes, hi Proc. Conf. Orientation Anim., Off. Nav. Res., W.ashington, D.C., Section 2: 103-130. FLOTSAM IN OFFSHORE WATERS 27 Besednov, L. X. 1960. Ni'kotovyo ilaiiuyo pu ikhtiofauno tikhookfiiu- skogo plnvaika (Some data on the ichthyofauna of Pacific Ocean flotsam^ Tr. Inst. Okoanol. (.\kad. Nauk SSSK). 41: 192-197. (.Transl. by W. G. Van Campcn, May 19GS, 7 pp., avail. HCF, Hiol. Lab.. Honolulu, or from OTS, SL.V.) Cvi.KiNS, Thomas P., and WiToi.n L. Ki.awe. 19G3. Synopsis of biological data on black skipjack Eulhyiinus limatiis Kishinouyc 1920. In I'roc. World Sci. Meeting Biol. Tunas Related Spec, V\0 Fish. Rep. 6, (2) Spec. Synopsis 2; 130-1-16. CARLISLE, John, Ju., Charles Ti'rner, and Eari. Ebert. 19G4. .\rtificial habitat in the marine environment. Calif. Kish and Came, Visli Hull. 121: 1-93. Dales, R. Phillics. 1957. Interrelations of organisms. .\. Commensalism. In Treatise on Marine KcoIokv und Paleoecology, ^'ol. I, Ecology. Geol. Soc. .\n\. Mem. 1(>7V 391-412. Ebelin'Q, Alfred W. 1961. Miigil galapagcnsis, a new mullet from the Galii- pagos Islands, with notes on related species and a key to the Mut;ilidae of the eastern Pacific. Copeia 1961 (3): 2*)5 30,">. F.voER, E. W., and J. .V. McC.owan. 1963. Zooplankton .species groups in the north Pacific. Science 1 40(3566 V. 453-460. Galea, J. .\. 1961. The "Kannizzati" fishery. Proc. Gen. Fish. Counc. Mediter., Tech. Pap. 7: 85-91. Hardenbero. J. D. F. 1950. Development of pelagic fisheries. Proc. Indo- Pac. Fish. Counc. Singapore. 1949, No. 1, Sec. TV: 13S-143. Klavber, L. M. 1935. The feeding habits of a -sea snake. Copeia 1935 (21: 182. Klawe, W. L. 1964. Food of the black-and-yellow sea snake, Pelamis platiirus from Ecuadorian coastal waters. Copeia 1904 (4): 712-713. KOJIMA, Shvmpel 1955. .\ study of dorado fishing condition in the western part of the Japan Sea — I. Hull. Jap. Soc Sci. Fish. 20 02): 1044-1049. (.Translation available. Bureau of Commercial Fisheries, Honolulu, Hawaii.) 1956. Fishing for dolphins in the western part of the Japan Sea — II. Why do the fish take shelter under floating materials? Bull. Jap. Soc. Sci. Fish. 21(10): 1049-1052. (Translation available. Bureau of Com- mercial Fisheries. Honolulu, Hawaii.) 1960a. Fishing for dolphins in the western part of the Japan Sea — ^'. Species of fishes attracted to bamboo Rifts. Bull. Jap. Soc. Sci. Fi.^h. 26(4): 379-3S2. tTranslation available. Bureau of Commercial Fish- eries. Honolulu, Hawaii.) UHiOb. Fishing for dolphins in the western part of the Japan Sea — \'I. Behaviours of fish gathering around bamlHio rafts. Bull. Jap. Soc. Sci. Fish. 26(4) : 3S3- 388. (Translation available. Bureau of Commercial Fisheries, Honolulu, Hawaii.) 1961. Studies on fishing conditions of dolphin, Cori/- l>hacnc liippurus L., in the western region of the Sea of Japan — III. On food contents of the dolphin. Bull. Jap. Soc. Sci. Fish. 27(7): 625-629. (Translation available, Bureau of Commercial F'isheries, Honolulu, Hawaii.) KKK.ISA, Richard J. 1960. The eastern, tropical Pacific fishes of the genus lilcnniolus, including a new island endemic. Copeia 1960(4): 322-336. LiMBAVHH, Conrad. 1955. Fish life in the kclplicils and the cITccts of kelp harvesting. I'niv. Calif. Inst. Mar. Resources. Ref. 55-9: 1-5S. MANst'KTi, Romeo. 1963. Symbiotic behavior between small fishes and jellyfi.-^hes, with new data on that between the stroma- teid, Pcprihts alepiilotns, and the sc>phomedusa, Chnjsaora quinquecirrha. Copeia 1963(1): 40-80. McNeely, Richard L. 1961. The purse seine revolution in tuna fishing. Pac Fish. 59(7): 27-58. Murray, John, and Johan Hjort. 1912. The depths of the ocean: a general account of the modern science of oceanography based largely on the scientific researches on the Norwegian steamer Michael Sars in the North .Atlantic. MacMillan i<; Co. Ltd., London, x.\-|-821 pp. Orange, Craig J., Milnkr B. Schakfkr. and Fred M. Larmie. 1957. Schooling habits of yellowfin tuna (Xeothunnus macropteru!^) anil skipjack {Knisuwonus pelnniis) in the eastern Pacific Ocean, as indicated by purse seine catch records, 1946-1955. Inter-.Vmer. Trop. Tuna Comm., Bull. 2(3): 8-126. Peterson, Clifford L. 1960. The physical oceanography of the Gulf of Nicoya, Costa Rica, a tropical estuary. Int('r-.\mer. Trop. Tuna Comm.. Bull. 4(4): 137-216. SOEMARTO. 1960. Fish behaviour with special reference to pelagic shoaling species: Lajang, (Dccapterus spp.). Proc. Indo-Pac Fish. Counc, Colombo, Ceylon, 1958, No. 8, Sec. 111:89-93. I'cHiDA, Keitaro, and Ygichi Shojima. 1958. Studies on the larvae and juveniles of fishes ac- companying floating algae — I. Research in the vicinity of Tsuyazaki, during Mar., 1957— Mar., 1958. Bull. Jap. Soc. Sci. I'ish. 24(6 and 7): 411-415. (Transla- tion available. Bureau of Commercial I'isheries, Hono- lulu. Hawaii.) I'da, Mkhitaka. 1933. Types of skipjack schools and their fishing quali- ties. Bull. Jap. Soc Sci. Fish. 2(3): 107-111. (Trans- lation. Van Campen, W. G., 1952. Five Japanese papers on skipjack. I'.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. a3. iii + 78 pp.1 2S U.S. FISH .\XD WILDLIFK SERVICE L'da, Michitaka, and Jiro Tsukushi. Yabe, Hiroshi, and Tokumi Mori. 1934. Local variations in the composition of various 1950. An observation on the habit of bonito, Katau- shoals of "katuwo," Eulhynnus vagans (Lesson), in wonus vagann and yellowfin, Neothunntis macropterua, several sea-districts of Japan. Bull. Jap. Soc. Sci. school under the drifting timber on the surface of ocean. Fish. 3(4): 196-202. (Translation, Van Campen, Bull. Jap. Soc. Sci. Fish. 16 (2): 35-39. (Translation W. G., 1952. Five Japanese papers on skipjack. available. Bureau of Commercial Fisheries, Honolulu, U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 83, iii + Hawaii.) 78 pp.) FLOTSAM IN OFFSHORE WATERS ^^ ANALOG COMPUTER MODELS OF FISH POPULATIONS By Ralph P. Silliman, Fishery Biologist (Research) Bureau of Commercial Fisheries Biological Laboratory, Seattle, Wash. 98102 ABSTRACT A modern analog computer, together with an X-Y plotter,' provides a means of constructing useful models of commercially exploited fish populations. The model combined conventional exponential fishing and natural mortality rates with a Gompertz curve of growth in weight. By combination of these rates and curve into a single differential equation, survival curves for suc- cessive year classes of fish were generated by the com- puter and plotted by the plotter. Weights for all year classes present in each season were summed graphi- cally. Recruitment was determined from a stock- recruitment curve, set into a function generator of the computer. Yields for each season were calculated by multiplying stock weight by rate of exploitation and were compared with actual yields to test the validity of the models. The technique was demonstrated by use of empirical and hypothetical data for the Atlantic cod (Gadus morhua). It is generally applicable to fisheries for which good measures of total catch, growth rate, natural and fishing mortality rates, and stock-recruit- ment relation are available. It is important to use reasonable values of biological variables in constructing the models because it was not possible to demonstrate beyond doubt that the set of parameters for any given model provides a unique solution. Compared with other techniques, the analog-graphic approach offers low cost of equipment, moderate com- putation speed, ready accessibility of equipment, and good visibility of results during computation. It is limited in accuracy (two or three digits) and in scaling requirements. Fishery biologists have devoted much effort in determining changes in fish populations as thej' respond to varying degrees of fishing intensity. Because stocks usually cannot be observed and measured directly, it has been necessary to use data of catch and fishing effort and biological data on relatively small samples of the stocks. These rec- ords, plus associated data on the environment, have composed most of the working materials of the fish- ery biologist. Limited to such materials, the fishery biologist has been forced to proceed largely by infer- ence. There has been no alternative. As work be- came quantitative, inference came to mean statistical or biometric inference, or a combination of both. One method of quantitative inference is that of simulation or modeling. Using the best empirical Note. — .Approved for publication Nov. 23, 1965. ' Trade names referred to in this publication do not imply endorsement of commercial products. data and biological judgment available, the biologist erects hypotheses concerning the additive and sub- tractive processes affecting the stocks. The former include recruitment, growth, and immigration; the latter, fishing mortality, natural mortality, and emi- gration. To test the validity of the hypotheses, characteristics of the models based on the hypotheses can be compared with what is knowTi of the real populations. Population models are used for the same purpose as models in ship or power-dam design: experiments are easier, quicker, and cheaper with the model than with the full-scale object. Any tool that will help the biologist in these processes should be valuable. This report describes such a tool in the form of an analog computer technique. Application of the analog computer as described herein has not, to the best of my knowledge, been made previously. Originality is not claimed, how- ever, for the technique of simulation in general or in FISHERY bulletin: VOLUME 66, NO. 1 31 the field of fisheries. Familiar applications outside the field of biology include aircraft and spacecraft design, ballistic studies for militarj- services, and management of systems such as power-generating pools and nuclear reactors. Among biologists, the ecologists especially lia\e made use of simulation. Notable examples include the simulation of ecologi- cal systems by tligital computer (Clarfinkel ami Sack, 1964) and by analog devices (Odum, 1960 and Mar- galef, 1962). In fisheries there have been three general types of simulation: analytical, digital, and analog. Analytical simulation i.s the oldest and best known. It is so well known that an extensive review is sui)er- fluous here. In it the biologist analytically deter- mines the nature of the mathematical relations controlling population structure and jjopulation response to exploitation. He then constructs mathe- matical models which predict what will occur under certain assumed conditions of the fishery and the environment. Most of the classical contributions to fishery dynamics are of this type. They include tlie work of such authors as Baranov (1918), Rus.sell (,1931), Thompson and Bell (1934), Graham (1935), Schaefer (1954), Beverton and Holt (1957), and Ricker (1958). During the history of the analytical techni(iue the complexitj- and sophistication of the formulations have generally increased; the work of Beverton and Holt unc[ucstionably represents the highest development in this respect to date. Some of the above authors have also used digital simulation. Thompson and Bell (1934) arithmeti- calh- constructed tables to simulate the catch and catch per unit of effort of the Pacific halibut under constant recruitment and certain assumed rates of growth and mortalitj*. They demonstrated a remarkable correspondence between the values pre- dicted by this method and the observed values for certain periods and areas. Ricker (1958) made somewhat similar calculations based on analytical functions and introduced the additional concept of relation between stock size and rate of recruitment. He expressed catches in terms of "equilibrium yields" which would be obtained when the additive processes affecting the populations just balanced the subtractive ones. An outstanding example of digital simulation of a fish stock is found in Larkin and Hourston (1964). To my knowledge, onlj' Doi (1957, 1962) has pro- jjosed api)lication of the analog computer to fishery dj'namics. He set forth mathematical formulations and computer block diagrams and made applications to Japanese fisheries. His formulas are similar to those used here, in general, but differ in detail. He also adapted the Volterra equations to analog treat- ment of predatory and competitive relations among lish populations. With some notable exceptions (Ricker, 1958; Larkin and Hourston, 1964; Larkin and Ricker, 19()4; International North Pacific Fislieries Com- mission, 1962), fishery-simulation attempts to date have dealt with eciuilibrium population conditions. This restriction is understandable in view of the mathematical complexities introduced by varying rates. Such an approach does, however, lead to models which are somewhat unreal compared with their counterparts in nature. For instance the number of recruits annually entering the stocks varies widely. This problem has been treated by considering recruitment constant for short periods (apparently close to the truth for some Pacific hali- but stocks during various 8- to lO-j'car periods be- tween 1918 and 1933) or circumvented by expressing results in terms of "yield per recruit." Likewise, fishing mortality varies with fishing intensity. Any realistic simulation of fished populations over con- siderable periods reciuires .^^ome provision for changes in recruitment and other vital rates, as can be accom- plished on the analog computer. The importance of this problem was recognized by Schaefer and Beverton (1963) who wrote: ". . . . the characteristics of the recruitment to marine fish populations — its degree of fluctuation, its relation to stock size and the influence on it of changing environ- mental conditions — are the key to the interpretation and prediction of the long-term dynamics of a fi.sliery . . ." antl : ".\ctual fisheries are, however, .seldom in steady states ..." The remainder of this report is devoted to a description of the plan of attack (Piatt, 1964 — strong inference) used in the analog computer approach. Briefly outlined, this plan is as follows: 1. Formulation of vital rates in a manner suit- able for analog solution, thus .setting up a hj^pothesis. 2. Simulation of populations and yields over the pcM'iod for which observational data are available. 3. Comparison of simulated and observed yields, testing the hypothesis. 32 U.S. FISH .\ND WILDLIFE SERVICE 4. Acceptance or rejection of the hypothesis. If the agreement of simulated with observed jnelds is good, the hypothesis is accepted. If it is rejected, a new hypothesis is erected by adjustment of parameters in the analog model, and steps 2 to 4 are repeated until satisfactorj' fit of simulated to actual yields is either at- tained or found unattainable. In practice, the process was never repeated more than a few times, since improvement fell off rapidlj'. Also, indefinite repetition would be out of keeping with the scientific method. BASIC FORMULATIONS For the initial trials of the analog technique, I adopted what seemed the simplest useful model of a fish population. This model includes rate of growth, rates of fishing and natural mortality, and a recruitment-stock relation. It does not take ac- count of immigration, emigration, or environmental effects. Symbols used have been adapted from Holt, Gulland, Taylor and Kurita (1959) in further- ance of their admirable attempt to secure uniformity in the terminology of fishery dynamics. Definitions are as follows: A', R f F 1 M Z t tr tc P, P,o W, U'oj K'r E = Xumber of recruits surviving at time /. = Initial number of recruits to fishable stock for a single year class. = Fishing effort. = Instantaneous rate of fishing mortality. = F f. = Instantaneous rate of natural mortality. = F + M. = Age of fish in years. = .\ge of fish at recruitment to fishable stock. = .\ge of fish when first vulnerable to capture by gear in use. = Weight of all fish of a given year class surviving at time /. = Weight of all fish present at beginning of season. = Weight of individual fish at time (. = Upper asymptotic limit of Wt. = Weight of individual recruit at time tr. = Estimated yield of fishable stock in weight, per year. = .\ctual yield of fishable stock in weight, per year, from official statistics. = Rate of e.xploitation, = ^ ^^ f 1 - «-<'' + '^'j = Subscript referring to individual year classes. Ifw = Initial weight of recruits to fishable stock for a single year class. 0,g = Constants of Gompertz growth curve. Because interest in this study is centered on the commercial catch, the model is limited to the fishable sizes and ages of fish. For a year class of fish passing through the fishable stock, numbers of fish surviving may be expressed according to the declining expo- nential formula, as set forth in Beverton and Holt (1957): iV, = /?p-(F+-w) c-g (1) In addition to the above, the following symbols have been adopted for the formulations here: To take account of the growth of individual fish, and to obtain yields in weight for comparison with commercial catches, it is necessary to introduce a formula for weight-at-age. Beverton and Holt em- ployed the von Bertalanffy equation for length-at- age, converting to weight-at-age by means of a cubic length-weight relation. Use of the cubic relation has been shown to lead to considerable error when the real relation between length and weight involves a power of length other than 3 (Paulik and Gales, 1964). Although this difficulty can be overcome by use of the Incomplete Beta Function (Wilimovsky and Wicklund, 1963) in the yield equation, the formulation still is not well adapted to analog computation. As an alternative to the von Bei'talanffy equation, I investigated the characteristics of the equation developed by Benjamin Gompertz. He applied it as an expression of human mortalitj' rates, but various forms of it have since been used as growth curves for both length and weight of animals. Its applicability in this respect was thoroughly dis- cussed b}' Winsor (1932). In the form used by Weymouth and Mc^NIillin (1931), it is seen to be an exponential curve in which the slope declines expo- nentially. They point out that the relative (as oppo.sed to absolute) growth of an animal declines with age because of an increasing proportion of inac- tive material, and other causes. The declining slope of the Gompertz curve is in accord with this phe- nomenon. Also, it provided a good fit to the empirical data of weight-at-age for several fishes. Beverton and Holt (1957) rejected the Gompertz curve on the basis that it deals with growth as an additive process only, ignoring the breakdo\\'n of protoplasm. The net effect, however, of anabolism and catabolism may well be the kind of declining relative growth described by the Gompertz curve. This curve thus did not appear to be rejected on ANALOG COMPUTER MODELS OF FISH POPULATIOXS 33 biological grounds, and since it was practical for analog computation, I employed it. It may be noted that growth rates of fisii typically decline throughout life and can be represented most simply by a positive instantaneous rate which de- creases exponentially with time, as in the Gompertz curve. The relations can be expressed in the follow- ing formulas (where G represents the initial exponen- tial growth and g governs the exponential rate of decline) : (2a) , = Wre''e-°'-''"-''^ iv, = u\e G-ae-«"-'T) (2b) This formula can be combined with formula (1) to express total weight of survivors at any time. It is of interest, also, that it has an upper asymptote u'„ similar to the "LJ' of the von Bertalanffy equation. Thus: as I > cc, e-""-''' >0 and e-o'-'"-''' 1.00 ca from ecjuation so that w, > wve," as t (2a) above. The limiting value of this expression is w^. If extended from w, = to !(',^m'„, the Gompertz curve has a point of inflection lacking in the von Bertalanffy curve. This inflection is found in age- weight curves of many fishes. The total weight of survivors from a single year class may be expressed ; P, = w',^V, (3) Substituting in (3) for w, and .V, their equivalents in (1) and (2b): Because I dealt with weight rather than numlier of recruits, I set R^ = RlVr and obtained as my working equation: For convenience, clarity, and ready comparability with other work, I have dealt with all relationships so far in algebraic form. Although the computer requires differential elc even though the equations describing the graph are unknown. The analog computer is primarily a device for solving differential ecjuations with time as the inde- pendent variable. It, therefore, becomes evident that if a biological process can be expressed as a dif- ferential equation, the etjuation can be mechanized by interconnecting analog computer components corresponding to the mathematical operations. I'liaiiE 1. — .\iiulog cdinputer and plotter. U.S. FISII ASD WILDLIFE SERVICE Analog computer programming is based essentially on the electrical principle of the feedback loop. For a simple illustration, let us return to the declining exponential curve, as expressed in equation (1). This expression may be further simplified by setting F+il/ = Z, as in international notation and assuming that our measurement of time begins at t, so that /r = and {t — tr)=t. Expression (1) then becomes: Remembering that the analog computer requires differential expressions, we differentiate the above to find: clN, dl -RZe-^' As in digital computation, we start with a "block diagram" (ordinarily this step would be omitted in sucli a simple circuit, but it is shown here to illustrate the process). A few symbols are needed (arrows indicate direction of information flow) : OPERATION PERFORMED BY COMPUTER SYMBOL EQUATION Integrolion with respect to time Inversion (Ctionqe of sign) Multiplication by o constant Summation "i" :|f^ y =y^«, dt ^ i c = y] y = Ax| y = X + X +x Using the first three of these symbols, we can now construct a block diagram for our differential expres- sion. We start by assuming that a voltage propor- tional to dN ,/dt exists: dN, dt I.e. I 'dt Nt + i.e. Considering only the rate at which ^V, changes, and disregarding its absolute value, we can omit the constant of integration (this constant will be added in the circuit diagram) : dN, "dT 'dt -^ Nt = Re' zt But we know that 3^=— RZe~^', so we simply dt assemble the derivative: dN| dt = -RZe Zt Nt = Re' ■Zt Vdt -ZRe-^^ Z The next step is to construct a circuit diagram showing the actual computing elements and taking account of "initial conditions" and any sign changes that may occur. For this step we use additional symbols : COMPUTING ELEMENT Integrating network witti Summing Junction Potentiometer SYMBOL ^ BLOCK DIAGRAM EQUIVALENT Summing amplifier Still disregarding i. c, we have: dNi dt = -RZe -zt -Nt=-Re' ■zt ■ZRe -zt Finally we must supply the "initial condition" voltage, i.e., which is equal to the constant of integra- tion. We noted above, in the definition of the symbol I.e. 1 A A\ ? that y= fo'.Tidt + i.e. and i.c.=y],=o. In our equation "y" is replaced byA'^,. Since we set /r = 0, by definition N,],^ = R = i.e. The completed diagram, assuming a computer ANALOG COMPUTER MODELS OF FISH POPULATIONS 35 reference voltage supply of —10 volts, becomes: -10 R/IO O z In the operation of the integrating network, i.e. only determines the value of N, (condenser charge) at the beginning of the computation. It is cut off the instant computation begins, and current flow is then confined to the feedback loop. That is why i.e. does not appear as a variable in the block dia- gram. The figure "1" by the integrator input refers to the input resistor. Operation of the integrator is such that the input is multiplied by a factor 100/7?f when /?, = input resistance in kilohms. With the knowledge now at hand we can proceed to set up a circuit in which Z is divided into its components F and M. The equation becomes .V, = /?fi~"^+"" and the diagrams are: BLOCK DIAGRAM CIRCUIT DIAGRAM -10 I f. r- -1 1 <—*— _ < F 4 — <— — IV E Study of the circuit u.sed here (fig. 2) reveals it to be essentially an elaboration of the basic feedback loop. A multiplier has been added to take care of the nonlinear Gompertz growth curve, and other elements have been added as described in the text. The "time-base" {t) is generated bj^ a simple inte- grator output. Biologists interested in further in- formation on analog computer programming will find an excellent brief introduction in Strong and Hannauer (1902) or a more extended treatment in Ashley (19(53). The plotter has a pen actuated by two servo- motors. One moves it along the "X" axis and the other along the "Y" axis in proportion to an input voltage supplied by the computer. Thus the pen moves to any point A', Y in a system of rectangular coordinates on the plotting surface, corresponding to input voltages T'^ and V^. Since the computer in- tegrates with respect to time, a voltage directly proportional to time is usually fed into "A'." The input for "!'" can be taken from any point on tlie computer circuit to plot the variable(s) desired against time. As an alternative method of output display, an oscilloscope can be used. This combination requires that the computer be modified for "repetitive opera- tion." Problem solutions are repeated 10 to 100 times per second, so that they appear as curves on the oscilloscope screen. For a permanent record the screen can be photographed as mentioned by Doi (1902). This oscilloscope display is particularly valuable for curve fitting, since points can be plotted on the face of the tube. In this manner the effect of the potentiometer adjustment in improving the fit is instantly seen. The work described below was performed on a Pace TR-10 analog computer in conjunction with an EAI Variplotter 1 1 10. The TR-10 is one of the smallest general-purpose analog machines. Since both computer and plotter are fully transistorized, they are small and can be used conveniently atop a desk or small table. The set of units available in the computer as I used it was as follows: Unit Number Amplifier (u.-^ed with integrator, multiplier, etc.) 10 Coefficient potentiometer (ussed as above) 18 Null potentiometer (used to set coefficient potentiometers) 1 Integrator (used as above) 4 Multiplier (used as above) 1 Dioile function generator (use described later in text) 1 Comparator (use described later in text) 1 GENERATION OF SURVIVAL CURVES The differential eciuation needed for generating the survival curve of a given weight of recruits, R^, is obtained by differentiating expression (4): (5) 3G U.S. FISH .\N"D WILDLIFE SERVICE + 10 Figure 2. — Analog computer circuit for generation of P, and W. Standard symbols for computing elements, scaling excluded. By integrating this expression, the computer gen- erates P, as a function of {t—tr) for any given values of /?^, g, G, F, and M. The circuit diagram, which includes conventional symbols for the computing elements, is shown within the broken line enclosures of figure 2. The elements outside the broken lines are used in the combination of analog and graphical computation employed to sum the weights of year classes for each season and to determine subsequent recruitment. Their nature and use are described later. As examples of analog computation, growth and survival curves from the plotter are shown in figures 3 and 4. The data are hypothetical except for the growth constants, which have been derived from data for the California sardine. Data of growth in weight were obtained by com- bining a curve of growth in length with a weight- length relation. A table of length-at-age was given in Phillips (1948) and a weight-length relation in Clark (1928). Fitting of the Gompertz relation to these data (fig. 3) was readily accomplished by suc- cessive trials, with appropriate adjustment of the potentiometers for G and g. The fitted curve fol- lowed expression (2b), with ?iv = 93 g., G = 0.825, g = 0.44.5, and ^ = 2 years. Starting with a hypothetical 1,000 fish, 7?,^ = 93 kg. when uv = 93 g. The upper curve (fig. 4) shows how, with no fishing mortality and low natural mortality, P, may increase for a year or two before mortality overcomes growth. In the two lower curves the effect of adding a substantial fishing mortality may be seen. Application of an increase in fishing mor- tality at t — tr = 2 resulted in the lower branched curve. This change is readily made on the analog computer by placing the machine in the "hold" mode. The potentiometer for "F" is then reset, the machine returned to "operate" mode and the com- putation resumed. The ability to change quickly the vital rates during a computation is one of the advantages of the analog machine. It may be done even more conveniently by presetting a number of ANALOG COMPUTER MODELS OF FISH POPULATIONS 37 225 200 175 ^150 < q: 125 100 75 50 25 ' 10 12 14 AGE ( YEARS 16 18 20 22 Figure 3.— The Gompertz curve fitted to weight-at-age data for the Pacific sardine following expression (2b) in text; u\ = 93 g., G = 0.825, g = 0.445, and (r = 2 years. 120 M 100 5 / V = .00M=.20 1 < '^ on \ \ ^ O \ ' rF = .3\ ! -J \t tlVI = .20\ \ 40 k \ •*^ 20 F = 501 Ix •^ M = .20i ^^ ::ii::~ ' 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Figure 4. — Computer curves of P, as a function of (t-tr)', growth data from the California sardine, according to expre.ssion (4) in text, \alues of constants are: li = 1,000 fish, »v = 93 g., R^ = 93 kg., O = 0.825, g = 0.445. 38 U.S. FISH AND WILDLIFE SERVICE potentiometers at needed values of "F" and shifting from one to tlie other. PROCEDURE OF SIMULATION Simulation of populations and yields by a com- bination analog-graphic approach required develop- ment of a standardized procedure. A chart was prepared for the plotter, with appropriately scaled coordinates for time (X-axis) and stock (Y-axis). Vertical lines on this chart marked the points for changes in mortality rates or recruitment relation. Because the commercial stock in any fishing season is made vip of survivors from individual year classes of various ages, that stock cannot be started "full blown," with all year classes present. I therefore started with a stock size such that the rate of exploi- tation (E) estimated to be in effect would produce a catch [Yir) eciual to the real catch for the first fishing- season, or the mean catch for the first two fishing- seasons of the study period. This stock was then built up by starting year classes at times 1, 2, 3, . . . n years before the beginning of the period ; n repre- sents the number of years for a year class to pass through the fishery. During this "prestuily" ])eriod, the mortality rates in effect at the beginning of the study period were assumed to be in effect. Each rth curve begins at the value of R,r (assumed the same for each year class) required to produce the 1 = n specified initial value of total stock Pto=^^P n- i=l This value of 7?„, can be quickly determined by a few computer trials. P, for each year class is gen- erated until it declines to a small arbitrary value close to zero, but only the portion extending into the study period is plotted. The initial and subsequent values of P,o are cal- culated by graphically summing the heights P„ of the 7th survival curves for each fishing^ season, by means of a pair of dividers. Once the initial value of Pio has been obtained, the calculation is self- sustaining. From the recruitment curves, values of Ru^i corresponding to P,o t, years before are obtained. Ry,i is generated as a function of P ,„ in the Diode Function Generator (fig. 2) by the simple device of setting the P,o potentiometer so that the plotter "F value" corresponds to the P,„ value for the particular year in question. 7?^( is then plotted by attaching the Y input to the Ry, point in the computer circuit. Any empirical or theoretical curve relating recruit- ment to spawning stock can be set into the Diode Function Generator, which, with an input Xi, pro- duces an output in the form of a curve //=/(xi), composed of 10 straight segments. SiU'vival curves as described above can be gen- erated for each year class entering the fishery during the study period. As in the case of the initial season just described, heights P„ of the individual survival curves for each year are summed graphically, and a mark made representing the total commercial stock, P,o=VP,,. The "P,„ set" potentiometer is ad- 1=0 justed to bring the "I'-value" of the plotter into conformance with the total stock value P^. The catch is calculated by setting potentiometer "E" (Fig. 2) at the value E = - jl_e-(F+.w]_ The catch or yield value proceeds from the simple rela- tion Yu- = EP,o. It is plotted by attaching the "Y input" of the plotter at the point F„ in the computer. After plotting of Y^ the process for P„j (above) is repeated, and the cycle recommenced. In outline, then, the process of simulating popula- tion and yield is as follows: 1. Set the initial value of Y,^ at the size of the actual catch for the initial year or two of the study period. i = n 2. Determine initial P,„— Z^Pn from the rela- i=o tion P,„ = \\/E. 3. By computer trial, find value R„i such that i = n ^ P„ = initial P,„. 1=0 4. Generate n curves of Ph, where n is the num- ber of years required for P,i to decline from Rwi to an arbitrary small value near zero. Start at 1, 2, 3, ... n years before the begin- ning of the study period. 5. By Diode Function Generator evaluate Rwt for each season from P,„ for season U years before. Generate curves P,j starting at Pit = Rwi for each fishing season. 6. For each fishing season, graphically deter- i = n mine P,„=%Pti. Calculate \\ = EP,„. ANALOG COMPUTER MODELS OF FISH POPULATIONS 39 7. Repeat cycle to end of study period, starting each cj'cle with step No. 5. EXAMPLE OF APPLICATION Since hypothetical data are seldom satisfactory to demonstrate the application of a technique, the following example of application to the fishery for Atlantic cod (Cad us morhua) is included. It has been used to achieve concreteness, not to make new discoveries about the cod. Catch data were summed for International Commission for the Northwest Atlantic Fisheries Divisions 5Y and 5Z, and the following parameters were assemblcfl for analog computation: 1. The central value of F = 0.35 used in Bever- ton and Hodder (19G2) was assumed to be the average {F) for the entire study period 1932-1958. From this figure, values were calculated for eight periods, from the relation F= qf, where/ was value of fishing effort from Beverton and Hodder and q = F/f: Moan pfTort lhous:m Figure 7. — Effect of varying F or jV/ in simulation trials with cod. The value r is the coefficient of correlation between calculated catches (?„) and actual catches (V'„). Since r is affected only by Z, and not the ratio of its com- ponents F and M, it has only one value for each pair of combinations. The fraction ?«,/)'«, represents the ratio of the mean calculated catch (vio) to the mean actual catch (Tu,). Vertical line of dashes indicates combination of values used in third cod simulation trial. ANALOG COMPUTER MODELS OF FISH POPULATIONS 43 ment. It is, of course, possible to perform work on digital machines under contract or rental arrange- ments at no initial cost. This arrangement is also possible for analog machines. 2. Time required for computation. — For the total operation as outlined above, limited tests indicated analog-graphic computation to be about four times as fast as desk calculation. With digital computers, the calculation time is a matter of minutes. If time required for preparation of data for computer cal- culation, programming the computer, and exchange of data with the computer center are considered, however, total time may well ajiproach that for the analog method. 3. Visihilitij of ivork during computation. — In the analog-graphic method, stock size, recruitment rate, and yield are all visible in graphic form as the computation proceeds. This advantage is impor- tant to the biologist, since it quickly reveals absurd results, or permits him to end a computation that is leading away from reality. Work is invisible during digital computation, and the final "readout" is usually in the form of a table that may have to be plotted for study. 4. Scale adjustment. — Quantities generated within an analog computer must be kept within the voltage limits of the machine. This limitation leads to a considerable amount of "fussing" to achieve proper scaling of the variables. This problem is only minor in digital calculation and therefore represents a comparative disadvantage for the analog computer. Fortunately, once scaling has been adopted for a given formula, it can usually be used with only one or two changes when shifting to a new set of empirical data for the same formula. "). Accuracy of results. — Because of the nature of components in an analog computer, the final results are usually accurate only to two or three significant digits. At the present stage of development of fishery science, the empirical data available are not such as to justify carrying more fligits. In fact, two-digit accuracy in fishery predictions would be considered more than satisfactory by most fishery administrators. Thus, the accuracy limitations of the analog machine as compared with digital com- putation do not at present represent a serious disadvantage. 6. Summary comparison of methods. — From the above brief listing, the analog technique is seen to have advantages in comparatively low initial cost of equipment, moderately rapid computation rate, and visibility of results. It has limitations in accu- racy and in scaling requirements and is slower than a digital comijuter. Decision as to which techni(|ue to use must depend on the situation of the individ- ual investigator. Factors liearing on the decision include the salaries of persons doing various parts of the work, the accessibility of the research station to a digital computer, anrl the types of emiiirical data availalile. UTILITY OF THE TECHNIQUE In this report I have described what may be a useful working tool for the fishery biologist. The example of application given demonstrated the types of l)asic data neeiled and the way in which they could be adjusted to improve "goodness of fit" of calculated to actual catches. As with any tech- niciue, its utility can be asses.sed only by those making use of it. Where extensive biological data are available, simulation is valuable in determining the effects of interaction among the varying mortality, growth, and recruitment rates. The accuracy with which actual catches can be rei)roduced should serve as a check on the validity of the sampling, analysis, and interpretation involved in the derivation of popvilation parameters. SUMMARY 1. The objective of this study was to develop :m analog-computer simulation techni(iue for niodoling exploited fish jiopulations. 2. The mathematical formula for survival of a year class expressed the effect of fishing and natural mortality rates and incorporated a Oompertz curve of growth. 3. Survival curves for successive year classes were generated on an analog computer through use of the differential form of the survival formula. A combined analog-graphic technique summed the weights of survivors in each season to give the weight of the fishable stock. 4. Yield was calculated by applying the rate of exploitation to the fishable stock. 5. Properly lagged recruitment was determined from the stock weight tluough a stock-recruitment curve. 6. Mechanics of the technique were demonstrated by application to the Atlantic cod. 7. This technique may be applied to any fishery for which good measures or estimates of catch, 44 U.S. FISH .\ND WTLDUKE SKUVICE growth rate, fishing and natural mortality, and stock-rpcriiitment relation can be obtained. 8. The problem of uni([ueness was studied from simulations in which F and .V were varied over a range of values. The failure of results to prove uniqueness brings out the importance of using reasonable values of parameters. 9. As compared with other techniques, the analog- graphic approach described here offers low initial cost of equipment, moderate computation speed, ready accessibility of equipment, and good visibility of results during computation. It has limitations in accuracy (two or three digits) and in requirements for scaling variables. ACKNOWLEDGMENTS Suggestions from my se\eral re\ie\vers were freely followed in revising the text. Reviews were fur- nished by Vaughn C. Anthony, Raymond J. H. Beverton, Anthony C. Burd, Paul H. Eschmeyer, Re>Tiold A. Fredin. David J. Garrod, John A. Gulland, Richard C. Hennemuth, Ralph Hile, George Hirschhorn, Peter A. Larkin, Garth I. Murphy, and William E. Ricker. Paul E. Huber, of Electronic Associates, Inc., kindly reviewed the sections on analog computation. .lohn L. McHugh provided guidance during final revision of the manuscript. LITERATURE CITED ASHI.KY, J. ROBEKT. lOCi:-!. Introduction to analog computation. John Wiley and Sons, Inc., New York, 294 pp. Bar.wov, Feodor I. 1918. K voprosu o biologicheskii osnovaniyakh rybnogo khozyaistva (On the question of the biological basis of fisheries). Moscow, Izv. Xauch. Is.sled. Ikhtiol. Inst., Izv. Otd. Rybovod. Xauch. -promysl. Issled. 1(1): 81- 128. In Russian. Beverton, R.wmond J. H. 1903. Maturation, growth and mortality of clupeid and engraulid stocks in relation to fishing. Cons. Perma. Int. Explor. Mer, Rapp. Proe.-Verb. Reun. 154; 44-67. Beverton, R.wmond J. H., and V. M. Hodder. 1962. Report on working group of scientists on fishery .assessment in relation to regulation problems. Int. Comm. Northwest Atl. Fish., .\nnu. Proc. 11 (suppl.): 1-81. Beverton, R.wmond J. H., and S. J. Holt. 1957. On the dynamics of exploited fish populations. Great Brit. Min. Agr. Fish. Food, Fish. Invest, ser. 2, 19: 1-533. 1959. A review of the lifespans and mortality rates of fish in nature, and their relation to growth and other physiological characteristics. Ciba Found. CoUoquia on Ageing, 5: 142-180. Little, Brown and Co., Bcston. BiGELow, Henry B., and Willi.\m C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53: viii + pp. 1-.577. Cl.^rk, Fr.\nces N. 1928. The weight-length relationship of the California sardine {Sardina cacrutea) at San Pedro. Calif. Div. Fish Game, Fish Bull. 12, 59 pp. Cl.\rk, Frances N., and John C. Marr. 1955. Popul.ation dynamics of the Pacific sardine. In Progress Report, 1 July 1953 to 31 March 1955: 11-48. Calif. Dep. Fish Game, Mar. Res. Comm., Calif. Coop. Oceanic Fish. Invest. Doi, Takeyuki. 1957. An analogue computer for analyzing the properties of marine resources and pretlicting catches. Joint Sci. Meet., Int. Comm. Northwest .\tl. Fish., Int. Counc. Explor. Sea, Food .\gr. Organ. UX, Lisbon, Doc. P. 12. 14 pp. 1962. The predator-prey and competitive relationships among fishes caught in waters adjacent to Japan. Bull. Tokai Regional Fish. Lab. 32: 49-118 (Fish. Res. Bd. Can., Transl. Ser. 461). Garfinkel, David, and Richard Sack. 1964. Digital computer simulation of an ecological system, based on a modified mass law. Ecology 45 (3): 502-507. Graham, Michael. 1935. Modern theory of exploiting a fishery, and appli- cations to the North Sea trawling. J. Cons. 10 (3) : 264-274. Holt, S. J., J. .A. Gulland, C. Taylor, and S. Kurita. 1959. A standard terminology and notation for fishery dynamics. J. Cons. 24(2): 239-242. International North Pacific Flsheries Commission. 1962. The exploitation, scientific investigation and man- agement of salmon (genus Oncorhynchiis) stocks on the Pacific Coast of the United States in relation to the abstention provisions of the Xorth Pacific Fi-sheries Convention. 6. Supplementary information on salmon stocks of the United States: Return escapement relationships: salmon. Int. N. Pac. Fish. Comm., Bull. 10: 67-72. Larkin, P. A., and A. S. Hourston. 1964. A model for simulation of the population biology of Pacific salmon. J. Fish. Res. Bd. Can. 21(5): 1245-1265. Larkin, P. A., and W. E. Ricker. 1964. Further information on sustained j-ields from fluctuating environments. J. Fish, Res. Bd. Can. 21(1): 1-7. Marc.alef, Ramon 1962. Simplified physical models of populations of organ- isms. Mem. Real Acad. Cien. Art., Bare. 34(5): 1-66. Odum, Howard T. 1960. Biological potential and analogue circuits for the ecosystem. Amer. Sci. 48(1): 1-8. ANALOG COMPUTER MODELS OF FISH POPUL.ATIONS 45 Paulik, Geuald J., and L. E. Gales. 1964. Allometric growth and the Beverton and Holt yield equation. Trans. Amer. Fish. Soc. 93(4'): 369-381. Phillips, Julius B. 104S. Growth of the sardine. Sardinops cacrulen, 1941-42 through 1940-47. Calif. Div. Fish Game, Fish Bull. 71: 1-32. Platt, John R. 1964. Strong inference. Science 146(3642): 347-353. RicKBR, William K. 1958. Handbook of computations for biological statis- tics of fish populations. Fi.sh. Res. Rd. Can., Bull. 119, 300 pp. Russell, Kdward S. 1931. Some theoretical considerations on the "over fishing" problem. J. Cons. 6(1): 3 20. SCHAEFER, MlI.NER B. 1954. Some aspects of the dynamics of jiopulations important to the management of the commercial marine fisheries. Inter-Amer. Trop. Tuna Conim., Bull. 1(2): 27-56. ScHAEFER, MiLNER B., and R. J. II. Beverton. 1963. Fishery dynamics — their analysis and interpreta- tion. In The sea 2 (ch. 21): 464-483. Inter.science Publishers (John Wiley and Sons), New York. ScHROEDER, William C. 1930. Migrations and other phases in the life history of the cod off southern New P'ogland. U.S. Bur. Fish., Bull. 46: 1-136. SiLLlMAN, Ralph 1'. and .1. P. Wise 1961. Effects of increased trawl cod-end mesh size on Georges Bank haddock yields. Int. ( 'onim. Northwest Atl. Fish., Annu. Meet., June 1961, Doc. 11, Ser. 829, pp. 1-14. Strong, John D., and Georc.e IIannaueh. 1962. .\ practical approach to analog computers. In- struments and control systems 35 (8), 10 pp. (Reprints available at no cost from Electronic .\ssociates, Inc., Long Branch, N.J. 07740.) Thompson, William F., and F. II. Bell. 1934. Biological statistics of the Pacific haliliut fishery. (2) EfTect of changes in intensity upon total yield and yield per unit of gear. Rep. Int. Fish. Connn. 8: 1-49. Weymouth, F. W., and H. C McMii.i.in. 1931. The relative growth and mortality of the Pacific razor clam {Siliqua patula, Dixon), and their bearing on the commercial fishery. U.S. Bur. Fi.sh., Bull. 46: 543-567. WiLLMOvsKY, NoHMAN J,, and Eric C. Wicklund. 1963. Tables of the incomplete beta function for the calculation of fi.sh population and yield. Inst. Fish., I'niv. Brit. Columbia, Vancouver, Can., 291 pp. WiNsOR, Charles P. 1932. The Gomi)ertz curve as a growth curve. Proc. Nat. Acad. Sci. 18(1): 1-8. 4G U.S. FISH AND WII.DI.IFK SKIIVICE A SEROLOGICALLY DETECTED SERUM FACTOR ASSOCIATED WITH MATURITY IN ENGLISH SOLE, PAROPHRYS VETULUS, AND PACIFIC HALIBUT, HIPPOGLOSSUS STENOLEPIS ' By Fred M. Utter and George J. RIDGWAY^ Chemists Bureau of Commercial Fisheries Biological Laboratory, Seattle, Washington 98102 ABSTRACT An antigenic serum component in maturing female pleuronectids was detected by immunodiffusion tech- niques. Its presence was related to age, length, and maturity in English sole {Parophrys vetulus) and Pacific halibut (Hippo^lossus stenolepis). The factor had a qualitative seasonal variation for English sole; the highest incidence was during the spawning season and the lowest at midsummer. The factor was detected in all mature female halibut sampled, but complete sea- sonal data are lacking for this species because samples were not available in summer and fall. The factor was detected in the serum of some immature females of both species during the spawning season. Evidence as- sociating the synthesis of this factor with the produc- tion of estrogenic hormones was obtained when estradiol was injected into male English sole and induced them to produce the factor. Maturity studies are an important aspect of bio- logical investigations of fish. Knowledge of age and size at maturity, fecundity, and duration and frequency of spawning is generally required in the management of a species. As a result of the impor- tance of information on maturity, a wealth of litera- ture exists on the subject covering a broad range of fish species. Most maturity investigations have been concerned with development of ovaries rather than testes because of the importance of egg produc- tion in population dynamics. Biological studies of maturity of the Pleuronec- tidae (flounders) are representative of the variety of approaches for fish generally. Pleuronectid ova- ries may be easily classified as developing or imma- ture by macroscopic inspection during the intervals before and after (and including) the spawning season. Harry (1959) used this method to study time of ' This material was included in a thesis submitted by Mr. Utter in partial fulfillment of the requirements for the M.S. degree from the Graduate School of the University of WashinRton ' Assistant Laboratory Director, Biological Laboratory, West Boothbay Harbor, Maine. Not« — Approved for publication Sejitember 21, 1965. spawning, length at maturity, and fecundity in three species of flounders. Use of a maturity scale, devised by Heincke (1898) for North Sea herring investigations, allows more quantitative estimates of annual ovarian variations. To calculate the spawning season of the Dover sole (Microstomus pacijicus) Hagerman (1952) used a typical modifi- cation of Heincke's scale, based on grossly discernible criteria such as size, transparency, and presence of macroscopic ova. More precise information can be gained by examining the interior of the ovary. By microscopic study of egg diameter and the develop- ment of ova, Thompson (1915) determined the presence of egg stocks for more than one season in the Pacific halibut (Hippoglossus stenolepis) and found an extended spawning interval for individual females. Through histological methods, Franz (1910) identified four distinct developmental stages and established a firmer understanding of the ma- turity process than could be done solely by e.xternal observations of the ovary in plaice (Pletironedes platessa). Characteristic changes in the blood of certain FISHERY bulletin: VOL. 60, NO. 1 47 female vertebrates arc related to maturity and can be studied by biofhemical and serological teclini(iues. Serum vitellin has been associated with maturation in hens (Roepke and Hughes, 1935) and related serologically to egp; vitellin in fowl (Roepke and Bushncll, 1936). Analogous conditions have been described in other oviparous vertebrate classes in- cluding teleosts. Uhlenluith and Kodama (1914), as quoted by Sasaki (1932), used antisora prepared from carp ovaries to distinguish serum of mature female carp from serum of immature female and male carp. Kidgway associated a serologically detected serum factor of the sockeye salmon (also found at high concentrations in the egg) with ma- turity in females (Ridgway, Klontz, and Matsumoto, 1902); subsequently he has found the antiserum to cross-react with mature females in all salmonid species tested (Ridgway, unpublished data). Con- currently, Vanstone and Ho (19G1), in studies of electrophoretic i)atteins of coho salmon sera at vari- ous stages of development, observed a component that was characteristic of maturing females. Fine and i:)rilhon (1903) identified a similar protein in Salmo salar by immunodiffusion. Existing evidence indicates that serum vitellin may be accounted for by the following secpience of events in oviparous vertebrates: under the control of the pituitary, estrogen produced in the ovary stimulates produc- tion by the liver of proteins that are passed through the blood to the ovary and there utilized in yolk formation. The present study attempts to unite the biological and serological approaches in investigations of the maturity of two pleuronected species. We origi- nally intended to study only Pacific halibut. When we found that collecting an adequate halibut sample was impractical, we included English sole {Parophri/s rrlvlus), a more available species. The study is based on a serum-vitellin component, called the HM factor, which occiu-s in mature female flounders. The objective is to demonstrate that serological methods may be advantageously applied in matur- ity studies of these species by showing the relation of the factor to various biological features. METHODS AND MATERIALS The serological mctiiods, preparation of the anti- genic substance used in the methods, and the collec- tion of samples for analysis as well as for deter- mination of age of fish from which samples were taken are described. SEROLOGICAL METHODS The procedure for the detection of the TIM factor was a microslide adaptation of the Ouchterlony method of double-diffusion prccijjitin analysis as described by Ridgway et al. (1902). The diffusion method provides a means of identifying antigenic components of a solution through diffusion of the solution and an antiserum towards one another in a .semisolid medium,'' If the antiserum contains anti- bodies specific for components of the solution, a precipitate line is formed in the zone where given antigen molecules meet specific antibodies in optimal proportions. If two antigen solutions that are placed adjacently diffuse towards a single antiserum source, precipitate lines for common antigen- antibody systems will fuse. Tests for the presence of the HM factor were made in the manner illustrated in figure 1. Sera from known HM-positive females were placed in positions 1 and 4, thus placing each unknown serum adjacent to an HM-positive individual. Distinct positive reactions are seen in positions 5 and 0; a weak positive reaction in position 2, and no reaction in [losition 3. This arrangement was particularly useful for positive individuals with low concentra- tions of HM antigen. Although a distinct line was not necessarily formed, a licnding of the control line towards the unknown position, as in position 2, indicated the presence of the HM factor and allowed a highly sensitive test for the HM factor's presence. Relative concentrations of the HM factor were determined by a single-diffusion method described by Hayward and Augustin (1957). In this method the antiserum is incorporated into the agar at 5 percent concentration and serial dilutions of the fluid bearing the HM factor are introduced into wells in the agar. The end point, the highest dilu- tion at which a visible ring can be observed around the well, is referred to as the titer of the solution for the HM factor. Figure 2 illustrates the reactions of serial dilutions of HM factor from an extract from eggs of starry flounder {I'latichtijs ddlalus) with the antiserum used in this study. (The pre- paration of both the extract and the antiserum is described below.) The end i)oiiit is at the 1/100 dilution. •The medium in this study had the followinK romposition: Difeo agar. I..'i iH-reent; sociiiiiii rliloride, 0.72 percent: sodium eitrato. 0.6 percent: merthiolatv, 001 percent; trypan blue. 0.01 percent. The pH was adjuatt^d to 6.7, 48 U.S. FISH AND WILDLIFE SEUVICE Figure 1. — Reactions on typical double-diffusion slide demonstrating tests for presence of HM antigen in kidney-tissue fluids of female English sole. Magnification 5X. Figure 2. — Demonstration of single-diffusion technique to estimate relative concen- trations of HM antigen. End point is seen at 1/160 dilution. Magnification 2.5X. SERUM FACTOR IN MATURITY 49 PRODUCTION OF ANTISERUM The antigenic substance used for production of antisera was a vitellin preparation from starry floun- der eggs. This prei)aratiou met the chissical bio- chemical criterion for vitelhn, since it was the water- insoluble fraction of the egg yolk (Jukes and Kay, 1932). The pr('])aration was made in tlie following manner: extraction was in a Waring Blendor ^ from 1 part of eggs to 3 parts 1 percent saline solution. One part of the supernatant oi)taincd after centri- fugation at 10,000 r.p.m. for 20 minutes was diluted with 11 parts distilled water, and the resulting pi'c- cipitate was dissolved in 1 percent saline, rcprecip- itated, and redissolved. The vitellin preparation was utilized in this final form for immunization procedures. Altiiough different injection routes were used and the vitellin preparation was injected both with and without iidjuvants, consistently uniform antisera were produced in the five rabbits that were stimulated. Because of this uniformity, a single pooled reagent composed of nvmierous bleedings from all of the rabbits was made and used through- out this study. The antibody specificities appeared to be directed toward one or more of the starry flounder's vitellin antigens. The pooled antiserum cross-reacted with serum of mature females of all pleuronectid species tested. Reactions with sera from males were ob- served only infrequently. These reactions were invariably weak and are discussed in a later .section. COLLECTION OF SAMPLES Collection dates and numbers of English sole sampled are listed in table 1. All samples were collected by the University of Washington research vessel Commando at the Port Orchard area of Paget Sound. Only female English sole were taken. These fish, with the exception of the March sample, were randomly sampled over a variety of lengths. This sample was biased toward smaller individuals with relatively undeveloped ovaries. This bias was based on the relatively high frequency of the HM factor in smaller females in the two previously col- lected samples; the data for the March sample reflect this bias. Random sampling was resumed through the remainder of the investigation. Collection of an adequate blood sample from h^ng- lish sole was found impractical, and all (lualitative determinations for the HM factor were made from the fluids expressed from kidney tissue. When quantitative data were required, blood samples were collected from larger individuals by cardiac i)uncture. The collection data for the halibut samples are presented in table 2. A sufficient blood volume was available from the individual halilnit to u.se serum for determination of the HM factor, although par- allel samples of kidney tissue were colle('ted from most fish. All samples were collected under the direction of the International Pacific Halibut Commission. 'l'.\BI,?; 1. — Collection (Uilix (111(1 number of Joiidh EiKjlish sole in samples, Port Orchard area, Wash. Collection date Fish Collection date Fish mil Dec. 27 Number 30 38 42 IMS Julys Number 25 Feb. 5.... July 23 41 Mar. 7 Octolier 22 .^5 November 23 58 Table 2. — Date and areas of collection, number of indindu(d samples, and source of HM antigen for halibut Fish Source or HM antigen Collection Area date Female Male Serum Kidney ism Number Number Mar. 25 Ciueen Char- lotte Sound.. 27 13 X _ WHS May 11 Cape Flattery, 6 10 X X June (■>'.-. .- Queen Char- lotte Sound.. 41 X X I9r,3 Feb. 2 Cape Flattery . 7 fi X X May 23 Capo Flattery . (i 15 X X * Trade names referred to in this publication do not iniitly cndorscnient of commercial i>roducts. ' Serum collected day of capture; fiah eviscerated and iced at tliis time and kidney fragments collected in port 4 days later. AGE DETERMINATIONS Age determinations in both halibut and English sole were made from otoliths. Personnel of the International Pacific Halibut Commission made age determinations of the halibut. English sole otoliths were treated in papain as described by Pruter and Alverson (l'J()2) and read by the senior author. COMPARISONS OF HM CONCENTRATION IN SERUM AND KIDNEY FLUIDS Tlie need to u.se kidney tissue to obtain ciualitative data on the HM factor in English sole required us to compare concentrations of the HM factor in kidney- tissue fluids and blood serum of the same individuals. 50 U.S. FISH AND WII.DLIFK SEUVICK Table 3 makes such a comparison in 10 females taken a few weeks preceding the spawning season. An end-point fluctuation of plus or minus one serial dilution can be anticipated as part of the experi- mental error inherent in the technique (Kabat and Mayer, 1961). Only one fish e.xceeded this range. Table 3. — Cotnparative titers of HM factor in kidney-tissue fluid and serum of 10 English sole Fish number Kidney fluid titer* Serum titer' 1 32 32 64 64 64 32 32 64 32 32 64 2..._ 3 4 5 6 64 64 64 128 32 7._ 8 32 32 9 10 128 32 *Reci|irocal of last positive dilution. In 19 female halibut, where kidney fluid and serum samples were obtained from freshly caught individ- uals, identical qualitative results were observed for every fish. In the halibut sample taken during June 1962, kidney fragments were obtained from carcasses that had been eviscerated and iced for 4 days, but serum samples were obtained from the same fish when freshly taken. Qualitative tests were made with both serum and kidney fluids. Quantitative tests were made with those sera which gave positive double-diffusion reactions (table 4). The only disagreements between the qualitative data for the kidnej^-tissue fluids and serum were the three individuals with the lowest serum concentra- tion. The 4-day icing of the cleaned halibut carcass doubtlessly diluted the HM concentration in the Table 4. — Comparison of HM titer of blood serum in halibut with double-diffusion reaction of kidney-fragment fluids col- lected from, the same individuals 3 to 4 days after evisceration and icing Fish Serum titer' Kidney fluid reaction Positive Negative Number 6 256 128 54 32 16 8 4 2 No reaction Nil mber 6 7 4 1 2 Number 7.. . . 4 1 2 1 1 1 1 1 1 18. 18 *ReciprocaI of last positive dilution. adhering kidney fragments; it seems likely all tests would have agreed if the kidney tissue had been fresh. The above evidence indicates that serum and freshl}' taken kidney-tissue fluids can be used inter- changeably with considerable confidence for detec- tion of the HM factor in these two species when qualitative data are desired. ANALYSIS OF DATA ON ENGLISH SOLE The six stages used by Hagerman (1952) to des- cribe development of the ovary in the Dover sole were modified in the following manner to describe the development in the English sole: Immature: A. Ovaries very small (generally less than 1 g.), white, transparent, and somewhat gelati- nous. Mature : B. Developing. Ovaries enlarging, becoming yellowish and opaque. Developing egg visible macroscopically. C. Gravid. Ovaries very full of yellowish granular eggs. D. Spawning. Ovaries full of transluscent eggs which run under slight pressure. E. Spent. Ovaries flaccid; ovarian membrane vascular and sac-like. F. Resting. Ovaries firm, white, transluscent, and somewhat gelatinous. Distinguished from stage A by the greater size. The scale was not universally applied in this study owing to the overlap among the various stages. Stages A and F, in particular, were often difficult to distinguish; however, in November through Feb- ruary, including the peak of spawning in December and January (Holland, 1954; Harry, 1959), the stage A ovaries were distinct because stage F ovaries were lacking. All individuals taken during this period with ovaries in stages B through E were HM positive. Certain individuals with stage A ovaries surpris- ingly were HM positive during the spawning season (table 5). Maturity classifications were made on these gonads, which were fixed in formalin, ^^'hen state of maturity is not listed, the ovaries had been sectioned for histological examination before any e.xternal maturity classification was attempted. The two individuals with mature gonads had ovaries in SERUM FACTOR IN MATURITY 51 stages B and C, and all gonads classified as immature represented stage A. The HM-positivc individuals with stage A gonads from the November samples might have matured later in the spawning season. It is unlikely, however, that the two HM-positive fish with immature gonads from the February sample would have spawned during that season, since all ovaries collected subsecjuently, through the July sample, were immature or spawned out. Preliminary results of histological studies being carried out on h^nglish sole gonads by Kathleen Ladue, of our staff, were available for the lower two fish from the February sample listed in table 5. Although both were HM positive, only the ovarj" of the last individual showed atretic follicles as evi- dence of having spawned ; the ovary of the other gave no histological evidence of maturing ova. This evidence supports the previous implications that the 11 M factor occurs in certain female English sole during the spawning season or seasons preceding that in which they are destined to spawn initially. Table 5. — HM-jmsitit'e female English sole from November, December, and February samples that had gonads weighing less than 1 g., Port Orchard, Wash. Month of capture and m.iturity classification Gonad weigllt standard body length Age group I9GS XoFfmhfT Immatnro _, Gm. 0.2 .5 .5 l . 7 .9 .8 .9 .5 .6 .8 .8 .8 A/m. 207 208 225 249 250 257 265 254 245 247 2:15 268 254 265 III III Immature HI Mature _ II Immature III III Immature in Will December III Mature . II 19G3 February II Immature . II II II III Table 6 lists the frequency of the HIM factor in the various samples by age. The samples are grouped by age to follow the ovaries from the devel- oping stage through the resting stage. Samples taken after the height of the spawning season (December and January) through July are assigned to the age they had during the spawning season, because up to July the ovaries are progressing to- wards the resting stage. The October and Novem- ber samples are assigned to the age that they would have had at the next spawning season, because the ovaries are maluring during these months. Thus a fish entering its fourth year of life in January would be included in age-group III if taken in July, and age-group IV if taken in October. The HM factor occurred initially in age-group II in the fish sampled in this study. The possibility of its presence in group-I fish is not ruled out because insufficient numbers of the group were sampled. At a given age the highest frequency of tlu; HM factor appears in the December sample at the peak of the spawning season. Frequency generally de- creases through July but increases again in October and November. The occurrence of a lower II M frequency in II- and Ill-group fish for the March sample than for either the June or July samples is very likely the result of the bias described under Methods and Materials regarding the collection of that sample. Table 6. — Frequency of the UM factor in the various groups of female English sole, Port Orchard, Wash. age Monlh and item Age group I II III IV V 1901 nec(*mber: 5 4 80 M 10 29 43 2 5 25 8 32 'I i:i 12 4 33 II 2 14 14 12 86 3 1 33 8 13 10 7 60 Ifi 3 19 37 26 60 39 29 74 8 8 100 Percentage positive ms February: Total number of fish I 9 Percentage positive. ,. March: Xuniber positive ... June: Total number of fish . I 100 20 1 1 100 5 3 60 3 Number positive 3 Percentage positive 100 July: Total number of fish 11 55 October: 5 Number positive 5 100 November: Total number of fish Number positive .\ge-group V and older. Figure 3 is a plot of the seasonal fluctuation of the HM factor with body length. Fish 22 to 25 cm. long were combined in the December sam()le because the separate groups had few individuals. U.S. FISH AND WILDLIFE SERVICE IffTVTITlltffT"*'™ — ttfln# ■ February ■ March June July October Navember December -21 22-23 24-25 26-27 28-29 30-31 32-33 -34 LENGTH INTERVALS (Cm ) Figure 3. — The relation of the frequency of the HM factor to length in female P^nglish sole in different months, 1962-63, Port Orchard, Wash. A distinct seasonal change took place between the spawning in December and the resting stage in July; during this period the HM frequency for a given length interval generally decreased. In the October and November samples, as tlie ovaries developed for the ne.xt spawning season, the HM frequencies in the various length intervals again increased. Figure 3 does not take into account any growth between the December sample and subsequent col- lections. If seasonal growth increments were con- sidered, most individuals 26 to 29 cm. long in Decem- ber would be 30 to 33 cm. long in July. This state- ment is based on the assumption that El Sayed's (1959) growth estimates for English sole from Holmes Harbor in Northern Puget Sound are applicable to the Port Orchard population. A comparison of the — Februory ■Morch June July ^=^ October - ■■■ — November oooooo December 0.5-0.9 1.0-1.4 1.5-1.9 2.0-2.9 3.0-3.9 GONAD WEIGHT (G.) 4.0 HM frequency in these two length intervals in December with the next higher interval in July still indicates a striking variation of freciuency. Ovarian weight and HM frequency are related (fig. 4) . Gonad weight and HM frequency increased simultaneously. A qualitative seasonal variation in the factor's presence is again indicated. All ovaries weighing more than 1 g. that were taken during or immediately following the spawning season were from HM-positive individuals. Even a con- siderable number of the largest ovaries taken during the resting stage were from HM-negative individ- uals; HM frequency decreased as the ovarian mass decreased. The relation between gono-somatic ratio (gonad weight expressed as percentage body weight) and HM frequency is generaly similar to that between gonad weight alone and HM frequency (fig. 5). 0.3 0.4 0.5 0.5 0.7 0.8 GONO-SOMATIC RATIO 0.9 2|0 Figure 4. — The relation of HM frequency to gonad weights in female English sole, 1962-63, Port Orchard, Wash. Figure 5. — The relation of HIM frequency to gono-somatic ratios in female English sole. The gonad weight shows the relation more clearly than the gono-somatic ratio, however, suggesting that the qualitative seasonal variations of the HM frequency are dependent more on the absolute weight of the gonad than on its mass relative to body weight. ANALYSIS OF DATA ON HALIBUT The HM frequencj^ of the various halibut samples is given according to length in table 7. Fish of various ages were combined in the length intervals because of the small samples. The criterion for maturity of halibut ovaries was the presence of macroscopic ova. Personnel of the International Pacific Halibut Commission made all maturity esti- mates. The HM factor was detected in the serum from all female halibut that were estimated to be mature. SERUM FACTOR IX MATURITY 53 In the sample of June 5, 1962, which gave the best representation of different ages within a gi\-en length interval, no indication of an age dependency was detectable. In spite of the small sample sizes, the fre(|uency of the HM factor seems to increase in length in all samples but that of February 23, 1963. This exceptional situation is discussed lielow. Tahi.e 7. — IIM frequencies in female hatibiil iiecnrding to lenijlh Length (cm.) Date of collection <85 85-100 101-110 >110 HM + HM— HM + HM— HM-I- HM— HM-I- HM— Niim- Num- Num- Num- Num- Numr Num- Num- ber of ber of ber of ber of ber of ber of ber of ber of i9es fish fish fish fish fish fish fish fish Feb. M 2 3 1 1 (1 n mo Apr. 20 1 1 3 1 7 (1 ;9«5 .^pr. 2.1 2 1 1 1 1 mm Miiy 11 2 2 2 n Junes - fi 10 8 8 9 Note. — Sampling areas shown in table 2. As pointed out previously, we detected the HM factor during the spawning season in tlie serum of certain indi\'idual English .sole with immature ova- ries; apparently a similar condition exists in halibut. Two 76-cm. females taken in the February 23, 1963, sample at the peak of the spawning season were HM positive. [Thompson (1915) defined the spawning season for halibut as extending from December to April with the peak in February.] These fish were smaller than the minimum length reported by Thompson for mature females captured from areas included in this study, and the small ovaries diil not give external indications of development. The rec- ords of scrum titers of the HM-positive females in the sample of February 23, 1963, show that the two HM-positive females which were judged immature had very low titers whereas the mature individuals had titers exceeding a serum dilution of 1/250 (table 8). These findings suggest that a quantita- tive means may lie used to distinguish maturing and mature females from immature females in which the HM factor may also be present. In the discu.ssion of table 4 in a preceding section, serum titers were comjiared with (lualitative reac- tions of kidney-tissue fluids from the halibut sample taken on June 6, 1962. The only failures of parallel reactions were in individuals lia\'ing .serum dilution titers of one-eighth or less. This information indi- cates that kidney-fragment fluids may l)e useful in researchers' obtaining data relative to maturity and sex by sampling the eviscerated fish in the commer- cial catch upon arrival in port. Tahi-E S. — Serum tilerx mid iiinliiritji eslimales of IIM- posilive fernalea from hiilihut soiiipir of February 2.-i, 1!)6S, taken at Cape Flattery, Wash. Number of flsh HM scrum titer' Estimate of maturity' 1 <2 <2 512 512 256 1.024 Immature . .do 2 Mature 6 do 8-. ...do 11 do ' Reciprocal of last positive diltition. PRODUCTION OF HM FACTOR BY MALES The link connecting estrogen with the i)roduction of .serum components related to maturity has been determined by a numi)cr of investigators of birds, through the production of these components in males and immature iemales after artificial stimula- tion with estrogenic hormones (McDonald and Rid- dle, 1945: Frist and Schjeide, 1961). Bailey (1957), Frist and Schjeide (19()1), and Ho and Vanstone (1961) have likewise demonstrated that artificial stimulation with estrogenic hormones produces a blooil serum situation similar to that of the mature female in the teleosts Carassius aiiratus, Paralabrax clalhratus, and Oncorhi/nchus ncrka, respectively. We attempted this procedure in this study with four English sole maintained in an aquarium at the Fniversity of Washington College of Fisheries. Each fish was injected intramuscularly with 1 ml. of an aqueous suspension of estrone (5 mg./cc). Two fish survived the initial handling (table 9). Data for the fish firmly establisii the presence of the HM factor as a consequence of introducing the estrogenic hormone. .Mthough a control bleeding was not made for fish Xo. 1, the rise in titer between the first and second bleeding establishes beyond doubt the efTect of the estrone injection. The HM factor was in low concentration in two male halibut taken during the spawning season. The possibility of coiitaniinat ion cannot be excluded becau.se both were taken immediately following the collection of a sample from an HM-positive female, although precautions were taken to minimize contamination. 54 U.S. FISH .\ND WILDLIFE SKRVICE Table 9. — Effect of estrone injections on the occurrence of the HM factor in the serum of male English sole Fish number and time of bleeding HM reaction Titer' No. 1: Prior to injection 48 hours after injection ___ 144 hours after injection No. 2: Prior to injection 48 hours after injection 144 hours after injection — 2 + + + + 8 256 <8 256 ' Reriproral dilution. 2 No control bleeding. The production of the HM factor in the serum of male English sole after estrogenic stimulations and the natural occurrence of the factor in male halibut during the spawning season indicate that males should be included in investigations of the HM factor. The natural occurrence of the factor in males may have a number of causes. Although the testes of the HM-positive male halibut appeared normal, the presence of the factor may have been due to low-level secretions of estrogenic hormones. Hermaphoditism has been reported in a diverse range of tcleosts, including clupeids (Fowler, 1912), salmonids (Ross, Yasutake, and White, 1963), silur- oids (Singh and Sathyaneson, 1961), cyprinodonts (Chidester, 1917), centrachids (James, 1946), and scombroids (Uchida, 1961). A further possibility is the stimulation by estrogenic hormones of exoge- nous origin through ingestion of mature females of smaller species of flatfish or possibly fish from other families. Estrogenic hormones are effective when administrated orally to mammals and presumably could be similarly effective in fish. BIOLOGICAL IMPLICATIONS OF THE PRESENCE OF THE HM FACTOR The English sole has been demonstrated to have a qualitative seasonal ^■ariation of the HM factor. The factor occurs first during the spawning season, at least as early as the second year in some individ- uals. After the spawning season through at least midsummer, fewer and fewer individuals retain the factor in the serum. As a new spawning season approaches, the factor gradually reappears. Both the disappearance and the reappearance of the factor are more pronounced with increase of body length and ovarian mass. This relation may be the result of resorption of residual vitellin retained with the ovary, since the large ovaries retain a greater volume of unspawned ova. A ciualitati\-e seasonal variation was not found in mature halibut, but we lack samples taken later than June. Thompson (1915) reported a continuous development of the ova which are to mature in the succeeding generation in the spent halibut ovary; vitellin synthesis may be a perennial process in the mature female halibut. The detection of serum vitellin in postspawning Atlantic salmon (Salmo salar) by Fine and Drilhon (1964) suggests its peren- nial occurrence in this species. The presence of the HM factor during spawning season in immature females of both species indicates that such an occurrence may be widespread among the Pleuronectidae. Incomplete maturation pre- ceding initial spawning in the Pleuronectidae has been reported previously. Thompson (1915) stated that contemporary investigators had found some ova in immature pleuronectid females which ap- peared ready to ripen but which failed to do so because the ovary, as a whole, was not yet ready. Franz (1910) reported finding this condition most marked in plaice during the last winter preceding initial spawning. PRACTICAL APPLICATIONS OF THE HM FACTOR As a practi<'al procedure, the determination of the HM factor appears to have its greatest potential value in the larger pleuronectid species. In large species such as halibut or starry flounder, the sex cannot be determined at sight, except in ripe indi- viduals. Small samples of blood taken at the time of tagging could yield information on sex and matur- ity without endangering the fish. Repeated bleed- ings of four starry flounders kept in captivity did not appear to endanger these fish. Routine prac- tical applications to smaller species seem less likely. The sexes of smaller flatfish species, such as English sole, are generally evident by external examination; and bleeding English sole, where required in this study, caused high mortality. Evisceration of the commercial halibut catch at sea does not preclude practical application, since analysis of kidney- fragment fluids can be made after the catch arrives in port. On the other hand, smaller species are brought to port in the round and sex information can be obtained directly. Smaller species, however, are frequently more readily available in greater SERUM FACTOR IN MATURITY 55 numbers; they are valuable for clarification of gen- eral principles wliicli may be applied to other species as well. SEPARATION OF MATURE FEMALES FROM OTHER HM-POSITIVE INDIVIDUALS Separation of HM-positive males and immature females from mature females appears possible by quantitative means. The HM serum levels of ma- ture female halibut taken during the spawning season had titers above 200, whereas the titers of HM- positive males and immature females, which were found only at this time, were less than 2. An extension of the single diffusion ciuautitation, as used in this study, may be applied where routine ciuantitation is required. From figure 2 it can be observed that the diameter of the precipitin ring decreases regularly as the HM concentration de- creases. A measurement of the diameter of the precipitin ring formed by the undiluted fish serum could give the approximate titer. AREAS FOR FURTHER INVESTIGATION Several areas for further study are evident. More freiiuent and larger samples are desirable. As indi- cated above, routine quantitation may l)e necessary during the spawning season, and a knowledge of the quantitative seasonal fluctuation of the HM factor in a given species would be useful. Perhaps the relative HM concentration can be related to such factors as age, weight, or fecundity. More exten- sive histological examinations of the ovary would help, and a similar examination of the pituitary gland may establish more fundamental criteria for the occurrence of the factor. A biochemical assay might indicate that the composition of the factor in fishes is related to analogous components in other vertebrates. THE BROADENING APPLICATION OF SEROLOGY IN FISHERY RESEARCH Serological techniques have had increasing appli- cation in fishery problems during recent years. This research has been directed mainly toward racial studies of serum antigens or red blood cell antigens. Manj- of the current approaches to serological in- vestigations of populations were discussed in a sym- posium moderated by Cushing (1902), and the sub- ject has been reviewed recently by Marr and Sprague (19t)3) and Cushing (H»(i4). Antigenic differences at species level have also been investigated. Ridgway and Klontz (unpul)- lishod data) and SindermaiHi (li)(i2) have found distinct species-specific antigenic characteristics in red blood cells and serum of species of Pacific salmon and .Vtlantic cluijciods, lespectively. Ridgway (1903) reported species-specific antigens in muscle tissue of certain tuna species, in addition to species-s[)ecific lilood serum components. This finding olTers a jios- sible serological means of distinguishing larva^ of these species. We hope that this study will help broaden the interest in ai)plication of serological methods to other areas of fishery biology. Recause components similar to the HM factor have been detected in a diverse range of teleosts, a similar approach presum- ably could be used throughout this class of verte- brates. We feel that this a|)proach can be a valuable supplement to in\'estigating maturity in fish, though perhaps not universally applicable. SUMMARY A serological investigation of a serum vitellin factor in mature and maturing female flatfish was made on English sole and Pacific halibut. Immuno- diffusion techniques with antisera prepared in rab- bits stimulated with egg vitellin extracted from starry flounder eggs were u.sed to detect the factor. In English sole the factor's occurrence was compared with age, length, gonad weight, and gono-somatic ratio. A qualitative sea.sonal variation was found; individuals with heavier ovaries during the summer were more likely to retain the factor in the scrum. The presence of the factor in female halibut was compared with length, age, and maturity. A ([uali- tative seasonal variation in mature halibut could not be studied because no samples were available tluring the summer or autumn. The factor was found in some immature females of both species during their spawning seasons. Production of the factor in male English sole by injections of estrogenic hormones associates synthesis of the factor in females with production of estrogen. Determination of the factor appears to have potential value as a supplement to other means of investigating maturity, particularly in large species. 56 U.S. FISH AND WILDLIFE SKUVICE LITERATURE CITED Bailey, R. E. 1957. The effect of estradiol on serum calcium, phos- phorus and protein of goldfish. J. E.xptl. Zool. 136(3) : 455-469. Chidester, F. \V. 1917. Hermaphroditism in Fiiiididus hrtiroclitus. Anat. Rec. 12(3): 389-396. CusHiNG, John E. 1962. Symposium on immunogenetic concepts in marine population research. Anier. Natur. 94(889): 193-256. 1964. The blood groups of marine animals. Adv. Mar. Biol. 2: 85-131. El-Sayed, Sayed Z. 1959. Population dynamics of Engli.sh sole {Paroiihnjs retulus Girard) in Puget Sound, Washington, with special reference to the problems of sampling. Ph.D. Thesis, University of Washington, 189 pp. Fi.\E, J. M. and A. Drilhon. 1963. Etude immunologique des protcines de S(''rum de Snimo salar. Etude par immunodifTusion. Compt. Rend. Soc. Biol. Paris 157: 1937-1940. 1964. Etude I'lectrophoretique et immunologique des protcine-s scriques de quelques especes de Salmonides. Compt. Rend. Soc. Biol. Paris 158: 1307-1310. Fowler, H. W. 1912. Hermaphrodite shad in the Delaware. Science 36(914): 18-19. Franz, Victor. 1910. Die Eiproduktion der Scholle. Wiss. Meeres- untersuchungen, Aht. Helgoland, N.F. 9: 62-137. Hagerman, Frederick B. 1952. The biology of the Dover sole, Micrnslotinis pacifwus (Lockington). Calif. Dept. Fish Game, Fi.sh Bull. 85, 48 pp. Harry, George Y., Jr. 1959. Time of spawning, length at maturity, and fecundity of the English, petrale, and Dover soles {Pnrnphrys vcluliis, Eopsetla jordnni and Microstomus pnciftcits, respectively). Oreg. Fish Comm. Res. Briefs 7(1): 5-1:?. Hayward, B. J., and R. Augustin. 1957. Quantitative gel diffusion methods for assay of antigens and antibodies. Int. Arch. Allergy 11: 192-205. Heincke, F. 1898. Naturgeschichte des Herings. Teil I : Die Lokalf- ormen und die Wanderungen des Herings in den europaischen Meeren. Abhandl. Deutsch. Seefi- scherei-Vereins 2(1), 238 pp. Ho, F. Chung-Wai, and W. E. Vaxstone. 1961. Effect of estradiol monobenzoate on some serum constituents of maturing sockeye salmon (Oncorliyn- chits nerka). J. Fish. Res. Bd. Can. 18(5): 859-864. Holland, Gilbert A. 1954. A preliminary study of the populations of English sole {Parophrys vduliis Girard) in Carr Inlet and other localities in Puget Sound. M.S. Thesis Uni- versity of Washington, Seattle, 139 pp. James, Marian F. 1946. Hermaphroditism in the largemouth bass. J. Morph. 79(1): 9:3-94. Jukes, T. H., and H. D. Kay. 1932. Egg-yolk proteins. J. Nutr. 5(1) : 81-101. Kab.\t, Elvin A., and Manfred M. Mayer. 1961. Experimental immunochemistry. Charles C. Thomas, Springfield, 111. 905 pp. Marr, John C, and Lucian M. Sprague. 1963. The use of blood group characteristics in studying subpopulations of fish. Int. Comm. N. Atl. Fish., Spec. Publ. No. 4: 308-313. McDonald, Margaret R., and O. Riddle. 1945. The effect of reproduction and estrogen admin- istration on the partition of calcium, phosphorus and nitrogen in pigeon plasma. J. Biol. Chem. 159(2): 445-464. Pruter, .\lonzo T., and D. L. Alverson. 1962. Abundance, distribution and growth of flounders in the Southeastern Chukchi Sea. J. Cons. 27(1): 81-99. Ridgway, George J. 1963. Distinguishing tuna species bj- immunochemical methods. U.S. Fish Wildl. Serv., Fish. Bull. 63: 205-211. Ridgway, George J., G. W. Klontz, and C. Matsumoto. 1962. Intraspecific differences in serum antigens of red salmon demonstrated by immunochemical methods. Int. N. Pac. Fish. Comm., Bull. 8: 1-13. Roepke, R. R., and L. D. Bushnell. 1936. A serological comparison of the phosphoprotein of the serum of the laying hen and the vitellin of the egg yolk. J. Immunol. 30(2): 109-113. Roepke, R. R., and J. S. Hughes. 1935. Phosphorus partition in the blood serum of laying hens. J. Biol. Chem. 108(1): 79-83. Ross, A. J., W. T. Yasutake, and G. R. White. 1963. Hermaphroditism in rainbow trout. Trans. Am. Fi.sh. Soc. 92(3): 313-315. S.^saki, Koyotsuna. 1932. Precipitation test for the se.xes of fowl blood .serum with special reference to egg-laying. J. Immunol. 23(1): 1-10. Sindehmann, Carl J. 1962. Serology of Atlantic clupeiod fishes. Amer. Natur. 96(899): 225-231. Singh, T. P., and A. G. Sathyanesan. 1961. An instance of hermaphroditism in the catfish, Mystus viltabts. Curr. Sci. India 30(8) : 302-303. Thompson, W. F. 1915. A preliminary report on the life history of the halibut. Rep. B.C. Comm. Fish., 1914(1915): 76-99. Uchida, Richard X. 1961. Hermaphroditic skip.iack. Pac. Sci. 15(2) : 294-296. Uhlenhuth, p., and T. Kodama. 1914. Studien uber die Geschlectsdifferenzierung. Kokka-Igakkai Zasshi no. 331 : 385. [Cited by Sasaki. K. 1932] SERUM FACTOR IX MATURITY 57 I'msT, Marshai.i, Tt., and Arne O. Sciueide. Vanstone, W. E., and V. Chuno-Wai Ho. 1901. The partition of culciuni and protein in the 1961. Plasma proteins of coho sahnon, Oncorhyxchvs blood of oviparous vertebrates during estrus. J. Gen. kisutch, as separated liy zone electrophoresis. J. Fish. Physiol. 44(4) : 743-756. Kes. Bd. Can. 1 8(3) : 393-399. 58 U.S. FISH AND WILDLIFE SERVICE EFFECT OF WATER VELOCITY ON PASSAGE OF SALMONIDS IN A TRANSPORTATION CHANNEL By Joseph R. Gauley, Fishery Biologist (Research) Bureau of Commercial Fisheries Fish-Passage Research Program, Seattle, Wash. 98102 ABSTRACT Passage times of fish at velocities of 1 and 2 feet per second were compared in a 4-foot wide transportation channel, with a water depth of 6 feet. The timing zone was about 100 feet long. Passage times did not differ significantly between water velocities for any one of three species: chinook salmon (Oncorhynchus tshawytscha), steelhead trout {Saltno gairdneri). and sockeye salmon (Oncorhynchus nerka). The two salmon species moved faster than steelhead trout at both water velocities. Transportation channels for migrating adult fish are part of the fish-passage facilities at many clams. These channels vary somewhat physically, but all have the primary purpose of providing a passage area leading either to or from the fish ladders or other passage facilities. Most of the large dams on the Columbia River — Bonneville, ]\IcXary, and The Dalles, for e.xample — have a multiple-entrance collection channel on the dowTistream side of the powerhouse which also serves as a transjDortation channel. In addition, independent channels are occasionally provided at some dams to pass fish from a single major entrance to a distant fishway. The Dalles Dam is ec^uipped with both types (U.S. Army Corps of Engineers, 1957). These channels make it possible for one fishway to serve two or more collection points. Some channels are nearly a quarter mile long and may require up to 1,000 cubic feet per second (c.f.s.) of water for operation. Water velocity in a transportation channel is important fi'om the standpoint of fish passage as well as water use. Clay (1961) reported that the accepted standard velocity for ensuring continuous migration of fish through open channels is near 2 feet per second (f.p.s.). Preliminary experiments at the Fisheries-Engineering Research Laboratory at Bon- Note — Approved for publication March 8, 1966- neville Dam indicated that a velocity considerably less than 2 f.p.s. might be satisfactory for passage of salmonids. If so, velocity standards for transporta- tion channels could be lowered and less water used without impeding the passage of migrating fish. The purpose of this sttidy ' was to determine if salmonids would move up a transportation channel as rapidly in a water velocity of 1 f.p.s. as in 2 f.p.s. EXPERIMENTAL EQUIPMENT The study was made in the Fisheries-Engineering Research Laboratory at Bonneville Dam on the Columbia River. Details of the laboratory were described by Collins and Filing (19G0). The experi- mental transportation channel (fig. 1) was 4 feet wide, 91 feet long, and operated at a water depth of 6 feet. Fish were timed over a distance of about 100 feet. (This included a short introductory area extending from a release compartment to the chan- nel.) Water velocity was controlled by regulating the head on a weir located between the flow-intro- duction pool and the test channel. Headwater elevations producing velocities of 1 and 2 f.p.s. were determined before the experiment was started. I Research financed by tlie U.S. Army Corps of Engineers as part of a broad program to provide design criteria for more economical and efficient fish-pasaage facilities at Corps projects on the Columbia River. FISHERY bulletin: VOLUME 66, NO. 1 59 Exit Fishway Flow- introduction pool Transportation ctianne Entronce fishway g' Rele Fish ■-Water control weir Observation chamber ase box 4 Uf ih entry(A),y Introductory area Excess woter drain qril FiGUKE 1. — Diagrammatic plan view of laboratory slunvint; transportation channel and timing zone (A to B). Velocities were measured with a cup-type current meter. The channel was lighted by 1,000-watt mercury- vapor lights placed feet apart and suspended G feet above the water. Light readings at the surface averaged about 700 foot-candles, which approxi- mates light intensity on a l)right cloudy day. PROCEDURE Salmonids used in these tests were diverted from the Washington shore fishway at Bonneville Dam. They ascended a short entrance fishway to the lab- oratory. Varying numbers of fish were u.sed in each test, depending on seasonal abundance of the different species. Each fish entering the laboratory was tested only once. After a fish had completed passage through the test facility, it left the lab- oratory through a small exit fishway and entered the main fishway about 200 feet tipstrearn from the laboratory. Fish of all sizes were u.sed; however, on a few | occasions, fish were rejected because of severe cuts or other obvious physical injuries. No distinction was made as to sex. l EXPERIMENT.\L DESIGN A 2 by 2 Latin Sciuare design u.'^ed in these tests allowed 4 days (2 davs at each velocity) for each test. The experiment consisted of four separate tests — two with cliinook salmon {Oncorhynchus tshnwytscha) and one each with steelhead trout (Salmo (jairdncri) and sockeye salmon {Oncorhynchus' nerka). TIMING FISH Fish entered a relea-se compartment (fig. 2) FicrnE 2. — Release compartment. Operator has raised gate (foreground) to allow fish to enter test .area. where they were identified as to species and released individually into the test area. Only one fish was permitted in the channel at any one time. CO U.S. FISH AND WILDLIKi: SEKVICE The timing zone (A to B, fig. 1) extended from the release compartment (A) to the exit area (B) at the upper end of the channel. A deflecting grillwork (fig. 3) directed fish toward an observation Figure 3. — Exit area of test channel viewed from above. Grill on left foreground deflects fish toward view'ing area (arrow) of submerged observation chamber. Flow is toward foreground. cent confidence intervals about the median (Dixon and Massey, 1957) were applied to test for signi- ficance of differences between passage times at the two velocities. As used here, the median is the passage time of the median fish of all fish tested in each group, including those fish that failed to com- plete passage within the arbitrary 45-minute time limit. Computations of the mean passage time include only those fish that completed passage of the test channel within 45 minutes. PASSAGE TIME IN RELATION TO WATER VELOCITY Comparisons of time ret[uired to pass through the test channel at the two velocities are given by species in the following subsections. CHINOOK SALMON Tests with chinook salmon were made during two periods— May 8-11 and June 12-15, 1962. Fish in the early period are normally called spring-run and those in the latter period, summer-run chinook salmon. Median passage times in the May test at water velocities of 1 and 2 f.p.s. were 3.4 and 3.9 minutes, respectively (table 1). Results of the June test were similar. Median passage times in the two velocities did not differ significantly between velocities or between tests. Mean passage times, given for comparison, suggest similar trends. In both tests, however, chinook salmon took slightly more time to pass through the channel at 2 f.p.s. than at 1 f.p.s. (fig. 4). This difference corresponds with observations by Weaver (1963), who found that chinook salmon moved more slowly as velocity in- creased in the range of 2 to 8 f.p.s. window at the point of exit. This arrangement ensured accurate observation and timing of the fish when visibility was limited owing to turbid water. The time of entry and exit for each fish was registered on a special time-event recorder. If a fish had not completed passage of the channel within 45 minutes after time of entry, it was removed and anotiier fish was introduced. ANALYSIS OF PASSAGE TIME The effect of water velocity on fish passage was determined by measiu'ing the time required for the fish to pass through the channel. Ninetj'-five per- T.\BLE 1. — Median and mean passage times of chinook salmon in an experimental transportation channel at Bonneville Dam at water velocities of 1 and 2 f.p.s., May and June 1962 Test Water velocity Fish tested Passage time period Median Lower limiti Upper limit' Mean May 8-11 ._ June 12-15 2 I __. 2 Numher 37 44 45 75 \finutes 3.4 3.9 3.1 3.6 Minutes 2.7 3.2 2.5 2.7 Minutes 5.5 4.5 4.4 4.2 Minutes 4.7 '4.9 4.8 6.3 ' 95-percent confidence intervals about the median. ' One fish failed to complete passage within the 4.5-minute time limit and was not included in computation of the mean. TR.\XSPORT.^TION CHANNELS FOR SALMONIDS 61 12 10 STEELHEAD < to < CHINOOK MAY 8-11 JUNE 12-15 SOCKEYE (37) (44) (45) (751 (29) (18) (28) (22) ' IN I I U 111^ 12 12 12 12 WATER VELOCITY ( f p.s ) Figure 4. — Median pa.s.'sage time.s of cliinook .salmon, steel- head trout, and sockeye salmon in a transportation channel at water velocities of 1 and 2 f.|).s., 1962. Numbers of fish tested are shown in parentheses near the base of each bar. STEELHEAD TROUT Median passage times of steelliead trout at water velocities of 1 and 2 f.p.s. were 10.6 and 8.8 minutes, respectively (table 2). This difference was not statistically significant. Mean passage times were similar to the medians. Steelliead trout moved somewhat faster at the higher velocity, in contrast to the difference in chinook salmon (fig. 4). Weaver (1963) observed similar performances among steel- head trout, i.e., faster movement as water velocity increased. In comparison with the other species tested, steel- head trout obviously spent considerable time in the test cliannel. Given suitable hydraulic conditions, steelhead trout frequently remain in favored pools or runs for varying iioriods of time before proceeding upstream. This characteristic possibly accounts for the relatively slow passage times of this species in the present tests. SOCKEYE SALMON Performances of sockej^e salmon at the two water velocities were similar to those of chinook salmon. Both the median and mean passage times (table 8) give evidence of a slightly faster passage at the lower velocity. The difference between median passage times, however, was not significant. T.vBi.E 2. — Median and mean passage times of sterthead trout in an experimental transportation channel at lionnemllc Dam at water mlocities of 1 and 2 f.p.s., Jidi/ SO-.iiK/iisf S, 1962 Water Fish tested Passage time velocity Median Lower limit' Upper limit' Mean F.-P.S. 1 Number 28 19 Minutes 10.6 8.8 Minutes .5.4 3.0 Minutes 15.0 11.0 Minutes »9.6 J8.1 2 ' 95-percent confidence intervals about the median. ■ Four fish failed to roniplet^- i>as8aKe within the •J.'i-minute time limit and wtTf not included in computation of tlie mean. ^ Excludes one fish that did not complete passaKO within 45 minutes. 1\\BLE 3. — Mf'fliari and mcati passage times of sockeye saltuo?l in an experimental transportation channel at Bonneville Dam at water velocities of 1 and 2 f.p.s., July 10-13, 1962 Water Fish tested Piissage time velocity Median Lower limit' Upper limit' Mean F.P.S. 1 Number 28 22 Minutes 1.9 2.7 Minutes 1.5 2.0 Minutes .3.3 .5.7 Minutes 2 =4, 1 ' 95-percent confidence intervals about the median. 2 One fisli failed to complete passaKft witliin the 45-minute time limit and was not included in computation of the mean. CONCLUSION A water velocity of 1 f.p.s. is as suitable as one of 2 f.p.s. for the passage of chinook salmon, steelhead trout, and sockeye salmon in a transportation channel. ACKNOWLEDGMENTS dcrald H. Collins and Carl II. Klling assisted in planning the e.xperiinents. William S. Davis sug- gested the e.xperimental design, and the staff at the Fisheries-Engineering Research Laboratory aided in carrying out the work. LITERATURE CITED Clav. C. EI. I'.Kil. Desi{!n of fishw.ays and other fish facilities. I)o[). I'ish. Can., (Jueen's Printer, Ottawa, Can., :{01 pp. 62 U.S. FLSH .VND WILDLIKK SKKVU'E Collins, Gerald B., and Carl H. Ellino. U.S. Army Corps of ENf;iNEERs, Portland and Walla 1960. Fishway research at the Fisheries-Engineering Walla Dlstricts. Research Laboratory. U.S. Fish Wildl. Serv., Circ. ^^^^- Annual fish passage report, North Pacific Divi- 9g. i_i7 ' sion; Bonneville, The Dalles, and McXary Dams, 49 pp. Dixon, Wilfrid J., and Frank J. Massev, Jr Weaver, Charles R. 1957. Introduction to statistical analysis 2d ed 1963. Influence of water velocity upon orientation and d along the liorizontal dor.sal edge to the point where it forms a slight depression before the si)ine. Total widtli (E) was measured from the anterior jioint of the horizontal dorsal edge along a ])roj('ctioii of the line of this eilge to the posterior margin of the spine. SOth vertebra. Length was measured along the axis of the centrum from the outer anterior rim to the outer posterior rim. Width was measured ver- tically on the lateral sui'face at the narrowest part of the centrum and does not include the neural or haemal arches. Bone terminology is that of de Sylva (1955), which has been used for tuna osteologj' by most other recent workers. A number of names differ from those accepted by many fish osteologists. PART 1. COMPARATIVE ANATOMY The morphological characters useful for distin- guishing the species of Thinmus fall into seven groups: osteology, viscera, vascular system, olfactor.v organ, meristic characters, morphometric characters, and coloration. These will be discussed in this order in this section of the paper. The first three groups, osteology, viscera, and vascular system, in- clude the most important characters. OSTEOLOGY Osteological characters are very important in dis- tinguisliing species of Thuiinufi. They have an ad- vantage over characters in the soft anatomy in that the bones can be saved so futui'e workers can reex- amine the material on which a study is based. We have used a large amount of the material Godsil and his co-workers reported on. Four groups of osteo- logical characters will be considered: neurocranium, liranchiocranium, pectoral and pelvic girdles, and verteliral column. The most us(>ful characters are in the skull and the vertebral column. Each of the four groups will be discussed separately, giving a general osteological description followed by an enu- meration of the specific charactei's. Neurocranium General characteristics. — Details of the general neurocranial structure of tunas have been illustrated by Masterman (1894), Kishinouye (1923), Frade (1932), Gregory (1933), Godsil and Byers (1944), de Sylva (1955), and Nakamura (1905). The ac- companying labeled figures of an albacore (Thunnus alalunga) skull show the bones of the neurocranium (figs. 1-3). Photographs of the skulls of six of the seven species of Tlniiimif: are jjresented in appendix figs. 1-3. The skull of a tuna, compared with that of most typical percoid fishes, is short and wide. In dorsal view (fig. 2), at its anterior end, the dermethmoid (= ethmoid) is wide and its anterior margin only slightly curved. The interorbital region is broad, and the otic region broader still. A i)rominent dor.solateral crest is formed on each side by the frontal, parietal, and epiotic bones; each of the epiotics bears a short, posteriorly directed process. The lateral edges of the frontal and pterotic, making up tlie sides of the roof of the neurocranium, form a more j^rominent and rather fiat sharp crest on each side which extends posteriorly as a long i)late-likc pterotic spine. The supraoccipital crest is high and extends posteriorly o\'er the first few vertebrae. \'eiitrally, the dentigerous vomer (= prevomer) is flanked l)y broad processes of the parethmoids (= lateral ethmoids). Most of the ba.se of the skull is formed by the parasphenoid, which is flat or slightly concave in its anterior two-fifths, bears a medial, ventrally directed crest in the next two- fifths, and posteriorly is first convex, then deeply conca\e, with dorsomesially curved lateral flanges that enclo.se a parasphenoid chamber. Lateral wings jjioject dor.sad from near the end of the ventral crest to form part of the posterior myodome. In lateral view, thealisphenoids (= pterosphenoids) form a partial interorbital septum extending vontrad from the roof of the skull. In extreme cases (large Tluinniis thijnnus) this septum may be fused with the j)arasphenoid, to form a bony septum partially separating the orl)its. In the posterior part of the 08 U.S. FISH AND WILDLIFIO SEHVICE SUPRAOCCIPITAl CREST EPIOTIC PARIETAL FRONTAL DERMETHMOID PTEROTIC VOMER PARETHMOID PARASPHENOIO ALISPHENOID OPISTHOTIC EXOCCIPITAL FIRST CENTRUM PROOTIC BASIOCCIPITAL SPHENOTIC BASISPHENOID Figure 1. — Skull of Thunnus alalunga. Lateral view. orbital region is a median, vertically oriented basi- sphenoid, which usually has an anteriorly directed process near its dorsal end. The posterior base of the cranium is formed by the end of the parasphenoid (ventral profile) and the lateral flanges of the basi- occipital (posterior profile). The first vertebra artic- ulates firmly, partly by a jagged suture, with the occipital region and forms an integral part of the skull. One end of Baudelot's ligament attaches to the basioccipital, the other end to the supracleithrum. The prootic pits (Godsil, 1954) are large pouchlike concavities on each side of the ventral surface of the cranium, opening posteriorly and separating the pterotic bones from the brain case. Part of the roof, floor, and sides of each prootic pit is formed by the prootic bone, and the anterior wall by the sphenotic. The pits function as areas of attachment for the branchial musculature. These pits are char- acteristic of the most advanced scombrids — Thunnus, Euthynnus, Katsuwonus, Auxis, and Allothunnus — and are incipient in Sarda (Starks, 1910; Godsil, 1954). The posterior myodome is a deep median depres- sion opening anteriad at the posterior end of the orbital region. Its anterolateral walls and roof are formed by the prootics, its floor and ventrolateral walls by the parasphenoid, and its posterior concave wall by the basioccipital. The posterior myodome functions as a place of attachment for the rectus muscles of the eyes. The narrow basisphenoid lies just anterior to the anterior opening of the posterior myodome. There is a posterior or parasphenoidal chamber (Kishinouye, 1923), which communicates with the posterior myodome and is formed by the upcurved walls of the posteriormost end of the parasphenoid. A large triangular fronto-parietal foramen (lateral parietal foramen of Masterman, 1894) is present on each side of the dorsal surface of the skull, at the junction of the frontal, parietal, and supraoccipital ANATOMY AND SYSTEMATICS OF TUNAS 69 OPISTHOTIC PTEROTIC PARIETAL SPHENOTIC PARASPHENOID DERMETHMOID PARETHMOID SUPRAOCCIPITAL EPIOTIC EXOCCIPITAL FIRST VERTEBRA FiouRK 2. — Skull of ThuiiniiH alaUinga. Dorsal view. jThe piiioal foramoii is iiiconcctly labeled as parietal foramen. bones. In life these foramina are covered by a tough membrane and are not passages for nerves or blood vessels. We were unable to determine their function. Fronto-parietal foramina arc characteristic of Thun- nus, Euthynnus, and Katsuwoniui, and the bone is thin in this area in several other scombrids. There is a prominent medial pineal foramen between the edges of the frontal bones, just anterior to the supraoccipital crest. Rivas (1954a) has sug- gested that in 7'. Ihijnnus light can pass through the transparent "window" in the skin over this foramen and then down to the brain through the carti- laginous lens that fills the foramen in life. He postulated that the pineal ajjparatus has a photo- tropic function involved in migration. Holmgren (1958) also studied the pineal apparatus of T. Ihynmts but could find no evidence of a photorecep- tive role for the pineal organ. The pineal foramen is characteristic of the more advanced members of the Scombridae and is absent or represented by only a small slit in the more primitive genera such as Scomber, Raslrdligcr, and Scomberomorus (Allis, 1903; Kishinouyo, 1923; Mago Leccia, 1958). Specific Characters. — Four neurocranial cliaracters ro U.S. FISH .\ND WILDLIFE SERVICK BASIOCCIPITAL OPISTHOTIC PROOTIC FRONTAL DERMETHMOID PARETHMOID SPHENOTIC ALISPHENOID f ,/■ . VOMER PARASPHENOID PTEROTIC EXOCCIPITAL PARASPHENOID Figure 3. — Skull of Thunnus alalunga. Ventral view. have been found useful in distinguishing the species of Thunnus: the ahsphenoids, posterior parasphenoid margin, supraoccipital crest, and ventral parasphe- noid shaft. They are characteristic of specimens from all oceans. Alisphenoids (fig. 4). The alisphenoids meet in the median line and extend ventrad into the orbit. They approach the parasphenoid more closely in T. ihijnnus and T. maccoyii than in the other tunas. The greatest height of the anterior part of the orbit, B, measured from dorsal parasphenoid to upper median part of parethmoid, was divided by the least distance between alisphenoid and para.sphenoid, A. In 46 skulls of T. thynnus, 16 have the alisphenoids fused to the parasphenoids; in the remaining 30, A goes into B 2-15 times; with a mean (I-) of 4.8, only in 6 specimens is the ratio less than 2.5. No fusion was observed in T. maccoyii; in 17 skulls the ratio was 2.0-10.3, x 4.8. By contra.st, in all other species, A goes into B 1-3 times. Among 122 skulls of the other species, only 3 T. albacares and 2 large T. tonggol have a ratio of 2.5 or greater. Mean ANATOMY AND SYSTEMATICS OF TUNAS 71 ALB ATL J TON FlouRK 4. — Skulls of Thunn Right lateral view of orbital ronion of neurocrimium showinf; conformation of basi- sphenoid (stippled) and ai.. , -J. Arrangement within each species in order of increasing skull lengths from left to right: ALA— T. alabmga, 131, 138, 150, 107 mm. ALB— T. albacimx, 103, 110, 122, 179 mm. XV\.— T. atUintiais, 51, SO, 97, 102 mm. OBE— 7". ohesus: upper-Atlantic, 170, ISl, 185, 207 mm.; lower-Pacific, 07, 142, 147, ca. 240 mm. THY— T. thynnus, 76, 139, 231, ca. 320 mm. TON— T. tonggol, 57, 61, 122, 128 mm. values arc: T. tonggol (.V=8) 2.0; T. alhacarcs (A^=43) 1.8; T. ohesus {N=32) 1.6; T. alnhmgn (.V=27) l.G; T. ailanlicus (iV=14) 1.2. Posterior parasphenoid margin (fig. 5). Tlie jjor- tion of the parasphenoid forming the walls of the parasphenoidal chamber is partialh- covered laterally by tlic basioccipital. Together tlie margins of these bones extend vcntrad from the first vertebra, either vertically or slanting forward or backward ; anteriorly the parasphenoid alone forms the margin. With due consideration for growth changes and individual variation, the profile formed at this part of the skull 72 U.S. FISH AND WILDLIFE SKUVICE Figure 5. — Skulls of T/ii(?!?H;s species. Right lateral view of postcroventral part of neurocranium showing conforma- tion of posterior parasphenoid margin. Arrangement with- in each species in order of increasing skull length from right to left. ALA— 131, 138, 150, 167 mm. ALB— 103, 107, 116, 122, 169, 179 mm. ATL— 51, 80, 88, 97, 102 mm. OBE— Pacific, 97, 141, 143, 146, 165, 189, 210 mm. THY— 76, 139, 230, 335 mm. TON— 57, 61, 122, 127 mm. is quite characteristic of somi species. In T. thynnus, T. maccoyii, and T. alabmga a decided angle is formed by the posterior parasphe- noid margins. The acuity of the angle and its posterior e.xtent generally increase with size in T. Ihynnns (fig. 5), and extreme development of the angle is found in very large specimens. Within its observed size range (skull length 88-167 mm.), T. alalunga, however, displays a more acute angle than does T. thynnus of similar size. T. maccoyii resembles T. thynnus in this respect. In relatively large specimens of T. obesus the angle is apparent but not as acute as in T. thynnus, T. maccoyii, and T. alalunga. Observations of Pacific specimens indicate probable changes with growth. The two smaller eastern Pacific specimens reported by Godsil and Byers (1944) and again by de Sylva (1955) have rounded margins, but other eastern Pacific specimens within the same size range show a definite angle. T. albacares, T. atlanticus, and T. tonggol exhibit great variation. Some have unmistakably rounded margins; others are somewhat angulate but have a very short distance from the first vertebral centrum to the apex of the angle, so that the angle itself is never obvious. Supraoccipital crest. In T. alalunga (fig. 1) the supraoccipital crest is relatively more slender than in any of the other species of Thunnus and is longer, nearly always reaching at least to the centrum of the third vertebra. In the other six species the crest rarely extends beyond the second vertebra. Ventral parasphenoid shaft. In T. atlanticus the anterior portion of the parasphenoid shaft is concave ventrally (de Sylva, 1955). In T. tonggol we found it concave in three small specimens (skull length 57-99 mm.) and flat in two larger specimens (skull length 122-128 mm.). It is most commonly flat or slightly convex in the other species, but a degree of concavity has been observed in individuals, especially young, of all except T. obesus. Other characters. — Godsil and Byers (1944) cited several additional neurocranial characters that are supposedly useful in distinguishing among the species. In our estimation, none of these is valid for the following reasons. The parietal crest in T. albacares was described as extending farther forward than in T. obesus so that a projection of the curvature of the lateral edge of the parethmoid (prefrontal of Godsil and Byers) would he continuous with the parietal crest in T. obesus but woidd run below it in T. albacares. Our material shows both conditions in all species. The angle of the long axis of the basisphenoid relative to the parasphenoid is highly variable and not reliable as a specific character. The width of the basisphenoid relative to its height is not only variable within any given size range but also changes with growth. The anteriorly directed process at the upper end of the basisphenoid was used by de Sylva as a distin- guishing character (1955: 32-35). He described the process in T. albacares and T. "sibi" (Pacific T. obesus) as being directed obliquely ventrad so that a line drawn through its axis would transect the parasphenoid at or behind the junction with the parethmoid; in T. thynnus, T. alalunga, and T. atlanticus such a line would more nearly parallel the parasphenoid and would not cross it. We find this character variable within a species. Furthermore, in larger fishes, the entire bone becomes relatively shorter and wider, whereas, the process becomes broader and more rounded. The head of the vomer in T. alalunga was described as having a thin bony ridge behind the dentigerous anterior portion, a similar ridge being present in some T. thynnus, but not in T. albacares or T. obesus. Actually, all the species may have such a ridge. In T. alalunga teeth are generally restricted to the anterior end; the posterior end is very thin. The other species usually, but not always, bear teeth along the entire ridge, and the posterior portion is wider. In T. atlanticus the ridge is usually absent. ANATOMY AND SYSTEMATICS OF TUNAS 73 These tendencies exist, but frequent exceptions ren- der the character uncertain. A depression in the dorsal profile just anterior to the supraoccipital, reported to be present in all species but most pronounced in T. thtjnnus, is related to the pineal foramen. We find this variable in all species and distinctive in none. In the contour of the posterior margin of the sphenotic as seen in ventral view, we can detect no difference among the species. A concave curvature of the margin, held by Godsil and Bj'ers to be characteristic of T. alalunga and T. thynnus, is not only slight but may be present or absent in all species. Branchiocranium General description. — The branchiocranium in- cludes the branchial bones, opercular apparatus, jaws, and associated bones. On each side the den- tigerous premaxilla forms the upper jaw, and the maxilla is located dorsomesial to it. A small sup- lamaxilla is attached to the posterior end of each maxilla. The lower jaw includes the dentary, which bears teeth; the articular, forming the rear end of the jaw and articulating with the condylar region of the quadrate; and a small angular at the poster- oventral corner of the articular. The suspensorium begins with the hyomandibula, which articulates at its upper end with the otic region of the neurocranium and with the opercle. The ventral limb of the hyomandibula articulates with the metapterygoid, and the ventral portion of the latter in turn articu- lates with the symplectic and quadrate. To the anteroventral part of the metapterygoid are joined the ba.'ial portions of the endopterygoid and ectop- terygoid. At their anterior ends, these are joined to the short, dentigerous palatine, which articulates with the condyle of the anterior end of the maxilla. In addition to the hyomandibula, the hyoid arch is composed of two median and four paired bones. A glossohyal supports the tongue, and a urohyal lies below and between the two sides of the arch. The paired bones include small basihyals, large cera- tohyals that articulate with smaller epihyals by jagged, toothlike sutures on the mesial side only, and small interhyals posteriorly joining the operculum. In the branchial arches are three median basi- branchials (a small cartilage posterior to the third may represent a fourth ba.sibranchial, see Iwai and Nakamura, 19r)4a) and on each side three hypo- branchials, five ceratobranchials, four epibranchials, and four pharyngobranchials. The posteriormost ceratobranchials and mesial three pairs of phar3'ngo- branciiials bear villiform teeth. The anterior sur- faces of the first four arches bear gill rakers, and, as supports for the gill filaments, so-called gill bars are found on the posterior surfaces (Iwai and Xakamura, 19G4a). Specific Characters. — Differences worthy of note have been described for only six bones by Kishinouyc (1923) or Godsil and Byers (1944). These are the hyomandibula, metapterygoid, quadrate, subopercle, interopcrcle, and ceratohyal. We concur in their observations on the first four only. The anterior (sphenotic) articulating head of the hyomandibula (fig. (!) is relatively longer and nar- FiornE 6. — Hyomriiidibula of (left) Thunnus thynnus, skull length 130 mm., (right) T. alalnruja, skull length 150 mm. Measurements of anterior articular head include A — least width, H— length. rower in T. alalunga than in the other species. The proportion of length to least width in our specimens ranged as follows: T. alalunga {N = 3o) 1.7-2.7, x='2.2; T. thynnus (iV=44) 1.3-2.1, x=1.7; T. mac- coyii (iV=17) 1.6-2.3, x=1.9; T. obesus (.V=36) 1.3-1.9, x= 1 .5; T. atlanticus {N= 18) 1.3-2.2, x= 1.7; T. albacares (.V=58) 1.2-1.9, x=1.6; T. tonggol (.V=4) 1.4-1.8, x= 1.6. These proportions are close to those given by Godsil and Byers (1944: 86), who reported 1.7-3.0 for Pacific T. alalunga and 1.2-1.5 for Pacific T. albacares, and stated that Pacific T. thynnus and T. obesus are similar to T. albacares. 74 U.S. FISH .\N'D WILDLIFE SERVICE Kishinouye (1923: 322) stated that the anterior head is "more or less roundish in cross-section in Thunnus; but more or less flattened in Parathiinnus and Neothunnus. . . ." We can find no significant dif- ference among any of the species in this character. Furthermore, we cannot confirm de Sylva's conten- tion (1955:14) that the process is oblicjue to the vertical limb in T. atlanticus but forms a right angle in the other species. In T. alalunga the metapterygoid (fig. 7) is nar- rower than in other species. This condition can be indicated by the proportion of the length of the anteroventral margin to the posteroventral margin Figure 8. — Quadrate of (left) Thunnus thynnus, (right) T. alalunga, same specimens as in fig. 6, showing measure- ments of E — total width, F — width of horizontal edge, G — length. Figure 7. — Metapterygoid of (left) Thunnus thynnus, (right) T. alalunga, .same specimens as in fig. 6, showing measure- ments of C — anteroventral margin, D — posteroventral margin. (measured in each case to the midpoint of the arc of the posteroventral edge.) In our material, this pro- portion is as follows: T. alalunga {N=32) 1.1-1.8, x=lA; T. thynnus (A^=43) 1.6-2.6, x=2.0; T. maccoyii {N=17) 1.4-2.7, x= 1.9; T. obesm {N=37) 1.5-3.1, x=2.0; T. atlanticus {N= 19) 1.4-2.1, x= 1.7; T. albacarcs {N=58) 1.5-3.1, x=2.1; T. tonggol {N=4) 1 .8-2.2, x= 2.0. These proportions are simi- lar to those given by Godsil and Byers (1944: 86) for Pacific tunas, but provide even better distinction. Godsil and Byers measured each margin to the "most ventral point," which seemed more nebulous to us than the midpoint of the arc. Their figures of 1.0-1.5 for T. alalunga and 1.3-1.9 for T. albacarcs and the other species include slightly lower propor- tions than were found in our specimens, but the conclusions are nevertheless similar. Again, in T. alalunga the quadrate (fig. 8) is slightly narrower than in the other species. The proportion of length to total width in our specimens is as follows: T. alalunga (A^=35) 1.5-2.1, 5=1.8; T. thynnus (.V=44) 1.2-1.6, x=lA; T. maccoyii (iV=16) 1.3-1.5, x=1.4; T. obesus (iV=37) 1.4-1.7, x=1.5; T. atlanticus {N=20) 1.3-1.8, 5=1.6; T. albacarcs {N=57) 1.4-1.8, 5=1.5; T. tonggol (iV=4) 1.4-1.6, x= 1.5. The proportions of the same length to width of the horizontal dorsal edge are : T. alalunga (A^ = 35) 2.1-2.7, 5=2.5; T. thynnus (Af=44) 1.6-2.1, 5=1.9; T. maccoyii (iV=16) 1.6-2.0, 5=1.9; T. obesus (N=37) 1.8-2.2, 5=2.0; T. atlanticus (N=20) 2.0-2.3, 5=2.1; r. albacarcs (A^=57) 1.8-2.2, 5=2.0; T. tonggol (A^=4) 1.8-2.2, 5=2.0. We are not certain where Godsil and Byers (1944: 87) measured the width, but they gave proportions of 1.8-2.3 for Pacific T. alalunga and 1.6-1.8 for other species. Thus there is agreement in order of magnitude, but their proportions are generally higher than our first and lower than our second. The subopercle of T. thynnus and T. maccoyii (fig. 9) differs, with few exceptions, from that of the other species in being relatively narrow and in having the anterodorsal margin almost vertical in its lower two-fifths to one-half, followed by a decided change in slope of the upper portion. In the other species there may be a very short perpendicular portion, less than one-fifth of the length, before the oblique slope begins, or, most often, there is an almost straight or very slightly convex oblique edge. This finding conforms with the observations of Godsil and Byers (1944: 101) for the Pacific forms and presumably also with the observations of Kishinouye (1923 : 325), although his statement is less clear. ANATOMY AND SYSTEMATICS OF TUNAS 75 Fku're 9. — Subopercle of (left) Thurmus thyrin}i.i, (right) T. (ilidunija. Same specimens as in fig. (i. We are unable to confirm two other characters mentioned by Kishinouye (1923: 325, 327). In Japanese T. thi/nnus the posterior margin of the interopercle was described as being convex, whereas in other Japanese tunas it is nearly straight. In our material the shape of the mai-gin is variaV)le; most of the species have both types and variants thereof. Kishinouye also described a groove for blood vessels along the dorsolateral edge of the ceratohyal. The groove was present in T. thi/nnus, T. alalunga, and T. obesus but hardly visible in T. albacarrs. This grove is present in all species and is generally more apparent in larger specimens. In small specimens of most species the groove is indistinct or absent. Pectoral and Pelvic Girdles The pectoral girdle is composed of a series of bones connecting the skull and the pectoral fin. The two- armed supratemporal, not really a functional part of the pectoral girdle, is closely applied to the skin beside the otic region of the neuroeranium. A larger, also two-armed, posttemporal articulates with the skull, followed by a supracleithrum and the large, curved, bladelike cleithrum. Baudelot's ligament runs from the supracleithrum to the basioccipital. From the posterior margin of the supracleithrum extend two flattened postcleithra, the second of which has an attenuated posteriorly directed process. The long curved blade of the cleithrum forms a thin- walled trough that opens posteriorly. The thick- ened, somewhat rectangular scapula is borne dorsallv, on the mesial side. Below the scapula, the blade- like coracoid is attached. The lower, posterior end of the scapula and the uppermost posterior edge of the coracoid are thickened and flattened; they form articular surfaces for four (iterygials, on which the fin rays are borne. The pelvic girdle includes a pair of winglike basi- pterygia that join posteriorly in the median line. Anteriorly each bone has a flattened lateral wing and a long narrow mesial process. Posteriorly a long mesial process extends between the fin rays. There are no pterygials. No differentiating characters are apparent in cither the pectoral or the pelvic girdle. Vertebral Column General descriplion. — Important papers describing the vertebral column in species of Thunnus include: Starks (1910), Kishinouye (1923), Frade (1932), Godsil and Byers (1944), Godsil and Holmberg (1950), Clothier (1950), de Sylva (1955), and Nakamura (1965). The vertebral column usually has 39 vertebrae, including the hypural plate. The first vertebra is articulated firmly with the rear of the skull. Neural arches and spines are present on all except tlio hypural plate. The spines are erect and laterally flattened on the first six vertebrae. On the seven vertebrae anterior to the hypural plate, both the neural and haemal spines arc wide and depressed and lie on top of the next posterior centrum, forming a strong and rigid tail section. Laterally directed transverse processes (parapophyses) appear as small projections on the third vertebra, become longest and broadest on about the sixth, shorter and more canted on the next two or so, and usually become both longer and ventrally directed on the eighth or ninth. By the 10th or 11th vertebra, the first closed haemal arch is formed by the meeting of the distal ends of the parapophyses. The ventral ends of the haemal arches become progressively longer, forming haemal spines. Ribs are attached, begin- ning with the 3d vertebra, to each parapophysis or to the end of each haemal arch or spine until the 18th or 19th vertebra. Posteriorly, haemal spines are present, but ribs are absent. Dorsal to the ribs, intermuscular bones (epipleurals) articulate either on the neural arch or the centrum of each of the \ertebrae from the 1st to the 31st. Beginning at the 12th to 18th vertebra, each haemal arch bears on each side a process directed obliquely ventrad which has been called a haemal prez.ygapophysis (de Sylva, 1955). On suc'ceeding vertebrae this process is longer, then shorter and re U.S. FI.SH .\ND WILDLIFE SERVICE more dorsally situated ; by the 20th to 25th vertebra it comes to arise from the anterior end of the centrum rather tlian from the haemal arch. Beginning at about the fourth vertebra, a process which has been called a haemal postzygapophysis (de Sylva, 1955) arises on each side from the poster- ior end of each centrum. On the anterior centra it is small and laterally directed. This process becomes ventrally directed on the eighth vertebra, and its distal end meets the parapophysis ; farther poster- iorly it meets the upper part of the haemal arch or the haemal prezygapophysis when it is formed. On approximately the last eight vertebrae the haemal spines are situated so posteriorly that they obliterate the haemal postzygapophyses. From some or all of the 20th to the 33d vertebrae, the blood vessels and nerves that emerge from the haemal canal exit through ventrolateral foramina. The anterior foramina are formed by struts running from the haemal arch to the centrum near the base of the haemal postzygapophyses. They become smaller posteriorly and are separate from and ante- rior to the haemal arches; the latter in this region gradually become located toward the posterior end of the centra. On the 32d to 36th vertebrae, flat- tened lateral processes form a horizontal bony keel. The sizes of the individual vertebrae vary consider- ably, and regional differences are emphasized in older specimens. The length increases regularly to the 35th vertebra; the 36th is slightly shorter, the 37th and 38th are very short, and the 39th is incorporated into the wide, triangular hypural plate. The depth of the vertebra increases regularly to about the 25th, beyond which there is a gradual decrease to the hypural plate. A simple splintlike hone (epural) is closely applied to the anterodorsal surface of the hypural plate. A similar bone (hypural) bearing a spinous process on each side is present along the anteroventral surface of the hypural plate. The terminology and derivation of these two bones are in doubt. Specific Characters. — The vertebrae typically total 39 in all species. Godsil and Byers (1944) reported, and we have reexamined, a California T. thyniius with only 38, in which 1 vertebra near the hypural is obviously missing. Among more than 200 skele- tons of the seven species, we found only three additional abnormalities, all due to recognizable fusion of two adjacent centra. Frade (1932) re- ported, among 110 T. thynnus, 8 with 38 vertebrae, 6 with 40, and 1 with 41. We doubt the counts of 40 and 41 but cannot explain them. All but one species have 18 precaudals and 21 caudals, the first long haemal spine occurring on the 19th vertebra. The same count was given by de Sylva (1955) for T. atla)}ticus, but, as Watson (1964) has shown, this species differs from all other Thunnus in having 19 precaudals and 20 caudals. Exceptions may be expected in all species; we have examined T. atlan- ticus with counts of 18-|-21 and 20+19, T. obesus and T. albacarcs with 17 + 22, and T. thynnus and T. obesus with 19 + 20; and Godsil and Byers (1944: 86) reported one T. alalunga with 20+19. The position of the first (anteriormost) ventrally directed parapophyses appears to show almost no variability within a species or subspecies; only one exception has been noted. These parapophyses occur on the 8th vertebra in T. thynnus, on the 10th in T. tonggol, and on the 9th in the other species. In T. alalunga none of the parapophyses is directed quite so obviously ventrad as in the other species; those on the ninth vertebra that we regarded as ventrally directed are much shorter than in any other species and seem almost twisted, never becom- ing completely ventrally oriented. In other species, there is variation in the ventral extent of the preced- ing parapophyses. As long as these were more or less flattened and rounded, their relative location was not considered. The first ventrally directed ones are definitely elongated in a ventral direction (compare the eighth vertebra in T. thynnus and T. niaccoyii, fig. 10). The first (closed) haemal arch usually occurs on the 11th vertebra in T. albacarcs. T. atlanticus, T. tonggol, and T. obesus, and usually on the 10th vertebra in T. alalunga, T. maccoyii, and T. thynnus. In all species except T. alalunga and T. maccoyii we observed the first closed arch occasionally either one vertebra anterior or one vertebra posterior to the usual position. Godsil and Byers (1944: 68, 101) observed notable variation in Pacific T. albacarcs and in T. thynnus. In most of the species the parapophyses on the vertebra preceding the one that bears the first haemal arch approach each other so closely in the median line that it appears to be a matter of chance whether or not they or the next pair fuse. In many specimens in which the haemal arch was formed anterior to its usual position, its shape was noticeably different (fig. 10, OBE). In T. alalunga the first haemal arch is directed forward, Avith an angle of about 45° or less between it and the vertebral axis. In all of the other species ANATOMY AND SYSTEMATICS OF TUNAS 77 ALA THY OBE ALB ATL TON TON FiGiRE 10. — Anterior view, first closed hiipiual arch and pre- ceding vertebrae of Thinwus species. ALA — vertebrae 0- 10, ALR— 9-11, ATL— 9-11, NL\C— 8-10, OBI<:-Ml (the one slightly below the main row shows shape of first closed haemal arch when located on 10th vertebrae), THY— 8-10, TON— 9-11. FiGi'RE 11. — Left lateral view of vertebrae of Thmiiius species .showing first ventrally directed parapophyses (left verte- bra) and first clo.sed haemal arch (right vertebra). ALA — vertebrae 9-10, ALB— 9-11, ATL— 9-11, OBE— 9-11, THY— 8-10, TON— 10-12. 78 U.S. FISH AND WILDLIFE SKKVICK ALA THY OBE ALB ATL TON Figure 12. — Left lateral view of vertebrae of Thunnus species, showing development of anteriorniost haemal pre- and post- zygapophyses. ALA— vertebrae 13-16, ALB— 14-16, ATL— 14-18, OBE— 14-17, THY-14-16, andTOX-14-17. it may range from almost perpendicular to a 60° angle (fig. 11). The shape of the first haemal arch and the dimensions of its bony parts vary consider- ably, but in T. atlanticus, T. tonggol, and T. maccoyii the bony portions are thinner and the sides more bowed than in the other species (fig. 10). As shown by Yabe et al. (1958), by Matsumoto (1963), and by Yoshida (1965), the heamal spine of the first caudal vertebra is always laterally flattened and winglike in T. alalunga but in the other species resembles the other haemal spines and is not flattened. The length of the haemal prezygapophyses and the distance of their origin from the centrum vary among the species of Thunnus. In T. alalunga the haemal prezygapophyses all originate at the centrum or extremely close to it. Correlated with this, the anterior haemal postzygapophyses are relatively short (fig. 12). In the other species the anterior haemal prezygapophyses arise from the sides of the haemal arches of 3 to 12 vertebrae before they begin to arise from the centra, and the posterior haemal postzygapophyses are relatively longer. In T. ala- lunga, T. ohesus, T. nmccoyii, and T. thynnus the haemal prezygapophyses arise high on the neural arch, so that only the first two or three at most can be regarded as clearly on the arch. Correspondingly, the associated haemal postzygapophyses hardly dif- fer in length from those on the posterior vertebrae. By contrast, in T. atlanticus, T. albacares, and T. tonggol the haemal prezygapophyses arise far more ventrad on the haemal arch, from one-fourth to one-half the distance to the ventral tip, and there is no question that at least five (usually more) are definitely on the arch, not on the centrum. The associated haemal postzygapophyses in these species are longer than in the other three, although less so in T. albacares than in T. atlanticus and T. tonggol. In T. atlanticus and T. tonggol the longest haemal postzygapophyses are equal to or longer than the length of the centra; in T. albacares they may be about three-fourths the centrum length. The species differ in the development of the ventro- lateral foramina that are found on some or all of the 20th to the 33d vertebrae (fig. 13). These foramina are best developed anteriorly and in this region appear to arise through the formation of a bony strut from the haemal postzygapophyses to the dorsal part of the haemal arch. They diminish in size posteriorly and are absent on the last several ver- tebrae. In T. atlanticus, T. albacares, and T. tonggol the anterior openings are large, longer than wide, ANATOMY AND SYSTEMATICS OF TUNAS 79 ALA THY OBE ALB ATL TON Figure 13. — Left lateral view of vertebrae 20-28 of Thurinus species, showing development of inferior foramina. the largest usually tliree times or more as long as the horizontal width of the base of the corresponding haemal spine. In T. ohesus, T. maccoyii, T. thynnns, and T. alalunga the size of the openings is variable, usuallj' small, and the largest opening is rarely more than aliout 1.5 times as long as the width of the adjoining bony neural arch. T. atlandcus, T. albacarcs, and T. tonggol are distinguishable by the development of vertebral processes and openings, which approach the ornate trelliswork seen in Auxis, Euthynnus, and Katsu- xvonus; T. alalunga shows the least development, and T. ohesus, T. maccoyii, and T. thynnus are inter- mediate. The first haemal prezygapophysis tends to be found more anteriorly in T. albacarcs and T. tonggol, most commonly on vertebra 13 or 14 (range 12-15) in T. albacarcs. In T. maccoyii it usually occurs on 14 or 15 (range 14-16), in T. alalunga and T. obesus on 15 or 16 (range 14-17), and in T. allanticus on 16 or 17 (range 15-18). T. thynnus displays a wider range of variation (12-17), and overlap is considerable between it and the other species. While this paper was in press, Nakaniura and Kikawa (1966) described specific differences in the infra-central grooves on the ventral side of the centra of vertebrae 10-30. These infra-central grooves were categorized into three types: type A with two separate infra-central grooves; type B with two grooves connected by a canal; type C witii a single, usually elongate, groove. We have reexamined our skeletal material, which comprises many more speci- mens than the total of 19 used by Nakamura and Kikawa, and we conclude that the infra-central grooves are a useful character, but show variation that often precludes positive species identification. Our conclusions are as follows: Thunnus albacai-es (24 specimens, fork length 280- 1,430 mm.) has type C infra-central grooves th;it tend to be divided by a thin septum anteriorly but that posteriorly are undivided or occupied by a honeycomb-like network of bony material. In this character, T. albacarcs is distinctive, but is ap- proached by some specimens of T. maccoyii. Thunnus maccoyii (16 specimens, 742-1,445 mm.) is the most variable of all the species. Very few have type C (undivided) grooves, as described by Nakamura and Kikawa. Typically, all the verte- brae have two grooves (type A) that are divided by very thin septa or a honeycomb of septa. This condition resembles that of the anterior grooves of T. albacarcs. Only an occasional undi\ide(l groove occurs among the divided ones in T. maccoyii, whereas most of the grooves are undivided in T. albacarcs. Some specimens, particularly tiie si.\ from Australia that were used by Godsil and Holmbcrg (1950), were almost impossible to distinguish from T. thynnus. Thunnus thynnus thynnus (7 specimens, .'U()-2,315 mm.) and T. t. orientalis (18, 530-1,450 mm.) have type A grooves, two grooves per centrum, that tend to be round or oval anteriorly and become longer and narrower posteriorly. The distance between the two grooves on the anterior centra is highly variable. When the grooves are rtarrowly separated, the specimens resemble some T. maccoyii. 80 U.S. FISH AND WII.DI.IKK SKRVICE Thunnus obesus (7 Atlantic specimens, 697-1,360 mm.; 3 Indian Ocean, 630-1,270 mm.; 1 eastern Pacific, 1,600 mm.) liave type A grooves and could not be distinguished from T. thy?mus. Thunnus ahdunga (20 specimens, 520-1250 mm.), like T. thi/nnus and T. obesus, have type A grooves, but, although the grooves vary greatly in width, they tend to be much narrower than in the latter two species. Usually the grooves are widely sepa- rated, but in a few specimens, the partitions are narrow. Thunnus tonggol (4 specimens, 373-924 mm.) have type B grooves, two grooves per centrum, quite variable in size, with a shallow canal connecting them. The grooves tend to be larger than those of T. atlnnticus and smaller than those of T. thynnus or T. obesus. In the two larger specimens, the anter- ior grooves strongly resemble those of T. thynnus or T. obesus in being larger and close together. Thunnus atlanticus (20 specimens, 322-665 mm.) have type B grooves, with two very small pits on each centrum connected by a very narrow canal. The canal is not evident in some specimens, giving the impression that the two grooves are separated by a sharp ridge. In this respect T. atlanticus is distinctive. The largest specimens, however, closely resemble small T. tonggol, with small grooves con- nected by an obvious canal, the grooves becoming elongate posteriorly. The proportion of length to depth of the 36th vertebra is relatively greatest in T. albacares (.¥=60) 1.2-1.9, x=1.4 and in T. alalunga (jV=40) 1.1-1.7, f=1.5. In T. alalunga the vertebrae are often considerably smaller anteriorly than posteriorly (fig. 14), accounting for the high ratio. In the other species, this proportion is: T. Ihyyinus (N = 58) 0.8-1.3, x=l.l;T. maccoyii {N= 16) 0.8-1.3, x= l.O; T. obesus (A^=30) 0.9-1.4, x=1.2; T. atlanticus {N=23) 1.0-1.4, f=1.3; T. tonggol {N=8) 1.1-1.5, 5=1.3. VISCERA The relative position, shape, and size of the various internal organs provide excellent diagnostic char- acters (fig. 15). These organs are treated here by systems. Important works on the viscera of Thunnus include those of Eschricht and Miiller (1837), Kishinouye (1923), Frade (1925), Serventy (1942), Godsil and Byers (1944), and Godsil and Holmberg (1950). Figure 14. — Left lateral view of vertebrae 35-39, showing differences in proportions of vertebra 36. (top) Thunnus albacares, (middle) T. alalunga, (bottom) T. obesus, typical of other species not illustrated. Digestive System and Associated Organs General Description. — At the anterior end of the body cavity the liver abuts against the transverse septum and caps the other organs. It is usually composed of three lobes, only the middle of which is plainly visible in ventral view; the other two lobes lie along the lateral body wall, hidden by the other organs. The ventral surface of the liver of some species appears striated owing to the parallel arrangement of blood vessels (both arterial and venous) near its surface. The species with striated livers also possess, on the dorsal surface of the liver, several large "vascular cones," each comprising numerous vessels bound in a common sheath. These are absent in species with unstriated livers. In all species (Morice, 1953, to the contrary) there are two efferent (venous) vessels leading directly from the anterior surface of the liver into the sinus venosus. The esophagus merges indistinguishably (in external view) into the stomach, which forms a blind sac posteriorly. The intestine rises from the anterior AN.VTOMY .\ND SYSTEMATICS OF TUN.\S 81 FT m URINARY BLADDER SWIMBLADDER GONAD LIVER CAECAL MASS SPLEEN STOMACH INTESTINE GALL BLADDER BODY WALL Figure 15. — Ventral view of in situ visceral patterns of Thunnus species. Arranged from left to right in order of increiusing fork length. ALA— 780, 875, 1,030 mm.; ALB— 637, 700, 808, 1,340, 82 U.S. FISH AND WILDLIFE SERVICE 1,420 mm.; ATL— 337, 438, 513, 563, 650 mm.; OBE— 1,250, 1,310, 13,50 1,450, 1,540 mm.; THY— 457, 750, 872, 1,050, 1,670 mm.; TON— 373, 392, 910, 924 mm. ANATOMY AND SYSTEMATICS OF TUNAS 83 pnd of the stomach, and a very large caecal mass is attached to its origin by several ducts that are not externally apparent. The intestine proceeds caudad for half or more the length of the body cavity (straight intestine), forms a loop, runs craniad (ascending portion) almost to the pylorus, then forms another loop and continues in a nearly straight line (descending portion) to the anus. The spleen is located between the straight and ascending portions. The gall bladder is a long, tubular sac rising from the riglit lobe of the liver, attached to the dorsal wall or the left side of the straight intestine. A swim- bladder, when present, is situated dorsad to the main visceral mass, and may be rudimentary or well developed. Specific Characters. — The ventral surface of the liver (fig. 16) is striated in T. alalunga, T. maccoyii, T. ihynnus, and T. obesus. These striations give the impression of being denser and extending farther toward the center of the middle lobe in T. alalunga, T. maccoyii, and T. thyiuui^ than in T. obcsus, but this difference is not easy to detect. We have seen only one instance of striations being limited to the peripheral margins, at least of the middle lobe, as described by Kishinouye (1923) and Godsil and Byers (1944) for Pacific, and by Morice (1953) for Atlantic T. obesus. The peripheral nature of these striations has been used as a major diagnostic char- acter of the nominal genus Paralhunnm^. In T. albacares, T. atlanticus, and T. tonggol the liver lacks striations. In T. alalunga, T. maccoyii, T. Ihynnus, and T. obesus the three liver lobes are subequal in length, the lateral lobes most often slightly shorter than the middle lobe. In T. albacares, T. atlanticus, and T. tonggol the right lobe is much longer and narrower than the middle or left lobe. Correlated with the ventral striations on the livers of T. alalunga, T. maccoyii, T. thynnus, and T. obesus, vascular cones are associated dorsally with each lobe. The middle lobe always has a single large cone; the other lobes usually have two or more, and these may be somewhat difficult to distinguish from ordinary l)lood vessels. In the left lobe, we rccortled two to five cones in T. alalunga; two to six, usually two or three, in T. thynnus; two to four in T. maccoyii; one to six, usually one or two, in T. obesus. In the right lobe we found: two to eight, usually two, in T. alalunga; one or two, usually two, in T. thynnus; two in T. maccoyii; one to four, usually one or two, in T. obesus. Figure 16. — Ventral view of livers of Thunnus species, showing shape and presence or absence of striations. Fork lengths (left, right): AL.\— 875, 1,030 mm.; .\LB— 700, 1,175 mm.; .\TL— 568, 650 mm.; OBE— 1,450, 1,.540 mm.; THY— 457, 757 mm.; .ami TON— '.)10, <»2I mm. Morice (1953) described the liver of T. albacares as having two efferent vessels at the anterior end, but noted only a single vessel in T. alalunga, T. thynnus, and T. obesus. He probably overlooked the vessel in the right lobe near the junction with the middle lobe, because this vessel is smaller than the one in the middle lobe. All specimens that we examined had two such efferent vessels. 84 U.S. FISH AND WILDLIFE SERVICE Thunnus alalunga has the spleen on the left side and the stomach on the right. In the other species these positions are reversed (fig. 15). Godsil and Byers (1944) attached considerable importance to their observation that tlie straight intestine in T. alalunga crosses from the right to the left side, and the descending portion lies on the left side. This course is obviously correlated with the position of the stomach in this species. We found that the intestine often does not cross over so obviously and that the descending portion is com- monly near the middle. Thus there is little or no difference between T. alalunga and the other species in position of the descending intestine, although the side on which the intestine originates is different. The relative position of the first loop of the intes- tine (where the straight intestine forms the ascending portion) differs to a degree among the species (fig. 15). In T. alalunga this loop is shortest and is located about one-half to two-thirds the distance between the posterior margin of the middle liver lobe and the anus. In the other six species the loop may reach from about three-fourths to nine-tenths of the liver-anus distance. In Pacific forms, T. alalunga is reported by Godsil and Byers (1944) to have a short "fold" (27-41 percent of body cavity), T. albacares a slightly longer "fold" (36-61 percent), and T. thynnus orientalis and T. obesus a long "fold." Their illustrations and measurements, and our ob- servations, indicate such a wide range of variation that only T. alalunga can be regarded as distinct in this character. One Atlantic specimen of T. obesus among our study material had two intestinal loops, a situation never before reported to our knowledge. The length of the spleen (fig. 15) is normallj^ short, seldom reaching beyond half the distance from caecal mass to end of body cavity in T. alalunga, T. obesus, and T. allanticus, but usually long, reaching at least three-fourths of this distance in T. thynnus, T. maccoyii, and T. albacares; T. tonggol is variable. In all species there are exceptions. The gall bladder (fig. 15) in T. alalunga is normally exposed along the entire right side of the straight intestine. In the other species it is usually either entirely hidden or a small portion may appear posterior to the first intestinal loop; in the few specimens in which it was largely exposed, the visceral mass seemed to be distorted. The swimbladder (fig. 17) appears to be invariable and distinctive only in T. obesus, in which it is long, usually slender, beginning near the transverse sep- tum, and tapering to a point that reaches the poste- rior end of the body cavity. i^\ n OBE ALB Figure 17. — \'entral view of swimbladder shapes of Thunnus species. Fork lengths (left to right): ALA— 875, 1,030 mm.; ALB— 637, 700, 1,17.5, 1,420 mm.; ATL— 650, 664 mm.; OBE— 1,250 mm.; and THY— 954, 1,050, 1,670 mm. ANATOMY AND SYSTEMATICS OF TUNAS 85 Absence of a functional swimbladdcr in T. longgol may be regarded as a useful diagnostic character, but in most specimens we discovered a tiny swim- bladder, about 4 mm. in diameter, which could easily have been overlooked. A very small swim- bladder (ca. 20 mm. long) also was present in two T. maccoyii (742 and 7()4 mm.). Variation in the other five species is considerable (fig. 17). In T. allanticiis the swimbladdcr may be short, oblate, lying far anterior in the body cavity; or it may approach the length of a poorly developed swimbladder of T. albarnrcs. In the latter case, the swimbladder is comprised of two chambers divided by a transverse membrane, with the anterior chamber probably lepresenting the small type of swimbladder, the posterior chamber an addition. In T. albacarcs the swimbladder is moderately long, usually reaching about the level of the 14th or 15 vertebra (range 12-17). In T. alalunga it varies from narrow, only moderately long, and deflated, to fully the length of the body cavity and inflated to fill much of the cavit.v. In T. Ihynnns and T. maccoyii variation in swim- bladder dimensions may be correlated with growth. The changes in swimbladder dimensions were recog- nized by Serventy (1956a: 10), and were suggested l)ut not recognized by Godsil and Holniberg (1950: 21). Smaller specimens (457-954 mm.) have rudi- mentary swimbladders that are slender, deflated, and short (about 6 vertebrae long), about a quarter of the length of the body cavity, and do not reach the depression anterior to the dorsal bulge. In two T. maccoyii the swimbladder was tiny and could easily have been missed in a casual dissection. In a 1,060- mm. specimen of T. I. Ihynnus the organ was inflated and occupied about half the body cavity (fig. 17). In specimens 1,390-mm. and larger the swimbladder extends from the depression anterior to the kidney almost or quite to the posterior end of the body cavity; it is as wide as the body cavity in its anterior half and posteriorly tapers almost to a point. Urogenital System General Description. — The paired gonads are fi(!- quently visible in ventral view. They lie along the dorsolateral body wall, their posterior ends forming ducts which open on each side of the urinary papilla. The kidney is anterior in position and dorsal to the layer of fibrous connective tissue overlying the swim- bladder. Its anterior margin usually follows the edges of the depression anterior to the hump formed by the haemal arches, and lateral extensions reach forward in a semicircle and sometimes nearly meet anteriorly. Depending upon the species, a posterior extension ("tail") may reach about as far as the level of the 16th vertebra; and T. alhacares some- times has accessory masses of kidney ti.ssue posterior to the main mass. In the anterolateral extensions of the kidney, the urinary ducts ("ureters") arise and join witliin or just posterior to the kidney substance, and the resulting single duct proceeds posteriorly. Just before the anus, the duct enters a small but prominent urinary bladder, which may lie within the mesenteries of the gonads or project freely into the body cavity, depending upon the species. The urinary bladder empties through a urinary papilla behind the anus. Specific Characters (fig. 18). — The kidney of T. alnbinga is uni(iue in lacking a "tail," the end reach- ing the level of the 7th to 9th vertebra (1 1th in one specimen). In T. thynnus and T. maccoyii the tail is relatively short, reaching the Sth to 1 Ith vertebra. Its configuration varies in Atlantic specimens from tapering to truncate, encompassing all the forms used by Ciodsil and Holmberg (1950) to differentiate T. maccoyii from Atlantic and Pacific T. Ihynnus. None of our T. maccoyii had truncate kidneys. In T. ohesns the tail is slightly longer than in T. thynnus, reaching the 11th to 14th vertebra, is narrower, and is moi'e distinctly delimited from the anterior kidney mass. The kidney of T. alhacares has a long tail, tapering gradually' from the anterior kidney mass and reaching the 12th to 16th vertebra; accessary kidney masses are often present posteriorly. In T. atlanticus the kidney mass is bulky anteriorly and has a long, narrow tail that reaches the r2th to 17th vertebra. In T. tonggol there is a large anterior portion and a long, narrow tail that reaches the 15th to 17th vertebra. The branching of the ureters varies considerably, but shows some general tendencies that, together with the shape of the kidney, are useful in distin- guishing species (fig. 18). In the tailless kidney of T. alalunga, the two main branches are widely divergent, joining at the posteriormost end of the kidn(>v sulistance. In 7'. Ihynnus and T. maccoyii the junction may occur at the posterior end, at some distance craniad, or just posterior to the kidney. The angle formed liy the branches is small when the junction is well posterior, large \\;hen far anterior. In T. obesus the branches converge to run close together and almost parallel for a considerable dis- 86 U.S. FISH AND WILDI.IKK SKRVICE Figure 18. — Ventral view of kidney and ureter of Thunnus species. Forward extent shown only in ALA and ATL. Fork lengths (left, right); ALA— 875 mm.; ALB— 700, 1,175 mm.; ATL— 533 mm.; OBE— 1,310, 1,350 mm.; THY— 457, 872 mm.; and TON— 910 mm. tance before they join at the posterior end of the kidney. In those specimens in which the junction occurs well forward, the angle between the branches is less than in T. thynnus or T. maccoyii. In T. albacares and T. longgol the branches usually con- verge gradually, at a very slight angle, and join at some distance craniad from the end of the kidney tail. When the junction is almost at the posterior end of the tail, the branches may be almost parallel for a short distance. The junction of the branches in T. atlanticus usually occurs far anteriorly, and the angle between the branches is usually large. Some specimens, however, resemble T. albacares. Our observations on the urinary bladder are few, but they suggest more variation than was implied by Godsil and Byers (1944). The extent to which the bladder is embedded in the dorsal body wall or projects freely into the body cavity seems to be partly a matter of interpretation. In all species, most of the bladder was contained in the membrane between the left and right gonads. Some T. thynnus and T. alalunga had much of the posterior part closely attached to the body wall, but the anterior part was separated from the body wall and contained in the membrane of the gonads. The anterior tip of the bladder actually projected free of the mem- brane in some specimens of T. obesus. In all three of these species, at least some specimens had the bladder, except for its posteriormost end, entirely within the membrane, not attached to the body wall, and without a freely projecting anterior tip; this was the only condition observed in the other four species. Dorsal Connective Tissue Covering most of the dorsal body wall, dorsal to the peritoneimi but ventral to the kidney, is a region of tough, white fibrous connective tissue. In T. alalunga the sheet becomes extremely thick poste- riorly. In T. albacares a thick raised median cord forms in the anterior half. The other species have a rather uniform thin sheet of tissue, perhaps slightly thickened posteriorly. VASCULAR SYSTEM Important papers on the circulatory system of Thunnus include those of Kishinouye (1923), Godsil and Byers (1944), and Godsil and Holmberg (1950). General Description The pericardial cavity is separated from the pleuro- peritoneal cavity by a transverse septum, the walls .\XATOMY AND SYSTEMATICS OF TUNAS 87 of whic-li are formed posteriorly by peritoneum, anteriorly by pericardium and the walls of the sinus \enosus and the ducts of Cuvier, which enter the sinus. No specific differences have been observed in these structures or in the heart itself. After leaving the heart, blood is carried in the ventral aorta, which sends an afferent branchial artery into each of the gill arches. After circulating in the capillaries of the gill lamellae, the blood from each gill arch enters an efferent branchial (epihranch- ial) artery. The anterior two epibranchials of each side unite to form a common trunk, and these trunks join as the "Y" of the aorta, usually beneath the second vertebra, to form the dorsal aorta (fig. 19). The posterior two epibranchials of each side also unite, and their short common trunks join the dorsal aorta, usually beneath the third vertebra. The dorsal aorta continues posteriorly along the dorsal body wall to the first haemal canal. The coeliaco- Khu'RE 19. — Representative piiUiTii.-i of anterior hranclie.s of dorsal aorta in 7'/ihh«».s species. (Left) T. alalunga; (nii(lille) T. (tttanticiit;; (riglit) T. tonggol. Y: Y of aorta. ANT KPI; anterior epil>ranehials. I'OST EIT: posterior epibranchials. C-M: coeliaco-mesenteric. SEG: segmen- tal. L. CUT and R. CUT: Left and right cutaneous. VI'^RT: vertebrae separated by dashed lines. CSOf^'" ,fnirp'',fi'p uuiiunintiimm iiy;jjjjjju//jjjjjj;jjjjj; c KiCiURE 20.— Cutaneous system of arteries (red) and veins (bhiel of Tbunnus ohcsus. (upper) Coursse of cutaneous vessels in superficial musculature. (A) Enlarged transverse section. (H) Enlarged partial view of C, with portions of the arterial walls cut away to show origin of arterioles and venules. (C) Enlarged lateral view of cutaneous vessels. (D) Posterior course of cutaneous vessels; no posterior commissure. From Godsil and Byers, 1944 (fig. 66). 88 U.S. FISH .\N'D WILDLIFE SERVICE mesenteric artery arises from the dorsal aorta beneath the third to fourth vertebra and forms two or three main branches that go to the hver, supply the vascular cones when present, and give off branches to the other visceral organs. The paired cutaneous arteries rise posterior to the origin of the coeliaco- mesenteric artery and course laterad almost or quite perpendicular to the aorta; they penetrate the lateral body musculature, and on each side form two branches beneath the skin that almost parallel one another to the caudal peduncle (fig. 20, 21). At about the level of the 30th vertebra, a posterior commissure may connect the dorsal and ventral branches of each side. Along their length, the lateral branches give off into their interspace arteri- oles, which are so dense that they seem to form a solid sheet penetrating the dark muscle (chiai). Cutaneous veins accompany the arteries; the two parallel lateral branches join anteriorly on each side (fig. 22) to form a large vein that enters the duct of Cuvier, which in turn enters the sinus venosus. If a post-cardinal vein is present it emerges from the first closed haemal arch, runs toward the right side in the kidney mass, and joins the right cutaneous vein. There is usuall.y a cross-connection between the post-cardinal and left cutaneous veins. Specific Characters The coeliaco-mesenteric artery usually has two branches in T. thynnus, T. maccoyii, T. alalunga, and T. atlaniicus, and three branches in T. albacares and T. tonggol. In T. obesus either two or three branches may be present in both Atlantic and Pacific speci- mens. Exceptions were noted in T. albacares and T. adanticus. A connecting branch near the liver between two of the coeliaco-mesenteric branches is present in T. maccoyii, ma}' be present or absent in T. thynnus, / / / muilNlHlltlllH Figure 21. — Cutaneous system of arteries (red) and veins (blue) of Thunnus albacares. (upper) Course of cutaneous vessels in superficial musculature. (A) Enlarged transverse section. (Bj Enlarged partial view of C, to show origin of venules (dorsal) and arterioles (ventral). (C) Enlarged lateral view of cutaneous vessels. (D) Posterior commissure. From Godsil and Byers, 1944 (fig. 31). ANATOMY AND SYSTEMATICS OF TUN.\S 89 KIDNEY URETER Figure 22. — Post-cardinal vein in relation to cutaneous veins and kidney in Thunnus atlanticiis. Also typical of T. obesus, T. albacarex, and T. tonggnl. T. obesus, and T. albacares; and appears to be absent in T. alalunga, T. allanticus, and T. ionggol. Godsil and Byers (1944) implied that this connection is always, or nearly always, present in Pacific T. albacares. Godsil and Holmberg (1950) used its supposed absence in Atlantic T. t. thynnus as one character that differentiates Atlantic from Pacific specimens {T. t. orienlalis), a conchision which our observations do not support. The cutaneous artery usually originates at the level of the tliird or fourth vertebra in T. thynnus, T. maccoyii, and T. alalunga, and at the sixth to eighth vertebra in T. albacares, T. obesus, T. allanti- cus, and T. tonggol. The cutaneous arteries pass laterally between the third and fourth ribs in all T. thynnus, T. maccoyii, and T. alalunga examined by us (also between the second and third according to Godsil and Holmberg, 19o0) ; in T. albacares, T. obesus, T. tonggol, and T. atlanticus they pass between the fifth and sixth ribs, or occasionally between the fourth and fifth. Branching occurs between the fourth and fifth inter- muscular bones in T. thynnus, T. maccoyii, and T. alalunga, and l)ctwcen the sixth and seventh in T. albacares, T. obesus, T. tonggol, and T. atlanticus. Godsil and Holmberg (1950) reported more T. mac- coyii with branching between the fifth and sixtli, and we observed this in one specimen of T. t. thynnus. In a significant number of our T. albacares, as well as two T. obesus, branching occurred between the seventh and eighth intermuscular bones. A posterior commissure is present in T. thynnus, T. maccoyii, T. albacares (fig. 21), T. tonggol, and T. allanticus, but absent in T. alalunga. In T. obesus it is present or absent (fig. 20). In all species except T. albacares and T. maccoyii, we encountered specimens in which we could not ascertain the pres- ence of a commissiue. We noted the number and position of the rows of arterioles and venules arising from the lateral cut- aneous vessels in relatively few si)ecimens of each species (see fig. 23). T. alalunga, T. albacares (fig. 21), T. atlanticus, and T. tonggol, have one row from each vessel; in T. alalunga it originates on the mesial side, and in the other species on the lateral side. In T. thynnus, T. maccoyii, and T. obesus (fig. 20) two rows, one mesial and one lateral, arise from each vessel. Kishinouye (1923) reported two rows of venules and a single row of arterioles in Japanese T. thynnus orientalis. California specimens have two rows of arterioles (Godsil and Byers, 1944, and our observations). We feel certain that Kishinouye either was mistaken or relied on an unusual specimen. In T. albacares large parallel trunks connect the posterior epibranchial and the cutaneous arteries on OBE ALB ALA Figure 23. — Rclation.'ihips of arterioles and venules to cuta- neous artery (light) and vein (dark). Three patterns represented by T. obesus (OBE), T. albacares (ALU), and T. alalunga (ALA). Lateral view. After Kishinouye (1923). 90 U.S. FISH AND WILDLIFE SERVICE each side ; they are absent in the other species. A post-cardinal vein joins the right cutaneous vein in T. albacares, T. obesiis, T. atlanticus, and T. tonggol; it is absent in T. thijnn us, T. maccoyii, and T. alalunga. OLFACTORY ORGAN As this manuscript was being completed, Iwai and Nakamura (1964b) described their use of the olfac- tory rosette to distinguish species of Thunnus. In each nasal cavity beneath the anterior naris is an olfactory rosette, consisting of numerous laminae arranged radially around a central axis. According to Iwai and Nakamura, T. alalunga is unique in having a pair of fleshy labia surrounding the short laminae. The laminae in T. obesus are described as smooth, greatly expanded distally with adipose tissue, and often partly fused distally. In T. thynnus the laminae in small specimens are smooth and of uni- form thickness to the distal edge, whereas in larger specimens the laminae are smooth and distally ex- panded, but with little evident adipose tissue. Thunnus maccoyii is differentiated from T. thynnus by the presence, in some specimens, of slight fringing in the distal ends of the laminae. In T. albacares and T. tonggol the laminae are entirely fringed on their distal edges; and in large T. albacares the basal half of the rosette is densely spotted with pigment; howe\'er, Iwai and Xakamura admit that tlie rosettes of T. tonggol and small T. albacares resemble each other, and, from their figure 3, also resemble those of T. maccoyii. On the basis of admittedly insufficient material, Iwai, Nakamura, and Matsubara (1965) indicate that the nasal rosettes of T. atlanticus closely resemble those of T. tonggol except that the laminae of the former tend to ha\'e the distal ends folded inward. We have examined very few nasal rosettes, but oiu- observations indicate a need for more careful scrutiny of this character, with respect to normal variation and growth changes, before it is used widely. A specimen of T. alalunga and two of T. obesus agreed with Iwai and Nakamura (1964b), but in one T. obesus the laminae were lightly pigmented and had short, flat fimbriae along their entire length. Two specimens of T. albacares were distinctive in the abundance of pigment in the laminae, but while in one the fimbriae were \'ery pronounced and finger- like, in the other they were less developed and flatter. Rosettes of one T. allanlicus and one T. tonggol were virtually identical and were similar to those of T. albacares, with dense, short, flat fimbriae on the laminae, but had less abundant pigment; folding of the distal ends of the laminae was not apparent in either species. In a larger specimen of T. thynnus from Cape Town, South Africa, the laminae were almost uniform in thickness to the distal edge and bore prominent flattened fimbriations along most of their length. These observations are enough at variance with those of Iwai and Nakamura that they clearly show the need for further study. MERISTIC CHARACTERS The species of Thunnus are essentially identical in the number of fin rays (table 1). The number of gill rakers is the only meristic character that we have found valuable in separating species of Thunnus (table 2). Species Differences. — T. atlanticus has fewer gill rakers (19-25) than any other species of Thunnus in the Atlantic, and T. tonggol has fewer (19-26, rarely to 28) than any other Thunnus in the Indo-Pacific (table 2). T. thynnus and T. maccoyii have the greatest number of gill rakers in the genus (31-43). The other three species fall between these two groups with a combined range 23-35. The overlap be- tween species with low, medium, and high numbers of gill rakers is very slight. The T. thynnus- maccoyii complex shows differences of some magni- T.^BLE 1. — Range of variation in fin-ray counts in the specie.s of Thunnus [Based on original and published data] Fin Species T. alalunga T. albacares T. atlanticus T. oliesus T. thunnus T. tonggol Numher 12-14 13-16 7-9 21-24 13-15 7-9 20-23 31-36 Number 12-14 13-16 8-10 22-24 12-15 7-10 21-23 33-36 Number 12-14 12-15 7-9 20-23 11-15 6-8 19-22 31-35 Nu inber 13-14 14-16 8-10 22-24 11-15 7-UI 21-23 31-35 Nu ml}er 12-14 13-15 S-10 22-24 13-16 7-9 21-24 30-36 Number 11-14 Second dorsal rays -- -- -- 14 9 Total second dorsal rays .. 23 14 Ana] finlets 8 21 23 Pectoral rays _ . 30-35 ANATOMY AND SYSTEMATICS OF TUNAS 91 Number of rakers 19.. 20.. 21.. 22.. 23.. 24-. 2S. 2f).. 27.. 2J!.. 29. 30. 31- Numberof flsh- Average. Sources 25.. 26.. 27.. 28.. 29.. 30.. 31.. 32.. 33.. 34.. 35- Numbcr of fish. Average... Sources 31.. 32.. 33.. 34.. 35.. 36.. 37.. 38.. 39.. 40.. 41.. 42.. 43.. Number of flsti. Average.. Sources Table 2. — Total ininiliiT of siU rakers on the fir.'it arch in the species of Thunmts T. longgot West Indian- Red Sea Number 25.1 fi, 15, 17 SE Asia Number 21.7 6 Australia Number 225 22. n fi, 13, 25 T. ttllantkus West Atlantic Number 120 21.9 2, fi, 12, 17 T. obesus West .\tl3ntic Number 27.3 6, 12 East Atlantic Number Central-West Pacific Number 1 5 27 159 147 70 23 14 2 448 2fi. ft 3, 5, .30 East Pacific Number 1 3 26 28 17 12 87 27.1 6, 6, 8, 30 T. alalunga West -Atlantic 55 28.3 fi, 12 East Atlantic 1.58 28.6 fi, 10. 11 Indian Ocean 42 28.8 14, 28, 29 Central-West Pacific fi 45 142 17fi 06 15 1 481 28.8 5. 8, 29, 30 East Pacific fi4 28.6 5, 8, 30 T. albaeares West Atlantic 127 29.8 fi. 12, 17 East Atlantic 2 3 11 51 88 12fi 80 23 7 1 392 30.8 fi, 11, 23 West Indian Ocean 171 29.5 fi, 19, 28 Central- West Pacific 1 3 24 ■3 194 242 202 93 21 2 855 30.0 5, 8, 21, 22,30 East Piwific 2 8 20 50 61 24 161 30.5 .5, 8, 20, .30 T. maecoyii South Africa 13 33.5 26 Australia 331 33.7 1, 5, fi, 9, 24 T. t. orienlalis Central-West Pacific 13 3fi. 4 5, Ea.st Pacific 45 35.8 fi, 9, 12 T. t. Ihynnus West Atlantic 38.8 6. 7, 9, 12 East Atlantic- South Africa 4 8 40 88 105 74 3« 7 4 3fi6 38.9 , 18, 26, 27 — Gibba and Collette. original data; Source : 1_ Abe 19.5.'-): 2— Beebe and Tee-Van. 1936; 3— Brock. 1949; 4— Crane, 1936; 5— Dung and Royc_e 1953; ., , r.^,.. ,„ M„,i,prinR4- 7-C,f„8burK 19.^3; 8-Oodsil and Hyers. 1944; 9-Godsil and Holmberg 19.50; ip---I,etaconn„ux, 19ol; >1-Marrhal, 19.^9 '2^^a h^J. \f^: l-t_Miinrn l9-)7- 14— Postel et al I960; 15— Ranade, 1961 ; 16— Rivas. l«o4b; 17— Rivas, 1961 . 18- Kobins. lya; . i.) no.vti, i.'".) lt.sJhaefer 948- 21-Schaefer 19.52° 22-.'- 7400 4- thil Hints _ 1,600 4- 7 At less than 500 mm., only T. thynnus is distinct, with pectorals 21 percent of fork length or less. Data for T. maccoyii are lacking. All other species overlap more or less in the 25-31 percent range, although between 400 and 500 mm. species distinc- tions begin to be apparent (viz. Pacific T. obcsun pectorals become relatively longer, those of T. tonggol shorter). The marked positively allomctric growth of the pectorals of T. alalunga and Pacific T. obesus makes these two forms clearly distinctive from all others between 500 and 1,200 mm. (fig. 20). Most speci- mens have fins 34-46 percent of fork length. I'p to 700 mm., T. alalunga has slightly shorter fins than Pacific T. obesus, but from 700-1,200 mm. they are virtually identical. Atlantic T. obesus in the 050- 1,200 mm, range (no data were available for smaller specimens) appear to have significantly shorter pec- torals than Pacific specimens : 29-35 percent at sizes of 0.50-1,000 mm., then gradually decreasing until no suggestion of difference is seen above 1,300 mm. In T. albacares a gradual negative allometry after 600 mm. keeps the pectorals shorter than in T. alalunga or either Atlantic or Pacific T. obesus until about 1,100 mm., wli(>n overlap with T. obesus begins to increase. Whereas the pectorals of T. allanticus are, at first, very similar in length to those of T. albacares, the more rapid decline in relative length makes them at sizes above 000 mm. even shorter than in 7'. albacares. The greatest negative allometry is seen in T. tonggol, which, at 500 mm., already shows the trend that brings the pectoral length into the ranges of T. maccoyii and T. Ihynnus between 650 and 900 mm. Th(> fins are the shortest of all the species at fork 94 U.S. FISH .\ND WILDMKK SKKVICE lengths from 900 to 1,050 mm. (the maximum size of T. (onggol for ^\•hich data were availal^le). Except for the size range where T. tonggol overlaps it, T. thynniis has pectorals consistently shorter than those of other species, the longest on record being about 23 percent of fork length. Although more data are needed for T. maccoyii, it appears that tliis species has a slightly longer pectoral fin than T. Ihynnus. From 650-7.50 mm. fork length, for which a fair amount of data is available, the fin of T. maccoyii is 20-24 percent of fork length, that of T. thynnus 17-21 percent. A few specimens of T. maccoyii between 900 and 1,000 mm., and one of 1,445 mm. have pectorals 22-23 percent of fork length, also slightly longer than those of similar sized T. thynnus. The ranges given by Iwai, Nakamura, and Matsubara (1965: 31, 33) of 4.8-6.0 in fork length (=16.7-21.7 percent) for T. thytinus and 4.4-4.5 in fork length ( = 22.2-22.7 percent) for T. maccoyii suggest a distinctness of separation that is not upheld by our data, although the basic species differences in pectoral length appear to be real. The distance from snout to second dorsal origin relative to fork length (fig. 27) shows a negative regression in all species of Thunnus over 400 mm. When size is taken into account, this measurement provides a reliable separation of some species, but a simple statement of range is inadeciuate. For ex- amjile, T. tonggol appears distinct throughout its size range (maximum around 1,000 mm.) but larger T. albacarcs, T. obesus, and T. thynnus all have a distance that is the same as that of smaller T. tonggol. Throughout the size ranges examined by us (fig. 27) the distance is greatest in T. alahtnga and least in T. tonggol. Two intermediate groups can be cate- gorized: one with a shorter distance that includes T. atlanticus and T. albacares, and one with a longer distance that includes T. obesus and T. thynnus. The meager data for T. maccoyii fall in the range of T. thynnus. Below a fork length of about 600 mm. there is so much overlap that the usefulness of snout- second dorsal distance in species distinction is doubt- ful, but above 600 mm. it appears to be useful. Greatest body depth is shown in fig. 28. This iiil' B= Ps _ a> nil' $>• \vvvV\ J" ' "^ ^^Bl ^p^^^^^ ^^- Y^ ^^ ^ 1 1 1 I •'' 1 1 1 1 Figure 27. — Relative distance from snout to second dorsal origin in Th minus species. Data, in additon to our own, include Dung and Royce (1953: tables 12, 21, 28, 42, 50-54). 1— T. atlanticus; 2-—T. tonggol; 3—T. albacarcs; i—T. obesus; 5 — T. alahinga; 6 — T. thynnus. Figure 28. — Relative greatest bod_y depth in Thunnus species. Data, in addition to our own, include Dung and Royce (1953: tables 27, 28, 42). 1— T. atlanticus; 2— T. tonggol; 3 — T. albacares; 4 — T. obesus; 5 — T. alahinga;^ — T. thynnus. character is so variable that it should not be used by itself. Rather, there are tendencies which, with other characters, can be helpful in species deter- mination. In specimens less than about 600 mm. fork length, overlap is particularly evident, but two categories can be based on greatest depth: deep-bodied species, including T. obesus, T. thynnus, and T. atlanticus, with depths usually 26-30 percent of fork length; and slender species, including T. albacares, T. tonggol, and T. alahtnga, with depths usually 23-26 percent of fork length. The few data for T. maccoyii fall in with T. thynnus. There appears to be little change in depth relative to fork length from 600 to 1,500 mm. in any species except T. alalunga, in which the relative depth increases gradually until specimens over 1,000 mm. are clearly in the deep-bodied category. The two species that commonly become larger than 1,500 mm., T. obesus and T. thynnus, exhibit A.NATOMY AND SYSTEM,\TICS OF TUNAS 95 increased variability at these larger sizes. In T. thijrmus this is particularly evident; specimens over 2,000 mm. fork length (not shown in fig. 28) appear randomly distributed over a range of body depths from 22-29 percent of fork length, which is almost the entire range of all species combined. The greatest body depth is found in individuals of T. obr.vis at all sizes over fiOO mm., but the .species overlaps with T. atlanticus, T. thynnus, or T. alahinga throughout its known size range. Eye size, in combination with other characters, is a useful species criterion, but the negative allometry must be considered. Because we measured the bony orbit, our data are not comparable with most other published data. We recommend that future work- ers use iris diameter. Fig. 29 compares the species. Both T. alalungn and T. atlanlicus exhibit wide variation in eye diame- ter, making categorical statements difficult. At less than fiOO mm. fork length the smallest orbit diameter is found in T. thijnnus and T. (oiiggol, the largest in T. atlanticus and T. alalungn, and intermediate in T. albacarcs; T. obesus is not represented. At sizes greater than (100 mm., T. obc.sus clearly has the largest orbit diameter. Variation in T. alalunga covers the range from largest to smallest. The other species have so much overlap with one another that species distinctions are impossible. COLORATION Colors and color patterns of tunas have limited vise in tuna systematics because they show great indi- vidual and age variation, and because they may be lost after death and preservation. Nevertheless, there are some excellent color characters, in spite of J i_ 400 &00 FORK LENGTH Figure 29.— Diameter of bony orbit relative to fork length in Thuitnus species. Only our data used. 1 — T. ntlaritictis; 2—T. tonggol; 3— T. albacares; A—T. obcsus; b—T. aUi- lunga; 6 — T. thynnus. the difficulty in verbal expression of many of them. Many of the descriptions are taken from Mather (19fi4). Body. Most Thunnus species are iridescent dark blue above and silvery below. T. albacares is the most brilliantly colored, with a shining golden lateral band. T. atlanticus also has a prominent gold lateral band, but its body is usually very dark compared with other species. T. obcsus may display a trace of a gold band, but the Imnd is apparently absent in T. thi/nnus and T. tonggol and is replaced by an iridescent blue band in T. alalunga. Small specimens of all species may dis'jilay a pattern of white spots or streaks ventrolaterally. In T. tonggol these markings consist of horizontally elongated spots. The other six species have rounded spots that are either randomly distributed or t«nd to l)ecome arranged in vertical rows that alternate with vertical white lines; horizontally elongated spots are sometimes seen on the caudal peduncle but rarely farther anteriorly. This pattern is usually lost in large individuals, although it may be retained in s])ecimens of T. albacarcs and T. thynnus up to 1,. 500-1, 600 mm. In T. maccoyii, alone among the species of Thunnus, the caudal keels are an unmistakable bright yellow. In the fish markets of Japan, we were able to recognize T. maccoyii from a consider- able distance on the basis of this charac^ter. How- ever, we suspect the keels may lo.se their yellow in larger adults. Fins. The color of the first dor.sal fin is variable. It may be entirely white, or there may be a yellow suffusion, and the distal margin may be black. Too few ob.servations have been made to enable us to characterize the species. The second dorsal and anal fins usually have yellow tips in all but T. alalunga and T. atlanticus, which have dark fins with white distal margins. The dorsal and anal finlets are usually bright yellow with black margins in all except T. alalunga and T. atlanlicus. In T. albacarcs the black margin is usually very narrow, while in T. obcsus it is wider. T. alalunga may have yellow in the dorsal and some anal finlets, but the anal finlets are commonly all silvery or dusky. Both the dorsal and anal finlets of T. atlanticus are almost invariably dusky with white margins; yellow has been ob.servod in these finlets only in frozen specimens. June (19.j2b) reported black dorsal and anal finlets in an unusual specimen of T. albacarcs from the central Pacific. 96 U.S. FI.SH .\.ND WILDLIFK SERVICE The caudal fin of T. alalunga has a narrow, white traihng margin that distinguishes it from all other Thunnns, in wiiich the white margin is lacking. Specific Characters. The uniformly white- margined dusky finlets of T. atlanticus, the white caudal margin of T. alalunga, the horizontally elon- gated ventrolateral spots of T. tonggol, and the yel- low caudal keels of T. maccoyii are the only color characters we regard as generally useful in distin- guishing species, and confusing examples of other species with these same characters have been ob- served. PART 2. SYSTEMATICS Workers have differed in their interpretations of the suprageneric relationship of tunas and the mackerel-like fishes. Regan (1909) and Starks (1910) placed all of these fishes in the single family Scom- bridae. Kishinouye (1915, 1917, 1923) recognized four families: Scombridae, Cybiidae, Katsuwonidae, and Thunnidae, the last two of which he (1917, 1923) recognized as an order Plecostei, separate from the Teleostei, in which he included all other higher bony fishes. Takahashi (1924, 1926) disagreed with the recognition of a distinct order but did not alter the four families. More recently, Fraser-Brimner (1950) placed the tuna-like and mackerel-like fishes back in the Scombridae. Berg (1940, 1955) is one of the few recent taxonomists who followed Kishinouye in placing the tunas in a separate order Thunniformes. For reasons outlined elsewhere (Collette and Gibbs, 1963), we follow Regan, Starks, and Fraser-Brunner in placing all of the tunas and other mackerel-like fishes in the family Scombridae. It is possible to divide the Scombridae into smaller units. Gasterochisma is so different from the other scombrids that it deserves at least subfamily status. Fraser-Brunner (1950) recognized only the subfamilies Gasterochismatinae and Scombrinae. Nakamura (1965) considered Thunnus and Eulhijnnus (includ- ing Katsuwomis) as a third subfamily, Thunninae. In the comparative diagnosis of the genus Thunnus which follows, we give suggestions of other possible subdivisions. Until a thorough anatomical study is completed, however, we do not wish to present formally a revised family classification. THUNNUS SOUTH, 1845 Thynnus Cuvier, 1817: 313 (type-species: Scomber thynnus Linnaeus, 1758, by absolute tautonymy; preoccupied by Thynnus Fabricius, 1775, a genus of Hymenoptera). Orcynus Cuvier, 1817: 314 (type-species: Scomber germo Lacepede, 1800 [= Scomber alalunga Bon- naterre, 1788], by subsequent designation of Jordan, 1888: 180; preoccupied by Orcynus Rafinesque, 1815, a substitute for Scomberoides Lacepede). Thinnus S. D. W., 1837 (emendation of Thynnus Cuvier, 1817, therefore taking the same type- species: Scomber thynnus Linnaeus, 1758; suppres- sion in favor of Thinnus South, 1845 requested by Collette and Gibbs, 1964). Thunnus South, 1845 (emendation of r^nnws Cuvier, 1817, therefore taking the same type-species: Scomber thynmis Linnaeus, 1758). Orycnus Cooper, 1863: 77 (substitute name for Thynnus Cuvier, 1817, and therefore taking the same type-species: Scomber thynnus Linnaeus, 1758; not Orycnus of Gill, 1861, a misprint for Orcynus Cuvier, 1817). Albacora Jordan, 1888: 180 (substitute name for Thynnus Cuvier, 1817, therefore taking the same type-species: Scomber thynnus Linnaeus, 1758). Germo Jordan, 1888: 180 (substitute name for Orcynus Cuvier, 1817, therefore taking the same type-species: Scomber germo Lacepede, 1800 [ = Scomber alalunga Bonnaterre, 1788]). Parathunnus Kishinouye, 1923: 442 (type-species: Thunnus mebachi Kishinouye, 1923 [= Thynnus obesus Lowe, 1839], by monotypy). Neothunnus Kishinouye, 1923: 445 (type-species: Thynnus macropterus Temminck and Schlegel, 1844 [= Scomber albacares Bonnaterre, 1788] by subsequent designation of Jordan and Hubbs, 1925:218). Kishinoella Jordan and Hubbs, 1925: 219 (type- species: Thunnus rarus Kishinouye, 1923 [= Thyn- nus tonggol Bleeker, 1851] by original designation). Semathunnus Yow'hr, 1933: 163 (type-species: Sema- thunnus guildi Fowler, 1933 [= Scomber albacares Bonnaterre, 1788] by original designation). Comparative Diagnosis The tunas, genus Thunnus, comprise a group of seven closely related species representing the most advanced members of the family Scombridae ANATOMY AND SYSTKMATICS OF TUNAS 97 (sensu Regan, 1909; Starks, 1910; and Fraser- Brunner, 1950). The subfamily Scombiinae of Frascr-Bninner, which includes all Scombridae ex- cept Gastprochisma, is divisible into two major groups (Collette and Gibbs, 19G3). The more primitive Scomber, Rastrelliger, Scomber omor us, Grammatorcy- 7n(.s, and Acanthocybium have a posterior notch in the hypural plate and lack a bony lateral keel on the caudal vertebrae. The more advanced group, con- sisting of <7;/m«o.san/a, Orcynop.tis, Sardn, Cybiosarda, Auxis, Euthynnus, Katsuwonus, Allothunrms, and Thunnus, lack a hypural notch and have a bony caudal keel. Within the latter group another cate- gory may be recognized as including Allothunnus, Auxis, Euthynnus, Kaisnwonus, and Thunnus (the Plcco.stei of Kishinouye, 1917, 1923), characterized by the presence of well-developed prootic pits and (except Allothunnus) a subcutaneous vascular sys- tem. Within this group of higher scombrids, the genus Thunnus is characterized by the presence of fronto-parietal foramina, a particularly well-devel- oped subcutaneous vascular system with two long lateral branches on each side, and the body fully covered with scales. Auxis and Allothunnus lack fronto-parietal foramina. Auxis, Euthynnus, and Katsuwonus have the body squamation limited to an anterior corselet and do not have the subcutaneous vascular system as well-developed &s in Thunnus: the lower lateral branch is either short, or, if long as in K. pelamis, it meets the upper branch mesial to the ribs. Validity of Nominal Genera Cuvier (1817: 312-314) was the first to divide the large Linnaean genus Scomber. For the tunas he proposed Thynnus for T. thynnus and Orcynus for T. alalunga. Later (in Cuvier and Valenciennes, 1831) he placed hissubgenusOrr^nusin thesynonymy of Thynnus. Several subsequent workers independ- ently realized that Thynnus Cuvier was preoccupied by Thynnus Fabricius in insects. Thus Cooper (1863) accepted Gill's (1861) Orycnus, a misspelling of Orcynus, as a replacement name for Thynnus Cuvier (see also Gill, 1889). Jordan (1888) over- looked this action and proposed Albacora to replace Thynnus, and Germo to replace Orcynus. Gill (1894) settled matters by showing that South (184.')) had previously suggested Thunnus to replace Thynnus Cuvier. Most subseciuent workers have used Thun- nus South either for T. thynnus alone or for several or all of the seven species we refer to Thunnus. Whitley (1955) recently discovered an earlier modifi- cation of Thynnus Cuvier, namely Thinnus S.D.W., [ 1837. S.D.W. (perhaps S. D. Wood, according to j Whitley) emended a numb(>r of names by changing y to i, ph to f, . . . . As far as we can determine, only Abe (1955) followed Whitley in the usage of Thinnus i S. D. W. In order to stabilize the consistent usage of Thunnus South from about 1890 to the present, we have applied to tlie International Commission of Zoological X'^omcnclature to suppress Thinnus S. D. W. (Collette and Gibbs, 1904). Other nominal genera have been proposed, based | on anatomical data. Kishinouye (1923) described two additional genera: Parathunnus based on T. obesus (as mebachi) and Neolhunnus which included albacares (as nuirropterus) and longyol (as rarus). He based this division on anatomical characters such as liver striations, the level at which the subcutaneous blood vessels pass through the myomere, and pres- ence or absence of the postcardinal vein. Jordan and Hubbs (1925) then proposed Kishinoella for T. tonggol (as raru.s), the only tuna that generallj' lacks a swim bladder. We have summarized these differ- ences and others that have been used to distinguish genera or subgenera (table 4). A large number of different arrangements can be made depending on which characters one wishes to emphasize as "basic." Thunnus can be divided into two groups using the area of origin of the cutaneous artery, the level at which it passes between the ribs, and the inter- muscular bones between which it divides : T. alalunga, T. maccoyii, and T. thynnus in one and T. obesus, T. albacares, T. tonggol, and T. atlanticus in the other. On the basis of number of arteriolar rows, T. thynnus, T. maccoyii, and T. oljcsus stand out from the other species. The presence of liver striations and vascular cones and the length of the liver lobes place T. alalunga, T. maccoyii, T. thynnus, and T. obesus in one group, the remaining three species in another. The absence of a swimbladder distin- guishes T. tonggol from the other species, but the swimbladder may be rudimentary in T. ?naccoyii, and a swimbladder has been observed in small T. tonggol. T. atlanticus is unique in Thunnus in hav- ing 19 instead of 18 precaudal vertebrae. T. atlanti- cus and T. tonggol fall together on the basis of their low number of gill rakers, and T. thynnus stands out with the highest number in the genus. T. alalunga is uniciue in the position of spleen and stomach, in the shape of the first ventrally directed parapophysis, 98 U.S. FISH AND WILDLIFE .SEUVICE Table 4. — Cojnparison of diagnostic characters of the species of Thunnus Character Cutaneous artery orginates at vertebra number Cutaneous artery passes between ribs number Cutaneous artery divides between intermuscular bones number Number of arteriolar rows from cutaneous artery Post-cardinal vein Liver striations and vascular cones Liver lobes Swimbladder Spleen position First haemal arch on vertebra number First ventrally directed parapophysis on vertebra Anterior haemal prezygapophysis position Anterior haemal postzygapophysis length Ventrolateral foramina size Number of precauda! vertebrae... Posterior parasphenoid margin Pectoral length. Gill raker number T. alalunga 3-4 3-4 4-5 1 absent present subequal present right 10 9 on centrum short small 18 angulate long 25-31 T. thynnus 3-4 3^ 4-5 2 absent present subequal present left 10(11) g near centrum short small 18 angulate short 34^3 T. maccoj/ii 3-4 3-1, 2-3 4-5, 3-4 2 absent present subequal present left near centrum short small 18 angulate short 31-40 T. obesus &-8 5-6 6-7 2 present present subequal present 11(10) near centrum short small 18 angulate long to medium 23-31 T. albacares 6-8 5-6 fi-7 1 present absent right long present left 11(10, 12) well ventrad long large 18 non-angulate medium 26-34 T. atlantkus 6-8 5-6 6-7 1 present absent right long present left 11(10) well ventrad long large 19 non-angulate medium 19-25 T. tonggot 6-8 5-6 6-7 1 present absent right long absent or rudimentary left 11(12) 10 well ventrad long large 18 non-angulate medium to short 19-28 and in the flattened haemal spine of its firet caudal vertebra. Not only is the subdivision of Thunnus into genera or subgenera an arbitrary matter, but such subdivi- sion obscures the close relationship among the species. In this concept we agree with such workers as Rivas (1951, but not 1961), de Sylva (1955), and Iwai, Nakamura, and Matsubara (1965). Thunnus can be divided into as many as six groups, but these are essentially species, not subgenera or genera (table 4). However, based on the 18 characters in table 4 (ex- cluding pectoral fin length), there do appear to be two groups of species. T. alalunga, T. thynnus, and T. maccoyii are similar to each other in 14-16 char- acters; T. albacares, T. atlanticus, and T. tonggol are similar to each other in 15-16 characters; and T. obesus is in between the two groups, sharing 12 characters with T. maccoyii and 10 with T. albacares. This agrees with the intra-generic relationships pre- sented by Iwai et al. (1965) and Xakamura (1965). It disagrees with Watson's (1964) groups where she placed T. obcstis in the first group. T. obesus is similar to the first group in several liver and verte- bral characters but fits with the second group in position of the cutaneous artery, presence of the post- cardinal vein, and position of the first haemal arch. The synonymy of each species includes all the combinations of names we have found, together with selected references containing information on anato- my, morphometry, and distribution. Readers wish- ing more references should consult tuna bibliog- raphies such as Corwin (1930), Shimada (1951), and volume 4 of the "Proceedings of the World Scinetific Meeting on the Biology of the Tunas and Related Species" (Bernabei, 1964). THUNNUS ALALUNGA (Bonnaterre, 1788) ALBACORE Scomber pinnis pectoralibus longissimis Cetti, 1777: 191-193 (Sardinia, alalunga in vernacular). Scomber alalunga Bonnaterre, 1788: 139 (original description based on Cetti). Walbaum, 1792: 222. Risso, 1810: 169-170. Scomber alalunga Gmelin, 1789: 1330 (original de- scription based on Cetti; "alalunga" a misprint for "alalunga" ; date of publication according to Cat. Books British Mus. is 1789, not 1788). Lacepede, 1800: 599 and 1802: 21-22. Scomber gcrmo Lacepede, 1800: 598 (original descrip- tion in table of species of Scomber; misspelled S. germon). Lacepede, 1802: 1-8 (description: S. Pa- cific Ocean, 17° S., 103° W. ; based on Commerson's manuscript). Orcynus germon, Cuvier, 1817: 314. Orcynus alalonga, Risso, 1826: 419-420 (Mediter- ranean). Thynnus alalonga, Cuvier in Cuvier and Valencien- nes, 1831 : 87-95 (Atlantic), fig. 215. Lowe, 1839: 78 and 1849: 2 (Madeira). Gtinther, 1860: 366. Cunningham, 1910: 109-110 (synonymy, descrip- tion; St. Helena), fig. 3. Thynnus pacificus Cuvier in Cuvier and Valenciennes, 1831: 96-97 (substitute name for Scoynbcr germo Lacepede, 1800). AN.\TOMV AND SYSTEM.\TICS OF TUNAS 99 Thunmis alalonga, South, 1845: 022. Thiinniis pacificus, South, 1845: 622. Orcyyjus pacificus (not of Cuvier) Cooper, 1863: 75-77 (original description; California), fig. 19. Orcynus germo, Ltitken, 1880: 408-472, 596 (synon- ymy in part), pi. 3, figs. 1-2 (young). Kitahara, 1897: 2 (description: Japan), pi. 2, fig. 3. Germo alalonga, Jordan, 1888: 180. Barnard, 1927: 799 (S. Africa). Morice, 1953: 68-69, fig. 3 (de- scription of liver; E. .Ulantic). Albacora alalonga, Dresslar and Fesler, 1889: 438- 439 (synonymy in part), pi. 6. Grrmo alalnngn, Jordan and Evermann, 1896: 871 (description; synonymy in part; Atlantic and Pacific). Jordan and Jordan, 1922: 33 (Hawaii). Meek and Hildebrand. 1923: 316-317 (description; sjTionymy in part). Jordan and Evermann, 1926: 15 (Atlantic). Buen, 1930: 48 (synonymy in part), fig. 6. Jordan, Evermann, and Clark, 1930: 260 (synonymy in part). Fowler, 1936: 621-623 (synonymy in part; description), fig. 280. Walford, 1937: 14-17 (description; a single world- wide species of albacore), color pi. 35. Fowler, 1944: 498 (W. of Chile). Tinker, 1944: 158-159 (Hawaii), pi. 1, fig. 6. Brock, 1949: 267 (in key to Hawaiian tunas). Le Gall, 1949 (synonymy, de- scription, biology). Smith, 1949, 1953: 299 (S. Africa), pi. 66, fig. 835. Fernandez-Yepez and Santaella, 1956: 12, fig. 3, pis. 1, 5, (Venezuela). Tucker, 1955 (British Seas). Otsu, 1960 (migra- tion, growth; X. Pacific). Frade and Yilela, 1962: 17-59 (morphology, biology; E. Atlantic). Postel, 1963 (description, biology; E. Atlantic). Thynnus alalunga, Clarke, 1900 (Scotland). Germo gcrmon. Fowler, 1905:701-763 (Sumatra). Germo germo, Jordan and Seale, 1906: 228 (Samoa). Jordan and Hubbs, 1925: 217 (Japan). Jordan and Evermann, 1920: 10 (Pacific), pi. 3, fig. 1. Jordan et al., 1930: 260 (synonymy). Thunnus alalunga, Jordan, Tanaka, and Snyder, 1913: 120 (Japan). Kishinouye, 1915: 18 (de- scription; Japan). Fraser-Bi-unner, 1950: 142 (key to Thunnus), 143 (synonymy in part ; distribution), fig. 5. Rivas, 1951: 222-223 (synonymy, descrip- tion; Atlantic), de Sylva, 1955: 33 (relation- ships, osteology), fig. 56 (neurocranium). Bullis and Mather. 1956 (counts, measurements; key to Caribbean Thunmis), fig. 3. Kurogane and Hi- yama, 19.i7b (morphometry; NW. Pacific). C.osline and Brock, 1900: 259-200 (description; Hawaii), 336-337 (sjmonjTny), fig. 257i. Mather and Gibbs, 1957: 242-243 (39° 45' N., 73° 00' W.). Jones and Silas, 1960: 382-383 (Indian Ocean), fig. 9. Talbot and Penrith, 1962: 558 (S. Africa). Jones and Silas, 1963: 1790-1791 (Indian Ocean). Rodrigues Lima and Wi.se, 1963 (ilistribution; W. tropical Atlantic). Squire, 1963 (distribution; NW. Atlantic). Talbot and Penrith, 1963: 609- 616 (description, biology; S. Africa). Iwai and I Nakamura, 1964: 6, fig. 3a (olfactory rosettes). Williams, 1964: 121 (E. Africa). Iwai et al., 1905: 3-5 (synonymy), 28-30 (description), figs. 13, 14, 15. Nakamura, 1905: 13-17, figs. 1, 2, • 3 A, 4, 5A (osteology). Merritt and Thorp, 1966: 377 (E. Africa). Nakamura and Kikawa, 1906 (infracentral grooves). Thunnus germo, Kishinouye, 1923: 434 (anatomy; Jai)an), figs. 20, 40, 52. Serventy, 1941: 23-24 (Australia), pi. 2. Godsil and Byers, 1944: 70-87 (anatomy; comparison of Pacific si)ecimens), figs. 30-47. Godsil, 1948 (morphometric comi)arison of Japanese, Hawaiian, and American specimens). Alverson, 1961 (distribution; NE. Pacific). Clemens, 1901 (migration, age, growth; N. Pacific). Clemens, 1963 (migration; N. Pacific). Yoshida and Otsu, 1963 (biology; Pacific and Indian oceans). Otsu and T'chida, 1963 (migration; Pacific). Jones and Silas, 1964: 34-36 (Indian Ocean). Germo gcrmon sleadi Whitley, 1933: 81-83 (original description; New South Wales), pi. 11, fig. 1. Thunnus gcrmon, Tortonese, 1939: 324-325 (W. coast S. America). Thunnus alalunga germo, Mimro, 1958: 111 (Aus- tralia). Types of Nominal Species Scomber alahinga Bonnaterre, 1788. No type specimens. Original description based on Cetti, 1777 (Scomber pinnis pccloralihus longissimus). Scomber alalunga Gmelin, 1789. No type speci- mens. Original description based on Cetti, 1777. Specific name, published as alalunga, a misprint. Scomber germo Lac<'>pede, 1800. No type speci- mens. Original description based on manuscript by Commerson. Specific name spelled germon in table of species of Scomber (1800), spelled germo in de- scription (1802). Thynnus pacificus Cuvier, in Cuvier and Valen- ciennes, 1831. No type specimens. Original de- scription based on Lacepede's Scomber germo (1800) and on Commerson's manuscript. The specimen. 100 U.S. FISH AND WILDLIFE SKRVICE MNHN A. 6862, considered as the holotype by Bauchot and Blanc (1961, p. 377) and Blanc and Bauchot (1964, p. 456) is, therefore, not a type (Collette, 1966). Orcynus pacificus Cooper, 1863. No type speci- men, although mention is made of "State collection, species 1033." Gcrmo gcrmon steadi Whitley, 1933. Holotype Australian Museum, Sydney, lA 2457, New South Wales, a misshapen skin, 960 mm. FL, preserved in "formalin with most of the fins broken. Pectoral fin about 45 percent of fork length. Figured by Whitley (1933, pi. xi, fig. 1). Characters Pectoral fin very long, usually reaching nearly or quite to second dorsal finlct, usually 31 percent of fork length or longer (similar to T. obcsus). Body depth greatest near dorsal and anal origins. A nar- row white posterior margin on caudal fin. Anal finlets silvery or dusky. Gill rakers 25-31 (similar to T. obesus and T. albacares). Liver with striations on ventral siu'face, its three lobes subequal in length, vascular cones present on its dorsal side (as in T. thynnus, T. maccoyii, and T. obesus). Spleen located on left side, stomach on right. Straight intestine short, the first loop located at about half to two-thirds »the distance between middle liver lobe and anus. Gall bladder exposed in ventral view along right side of straight intestine. Connective tissue on dorsal wall of body cavity much thickened posteriorly. Kidney short, without pos- terior "tail," reaching level of vertebrae 7-9. Cutaneous arteries usually originating at level of vertebra 3-4, passing laterally between ribs 3 and 4. and branching between intermuscular bones 4 and 5 (as in T. Ihy units and T. maccoyii) ; no posterior com- missure. A single row of arterioles and venules arising from each main lateral cutaneous branch (as in T. albacares, T. tonggol, and T. atlanticus) but from vertebral side of vessels. Post-cardinal vein absent (as in T. thynnus and T. maccoyii). Posterior parasphenoid margin forming an acute angle (not as extreme as in large T. thynn us but more acute than in large T. obesus). Supraoccipital crest relatively slender and long, reaching at least to centrum of vertebra 3. Anterior articulating (sphenotic) head of hyoman- dibula relatively long and narrow, proportion of length to least width 1.7-2.7. Metapterygoid rela- tively narrow, proportion of length of anteroventral margin to postero ventral margin 1.1-1.8. Quadrate relatively narrow, proportion of length to width of horizontal dorsal edge 2.1-2.7. Vertebrae 18-(-21 (as in all Thunnus except T. atlanticus). First ventrally directed parapophysis on vertebra 9 (as in all except T. thynnus and T. tonggol), appearing twisted and not extending strongly ven- trad. First closed haemal arch on vertebra 10 (as in T. thynnus, T. maccoyii, and occasionally in other species), forming an angle of 45 degrees or less with the vertebral axis. All haemal prezygapophyses arising from centra, not from haemal arches. All haemal postzygapophyses less than one-fourth cen- trum length. Anteriormost ventrolateral foramina small, their width not greater than basal width of haemal spine. Least height of centrum of 36th vertebra 1.1-1.7, usually 1.4-1.6, in centrum length (similar to T. albacai'cs), centrum commonly taper- ing, with least depth at anterior end (in the other species the vertebra is of nearly equal height throughout). Haemal spine of first caudal vertebra flattened, wing-like. Nominal Species Although no one seems to have reported any im- portant differences l)etween Atlantic and Pacific populations of T. alalunga, at least since Jordan and Evermann (1926), many recent authors still refer to the Pacific populations as T. gcrmo. Even Jordan and Evermann (1926) admitted that the slight diffei'cnces they noted in body proportions and coloration would probably not be valid when more specimens were examined. Our data on T. alalunga confirm the view that the Atlantic and Pacific popu- lations belong to the same species. Godsil (1948) and Kurogane and Hiyama (1958a, 1959) found slight population differences within the Pacific, but intermingling of a significant portion of the eastern and western Pacific populations of T. alalunga was indicated by tag returns reported by Ganssle and Clemens (1953) and Blunt (1954), and demonstrated by more recent works, including those of Otsu (1960), McGray, Graham, and Otsu (1961), Clemens (1961, 1963) and Otsu and Uchida (1963). Range In the western Atlantic, T. alalunga is known from south of New England to southern Brazil. Squire (1963) presented seven records north of 40° N., the most northerly 42°18' N., 64°02' W. Le Danois ANATOMY AND SYSTEM.^TICS OF TUNAS 101 (1951) reported the species off the coast of Vene- zuela. Rodrigues Lima and Wise (1903) reported catches from 10° N. to 32° S. off the coast of Brazil, with a concentration near 15° S. There are no records for the Gulf of Mexico. In the eastern Atlantic, it has been found from the Orkney Islands north of Scotland (Clarke, 1900; Tucker, 19.55), south to Angola off west African coast (Vilela and Monteiro, 1959) and in the Mediterranean Sea. The range may extend south to South Africa, because Talbot and Penrith (1902, 1903) have found a con- tinuous distribution around South Africa. On the other hand, the South African population may be of Indian Ocean origin. The distribution in the Indian and Pacific oceans was mapped by Yoshida and Otsu (1903) and liy Suda, Koto, and Kumc (1903). T. alalunga is found across the Indian Ocean from East Africa to Australia between about 10° N. and 30° S. Its range in the western Pacific extends from about 45° X., off the coast of Hokkaido, south to 40° S., off the southern tip of Australia. Longline fishing has indicated a fairly continuous distribution be- tween 30° N. and 20° S., eastward past the Hawaiian Islands. In the eastern Pacific, it is known from about 50° X., off Vancouver Island, British Columbia (Cowan, 1938; Samson, 1940), south to about 42° S. (Japan Fishery Agency, 1904, 1905). THUNNUS ALBACARES (Bonnaterre, 1788) YELLOWFIN TUNA Albacores or Thi/nni Sloane, 1707: 11 (description; Madeira), fig. 1. Scomber albacares Bonnaterre, 1788: 140 (original description based on drawing by Sloane). Scomber albacorus Lacepetle, 1800: 599 and 1802: 48-49 (substitute name for Scomber albacares Bonnaterre, 1788). Thynnus argentiviilatus Cuvicr in Cuvier and Valen- ciennes, 1831 : 97-98 (original description; Atlantic and Pacific). Gunthcr, 1860: 300. Scomber Sloanci Cuvier in Cuvier and Valenciennes, 1831 : 148 (original description based on Sloane). Thynnus albacora Lowe, 1839: 77-78 (original de- scription; Madeira) and 1849: 2 (repeat of original description). Giinther, 1800: 305. Cunning- ham, 1910: 110-112 (synonymy, description; St. Helena), fig. 4. Thynnus macropterus Temminck and Schlegel, 1844: 98-99 (original description; Japan), pi. 51. Kishinouye, 1915: 19 (description, anatomy; Japan), pi. 1, fig. 12. Thunnus argentivittalus, South, 1845: 022. Rivas, 1951: 221-222 (synonymy). Orcynus subulatus Poey, 1875: 145-140 (original description; Cuba), pi. 3, fig. 4 (head), fig. 5 (scale). Orcynus albacora, Poej', 1875: 145. Orcynus macropterus, Kitahara, 1897: 2 (de.scri])- tion; Japan), pi. 2, fig. 3. Gcrmo macropterus, Jordan and Snyder, 1901: 04 (Nagasaki). Jordan and Seale, 1900: 228 (Samoa). .Ionian and Jordan, 1922: 32-33 (Ha- waii). Thunnus macropterus, .Ionian et al., 1913: 121 (.lapan). Kishinouye, 1915 (description, anatomy; Japan), dc Beaufort, 1951: 223-225 (synonymy, description), fig. 39. Ginsburg, 1953: 8-10 (re- striction of name macropterus to W. Pacific yellow- fin). Thunnus allisoni Mowbray, 1920: 9-10 (original description; Miami, Fla.), figure (uniumibered). Gcrmo argentiviltatus, Xichols and Murphy, 1922: 507 (Peru). Gcrmo allisoni, Nichols, 1923: 3 (Christmas Island). Neothunnus macropterus, Kishinouye, 1923: 440-448 (anatomy; Japan; placed in new genus Neothun- nus), figs. 13, 19, 23, 45, 51. Jordan and Ilubbs, 1925: 219 (Japan). Jordan and iMcnnann, 1920: 20-21 (description), pi. 5. Herre, 193(5: 100-107 (synonymy; Galapagos, Philippines, Jajian; no species diflcrcnccs between long- and short-finned yellowfin). Walfonl, 1937: 3-7 (description: Pa- cific Allison tuna merely old yellowfin), color pi. 33. Serventy, 1941: 25-20 (description; Aus- tralia), pi. 2. Godsil and Byers, 1944: 46-09 (anatomy; comparison of Pacific sijocimens), figs. 20-35, 70-70. Tinker, 1944: l.W-lOO (Hawaii), 1)1. 1. fig. 5. Godsil, 1948 (morphometry; Japan, Hawaii, E. Pacific). Schaefer, 1948 (morphom- etry; Pacific Costa Rica). Brock, 1949: 27(5 (key to Hawaiian tunas). Schaefer and Walford, 1950 (compai'ison of yellowfin from Angola and Pacific coast of Central America). Godsil and Green- hood, 1951 (comparison of E. and central Pacific specimens). Schaefer, 1952 (comparison of Ha- waiian and W. Pacific specimens). Roj-ce, 1953 (morphometry; Pacific; an east-west cline across the Pacific in some characters). Tsuruta, 1954 (morphometry; SW. Pacific). Schaefer, 1955 (comparisons of specimens from SJ'l Polj'nesia, 102 U.S. FISH AND WILDLIFE SERVICE Central America, and Hawaii). Tsuruta, 1955 (morphometry; southwest Great Sunda Island; yellowfins probably a single \:orldwide species with many sub-populations). Kurogane and Hi- yama, 1957a (morphometry; equatorial Pacific). Kurogane and Hiyama, 1958b (morphometry; Indian Ocean). Munro, 1958: 111 (Australia). Nakagome, 1958 (morphometry; Indian Ocean). Broadhead, 1959 (morphometry; E. tropical Paci- fic). Klawe, 1959 (reidentification of juvenile called T. thynnus by Fowler, 1944). Gosline and Brock, 1960: 260-261 (description; Hawaii), 337 (synonymy), fig. 257j. Jones and Silas, 1960: 385-386 (Laccadive Sea, Gulf of Mannar, Ratna- giri), fig. 12. Legand, 1960 (measurements, counts; New Caledonia; eas*^-west cline in gill rakers across Pacific). Tsuruta and Tsunoda, 1960 (morphometry; Indian Ocean). Talbot and Penrith, 1962: 558 (S. Africa). Mimura et al., 1963a (biology; Indian Ocear). Talbot and Pen- rith, 1963:617-623 (description, biology ;S. Africa). Thunnus subulatus, Jordan and Evermaim, 1926: 11-12 (repeat of Poey's original description). Jordan et al., 1930: 260. Ginsburg, 1953: 6-8 (synonymy, description; the name subulatus used for W. Atlantic yellowfin). Fernandez-Yepez and Santaella, 1956: 6 (Venezuela; in key as a species of bluefin). Neothunnus cataUnae Jordan and Evermann, 1926: 19 (original description; Santa Barbara Islands, S. California), pi. 4. Jordan et al., 1930: 260. Nichols and La Monte, 1941: 31, fig. 1. Neothunnus albacora, Jordan and Evermann, 1926: 21-22. Frade, 1929: 235-241 (morphometry, swimbladder; Canary Islands), pi. 5, fig. 2. Buen, 1930: 49-50, fig. 8. Bini, 1931: 31-36 (morpho- metry; Canary Islands), figs. 12, 13. Frade, 1931a: 123-126 (synonymy, morphometry; E. Atlantic). Nichols and La Monte, 1941 : 30 (syn- onymy in part), fig. 2. Barnard, 1948: 378-380 (S. Africa), pi. 11. Bellon and Bardan de Bellon, 1949 (morphometry; Canary Islands). Morice, 1953: 71-73, fig. 5 (liver; E. Atlantic). Postel, 1955 (biology, morphometry; E. Atlantic). Fer- nandez-Yepez and Santaella, 1956: 15 (in key to Atlantic tunas). Marchal, 1959 (morphometry; E. Atlantic). Vilela and Monteiro, 1959: 30-53 (morphometry; Angola). Tsuruta, 1961 (morpho- metry; SW. Indian Ocean). Vilela and Frade, 1963 (biology; E. Atlantic). Neothunnus itosibi Jordan and Evermann, 1926: 22-23 (original description; Hawaii), pi. 6. Smith, 1935: 207-209 (S. Africa), fig. 4. Phillips, 1932: 231 (New Zealand). Powell, 1937: 80-81 (New Zealand), pi. 17, figs. 2, 3. Jones and Silas, 1900: 387-388 (Madras; recognized as distinct from A''. macropterus), fig. 13. Neothunnus albacor"s, Jordan and Evermann, 1926: 23-24 (description). .lordan et al. 1930: 260. Fernandez-Yepez and Santaella, 1956: 17-18 (Venezuela), fig. 6, pi. 8. Neothunnus aUisoni, Jordan and Evermann, 1926: 24 (description). Jordan et al., 1930: 260. Nichols and La Monte, 1941: 30-31 (synonymy), fig. 3. Fernandez-Yepez and Santaella, 1956: 16 (Venezuela), fig. 5, pi. 7. Kishinoella zacalles Jordan and Evermann, 1926: 27 (original description, Honolulu fish market), pi. 7. Semathunnus guildi Fowler, 1933: 163-164 (original description; Tahiti), pi. 12. Semathunnus itosibi, Fowler, 1933: 164. Tinker, 1944: 160 (Hawaii). Neothunnus argentivittatus, Beebe and Tee-Van, 1936: 184-192 (synonymy, description; West Indies), fig. 5 (copy of fig. in Cunningham), figs. 6-12 (photographs), fig. 13 (copy of fig. in Sloane). Fowler, 1944: 498 (Mexico, Ecuador, Peru). Germo albacora, Fowler, 1936: 623-624 (synonymy, description), fig. 282. Smith, 1949, 1953: 299 (S. Africa), pi. 66, fig. 835. Thunnus albacora, Tortonese, 1939: 326 (off Brazil). Fraser-Brunner, 1950: 142 (key to Thunnus), 144-145 (synonymy), fig. 7. Morrow, 1954: 16 (29 E. African specimens similar to Pacific speci- mens). Germo itosibi, Smith, 1949, 1953: 299 (S. Africa) pi. 65, fig. 834. Neothunnus albacora brevipinna Bellon and Bardan de Bellon, 1949: 12-19 (original description; as Neothunnus albacora forma brevipinna; Canary Islands). Neothunnus albacora longipiruia Bellon and Bardan de Bellon, 1949: 12-19 (new name for long-finned T. albacares of East Atlantic; as Neothunnus albacora forma longipinna; Canary Islands). Neothunnus macropterus macropterus, Bellon and Bardan de Bellon, 1949: 15 (Pacific short-finned form; as Neothunnus macropterus forma macrop- terus). Neothunnus macropterus itosibi, Bellon and Bardan de Bellon, 1949: 15 (Pacific long-finned form; as Neothunnus macropterus forma itosibi). lN.\TOMY AND SYSTEMATICS OF TUNAS 103 NeoOuinnus hrcripinna, Postel, 19o0: C7-74 (descrip- tion, biology; considered a good species distinct from A'', albacora). Thimnus zacalles, Frascr-B runner, 1950: 142 (key to Thunnus), 146, fig. 9. Thunnus alhacares, Ginsburg, 1953: 3-6 (synonymy, description; the name alhacares restricted to the E. Atlantic yellowfin). de Sylva, 1955: 33-40 (osteology, relationships), fig. 58 (neurocranium). Bullis and Mather, 195G (counts, measurements, key to Caribbean Thunnus), fig. 2. ^Mather and Gibbs, 1957 : 242 (off New England). Rivas, 1961 : 136-139 (synon^Tuy, range). Schaefer, Broad- head, and Orange, 1963 (biology; Pacific). Scjuire, 1963 (distribution; NW. Atlantic). Iwai and Nakamura, 1964: 6, figs. 3G, H (olfactory ro- settes). Tsuruta, 1964: 59-66 (morphometry; Pacific and Indian oceans) . Williams, 1964: 115- 120 (E. Africa). Iwai et al, 1965: 11-15 (sjm- onymy), 36-38 (description), figs. 20, 21. Naka- mura, 1965: 20-22, figs. 3E, 9B, 10 (osteology). Koyce, 1965 (morphometry). Mcrritt and Thorp, 196(): 375-376 (E. Africa). Nakamura and Ki- kawa, 1966 (infracentral grooves). Thunnus calaUnac, Ginsburg, 1953: 8 (name used for E. Pacific yellowfin). Neothunnus alhacares, Mather, 1954: 292 (SE. of New York). Mather and Day, 1954: 184-185 (N. Brazil and W. Africa). Thunnus alhacares, Le Danois, 1954: 283-287 (his- tory of nomenclature), 285-286 (partial syn- onymy), 288-294 (biology; Pacific yellowfin recog- nized as Thunnus albacores variety argentiviUatus) . Neothunnus albacora macropfrrus, Schultz, 1960: 414-415 (description of Bikini and Marianas specimens), pi. 122 A. Thunnus alhacares macropterus, .Tones and Silas, 1 963 : 1793-1794 and 1964: 40-42 (Indian Ocean). Thunnus ilosihi, Jones and Silas, 1963: 1794-1795 and 1964: 42-43 (Indian Ocean). Types of Nominal Species Scomher alhacares Boniiaterre, 1788. No type specimens. Original description based on Sloane (1707, pp. 11-12; table 1, fig. 1). Scomber albacorus Lacepede, 1800. Substitute name for Scomher alhacares Bonnaterre, 1788. Thynnus argentiviUatus Cuvier in Cuvier and Valenciennes, 1831. Syntypes :\INHN A.5567 (a stuffed wliole skin; collected in the Atlantic Ocean by Quoy and Gaimard) and A.o572 (a half skin, with glass eye, mounted on a lioard; sent by Dussumier from the Indian Ocean). \ third specimen, A. 5814, designated by Schaefer and Walford (1950) as lecto- type, is not a syntypo liecause it was not mentioned by Cuvier in the original description (cf. Bauchot and Blanc, 1961, p. 376; Blanc and Bauchot, 1964, p. 454). We have examinefl lioth syntypes and can- not be certain what species they represent (see discussion under Nominal Species). Scomber sloaneiCuv\cr in Cuvier and Valenciennes, 1831. No type specimens. Original descri|)tion clearly based on Sloane (1707), plate 1, fig. 1, but also referring to page 28, where Sloane refers to a different fish (Scomhrus major torosus). Cuvier stated that Scomber albacorus Lacepede, 1800, is not the same as Scotnher sloanei, because Lacepede's descrip- tion refers to page 1 1 of Sloane. This, however, is the description of the fish, from the illustration of which Cuvier drew his description. Thijnnus albacora Lowe, 1839. No type specimens. Thynnus macropterus Temminck and Schlegel, 1841. Original description clearly based on the specimen figured in plate 51 , and not on the specimen in the Kijksmuseum van Natuurlijke Historie, I>eiden, number 2552, considered by Boeseman (1947, 1961) as the holotype. In particular, Temminck and Schlegel refer to the long second dorsal and anal fins, which the pi-esumed holotype (fork length 670 mm. as measuretl bj' Gibbs in 1962) is too small to have developed. We believe that this specimen shoukl not have been considered as holotype of this species. The specimen (a stuffed skin) is not a j^ellowfin tuna at all, but is T. tongqol. The pectoral fin is 22 percent of fork length and the snout to seconil dorsal dis- tance is 50.7 percent, both characteristic of T. tonggol. Since we do not believe this specimen was used in the original description and, therefore, is not a type, we are saved the necessity of having to consider the name tonggol Bleeker, 1851 as a junior synonym of macropterus, which has been used more often than has any other name for Pacific yellowfin tuna. Orci/nus suhulatus Poey, 1875. No type speci- mens known to us. Original descrijjtion from an 1,800-mni. specimen, of which the head is figured and might have been saved. Thunnus allisoni Mowbray, 1920. No type speci- men known to us. Original description from three specimens: one taken by sjiearing in Biscayne Bay, Miami, Fla. for which counts, proportions of body parts, and color are given, but the length noted as a 104 U.S. FISH .\ND WILDLIFK 8E11VP 1. little larger than the second specimen; a second specimen, 5 feet 9 inches long (1,753 mm.), "taken in the Gulf Stream," but "badly torn by sharks"; and a third specimen weighing 135 pounds (61 kg.). Neoihimnus catalinac .Jordan and Evermann, 1926. Type originally designated as "Xo. 597, Mus. Calif. Acad. Sci., a photograph of a fish taken off Santa Catalina Island, California," weight 157Ji pounds (71 kg.). This photograph was published earlier as Germo macropterm by Jordan and Starks (1907: 69). - The fish appears to be a mounted specimen. Neothunnus itosibi Jordan and Evermann, 1926. Type originally designated as "No. 598, Mus. Calif. Acad. Sci., a photograph ... of a specimen weighing 321 pounds in Honolulu market." The specimen is no longer extant. Kishinoella zacalles Jordan and Evermann, 1926. Type originally designated as "Xo. 599, Mus. Calif. Acad. Sci., a photograph of a specimen e.xamined in the Honolulu market . . ., 23^ feet long, . . . weighing 14 pounds." The characters given in the key to species (p. 26) are based on the specimen photo- graphed; the te.xt description is based on another specimen. Jordan and Evermann described zacalles as lacking a swimbladder, and they and subsequent workers (Serventy, 1942; Fraser-Brunner, 1950) have placed it close to T. tonggol. Jordan and Evermann, however, gave for their zacalles a gill-raker count of 30, which is completely outside the known range for T. tonggol (19-28, Table 2). It is our experience that the swimbladder may be quite difficult to find in some specimens of most species of Thunnus, and we believe that Jordan and Evermann probably overlooked it in their specimens of zacalles. They can not have been describing T. thijnnus, as this species has more gill rakers and a much shorter pectoral fin than they show in their photograph of zacalles. Of the three remaining Pacific species, T. alaluTiga may be quickly eliminated because it has a much longer pectoral fin and an entirely different coloration. T. obesus has a much larger eye than that shown for zacalles, and the swimbladder is well developed in all specimens that we ob.served. The description of Kishinoella zacalles fits T. albacarcs in number of gill rakers (mean for Pacific T. albacarcs 30.2, table 2), length of pectoral fin, coloration, and general body proportions. Also, Jordan and Ever- mann, in their original description, reported about a dozen specimens of zacalles, all from Hawaii, and no specimen of it (or of T. tonggol) has since been reported from there. In view of the great fishery and research program on tunas in the Pacific, it seems highly unlikely that a valid species has been overlooked. A'colhinnus albacora brevipinna Bellon and Bardan de Bellon, 1949. No type specimens. Original de- scription based on 11 specimens from the Canary Islands. Neothunnus albacora longipinna Bellon and Bardan de Bellon, 1949. No typo specimens. Presumed to be a new subspecific designation for the typical subspecies of .V. albacora Lowe (1839). Semathunmis guildi Fowler, 1933. Holotype AX'SP 55982, a dried skin with skull intact, from Tahiti. Fowler stated, "Length 1,830 mm." Our measurement of fork length was about 1,460 mm., of length to end of caudal lobes about 1,680 mm. The specimen is obviously a yellowfin tuna, with high dorsal and anal fins, and the pectoral reaching to the middle of the second dorsal base. Characters Pectoral fin intermediate in length, usually reach- ing bej^ond second dorsal origin but not beyond end of its base, usually 22-31 percent of fork length (generally similar to T. atlanticus and large T. obesus). Dorsal and anal fins very long in large specimens, becoming well over 20 percent of fork length. Gill rakers 26-35 (overlapping with T. alalunga and T. obesus). Liver without striations on ventral surface, its right lobe longer and narrower than the others; vascular cones not present on dorsal side (as in T. atlanticus and T. tonggol). Spleen located on right side, and stomach on left (as in all except T. alalunga). Connective tis.sue on dorsal wall of body cavity thickened at anterior end to form a prominent rai.sed cord. Kidney long, tapering, reaching level of vertebra 12-14, often with accessory masses posterior to main kidney. Cutaneous artery usually originating at level of vertebra 6-8, passing laterally between ribs 5 and 6, and branching between intermuscular bones 6 and 7 (as in T. atlariticus and T. tonggol) or 7-8. A single row of arterioles and venules arising from each main lateral cutaneous branch (as in T. alalunga, T. tonggol, and T. atlanticus) but from lateral sides of vessels (as in T. tonggol and T. atlanticus). Vessels present on each side parallel to dorsal aorta connecting posterior epibranchial to cutaneous artery. Post- cardinal vein present, joining right cutaneous vein (as in T. atlanticus, T. tonggol, and T. obesus). AX.\TO.MY AND SYSTEMATICS OF TUNAS 105 Posterior parasplienoid margin variable in shape, rounded, concave, or somewhat angulate (as in T. atlanticus and T. tonggol) hut never with a i)ro- noiniced angle. Vertebrae 18+21 (as in all Thnnnus except T. atlanticus). First ventrally directed parapophysis on vertebra (as in all Thutnnis excejit T. thj/nn)is and T. tonggol). First closed haemal arch usually on vertebra 11 (as in T. atlanticus, T. tonggol, T. ohcsus and often in T. thi/nnus). Anteriormost haemal prezygapophyses arising far ventrad on liacnial arch (as in T. atlanticus and T. tonggol). Haemal ])ostzygai)ophyses long (as in T. atlanticus and T. tonggol), the longest about 75 percent of its c(>ntrum length (somewhat shorter than in T. atlan- //(■(/.< and T. tonggol). Anteriormost ventrolateral foramina large, their width three or more times that of haemal spine (as in T. atlanticus and T. tonggol). Least height of centrum of 3nth vertebra 1.2-1.9, usually 1.3-1.') in centrum length (resembling T. alahmga, but in that species the vertebrae taper, whereas in T. albacarrs they are of nearly etiual width throughout). Nominal Species More names have been proposed for sujjposedly different ])opulations and individual variants of T. albacarcs than for all other species in the genus. Jordan and Evermann (1926) took the most extreme position in recognizing seven species: catalinar, from the Cf.'ifornia coast; vjacroptcrus, from the central and western Pacific; itosibi, a long-finned form from Hawaii and Japan; alhacora, from the eastern Atlan- tic; albacores, from Madeira and the West Indies: allisoni, a western Atlantic long-tinned form; and zacallcs from Hawaii (wtiich has heretofore been considered as most clo.sely related to T. tonggol, see above). The main characters that they used to separate tiiese forms were the lengths of the s(>cond dorsal and anal lobes. I'sing the same characters, Ginsburg (1953) distinguished an eastern Atlantic alhacares, a western Atlantic .suhulatus, an eastern Pacific catalinar, and a central and western Pacific macroplerus. It became apparent to us that T. albacarrs is an extremely variable species mori)ho- metrically, from our own data and from the many detailed moriihometric studies on pojiulations of T. albacarcs, especially in the Pacific, by workers such as Godsil (1948), Schaefer (1948, 1952, 1955), Schaeferand Walford (1950), Godsil and Creenhood (1951), Poyce (1953), Tsuruta (1954, 1955, 1961), Km-ogane and Hiyama (1957a, 1958b), Nakagome (1958), Broadhead (1959), Legand (19(10), and Frade (1931a, for the eastern Atlantic). Statistical analysis of morphometric data indi- cates that many sul)popuIations of 7'. albacarcs are differentiated, but certainly not to a species or sub- species level. Royce (1965), in a monumental study of the morphometry of T. alhacares, showed con- clusively that it is a single, locally variable, pantrop- ical species. He found that the dilTerences between eastern Atlantic and eastern Pacific specimens were less than the differences between eastern Pacific and Caroline Islands specimens, and that several char- acters change clinally from west to east in the e(iua- torial Pacific. Tiiere has been considerable confusion concerning the name Thgnnns argentivitlatns Cuvier. The original description (Cuvier, in Cuvier and Valen- ciennes, 1831: 97-98) was based on two specimens now at the M\HX in Paris: one from the Atlantic, collected by Quoy and Gaimard (MNHN A.5572) and one from the Indian Ocean, sent by Dussumier (MNHX A.5567). Schaefer and AValford (1950) reported that, according to information received from L. Bertin, the description was based on three specimens: the two already mentioned and a third from the Indian Ocean, coast of Malabar, sent by Dussumier (MNHN A. 581 4; given erroneously as A.5816 by Schaefer and Walford). .\.5814, a speci- men in alcohol, was designated the lectotype by Schaefer and Walford (1950, p. 441), who thus recog- nized the Indian Ocean yellowfin as Xrothunnus argentivittatns, the Pacific form as .V. macropirrus, and the Atlantic form as .V. albacora. Based on A.5814 being the lectotype, Uivas (1961) used the name argentivittatns for an Indian Ocean tuna which he tentatively placed in the subgenus Paralhunnus, and regard(>d as different from Neothunnus albacarrs, (lie yellowfin (una, which he considered to be a single, paiitropical species. We have examined the supi)os('d lec(otyi)e (A.5814) and believe it, and the other specimens in Rivas' (19GI) account, actually to be T. tonggol. Watson (1964) reached the same conclusion, and suggested that 7'. argrntiritlalus be synonymized with T. tonggol. This aclion, to begin with, is inappro- priate, for argrnliriltatus has ijriority over tonggol. A.581 1, however, (■aiiiio( lie considered as the lecto- type of Thgnnu.^ argrntirillatns, as it is nowhere men- 106 U.S. FISII .\ND WII.DI.IFK SKItVICE tioned by Cuvier in the original description, whereas the two proper syntypes are noted (Bauchot and Blanc, 19G1, p. 376; Blanc and Bauchot, 19G4, p. 454). The lectotype must be selected from A. 5567, a stuffed whole specimen, and A. 5572, a dried half specimen mounted on a board. Although both of us examined the two syntypes and independently made counts and measurements, we do not feel that we can make a selection. Even if the appropriate measurements could be considered accurate, which they certainly cannot, they do not indicate that the syntypes are T. alhacarcfi, but rather leave the possi- bility that they could be T. tonggol or T. allanticus. The distance from snout to second dorsal origin appears to eliminate T. Ihynnus, T. alalunga, and T. obesus fi'om consideration. We do not believe that these specimens can be definitely identified, unless a new and better character is found. Range As Royce (1965) has shown, T. albacares is a pantropical species. In the western Atlantic, it is known from about 42° N. (Squire, 1963) south through the Sargasso Sea to the Gulf of Mexico and the Caribbean Sea (Wathne, 1959) and off the coast of South America from about 10° N. to 32° S. (Rodrigues I.ima and ^Mse, 1963). In the eastern Atlantic, it is recorded from the coasts of Spain and Portugal south to .\ngola (Vilela and Monteiro, 1959; ^'ilela and Frade, 1963) but not from the Mediterranean Sea. Talbot and Penrith (1962, 1963) have shown that T. albacares has a continuous distribution around South Africa, but the origin of these fish is uncertain. It is abundant in East African waters (Williams, 1964) and is known from 20° X. to 30° S. in the Indian Ocean (Mimura et al., 1963a). In the west- ern Pacific T. albacares occurs from 40° X., off the coast of Japan, to 30° S., off the coast of Australia, between the 70° F. September isotherm to the north anil 75° F. February isotherm to the south (Schaefer et al., 1963). The distribution extends across the Pacific in a broad belt from about 30° X. to 20° S., between the same isotherms, and as far as 40° S. (.Japan Fishery Agency, 1965). THUNNUS ATLANTICUS (Lesson, 1830) BLACKFIN TUNA Thynnus allanticus Lesson, 1830: 165-166 (original description; Trinidade Is. off Brazil). Giintlier, 1860: 362 (in footnote as dubious species). Thynnus coretta Cuvier in Cuvier and Valenciennes- 1831: 102-104 (original description; Martinique)- Giinther, 1860:363. Thynnus balteatus Cuvier in Cuvier and Valen- ciennes, 1831: 136-137 (original description based on Lesson's unpublished drawing). Thunnus balteatus, South, 1845: 622. Thunnus coretta, South, 1845: 622 (description). Jordan and Evermann, 1926: 11 (description). Jordan etal., 1930: 260. Orcynus balteatus, Poey, 1868: 361-362 (Cuba). Poey, 1875: 145 (Cuba). Parathunnus rosengarteni Fowler, 1934: 354, 356 (original description; Key Largo, Fla.), figs. 3-5. Parathunnus ambiguus Mowbray, 1935 (original description; Bermuda). Parathunnus allanticus, Beebe and Hollister, 1935: 213-214 (Union Is., British West Indies). Beebe and Tee-Van, 1936: 178-184 (sjTionymy, descrip- tion; Bermuda and West Indies), figs. 1-4. Fowler, 1944: 102-103 (synonymy, description; AV. Caribbean), fig. 149. Schuck and Mather, 1951: 248 (X. Carolina). Mather and Schuck, 1952: 267 (Martha's Vineyard; XW. Caribbean). Mather and Day, 1954: 183-184 (off coasts of Brazil and Bermuda). Thunnus allanticus, Rivas, 1951: 219-220 (syn- onymy, description), de Sylva, 1955 (osteology, relationships, generic status), figs. 1-54, 57 (osteol- ogy). Bullis and Mather, 1956 (counts, measure- ments, key to Caribbean Thunnus). Rivas, 1961 : 129-131 (synonymy, description). Iwai et al., 1965: 15-16 (synonymy), 38-39 (description), fig. 22. Xakamura, 1965: 23-24, figs. 3F, 9C, 11 (osteology). Xakamura and Kikawa, 1966 (in- fracentral grooves). Misidentifications The specimen reported as Parathunnus obesus by Beebe and Tee- Van (1928: 100) from Haiti is T. allanticus as they (Beebe and Tee- Van, 1936: 181) later pointed out. Fernandez- Yepez and Santaella (1956: 19) reported specimens from Venezuela as Parathunnus obesus, but these are probably T. allanticus as indicated by Rivas (1961: 130). The International Game Fish Association (1965) listed the world record T. allanticus as a 44 pound, 8 ounce, specimen from Cape Town, South Africa. This record is obviously in error and has been cor- rected (1966). ANATOMY AXD SYSTEMATICS OF TUNAS 107 Types of Nominal Species Thijrmus atlanlicus Lesson, 1830. Xo tyjje speci- mens. Original description based on a specimen 28 inches total length (711 mm.), with a pectoral fin G inches long (l.Vi mm.). Subtracting .')() mm., wo obtain a fork length of about GGO mm. The pectoral is about 23 percent of fork length; too sliort for either T. albacarcs or T. atlanlicus (see fig. 2G), but is nearer the latter. Lesson mentioned only two other characters useful in identifying the species: a cop- pery-red lateral band, and blue-slate coloied fins (presumably also finlets). These appear sufficient to associate the name alhinlicitf! with the blackfin tuna, and we follow I^cebe and Tee-^'an (193()) and later autliors in doing so. Scomber corctla Cuvier, 1820. No type specimen. The first use of the name corctla for a tuna is usually credited to Cuvier in Cuvier and Valenciennes (1831), where he described Tlii/nnus corctla. The original description, however, consists of a footnote on page 198 of the second edition of Regne Animal (1829), which refers only to Sloane (1707, vol. 1, plate 1, fig. 3). Sloane 's figure is of "Scomber Major torosus" and there is no way of associating it with any known species, but this indication prevents the name from being considered a nomcn nudum. Scomber corctla Cuvier, 1829 must be regarded as a nomen dubium. Thynnus corctla Cuvier in Cuvier and Valen- ciennes, 1831. This later use of the name corctla is based on a preserved specimen, MXHN A. 5380, 2G3 mm. fork length from Martinique. It is a black- fin tuna with 19-1-20 vertebrae and a gill-raker count of G -hi 7 (left) and 7-1-17 (right). Thynnus baltcatus Cuvier in Cuvier and \'alen- ciennes, 1831. No type spe(nmens. Original de- scription based on an unpublished illustration by Lesson of the same specimen from which Thynnus atlanticus was described, and, therefore, a synonym of that species. Parathunnus rosengarteni Fowler, 1934. Holo- type ANSP G0174, a stuffed skin G29 mm. fork length. A count of gill rakers was impossible, but our measurements show the pectoral fin to be 25.8 percent of fork length, characteristic of T. aUanlicus. Parathunnus ambiguns Mowbray, 1935. No type specimens. Original description based on Bermuda specimens; gill rakers noted as G-|-17, swimbladder "simple, broader than long, well forward," finlets dusky with a trace of yellow. These characters unquestionably refer this nominal species to the synonymy of T. atlanlicus. Characters Pectoral fin intermediate in length (generally simi- lar to T. albacarcs and large T. obcsus), usually 22-31 percent of fork length. Dorsal and anal finlets in fresh specimens lacking yellow. Gill rakers 19-25, resembling only T. longgol. I.,iver without striations on ventral surface, right lobe long and narrow, no vascular cones on dorsal surface (as in T. albacares and T. tonqgol). Spleen located on right side, and stomach on left (as in all except T. alalunga). Swimbladder either small, oblate, situated far anteriorly, or resembling a poorly developed T. albacares; when long, anterior and posterior cham- bers divided by a membrane. Cutaneous arteries usually originating at level of vertebra (J-8, passing laterally between ribs 5 and G, branching between intermuscular l)ones G and 7 (as in T. albacarcs, T. longgol, and T. obcsus). A single row of arterioles and venules arises from each main lateral cutaneous branch (as in T. albacarcs, T. longgol, and T. alalunga), but from the lateral side of each vessel (as in T. albacares and T. longgol). Post- cardinal vein present, joining right cutaneous vein (as in T. albacarcs, T. longgol, and T. obesus). Posterior parasi)lienoid margin rounded, concave, or somewhat angulate (as in T. albacarcs and T. ionggol), never with a pronouiu'ed angle. Parasphe- noid concave in its anterior portion (seen occasionally in small specimens of all other species). Vertebrae 19-1-20, with rare exceptions. First ventrally dii-ected parapophysis on \ertebra 9 (as in all except T. longgol and T. thynnus). First closed haemal arch usually on vertebra 1 1 (as in all except T. alalunga and some T. thynnus). Haemal arches narrow, bowing widely, forming a large, ovate canal (as in T. longgol). Anterior haemal prezygapo-, phj'ses arising far ventrad on haemal arch (as in T. | longgol and T. albacares). Longest haemal post- 1 zygapophyses equal to or longer than centrum (only T. longgol and T. albacarcs approach this). Ante- riormost ventrolateral foramina large, more than; three times width of haemal spine (as in T. albacares and T. longgol). Nominal species Beebe and Tee- Van (193G) established the validity of T. atlanticus and placed Thjnnus baltcatus, Para- 108 U.S. FISH .VND WILDLIKK SERVICE thunnus rosengartem, and P. atiMguus in its synon- ymy. Thynnus coretta, wliich was placed in the synonymy of T. thijnnus Ijy Fraser-Brunner (1950) and Rivas (1951), is also a synonym of T. allanticus, as Rivas (19(il) has recently shown. Alorice (1953), Frade (1960), and others have mistakenly plac Ted. allanticus in the synonymy of T. obesus. Range Thunnus ailanticus is known only from the western Atlantic, from off Martha's Vineyard, Mass. (Mather and Schuck, 1952), and Cape Hatteras (Schuck and Mather, 1951), throughout the West Indies (Beebe and Tee-Van, 1936) and the northern Caribbean (Bullis and Mather, 1956), south to Trinidade Island off the coast of Brazil (Lesson, 1830) and off Rio de Janeiro at 22°21' S., 37°37' W. (Mather and Day, 1954). THUNNUS OBESUS (Lowe, 1839) BIGEYE TUNA Thynnus ohcsus Lowe, 1839: 78 (original description; Madeira). Lowe, 1849: 2 (copy of original de- scription). Giinther, 1860: 362 (in footnote as dubious species). Cunningham, 1910: 112 (syn- onymy, description; St. Helena), fig. 5. Thynnus sibi Temminck and Schlegel, 1844: 97-98 (original description; Japan), pi. 50. Giinther, 1860: 362 (in footnote as dubious species). Orcynus sibi, Kitahara, 1897: 1-2 (description; Japan), pi. 1, fig. 2. Thunnus sibi, Jordan and Snyder, 1901: 64 (Germo sibi; Nagasaki), 125 (supplementary note: the "Shibi" should be a species of Thunnus, T. sibi). de Beaufort, 1951: 222-223 (synonymy, descrip- tion), de Sylva, 1955: 34-40 (osteology, relation- ships), fig. 59 (neurocranium). Rivas, 1961 : 135- 136 (synonymy, description; a valid Indo-Pacific species). 3ermo sibi, Jordan and Snyder, 1901: 64' (listed; in supplementary note, p. 125 as Thunnus sibi). Jordan and Jordan, 1922: 33 (Hawaii). Thunnus mcbachi Kishinouye, 1915: 19 (original description; Japan). Paralhunnus mebachi, Kishinouye, 1923: 442-445 (description, anatomy; placed in the new genus Paralhunnus), figs. 4, 22, 47, 49. Godsil and Byers, 1944: 104-119 (anatomy; E. Pacific). Mimura et al., 1963b (biology; Indian Ocean). ^arathunnus sibi, Jordan and Hubbs, 1925: 218 (description; Japan). Jordan and Evermann, 1926: 17 (description), pi. 3, fig. 2. Tinker, 1944: 159 (Hawaii). Brock, 1949 (description; Hawaii). Shimada, 1954 (distribution in Pacific). Gosline and Brock, 1960; 261 (description; Hawaii), 337 (synonymy). Alverson and Peterson, 1963 (biol- ogy; Pacific). Parathunnus obesus, Jordan and Evermann, 1926: 17 (description). Frade, 1929: 229-235 (mor- phometry, swimbladder; Canary Is.), pi. 5, fig. 1. Buen, 1930: 50, (Spain) fig. 9. Bini, 1931 :' 27-30 (morphometry; Canaiy Is.). Frade, 1931a (mor- phometry, swimbladder; E. Atlantic). Beebe and Tee- Van, 1936: 181 (comparison with T. allanti- cus). Morice, 1953: 70-71, fig. 4 (liver; E. Atlan- tic.) Frade, 1960: 1-74 (description, distribution, biology, bibliography), pi. 1-7. Thunnus obesus, Fraser-Brunner, 1950: 142 (key to Thunnus), 144 (synonymy, in part), fig. 6. Rivas, 1951: 220 (comparison with T. allanticus and T. alalunga). Bullis and Mather, 1956 (coimts, mor- phometry, key to Caribbean species of Thunnus), fig. 2. Mather and Gibbs, 1958: 23 (NW. At- lantic). Rivas, 1961: 133-135 (description, .syn- onymy; restricted to Atlantic). Talbot and Pen- rith, 1961 : 240 and 1962: 558 (S. Africa). Talbot and Penrith, 1963: 624-629 (description, biology; S. Africa). Iwai and Nakamura, 1964: 6, fig. 3B (olfactory rosettes). Iwaietal., 1965: 9-11 (syn- onymy), 34-36 (description), fig. 19. Nakamura, 1965: 18-19, figs. 3D, 8, 9A (osteology). Merritt and Thorp, 1966: 376-377 (E. Africa). Naka- mura and Kikawa, 1966 (infracentral grooves). Paralhunnus obesus mebachi, .lones and Silas I960- 383-384 (Indian Ocean), fig. 10. Thunnus obesus sibi, Jones and Silas, 1963: 1791- 1792 (Indian Ocean). Thunnus obesus mebachi, .Jones and Silas, 1964: 36- 38 (Indian Ocean). Misidentification The 1,450-mm. specimen reported as T. thynmis by Fernandez- Yepez and Santaella (1956) is prob- ably T. obesus as indicated by Mather and Gibbs (1958: 238) and by Rivas (1961: 134). Types of Nominal Species Thynnus obesus Lowe, 1839. No type specimens. Original description rather vague, but definitely referring to the species as now recognized. Appar- ently based on large specimens from Madeira. Dif- ferentiated from T. thynnus (as Thynnus vulgaris) VXATOMY AND SYSTEMATICS OF TUNAS 109 "by the larger eye and shorter thickset figure." Pectoral fins described as reaching end of second dorsal fin, longer than in T. albacares (as Thijnnus albacora). Thy7inus sibiTcmm'mck and SMcgcl, 1844. Lec- totype, RMNH 2327 (a mounted skin, 600 mm. fork length), and paralectotype, RMNH 799 (right side of mounted skin, backed bj' cardboard, 5o7 mm. fork length) designated by Boeseman (1947). Measure- ments made l)y Gibbs in 1902 fall in tlie range of T. albacares rather than T. ohcsus, but mounted speci- mens could be expected to be unreliable for this pur- po.se. The description by Tcmminck and Schlegel likewise offers little aid in identifying the species. They note that the pectoral fin is shorter than in T. alalunga (as Thynnus alnlonga or T. pacijicns) and approaches in length that of T. albacares (as T. argcnlivittalus), and their illustration shows a pec- toral fin resembling that of a fairly small Pacific T. obcxus. On this basis, we follow other authors in considering T. sibi a synonym of T. obcsus. If the measurements of the lectotype and paralectotype were taken at face value, T. sibi would have to be regarded as a synonym of T. albacares, but we prefer for the present to disregard these specimens. Thun/ms mcbachi Kishinouyc, 1915. No t.ypo specimens. Original description clearly referable to T. obesus, apparently based on a number of specimens. Characters Pectoral fin intermediate in length (22-31 percent of fork length) in specimens longer than 1,100 mm. (as in T. albacares and T. allanticxs), as long as in T. alalunga (greater than 31 percent) in specimens less than 1,100 mm. from the Pacific. Gill rakers 23-31 (generally similar to T. albacares and T. alalunga). Liver with striations on ventral surface (not restricted to margins, fig. 30), its three lobes subeciual in length, vascular cones present on its dorsal side (as in T. thynnus, T. maccoyii, and T. alalunga). Spleen on right side, stomach on left (as in all except T. alalunga). Swimbladder as long as body cavity, with two globular dorsal heads anteriorly, tapering gradually to a jwint posteriorly. Kidney with a short tail, reaching the level of vertebra 11-13. Cutaneous artery usually originating at level of vertebra 0-8, passing laterally between ribs 5 and G, branching between intermuscular bones G anil 7 (as in T. albacares, T. tonggol, and T. atlanlicus). Two rows of arterioles and venides arising from each main lateral cutaneous branch (as in T. thynnus aiid T. maccoyii). Post-cardinal vein present, joining right cutaneous vein (as in 7'. albacares, T. tonggol, and T. atlanlicus). Posterior parasphcnoid margin either rounded (in some small specimens) or forming a slightly olituse angle (not as acute as in T. alalunga, T. maccoyii, or T. thynnus). Vertebrae 18 + 21 (as in all Thunnus except T. atlanlicus). First \-ent rally directed parapo|)hysis on vertebra 9 (as in all excciit T. tonggol ami T. tliynnu.e the posi- tion of the first ventrally directed parapophysis (on the 9th vertebra in T. maccoyii, as opposed to the 10th, in T. thynnus) and the color of the fleshy caudal keels (yellow in T. maccoyii, dark in T. thynnus). The few other characters, none of them affording complete separation, are given in table 5. The presence of T. thynnus orientalis in the south- eastern Pacific and the northeastern Indian Ocean (Nakamura and Warashina, 1965), and of T. thynnus thynnus off Cape Town (Talbot and Penrith, 1963) in the same geographical areas as T. maccoyii gives biological support to considering T. maccoyii as a separate species, although it is not known whether the two actually spawn in the same areas. Differences in the configuration of the dorsal wall of the body cavity are not apparent in specimens less than about 1,300 mm. As described by Godsil and Holmberg (1950), numerous western Atlantic speci- mens (7". t. thynnus) examined by us in the field and laboratory had a wide anterior bulge without a lat- eral concavity and had a deep, narrow trough lateral to the bulge (fig. 32). Our only large specimen of T. maccoyii (1,450 mm.) was similar to the western Atlantic forms. Eastern Pacific specimens of T. t. orientalis, 1,390 and 1,450 mm., confirm the dif- ferences described by Godsil and Holmberg. The anterior bulge is comparatively narrow, with a lateral concavity, and with a wide trough lateral to the bulge (fig. 32). Although we have dissected no large Japanese specimens, we are confident they will resemble those from the eastern Pacific. Godsil and Holmberg (1950) eliminated a large number of characters from systematic consideration. We can substantiate almost all of their conclusions, and our observations invalidate most of their remain- ing differential characters. The tubules of the caecal mass of T. t. thynnus and T. t. orientalis were said to be relatively large and T.\BLE 5. — Comparison of Thunnus maccoyii and the subspecies of T. thj'iinus IMean values given in parentheses] Character T. t. thynnits T. t. orientalis T. maccoyii Number of pill TLikers . _ 34-!3 (38. 9) 8 1.2-29.6 (8.1) 1.9-5.4 (3. ,'i) 1.5-3.2 (2.2) first increase, then decrea.se wide bulge with no lateral concavity; deep, narrow lateral trough 17. 0-21. 7 dark 32-40 (35. 9) 8 1.1-12.7 (4.1) 1.0-9.0 (4.1) 1.4-4.8 (2.0) first increase, then decrease narrow bulge with lateral concavity; wide lateral trough 16.8-20.8 dark 31^0 (33. 7) 9 0. 8-3. 2 (1.6) 6.0-1.5.9 (11.3) 0.9-1.7 (1.3) decrease wide bulge with no lateral concavity; deep, narrow lateral trough 20. 2-23. First ventrally directed parapophysis on vertebra number . 9th vertebra: p.arapophysis height/least distance apart. 10th vertebra; canal height/least width of processes.. inth vertebra: canal height/ranal width Depth of anterior haemal canals Shape of dorsal wall of body cavity in large specimens Pectoral length as percent fork length (600-1,000 mm.) Color of caudal keels AX.\TOMY AND SYSTEM.\TICS OF TUNAS 113 Figure 32. — Dorsal wall of botly cavity of Thuimus thynnus. Ventral view with viscera removed and head end to the left. Left: 7'. /. orientalis, 1,450 mm. fork length, from California, showing the comparatively narrow anterior bulge with lateral concavity and wide lateral trough. Right: T. t. thynnus, 1,850 mm. fork length, from the western \orth .Atlantic, showing the wide anterior bulge without a lateral concavity. coarse compared with T. maccoyii. We could detect no differences. The caecal mass is so variable in size that its di- mensions cannot be used to differentiate populations. The relative length of the lateral liver lobes of western Atlantic specimens encompasses the differ- ences in lobe lengths suggested by Codsil and Holmberg. The stomach length cannot logically be u.sed as a specific character, since this is a highly distensible organ, the dimensions of which will vary under different physiological states. Swimbladder dimensions vary with size, becoming larger with growth, as shown by Serventj' (1956a) for T. maccoyii from Australia and bj' us for western Atlantic T. t. thynnus and eastern Pacific T. t. orien- talis. Abe (1955) reported that the swimbladder was well developed in a 1,470-mm. specimen thought to be T. maccoyii from the eastern Indian Ocean. Kishinouye (1923) and Fradc (1925) illustrated swimbladders for western Pacific and eastern At- lantic specimens, respectively, that are ver.y similar to those of larger specimens from other regions, and Kishinouye noted that the swimbladder is short and very narrow in immature specimens of .Japanese T. thynnus orientalis, but short and wide in adults. Godsil and Holmberg described the posterior end of the kidney of T. maccoyii specimens as truncate. This condition was observed in several specimens of western Atlantic T. t. thynnus, which displayed all variations that have been described, but our speci- mens of 7'. maccoyii did not show this condition. The branching of the ureter of T. maccoyii was said to differ in that the branching occurred well anterior in the kidney mass. We observed this con- dition in both T. t. thynnus and T. t. orientalis; in 114 U.S. FISH .\XI) WILDLIFE SKHVICE two of our three specimens of T. maccoyii the branch- ing occurred near the end of the tail of the kidney. In T. maccoyii, the dorsal aorta was reported to be usually conspicuously constricted behind the ori- gin of the cutaneous arteries. We observed this condition in both T. i. thynnus and T. t. orientalis, but size did not decrease in our specimens of T. maccoyii. The presence of a connecting branch between the two main branches of the coeliacomesenteric artery was said to distinguish T. t. orientalis and T. maccoyii from T. t. thynnxis; however, Godsil and Holmberg (1950) did not find the branch in one Australian specimen and they were uncertain as to its presence in another. Furthermore, we observed this connec- tion in several western Atlantic specimens. The cutaneous artery in T. maccoyii was said to pass laterally most often between ribs 2 and 3 (rather than 3 and 4) and to divide usually between inter- muscular bones 5 and 6 (rather than 4 and 5). Godsil and Byers (1944) recorded this condition as rare in T. t. orientalis, and Godsil and Holmberg (1950) noted the same for T. t. thynnus. In all our material, including T. maccoyii, the artery passed between ribs 3 and 4, and divided between intermus- cular bones 4 and 5 (between 5 and 6 in one specimen of T. t. thynnus). The place of attachment of the internal wing of the pelvic girdle was said to be different in each of the three forms (Godsil and Holmberg, 1950). AVe found the condition in all three similar to their descriptions of T. maccoyii. THUNNUS MACCOYII (Castelnau, 1872) SOUTHERN BLUEFIN TUNA Thynnus maccoyii Castolnan, 1872: 104-105 (original description; Melljourne market). Thunnus phillipsi Jordan and Evermann, 1926: 13 (original description; New Zealand), "pi. 2, fig. 4. Thunnus maccoyii, Jordan and Evermann, 1926: 13 (description). Serventy, 1941: 27-33 (descrip- tion; Australia), fig. 5, pi. 2. Godsil and Holm- berg, 1950 (comparison of Australian with New England and California specimens; anatomy). Mimura and Warashina, 1902. Iwai and Naka- mura, 1964: 6, figs. 3E, F (olfactory rosettes). Iwai et al., 1965: 9 (synonymy), 33-34 (descrip- tion), fig. 18. Nakamura, 1965: 18, figs. 3C, 5C, 7 (osteolog.y). Nakamura and Kikawa, 1966 (infracentral grooves). Thunnus obesris, Fraser-Brunner, 1950: 144 (T. maccoyii in synonym^')- Thunnus thynnus maccoyii, Serventy, 1956a (coimts, distribution around Australia). Monro, 1958: 111 (Australia). Robins, 1963 (biologj'; Aus- tralia) . Th%innus thynnus subspecies, Serventy, 1956a: 13 (probably a separate subspecies in S. Africa). Thunnus thynnus orientalis, Jones and Silas, 1960: 381-382 (Indian Ocean), fig. 8. Collette and Gibbs, 1963: 28. Jones and Silas, 1963: 1788- 1790 (Indian Ocean). Talbot and Penrith, 1963: 630-636 (description, biology; S. Africa). Jones and Silas, 1964: 30-34 (Indian Ocean). Types of Nominal Species Thynnus maccoyii Castelnau, 1872. No type specimens. Bauchot and Blanc (1961: 377) re- ported that a type specimen was catalogued in the collections of the Museum National d'Histoire Naturelle, Paris, in 1877, as number 515, but that the specimen cannot be located. Original descrip- tion based on several specimens, fresh and dried, from the Melbourne, Australia, market, the largest 23 inches (585 mm.) long. This description is inade- quate, but the short pectoral (two-thirds of head) suggests one of the bluefin tunas or T. tonggol, and the locality rules out all except T. maccoyii as now recognized. Thunnus phillipsi Jordan and Evermann, 1926. Type originally designated as "A photograph. No. 596, Mus. Calif. Acad. Sci. ... of a specimen taken in the Bay of Islands, New Zealand." This photo- graph is of a pug-headed mounted specimen in the Dominion Museum, Wellington. The cast is 1,575 mm. FL. The pectoral fin is short (295 mm.), which makes T. maccoyii the only reasonable assignment for this nominal species, T. tonggol not being knowTi to occur in New Zealand. According to J. Moreland (pers. comm.), the pug-headedness appears to be the result of the fish being stood on its head forcing the processes of the premaxillaries up over the frontals where they remained when the cast was made. Characters Pectoral fin short, not more than 80 percent of head length, 20-23 percent of fork length in speci- mens 650-1,450 mm. (overlapping T. tonggol; slightly longer than T. thynnus). Caudal keels yel- low in most specimens: this color possibly lost in larger adults. ANATOMY AND SYSTEMATICS OF TUNAS 115 Gill rakers 31-40, more numerous than in any other species of Thunnus except T. thynnus. Liver with striations on ventral surface, its three lobes subcfjual in length, and with vascular cones on its dorsal side (as in T. alalunga, T. obrsus, and T. Ihynnvs). Spleen located on right side, and stomac-h on left (as in all except T. alalunga). Kidney with a very short "tail," reaching to the level of vertebra 8-12 (as in T. Ihynnus). Cutaneous arteries originating at level of vertebra 4-5, passing laterally between ribs 2 and 3 or 3 and 4, and dividing between intermuscular bones 4 and 5 (as in T. alalunga and T. thynnus). Two rows of arterioles and venules arising from each main lateral cutaneous branch (as in T. obcsus and T. thynnus). Post-cardinal vein absent (as in T. alalunga and T. thynnus). Posterior parasplienoid margin forming an angle, becoming acute in large specimens (as in T. alalunga, T. thynnus, and, to a lesser tlegree, in T. obcsus), occasionallj- rounded in small specimens. Alisphe- noids extending far ventrad into orbital cavity; dis- tance from most ventral part of alisphenoid to near- est point on parasphenoid goes into greatest height of anterior part of orbit two times or more (only T. thynnus and larger specimens of T. tonggol have a similar condition). Alisphenoids not known to fuse with parasphenoid as is the case in some T. thynnus. Subopercle relatively slender, its upper anterior mar- gin usually almost vertical in its lower two-fifths or more, sloping jjosteriad in its upper portion (as in T. thynnus); rarely, there is no vertical portion. Vertebrae 18 + 21 (as in all except T. atlanlicus). First ventrally directed parapophysis on vertebra 9 (as in all except T. thynnus and T. tonggol). First closed haemal arch on vertebra 10 (as in T. alalunga, T. thynnus, and, rarely, all others except T. tonggol). Anterior haemal prezygapophyses arising high on haemal arch (as in T. alalunga, T. thynnus, and T. obcsus). All haemal postzygapophyses short, less than half centrum length (as in T. alalunga, T. thynnus, and T. obesu.s). \>ntrolateral foramina small, not more than one and one-half times width of haemal spine (as in T. alalunga, T. thynnus, and 7'. obcsus). C'omijarisons with T. thynnus are given in table 5. Range Thunnus maccoyii is apparently mainly restricted to the Southern Ocean, although it is impossible to evaluate many records. This species is best recog- nized at present by skeletal characters, although the yellow caudal keel is also diagnostic. We have examined skeletal material from Tasmanian waters reported bj' Godsil and ?Iolmbcrg (1950); from west- ern South Africa, reported by Talbot and Penrith (19(33); additional specimens from the Sydney, Aus- tralia, market; from west of southern Australia; and from off the coast of Chile. The presence of T. maccoyii in the Pacific and Indian oceans off both sides of southern Australia and off Chile, and in the Atlantic off South .\frica is thus definitely estal)- lished. The geographic distril)ution summarized by Robins (19G3) included waters north of New Zealand and areas off western Australia north almost to the Indonesian Islands. These records are prob- ably correct, but need confirmation through osteo- logical studies. Southeastern Pacific catches arc reported by Jajmnese expeditions (Japan Fishery Agency, 1964). If these unconfirmed records are accepted, it seems likely that T. maccoyii will be found throughout the Southern Ocean south of 30° S. THUNNUS THYNNUS (Linnaeus, 1758) BLUEFIN TUN.\ The synonymy of each of the two subspecies is presented separately. The diagnosis of the species, discussion of nominal species and subspecies, and summary of the range consider both subspecies. THUNNUS THYNNUS THYNNUS (Linnaeus, 1758) ATLANTIC BLUEFIN TUNA Scomber pinnulis oclo ncl novem, in extremo dorso ex sulco ad pinnus vrntralcs Artedi, 1738a: 31 (descrip- tion) and 1738b: 141-142 (references to Aristotle, Ovid, Pliny, etc.). Scomber thynnus Linnaeus, 1758: 297-298 (original description; based on Artedi, 1738). Bonnaterre, 1788: 139, pi. 58, fig. 228. Gmelin, 1789: 1330- 1331 (description, synonymy). LacC'pede, 1800: 598, ()05-()32 (description, synonymy). Ilisso, 1810: 163 (Nice). Thymius thynnus, Cuvier, 1817: 313 (Mediter- ranean). Giinther. 1860: 362-363 (synonymy, description; Atlantic and Mediterranean). Thynnus mediterraneus Risso, 1826: 414-415 (substi- tute name for Scomber thynnus Linnaeus, 1758; Nice). Thynnus vulgaris Cuvier in Cuvier and Valenciennes, 1831 : 42-71 (substitution of new name for Scomber thynnus Linnaeus, 1758), pi. 210. 116 U.S. FISH AND WILDLIFE SERVICE Thunnus vulgaris, South, 1845: 620-621 (descrip- tion, natural history). Thynmis secundo-dorsalis Storer, 1967: 65-67 (origi- nal description; Massachusetts Bay), ph 12, fig. 4. Orcynus thynnus, Poey, 1875: 144-145. Liitken, 1880: 460-464, 595-596 (in part; development). Buen, 1925 (migrations, biology; E. Atlantic). Orcyniis secondidorsalis, Poey, 1875: 145 (Cuba). Albacora thynnns, Jordan, 1888. Dresslar and Fesler, 1889: 439-440 (s.ynonymy in part), pi. 7. Thunnus thynnus, Jordan and Evermann, 1896: 870 (description, synonymy in part; a single worldwide species of bluefin). Meek and Hildebrand, 1923: 314-315 (description, synonymy in part). Jordan and Evermann, 1926: 10 (synonymy; Europe). Barnard, 1927: 798-799 (S. Africa). Buen, 1930: 49 (synonymy), fig. 7. Frade, 1931b (biometrics; Portugal). Frade, 1931c (meristics; E. Atlantic). Crane, 193G (description; Gulf of Maine). Fowler, 1936: 619-630 (synonymy, description). Tortonese, 1939: 324 (Yokohama). Bellon and Bardan de Bcllon, 1949: 8-11 (Canary Is.). Smith, 1949, 1953: 298 (S. Africa), pi. 66, fig. 831. Godsil and Holmberg, 1950 (anatomy; New Eng- land). Fraser-Brunner, 1950: 142 (key to Thun- nus), 143 (synonymy in part), fig. 4. Rivas, 1951 : 217-219 (description, synonymy). Ginsburg, 1953: 1 (the name thynnus restricted to the E. Atlantic population of bluefin). Morice, 1953: 67-68, figs. 1, 2 (liver; E. Atlantic). Bellon, 1954 (description, relationships, biology, anatomy, dis- tribution). Mather and Day, 1954: 181 (W. At- lantic). Rivas, 1954b: 302-322 (spawning in straits of Florida), figs. 1-3. Rivas, 1955 (com- parison between Gulf of Maine and Florida specimens), de Sylva, 1955: 33-40 (osteology, relationships), fig. 55 (neurocranium). Bullis and Mather, 1956 (key to Caribbean species of Thun- nus). Robins, 1957 (counts on dorsal and anal fins, gill rakers; one species of bluefin in the At- lantic). Mather and Schuck, 1960 (growth; NW. Atlantic). Frade and Vilela, 1962: 17-58 (mor- phology, biology; E. Atlantic). Tiews, 1963 (biology; Atlantic.) Thunnus secundodorsalis, Jordan and Evermann, 1926: 12 (description). Jordan et al., 1930: 260. Ginsburg, 1953: 1-3 (W. Atlantic; summary of meristics from various authors). Thunnus thynnus thynnus, Serventy, 1956a: 11-13 (subspecies found along Atlantic coast of Europe). Talbot and Penrith, 1963: 633-640 (description, biology; S. Africa). Thunnus thynnus coretta, Serventy, 1956a: 11-13 (subspecies found along Atlantic coast of America). Misidentification Thynnus brachypterus Cuvier (1829) was based on illustrations by Rondelet (1554) and Duhamel du Monceau (1769). Collette (1966) has indicated that this name is a synonym of Sarda sarda (Bloch) . Al- though Cuvier (in Cuvier and Valenciennes, 1831) based his later description of T. brachypterus on speci- mens, four of which are T. thynnus and one Euthyn- nus alletteratus, this can not be regarded as the origi- nal description, and these specimens are not types. Types of Nominal Species Scomber thynnus Linnaeus, 1758. No type speci- mens. Original description not diagnostic, but based on Artedi (1738a, p. 31), who stated: "Longi- tude 7 pedum circiter." This could only refer to the bluefin tuna. Thynnus mediterrancus Risso, 1826. Substitute name for Scomber thynnus Linnaeus, 1758, and tak- ing the same type. Thynnus vulgaris Cuvier in Cuvier and Valen- ciennes, 1831 . Sub-stitute name for Scomber thynnus Linnaeus, 1758, and taking the same type. Thynnus secundodorsalis Storer, 1867. No type specimens. Original description based on two speci- mens, 8 feet, 6 inches (1,590 mm.) and 9 feet, 3 inches (1,820 mm.) total length. The pectorals "about one seventh of length of fish," the size and the locality (Mass.) unquestionably assign this nominal species to the synonymy of Thunnus thyn- nus thynnus. THUNNUS THYNNUS ORIENTALIS (Temminck and Schlegel, 1844) PACIFIC BLUEFIN TUNA Thynnus orientalis Temminck and Schlegel, 1844: 94-95 (original description; Japan). Giinther, 1860: 362 (in footnote as dubious species). Orcynus schlegelii Steindachner in Steindachner and Doderlein, 1884: 10-11 (original description; Tokyo), pi. 3, fig. 1. Thunnus thynnus, Jordan and Evermann, 1896: 870 (description and synonymy in part). Jordan et al., 1913: 121 (Japan). Walford, 1937: 7-13 (description; Pacific specimens; possibility of a single worldwide species of bluefin), color pi. 34. ANATOMY AND SY.STEM.ATICS OF TUNAS 117 Brock, 1938 (Washington). C.odsil and Byers, 1944: 88-102 (anatomy; E. Pacific), figs. 48-58. Tinker, 1944: 151 (Hawaii), pi. 1, fig. 8. Brock, 1949: 276 (key to Hawaiian tunas). Fraser- Brunner, 1950: 142-143 (synonymy in part), fig. 4. Godsil and Holmherg, 1950 (anatomy; Cali- fornia). June, 1952a (Hawaii). Buen, 1953 (Chile; but might be T. maccoyii). Iwai and Nakamura, 1964: 6, figs. 3C, D (olfactory ro- settes). Iwai et al., 1965: 3, 6-8 (synonymy), 31-33 (description), fig. 16. Nakamura, 1965: 17-18, figs. 3B, 5B, 6 (osteology). Nakamura and Warashina, 1965: 9-10 (E. Indian and SE. Pacific oceans). Nakamura and Kikawa, 19(')6 (infracentral grooves). Orcynus thynnus, Kitahara, 1897: 1 (description; Japan), pi. 1, fig. 1. Thvnmis schlegelii, Jordan and Snyder, 1900: 352 (Tokyo). Jordan and Snyder, 1901: 64 (Yoko- hama). Thvnnus oricninlis, Kishinouye, 1915: 17 (descrip- tion, anatomy; Japan), pi. 1, fig. 9. Kishinouye, 1923: 437-442 (anatomy; Japan), figs. 3, 21, 43, 44, 50. Jordan and Hubbs, 1925: 216-217 (Japan). Jordan and Evermann, 1926: 14 (de- scription). Tinker, 1944: 157-158 (Hawaii). Brock, 1949: 276 (key to Hawaiian tvmas). Gos- line and Brock, 1960: 259 (description; Hawaii), 336 (synonymy), fig. 257h. Yamanaka et al., 1963 (biology; Japan). Thunnus aaliens Jordan and Evermann, 1926: 10-11 (original description; California), pis. 1-2, figs. 1-3. Jordan et al, 1930: 259. Gin.sburg, 1953: 3 (saliens recognized as American Pacific species of bluefin). Neave, 1959 (N. end Vancouver Is.). Bell, 1963 (biology; E. Pacific). Thunnus Ihynnus oricntalis, Serventy, 1956a: 11- 13 (the subspecies found along Asiatic coast of N. Pacific). Thunnus thynrms saliens, Serventy, 1956a: 11-13 (the subspecies found along Pacific coast of N. America). Buen, 1958: 24-25 (Chile; but might be T. maccoyii). Types of Nominal Species Thynnus oricnlalis Temminck and Schlegel, 1844. Holotype RMNH 794, 450 mm. fork length, a movmted si)ecimen from .Japan with a pectoral fin 18.4 percent of fork length. Orcynus schlegelii Steindachner, 1884. Holotype (not seen by us) presumably in Vienna Museum, 360 mm. fork length. The pectoral of barely more than half the head length and the locality (Japan) enable referral of this nominal species to the synonymy of T. thynnus oricnlalis (Temminck and Schlegel, 1844). Thunnus saliens Jordan and Evermann, 1926. Type originally designated as "Xo. 595, Mus. Calif. Acad. Sci., a photograph of a specimen weighing I57I2 pounds taken ... off Catalina, California." The i)hotograph is clearly of a tuna with short pec- toral fins; Jordan and I^vermann (1926) recorded the fin length as 53^ (p. 9) or 5 (p. 10) in (standard) length, or about 20 percent. The locality allows referral to the synonymy of T. thy)inus oricnlalis. Characters Pectoral fin short, not more than 80 percent of head length, less than 23 percent of fork length, slightly shorter than in T. maccoyii at a given size (fig. 33), overlapped by T. longgol. Gill rakers 34-43 in T. I. thynnus, 32-40 in T. t. oricnlalis, more numerous than in any other species of Thunnus except T. maccoyii. T Z 1: » r^ . • .., > / 25 _ ''• 1 1 1 1 1 9* IBM HKO 210Q 2*00 2600 z z J * r- IS 1 1 1 1 1 1 OiK sou 1000 IIOO Figure .33.— Relative length of pectoral fin in Thunnus thynnus (dots) ami T. maccoyii (open circles). Liver with striations on ventral surface, its three lobes subequal in length, and with va.scular cones on its dorsal side (as in T. alalunga, T. obcsus, and T. maccoyii). Spleen located on right side, stomach on left (as in all except T. alalunga). Kidney with a very short tail, reaching to level of vertebra 8-11 (as in T. maccoyii). Cutaneous arteries originating at level of vertebra 3-6 (usually 4 or 5), passing laterally between ribs 3 and 4 (occasionally 2 and 3) and dividing between intermuscular bones 4 and 5 or 5 and 6 (as in T. alalunga and T. maccoyii). Two rows of arterioles and venules arising from each main lateral cutaneous branch (as in T. obcsus and T. maccoyii). Post- 118 U.S. FISH AND WILDLIFE SERVICE cardinal vein absent (as in T. alalunga and T. maccoyii). Posterior parasphenoid margin forming an angle, becoming acute in large specimens (as in T. alalunga, T. maccoyii, and, to a lesser degree, in T. obesus), occasionally rounded in small specimens. Alisphe- noids extending far ventrad into orbital cavity; dis- tance from most ventral part of alisphenoid to near- est point on parasphenoid goes into greatest height of anterior part of orbit two times or more (only T. maccoyii and larger specimens of T. tonggol have a similar condition). Alisphenoids fused to para- sphenoid in some larger specimens. Subopercle (see fig. 9) relatively slender, its upper anterior margin usually almost vertical in its lower two-fifths or more, sloping posteriad in its upper portion (as in T. maccoyii) ; rarely there is no vertical portion. Vertebrae 184-21 (as in all except T. atlaniicus). First ventrally directed parapophysis on vertebra 8. First closed haemal arch usually on vertebra 10 (as in T. alalunga and T. maccoyii), sometimes on 11 (as in the other five species). Anterior haemal prezyga- pophyses arising high on haemal arch (as in T". maccoyii and T. obesus). All haemal postzygapo- physes short, less than half the centrum length (as in T. maccoyii and T. obesus). Ventrolateral foramina small, not more than one and one-half times width of haemal spine (as in T. alalunga, T. maccoyii, and T. obesus). Comparisons with T. maccoyii are given in table 5. Range T. thynnus thynnus has been found in the western Atlantic from Hamilton Inlet, Labrador (La Monte, 1946 : 22) , and Newfoundland, south along the Atlan- tic coast of the United States into the Gulf of Mexico and Caribbean Sea (Wathne, 1959). It is known off Venezuela ( Fernandez- Yepez and Santaella, 1956), and south to northeastern Brazil. In the eastern Atlantic, T. t. thynnus is found from the Lofoten Islands of Norway (about 70° N.), south along the coast of Europe and north Africa, south to the Canary Islands. Records from near Cape Verde Islands, Angola, and Republic of South .\frica have been questioned (Tiews, 1963), but gill-raker counts given by Talbot and Penrith (1963) for large speci- mens caught from January to March suggest con- vincingly that T. t. thynnus does occur west of the Cape Peninsula of South Africa, and we have ex- amined one specimen from there. Tag returns have shown that there is at least some interchange between eastern and western North Atlantic T. i. thynnus. Mather (1960) reported two specimens tagged off Martha's Vineyard, Mass., and recaptured in the Bay of Biscay 2 to 5 years later. Two large specimens tagged off Cat Cay, Bahamas, were recaptured off Bergen, Norway, a distance of over 4,000 miles, after 118 and 119 days at large (Mather, 1962). T. thynnus orientalis has been reported in the eastern north Pacific from the Shelikof Straits, north of Kodiak Island, in the Gulf of Alaska (Radovich, 1961), off Vancouver Island (Neave, 1959), off Willapa Ba}' and the mouth of the Columbia River (Brock, 1938), regularly off southern California and the length of Baja Calif. (Bell, 1963). In the west- ern north Pacific, T. t. orientalis is known from the island of Sakhalin in the southern Okhotsk Sea, southward on both sides of Japan, to the northern Philippines; eastward from Japan between about 30°-40° N. to about 160° W.; and eastward between about 5°-10° N. from about 135°-175° E. (Yama- naka et al., 1963). It is taken occasionally in Hawaiian waters (Jordan and Jordan, 1922; Fowler, 1928; June, 1952a). The contention that both eastern and western north Pacific T. thynnus constitute a single sub- species is supported by the recapture off Japan of at least three specimens that had been tagged 2 to 5 years previously near Guadalupe Island, Mexico (Orange and Fink, 1963; Anonymous, 1964). Thunnus thynnus has been recorded from the Galapagos area (Snodgrass and Heller, 1905; Herre, 1930), but there is no supporting evidence which would eliminate T. maccoyii or anj^ other species from consideration. Nakamura and Warashina (1965) reported T. thynnus orientalis (as T. thynnus) from two areas previously not verified. Two specimens, 2,657 mm. and 2,200 mm., were taken in the Indian Ocean off western Australia at 28°24' S., 105°56' E. and 27°43' S., 102°25' E., respectively. .Another, 2,206 mm., was captured in the southeastern Pacific at about 37°11' S., 114°41' W. Specimens from Chile had previously been reported by Buen (1953, 1958), as T. thynnus saliens. These are areas from which T. maccoyii is known. Nakamura and Warashina gave measurements of one specimen from each locality. Converting their figures for pectoral length into percent of fork length (their "total length") gives 18.6 and 17.5 percent, falling below our data for smaller T. maccoyii and agreeing well with T. AX.\TOMY AND SYSTAMATICS OF TUNAS 119 thijnnus orientalis (table 5). All throo, however, are much larger tlian the largest roliahly identified T. 7naccoijn{\, 7 48 mm.; Iwai and Nakamura, lU()4b: 2). It is entirely possible that the two best external diagnostic chai-acters — color of caudal keel and length of pectoral fin — may no longer l)e distinct at large sizes. At the present time only examination of vertebral characters can offer assurance of their identity. We examined the skull and vertebral column of the specimen from 37°11' S., 114°41' W. The skull (L'90 mm.) is laiger than any we have examined of T. maccoyii and has the alisphenoids fused to the parasphenoids, a condition we have found only in large specimens of T. thijnnus. The first ventrally directed parapophyses are on tiie eighth vertebra and the first closed haemal arch is on the tenth vertebra as in T. Ihijnnufi. Three other vertebral characters useful in distinguishing T. tlu/nnu.^ from T. maccoijii have the following values: 9th vertebra: parapoi)hysis height divided by least distance apart — 4.2; 10th vertebra: canal height divided by least width of processes — 2.9; and 10th vertebra: canal height divided by canal width — 1.8. The first and third are higher and the second is well below the range we have found for T. macroi/n, and all agree well with our data for T. I. orientalis (table 5). I 'nfortvmately, skeletons of the suspect Indian Ocean specimens are not available, bvit specimens from this region, observed by us in the Yaizu market, appeared to have the dorsal bulge of the body cavity as in T. thynnus. THVNNUS TONGGOL (Bleeker, 1851) LONGTAIL TUNA Thynnus tonggol Bleeker, 1851 : 356-357 (original description; Batavia Sea). Giinther, 1860: 364. Thunnus rarus Kishinouye, 1915: 28 (original de- scription; Tokyo market), pi. 1, fig. 13. Neothunnus rarus, Kishinouye, 1923:448-450 (anat- omy), figs. 24-48,64. Herre, 1940:39 (Malaya). Nichols and La Monte, 1941: 32 (synonymy in part). Kishinoella rara, Jordan and Hubbs, 1925: 219 (placed in the new genus Kishinoella). Jordan and Evcrmann, 1926: 26 (description). Herre, 1945: 148 (Zamboanga, Philippines). .Veo//iimni/A/o/i(/^o/, .Jordan and Kvermann, 1926: 22. Thunnus nicolsoni Whitley, 1936: 30-31 (original description; Queensland), fig. 2. Thunnus tonggol, Tortonese, 1939: 326 (Java Sea). Fraser-Brunner, 1950: 142 (key to Thunnus), 145- 146 (synonymy), fig. 8. de Beaufort, 1951: 225- 226 (synonymy; description; Bleeker's types checked). Iwai and Nakamura, 1964: 6, fig. 31 (olfactory rosettes) . Jones and Silas, 1 964 : 38-40 (Indian Ocean). Iwai et al., 1965: 16-17 (syno- nymy), 39-40 (description), fig. 23. Nakamura, ]9()5: 24, figs. 3(i, 12, 13A (osteology). Nakamura and Kikawa, 19()() (infracentral grooves). Kishinoella tonggol, Serventy, 1941: 33-38 (descrip- tion; Australia), figs. 6-9, pi. 2. Serventy, 1942 (descrii)tion, anatomy, synonymy; Australia), fig. 1, pis. 3-5. Serventy, 1956b (counts, distribu- tion; Australia). Munro, 1958: 111 (Australia.) Jones and Silas, 1960: 384-385 (west coast of India), fig. 11. Ranade, 1961 (description; Ara- bian Sea). Jones, 1963 (biology; Indian Ocean). Jones and Silas, 1963: 1792-1793 (Indian Ocean). Misidentifications Munro (1957) reported a specimen of tuna as Parathunnus mebachi from southern Queensland. Rivas (1961) considered this specimen to be the same as his T. argentivittalus, but as we have shown under the account of T. albacares, Rivas' account and that of Schaefer and Walford (1950) is based on a specimen of T. tonggol. Judging from the low number of gill rakers (7 -f 16 = 23), pectoral length, and distance from snout to second dorsal origin rei)orted by Alunro (1957), his specimen was also T. tonggol. His later account (Munro, 1958) con- firms this opinion. Serventy (1942, 1956b), Fraser- lirunner (1950), and others have considered Kishi- noella zacalles Jordan and Evermann (1926) as close to or a synonym of T. tonggol. but zacallrs is a syno- nym of T. albacares, as we show under the account of that species. Types of Nominal Species Thynnus tonggol Bleeker, 1851. No type speci- mens known to us. The designation of a neotype by Bocseman (1964; was not in accordance with the International Code of Zoological Nomenclature (1964, .\rticle 75), which .states, among other things, that a neotype is to be designated only in connection with revisionary work, and that the designator of a neotype must give his reasons for believing all origi- nal type material to be lost or destroyed and the steps that have been taken to trace it. Since desig- nation of a neotype would .solve no nomendatorial problems, and since we have not exhaustively sought 120 U.S. FISH AND ■WILDLIFE SF.RVICE type material, we do not deem it necessary or proper to taive this action ourselves. Bleeker's original description was based on a single specimen, 650 mm. long, from "Batavia, in mari," with a pectoral fin shorter than the head and no swimbladder. The description obviously applies to the species for which the name is now used. Thunnus rarus Kishinouye, 1915. No type specimens. Original description based on a single specimen, 28.8 inches (ca. 730 mm.) long, from Nagasaki. The gill-raker count of 0+17, short pectoral fins (no measurements given), and lack of swimbladder show this nominal species to be a synonym of T. tonggol. Thmnusnicolsoni Whitley, 1936. Holotype Aus- tralian Museum lA. 6553, a 189 mm. head of a specimen originally 30 inches (762 mm.) total length caught between Lindeman and Maher Islands, Cumberland Group, North Queensland, Au.stralia. The gill raker count of 6+16 and pectoral shorter than head establish this as a synonj^m of T. tonggol. Characters Pectoral fin (see fig. 26) varying in length from medium (22-31 percent of fork length) in specimens less than 600 mm. to short (16-22 percent) in those over 600 mm. (the latter resembling only T. thijnnus and T. maccoyii). Tail region comparatively long, longest in large specimens; distance from snout to second dorsal origin 49-55 percent of fork length, decreasing with size (consistently lower than in any other Thunnus species). Gill rakers 19-26 (rarely to 28), fewer than in any other Thunnus species except T. atlanlicus. Liver without striations on ventral surface, its right lobe long and narrow, without vascular cones on its dorsal side (as in T. albacares and T. allanticus). Spleen on right side, stomach on left (as in all e.xcept T. alalunga). Kidney with a bulky anterior mass and a long, narrow tail, reaching vertebra 15-17. Swimbladder absent or rudimentary. Cutaneous arteries originating at the level of vertebra 7-8, passing laterally between ribs 4 and 5 or 5 and 6, dividing between intermuscular bones 6 and 7 (as in T. albacares, T. ohesus, and T. allanticus). A single row of arterioles and venules arising from each cutaneous branch (as in T. alalunga, T. alba- cares, and T. atlanticus), but arising from the lateral side of each vessel (as in T. albacares and T. atlan- ticus). Post-cardinal vein present (as in T. albacares, T. atlanticus, and T. obesus). Posterior parasphenoid margin not angulate (simi- lar to T. albacares and T. atlanticus). Vertebrae 18 + 21 (as in all except T. atlanticus). First ventrally directed parapophysis usually on vertebra 10. First closed haemal arch usually on vertebra 11 (as in T. atlanticus, T. albacares, T. obesus, occasionally T. thynnus) or 12. Anterior haemal prezygapophyses arising well ventrad on haemal spines (as in T. albacares and T. atlanticus). Haemal postzygapophyses long, the longest about equal to or longer than centrum length (as in T. atlanticvs, slightly longer than in T. albacares). Anteriormost ventrolateral foramina large, more than three times as wide as haemal spine (as in T. albacares and T. atlanticus). Nominal Species There appear to be only two synonyms of T. tonggol: Thunnus rarus Kishinouye from .Japan and T. nicolsoni Whitley from Queensland. Rivas (1961) placed T. nicolsoni in the synonymy of T. albacares but the gill raker count of 6+16 = 22 alone (Whitley, 1936) is enough to show that this is in- correct. Range T. tonggol is limited to the Indo-West Pacific. It is found from the western and southern coasts of Kyushyu and the southwestern part of the Japan Sea (Kishinouye, 1923, p. 449), south through the Batavia Sea (Bleeker, 1851, p. 356) to New Guinea, New Britain, and the entire north coast of Australia (Serventy, 1956b). On the Australian east coast, it is reported at least as far south as Twofold Bay, New South Wales; on the west coast it reaches at least to Cockburn Sound in the Freemantle area. Its range in the Indian Ocean (Jones, 1963) includes the Indo-Australian Archipelago, Andaman Islands, both coasts of India, southern Arabia, the Somalia coast, and the Red Sea, but it was not reported from East African waters by Williams (1964) or Merrett and Thorp (1966). Gill-raker counts indicate differences between populations in the western Indian Ocean, with a modal number of 26, and those in the eastern Indian Ocean and western Pacific, with a mode of 23. More data are necessary to corroborate this. ANATO.MY .\ND SYSTEMATICS OF TUNAS 121 LITERATURE CITED Abe, Tokiharu. 1955. Proliininary notes on the "indo-iuaKuro" (a kind of bluefin tuna) taken commercially from the eastern part of the Indian Ocean. Bull. Jap. Soc. Sci. Fish. 21(1): 20-23. Allis, EDw.\nu I'., Jr. 1903. The skull and the cranial and first spinal muscles and nerves in Scomber scomber. J. Morph. 18(1-2) : 45-328. Alverson, D.wto.v L. 1961 . Ocean temperatures and their relation to albacore tuna (Thunniis germo) distribution in waters off the coast of Oregon, Washington and British Columbia. J. Fish. Res. Bd. Canada 18 (6): 1145-1152. Alverson, Franklin G., and Clifford L. Peterson. 1963. Synopsis of biological data on bigeye tuna Parathurmvs sihi (Temminck and Schlegel) 1844. FAO Fish. Rep. 6, 2: 482-514. Anonymous. 1964. Bluefin tuna tagged off Mexico recaptured near Japan. Com. Fish. Rev. 26(12); 62-63. Artedi, Petri. 1738a. Genera piscium m Ichthyologia, sive opera omnia de piscibus. Pars III: 1-88. Lugduni Batavorum. Reprinted, 1962, Cramer, Weinheim. 1738b. Synonymia nominuni piscium in Ichthyologia, sive opera omnia de piscibus, Pars IV: 1-139. Lug- duni Batavorum. Reprinted, 1962, Cramer, Weinheim. Barnard, K. H. 1927. A monograph of the marine fishes of South Africa. Ann. S. Afr. Mus. 21(2): 419-1065. 1948. Further notes on South African marine fishes. Ann. S. Afr. Mus. 36(7): 341-406. Bauchot, M. L., and M. Blanc. 1961. Catalogue des types de Scombroidei (poissons t(516ostcens Perciformes) des collections du Mu-sOum National d'llistoire Naturelle de Paris. Mus. Nat. Hist. Natur. Paris, ser. 2, Bull. 33(4): 369-379. Beebe, William, and Gloria IIollister. 19.35. The fishes of Union Island, Grenadines, British West Indies, with the description of a new species of star-gazer. Zoologica 19(6): 209-224. Beebe, William, and John Tee-Van. 1928. The fishes of Port-au-Prince Bay, Haiti. Zoo- logica 10(1): 1-279. 1936. Systematic notes on Bermudian and West Indian tunas of the genera Parathunnus and Neothunnus. Zoologica 21(3): 177-194. Bell, Robert R. 1963. Synopsis of biological data on California bluefin tuna Thunniis saliens Jordan and Evermann 1926. FAO Fish. Rep. 6, 2: 380-421. Bell6n, Luis. 1954. Historia natural del atiin, Thunnns Ihynnus (L.). Bol. Inst. Esp. Oceanogr. 67: 1-88. Bell6n, Luis, and Emma Bardan De Bell6n. 1949. Algunos datos sobre los "Thunnidae" de Can- arias. Bol. Inst. Esp. Oceanogr. 19: 1-28. Berg, Leo S. 1940. Classification of fishes and fishlike animals, both recent and fossil. Trudy Zool. In.st. Akad. Nauk SSSR 5(2): 87-345. Lithoprinted, 1947, J. W. Edwards, Ann Arbor, Mich. |Iii Russian and English). 1955. Classification of fishes and fi.shlike animals, both recent and fossil. 2d ed., Trudy Zool. Inst. Akad. Xauk SSSH 20. 286 pp. |In Russian.! Bernabei, H. (editor). 1964. Bibliography. F.\0 Fish. Rep. 6, 4: 1853-2272. Bini, Giorgi. 1931. Rapporto sulla crociera di pesca del tonno compiuta a bordo del piropeschereccio "Orata" nell' Atlantico. Boll. Pesca, Piacicolt., Idrobiol. 7(1): 57-90. Blanc, M., and M. L. Bauchot. 1964. Les Scombroidei (Poissons Teleosteens Perci- formes) du Museum National d'Histoire Naturelle de Paris. Symposium Scombroid Fishes, Mar. Biol. Assoc. India, ser. 1 : 443-458. Bleeker, Pieter. 1851. Over eenige nieuwe geslachten en soorten van makrellachtige visschen van der Indischen Archipel. Natur. Tijdschr. Ned. Iml., 1: 341-372. Blunt, C.E., Jr. 1954. Two mid-Pacific recoveries of California-tagged albacore. Calif. Fish Game 40(3) : 339. BOESEMAN, MaRINUS. 1947. Revision of the fishes collected by Burger and von Siebold in Japan. Zool. Meded. 28: 242 pp. 1964. Scombroid types in th" Leiden Museum collec- tion. Symposium Scombroid Fishes, Mar. Biol. Assoc. India, ser. 1 : 461-468. Bonnaterre, Joseph P. 1788. Tableau encyclopcdi(|ue ct m^thodique des trois regnes de la nature. Ichthyologie. Paris, 215 pp. Broadhead, Gordon C 1959. Morphometric comparisons among yellowfin tuna, Ncothiintnis mncroplerua, from the eastern tropical Pa- cific Ocean. Inter-Amer. Trop. Tuna Comm., Bull. 3(8): 355-382 [Knglish], 383-391 [Spanish]. Brock, Vernon E. 1938. A new tuna record from Washington. Copeia 1938(2): 98. 1949. A preliminary report on Parathunnus sibi in Hawaiian waters and a key to the tunas and tuna-like fishes of Hawaii. Pacific Sci. 3(3) : 271-277. BuEN, Fernando de. 1925. Biologia de atiin Orcynus Ihynnus (L.). Result. Campan. Real. Acuer. Inter., Inst. Espaii. Oceanogr., Madrid, 1, 118 pp. 1930. Ictiologia Espanola. — I. Scombriformes y Thun- niformes. Bol. Oceanogr. Pe.scas 15(2) : 34-53. 1953. Contribuciones a la ictiologia. VII. Sobre la presencia del atun (Thunnus Ihynnus) en la costa de Chile central. Inves. Zool. Chil. 1(10): 3-4. 1958. Feces del suborden Scombroidei en aguas de Chile. Rev. Biol. Mar. Valparaiso, 7(1-3): 3-38. BuLLis, H.\RVEY R., Jr., and Frank J. Mather III. 1956. Tunas of the genus Thunnus of the northern 122 U.S. FISH AND WILDLIFE .SKUVICK Caribbean. Anier. Mus. Xatur. Hist. Novit. No. 1765: 1-12. C,\STELN.\U, Fr.\NCIS L. DE. 1872. Contribution to the ichthyology of Australia. Proc. Zool. Acclim. Soc. Victoria, 1: 29-247. Cetti, Francisco. 1777. Amfibi e pesci di Sardegna. Sassari, 208 pp. Cl.\rke, William E. 1900. Occurrence of the long-finned tunny off the Orkney Islands. Ann. Scot. Natur. Hist., 1900: 248. Clemens, Harold B. 1961 . The migration, age, and growth of Pacific albacore (Tliiimiiis germo). 1951-1958. Calif. Div. Fish Game, Fish Bull. 115, 128 pp. 1963. A model of albacore migration in the North Pacific Ocean. FAO Fish. Rep. 6, 3; 1537-1548. Clothier, Charles R. 1950. A key to some southern California fishes based on vertebral characters. Calif. Div. Fish Game, Fish Bull. 79, 83 pp. CoLLETTE, Bruce B. 1966. Revue critique des types de Scombridae des collections du Museum National d'Histoire Naturelle de Paris. Bull. Mus. Nat. Hist. Natur. Paris. In press. CoLLETTE, Bruce B., and Robert H. Gibes, Jr. 1963. A preliminary review of the fishes of the family Scombridae. FAO Fish. Rep. 6, 1 : 23-32. 1964. Thunnus South, 1845 (Pisces): proposed valida- tion under the plenary powers. Z.N. (S.) 1652. Bull. Zool. Nomencl. 21(6): 442-443. Cooper, James G. 1863. On new genera and species of California fishes — No. 1. Proc. Cahf. Acad. Sci., 3(1): 70-77. CoRwiN, Genevieve. 1930. A bibliography of the tunas. Calif. Div. Fish Game, Fish Bull. 22, 103 pp. Cowan, I.^n M. 1938. Some fish records from the coast of British Columbia. Copeia 1938(2): 97. Crane, Jocelyn. 1936. Notes on the biology and ecology of the giant tuna, Th minus thyunus Linnaeus, observed at Port- land, Maine. Zoologica 21(3): 207-212. Cunningham, J. T. 1910. On the marine fishes and invertebrates of St. Helena. Proc. Zool. Soc. London, 1910, 1: 86-127. CuviER, Georoes. 1817. La regne animal. Les reptiles, les poissons, les moUusques, et les annelides. Paris. 2, 532 pp. 1829. La regne animal. Les reptiles, les poissons, les moUusques, et les annelides. 2d ed., Paris. 2, 406 pp. CuviER, Georges, and Achille Valenciennes. 1831. Histoire naturelle des poissons. Paris. 8, 375 pp. DE Beaufort, L. F. 1951. The fishes of the Indo-Australian Archipelago. IX. Percomorphi (concluded), Blennioidea. Brill, Leiden. 484 pp. de Svlva, Donald P. 1955. The osteology and phylogenetic relationships of the blackfin tuna, Thinmus allanticus (Lesson). Bull. Mar. Sci. Gulf Carib. 5(1): 1-41. Dresslar, Fletcher B., and Bert Fbsler. 1889. A review of the mackerels (Scombrinae) of America and Europe. U.S. Fish. Comm., Bull. 7 (1887): 429-446. Duhamel du Monceau, H. L. 1769. Traite general des pfiches. Paris. Dung, Dorothy I. Y., and William F. Roy'ce. 1953. Morphometric measurements of Pacific scom- brids. [LT.S.] Fish Wildl. Serv., Spec. Sci. Rep., F'ish. 95, 170 pp. Eschricht, D. F., and J. Muller. 1837. t)ber die arteriosen und veniisen Wundernetze an der Leber und einen merkwiirdigen Bau dieses Organes beim Thunfische, Thynnus vulgaris. Abhand. Konig. Akad. Wissen. Berlin, 1835: 1-32. Fernandez-Yepez, Agustin, and Eduardo J. Sant.\ella. 1956. Presencia de atunes en Venezuela. Soc. Venez. Activ. Submarin., Pub. Ocas. Bol. 1(1): 29 pp. Fowler, Henry W. 1905. New, rare or little known scombroids. No. 1. Proc. Acad. Nat. Sci. Phila. 56: 757-771. 1928. The fishes of Oceania. Mem. Bernice P. Bishop Mus. 10, 540 pp. 1933. Description of a new long-finned tuna (Sema- thuniius guihti) from Tahiti. Proc. Acad. Nat. Sci. Phila. 85: 16:5-164. 1934. The buckler dory and descriptions of three new fishes from off New Jersey and Florida. Proc. Acad. Nat. Sci. Phila. 86: 353-361. 1936. The marine fishes of West Africa. Amer. Mus. Natur. Hist. Bull. 70, pt. 2: 607-1493. 1944. The fishes. In Results of the Fifth George Vanderbilt Expedition (1941). Acad. Nat. Sci. Phila., Monogr. 6: 57-529. Frade, Fernando. 1925. Sur I'anatomie de deux poissons scomberiformes: Thunnus thytjnus (L.) et Auxis thazard (Lacep.). Bull. Soc. Portugaise Sci. Nat. 10(1): 1-13. 1929. Sur quelques thons peu connus de I'Atlantique. Bull. Soc. Portugaise Sci. Nat. 10(20) : 227-243. 1931a. Donnces biometriques sur trois especes de thon de I'Atlantique oriental. Cons. Perm. Int. Explor. Mer, Rapp. Proc.-Verb. Reun. 70: 117-126. 1931b. Donnees biometriques pour I'etude du thon rouge de I'Algarve (cote m^ridionale du Portugal). Bull. Soc. Portugaise Sci. Nat. 11(7): 89-130. 1931c. Sur le nombre de rayons des nageoires et de pinnules branchiales chez le thon rouge Atlantique. Bull. Soc. Portugaise Sci. Nat. 11(10): 139-144. 1932. Sur les caracteres ostcologiques a utilizer pour la determination des thonides de I'Atlantique oriental et de la Mediterranoe, Comm. Int. Explor. Sci. Mer. Mediterrance Rapp. et Proc.-Verb. Reun. 7, Annexe C: 79-90. 1960. Le thon patudo Parathunnus obcsus (Lowe) et sa peche. Junta Inves. Ultramar, Estud., Ensaios, Doc. 69, 74 pp. Lisbon. Frade, Fernando, and H. Vilela. 1962. Le thon rouge et le germon, Thunniis thynnus (L.) et Germo alalunga (Bonn.). Morphologic, biologie ANATOMY AND SYSTEMATICS OF TUNAS 123 et p^che. Junta Inves. Ultramar, Estud., Ensaios, Doc. 98, 92 pp. Lisbon. FR.\SER-BnUNNER, A. 1950. The fishe.'i of the family Scombridae. .\nn. Mag. Natur. Hist., ser. 12, 3: 131-163. Ganssle, David, and Harold B. Clemens. 1953. California-tagged albacore recovered off Japan. Calif. Fish Game 39(4) : 443. Gill, Theodore. 1861. Catalogue of the fishes of the eastern coast of North America from (^ireenland to Georgia. Proc. Acad. Nat. Sci. I'hila. 1861 (Suppl.): 1-63. 1889. On the proper generic name of the tunny and albicore [sie]. Proc. U.S. Nat. Mus. 11(716) : 319-320. 1894. The proper generic name of the tunnies. Proc. U.S. Nat. Mus. 16(965): 693-694. jiNSBURG, Isaac. 1953. The taxonomic status and nomenclature of ."some Atlantic and Pacific populations of yellowfin and bluefin tunas. Copeia 1953(1): 1-10. Gmelin, Johann F. 1789. Carolia a Linn6 . . . Systema naturae. Pisces. 1(3): 1126-1516. GoDsiL, Harry C. 1948. A preliminary population study of the >-eIlo\vfin tuna and the albacore. Calif. Div. Fish Game, Fish Bull. 70: 90 pp. 1954. A descriptive study of certain tuna-like fishes. Calif. Div. Fish Game, Fish Bull. 97: 185 pp. GoDsiL, Harry C, and Robert D. Byers. 1944. A systematic study of the Pacific tunas. Calif. Div. Fish Game, Fish Bull. 60: 131 pp. GoDsiL, Harry C, and E. C. Greenhood. 1951. A comparison of the populations of yellowfin tuna, Neothuiunis tnacropteriis, from the eastern and central Pacific. CaHf. Div. Fish Game, Fish Bull. 82: 33 pp. GoDsiL, Harry C, and Edwin K. Holmberg. 1950. A comparison of the bluefin tunas, genus Tliniin i/.s\ from New England, Australia, and California. Calif. Div. Fish Game, Fish Bull. 77: 55 pp. GosLiNE, William A., and Vernon E. Brock. 1960. Handbook of Hawaiian fishes. L'niv. Hawaii Press, Honolulu, 372 pp. Gregory, William K. 1933. Fish skulls: A study of the evolution of natural mechanisms. Trans. Amer. Philos. Soc, n. s. 23(2): 75-481. GtJNTHER, Albert. 1860. Catalogue of the aeanthopterygian fishes in the collection of the British Museum. 2:1-548. London. Herre, Albert W. 1936. Fishes of the Crane Pacific Expedition. Field Mus. Natur. Hist., Zool. Ser., 21: 1-472. 1940. Additions to the fi.sh fauna of Malaya with notes on rare or little known Malayan and Bornean fi.shes. Bull. Rallies Mus. [Singapore] 16: 27-61. 1945. Additions to the fish fauna of the Philippine Islands. Copeia 1945(3): 146-149. Holmgren, Uno. 1958. On the pineal organ of the tuna, Thyyinvs thyrimis L. Mus. ( 'omp. Zool. Breviora 100, 5 pp. 1ntern.\tional Game Fish As.soci.\tion. 1965. World record marine fishes. 16 pp., Miami. 1966. World record marine fishes. 16 pp., Miami. IwAi, Tamotsu, and Izumi Naka.mura. 1964a. Branchial skeleton of the bluefin tuna, with special references to the gill rays. Bull. Misaki Mar. Biol. Inst. (6): 21-25. 1964b. Olfactory organs of tunas with special reference to their systematic significance. Hull. Misaki Mar. Biol. In.st.'(7): 1-8. IwAi, Tamotsu, Izumi Nakamura, and Kiyomatsu M.VTSUBARA. 1965. Taxonomic study of the tunas. Mi.saki Mar. Biol. Inst., Spec. Hep. 2: 1-51 [In Japanese with Eng- lish summary; Bur. Comm. Fish. Ichthyological Lab. Translation no. 3S|. Japan Fishery Agency. 1964. Report on investigations of the Shoyo Maru in the eastern Pacific (Chosasen Shoyo Maru Hokokusho). 465 pp. [In Japanese.] 1965. Report on investigations of the Shoyo Maru in the eastern Pacific (Chosasen Shoyo Maru Hoko- kusho). 417 pp. [In Japanese.] Jones, S. 1963. Synopsis of biological data on the northern bluefin tuna. Kishinoelln tonggol (Bleeker) 1851 (Indian Ocean). FAO Fish. Rep. 6, 2: 862-876. Jones, S., and E. G. Silas. 1960. Lulian tunas — a preliminary review, with a key for their identification. Indian J. Fish. 7(2) : 369-393. 1963. Tuna and tuna-like fishes from the Indian seas. F.VO Fish. Rep. 6, 3: 1775-1796. 1964. A systematic review of the scombroid fishes of India. Symposium Scombroid Fishes, Mar. Biol. Assoc. India, ser. 1: 1-105. Jordan, David S. 1888. On the generic name of the tunny. Proc. Acad. Nat. Sci. Phila. 40: ISO. Jordan, David S., and Barton W. Evermann. 1896. The fi.-ihi's of North and Middle America. U.S. Nat. Mus., Bull. 47(1), 1240 pp. 1926. A review of the giant mackerel-like fishes, tunnies, spearfishes and swordfishes. Calif. Acad. Sci., Occas. Pap. 12: 1-113. Jordan, David S., Barton W. Evermann, and Howard W. Clark. 1930. Check list of the fi.shes and fishlike vertebrates of North and Middle -Vmerica north of the northern boundary of Venezuela and Colombia. U.S. Comm. Fish., Rep. for 1928, pt. 2, 670 pp. Jordan, David S., and Carl L. Hubbs. 1925. Record of fishes obtained by David Starr Jordan in Japan, 1922. Mem. Carnegie xMus. 10(2) : 93-346. Jordan, David S., and Eric Knight Jordan. 1922. A list of the fishes of Hawaii, with notes and descriptions of new species. Mem. Carnegie Mus. 10(1): 1-92. 124 V.S. FISH .\ND WILDLIFE SKRVICE Jordan, David S., and Alvin Seale. 1906. The fishes of Samoa. U.S. Bur. Fish. Bull. 25: 173-455. Jordan, David S., and John O. Snyder. 1900. A list of fishes collected in Japan by Keinosuke Otaki, and by the United States steamer Albatross, with descriptions of fourteen new species. Proc. U.S. Nat. Mus. 2.3(1213): 335-380. 1901. A preliminary check list of the fishes of Japan. Annot. Zool. Jap. 3(2-3): 31-159. Jordan, David S., and Edwin Chapin Starks. 1907. Notes on fishes from the island of Santa Catalina, southern California. Proc. U.S. Nat. Mus. 32(1510) : 67-77. Jordan, David S., Shigeho Tanaka, and John O. Snyder. 1913. A catalogue of the fishes of Japan. J. College Sci. Tokyo Imp. Univ. 33(1): 1-497. June, Fred C. 1952a. Observations on a specimen of bluefin tuna {Thunnus tkynnus) taken in Hawaiian waters. Pacific Sci. 6(1) : 75-76. 1952b. An "unusual" yellowfin tuna {Neothininus mac- ropterus) from the waters of the northern Line Islands in the Central Pacific Ocean. Copeia 1952(3): 210-211. KiSHINOUYE, KaMAKICHI. 1915. A study of the mackerels, cybiids, and tunas. Suisan Gakkai Ho 1(1): 1-24. |In Japanese, trans- lation by W. G. Van Campen published in U.S. Fish and Wildl. Serv., Spec. Sci. Rep. Fish. 24, 14 pp.]. 1917. A new order of the Teleostomi. Suisan Gakkai Ho 2(2): 1-4. (In Japanese, translation by W. G. Van Campen published in U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 50: 1-3]. 1923. Contributions to the comparative study of the so-called scombroid fishes. J. Coll. Agr., Imp. Univ. Tokyo 8(3) : 293-475. Kitahara, Ta.saku. 1897. Scombridae of Japan. J. Imp. Fish. Bur., Tokyo, 6: 1-18. Klawe, W. L. 1959. Nuevo examen de atunes jovenes capturados frente a la costa occidental de Mexico durante la quinta expedicion George Vanderbilt (1941). Ciencia, Mexico, 18(11-12): 245-247. Kurogane, Kenji, and Yoshio Hiyama. 1957a. Morphometric comparison of the yellowfin tuna taken from the equatorial Pacific. Bull. Jap. Soc. Sci. Fish. 23(7-8): 388-393. |In Japanese with English summary.] 1957b. Morphometric characteristics of the albacore, Germo germo, from the northwest Pacific. Rec. Oceanogr. Wks. Japan 4(1): 67-75. 1958a. Morphometric comparison of the albacore from the northwest, the equatorial and the southwest Pacific. Rec. Oceanogr. Wks. Japan 4(2) : 200-209. 1958b. Morphometric comparison of the yellowfin tuna from six grounds in the Indian Ocean. Bull. Jap. Soc. Sci. Fish. 24(6-7): 487-494. 1959. Morphometric comparison of the albacore from the Indian and the Pacific Ocean. Rec. Oceanogr. Wks. Japan 5(1): 68-84. LAC^PfeOE, Le ClTOYEN. 1800-1802. Histoire naturelle des poissons. 2: 632 pp. (1800); 3: 558 pp. (1802), Plasson, Paris. La Monte, Francesca. 1946. North American game fishes. Doubleday and Co., Garden City, N.Y., 202 pp. Le Danois, Edouard. 1951. Sur la presence du thon blanc ou germon sur les cotes du Venezuela et sur le lieu de ponte de cette espece dans 1' Atlantique Nord. C. R. Acad. Sci. Paris 232: 1029-1030. 1954. Resume de nos connaissances actuelles sur I'alba- core ou yellow fin tuna (Thunnus albacores Bonnaterre). Bull. Inst. Fran?. Afr. Noire 16A(1): 283-294. Le Gall, Jean. 1949. Germon. Resume des connaissances acquises sur la biologie du germon. Rev. Trav. Off. Pechea Marit. 15(1-4): 1-42. Legand, M. 1960. Donnees biometriques sur les thons a nageoires jaunes en Nouvelle-Cal(5donie. Inst. Frang. Oceanic, Cent. Oceanogr., Noumea, Nouvelle-Caledonie, Rapp. Sci. 11: 32-54, plus 24 pp. append. Legendre, R. 1936. Presence de Parathunnus obesus dans le Golfe de Ga?cogne. Comp. Rend. Soc. Biogeog., Paris, 13 (113): 66-67. Lesson, Rene P. 1830. Les poissons. In L. I. Duperry (editor) "Voyage autour du monde execute sur la corvette "La Coquille," pendant les annees 1822-25. Vol. 2, Zoologie, Chap. 10: 66-238. Paris. Letaconnoux, R. 1951. Le germon (Germo alalunga Gmelin). Examen du nombre de branchiospines. Ann. Biol., Copen- hagen 8: 63. Linnaeus, Carolus. 1758. Systema naturae. Tom. I (Regnum animale), 10th ed., Holmiae. Lithoprinted, 1956, Brit. Mus. Natur. Hist., 823 pp. Lowe, Richard T. 1839. A supplement to a synopsis of the fishes of Ma- deira. Proc. Zool. Soc. London 7: 76-92. Reprinted 1849, Trans. Zool. Soc. London 3(1): 1-20. LtJTKEN, Christian F. 1880. Spolia Atlantica. Bidrag til kundskab om formforandringer hos fiske under deres vaext og un- vikling, saerligt hos nogle af Atlanterhavets H0js0fiske. Dansk. Vid. Selsk. Skrift., Kj0benhavn, ser. 12: 409- 613. McGary, James W., Joseph J. Graham, and Tamio Otsu. 1961. Oceanography and North Pacific albacore. Calif. Coop. Oceanic Fish. Invest., Rep. 8: 45-53. Mago Leccia, Francisco. 1958. The comparative osteology of the scombroid fishes of the genus Scomberomorus from Florida. Bull. Mar. Sci. Gulf Carib. 8(4): 299-341. ANATOMY AND SYSTEMATICS OF TUNAS 125 Marchal, Emile. 1959. Observations svir quolques Niotbuiinus albncora (Lowe) captures au large des cotes de Guin(5e. Bull. Inst. Fran?. Afr. Noire 21A(4): 1253-1274. Makr, John C, and Mii-ner B. Schaefer. 1949. Definitions of body dimensions used in describing tunas. It'.S.] Fish Wildl. Serv., Fish. Bull. 51: 241- 244. Masterman, .Arthir T. 1894. f)n the skeleton of the tunny. r2t)i .\nn. Rep. Fi.sh. Bd. Scotland: 272-283. M.\tiier, Frank J. III. 1954. Northerly occurrences of warmwater fishes in the western Atlantic. Copeia 1954(4): 292-293. 1960. Recaptures of tuna, marlin and sailfish tagged in the western North Athiiitic. C'opciu 1960(2); 149- 151. 1962. Transatlantic migration of two large bluefin tuna, J. Cons. 27(3): 325-327. 1964. Tunas (genus Thunrius) of the western North Atlantic. Part II. Description, comparison and identification of species of Thunnus based on external characters. Symposium Scombroid Fishes, Mar. Biol. As.soc. India, ser. 1: 395-410. Mather, Frank J. Ill, and C. Godfrey Day. 1954. Observations of pelagic fishes of the trojjical Atlantic. Copeia 1954(3): 179-188. Mather, Frank .1. Ill, and Robert H. Gibbs, .Jr. 1957. Distributional records of fishes from waters off New Kngland and the middle Atlantic states. Copeia 1957(3): 242-244. 1958. Distribution of the Atlantic bigeye tuna, Tlnmnus ohesiis, in the western North .\tlantic and the Caribbean Sea. Copeia 1958(3) : 237-239. Mather, Frank J. Ill, and Howard A. Schitck. 1952. Additional notes on the distribution of the l)lack- fin tuna (I'drdlhinitnix nthinticiis). Copeia 1952(4): 267. 1960. Cirowth of bluefin tinia of the western North M- lantic. I'.S. Fish Wildl. Serv., Fish. Bull, (il: 39-52. Matsumoto, Walter M. 1963. I'nique shape of the first elongate haemal spine of albacore, Thunnus nintuiiga (Bonnaterre). Copeia 1963(2) : 460-462. Meek, Seth K., ai\d Sami ei. 1'. Hii.debrand. 1923. The marine fishes of Panama. Field Mus. Natur. Hist., Publ. 215, Zool. Ser. 15, pt. I, :«0 pp. Merrett, N. R., and C. H. Thorp. 1966. A revised key to the scombroid fishes of East Africa with new observations on their biology. Ann. Mag. Natur. Hi.st., ser. 13, 8: 367-384. MiMiRA, KovA, and Ikuo Warashina. 1962. Studies on indomaguro (Thunnus mnccoyii!). Description of the development of the fishery, geo- graphical dilTerence and seasonal change of distribution and relation which is seen among indomaguro, southern bluefin, and goshomaguro distribution. Rep. Nankai Reg. Fish. Res. Lab. (16): 135-154. |ln Japanese with English summary.) MiMVRA, KoYA, and staff. 1963a. Synopsis of biological data on yellowfin tuna, Neothunnus marroptcrus Temminck and Schlegel 1842 (Indian Ocean). FAO Fi.sh. Rep. 6, 2: 319-349. 1963b. Synopsis on the biology of bigeye tuna I'arii- thunnus nicbnrhi Kishinouye 1923 (Indian Ocean). FAO Fish. Rep. 6, 2: 350-379. Morice, Jean. 1953. Un caractere systematique pouvant servir i\ separer Ics e.speces de Thunnidae atlantiques. Rev. Trav. Inst. Sci. Tech. Pcches Marit. 18, fasc. 1, no. 69: 65-74. Morrow, James K. 1954. Data on dolphin, yellowlin tuna and little tuna from East Africa. Copeia 1954(1): 14-16. Mowbray, Louis L. 1920. Description of a Thuuuua believed to be new. Copeia no. 78: 9-10. 1935. Description of the Bermuda large-eyed tuna I'lirii- Ihuunus auilniiuux, n. sp. 3 ])]). privately printed pani- [)hlet. Government Aquarium, Bermuda. MuNRO, Ian S. R. 1957. Occurrence of the big-eyed tuna, Parathunnus mebtichi (Kishinouye) in (Jueensland. Ichthyol. Notes Queensland 1: 14,5-148. 1958. Handbook of Australian fishes. (27): 109-112. Nagai, Mitsuo, and Atsushi Nakaoome. 1958. Tuna resources of the equatorial .\tlantic. Com. Fish. Rev. 20(2): 67-68. (Translated from the Japa- nese by W. G. Van Campen; origin.ally published in Tuna Fishing 42: 21-26, Kanagawa Pref. Fish. I^xjier. Sta.l Nakaoome, Jun. 1958. Morphometric comparison of yellowfin tuna for the western, middle and eastern parts of the Indian Ocean. Bull. Jap. Soc. Sci. Fish. 24 (3): 165-168. [In Japanese with Knglish summary.] Nakami'ra, IziMI. 1965. Relationships of fishes referable to the subfamily Thunninae on the basis of the axial .skeleton. Bull. Misaki Mar. Biol. Inst. (8) : 7-38. Nakamura, IztMi, and Shoii Kikawa. 1966. Infra-central grooves of tunas with special ref- erence to the identification of young tunas found in the .stomachs of large predators. Rep. Nankai Reg. I'ish. Res. Lab. (23): 55-56. Nakamura, Izi'mi, and Yukio Warashina. 1965. Occurrence of bluefin tuna, Thunnus thynnus (Linnaeus) in the eastern Indian Ocean and the eastern South Pacific Ocean. Rep. Nankai Reg. Fish. Res. Lab. (22): 9-12. Neave, Ferris. 1959. Records of fishes from waters o(T the British Columbia coa-st. J. Fi.sh. Res. Bd. Canada 16(3): 383-384. Nichols, John T. 1923. Two new fishes from the Pacific Ocean. .\mer. Mus. Natur. Hist. Novit. 94, 3 pp. Nichols, John T., and Francesca R. La Monte. 1941. Yellowfin, Allison's and related tunas. Ichthyol. Contrib. Int. Game Fish. .Vss. 1(3): 2.5-32. 126 r..'i. FISH ASD WILDIJFE SKRVICE Nichols, John T., and Robert C. Murphy. 1922. On a collection of marine fishe.s from Peru. Amer. Mus. Natur. Hist. Bull. 46(9): 501-516. Orange, Cr.\ig J., and Bernard D. Fink. 1963. Migration of a tagged bluefin tuna across the Pacific Ocean. Calif. Fish Game 49(4): 307-308. Otsu, Tamio. 1960. Albacore migration and growth in the North Pacific Ocean as estimated from tag recoveries. Pa- cific Sci. 14(3): 257-266. Otsu, Tamio, and Richard N. Uchida. 1963. Model of the migration of albacore in the North Pacific Ocean. U.S. Fish Wildl. Serv., Fish. Bull. 63: 33-44. Phillips, W. J. 1932. Notes on new fishes from New Zealand. New Zealand J. Sci. Technol. 13(4) : 226-234. PoEV, Felipe. 1868. Synopsis piscium Cubensium. Repert. Fis.-Nat. Cuba 2: 279-484. 1875. Enumeratio piscium Cubensium. Part 1. An. Soc. Esp. Hist. Nat. 4: 113-161. POSTEL, EmILE. 1950 . Note sur les thonidfis de la presqu'ile du Cap Vert. Bull. Serv. Elev. Industr. Anim. Afr. Occ. Fran?. 3(2-3): 41-76. 1955. Contribution a I'dtude de la biologie de quelques Scombridae de I'Atlantique tropico-oriental. Sta. Oceanogr. Salammbo, Ann. 10, 167 pp. 1963. Expose synoptique sur la biologie du germon Germo alalunga (Cetti) 1777 (Atlantique oriental). FAO Fish. Rep. 6, 2: 931-975. Postel, Emile, p. Fourmanoir, and P. Gueze. 1960. Sur quelques thons (.sensu lato) de la Reunion, en collection au Laboratoire des Peches Outre-Mer. Bull. Mus. Nat. Hist. Natur., ser. 2, 32(5): 392-394. Powell, A. W. B. 1937. Marine fishes new to New Zealand; including the description of a new species of Hnlieutnen. Trans. Roy. Soc. New Zealand, 67: 80-92. Radovich, John. 1961. Relationships of some marine organisms of the northeast Pacific to water temperatures particularly during 1957 through 1959. Calif. Div. Fish Game, Fish Bull. 112, 62 pp. Ranade, M. R. 1961. Notes on the tuna and frigate mackerel from Ratnagiri. J. Bombay Natur. Ili.st, Soc, 58(2); 351- 354. Regan, C. Tate. 1909. On the anatomy and classification of the scom- broid fishes. Ann. Mag. Natur. Hist. ser. 8, 3: 66-75. Risso, Antoine. 1810. Ichthyologie de Nice ou histoire naturelle des poissons de departement des Alpes Maritimes. Paris, 388 pp. 1826. Histoire naturelle des poissons de la Mediter- rance que frequentent les cotes des Alpes Maritimes, et qui vivent dans le Golfe de Nice. In Histoire natu- relle des principales productions de I'Europe m^ri- dionale. Vol. 3: 97-480. Paris and Strasbourg. ANATOMY AND SYSTEMATICS OF TUNAS RivAs, Luis R. 1951. A preliminary review of the western North Atlantic fishes of the family Scombridae. Bull. Mar. Sci. Gulf Carib. 1(3): 209-230. 1954a. [dated 1953). The pineal apparatus of tunas and related seombrid fishes as a possible light receptor controlling phototactic movements. Bull. Mar. Sci. Gulf Carib. 3(3): 168-180. 1954b. A preliminary report on the spawning of the western North Atlantic bluefin tuna (Tliunnus thyn- nus) in the Straits of Florida. Bull. Mar. Sci. Gulf. Carib. 4(4): 302-322. 1955. A comparison between giant bluefin tuna (Thun- nus thyrnivs) from the Straits of Florida and the Gulf of Maine, with reference to migration and population identity. Proc. Gulf Carib. Fish. Inst. 7th sess.: 133- 149. 1961. A review of the tuna fishes of the subgenera Parathunnus and Ncothiiniins (genus Thunnus). An. Mus. Civico Storia, Natur. Genova 72: 126-148. Robins, C. Richard. 1957. A study of bluefin tuna populations from the eastern and western North Atlantic based on an analy- sis of fin-ray and gill-raker counts. Charles F. Johnson Ocean. Gamefish Invest., Univ. Miami Mar. Lab. (mimeo) 57-13: 1-7. Robins, J. P. 1963. Synopsis of biological data on bluefin tuna Thun- nus thynnus miiccoyii (Castelnau) 1872. FAO Fish. Rep. 6, 2: 562-587. RoDBiGUES Lima, Flavio, and John P. Wise. 1963. A first examination of the abundance and distribu- tion of yellowfin and albacore tuna in the western tropical Atlantic, 1957 to 1961. FAO Fish. Rep. 6, 3: 1515-1521. Rondblet, G. 1554. Libri de piscibus marinis. Lyons. 583 pp. RoYCE, William F. 1953. Preliminary report on a comparison of the stocks of yellowfin tuna. Proc. Indo-Pac. Fi.sh. Counc. 1952, 4 (sec. 2): 1.30-145. 1965 [dated 1964]. A morphomctric study of yellowfin tuna ThunnuB albacares (Bonnaterre). U.S. Fish Wildl. Serv., Fish. Bull. 63: .395-443. Samson, V. J. 1940. Notes on the occurrence of albacore (Germo ala- lunga) in the North Pacific. Copeia 1940(4): 271. Schaefer, Milner B. 1948. Morphomctric characteristics and relative growth of yellowfin tunas {Neothurmus macroptenis) from Cen- tral America. Pacific Sci. 2(2): 114-120. 1952. Comparison of yellowfin tuna of Hawaiian waters and of the American west coast. U.S. Fish Wildl. Serv., Fish Bull. 52: 353-373. 1955. Morphometrie comparison of yellowfin tuna from southeast Polynesia, Central America, and Hawaii. Inter-Anier. Trop. Tuna Comm. Bull. 1 (4) : 91-116 [English]: 117-136 [Spanish]. Schaefer, Milner B., Gordon C. Broadhead, and Craig J. Orange. 1963. Synopsis on the biology of yellowfin tuna Thunnus 127 (Neothunmts) albacnres (Bonnaterre) 1788 (Pacific Ocean). FAO Fish. Rep. (i, 2: 538-561. ScHAEFER, MiLNEU B., and Lionel A. W.m.ford. 1950. Biometiic comijarison Ijetwecn yellowfin tunas (Neothuntms) of Angola and of tlie Pacific coast of Central America. U.S. Fish Wildl. Serv. Fish. Bull. 51 : 425-413. ScHUOK, IIow.\RD .\., and Fr.vnk J. M.vther III. 1951. A Wackfin tuna (Farnthuntius allantictis) from North Carolina waters. Copeia 1951(3): 248. ScHULTZ, Leonard P. 1960. Family Scomhridae: tunas. In L. P. Schultz and collaborators, Fislies of the Marshall and Marianas Islands. Bull. U.S. Nat. Mus. 202, 2: 410-417. Serventy, D. L. 1941. The Australian tunas. Counc. Sci., Ind. Res., Div. Fish. Rep. 104, 48 i)p. 1942. The tuna Kixhiiinclla tonggol Bleeker in Australia. J. Counc. Sci., Ind. Res. 15(2): 101-112. 1956a. The southern hluefin tuna, Thunnus Ihynnns maccoyii ((^astelnau), in Australian waters. Aus- tralian J. Mar. Fre.shwater Res. 7(1): 1-43. 1956b. Additional observations on the biology of the northern bluefin tuna, Kishinoella tonggol (Bleeker) in .Australia. Australian J. Mar. Freshwater Res. 7(1): 44-63. Shimada, Bei.i, M. 1951. An annotated bibliography on the biology of Pacific tunas. I'.S. Fish Wildl. Serv., Fish. Bull. 52: 1-58. 1954. On the distribution of the big-eyed tuna, Para- thunniis sihi, in the tropical eastern Pacific Ocean. Pacific Sci. S(l): 234-235. Sloane, Hans. 1707. A voyage to the islands Madeira, Barbados, Nieves, S. Christophers and Jamaica, w^ith the natural history of the herbs and trees, four-footed beasts, fishes, birds, insects, reptiles, etc. London, 264 pp. Smith, J. L. B. 1935. New and little known fishes from South Africa. Rec. Albany Mus. 5: 169-235. 1949. Revised 1953. The sea fishes of southern .\frica. Central News .\gency, Ltd., South ,\frica, 564 pp. Snodgra-ss, Robert Evans, and Kdmund Heller. 1905. Papers from tiie Hopkins-Stanford Galapagos Ex- pedition, 1898-1899. XVII. Shore fishes of the Revillagigedo, Clipperton, Cocos and Galapagos Is- lands. Proc. Wash. Acad. Sci. 6: 333-427. South, J. F. 1845. Thunnus. /» Smedley, Rose, and Rose (editors), Encyclopaedia Metropolitana, 25: 620-622. Squire, James L., Jr. 1963. Distribution of tuna species in the oceanic north- west Atlantic. FAO Fish. Rep. 6, 3: 1477-1514. Stakks, Edwin C. 1910. The osteology and mutual relationships of the fishes belonging to the family Sconibridae. J. Morph. 21(1): 77-99. Steindachner, Franz, and Li-nwic Doderlein. 1884. Beitrage zur Kenntnis.s dcr Fisclie Japan's. |I1I1 Denkschr. Mathemat.-Naturwiss. Kaiserlichen .\kad. Wiss. Wien, 49: 171-212. Storer, David H. 1867. A history of the fishes of Massachusetts. Cam- bridge and Boston, 287 pp. SuDA, Akira, Tsutomu Koto, and Susumu Kume. 1963. An outline of the tuna longline grounds in tlie Indo-Pacific (preliminary report). F.\0 Fish. Rep. 6, 3: 1103-1176. Takahashi, Nisuke. 1924. On the order Plecostei established by Dr. Kislii- nouye. Dobutsugaku Zasshi [Zool. Mag.l 36: 397- 408. |In Japanese, translation by W. G. Van Campen published in U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 50: 3-16.] 1926. On the Plecostei, an order of the Teleostoini, established by Prof. Kishinouye. J. College Agric, Imp. Univ. Tokyo 7(4): 383-398. Talbot, Frank H. 1964. The South African tunas. Symposium Scom- broid Fishes, Mar. Biol. As.soc. India, ser. 1: 187-209. Talbot, Frank H., and M. J. Penrith. 1961. Thunnus obcsus Lowe in South Africa. S. Afr. J. Sci. 57(9): 240. 1962. Tunnies and marlins of South .\frica. Nature 193(4815): 558-559. 1963. Synopsis of biological data on species of the genus Thunnus {sensu laio) (South Africa). FAO Fish. Rep. 6, 2: 608-646. Temminck, C. J., and H. Schlecel. 1844. Pisces. In Siebold, Fauna Japonica. 323 pp. Tiews, K. 1963. Synojisis of biological data on bluefin tuna Thunnus thynnus (Linnaeus) 1758 (.\tlantie and Medi- terranean). FAO Fish. Rep. 6, 2; 422-481. Tinker, Spencer W. 1944. Hawaiian fishes. Tongg Publ. Co., Honolulu. 404 pp. Tortonese, Enrico. 1939. Risultati ittiologici del viaggio di circumnaviga- zione del globo della R. N. "Magenta" (1865-68). Boll. Mus. Zool. .\nat. Comp. Univ. Torino, 47, ser. 3, no. 100: 177-421. TsURUTA, SaBURO. 1954. Morphometric comparison of yellowfin tuna of soutliwestern Pacific, off southwest of Gilbert Islands and of Hawaiian waters. J. Shimonoseki Coll. Fish. 3(3): 217 228. lln Japanese with English summary. 1 1955. Morphometric comparison of yellowfin tuna of southwestern Great Sunda Islands and of the Pacific waters. J. Shimonoseki Coll. Fish. 4(3): 311-319. [In Japanese with English summary.] 1961. Morphometric characters of yellowfin tuna, Neothunnus albacora (Lowe), in the southwestern waters of the Indian Ocean. J. Shimonoseki Coll. Fish. 11(2): 371-390. • 1964. Mori)hometric comparison of the yellowfin tuna, Thunnus alhacarcs (Bonnaterre), from western waters of the Line I.slands and those of the Gilbert Islands 128 U.S. FISH .\ND WII.DMFK SKKVICE and southeast of the Indian Ocean. J. Shimonoseki I'niv. Fish. 13(2): 133-140. |In Japanese with English summary.] TsuRUTA, Saburo, and Sehchi Tsunoda. 1960. Morphometrical characters and fishing condi- tions of yellowfin tuna NeoUtmnius macroplerus in the Indian Ocean. J. Shimonoeski Coll. Fish. 10(1): 1-34. |In Japane.se with English .summary.] Tucker, Denys W. 1955. The long-finned tunny, Germo alalunga (Bonna- terre), in British seas. Nature 175: 174. ViLELA, H., and Fernando Frade. 1963. E.xpose synoptique sur la biologic du thon i nageoires jaunes Neothunnus albacora (Lowe) 1839 (Atlantique oriental). FAO Fi.sh Rep. 6, 2: 900-930. ViLELA, H., and R. Monteiro. 1959. Sobre atuns de Angola. Bol. Pesca, Portugal 64: 11-54. W., S. D. [perhaps S. D. Wood], 1837. The fishes (Pisces) of Britain, systematically arranged. Analyst 5: 204-215. Walb.U'm, Johann J. 1792. Petri Artedi sueci genera piscium in quibus sys- tema totum ichthyologiae proponitur cum classibus, ordinibus, generum characteribus, specierum differen- tiis, observationibus plurimis. Grypeswaldiae. 723 pp. Walford, Lionel A. 1937. Marine game fishes of the Pacific coast from Alaska to the equator. Univ. Calif. Press, Berkeley. 205 pp. Wathne, Frederick. 1959. Summary report of exploratory long-line fishing for tuna in Gulf of Mexico and Caribbean Sea, 1954- 1957. Com. Fish. Rev. 21(4): 1-26. Watson, Marg.\ret E. 1964. Tunas (genus Thunrxus) of the western North Atlantic. Part 1. Key to the species of Thunnus based on skeletal and visceral anatomy. Symposium Scombroid Fishes, Mar. Biol. Assoc. India, ser. 1: 389-394. Whitley, Gilbert P. 1933. Studies in ichthyology. No. 7. Rec. Australian Mus. 19(1): 60-112. 1936. More ichthyological miscellanea. ^lem. Queens- land Mus. 11(1): 23-51. 1955. Taxonomic notes on fishes. Proc. Roy. Zool. Soc. New South Wales, 1953-54: 44-57. Williams, Frank. 1964. The scombroid fishes of east Africa. Sympo- sium Scombroid Fishes, Mar. Biol. Assoc. India, ser. 1 : 107-164. Yabe, Hiroshi, Shoji L'eyanagi, Shoji Kikawa, and Hisaya W.\tanabe. 1958. Young tunas fouml in the stomach contents. Rep. Nankai Reg. Fish. Res. Lab., (8): 31-48. |In Japanese, translation by Sigeru Motoda, U.S. Bureau of Commercial Fisheries, Hawaii Area Office, Hono- lulu.] Yamanaka, Ha.iime, et al. 1963. Synopsis of biological data on kuromaguro Thun- nus orienlalis (Temminck and Schlegel) 1842 (Pacific Ocean). FAO Fish. Rep. 6, 2: 180-217. YosHiDA, Howard O. 1965. New Pacific records of juvenile albacore Thunnus alalunga (Bonnaterre) from stomach contents. Pa- cific Sci. 19(4): 442-450. Yoshida, Howard O., and Tamio Otsu. 1963. Synopsis of biological data on albacore Thunnus germo (Lacepede) 1800 (Pacific and Indian oceans). FAO Fish Rep. 6, 2: 274-318. APPENDIX TON '■■■f-^.-i Figure A-1. — Lateral view of skulls of Thunnus species. (top) T. atlanticus, skull length 88 mm. ; (middle) T. tong- gol, 56 mm.; (bottom) T. tonggol, 122 mm. ANATOMY AND SYSTEMATICS OF TUNAS 129 .■< THY ALB Figure A-2.— Lateral view of skulls of Thwmus species, (top) T. alalunga, skull length 151 mm.; (middle) T. thynnm, 103 mm.; (bottom) T. thynnus, 335 mm. Figure A-3.— Lateral view of skulls of riuainits species, (top) T. obesus, skull length 200 mm.; (middle) T. alba- cares, 113 mm.; (bottom) T. alhacarcs, 164 mm. 130 U.S. FISH AND WILDLIFE SERVICE INFLUENCE OF ROCKY REACH DAM AND THE TEMPERATURE OF THE OKANOGAN RIVER ON THE UPSTREAM MIGRATION OF SOCKEYE SALMON By Richard L. Major and James L. Mighell Fishery Biologists (Research), Bureau of Commercial Fisheries Biological Laboratory Seattle, Wash. 98115 ABSTRACT Tagging experiments show that Rocky Reach Dam, constructed on the Columbia River 7 miles above Wenatchee, Wash., in 1957-61, has not appreciably increased the time required for adult sockeye salmon (Oncer hynchns nerka) to migrate to Zosel Dam on the Okanogan River (a tributary to the Columbia River above Rocky Reach Dam). Water temperature of the Okanogan River is, however, a major cause of delay. .\bove 70° F., rising or stable Okanogan River tempera- tures block the entry of the fish from the Columbia River into the Okanogan River; falling temperatures allow the migration to resume. Below 70° F., migra- tion is not blocked by rising or stable temperatures. Delay may reduce survival because it increases the exposure of the sockeye salmon to other factors that affect them adversely. INTRODUCTION If Pacific salmon {Oncorhynchus spp.) and steel- head trout (Salmo gairdneri) are to reproduce successfully, sufficient adults in spawning condition must reach the spawning grounds. Consequences can be serious if the migrants are delayed en route. Thompson (1945) showed, for example, that of the tagged sockeye salmon (Oncorhynchus nerka) that had been delayed longer than 12 days at the Hell's Gate rock slide on the Fraser River, British Colum- bia, in 1941, practically none reached their spawning grounds. He also suggested that lesser delays re- duced the reproductive capacit.y of the fish. Although thej' are equipped with facilities for passing fish, hydroelectric dams on the migration loutes constitute another type of barrier which can delay adults en route to their upstream spawning grounds. To assess and find ways to minimize the effects of the.se dams as they are built on the Colum- bia River is a most important aim of the agencies concerned with the salmon and steelhead resources of the stream. One facet of this work is to detect Note. — .Approved for publication .\pril 28, 1966. and minimize any delay of the adults as they migrate upstream. The time required for adults to locate and ascend fish ladders is sometimes reduced, for example, by altering the spill pattern to improve attraction to the ladders or even by modifying the ladders themselves. In this paper we show that Rocky Reach Dam, constructed on the Columbia River, 7 miles above Wenatchee, Wash., in 1957-61, has not appreciably increased the time required for sockeye salmon to migrate from Rock Island Dam (below Rocky Reach Dam) to Zosel Dam on the Okanogan River, a tributary to the Columbia River above Rocky Reach Dam (fig. 1). We also illustrate how the temperature of the Okanogan River periodically blocks the upstream migration of the sockeye salmon at the confluence of the Okanogan and Columbia Rivers. This studj' originated as part of a broad program to assess the effects of Rocky Reach Dam on the fish and wildlife resources of the upper Columbia River. The program was developed by repi-esentatives of interested State and Federal agencies and financed by Public Utility District Number 1 of Chelan FISHERY bulletin: VOLUME 66, NO. 1 131 5 10 15 20 25 Scale in Mi les BRITISH COLUMBIA WASH I NGTON k.1 ,Lake Vaseaux \ Similkameen R.' lOsoyoos Lake c, O v_ l<^. Lake Wenatchee Lake Chelan <9>. .Brewster -JMonse WellsDom i»V^ Chief Joseph Dam Grand Coulee Dam ^\^><<> ^5 STumwaterf ^ Dam/ \ .^-^^ ^O 'Rocky Reach Dam k,Wenatchee Rock Island Dam 132 Figure 1. — The Columbia River and the locations important to the present study. U.S. FISH AND WILDLIFE .SERVICE County, the builders of Rocky Reach Dam, in accordance with the terms of Federal Power Com- mission license number 2145. Biologists of the Bureau of Commercial Fisheries were designated to study possible delay because of their experience with a similar investigation at Rock Island Dam in 1953-5G (French and Wahle, 19G(i). The study underwent a major expansion as it proceeded. Originally, the sole aim was to measure delay, if any, caused to upstream-migrating adult salmonids by Rocky Reach Dam. This aim was to be accomplished by comparing the time required for tagged sockeye salmon to migrate from the forebay (reservoir) of Rock Island Dam to Zosel Dam, before (1957) and after (1962 and 1963) the completion of Rocky Reach Dam. If minimal influence by other factors were assumed, any major change in travel time could be attributed to the new structure. French and Wahle (1960) summarized the "pre- dam" work (performed before the dam was built). Because too few tagged fish were observed at Zosel Dam in 1957, they estimated the travel time (10.7 days) from tagged sockej^e salmon tliat had been released in the Rock Island Dam forebay and later observed at Zosel Dam during the earlier (1953-56) study. The tagging results in the first '"postdam" year (1962) had a profound impact on our investigation. Sockeye salmon migrating through the study area were noticeably delayed — apparently by high water temperatures of the Okanogan River. If a tem- perature block were shown to exist, the assumption of minimal influence by factors other than Rocky Reach Dam would have been invalidated and the straightforward comparison of "predam" and "post- dam" travel times as a measure of delay ruled out — unless the influence of the various factors could be examined separately. It was necessary, therefore, to confirm the existence of the temperature block, to ree.xamine the estimated "predam" travel time, and finally to evaluate Okanogan River temperatures as well as Rocky Reach Dam as sources of delay to sockeye salmon in their upstream migration. These expanded objectives greatly increased the complexity of the "postdam" phase of the study. METHODS AND MATERIALS The experimental procedure was as follows; (1) determine, both before and after the completion of Rocky Reach Dam, the time required for tagged sockeye salmon to migrate from the Rock Island forebay to Zosel Dam; (2) determine the time re- quired for tagged sockeye salmon to migrate from the Rock Island forebay to Rocky Reach Dam; (3) examine the variability in passage time in relation to Rocky Reach Dam and the flows and temperatures of the Okanogan and Columbia Rivers. Tagging experiments provided the answers to items (1) and (2). For item (3), these same tagging data were supplemented by counts of sockeye salmon made at Rock Island and Zosel Dams in years when there was no tagging. SOCKEYE SALMON AS THE STUDY SPECIES Observations were confined to sockeye, the only salmon that can be intercepted in significant num- bers above Rocky Reach Dam while still actively migrating to the spawning grounds. Most sockeye salmon that pass Rock Island Dam are bound for spawning areas in the Wenatchee and Okanogan River systems. Those that pass Rocky Reach Dam are, on the other hand, mostly Okanogan- bound fish, because the Wenatchee population leaves the main Columbia River below Rocky Reach Dam. After sockeye salmon pass Rocky Reach Dam, they move to the mouth of the Okanogan River near Brewster, Wash., and continue up the Okanogan into Lake Osoyoos, where they remain until they migrate to the spawning grounds, 10 to 15 miles above the lake, in late September. We have as- sumed that the Okanogan and Wenatchee popula- tions pass Rock Island Dam simultaneously. The close similarity in the shapes of the graphs of sockeye salmon counts for Rock Island and Rocky Reach Dams (examples of which are shown in figure 10, in conjunction with other data) suggests that this assumption is reasonable. TAGGING Sockeye salmon were tagged at Rock Island Dam in 1953-57, before Rocky Reach Dam was con- structed, and in 1962 and 1963, after construction. The numbers of tagged sockeye salmon that were later observed at Zosel Dam in 1953, 1954, 1962, and 1963 are presented in table 1. The effort to observe tagged sockeye salmon at Zosel Dam was so variable and so ineffective in 1955-57 (for reasons given later) that data for these years are not included. The tagging procedure was the same each year. Tagging was started when the daily count reached about one thousand fish and continued until it dropped to about one thousand near the end of the MIGRATION OF SOCKEYE SALMON 133 Table 1. — Number of sockeye salmon tagged at Rock Island Dam and obseried at Zoset Dam in 1933, 1!)S^, 1.962 and 1963 Year Tagged at Rock Island Dam Observed at Zosel Dam 1953 Number 710 1,234 1,009 730 Number 334 1954 215 1962 89 1963 193 run. In 1954 and 19G3, tagging was continued for periods of 2 to 4 consecutive days, separated by intervening 3 to 4 day periods of no tagging. Tag- ging was all at the end of the run in 1953 and was confined to the middle portion of the small run in 1 9()L'. The method of tagging was standard for these studies. Fish were trapped at Rock Island Dam in either the fish way or forebay (figs. 2 and 3), trans- ported by tank truck to the release sites, tagged, and released. Tagging time seldom exceeded 30 seconds per fish. Several types and colors of tags were used. Petersen plastic disks were used alone in 19G2 and 1963, but were used in combination with plastic bars and vinyl streamers in 1953 and 1954. Nickel pins, inserted through the body just below the dorsal fin, provided the attachment. Tags were always ap- plied in pairs, so that the same color and type of tag showed on both sides of the fish. STATIONS FOR COUNTING SOCKEYE SALMON AND OBSERVING TAGS Zosel Dam, which lies on the Okanogan River at Oroville, Wash., 1 mile below Lake Osoyoos, is the principal upstream location for the observation of tagged lish. The dam, which forms a sawmill pond, is provided with two fishways, each with a trap at its exit for the capture of upstream-migrating fish. Sockeye salmon were counted at Zosel Dam in 1935-37, 1944, 1952-57, 1962, and 1963. Since the dam was modified in 1948, however, fish have been able to i)ass upstream at certain water levels without using the fishways. When stream flow exceeds the l''i(;uiiK 2. — Fislnv:iy trap, Uock Island Dam. 134 U.S. FISH .\ND WILDLIFE .SKUVICK amount required to maintain the desired pond level, the surplus water flows either over the top of the dam or is released under lifting gates. Increased flow raises the water depth just below the dam and thereby decreases the velocity of water flowing under the gates. Under the.se conditions, we saw sockeye salmon swim upstream under the gates, especially when the water depth on the wooden apron just below the dam e.xceeded 12-18 inches. Conse- quently, the number of fish passing through the fishways was not always a reliable inde.x of the number passing upstream. Sockeye salmon have been counted and tags ob- served (when present) at Rock Island Dam since 1933 and at Rocky Reach Dam since 1961. Com- plete counts are obtained at Rock Island Dam. During midsummer, when sockeye salmon are migrating, the counting gates near the exits of the three fish ladders are open during daylight but closed at night. At Rocky Reach Dam, on the other hand, the counting gate near the exit of the single fish ladder is open 24 hours daily. Fish are counted 50 minutes per hour from 5 a.m. to 9 p.m. The 50- minute counts are multiplied by 1.2 to estimate the total hourly count. A nighttime correction factor is obtained by counting 24 hours per day once a week. All fish-count data from Rocky Reach Dam used in this report have been corrected by both the hourly and nighttime factors. STREAM FLOW AND TEMPERATURE The sockeye salmon migration between Rock Island and Zosel Dams is marked by movement from a larger, cooler river to a smaller, warmer stream. Comparative data on stream flow and temperature are, therefore, potentially important to this study. Data provided by the annual surface-water reports of the U.S. Geological Survey show, however, that the flow of both the Columbia and Okanogan Rivers generally decreases during July and August and has no apparent effect on the migration of sockeye salmon from Rock Island Dam to Zosel Dam. Figure 3. — Forebay trap, Rock Island Dam. MIGRATION OF SOCKEYE SALMON 135 Consequently, stream flow is not considered further in our analysis. It was immediately apparent, on the other hmnl, from our first observations in 1002 that the water temperature of the Okanogan River greatly influ- ences the migration of sockeye salmon bound for Lake O.soyoos. To understand the effects of tem- perature better, we have assembled migration-route water temperatures dating i)ack to 1937. Tem- jjerature was originally recorded l>y thermographs or by hand-hckl thermometers. Daily averages have been computed from the highs and lows on the thermograph charts or from the 8 a.m. and 4 i).m. thermometer readings. LIMITATIONS AND ADJUSTMENTS OF TUF. DATA Before proceeding, it is appropriate to review certain limitations and adjustments of (he data that are potential sources of error. Reliability of the Fish Counts and Tag Observations at Zosel Dam After 1948, varying proixutions of the Okanogan run passed Zosel Dam by means other than the fishways, and the reliability of the counts and tag observations recorded at Zosel Dam since 1948 varies accordingly. We used two criteria to deter- mine which data were adetiuate for this study. First, b}^ comparing the covmts at Zo.-d notes maintained by the fish coimters at Zosel Dam whether the modal counts correspontled to the modal numbers estimated to be passing the dam. Estimates of trapping eflTiciency are shown in table 2. The overall accuracy of the fish counts in 1955-57 was severely limited. Equally important, the modal counts for those years did not correspond to the modal luimbers estimated to be passing the dam. We have accordingly deleted the 1955-57 data from our analysis. Although trapping effici- ency was not much higher in 1952, 1954, and 19(12, the modal counts agreed closely with the numbers estimated by the counters to be available below the dam. The data for 19.52, 19.54, and 19(i2 have been retained, therefore, and, together with the data of high-efliciency years, 1937, 1944, 1953, and 1963, constitute the l)asis of our evaluation of the effect of water temperature on the migration of sockeye between Rock Island and Zosel Dams. Although no spawning ground estimate was made in 1963, our regular observations indicated a relatively high trapping efficiency at Zosel Dam for that j-ear. A fish-tight weir enabled the counters to make complete fish counts in 1935-37 and 1944. Although the counts of fish were comi)lete in 1935 and 1936, we cannot use them here because there are no water temperature records for those years. Table 2. — I nfonnalion on the efficiency of the trajiping system at Zosel Dam Year Count of sockeye at Zosel Dam Spawning- uround estimate' Trapping ctnciency 1952 Xitmher .■!,217 67, .542 :!, 760 4, i:io 668 2,019 944 Nu mher 25,000 :!4, 260 i:!,206 47, 930 39, 2.W 25.3.50 6,405 Perofjit 13 19.M High 1954 28 1955 9 1956 2 19.57 8 1962 15 ' Tufts, Dennis F . ami Donovan R. Craddock. 19G3. Spawnins esrapenient of Coluiuhia River sockeye salmon (O. nerka). 1902. U.S. Bur. of Conini. Fish Biol. I.ab.. Seattle, Wasli., 17 pp , Jan. 1963 (Processed). Adjustment of the Tagging Data French and Wahle (1900) estimated that, before the construction of Rocky Reach Dam, 10.7 days were required for sockeye salmon to migrate from the forebay of Rock Island Dam to Zosel Dam. Their estimate was based on a small sample of 30 tagged .sockeye salmon that had been released just above Rock Island Dam in 19.54 and 1955 and later observed at Zosel Dam. To estimate migration time, we have used the large numliers of tagged sockeye salmon (334 and 215) that were released just brlow Rock Island Dam in 19.53 and 1951, and later observed at Zosel Dam. Tagged .sockeye salmon were relea.sed above Rock Island Dam in 1962 and 1963. An adjustment was oln iously necessary before the travel time to Zosel Dam of tagged fish that had been released lielow Rock Island Dam could be compared with the travel time of tagged fish that had been released above Pvock Island Dam. We adjusted the data according to the day or days when the number of tagged fish which had been released below Rock Island Dam peaked at the counting stations of Rock Island- Dam. The dates of these i)eaks were treatetl as dates of release in the forebay. For examjile, sockeye that 136 U.S. FISH ASD WII.OI.IFIC SKRVHi; had been released below Rock Island Dam on August 2, 1953, peaked at the Rock Island counting stations on August 4, 1953. In our anal3'sis, there- fore, we treated this tagged lot as though it had been released in the Rock Island foreba.v on August 4. This adjustment made it possible to treat all tagged lots in all years as if they had been released in a common location — the forebay of Rock Island Dam. Water Temperatures Along the Migration Route Records of temperatures are frequently lacking from the Okanogan River at Mouse and the Colum- bia River at Brewster, near their confluence, but are available instead from Oroville on the Okanogan River and from Rock Island Dam and Bridgeport on the Columbia River (fig. 4). Certain features are clearly evident in figure 4. First, in July and 75 70 - r. G5 60 =- 55 75 70 1945 ,1 - ,- 'yj V ' '•- ,'' «' ;,■ T .•••r-'-^.vv';--;: Okonoqon River Oroville Monsc Columbio River . Bridgeport Brewster Rock Is Dom 1 1 t 1 1 1951 65 60 55 1952 963 A\ .'h 10 20 31 10 20 JULV AUGUST 10 20 31 10 20 JULY AUGUST Figure 4. — Comparison of Columbia and Okanogan River temperatures, 1945, 1951, 1952, and 1963. August the Okanogan River is considerably warmer than the Columbia River. Second, in terms of trends (rises and falls over a period of several days) the readings at Oroville and Rock Island (or Bridge- port) reflect the situation in the two streams near MIGRATION OF SOCKEYE SALMON their confluence. The absolute readings vary con- siderably within a river, however, particularly in the Okanogan River, where daily differences occasion- ally reach 4 or 5° F. (neither station was regularly the higher). Caution must be used, therefore, in comparing the absolute readings from the Oroville and Monse stations. The methods we describe are not rigorous, and our data are not precise. Yet we believe them to be adequate for the purposes of this report. TRAVEL TIME BETWEEN DAMS Travel time of sockeye salmon from Rock Island Dam to Zosel Dam was estimated in 1953 and 1954 before Rocky Reach Dam was built and in 1962 and 1963 after construction. Travel time from Rock Island Dam to Rocky Reach Dam was measured in 1962 and 1963, after Rocky Reach Dam was built. TRAVEL TIME BETWEEN ROCK ISLAND AND ZOSEL DAMS In the examination of the basic tagging data for this phase of the study (tables 3-4 and figs. 5-6), J^ 1953 ^ JlniL^ _ a: o _^ - \ ( A- = 2061 Ml. ■Z43I u = 2611 so* lb 20 30 J** -A^ "iS^^^ a^ElW. tWjn. iO 10 20 ' AUGUST Figure 5. — Number of tagged sockej'e observed at Rock Island and Zosel Dams after release below Rock Island Dam in 195.3 and 1954. The dates of release are designated by triangles below the base lines and the number of fish released is given in parentheses. Daily average tempera- ture of the Okanogan River is given in the center panel. attention should be given first to the years before Rocky Reach Dam was constructed. In 1953 the sockeye salmon that had been tagged below Rock Island Dam on July 31, August 1, and August 2, peaked at Rock Island Dam on August 2, 3, and 4 in that order, and at Zosel Dam 6, 9, and 9 days later. The tagged fish that were released below 137 JUL! auGUST _ JUL'' AUGUST Figure 6. — Number of tagged aockeye observed at Rockj- Reach and Zosel Dams after release in the Rock. Island forebay in 1962 and 1963. The dates of release are desig- nated by triangles below the base lines and the number of fish released is given in jiarentheses. Daily average tem- perature of the Okanogan River is given in the center panel. Rock Island Dam in 1954 have been groupod be- cause of the small numbers wliich were observed at Zosel Dam. These grouped releases peaked at Rock Island Dam on July 25, July 31, and .\up;ust 7 and at Zosel Dam 8 to 9, 7 to 8, and 7 to 8 days later. After Rocky Reach Dam was built, certain changes ai)pcar(>d in the travel time. 'I'agKcd sockeye salmon tliat had been released in the Rock Island forebay on July 19-20, 21-23, and 24-26, I !)()'_*, appeared in greatest numbers at Zosel Dam on August 11-13, 7-13, and 7-12. Travel times for the three groups were 22 to 25, 15 to 23, and 12 to 1(1 days, respectively. The travel times in 19()3, on the other hand, were similar to those before Rocky Reach Dam was built — the releases of July 10-12, 17-18, and 23-24 appeared in greatest numl)ers at Zosel Dam after 7 to 11,8 to 12, and 8 to 9 days, respectivelJ^ Table 3. — Xumber of tugged sockcyc salmon ohseri'cd at Rock Idaml and Zosel Dams after release lietoir Hock Island Dam, 1953 and 1!);14 Number tagged Point of observation.s Date of observation Date of tagging July August 21 22 23 24 25 26 27 28 29 30 31 1 2 3 4 5 6 7 8 9 10 19SS 2G1 243 20f. 490 391 353 /Rock Islandl 12 39 45 27 13 9 4 1 24 2 8 9 5 28 5 4 5 1 3 3 2 17 84 July 31 26 4 11 2 8 1 1 1 10 26 11 [Rock Islandl 9 36 40 30 IS 3 Aug. 1 12 'Rock Island 12 24 46 22 16 3 4 Aug. 2 \ZoseI Dam ._ /Rock Island ... 3 lOSi July 20-23.. .-. 2 C 19 .'»7 IM 73 47 20 2 2 22 10 11 2 35 11 6 34 16 25 3 14 23 1 1 9 3 1 7 1 4 8 1 19 5 2 48 2 I 9 96 1 /Hock I.sland 16 25 1 July 27-30 -. i;i fRock Island IS Aug. 3-6 - Table 3. — Number of tagged sockeye salmon obsereed at Rock Island and Zasil Dams after riUase below Rack Island Dnni. tn/iS and 195/f — Continued Number Tagged Point of observation Date of observation Date of tagging August 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 2S 29 30 .31 I9S3 261 243 206 490 391 353 /Rock Islandl 1 17 1 21 1 24 4 1 July 31 IZosel Dam /... Rock Islandl 10 16 10 1 10 2 5 2 3 1 4 1 4 3 2 .... 1 1 1 2 2 .... *""" *""* --.- Aug. 1 Zosel Dam /... /Rock IslandL-. \ZoseI Dam /.-- (Rock Islandl... iZosel Dam /.-. /Rock Island -.. IZosel Dam ... (Rock Island ... 1 Zosel Dam ... 10 3 5 1 .... 5 4 24 2 21 "2 .... 4 1 1 3 I 2 "' Aug. 2 1 I 3 1 1 2 2 1 I95i July 20-23 1 .... 1 July 27-30 5 4 5 4 3 2 4 10 I "20" I 5 2 1 I Aug. 3-6 - 8 5 ' .... 2 138 U.S. FI.su AM) V.Il.DI.IFK SKKVUIE The travel times between Rock Island and Zosel Dams are summarized in table 5. In view of the discrepant results of 1962 and the confinement of the 1953 tagging to the end of the run, the effect of Eocky Reach Dam on the travel time between Rock Island and Zosel Dams is best obtained from the records for 1954 and 1903. The travel times tended to be slightly longer in 1903 than in 1954, but the significance of this small difference is questionable. In any event, the difference between travel times in 1963 and 1954 is greatly overshadowed by the difference between 1962 and 1963, after Rocky Reach Dam was completed. The travel time in 1962 was much the longer and varied erratically within the season — decreasing as the season pro- gressed. This variability was not evident at Rocky Reach Dam, where the appearance of the various T.\BLE 5.— Travel time of tanged sockeye salmon from Rock Island Dam to Zosel Dam, 1953, 1954, 1062, and 1983 Year Date of tagging' Best estimate of travel time Pre-Bocky Reach: Aug. 2 Days 1953 „_ Aug 3 9 Aug. 4 .Iuly25 1954 ■ Julv 31 7-8 Aug. 7.'.".";""";;: Post-Rocky Reach: Julv 19-20.... 1962 Tlllv 21 9*^ 15-23 Julv 24-26 [July 10-12 1963 J Tlllv 17 18 8-12 IJuly23-24]""i;;; For 195.3 and 1954. the dates are those when tasged fish which had been released below Rock Island Dam reappeared in peak numbers at the countmg stations of Rock Island Dam tagged lots was orderly in lioth years (fig. 6). It is appropriate, however, to examine in greater detail T.\BLE i.—yumher of tagged sockeye salmon ohxcrved at Rocky Reach and Zosel Da 1062 and 1963 ms after release in the Rock Island Dam forehay, Date of tagging ■Number tagged Point of observation Date of observation July 11 12 13 14 15 16 17 18 19 20 4 21 10 22 24 23 16 24 8 25 5 26 5 27 1 28 29 1 30 31 19m July 19-20 177 483 349 33S 219 173 (Rocky Reach . IZosel Dam fRocky Reach . Zosel Dam f. 12 26 66 66 39 18 13 2 1 1 IZosel'llam /; fRocky ReachI. 1 Zosel Dam j /Rocky ReachI. IZo.sel Dam f. jRockv ReachI. IZosel Dam / .... -- — - — 26 38 48 43 22 6 19F,3 July 10-12 18 24 82 73 50 [J 2 6 1 20 34 1 23 53 1 16 39 1 7 7 3 July 17-18 ;::; "io" S 6 1 9 2 3 43 3 8 ' 1 -— July 23-24 --- '20' 13 13 5 12 15 5 12 1 2 2 1 — - — - — - — - — - 3 T..BI.E i.—\umher of tagged sockeye sahnon ahserred at Rocky Reach and Zosel Dams after release in the Rock Island Dam forehay 1962 and 1963— Continued Date of tagging Number tagged Point of observation Date of observation August 1 2 3 4 5 '5' 6 7 "i' 8 9 10 "i' 11 '3" 1 12 "4 13 '5' 14 15 16 17 18 19 20 21 '2 22 23 24 25 26 27 28 July 19-20 luly 21-23 177 483 349 338 219 173 fRocky Reach).... IZosel Dam / fRocky ReachI "\ ... ... — luly 24-26 IBfiS 'ulv 10-12 Rocky ReachI \Zosel Dam / fRocky ReachI... T 1 "s 2 6 3 "3' i 6 '3" 7 "4" 5 1 '2 — '" '2' '-'-'- ... 1 "i" --- — — --- — --- Fuly 17-18. \Zosel Dam / 1 Rocky ReachI 1 Zosel Dam i 1 ^ -- — ... 'i' ... — v --- ... "i" -.:: luly 23-24 j Rocky ReachI... IZosel Dam /,... 2 9 '5' "2 "2 'i' -- ... _:: ... ... ... — -- — i -■ --- "i" -- i "2' MIGRATION OF SOCKEYE SALMON 139 the movement of tagged fish between Rock Island Dam forebay and Rocky Reach Dam in 1062 and 1903. TRAVEL TIME BETWEEN ROCK ISLAND AND ROCKY REACH DAMS Fifteen lots of tagf!;o(l fish were iclcasod in the forebay of Rock Island Dam in the combined tag- ging seasons of 19fi2 and 19()3. The reai)i)earance of these lots at Rocky Reach Dam 15 miles upstream varied little; 12 peaked on the seconil day, 1 on the first day, and 2 on the third day after release. We plotted the observations of tags at Rocky Reach Dam by 4-hoiir intervals (fig. 7). Only the i n nii n n , o ii ^ (^ 60 o 50 i n In. Sow- 9am- lorn Sd"" 9gm lo» 5pm 9d" Isf Oflt OUT ^om 9om Ipm 5o* 9om Ipm 5pm 9pn> 2d DAY OUT 5om- 9Dn*- Ipm- 5pii 9om 1pm 5pm 9pm 3d DAY OUT pn bom- 90"^ IP""' ^C" ^om lO"" 5 pm 9pm 4lh DAY OUT Figure 7. — Xumbor of tagged sockeye eountoil over Rocky Reach Dam in different 4-hour periods during the 4 days after release in the Hock Island forebay in l'.)t)2 and KltiH. The periods (shown at the bottom of the graph) were 5 a.m.-9 a.m., 9 a.m.-l p.m., 1 p.m.-5 p.m., and 5 p.m.- 9 p.m. Counting was discontinued between 9 p.m. and 5.a.m. tags observed during the first 4 days after release have been included because tagged fi.sh became scarce after the fourth day. Fish tagged on .luly 17 and 19-21, 1902, have not been included because the precise time that these fish passed Rocky Reach Dam was not recorded. The agreement of tlie data for the 2 years is extremely close. For each release the greatest numbers of tags were observed at 1 p.m. to 5 p.m. on the second day after release. If we consider 10 a.m. as the average release time and 3 p.m. (the midpoint of the 1 p.m. to o p.m. period) on the second day out as the average time when tagged fish passed Rocky Reach Dam, the modal travel time was 53 hours. Diurnal consistency is 140 also evident: seven times in eight the numbers of tags observed increased from the first 4-hour period (5 a.m. to 9 a.m.) to the second (9 a.m. to 1 p.m.), peaked during the third period (1 p.m. to 5 p.m.), and decreased during the last (5 p.m. to 9 p.m.). The consistency of the data indicates that fish passage was uniform and orderly at Rocky Reach Dam in 1902 and 1903, ami that the much greater travel time from Rock Island Dam to Zosel Dam in 1902 must be attributed to longer travel time above Rocky Reach. Thus, it is necessary to look to the stretch between Rocky Reach Dam and Zosel Dam for the cau.ses of the slow travel time. EFFECTS OF W.\TER TEMPERATURE ON THE MIGRATION OF SOCKEYE SALMON On .\ugiist 1, 19()2, after sockeye had failed to appear at Zosel Dam despite high coimts at Rocky Reacii Dam, the 133-mile migration route between Rocky Reach and Zo.sel Dams was .searched by plane for schools of salmon. Despite optimum aerial-survey conditions, not a single sockeye salmon was sighted — evidence that the run had not yet entered the Okanogan River. On August 2, the following day, I (Major) visited the area on the bank of the Columbia River im- mediately adjacent to the confluence of the Okanogan and Columbia Rivers — a traditional fishing site of the Colville Indians. Of the 8 to 10 Indians present, only 1 responded to questions. lie ans- wered that "blueback (sockeye) were milling in the area and fishing was getting i)etter every day." So('keye salmon did not reach Zo.sel Dam imtil August 7 : at that time the counter leported several hundred below the dam and captured 155 in the traps, including 15 with tags. Tag recoveries in- cluded individuals from six of the .seven lots. This l)reakdown of the u.sual chronological order, and the resultant mixing and accumulation of the various segments of the rtni, indicated that the run had been delayed. Information from the aerial search and from the Indian's report pinpoints the delay at the confluence of the Okanogan and Columbia Rivers. We hypothesized that the sockeye .salmon had been blocked from the Okanogan River by unfavor- ably high water temperatures until a sharj) tem- perature drop on August 2-3 jiermitted them to enter the stream on or about .Viigust 3 and to reach Zosel Dam on August 7. To examine the validity of the general hypothesis U.S. FI.SH A.ND WILDLIFE SKRVICE as it applies to all years, we shall use fish counts and water temperatures along the migration route for 1937, 1944, 1952-54, 1962, and 1963 (table 6 and figs. 8-10). For years when sockeye salmon were tagged (1953, 1954, 19G2, and 1963), we will re- examine as part of the analysis the tagging data in figures 5 and 6. Although the data, particularly the fish counts at Zosel Dam, are not precise enough for us to make 85 - . '5 - -65 - 55I: M^ 20 30 10 20 30 '^y^- gd 1500 1962 Ms ^£ 1000 NUMBER COUNTED 8 . . . AW_. 10 20 30 10 20 30 JULf AUGUST 10 20 30 10 20 30 JULY AUGUST 10 20 30 10 20 30 Figure 10. — Numbers of sockeye salmon counted at Rock Island, Rocky Reach, and Zo?el Dams in 1962 and 1963. -Average temperatures of the Okanogan (dotted line) and Columbia (solid line) Rivers are given in the second panel from the top. Figure 8. — Number of sockej-e salmon counted at Rock Island and Zosel Dams, July and August, 1937 and 1944. Average temperatures of the Okanogan (dotted line) and Columbia (solid line) Rivers are given in the middle panel. Figure 9. — Number of sockeye .salmon counted at Rock Island and Zosel Dams, July and August, 1952, 1953, and 1954. Average temperatures of the Okanogan (dotted line) and Columbia (solid line) Rivers are given in the middle panel. a ciuantitative analysis of the relations invoh^ed, we do have evidence pertinent to our hypothesis that imfa\orable water temperatiue (or related factors) blocks the sockeye salmon from the Okanogan River until falling temperatures allow the migration to continue. Figures 8 to 10 establish clearly the general relation between temperatiu'e and movement of fish in the Okanogan River. When fish are presumably available to the river, highs and lows in the count at Zosel Dam not evident in the counts at Rock Island Dam or Rocky Reach Dam, regularly follow falling and rising temperature in the Okanogan River. This relation is particularly striking when the tem- perature drop is precedeti In' prolonged high tem- peratures. Xo influence of the temperature of the Columbia River on the upstream migration of sockeye salmon is apparent. Knowledge of the normal travel time from Rock Island Dam to Zosel Dam under favorable tem- perature conditions is highly relevant to an under- standing of the effects of unfavorable temperatures MIGRATIOX OF SOCKEYE SALMON 141 Table 6. Counts of sockeye salmon jmssing Rock Island, Zosel. mul Rocky Reach Dams, ami Okanogan and Columhia River temperatures, 1037, 1944, 1052-54, 1962, and 196.i 1937 1944 1952 1953 Date Sockeyo Temperature Sockeye Temperature Sockeye Temperature Sockeye Temperature Rock Is. Dam Zosel Dam Okan. River Col. River Rock Is. Dam Zosel Dam Okan. River Col. River Rock Is. Dam Zosel Dam Okan. River Col. River Rock Is. Dam Zosel Dam Okan. River Col. River July 1.-. 2 3 4 5.... 6 7 8 9 10 Number 1 1 I 1 2 1 8 13 1.56 425 962 1,317 1,171 1,288 925 670 738 929 595 1,125 756 479 622 461 366 268 299 Number "F. °F. 60.0 60.5 Number 3 4 8 148 56 72 60 85 128 126 110 .58 255 3.55 402 4.55 170 378 2;)7 460 136 398 137 249 83 47 66 59 34 38 17 Number "F. "F. 57. 5 56. 5 ,57.0 ,58. 5 .58.0 59.0 58.0 58.5 ,59. 5 60.5 61.0 60.5 62,0 61 . ,59. 5 60.5 62.0 61.5 62. 5 6:i.5 64.0 63. 5 63.0 64.5 64.5 64.0 04. 65. 65. 5 65. 5 6:i.o Number 68 132 329 998 739 1,019 l,2:i2 1,143 4,777 6,810 8,135 7,981 8,293 7, 200 9, .581 7, 728 3, 797 4, 790 3, 544 5, 435 .5,490 2,644 2, 8.50 2,851 2, 037 847 882 904 490 403 914 Number °F. "F. .56.5 56.5 .57.5 58.5 57. 5 57.0 .57.0 58.5 .58.5 59.5 60.5 .59.5 .59. 5 61.0 61.5 60.5 .59. 5 59.5 59.5 60.0 .59.5 59.5 60. 5 61.5 61.5 61.5 62.0 62.0 62.6 Number Number °F. 66. 5 68.0 70.0 69.0 70. 71.5 73. 5 73. 5 74.5 74.5 74.5 74.5 73.0 72.5 73.0 74.5 76.0 76.0 73. 5 73. 5 74.5 73. 72.0 71.0 70.0 70.5 71.0 71.0 72.5 71.5 70.0 "F. .56. fl .56. 5 .57.5 62.5 61.5 62.0 62.0 62.5 64.0 64.0 64.0 64.0 64.5 6.5.5 65.5 65. 5 65. 5 66.0 66. 5 M. 5 66.5 66. 5 67.5 67.5 67. 5 68. 5 68.5 68.5 68.0 68.0 67.0 .57. 5 .57.0 5 31 % 146 232 375 1,626 2,094 2, 732 7, 332 8,201 11,984 15,355 11,348 6, 240 7, 403 8,1.58 7,464 4, 106 6, 327 6, 397 5, 191 4,929 3,419 .57. 5 .58.0 .58. 5 .58. 5 71.5 72.0 71.5 70.5 71.5 71.5 72.0 72.5 73. 5 "'72' .5' 75.5 76.5 77.0 78.0 78.0 78.0 79.5 80.5 79.0 74.5 74.0 74.5 7.5. 5 76.0 76.0 74.5 74.0 73.0 72.5 72.5 71.0 72.5 69.5 69.5 70.5 72.0 72.5 73.5 75.5 ,58. 5 72.5 .59. 12 13.... 14 59. 7.3.0 73.0 73.0 74.5 75.0 74.5 73.5 74.0 73.5 73, 5 74.0 76. 77.5 77.0 78.0 7.5. 5 74.0 71.0 59. 5 5" .59.5 7 287 70 107 125 63 99 90 74 52 34 10 .58. 5 17.... 1 10 15 3 23 8 3 3 3 393 60.5 19.... 20.... 21 22.... 23.... 24.... 25.... 26.... 27 28.... 29.... 30.... 31.... 60.0 .59. 5 10 2,090 4,175 3,293 3, ,576 4,463 3, 0.59 5, 765 5,180 3,396 2,153 .59. 5 60. 5 61.0 60.5 61.0 61.0 61.5 61.5 61.5 62. 5 1 3, 692 62. 5 74.0 1 62.5 2,95:1 62. 5 Table 6.— Coum(s 0/ sockeye salmon passing Rock Island, Zosel, and Rocky Reach Dams, and Okanoijiin and Columhia River temperatures, 1937, 1944, 1952-54, 1!)S2, and 1963—Continiied 1937 1944 1952 1953 Date Sock eye Temperature Sockeye Temperature Sockeye Temperature Sock eye Temperature Rock Is Zosel Okan. Col. Rock Is. Zosel Okan. Col. Rock Is. Zosel Okan. Col. Rock Is. ZoscI Okan. Col. Dam Dam River River Dam Dam River River Dam Dam River River Dam Dam River River Number Number °F. "F. Number Number °F. °F. Number Number "F. °F. Number Number °F. "F. Aug. 1...- 2 70. 5 66.5 9 73.0 fa.i 492 11 76 5 63.5 2,161 1,891 70.5 62. 5 62 343 71 65.5 28 4 7.5.0 65.5 1.324 3 76.0 63.5 2, 499 2,823 7:i. 62.5 3... 4 200 71.5 66. 12 176 74.0 65. 5 907 1 77.0 63.5 2, 875 3,959 7:).o (a. 5 100 336 71 5 66.5 14 1.57 74.5 66. 874 77.0 63.5 2,433 1.984 69.0 63. 5 S-... 6 81 70.5 66.5 11 95 75.5 65. 5 720 2 76.0 6,3.5 1,809 2,310 69. 5 (A. 79 60 70 66. 5 6 20 66.5 743 4 74.0 6:). 5 1,651 2,076 72.6 63. 5 139 70.0 66. 5 16 39 71.5 66.5 6X6 1 73. 5 64.0 1,899 1,314 74.5 6.3. 5 g n .59 70 66. 20 17 71.0 788 74.0 64.0 1,462 1,864 71.0 63. 5 9 35 71.0 66.0 4 2 69.0 66.6 587 13 76.0 64.5 1,177 1,676 72.0 64.0 10 48 14 71.5 66.0 .54 70.5 65.0 392 68 76.5 M. 5 746 9.50 72.0 64.5 47 71.5 66.5 1 ,55 70.0 65. 5 250 136 75.5 64.5 444 882 73. 12 21 13 74.0 67.5 2 117 65. 408 55 76. 5 65. 5 436 1,676 7:1.5 (A. 5 g 72.5 68.0 6 ,57 70.0 65. 218 21 77.5 65. 807 1.551 7,5.0 64.5 14 53 22 44 32 69.0 70.5 71.5 72.5 66.0 65.5 05. 5 65. 5 5 3 4 19 31 12 11 70.0 70.0 69.5 69.5 64.5 64.0 64.5 65. 187 18:! 116 101 2 1 75.0 74.5 73.0 73. 5 W. 64.0 fa. 5 63.0 758 487 361 ;i33 1,1,52 698 376 610 75. 76.0 76. 5 72.0 64. 5 15 64.5 16 .. 17 18 .. 38 70 32 19 11 18 12 8 6 3 79 24 72.5 71.5 70.5 6.5.5 66.0 ""'w.'.V 67.0 66. 65. 5 61.5 64.5 64.0 64.0 64.0 64.0 6:1.0 3 1 2 2 1 1 3 1 1 1 13 4 1 5 2 1 1 2 70.0 69. 69.5 70.0 70.5 69. 5 70. 68.5 69.0 69.5 70.5 70.0 72.5 70.5 65. 5 65. 65. 65.5 "'«).' 6' «!. 66.0 6(i. 5 66. 5 66.5 """65.5' 138 114 12:i 175 79 42 76 34 46 :)8 34 12 14 9 148 124 104 201 93 115 169 148 186 164 119 13 60 26 71.5 71.5 71.5 6,3.0 63.0 63.0 6:i.O 63.5 63.0 62.5 62.0 62.5 62.5 62.5 6:i.o 62.5 62.5 287 226 172 74 68 146 69 47 37 29 22 60 12 9 311 209 214 .525 292 191 335 346 65 28 74 72.5 73. 5 72.0 71.0 72.0 69.0 67.5 67.0 67.0 67.0 67.0 67. 5 68.0 70.0 65. 19 20 .. 21 22 67.5 66. 66.5 66.0 65.3 6.5.3 65.3 23 24 25 . 26 27 .. 28 29... 30... 31.... 63. S 64.5 142 U.S. FI.-^H AND WILDLIFE SERVICE Table 6.- -Coiints of sochcye salmon passirig Rock Island, Zosel, and Rocky Reach Dams, and Okannrian and Columbia River teniperalures, 1937, 1944, 1952-54, 1962, and 1 963— Continued 1954 1962 1963 Date Soekeye Temperature Soekeye Temperature Soekeye Temperature Rock Is. Dam Zosel Dam Okan. River Col. River Rock Is. Dam Rocky Reach Dam Zosel Dam Okan. River Col. River Rock Is. Dam Rocky Reach Dam Zosel Dam Okan. River Col. River July 1 Number 8 15 27 39 93 136 182 272 375 398 768 715 843 1,158 2,383 3,396 4,368 5,440 3,417 2,136 3,778 4,178 5.908 5,811 5,584 4,137 3. .347 3, 753 2.815 2,913 Number °F. 64.5 64.0 66.5 68.6 70.0 69.5 68.5 68.0 69.0 65.5 67.0 68.0 70.0 72.0 73.5 74.5 74.5 74.0 72.5 68.0 68.5 69.0 69.0 70.5 72.0 71.5 69.0 68.0 69.5 71.0 73.5 "F. Number 40 16 23 36 54 69 90 163 96 140 292 389 416 638 932 745 788 1,272 1,587 1,733 2,541 2,133 1,702 1,956 1,237 1,544 1,160 880 8.53 573 666 Number 4 8 12 26 28 22 43 66 66 66 140 166 235 415 334 474 553 454 671 800 708 803 856 683 793 670 480 460 315 Number "F. °F. 58.5 58.5 58.5 58.5 58.0 58.5 59.0 59.5 60.0 60.5 60.5 61.5 61.0 61.0 61.5 61.6 61.0 60.5 60.6 61.0 61.0 61.5 62.5 62.5 62.5 63.0 63.6 64.0 64.0 63.0 62.5 Number 285 287 484 521 1,02:! 1,401 1,719 1,291 1,238 1,819 1,971 2,893 4,567 4,548 4,601 2,986 3, 300 2, 653 3,572 3,116 2, 734 2, 725 1,814 886 841 475 633 1,015 1,267 1,138 753 Number 121 131 196 299 406 732 1,123 602 1,277 807 1,426 1,337 2.042 2,981 3, 636 2,891 2,367 2,321 2,171 2,888 2,783 2,127 2,308 1,901 932 438 315 307 215 479 546 Number °F. °F. 57.5 59.0 69.0 59.5 59.5 69.5 69.0 59.5 59.0 58.5 59.0 59.5 59.5 59.6 59.6 60.5 60.5 60.5 61.0 61.0 61.5 61.5 61.6 60.6 60.6 60.5 60.5 61.0 62.0 62.0 62.0 2 3 64.0 66.5 67.5 67.0 66.0 64.0 62.6 63.5 64.0 62.0 63.0 65.6 65.5 66.0 67.0 67.0 68.0 68.5 68.6 68.0 66.0 65.5 64.0 65.6 68.0 68.6 69.5 69.6 70.0 4 5 6 7 56.5 57.0 56.5 57.0 56.5 57.0 57.0 57.5 58.5 58.5 58.5 58.5 58.5 58.5 68.5 58.5 59.5 59.5 59.5 60.5 60.5 60.5 60.5 60.5 61.0 8 9 10 11 12 13 14 15 16 114 148 187 810 898 1,018 766 626 776 971 968 934 1,900 1,233 768 482 17 18 19 20 21 22 23 9 117 .101 110 87 74 61 117 226 24 51 32 1 1 8 25 26 27 _ 28 29 30 31 Table 6.- -Coutns of soekeye salmoti passing Rock Island, Zosel, and Rocky Reach Rams, and Okanogan and Columbia Rirer iemperaturcs, 1937, 1944, 1952-54, 1962, and 1963— Continued 1964 1962 1963 Date Soekeye Temperature Soekeye ' Temperature Soekeye Temperature Rock Is. Dam Zosel Dam Okan. River Col. River Rock Is. Dam Rocky Reach Dam Zosel Dam Okan. River Col. River Rock Is. Dam Number 490 380 491 706 513 469 360 222 242 197 152 132 154 192 189 58 74 70 67 51 71 48 55 33 38 53 28 32 28 13 12 Rocky Reach Dam Zosel Dam Okan. River Col. River Auir. 1 2 3 4 5 6 7 8 9 10 11 Number 2.158 2. 573 1.654 1.506 1,542 1,415 1,832 1,575 1,317 975 913 762 627 718 502 332 320 349 279 177 129 173 106 144 67 78 54 44 40 39 33 Number 386 311 159 88 130 146 234 280 234 126 106 103 116 118 104 47 57 45 42 46 "F. 74.5 73.0 69.0 68.0 69.5 71.0 72.0 71.0 68.5 70.0 70.0 70.0 72.0 71.5 70.0 68.0 68.5 69.0 68.0 68.0 69.5 69.5 66.5 65.5 65.5 65.6 66.0 67.0 68.0 69.6 67.5 "F. 61.5 61.5 61.6 61.0 61.0 61.5 61.5 62.0 62.5 62.0 61.6 61.5 62.5 62.0 61.5 61.5 61.5 62.0 62.0 62.5 62.6 62.0 61.6 62.0 61.5 60.5 61.0 61.0 61.5 61.5 62.0 Number 503 583 489 360 273 196 174 168 94 76 116 116 100 67 65 101 59 66 52 63 43 61 38 30 38 29 54 39 38 36 17 Number 344 365 298 232 400 139 142 112 109 73 61 73 81 61 50 46 45 64 18 18 17 22 15 42 14 8 8 18 9 11 8 Number 2 2 16 155 102 100 61 1.35 96 46 11 31 17 15 3 14 6 24 16 "F. 76.5 76.0 70.0 70.5 69.5 69.5 67.5 68.0 70.0 69.5 70.5 70.0 70.0 71.0 73.0 71.5 70.0 69.5 71.0 70.0 08.5 °F. 62.5 61.5 62.0 62.5 61.5 62.0 62.0 62.5 61.5 61.0 60.5 60.5 61.5 61.5 62.0 62.0 62.5 62.0 62.0 62.5 62.5 62.5 62.5 62.5 62.5 62.5 61.5 61.5 61.5 62.0 62.5 Number 594 198 484 327 309 339 241 219 183 172 128 78 68 61 42 55 37 10 9 12 16 8 4 4 29 5 26 52 9 49 25 Number 366 281 306 248 83 20 101 149 31 29 .38 43 116 42 47 12 159 257 107 36 117 141 166 170 126 108 99 48 °F. 69.5 71.0 71.6 73.0 73.0 74.0 74.0 7.5.5 76 76.0 74.0 7.5.5 74.5 74.6 71.0 71.5 71.5 72.0 69.5 69.0 68.0 69.0 67.0 66.0 "F. 62.0 62.0 62.5 63.0 63.5 64.0 64.0 64.5 64.5 64.5 14 15 16 17 18 19 20 21 64.5 63.5 63.5 64.0 63.5 63.0 63.5 62.5 62.5 62.8 62.5 62.5 62.5 62.0 63.0 62.5 62.5 63.0 22 23 24 3 12 26 10 10 9 14 17 25 26 27 28 29 30 31 MIGRATION OF SOCKEYE SALMON 143 on the migration of sockeye salmon. As has l)een brought out earlier, the travel time was 8 to 9, 7 to 8, and 7 to 8 days for the three groups taggcMl in 1954 before Rocky Reach Dam was constructed. In 1963, after Rocky Reach Dam was completed, the most frequent time was 9 days. On the basis of these travel times, we may assume that the sockeye salmon arrive at the mouth of the Okanogan Rivor, SO miles above Rock Island Dam or roughly hallway on the 154-mile distance from Rock Island Dam to Zosel Dam, on the fourth or fifth day after I)assing Rock Island Dam. Accordingly, the re- maining 3 to 5 days of the typical 8- or 9-day total migration time are spent negotiating the 74-mile route from the confluence to Zosel Dam. Under this assumption, we can retrace the migra- tion of certain segments of the runs in several years. In 1937, for example, fish were abundant at Rock Island Dam beginning July 13. Had the Okanogan- bound segment of this run moved without delay, we would have expected it at Zosel Dam beginning July 21. Yet, the counts at Zosel Dam remained practically nil until July 31—3 days after the temperature of the Okanogan River began a sharp decline. Prior to July 28, the temperature had been either relatively stable or climbing sharply; either condition apparently delayed the fish — some as long as 10 days. The events of 1937 were essentially repeatet! in 1944; fish counts increased significantly at Rock Island beginning July 13, 1944, but not at Zosel Dam until August 3 — 5 days after the water tem- perature decreased on July 29 and up to 13 days later than expected. Then, following several days of high counts, the number of sockeye salmon arriving at Zosel Dam dropped sharply, finally reaching a low of two fish on August 9. These low counts corresponded with rising or stable water temperatures. The count surged again on August 10 — 3 days after a sharp drop in water temperature. Fish migration was similar in 1952. Counting of fish began at Zosel Dam on July 15, but no sockeye salmon were seen until the evening of July 20, when seven were taken in the traps. The count increased markedly the next day, and sockeye salmon were relatively abundant the following 9 days. This increased abundance of fish began 5 days after the beginning of a temperature drop which eventually lasted 9 days. Then, beginning July 25, the tem- perature rose steadily until .\ugust 5, when it began to decline. This second rise in temperature brought a second period of low counts which lasted until August 10 — 5 days after the temperature fell on August 5. Movement was suppressed a third time by a general 6-day rise in temperature from August 8 to 13. A decrease of tempiM-ature on .\ugust 14 resulted in a surge of fish at Zosel Dam on August 18 — 4 days later. Thereafter, the temperature tlioppcd steadily and, judging by the counts, fish migration through tiie area was unimpeded. The counts at Zosel Dam in 1953 generally re- flected the counts at Rock Island. This agreement probably occurred because increases in water tem- perature were short (2-4 days) and were followed by falling temperatures during the time when the greater portion of the run was migrating through the critical area. A late-season temperature rise on August 6-7, 1953, coincided with the arrival of tagged fish. This event provides the first opportunity to study the effects of water temjicrature in terms of a marked segment of the population. For this analysis, we need to refer to figure 5, which depicts the Okanogan River temperatures and the movement of tagged fish from Rock Island to Zosel Dam. On the as- sumption that the normal travel time from Rock Island Dam to the confluence of the two rivers is 4 to 5 days, we reason that some individuals from the lot tagged on July 31 reached the confluence on August 4 or 5, liefore tlie temperature rise of August (1-7. These early arrivals peaked initially at Zosel Dam on August 8. The lising water tem- peratures on August 6-7 suppressed entry of the later arrivals into the Okanogan River until a drop of temperature on August 8. This decrease of temperature resulted in another surge of tagged fish at Zosel Dam on August II to 12, and gave the count of tagged fish at Zosel a bimodal distribution not evident at Rock Island. Apparently, few fish from the lot tagged on August 1 arrived at the river mouth before the temperature rose; most arrived during the rise of August 6 to 7 and therefore did not appear at Zosel Dam until August 12—4 days after the temperature fell on August 8. Similarly, few fish from the lot tagged on August 2 arrived at the confluence l)cfore the temperature rise suppressed their entry. Most of the.se fish arrived at Zosel Dam on August 12 to 13 — 4 or 5 days after the tem- perature drop. The movement of the various tagged segments of the 1951 run can also be retraced from figure 5. 144 U.S. I'isn AND WII.DUFK SERVIfn Fish released below Rock Island Dam on July 20-23 peaked at Rock Island Dam on July 25, and at Zosel Dam on August 2-3, 8 to 9 days later. From these records we infer that most of the tagged fish reached the Okanogan River on July 29, when the temperature was just beginning to increase, and that this increase did not suppress entry to the stream. Similarly, the collective releases of July 27-30 peaked at Rock Island Dam on July 31 to August 1 and at Zosel Dam on August 8. On the assumption of 4 days for travel, we estimate that most of these fish arrived at the mouth of the Okanogan on August 5-6 when the temperature began a new rise. Again, however, migration was not affected. Final- ly, the August 3-6 releases of tagged fish peaked at Rock Island Dam on August 7-8 and probaljly arrived at the mouth of the Okanogan River on August 11-12, when temperatures were fairly stable. These fish also appear to have migrated freely through the Okanogan River. A possible explanation of the normal progress of the migration despite rising temperatures is given below in our discussion of the 1963 migrations. Temperatures of the Okanogan River were not recorded in 1962 until August 1. High air tem- peratures indicate, however, that the water tem- perature almost surely had been rising prior to August 1. Rising temperature would account for the July 26 to August 5 lull in the count at Zosel Dam (figs. 6 and 10). On the basis of counts at Rocky Reach Dam, we would have expected the arrival of sockeye salmon, both tagged and untagged, at Zosel Dam during this interval. The arrival of the fish at the mouth of the Okanogan River coin- cided, however, with rising water temperatures, a condition which apparently blocked their entry. The 1963 migration progressed from Rocky Reach Dam to Zosel Dam without major delay (figs. 6 and 10). Tag recoveries were orderly in contrast to those of 19()2; peaks at Zosel Dam followed com- parable peaks at Rocky Reach Dam by 6, 6 to 8, 6, 7 to 9, 6 to 9, and 6 to 10 days. Thus, migration was normal in 1963, despite generally rising water temperatures. The ab.sence of delay in 1954 and 19(53, despite rising water temperatures, focuses attention on the importance of the level at which the temperature is changing. For example, migration was unimpeded by rising temperatures in the 62° to 69° F. range in 1963, but was halted by rises in the 75° to 78°, 70° to 77°, and 74° to 78° F. ranges in 1952.= Furthermore, a temjierature rise in the 73° to 78° F. range at Oroville interrupted the migration in 1937, a j'ear in which occasional tcmperatin-e readings taken at Monse on the lower Okanogan River were even higher than those at Oroville (Chapman, 1941). On the other hand, the 1954 migration was ap- parently' unaffected l\v rises in the 68° to 70° F. range. The dependability of the latter example is subject to some question, however, because the temperature readings were recorded at Oroville, not at Monse. These sevei-al examples suggest a threshokl tem- perature of about 70° F., below which migration is not affected, but above which rising or stable tem- peratures inhibit migration — a condition which endures until a sharp drop allows the migration to resume. We have not considered here a situation in which fish enter the Okanogan River under favorable con- ditions only to be confronted enroute by sharply rising water temperatures. We have no data on this aspect of the problem, but suspect that the behavior and survival of the fish depend on several factors, including: (1) their location at the time they are confronted by rising temperatures; (2) their ability to acclimate: (3) their size, general health, and stage of maturity; and (4) the level to which the water temperature rises. RESULTS OF OTHER STUDIES AND THEIR POSSIBLE BEARING ON THE PROBLEM IN THE OKANOGAN RIVER The environmental factors that control the migra- tions of adult Pacific salmon have long been of practical and theoretical interest to fishery biologists. The literature gives many examples of environ- mental influences that affect different populations in different ways. Rather than present another review of the extensive literature on this broad subject, a matter so capably handled by Hoar (1953) and Allen (1956), we refer here only to the more important papers that deal with the environmental factors that influence the sudden mass movement of migrating salmon. Several investigators have found that rainfall and streamflow afi'ect the migration of adult salmon. Pritchard (1936), and Davidson, Vaughan, Hutch- inson, and Pritchard (1943), who studied pink ■^ These temperatures, recorded at Monse. are not subject to the possible error of estimating temperatures in the lower river from actual readings at C)roville. MIGRATION OF SOCKEYE SALMON 145 salmon (0. gorhischa), and Hunter (1959), who worked witli pink salmon and chum salmon (0. kein), coni-luded that entry into a river follows increases in stream flow. Shapavalov and Taft (1954) noted a correlation between the general periods of the spawning runs of silver salmon (0. k-isiilch) antl rainfall. They further believed, but were unable to demonstrate ([uantitatively. that fish movement increased with a rise of stream flow. .Mien (195(1), on the other hand, related the movement of silver salmon and chinook salmon (0. Ishawi/I.srha) to nighttime rainfall and low barometric jjressure, respectively. Ellis (1963) believed that the entry of silver salmon and sockeye salmon into rivers was associated with the appearance of atmosplieric warm fronts over the estuary. Only Foerster (1929) and Cramer and Ilammack (1952) attributed the sudtk'n movement of salmon to changes in water temperature. Foerster, who reported on sockeye salmon at Cultus T.ake, Rritish Columbia, noted that the numbers of fish arriving at a counting fence synchronized closely with tem- perature change: increases in the daily nui accom- panied declines in temperature. Cramer and Ham- mack, who studied chinook salmon in Deer Creek, a tributary to the Sacramento River, Calif., con- cluded that at the close of a period of clear weather and relatively cool water, sudden increases in water temperature to 75° F. caused an upsurge of fish. Andrew and Cleen (19G0), in their ajjpraisal of all available information on the possible effects of dam construction on the Fraser River, British Columbia, devoted considerable attention to the efTects of tem- perature and delay on upstream-migrating salmon. High water temperatures, they concluded, are detri- mental to salmon in their upstream migration be- cause they increase the rate of energy consumption and the incidence of disease and parasites, and may be directly lethal. In the matter of delay, Andrew and Geen (1900) cited Thompson (1945) as having shown that a delay of 12 days at Hell's Cate (before construction of the fishways) was sufficient to prexent sockeye salmon from reaching their spawning grounds, and that lesser delays reduced the reproductive cajiacity of the fish. The same authors also referred to an incident in the Fraser Canyon at Vale, British Columbia, in 1955 where the early run to the Stuart River was blocked (J days by higli water. Of an estimated .30,000 to 35,000 sockeye salmon, only 2,170 reached the spawning grounds. The effect of delay on the productivity of salmon has been illustrated by studies of fish passage at a rock slide on the Habine Rixcr, British Columbia (Godfrey, Hourston, Stokes, and Withler, 1954). Concerning this study, Andrew and Geen (19(i0) stated: .... T;iKK''>H of ti'*'' liclow tlio pciiiit of diftu'ult passane and rcrovory of the tagged fisli at h countiiiK fciu-c -!0 miles up- stream showed tliat ."iome of tlie fisli delayed helow the ob- struetion were able to migrate to their spawning grounds but relatively few were able to spawn successfully. Heeause fish were delayed and weakened below the obstruction they were not able to migrate at a normal rate after passing the obstruc- tion. The effective spawning in 1952, when some .sockeye were delayed for extended periods, was estimated as :50 to 42 percent of the numbers of female sockeye that rea<'hed the spawning grounds or 7 to 10 percent of the total escapement. From 'M to 40 percent of tlie female sockeye examined on the sp.'iwning ground died unspawned and others die2-()3. These exiw-riments were used to detect any changes in tl;e migration time caused by Rocky Reach Dam, which was constructed on the migration route diuing the inter\-ening years. Travel time varied greatly, both between and within years. The difTerence between 19()2 and 19()3 exc(>ede(l the difference between 1963 and 1953-54. The best estimate of the time required for .sockeye salmon to migrate from Rock Island Dam to Zosel Dam is 7 to 9 days, barring major delay due to environment. This travel time has not been increased by Rocky Reach Dam. ^^'ate^ temperature of the lower Okanogan River, or factors linked with it, is by far the greatest sotnce of delay. Below 70° F.. entry into the Okanogan River is relatively unimpeded. Above 70° F., relatively stable or rising temperatures delay entry until a sharp droj) occurs. 146 U.S. FISH .\ND WILDLIFE SEKVICE LITERATURE CITED Allen, George Herbert. 1956. Migration, distribution, and movement of Puget Sound silver salmon. Ph.D. thesis, Univ. of Wash., Seattle, 295 pp. Andrew, F. J., and G. H. Geen. 1960. Sockeye and pink salmon production in relation to proposed dams in the Fraser River system. Int. Pac. Sal. Fish. Comm. Bull. 11, 259 pp. Chapm.\n, Wilbert McLeod. 1941. Observations on the migration of salmonoid fishes in the upper Columbia River. Copeia 1941(4): 240- 242. Cramer, Frederick K., and David F. Hammack. 1952. Salmon research at Deer Creek, Calif. Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 67, 16 pp. David.sox, F. a., Elizabeth Vaughan, S. J. Hutchinson, and A. L. Pritchard. 1943. Factors influencing the upstream migration of the pink salmon {Oncorhynchus gorbuscha). Ecology 24 (2) : 149-168. Ellis, D. V. 1962. Preliminary studies on the visible migrations of adult salmon. J. Fish. Res. Bd. Can. 19(1) ; 137-148. FOERSTER, R. E. 1929. An investigation of the life history and propaga- tion of the sockeye salmon {Oncorhynchus nerka) at Cultus Lake, British Columbia. No. 1. Introduc- tion and the run of 1925. Contr. Canad. Biol. N.S. 5(1): 1-35. French, R. R., and R. J. Wahle. 1960. Salmon runs — upper Columbia River, 1956-57. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 364, 15 pp. 1966. Study of loss and delay of salmon at Rock Island Dam, Columbia River, 1954-56. U.S. Fish Wildl. Serv., Fish. Bull. 65: 339-368. Gangmark, H.\rold a., and Leonard A. Fulton. 1952. Status of Columbia River blueback salmon runs, 1951. Fi.«h Wildl. Serv., Spec. Sci. Rep. Fish. 74, 29 pp. Godfrey, H., W. R. Hourston, J. W. Stokes, and F. C. WiTHLEB. 1954. Effects of a rock slide on Babine River salmon. Fish. Res. Bd. Can., Bull. 101, 100 pp. Hoar, W. S. 1953. Control and timing of fish migration. Cambridge Phil. Soc, Biol. Rev. 28(4): 437-452. Hunter, J. G. 1959. Survival and production of pink and chum salmon in a coastal stream. J. Fish. Res. Bd. Can. 16(6) : 835-886. Pritchard, A. L. 1936. Factors influencing the upstream spawning migra- tion of the pink .salmon, Oncorhynchus gorbuscha (Wal- baumi. J. Biol. Bd. Can. 2(4) : 383-389. Shapavalov, Leo. and Alan C. Taft. 1954. The life histories of the steelhead rainbow trout (Salmo gairdneri gairdneri) and silver salmon (Oncor- hynchus kisutch) with special reference to Waddell Creek, California, and recommendations regarding their management. Calif. Dep. Fish. Game, Fish. Bull. 98, 375 pp. Thompson, William F. 1945. Effect of the obstruction at Hell's Gate on the sockeye salmon of the Fraser River. Int. Pac. Sal. Fish. Comm., Bull. 1, 175 pp. U. S. Geological Survey. 1937-1965. Surface water supply of the United States, (1935-1963). Part 12. Pacific slope basins in Wash- ington and Upper Columbia River Basin. Geological Survey Water Supply Papers, U.S. Government Print- ing Office, Washington, D.C. ^'a^ious pagination. MIGRATION OF SOCKEYE SALMON 147 SEASONAL OCCURRENCE AND SIZE DISTRIBUTION OF POST LARVAL BROWN AND WHITE SHRIMP NEAR GALVESTON, TEXAS, WITH NOTES ON SPECIES IDENTIFICATION ' By KE>fNETH N. Baxter and William C. Renfro% Fishery Biologist, (Research) Bureau of Commercial Fisheries Biological Laboratory, Galveston, Tex. 77552 Postlarvae of the genus Penaeus were collected at the entrance to Galveston Bay, Tex., over a 4-year period and along Galveston Island's beach during a I-year period. Postlarval brown shrimp, P. aztecus, and white shrimp, P. setiferus, were the predominant penaeids caught. Morphological characters, seasonal size differ- ences, and occurrence of juveniles in adjacent nursery ABSTRACT areas were used to identify these species. Seasonal occurrence, size distribution, and measures of relative abundance are given for postlarvae of the two species. The uniformity in size of postlarvae from collections along the beach and at the bay entrance indicated that small shrimp do not grow much when they are along the beach. Shrimp are the most vahiable marine fishery re- source of the Gulf of Mexico, where commercial landings annually exceed 170 million pounds and are valued at nearly .ffiO million. Many aspects of the biology and early life history of these crustaceans have been examined; however, the factors causing fluctuations in their abundance must be better de- fined before optimum management of the shrimp fishery can be realized. The early life histories of commercially important species of the genus Penaeus inhabiting the north- western Gulf of Mexico are similar. Each spawns in offshore waters, where the planktonic larvae hatch after several hours. During ensuing weeks, the larvae pass through a series of metamorphoses and reach near-shore areas as postlarvae. The young shrimp grow rapidly after moving into estuarine nursery areas, and return to offshore waters to com- plete their life cycle. As Bearden (1961) has pointed out, the postlarvae that reach inshore waters represent the success of ' Contribution No. 212, Bureau of Comniercial Fisheries Biologica Laboratory, CJalveston, Tex. = Present address: Department of Oceanograpiiy, Orecon State Univer- sity, Corvallis, Greg. Note. — .Approved for publication .\pril 28, 1966. the spawning season and, after several months of growth, will make up the bulk of the commercial shrimp catch for a given year. Baxter (1963) has shown that .systematic sampling of postlarvae enter- ing the major nursery areas can provide an index that is useful for predicting the subsequent abun- dance of juvenile and adult shrimp on inshore and offshore fishing grounds. The objectives of this report are to describe trends in the seasonal abundance and size composition of commercial shrimp postlarvae near Galveston Island, and to evaluate the u.se of seasonal differences in their body lengths as an aid in identifying the various species. Also examined is the question : Do young shrimp use the surf zone as a nursery area? The results of this 4-ycar study form a basis for cur- rent research on the biology and dynamics of the postlarval pha.se of commercial shrimp populations in the Gulf of Mexico. SAMPLING PROCEDURE Studies of postlarval shiimp began as part of a developing investigation of the life history of penaeid shrimp outlined in detail by Kutkuhn (1963). Knowing that shrimp reach shore as postlarvae and FISHERY bulletin: VOLUME 06, NO. 1 149 enter nursery areas through tidal passes, we estab- hshed a sampling station at the entrance to Galves- ton Bay in Xovember 1059. Additional stations along Galveston Island's Gulf beach were added later. GALVESTON ENTRANCE The initial sampling site was on the south side of the entrance to Galveston Bay (station A, fig. 1), Galveston Entrance 'GALVESTON 95 90 TEXAS i LA Figure 1. — Galveston Island and environs, showing sampling stations. where we collected postlarval shrimp twice each week. This location was not suitable as a sampling station after Hurricane Carla in September 19G1. Thereafter, semiweckly samples were obtained from station B, near the base of tlie north jettJ^ Bottom materials at both stations consisted of well- compacted sand. Collections of postlarvae were made with a 5-foot, hand-drawn beam trawl fitted with a plankton net at its cod end (Renfro, 1903). The wings of the trawl consisted of nylon netting having 50 holes per square centimeter. We believe that escapement of postlarval shrimp was negligible, because most collections contained an abundance of organisms more minute than the smallest postlarvae captured. To test whether or not large shrimp were evading the small beam trawl we towed a fine-mesh, 20-foot seine on several occasions. A standard procedure was followed during each collection. One end of a 1 50-foot line was tied to a stake driven into the sand at the wafer's edge. The collector held the free entl of this line in one hand and the bridle of the trawl in the other and pulled the gear along the bottom in a semicircular i)ath from the shoreline. GULF BEACH Collections of postlarval shrimp were made twice each month between April 1960 and April 19()1 at 5- mile intervals along Galveston Island's 25-miIe beach (stations C, D, E, and F, fig. 1). The same beam trawl was u-sed at beach stations, but because of the surf, the sampling procedure was altered from that used at stations A and B. The collector waded a measured 75 yards directly offshore, set the gear, and towed it back to shore. Computations of bot- tom areas sampled were based on distance towed and the dimensions of the net. At all stations we made meteorological and hydro- graphic observations. Those that we consider to be pertinent, namely water temperature, salinity, and tidal stage, are listed in appendi.\ tables 1 and 2 along with the numbers of postlarval brown and white shrimp collected on each sampling date. SEASONAL OCCURRENCE GALVESTON ENTRANCE Postlarval brown shrimp, P. aztecus Ives, appeared at Galveston I'^ntrance and migrated to the nursery areas within Galveston Bay at about the same time WHITE SHRIMP t«0 SAMPLINC '/\ A, BROWN SHRIMP ^ UX Yir.vRK 2. — Seasonal abundance of postlarval brown and white shrimp at Galveston Entrance, 1000-6.'5. 150 U.S. FISH .A.ND WII.DLIKE t^?;RVICE during each year of the study (fig. 2). The greatest numbers occurred in the spring; usually peak abun- dance was reached between mid-March and mid- April. Following the spring peak, comparatively few postlar\-ae were caught until about mid-June. Thereafter, the number of postlarvae in the collec- tions increased through July and reached a second peak in August or September. In each year, the numbers of brown shrimp postlarvae present at Galveston Entrance diminished rapidly after the second peak and remained low throughout the win- ter. During 1961, peak abundance appeared to develop in late April and earlj' May, but because sampling was suspended from May 8 to August 11, the actual time of the peak for that year is unknown. The first postlarval white shrimp, P. setifcnis (Linnaeus), were taken in early May of each year at Galveston Entrance (fig. 2). Seasonal distribution of postlarval white shrimp suggests that two peaks in abundance may occur each summer and that the relative strength of these peaks is variable. GALVESTON ISLAND BEACH Trends in seasonal occurrence of postlarval brown and white shrimp at Galveston Island beach stations were similar to those at Galveston Entrance stations (table 1). Brown shrimp postlarvae were numerous in mid-April 1960, from late June through August, and again during April 1961. In contrast to Gal- veston Entrance, a few brown shrimp postlarvae were present along the beach during late December and January. In 1961 brown postlarvae did not appear in significant numbers until early March. Postlar\al white shrimp were caught in beach sam- ples from mid-May through November 1960 and were most abundant from late June through July. None was taken from December 1960 through April 1961. Samples of postlarvae were collected along Gal- veston Island beach to determine if young shrimp use the littoral zone along beaches as nursery areas. Should they use this zone, advanced stages of post- larval shrimp could be expected in collections from beach stations. Agreement as to general size of postlar\'ae from the beach and from Galveston En- trance (table 2), indicates, however, that postlarvae spend little time in the beach area. Repeated tows with a fine-mesh seine at beach stations caught no shrimp larger than those taken in the beam trawl. Table 1. — Average monthly densities of postlarval shrimp at Galveston Entrance and Galveston Island beach stations, April 1960-61 IFigures represent the average number of postlarvae per 100 m.' of bottom in 7 to 12 collections each month] Brown shrimp postlarvae White shrimp postlarvae Month Galveston Entrance Gulf beach Galveston Entrance Gulf beach 1960: Apr May June July Aug Sept Oct Nov Dec 1961: Jan Feb Mar Apr 294 2 23 35 SI 1 1 13 72 52 15 54 234 153 3 3 3 8 1 1 70 760 6 40 14 29 2 1 9 52 133 26 39 3 3 Table 2. — Mean total lengths of postlarral shrimp collected concurrently along the Galveston Island beach and in Galveston Entrance, 1960-61 (Figures in parentheses indicate number of specimens me^ured] Bro^vn shrimp postlarvae White shrimp postlarvae Month Beach Entrance Beach Entrance 1960: Apr Mm. 11.4 (82) 10. 4 (52) 8.9 (113) 8.7 (181) 8.6. (241) 9. 5 (25) 10. 1 (27) 10. 9 (23) 11.9 (59) 11.7 (6) 11.0 (11) 11.6 (165) 11.3 (200) Mm. 11.5 (167) 10. 5 (34) 8.8 (101) 8.4 (155) 8.5 (146) 10.0 (10) 11.0 (4) 11. 2 (6) Mm. Mm. May June July Aug Sept Oct Nov Dec 6. 3 (47) 5.9 (115) 7. 2 (186) 6.7 (177) 7. 5 (77) 6. 8 (24) 7. 5 (23) 6.4 (51) 6.5 (149) 6. 3 (129) 6.3 (163) 7. 1 (35) 7.2 (10) 7. 5 (8) 1961: Jan Feb 12.0 (6) 11.6 (86) 11.6 (112) Mar Apr _ IDENTIFICATION AND SEASONAL SIZE DISTRIBUTION Of the three commercially important species of the genus Penaeus in the northern Gulf of Me.xico, the pink shrimp, P. duorarum, is the least abundant. Small numbers of adult pink shrimp are commonly caught off Galveston Island (15-20 fathoms), but landing data compiled by the Bureau of Commercial Fisheries Branch of Statistics^ included no pink shrimp in landings of 3.7 million pounds taken from Galveston Bay during 1960-63. A few pink shrimp, however, may have been landed and reported as ' "Gulf Coast Shrimp Catch by .\rea. Depth, Variety, and Size," Annual Summaries, 1960-63. DISTRIBUTIOX OF SHRIMP NEAR GALVESTON 151 brown shrimp. Of about 47,000 juvenile shriinj) examined from Galveston Bay bait landings between January 1900 and December 19(13, only 17 (less than 0.04 percent) were pink shrimp. In earlier work, the second author (1958-59) found no pink shrimp among more than 10,000 juvenile penaeid shrimj) taken from upper Galveston liay. Although post- larval pink shrimp obviously occur in the Galveston area they evidently are scarce; all postlarvae we caught were classified as brown or white shrimp. MORPHOLOGY No single criterion is sufficient to distinguish brown and white shrimp postlarvae, but they can be separated by taking into account various morpho- mctric characters, relative size, and seasonal occur- rence as juveniles in the estuary. Mori)hological and morphometric differences between postlarval brown and white shrimp provided by Pearson (1939) and Williams (1959) are sufficient to separate these species during most seasons. Williams, working with shrimp from North Carolina, developed a pro- visional key to early postlarvae. He stated that the tip of the rostrum and the extended third pereiopod on postlarval white shrimp do not e.xtend to the dis- tal edge of the eye. Conversely, in the brown shrimp, both the tip of the rostrum and extended third pereiopod reach to or beyond the edge of the eye. In postlarvae from the Galveston area, these characteristics suffice only to separate postlarval white and brown shrimp with a total length of 10 mm. or less, whereas Williams was able to use them in North Carolina for separating postlarvae up to 12 mm. total length. OCCURRENCE ON GALVESTON BAY NURSERY GROUNDS According to our records, brown shrimp are the only postlarval Pcnacus that enter Galveston Bay during the first 4 months of the year. This observa- tion agrees with findings from several previous studies conducted in the bay. Rcnfro (1959) found only brown shrimp postlarvae and juveniles (17 mm. and above) in upper Galveston Bay during April and May 1959. Guntcr (1960) also found brown shrimp to be the only species at the juvenile stage present in Galveston Bay during April and May ]9(>0. I.ater reports by biologists of the Texas Game and Fish Commission corroborate the observations of Renfro and Gunter (Pullen, 1962). By June, advanced postlar\al and early juvenile white shrimp (18-28 mm.) become abundant in Galveston Baj% and both brown and white shrimp are present throughout the summer (Gunter, 1960). Additional evidence regarding the identity of the winter and early spring postlarvae was provided in 1960 when 1,200 postlarvae, taken on April 12 at Galveston Entrance, were brought into the labora- tory to be reared. Ail that grew to identifiable size (150) were brown shrimp. SEASONAL SIZE DLSTRIBUTION The size of postlarvae caught at the entrance to Galveston Bay provides a strong clue to species identity during .some seasons (fig. 3). During the winter, the total length of brown shrimp postlarvae ranged from 10 to 14 mm. and averaged 12 mm. (fig. 3). lieginning in May of each year, a second group of much smaller (6.0 to 8.0 mm.) postlarvae appeared in the samples. These shrimp possessed the external morphological characteristics of post- larval white shrimp described by Pearson (1939) and Williams (1959). By late June the length distribu- tions of the two groups of postlarvae began to over- lap. The modes of the length distribution of brown postlarvae decreased, possibly because adult brown shrimp were spawning near shore in spring and sum- mer, or because warm water temperatures increased the developmental rates of larvae. During the same period, some white shrimp postlarvae as long as 10.5 mm. entered the estuary. Most of the larger post- larvae, however, exhibited the characteristics as- cribed to brown shrimj) by Williams (1959). The overlap in length distributions persisted throughout the summer, but the mean length of brown shrimp postlarvae always exceeded that of white shrimp in the same samples (fig. 3). In the latter part of each year, the modal length of brown shrimp postlarvae increased, and b}- October in some years the overlap in length distributions had ended. Postlarvae of brown and white shrimj) caught at beach stations and at Galveston I^nt ranee were of similar sizes (table 2). The total length of po.st- larval brown shrimp ranged from 8.5 to 12.0 mm. (mean, 11.5 mm.). White shrimp ranged from 5.0 to 9.5 mm. (mean, 7.0 mm.). No significant differ- ence existed among the mean lengths of postlarvae taken at the various beach stations on the same day. SUMMARY Collections of penaeid postlarvae were obtained semiweeklv at Galveston I^ntrance over a 4-year 152 U.S. FISH AND WILDLIFE SERVICE 50 r \ NM42 - /■ \..y. JULY - N=260 AUG. \ / \ - •■" ■•. / \- N= 384 .. \^ -1 1 1 i 1 1 — r~ 5 6 7 8 9 10 II 12 13 14 5 6 7 8 9 10 II 12 13 14 TOTAL LENGTH (MM. BROWN SHRIMP ■WHITE SHRIMP Figure 3. — Seasonal size distribution of postlarval brown and white shrimp at Galveston Entrance, 1960-63. (N indicates sample size.) period and twice each month at four stations along Galveston Island's Gulf beach for 1 year. Postlarval brown shrimp were collected at Gal- veston Entrance from February until mid-December of each year. At Galveston beach stations, they were found throughout the year but in smaller num- bers during the winter. Numbers of brown shrimp postlarvae reached an annual peak between mid- March and mid-April. Postlarval white shrimp were first caught at Gal- veston Entrance and along the beach in May and were most abundant through the summer. Postlarvae of browii and white shrimp were sep- arated by morphometric characters and by the sea- sonal occurrence of each species in the adjacent estuary. The brown shrimp was the only Penaeus species at the postlarval stage present along the Gal- veston Island beach and at the entrance of Galveston Bay from December through April. All individuals were relatively large (11 mm. or longer) during this period. After April, their average size decreased. remained relatively small throughout the summer, and then increased again in the fall. White shrimp postlarvae first appeared in May at lengths much shorter than those of brown postlarvae in the same collections; the total lengths of the majority ranged from 6.0 to 8.0 mm. During the summer, the length distributions of postlarvae of brown and white shrimp overlap in the 8- to 10-mm. length range. The two species at this stage of development may, however, be separated by the morphological char- acteristics described by Pearson (1939) and Williams (1959). At times, the largest white shrimp post- larvae in a sample were longer than the smallest postlarvae of brown shrimp, but the mean lengths of the white postlarvae were always less than those of the brown postlarvae. The similarity of mean lengths of postlarvae col- lected along the beach and at Galveston Entrance suggests that significant growth does not occur along the beaches and that the surf zone is not an impor- tant nursery area for small slirimp. DISTRIBUTION OF SHRIMP NEAR GALVESTON 153 of commercial shrimp Contr. Bears Bluff Lab. LITERATURE CITED Baxter, Kenneth N. 1963. Abundance of postlarval shrimp — one index of future shrimping success. Proc. Gulf Carib. Fish. Inst. 15th Annu. Sess.: 79-87. Bearden, Chakles M. 1961. Notes on postlarvae (Penaeus) in South Carolina. 33, pp. 3-8. GuNTF.R, Gordon. 1960. The field program (shrimp). Tex. Game Fish Comm. Mar. Fish, Div., I'roj. Rep. 1959-60, Spec. Rep., 14 pp. KuTKUHN, Joseph H. 1963. Expanded research on Gulf of Mexico shrimp resources. Proc. Gulf. Carib. Fish. Inst. 15th Annu. Sess.: 65-78. Pearson, John C. 1939. The oariy life histories of some American Pe- naeidae, chiefly the commercial shrimp, Pouicus seti- ferus (Linn.). " U.S. Bur. Fish. Hull. 49; 1-73. PtTLLEN, Edward J. 1962. A study of the juvenile .shrimp populations, Penaeus azlccus and Penaeus setifervs, of Galveston Bay. Tex. Game Fish Comm. Mar. Fish. Div., Proj. Rep. 1961-62, Proj. MS-R-4, 23 p. Renfro, William C. 1959. Basic ecological survey of Area M-2. Check list of the fishes and commercial shrimp of Area M-2. Tex. Game Fish Comm. Mar. Fish. Div., Proj. Rep. 1958-59, Proj. M-2-H-1, 30 pp. 1963. Small beam net for sami)ling postlarval .shrimp. In Biological Laboratory, Galveston, Tex. fishery re- search for the year ending June 30, 1962, pp. 86-87. U.S. Fish Wildl. Serv., Circ. 101. Williams, Au.stin B. 1959. Spotted and brown shrimp postlarvae (Penaeus) in North Carolina. Bull. Mar. Sci. Gulf Carib, 9(3): 281-290. APPENDIX Table A-1. — Numbers of imstlnrml shrimp collected and asso- ciated hydrographic observations, Galveston Entrance, 1959-63 Tablf, .\-1. — \umbers of postlarval shrimp collected and asso- ciated hydrographic observations, Galveston Entrance, 1959- 63 — Continued Date Time Postlarvae per standard tow Water temperature Salinity Tidal P. tiHecus P. tetiferut stage' 1959: Nov. 9 16 Dec. 11 1000 1330 1500 Number Number » °C. o/oo 21 31 Postlarvae per standard tow Date rime . Water emperatiire Salinity Tidal stage' P. aztecui P. setiferus Number Number "C. o/oo mo: Jan 8 1 4 13 15:i0 15,5 25,3 HWS l.-i lion 14,9 2.5,3 F 19 1400 9,8 9.4 E 22 0900 12 9,0 2;i,9 F 25 1100 1 10,2 26,0 F 28 1115 1 12,5 18,3 F Feb. 1 1300 1 12,2 28,4 5 1100 10,2 14,0 F 9 0900 2 15,0 31,2 11 0930 13,0 27,3 E 16 18 3 16,9 1415 3 12,5 12,1 E 24 1000 2 10,0 23,6 F 25 1500 2 12,0 F Mar. 1 1345 10,0 28,6 F 3 1405 10,5 11,1 E 7 1620 6 10,0 26,2 F 11 1330 5.3 13,0 15,9 E 15 1400 39 15,0 25,4 18 1420 72 14,2 23,2 K 22 1400 39 18,5 28,9 F 28 1100 4,710 20,8 26.2 Apr. 1 1120 3,680 19,5 26.9 F 5 1045 86 18,2 16.1 F 8 08:i0 5 18,5 2,5.7 LWS 12 1330 i.ono 21,0 .30,1 F 15 0900 100 23,5 28,2 F 19 1115 9 22.2 24,1 F 21 1330 .50 26,0 24,0 HW.S 26 1515 56 27,0 24,0 F 29 1330 3 24,0 23,8 E May 3 0830 22,0 22,2 F 6 1400 4 24,0 23,0 F 10 - 0900 6 4 22,2 24,2 E 13 1,)30 1 1 24,2 29,0 F 17 0845 2 2 24,8 :i0,5 V 20 0900 9 12 2,5, 6 29,5 F 23 08.30 7 82 2.5.4 27,8 E 26 1.545 5 6 29,0 26,9 E 31 1500 1 31,2 HWS June 3 1400 12 30,7 27,2 LWS 6 1030 23 29,0 29,3 F 9 1030 29,5 32,5 F 14 0830 8 28,0 31,2 F 17 0900 167 428 28,4 31,8 E 21 0930 1 6 29,0 32.7 F 24 1515 65 25 28,2 32,5 F 27 1400 108 60 30,0 31,1 F 30 0900 38 98 30,0 26,8 F July 5 8 1500 "4 148 31,9 31,4 LWS 1445 4 1 33,2 25,8 F 12 0930 61 28 30,0 29,4 E 15 1300 .30 C 32.0 25,8 E 19 18 73 71 35 29,0 32.0 29,6 31,3 F 21 - 1330 F 25 0840 241 21 30,0 35,5 F 29 1410 8 1 34,0 28,0 E Aug. 1 148 21 117 27 29,5 32.0 35,6 36,1 E 6 1345 F 9 . 1330 16 24 .33, 3 :i4,2 F 12 1430 4 29,5 26,3 E 15 .. 1330 257 202 31,0 ;i3,3 E 19 - 1600 2 30,5 27,1 E 22 1000 81 20 28,0 27,8 E 25 0915 306 77 29,5 28,8 LWS 29 1220 8 3 29,0 F Sept. 2 1130 4 9 31,4 26,5 F 1400 3 10 32,0 24,6 9 1010 3 30,1 23,2 E 12 0900 1 27,0 2.5,9 16 0930 27,0 21,3 E 20 0950 1 4 29,0 27.0 E Sec footnote at end of table. 154 See footnote at end of table. U.S. FISH AND WILDMFK SERVICE Table A-1. — Numbers of postlarral shrimp collected and asso- ciated hydrographic observatio7is, Galveston Entrance 1959- 63 — Continued Table A-l.— Numbers of pnstlan-al shrimp collected and asso- ciated hydrographic observations, Galveston Entrance, 1959- 63 — Continued Date Time Postlarvae per standard tow P. aziecus P. setiferus /Seo.— Continued Sept. 23. 27. 30. Oct. 3. 7_ 11_ 14. 17-. 20. 25.. 28.. 31.. Nov. 3. 10." 15. 18. 21. 25. 28. Dec. 2.. 9. 13.. 16.. 19.. 22.. 27.. 29.. Jan. 3.. 6.. 10.. 13.. 16.. 20.. 25.. 27.. 30.. Feb. 3. 7. 10. 17. 21.. 24.. 27.. Mar. 3. 6. 10. 14. 17. 21. 27. 31. Apr. 5.. 7.. 11.. 17.. 21.. 25.. 28.. May 2.. Aug. 11. 15. 17. 21. 2». 28. 1310 1400 1615 1400 1100 1400 1400 0845 0930 0900 1415 1430 1415 1430 1600 1145 1416 16.30 1430 1545 1545 15-30 1500 1546 1445 1440 1600 1100 1145 1330 1135 1530 1405 1115 1500 1410 1340 1340 1420 1430 1416 1406 1420 1100 1115 1350 1510 1520 1430 1400 1520 1440 1040 0845 1515 1045 1555 1600 1600 1550 1356 1500 1720 1410 0845 1400 0915 1400 Number 12 4 209 10 3 2 Water temperature Number 10 °c. .32.0 28.0 30.0 28.0 27.2 30.0 26. S 2.5.0 21.0 2,3.5 27.0 21.3 24.0 21.0 12.5 24.0 18.5 18.0 20.0 22.5 16.0 19.0 15.5 13.5 13.0 12.0 12.2 13.0 12.5 11.8 12.2 13.0 12.4 12.5 14.8 10.2 8.0 11.8 13.0 11.0 13. 19.2 15.0 17.0 16.0 19.0 19.4 17.8 19.2 20.6 19.7 27. :j 20.1 19.2 19.6 19.1 20.0 23.0 24.0 24.1 25.8 26.9 Salinity —SAMPLING INTERRUPTED- SI.? 33.8 29.8 31.5 29.8 29.8 o/oo 27.8 24.1 26.5 28.6 24.6 26.7 28.6 27.0 16.3 27.3 2.'i. 5 17.1 20.0 22.9 17.9 26.2 26.9 24.9 11.2 24.5 27.8 27. 1 29.5 31.0 19.2 2.5.4 19.4 11.1 14.4 24.9 8.3 15.2 7.8 26.8 12.7 11.9 13.8 10.8 24.4 28.6 7.6 8.9 10.9 30.2 21 8 26.7 26.7 26.7 25.9 25.8 18.2 25.3 30.9 29.2 27.8 29.4 25.6 15.0 16.6 19.1 30.5 30.3 28.4 20.6 24.6 22.2 Tidal stage' LWS E LWS F E E F F E F E E F F E F F F E F F F F E E E E F F E F E E E E F E E E E F E E E F E E E F E E E E F E F E E F E E E Date Time ;9S;;— Continued Sept. 1. 26. 27. Oct. 2.. 5.. 10.. 12... 16... 19... 23... 25... 27... 30... Nov. 3. 14. 16. 21. 24. 27. 30. Dec. 5.- 11. 14. 19. 22.. 26.. 29.. Jan. 2. 4.. 12. 15. 17.. 23. 26.. 29.. Feb. 1.. 6.. 9.. 12.. 15.. 19.. 23.. 26.. Mar. 1. 6. 9. 12. 16. 20. 23. 26. 29.. Apr. 4.. 9. 12.. 17.. 20.. 23.. 26.. May 1. 3. 7. 10. 1416 1416 1200 1510 0925 0900 1005 1400 1.545 1520 0925 0905 0920 1426 1345 1620 0910 0940 0945 0940 1406 1020 0946 0930 1430 0920 0920 0925 1400 0945 0920 1410 0926 0925 0906 0905 1400 1410 1045 0900 1650 0845 1400 1030 1400 0925 1340 0900 1625 1030 1415 1100 1415 0925 1440 10.35 1420 0925 1445 09O0 1445 0925 1430 0905 1400 1000 1425 0910 0925 Postlarvae per standard tow P. aztecus P. setiferus 520 65 Number 644 44 Water temperature 30.9 33.9 -HURRICANE CARLA- 5 4 2 6 32 5 61 6 2 144 11 9 3 45 13 1 4 21 11 2 1 12 ■224 6 1 73 34 196 48 222 53 1,220 40 368 66 8 506 626 140 76 1,682 234 24 136 192 44 103 3 4 260 23 .32.0 34.0 29.0 20.0 29.0 ,30.2 24.5 27.0 23.5 24.9 20.5 30.0 19.6 12.0 13.0 19.0 17.0 17.5 19.0 21.5 12.0 20.0 16.0 20.5 8.6 13.5 18.0 13.5 10.0 16.0 16.0 12.0 -2.0 6.0 9.0 9.0 16.0 10.0 14.0 9.0 22 19.0 21.0 1,6.0 22.0 22.0 ,5.5 12.0 24.0 17.6 17.6 19.5 20.6 19.0 24.0 16.0 23.0 21.0 24.0 20.5 27.6 24.0 26.0 Salinity .See footnote at end of table. o/oo 24.3 22.3 17.2 17.2 18.1 27.8 27.3 24.8 27.4 2,6.7 24.7 26.3 28.6 29.1 27.0 24.0 24.7 23.5 15.0 28.7 28.7 27.6 28.4 27.5 27.0 25.6 23.1 24.4 26.3 27.9 24.4 27.5 31.2 31.0 ,30.5 29.9 29.8 27.7 27.2 25.7 26.5 31.4 29.6 26.3 27.6 27.6 22.3 21.7 23.0 23.3 24.7 25.5 19.5 26.9 26.7 22.3 28.9 27.0 25.5 24.4 25.3 24.9 24.4 25.2 24.6 26.0 23.3 24.0 18.0 24.1 17.7 25.0 23.6 Tidal stage' F F F F E E E E E E E E F F E E E F F E LWS E P E HWS E E E E LWS E E F E F E F LWS F E E F F E E F F E E F F LWS F F F E E F E E F F F E F F F See footnote at end of table. DISTRIBUTION OF SHRIMP NEAR GALVESTON 155 Table A-1. — Numbers of posllarval shrimp collected and asso- ciated hydrogfaphic observations, Galveston Entrance, 1959- 63 — Continued Table A-1. — Numbers of posllarval shrimp collected and asso- ciated hydrographic observations, Galveston Entrance, 1959- es— Continued Date Time /96*.-— Continued May 15 18 21 24 29 June 1 4 8 12 16 18 21 26 29 July 2 5 - 10 13 16 19 - 24 27 30 AuR . 2 8 10 13 16 21 24 27 30. Sept. 4. 7. 10- 13. 18. 21. 24. 27. Oct. 2.. 5.. 8.. 11.. 16.. 19.. 22.. 25.. 30.. Nov. 2. 5. 8. 13. 16. 19. Zi. 26. 29. i9es Jan. 4. 8. 11. 14. Postlarvae per standard tow P. azUcua P. tetiferus Dec. 3 6 11 14 17 20 26 28 31 0925 1500 0915 1500 0905 1430 0925 1510 0950 1100 0855 1423 0945 1450 0905 1100 0905 1410 0845 1445 0900 1410 0910 1415 0900 1430 0905 1400 0915 1405 0925 1420 0900 1410 0925 1425 0905 1505 1045 1410 1500 1430 0920 1420 0910 1440 0845 1410 0905 1505 0845 1405 0930 1445 0915 0900 0905 1450 0845 1405 0915 1315 0915 1410 0905 1600 1035 1445 09.30 1430 0845 Water temperature Number 2 3 6 8 32 16 6 11 17 13 48 3 1 14 19 4 145 16 25 76 6 15 3 25 2 37 11 2 367 24 2 6 4 12 6 1 2 Salinity NumbfT 1 2 6 4 6 28 6 13 15 1 61 116 4 6 4 2 4 46 29 31 36 29 ,06 10 38 2 78 42 7 1,227 96 6 17 1 11 150 29 46 13 4 I 26.0 27.9 27.0 29.0 26.0 26.0 26.0 31.5 29.5 31.5 31.0 34.0 30.5 32.0 30.0 33.0 32.0 3.3.0 30.0 34.0 31.0 32.5 30.0 33.0 31.0 35.0 30.0 32.0 31.0 33.5 29.0 31.0 31.0 31.0 .10. 5 33.0 30.0 27.0 28.0 26.0 29.0 27.0 29.5 31.0 29.0 29.0 2.1.0 22.0 18.0 22.0 17.0 20.0 14.0 22.0 12.0 16.0 18.5 11.0 16.5 16.5 16.0 12.0 9.0 20.0 11.0 13.0 11.5 15.0 U.O 16.0 0.0 Tidal stage' o/oo 31.7 24.2 23.8 24.7 18.9 I.'). 7 14.7 17.9 18.6 26.8 26.1 24.4 24.6 23.8 2.1.2 19.9 20.0 31.5 29.8 33.6 35.5 37.4 33.8 34.4 31.4 35.6 .14.4 .15.8 .16.1 35. 6 36.1 29.8 .10. 5 28.6 31.0 31.5 28.7 26.8 25.3 27.0 27.8 26.9 27.0 28.7 30.1 29.8 28.8 30. 2 26.5 29.2 29.1 29.3 30.8 30.9 31.0 29.7 29.9 22.9 29.1 20.8 .10.6 32.4 31.5 31.8 26.3 22.7 28.1 29.1 218.2 31.3 32.4 HWS F F F E HWS F HWS E HWS F F F E F F E E F E F HWS E E E E E F E E E F E E E F E E LWS F F E E F E E LWS F E E E E E E F E E F E F E LWS F F E F E Time /fl()5.'— Continued Jan. 17. 22.. 25. 28. 31. Feb. 5 IK- 14 . 19.. 21. 25- 28. Mar. 5.. 8-. 11.. 14.. 10.. 22.. 25.. 28.. Apr. 2.. 5 - 8-. 11.- 16.. 19.- 22- 25.- 30.. May 3. 6. 9. 14. 17. 20. 23. 28. 31. June 3.. 6.. 11.. 15.. 17.. 20.. 25-. 28. July I- 6-. 9,. 12.. 15.. 18-. 23.. 26.. 29.. Aug. 1.. 6.. 9.. 12.. 15.. 20.. 23.. 26.. 29.. Sept. 3.. 6.. 9. 12. 18. 20. 23. 26. Postlarvae per standard tow P. aztecus P. tetiferus Water temperature Number Number 1310 0900 1430 0935 1410 1115 I4.)5 0940 1430 0935 1410 0930 1415 0920 1415 0925 1400 08.50 1425 0840 1350 09,58 1146 0925 1415 0910 1415 0910 1410 08.55 1410 0910 1400 0910 1410 0950 1435 0915 1405 0910 1415 0915 0915 1410 1440 0915 1410 0910 1415 0910 1415 0910 1420 0910 1420 0935 1415 0915 1415 0915 1430 0910 1415 0910 1430 09.10 1430 0910 1410 0910 1415 0935 1420 441 16 288 21 280 286 986 114 360 3,521 147 54 167 44 103 93 41 68 181 71 1(J 16 17 134 29 28 381 5 6 38 24 882 211 16 62 59 2;t 33 U 32 2.1 14 51 94 19 27 48 93 4 10 41 10 I 6 24 206 60 68 Salinity 3 5 272 9 1 2 70 26 115 3,407 117 19 21 18 10 548 4 9 29 2 2 3 6 5 21 30 36 26 12 38 5 28 14 18 35 264 1.12 94 167 9.0 9.0 7.0 I.O 15.0 12.0 16.0 9.0 11.0 8.0 16.0 11.0 15.0 14.5 16.0 17.0 18.0 21.0 1,5.0 20.0 27.0 22.5 20.0 21.0 28.0 23.8 27.0 2,5.0 29.0 24.0 30.0 25.0 27.0 26.0 29.0 28.0 27.0 27.5 32.0 34.0 32.0 30.0 28.0 35.0 31.0 29 35 30.0 32.5 30.0 31.0 29.0 .13. 31.0 26.0 31.0 34.0 31.0 3^1.5 29.0 29.5 30.0 35.0 30.0 34.0 .30.0 29.5 30.0 32.0 24.0 29.0 2.5.0 26.0 Tidal stage* o/oo 29.5 24.9 28.5 27.9 26.9 29.8 29.6 31.5 2.5.8 27.2 29.6 28.7 29.7 .10.4 30. 6 30.4 29 29.2 27.4 26.9 25.9 27.6 27.6 27.6 30. 6 33.0 32. 28.0 21.0 21.6 20.3 21.7 24.2 28.8 29.0 31.3 34.5 32.3 33.2 34.8 32. 6 .10.8 .12.9 32. 3 33. 1 29.9 27.1 31.2 31.3 31.1 34.3 31.8 34.9 33. 6 29.1 36.1 3.5.2 34.6 .16.5 35.0 3.5. 2 35.9 ,16.9 .16.4 37.1 37.6 3.1.3 37.1 36.7 28.2 27.7 27.4 24.9 F £ F E F E F E E E F E F E F E F HWS E F HWS F F LWS F F F F E F F E F F F E F F F E E F F F F F F F HWS E F F E F LWS E E E LWS E E F E E E F LWS HWS E F HWS E LWS See footnote at end of table. See footnote at end of table. 156 U.S. FISH AXD WILDLIFE SERVICE Table A-1. — Numbers of posUarial shrimp collected ami asso- ciated hydrographic obserralior^s, Gaheston Entrance, 1959- 63 — Continued Table A-2. — Numbers of postlamil shrimp and associaf-ed hydrographic observations, Galveston Island beach stations, 1960-61— Continued Date Time MSJ;— Continued Oct. 1. 4. 7_ 10- 15_ 18. 21. 24. 29. Nov. 1 4 12 15 18 26 Dec. 2_ 5. 10. 13. 16. 19. 24. 27. 30. Postlarvae per standard tow P. aztecus P. seliferus 0910 1415 0910 1415 0910 1405 0915 1415 0915 1505 0925 1430 0925 1410 1505 1100 1420 1415 0950 1410 0930 1430 1430 1515 1005 A^umbfr Nuvtbej 177 2 141 47 76 44 14 14 309 19 39 15 Water temperature 21.0 24.0 26.0 30.0 26.0 26.5 24.0 26.0 21.0 19.5 21.0 25.0 18.0 26.0 17.0 20.0 17.0 15.0 9.0 5 9.0 10.0 18.0 10.0 Salinity o/oo 29.3 29.6 29.5 30.5 31.4 32.6 31.7 31.8 31.4 30.7 30.8 30.6 30.0 30.3 31.3 ■30.7 30.4 32.5 32 8 29.9 31.0 30.3 29.9 31.1 32.8 Tidal stage' E LWS E E E F E E HWS F E E E F F F E E F F E E E F E ' F=F]ood;E=Ebb; HWS = High-water slack: I.WS =Low-water slack. Table A-2. — Numbers of postlarval shrimp and associated hydrographic observations, Galveston Island beach stations, 1960-61 Postlarvae per standard tow Water Salinity Tidal Date Station Time temperature stage! P.aztecus P. setiferus Number Number °C. o/oo I960: Apr. 14 C D 1200 18 21.2 29.8 HWS E 1245 137 21.6 30.2 HWS F 1345 161 22 2 30.7 E 27 C 0900 1 24.2 2.3.9 F D 0940 3 24.^ 24.0 F E 1030 5 25.2 24.2 F F 1115 5 25.5 23.3 F May 11 C 1045 23.2 30.1 LWS D 1130 3 23.6 .30.3 LWS 12.30 4 2:j.2 30.8 F F 1330 3 1 23.0 31.0 F 26 C 0830 1 4 26.3 29.2 E D 0915 11 3 26.8 29.7 E E 1000 89 50 27.0 29.8 E F 1045 5 14 27.0 29.7 E June 8 C 0900 2 9 28.0 32 7 F D 0945 3 3 29 33.1 F E 1030 6 2 29.1 33.5 F F 1100 2 1 29 33.9 F 22 C 0840 191 64 29.6 32.9 E D 0940 51 107 30.0 32 9 E E 1020 113 122 .30.1 33.1 F F 1100 28 73 30. 8 33.1 F Date Station Time /SSO.— Continued July 6 C D E F 20 Aug. Sept. 15 Sept. 28 Oct. 12 Nov. 9 Dec. 8 Dec. 21 Jan. Feb. 1 C D E F C D E F C D E F C D E F C D E F C D E F C D E F C D E F c D E F C D E F C D E F C D E F C D E F C D E F C D E F 0930 1000 1030 1200 0840 092O 1045 1200 0830 0930 1000 1120 0830 0910 1000 1130 0840 0930 1046 1205 0840 0935 1110 1305 0845 0930 1200 1010 0840 0930 1030 1110 0845 0945 1040 1150 1.330 1110 1445 1520 1320 1405 1435 1540 1405 1435 1516 1600 1355 1430 1516 1545 0930 1015 1100 1130 0930 1030 1116 1200 1.330 1420 1440 1530 Postlarvae per standard tow P.aztecus P. setiferus Number 6 135 112 126 39 432 392 390 59 97 168 10 166 897 160 10 20 28 66 6 8 3 2 C 1 n 5 1 1 6 3 2 6 Number II 90 168 62 36 288 54 260 63 24 15 10 14 78 51 13 14 12 74 189 12 10 2 1 2 2 7 5 4 1 1 4 4 4 1 11 3 Water temperature 29.8 30.8 30.6 33.0 29.4 29.3 30.0 30.4 29.7 29.8 30.0 30.5 28.5 28.9 29.2 30.0 29 29 30.2 31.2 27.8 28.0 28.5 31.3 24.0 24.0 26.0 25.0 27.0 27.4 27.8 28.0 23.8 24.4 24.2 23.8 21.3 21.5 21.8 22.0 19.5 20.0 19 5 20.0 17.0 16.5 17.0 16.5 10.2 10.2 11.0 12 10.8 10.9 11.7 11.6 14.6 16.0 16.5 16,0 13.0 13.0 14,0 14.0 Salinity o/oo 31.6 31.9 31.3 31.7 33.4 33.4 33.9 33.3 36 1 36.0 35.9 35.9 32.4 32.1 32.2 27,5 29,4 Tidal stage' 2,6,9 25.5 25,4 25,7 28,1 28,2 28,6 28,6 28, 1 28,8 28,9 29,2 27,3 27,6 27,2 26 6 26,9 27,6 27,5 28 8 25,5 27.3 25.7 25.4 28.2 28.1 28.6 29.3 29.1 28.7 27.3 26.5 27.3 26 8 27.9 32.7 32 9 32 4 26.6 24.6 24.9 24.9 LWS F F F E HWS E F F F F F F F F E F F F E F E E E E E E E E E E E E E E E E E E E E E E F F F F F E E E E E E E E E F F F E E F F See footnote at end of table. See footnote at end of table. DISTRIBUTION OF SHRIMP NEAR G.\LVESTON 157 Table A-2. — Numbers of posllarval shrimp ami associated hydrographic observations, Galveston Island beach stations, 1960-61 Postlarvae per standard tow Water Salinity Tidal Date Station Time temperature stage' P.mucxts P. sttiferus Number Number "C. o/oo /9S/.— Continued Feb. 20 C 1320 17.4 26.9 F D HOO 2 16.9 27.4 F E 1435 7 17.2 26.9 F F 1500 2 16.8 26.8 F Mar. 8 C \?af, 14 18.0 28.8 E D Nl.'i 60 16.7 28.8 E E 144.5 18 17.5 29.1 E F 1510 138 17.9 29.3 E Mar. 23 C 0900 8 17.9 27.2 F D 0937 145 18.4 27.2 F E 1020 72 19,0 27 4 F F 1037 69 19.8 27.4 F Apr. 6 C 1315 141 20.0 29.2 F D 1355 217 19.7 27.7 F E 1425 1,040 20.0 30.7 F F 1455 2,662 ,33.6 E 20 C 1.3.35 173 1) 24.2 29.1 E n 1410 254 24.0 29.6 E E 1430 196 23.6 .30.4 E F 1515 850 24.0 31.0 E ' F= Flood; F=Ebb; HWS =High-water slack: LW.S = Low-water slack. 168 U.S. FISH AND WILDLIFE SERVICE CODIUM ENTERS MAINE WATERS By Gareth W. Coffin, Fishery Technician and Alden P. Stickney, Fishery Biologist (Research) Bureau of Commercial Fisheries Biological Laboratory, Boothbay Harbor, Maine 04538 An exotic species of marine algae, Codium fragile (Sur.) subsp. tomentosoides (Hariot) (fig. 1) was found growing in Bootlibay Harbor, Maine, near the Bureau of Commercial Fisheries Biological Labora- tory on July 17, 1964. Although the species has been established in the Long Island (N.Y.) and Cape Cod regions for se\'eral years, this is its first record from the Atlantic coast north of Cape Cod. In many oyster producing areas, Codium grows luxu- riantly on the oyster shells and is considered to be a serious pest by the oyster growers. The specimens from Boothbay Harbor were all collected within 400 m. of the Biological Laboratory in a sheltered cove, and all but one were attached to various objects just below the low tide .mark; the single exception was unattached and entangled in some fronds of rock weed (Ascophyllum nodosum). Among the substrata to which Codium was attached were stones. Modiolus modiolus shells, seaweeds, and waterlogged timbers. Because Galtsoff' had re- ported Codium from depths down to 12 m., SCl'BA divers surveyed the same general area for subtidal specimens. Their survey, as well as littoral surveys in other parts of the harbor, yielded rio additional specimens, although more were subsequently found near the site of the original discovery. Table 1 summarizes the data on all plants col- lected from July 20, 1964 to August 5, 196.5. Game- tangia were found on plants only during July and August. Nine specimens bearing gametangia were examined histologically to determine their sex: five bore mostly male, three bore mostly female, and one Note. — .Approved for publication .\pril 20. 1966. 'P. S. Galtso£f in a manuscript on file at the Bureau ol Commercial Fisheries Biological Laboratory, Woods Hole. Mass., first called attention to Codium on Cape Cod in January 1962. had about equal numbers of male and female gametangia. Growth in the Boothbay Harbor area appears to be rapid even during-the cold part of the j-ear. For example, plants collected in May 1965 apparently had grown as much as 34 cm. in length since the previous November when they were so small as to be barely visible. Table 1.— Codium fragile var. tomentosoides collected in the Boothbay Harbor area, 1964-65 Date collected Specimens Length range Mean length With gametangia 1964: Julv 20 Number 18 10 7 1 2 6 16 2 4 Cm. 9-54 8-20 10-16 20 17-20 11-34 3-24 13-14 18-43 Cm. 25.6 13.2 13. 5 20.0 18.5 20.3 10.6 13.5 26 Number Aug. 11 Sept, 9 Nov. 23 Do. Do. Do. Do. Nov. 24.. 1965- Mav5 July7_.. Aug. 5.. In the past 60 years, Codium fragile appears to have spread widely throughout the world. Silva (19.55) presumed its original center of distribution to be the Pacific and Subantarctic regions, perhaps in Japan. It appeared in Holland about 1900 (Van Goor, 192.3) and spread to Denmark, Norway, Sweden, England, and France (Silva, 1955). The introduction of Codium in Cape Cod was described by Wood (1962), as well as Galtsoff. No certain evidence is available to explain the source of its introduction in Long Island, N.Y., or that of its recent appearance in Maine. One possible explana- tion of the Maine introduction might be that the FISHERY BULLETIX : Vol. 66, No. 1 159 Figure l.-Two typical specimens of C odium fragile subsp. t.menlosoid.s cllected in Bootl.l..i>- Ilarl.or, Maine. ^- Figure 2.— Utricles and attached gametanRia from a specimen of Codium fragile subsp. tomentosoides. 160 plant arrived on oysters {Crassostrra vhgimca) shipped from Long Island to Bootlibay Harbor to delay their spawning by holding them in the coUlcr Maine waters. These oysters were customarily in- spected, however, both upon arrival and again l)efore return, to prevent the possible introduction of unde- sirable species. Codium in very early stages of development may have been overlooked when the oysters arrived from Long Island, because at the time Codium was not one of the undesirable species being checked; nevertheless, the reexamination before returning the oysters should have brought to light the plants at a larger stage of development. Codium could possibly have been introduced on the hulls of some of the yachts visiting the area. Such an e.xplanation was not favored by Roscnvinge (1920) who l)elieved the spread in Europe to be due to the breaking loose and drifting of plants or to their being transported along with oysters or other shell- fish. U.S. FISH .\NI3 WILDLIFE SERVICE LITERATURE CITED Van Goor, A. C. J. Galtsoff, Paul S. ^^^^- ^^^ holliindischen Meeresalgen (Rhodophyceae, 1962. Introduction of new seaweed, Codium fragile into Phaeophyceae, und Chlorophyceae) ins besonderc der Cape Cod waters. Unpublished report to Bureau of Umgebung von Helder, des Wattenmeeres, und der Commercial Fisheries, Branch of Shellfisherie^ Janu- Zuidersee. Verhandl. Akad. Wetensch. Amsterdam, arvl5, 1962. " ' Sect. 2, 23(2) : 232 pp. ROSENVINGE, L. K. '^'°°°' K- D. 1920. Om nogle i tid indvandrede havalger i de danske ^'^f' 5!f*"'" '' "^"'^"^ ^° ^''P'^ ^°^- ^^"- ^"""^y farvande. Bot. Tidsskr. 37: 125-135. Bot. Club 89(3): 178-180. Silva, P.iul C. 1955. The dichotomous species of Codium in Britain. J. Mar. Biol. Assoc. U.K. 34: 565-577. CODIUM ENTERS M.\INE WATERS lol LABORATORY EVALUATION OF RED-TIDE CONTROL AGENTS By Kenneth T. Marvin and Raphael R. Proctor, Jr., Chemists Bureau of Commercial Fisheries Biological Laboratory, Galveston, Texas 77550 Intense blooms of the dinoflagellate Gymnodinium breve Davis that occur at irregular intervals along the west coast of Florida (Feinstein, Ceurvels, Hutton, and Snoek, 1955) may cause extensive mor- tality of marine organisms. The blooms are popu- larly knowTi as red tides because of the amber to red discoloration they impart to the water. The Fish and Wildlife Service initiated studies in 1948 to determine the possibility of artificial means to reduce the occurrence or intensity, or both, of the red tides. Early tests indicated that copper, in con- centrations as low as 0.03 p. p.m., is lethal to labora- tory cultures of the red-tide organism. Rounsefell and Evans (1958), and Alarvin, Lansford, and Wheeler (1961) demonstrated, however, that control by copper was not feasible under field conditions. The copper precipitated from solution after a few days and, consequently, was ineffective for control. In 1959, scientists of the Bureau of Commercial Fisheries Biological Laboratory in Galveston, Te.\-., began a systematic evaluation of 4,30G compounds 'Contribution No. 215, Bureau of Commercial Fisheries Biological Laboratory. Galveston, Tex. Note. — .\pproved for publication May 6. 1966. as red-tide to.xicants. The initial phase of the study (Marvin and Proctor, 1964) involved testing each compound to determine its toxicity to G. breve. The final phase of the study, described here, evaluated some of the more toxic materials in the laboratory. We investigated only the compounds that we deter- mined to be 100-percent lethal to G. breve within 24 hours at concentrations of 0.01 p. p.m. or less. A red-tide control agent must also be selectively toxic; it must kill the red-tide organism without harming other species. The chemicals fulfilling the toxic requirement for red-tide control were tested for selectivity by deter- mining their effects on juvenile forms of marine species living in Galveston Bay and adjacent coastal waters. The selectivity threshold concentration was set arbitrarily at 0.1 p.p.m. Chemicals that killed 50 percent or more of any test organism within 24 hours at or below this concentration were rejected. The five chemicals that passed the selectivity tests, their effects on the test organisms at the threshold concentration, and the species tested are noted in table 1. T.\BLE l.—Pcrce>itage mortality of test organisms held 24 hours at toxicant concentration levels of 0.10 p.p., Chemical Carbamic acid, diethyldithio-; tellurium salt... Carbamic acid, dimethyldithio-: ferric salt Disulfide, bis(diethylthiocarbamvl) Sulfide, bis(2-hydrosy-3-bromo-5-chlorophenyi)-; bi.s dimethylamino butyne monosalt Sulfide, bis(2-hydro.\y-.3-bromo-5-chlorophenyl)-; cyclohexylamine mono salt Species' Blue crab (megalops) striped mullet Brown shrimp (postlarval) Sailfln molly Marsh periwinkle Shccpshead minnow Hermit crab Atlantic croaker 20 10 ^SSSEf------ --^^ FISHERY bulletin: VOL. 66, NO. 1 163 Table 2.— Results of six toxicity tests in terms of percentage mortality of G. breve after 2.', hours exposure Chemical Carh;miic acid, dicllivlililliio-: Ipllurium salt — Curhamic aid. diimthyWitliio-; ferric salt Disulfide. bis(diet hyUliioearbamyl) Sulndo. bi5r2-hydroxy-.'Miromo-.Vclilon)i)llenyl)-; bis dimithvlaniinn butyne mono salt Sulfide. ois(2-bydroxy-:i-broino-5-chloroplieiiyl)- cycloiictylamiiic mono salt - -- Tost numbers for concentration of 0.01 p. p.m. 100 10(1 100 25 25 100 100 100 100 .■so 100 75 75 50 25 100 100 100 100 100 100 100 100 75 100 100 25 25 Test numbers for concentrations ofO.OOli p.p.m. The selective chemicals were tested to determine their minimum toxic concentration levels to G. breve. Each toxicant was tested six times at 0.01 and 0.003 p.p.m. The results, in terms of mortality of G. breve, apjicar in table 2. Variation was con- siderable among the supposedly replicate sets of four of the chemicals. This suggests that the concentra- tion of the.se four chemicals was close to the toxic threshokl. At or close to the toxic threshold level, a slight variation in the concentration of a toxicant can have a pronounced effect on the mortality of organisms in cultures containing the toxicant. Only one of the selective toxicants, carbamic acid, diethyldithio-; tellurium salt, consistently met the toxic requirement arbitrarily established for a con- trol agent (R. T. Vanderbilt Co., Inc., 230 Park Avenue, Xew York City, N.Y. 10017; $2.13 per pound in 100-pound containers). This compound has two shortcomings, however: it killed 10 jiercent of the test organisms of two species (table 1); and its cost is proliiliitive for massive u.se in the field. LITERATURE CITED Feinstein, A.nit.\, a. Russei, Ceuuvels, HonEiiT F. lIcTTON, and Edw.^rd Snoek. 1955. Red tide outbreak.s oft the Florida west eoa.^t. Univ. Miami, Mar. Lab., Rep. 55-15 to Fla. State Bd. Conserv., 44 pp. M.\nviN, Ke.nxeth T., L.vnENcE M. L.\nsford, and R.\y S. Wheeler. 1961. lOffects of copper ore on the ecology of a lagoon. I'.S. Fish Wildl. Serv., Fish. liull. 61: 15:5-100. Marvin, Ke.\neth T., and Raphaei, R. Proctor, Jr. 1964. Preliminary results of the systematic screening of 4,306 compounds as "red-tide" toxicants. U.S. Fish Wildl. Serv., Data Rep. 2, 3 microfiches CU- SS pp.). Rou.NsEFELL, Geor<;e .\., and Joii.\ E. Evans. 1958. Large-scale experimental test of copper sulfate as a control for the Florida red tide. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 270, vi-l-57 \)\i. 164 U.S. FISH AND WILDLIFE SERVICE li- U S. GOVERNMENT PRINTrNG OFFICE: 1967—0 219-992 OBJECTIVE STUDIES OF SCALES OF COLUMBIA RIVER CHINOOK SALMON, ONCORHYNCHUS TSHAW^YTSCHA (WALBAUM)' BY Ted S. Y. Koo,= Research Associate Professor, and Andhi Isarankura,^ Fisheries Biologist, FISHERY RESEARCH INSTITUTE, College of Fisheries, University of Washington, Seattle,Washington 98105 ABSTRACT This study uses an objective method that measures and graphs the spacings of circuli. It also introduces a new method of diflFerentiating ocean nucleus from stream nucleus. Four groups of chinook salnlon scales were studied, each with a specific purpose. First, scales from recoveries of two kinds of marked fall chinook at Spring Creek National Fish Hatchery were compared: one kind was from fish released as fry, and the other from fish released as fingerlings. In the nuclear part of scale growth, the group released as fry showed a larger variance in spacing of circuli than the group released as fingerlings but the difference in the mean values between these two groups was not significant. In the first marine part of scale growth, circulus spacing was significantly wider in the group released as fry than in the group re- leased as fingerlings. It was not possible, however, to identify individual scales as coming from fish released as fry or as fingerlings. Second, scales from marked and unmarked fall chinook salmon at Spring Creek Hatchery were compared to see if any effect of marking could be detected. Significant differ- ences in circulus spacing in marine growth existed between marked and unmarked fish, the latter having wider spac- ings. Marking was in the removal of adipose and right pectoral fins from chinook fingerlings. This technique was therefore regarded as having unfavorably affected the growth of marked fish. Third, scales from marked fall chinook that had been released at various times of the year at Little White Salmon National Fish Hatchery were studied. The scales showed that young chinook salmon released in May and July of the first year grew an ocean nucleus typical of fall chinook; those released in February of the second year grew a stream nucleus typical of spring chinook; and those released in September and October of the first year grew a nucleus intermediate in character. Fourth, scales of fall and spring chinook salmon were studied to see how these two groups could be identified by their scales. Measurements of circulus spacing in the first and second summer of marine growth revealed that, in the spring chinook, marine circuli in both summers were about equally wide; whereas, in the fall chinook, marine circuli of the second summer were nearly one and one-half times wider than those of the first summer. Thus, these scales can be distinguished, not by nuclear growth as is normally done by subjective judgment, but by relative marine growth as measured by objective means. Since the early studies on the scales of chinook salmon, Oncorhynchus tshawytscha (Walbaum), by Gilbert (1914), Rich (1922), Rich and Holmes (1929), and othei-s, little has been published on the subject. Many problems still deserve further study. The most important and interesting problem is the classification and identification of nuclear growth zones, the cen- tral part of scale growth. Gilbert (1914) classi- fies chinook scales into two types : those with an ocean nucleus and those with a stream nucleus. The ocean nucleus type originates from fish that 1 Contribution No. 235, College of Fisheries, University of ^Vashington. 2 Present address: Chesapeake Biological Laboratory, University of Maryland, Solomons, Md. 20688. 3 Present address: Department of Fisheries, Bangkok, Thailand. FISHERY BULLETIN: VOL. 66, NO. 2 Ptiblished April 1967. migrate seaward in their first year and thus has the first annulus at the end of the first year's marine growth; the stream nucleus type originates from fish that do not migrate sea- ward until the early months of the second year and thus has the first annulus at the end of fresh- water growth. To the former group be- longs the fall run of chinook, which enters the river from July through November; to the latter, the spring run. which enters the river from March to June. Classification of nuclear growth zones is very useful and today still serves as the foundation of age study of chinook salmon. This method is most useful when only two groups of chinook salmon are involved and their nuclear zones 165 are clearly defined. Its application becomes limited, however, when the boundaries of nu- clear growth are not clear-cut. The chinook salmon young of the Columbia River, for in- stance, migrate seaward throughout most of the year (Rich, 1922) ; consequently, the first year's growth is subject to numerous varia- tions that intergrade so completely that it is impossible to draw any sharp line of distinction (Rich and Holmes, 1929). Most Columbia River chinooks, according to Rich and Holmes (1929), have neither typical stream nor typical ocean nuclei, but apparently have spent part of the first year in fi'esh water and part in the ocean. The result has been a nuclear area composed in part of stream growth with narrowly spaced circuli and in part of ocean growth with widely spaced circuli to form what these authors term "composite nucleus." The composite nucleus makes age determina- tion difl^cult. In a composite nucleus, the amount of stream growth varies inversely with the amount of ocean growth. At one extreme is the type with only a small amount of stream growth accompanied by a large amount of ocean growth. At the other extreme is the type with a great amount of stream growth ac- companied by a small amount of ocean growth. The first type of nuclear growth approaches the ocean nucleus, and the second type ap- proaches the stream nucleus. Between these two extremes there are complete intergrada- tions. This poses the question : "Where should the annuli be placed, and how many?" The question is further complicated by the formation of the so-called "intermediate growth," that is, growth of circuli in the estuary while the fish is migrating seaward. Circuli of this growth cannot be distinguished with certainty from either the stream or the ocean circuli, and they often form a check which, in the words of Rich and Holmes, "might easily be mistaken for an annulus by an inexpe- rienced observer." These same authors main- tain that with experience this kind of error may be eliminated almost completely, and that their own experience with the scales of fish of known history has provided sufficient information for correct age determination. The prerequisite of experience in scale read- ing cannot be denied, but the dependence upon experience can be lessened and the accuracy of age determination improved if some mechanical method in scale work can be developed so that the scale growth and marks can be interpreted more objectively. The development of an objective method is the major purpose of the present work. From a large number of scales collected from chinook .salmon of known ages through recoveries of marked fish, we were able to establish some definite criteria and methods whereby one can objectively interpret scale marks with a minimum amount of guess work. The present study comprises four parts. First, scales from adult fall chinook that have migrated seaward as unfed fry and as fed fin- gerlings ' were compared in an attempt to find characteristics that might serve to identify fish of unknown origin; i.e., whether they come from fry migrants or from fingerling migrants. Second, comparative studies were made be- tween scales from fall chinook that had been fin-clipped when released as fingerlings and those that had not been marked. This was to see if marking had any adverse effect on growth that could be detected by scale measure- ments. Third, marking experiments on young fall chinook performed by U.S. Fish and Wildlife Service personnel at the Little White Salmon Hatchery provided an unusually valuable series of adult scale samples for age and growth study. Young chinook salmon were released over a wide range of time (May to February), and each release had a different mark. Scales from returned adults originating from different re- leases were studied to gain insight into the formation of a fresh-water annulus and to assess the relative amount of first and second year's ocean growth due to different release dates. This provided valuable information for understanding scale growth patterns in fall and spring chinooks. Fourth, the relative amount of the first and second year's ocean growth on scales in known stocks of fall and spring chinooks was studied and compared. An objective method of determining the presence or absence of an an- 1 "Fed finKerlinKs" refers to young chinook salmon that have been fed for about 3 months. 166 U.S. FISH AND WILDLIFE SERVICE niilus in nuclear growth and therefore in dis- tinguishing fall and spring chinooks was de- veloped, independent of fresh-water growth itself. MATERIALS AND METHODS Materials for the present study were supplied by the Fish Commission of Oregon and by the Portland Program Office of the Bureau of Com- mercial Fisheries of the U.S. Fish and Wildlife Sei'vice. Scale impressions on cellulose acetate cards of the following were available for study. 1. Returns of marked fall chinook to Spring Creek Hatchery: a. 1958 returns — released as fry, 1 fish ; released as fingerlings, 8 fish. b. 1959 returns— released as fry, 8 fish; released as fingerlings, 173 fish. c. 1960 returns — released as fry, 28 fish; released as fingerlings, 158 fish. 2. Returns of unmarked fall chinook to Spring Creek Hatchery: a. 1959 returns— 925 fish. b. 1960 returns— 898 fish. 3. Returns of marked fall chinook to Little White Salmon Hatchery: Mark* Date released Fish returns Fish returns Xo. Xo, Xo. Xo. Xo. Xo. LP— May S-9 (1957, 1958) 10 3 13 12 6 18 RP— Jiily 1-2 (1957, 1958) 27 6 33 28 18 46 D-LP— Sept, . 4 (1957, 1958) 3 3 4 2 fi D-RP— Oct. 15 (1957, 1958) 6 1 7 2 2 4 An-RP— Oct. 15 (1957, 1958) 20 20 20 11 31 An-LP— Feb. 13-15 (1958, 1959) 43 10 53 48 24 72 Total .... 109 20 129 114 63 177 *L=left; P=pectoral; R — right; D = dorsal; An=anal. 4. Returns of unmarked spring chinook to Carson National Fish Hatchery: Samples of several hundred scales (one from each fish) each year collected during 1955—57 and 1959-60. In addition to the above impressions of scales of adult chinooks, specimens of young fall chinook, preserved at the time of release at several Federal hatcheries, were also available. Scales from these young fish were studied for comparison with the nuclear zone of adult scales. The study was based on objective means as much as possible. Scales were studied under a microprojector at magnifications of 92, 140, or 400 times, depending on the magnification de- sired. The image of a scale was projected di- rectly on millimeter graph paper, and the posi- tions of circuli along the antero-lateral radius of the scale were marked on the paper. The center of the central plate was always used as the starting point, and the edge of the central plate became the first mark. In counting and measuring the circuli, we regarded the mark next to the central plate as the first circulus. All subsequent studies of the scale growth were made from the markings on the graph paper. Distances were measured in terms of millimeters, and the actual dimensions deter- mined by the magnifications used. The various methods of counting, measuring, and graphing will be described under individual sections. SCALE GROWTH IN FALL CHINOOK SALMON RELEASED AS FRY AND FINGERLINGS Spring Creek Hatchery (fig. 1), a Federal installation located about 175 miles from the mouth of the Columbia River on the Washing- ton side, produces primarily fall chinook sal- mon. In the past, young chinooks were released either as unfed fry during the first week of February or as fed fingerlings during the first week of May. To evaluate the relative merits of fry and fingerling releases, the Bureau of Commercial Fisheries marked young chinook salmon of brood years 1956, 1957, and 1958. Among the young released each year during 1957-59, some fish were marked, consisting of about equal numbers of fry and fingerlings. Two combinations of fin marks were used: adipose and left pectoral fins on fry and adipose and right pectoral on fingerlings. Fish with both marks were recovered in years 1958-60,^ and scales were collected from all returned fish. An interesting question here is : "Can the scales of adults that were released as fry be differentiated from those that were released as fingerlings?" This problem is of both theoretical and practical importance. Later recoveries are not included in the present study. SCALES OF CHINOOK SALMON 167 LOWER COLUMBIA RIVER WATERSHED SCALE SruOT AREAS ^ fISM MArCHCHlES 10 20 MILES Figure 1. — Map of lower Columbia River watershed, showing locations of the three National Fish Hatcheries from which study materials were obtained. Theoretically speaking, these two groups of fish should have differences in the growth pat- tern of scales, because at the time of release the fry have not started to grow scales, where- as the fingerlings have already grown scales with circuli. The nuclear area (the central portion of the scale) , or at least the initial part of it, of adult scales originating from these two groups of fish must have then grown under different conditions: that from the fry group in river water under lower but variable tem- perature '■ and feeding conditions, and that from the fingerling group in hatchery water with higher and nearly constant temperature ' and ample food. It may be expected then that scales Water temperatures of Columbia River at Bonneville Dam in February were 2.8-4.4 T. (1957); 6. 1-8. ST. (19.">8); 3.9-.i.0<>C. (1959). U.S. Army Corp.s of Kngineers. Annual Fish Passape Reports. 7 Water temperature in Spring Creek Hatchery is about 7.8='C. year-round. from fish released as fry should have more closely spaced circuli in the nuclear zone than those released as fingerlings. Further, because the fry were released 3 months before the fingerlings, they should reach the ocean earlier and consequently may have a different pattern of ocean growth than have the fingerlings. Practically speaking, if the origin of release — whether fry or fingerling — of returning adults can be identified through .scale charac- ters, then the two methods of release can be evaluated without having to mark young fish. Junge and Phinney (1963) indicate that fish released as fingerlings have a much greater survival rate than have fish released as fry. Therefore, the elimination of mai-king would not only save the costs of marking but also eliminate any possible harm that marking may cause the fi.sh. 168 U.S. FISH AND WILDLIFE SERVICE To find out whether actual differences existed between scales of adult fish released as fry and those from adults released as fingerlings, we selected scale samples from brood year 1956 because that year had the largest number of specimens that were released as fry. Returns of fish released as fingerlings are plentiful for analyzing this group in any brood year. In 1959, eight fall chinook salmon with Ad- LP mark (released as fry) were recaptured. Of these, seven were 3 years old and therefore came from 1956 brood. In 1960, 28 such marked chinooks were recaptured and 16 of these were 4 years old and of the 1956 brood. This total of 23 scales that belonged to the 1956 brood, plus 102 3-year-olds that were recaptured in 1959 with Ad-RP marks (released as fingerlings) and 167 4-year-olds that were recaptured in 1960 with the same mark, provide the samples for the following study. Based on theoretical considerations given earlier, we used two purely objective methods aimed at detecting any diff"erence these two groups of scales might have in growth in fresh water or the first year of growth in the sea. The first objective method was that of com- paring growth patterns revealed by scale graphs based on spacing of circuli. Under a magnification of 140 times, the circuli were marked along the antero-lateral radius on a millimeter graph paper. We then divided the radius into 20-mm. units and calculated the mean spacing of circuli of each unit. For each group of scales, the means of circulus spacing of a unit were summed and averaged to give the mean of the group. When the group means were plotted on the ordinate against the radius units on the abscissa, we obtained a scale graph which shows the growth pattern. Figure 2 shows information on groups re- leased as fry and as fingerlings. The fresh-water growth part of figure 2 shows a similar pattern for the two groups, namely, circuli are wide at the start but rapidly narrow down : the extent of growth covers about the same distance on scale radius. Also, there is only a slight difl'erence in the mean spacing of circuli. Such difference, as will be shown in the second method, is not statistically significant. S 6 ?. 4- ■ MARKED FRY. N = 16 -MARKED FINGERLING. N=17 12 3 4 5 6 7 8 DISTANCE FROM FOCUS ALONG ANTERO-LATERAL AXIS IN 20-MM UNITS (XMO) FIGURE 2 Figure 2. — Spring Creek Hatchery chinook salmon: Mean scale graphs showing pattern of fresh-water growth and the major portion of first year's marine growth of group marked as fry (solid line) and of group marked as fingerlings (dash line). Marked difference, however, is evident in the marine growth section of figure 2 (units 4 to 8). The group released as fry has much wider cir- culi at every unit than has the group released as fingerlings. This is, of course, only a reflec- tion of group difference, as the values plotted are mean widths. At each unit, the mean cir- culus widths of the two groups of scales overlap widely so that we could not identify the group origin of individual scales on that basis. Ex- amples of scales of adults that were released as marked fry and as marked fingerlings are shown in figures 3 and 4. The second objective method, aimed at detect- ing differences in first year growth of fish released as fry and as fingerlings, was to meas- ure and compare the total distance of the first 5 circuli, of the first 10 circuli, and of 10 circuli counted from the 16th through 25th circulus. The reasons for the selection of these three measurements are as follows: Fish released as fingerlings have developed, in the hatchery, the first 5 circuli and most, if not all, of the first 10 circuli ; but the fry that are released develop all circuli in the natural environment. We measured the first 5 and first 10 circuli, there- fore, to detect differences in initial fresh-water growth. The third measurement, distance from the 16th through 25th circulus, was made to SCALES OF CHINOOK SALMON 169 Figure 3. — A scale of Spring Creek Hatchery chinook salmon that was released as marked fry. Note the relatively wider spacing between circuli in the first year of marineg rowth. Figure 4. — A scale of Spring Creek Hatchery chinook salmon that was released as a marked fingerling. Note the relatively narrower spacing between circuli in the first year of marineg rowth. study first-year marine growth, because these 10 circuli always represent the major but not the entire part of the first summer growth in the ocean. Using the 10 circuli enables us to have more consistent measurements than we would obtain by measuring the entire first sum- mer growth, because we cannot delimit exactly the first and last circuli of summer growth. Circuli 11 through 15 were purposely skipped, for they may represent some transitional growth and therefore are quite variable as a group. In reference to scale graphs, the initial 10 cir- culi are represented by the first two and a half units on the abscissa; and circuli 16 to 25, by units 4 to 6. In essence, the measuring method enables us to check on the graphing method, for we can tabulate the data and subject the results to statistical tests. The results of the measurements and statis- tical tests are shown in table 1. In all the tests between the paired .sample means, we first tested for the variances (s-) and then applied the appropriate f-test. In the comparisons of the fii\st 5 circuli and of the first 10 circuli, the variances of the paired samples are significantly difi'erent, and f-test shows that the sample means are not sig- nificantly different. This is to say that although circulus spacing in the initial 5 or 10 circuli is more variable in the group released as fry (larger variance) than in the group released as fingerlings, the average values do not differ significantly between these two groups. The latter point confirms the results of the scale graph method In the comparison of the 10 circuli counted from 16th to 25th circulus, the variances are not significantly different, and t-test shows that the sample mean of "A" (fry relea.ses) is sig- nificantly larger than that of "B" (fingerling releases). This is to say that the groups re- leased as fry and as fingerlings have a similar amount of variation in circulus spacing for circuli 16 to 25, but that the average spacing of circuli in the group released as fry is larger than that in the group released as fingerlings. The latter point also confirms the results of the scale graph method. 170 U.S. FISH AND WILDLIFE SERVICE Table 1. — Frequency and stalislics of Mai distance of circnli of two groups of scales from salmon returning lu Spring Creek Hatchery: A — adult chinooks that were marked and released as fry; B — adult chinooks that were marked and released as fingerlings Distance First 5 circuli First 10 circuli Circuli 16-25 A B A B A B Mm. X 10 12-14 Number Number 2 12 66 70 74 31 10 3 1 Number Number Number Number 6 5 6 4 1 1 16-18 18-20 20-22 1 24 26 1 3 14 22 47 57 47 43 22 1 4 1 1 2 1 28-30 1 5 2 1 1 2 1 2 1 1 2 1 5 3 1 3 2 3 9 30-32 13 32-34 36 34-36 41 36-38 45 38-40 36 40-42 36 42-44 24 44-46 14 46-48 10 48 50 1 2 62-54 54-56 I N 23 269 18.8 19.6 12.69 6.94 1.09<^<3.66 SB'- -1.05 53 23 269 34.3 35.8 32.68 ^ 14.97 1.31<^<4.36 SB- -1.24 24 23 269 X 39.8 37.7 32.63 23.12 95% confi- dence limits for ratio of population variances. (-statistic d.f.. 0.84<^<2.82 SB' 2.10* 290 Value of ((] at0.05signifi- cance level. 95% confi- dence limits for differ- ence be- tween popu- lation means. 2. -2.34<,i Dl -M!<0.74 2. -3.99 siRiiifi- cance level. 9.T percent confi- dence limits for (lifTerence lir'tueen population means. Marked fish 33-34 inches long Nu mber I'mnarked fish 33-34 inches long Number Marked fish 33-34 inches long Number Unmarked fish 3.i-3« inches long Number 17 40.88 30.23 4.'i.47 23.89 0.29<-'A;<2.I9 SB- -3.58* 32 2.04 -9.95<;i,-;i!<-2.73 17 40.88 30.23 4.12* 26 47.27 22.12 41 2.02 3.25<^i-jii<9..'i3 can be accepted ; therefore, a simple t-test was used. The width of ten marine circuli was sig- nificantly greater in the unmarked groups than in the marked group, compared either between two modal lengths of fish or between fish of the same length. These tests thus confirm the re- sults obtained by the scale graph method. 172 U.S. FISH AND WILDLIFE SERVICE Growth rates of fishes are reflected in the spacing of scale circuli: the faster the growth rate, the wider the spacing between circuli. The present findings, therefore, suggest that mark- ing through the excision of adipose and right pectoral fins in chinook salmon may have been responsible for the slower growth rate of the marked fish. Biologists, using various fin marks, working on various species of fish, and experi- menting under various conditions, have obtain- ed contradictory results in this respect. Ricker (1949), for instance, excised the pectoral, both ventrals, or one pectoral and both ventrals of the largemouth bass and found that recoveries of these and unmarked fish indicated that the marked fish were significantly smaller than the unmarked ones. He believes that marking pos- sibly affected the growth rate directly; how- ever, when he marked 2-year-old bluegills, the growth of marked and unmarked fish was the same. Armstrong (1949) studied lake trout fingerlings and found no appreciable difference in length and weight between those that were unmarked and those that had had the adipose removed. Shetter (1951) also shows that re- moval of the dorsal and adipose fins, right pec- toral fin, or right pelvic fin from the fingerling lake trout had no effect on the growth of the marked fish but that removal of the left pectoral appeared to have slowed the growth of the fish. Again, on a study of growth of marked and unmarked lake trout fingerlings in the presence of predatory fish, Shetter (1952) found no significant difference in the growth rate be- tween marked and unmarked groups. In the Cultus Lake experiments on the sock- eye salmon, Foerster (1934, 1936a, 1936b) shows that unmarked smolts had a return rate two and one-half times greater than marked smolts that had both pelvics and adipose or both pelvics and dorsal removed. He shows further that this differential mortality was due to the effect of marking upon marine survival, since marking did not affect lake survival. No data on fish length or scale growth were given, however, so it is not known whether marking did have an adverse effect on growth. The reasons for the apparent paradoxical results on the effect of marking on the growth rate of fish by various workers may be quite SCALES OF CHINOOK SALMON varied. The different results could be due to different fins being clipped, different species be- ing experimented on, different techniques being applied, or different conditions under which the experiments were made. SCALE GROWTH OF FALL CHINOOK SALMON, RELEASED BETWEEN MAY AND FEBRUARY At Little White Salmon Hatchery (fig. 1), another Federal installation some 10 miles downriver from the Spring Creek Hatchery, the Bureau of Commercial Fisheries has con- ducted further marking experiments on fall Chinook salmon. Here, for the brood years 1956-58, young chinook salmon were reared for various lengths of time and released at five dif- ferent times of the year from May to February (see under "Materials and Methods"). A dif- ferent mark was applied for each release, so that at return a marked fish could be positively identified as to its date of release. The returns from these experiments offer excellent scale samples for studying the growth of nuclear zones. Fish released earliest (May) should go to sea during the first year, and their scales should show a typical ocean nucleus. Those released latest (February of the follow- ing year) spent the first winter in the hatchery, and their scales should therefore have a stream nucleus. Fish released between the above pe- FlGURE 7. — A scale of adult chinook salmon that was released as a fingerling in May at the Little White Salmon Hatchery (May 1957 release, 1959 return). 173 riods (July, September, and October) should have scale growth of intermediate nature. First, let us examine a typical scale of an adult Chinook that originated from May release (fig. 7). At the center of the scale, there are 14 closely placed fine circuli, which are followed by more widely spaced coarser circuli. A check- like structure (C) is present at the border be- tween the two zones. Most of the fine circuli represent intei-mediate growth that took place after the fish was released, because young Chinook released in May average only two to three circuli on their scales. The zone of more widely spaced coarser circuli that follows the check represents what is generally regarded as marine growth. It is bounded by a distinct band of closely placed circuli (fig. 7, I). Both the check (C) and the band (I) have the appear- ance of an annulus. But since this is known to be an age II fish (1957 release, 1959 returns), and since the second annulus (II) is evident near the resorbed margin of the scale, only one of the two marks can be regarded as a genuine annulus. Based on relative distance, the band (I) should be regarded as the first annulus. "C," therefore, is a sort of migration check. The entire growth up to and including the band (I), forms what is known as the ocean nucleus. In the ocean nucleus, then, an annulus in the fresh-water growth part is lack- ing, and that gives rise to the age terminology of "sub-one" for this group," or "O.", to use the terminology of Koo (1962). Next, let us examine a tipical scale of an adult Chinook that returned in 1960 from a February 1958 release (fig. 8). Here, there is also the central crowded area of fine circuli (I) and the surrounding wide marine growth that is bound- ed by a band of closely placed narrow circuli (II). Although "I" and "11" in this figure ap- pear to be corresponding respectively to "C" and "I" in figure 7, they are different in sig- nificance. Because the fish was held in the hatchery over the winter and was not released until February, "I" in figure 8 is a true annulus, not a mere check, as the "C" in figure 7. The central area up to "I" forms what is known as the stream nucleus and because the young fish s The term "sub-one" is derived from the subscript of Gilbert- Rich's (1927) scale formula, for example, 3i, 4j. Figure 8. — A scale of adult chinook salmon that was released as a finfrerling in February at the Little White Salmon Hatchery (February 1958 release, 1960 return). left fresh water during its second year, it is also referred to as "sub-two age," or "1.", meaning one annulus in fresh-water growth. This fish is known to be age III, so there can be only two marine annuli, which are labeled as II and III in figure 8. The narrow band (i) between these two annuli must therefore be regarded as an incidental check. From the standpoint of age determination, it is imperative that an ocean nucleus and a stream nucleus can be positively identified, for it will make a difference of 1 year in age, de- pending upon whether an annulus or a check is assigned to the central fine circuli area. No definite criteria can be found in literature that positively differentiate a mere check from a genuine annulus in this nuclear area of growth in chinook scales. Determination of age is usu- ally based on the appearance of the nuclear zone and is highly dependent upon personal judg- ment. Thus, a stream nucleus has been de- scribed as an area of many closely placed circuli bounded by a distinct narrow band of more 174 U.S. FISH AND WILDLIFE SERVICE closely spaced circuli, the annulus. An ocean nucleus, on the other hand, is recognized when the nuclear zone consists of relatively few but wider circuli that are not marked off by a dis- tinct check from the ensuing widely spaced ma- rine growth. Unfortunately, nuclear zones of many Chi- nook salmon scales are not clearly defined so that the morphology of the nuclear zones alone does not enable us always to differentiate with certainty the ocean nuclei from the stream nuclei. If, for example, the scales in figures 7 and 8 had come from fish of unknown age, we would have no real basis for calling one mark a mere check (C) and the other a true annulus (I). Obviously, something other than visual deter- mination must be devised. As we had avail- able a large number of scale samples from re- captured marked chinook salmon comprising both the stream type and the ocean type of nuclear growth, we were able to compare the characters of these two groups of scales on a quantitative basis. Because the nuclear zones of circuli failed to show significant differences, our study was extended to cover marine growth as well, and we have developed some criteria that help to guide the chinook scale reader to differentiate ocean from stream nuclei on a more objective basis. Before we discuss quantitative measure- ments, let us examine some scales of adult chinook that originated from releases during the intermediate period between May of the first year and February of the next year, to observe the transition from ocean nucleus growth type to stream nucleus growth type. A scale of a July release origin is shown in figure 9. Being similar to the scale of a May release origin (fig. 7), it also shows an ocean nucleus (I) and a strong check (C) for the nuclear area. Based on the known age of this Figure 10. — A scale of adult chinook salmon that was released as a fingerling in September at the Little White Salmon Hatchery (September 1958 release, 1960 return). Figure 9. — A scale of adult chinook salmon that was released as a fingerling in July at the Little White Salmon Hatchery (July 1957 release, 1959 return. Figure 11. — A scale of adult chinook salmon that was released as a fingerling in September at the Little White Salmon Hatchery (September 1957 release, 1960 return). SCALES OF CHINOOK SALMON 175 fish, we know that "I" marks an annulus and "C" is merely a check. An incidental check (i) is also present between annuli I and II. Two scales of adults that came from Septem- ber release are shown in figures 10 and 11. In figure 10, the marine growth of the first year (C to I) is much reduced as compared with the scale of May or July release origin (figs. 7 and 9). Consequently, the annulus (I) is getting closer to the check (C), and the entire ocean nucleus becomes much smaller in size. Because of this, it is easy to determine that the check (C) here is not an annulus. Further reduction in the first year's marine growth is seen in the second example of a September release (fig. 11). Here the entire nuclear zone assumes the appearance of a stream nucleus. Indeed, it is questionable whether there is any amount of true marine growth inside the first annulus (I). A scale of the October release origin (fig. 12) shows the same characteristics, i.e., a much reduced zone between "C" and "I," and a nu- clear zone that assumes the look of a stream nucleus. At least, as far as age determination is concerned, because the total age of this fish is known to be III, it is certain that "I" is the first and only annulus up to that point, much as "I" in a typical stream nucleus such as that Figure 12. — A scale of adult chinook salmon that was released as a fingerlinp in October at the Little White Salmon Hatchery (October 1957 release, 1960 return). of a February release origin (fig. 8). From the above series of examples, it is evi- dent that when the young chinook salmon were released as hatch-of-the-year from May through July, they entered the ocean during the grow- ing season of the first year after some sojourn in the river. As a result, there was a large number of wide mai'ine circuli outside the cen- tral zone of narrow fresh-water circuli, result- ing in a large ocean nucleus. As the release date became later and later in the year (Sep- tember and October), however, the chinook sal- mon would miss more and more of the current season's marine growth, and the result was a nuclear type similar in appearance to a stream nucleus. Finally, when the young chinook sal- mon were reared in fresh water over winter and were not released until February of the second year, the nuclear zone was composed solely of fresh-water growth, and any marine growth belonged to the following year. For all practical purposes, scales from September and October releases should be treated as stream nucleus type, for there is no way of knowing that "I" is not a stream annulus without the knowledge of release date. Because the fresh-water growth part in an ocean nucleus may not be distinguishable from that of a stream nucleus, we extended our study into the marine growth of the first and second years of ocean life to find diff'erences between these two types. In this study, 72 returns from May and July releases were treated as one group representing the ocean nucleus type, and 85 returns from October and February releases were treated as another group representing the stream nucleus type. The method consists of first locating the ap- parent first marine annulus, i.e., a band of nar- row circuli after a zone of wide circuli. This is "I" in figures 7 and 9 and "11" in figures 8, 11, and 12. Then, from the midpoint of this annulus band 20 circuli were counted out- ward toward the edge of the scale along an antero-lateral radius, and the total distance of these 20 circuli was measured and repi'esented by "A." This represents the major part of the second year growth in ocean for both gi-oups of scales. Similarly, 20 circuli were counted in- ward toward the focus and the total distance 176 U.S. FISH AND WILDLIFE SERVICE was measured as "B" (fig. 13), which repre- sents the first year growth in ocean for both groups of scales. Then we computed the ratio Figure 13. — The measurement of circulus spacing. Left, Chinook scale with an ocean nucleus; right, Chinook scale with a stream nucleus. 30 20 10 30 20 10 LITTLE WHITE SALMON RIVER HATCHERY MAY AND JULY RELEASES N = 72. x= i,;o ^ -^- ■ -ES- LITTLE WHITE SALMON RI\ER HATCHERY OCTOBER AND FEBRUARY RELEASES -^ I i N= 85, X=l 22 f^^ , 1.0 1.1 1.2 1.3 1.4 1,5 16 1.7 1.8 RATIO OF SECOND TO FIRST YEAR OCEAN GROWTH (j) of A/B, which is the ratio of second year marine growth to the first year marine growth. We found that in the May to July release group, circuli of the second year marine growth (A) were, on the average, nearly 50 percent wider than the first year marine growth (B) ; whereas in the October to February release group, "A" was only 22 percent wider than "B". The frequency distribution of the ratio A/B of these two groups of scales is shown in figure 14. It is obvious that the two groups are distinctly different in the value of A/B, but there is also enough overlap so that not all scales can be identified to their nuclear growth type on this character alone. DIFFERENTIATION OF FALL CHINOOK AND SPRING CHINOOK SCALES BY MARINE GROWTH The fall chinook scales normally have a typical ocean nucleus (sub-one age), and the spring chinook scales normally have a typical stream nucleus (sub-two age). The nuclear growth part of these two types of scales cannot always be distinguished. So in order to identify these two groups of fish, we applied the method of comparing first and second year's marine 15 I :: 10 o 5 z 60 40 20 FALL CHINOOK FROM SPRING CREEK HATCHERY N= SO. X= 142 -Ea i f^ [ M I -PSSJ- SPRING CHINOOK FROM CARSON HATCHERY N = 109. X= 1.04 I £3- 0.9 1.0 11 1.2 13 1.4 15 1.6 1.7 1.8 1.9 Figure 14. — Frequency distribution of the ratio of second to first year marine growth (A/B) of little white salmon Hatchery recaptures of May and July releases, and October and February releases. FIGURE IS Figure 15. — Frequency distribution of the ratio of second to first year's marine growth (—A/B) of fall chinook and spring chinook. SCALES OF CHINOOK SALMON 177 growth as developed from the study of Little White Salmon Hatchery mark recovery speci- mens. For study material, we used scales col- lected from unmarked fall chinook at Spring Creek Hatchery and those collected from un- marked spring chinook at Carson Hatchery. Scales of 50 fall chinook and 109 spring chinook were measured. The frequency dis- tributions of the ratio of second year's to first year's marine growth A/B of these two groups of fish are shown in figure 15. These distribu- tions show clearly that fall and .spring chinooks are distinctly separate groups as far as the character of marine growth is concerned. The difference between the two groups is similar to that between May to July release group and October to February release group of Little White Salmon Hatchery fall chinook. In other words, the fall chinook are similar to May to July release group in having a large A/B ratio, and the spring chinook are similar to October to February release group of fall chinook in having a small A/B ratio. The outstanding feature of spring chinook scales is that the marine growth of the fir.st year is nearly as good as that of the second year, so that its A B ratio approximates 1.0, as compared with 1.2 for the fall chinook released Figure 16. — A scale of a spring chinook salmon. 178 in October to February. In fact, this character alone is often sufl!icient to distinguish a spring run from a fall run of chinook salmon. An example of a spring chinook scale is shown in figure 16, in which the circuli inside of the first marine annulus (II) are as widely spaced as circuli outside of it. SUMMARY AND CONCLUSIONS Scales of Columbia River chinook salmon were studied to find answers to the following questions : 1. Is it possible, from structures of adult chinook scales, to identify whether a fish has originated from fry or fingerling migrant ? The answer is negative. Scales from marked fish recoveries showed that there was no sig- nificant difi'erence in the mean values of cir- culus spacing in nuclear growth part between chinook salmon released as fry and those released as fingerlings, although the spacing is more vai'iable in the fry than in the fingerling group. In the first marine growth, circuli in the group released as fry are more widely spaced than in the group released as finger- lings. While the difference is statistically sig- nificant, there was too much overlap so that identification of individual scales was not pos- sible. 2. Can the effect of marking, if any, on growth of chinook salmon be detected by scale studies? The an.swer is positive. At the Spring Creek Hatchery, fall chinook fingerlings were marked by removal of adipose and right pectoral fins. When scales from marked fish recoveries were compared with those from unmarked fish re- turns, circulus spacing in marine growth of the marked group was found narrower than the unmarked group. 3. How do the scales of early season (May- July) releases of fall chinook differ from those of later season (October-February) releases? In answering the above question, we found some interesting relations between scale growth patterns and times of release. Early season releases of fall chinook resulted in an ocean nucleus type (sub-one age) that is typical of fall chinook scales. Late season releases, how- ever, resulted in a stream nucleus type (sub- U.S. FISH AND WILDLIFE SERVICE two age) resembling that of spring chinook scales. Moreover, these two groups of scales are different in marine growth patterns. When circulus spacing in the second year marine growth is compared with that in the first year marine growth, the ratio is far greater for the sub-one group than for the sub-two group. 4. Can fall chinook scales be separated from spring chinook scales by objective means? The answer is positive. Fall and spring chinooks can be differentiated by their scales. Differentiation, however, is not made from nuclear growth patterns as is usually done visually, but is achieved objectively by compar- ing marine growth circuli of the first 2 years, the same technique as used for eai-ly-season and late-season releases of fall chinook. In the spring chinook, circuli in the second year of marine growth are nearly 50 percent more wide- ly spaced than those in the first year ; whereas in the fall chinook, they are about the same. ACKNOWLEDGMENTS The most important ground work in con- nection with the present paper had been done before we started our studies, for marking experiments on the Columbia River chinook salmon and collecting scale samples and perti- nent data had been performed by the Bureau of Commercial Fisheries of the U.S. Fish and Wildlife Service. Paul Zimmer, Harlan E. John- son, and Roy Wahle of the Bureau provided the study material and data. The Fish Commission of Oregon and the Washington State Depart- ment of Fisheries, which are also studying the Columbia River chinook salmon, supplied addi- tional material and information, and in this con- nection, I was assisted by Sigurd J. Westrheim, Raymond A. Willis, and Robert N. Thompson of the former organization and Peter Bergman of the latter agency. Charles E. Walker of the Canadian Department of Fisheries in Van- couver, British Columbia, sent us some dupli- cate scale impressions of Big Qualicum chinook salmon for our study, and Gei-ald J. Paulik of the Fisheries Research Institute of the Univer- sity of Washington advised us on the statistical treatment of the data. LITERATURE CITED Armstrong, George C. 1949. Mortality, rate of growth, and fin regenera- tion of marked and unmarked lake trout finger- lings at the Provincial Fish Hatchery, Port Arthur. Ontario. Trans. Amer. Fish. Soc. 77: 129-131. FOERSTER, R. E. 1934. An investigation of the life history and propagation of the sockeye salmon (Oncorhyn- chus nerha) at Cultus Lake, British Columbia. No. 4. The life history cycle of the 1925 year class with natural propagation. Contrib. Canada. Biol. Fish., Series A 8 (27) : 347-355. 1936a. An investigation of the life history and propagation of the sockeye salmon (Oncorhyn- chus nerka) at Cultus Lake, British Columbia. No. 5. The life history cycle of the 1926 year class with artificial propagation involving the liberation of free-swimming fry. J. Biol. Bd. Can. 2(3) : 311-333. 1936b. The return from the sea of sockeye salmon (Oncorhynchus nerka) with special reference to percentage survival, sex proportions and prog- ress of migration. J. Biol. Bd. Can. 3(1) : 26-42. Gilbert, Charles H. 1914. Age at maturity of the Pacific Coast salmon of the genus Oncorhynchus. U.S. Bur. Fish., Bull. 32: 1-22. Gilbert, C. H. and W. H. Rich. 1927. Second experiment in tagging salmon in the Alaska Peninsula fisheries reservation, sum- mer of 1923. Ibid. 42: 27-75. JUNGE, Charles O., Jr., & Lloyd A. Phinney. 1963. Factors influencing the return of fall chinook salmon (Oncorhynchus tshan-iitscha) to Spring Creek Hatchery. U.S. Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 445, iv-32 pp. Koo, Ted S. Y. 1962. Age designation in salmon. Univ. Wash., Publ. Fish., New Series 1: 41-48. Rich, Willis H. 1922. Early history and seaward migration of chinook salmon in the Columbia and Sacramento Rivers. U.S. Bur. Fish., Bull. 37: 1-73. 1926. Growth and degree of maturity of chinook salmon in the ocean. U.S. Bur. Fish., Bull. 41: 15-90. Rich, Willis H., and Harlan B. Holmes. 1929. Experiments in marking young chinook salmon on the Columbia River, 1916 to 1927. U.S. Bur. Fish., Bull. 44: 215-264. Richer, W. E. 1949. Effects of removal of fins upon the growth and survival of spiny-rayed fishes. J. Wildl. Manage. 13(1): 29-40. Shetter, David S. 1951. The eff"ect of fin removal on fingerling lake trout (Cristivomer namaycnsh). Trans. Amer. Fish. Soc. 80: 260-277. SCALES OF CHINOOK SALMON 179 1952. The mortality and growth of marked and Pacific Division, Bonneville, The Dalles and unmarked lake trout fingerlings in the presence McNary Dams, Columbia River, Oregon and of predators. Trans. Amer. Fish. Soc. 81: 17-34. Washington, 1957-59. Prepared by the U.S. Army Engineer Districts, Portland and Walla Walla, U.S. Army Corps of Engineers. Corps of Engineers. Portland, Oregon. [Each 1957-59. Annual fish passage report, North year published separately.] 180 U.S. FISH AND WILDLIFE SERVICE CATCH AND ESTIMATES OF FISHING EFFORT AND APPARENT ABUNDANCE IN THE FISHERY FOR SKIPJACK TUNA {KATSUWONUS PELAMIS) IN HAWAIIAN WATERS, 1952-62 By Richard N. Uchida. Fishery Biologist. Bureau of Commercial Fisheries, Biological Laboratory Honolulu, Hawaii SXj812 ABSTRACT Detailed data on catch and effort are obtained each year from all vessels that fish full time in the Hawaiian skipjack tuna fleet. These data permit description of the fishery and inferences about the abundance of skipjack. Our past measures of abundance have been stated in terms of total catch and catch per unit of effort calculated in terms of productive trips of all sizes of vessel. This study offers information on changes in the apparent abundance of skipjack in Hawaiian waters calculated from standardized units that will be unaffected by changes in the numbers of small and large vessels in the fishing fleet. Effort was measured in terms of an "effective" trip, which was defined as a trip in which skipjack were caught. Bias introduced by the lack of data on zero-catch trips is dis- cussed. The number of men hooking per trip declined in 1950- 60; however, those that remained in the fishery increased their catch rate by shifting their emphasis in fishing tech- nique from "grasping and unhooking" each fish to "flip- ping," a method in which the fisherman swings the fish aboard and by relaxing the tension on the pole, permits the hook to fall clear of the fish's mouth. The higher catch rate from the "flipping ' method was one of the factors that offset the effect of the decline in the number of men. Another factor that appeared to increase the catch rate was the reduction in the number of small vessels that did poorly. The vessels were separated into two size classes: Class 1, with bait-carrying capacities of less than 800 gallons per baitwell; Class 2, with bait capacities of more than 800 gallons per baitwell. The catch was standardized to Class 2 vessels. The catch per standard effective trip (Y/f) and the total catch fluc- tuated similarly in all years. The Y/f had no apparent trend and averaged about 5,700 pounds. The Y/f and the relative effective fishing intensity were not correlated significantly over the 11-year period. Year-to-year changes in apparent abundance seem to be independent of changes in fishing effort. I concluded that variations in the availability and vul- nerability of skipjack contribute to fluctuations in landings. The variations in strength of year classes also may have contributed importantly to fluctuations in the landings. In Hawaii, the skipjack tuna, Katsuivonus pelamis (Linnaeus), or aku, as it is called locally, supports the State's most important commercial fishery, contributing about 66 per- cent by weight to the total Hawaiian marine catch, and accounting for about 40 percent of the total annual ex-vessel value. The fish are caught exclusively by pole and line from schools which are concentrated at the stern of the vessel by chumming with live bait. The fishery is highly seasonal. Landings have ranged from about 29,000 pounds in January, typically a poor month, to about 3.7 million pounds in July, when the catch usually is large. Four- to 5-pound fish usually are caught throughout the year, but between May and September, larger fish, ranging between 13 and 25 pounds, are also taken. The latter contribute a large percentage by weight to the total annual catch. Not only does the catch fluctuate by month but also by year. In 1952-62, the yearly landings ranged between 6.1 and 14.0 million pounds, apparently with changes in the numbers of the larger fish at the islands. In the past, abundance of skipjack tuna in Hawaiian waters has been measured in terms of total catch and catch per unit of effort cal- culated in productive trips of all sizes of vessels — uncorrected fishing effort (Yamashita, 1958 ; Shippen, 1961). In general, total catch is not a dependable measure of abundance, because it is affected seriously by changes in the amount FISHERY BULLETIN: VOLUME 66, NO. 2 Published April 1967. 181 of fishing effort and by weather and sea con- ditions. Catch per unit of effort calculated in terms of uncorrected fishing effort is also un- reliable because it varies from year to year with changes in the fishing fleet. The fleet is made up of vessels of different sizes, and num- bers of these change as some enter and some leave the fishery. Both size and number affect the catch per unit of eft'ort. An important phase of this study is the derivation of a measure of abundance of skipjack in Hawaiian waters based on catch per unit of effort in standard- ized units that are unaffected by changes in the fishing fleet. SOURCES OF MATERIALS The basic data for this study were obtained from Fish Catch Reports (January 1952 to June 1954) and Aku Catch Reports (July 1954 to December 1962) submitted by the fishermen to the Hawaii Division of Fish and Game. Catch reports of only those vessels that fished for skipjack tuna full time were u.sed. The report form has undergone several revisions through the years, but all versions have carried spaces for the following information: The date of landing, the pounds of skipjack caught, and the fishing area. Yamashita (1958) described the method of reporting the areas fished by a skipjack vessel. Briefly, a fisherman reports only the code number corresponding to the statistical area where the catch was made. These areas are indicated on the Division of Fish and Game's Fisheries Chart No. 2 (see Yamashita, 1958: fig. 2). Data on number of men hooking per trip were obtained from Aku Boat Interview Sheets (January 1950 to July 1956), which were col- lected and checked by the personnel of the Division of Fish and Game, from logbook records (1957-59), and from Sampan Interview Records (August 1959 to June 1961). DESCRIPTION OF THE FISHERY The present brief de.scription of the fisher>' and review of fishing operations is based on June (1951). The number of skipjack tuna sampans fishing full time reached a maximum of 28 in 1951, but since then has declined; in 1963 only 20 vessels were fishing full time for skipjack. The vessels, generally of wooden con- struction, range from 58.3 to 80.5 feet in regis- tered length and from 27 to 77 in gross tonnage. These vessels carry 6 to 14 men per fishing trip. The nehu or anchovy, Sfolcphorus purpiireus Fowler, makes up about 92 percent of the bait catch; a second bait is the iao or silverside, Prnncsus insularum (Jordan and Evermann). Each vessel catches its own bait, fishing day pnd night until a sufficient supply is obtained. All the vessels have si.x baitwells with screened holes at the bottom through which sea water circulates. The Hawaiian skipjack tuna fishermen usual- ly confine their fishing and scouting operations to waters within 90 miles of the main islands. Skipjack on the fishing ground are indicated to the fishermen almost exclusively by bird flocks which are often associated with schools of fish. When a school has been sighted the captain attempts to intercept it. Once the head of the school is reached, water sprays are turned on and the "chummer" scatters live bait into the water. If the skipjack bite, the fishermen begin fishing off the stern. Fishing continues until the bait supply is exhausted or until the captain decides that further fishing is not worthwhile. If chumming is unsuccess- ful, the school is abandoned and scouting is re- sumed. The sampans may encounter several .skipjack tuna schools during the day, but the fish may bite in only about half of them. Scouting and fishing are discontinued as dark approaches, and the ves.sels usually head for port to unload the day's catch. TRENDS IN CATCHES OF SKIPJACK TUNA To show the trends in catches of skipjack tuna from Hawaiian waters, catch statistics were summarized by months and quarters for each year and by two broad geographical areas. Comments on the trends of catches are based on tabulation of data for 1952-62; therefore, the results may not be in complete agreement with those published for 1948-53 by Yamashita (1958). 182 SKIPJACK IN HAWAIIAN WATERS Table 1. — Monthly, rpiarlerhj, and annual catches of skipjack tuna in Haicaii, 1952- IThousands of poundsl Period of time Month: January. -. February,. March April May June July August September October. .- November. December. Quarter: First Second Third Fourth Annual 29 90 56 387 678 818 1,654 1,758 987 576 110 249 175 1.783 4.399 935 7,292 1953 200 204 576 864 1,240 2.241 1.510 2.142 1,281 1.199 218 384 980 4.345 4.933 1.801 22,059 271 185 359 7.59 1.132 2.804 3.705 2.178 1.049 1.121 390 815 4.095 1 6,932 1.579 14,021 1955 118 263 681 1,399 2,197 1,824 1,170 993 440 272 239 479 4.277 3.987 951 9.694 1956 322 281 230 433 1.375 1.926 2.321 1 . ,586 1.065 725 635 243 833 3.734 4.962 1,603 11.132 524 133 4,54 331 816 812 919 698 489 530 282 142 1,111 1,959 2,106 954 6.130 638 236 15! 792 277 948 1,405 1,191 631 186 144 235 1.025 2.017 3,227 565 6,834 215 107 397 839 1,757 1.679 2.382 1.815 1.377 1,064 626 155 719 4,275 5,574 1,845 12,413 1960 242 179 327 411 842 776 1.430 1.396 690 439 219 409 748 2,029 3.516 1,067 7,360 1961 492 247 600 620 930 2,721 2,288 1,359 705 516 213 203 1.339 4,271 4,352 932 10,894 459 412 199 480 1,178 2.308 1.809 922 568 525 170 385 1.070 3.966 3.299 1.080 9.415 Range 29-638 90-412 56-600 331-864 277-1 . 757 776-2.804 919-3.705 698-2.178 489-1.377 186-1.199 110-635 68-409 175-1.339 1.783-4.695 2.106-6.932 565-1.844 6.130-14.021 Average 317 199 328 600 1.048 1.748 1.932 1.474 893 666 298 246 845 3.395 4.299 1.210 9.749 MONTHLY, QUARTERLY, AND ANNUAL CATCHES The seasonal character of the Hawaiian skip- jack tuna fishery is shown by the monthly and quarterly catches in 1952-62 (table 1). The catch usually increased gradually from April to a peak in June, July, or August and then declined progressively to a low level in Decem- ber. Usually February had the smallest catch and July the largest. Also, the catch usually rose in January following the progressive de- cline from the summer peak to December. The quarterly catches showed the same trend as the monthly landings. First-quarter catches were usually the smallest and averaged 0.8 million pounds. Second-quarter catches re- flected the increased fishing activity during the spring and averaged about 3.4 million pounds. Third-quarter landings were rather consistently the largest and averaged 4.3 million pounds ; only in 1955 and 1962 did second-quarter catches exceed those of the third quarter. The fourth-quarter catches declined to an average of 1.2 million pounds. The variations of the annual catches were large. In 1952-62, there were 4 poor years — 1952, 1957, 1958, and 1960— in which the matches were far below the 11-year average of 9.7 million pounds. The catches in 1955 and 1962 were close to the 11-year average, and those of the remaining years were above av- erage. The maximum catch of 14.0 million pounds occurred in 1954; the minimum of 6.1 million pounds was in 1957. INSHORE AND OFFSHORE CATCHES For this study, I consider the inshore area to extend from the coastline to 20 miles at sea and the offshore area to include all statistical areas beyond 20 miles from the coastline. The catch reports used in this study were from vessels that fished for skipjack full time. The total weight landed by these vessels and the effort expended to produce it are hereafter called sample catch and sample effort. In addi- tion to the catches made by these vessels, catches were made by vessels that fished for skipjack tuna only part time. The total weight landed by vessels that fished full time and those that fished part time is hereafter called total catch and the effort expended to produce it is total effort. Data on total catch were ob- tained from annual summaries of catch issued by the Hawaii Division of Fish and Game. I obtained the sample catch (all areas) and the sample inshore catch from the catch reports, and from these data, I calculated the percentage of the catch made inshore. The total inshore catch was estimated by applying the percentage of the catch made inshore to the total catch. The estimated annual inshore catches (table 2) are shown in relation to the total catch and the estimated total oflFshore catch in figure 1. The percentage of the catch made inshore ranged from 63 percent in 1954 to 90 percent in 1960. During the poor years — 1952, 1957, 1958, and 1960 — the inshore catch averaged 83 percent of the total catch, whereas in U.S. FISH AND WILDLIFE SERVICE 183 Table 2. — Estimated total inshore and total offshore catches of skipjack tuna in Hawaiian waters, 1952-6$ Year 1952. 1953. 1954. 1955 1956 1957. 1958. 1959 1960 1961 1962 Sample Percentage Estimated catch of sample Actual inshore •\U catch total catch catch ir. 1- inshore Thousand Thousand Thousand pounds Percent pounds pou nds 6,277 76 7,292 5,542 10,543 66 12,059 7.959 11,229 63 14.021 8.833 8.257 83 9.694 8.046 10,937 73 11.132 8.126 6,075 80 6.130 4,904 6,494 86 6.834 5,877 11,945 83 12.413 10.303 7,107 90 7,360 6,624 10,780 78 10,894 8,497 9,086 82 9,415 7,720 Estimated offshore catch Thousand pounds 1,750 4,100 5,188 1.648 3.006 1.220 957 2.110 736 2.397 1.695 average and good years the inshore catch averaged 75 percent. Yamashita (1958) who examined the 1948- 53 catches of skipjack tuna suggested that about 8.0 million pounds may be nearly the Q 4 o a. S .J ^ 8 ^H lUIAL I.AI«.n INSHORE CATCH H iiiiiidni OFFSHORE CATCH 1952 I9S3 I9,S1 1955 1956 1957 1958 1959 1960 196J 1962 YEAR Figure 1. — Total catch of skipjack tuna (all areas) and the estimated inshore and offshore catches in the Hawaiian fishery, 1952-62. 184 maximum that can be obtained in the inshore area. The present study indicates, however, that the inshore catch can be well above this level. The 1959 landings, for example, were 10.3 million pounds and were caught by a fleet of 21 full-time vessels, although in 1949-53 26 to 28 vessels were fishing full time for skip- jack. The offshore catch increased gradually from 1.8 million pounds in 1952 to a peak of about 5.2 million pounds in 1954 (fig. 1). After a sharp decline to about 1.6 million pounds in the following year, the offshore take fluctuated between 0.8 and 3.0 million pounds from 1956 to 1962. FISHING INTENSITY Fishing intensity is the total amount of ef- fort expended in catching fish. Eflfort changes with time in difl'erent ways. For example, in a fishery where a trip is considered a unit of fishing eflfort, an increase in the duration of trips or an increase in the fishing power of vessels alters the unit of effort. These changes complicate the analysis of catch and eflfort data ; therefore it becomes necessary to obtain and examine information on size of vessels, on modification of or improvement to fishing gear, and on changes in fishing time. SIZE CLASSES OF VESSELS It may be expected that size of a vessel in- fluences its potential eflficiency as a fishing unit in a pole-and-line flshery because the larger crews give the larger vessels greater flshing power. One measure of effort is the number of men aboard per trip ; this number may vary among vessels and with the years. The inter- view records for 1950-56 indicated that the number of men aboard per trip varied between 6 and 14. The eflfects of this crew variability were reduced by separating the vessels arbitrarily into two size classes according to their bait- carrying capacities. The bait-carrying capacity was a good measure of the vessel's fish capac- ity, because on the return to port the empty baitwells were used to store the catch. The vessels with large bait capacities were the large ones that usually carried more men. Data SKIPJACK IN HAWAIIAN WATERS on the bait capacity of most of tiie vessels in the fleet were given by Yamashita (1958; appendix table 1 ) . The bait capacity of a vessel is stated in terms of the average eff'ective volume in gal- lons per baitwell and is derived from the length of the baitwell, its width, and its depth up to the water level. The size classes of vessels used in this study are as follows: Class 1. — Bait capacity up to 800 gallons per baitwell; registered length, 58.3 to 71.9 feet; gross tonnage, 27 to 54 tons; engine, 110 to 450 horsepower. Their number ranged from 8 to 16 in 1952-62. Class 2. — Bait capacity more than 800 gal- lons per baitwell ; registered length, 65.0 to 80.5 feet ; gross tonnage, 45 to 77 tons ; engine, 160 to 600 horsepower. Their number ranged from 11 to 14 in 1952-62. It was necessary to estimate the bait-carry- ing capacity of four vessels for which Yama- shita (1958) gave no records. To determine the most dependable procedure, characteristics such as gross tonnage, net tonnage, registered length, and engine horsepower, were examined in relation to average effective volume of bait- wells. The regression of average effective volume per baitwell (Y) on gross tonnage (X) (fig. 2) proved to have the smallest error of estimate. This relation was used therefore, to estimate the average effective volume of the four vessels. THE EFFECTIVE TRIP AS A MEASURE OF EFFORT The records used carried three types of statistics from which one might estimate ef- fort : The number of men hooking per trip, the number of men aboard per trip, and the num- ber of trips. The number of men hooking per trip and the number of men aboard per trip were not consistently entered ; therefore, I selected the number of fishing trips as the unit of eflfort. The catch reports showed all trips on which a catch was made, but gave no indication of zero-catch trips. For this study I define efl'ort as an efl:ective trip (a trip on which skipjack tuna were caught). Because zero-catch trips were not recorded, effort always is underestimated. The extent of _ 1.600 J 1.400 < o J 1.200 u 03 u s 3 § 800 ^/« •/ • • / y • /. • y f • 400 10 20 30 40 50 60 70 80 90 GROSS TONNAGE Figure 2. -Regression of average effective volume per baitwell on gross tonnage of Hawaiian skipjack tuna vessels. the underestimate may not be serious in some years, because zero-catch trips are fewer when fishing is good (Shippen, 1961). This source of error can weigh heavily, however, in a year of poor fishing, when zero-catch trips become numerous. ESTIMATES OF ZERO-CATCH TRIPS Because only the number of effective trips is known from the catch reports, estimates of catch per effective trip (Y/g) are larger than if total effort had been used, assuming that zero-catch trips occur from time to time. (The notation Y refers to the total weight of fish in the catch and g to the fishing effort or effective trip as recorded. In a later section of this paper the notation f is used to refer to fishing effort expressed in standard effective trips.) A measure of effort, however, should reflect zero- catch trips as well as those on which fish were caught. Logbook records available for a few vessels in 1957-59 provided some data on zero-catch trips (table 3). Five vessels kept logbooks in 1957, seven in 1958 and five in 1959 ; the vessels represented 20, 29, and 24 percent of the fleet. U.S. FISH AND WILDLIFE SERVICE 185 Table 3. — The lotaJ catch, number of Class 1 and Class 3 vessels sampled, their total recorded trips, and number and percentage of zero-catch trips, Hawaii, 1957-59 Total catch Class 1 Class 2 Year Ves- sels Trips Zero-catch trips Ves- sels Trips Zero-catch trips 1957 -- Thou- sand pounds 6.130 6.834 12.413 Num- ber 2 Num- ber 181 Num- ber 51 31 2 Per- cent 28 21 2 Num- ber 3 5 4 Num- ber 357 482 336 Num- ber 136 131 32 Per- cent 38 1958 2 145 I 93 1959 10 Because no additional data are available, I as- sumed that this sample represents the fleet's activities for these years. Zero-catch trips were more frequent among Class 2 than among Class 1 vessels (table 3). The higher rate of occurrence in poor years (1957-58) than in the good year (1959) indi- cates that the percentage of zero-catch trips tends to decrease as total catch increases. The apparent negative correlation between these variables is supported to some extent by Ship- pen (1961: table 2) who analyzed the logbooks of two vessels. (The original records show that these were Class 2 vessels.) He found that in a poor year (1952), zero-catch trips ac- counted for 10 percent of the total trips made by Boat A and 14 percent of those made by Boat B. In a good year (1953), zero-catch trips were 8 and 10 percent for Boat A and Boat B, respectively. The numbers of effective trips and zero-catch trips were used to estimate the total effort for 1957-59. For example, in 1957, the total num- ber of zero-catch trips among Class 1 vessels was estimated by simple proportion to be 262. The estimated total effort for both size classes for 1957-59 is given in table 4. The number of zero-catch trips was large in 1957 and 1958. In 1957 the estimated total number of unreported zero-catch trips was 822, or an average of about 33 per vessel ; in 1958 the estimated number was 502 or about 21 per vessel. In 1959, a good year in the fishery, the estimated number of zero-catch trips for the fleet was only 128 — about 6 per vessel. The results indicate that catch per effective trip should be regarded with caution. Effective effort is a biased measure of fishing pressure, but it has been used becau.se information on zero-catch trips was not available from the catch reports used. This condition has been remedied ; in July 1964, the Hawaii Divisi()n of Fish and Game issued revised catch-report forms which have spaces for recording zero- catch trips. DURATION OF AN EFFECTIVE T^IP Most of the vessels in the fleet made short runs. They left for the fishing grounds in the early morning and returned to port in the evening. On occasion, however, trips of 2, 3, or 4 days have been recorded. Only a small proportion of the day was devoted to actual fishing ; the greater part was spent scouting for bird flocks that follow schools of skipjack tuna. To judge the possible effects of longer trips, the frequency of occurrence of 1- and multiple- day trips (2 or more days per trip) was de- termined from 1960 interview records for 16 vessels (table 5). Records for a total of 329 trips showed 315 of 1 day (95.7 percent), 13 of 2 days (4.0 percent), and 1 of 3 days (0.3 percent). Of the thirteen 2-day trips, 9 had catches during both days at sea, but each of the Table 4. — Xumlier of effective trips and estimated number of zero-catch trips of Class 1 and Class 2 vessels, Hawaii, 1957-59 Table 5. — Xumber and percentage of /-, 2-. and S-dnij trips, total catch, catch per effective trip, and range of catches of 16 Hawaiian skipjack tuna vessels in 1960 Vessels Class 1 Class 2 Days per trip Effective trips Catch Effec- tive trips Esti- mated zero- catch trips Esti- mated total trips Effec- tive trips Esti- mated zero- catch trips Esti- mated total trips Year Total Per effective trip Range Arttni6er 1 Number 315 13 1 329 Percent 95.7 Pounds 1.636.185 121.462 35.000 1.792.647 Pounds 5.194 9.343 35.000 Pounds 37.1-39,. '..'B 2.00O-I7.000 Number 25 24 21 Number 668 659 779 Numt}er 262 179 17 Number Number Number 560 323 111 Number 1.470 1.188 1 166 1957 930 910 838 ' 885 796 I .n.Vl 2 3 Totals 1958. 1959... 186 SKIPJACK IN HAWAIIAN WATERS remaining 4 had only 1 day in which skipjack tuna were caught. We may conclude that a trip usually represents 1 day's fishing. AVERAGE NUMBER OF MEN HOOKING PER EFFECTIVE TRIP The number of hooks fishing on a Hawaiian skipjack tuna vessel depends on the number of fishermen that take fishing positions along the stern during the fishing operation, since each man fishes a single pole to which a line and feathered jig is attached. Yuen (1959), in a study of the response of skipjack tuna to live bait, pointed out that the number of men hooking was one of the factors that affect the catch per school. The catch per effective trip is also related to the number of men hooking ; it is important, therefore, to examine the year- to-year variation in this number. Data on the number of men hooking were available only from records collected between 1950 and 1960 ; those for 1950-56 and 1960 were from inter- Tabi.e 6. — The monthly and annual average of the number of men hooking per effective trip on Class 1 and Class 2 Hawaiian skipjack tuna vessels, 1950-60 Month 1950 1951 19.12 19,'vt 19,M 19.55 19.56 19.57 19.58 1959 1960 and class Jan. 1 . 6 n 6 n 7.1 7.0 2 11 8 3 7 ? 7 ? 7.9 Feb. 1 8 6 n 5 7 6.7 8.2 ? 8.3 6.3 6 8 6.8 fi 9 Mar. 1 8 8 9 9 9 3 7.1 ?, 9 fi 10 10 3 10 fi 6 7 3 7.4 Apr. 1-. 9.5 9.3 9.2 7.3 6.5 7,1 7,4 6,7 2-. 9.3 9.7 9.6 10. 7,2 7,7 7,1 7.0 May 1.- S.4 8.2 9.3 9.3 9.4 9.2 7.2 6.7 fi 6,8 6,6 2.. 10.6 10. 9.0 9.4 10.5 9,7 8,1 8 ?. 8 7.0 June I-. 8.8 8.0 8.0 8.8 9.2 9.3 7.1 6,7 7,1 8,6 7,4 2-. I-, 10.2 8.8 10.7 7.9 9.7 8.2 9.4 8,0 9.7 8.5 8.2 7.0 8,7 8,9 7.8 8,9 7.3 July 8.8 7.7 6,8 2- 9.V 8.5 9.7 9.8 9.5 9.2 8.2 8,7 73 7.7 Aug. I-. 8.6 8.7 8.5 8.6 8.4 8.2 6.3 7,9 8,4 6,5 2- 1U.8 9.7 9.1 10.3 10.1 8.3 7,4 8 5 7,2 7.6 Sept. 1.- 7.4 8.8 9.1 6.5 9.5 7.5 5,5 5,7 7,5 6.5 2.. 10. V 9.1 9.1 10.2 9.4 7.4 7,8 7,8 6.9 6.8 Oct. 1-- 6.2 7.8 7.9 8.5 8.0 7,2 6 1 6,9 7 3 2.. 8.7 8.6 9.4 7,0 6 7 8,0 6,6 6.7 Nov. 1-- 8.7 9.9 10. 8.0 7.3^ 6,4 6.2 5.0 2.. 7.6 7.0 9.9 11.9 6.0 6,5 8.3 6.9 5.0 Dec. 1 8.0 6 8 6.6 7 9 7.0 fi 7 6 5 2.. 9,3 11 6 1 6.4 Annual average 1 8.4 8.2 8.4 8.6 9.2 8.9 7.1 6.7 7.1 7,6 6,9 2.- 10.4 9.6 9.2 9.7 9.8 8.3 8.7 7.4 8.1 7.2 7.4 view records and those for 1957-59 were from logbooks. These data were used to calculate monthly and annual averages by size classes of vessels (table 6). I expected that the number of men hooking per effective trip would be greater during the season months (May to September) than dur- ing the off-season months (Shippen, 1961). Despite the incompleteness of the data for some years, the trend of change, discernible from the data for those years where informa- tion was adequate, indicates no pronounced in- crease in the number of men hooking in May to September (fig. 3). More men fished per ef- ::: ^ . 1950 ( 1 \ ■^ 1 H. n JAN: .\UR may JUl-T SEPT. MONTH .-- ^ .... ». .IMC '^ ■^ -Y 4 — A ..-' , ■■•• > — ^ -r •t -^ rfta rrr X..- ' \ ■- •- laea <. /' .i; -sr ;:»- - w •^ ■<■■. .... -; V JAN ,M-\R ,M,\y JULY SFPT- NOV .MO.N'TII Figure 3. — Monthly averages of number of men hook- ing per effective trip on Class 1 and Class 2 Hawaiian skipjack tuna vessels, 1950-60. Class 1 vessels, broken line; Class 2 vessels, solid line. fective trip in Class 2 than in Class 1 vessels, although the 1959-60 data indicate that the differences between the two classes were small. Figure 4 illustrates the decline in the annual average. The average number of men hooking per effective trip on Class 1 vessels was fairly steady from 1950 to 1955, then dropped and remained at a lower level in 1956-60. The average for Class 2 vessels declined almost steadily from 1950 to 1960. This decrease in the number of men hooking from 1950 to 1960 was not, however, accompanied by a decline in the catch per effective trip. An explanation is given in the section on Apparent Abundance. U.S. FISH AND WILDLIFE SERVICE 187 k 1 ^<___ ' ,.-■•■"" ■••■\- V' ^r cuss ; 2 ° 1 7 cuss r \ \ \ ...>^ ■•■., o i " s b S A u after 1953 (fig. 5). The percentage of effective inshore trips did not differ greatly between Class 1 and Class 2 vessels. For Class 1 vessels, 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 YEAR Figure 4. — Annual averagres of number of men hook- ing per effective trip on Class 1 and Class 2 Hawaiian skipjack tuna vessels, 1950-60. DISTRIBUTION OF EFFORT BY AREA About 80 percent of the effective trips during any given year were in the inshore area (table 7). The percentage of effective inshore trips for both classes of vessels declined from 1952 to 1953. Class 1 vessels showed a further decline in 1954, then a gradual increase, where- as Class 2 vessels showed a gradual increase Table 7. — The number and (in parentheses) percentage of effective trips by Class 1 and (lass 2 Hawaiian skipjack tuna vessels in inshore and offshore areas, 1952-62 Effective trips Year Class 1 vessels Class 2 vessels Inshore Offshore Inshore Offshore 1952 1953 -. -- Num- ber 658 799 696 817 709 MO • .187 658 563 608 646 662 Per- cent (77) (72) (70) (88) (76) (81) (89) (84) (92) (87) (89) (81) Num- ber 197 30S 291 llfi 222 128 72 121 51 88 80 1.52 Per- cent (23) (28) (30) Num- ber 555 738 690 Per- cent (80) (76) (77) (86) (81) (82) (90) (84) (92) (84) (86) (83) Num- ber 137 232 209 124 184 168 82 171 73 179 131 154 Per- cent (20) (24) 1954 ..; (23) 1955 (12) I 748 (24) 1 772 (19) ; 742 (ID 783 (14) 1956 (19) 1957 (18) 1968 (10) 19S9 (16) (8) (13) (11) (19) 884 858 939 804 774 (16) 1960 (8) 1961 (16) 1962 (14) Average '- (17) ' Percentages were computed from the average annual numbers of effective trips. 1952 1953 1954 1955 1956 1957 19SB 1959 I960 1961 1962 VKAH Figure 5. — Percentage of effective trips inshore by Class 1 and Class 2 Hawaiian skipjack tuna vessels, 1952-62. the percentage of inshore trips, on the average, was 81 percent, whereas for Class 2 vessels, the average was 83 percent over the 11 years. A summary of effective trips by areas for each size class of vessels with respect to poor years and average and good years showed that in poor years, the percentage of effective in- shore trips by Class 1 vessels was 84 percent ; in average and good years, it was 81 percent. For Class 2 vessels, the values were 86 percent in poor years and 82 percent in average and good years. One may wonder if skipjack tuna are more abundant inshore, since a larger percentage of the total trips is made within 20 miles from land. Observations of skipjack .schools in Ha- waiian waters in 1953 indicated that sightings of tuna schools were equally numerous offshore and inshore except for .sectors to the northeast and southwest of Oahu (Royce and Otsu, 1955). If schools are equally abundant offshore and inshore, the question arises as to why effort 188 SKIPJACK IN HAWAIIAN WATERS has been concentrated in the inshore grounds. There are several possible answers. Fishermen reduce costs by remaining close to port as long as they can make profitable catches. The con- centration of effort inshore also may be dictated by the quality and quantity of live bait. Even though it may occasionally survive as long as a week, the delicate nehu may die w^ithin a few hours. The fishermen logically would fish inshore to use the bait quickly before mortality becomes heavy. Furthermore, the need to re- plenish live-bait supplies to some extent re- stricts trips to the distant offshore grounds, where live bait is unavailable. EFFECTIVE TRIPS BY SIZE CLASSES OF VESSELS The average number of effective trips per vessel per year fluctuated widely in 1952-62 (table 8). The average number of effective trips per Class 1 vessel per year (fig. 6) Table S. — The total nvmber nf effective trips and the average niimlier of effective trips per vessel by Class 1 and Class 2 Hawaiian skipjack tuna vessels, V>r>'2-62 CIA , , 1 , , « 1 uzttztt ! 1 ' \ V ~^/ \J .'^ <^, / -■•■ 1 S^ '\ i n!/ 1/ :iAss r 2 VESSELS \ ...-• ,.•* 1952 1953 1954 1955 1956 1957 1958 1959 19G0 19G1 1962 Figure 6. — Average number of effective trips per vessel per year by Class 1 and Class 2 Hawaiian skipjack tuna vessels, 1952-62. fluctuated between 51 and 72 in 1952-58, rose sharply to 97 in 1959, and ranged between 77 and 91 in 1960-62. For Class 2 vessels, the average rose very sharply from 58 to 88 in 1952-53, then declined to 75 in 1954 and to 62 in 1955. After 1955 the average appeared to increase gradually. The reason for the increase in the number of effective trips, particularly lass 1 vessels Class 2 vessels Year Vessels Total Average Vessels Total Average fished per vessel fished per vessel Number Number Number Number Number Number 1952... 16 855 53 12 692 68 1953 16 1.107 69 11 970 88 1954 15 987 66 12 899 75 1955 14 933 67 14 872 62 1956--- 13 931 72 13 956 74 1957 13 668 51 12 910 76 1958 11 659 60 13 865 66 1959 ^-- 8 779 97 13 1,055 81 1960 8 614 77 13 931 72 1961 8 696 87 13 1.118 86 1962 8 726 91 12 935 78 the sharp increase since 1959 among Class 1 vessels, is not known. APPARENT ABUNDANCE The catch per unit of effort does not provide estimates of true abundance but of apparent abundance, since it is affected by availability^ and vulnerability- to the fishing gear. In the section that follows, I discuss the catch per effective trip, the factors affecting it, and the method used to obtain a standard unit of effort. CATCH PER EFFECTIVE TRIP BY SIZE CLASSES OF VESSELS AND AREAS Data on Y/g (catch per effective trip) by size classes of vessels and areas are given in table 9 and plotted in figure 7. The inshore Y/g for Class 1 vessels fluctuated within a relatively narrow range, whereas that for off- shore fishing fluctuated more widely. The curves for Class 1 vessels offshore and Class 2 vessels inshore were similar. Catches of Class 1 vessels that fished offshore fluctuated widely and followed the curve for the total catch. The inshore and offshore Y/g for Class 1 vessels and total catch were significantly correlated (r = 0.675; df ^ 9; p = 0.03 and r = 0.923; df — 9; 7J<0.001, respectively). A similar com- parison of data for Class 2 vessels showed that both the inshore and offshore Y/g were sig- 1 "Availabilil\' is the iiorlion (a percentage) of the recruited population that is physirally within the geographic range of the fishery durintr the fishing season." (Ahlstroni. 1960: p. 1361.) - "Vtilnerabijity is tlie accessibility of the fish within the geographic range of tlie fishery to the efforts of a fishery." (Ahlstroni, 1960: p. 1361.) U.S. FISH AND WILDLIFE SERVICE 189 1953 1953 1954 1955 I95G 1957 1958 1959 1960 1961 1962 YKAR Figure 7. — Total catch and catch per effective trip by areas for Class 1 and Class 2 Hawaiian skipjack tuna vessels, 1952-62. nificantly correlated with the total catch (r = 0.762; df = 9; p<0.01, and r = 0.697; df = 9; p = 0.02, respectively) . FACTORS AFFECTING ESTIMATES OF THE CATCH PER EFFECTIVE TRIP Proper interpretation of statistics on catch and effort requires information about factors that contribute to variations in catch per unit of effort. For the skipjack tuna fishery in Hawaii, a number of factors have been isolated as causes of variation in catch per effective trip. Some of these are discussed here. Changes in the Availability of Skipjack Tuna In a study on the oceanography and skipjack fishery in the Hawaiian region, Seckel and Waldron (1960) pointed out that the time of initial warming of the surface water at Koko Head, Oahu, appears to be related to the annual skipjack landings. When the initial warming occurred in February, this implied that the California Current System was well developed and average or better-than-average fishing years occurred. When the initial warming oc- curred in March, fishing was poor. The relation between the time of initial warming and skipjack landings later in the season appeared to have some predictive value concerning the availability of skipjack tuna. According to Seckel (1963: fig. 4), the initial warming occurred in March in 1952, 1955, 1957, 1958, and 1960 ; skipjack availability should be low during these years. As was pointed out earlier, however, the catch in 1955 was close to the average catch of 9.7 million pounds for 1952-62 (table 1) and was considered an av- erage year for the purpose of this report. The forecasts made from 1959 to 1962 were de- pendable, but the prediction of favorable fish- ing conditions and catch levels could not be made with assurance because only partial understanding of the relation has been achieved. At present, variations in skipjack availability appear to be one of the most im- portant factors causing fluctuations in the total catch of skipjack in the Hawaiian fishery. Changes in the Number of Men Hooking and in Fishing Technique Since the average number of men hooking per effective trip declined between 1950 and 1960, I examined the data to .see if the Y/g showed a similar decline. The results (table 9 and fig. 7) showed that Y/g did not decline during 1952-62. For example, if the poor years of 1952, 1957, 1958, and 1960 are omitted to simplify the comparison, Y/g for the remaining years appears to be approximately the same before and after 1955, that is, no indication exists of a decline in Y/g. An excellent 2-year comparison is provided by the data of both size classes in 1954 and 1959, in which years T.Mii.F, 9. — The catch per ejfeclive trip of Class 1 and Class 2 Hawaiian skipjack tuna vessels in inshore and offshore areas, 1962-62 Catch p?r elTective trip Year Class! Class 2 Inshore onshore Inshore Oflshore 1952 Pounds 3,586 4,0,55 3,867 3,762 4,914 3,219 3,238 5,133 3,573 4,775 4,019 Pounds 3,728 6.337 7.758 5.704 6.338 3.790 4.928 0. .551 Pounds 4.323 5.069 6.312 5.048 5.871 4.210 4.723 7 4in Pounds 5. 722 1953 6 943 1954 9.210 19.55 ft 0'*3 1956 8,228 1957 4 333 1958- 6,,')93 1959 7 129 1960 5.283 5,093 0,490 5,891 6.071 6.000 6,244 1961.. 1962 H.SMO 8 602 1 190 SKIPJACK IN HAWAIIAN WATERS oceanographic conditions were similar. In both years the water in the Hawaiian Islands area warmed early (Seckel and Waldron, 1960) ; the total catch was above average ; and Y/g was also large (table 9) even though the av- erage number of men hooking per effective trip (both classes of vessels) was 9.6 in 1954 and 7.3 in 1959. Richard S. Shomura (personal communica- tion) has suggested that failure of this decline in Y/g to appear may have been the result of a change in the fishing techniques necessitated by a decrease in the number of skilled fisher- men. In the past, when a fisherman caught a fish, he grasped it under his arm so that he could remove the barbless hook ; he then dropped the fish on the deck. (Among local fishermen, the method is called "catch" or its Japanese equivalent "daku" which means to hold in one's arm.) A considerable amount of practice and experience is required before one develops the skill necessary to fish by this method. Another method used only occasionally in the past was "flipping," in which the fisher- man swings the fish aboard and suddenly re- laxes the tension of the pole to permit the hook to fall clear of the fish's mouth before it drops on the deck. (This method is sometimes called "mochikomu" which in Japanese means to bring in.) Interviews with fishermen indi- cated that flipping allows them to catch fish faster and does not require the same degree of skill as "catching fish under the arm," but the fishermen tire more rapidly. The shift in emphasis from catching under the arm to flipping permitted a short-handed crew of limited experience to equal or better the catch of a larger and more experienced group of fishermen. Because flipping bruises the fish, some fishermen still catch under the arm when fishing large skipjack, which bring a premium price on the fresh-fish market. Damaged fish bring lower prices. Changes in the Efficiency of Class 1 Vessels Fishing efl^ciency per vessel also increases when vessels that do poorly in the fishery are forced to stop fishing. In 1952-62, the number of Class 1 vessels actively fishing decreased from 16 to 8 (2 were wrecked and 6 stopped fishing). When Class 1 vessels were ranked according to the total catch of each vessel, the results showed that the eight vessels fishing in 1962 were usually among those that were ranked high in previous years. Because those that did pooi-ly stopped fishing, the fishing efficiency (as measured by the average catch per efl'ective trip) of all the remaining vessels increased. This increase in fishing efficiency also may have off'.set the effect of the decline in the number of men hooking per trip. The number of Class 2 vessels reached a maximum of 14 in 1955 and declined to 12 in 1962 (1 was wrecked and 1 stopped fishing). Since only one vessel in this size class has stopped fishing, the efficiency of the class could not be expected to change markedly. Amount of Bait Used per Effective Trip Bait supply as well as the number of men hooking per effective trip may affect catch. It is important, then, to determine whether the larger vessels do carry and use more bait than the smaller ones. Data on bait catch (table 10) permitted investigation of this problem. Table 10. — Total buckets of bait used and amount used per effective trip by Class 1 and Class 2 Hawaiian skipjack tuna vessels, 1952-62 Class 1 Class 1 Total 1952 1953 1954 1956 1956 1967 1968 1969 1960 1961 1962 Average 1 otai buckets Effective of bait used Number 12,202 12,932 12,592 13,144 12,092 8,187 5,721 10,233 5.748 9.462 9,315 10.148 trips Buckets I Total used per effec- tive trip loiai buckets Effective of bait used trips Buckets used per effec- tive trip Nitmber Number Number Number Number 855 14.3 11.319 692 16.4 1.107 ' 11.7 14.163 I 970 14.6 987 12.8 15.503 | 899 17.2 933 ] 14.1 16.092 I 872 18.4 931 I 13.0 14.688 1 956 ' 15.2 668 12.2 13.497 910 I 14.8 659 8.7 10.966 , 865 i 12.7 779 13.1 17.961 1,055 I 17.0 614 9.4 10.593 931 I 11.4 696 13.6 18.051 1.118 16.1 726 12.8 14.733 935 15.8 814 12.3 14.315 928 15.4 Class 2 vessels used more bait per effective trip in all years for which there were records. The 11-year unweighted averages indicated that Class 1 vessels used 12.3 buckets per effective trip compared with 15.4 buckets per effective trip by Class 2 vessels. The difference in the average number of buckets of bait used per effective trip between class 1 and Class 2 vessels was statistically significant U.S. FISH AND WILDLIFE SERVICE 191 (t = -11.31; df = 10; p <0.001). We can conclude, therefore, that the amount of bait used in fishing contributed to the larger catch per effective trip among Class 2 vessels. STANDARDIZATION OF CATCH PER EFFECTIVE TRIP Differences between the large and the small vessels in numbers of men and in quantity of bait and hence in catching ability can compli- cate the estimation of apparent abundance. Rather than analyze data for the two classes of vessels separately, I have employed only one index, based on a "standard" unit of fishing effort. This unit is derived from a set of con- version factors which translate unequal fishing practices and capacities into a standard unit. For example, under conditions of equal abundance, when a small vessel makes a smaller catch than a large vessel, standardi- zation of the effort units takes into account the differences in their fishing power. A gen- eral discussion of the problems in standardiz- ing fishing effort may be found in Gulland (1955, 1956), in Shimada and Schaefer (1956), and in Schaefer (1963). Efficiency Factors The yearly Y/g of the two classes of vessels by areas permits the calculation of efficiency factors (Shimada and Schaefer, 1956). For each area the ratio of the yearly Y/g of Class 1 to that of Class 2 was computed. For example, from table 9, values of Y/g for 1952 were as follows: Class 1: Inshore —3,586 pounds /effective trip Offshore — 3,728 pounds/effective trip Class 2: Inshore —4,323 pounds/effective trip Offshore — 5,722 pounds/effective trip For Class 1 vessels, the efficiency factor for inshore was 3,586/4,323 = 0.83; for off"shore, it was 3,728/5,722 = 0.65. The efliciency fac- tors for Class 2 vessels are fixed at 1.00 for all years. The mean efficiency factor for the year is the geometric mean of the inshore and off-- shore values. The geometric mean is appropri- ate for averaging ratios. The mean efficiency factors for Class 1 ves- sels and the average for the 11-year period not only demonstrate the greater capability of Class 2 vessels, but also the variability of the 192 factor (table 11). For example, if the Y/g of Cla.ss 1 vessels were some constant proportion of that for Class 2 vessels, one would expect an almost constant eflJiciency factor. The efficiency factors of Class 1 vessels, however, were as Tabi.k 11.— Tn/we.s of efficiency factors for Class 1 Hawaiian skipjack tuna vessels in terms of a fixed value of 1.00 for Class 2 vessels [These factors were iLsed to standardize the unit of effort in 1952-62] Year Class 1 Year Class 1 1952... 0.74 .86 .72 .84 .80 .82 1958 0.72 .80 .77 .73 .68 .77 1953 1959 1954 1960 1955 1961 1956 1962 1957 high as 0.86 and as low as 0.68. These values show no trend, and apparently are not related to good and poor years. The efficiency factors by area (computed in terms of a fixed value of 1.00 for Class 2 ves- sels, off-shore) for each vessel class (table 12) show that the values for both Class 1 and Class 2 were almost consistently smaller for T.^Bi.E 12.— Va/i/ra of efficiency factors for Class 1 Hawaiian skipjack tuna vessels inshore ami offshore and for Class 2 vessels inshore in terms of a fixed value of I.OO'for Class 3 vessels offshore Year Class 1 Class 2 Inshore Offshore Inshore 1952 0.63 .58 .42 .62 .60 .74 .49 .72 .57 .48 .47 0.65 .91 .84 .95 .77 .87 .75 0.76 .73 .68 .84 .71 .97 .72 1953 1954 1955 1956 19.57... 1958 1959 1960 .85 .82 .65 ; .59 .70 1 .70 1961 1962 Average .67 .80 t .78 inshore than for offshore fishing. The average for the 11-year period indicates that the off"- shore values of efficiency factors are higher than their respective inshore values. Further- more, the mean efficiency factor for Class 1 offshore is slightly larger than that for Class 2 inshore. This result was not unexpected be- cause efficiency factors do not take into ac- count the ability of a vessel to visit distant areas where fish density may be higher. Al- SKIP.IACK IN HAWAIIAN WATERS though it may appear from the values of effi- ciency factors in table 11 that Class 2 vessels always have better results than Class 1 vessels, the data indicate that, on the average, the offshore catches of Class 1 vessels are likely to be larger than those of Class 2 vessels fish- ing in inshore waters. Catch per Standard Effective Trip The efficiency factors, given in table 11, were used in calculating the standard unit of effort. For example, in 1952 there were 855 effective trips by Class 1 vessels and 692 by Class 2 vessels. The standard effective trip is the sum of the products of the mean effi- ciency factor and total number of effective trips of the size classes: 0.74(855) + 1.00(692) = 1,325 standard effective trips. The catch per standard effective trip (Y/f ; the notation f refers to fishing ef- fort expressed in standard effective trips) is found by dividing the sample catch by the standard effective trips: 6,277,046 = 4,737 pounds per standard ef- 1,325 fective trip ; and the fishing intensity is ob- tained from the total catch and Y/f: 7,291,851 1,539 standard effective trips. 4,737 The Y/f reflects only apparent abundance based on the trips on which fish were caught. The total catch of skipjack tuna in pounds, Y/f, and relative effective fishing intensity per Class 2 trip are presented in table 13 and the index curves are illustrated in figure 8. Table 13. — Tolal landings of skipjack tuna in Hawaii, catch per standard effective trip, and relative effective fishing in- tensity, 1952-62 Year Total catch Catch per stand.ird effective trip (Class 2 trip) Relative effective fishing intensity (in Class 2 trips) 1952 nousand pou lids 7.292 12.059 14,021 9,694 11,132 6.130 6.834 12,413 7,360 10,894 9,415 Pounds 4,737 5.486 6,983 4,986 6,430 4.166 4.850 7.119 5.062 6.629 6,358 Trips 1.539 1953 2.198 1954 -.- 2.008 1955 1,944 1956 - . 1,731 1957 1,471 1958 1.409 1959 . 1,744 1960 1,4.14 1961 -- 1,643 1962 1.481 19S2 1953 1954 1955 195G 1957 1958 1959 1960 1961 1962 YEAR Figure S. — Total catch, catch per standard effective trip, and the relative effective fishing intensity for skipjack tuna in Hawaii, 1952-62. The curve for Y/f was at 4,737 pounds in 1952, rose in 1953 and again in 1954 to reach a peak of 6,983 pounds, then dropped to about the 1952 level in 1955. Another peak in 1956 was followed by a decline to its lowest level in 1957. The Y/f reached another peak in 1959, surpassing that of 1954. The Y/f had no trend; rather, the values varied around an average of about 5,700 pounds per standard or Class 2 trip. The relative effective fishing intensity exhibited a slight decreasing trend from about 1953 to 1958 and leveled oflf at about 1,600 standard trips in 1959-62. INTERRELATION OF TOTAL CATCH, FISHING INTENSITY, AND APPARENT ABUNDANCE The catch per standard effective trip (Y/f) and the total catch fluctuated in a similar fashion in 1952-62 (r = 0.851; df = 9; p <0.001). For the 11-year period, then, the total catch may be used as an index of apparent abundance, but it should by no means be con- sidered an appropriate index in other years. The situation may change in other years, be- cause of the sensitivity of the total landings to various influences such as weather, sea con- ditions, the amount of effort expended, and the market for skipjack tuna. The Y/f and the relative effective fishing U.S. FISH AND WILDLIFE SERVICE 193 intensity were not correlated significantly over the 11-year period (r -0.343; df = 9; p = 0.32). The lack of correlation suggested that changes in the size of the Y/f were not influenced by changes in the amount of fishing, but by other fishery-independent factors, such as variation in availability and vulnerability; the strength of year classes also may be im- portant (Rothschild, 1965). The eflfective fish- ing intensity tended to decline over the years under study, largely because of a decrease in the number of vessels in the fleet. ACKNOWLEDGMENTS Two Hawaii State Government agencies as- sisted in the preparation of this paper: The Division of Fish and Game of the Department of Land and Natural Resources made available all its catch statistics and interview records and the Division of Archives of the Depart- ment of Accounting and General Services per- mitted the use of its facilities. The Computing Center of the University of Hawaii provided technical assistance. Robert R. Parker reviewed the manuscript. LITERATURE CITED Ahlstrom, E. H. 1960. Fluctuations and fishing. In H. Rosa, Jr., and Garth Murphy (editors), Proc. World Sci. Meet. Biol. Sardines Related Species 3: 1353~ 1371. Food Agr. Organ. United Nat., Rome, Ital. GULLAND, J. A. 1955. Estimation of growth and mortality in com- mercial fish populations. Fish. Invest., London, Min. Agr. Fish. Food, Ser. II, 18(9), 46 pp. 1956. On the fishing eff'ort in English demersal fisheries. Fish. Invest, London, Min. Agr. Pish. Food, Ser. II, 20 (5), 41 pp. June, Fred C. 1951. Preliminary fisheries survey of the Ha- waiian-Line Islands area. Part III. The live-bait skipjack fishery of the Hawaiian Islands. Com. Fish. Rev. 13(2) : 1-18. Rothschild, Brian J. 1965. Hypotheses on the origin of exploited skip- jack tuna {Katsinvonus pclamis) in the eastern and central Pacific Ocean. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 512, iii -f 20 pp. RoYCE, William R., and Tamio Otsu. 1955. Observations of skipjack schools in Ha- waiian waters, 1953. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 147, v + 31 pp. Schaefer, Milner B. 1963. Statistics of catch and effort required for scientific research on the tuna fisheries. In H. Rosa, Jr. (editor), Proc. World Sci. Meet. Biol. Tunas Related Species 6(3): 1077-1086. Food Agr. Organ. United Nat., Rome, Ital. Seckel, Gunter R. 1963. Climatic parameters and the Hawaiian skipjack fishery. In H. Rosa, Jr. (editor), Proc. World Sci. Meet. Biol. Tunas Related Species 6(2): 1201-1208. Food Agr. Organ. United Nat., Rome, Ital. Seckel, Gunter R., and Kenneth D. Waldron. 1960. Oceanography and the Hawaiian skipjack fishery. Pac. Fisherman 58(3): 11-13. Shimada, Bell M., and Milner B. Schaeffer. 1956. A study of changes in fishing eff'ort, abundance, and yield for yellowfin and skipjack tuna in the eastern tropical Pacific Ocean. Inter- Amer. Trop. Tuna Comm., Bull. 1(7) : 351-469. Shippen, Herbert H. 1961. Distribution and abundance of skipjack in the Hawaiian fishery, 1952-53. U.S. Fish Wildl. Serv., Fi.sh., Bull. 61: 281-300. Yamashita, Daniel T. 1958. Analysis of catch statistics of the Hawaiian skipjack fishery. U.S. Fish Wildl. Serv., Fish. Bull. 58: 253-278. Yuen, Heeny S. H. 1959. Variability of skipjack response to live bait. U.S. Fish Wildl. Serv., Fish. Bull. 60: 147-160. 194 SKIPJACK IN HAWAIIAN WATERS CHARACTERISTICS OF THE BLOOD OF ADULT PINK SALMON AT THREE STAGES OF MATURITY By Kenneth E. Hutton, Department of Biological Sciences San Jose State College, San Jose, California 95114 ABSTRACT Selected characteristics of the blood of adult pink salmon (Oncorhynchus gorbuscha) were studied in fish at three stages of maturity — migrating fish approaching the general area of spawning streams but still in the open ocean, fish in the immediate vicinity of the spawning stream but in the estuary, and fish in the spawning stream. Although some hematological characteristics changed little, blood proteins, glucose, and cholesterol decreased progressively, and lipid phosphorus increased. The blood chemistry of salmon of the genus Oncorhynchus is especially interesting because of physiological changes that occur during the spawning migration from sea water to estua- rine waters of reduced salinity and then into fresh water. This change in the environment is concurrent with the final stages of matura- tion. Some information is already available on changes in blood characteristics at this time of the life cycle. Lysaya (1951) found several physiological changes in the blood with ad- vancing sexual maturity in the Asiatic pink salmon (0. gorbuscha) and chum salmon (0. kefa). The erythrocyte count, the hemoglobin concentration, and the blood glucose, chloride, and calcium levels fell; and the erythrocyte sedimentation rate and the blood urea and nonprotein nitrogen concentrations increased. Biologists of the Fisheries Research Board of Canada found that adult sockeye salmon (0. nerka) on their spawning migration up the Eraser River lost 11 to 30 percent of their body weight and had decreasing blood choles- terol (Idler and Tsuyuki, 1958) ; liver gly- cogen decreased, except for a terminal in- crease (Chang and Idler, 1960) ; and concen- trations of adrenal corticosteroid hormones increased (Idler, Ronald, and Schmidt, 1959). Chinook salmon (0. tshawytscka) during their spawning migration up the Sacramento River and its tributaries in California showed : FISHERY BULLETIN: VOLUME 66, NO. 2 increased activity of the pituitary with term- inal degeneration ; hypertrophy of the islets of Langerhans; hyperplasia of the adrenal cor- tices (a rise in concentration of 17-hydroxy- corticosteroids ended with degeneration of the adrenal glands) ; and the deterioration of the stomach, liver, spleen, thymus, kidneys, thy- roid, and cardiovascular system (Robertson and Wexler, 1960, 1962; Robertson, Krupp, Favour, Hane, and Thomas, 1961; Robertson, Wexler, and Miller, 1961; and Robertson, I^rupp, Thomas, Favour, Hane, and Wexler, 1961). In 1963, under the sponsorship of the Bu- reau of Commercial Fisheries, I had the op- portunity to study the hematology and blood chemistry of adult pink salmon in three stages of maturity in Alaska: (1) maturing fish in salt water migrating toward the spawning areas; (2) nearly mature fish milling in the estuary of a small creek; and (3) mature fish spawning in a fresh-water stream. This paper reports the results of these studies. COLLECTION OF SAMPLES Pink salmon in the three stages of maturity were taken from three stocks on different dates. Those migrating toward the spawning grounds (termed "migrating"), were taken from the open ocean near the community of Elfin Cove, southeastern Alaska. They were captured on August 5, about a month before 195 Published April 1967 they would have spawned; only males were sampled. Salmon milling at the mouth of a creek (called "prespawning-") were taken August 9, about 2 weeks before the start of movement into fresh water, from a bay at the mouth of a stream at Little Port Walter on the southern end of Baranof Island, south- eastern Alaska; equal numbers of males and females were sampled. Salmon spawning in the stream (termed "spawning") were taken from Olsen Creek, which empties into Olsen Bay on Port Gravina, Prince William Sound. They were taken on July 19 (males only) and September 2 (males and females). The Olsen Creek fish, which made up more than half of all the pink salmon sampled, were sampled on two dates because they arrive in two distinct runs. The early run typically la.sts from mid- July to mid-August and the late one from late August to middle or late September. These populations may be genetically distinct. One sample of blood was taken from each specimen while the fish was held on its back, in a wooden trough. A no. I8-1/2 needle on a syringe was inserted into the dorsal aorta above the roof of the pharynx, in the region of the second gill arch, and 12 ml. of blood were withdrawn. About 0.2 g. (a pinch) of potassium oxalate, an anticoagulant chosen be- cause it is dry and hence does not cause dilu- tion, was placed in the syringe before the sample was taken. The blood was transferred to a capped vial that also contained a pinch of the oxalate and was placed in an iced, insu- lated chest and transported by plane to the Bureau of Commercial Fisheries Biological Laboratory at Auke Bay. About 24 hours elapsed between collection and analysis of blood. HEMATOLOGY Certain hematological characteristics were determined. Specific gravity was measured by standard methods ; packed cell volume was estimated after the samples were centrifuged in Wintrobe tubes at 2,700 r.p.m. for 15 min- utes ; erythrocytes were counted by standard techniques (Wintrobe, 1933) with 0.85 per- cent saline as a diluent ; and hemoglobin was determined with Hycel' cyanomethemoglobin reagents. Mean corpuscular volume, mean corpuscular hemoglobin, and mean corpuscular hemoglobin concentration were calculated by the formulas of Wintrobe (1933). A current general reference for work of this type is that of Hesser (1960). The results of the analyses (table 1) are discu.ssed in comparison with the results of other workers. Because most of the charac- teristics did not vary among the three stages "^ Trade name referred to in Iliis pnbliration Hoe.s not imply endorsement of t-omtnercial product. Table 1. — Certain hemalohqical characteristics of adult pink salmon in three stages of maturity from three areas of Alaska in 106.1 (Numbers in parentheses are numl>ers of samples analyzed] Stages of maturity Mean specific gravity Mean packed cell volume Mean erythrocytes Mean hemoglobin Mean corpuscular volume ' Mean corpuscular hemoglobin ' Mean corpuscular hemoi^Iobin concentration = Migrating (Elfin Cove) Males.- ... - Prespawning (Little Port Walter) Males and females.. ... 1.061(16) 1.057(15) 1.059(31) 1.058(8) 1.059(62) .004 1.050-1.065 Percent 42(1.5) 38(15) 38(29) 32(8) 39(59) 4.5 32-52 Number/ mm.'XIO' 0.98(15) .97(15) 1.01(29) .94(8) .97(59) .17 0.. 53-1. 35 Orams/lOO ml. 11.3(1.5) 10.7(1.5) 11.1(29) 10.5(8) 11.0(59) .9 8.7-13 439(15) 396(15) 403(29) 354(8) 410(59) 69 270-641 Micro micrograms 118(15) 114(15) 117(29) 114(8) 116(59) 22 84-196 Percent 27(15) 28(15) Spawning (Olsen Creek) Males 30(29) 32(8) Coml>ined * 28(59) Standard deviation 3 Range 21-38 1 virv- ^'Q*""^*^ of red hlood cells in 1.000 ml. hlood ~ Red V)lood cell count in mill ion /mm. ' Hemogl obin in grams per 1.000 m l. blood ~ Red blood cell count in million/mm.' >MCHC: ITcmoglobin in grams percent X 100 Packed cell volume * Female pink salmon from Olsen Hay not inrluded. 196 U.S. FISH AND WILDLIFE SERVICE of maturity, only the mean from each stage and the grand average, range, and standard deviation for the three stages combined are given in table 1. Statistical comparisons were made by the one-way analysis of variance, or F-test (Li, 1957). No distinction is made here between early- and late-run salmon at Olsen Creek. The specific gravities, erythrocyte counts, and hemoglobin concentrations fall within the ranges of those listed by Wintrobe (1933) for oceanic bony fishes, although the mean cor- puscular volume and mean corpuscular hemo- globin were high and are comparable with the more primitive fishes. In the comparison of my present findings on hematology of pink salmon with those of other workers, several points are of interest. In California, Robertson, Krupp, Favour, Hane, and Thomas (1961) found for chinook salmon that erythrocyte counts, hemoglobin levels, and packed-cell volumes increased dur- ing the migration and decreased during the spawning stage (to levels similar to those in animals in the open sea). My findings agree with those of Robertson and his associates in that packed-cell volumes (fig. 1) were higher MIGRATING M PRESPAWNING C SPAWNING C SPAWNING M SPAWNING F 1,1,1 1 1 10 20 30 40 50 PACKED CELL VOLUME (MILLIGRAMS PERCENT) Figure 1. — Packed-cell volume of blood of adult pink salmon in three stages of maturity. (C, sexes com- bined; F, females; M, males.) in migrating males than in spawning males. The packed-cell volumes in combined male and female samples were also higher in prespawn- ing populations than in spawning populations. Within the spawning population, the males had greater packed-cell volumes than the fe- males. A significant increase (at the 2.5-per- cent level) in mean corpuscular hemoglobin concentration between the prespawning and the spawning stages was concurrent with the small decrease in packed-cell volume. In his studies of pink salmon in Asia, Lysaya (1951) found that erythrocyte counts and hemoglobin levels fell noticeably between the time fish entered the estuary and the time they arrived on the spawning grounds. Such a trend is clearly evidenced by the decrease in packed-cell volume in my study, although it is not noticeable in the erythrocyte and hemo- globin values. The absence of a difference in hemoglobin concentrations between prespawn- ing and spawning stages in my work, was also reported by Sinderman and Mairs (1961) for the alewife, Alosa pseudoharengus, a fish that returns to the sea after spawning in fresh water. Benditt, Morrison, and Irving (1941) found that in Atlantic salmon (Salmo salar) afl!inity of hemoglobin for oxygen was greater while fish were in the spawning stage in fresh water than in the prespawning or migrating stage in salt water. This last phenomenon would com- pensate those changes mentioned above that would tend to decrease the oxygen-carrying efliciency of the blood. Perhaps an under- standing of these points will be possible when larger numbers of fish are analyzed at all stages of migration. BLOOD CHEMISTRY The concentrations of several components of blood (table 2) were determined by the tech- niques given in Fister (1950). As with the hematology and corpuscular indices, the meas- ured values of some of the characteristics of blood did not vary significantly among the three stages of maturity; only the mean from each stage and the mean, range, and standard devia- tion for the three stages combined are given in table 2. BLOOD OF ADULT PINK SALMON 197 Table 2. — Average values in blood chemistry of adult pink .tahiion at three stages of maturity from three areas of Alaska in 1963 (Numbers In parentheses are numbers o( samples analyzed! Stage of maturity 1 Albumin" 1 tllobulin ' 1 Glucose ■ ' Cholesterol ' Lipid Uric acid ' phosphorus ' 1 Urea ■ Creatinine - Migrating (Elfin Cove) Males Grams percent 1..5(11) 1.3(8) .7(8) Grams Miltigrams percent ! percent 0.4(12) lOirifi) AfiUigrams percent 835(10) Milligrams percent Milligrams percent 2.2(16) 1.4(S) 2.1(8) Milligrams percent 4.1(15) 5.1(15) Milliorajnit per cent 1.0(16) I'respawning (Little Port Walter) Males .7(8) 1.0(8) Females Combined 68(16) 656(14) 5.8(14) Spawning (Olsen Creek) Males - 1.8(26) .6(23) 7.4(23) 1.2(16) Early run 78(16) 43(11) 41(8) 75(.59) 30 23-167 494(16) 580(14) 542(7) 621 (54) 136 3R4 1 y50 15.9(14) 11.3(16) .7(16) Late run Females Late run . ... .6(5) 1.5(63) .6 .5-3.4 .4(5) .7(51) .4 .2-1.2 17.5(8) 1.7(5) 11.0(44) 1.6(04) 3.3 1 .5 6.7(6) 5.8(53) Combined ' Mean . 1 . 1 (32) Range ■ " Plasma. 1 Whole blood. " Female pmk saltnon from Olsen Bay not included in combined values. ALBUMIN AND GLOBULIN Albumin antJ globulin are discussed together because both are blood proteins. Comparisons are made between males and females (table 2) in the three spawning stages (figs. 2 and 3). MIGRATING M PRESPAWNING M PRESPAWNING f SPAWNING M SPAWNING F 1 1 1 1 ' 0.0 1.0 ALBUMIN IMILLtGRAMS PERCENT) 2.0 MIGRATING M PRESPAWNING M PRESPAWNING F SPAWNING M SPAWNING F J 1 I L. 0.0 0.5 GLOBULIN (MILLIGRAMS PERCENT) Figure 3. — Globulin of blood of adult pink salmon in three stages of maturity. (F, females; M, males.) Figure 2. — Albumin of blood of adult pink salmon in three stages of maturity. (F, females; M, males.) Although the range in values was large, the average concentrations of components in males showed little change from the migrating through the prespawning and spawning stages. The albumin-globulin ratio was greater than 1:1 — the ratio considered normal for mammals and most fishes (Shell, 1961). The albumin values for prespawning females were about half of those for prespawning males, whereas the globulin values for prespawning females averaged higher than tho.se for males (table 2). The albumin-globulin ratio was 0.7:1 for prespawning females. The globulin was great- ly reduced in spawning females, and the albu- min-globulin ratio (1.5:1) was more nearly 198 U.S. FISH AND WILDLIFE SERVICE like the ratio for spawning males (3:1). The results of my analysis of albumin and globulin are consistent with those of Robertson, Krupp, Favour, Hane, and Thomas (1961), who found that the normal ratio of albumin to globulin of 1 :2 in chinook salmon living in the sea was reversed in both sexes during migration but tended to revert to the original during spawn- ing. The greater reduction of albumin and glo- bulin in females than in males by spawning time probably indicates a greater depletion of body protein in egg formation. Shell (1961), who surveyed the nutritive, osmotic, and other functions of blood proteins in fish, found a cyclic reversal of the albumin-globulin ratio in small- mouth bass, Micropterus dolomieui, and in his review of the literature stated that "Results of determinations of the A:G ratio in fish are confusing." GLUCOSE My discussion of glucose levels includes com- parisons between pink salmon in the migrat- ing and the prespawning stages and between MIGRATING M PRESPAWNING C SPAWNING M' SPAWNING M" SPAWNING F" 1 1 — 1 L 1 1 1 1 , . 1 20 40 60 80 GLUCOSE (MILLIGRAMS PERCENIl 100 Figure 4. — Glucose of blood of adult pink salmon in three stages of maturity. (C, sexes combined; F, females; M, males; single asterisk, early run; double asterisk, late run.) pink salmon in early and late runs (fig. 4 and table 2) . The drop in glucose between the migrating and prespawning fish is significant at the 1-percent level, and the decrease from early to late spawners is significant at the 2.5- percent level. The different levels of glucose in spawning salmon may be attributed to the fact that the salmon of the early run at Olsen Creek were not completely ready to spawn, whereas the fish of the late run were actually spawning. The findings are in accord with those of Lysaya (1951). It is commonly assumed that carbohydrate metabolism in fish is inefficient. Robertson and Wexler (1960), however, found an increase in the number and size of islets of Langerhans in chinook salmon during the spawning mi- gration. Robertson, Krupp, Favour, Hane, and Thomas (1961) found an increase in blood glucose while fish were migrating from the open sea, followed by a tendency toward a decrease during spawning. They suggested that rising levels of blood glucose are due to glu- coneogenesis that results from the action of increasing adrenal corticoids on muscle and fat deposits and a simultaneous increase in insulin production to utilize the product. This viewpoint is somewhat corroborated by studies on the Fraser River in which sock- eye salmon have shown an 11- to 30-percent loss of body flesh (Idler and Tsuyuki, 1958) accompanying increased production of adrenal corticosteroid hormones (Idler et al., 1959). Chang and Idler (1960) observed that liver glycogen gradually decreased during migration in fresh water but increased at spawning. These changing glycogen levels were attributed to changing hormone balances. CHOLESTEROL Cholesterol levels are compared among the three stages of maturity (fig 5). A downward trend in cholesterol levels from the migrating to the spawning stage was consistent, i.e. sig- nificantly lower (at the 1-percent level) in the prespawning than the migrating fish and in the spawning than the prespawning fish. Robert- son, Krupp, Favour, Hane, and Thomas (1961) and Idler and Tsuyuki (1958) observed this same consistent downward trend in chinook salmon. Although I found this downward trend in cholesterol levels from the migrating to the spawning stage, levels in pink salmon within the spawning group were higher in the late run than in the early run. BLOOD OF ADULT PINK SALMON 199 MIGRATING M PRESPAWNING C SPAWNING M- SPAWNING M" SPAWNING F" 1 1 1 1 1 ._ J . 1 1. 500 CHOLESTEROL (MILLIGRAMS PERCENT) 1,000 PRESPAWNING C SPAWNING M* SPAWNING M" SPAWNING E" 1 1 r 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 LIPID PHOSPHORUS (MILLIGRAMS PERCENT) 20 Figure 5. — Cholesterol of blood of adult pink salmon in three stages of maturity. (C, sexes combined; F, females; M, males; single asterisk, early rim; double asterisk, late run.) Figure 6. — Lipid phosphorus of blood of adult pink salmon in two stages of maturity. (C, sexes com- bined; F, females; M, males; single asterisk, early run; double asterisk, late run.) The function of cholesterol in the metabol- ism of fishes (reviewed by Shell, 1961) re- mains obscure. In my study, however, choles- terol showed an inverse correlation with lipid phosphorus (significant at the 1-percent level). LIPID PHOSPHORUS Concentrations of lipid phosphorus in sam- ples from the prespawning and spawning stages and the early and late spawning runs are compared (fig. 6). No data are available from the migi-ating group. The increase in lipid phosphorus levels from the prespawning to the spawning stage was significant at the 1-percent level. The values for males in the spawning stage in the early part of the run (fig. 6) were also significantly higher than in the late run (at the 1-percent level.) The high values for lipid phosphorus in the females sam- pled in the late run may be due to a terminal increase in 17-hydroxycorticosteroids in fe- males as values at that time are dropping in males (Hane and Robertson, 1959). Although Shell (1961) found a direct correlation be- tween lipid phosphorus and the blood proteins (albumin and globulin), no such correlation is evident in my data. Comparisons between in- dividual animals, however, indicated a positive correlation (1-percent level of significance) with glucose. The results suggest a mechanism whereby the concentration of lipid phosphorus increases as cholesterol and glucose decrease. URIC ACID The values for uric acid are discussed for males and females in the three stages of ma- turity. The decline in uric acid concentration in the blood of males from the migrating to the prespawning stage is significant at the 1-per- cent level (fig. 7). The further drop in uric acid from the prespawning to the early part of the spawning stage (in the early run only) is also significant at the 1-percent level. No such drop is apparent, however, in the comparison of the males of the prespawning and the late part of the spawning stage (fig. 7). Within the prespawning stage, uric acid values were high- er in the females than in the males (sig- nificant at the 5-percent level). Uric acid con- centrations in females from the late spawning stage average only slightly higher than tho.se in the males (table 2). If uric acid is accepted 200 U.S. FISH AND WILDLIFE SERVICE MIGRATING M PRESPAWNINGM PRESPAWNING F SPAWNING W SPAWNING M" SPAWNING F" , 1,1,1 1 1 1 1 0.0 0.5 .0 2.0 2.5 URIC ACID (MILLIGRAMS PERCENT) Figure 7. — Uric acid of blood of adult pink salmon ir three stages of maturity. (F, females; M, males; single asterisk, early run; double asterisk, late run.) as an end product of purine metabolism, a de- cline in purine metabolism at spawning is indi- cated, and females maintain a higher level longer than males. UREA The variations in the concentration of urea among individual specimens from an area were so great that no trend is apparent for this blood component, which is the end product of nitrogen metabolism. Lysaya (1951) noted increasing concentrations as spawning ap- proached and attributed death of pink and chum salmon to urea poisoning. CREATININE No data are available on the concentrations of creatinine in blood samples from fish in the prespawning stage, and no trend is indicated by the values for the other groups. Creatinine is sometimes considered an end product of tissue catabolism, which is a dominant process in the fish sampled here. The values I found, however (table 2), are similar to those de- termined for smallmouth bass by Shell (1961) and for carp (Cyprinus carpio) and brook trout {Salvelinus fontinalis) by Field, Elveh- jem, and Juday (1943) for fish in which cata- bolism was not high. SUMMARY AND CONCLUSIONS Blood samples were taken from adult pink salmon collected at three stages of maturation during their migration to the spawning grounds — in the ocean actively migrating, mill- ing in the estuary of a spawning stream, and in fresh water on the spawning grounds. Basic hematological characteristics, includ- ing specific gravity, packed-cell volume, ery- throcytes, hemoglobin, corpuscular volume, corpuscular hemoglobin, and corpuscular hemo- globin concentration, were determined. Statis- tical analyses indicated no significant differ- ence among groups of fish. The concentrations of several components of blood indicate that several changes accompany migration and maturation. As pink salmon mature, utilization of protein reserves (evi- denced by lowered albumin and globulin levels in females) may result from rapid building of egg ti.ssue. Glucose levels declined, especially in females. Cholesterol concentrations also de- clined, although lipid phosphorus rose in both sexes ; the increase was especially noticeable in females. Lipid phosphorus may play an in- creasingly important part in energy transfer as salmon mature. The pink salmon from the spawning stream were from two distinct components of the run — the early and the late. The late spawners had significantly higher concentrations of choles- terol and uric acid, but lower levels of glucose and lipid phosphorus. I do not know if these differences are due to intrinsic genetic factors or are induced by extrinsic environmental factors. I could see no trend in urea or creatinine concentrations at the three stages of maturity. ACKNOWLEDGMENTS David Boye and Mrs. Frances Pierce helped with the statistical comparisons. Theodore R. Merrell, Jr., and several other staff members at the Bureau of Commercial Fisheries Bio- logical Laboratory, Auke Bay, Alaska, aided in collecting samples. This work was financed by BLOOD OF ADULT PINK SALMON 201 the Bureau of Commercial Fisheries under Contract No. 14-17-0005-46 with funds made available under the Act of July 1, 1954 (68 Stat. 376) commonly known as the Saltonstall- Kennedy Act. LITERATURE CITED He.nditt, E., p. Morrison, and L. Irving. 1941. The blood of the Atlantic salmon during migration. Biol. Bull. 80: 429-440. Chang, V. M., and D. R. Idler. 1960. Biochemical studies on sockeye salmon dur- ing spawning migration. XII. Liver glycogen. Can. J. Biochem. Physiol. 38: 553-558. Field, J. B., C. A. Elvehjem, and C. Juday. 1943. A study of the blood constituents of carp and trout. J. Biol. Chem. 148: 261-269. FiSTER, H. .1. 1950. Manual of standardized procedures for spectrophotometric chemistry. Standard Scien- tific Supply Corp., New York. [Pagination varied.] Hane, S., and 0. H. Robertson. 1959. Changes in plasma 17-hydroxycorticoster- oids accompanying sexual maturation and spawn- ing of the Pacific salmon (Oncorhynchi(s tsharvytseha) and rainbow trout {Salnw gaird- nerii) Proc. Nat. Acad. Sci. 45: 886893. Hesser, E. F. 1960. Methods for routine fish hematology. Progr. Fish-Cult. 22: 164-171. Idler, D. R., A. P. Ronald, and P. J. Schmidt. 1959. Biochemical studies on sockeye salmon dur- ing spawning migration. VII. Steroid hor- mones in plasma. Can. J. Biochem. Physiol. 37: 1227-1238. Idler, D. R., and H. Tsuyuk:. 1958. Biochemical studies on sockeye salmon dur- ing spawning migration. I. Physical measure- ments, plasma cholesterol, and electrolyte levels. Can. J. Biochem. Physiol. 36: 783-791. Li. J. C. R. 1957. Introduction to statistical inference. Edwards Bros., Ann Arbor, Mich. Lysaya, N. M. 1951. Oh izmenenii sostava krovi lososei v period nerestovykh migratsii. (Changes in the blood composition of salmon during the spawning migration.) Izv. Tikhook. Nauch.-Issled. Inst. Rybn. Khoz. Okeanogr. 35: 47-60. [Translated in Pacific salmon, pp. 199-215. Nat. Sci. Found, and U.S. Dep. Interior Israel Program Sci. Transl.] Robertson, O. H., M. A. Krupp, C. B. Favour, S. Hane, and S. F. Thomas. 1961. Physiological changes occurring in the blood of the Pacific salmon {Oncorhyncluis tshawy- tscha) accompanying sexual maturation and spawning. Endocrinology 68: 733-746. Robertson, 0. H., M. A. Krupp, S. F. Thomas, C. B. Favour, S. Hane, and B. C. Wexler. 1961. Hyperadrenocorticism in spawning and nonmigratory rainbow trout {Satmo gairdncrii) ; comparison with Pacific salmon (genus Oncor- hynchus). Gen. Comp. Endocrinol. 1: 473-484. Robertson, O. H., and B. C. Wexler. 1960. Histological changes in the organs and tis- sues of migrating and spawning Pacific salmon (genus 0»co)7ii/Hc/n(s). Endocrinology 66: 222- 239. 1962. Histological changes in the pituitary gland of the Pacific salmon (genus Oncorhynchns) ac- companying sexual maturation and spawning. J. Morphol. 110: 171-185. Robertson, 0. H., B. C. Wexler, and B. F. Miller. 1961. Degenerative changes in the cardiovascular system of the spawning Pacific salmon (Oncor- hynchns tshawytRcha) ■ Circulation Res. 9: 826- 834. Shell, E. W. 1961. Chemical composition of blood of small- mouth bass. U.S. Fish Wildl. Serv., Res. Rep. 57, 36 pp. Sinderman, Carl J., and Donald F. Mairs. 1961. Blood properties of prespawning and post- spawning anadromous alewives {Alosa pscudo- harengus). U.S. Fish Wildl. Serv., Fish. Bull. 61: 145-151. Wintrobe, M. M. 1933. Variations in the size and hemoglobin con- tent of erythrocytes in the blood of various verte- brates. Folia Haematol. 51: 32-49. 202 U.S. FISH AND WILDLIFE SERVICE CROSS-REACTIVE PROPERTIES OF ANTISERA PREPARED IN RABBITS BY STIMULATIONS WITH TELEOST VITELLINS By Fred M. Utter. Chemist and George J. Ridgvc ay. Biochemist.^ Bureau of Commercial Fisheries Biological Laboratory. Seattle. Wash.. 98102 ABSTRACT Antisera were prepared by injecting rabbits with egg or serum vitellin preparations from seven teleost species be- longing to different families. The ranges of reactivity of these antisera were tested with sera from mature females of nine teleost families as well as with sera from females of spiny dogfish. Pacific lamprey, and white sturgeon. All of these antisera reacted with vitellins from all species tested from the homologous families. Antisera prepared against rockfish and flounder vitellins cross-reacted with sera from mature females of all teleost species tested. A greater anti- genic complexity in the vitellins of more taxonomically advanced species than more primitive species is indicated by the results of the reactions and absorption tests. The results are of practical importance in studies on maturity of fishes and have theoretical implications in the field of systematics. Fishery researcher.'^ have used serological techniques with increasing frequency since 1950. Much of the work has been directed toward identification of populations either through blood-grouping techniques or studies of variable serum antigens (Gushing, 1964). Ridgway. Klontz, and Matsumoto (1962) ob- served a characteristic antigen in the serum of maturing and mature female sockeye salmon {OncorJiiiiichus iierka) . A review of the litera- ture and subsequent studies by our group has revealed that similar components occur not only in all the teleosts but also in all vertebrate classes were oviparity occurs (Urist and Schjeide, 1961; Drilhon and Fine, 1963), These antigens appear to have considerable practical value in investigations of maturity in female fish because of their connection with the proc- ess of sexual maturation (Ho and Vanstone, 1961; Olivereau and Ridgway, 1962; Ridgway, 1964; Utter and Ridgway, 1966). The function of the blood serum as a trans- porting medium between the site of synthesis in the liver and the site of storage in the ovary appears to explain the presence of yolk com- ponents in the blood (Vanstone. Maw, and Common, 1955), The serum vitellins studied have displayed similar biochemical properties, are characterized as phospholipoproteins, and I'revf. .Tl.ir,. Wfvr l*,ontliti.-H> Harbor. Elaine. FISHERY BULLETIN: VOLUME 66, NO, 2 Published April 1967. conform to the classical de.scription of avian vitellin as the water-insoluble fraction of the egg yolk (Jukes and Kay, 1932; Vanstone, Maw, and Common, 1955), This report is based on data obtained through testing numerous antisera for cross- reactive properties. The antisera were pre- pared in rabbits against vitellins of teleosts. We intend to bring out points of both practical significance and theoretical interest, METHODS AND MATERIALS IMMUNOLOGICAL TESTS A microslide adaptation of the double dif- fusion method of Ouchterlony as modified by Ridgway, Klontz, and Matsumoto (1962) was used for all serological tests. The agar medium consisted of 1.5 percent Difco agar, 0.72 per- cent sodium chloride, 0,60 percent sodium ci- trate, 0,01 percent merthiolate. and 0,01 per- cent trypan blue. Wells were punched in the agar at 8-mm. intervals and filled to a volume of about 0.01 cc. of reactant. Slides were eval- uated after 24 hours of incubation at 37'- C. PRODUCTION OF ANTISERA Table 1 lists the antisera used in this study. Egg vitellin preparations were made by blend- ing and centrifuging one part eggs with three 203 Table 1. — Aniisera iiscrl in this stiidij Designation Number Vitcllin source of 1 rabbits AntiCI.M... Anti-SM AntiCM.... Anti-GM.... Anti-TM Anti-RM Anti-HM.... Pacific herring (Cliipea harengus paUasi) qh^s I Cllinooli salmon {(liicorliynchii.i Ixhairulscha) eggs. :i Norlliern stiuawfisli {Ptychocheiliis oregotteihiis) ' 1 eSBS. I*aciflc cod lOadiix mncrocephaliis) scrum 1 Copper roclifisli i Selia.stodefi cauriiius) eggs '.i Starry (louiKler (I'lalkMhys slellaliis) eggs... .i parts 1 percent saline in a Waring Blender. - After centrifugation, the addition of 11 parts of distilled water precipitated the vitellins from the supernatant fluid. The precipitate was dissolved again in saline, reprecipitated and redissolved, and used for injections. Whole serum from a mature female Pacific cod was u.sed to produce the anticod vitellin reagent. The resulting antiserum was absorbed at a 1 :1 ratio with male cod serum before testing. Usually the vitellin-bearing materials were suspended in a bayol-arlacel mixture and in- jected into the rabbits intraperitioneally. Con- sistently uniform results were obtained when other injection procedures were used, but a greater number of injections was usually re- quired. Single bleedings were used for testing with the exception of the reagent prepared against starry flounder vitellin which was a pool of numerous bleedings from five rabbits. The antisera produced in different rabbits in- jected with the same vitellin material were qualitatively very similar. This uniformity of reagent indicates that the differences reported later are not due to variations in the immune response of individual rabbits. COLLECTION OF SERUM SAMPLES Samples of fish serum were taken from whole blood that had been processed within 48 hours after collection ; the samples were then stored at -35° C, a temperature at which the vitellin fraction appeared to be stable. Some sera had been stored as long as 8 years when tested. REACTIVE PROPERTIES OF THE ANTISERA AND VITELLINS TESTED Table 2 summarizes the data obtained l: Trade names referred lo in lliis piibli<-atioii do not imply en«Iorsement »»f coMilriercia! iinidui-ts. through testing of sera from mature females of various fish species. All antisera were tested with the same fish sera ; this testing included males as well as females from most species. The only reaction with male serum occurred between the antirockfish reagent and male rock- fish serum. This reaction was very weak and was most likely the result of nonvitellin anti- gens present in the injected material. The re- action with male rockfish sera could not be confused with the reaction with female rockfish sera. Table 2. — Cross-reacliriti/ of rabbit antilelcnal vitellin sera tcilh vitellins of fish representing various laxonomic groups ' |X = strong reaction ; W= = weak reaction; 0= no reaction! Nonteleosts Teleosts M w Keagcnt a 2 s ■a In s 2 o p s 2 o 2 1 2 ■5 1 a c ■a c o 3 O e a -5 J3 6 Q. 2 s 2 bc £ >, H o i! 53 < O o u CO u a U} CO U Ph ca Anti- Clupeid (CLM)... X w u u Sulmoiiid (SM).... II (1 II X II II \v Cvprimd (CM).... X w X u liadid Ki.M) (1 II 11 X w u u II Scombrid (TM).... n w w w X X X X X II Scorpaeiiid (KM),, X X X X X X X X X X Pleuroncctid(HM). X X X X X X X X X (1) P.'jcific lamprey. Lamprira iTidnlnln: (2) spiny dogfish, Squalus acanthias: (31 White sturgeon. Aripni:^fr trnit^innntmius: (4) sh;id. Alona .yapidissima: Pacific herring, Cliipca hareiigiis pallasi: (.l) sockeye salmon, Oncorhynchus nerkii: (6) carp. Cyprinus carpin: northern sqiiawfish I'lychochrilus oregon- etisis: largescalc sucker, Catostomns macrochfitus; (7) Pacific hake, \ferlucciux prnduclux: Pacific cod. Gadus macrocephaliis: (8) bigeye t(nia, Thuniius oUfnuH; Pacific mackerel. Scomber jopajticux; (9) copper rockfish, Sehaslodes canrinus: (101 cabczon, !icorparniclilhy!i marmnrnlnx: staghorn sculpln, Lcplocotlus armalus; (11) sand sole, I'srilichthys mdnnoxlirliix: English sole, Parophrya veluhis: starry flounder, Plnlkhlhys xirllninx: (12) northern mid- sliipman, Porkhibys notatus. All antisera reacted strongly with sera from mature females within the families that pro- vided the vitellin for antibody stimulation. Because of this high degree of cros.s-reactivity within families, the reactions of the antisera may be considered mainly with regard to the family rather than the species from which the vitellin used for antibody stimulation origin- ated. The arbitrary designations given in tables 1 and 2 refer to vitellin of any species of that family. Four of the reagents also reacted strongly beyond the immediate family group; two of 204 U.S. FISH AND WILDLIFE SERVICE them reacted distinctly with sera from mature females of all teleost species tested. Figure 1 illustrates the reactions of sera from mature Figure 1. — Reactions of sera from mature females of six teleost species with rabbit anti-RM serum. Peri- pheral wells contain sera from 1, copper rockfish; 2, sand sole; 3, bigeye tuna; 4, Pacific cod; 5, carp; 6, sockeye salmon. female teleosts of six different families when tested with the anti-RM reagent. The strongest reaction was with the female rockfish serum. The degree of cross-reactivity regularly de- creased through the somewhat distantly related salmonoids and cyprinoids. Even in these groups, however, the reaction was clear. ponents. The SM vitellin cross-reacts com- pletely with the antibodies directed against one of these components. Two components are visible in figure 2b that react with the anti- SM antiserum. The RM vitellin cross-reacts partially and very weakly with the antibodies directed against one of these components. Table 3. — Results of ahsorption^ of rahhil anlirochfish riieUin (RM) serum with sera of mature female teleosts (X = strong reaction: \V= weak reaction: 0= 10 reaction. See figure 1 ] Fish sera used Fish sera tested tor absorption Rockfish Sole Tuna Cod Carp Salmon Unabsorbed Rockflsli X X X X (1 i X X X w X X X X X X X X Sole n Tuna Cod Carp Salmon XXX X Comparisons similar to those presented in figure 2 were made between the broadly cross- reactive antisera (anti-RM and anti-HM), the less cross-reactive antisera (anti-CLM, anti- SM, anti-CM, and anti-GM), and the cor- responding vitellins. All results were similar. The weak or negative reactions of the RM and HM vitellins with the less cross-reactive or group-specific antisera contrasted with the strong reactions of the anti-RM and anti-HM antisera with the vitellins which elicited the less cross-reactive antisera present an interest- ing serological phenomenon. The cross-reactive antibodies of the anti-RM and anti-HM anti- sera apjear to have a considerably greater avidity for the vitellins which elicit group- specific antisera than the group-specific anti- bodies have for the RM and HM vitellins. Table 3 gives the results of absorptions of the anti-RM reagent with the fish sera of figure 1. It is evident from both figure 1 and table 3 that the anti-RM reagent contains antibodies of numerous specificities. Figure 2 presents a more detailed examina- tion of the relationship between SM and RM vitellins. It is evident from figure 2a that the RM vitellin has at least three distinct com- Both the homologous and cross-reaching heterologous antisera gave uniform results where tested with a larger number of individ- uals. Randomly selected sera from 48 sockeye salmon were tested with anto-SM, anti-RM, and anti-HM reagents; 48 halibut sera were tested with the anti-RM and anti-HM reagents. Re- sults were identical regardless of the antiserum used, including two weak but positively react- ing halibut sera. CROSS-REACTIVE PROPERTIES OF ANTISERA 205 SM + RM Fiai'RE 2. — The relation between SM and RM vitellins detected (a) by rabbit anti-RM serum and (b) by rabbit anti-SM serum. Arrow in (b) indicates the partial cross reaction between RM vitellin and the anti-SM reagent. PRACTICAL APPLICATIONS FOR STUDIES OF MATURITY The broad cross-reactive range of antisera produced against the vitellins of rockfish and starry flounder is of practical importance. It is likely that these reagents react with serum vitellins from mature female teleosts at least through the taxonomic range of this study. An investigator wishing to include serological data in maturity studies may therefore use a single reagent throughout a I'ange of teleost species rather than produce different antisera for relatively limited taxonomic groupings. The necessity of obtaining vitellin-bearing material for immunizations from species where such materials would be difficult to obtain or process is also eliminated. The data suggest that vitellins of the most taxonomically advanced species stimulate the highly cross-reactive antisera. Some theoretical implications of this apparent trend are dis- cussed below. As a practical consideration, however, it appears that vitellin from Perci- form or closely allied species may be most likely to stimulate antisera which have broad cross- reactive properties. SYSTEMATIC CONSIDERATIONS Nuttall (1904) in an early immunological study, observed that the quantities of precipi- tates formed by specific antigen-antibody inter- actions decrease as the taxonomic relationships become more distant from the materials used in antibody stimulation. The present study agrees generally with this observation. As illustrated by figure 1, the broadly cross-reac- tive antisera reacted most strongly with the homologous vitellin and least strongly with the most distantly related cpyrinid and salmonid vitellins. A notable exception is the reaction of sturgeon vitellin with anti-cyprinid vitellin, an antiserum which fails to react with vitellins of numerous, more closely related, groups. The other exceptions include the weak cross-reac- tions of antisalmonid vitellin with rockfish vitellin but not with herring or carp vitellin, and the similarly weak cross-reaction of anti- clupeid vitellin with rockfish vitellin but not with salmon vitellin. Fine, Buffa, and Drilhon (1964) found a component in mature female marine lampreys analogous to the teleost vitellins described in this report. The spiny dogfish egg, unlike those 206 U.S. FISH AND WILDLIFE SERVICE of many sharks, is provided with an abundance of yolk material. A Saline extract of the dogfish yolk material was tested in addition to the sera from numerous adult dogfish; this extract also failed to react with any of the antisera used in this study. The lack of reactivity observed here with yolk materials from females of dogfish or lampreys appears to reflect the phylogenetic gap between the teleosts and these more primi- tive vertebrates. The vitellin substances of the advanced tele- osts that stimulate production of the broadly cross-reactive antisera appears to be biochemi- cally and antigenically more complex than those of the more primitive teleosts. It is evident from figure 1 and table 3 that only a small fraction of the total number of anti-RM anti- bodies react with SM vitellin; the major anti- genic vitellin component of SM is detected, however, by the anti-RM reagent (Figure 2). Possibly the vitellin antigens of more primitive teleost species have been retained in certain advanced species without extensive modifica- tion during the evolution of additional vitellin substances. This study further demonstrates the useful- ness of serological methods to determine ma- turity in oviparous vertebrates. The results are also of significance in systematics. The exist- ence of antigens in the sera of maturing fe- males which do not occur in the sera of males and immature females must be taken into ac- count in studies that attempt to apply serology to problems of taxonomy. These antigens them- selves, as was demonstrated here, also offer additional materials for more detailed examina- tions of systematic relationships. LITERATURE CITED Gushing, John E. 1964. The blood groups of marine animals. Adv. Mar. Biol. 2: 85-131. Drilhon, a., and J. M. Fine. 1963. Dimorphisme sexuel dans les proteines se- riques de Salmo salar: Etude electrophoretique Compt. Rend. Soc. Biol. 157(9): 1897-2000. Fine, J. M., G. A. Boffa, and A. Drilhon. 1964. Etude electrophoretique et immunologique des proteines seriques de la lamproie marine (Petromyzoon marinus L.) Compt. Rend. Soc. Biol. 158 (10) : 2021-2025. Ho, F. Ghung-Wai, and W. E. Vanstone. 1961. Effect of estradiol monobenzoate on some serum constituents of maturing sockeye salmon (Oncorhynchus ncrka). J. Fish. Res. Bd. Can. 18: 859-864. Jukes, T. H., and H. D. Kay. 1932. Egg-yolk proteins. J. Nutr. 5: 81-101. NUTTALL, G. H. F. 1904. Blood immunity and blood relationship. Cambridge Univ. Press, New York, 444 pp. Olivereau, Madeleine, and George Ridgway. 1962. Cytologie hypophysaire et antigene serique en relation avec la maturation sexuelle chex Oncorhynchus species. Compt. Rend. Seances Acad. Sci. 254: 753-755. Ridgway, George J. 1964. Salmon serology. In U.S. Bureau of Com- mercial Fisheries, Report on the Investigations by the United States for the International North Pacific Fisheries Commission-1962, pp. 107-110. Int. N. Pac. Comm. Annu. Rep. 1962. Ridgway, George J., G. W. Klontz, and C. Matsumoto. 1962. Intraspecific differences in serum antigens of red salmon demonstrated by immunochemical methods. Int. N. Pac. Fish Comm., Bui. 8: 1-13. Urist, Marshall R., and Arne 0. Schjeide. 1961. The partition of calcium and protein in the blood of oviparous vertebrates during estrus. J. Gen. Physiol. 44: 743-756. Utter, Fred M., and George J. Ridgway. 1967. A serologically detected serum factor as- sociated with maturity in English sole, Paro- phyrys vetulus, and Pacific halibut, Hippoglos- sus stenolepis. U.S. Fish. Wildl. Serv., Fish. Bull. 66: 47-58. Vanstone, W. E., W. A. Maw, and R. H. Common. 1955. Levels and partition of the fowl's serum proteins in relation to age and egg production. Can. J. Biochem. Physiol. 33(6): 891-903. CROSS-REACTIVE PROPERTIES OP ANTISERA 207 OCCURRENCE OF MACROZOOPLANKTON IN TAMPA BAY, FLORIDA, AND THE ADJACENT GULF OF MEXICO ' By John A. Kellv, Jr., and Alexander Dragovich, Fishery Biologists. Bureau of Commercial Fisheries Biological Laboratory, St. Petersburg Beach, Florida, 33706 ABSTRACT This report describes a 12-month (September 1961 through August 1962) study. Plankton was collected at 14 locations with a No. 000, one-half meter net, which strained an estimated 35 m.' of water per tow. Wet plank- ton volumes varied from < 0.5 to 92.0 ml. and averaged 7.0 ml. per tow. Fift>-two percent (by volume) of the plankton was collened in the summer, 25 percent in the fall, 18 percent in the spring, and 5 percent in the winter. Lucifer faxoni, the most numerous organism, accounted for 18.5 percent of the total plankton volume. Sixteen species, 24 genera, 30 families, and 21 taxonomic categories higher than family were identified. Decapod crustaceans accounted for 87 percent of the total number of zooplankters collected. The most numerous organisms, in descending order were Lucifer faxoni, larval porcellanids, brachyurans, chaetognaths, copepods, larval polychaetes, carideans, appendicularids, larval fish, fish eggs, thallas- sinids, cladocerans, and larval stomatopods. Larval forms of commercially important species were Penaeus duorarum, Bretoortia spp., Anchoa sp., Trachinotus spp., Leiostomus xanthurus, Cynoscion spp., and Soleidae. Observed temperature ranged from 12.8° to 32.0° C. and salinity from 19.00 to 36.00 p.p.t. In relating the abundance of zooplankton to temperature and salinity the data sug- gested that low temperatures and salinity values were more restrictive than high ones to most of the organisms. A study of macrozooplankton was under- taken as part of an investigation of estuarine biology in the eastern Gulf of Mexico. The pri- mary aim was to determine temporal and spatial variations in the abundance of macro- zooplankton in the surface waters of Tampa Bay and the adjacent Gulf of Mexico, and to relate the occurrence of frequently collected taxa to water temperature and salinity. The abundance and composition of zooplank- ton provide an important index of biological production in estuaries, because zooplankters are the basic food of many marine organisms. Mysids, euphausids, amphipods, larval stomato- pods, and fish larvae are frequent in stomachs of commercially important fishes (King, 1954). The bulk of this plankton, however, reaches large fish indirectly through their consumption of foraging organisms. The literature on zooplankton in the coastal waters of west Florida is limited. No reports deal with the seasonal composition of zooplank- ton throughout Tampa Bay. Published material ^ Contribution No. 27. Bureau of Commercial Fisheries Biological Laboratory. St. Petersburg Beach. Fla. FISHERY BULLETIN: VOLUME 66, NO. 2 Published April 1967. includes: a description of certain biological, taxonomic, and ecological aspects of the chae- tognaths of the west coast of Florida (Pierce, 1951) ; notes on chaetognaths from the Gulf of Mexico (Tokioka, 1955) ; the seasonal distribu- tion of penaeid larvae from the lower portion of Tampa Bay, Fla., and the adjacent Gulf of Mexico waters (Eldred, Williams, Martin, and Joyce, 1965) ; a qualitative and quantitative seasonal study of the copepods of Alligator Harbor (Grice, 1956) ; studies of the taxonomy of several calanoid copepods in the eastern Gulf of Mexico (Fleminger, 1957a and 1957b) ; a preliminary report on the plankton of the west coast of Florida with a discussion of the distribution and occurrence of copepods and other crustaceans (King, 1949) ; and records of various taxa from the marine and brackish waters of south Florida (Davis, 1947, 1948, 1949, 1950; Davis and Williams, 1950; and Dragovich, 1963). DESCRIPTION OF THE AREA Tampa Bay is a shallow embayment consist- ing of five sub-areas, also identified as bays — 209 Old Tampa Bay, Hillsborough Bay, Tampa Bay, Boca Ciega Bay, and Terra Ceia Bay. Collectively, these areas have a shoreline of 341 km. and cover an area of 896 km.-, 90 percent of which is less than 6.7 m. deep (Olson and Morrill, 1955).= The principal tributaries of Tampa Bay are the Hillsborough, Alafia, Mana- tee, and Little Manatee Rivers. Their discharge i.s largely influenced by rainfall (Dragovich and May, 1962) and is subordinate to tidal exchange in the circulation of Bay water (Goodell and Gorsline, 1961). The climate of the Bay area is subtropical. The mean monthly air temperature at Tampa, Fla., averages 22.3° C. annually and varies from 16.2° C. (January) to 27.8° C. (August) \ The rainy season in the Tampa Bay area usu- ally extends from June to October. Mean rain- fall varies monthly from 3.7 cm. (November) to 21.9 cm. (July) and totals 131.0 cm. an- nually. = Olson, F. C. W.. and John B. Morrill, Jr. 195.5. Literature survey of the Tampa Bay area. Armed Serv. Tech. Info. Agenov, AD 81621 (Pt. 1) : 66 p.p. ^ The rainfall and temperature data used in this section are climatological normals (1931-60) compiled by the U.S. Depart- ment of Commerce, Weather Bureau, and published in the 1964 Annual Summary of riimatological Data For Tampa, Fla. APPARATUS AND METHODS FIELD PROCEDURES Plankton was sampled monthly in Tampa Bay and adjacent waters of the Gulf of Mexico from September 1961 through August 1962 (table 1). Surface samples were collected at 14 stations (fig. 1 and table 2) with a No. 000, Table 2, — Sampling tocalions in Tampa Hay and the adjacent Gulf of Mexico, September 1961-Augusl 1962 Station 2. 3- 4. 5. 6. 7. 8. 9. 10 11 12 13 14 Latitude N. Longitude W. Area description 27°35.8' 82°67.r 10 miles (18.5 km.) offshore 27°36.0' 82°50.0' 3M miles (6.5 km.) offshore 27<'35.5' 82°45.5' Egmont Key 27°32.7' 82°43.7' Lower Tampa Bay 27''38.8' 82°42.7' Boca Ciega Bay 27''41.5' 82°44.r Boca Ciega Hay 27°42.2' 82">40.5' Boca Ciega Ray 27<'33.0' 82°35.7' Terra Ceia Bay 27°41.3' 82°32.9' Central Tampa Bay 27°47.6' 82°34.4' Upper Tampa Bay 27''56.6' 82°37.0' Central Old Tampa Bay 28°00.9' 82<'40.7' Upper Old Tampa Bay 27°48.7' 82°26.8' Lower Ilillsljorough Bay 27°53.7' 82°26.4' - Upper Ilillsborough Bay '/2-ni., nylon plankton net (mesh size 1.024 mm.). This net was selected primarily for the collection of larval fishes and invertebrates. Tows were made at 5.6 km. per hour (3 knots) for 2 minutes and randomly with respect to tidal stage. Vessel speed was determined before sampling by clocking elapsed time over a known distance. Net-towing rates were held Table \.— Dates and numbers of plankton tows in Tampa Ray and the adjacent (1 ulf of Mexico, September 1961 thro>i< /h Aug} ,■!( IWJ Dates Stations All 1 2 3 4 5 6 7 8 9 10 11 12 13 14 stations Number of plankton tows Sept. 18-27, 1961 1 1 2 1 1 2 1 1 2 1 1 1 2 2 1 2 2 1 2 2 1 2 1 1 2 1 1 2 I 2 1 1 3 1 2 1 1 2 Oct. 2-30... Nov. 1-30 17 28 Season total . 4 * 4 3 c 5 * 4 4 4 5 4 4 ■ " Winter Dec. 11-12, 1961 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 I 2 2 1 2 2 2 2 1 2 1 1 Jan. 8-26. 1962-. Feb. ,5-10 25 Season total.. . 4 4 4 4 6 6 6 4 3 S 5 6 4 4 Spring Mar. 8-29. 1962 1 2 1 1 2 1 1 2 1 1 2 1 2 2 2 2 2 1 2 2 2 1 2 1 1 2 2 1 2 2 1 2 2 2 2 1 1 2 2 I 2 2 Apr. 9-26 18 May 7-29 28 21 Season total 4 4 4 4 6 5 2 2 2 6 4 5 5 5 5 5 5 Summer Jun. 11-28, 1962 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 1 2 2 2 2 2 2 2 1 1 2 2 2 2 Jul. 9-25 28 Aug. 6-27. 26 22 Season total 6 6 6 6 6 6 6 5 5 6 6 4 4 4 Total for 12 months.. 18 18 18 17 23 22 23 17 17 20 20 20 17 17 267 210 U.S. FISH AND WILIFE SERVICE 28' 00' 27° 30' 33' 00' I »1 Figure 1. — Sampling locations in Tampa Bay and the adjacent Gulf of Mexico. nearly constant by maintaining a fixed engine speed. The volume of the cylinder of water strained through the net was determined to be 35 m.^ (calculated from the towing distance and the area of the net mouth). Since no cor- rection was made to adjust for the effects of currents and clogging on the flow of water through the net, the quantitative data are not exact. Plankton samples were preserved immed- iately after their collection in 5 percent neu- tralized formalin and stored in 30-oz. jars. Water temperature and water sample (for salinity titration) were taken at the beginning of each plankton tow. Temperature was read to the nearest 0.1° C. with a thermister (Whit- ney^ underwater thermometer, Model TC-5) . * References to trade names in this publication do not imply endorsement of commercial i>rodncts. MACRO-ZOOPLANKTON IN TAMPA BAY AND ADJACENT GULF 211 LABORATORY PROCEDURES Plankton samples were placed in enameled photographic trays. Seaweeds and jelly fishes were removed from the samples manually, and the remaining volume was determined in the laboratory by the displacement method de- scribed by Thrailkill (1957). Plankton counts were made from aliquots whenever the wet plankton volume of the sample exceeded 0.5 ml. After the sample had been diluted to a known volume, usually 500 ml, four 5-ml. aliquots were withdrawn with a calibrated pipette. They were then transferred into a quadripar- titioned petri dish for examination and count- ing under a binocular dissecting microscope. Samples having a wet volume of 0.5 ml. or less were transferred directly to petri di.shes for counting. All samples were examined routinely for unusual organisms that might have been excluded from the aliquots. The mean number of organisms per cubic meter of water was calculated for each taxonomic group. Body lengths of chaetognaths and fish larvae were measured-chaetognaths from the anterior extermity of the head to the tip of the caudal segment, excluding the caudal fin (Owre, 1960), and fish larvae from the snout to the base of the hypural plate (standard length). ENVIRONMENTAL FACTORS HYDROLOGY The minimum and maximum water tempera- tures observed were 12.8° and 32.0° C. The smallest range in temperature at individual stations occurred in Boca Ciega Bay (15.3° C), and the greatest (18.4° C.) in Old Tampa Bay (table 3). Seasonally, the range in mean temperatures between stations located on a traverse from offshore to Hillsborough Bay (.stations 1, 2, 3, 4, 9, 10, 13. 14) was greatest in the spring (3.9° C.) and lowest in the win- ter (1.0° C). These ranges in the fall and summer were 1.7° and 1.2° C. respectively. Salinity was determined by Mohr-Knudsen method (Knudsen, 1901). Lowest salinities were usually in the upper area of Hillsborough Bay and highest 18.5 km. (10 nautical miles) offshore (table 3). The seasonal differences in mean salinities between these two areas (sta- tions 1 and 14) were 8.62 p.p.t. (fall), 6.75 p.p.t. (winter), 8.03 p.p.t. (spring), and 13.58 p.p.t. (summer). The range in salinity at in- dividual stations decreased progressively from upper Hillsborough Bay seaward. The smallest range was 18.5 km. offshore and the greatest in upper Hillsborough Bay, where temporal changes in salinity generally followed the dis- Table S.—Mean surface water temperature and saliniti/ for Tampa Bay and the adjacent Gulf of Mexico. September 1961 through August 1962 Seasons Temperature at stations 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fall -. ° C. 24.9 17.7 21.2 29.3 24.0 °C. 24.7 17.2 21.7 30.2 24.2 ° C. 24.5 17.4 22.4 30.0 24.3 "C. 26.2 17.5 22.0 30.3 24.6 ° C. 24.5 17.9 21.0 30.3 23.4 24.2 18.0 19.8 30.4 23.2 " C. 23.9 18.2 20.9 29.7 23.2 "C. 25.6 18.3 23.2 30.6 24.8 " C. 25.5 17.7 22.7 30.2 24.7 "C. 24.7 17.2 23.4 29.7 24.0 °C. 24.6 17.3 23.9 30.2 24.3 "C. 23.5 18.7 21.7 30.0 22.9 24.8 17.2 24.1 30.4 24.1 "C. 25.4 18.2 Winter i... - Spring 30.5 24.8 12 months --- 15.3 30.9 15.1 31.2 15.4 31.4 15.4 31.4 14.3 31.9 15.2 31.4 15.4 30.7 15.1 31.5 15.4 31.6 13.8 30.6 13.2 31.6 12.8 31.2 13.4 31.2 13.8 32.0 Maximum I2-nionth range 15.6 16.1 16.0 16.0 17.6 16.2 15.3 16.4 16.2 16.8 18.4 18.4 17.8 18.2 Salinity at station s 1 2 3 4 6 6 7 8 9 10 11 12 13 14 Fall Winter 34.96 34.45 34.89 35.67 35.07 %, 34. M 33.64 34.30 35.31 34.52 %o 34.03 33.82 33.76 34.99 34.23 %. 33.80 33.20 33.97 35.16 34.18 %o 32.30 33.01 33.90 33.98 33.34 32.94 33.22 33.79 34.62 33.67 %, 31.82 32.01 32.44 32.59 32.23 %. 31.15 31.96 31.57 30.56 31.26 %. 29.53 30.52 31.19 29.46 30.17 %. 28.29 29.39 29.45 27.90 28.74 %. 25.73 27.76 29.06 28.46 27.89 %. 24.66 26.86 28.52 28.05 26.96 %. 27.27 28.38 28.10 25.96 27.46 26,34 27.70 26.86 22.09 25.81 Summer - 12 month.s , _ 33.96 35.93 33.04 36.00 32.63 35.58 32.74 36.00 31.29 35.35 32.38 30.05 27.63 33.68 27.45 33.13 24.78 30.79 24.11 29.83 21.82 29.67 24.36 29.85 19.00 28.51 Maximum - 35.39 34.33 12-month range 1.97 2.96 2.95 3.26 4.06 3.01 4.28 6.05 5.68 6.01 5.72 7.85 5.49 9.61 212 U.S. FISH AND WILIFE SERVICE charges of the Hillsborough River ^ — the major source of river water to the bay (fig. 2) . ■^ River discharge data (tig. 2) and rainfall data used in this section were taken from the 1961 and 1962 Surface Water Records of Florida, Vol. 1 : Streams, compiled by the U.S. Department of the Interior. Geological Surve.v, and from the 1961 and 1962 Annual Summaries of Local Climatological Data for Tampa. Florida, pub- lished by the U.S. Department of Commerce. Weather Biireau. J 3 30 27 24 2 I 1 8 - I 5 1 2 36 33 ; 30 27 24 21 1 8 WtlER lEMPEXTURE !I0-I193 _ HO- \}1i - J20-(J) - l(0-JU - 10- I5S - - - 1- - JO ^ HI ItSeOROUCH RIVEII DISCHARGE WINTER I SPRING ' SUMMER SEPI IJSI — AUC IIE2 Figure 2. — Mean monthly discharge rate for the Hills- borough River and monthly surface water tempera- ture and salinity for upper Hillsborough Bay (station 14) and 10 nautical miles offshore (station 1), Sep- tember 1961 through August 1962. Mean values for temperature and salinity are given when two meas- urements of these variables were made in a month. CLIMATOLOGY Climatological data were taken from the records of the U.S. Weather Bureau for Tampa, Fla. Rainfall was abnormally low during the study. From September 1961 through August 1962 total rainfall at Tampa, Fla.. was 95.3 cm., 35.7 cm. below the climatological normal. Half of this amount (47.7 cm.) fell during the summer. Mean monthly air temperatures at Tampa from September 1961 through August 1962 varied from 15.8° C. in January to 28.3° C. in July. Seasonally, mean air temperatures were 23.3° C. (fall), 16.7° C. (winter), 21.9° C. (spring), and 27.4° C. (summer). ZOOPLANKTON VOLUMES In 267 plankton tows, the volume of zoo- plankton per tow ranged from < 0.5 (consider- ed as 0.25 ml. in all statistical treatments) to 92.0 ml. and averaged 7.0 ml. per sample. The greatest concentrations of macrozooplankton were in upper Tampa Bay, central Tampa Bay, lower Hillsborough Bay, central Old Tampa Bay, and 6.5 km. 3-i 2 nautical miles) offshore (fig. 3. The abundance and composition of zooplank- ton varied widely by season and location. Twenty-five percent of the total volume was collected in the fall, 5 percent in the winter, 18 percent in the spring, and 52 percent in the summer (values adjusted for different num- bers of tows per season). Coefllicients of variation in zooplankton volume were calculat- ed for each station to compare the areal vari- ability of volumes (table 4). These coeflRcients Table 4.- -Mean zooplankton volumes and the coefficients of variation of individual zooplankton volumes taken in Tampa Bau and the adjacent Gulf of Mexico. September 1H61 through Augu.'st 1962 Seasons Stations 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fall ml. 3.1 0.7 1.1 4.5 2.6 <0.5 20.0 170 ml. 10.4 10.1 3.8 7.4 8.3 <0.5 35.0 120 ml. 9.1 2.4 2.4 7.8 5.6 <0.5 26.0 130 ml. 1.3 1.2 4.8 6.0 3.3 <0.5 21.5 140 ml. 2.2 0.7 3.8 2.4 2.3 Slo 14 Figure 3. — Monthly plankton volumes for the surface waters of Tampa Bay and the adjacent Gulf of Mexico, September 19G1 through August 1962. Plankton volume is expressed as milliliters per 35 m.' of sea water. 214 U.S. FISH AND WILIFE SERVICE varied from 100 to 190 percent. Greatest varia- tions were at Hillsborough Bay and 18.5 km. offshore; minimum variations were in Terra Ceia Bay and in upper Tampa Bay. CONSTITUENTS OF ZOOPLANKTON The zooplankton consisted of holoplankton (53.5 percent), meroplankton (46.2 percent), and hypoplankton (0.3 percent). Most (87 per- cent) of the zooplankters in these categories were decapod crustaceans. Sixteen species, 24 genera, 30 families, and 21 taxonomic divisions higher than family were identified. On the basis of abundance and frequency of occurrence the plankton is treated in three groups: major plankton; less abundant but frequently occurring and widely distributed organisms ; and forms caught rarely. MAJOR PLANKTON Lucifer faxoni, larval porcellanids, and lar- val brachyurans, each of which accounted for 10 percent or more of the total number of organisms, were classified as major plankters. Collectively these taxa represented 83.5 percent of the zooplankton. L. faxoni constituted 45.6 percent of the total number of zooplankters (table 5). It was the dominant zooplankter in Tampa Bay and was the only sergestid found. Most of the speci- mens were in the mastigopus phase, although protozoea and acanthosoma types were seen. It was collected at all stations and taken in num- bers up to 1,051 per m.^*; 52 percent of L. faxoni were collected from the upper and cen- tral areas of Tampa Bay. This species was the most numerous organism in the fall, winter, and summer (fig. 4). As a result of its large size and numbers, L. faxoni accounted for 18.5 percent of plankton biomass. The monthly peaks in its displacement volume corresponded generally with the monthly peaks in the total volume of plankton. Porcellanid larvae (zoea and megalops stages) formed the second most abundant group of organisms. They accounted for 27.4 percent of the total number of zooplankters and were collected in numbers up to 2,634 per m.3 They were most numerous in upper Tampa Bay and lower Hillsborough Bay and during 1 ■s i FALL WINTER SPRING SUMMER I2M0NTHS SEPI 1961 — AUG 19BZ ■ Lucifer taioni 2 Porcellanid larvae [^ Brachyuran larvae ■ Sagitia hrspida § CopepDds Figure 4. — Zooplankton taxa from Tampa Bay and the adjacent Gulf of Mexico that accounted for 5 percent or more of the organisms found during the periods shown, September 1961 through August 1962. the spring were the dominant organism in the area of investigation. Except for Dromiidae, larval brachyurans were not identified to family and were classed only as zoea or megalops. Collectively they constituted 10.5 percent of the total number of zooplankters, and were the third-most-abun- dant taxon. Zoea and megalops were collected in numbers up to 251 per m.' and 51 m.^ re- spectively. They were found at every station but appeared most abundantly in upper Tampa Bay and least abundantly in upper Old Tampa Bay. During the winter, megalops were absent at all stations in the upper portion of the area (stations 9 through 14). In another and concurrent study, Dragovich and Kelly (1964) noted 2 species of adult Por- cellanidae (Petrolisthes galathinus and P. armatus), 23 species of adult brachyurans (many of which were gravid), and a large number of juvenile portunids. Of the com- MACRO-ZOOPLANKTON IN TAMPA BAY AND ADJACENT GULF 215 Table 5. — Frequency of occurrence and ahumlnnce (number per cubic meter in parenthef:e.i) of major zooplanklern accounting for 10 percent or more of the total number of organisms collected in Tampa liaij and the adjacent Clulf of Mexico, September lUOl through August 1962 Frequency of occurrence and abundance per cubic meter (in parentheses)' Taxon and seasons Stations 1 2 3 4 5 6 7 8 9 Luci/er faxoni: Fall No. ([4) (1.1) 4 (4.4) 6 (2.4) 17 (2.3) m. 4 (107.8) 4 (145.1) 4 (55.6) 5 (75.2) 17 (93.6) No. 4 (56.5) 4 (14.8) 4 (17.2) 6 (23.7) 18 (27.6) No. 2 (3.7) 3 (11.5) 4 (14.8) 6 (39.7) 15 (20.8) No. 5 (21.0) 5 (1.8) 6 (16.7) 6 (51.7) 22 (22.9) No. 4 (15.6) 6 (2.4) 4 (1.5) 6 (19.2) 20 (9.8) No. 4 (50.9) 6 (7.8) 5 (61.8) (75.8) 21 (49.0) No. 4 (29.2) 4 (24.2) 4 (15.2) 5 (10.0) 17 (19.1) No. Winter (343.0) Spring (44.6) Summer . (97.6) 12 months (273.4) I'orcellanidae; (197.7) Fall Winter 4 (1.4) 1 (0.1) 4 (1.2) 4 (0.3) 13 (0.7) 4 (81.4) 3 (4.1) 4 (6.4) 5 (5.9) 16 (22.4) 3 (6.0) 4 (1.2) 4 (24.2) 6 (72.5) 17 (31.2) 3 (0.8) 2 (0.8) 4 (3.8) 6 (4.2) 15 (2.7) 3 (5.5) 5 (0.9) 6 (14.4) 6 (4.9) 20 (9.1) 3 (0.5) 6 (1.3) 3 (14.4) 5 (3.4) 17 (4.7) 4 (4.6) 4 (0.4) 5 (17.9) 5 (3.8) 18 (6.8) 4 (3.5) 3 (1.2) 4 (30.2) S (25.4) 16 (15.7) 4 (8.8) Spring _ (0.2) Summer (85.0) (6.0) Brachynra: (28.8) Fall 4 (2.1) 3 (0.4) 4 (3.0) 6 (3.3) 17 (2.3) 4 (34.6) 3 (3.1) 3 (8.2) 6 (19.9) 16 (16.9) 4 (3.9) 4 (1.2) 4 (8.2) 6 (19.5) 18 (9.5) 3 (4.7) 2 (1.9) (7.2) (8.8) 15 (6.1) 5 (23.8) 3 (0.6) 6 (5.1) 6 (4.2) 20 (7.8) 4 (5.7) 5 (0.8) 2 (1.8) 6 (15.9) 17 (6.2) 5 (38.2) 4 (1.0) 5 (8.2) 6 (7.5) 20 (12.6) 4 (39.9) 4 (1.1) 4 (30.1) 5 (6.8) 17 (18.7) Winter (17.0) Spring (0.3) (35.0) (11.2) (17. r,) Frequency )f occurrence and abundance per cubic meter (in parentheses)' Percentage of total number of organisms collected Taxon and seasons Stations All stations Maximum abundance 10 11 12 13 14 Lucifer faxoni: Fall No. 4 (315.2) 2 (4.8) 5 (167.4) 6 (492.5) 17 (253.8) No. 4 (38.6) 3 (1.0) 5 (11.3) 6 (107.3) 18 (43.0) No. 2 (0.3) (0) 1 (0.2) 3 (0.4) 6 (0.2) No. 4 (40.5) 3 (0.8) 5 (174.6) 3 (97.2) 15 (83.9) No. 4 (138.6) 2 (1.1) 4 (7.0) 3 (8.0) 13 (36.8) A^o. 53 (80.2) 48 (15.8) 60 (47.7) 72 (94.3) 233 (60.4) % 61.6 64.3 23.2 58.1 45.6 No./m.' Winter - Spring Summer 12 months Porcellanidae: Fall 4 (9.6) 2 (0.3) 5 (386.0) 6 (128.5) 17 (137.0) 1 (0.1) 3 (0.3) 5 (8.2) 5 (7.7) 14 (4.4) 1 (0.1) (0) 4 (1.1) 4 (0.8) 9 (0.4) 4 (3.9) 3 (0.4) 4 (774.6) 4 (104.2) 15 (253.4) 4 (7.9) 2 (0.4) 5 (44.6) 4 (14.6) 15 (18.5) 46 (9.2) 39 (0.8) 62 (105.8) 69 (26.6) 216 7.1 3.3 51.5 16.3 27.4 283 14 2634 471 Spring Summer ._ Brachvura: Fall 4 (65.4) 3 (0.4) 5 (109.8) 6 (61.3) 18 (59.0) 3 (26.2) 3 (0.3) 5 (18.1) 6 (28.1) 17 (18.3) 1 (0.2) (0) 2 (4.7) 3 (2.0) 6 (1.6) 4 (17.9) 2 (0.4) 5 (17.4) 2 (4.5) 14 (10.5) 4 (23.4) 3 (0.4) 4 (1.6) 3 (3.5) 13 (6.9) 40 (0.8) 58 (18.6) 72 (15.0) 223 (13.9) 16.6 3.3 9.1 9.2 10.5 202 11 2.i6 2112 Winter Spring Summer 12 months I Sec table 1 for number of samples collected. 216 U.S. FISH AND WILIFE SERVICE mercially important species they caught Calli- nectes sapidus and Menippe mercenana. LESS ABUNDANT BUT WIDELY DISTRIBUTED AND FREQUENTLY OCCURRING ORGANISMS This group of organisms consisted of taxa that occurred in 10 percent or more of the samples (table 6). Collectively they accounted for 14 percent of the total number of zooplank- ters. Most taxa in this category appeared dur- ing every season. Copepods were the fourth-most-numerous group collected. They occurred in highest num- bers in the spring, and were the third-most- abundant taxon in the winter (fig. 4). Because of the coarse mesh of the plankton net, only larger specimens were retained consistently. Some of the nauplius and copepodite stages and smaller adults were held in the net only when trapped among larger plankton and detritus. Labidocera aestiva appeared to be the dominant form. The caligoids formed only a small part (1.8 percent) of this group. Caridean shrimp constituted 2.4 percent of the zooplankters. Most of the specimens were advanced postlarvae and were classified only to family. Identified palaemonids were represent- ed by the subfamilies Palaemoninae and Pon- tiniinae and the genera Palaemonetes and Peri- climines; alpheids were represented by Alpheus and Synalpheus and hippolytides by Tozeuma, Hippohjsmata, and Latreutes fucorum. Toze- uma ssp., found in stages from mysis to adult, accounted for 64 percent of the hippolytides. Hippohjsmata sp. appeared only as advanced postlarvae and Latreutes fucorum only as adults. Thalassinids were mostly advanced post- larval stages. Larvae of Upogebia sp. and Callianassa sp. also appeared in the samples. Some of these larvae possibly were Upogebia affinis and Callianassa atlantica, for both spe- cies are found in Tampa Bay. Larval stomatopods were collected at every station during the summer. Antizoea, pseu- dozoea, erichtus, and alima types were in most of the samples. Possibly many of the larvae were Squilla empusa, a prominent organism in Tampa Bay (Dragovich and Kelly, 1964). Twelve percent of the amphipods belonged to the suborder Caprellidea; the remainder be- longed to one of the suborders, Gammaridea or Hyperiidea. Table ^.—Frequency of occurrence oj zooplankters found in 10 percent or more of the samples from Tampa Bay and the adjacent Uulf of Mexico (excluding the three most abundant forms shown in Table 5), September 1961 through August 1962 Taion (Number of tows shown in parentheses] ■ Annelida: Terebellidae. No. 47 Spionidae 42 Arthropoda: Copepoda.,. __ 177 Palaemonidae 158 Alpheidae... . 116 Stomatopoda Hippolytidae.- 85 Thalassinidea 79 Amphipoda... 48 Isopoda Penaeidae.. .. 26 Chaetognatha; Sagitta hispida 133 Chordata: Fish eggs 85 Appendiculariidae 83 Sciaenidae. . 42 Syngnatheidae 28 Frequency of occurrence ' During 12 months (267) F (59) No. 24 19 Season » W (65) JVo. 4 11 42 30 13 9 10 12 17 13 1 31 11 21 13 2 7 15 4 ' Fall, winter, spring, and summer. ' Immature Sagitta less than 5 mm long. (67) JVo. 12 8 45 52 36 36 25 24 15 7 I S (76) JVo. 25 14 52 41 36 32 29 24 9 8 20 40 43 2S 25 28 12 2 10 Stations 1 (18) JVo. 1 2 13 16 6 2 2 (18) 3 (18) JVo. JVo. 3 3 4 4 (17) JVo. 5 (23) JVo. 9 3 16 20 6 4 14 9 12 2 2 10 II 6 (22) 7 (23) JVo. 4 2 12 13 5 2 6 5 6 2 6 8 7 6 4 5 4 JVo. 9 13 11 5 8 2 1 2 1 8 (17) JVo. 1 11 15 17 13 6 11 14 6 6 1 13 9 7 14 4 5 2 2 (17) JVo. 10 (20) JVo. 4 4 19 11 11 13 9 7 5 1 2 11 10 10 4 7 4 2 6 11 (20) JVo. 6 7 16 14 8 7 7 13 4 12 (20) JVo. 13 (17) JVo. 4 4 14 11 5 2 3 5 1 2 1 3 5 7 6 1 JVo. 10 3 3 2 I 4 4 2 5 1 Percent- age of total number of or ganisms collected % 1.8 0.4 2.7 1.3 0.6 0.4 0.5 0.8 0.1 <0.1 0.2 1.2 1.2 1.0 1.3 0.4 0.1 0.1 <0.1 Maxi- mum abun- dance JVo./m." 128.0 46.0 188.0 37.0 II. 37.0 23.0 28.0 3.0 I.O 34.0 34.0 97.0 177.0 37.0 23.0 1.0 3.0 0.5 MACRO-ZOOPLANKTON IN TAMPA BAY AND ADJACENT GULF 217 Most of the isopods were free-swimming cymothoids and were grouped in the genus Aegathoa. Penaeids were represented by small numbers of larvae of Sicyonia spp., Traclujpeneus spp., and Penaeus duoraruvx; Sicyonia (mainly mysis I and mysis III stages) constituted 36 percent of this family. They were restricted to the offshore area and lower areas of Tampa Bay (stations 1-4) . Only two samples con- tained Trachypeneus larvae. The pink shrimp, P. dnorarum, contributed 16 percent of the total penaeids. Postlarvae III stages of P. duo- rarum appeared most frequently; only occas- ional postlarvae I and II and mysis III were taken. These larvae were most abundant in the summer and were collected primarily in Boca Ciega Bay (station 6) and the immediately adjacent Gulf waters. Our observation of the temporal occurrence of larval stages of pink shrimp in Tampa Bay agrees generally with the findings of Eldred et al. (1965). Appendicularia spp. and Oikopleura spp. were common appendiculariids and were found at most of the sampling locations. A number of eggs and larval fish were col- lected. Many of the fishes were identified as commercially important species. The role of Tampa Bay in the production of species im- portant in Gulf fisheries was discussed by Sykes and Finucane (1965). Fish eggs were taken most frequently 18.5 km. offshore, but were most abundant at Eg- mont Key where 54 percent of the total number were collected. They were not identified. Larval fish accounted for 0.8 percent of the total number of zooplankters. All engraulids were identified as Anchoa spp. Identified sciaenids were Cynoscion spp. and Leiostomus xanthurus. Larvae of L. xanthtirus, 6 to 15 mm. long, were taken from late fall through early spring in Boca Ciega Bay and 6.5 km. offshore. Cynoscion spp. appeared infrequently during the spring and summer at most of the bay stations but were not found in Hills- borough Bay or offshore. Seventy-seven per- cent of the clupeids (3 to 20 mm. long) were identified as Brcvoortia. All syngnathid larvae (5 to 44 mm. long) were of the genus Syngna- thus. Chaetognaths made up 3 percent of the total number of zooplankters. All undamaged speci- mens more than 5 mm. long were identified as Sagitta. S. hispida was the only chaetognath found throughout the area of investigation. It was plentiful in all seasons and was the second most abundant taxon during the winter (fig. 4). The broad dispersal and numerical abun- dance of immature Sagitta less than 5 mm. long suggest that Sagitta breed both in Tampa Bay and the adjacent offshore waters. The smallest chaetognath was 2.5 mm. long, but it is likely that smaller ones escaped through the net. Polychaete larvae made up 2.4 percent of the total number of zooplankters. Terebellids (0.4 to 4 mm. long) were numerous in samples that contained a high proportion of Bellerochea malleus. The gut always contained large quan- tities of chlorophyll. None was identified to genus. Spionids (0.4 to 4 mm. long) were col- lected at 11 to the 14 sampling locations, but the genera Polydora and Prionospio were col- lected only off Egmont Key, in Boca Ciega Bay (station 6), and in Terra Ceia Bay. FORMS RARELY CAUGHT This group of organisms consisted of taxa which were in less than 10 percent of the samples (table 7). Only 10 of these taxa ac- counted for 0.1 percent or more of the total number of zooplankters, though many of them (e.g. pagurids, mollusks, and echinoderms) are common as adults of this area. The paucity of planktonic stages in this study may be ascribed partially to the large mesh of the collecting net and to the fact that only surface samples were taken. The areal distribution of most of the plank- ters in this group was limited. Most of the cladocerans (66 percent), cirripedians (61 percent), and lancelets (60 percent) were col- lected in one sample taken during May from lower Hillsborough Bay. Fifty-three percent of the pagurids were taken in August in a single sample from upper Tampa Bay. Forty-two per- cent of the larval blennies were collected from the .same area ; they were present throughout the year but were most abundant in September. Sagitta helenae and S. enflata occurred fre- 218 U.S. FISH AND WILIFE SERVICE T.\BLE 7.- -Frequency of occurrence of zooplankters found in less than 10 percent of the samples fron Gulf of Mexico, September 1961 through August 1962 [Number of tows shown in parentheses) Tampa Bay and the adjacent Frequency of occurrence ' Percent- age of total number of or- ganisms collected Taxon During 12 months (267) Season i Stations Maxi- mum abun- F (59) W (65) S (67) s (76) 1 (18) 2 (18) 3 (18) 4 (17) 5 (23) 6 (22) 7 (23) 8 (17) 9 (17) 10 (20) n (20) 12 (20) 13 (17) 14 (17) dance Aschelminthes: No. 2 17 1 9 4 4 3 2 1 5 21 17 7 4 4 3 3 2 2 2 2 1 24 13 8 1 1 25 15 8 8 8 I 4 2 2 I 38 No. No. No. 2 1 No. No. No. No. No. No. 1 1 No. No. No. No. No. 1 iVo. No. No. No. ■ I ** ».**, V FALL 1963 '.' ¥.'.'. ¥•' . 'i.'SJ?.' • oceanochaphic cast • batmttmermograph lowering • surface bucket sample 2J 26 25 24 2122 _ • ■•••• • •• • } POINT GRENViLLE n s 20 15 14 i: ; COLUMBIA Figure 1.— Locations of oceanographic stations, RV George B. Kelez, April 30, to May 17 and October 23 to November 24, 1963. (The 183- and 1,829-ni. depth contours are shown.) 224 U.S. FISH AND WILDLIFE SERVICE The cause of the Davidson Current is not clearly understood. Off the coast of Oregon it appears to result from local wind stress, but direct measurements during October 1958 and January 1959 (Reid and Schwartzlose, 1963) indicate that the driving force of this current is not local winds ; it may be a surface manifesta- tion of a deeper northward-flowing countercur- rent that develops when winds weaken season- ally (Sverdrup, Johnson, and Fleming, 1942). This northward countercurrent which opposes the offshore California Current has been re- ported off central California below 200 m. throughout the year (Reid, Roden, and Wyllie, 1958). Northward flow also was reported off Washington and Oregon below 200 m. during the summers of 1955-57 and 1959 ; this report was based upon limited observations (Dodi- mead, Favorite, and Hirano, 1963). Our closely spaced observations during 1963 permit a more detailed evaluation of the size and con- tinuity of the surface Davidson Current, the subsurface California Undercurrent, and other major features of the circulation off the coasts of British Columbia and Washington. GEOSTROPHIC CURRENTS Geostrophic currents reflect the general cir- culation associated with the distribution of mass. They are calculated from an arbitrarily selected reference depth and are, therefore, relative currents. The 1,000-db. (decibar) sur- face has been used as a reference surface in the North Pacific Ocean by Reid (1961) , Dodimead et al. (1963), Budinger, Coachman, and Barnes (1964), and Favorite (1966) because they had Figure 2.— Geopotential topography, 0/1,500 m., spring 1963. (The 183- and 1,829-m. depth con- tours are shown.) GEOSTROPHIC CIRCULATION 225 sufficient data only to 1,000 m. Bennett (1959), using a deeper reference level, obtained greater surface velocities for the Gulf of Alaska. In the absence of a known depth of no motion, the deepest level compatible with all the data (1,500 db.) was selected for the reference sur- face in this study. Other assumptions that limit the accuracy of geostrophic currents are: (1) synoptic data, (2) unaccelerated flow, (3) lack of internal wave or tidal influence, and (4) absence of friction. Caution must be used in the estimation of the surface velocity from geostrophic currents alone. The Ekman Current (Sverdrup et al., 1942) caused by local, variable wind stress must be added to the geostrophic current, for it is reasonable to assume in the absence of di- rect measurements of current that the Ekman velocities at the surface may exceed the surface geostrophic velocities. Examination of trans- port values computed from mean monthly pres- sure charts (FofonoiT and Ross, 1961) suggests that Ekman velocities of 3 to 10 cm. /sec. gen- erally toward the southeast may be expected in this coastal area during spring and fall. Al- though the short-term Ekman Currents, aver- aged on a daily basis, may be even greater, they are negligible below 200 m. They were neglect- ed in this discussion which is concerned pri- marily with the main portion of the water column below 200 m. The data from the cruises of the Kelez per- mit construction of the first geostrophic cur- rent charts off the Washington coast from a reference level of 1,500 db. Relative currents flow along contour lines of equal geopotential r" M|li Tl II 1. III z 20 15 10 y \ lij Ul h- a: CDO oz zo 5" 20 15 10 5 \ < Q. III Mill 11 ll 11 1 1 1 1 1 in Figure .3.— Geopotential topography, 0/1,500 db., fall lOfiS. (The 183- and 1,829-m. depth contours are shown.) 226 U.S. FISH AND WILDLIFE SERVICE depth with speed proportional to the gradient across them. Broken lines (figs. 2-5) represent currents in water shallower than the reference depths, which are calculated by the method used by Bennett (1959). Surface Currents, to 1,500 db. Eddies complicated the pattern of surface geostrophic currents during the spring, but the predominant surface current within 185 km. of shore was generally toward the north (fig. 2). A major feature was the apparent divergence of onshore flow near southern Vancouver Is- land. North of the divergence, water from off- shore veered toward the northwest and flowed generally parallel to the coast. Northwestward velocities were 10 and 12 cm. /sec. at two locations on the northernmost line of stations. and a speed of 18 cm. /sec. occurred off the Washington coast near lat. 47° N. The latter flow turned eastward toward shore and was not evident north of lat. 47"30' N. Although large eddies were also present off the coast of Washington during fall, they were absent off Vancouver Island (fig. 3). Maximum speed off the coast of Washington was 11 cm./sec. in the anticyclonic eddy near lat. 47° N., long. 127" W. The northward flow of 10 cm. sec. off the Columbia River in the vicinity of the 1,829-m. depth contour appeared to be dissipated by eddies as it proceeded north. Off- shore water near lat. 48° N. flowed northeast- erly toward southern Vancouver Island as it had during spring, but the distinct divergence over the continental slope was absent in the fall, and most of the water appeared to flow Figure 4.— Geopotential topography, 200/1,500 m., spring 1963. (The 183- and 1,829-m. depth con- tours are shown.) GEOSTROPHIC CIRCULATION 227 northwesterly along the coast. A maximum speed of 20 cm. sec. occurred close to shore near Cape Cook. Results from additional stations along the Willapa Bay line at 30-mile intervals to Cobb Seamount indicated that just east of Cobb Sea- mount was a weak southerly flow which sug- gests a meander in the general onshore move- ment. In summary, the gross aspects of the surface geostrophic currents during spring and fall were similar. A major, recurring feature was the broad northeasterly movement of offshore water toward southern Vancouver Island ; this flow veered northwesterly generally parallel to the coastline. Most characteristic features of the circulation off Washington were the many eddies and the apparent lack of strong north- ward flow of near-shore water, the Davidson Current, across lat. 48° N. Drift bottle experi- ments during the winter of 1965, however, indicated a significant northward flow of water over the Continental Shelf off Washington and Vancouver Island. The onshore flow which restricts the northward movement of water along the coast of Washington suggests a cause for the formation of eddies. Lower Zone Currents, 200 to 1,500 db. Data collected during the spring and fall of 1963 indicate the bottom of the halocline did not extend to a depth of 200 m. in the coastal area. Geostrophic currents at 200 m., therefore, represent the movement of water which pos- sesses nearly constant properties below the halocline and is isolated from the direct in- FlGURE 5.— Geopotential topography, 200/1,500 db., fall 1963. (The 183- and 1,829-m. depth con- tours are shown.) 228 U.S. FISH AND WILDLIFE SERVICE fluence of seasonal processes (Tully and Bar- ber, 1960). During spring the directions of geostrophic flow at 200 m. and at the surface were nearly identical (fig. 4). An exception occurred near lat. 47° N., where water on the Continental Slope veered more sharply offshore than did the surface water. The deeper flow followed the 1.829-m. (1,000-fathom) depth contour and appeared to be influenced by the local bottom topography. The maximum cur- rent of 10 cm. /sec. was at 200 m., whereas the surface flow was 6 cm. /sec. This condition is contrary to the characteristic decrease in speed with depth throughout most of the Subarctic Region and occurred only in this one area off the coast of Washington. Geostrophic currents at 200 m. during fall also followed closely the direction of the sur- face currents (fig. 5) . The speed at 200 m. was generally one-half that of the surface current north of lat. 48° N., but speeds significantly greater than the surface flow were again pres- ent off Washington. The speed of the pro- nounced anticyclonic eddy just south of lat. 48° N. near the 1,829-m. depth contour was 15 cm. /sec, at least three times the speed at the surface. When Dodimead et al. (1963) showed the California Undercurrent flowing northward below 200 m., the surface water flowed south, opposing the Undercurrent. Because there was no southerly surface flow during the spring and fall of 1963, this apparent reversal did not exist ; but if the surface current was the slow- er, the Undercurrent would be evident as a relative maximum at 200 m. in the velocity profile. The vertical distribution of velocity in the upper 1,500 m. during the fall, seaward from Willapa Bay, Hoh Head, and Esperanza Inlet, showed an area within 165 to 220 km. of the Washington coast in which pronounced maxima in the velocity did occur between depths of 200 m. and 300 m. (fig. 6) . The direc- tion of flow in adjacent maxima opposed each other, apparently forming eddies. The resultant current across any line normal to the coast of Washington, although northward, was very small. Volume transport calculations indicate the magnitude of the net flow. FULL 1963 64 70 71 72 73 STATION NO ■ipoo; I30'W. 128° 126' 124" 51 34 33 32 31 5T1TI0N NO 16 17 18 19 20 15 14 13 1211 STATION NO Figure 6. — Vertical sections of geostrophic velocity (cm. /sec.) relative to 1,500 db. seaward from Willapa Bay, Hoh Head, and Esperanza Inlet, fall 1963. (Light shading indicates regions of northward flow.) VOLUME TRANSPORT Volume transports are calculated by inte- grating the geostrophic currents throughout the water column (Sverdrup et al., 1942). The volume transports indicate the resultant rela- tive flow through the selected cross-sectional area, and are, therefore, a more reliable repre- sentation of the net flow in an area than a chart of the geostrophic currents at a par- ticular depth. In the previous section on geo- strophic currents, I pointed out that surface Ekman currents which were neglected may be in the same order of magnitude as the surface geostrophic velocities. In terms of the net transport in this coastal area during spring GEOSTROPHIC CIRCULATION 229 "• COBB SEAMOUNT Figure 7.— Volume transport in lO'^.Vsec., to 1,500 m., spring 1963. (The 183- and 1,829-m. depth contours are shown.) and fall the zonal and meridional components of Ekman Transport computed by Fof onoflf and Ross (1961) appear to contribute only 0.01 x lO^m.'/sec. and thu.s may be neglected. During spring the net transport of water across each of the eight lines normal to the coast was directed toward the north and aver- aged approximately 2 x 10"m.^/sec. (fig. 7). This estimate appeared to give considerable credence to the existence of both the Davidson Current and California Undercurrent. On the other hand, the large northward flow of 5.3 X lO'm.' sec. off the coast of Washington was part of an anticyclonic eddy ; only a very weak net transport of less than 1 x 10'"'m^ sec con- tinued northward across lat. 48" N. A rela- tively large volume 6.7 x ICm.Vsec. entered the area from offshore, of which 4.1 x IC-m.V sec. apparently flowed onshore across the Con- tinental Slope where calculation of volume transport to 1,500 m. is less meaningful. The net northward transport increased to 3.7 x 10'' m.' sec. across the northernmost line and had the same direction and magnitude as that re- ported by Bennett (1959) during August 1955 for the near-shore area between lat. 50° and 55° N., just north of this study area. During fall the greatest northward transport again occurred off the northern coast of Van- couver Island; no significant change appeared in the volume of water flowing northward past Cape Cook (fig. 8). Off the Washington coast the net transport was again about 1 x 10''m.V sec, but the direction reversed across succes- 230 U.S. FISH AND WILDLIFE SERVICE Figure 8.— Volume transport in lO'm.Vsec, to 1,500 db., fall 1963. (The 183- and 1,829-m. depth contours are shown.) sive lines normal to shore. The onshore move- ment evident during spring had reversed to 0.4 X lO^m.V'sec. in the offshore part of the area and 2.7 x lO'^m.' sec. Seaward across the Con- tinental Slope. The net transport across the line from Willapa Bay to Cobb Seamourit was less than 1 X 10''m.^/sec. Compared with a trans- port of 14 X lO^m.' sec. for the Gulf of Alaska (Bennett, 1959), these results indicate a lack of significant net transport along the coast of Washington within 500 km. of shore. The surface Davidson Current and the sub- surface California Undercurrent reported by previous authors, therefore, did not contribute more than 1 x lO'm.'/sec. to the net north- ward transport of water along the coast of Washington. Although the total volume of transport was the same in spring and fall, an increase in the California Undercurrent was implied by the distribution of properties at and below 200 m. south of lat. 48° N. DISTRIBUTION OF PROPERTIES Although the most common method of de- termining oceanic circulation is the calculation of geostrophic currents and transports from observed values of temperature and salinity at standard depths, deductions concerning flow can also be made directly from the ob- served distributions of these water properties. Reasonable confidence may be placed in the interpretation of the circulation, particularly when the direction of flow suggested from the distribution of properties supports the calcu- GEOSTROPHIC CIRCULATION 231 lated geostrophic currents. The following items are discussed : features of the distributions of salinity, temperature, and dissolved oxygen ; changes in these properties near the bottom along the continental terrace ; and water mass movements implied by temperature-salinity relationships. SALINITY Throughout most of the Subarctic Region, the salinity structure consists of three distinct permanent zones: (1) an isohaline upper zone, which extends from the surface to about 100 m. ; (2) a halocline, in which the salinity in- creases about l%o between 100 and 200 m.; and (3) a lower zone, in which the salinity gradually increases with depth. The mechan- ism for the maintenance of this structure was discussed by Dodimead et al. (1963). Perhaps the most striking changes in the distribution of properties within the coastal areas occur in the salinity distribution in the upper zone and are due to the intrusions of fresh-water runoff from coastal rivers. Various authors have attempted to distinguish oceanic and coastal water on the basis of the salinity distribution near the surface. In the North Pacific Ocean, Doe (1955) used the 32.5%o isohaline as the boundary between offshore and coastal water masses in the upper zone; Dodi- mead et al. (1963) defined the extent of a coastal domain by the 32.4%. isohaline; and Budinger et al. (1964) suggested the Columbia River effluent could be traced by salinities less than 32.5°oo. Good agreement has thus been reached concerning a definable boundary be- FlGURE 9.— Surface salinity (%.), spring 1963. (The 18.3- and 1,829-m. depth contours are shown.) 232 U.S. FISH AND WILDLIFE SERVICE tween oceanic and coastal water, and the ef- fects of dilution have been shown to extend over several hundred kilometers from shore (Favorite, 1961). Sharp gradients, or fronts, found closer to shore, however, are more in- teresting and much more complex. The distribution of surface salinity during spring showed that the 32.5%o isohaline ap- proached within 160 km. of shore off the coast of Washington, but the most significant fea- ture was the front associated with the 32.0%o isohaline (fig. 9). The controversy regarding the precise definition of the term front in oceanographic usage has been discussed by Griffiths (1965). Front is used here in the sense that Cromwell and Reid (1956) defined the term, "... a band along the sea surface across which the density changes abruptly." The change of surface temperature near the front was not appreciable compared with the salinity change ; thus the density change at the front was dominated by the relatively sharp decrease in surface salinity. Although no particular isohaline appeared to define the exact extent of the front throughout the area, gross changes in the position of the front may be seen by tracing, the extent of the 32.0%o isohaline. The largest gradient of surface salinity was about 80 km. from shore near lat. 46^ N. where the front was apparently being maintained by effluent less was 22%o from the Columbia River. The maximum seaward extent of the 32.0%o isohaline was 112 km. near lat. 47° N. ; at lat. 48° N. it had decreased to 64 km., and all along the coast of Vancouver Island it was confined to within 48 km. of shore. It is not clear whether the large tongue of dilute water off the central coast of Wash- ington was a remnant of water from the Co- lumbia River which had proceeded north along the coast during the winter or if it came di- rectly from the Strait of Juan de Fuca. A patch of relatively high salinity water (> 32.2%o) about 11 km. seaward of Hoh Head indicated an area of local upwelling. A vertical section of salinity extending sea- ward from Willapa Bay illustrates the major changes in salinity with depth and distance from shore during spring (fig. 10). Dilute water of less than 32.0%o in which the iso- 43 STATION NO O -ZpOO Figure 10. — Vertical sections of salinity (%„), to 200 m. and to 2,500 m., along Willapa Bay line, spring 1963. lines were closely spaced, appeared to protrude seaward in the form of a tongue. Although the 32.0%<. isohaline that occurred at the leading edge of this tongue underwent large fluctuations in its seaward extent along the coast, the nearly constant depth of the 32.0%o isohaline near shore shows that the GEOSTROPHIC CIRCULATION 233 major effect of the dilution off the coast of Washinglon was limited to the upper 30 to 40 m. Offshore, the three vertical zones charac- teristic of the Subarctic Region were present; although the boundaries between zones gener- ally rose toward shore, the halocline and lower zone could be traced continuously inshore until they ended at the continental terrace. Tully and Barber (1960) suggested that across the boundary of the halocline-lower zone 33.8 ± 0.1%=, only upward transfer of water existed ; thus the depth of this surface forms the ultimate limit of downward transfer of water from the surface. Changes in prop- erties below this surface are, therefore, pri- marily due to advection, not directly influenced by seasonal changes near the surface. The 33.8%o surface was about 170 m. deep off- shore but rose to about 130 m. near the 183-m. depth contour. A horizontal section of salinity at 200 m. during spring showed uniform values of salin- ity just below the halocline; the range was from about 33.86%o to 33.93% (fig. 11) — a marked contrast to the i-ange of surface salinity, 22.0%o to 32.5%o. A second im- portant feature was the tongue of relatively high salinity (> 33.92%o) which appeared to point northward near the 1,829-m. depth contour off Willapa Bay. Although many of the features in the coastal area during the fall were similar to those during the preceding spring, important changes occurred near the surface between spring and fall. The salinity front was consistently nearer shore, 48 km. to 64 km. (fig. 12). The tongue of dilute water 234 isrw. oo* iM' i2a" liT izi' i25* Figure 11. — Salinity (%„) at 200 m., spring 1963. (The 18.3- and 1,829-m. depth contours are shown.) U.S. FISH AND WILDLIFE SERVICE Figure 12.-Surface salinity (%o), fall 1963. (The 183- and 1,829-m. depth contours are shown.) which had protruded twice as far seaward from the coast of Washington during the spring was absent in the fall. The 32.5%o isohaline had shifted seaward from 160 km. to 400 km. by fall. Although the boundary between the halo- cline and lower zone fluctuated over a greater depth range during the fall, the major struc- tural zones were again present and continu- ously defined along each section normal to shore. At 200 m. the maximum salinity in- creased to 33.96%o between Willapa Bay and lat. 48° N. (fig. 13). The small tongue of gi-eater salinity present during spring had en- larged to form a continuous ridge of high salinity along the coast with an axis about 140 km. from shore. This feature was more com- plex north of lat. 48° N. where the salinity decreased. TEMPERATURE In the Subarctic Pacific Region the water above the halocline begins to receive a net gain in heat in April and continues to warm into September (Dodimead et al., 1963) . During the spring the surface-temperature gradient along the coast was uniform ; temperatures from Vancouver Island to the Columbia River in- creased from 9.0° C. to 13.5^ C. (fig. 14). Off the coast of Vancouver Island the surface isotherms were generally oriented northeast- southwest, normal to the shore, and showed no apparent relation to the surface-salinity front. Off the coast of Washington, however, the isotherms generally ran from north to south, parallel to shore. Their configuration agreed closely with the surface isohalines. Vertical sections of temperature during GEOSTROPHIC CIRCULATION 235 Figure 13.— Salinity (%) at 200 ni., fall 1963. (The 183- and 1,829-m. depth contours are shown.) spring, one seaward from Cape Cook and the other seaward from Willapa Bay, illustrate the changes in temperature with depth and dis- tance from shore as well as difference in tem- perature along the coast between the northern and southern parts of the area (fig. 15). Characteristically the decrease of tempera- ture with depth throughout the water column was inconsistent only within the halocline which contained sporadic inversions not in ex- cess of O.S'' C. Below the halocline the tem- perature decreased logarithmically toward the bottom. As was true with salinity, the most pro- nounced changes within the area took place in the upper layers. The dilute water near shore had a weak vertical gradient. Offshore from Cape Cook the upper 50 m. was isothermal, but toward the south, the magnitude of the sea- sonal thermocline increased between the sur- face and 30 m. Within the halocline off the Washington coa.st the i.sotherms rose toward shore over the Continental Shelf, but beyond the shelf they sloped slightly downward toward shore. The temperature increase from Cape Cook to Willapa Bay extended to a depth of at least 200 m. In the lower zone, the isotherms were relatively level. Although the variations of temperature at a particular depth below the halocline were small, the tempei-ature distribu- tion at 200 m. did show an unusual feature. A tongue of cold water (<6.8° C.) extended shoreward near the middle of Vancouver Is- land, interrupting a band of warmer water 236 U.S. FISH AND WILDLIFE SERVICE Figure 14.— Surface temperature ("C), spring 1963. (The 183- and 1,829-ni. depth contours are shown.) (>7.2<^ C.) near the 1,829-m. depth contour (fig. 16). Differences between temperatures during spring and fall were most pronounced above the halocline, in the zone affected by seasonal heating. Significant changes also occurred, however, within the lower zone which is not influenced directly by the seasonal heating; changes in circulation are implied. Although the gradient of the surface tem- perature between northern Vancouver Island and the Columbia River was the same during each season, temperatures were generally 1.0° C. higher at each location during the fall (fig. 17). The most pronounced changes were near shore. During spring, warming was ap- preciable only near the surface in the dilute water off the coast of Washington, but in the fall the water was distinctly warmer over the entire Continental Shelf than offshore. The resulting temperature distribution shows a tongue of warm water extending northward along the coast and the maximum temperature along any line normal to shore near the edge of the Continental Shelf. Comparison of fall and spring conditions along the same two vertical sections indicated many differences and similarities with depth and distance from shore. The most pronounced differences were in the upper 70 m. In contrast to the gradual decrease of temperature sea- ward during spring was the presence of two maxima during the fall — one at the surface near the edge of the Continental Shelf and the GEOSTROPHIC CIRCULATION 237 5 6 STATION NO J : Ik V\i ^U : J' 8.6 — ^~^*°-^ . 80—' • ^^i\^v — ■'! ' ""~\ V^ -76^' ^ -\^ ^\v 7, ^ ^T— " N \l\\ 1 J. \ ^— — ^ \^// ■"■■. . 43 STATION NO I I o 4 5 6 STATION NO, apoo Figure 15. — Vertical sections of temperature ("O, to 200 m. and to 2,.500 ni., along Cape Cook and Willapa Bay lines, spring 1963. other at depth between 20 and 50 m. near shore. Apparently both maxima resulted from surface cooling and vertical mixing which had affected only the dilute water in the upper 20 m. Seaward of the .salinity front, the upper 50 m. of the water column wa.s isothermal off Cape Cook during both seasons. Southward of Cape Cook the magnitude of the thermocline again increased but by fall had deepened from the range to 30 m. to between 50 and 70 m. in the top of the halocline. Temperature in- versions were more frequent in the fall within the halocline, and the near-shore isotherms showed a slight depression or convergence. Vertical sections normal to .shore during fall as in the spring showed the logarithmic de- crease of temperature with depth and the level isotherms in the lower zone extending seaward from the Continental Slope. Despite this general uniformity in tempera- ture structure, minor changes did occur in the temperature distribution at 200 m. (fig. 18). The tongue of warmer water (> 7.2° C.) noted during spring changed: it became continuous along the coast, occupied a much larger area, and had a greater maximum temperature dur- 238 U.S. FISH AND WILDLIFE SERVICE Figure 16.— Temperature ("C.) at 200 m., spring 1963. (The 183- and 1,829-ni. depth contours are shown.) ing fall — 7.6° C. Although this increase in temperature between spring and fall was slight, it may be significant compared with the small range of values at 200 m. The associa- tion of the warm water near the coast of north- ern Vancouver Island with the lowest salinity values implied a local convergence. Offshore the cold water was also of low salinity. The warm water off the coast of Washington near the 1,829-m. depth contour, however, was as- sociated with the high-salinity ridge; a sig- nificant change in water mass is indicated. DISSOLVED OXYGEN The distribution of dissolved oxygen during the spring was obtained only over the con- tinental terrace between the depths of 55 m. and 1,829 m. As with the distributions of tem- perature and salinity, the sharpest vertical gradient occurred within the halocline. Below the saturated or mixed layer, about 50 m. deep, values decreased sharply to about 300 m. Below 300 m., concentrations decreased gradually to a minimum near 900 m., below which values gradually increased toward the bottom. Samples obtained at each station during the fall permit comparison of conditions near shore and offshore. The vertical section off Willapa Bay shows the complex distribution of dissolved oxygen in the upper 300 m. (fig. 19). Deeper isolines were relatively level and the oxygen minimum near 900 m. extended off- shore without significant change in depth. The isolines within the upper 300 m. usually fol- GEOSTROPHIC CIRCULATION 239 Figure 17.— Surface temperature ("C.) , fall 1963. (The 183- and 1,829-m. depth contours are shown.) lowed the configuration of the isohalines or isotherms, but were inclined generally upward toward shore. The rise reflected lower oxygen values near shore during spring and fall; minor inversions of dissolved oxygen were more frequent during fall, just below the bot- tom of the halocline near 200 m. A plot of dis- solved oxygen at 200 m. during the fall (fig. 20) showed that this band of low oxygen concentration (< .20 mg. at./l.) was continu- ous along the entire coast over the Continental Slope and closely followed the high-salinitv ridge (fig. 13). Comparison of fall conditions with those over the Continental Slope during the preceding spring indicated that oxygen values at 200 m. had decreased on the average, by about 0.05 mg.at. 1. If we assume that the seasonal change in biological utilization of dissolved oxygen was negligible, this decrease in dissolv- ed oxygen concentration corroborates the change in water-mass characteristics between spring and fall previously indicated by the increase in temperature and salinity at 200 m. off the coast of Washington. CONDITIONS NEAR THE BOTTOM To determine changes in salinity, tempera- ture, and dissolved-oxygen concentrations at a particular depth close to the sea floor along the continental terrace, samples were obtained as near the bottom as feasible — 55 m., 183 m., 914 m., and 1,829 m. along each of the nine lines normal to shore. At the 55-m. and 183-m. sta- 240 U.S. FISH AND WILDLIFE SERVICE 51- N Figure 18. — Temperature ("C), at 200 m., fall 1963. (The 183- and 1,829-m. depth contours are shown.) tions, Nansen bottles were tripped 5 m. from the bottom by two methods. In the first method, the Nansen bottle is placed 5 m. above a weight suspended on the end of the wire, and the bottle is tripped by a messenger when a change in wire tension indicates the weight is striking the bottom. In the second method, a Nansen bottle attached to a tripping mechan- ism is reversed when a weight suspended 5 m. below the device strikes the bottom. Although agreement of results from both methods was good, values obtained by the first method were used for most stations. Because of the limited depth range of the vessel's echo sounder (about 550 m.) and the inability to detect the bottom by wire tension at great depths, the locations of the 914-m. and 1,829-m. stations were de- termined from charted depths ; thus, the in- terval between the deepest bottle and the bot- tom depended upon the accuracy of the charted soundings and the vessel's position. The spring values of salinity and tempera- ture near the bottom varied most at shallow depths along the Continental Shelf and were uniform along the Continental Slope (fig. 21). At 55 m. salinity values were uniform between '31.9%o and 32.0%o north of the Strait of Juan de Fuca, but increased off Washington. The maximum of 33.4%o was near the mouth of the Columbia River. The range of tempera- ture values at 55 m. was about 1.0° C. The minimum value occurred off the Columbia River. The maximum salinity and minimum temperature indicated that water which is GEOSTROPHIC CIRCULATION 241 Figure 19. — Vertical sections of dissolved oxygen (mg.at./I.), to 300 m. and to 3,000 ni., along Willapa Bay line, fall 1963. 242 U.S. FISH AND WILDLIFE SERVICE Figure 20— Dissolved oxygen (mg.at./I.) at 200 m., fall 1963. (The 183- and 1,829-m. depth con- tours are shown.) normally found at a greater depth had moved shoreward in this area. Conditions near the bottom were reversed from those at the surface where the salinity was at a minimum and the temperature was at a maximum. At 183 m. the range of salinity was much smaller — between 33.67%= and 33.95%o — but the range of temperature at 183 and 55 m. was the same, 1.0= C. The salinity did not vary significantly along the coast and temperatures increased only slightly toward the south — about 0.1° C. Upwelling, therefore, was not taking place at 183 m. The range of values continued to de- crease with depth. Changes in salinity ranged ifrom 0.08%o at 914 m. to 0.07%<. at 1,829 m., and differences in temperature ranged from 0.15° C. at 914 m. to 0.10° C. at 1,829 m. These minor variations indicated no significant change in salinity or temperature near the bottom along the Continental Slope during spring between the Columbia River and Cape Cook. Values of dissolved oxygen near the bottom at 55 m. during spring were lowest off the mouth of the Columbia River (fig. 22). This situation appeared to corroborate the upwell- ing of deeper water, although biological utili- zation may have contributed to the low values. Oxygen, like salinity and temperature, fol- lowed no significant trend at a particular depth along the Continental Slope. At 55 m., values of salinity, temperature, and dissolved oxygen in the fall were signifi- cantly different from those during spring. GEOSTROPHIC CIRCULATION 243 Figure 21. — Temperature (°C.) and salinity {°D near the bottom at 55, 183, 914, and 1,829 m. along the continental terrace, spring 1963. (The 183- and 1,829-m. depth contours are shown, and the values in parentheses are interpolated. ) Thus, more uniformity of salinity values — range from 31.6%o to 32.3%» — indicated absence of upwelling. By fall, temperatures at 55 m. had increased 3" C. off Vancouver Island, and 4° to 6= C. off the Washington coast. The increase was about 0.6- C. even at 183 m. DISTRIBUTION OF WATER MASS BELOW 200 METERS Analysis of distribution of temperature, salinity, and dissolved oxygen indicated sig- nificant changes in characteristics of water mass in near-shore areas and also between seasons. Water masses of different character conventionally have been defined by the tem- perature vs. salinity (T-S) curve plotted from serial oceanographic data (Sverdrup et al., 1942). All T-S curves from spring data were grossly similar ; each had a characteristic s-shape and occupied a narrow envelope. The T-S curves of stations farthest offshore were consistently displaced downward toward the left, however. Waters here were colder and less saline than near the coast. The T-S curves from 10 stations within 500 km. of shore along the Willapa Bay-Cobb Seamount line during fall illustrate this displacement (fig. 23). The heavy curves on the lower and right-hand sides are general curves that represent the extreme water masses in the North Pacific Ocean — the Subarctic and Equatorial Pacific Water Masses ; they indicate that the coastal water is a mixture of two water masses. The separation 244 U.S. FISH AND WILDLIFE SERVICE Figure 22. — Dissolved oxygen (mg.at./l.) near the bottom at 55, 183, 914, and 1,829 m. along the continental terrace, spring 1963. (The 183- and 1,829-m. depth contours are shown, and values in parentheses are interpolated.) between the offshore water mass that intrudes from the west and the coastal water mass that intrudes from the south was not distinct at all depths. Near 200 m. three distinct groups of curves existed. Stations 11 to 13 near shore had the most southern characteristics ; stations 16 to 19 offshore had the most northern charac- teristics; and stations 14, 15, and 20 in the center were intermediate between the coastal and offshore water masses. Below 400 m. the boundary between the coastal and offshore water masses was distinct and lay between stations 14 and 15, about 165 to 220 km. from shore. The study of horizontal changes in the characteristics of water masses throughout the coastal area showed that the differences in temperature were greatest on the salinity sur- face 34.0%=. Dodimead et al. (1963) sug- gested that temperatures greater than 6.0' C. on this surface defined the extent of the Cali- fornia Undercurrent Domain which appeared to originate south of lat. 35" N. Their geo- strophic calculations, however, indicated only a weak northward flow below 200 m. during 4 of the 5 summers in 1955-59. The temperature distribution on the 34.0%= salinity surface during spring and fall of 1963 showed the isolines were predominantly parallel to shore although a tongue of warm water apparently entered the area over the Continental Slope from the south. The boundary between the coastal and offshore water masses was marked by a temperature gradient on the seaward side GEOSTROPHIC CIRCULATION 245 SALINITY (%o) 10.0 33.8 __l 34.0 I 34.2 I 34.4 I 34.6 l__ STATION NUMBERS COBB SEAMOUNT '? '.^ f '.^^ '.^ "l^l^I^" W.LLAPA BAY Figure 23. — Temperature vs. salinity curves for water of salinity greater than 33.8%o Willapa Bay and Cobb Seamount, fall 1963. between of the tongue. During spring the 6.0° C. iso- therm was discontinuous off southern Van couver Island (fig. 24) where the geostrophic currents indicated onshore movement (fig. 4). The distribution of temperature greater than 6.0° C. showed a band of more southern water about 40 km. wide seaward of the Continental Shelf. The configuration of the isotherms on 34.0%o surface was generally similar to the geostrophic currents at 200 m., and the south- ern water mass was located to the right (if one faces downstream), even when the current was southbound. The greater area encompassed by the south- ern water (> 6.0° C.) off the coast of Wash- ington during the fall suggests that the north- ward flow was greater during fall (fig. 25), but the increased flow was not reflected in the net volume transport. A more quantitative approach to the de- scription of the distribution of water masses along the Pacific Coast of the United States was made by Tibby (1941) who applied the method of Sverdrup and Fleming (1941) to data obtained by the E. W. Scripps in 1939. 246 U.S. FISH AND WILDLIFE SERVICE Figure 24. — Temperature ("C.) on salinity surface 34.0%„, spring 1963. Shaded portion is above 6.00 c. (The 183- and 1,829-m. depth contours are shown.) He showed a relatively higher percentage of Equatorial Pacific Water to be present along the Pacific Coast near shore from lat. 25° N. to lat. 45° N. All vertical sections normal to shore showed a greater percentage of southern water toward the bottom and toward shore. The low percentages of southern water in the northernmost sections (between 20 and 40 per- cent) suggested increased mixing with Sub- arctic Water to the north. The percentage of Equatorial Pacific Water off the Washington coast along lat. 46°45' N. during the fall of 1963 agreed closely with Tibby's results (fig. 26). Because this location is 222 km. farther north than Tibby's most northern line, however, a slightly lower per- centage was found. Percentages were not only relatively high near the bottom and near shore, but the high percentages between 200 m. and 400 m. immediately below the halocline ex- tended offshore as far as 500 km. This situa- tion resulted in a pronounced minimum, from 10 to 20 percent, between 600 and 900 m. Val- ues increased sharply within 220 km. of shore. The depth of this minimum percentage of Equatorial Pacific Water coincided surpris- ingly well with the depth of the oxygen mini- mum. This determination of distribution of water masses suggests a change in circulation with depth; between 400 m. and 1,000 m., Sub- arctic Water may move onshore from the west ; whereas, between 200 m. and 400 m. and be- tween 1,000 m. and 1,300 m., southern water may move northward along the coast. GEOSTROPHIC CIRCULATION 247 Figure 25. — Temperature (°C.) on salinity surface 34.0%o, fall 1963. 6.0» C. (The 183- and 1,829-m. depth contours are shown.) Shaded portion is above Figure 26. — Vertical section of the percentag:e of Equa- torial Pacific Water for water greater than 33.8%o between Willapa Bay and Cobb Seamount, fall 1963. SIGNIFICANT OCEANOGRAPHIC FEATURES OF COASTAL WATER Data obtained during the spring and fall 1963 at cio.sely spaced stations along nine lines normal to shore between the Columbia River and Cape Cook, Vancouver Island, have per- mitted a description of the significant ocean- ographic features in the spring within 220 km. of the Vancouver Island and Washington coasts and changes that occurred by the suc- ceeding fall. Surface geostrophic currents, 0/1,500 db., were similar during l)oth spring and fall. Off- shore water flowed northea.sterly toward the middle of Vancouver Island and then turned toward the northwest generally parallel to the 248 U.S. FISH AND WILDLIFE SERVICE coast. Eddies off the coast of Washington were such that the northward flow over the Con- tinental Slope, perhaps the Davidson Current, did not appear to continue north of lat. 48° N. Geostrophic currents, 200/1,500 db., fol- lowed the same direction as the surface cur- rents, and showed maxima of subsurface velocity in the eddies off the Washington coast. The net volume transport, based on a 1,500 db. reference level, was 3 x lO^m.Vsec. north- ward past Cape Cook during both spring and fall. A shoreward component of total transport 6.7 X 10"m.^/sec., was present in the spring, but by fall the transport had reversed to the seaward at 0.4 x lO-'m.Vsec. Northward trans- port of 4 to 5 X lO^m.Vsec. occurred locally off the coast of Washington, but was associated with strong anticyclonic eddies which had nearly an equivalent southward transport. Although the existence of the California Un- dercurrent may be implied by the distribution of properties and supported by the direction of the geostrophic currents at 200 m., the Current did not appear to contribute more than 1 x lO^m.^/sec. to the net northward flow along the coast of Washington. The most striking and permanent feature of the distribution of properties within the area was the surface salinity front which extended to a maximum distance of 112 km. seaward from the coast of Washington during the spring, but was confined within 64 km. of the coast in the fall. Vertical sections normal to the shore showed that the major structural features of salinity, temperature, and dissolved oxygen were consistent along each of the nine lines. The distribution of properties varied considerably above 200 m. within the main halocline, thermocline, and oxycline. Below 200 m. the range of values of salinity, temperature, and dissolved oxygen at a given depth was comparatively small. The major feature below the halocline was the nearly horizontal isolines of temperature, salinity, and dissolved oxygen along each vertical section. The concentration of dissolved oxygen had a pronounced mini- mum at about 900 m. throughout the study area. Minor variations at 200 m. in the fall indicated that a ridge of high-salinity water, also associated with high temperature and low dissolved oxygen, was especially well developed along the Continental Slope of Washington. Samples obtained near the bottom confirmed the absence of any significant change in the salinity, temperature, or dissolved oxygen at a particular depth along the Continental Slope between the Columbia River and Cape Cook. Thus, the only significant variations in water properties occur in the upper 200 m. The T-S curves indicated that a water mass of high salinity and high temperature was present over the Continental Slope. Off the coast of Washington the boundary between water masses of slightly different characteristics was distinct below 400 m. between 165 and 220 km. from shore. The more southern water mass near shore occupied a greater portion of the coastal area during the fall than in the spring. Although this implied an increase in north- ward flow at depth, the California Undercur- rent, increased flow was not reflected in the net volume transport. LITERATURE CITED Bennett, E. B. 1959. Some oceanographie features of the north- east Pacific Ocean during August 1955. J. Fish. Res. Bd. Can. 16(5): 565-633. BuDiNGER, Thomas F., Lawrence K. Coachman, and Clifford A. Barnes. 1964. Columbia River effluent in the northeast Pacific Ocean, 1961, 1962: selected aspects of physical oceanography. Dep. Oceanogr. Univ., Wash., Seattle, Tech. Rep. No. 99, 78 pp. Burt, Wayne V., and Bruce Wyatt. 1964. Drift bottle observations of the Davidson Current off Oregon. In Studies on Oceanog- raphy, Tokyo, Dedicated to Prof. Hidaka: 156- 165. Univ. Wash. Press, Seattle, Wash. Cromwell, Townsend, and Joseph L. Reid, Jr. 1956. A study of oceanic fronts. Tellus 8(1): 94-101. DoDiMEAD, A. J., F. Favorite, and T. Hirano. 1963. Salmon of the North Pacific Ocean, Part II, Review of the oceanography of the Subarctic Pacific Region. Int. N. Pac. Fish. Comm., Bull. 13, 195 pp. Doe, L. a. E. 1955. Offshore waters of the Canadian Pacific Coast. J. Fish. Res. Bd. Can. 12(1): 1-34. Favorite, Felix. 1961. Surface temperature and salinity off the Washington and British Columbia coasts, Au- GEOSTROPHIC CIRCULATION 249 gust 1958 and 1959. J. Fish. Res. Bd. Can. 18(3): 311-319. (in press). The Alaskan Stream. Int. N. Pac. Fish. Comm., Bull. FoFONOFF, N. P., and C. K. Ross. 1961. Transport computations for the North Pacific Ocean 1961. Fish. Res. Bd. Can., MS. Rep. Ser. (Oceanopr. and Lininol.), No. 128, 5 pp. Griffiths, Raymond C. 1965. A study of ocean fronts off Cape San Lucas, Lower California. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 499, 54 pp. Ingraham, W. James, Jr. 1964. Distribution of physical-chemical properties and tabulations of station data, Washington and British Columbia coasts, May 1963. U.S. Fish Wildl. Serv., Data Rep. 5, 3 microfiches, 88 pp. Reid, J. L., Jr. 1961. On the geostrophic flow at the surface of the Pacific Ocean with respect to the 1,000- decibar surface. Tellus 13(4): 489-502. Reid, Joseph L., Jr., Gunnar L Roden, and John G. Wyllie. 1958. Studies of the California Current System. Calif. Coop. Oceanic Fish. Invest. Rep., July 1, 1956-January 1, 1958: 28-5G. Reid, Joseph L., Jr.. and Richard A. Schwartzlose. 1963. Direct measurements of the Davidson Cur- rent off Central California. J. Geophys. Res. 67(6) : 2491-2497. Schwartzlose, Richard A. 1963. Nearshore currents of the western United States and Baja California, as measured by drift bottles. Calif. Coop. Oceanic Fish. Invest. Rep., July 1, 1960 to June 1, 1962: 15-22. Sette, Oscar E. 1947. South Pacific fishery investigations, hi U.S. Fish. Wildl. Serv., Div. Fish. Biol., Annu. Rep. Fiscal Year 1947, pp. 60-62. Washing- ton, D. C. [Processed.] SvERDRUP, H. U., and R. H. Fleming. 1941. The waters off the coast of southern Cali- fornia, March to July, 1937. Scripps Inst. Oceanogr., Bull. 4: 261-378. SvERDRUP, H. U., Martin W. Johnson, and Richard H. Fleming. 1942. The oceans, their physics, chemistry, and general biology. Prentice-Hall, Inc., N.Y., 1087 pp. TiBBY, Richard B. 1941. The water masses off the west coast of North America. J. Mar. Res. 4(2): 112-121. TuLLY, J. P., and F. G. Barber. 1960. An estuarine analogy in the Subarctic Pacific Ocean. J. Fish. Res. Bd. Can. 17(1): 91-112. 250 U.S. FISH AND WILDLIFE SERVICE SYSTEMATICS AND BIOLOGY OF THE BONEFISH, ALBULA NEMOPTERA (FOWLER)' By Luis R. Rivas - and Stanley M. Warlen,' Fishery Biologists Bureau of Commercial Fisheries Exploratory Fishing Base, Pascagoula, Miss. 39567 ABSTRACT This study is a review of the taxonomic status of the bonefish, Albula nemoptera, formerly placed in the genus Dixonina. Reasons for synonymizing Dixonina with Albula are discussed, and it is shown that pacifica is conspecific with nemoptera. The Atlantic and Pacific populations of nemoptera are compared with each other and with the common bonefish, A. vulpes. Presumed larval and juvenile stages of nemoptera are described and compared with those of vulpes. The ecology and distribution of nemoptera and vulpes is discussed. Prior to 1911 the family Albulidae wag known from several fossil forms and one living species, Albula vulpes (Linnaeus). Fowler (1911) described the second living species, Dixonina nemoptera, from a single specimen from Hispaniola. Eight years later a second specimen was recorded by Metzelaar (1919) from Venezuela and a third, from the Pacific coast of Mexico, by Myers (1936). A drawing of a specimen from the Pacific coast of Mexico identified as "Albula vulpes," was published by Kumada and Hiyama (1937). According to Walford (1939), apparently several specimens were available to these authors. Beebe (1942), on the basis of 19 specimens from Costa Rica, proposed the name "Dixoyiina pacifica" for the Pacific coast population. The third Atlantic record (Rivas, 1952) was based on two speci- mens from Jamaica. Recently Caldwell and Caldwell (1964) recorded, tentatively as "Albula vulpes," 14 larvae and juveniles from the Atlantic coast of Panama. According to the literature, therefore, this apparently rare species of albulid was hitherto known only from four Atlantic records (7 lar- vae, 6 juveniles, and 4 adults) and the three Pacific records (21 specimens of which 14 are not traceable). > Contribution No. 65 from the Ichthyological Laboratory and Museum, Department of Biology, University of Miami. 2 Permanent address: University of Miami, Department of Biology. " Present address : BCF Biological Laboratory, Beaufort. N.C. FISHERY BULLETIN: VOLUME 66, NO. 2 Published May 1967. During its cruise No. 92, May 5 thro"gh June 17, 1964, the Bureau of Commercial Fish- eries exploratory fishing vessel Oregon col- lected 21 adult specimens of Albula nemoptera along the Atlantic coast of Colombia. Nineteen of these are available for the present study (see materials and acknowledgments) ; one was deposited at the Santa Marta Marine Laboratory, Santa Marta, Colombia, and an- other at the Bureau of Commercial Fisheries Tropical Atlantic Biological Laboratory in Miami, Fla. Additional specimens from the Atlantic and Pacific, not previously reported in the litera- ture, were located in various institutions. The fairly adequate material of A. nemop- tera, now at hand, prompted this study, par- ticularly because the most recent account of the species was based on a single specimen (Hilde- brand, 1963) and the conclusions reached therein are open to question (Berry, 1964). In due fairness to the late S. F. Hildebrand, how- ever, it should be remembered that his study was published 14 years after his death in 1949. We performed this research at the Bureau of Commercial Fisheries Exploratory Fishing and Gear Research Base, Pascagoula, Miss. MATERIALS This paper is based on 56 specimens (35 Atlantic; 21 Pacific) of A. nemoptera and 43 Atlantic specimens of A. vulpes from the 251 collections of the U.S. National Museum (US- NM), Stanford University (SU), Field Museum of Natural History, Chicago (FMNH), Los An- geles County Museum (LACM), University of California at Los Angeles (UCLA), University of Miami Institute of Marine Science (UMML) , and University of Miami Ichthyological Mu- seum (UMIM). This material is distributed as follows : A. nemoptera (Atlantic). — Colombia: about 40 km. (22 nautical miles) NW. of Punta San Bernardo, USNM 199530 (3 adults), FMNH 66796 (2 adults); about 22 km. (12 nautical miles) WNW. of Puerto Colombia, USNM 199474 (6 adults), UMIM 5926 (2 adults); about 23 km. (13 nautical miles) NE. of Santa Marta, FMNH 66795 (2 adults), UMIM 5927 (2 adults); about 18 km. (10 nautical miles). WSW. of Puerto Colombia, UMIM 5925 (2 adults). Panama: Caledonia Bay, LACM 20467 (7 larvae, 4 juveniles), LACM 20468 (2 juve- niles). Jamaica: Port Antonio, LACM 5802 (1 adult), UMIM 1028 (2 adults). A. nemoptera (Pacific). — Mexico: Guerrero, Acapulco, USNM 75547 (1 adult); Sinaloa, Mazatlan Playa Camaron, UCLA W51-22 (13 young). Costa Rica: Potrero Grande, SU 46385 (5 young to adult); Gulf of Nicoya, Quepos, UCLA W54-55 (1 adult). Panama: Perlas Is- lands, Isla del Rey, Punta de Cocos, UCLA W53-285 (1 young). A. viilpes (Atlantic). — Florida: Monroe Co., Flamingo, Buttonwood Canal bridge, UMML 16775 (20 young); Dade Co., Miami, UMIM 5917 (2 adults). Bahamas; Cay Sal Bank, Cotton Cay, UNIM 5916 (7 adults). Cuba: Havana, estuary of Guanabo River, UMIM 758 (5 juvenile and young). Jamaica: Port Antonio, UMIM 5918 (1 adult). Colom- bia: St. Andrews Island, UMIM 5928 (8 larvae) . METHODS Measurements and counts were made accord- ing to methods described by the senior author (Rivas, 1960) with the following modifications and additions. Standard length was measured from the tip of the snout (not the upper lip) to the middle of the caudal base. Prcpectoral length was measured from the tip of the snout to the insertion of the appressed left pectoral fin. Head length is the longest distance be- tween the tip of the snout and the margin of the left opercular membrane. Mandible length comprises the distance between the anterior tip of the dentary and the posterior tip of the left articular. Preoral length is the median ventral distance between the tip of the snout and the anterior tip of the dentary with the mouth closed. Body depth was measured at the origin of the dorsal fin. Dorsal and anal fin heights were measured from the origin of the erect fin to the upper tip. Last dorsal and last anal ray lengths were measured between the end of the fin base and the tip of the ray. All the dorsal and anal rays were counted, including the an- teriormost short, closely approximated ele- ments. The last two dorsal and anal rays were counted separately. AH counts were made from the fish's left side. All pectoral and pelvic rays and all branched caudal rays were counted. All pored scales were counted including those beyond the caudal base. The scales above the lateral line were counted downward and back- ward from the dorsal fin origin to, but not in- cluding, the lateral line. Those below the lateral line were counted upward and forward from the anal fin origin to, but not including, the lateral line. Only the modified predorsal scales, along the midline of the back anterior to the dorsal fin, were counted. The scales around the caudal peduncle were counted at the region of the least depth. All the gill rakers on the first arch were counted, including rudiments; the count for the lower limb includes the gill raker at the angle. All branchiostegal rays were counted. The vertebral counts include the hypural. GENERIC STATUS OF THE BONEFISH Largely on subjective grounds the genus Dixonina Fowler ( 1911) is here considered as a synonym of the genus Albulu Scopoli (1777). The only two living species of the family Albulidae (Pterothris^us not included), A. vul- pes and A. nemoptera, are much more closely related to each other than previously suspected. The differences between them are only of de- gree and not of the order that would merit generic separation (tables 1-11). Their great superficial similarity is further emphasized by 252 U.S. FISH AND WILDLIFE SERVICE Table 1. — Comparison of £1 Altanlic AlbuUi iieiiKiptera ami 10 A. vulpes of similar mean length on the basis of differential proportional characters (in thousandths of the standard length ) [Ontogenetic variation of cliaracters is indicated Iiy syniliols in part'ntliescs: (I) isometric, CA+) positively allometric, (A — ) negatively allonietrici Character Standard length inini.) Prepectoral lengtli (A — ) I'reanal length (A+)-- - Head length ID Maxillary length (I) Mandible length (I) I'reiira: lenglli (11-. Orliit diameter ll) Caudal pedniiele depth (I) Dorsal base length (I) Dorsal nil heiglil (A-) Last dorsal r;iv length (A+) Last aiml ray ieliglh (A+) Upper eaiidal lobe length (A — ) Lower caudal lobe length (A — ) . 1. nemoptera Range 234-341 270-301 824-847 289-312 133-142 118-128 43- 49 46- 53 60- 68 173-190 161-177 152-193 80- 99 204-234 187-208 Mean 289 286 836 299 138 123 46 m 63 183 169 170 89 223 200 , 1. vntites Rantje 204-387 255-280 835-856 267-296 91-103 83- 97 26- 35 52- 60 70- 78 138-17(i 182-197 54- 66 54- 65 232-275 216-254 Mean 293 268 844 286 94 93 29 65 74 164 IKK 60 5K 256 239 Table 4. — Fref niyonieres .1. nemoptera? Caledonia May, I'unama. LACM 204(17 (7 specimens) liaitge *.;)- 4«.» i -S15 922 -971 119 - 74 .1. vnlfiet Alexander, 19(il: 38-40 (S specinu'ns) Mean- Range \\tean .11.3! 57.0- 4S.5' 51.4 796 1 81)5 -S3U 818 70.8 St. Andrews I.sland. Colombia. r.MIM .5928 (0 specimens) Kan tie .111.2- 48.0 S07 -825 9.111 -975 (17 - (19 I Mean .12.2 814 9(18 1)8.2 comparison of juveniles of A. vulpes and A. nemoptera (table 13). In both species the num- ber of vertebrae is higher than the observed TMti.iv i:i. ('(iiiifiiirisdu of jiifenitcH of .\\\iii\:i Mciiin|itiT.i dii'l .\. vulpes of similar size I Proportions in tliiMi.sandlhs of the standard lontitlil Cliiirai-IiT Standard lenKth (mm.) Predorsal lennth__ I'ri'iiial leniilh NiinUjer of vertebrae... .1. nemoiitera Caledortia Bay, Panama. I.AC.M 204(17, 204(18 (5 specimens) Range Mean Range 31)- 49 43 30- 49 454-407 4riO 472-490 781-793 787 800-817 80- 81 80.5 72- 74 . vnlftcs (illaiialio. Ciiha. r.MIM 7.18 15 specimens) Mean 42 481 812 73.0 number of myomeres, and this difference is probably due to the difl^culty in discerning the last, very closely api^roximated myomeres especially in A. nemoptera. The apparent greater number of vertebrae in juveniles (table 13) is, on the average, three or four units greater than in adults (table 11) as counted from the radiographs, and this diflference may U.S. FISH AND WILUIJKE SERVICE be explained by the fusion of three or four terminal vertebral centra in adults as pointed out by Hollister (1936). The presumed leptocephali of A. nemoptera are identical to those of A. vulpes in general appearance. The juveniles, however, are readily distinguished from those of A. vulpes by the much larger mouth. The smallest juvenile A. nemoptera (Caledonia Bay, Panama, LACM 20468) was 36 mm. standard length. Four other juveniles from the same general locality (LACM 20467) were 42 to 49 mm. Alexander (1961) stated that variation in total myomere counts (65 to 72) might indicate subspeciation or even separate species; some of her larvae with 69 or more myomeres may be A. nemoptera. ECOLOGICAL IMPLICATIONS Frizzell (1965) discussed the ecology and distribution of recent and fossil albulids and suggested that competition between A. vulpes and A. nemoptera drove the latter to deeper water. This conclusion was based on the study of fossils. In agreement with the above suggestion all adults (about 200 mm. standard length or larger) of A. nemoptera for which capture data are available, were collected in relatively deep water. The 21 Oregon specimens (234 to 341 mm.) were collected in trawls in depths of 27 to 110 m. The three Jamaican specimens (197 to 265 mm.) were taken with handline in about 37 m. The Pacific specimens reported by Kumada and Hiyama (1937) were taken by a trawl but no exact depth of capture was given. Of the 19 Pacific specimens reported by Beebe (1942: 44), 5 (220 to 365 mm.) were taken with handline from the ship (Zaca) at unde- termined depths; the other 14 (80 to 200 mm.) were collected with a seine presumably in shal- low water close to the beach. The senior author has been watching for A. nemoptera since 1938. and he has examined hundreds of bonefish in museums and especially in the field throughout southern Florida, the Bahamas, and the Caribbean area. All adult A. vulpes came from depths less than 2 m. except one (231 mm., UMIM 5918) taken with three A. nemoptera from Jamaica. No speci- mens of A. vulpes were taken with the Oregon collections of A. nemoptera. The available evidence suggests that there could be a bathic segregation of adult popula- tions of A. vulpes and A. nemoptera where the latter occupies the deeper stratum. Overlap in their depth ranges is also suggested, but the depth and width of the overlap zone cannot be determined now. ACKNOWLEDGMENTS The curators of several institutions loaned specimens. David K. Caldwell, Curator of Fishes, Los Angeles County Museum, Calif., and Frederick H. Berry, now at the Bureau of Commercial Fisheries Tropical Atlantic Bio- logical Laboratory, Miami, Fla., helped speed the loan of materials. Paul Moore and his assistant Gail Gullette, Singing River Hospital. Pascagoula, Miss., and Charles E. Dawson, Curator of Fishes, Gulf Coast Research Lab- oratory, Ocean Springs, Miss., made the radio- graphs. LITERATURE CITED Alexander, Elizabeth C. 1961. A contribution to the life history, biology and geographical distribution of the bonefish, Albida x'xipes (Linnaeus). Dana-Rep. 53, 51 pp. Beebe, William. 1942. Eastern Pacific expeditions of the New York Zoological Society. XXX. Atlantic and Pacific fishes of the genus Dixonina. Zoologia 27(1): 43-48. Berry, Frederick H. 1964. Review and emendation of: family Clupei- dae, pp. 257-454. Copeia 1964(4): 720-730. Caldwell, David K., and Melba C. Caldwell. 1964. Fishes from the southern Caribbean col- lected by Velero III in 1939. Allan Hancock Atl. Exped. Rep. 10, 61 pp. Fowler, Henry W. 1911. A new albuloid fish from Santo Domingo. Proc. Acad. Nat. Sci. Philadelphia 62: 651-654. Frizzell, Don L. 1965. Otolith-based genera and linages of fossil bonefishes (Clupeiforms, Albulidae). Weiler- Festschrift, pp. 85-110, 2 text figs., pi. 4. Hilderbrand, Samuel F. 1963. Family Albulidae. In Yngve H. Olsen (edi- tor) , Fishes of the Western North Atlantic. Mem. Sears Found. Mar. Res. 1(3): 132-147. BONEFISH SYSTEMATICS AND BIOLOGY 257 HoLLisTER, Gloria. 1936. Caudal skeleton of Bermuda shallow water fishes. I. Order Isospondyli: Elopidae, Me- galopidae, Albulidae, Clupeidae, Dussumieriidae, EriKraulidae. Zoologica 21(4): 257-290. KUMADA, Tosio, and Yosio Hiyama. 1937. Marine fishes of the Pacific coast of Mexico. Nissan Fisheries Institute & Co., Ltd., Odawara, Japan, 75 pp. Metzelaar, J. 1919. Over tropisch Atlantische Visschen, Pt. 1 (West Indian fishes). A. H. Kruyt publ., Am- sterdam, 179 pp. Myers, George S. 1936. A third record of the albulid fish Dixonina 7iemoptera Fowler, with notes on an albulid from the Eocene of Maryland. Copeia 1936(2) : 83-85. NVBELIN, ORVAR. 1960. A gular plate in Albula vulpes (L.). Nature (London) 188(4744) : 78. RivAs, Luis R. 1952. Little known fish facts. So you know all about bonefish? Fla. Fisherman March 7: 3. 1960. The fishes of the genus Pomaccntrus in Florida and the western Bahamas. Quart. J. Fla. Acad. Sci. 23(2): 130-162. ScopoLi, Giovanni A. 1777. Introductio ad historiam naturalem, sistens genera lapidum, plantarum et animalium hac- tenus detecta, caracteribus essentialibus donata, in tribus divisa, subinde ad leges naturae. Prague, x + 506 pp. Walford, L. a. 1939. [Review of] "Marine fishes of the Pacific coast of Mexico," by Kumada and Hiyama (see ref. above). Copeia 1939(2) : 119. 258 U.S. FISH AND WILDLIFE SERVICE RESPONSES OF MARINE ORGANISMS DURING THE SOLAR ECLIPSE OF JULY 1963 By Bernard E. Skud, Fishery Biologist Bureau of Commercial Fisheries Biological Laboratory BooTHBAY Harbor, Maine 04575 ABSTRACT Biological and physical observations and measurements were made on the day before and on the day of a total eclipse of the sun, July 19-20, 1963. Totality occurred at 1745 hours (e.d.t. ), and observations continued through sunset on both days and in two locations — Bar Harbor and Boothbay Harbor, Maine. Plankton was collected at half- hour intervals, and the activity of herring ( Clupea haren- gus harengus ) and green crabs ( Carcinus maenas ) was recorded every 15 minutes. Collections of physical data included surface and subsurface illuminance, air and water temperature, salinity, barometric pressure, cloud cover, visi- bility, and tidal data. At totality and sunset, the volumes of zooplankton in the surface waters decreased. The responses of copepods varied with the species. Pseudocalanus minutus and Acarlia longiremis showed the most pronounced response and moved towards the surface. Females of Acartia were more active than the males. The reactions of other zooplankters were either weak or ill-defined. Herring began schooling near the surface at totality; this behavior, though not as strong, was comparable to that observed at sunset. Green crabs were not active during the eclipse, but were very active after sunset. Similarly, strong echo-tracings were documented after sunset but none were recorded at totality. Apparently the duration of the eclipse was too short or the light intensity too high, or both, to elicit responses from some organisms. Observations from earlier eclipse studies of marine organisms are discussed and comparisons are made with other field and laboratory studies of behavior in relation to environmental changes. Animal behavior during solar eclipses has attracted the interest of scientists and natural- ists, alike, but relatively few observations of aquatic organisms, particularly marine an- imals, have been published. A discussion of this lack at a meeting of the Oceanographic Committee of the National Academy of Sciences in 1962 provided the impetus to un- dertake the observations reported in this paper. The fevi^ specific references to aquatic ob- servations during solar ecHpses generally failed to include adequate definition of the physical conditions. Wheeler, MacCoy, Griscom, Allen, and Coolidge (1935) reported on the behavior of fishes and amphibians as observed by game wardens and the interested public during the eclipse of 1932 in the United States. These reports included remarks about feeding habits of freshwater "trout" and "minnows," re- sponses to angling lures, and unusual activity such as pickerel jumping out of the water, and a goldfish eating the tail of another in an MARINE ORGANISMS DURING SOLAR ECLIPSE aquarium. E. E. Dissell (personal communica- tion, Portland, Maine) reported that a school of pollock (Pollachius virens) surfaced during the 1932 eclipse — the earliest observation I located for a marine fish. Some of these reports suggest a suppression of activity at totality and others an increased level of activity; but most of the reports were casual observations by lay- men and the significance of the observations is limited. Probably the first carefully planned series of observations was made by Mori (1939), dur- ing the 1936 eclipse in Japan. He studied the responses of insects and birds at totality and mentioned the behavior of the sandhopper, Orchestia sp., and the migration of eye pigment in the crayfish, Cambaroides japonicus. He also included a brief reference to responses of other crustaceans and several fishes. Weber (1952), though mostly concerned with terres- trial organisms, concentrated his efforts on species whose normal behavior was well known and recorded changes in temperature, light, and 259 Published June 1967. humidity during a 1952 eclipse in Iraq. Petipa (1955) sampled zooplankton during the 1954 eclipse in the Black Sea, U.S.S.R., and reported that most of these organisms i-esponded by rising towards the surface at totality. K. F. Wiborg (personal communication, Bergen, Norway) made studies off the Norwegian coast during the 1954 eclipse, but poor weather — overcast with strong winds — hampered the collection of zooplankton and interpretation of results. Some of the observations made during the July 1963 eclipse have already been report- ed. Skud (1964) recorded the responses of herring, Clupea harengus, and the green crab. Carcinus maenas; and Backus, Clark, and Wing (1965) described changes in depth of the scattering layers and the occurrence of bio- luminescent flashes. Though the number of references dealing with responses of marine organisms to solar eclipses is small, there is a considerable back- ground of information concerning reactions to light — both in nature and in the laboratory. The purposes of this article are to present the more detailed observations made during the eclipse of July 1963, to compare these observa- tions with pertinent information from similar .studies, and to add to the general knowledge of phototactic responses and rhythmic behavior patterns. OBSERVATIONS AND COLLECTING METHODS The total eclipse of the sun occurred on Sat- urday, July 20, 1963, and the path of totality bisected the State of Maine (fig. 1). At Bar Harbor, mideclip.se occurred at 21:45:00 Ephemeris time (17:44:25 ea.stern daylight time), the sun's altitude was 25 degrees, the path of totality was 53 miles wide, and the duration was 59 .seconds (U.S. Naval Observa- tory, 1961). All of the State experienced at least 98 percent totality. At totality, cloud cover varied considerably along the coast and obstructed viewing in some areas, as did fog patches in certain offshore areas. At Bar Har- bor, the 20-m. Fish and Wildlife Service re- search vessel Rorqual was used as an observa- tion and collecting platform; though a light fog reduced visibility on the water surface to PORTLAND NEW YORK/ Maine SCALE IN MILES 10 20 30 BOOTHBAY HARBOR Figure 1. — Path of totality and vicinity of sampling areas during solar eclipse of July 20, \06:i. a few miles, the eclipse was fully visible. At Boothbay Harbor, the 14-m. FWS vessel Phalarope was used to collect samples. Holding tanks were arranged at the Laboratory dock to study responses of organisms held in cap- tivity. Activity of the captive animals was recorded every 15 minutes. All observations are reported as eastern daylight time unless otherwise specified. Observations and collections were made on the day previous to the eclipse and on the day of eclipse, beginning at 1600 hours and con- tinuing at intervals through totality (1745) until 2300. The 2-day .sequence was intended to provide a test and control in detecting dif- ferences in the behavior of animals, as was the extension of observations through sunset and early evening. Light measurements at Boothbay Harbor were made at the surface with a Gos.sen Lunasix electronic exposure meter.' This meter lacked the flat interception ' Trade names refiTred ki in this publicalion rln not im|j|y endorsement of the products. 260 U.S. FISH AND WILDLIFE SERVICE screen necessary for precise measurement of illumination, but the incident light readings did provide an index for comparing the illumina- tion between and within days. Aboard the Rorqual an irradiance meter (Model C-la, Ma- rine Advisors, Inc.) equipped with Weston pho- tronic cells measured subsurface changes in light penetration. This unit had a filter with a peak sensitivity of 550 millimicrons and a range of 390 to 760 millimicrons. The irra- diance meter provided the ratio of the amount of radiation at the depth of the submerged cell to a reference cell on deck. Secchi-disk readings also were taken at regular intervals. Plankton samples were taken at half-hourly intervals with Miller high-speed samplers at Bar Har- bor, and at hourly intervals with the Clarke- Bumpus sampler at Boothbay Harbor. Tem- perature and salinity were recorded and echo soundings were made continuously during the sampling period. PHYSICAL CHANGES Measurements at Boothbay Harbor showed that illuminances on the day before the eclipse were (control day) and the day of the eclipse was not closely comparable (table 1). Except for the period of totality, surface illumination was far greater on the day of the eclipse than on the control day. This difference was also evi- dent from other data. Visibility on July 19 was limited to 9 km., and nine-tenths of the sky was covered by cirrostratus clouds ; on July 20, objects were visible at 16 km. and the stratus 17 Wednesday -I 18 Thursda y JULY 19 Fridau cloud cover of seven-tenths was generally dis- solving. These differences limited the com- parisons which could be made between test and control days. The decrease in surface illuminance before and at totality and the subsequent increase are documented in table 1. An hour before totality. Table I. — Surface illuminnnce nt Boothbiy Il/irbir, July V) end .?(\ 1063 Time (e.d.t.) July 19 July 20 1546-1600-- Luxes 38.000 38.000 33,000 75.000 93.000 64,000 Luxes >100.000 1601-1615 .. . >100.000 1616-1630 >100.000 1631-1645 75.000 1646-1700 1701-1715 75.000 1716-1730 .-. 55.000 1731-1745 24.000 24.000 11.500 9.500 16.750 4.800 3.600 3.600 900 265 110 20 <10 <10 ' 900 1746-1800 - 1.050 1S01-1K15 24.000 1S16-1S30 .- 2S.000 1S31-IS45 19,000 lMt'i-1900 _ 16.600 1901-1915 16.600 1916-1930-.- -- .- -- 3.600 1931-1945 825 1946-2000 - - = 265 2001-2015 2016-2030 50 2031-2045. . - <10 2046-2100 <10 > Totality. '^ Sunset. 75,000 luxes were recorded, 900 at totality, and 28,000 within the hour after totality. Darkness at totality (1745) approximated that which oc- curred one-half hour before sunset. Though the primary purpose of Secchi-disk observations was to measure water clarity, the results also provided information on the submarine light penetration during the eclipse. The extinction depth of the Secchi-disk was 7.0 m. at 1600, /- 20 Saturday 4- 21 Sundau Twi a 8 10 1^? 2 4 6 8 10 M' ? « 6 a 10 1,; 2 4 g B 10 M' ? 4 6 B 10 1.2 2 4 « 8 10 M' 2 4 6 8 10 1.2 2 4 6 8 10 M' 2 4 6 8 10 1,2 2 4 6 8 10 M' 2 4 BAROMETRIC PRESSURE ^3Q50 / f^^ t / -i-fe30J5(t: Figure 2. — Barometric pressure, mm. of mercury, adjusted to sea level, at Boothbay Harbor. The arrow indicates time of totality. MARINE ORGANISMS DURING SOLAR ECLIPSE 261 5.5 m. at totality, 6.0 m. at 1 hour after totality, and 5.5 m. at sunset. In conjunction with these observations, baro- metric pressure was measured from Wednes- day, July 17 through Sunday, July 21 (fig. 2). Totality coincided with the low pressure read- + 1.0 3000 ■..■I....I....I... ■!.... I,,..l,.,il..,.l.,,.l,... I 1700 1730 1800 TIME (ed.l) 1830 ing (29.45 mm.) for the period of observation. The consistent decline in pressure before the eclipse and the rise after totality may or may not be coincidental, but I have been unable to locate similar records from other eclipses. Observations aboard the Rorqual in Bar Harbor were supplemented by land-based ob- servations on Mount Cadillac. LFE Electronics (Boston) conducted a series of tests and sup- plied me with comparative measurements of a time-light series (fig. 3). The differences among the curves are largely due to the dif- ferent spectral responses of the photo cells and filters. This information provided an inde- pendent comparison of our own measurements with the irradiance meter (table 2). Ten min- Table 2. — Surface and submarine illuminanre (luxes) during the eclipse at Bar Harbor, Maine Depth Time (e.d.t.) 1612 1657 1735 > 1804 1840 Meters Lutes 36.000 7,600 2.600 1,100 500 Luxes 27.800 5.800 1..W0 700 300 Luxes 4.000 800 400 100 Luxes 4.800 1.100 700 200 100 Luxes 6.200 ^ 1 500 10 rm 15 150 18 - 100 Figure 3. — Time-light curve from: A. photometer; B. Gossen light meter; and C. Gossen Sixticolor meter. (One foot-candle = 10.76 luxes.) ' Totality 1745 e.d.t. utes before totality, the illumination at the surface registered 4,000 lu.xes, the lowest value in the series of measurements before and after totality. Subsurface values were also lowest at this time. As is evident from these data, the eclipse occurred during a normal period of declining brightness, but the substantial re- duction in illumination at or near totality and the subsequent increase clearly distinguishes the influence of the eclipse. During the eclipse, air temperature declined from 15.3° to 12.5° C, and water temperature at the surface de- clined from 12.8° to 11.3° C. Though the eclipse may have accentuated the temperature change, the late afternoon decline was antici- pated. Water temperature at depth remained nearly constant; 1 hour before totality it was 10.3° C. at 10 meters; 9.3° at 20 m.; 8.3° at 30 m. ; and 7.8° at 60 m. Salinity ranged from 31.35 to 31.74 %o at the surface and from 32.09 to 32.23 at 60 m. These differences in salinity were assumed to be caused by tidal 262 U.S. FISH AND WILDLIFE SERVICE movements. Low tide was at 1646 hours and was -0.08 m. ; high tide occurred at 2256, and was 4 m. A light fog sometimes hampered vertical visibility but generally did not obscure the sun. BEHAVIOR OF ZOOPLANKTON Cladocerans and copepods were the most numerous zooplankton, accounting for more than 90 percent of the plankters. Less abun- dant groups included gastropods, brachyurans, and decapod larvae, cirriped nauplii, and chae- tognaths. Total volumes of zooplankton and the distribution of seven species of copepods were examined to determine any behavioral changes during the eclipse. Though the light intensity on July 19 was lower and more variable than on the day of the eclipse, this difference did not totally negate the comparison of zooplankton distribution on the test and control days. Miller high-speed samplers (without meters) were used at Bar Harbor; and Clarke-Bumpus samplers (with JULY 19 1800 1900 2000 2100 TIME led! I Figure 4. — -Surface volumes of zooplankton from two sampling locations compared with surface illumina- tion at Boothbay Harbor. meters) at Boothbay Harbor. The zooplankton volumes from surface tows in the two locations are compared in figure 4, along with the changes in light intensity during the 2 days. Although this comparison does not account for amounts of water strained for either gear, the duration of tows in each locale were nearly the same, and the changes in zooplankton abun- dance which were recorded for the different gears followed similar trends. On both days, the surface volumes of zooplankton decreased at or near sunset and then increased rapidly during the following hour. On the day of the eclipse, a similar decrease was noted at totality, both at Bar Harbor and Boothbay Harbor. This phenomenon was also evident in the quantita- tive data from the Clarke-Bumpus samplers. Volumes of collections made at the surface were 130 cc./lOm.^ an hour before the eclipse, 67 cc./lOm.' at totality, and 105 cc./lOm.^ an hour after totality. The decrease in volume also was noted at 20 m. — 231 cc. lOm.^ an hour before totality, 165 cc./lOm.^ at totality, and 222 cc./lOm.^ an hour later. At intermediate depths, 3 to 10 m., there was no pronounced change in volume at totality. The distribution of seven species of copepods differed measurably on both the test and con- trol days and during the period of the eclipse. On July 19, Pseudocalanus minuhis and Acar- tia longiremis occupied shallower strata of water than on July 20, which was the brighter day. These species also showed the most pro- nounced response during the eclipse. The up- ward movement of these two species and the differences in vertical distribution between days are shown in figure 5. The responses of Centropages hamatus, Tortanns discaudatus, Calanus finmarchicus, Temora longicornis, and Eti-rytemora herdmani were not as well defined as those of P. mimitus and A. longiremis, and the responses of some species differed at the two sampling locations. For example, the numbers of C. finmarchicus and T. longicornis from the surface to 10 m. increased during the eclipse at Boothbay Harbor, but declined at Bar Harbor. The abundance of E. herdmani was so limited in Bar Harbor that its distribution could not be plotted reliably; in Boothbay Har- bor this species was one of the most abundant MARINE ORGANISMS DURING SOLAR ECLIPSE 263 TIME (ED.T) 1800 PRE-ECLIPSE 1703 ECLIPSE POST-ECLIPSE 1755 1903 10 20 10 2 20- I t— a. UJ Q 10 20- Pseudocolonus minutus PRE-ECLIPSE JULY 20 POST-ECLIPSE Acortia longiremus PRE-ECLIPSE JULY 20 POST-ECLIPSE 50 50 50 50 PERCENT OCCURRENCE 50 50 Figure 5. — Depth distribution of two species of cope- pods before, during, and after the eclipse. forms, but showed no appreciable change in distribution during the eclipse. Though more frequent sampling might have established a basis for understanding these differences, it is evident that all species did not respond in the same degree or in the same manner. The lack of uniformity among species and within species was not peculiar to the eclipse study. Wynne-Edwards (1962) summarized several early works which demonstrated the differences in behavior of copepods. Clark (1933 and 1934) discussed diurnal changes in vertical distribution relative to sex and age- groups and reported the stronger migratory habits of adult female C. fi)imarchicus, which rose much nearer to the surface at night than did the male. During the eclipse, female A. longi- remus were more active than males and were more prevalent at the surface (fig. 6). There was a suggestion of a reversal of this phe- nomenon in P. minutus, and no difference in the distribution of sexes in T. longicornis. Dur- ing the eclipse of 1954, Petipa (1955) reported that vertical migration in the Copepoda was Or S X I— LiJ Q Acortia longiremis cf> Acortia longiremis o 100 50 50 100 PERCENT OCCURRENCE Figure 6. — Depth distribution in percent of male and female Acartia longiremis during the eclipse. restricted almost entirely to adult females. In all of his samples, females of A. clausi were concentrated in the upper layers and the males at lower depths. BEHAVIOR OF THE GREEN CRAB Naylor (1958) observed that the rhythmic activity of green crabs iCarcinus maenas) could be divided into two components, one of diurnal frequency with a peak at night, and the other, a tidal frequency with a peak at high tide. On the day of the eclipse, low tide oc- curred an hour before totality and high tide three hours after sunset; con.sequently the be- havior of the crabs could be judged inde- pendently of their response to tides. Thirty crabs, 15 of each sex, were u.sed as test animals. The mean width was 62 mm. (range, 42-79 mm.) for the males, and 60 mm. (range. 51-69 mm.) for the females. The males were banded so that the sexes could be readily distinguished during the experiment. The crabs were placed in a fiberglass tank 1.3 by 1.0 by 0.5 m. that had a 7 cm. layer of sand and gravel on the bottom (fig. 7). One corner of the tank was covered with fiberboard to provide a darkened shelter. Running water 264 U.S. FISH AND WILDLIFE SERVICE Figure 7. — Fiberglass tank for holding green crabs. was supplied from the laboratory salt-water system and the tank was placed out-of-doors in an area free from shadows. Naylor (1958) found that the normal precision and level of the rhythmic activity declined after the crabs had been held for 3 to 4 days. For my study the crabs were placed in the tank two days before the eclipse and took refuge in the shelter immediately. Activity was recorded by count- ing the crabs that left this cover. A control was established by making observations period- ically during the day before the eclipse and at 15-minute intervals from 1600 to 2200 hours, well past sunset. On July 19 no crabs left the sheltered area until 2035. From that time until the last ob- servation at 2205 the activity generally in- creased; as many as 11 males and 10 females left the covered area ; the average number of active males was 5.7 and the average number of females 5.2. On the day of the eclipse, July 20, no activity was observed until 2050, more than 3 hours after totality; the greatest num- ber of males in the unsheltered area at any time was 13, and the highest number of fe- males 5; between 2050 and 2200 the average number of active males was 9.2 and females, 3.6. Because the crabs were not active during the eclipse, other experiments were conducted the following day to determine light conditions which would elicit a response. When the tank was in daylight (27,000 luxes) and then cov- ered by a heavy tarpaulin that reduced the light to less than 10 luxes, 2 minutes elapsed before any activity was noted. Under condi- tions of subdued artificial light of 500 luxes, which was then reduced to less than 10 luxes, the response was more rapid; 3 crabs were active within 30 seconds and as many as 10 came out of the shelter within 2 minutes. As in the observations made on the date of the eclipse, the males were the first to respond. The light intensity in the half hour before totality was considerably greater than that of the artificial light; apparently the duration of subdued light during the eclipse was too short or the in- tensity too high to elicit a response from the crabs. BEHAVIOR OF HERRING Generally, the behavior of Atlantic herring (Clupea harengus harengus) is well document- ed, but specific responses are extremely vari- able. Blaxter and Parrish (1965), studying vertical movement, concluded that it was not possible to show any relationship between the preferred depth, or the extent of upward move- ment, and such factors as gradients of salinity, temperature, or food. The herring used in the eclipse study had been held in large tanks for several weeks. Though their behavior could not be considered comparable to that of herring in their natural environment, the fish were ac- climated to confined conditions which were nec- essary for the observations made during the eclipse. About 75 two-year-old herring were placed in a small-meshed holding pen during early morning of July 19. Observations were made at 15-minute intervals from 1600 to 2200 hours on July 19 and 20. The pen (dimensions, 3 by 3 by 2 m.) was visually separated into quadrants A, B, C, and D (fig. 8) . Each quadrant was divided into two sections by an imaginary plane midway between the surface and the bottom of the pen. The presence or absence of fish in these quadrants and the depth divisions were recorded, along with remarks on schooling and directional movement. On both days, thei-e MARINE ORGANISMS DURING SOLAR ECLIPSE 265 Figure 8. — Impoundment for herring, showing quad- rants used to record observations. was only one period (30 to 45 minutes) during which no fish were in the upper layer; this distribution occurred about an hour before sun- set (2000). Similarly, there was only one period during which no fish were located in the lower layer. This period began one-half hour after sunset and continued for 45 to 60 minutes, after which herring were dispersed throughout the holding pen. During daylight herring were more numerous in the upper area of quadrants A and B than in C and D. This difference probably is ex- plained by the uneven di.stribution of light. Quadrants C and D were located closest to the vessel float which was used to anchor the hold- ing net. The upper areas of these quadrants were shaded by the float and were avoided by the fish. The distribution of herring in the lower layers of A-B and C-D was relatively uniform. After dark, fish were equally dispersed in the upper and lower layers of the pen. On the day of the eclipse, fish were distrib- uted in the upper layers of quadrants A and B and in the lower layers of quadrants A, B, C, and D. Fi.sh were absent from the upper layer of quadrant D until sunset. No fish were ob- served in the upper layer of quadrant C from the start of observations at 1600 until 1730 — 15 minutes before totality. Fish were active in this quadrant from 17.S0 until 1800; were absent at 1815; and reappeared from 1830 to 1900. This movement of fi.sh into quadrant C at the approach of totality was coupled with a change in behavior of herring in the other quadrants. Some fish began to school and moved to the surface of the water. This was in contrast to the preeclipse behavior of gen- eral dispersion without movement at the sur- face, and apparently was in response to the reduced light during the eclipse. The response was not strong, and not all fish reacted to the change. The data on subsurface illumination (mentioned above) suggests that the duration of lowered light intensity at totality was not enough to stimulate a stronger response. This conclusion is supported by the observations at sunset, when the decrease in light approximat- ed that of the eclipse, but for a longer period, and elicited a stronger schooling response from the herring. ECHO SOUNDINGS Echo sounders aboard the two ve.ssels were run continuously during the study, but none of the records showed any change or movement of organisms during the eclipse. Echo tracings in Boothbay Harbor documented considerable activity in the early evening on July 19 and 20 (fig. 9). On July 19, activity was first detected at 1854 when minor peaks and streaks extended up from the bottom. By 1955, some of these streaks and dots were no longer in contact with the bottom ; others remained in contact but were extended and more pronounced. By 2100, only a few of these marks were in contact with the bottom ; the rest were scattered from the bottom to the surface. This phenomenon was also recorded on the evening of July 20, though somewhat later and less pronounced than on the previous evening. None of the organisms taken in the plankton nets was large enough to account for the mark- ings observed on the recording paper, and the scheduled collections did not allow time to utilize other gear on July 19 or 20. On July 21, a trawl was fished during late evening in the same area. Echo soundings were similar to those on July 19 and 20. Large catches of jellyfish (Aurelia aurita) were taken in the net; presumably they were the animals detect- ed by the sounder, but the possibility exists that the tracings were from herring or other fish that escaped capture. In any case, the lack 266 U.S. FISH AND WILDLIFE SERVICE of response from these organisms during- the eclipse indicated that the lowered light inten- sity or duration of totality, or both, were not sufficient to stimulate the kind of movement observed after dark. 30m. i^M^fivi '^'^^sBii \^Mi Figure 9. — Echo soundings before and after sunset (2000 hr.). COMPARISON WITH OBSERVATIONS FROM OTHER ECLIPSES The observations presented in this paper show some agreement with those of previous workers, but the responses recorded for some species were not the same. Considering the variables such as light intensity, duration of totality, and the variation in experimental de- sign, these differences are not surprising, but they should be equated. Mori (1939) conducted carefully designed experiments on several species, and made de- tailed observation on others, including the sandhopper, Orchestia sp. He concluded that sandhoppers "were apparently not affected by the eclipse," yet he does mention that a few individuals were exposed "towards the end of totality ....," exhibiting their normal cre- puscular behavior, but this activity lasted only a few seconds before the animals retreated into hiding. Totality during the 1936 eclipse in Japan lasted 2 minutes, and began at 1519 when illuminance under normal conditions is high. As Mori stated, the inactivity may be explained by the fact that the change of light intensity at totality was too rapid ; but he also cautioned that factors such as humidity and atmospheric pressure might have been the con- trolling influences. He also reported on ob- servations from the aquarium of the Akkesi Marine Biological Laboratory, ". . . shrimp, a flat fish, a young salmon, a trout, and a herring were all indifferent to the eclipse, whereas a crab, which is quiet on ordinary days, began to move, and a bullhead appeared from the shady tangle of weeds when it became darker and hid again when it became lighter in just the same way as seen on ordinary days and nights." In reference to light-dark cycles, Biinning (1964) stated that deviations from the natural frequency of an organism could "necessarily have an entirely different relationship to the light and dark period than normally. For ex- ample, if the dark period is too short, the organism with its own cycle length does not have time enough within the dark period to reach the usual physiological state typical of night." He also stressed that sometimes the beginning of the light period has a greater influence on the timing of responses during dark than the beginning of the dark period itself. MARINE ORGANISMS DURING SOLAR ECLIPSE 267 other factors could also contribute to the degree of response. Of particular interest is the barometric pressure. As shown in figure 2, the pressure declined continuously for the 2 days preceding the 1963 eclipse. Brown (1958) showed that cycles of activity in certain organ- isms fluctuate with changes in pressure. These species included the fiddler crab (Uca), the oyster (Ostrea), and the quahog (Venus). Though these animals also exhibit daily and lunar cycles of activity, mean hourly rates of activity were correlated with the rates of rise or fall in barometric pressure. Activity in- creased with the hourly rate of fall and de- creased with the rate of rise. The importance of this phenomenon to the observations during the 1963 eclipse is uncertain; it could be of significance during an eclipse with a relatively long period of totality. In comparing the responses of zooplankton during the eclipses of 1954 and 1963, the im- portance of documenting the environmental differences is readily apparent. Petipa (1955) reported that most species reacted by rising to the upper water layer (0-5 m.) during the eclipse and by descending to lower depths (5-14 m.) after the eclipse. The strongest re- sponse was from Sagitta. and larvae of De- capoda, Lamellibranchia, and Gastropoda. His work was done in the Black Sea at Sevastopol Bay where surface temperatures during the eclipse (June) were at least 20° C. Zenkevitch (1963) described the general hydrological fea- tures of the Black Sea : .salinity varied between 17 and 18 %^ at the surface and was only 22 to 23 %o in deep water; temperature at 25 m. was 14° C. in summer and 6° C. in winter; dissolved oxygen content ranged from 1.05 to 7.76 cm.r 1. at 50 m. and declined rapidly with increasing depth below 50 m. ; water deeper than 150 m. was contaminated with hydrogen sulphide; Secchi disks di.sappeared between 18 and 21 m. ; and most species of zooplankton were found at depths above 50 m. and were concentrated between the surface and 25 m. Many of these characteristics are strikingly different from those in the coastal waters of the Gulf of Maine: during the eclipse of July 20 the temperature was less than 15° C. at the surface and was 10° C. at 10 m. ; .salinity was about 32 %„; and the Secchi disk disappeared at less than 10 m. The dissolved oxygen con- tent in the Gulf of Maine was reported by Gran and Braarud (1935) to vary between 5.5 and 7.8 cm. 1. at 40 m.. and Bigelow (1926), in contrast to the conditions in the Black Sea, reported many species of zooplankton below 50 m., some of which had their densest con- centrations below 100 m. Other differences to consider include the characteristics of the eclipse and the location of sampling in relation to the path of totality. The sampling sites in Maine were selected be- cause they lay in or near the path of totality. In contrast, Sevastopol Bay was about 400 miles from the path of totality in 1954. (This figure was estimated from eclipse data pre- sented by Oppolzer, 1962.) As Petipa (1955) did not provide any measure of light intensity, one can only a.ssume, other things being equal, that the illuminance during the eclipse was higher at Sevastopol than at Bar Harbor. Yet Petipa recorded more activity of zooplankton than was noted in the Gulf of Maine. Differ- ences in species were important, but I suspect that the differences in the two environments were more critical. Though not a species encountered in this study, experimentation on Daphnia offers sev- eral plausible explanations for the zooplankton behavior ob.served during the eclipse. Harris and Wolfe (1955) found that Daphnia re- sponded rapidly to changes in light intensity, moving in the direction of the original optimum intensity, but this was followed by movement towards an adapted optimum and resulted in little change of position. In essence, a high change of intensity produced an alteration of photonegative and photopositive phases, and the net result had relatively little effect on the depth at which the animal was located. When changes in light intensity were slow, however, the animals simply followed the movement of the original optimum zone. Ringelberg (1964) disagreed with the explanation of the photo- tactic response offered by earlier workers; he concluded, on the basis of a very thorough laboratory and field study, "that the directing stimulus for the phototactic reaction is a con- trast or a gradient present in the angular light 268 U.S. FISH AND WILDLIFE SERVICE distribution." Schallek (1943) reported that Acartia tonsa in a glass cylinder would move upwards when illuminated from above and downward when illuminated obliquely. He con- sidered the reaction of A. tonsa to diffuse light in the cylinder to be in accord with the down- ward movement in the ocean during the day, but that the reaction to direct light under experimental conditions had no bearing on its behavior in nature. These experiments emphasize the importance of other variables that one must consider in attempting to compare and evaluate observa- tions during an eclipse. The time of day and resultant attitude of the sun in relation to water clarity are of particular concern. Holmes (1957) discussed the penetration of water by light and explained that "The extinction of daylight in the sea is caused by absorption (by the water itself, by particles, and by dissolved substances) and by scattering (by the water and by particles)." Ringelberg (1964) and Schwassman and Hasler (1964) have recog- nized the importance of absorption and scat- tering on the phototactic behavior of aquatic organisms ; the former paper referred to Daphnia, especially the orientation of the eye axis and the body axis, and the latter referred to sun orientation of fishes. The responses of herring observed during the 1963 eclipse were in general agreement with reports of other observations under vary- ing conditions of light intensity. Johnson (1939) studied captive herring in southern New Brunswick and concluded that, in the absence of direct sunlight, these fish "extended to the surface at all times — dawn, sunrise, cloudy days, sunset, dusk, moonlight, starlight, and cloudy nights." He also found that during daylight, the depth of the fish was greatest when the sun's altitude was highest and the largest fish were in the deepest water. Blaxter and Holliday (1963) summarized the work of European scientists, particularly in the North Sea where the diurnal migration pattern of herring is well documented; and the depth of herring shoals has been correlated with isolux lines to estimate the optimum depth for setting gill nets. In studying diurnal changes in behavior of adult herring, Blaxter and Parrish (1965) found that the depth (and light intensity) at which the fish occurred during the day was extremely variable ; and demonstrated that fish did not move towards the surface until illumi- nance decreased to 10 luxes. These authors also reported that "recruit fish (21/2-3 years old)" remained in higher light intensities by day. The eclipse study lends support to this latter conclusion. Though the response of 2-year-old herring at totality and at sunset was limited, the light intensity was above the level that Blaxter and Parrish observed as necessary to elicit a surface movement by adults. Breder (1951 and 1959) discussed the influence of light on the social grouping of many species of fish and provided a thorough summary of other scientists' work in this field. He stressed the differences in responses by individuals, by sex, and by species. This emphasizes the need to select species whose behavior patterns are well known, when attempting to evaluate the eflfects of a solar eclipse. The Atlantic herring, in this regard, is a suitable species, except that the sexes cannot be distinguished readily through external examination. Mention should be made of the types of periodic activity and their importance to the observations made during the eclipse. Allee, Emerson, Park, Park, and Schmidt (1949) classified successive diel periods into two types : exogenous, "in which the pattern is directly induced and controlled by periodic environmen- tal influences" and endogenous, "in which the pattern is resident in the organism." Aschoff (1960) elaborated on the definitions, explain- ing that an environmentally controlled perio- dicity (exogenous) will cease under artificially constant conditions; whereas, periodic factors of the environment only serve as synchronizing agents (Zeitgeber) for circadian or endogenous periodicity. He pointed out that a single en- vironmental event can never synchronize con- tinuously and therefore cannot operate as a Zeitgeber. This implies that observations made during an eclipse should not, of themselves, be used to determine whether a response or lack thereof is indicative of either an exogenous or endogenous rhythm. On the other hand, these observations can provide supporting evidence MARINE ORGANISMS DURING SOLAR ECLIPSE 269 for laboratory or other field experiments con- cerned with rhythmic behavior patterns. Cloudsley-Thompson (1961) cautioned that rhythmical activities of an animal are not necessarily all of one type and stated that rhythms solely dependent on the environment are rare and probably represent rhythms which are independent but out of phase with the en- vii-onment. In regard to field observations dur- ing solar eclipses, he concluded that the results agree with those of laboratory experiments, in that certain animals exhibit some periodic activities that appear to be dependent on the environment and others that are more mark- edly independent. SUMMARY 1. A total eclipse of the sun occurred in Maine on July 20. 1963. Totality lasted 59 seconds. 2. Biological and physical observations were made on the day before the eclipse and the day of the eclipse and were continued through sun- set each day to provide a comparison with regular light-dark cycles. 3. Surface and subsurface illuminance de- clined markedly at totality, approximating conditions at sunset. 4. Barometric pre.ssure declined steadily for the 2 days prior to the eclipse, reached a low point 29.45 mm. at totality, and then increa.sed. 5. Zooplankton volumes from surface waters decreased during the eclipse and at sunset at both of the .sampling areas. 6. Of the dominant copepods, Pseudocalanus minutus and Acarfia longiremis exhibited the most pronounced rei5pon.se to the eclipse and moved toward the surface. The reactions of other species were either weak or ill-defined. 7. Female Acarfia longiremis were more ac- tive than males during the eclipse, and moved toward the surface at totality. 8. No change was ob.served in the behavior of green crabs during the eclipse. Apparently, the duration of the eclipse was too short or the light intensity too high, or both, to elicit a re- sponse. 9. At totality, Atlantic herring held in a pen responded in a manner comparable to that observed at sunset. The response was not 270 strong, but some fi.sh began .schooling and moved into the surface waters. 10. Echo tracings documented a movement toward the surface after sunset, but tracings during the eclipse showed none of this activity. Though large catches of jellyfish were taken, the traces could have been made by fi.shes which escaped the net. 11. Comparisons of my observations with tho-se made during other eclipses emphasize the importance of designing experiments carefully to assess properly the behavioral respon.ses in relation to environmental changes. ACKNOWLEDGMENTS Scientists and technicians from the Bureau of Commercial Fisheries Biological Laboratory, Boothbay Harbor, Maine, a.ssisted in the col- lection and preparation of the data; and S. R. Studenet.sky, Atlant-NIRO, Kaliningrad, U.S.S.R., arranged for the translation of Petipa's paper. LITERATURE CITED Allee, W. C, a. E. Emerson, O. Park, T. Park, and K. P. Schmidt. 1949. Principles of animal ecology. W. B. .Saun- ders Co., Philadelphia and London, 8,37 pp. ASCHOFF, JURGEN. 1960. Exogenous and endogenous components in circadian rhythms. Cold Spring Harbor Symp. Quant. Biol. 25: 11-28. Backus, Richard H., Robert C. Ciark, and Asa S. WiNR. 19fi5. Behavior of certain marine organisms dur- ing the solar eclipse of July 20, 1963. Nature 205; 989-991. BiGELOw, Henry B. 1926. Plankton of the offshore waters of the Gulf of Maine. Bull. Bur. Fish. XI., Part II: 500 pp. Blaxter, J. H. S., and B. B. Parrish. 1965. The importance of light in shoaling, avoid- ance of nets and vertical migration by herring. J. Cons. 30(1) : 40-57. Blaxter, J. H. S., and F. G. T. Holliday. 1963. The behavior and physiology of herring and other Clupeids. In F. S. Russell (editor). Ad- vances in marine biology 1, pp. 261-393. Aca- demic Press, London and New York. Breder, C. M., Jr. 1951. Studies on the structure of the fish school. Bull. Amer. Mus. Natur. Hist. 98, Art. 1: 1-27. 1959. Studies on social groupings in fishes. Bull. Amer. Mus. Natur. Hist. 117, Art. 6: .393 482. U.S. FISH AND WILDLIFE SERVICE Brown, Frank A., Jr. 1958. Studies of the timing mechanisms of daily, tidal, and lunar periodicities in organisms. In A. A. Buzzati-Traverso (editor). Perspectives in marine biology, pp. 269-282. Univ. of Cali- fornia Press, Berkeley and Los Angeles. BUNNING, Erwin. 1964. The physiological clock. Endogenous diur- nal rhythms and biological chronometry. Aca- demic Press, New York, 145 pp. Clarke, George L. 1933. Diurnal migration of plankton in the Gulf of Maine and its correlation with changes in submarine irradiation. Biol. Bull. 65(3): 402- 436. 1934. Further observations on the diurnal migra- tion of copepods in the Gulf of Maine. Biol. Bull. 67(3) : 432-448. Cloudsley-Thompson, J. L. 1961. Rhythmic activity in animal physiology and behavior. Academic Press, New York and Lon- don, 236 pp. Gran, H. H., and Trygve Braarud. 1935. A quantitative study of the phytoplankton in the Bay of Fundy and the Gulf of Maine (including observations on hydrography, chemis- try and turbidity). J. Biol. Bd. Can. 1(5): 279-467. Harris, J. E., and Ursula K. Wolfe. 1955. A laboratory study of vertical migration. Roy. Soc. (London), Proe. B, 144: 329-354. Holmes, Robert W. 1957. Solar radiation, submarine daylight and photosynthesis. Geol. Soc. Amer., Mem. 67(1): 109-128. Johnson, W. H. 1939. Effects of light on movements of herring. J. Fish. Res. Bd. Can. 4: 349-354. Mori, Syuiti. 1939. Effects of the total solar eclipse on the rhythmic diurnal activities of some animals. Annot. Zool. Jap. 18: 115-132. Naylor, E. 1958. Tidal and diurnal rhythms of locomotory activity in Carcinus maenas (L.). J. Exp. Biol. 35(3) : 602-610. Oppolzer, Theodor R. von. 1962. Canon of eclipses (Canon der Finster- nisse). [Translated by Owen Gingerich.] Dover Publications, Inc., New York, 376 pp. Petipa, T. S. 1955. Nablyudenie nad povedenium zooplanktona vo vremya solnechnogo zatmeniya (Observa- tions on the behavior of zooplankton during a solar eclipse.) Dokl. Akad. Nauk SSSR 104: 323-325. Ringelberg, J. 1964. The positively phototactic reaction of Daph- nia magna Straus: a contribution to the under- standing of diurnal vertical migration. Neth. J. Sea Res. 2: 319-406. Schallek, William. 1943. The reaction of certain Crustacea to direct and diffuse light. Biol. Bull. 84: 98-105. ScHWASSMANN, HoRST 0., and Arthur D. Hasler. 1964. The role of the sun's altitude in sun orienta- tion of fish. Physiol. Zool. 37: 163-178. Skud, Bernard E. 1964. Behavior of marine organisms during a solar eclipse. Amer. Zool. 4: 300. (Abstr.) U.S. Naval Observatory. 1961. The American Ephemeris and nautical almanac. Nautical Almanac Office, Department of the Navy, Washington, D.C., 518 pp. Weber, Neal A. 1952. The 1952 animal behavior eclipse expedition of the College of Arts and Science. Baghdad Coll. Arts Sci., Publ. 2: 1-23. Wheeler, William Morton, Clinton V. MacCoy, Ludlow Griscom, Glover M. Allen, and Harold J. Coolidge, Jr. 1935. Observations on the behavior of animals during the total solar eclipse of August 31, 1932. Amer. Acad. Arts Sci., Proc. 70: 33-70. Wynne-Edwards, V. C. 1962. Animal dispersion in relation to social be- haviour. Hafner Publishing Co., New York, 653 pp. Zenkevitch, L. 1963. Biology of the Seas of U.S.S.R. [Translated by S. Botcharskaya.] George Allen and Unwin Ltd., London, 955 pp. MARINE ORGANISMS DURING SOLAR ECLIPSE 271 SHELL DEFORMITY OF MOLLUSKS ATTRIBUTABLE TO THE HYDROID, HYDRACTINIA ECHINATA By Arthur S. Merrill, Fishery Biologist Bureau of Commercial Fisheries Biological Laboratory Oxford, Maryland 21654 ABSTRACT The colonial hydroid, a common epizoon on the external surface of the shell of the sea scallop, Placopecten magel- lunictis, sometimes becomes established on the internal shell surface. This intrusion interferes with the normal ac- tivities of the scallop's mantle, often causing shell de- formity. The scallop reacts by producing a new shell edge within the existing perimeter of the shell and by- passing the hydroid colony. This relation is amensal — one organism is inhibited, and the other is not a£Fected. No proof was found that the hydroid may ultimately cause the death of the scallop. These same hydroids, in symbiotic association with pagurid crabs, deform and enlarge the apertures of empty gastropod shells. Enlargement of the "house" is to their mutual advantage. The availability of surface on which to at- tach and grow is vital for sessile fouling or- ganisms. The larval forms of these species must eventually settle on firm substrate or perish. The organisms are often able to aug- ment available substrate by settling on the surface of other organisms. The upper valve of the sea scallop, Placopecten magellanicus (Gmelin), provides such a surface. Organisms found on shells of sea scallops are those which are found on any suitable sub- strate in the vicinity. They are commensals, typical epizoa competing for space. They may be pelagic and settle by chance in a particular area, or benthic and possess mobility during larval and early postlarval stages. Postlarval commensal species found on the sea scallop include representatives of most marine phyla (Merrill, 1961). Most abundant are the boring sponges, sea anemones, branch- ing and encrusting hydroids and bryozoans, pelecypods, barnacles, tubeworms, and simple and colonial ascidians. Wells and Wells (1964), reporting on a calico scallop, Aeqrdpecten gihhus (Linnaeus), community study, made similar observations. They found that of the major taxonomic groups only Asteroidea and Ophi- noidea were lacking on calico shells. FISHERY bulletin: VOLUME 66 NO. 2 Published April 1967. Most of these animals have a casual associ- ation with the sea scallop. They may gain to some extent in having a shell substrate on which to live and possibly by having particles of food brought to them by water currents produced by the scallop. Although the com- mensal may cause little inconvenience to the host, the scallop certainly does not appear to benefit from the association. In fact, the asso- ciation may be detrimental to the scallop if extensive fouling of the shell hinders swim- ming, or if marine borers excavate excessive quantities of the shell ; but usually the associa- tion provides neither advantage nor harm to the participants. An association in which the sea scallop is placed at a distinct disadvantage was observed during routine sea scallop studies. The colonial hydroid, Hydractinia echinata (Fleming), which grows frequently on the external shell surface of the sea scallop, was observed occa- sionally to expand around the shell periphery and invade the internal shell surface. The for- ward elements of the colony, coming in contact with the mantle of the scallop, caused certain inhibitory reactions by the scallop. This paper describes the association, empnasizing the means by which the scallop reacts to internal 273 shell invasion. This type of relation in which one of the associates is inhibited while the other is not affected is best described by the term amensal as defined by Odum (1953). BRIEF REVIEW OF THE LIFE HISTORY OF HYDRACTINIA ECHINATA Hydmctinia echinata is far less exclusive in its choice of habitat than earlier observers indi- cated. It has been dredged "on every sort of bottom" (Sumner, Osburn, and Cole, 1913), and has been found on a wide variety of substrate (Hargitt, 1908). Bunting (1894) mentioned that the hydroid lives on the sea mussel, Mytilns rdulis (Linnaeus), and Moore (1937) recorded its occurrence on the shell of another bivalve, Pectunculus {^Glycymeris) , but I have seen no other reports of the association of this hydroid with a specific bivalve. Hydractinia echinata is the well-known hy- droid which served formerly as the classic example of symbiosis with the hermit crab, Pagurus bernhardus (Linnaeus). Many papers have dwelt on the symbiosis of pagurids and actinians (see Balss, 1924, and Dales, 1957 for a summary of the literature). From experi- mental studies, however, Schijfsma (1935) concluded that the association of H. echinata and P. bernhardus could not be defined as true symbiosi.s — that the hydroid is merely an epizoon. Schijfsma (1939) described the early stages of growth of colonies of the hydroids, and Fraser (1944) and others have reported in de- tail the specialization of individuals that make up a colony of Hydractinia echinata. This hydroid is polymorphic. Several types of zooids develop, including special generative zooids that produce male and female sporosacs. The ova are fertilized in situ, and after being discharged sink to the bottom where they de- velop into mobile planulae in 24 to 48 hours. The planulae are never free-swimming; rather they crawl or glide in the manner of turbella- rians. The mobile phase lasts at least 24 hours, ending when the planulae fix themselves to a substrate and develop into typical tentacular zooids. The surface of a sea scallop shell can act as a base for settling planulae. Once attached. the zooids grow from a stoloniferous network of anastomosing tubes to develop into a colony. The coenosarcal expansion is covered with a heavy, chitinous perisarc from which rise the ridged and jagged spines characteristic of the encrusting colony. Nutritive zooids are the most numerous ; other types include defensive, sensory, and generative zooids. The mature colony appears as a reddish velvety covering, but feels rough to the touch because of the numerous spines. Batteries of nematocysts in the zooids protect the colony. ASSOCIATION OF HYDRACTINIA ECHINATA AND PLACOPECTEN MAGELLANICUS The size of a colony of Hydractinia echinata on the sea scallop varies with the length of time the colony has been established and with the surface area available. Where there is ample surface for expansion, the zooids in the advancing front of the colony tend to be ar- ranged in concentric rows corresponding to the growth marks on the shell. This arrangement was observed by Frederick M. Bayer (personal communication) of the University of Miami, Fla., to whom material was sent for identifica- tion ; his finding agrees with remarks by Schijfsma (1939) that the course of stolons is influenced by the surface sculpture of the substrate. The colony, by advancing more rapid- ly than the scallop grows, may eventually arrive at the periphery of the shell and con- tinue around the shell edge. Long, slender zooids, armed heavily with nematocysts, are especially abundant at the advancing edge. Ultimately, the forward elements of the colony come in contact with the extended mantle of the scallop. The mantle withdraws, presumably because of the nematocysts discharged into it.' The mantle retreats steadily as the hydroid colony encroaches inward (fig. 1, a-d). Evi- dence of this sequence is seen clearly in figure 1, a. To be noted are the numerous relatively • In section, the nematocysts were found to be about 10 microns lone and 4 microns wide and the thread was about at the limit of visibility with an ordinary littht microscope. Thus a thread in the scallop mantle cannot be disliutjuished with the histological technique used. The scallop mantle did, however, show evidence of an in- flammatory response (personal communication. Clyde Dawe. Nation- al Institutes of Health. Bethesda, Md.l. This response was indicated by the greatly increased number of cells under the epithelial margin. Dawe considered these to be wandering amoebocytcs coming to the site of the insult. 274 HYDRACTINA ECHINATA AND SHELL DEFORMITIES OOc: d Figure 1. — Interior and exterior views of a group of upper (left) valves of Placopecten mageUanicus showing shell malformation resulting from inner shell invasion by Hydractinia echinata. Interior views (a-d) illustrate the persistence of the invading hydroid which causes the scallop to retire further within the shell. External views (e-h) show several scallops that have managed to pass over and grow beyond the hydroids. thin, slightly irregular, concentric margins secreted by the mantle. Each margin repre- sents successive retreats of the mantle edge. To resume a normal existence the scallop must successfully overgrow the impinging hydroid colony (fig. 1, eh). Each time the mantle extends towards the margin it secretes conchiolin, and over this foundation the mantle attempts to produce a new edge of shell over the fringe of the hydroid colony. If it succeeds, shell growth resumes once more. Figure lb shows an edge of new shell which indicates a successful bridging by the scallop. A pigment is usually produced in the outer shell layer of the left (upper) valve of the sea scallop. Under some conditions, such as a serious break at the edge of the shell, a scallop secretes new shell material quickly, omitting pigment production until a repair is made and growth becomes normal. The reduc- tion or lack of pigment in scallops as they grow a new lip (fig. 1, e-h) indicates faster rate of shell deposition than usual. Sometimes a scallop is forced to produce a U.S. FISH AND WILDLIFE SERVICE new edge over the hydroid more than once. Figure 2 shows a scallop that had grown several new shell margins. This is an extreme example but it does illustrate the result of difficulties that sometimes confront scallops. A young scallop increases the periphery of its shell faster than an older one (fig. 1, e) and thus has a better chance to stay ahead of the advancing hydroid. Hydractinia echinata does not always occupy the entire external shell surface ; other epi- zoons compete for this space as well. Most colonial epizoons live to the severe exclusion of others ; it is unusual to see one colony over- growing another. Usually a distinct zone of demarcation is formed. This also occurs when two colonies of Hydractinia echinata meet on the same surface (Schijfsma, 1939). I have examined hundreds of colonies of Hydractinia echinata but only once did I find another colonial species growing over the hydroid. This was a granular, encrusting type of unidentified bryozoan spread over the older part of a well- established colony of hydroids. Never was a 275 Figure 2. — Exterior view of a scallop showing the result when a hydroid colony grows faster than the scallop at successive periods and repeatedly overgrows the shell perimeter. -\ c^4.. i r- Figure 3. — Exterior and interior view of the upper (left) valve of a scallop showing the hydroid partially covering and overgrowing the anterior part of the shell (arrows) causing the scallop to change its shape as it grows in a direction away from the hydroid. (Most fouling organisms were cleaned from the shell before the photo was taken; only chitinous growths remain, including the colonial hydroid, worm tubing, and filamentous hydroids.) 276 HYDRACTINA ECHINATA AND SHELL DEFORMITIES colony of Hydractinia echinata seen expanding over another colonial species. Because of competition with other epizoons, the hydroid often can occupy only a relatively small area of scallop shell surface, hence can inhibit scallop growth over only a relatively small area of the mantle periphery. Limited invasion along the anterior or posterior edge of the shell results frequently in a change in symmetry of the growing scallop; the shell tends to grow more rapidly in a direction away from the disturbed region. Figure 3 shows the interior and exterior view of a scallop shell on which a hydroid that overgrew a short segment of the edge produced a shift in the growth axis. (Note the excavations and deterioration of the shell caused by boring annelids.) ■ In several instances, expanding colonies of hydroids extended up the plates of barnacles which were also attached to scallop shells. Were they simply seeking additional substrate or did the water currents, created by the feed- ing barnacles, attract them? Schijfsma (1939) concluded that water currents influence the direction of growth of the colony. Certainly water currents created by a scallop effectively direct growth of a colony toward the shell periphery. Numerous scallops were seen on which the hydroid colony, with equal oppor- tunity to expand in any direction, grew towards the periphery, even obliquely crossing growth lines on the shell in the process. EFFECTS OF THE ASSOCIATION Natural mortality of sea scallops, as deter- mined by the ratio of clapper^ shells to live shells, was uncommonly high in 1960-62 on parts of the important commercial grounds of Georges Bank, off Cape Cod, Mass. (Merrill and Posgay, 1964). We were naturally con- cerned with the possibility that the hydroid might be responsible for part of this mortality ; therefore, during research cruises to Georges Bank in August and September 1961, frequency estimates of Placopecten magellanicus and Hydractinia echinata were made at all stations where the two species occurred together. Also samples of quick-frozen material from areas of high natural mortality and from areas of high hydroid occurrence were taken to the labora- tory for further analyses. A summary of the results of the analyses is shown in table 1. Samples 1 and 2 are from areas of high hydroid-scallop frequency as re- flected by the numbers of hydroids on live and clapper scallops. Samples 3 and 4 are from areas of high clapper-live shell ratios and show a much lower incidence of hydroids. These data make it obvious that the incidence of hy- droids and clappers in the same population of sea scallops is not necessarily correlated. T.\Bi.E 1. — hicidcnce of occurrence of Hydractinia echinata on live and clapper shells nf Placopecten magellanicus [Samples taken during M/V Delaware cruises 61-13 and 61-16] Date Sample number Location Number of scallops with- out hydroids Number of scallops with hydroids Lati- tude W. Longi- tude N. Live Clapper Live Clapper wm Xug. 16 Sept. 24 Aug. 18 Sept. 25 1 2 3 4 41°54.0' 41°53.0' 42°06.r 41°47.1' 6B°39.8' 66°45.9' fi6''40.1' 66''22.4' No. 634 300 1.123 341 No. 4 515 42 No. 208 272 2 30 No. 25 16 1 12 Still more apparent is the lack of correlation in the distribution of Hydractinia echinata and clappers (fig. 4). As indicated by the dark- ened area on the chart in Figure 4, areas of high clapper concentration (clapper ratios over 10 percent) in 1961 generally rimmed the 1 / 1 N • * : •.mo. 42 ''■. ^. A ■' ^ A • ' 41 • ,--- ..< .,-' PERCENT • under S A 5-25 a over 25 t 1 -The ligament iresilium) holds together the upper and lower valves of a scallop for a period of time after the scallop dies. In this state the shell is referred to as a "clapper." 70 W. 68 66 Figure 4. — Chart of Georges Bank showing the distri- bution and density of Hydractinia echinata on Placo- pecten magellanicus (symbols) at Georges Bank in relation to the area encompassing high natural mor- tality in the sea scallop in 1961 (darkened area). Percentage of hydroids on sea scallops indicated. U.S. FISH AND WILDLIFE SERVICE 277 perimeter from the northern to the lower eastern part of Georges Bank, mostly in depths of 40 to 50 fathoms (73 to 92 m.). Hydroids on scallops in this area, however, were dis- tributed in a more or less straight line from the northern to the eastern part. Furthermore, a high incidence of scallop-hydroid association was found in the western part of Georges Bank where natural mortality was relatively low (fig. 4). In summary, it can be stated that Hydrac- tinia echinata inhibits the sea scallop, but no proof has been found that the hydroid causes the death of scallops. ASSOCIATION OF HYDRACTINIA ECHINATA AND OTHER MOLLUSKS One colony of hydroids was found partially covering the shell of a 10 mm. pelecypod, Anomia aculcata Gmelin, that was attached to the shell of the sea scallop. Careful dissection of the perisarc in this area revealed the shell remains of two other A. aculcata, both less- than 2 mm. in diameter, completely covered by the colony. Anomia attaches to a surface by a partly calcified byssal plug that passes through the bottom shell valve. After the animal dies and the soft parts decompose, the valves soon separate and wash away. As the two valves of both animals were still intact, it appears that the young Anomia may have been completely enveloped and smothered by the invader. This circumstantial evidence suggests that mortality of a pelecypod can be attributed to Hydractinia echinata. Gastropod shells also may be occupied by Hydractinia echinata. Dead shells of Nassarius trivittatus (Say), Lunatia hcros (Say), Bucci- num undatum Linnaeus, Acirsa costidata (Mighels and Adams), Cobis pygmaca (Gould) , and Epitoninm greenlandicus (Perry) dredged from the Northern Edge of Georges Bank were covered with Hydractinia echinata. The shell apertures were badly deformed, greatly en- larged, and globose. A survey of a large assort- ment of Nassarius trivittatus in the mollusk collection of the Museum of Comparative Zoology at Harvard University revealed several other specimens almost identical to the dis- figured ones described above. Apparently this phenomenon is not particularly unusual in nature. Until recently I believed that these deformities of gastropod shell apertures were due to adverse relation between gastropods and hydroids (Merrill, 1964). Further field obser- vations and literature research show, however, that I was wrong. The hydroid grows out on the mouth of the shell only after the snail is dead and only when a pagurid crab inhabits the shell. The deformed and enlarged portion of the shell is made of two layers, the lower by glands of the pagurid, the upper by the hydroid (Aurivillius, 1891). This, then, is a symbiotic association. Enlarging the domain is advan- tageous to both animals. One other interesting observation was made regarding the association of Nas.'^ariiitf tririfta- tiis and Hydractinia echinata. In one dredge haul during R/V DELAWARE Cruise 62-7 (station 27, south of Nantucket (lat. 41°11' N.; long. 70 T6' W.) on June 16, 1962, in 15 fathoms (27 m.)) many live specimens of Buccinum undatum were taken. Most of the specimens had colonies of Hydractinia echinata. On top of some of the colonies, A^ tririftatus had deposited masses of egg ca.ses. Evidently the organs necessary for locomotion and egg laying in gastropods are not as sensitive as the mantle in pelecypods to the defensive elements of the hydroid. CONCLUSIONS AND SUMMARY Hydractinia echinata frequently lives as an epizoon on the shell of the sea scallop, Placopccten magcllanicm^. It often expands over and around the margin of the shell and interferes with the normal mantle activity of the host. This interference in turn affects noi'mal metabolism and can cause shell mal- formation. Mortality possibly attributable to this hydroid was noted in a pelecypod, Anomia aculcata, as was deformity of many gastropod shells. Thus, Hydractinia echinata, which normally uses a shell only as a substrate, is capable of becoming a harmful epizoon. Geographic distribution of .scallops in areas of known high natural mortality and areas of high hydroid occurrence were analyzed to de- termine the possibility that the hydroid is an epizootic agent. A lack of correlation indicated 278 HYDRACTINA ECHINATA AND SHELL DEFORMITIES that the hydroid was not the cause of the heavy natural mortality of scallops. LITERATURE CITED AuRiviLLius, Carl W. S. 1891. Uber Symbiose als Grund accessorisches Bildungen bei niarinen Gastropodengehausen. Kongl. Sven. Vetensk.-Akad. Handl. 24: 1-38. Balss, H. 1924. Uber Anpassunger and Symbiose der Paguriden. A. Morph. Okol. Tiere 1 : 752-792. Bunting, Martha. 1894. The origin of sex-cells in Hydractinia and Podocorync; and the development of Hydractinia. J. Morph. 9: 203-236. Dales, R. Phillips. 1957. The interrelation of organisms. A. Com- mensalism. Geol. Soc. Amer., Memoir 67, 1: 391-412. Fraser, C. McLean. 1944. Hydroids of the Atlantic Coast of North America. Univ. Toronto Press, Toronto, 451 pp. Hargitt, Chas. W. 1908. Notes on a few coelenterates of Woods Hole. Biol. Bull. (Woods Hole) 14: 95-120. Merrill, Arthur S. 1961. Some observations on the growth and sur- vival of organisms on the shell of Placopecten magellaniciis. Amer. Malacol. Un., Annu. Rep. 1961, Bull. 28: 4-5. 1964. Observations on adverse relations between the hydroid, Hydractinia echinata, and certain mollusks. Amer. Malacol. Un., Annu. Rep. 1964, Bull. 31: 2. Merrill, Arthur S., and J. A. Posgay. 1964. Estimating the natural mortality rate of the sea scallop (Placopecten jnagellanicus). Int. Comm. NW. Atl. Fish., Res. Bull. 1 : 88-106. Moore, Hilary B. 1937. Marine fauna of the Isle of Man. Proc. Trans. Liverpool Biol. Soc. 50: 1-293. Odum, Eugene P. 1953. Fundamentals of ecology. W. B. Saunders Company, Philadelphia, 384 pp. SCHIJFSMA, KAATJE. 1935. Observations on Hydractinia echinata (Flem.) and Eupaguriis bernhardus (L.). Arch. Neerl. Zool. 1: 261-314. 1939. Preliminary notes on early stages in the growth of colonies of Hydractinia echinata (Flem.). Arch. Neerl. Zool. 4: 93-102. Sumner, Francis B., Raymond C. Osburn, and Leon J. Cole. 1913. A biological survey of the waters of Woods Hole and vicinity. Bull. [U.S.] Bur. Fish. 31: 547-794. Wells, Harry W., and Mary Jane Wells. 1964. The calico scallop community in North Carolina. Bull. Mar. Sci. Gulf Carib. 14: 561-593. GPO 925721 U.S. FISH AND WILDLIFE SERVICE 279 OFFSHORE DISTRIBUTION OF HYDRACTINIA ECHINATA :^^ ^'.\caT cs03 Ho, I"-!-/ I LIBRARY i'.i. By Arthur S. Merrill, Fishery Biologist Bureau of Commercial Fisheries Biological Laboratory Oxford, Maryland 21634 ABSTRACT New distributional records for this hydroid are listed from 81 offshore locations along the Middle Atlantic region and from Georges Bank, off Massachusetts. The species was commonly found living on the shells of ani- mals having some degree of mobility — sea scallops (Placo- pecten magellanicus ) , and gastropod shells inhabited by predominantly gravelly — a mixture of sand, pebble, and shell. Depths at sampling stations ranged from 16 to 80 fathoms; depths where the hydroid occurred ranged from 16 to 62 fathoms. Deeper stations had soft substrates con- taining silt and clay, unsuitable for semimotile animals and lacking the hard substrate necessary for hydroid pagurid crabs. The bottom substrate for these aninoals is colonization. Hydractinia echinata (Fleming:) is circum- polar in its principal distribution (Fraser, 1946). Fraser (1944) documented the distribu- tion of this hydroid in the western Atlantic, mostly in shoal waters from Salmon Bay, Labrador, to Charleston Harbor, S.C. He re- corded numerous locations around the New England coast but none from Georges Bank, although he did include a locality from Georges Basin, near Georges Bank. Only a few scattered and inshore localities were listed for the great Middle Atlantic region. Deevey (1950) ex- tended the range to the Gulf of Mexico. The purpose of this paper is to describe the bathymetric distribution of Hydractinia echi- nata. Eighty-one new locations on Georges Bank, off Massachusetts, and on the continental shelf of the middle Atlantic bight from depths 3f 16 to 62 fathoms (fig. 1), complement Fraser's (1944) records from Cape Cod north- ward. I acquired these data while on research cruises relating to the sea scallop, Placopecten rnagellanicus (Gmelin) (Merrill, 1962; Merrill and Posgay, 1964). This paper is part of a general study to evaluate the significance of an adverse effect of this hydroid epizoon on the commercially important sea scallop (Mer- rill, 1967). The hydroid was found colonizing the shells of live sea scallops and the shells of gastropods FISHERY BULLETIN: VOLUME 66, NO. 2 occupied by pagurid crabs. The scallops and crabs are mobile and inhabit hard rather than soft bottom. Generally, the substrates of sta- tions shallower than 60 fathoms were pre- dominantly sand, pebble, and shell, whereas those deeper than 60 fathoms contained much silt and clay. Depths of the sampling stations ranged from 16 to 80 fathoms. The deepest record for the hydroid was 62 fathoms, which coincided with the greatest depth at which scallops and crabs were taken. Data on the bathymetric range of Hydrac- tinia echinata are S'pa.rse. Verrill (1885) stated it was common from low water to 60 fathoms. Smith and Harger's (1874) greatest depth was 65 fathoms. Fraser (1944, 1946) reported the following deepwater locations: 42''02'15" N., 70°15' N. [sic]. Cape Cod Bay, 362 fathoms; 42°03' N., 70°37' W., 30 miles off Cape Cod light, 106 fathoms ; 52°01' N., 68°00'30" W., off Cape Cod, 86 fathoms. The first two positions are shoal waters, under 25 fathoms. Furthermore, Cape Cod has no water 362 fathoms deep, nor has the w^hole Gulf of Maine. His third location in 86 fathoms is possible. A few records from Georges Bank have been noted in the literature. Smith and Harger (1874) listed Hydractinia polyclina (= Hy- dractinia echinata) from five stations on the Bank near Cultivator Shoal, the Northern 281 Published July 1967 Edge, and Corsair Canyon. The stations on Georges Bank (fig. 1) were made during M/V Delau-are cruise 61-16, September 22-30, 1961. where Hudractinia echinata was found on the shell of Placopecten magcUanicHs (Merrill and Posgay, 1964) in depths from 77 to 136 m. The stations covered most of Georges Bank except the northwest part where few scallops are taken. Several other locations on Georges Bank where hydroids and sea scallops have been found as.sociated are on figure 1. The records are mostly the result of miscellaneous samples brought to the laboratory. I looked for Hydractinia echinata on gastro- pod shells inhabited by hermit crabs during a cruise (M/V Delaware cruise 60-7) along the middle Atlantic from Block Island to Cape Hatteras, mostly in depths of 20 to 80 fathoms. May 11-21, 1960. The purpose of the cruise was to determine the distribution of sea scal- lops and other invertebrates. Pertinent infor- mation regarding the cruise was given by Merrill (1962). Stations where gastropod shells and hydroids occurred together are plotted in figure 1. The hydroids were found in depths from 16 to 40 fathoms on shells of Nassariiis trivittatus (Say), Lutmtia friseriata (Say), Lunatia heros (Say), Colus pygmaeus (Gould), Colus stimpsoni (Morch), and Bucci- MASSACHUSETTS 400 fm- 1 (^ GEORCES «• BANK • ••• * V ^CAPE HATT ERAS _| ± ± k2»N. 40° 38° 36° 76° W. 74° 72° 70° 68° 66° Figure 1. — Distribution of the hydroid, Hydractinia, echinata, in the Georges Bank and middle Atlantic areas. Circles represent locations where the hyroid was taken. 282 OFFSHORE DISTRIBUTION OF HYDRACTINA ECHINATA num undatum Linne which were occupied either by Pagurus bernhardus acadianus (Benedict) or by Pagiirus pollicaris (Say).i ' PaRurids identified by Anthony J. Provenzano, Jr., University of Miami, Miami, Fla. LITERATURE CITED Deevey, Edward S., Jr. 1950. Hydroids from Louisiana and Texas, with remarks on the Pleistocene biogeography of the western Gulf of Mexico. Ecology 31: 334-367. Eraser, C. McLean. 1944. Hydroids of the Atlantic Coast of North America. Univ. Toronto Press, Toronto, 451 pp. 1946. Distribution and relationship in American hydroids. Univ. Toronto Press, Toronto, 464 pp. Merrill, Arthur S. 1962. Abundance and distribution of sea scallops off the middle Atlantic coast. Proc. Nat. Shell- fish. Ass. 1960, 51: 74-80. 1964. Observations on adverse relations between the hydroid, Hydractinia echinata, and certain mollusks. Amer. Malacol. Un. Annu. Rep. 1964, 31: 2. 1967. Shell deformity of mollusks attributable to the hydroid, Hydractinia echinata. U.S. Fish Wildl. Serv., Fish. Bull. 66: 273-279. Merrill, Arthur S., and J. A. Posgay. 1964. Estimating the natural mortality rate of the sea scallop {Placopecten inagcllanicus). Int. Comm. N.W. Atl. Fish., Res. Bull. 1: 88-106. Smith, S. L, and 0. Harger. 1874. Report on the dredging in the region of St. George's Banks, in 1872. Trans. Conn. Acad. 3: 1-57. Verrill, a. E. 1885. Results of the explorations made by the steamer "Albatross," off the northern coast of the United States, in 1883. U.S. Comm. Fish Fish., pt. 11, Rep. Comm. 1883, append. D, art. 16: 503-699. U.S. FISH AND WILDLIFE SERVICE 283 CYCLOPOID COPEPODS OF THE GENUS TUCCA (TUCCIDAE), PARASITIC ON DIODONTID AND TETRAODONTID FISHES By Ju-shey Ho, B.Sc, M.A. Department of Biology, Boston University, Boston, Massachusetts 02215 ABSTRACT The female of Tucca impressus KrjJyer is redescribed, on the basis of specimens taken from Chilomycterus schoepfi (Walbaum) in the Gulf of Mexico. Both genus Tucca Kr(jyer and family Tuccidae Vervoort are redefined, and the genus is treated as monotypic. A restudy of the speci- mens in the U.S. National Museum revealed that T. corpulentus Wilson should be synonymized with T. im- pressus and that the males of T. impressus described by Wilson (1911) are actually some immature adult females of the same species before complete metamorphosis. Metamorphosis occurs only in the cephalothorax and the last two segments of the metasome; the second pedig- erous segment and the urosome remain unchanged. The metamorphosis is widening rather than lengthening in the head, but more lengthening than widening in the trunk. Some geographical variation in size and shape is ob- served in the metamorphosed parts of the body. The three recognized geographical types are: Atlantic type (with slightly bilobed lateral wings of head and less prominent posterior lobes in trunk). Gulf type (with unlobed lateral wings of head and less prominent posterior lobes in trunk), and Caribbean type (with prominent bilobed lateral wings of head and posterior lobes in trunk). This variation is not strictly expressed, however, by every individual in a given geographical range. This study was developed from the identifica- tion of two specimens of immature adult fe- males of Tucca impressus Kr0yer, which were collected from the caudal fin of a spiny boxfish, Chilomycterus schoepfi (Walbaum), at Alliga- tor Harbor, Fla. The specimens were collected by Jack Rudloe and sent to William A. Newman, Museum of Comparative Zoology, Harvard Uni- versity, for identification, and subsequently were passed to me through Arthur G. Humes, Department of Biology, Boston University, in May 1965. Because my observations on these two parasites were so different from the de- scription by Wilson (1911), five more collec- tions were obtained and studied. In addition, I reexamined the specimens in the USNM (U. S. National Museum) which were studied by Wil- son. This reexamination revealed that Wilson (1911) had introduced errors into our knowl- edge of the species of the genus Tucca Kr0yer. The later establishment of a subfamily (by Vervoort, 1962) and family (by Yamaguti, 1963) to contain Tucca is based on the in- formation supplied by Wilson. FISHERY BULLETIN: VOLUME 66, NO. 2 Published May 1967. A redescription of the species and redefini- tion of the genus and the family are given here. Observations on metamorphosis and geograph- ical variation in morphology are also included. The redescription of the female of T. impres- sus given below is mainly based on specimens collected off Cape San Bias, in the Gulf of Mexico, because this collection is the largest of my collections, contains numerous females in various stages of growth, and indicates a cer- tain pattern of metamorphosis. The data given in tables 2, 3, and 4 were prepared from this collection to aid in the explanation of meta- morphosis. After the discovery of a certain degree of geographical variation of T. impressus, tables 5 and 6 were prepared from the two largest col- lections in the USNM, one from North Carolina and the other from Jamaica. Table 4 gives data on the specimens taken from the Gulf of Mexico, which also helps to explain geograph- ical variation. The specimens were dissected and examined in lactic acid, and the figures were drawn with the aid of a camera lucida. 285 SYSTEMATIC ACCOUNT FAMILY TUCCIDAE VERVOORT, 1962 Diagnosis Female. — Metamorphosed cyclopoid. Body composed of head, neck, trunk, and "tail." Head formed by fusion of cephalosome and first pedigerous segment, globular doraally, flattened and hollowed ventrally, and winged laterally. Cephalic appendages and first leg housed in ventral concavity. Neck short, wider than long, formed by second pedigerous segment, dis- tinctly separated from trunk posteriorly. Trunk composed of fused third and fourth pedigerous segments, inflated, much wider than head and neck. "Tail" composed of transformed urosome with all segments completely fused, flattened, wider than long, attached posteroventrally to trunk. Caudal rami small. Eggs multiserate; egg sacs elongate, cylindrical. First ant€nna 5 or 6 segmented, with nu- merous setae. Second antenna 3-segmented; terminal segment armed, in addition to claws and setae, with pectinate, lamelliform process at tip and several rows of teeth or scales over posterior surface. Labrum with marginal teeth ; labium weakly developed. Mandible elongate, with two denticulated spines. Paragnath pres- ent. First maxilla a small, rounded protrusion, bearing four setae. Second maxilla 2-seg- mented, tipped with three denticulated spines. Maxilliped indistinctly 3-segmented, terminal segment strongly bent and pointed. Four pairs of biramous legs ; rami with reduced segments. Leg five, 1-segmented, segment very small, tipped with three setae. Leg 6 absent. Male. — Unknown. Remarks This family contains but a single genus, Tucca Kr0yer, 1837. The genus Tuccopsis Pearse, 1952, which was included in the family by Yamaguti (1963), is synonymous with Blias Kniyer, 1864, of the family Chondracanthidae. This synonymy was first pointed out by Causey (1955: 7) and followed by Vervoort (1962: 93). When Vervoort (1962) reviewed the family Bomolochidae, he included the genus Tucca, fol- lowing Wilson's (1911) opinion, but he set the genus in a new subfamily Tuccinae. Since Ver- voort did not himself examine specimens of the genus Tucca, his accounts on the Tuccinae Ver- voort, Tucca Kr0yer, T. impressus Kr0yer, T. corpulentus Wilson, and T. verrucosus Wilson were wholly based on Wilson's inaccurate ob- servations (see Remarks in the following two sections). Yamaguti's (1963) account was also based entirely on Wilson's descriptions. There- fore, neither the diagnosis of the family Tucci- dae given by Yamaguti (1963: 42) nor the diagnosis of the subfamily Tuccinae given by Vervoort (1962: 92) can be adopted here. The status of the family is then : a redefined family Tuccidae Yamaguti, 1963, embracing within it the redefined and promoted subfamily Tuccinae Vervoort, 1962. Wilson (1911: 353) pointed out that the copepods of the genus Tucca are closely related to the bomolochid copepods, a relation.ship especially suggested by the mouth parts and other cephalic appendages. I consider the fol- lowing characteristics of the female of the genus Tucca, however, so different from those of the bomolochids that Tucca should be placed in a different family: 1. The female undergoes metamorphosis after the last copepodid stage. All known bomolochids (this means all the copepods at- tributed to the subfamily Bomolochinae by Vervoort in 1962) have no metamorphosis, and all have a cyclopoid form of body. In the tuc- cids, however, a metamorphosed adult female has its body distinctly separated into head, neck, and "tail ;" the appearance is not at all cyclopoid. 2. The urosome of the female is rudimentary, its length less than one tenth of the body. The urosome of the bomolochids is always at least one third as long as the body and distinctly 5-segmented ; it comprises a fifth pedigerous segment, a genital segment, and three post- genital segments. Tuccids have a rudimentary fifth pedigerous segment, a genital segment, and a single po.stgenital segment, all fused into one unit and unsegmented. 3. The fifth leg is very rudimentary, merely a small, single segment armed with three .setae. The fifth leg of a typical bomolochid is 2-.seg- mented and consists of a small intermediate 286 U.S. FISH AND WILDLIFE SERVICE segment and a large spatulate, terminal seg- ment ; even in those with a 1-segmented fifth leg, such as the species of Pseudocucanthiis Brian and Orbitacola.r Shen, the free segment is still well developed and spatulate. The terminal, spatulate segment of the bomolochids is usually armed with one spine on the outer surface and two spines and one seta at the distal end. GENUS TUCCA KR0YER, 1837 Diagnosis Type species is Tucca impressus Kr^yer, 1837. Female. — Body form and mouth parts as de- fined for the family. Eggs multiserate ; egg sacs cylindrical, longer than body. First an- tenna 5- or indistinctly 6-segmented, basal segment armed with a strong hook on ventral surface. Second antenna 3-segmented, bearing terminally five weak claws, three setae, and one pectinate, lamelliform process; distal segment covered with teeth posteriorly. Leg 1 biramous, flattened, and 3-segmented, located on posterior wall of ventral concavity in head. Leg 2 bira- mous, 2-segmented. Leg 3 and leg 4 with 2- segmented exopod and 1-segmented endopod ; intercoxal plate missing. Leg 5 very small, a single segment tipped with three setae. Leg 6 absent. Male. — Unknown. Remarks When Kroyer (1837) established this genus, he gave almost no account of the appendages, and neither did Nordmann (1864) in his de- scription of West African specimens that he called T. impressus. Consequently, lacking such information, these authors were inconsistent in the familial attribution of the genus Tucca. Kr0yer placed it in the family Dichelestiidae and Nordmann in the family Chondracanthi- dae. Both Milne-Edwards (1840) and Bassett- Smith (1899) followed Kr0yer's opinion. The nature of the mouth parts of Tucca was not known until 1911, when Wilson studied the specimens of Tucca in the collections of the U.S. National Museum. According to his ob- servations, he placed the genus in the subfamily Bomolochinae of the family Ergasilidae, but later, in 1932, he promoted the subfamily to the familial level. Wilson's additional information on the mor- phology of the .species of Tucca was, however, correct only in the gross anatomy of the mouth parts and not entirely right in the fine struc- tures of the mouth parts and other appendages. I discovered these eri'ors after restudying the specimens of Tucca that had been studied by Wilson in 1911 (the collections from Woods Hole, Mass., and Beaufort, N.C.), in 1913 (the collections from Montego Bay, Jamaica), and in 1932 (the collections from Woods Hole, Mass.). The new species, Tucca corpulentus, described by him, is only a deformed specimen of T. impressus; and some immature adult females of T. impressus were mistaken by him for adult males. As Vervoort (1962: 93-96) and Yamaguti (1963: 43-44) were misled by Wilson's inaccurate observations, their ac- counts of the species of the genus Tucca should be used with reservations. This problem is discussed in more detail in a later section. The specimens described by Nordmann (1864: 491-494, pi. VI, figs. 7-10) as T. im- pressus were claimed by Wilson (1911: 359- 360) to be a new species, to which he gave the name T. verrucosus. I refrain from making any decision on the validity of T. verrucosus without consulting either the original material studied by Nordmann or other specimens col- lected from the same locality (west coast of Africa) and the .same host (Diodon sp.). If Wilson's assumption is correct, then T. ver- rucosus would naturally be the second species of the genus ; however, I now prefer to treat the genus as monotypic. A doubtful form, Tucca sp., was introduced to the genus by Pearse (1952: 12, figs. 23-27). This species, however, has been questioned by Causey (1955: 11) as being probably a muti- lated specimen of Blias prionoti Krttyer, 1864. The mandible of Pearse's Tucca sp. is very convincing evidence that it is not a tuccid. Its foi'm of a "slightly curved hook" indicates a chondracanthid type of mandible rather than a tuccid type. TUCCA IMPRESSUS KROYER, 1837 Tucca impressus Kr0yer, 1837, pp. 479-482, pi. V, fig. 2(a-h). Milne-Edwards, 1840, p. 496. Bassett-Smith, 1899, p, 469. Wilson, 1908, p. CYCLOPOID COPEPODS OF GENUS TUCCA 287 625; 1911, pp. 354-387, pi. 48, figs. 102-108, pi. 49, figs. 109-115, 118-120; 1913, p. 200; 1932, pp. 379-380, fig. 243 (a,b). Bere, 1936, p. 582. Heegaard, 1947, pi. 25, fig. 195. Sewell, 1949, p. 157. Carvalho, 1951, p. 136. Pear.se, 1952, p. 191. Causey, 1955, p. 3. Vervoort, 1962, pp. 93-95. Yamaguti, 1963, p. 43, pi. 47, figs. l(a-k). Tucca corpulentus Wilson, 1911, pp. 358-359, pi. 49, figs. 116, 117, pi. 50, figs. 121-127; 1932, pp. 380-381, fig. 235 (a.b). Heegaard, 1947, pi. 25, fig. 194. Sewell, 1949, p. 157. Veervoort, 1962, pp. 95-96. Yamaguti, 1963, p. 43, pi. 46, fig. l(a-g). Material Examined Two immature adult females from caudal fins of 2 CkUnmyctcrus schoepfi. caught in mullet seine, at Alligator Harbor, Fla., March 1965; 7 ovigerous females, 3 immature adult females, and 1 copepodid from fins of three C. schoepfi, caught in gill net, at Panacea, Fla., May 14, 1965; 44 ovigerous females, 9 immature adult females, and 2 copepodids taken from 14 C. schoepfi, caught in shrimp trawl, off Cape San Bias, Fla., May 16, 1965; 6 ovigerous females on dorsal and pectoral fins of 2 C. schoepfi, caught in shrimp trawl, off Carrabelle, Fla., July 18, 1965; 9 ovigerous females from fins and body surface of C. schoepfi, caught in shrimp trawl by R/V Oregon, off St. Simons Island, Ga., November 17, 1965. In addition to the above collections, I examined the following 16 collections in the USNM (the host names for USNM 38619, 38628, 47748, and 74375 are here changed from C. geometricus to C. schoepfi) : 6090 — 3 adult females "from exterior surface of rouRh swellfish, P. Stewart's pound," Woods Hole, Mass., July 26, 1882. 38369 — 3 adult females and 3 immature adult females from fins of C. schoepfi, collected in Louisi- ana by M. H. Spauldinp, August 10, 1907. 38625 — 8 adult females and 3 immature adult fe- males from fins of C. schoepfi, collected at Beaufort, N.C., in 1904. 38627 — 8 adult females from pectoral fins of C. schoepfi, collected at Beaufort, N.C., in 1905. 38628 — 11 adult females from pectoral fins of C. schoepfi, collected at Beaufort, N.C., in 1902. 42251 — 7 adult females from fins of C. antennatus (Cuvier), collected at Montego Bay, Jamaica, June 22, 1910. 42264 — 5 adult females from pectoral fins of Diodon hystrix Linnaeus, collected at Montego Bay, Jamaica, June 22, 1910. 42265 — 2 immature adult females from pectoral fins of Spheroidcs »iar»u>ratt(S (Ranzani), col- lected at Montego Bay, Jamaica, June 20, 1910. 42269 — 1 deformed immature female on fin of S. marmoratus, collected at Montego Bay, Jamaica, September 15, 1910. 42273 — 57 adult females, 1 immature adult female, and 1 copepodid from fins of C. atitennatus, collected at Montego Bay, Jamaica, June 15, 1910. 47748 — 19 adult females from C. schoepfi, collected at Morehead City, N.C., April 7, 1891. 53525—5 adult females from Beaufort, N.C. (no host or date given). 74375 — 7 adult females on fins of C. schoepfi, collect- ed at Beaufort, N.C, August 1905. 79089 — 2 adult females under pectoral fin of C. spitiosus (Linnaeus), collected at Lemon Bay, Fla., in 1934-35. The following two collections from USNM are labeled as Tiicca corpulentus: 38619— "Type," 2 adult females (1 decapitated) from fins of C. schoepfi, collected at Woods Hole, Mass., in 1887 (see Remarks). 79595 — 3 adult females on gill of S. maculatus (Bloch and Schneider), collected at Woods Hole, Mass., by G. A. Maccallum (no date given; see Remarks). Distribution See table 1. Table 1. — Hosts and distribution of Tucca impressus ' Host Family Tetraodontidae Spheroides maculatus. S. marmoratus Family Dlodontldae Diodon hystrix. Ctiilomyctfrus antennatus. C\ schoepfi Locality C. spinosus.. Woods IIolc, Mas.s Montego Bay, Jamaica, Danish West Indies -Montego Hay, Jamaica. Montego Bay, Jamaica. Woods Hole, Mass. Beaufort, N.C Louisiana Moreliead City, N.C... Beaufort, N.C Sao I*auIo, Brazil Alligator Harlior, Fla... rascagoula. .Miss Panacea, Fla Cape San Bias, Fla Carral)elle, Fla St. Simons Island. Ga... Lemon Bay, Fla Collection of USNM 6090-. USN.M 422fi5. USNM 42269. Authority USNM USNM USN.M USNM US.NM USNM USNM USN.M USNM USNM 422f)4. 422.'il . 42273. 38fil(l. 3862.'!. 3sr,'.!7. 3SfV_'H 3H3li9. 47748. 74375. Author. do.. do... do.. do... USNM Wilson (?) Uo. (1913) Do. (1913) Kr0ver (1837) Wilson (1913) Do. (1913) Do. (1913) Do. (1911) Do. (1911) Iin. (1911) Do. (lull) Do. (71 Do. (?) Do. (?) Carvalho (1951) I'earse (19.52) Causey (1955) Present paper Do. paper Do. paper Do. pajier Bere (1936) ' Nordmann's record of Tucca impressus from the west coast of Africa is excluded, because of its uncertain identification. USNM 53525 Is also not Included because the host was not Identified. 288 U.S. FISH AND WILDLIFE SERVICE Figures 1-8. Tucca impresstis, female, from Chilomyctenis schoepfi taken off Cape San Bias, Fla., in the Gulf of Mexico. The letter in the parentheses after the explanation of each figure refers to the scale at which the figure was drawn. 1. Entire, dorsal (A). 2. Head, neck, and anterior end of trunk, showing relative position of various cephalic appendages, ventral (B). 3. "Tail" (= urosome), ventral (C). 4. Genital segment, showing egg sac attachment area, dorsal (D). 5. Caudal ramus, ven- tral (E). 6. First antenna, exterior (E). 7. Second antenna, exterior (E). 8. Lamelliform process at tip of second antenna, interior (F). (a'. =first antenna; a". = second antenna; md.=mandible; mx'. =first maxilla; mx".=second maxilla; mxpd.=maxilliped; p. = paragnath; p,. = leg 1; p2.=leg 2). 3YCL0P0ID COPEPODS OF GENUS TUCCA 289 Description of Stages Three stages are described below: mature ovigerous female, immature adult female, and female copepodid. Mature ovigerous female: — Body (fig. 1) noncyclopoid, 1.51 to 2.92 mm. long, composed of head, neck, trunk, and "tail." Head (fig. 2) small, 0.46 by 0.71 mm., representing a fusion of cephalosome and first pedigerous segment; inflated dorsally, flattened ventrally (fig. 25), and with two wide lobed wings laterally, which in a fully grown adult female protrude beyond anterior margin of cephalosome. Ventral sur- face of head deeply invaginated at center, form- ing a hollow disk (fig. 2) which is reinforced anteriorly by rostrum and bases on first anten- nae and posteriorly by flattened leg 1. Second antennae, mouth parts, and maxillipeds found on bottom of this disk. Rostrum (fig. 2) well de- veloped, bearing some refractile points, two sclerotic protrusions, and one fairly strong hook pointing posteroventrally. Head jointed to trunk by a short neck (fig. 2) which is formed by second pedigerous segment ; this segment completely fused with head anteri- orly. This portion of body highly variable in length in diflferent individuals, fully extended in some specimens and completely contracted in others, leaving practically no space between head and trunk. Trunk (fig. 1) made up of third and fourth pedigerous segments, 1.54 by 1.31 mm., nearly square, with rounded corners. Posterior corners slightly produced on each side into one dorsal and one ventral lobe (fig. 25), and a third lobe produced from po.steromedial end of trunk between two dorsal lobes. Four depres- sions (fig. 1) on dorsal surface of trunk. "Tail" (fig. 3) flattened, 0.17 by 0.26 mm., attached to trunk posteroventrally (fig. 25), totally or par- tially concealed by dorsal posteromedial lobe. No appreciable segmentation seen in "tail," which apparently represents a fusion of the narrow fifth pedigerous segment, the circular genital segment, and one small postgenital seg- ment. Egg sac attachment area (fig. 4) well developed, occupying about two-thirds of lat- eral surface of "tail." Caudal ramus (fig. 5) small, 23 by 20 ,i, armed with five short setules and one long seta 114 ,* long. Egg sac elongate, cylindrical; fully grown sac longer than body, containing numerous small eggs with a dia- meter of 90 /!(. Many micropits on surfaces of head and trunk, as shown in detail in figs. 2 and 4, penetrating deeply into sclerotic cover- ing of body. First antenna (fig. 6) distinctly 5-segmented, but second segment suggesting a division into two segments. Armature on these five seg- ments, from proximal to distal: 15 + 1 hook (on ventral surface), 8, 3, 3, + 1 aesthete, and 7 + 1 aesthete. Second antenna (fig. 7) 3-segmented ; basal segment longest, naked; second segment bear- ing one seta. Terminal segment having sub- terminally a rod-shaped process and one pec- tinate, lamelliform process (fig. 8), and carry- ing terminally three setae and five weak claws ; several rows of teeth on posterior surface. Labrum (fig. 9) well developed, with fine teeth on posterior free margin; labium weak. Mandible (fig. 10) compo.sed of a large plate produced into long process, armed with one terminal masticatory process and one sub- terminal, bilaterally denticulated spine. Parag- nath (fig. 11) bearing setules, located postero- medially to mandible. First maxilla (fig. 12) a small rounded protrusion, located laterally to paragnath, bearing four setae, one of which is fairly long. Second maxilla (fig. 13) 2-seg- mented, terminal segment armed with three bilaterally denticulated spines. Maxilliped (fig. 14) powerfully developed, indistinctly 3-segmented; terminal segment strongly bent inward and almost perpendicular to first two segments. Last segment sharply pointed, with well-developed sclerites cutting into ventral and posterior surfaces, thus mak- ing these surfaces corrugated. In many speci- mens, the terminal, pointed process, broken when the parasite was removed from the host, appeared as a blunt process. Formula of spines and setae on first four pairs of legs as follows (Arabic numerals rep- resent setae and Roman numerals spines) : LCK 1 Leg 2 Leg 3 Leg 4 Protopod 00 10 00 l-I 00 10 00 10 Exopod 10 1-1 7 10 IIM-5 10 1111-5 10 III-5 Endopod 0-1 0-1 5 0-1 7 1 1 Leg 1 (fig. 15) strongly flattened, its setae 290 U.S. FISH AND WILDLIFE SERVICE Figures 9-20. — Tucca impressus, female, from Chilomycterus schoepfi taken off Cape San Bias, Fla. in the Gulf of Mexico. 9. Labrum and labium, ventral (G). 10. Mandible, posterior (G). 11. Para^ath, ventral (G). 12. First maxilla, anterior (G). 13. Second maxilla, anterior (G). 14. Maxilliped, posteroventral (D). 15. Leg 1, anterior (D). 16. Leg 2 and intercoxal plate, anterior (E). 17. Leg 3, anterior (G). 18. Leg 4, anterior (G). 19. Leg 5, ventral (F). 20. Copepodid, dorsal (B). CYCLOPOID COPEPODS OF GENUS TUCCA 291 Figures 21-27. — Tiwca hnpressus, female. Figures 21-23, from Chilomycterus sehoepfi taken oflf Cape San Bias, in the Gulf of Mexico; 24 25, from C. sehoepfi taken off St. Simons Island, Ga.; 26-27, from C. antennatus at Montego Bay, Jamaica. 21. Immature adult with short trunk, dor.sal (A). 22. Imma- ture adult with guitar-shaped trunk, dorsal (A). 23. Immature adult with circular trunk, dorsal (A). 24. Entire, dorsal (A). 25. Same, lateral (egg sacs omitted) (A). 26. Entire, dorsal (egg sacs omitted) (A). 27. Entire, showing absence of anterior lobes in trunk (egg sacs omitted) (A). 292 U.S. FISH AND WILDLIFE SERVICE densely haired. Medial surface of basis produced into blunt process, covered with hairs; mar- ginal surfaces of every segment in each ramus also covered with hairs. Spines on outer sur- faces of exopods of leg 2 (fig. 16), leg 3 (fig. 17), and leg 4 (fig. 18) weakly developed, seti- form, and naked. Intercoxal plate absent in leg 1, leg 3, and leg 4; coxa and basis not com- pletely separated in leg 3 and leg 4. Leg 5 (fig. 19) uniramous, very rudimentary, 10 by 9 /x; located at junction of "tail" and trunk and armed with three long setae. Leg 6 absent. hnmature adult female. — Body (figs. 21-23) noncyclopid, shaped differently in different stages of metamorphosis; proportions of vari- ous body regions also different in different stages (see table 2) . Table 2. — Measurements of immature adnlt females taken from Chilomycterus schoepfi off Cape San Bias, Fla., in the Gulf of Mexico. Specimen number Head Neck Trunk "Tall" Total length 1 Mm. 0.28 by 0.44 .36 by 0.49 .34 by 0.49 .31 by 0.49 .33 by 0.52 .30 by 0.47 .34 by 0.55 .34 by 0.57 .31 bv 0.56 Mm. 0.08 .09 .10 .10 .10 !io .09 11 Mm. 0.47 by 0.47 .45 by 0.49 .51 by 0.57 .74 by 0.56 .70by0.54 .82 by 0.60 .73 by 0.56 .75 by 0.60 .85 by 0.86 Mm. 0.17 by 0.24 .17 by 0.23 .17 by 0.27 .17 by 0.25 .18 by 24 Mm. 96 2 1.07 3- 4 S 1.12 1.13 1 M 6 .17by0.23 1.31 .18 by 0.23 , 1.33 .17 bv 0.26 1 1.35 7 8 9 . .16 by 25 1 1 43 Average. 0.32 by 0.51 0.10 0.67 by 0.64 0.17 by 0.25 1.21 Structure of appendages similar to mature ovigerous female. Details of these immature adult females are given in following section in discussion of metamorphosis. Female copepodid. — Body (fig. 20) cyclopoid, 0.70 by 0.37 mm. (excluding setae on caudal rami) ; no segmentation on cephalothorax and urosome, but with clear distinction between each two adjoining regions of four body regions. Cephalothorax semicircular anteriorly and rather truncated posteriorly; posterior sur- face roughly separated into dorsal and ventral portions. Second pedigerous segment (— neck of adult), 0.08 by 0.26 mm., attached to center of posterior surface of cephalothorax, carrying leg 2 ventrally at anterior margin. Third pedi- gerous segment, 0.13 by 0.26 mm., slightly in- vaginated on both sides and incompletely sepa- rated from fourth pedigerous segment on dor- sal surface. Fourth pedigerous segment, 0.12 by 0.26 mm., with posterior margin protruding over about one-third of urosome. Urosome (= "tail" of adult) 0.17 by 0.23 mm., carrying inside a pair of seminal receptacles (or cement glands ?) . Egg sac attachment area similar to that in adult. Caudal ramus attached to posteroventral surface of urosome, its armature as in adult. Two sclerites on dorsal surface of third and fourth pedigerous segment. Micropits present on body surface (omitted in fig. 20). All appendages similar to those in adult ovigerous female and immature adult female. Remarks In the vial labeled Cat. No. 38619 in the collection of USNM are two specimens (one decapitated) designated by Wilson (1911) as the type specimens of Tucca corpulentus. The trunk of the headless specimen appears like the one shown in fig. 24, namely, squarish and distinctly 3-lobed on its posterodorsal surface. The head of this specimen was supposedly dis- sected by Wilson for study of the mouth parts and other cephalic appendages, and probably was the source of his figs. 122-125. The other specimen (with head) is, doubtlessly, the source of his fig. 121. I have examined the latter speci- men with great care in lactic acid. Neverthe- less, I was not able to find any appendages that are significantly different from those described above. In addition, the posterodorsal surface of the trunk is also 3-lobed, not as smooth as illustrated by Wilson in his fig. 121. The cir- cular appearance of this specimen is possibly due to the fact that the parasite was somewhat pressed (by the fin, on which the parasite was attached, pressing against the body surface) before preservation, because its trunk appears unusually thin. The absence of pits or impres- sions on body surface, one of the characters cited by Wilson for establishing the new species, is conceivably also due to mechanical deformation prior to preservation. Consequently, as far as these two type speci- mens are concerned, T. corpulentus does not differ from T. impressus and should be synony- mized with it. There are some inconsistencies CYCLOPOID COPEPODS OF GENUS TUCCA 293 between the label of Cat. No. 38619 in the USNM and the statement of Wilson (1911: 359) : "There is but a single lot of this species, which was taken from the northern swell-toad, Spheroides maculatus, at Woods Hole, Massa- chusetts, and is numbered 38619, U.S.N.M. It includes three females, two of which bear egg- strings." The label of Cat. No. 38619 clearly says, however, that there are only "2 5 speci- mens," the host is "Chilomycterus geometri- cus." and no egfr strings were found in the vial. Another USNM collection of T. corpulentus is Cat. No. 79595. The label of this collection fits better with Wilson's statement. It says that there are "3 specimens" and the host is "Gills, Spheroides maculatus," but the three specimens of this collection are Pseudochondracanthus diceraus Wilson. They are mature adult females and all carry a pygmy male on their poster- oventral surface. This collection was not men- tioned by Wilson in any of his reports, not even in his reports of P. dicemus (1908: 436; 1932: 496), but the label says "Identified by C. B. Wilson." I have taken the two specimens kept in the vial of Cat. No. 38619 as Wilson's type specimens of T. corpulentus and synonymized the species with T. impressus. One of the three immature adult females in the vial of Cat. No. 38625 was obviously mis- taken by Wilson for an adult male. The rather small size, the different shape and proportion of various body regions, and the two bean-shaped reproductive organs inside the urosome might suggest incorrectly a male, if the process of metamorphosis in the female is unknown. The pair of stout hooks described by Wilson on the ventral surface at the posterior corners of the genital segment of this "male" specimen are merely two sclerotic protrusions (see fig. 3). NOTES ON METAMORPHOSIS The absence of the male parasites on the diodontid and tetraodontid fishes perhaps oc- curs because males do not grow beyond the copepodid stage. They probably die after copu- lation as do the males in the families Lernae- idae, Lernaeoceridae, and Pennellidae, in which only the female copepodid (after copulation) attaches to the fish host and metamorpho.ses into an adult. The two youngest females recovered from the diodontid fish caught off Cape San Bias, Fla., still show a cyclopoid form of body; they are particularly remini.scent of bomolochid and taeniacanthid copepods (see fig. 20). The cephalothorax is the widest part of the body, and the metasomal segments are still distin- guishable (see table 3 for measurements). These features, in comparison with the meta- morphosed adult female, indicate that they are either still in the last copepodid stage or, at most, just on the way to metamorphosis. The somewhat older females that I have in the same collection are the nine copepods that show no segmentation in the metasomal region, have swollen trunks as wide as the head or a little wider, and carry no egg sacs (see figs. 21, 22). T.iBLE 3. — Measurements of female copepodid from three collections. Record and body part measured Specimen 1 Host Locality Date Cephalothorax ihead) Thorax (neck + trLink) Urosome ("tail") Total length.-. Mm. Chilo- mycterus schoepfi Panacea, Fla. May 14, 1965 0.27 by 0.37 .30 by 0.26 .17 by 0.23 .70 Specimen 2 I Specimen 3 Specimen 4 Mm. I Mm. Mm. Chilo- Chilo- I Cliilo- jjiycteriis \ myclerus myctfrus schofpfi I schofpft j anifnnntus Cape San ' Cape San Montego ma.s, Fla, Hlas, Fla. May 16, 196,1 0.31 by 0,41 ,34 by 0.32 ,16 by 0.23 ,81 May 16, 1965 0.30 by 0.42 .36 by 0.31 ,17 by 0,21 .83 Hay, Jamaica June 15, 1910 0.31 by 0.47 ,41 by 0,37 .15 by 0,23 These females I have considered as the imma- ture adults inasmuch as they have attained ■sexual maturity and have copulated but have not yet produced ef^g sacs. Metamorphosis occurs only in the cephalo- thorax and the last two segments of the meta- some. As far as the size and shape are con- cerned, the second pedigerous segment and the urosome in the copepodid are not significantly different from the neck and the "tail" in the immature adult female, nor in the ovigerous female. The second thoracic segment, urosome, and all appendages are not transformed during metamorphosis, but the cephalothorax and the third and fourth pedigerous segment are tre- mendously changed. The size of the head of an immature adult 294 U.S. FISH AND WILDLIFE SERVICE female (0.32 by 0.51 mm.) is not much differ- ent from that of the cephalothorax in the cope- podid (0.31 by 0.42 mm.) ; the shape, however, is markedly different. The expansion is seen mostly in anterior corners, posterior subcorn- ers, and the dorsal surface of the head. The head of an ovigerous female (0.33 by 0.59 mm.) differs from that of the immature adult female chiefly in the more globular appearance of the middorsal surface; it is not lengthened but definitely widened. The metamorphosis in the head involves changes in form from semicir- cular to rectangular (in dorsal view) and from slightly convex to globular (in lateral view of the dorsal surface) . The amount of increase in proportions of the head is about 10 percent in the length and 45 percent in the width ; this widening rather then lengthening during meta- morphosis is due to the formation of the lateral wings. The second and third pedigerous segments are completely fused into a unit at the onset of metamorphosis (fig. 21). This fused trunk is then enlarged in three dimensions, the shape (in dorsal view) changes from oval (as in fig. 21) to guitar-shaped (as in fig. 22) or nearly circular (as in fig. 23) and then to squarish (as in fig. 1 ) . The posterior lobes, three on the dorsal and two on the ventral surface, are not formed in the immature adult female. The four chitinized platelets on the dorsal surface of the thorax of the copepodid are retained through- out metamorphosis. As these platelets are the points of attachment of trunk muscles on the tergum of the second and third pedigerous seg- ment, they have not been elevated by the en- larging action in the course of metamorphosis. Thus, the four platelets form the bottom of the "four pits" on the dorsal surface of the oviger- ous female. The amount of increase in pro- portions of the trunk is about 270 percent in the length and 120 percent in the width. The metamorphosis of the trunk, contrary to that of the head, involves more lengthening than widening. Specimens 4, 5, 6, 7, and 8 of table 2 have trunks distinctly longer than wide; they look like that in fig. 22. The remaining four speci- mens (1, 2, 3, and 9) of the immature adult females in the same collection have trunks nearly as long as wide and resemble fig. 23. In the present state of knowledge, we can say only that immature adult females have two forms. Which form comes first in the process of meta- morphosis is unknown. A comparison between table 2 and table 4 shows that the maturity of the females can be judged by the size of the trunk, in addition to the presence or absence of egg sacs. The trunk is definitely longer and wider in ovigerous fe- males than in immature adults, although the size of the head overlaps broadly in the two stages. The first ovigerous female in table 4 has smaller body length, but a definitely larger trunk, than the largest immature adult female in table 2. As noted in the previous section, the neck of this ovigerous female is unusually shrunken ; therefore, body length alone is not a good measure for determining the maturity of a female. GEOGRAPHICAL VARIATION According to our present knowledge of para- sitic copepods of fishes, Tucca impressus is para- sitic exclusively on two families of fishes, Tetra- odontidae and Diodontidae — especially the fishes of the latter family (porcupine fish or boxfish) . Our past records show that it is most abundant on the fishes of the genus Chilomycterus (Dio- dontidae) and always found either on the fins or on the body surface. A certain degree of variation is observed in the head and the trunk of the ovigerous females collected from three different areas, namely the west coast of North Atlantic Ocean, the Gulf of Table 4. — Measurements of smallest, largest, and eight randomly selected ovigerous females taken from Chilo- mycterus schoepfi off Cape San Bias, Fla., in the Gulf of Mexico Specimen number 1... 2 3 4 5 6 7 8 9 10 Average, Head Mm. 0.29 by 0.50 .30 by 0.61 .33 by 0.55 .31 by 0.56 .31 by 0.64 .30 by 0.60 .35 by 0.56 .36 by 0.65 .38 by 0.59 .33 by 0.65 Neck Mm. 0.06 .11 .09 .09 Trunk •Tall" Mm. 1.30 by 0.95 1.10 by 1.22 1.12 by 0.93! 1.19 by 1.02i .09;1.14by 0.87j .08|1.33by 1.12 .10 1.30 by 1.40 .09|l.31 by 1.31 .11 1.28 by 1.48 .10 1.59 by 1.14 0.33 by 0.59 0.09 1.24 by 1.17 0.17 by 0.26 Mm. 17by 0. 16by0. 17by 0. 18by0. 19by0. 18by 0, 18by 0. 18by 0. 16by 0. 18by 0. Egg sac Mm. 0.84 1.29 .75 1.10 broken broken 1.34 1.84 broken 1.41 Total length Mm. 1.38 1.51 1.54 1.59 1.68 1.71 1.75 1.77 1.78 2.02 1.67 CYCLOPOID COPEPODS OF GENUS TUCCA 295 Table 5. — .\feasurements of smallest, largest, and eight randomly selected ovigerous females taken from Chilo- mvcterus schoepfi at Morehead City, North Carolina (from USXM 4774s) 1 Specimen number Average. Head Mm. 0.38 by 0.67 .30 by 0.78 .35 by 0.96 .37 by 0.91 .31 by 0.84 .34 by 0.88 .36 by 0.85 .37 by 0.91 .38 by 0.94 .40 by 0.89 Neck Trunk Mm. 0.10 1, .09 2. .08,2, .12 2. .13 2. .11 2. .12 2. .09 2. .112. .12 2. Mm. "Tail" Mm. 76 by 1.65 0.18 by 0.24 Egg sac 02 by 1.60 15 by 2.06 30 by 1 . 79 29 by 1.89; 50 by 1.98' 39 by 1.95 51 by 1.86: 49 by 2.42, 46 bv 1.95 (2.79) 17 by 0.23 19 by 0.24 16 by 0.25 .16 by 0.261 .16 by 0.24' .17 by 0.27 .18 by 0.26 .19by 0.27i .17by0.25| (4.61) Total length 0.36 by 0.86; 0.1112.39 by 1.92 0.17 by 0.25 . Mm. 2.24 2.47 2.63 2.79 2.80 2.90 2.97 2.98 3.09 3.16 2.80 ' The egg sacs were found free in the vial. Since there is no way to identify each sac with its female, only the shortest and the longest sacs were mea,sured. T.\BLE 6. — ^feasurements of smalle.'it, largest, and eight randomly selected ovigerous females taken from Chilo- mvcterus antennatus at Montego Bail, Jamaica (from USXM 4227S) Specimen number Head Mm. 0.31 by 0.48 .32 by 0.54 .33 by 0.60 .35 by 0.56 .34 by 0.58 .35 by 0.59 .34 by 0.56 .33 by 0.66 .36 by 0.66 .34 by 0.68 Average.. 0.34 by 0.59 Neck Trunk •Tail" Egg sac Total length Mm. Mm. Mm. Mm. Mm. O.irO. 93 by 0.83 0.17 by 0.26 0.68 1.42 .09 .96 by 0.78 .19 by 0.27 broken 1.46 .091.15l)yl.l3 .18by0.25 1.02 1.S9 .10 1.29 by 1.10 .16 by 0.27 1.77, 1.75 .08 1.42 by 1.22 .18 by 0.26 broken! 1.78 .11 1.35 by 1.29 .16 byO. 23 1.59 1.81 .12 1.41byl.26 .15by0.24' 2.051 1.83 .09 1.47 by 1.27i .18 by 0,26 2.14' 1.95 .10 1.46 by 1.36 .18 by 0.25 broken' 2.00 .11 1.67by 1.41J .17 by 0.26 broken! 2.14 0.10!l.31by 1 ,17,0. 17 by 0.26 1.77 Mexico, and the Caribbean Sea. This variation occurs only in the metamorphosed parts of the body, and is in the size and the shape. A com- parison of fig. 1 (a representative from the Gulf of Mexico), fig. 24 (a representative from the west coast of North Atlantic Ocean), and fig. 26 (a representative from the Carib- bean Sea) together with reference to tables 4, 5, and 6 shows this picture of geographical variation. In the following discussion, for the sake of convenience, the specimens from Georgia, North Carolina, and Massachusetts are termed as the Atlantic type ; the specimens from Florida (west coast), Mississippi, and Louisiana, the Gulf type; and the specimens from Jamaica, the Caribbean type. The bilobed condition of the lateral wings of the head is generally most pronounced in the Caribbean type (figs. 26, 27), but the wings are almo.st unlobed in the Gulf type (fig. 1). The lateral wing of the Atlantic type (fig. 24) is only slightly bilobed; the posterior lobe is larger than the anterior lobe and is wider than those in the other two geographical types. In both Gulf type and Atlantic type, the pos- terior lobes in the trunk are usually less pro- nounced, and there are no anterior lobes. These anterior and posterior lobes are, however, present and well formed in the Caribbean type. A fully grown ovigerous female of an Atlantic type is much larger than those of the Gulf type and the Caribbean type. The following data were derived by considering all collections from a general geographical region as a whole to show the size ranges (in millimeters) of the ovigerous females of the three different geo- graphical types : Smallest., Largest. Longest egg sac. Caribbean type Mm. 1.36 (in USNM 42251) 2.14 (in USNM 42273) 2.78 Gulf type Atlantic type Mm. 1.51 (oil Cape San Bias, Fla.) 2.51 (Carrabelle, Fla.) 4.09 Mm. 1.59 (in USNM 38625) 3.16 (In USN.M 47748) 4.61 Thus, the shape of the trunk indicates that the Gulf type is closer to the Atlantic type than to the Caribbean type, but the size of the trunk indicates that the Gulf type is, on the contrary, closer to the Caribbean type than to the Atlan- tic type. In other words, comparisons of the trunk show that the Gulf type is intermediate between the Atlantic type and the Caribbean type. The variation of the head, in the Atlantic type, instead of the Gulf type, shows the inter- mediate character in the bilobed condition of the lateral wings. I have found specimens in USNM collections (Cat. No. 38625 and 74375), from Beaufort, N.C., which, instead of having the Atlantic type trunk, have the po.sterior lobes of the trunk fairly well defined as in the Caribbean type. Moreover, in the collections from Jamaica, some individuals lack the anterior lobes in the trunk, as shown in fig. 27. It appears, therefore, that the variation in the head and the trunk is not absolute, or, in other words, that this variation is merely a general tendency of modification that exists in a certain geographical area but is not strictly expressed by every individual of 296 U.S. FISH AND WILDLIFE SERVICE this species found in a given geographical range. As the hosts of this parasitic copepod are mostly inshore fishes and not powerful swimmers, considerable distant movement probably is accomplished only by drifting with the current. At present, however, it is impos- sible to determine whether the Gulf Stream has influenced this picture of geographical varia- tion. The single specimen of T. imipressns de- scribed by Krdyer (1837) is a female taken from the inner surface of the pectoral fin of a Diodon hystrix in the Danish West Indies. It definitely belongs to the Caribbean type, since in Kroyer's fig. 2a (dorsal view) and fig. 2b (lateral view) the posterior lobes, anterior lobes, and the bilobed condition in the head are of that type. According to Kroyer's description (p. 479) this Danish West Indian specimen measures 2 lines, of which the egg sac is about half. In other words, the length of the para- site's body is about 2.11 mm., which falls within the range of the Caribbean type (see table 6). The 37 specimens of T. impresses described by Carvalho (1951) from Brazilian C. schoepfi measure from 1.52 to 1.80 mm., and so fall with- in the range of the Caribbean type. In his discussion of the validity of Nord- mann's T. impressus, Wilson (1911: 359) ex- pressed his doubt upon the variation of the specimens of T. impressus: "either Nordmann's species or that of the present author is new to science. They can not both be identical with Kroyer's T. impressus." This implies that Wil- son's specimens are different from Kroyer's T. irnpressiis to a certain degree, but this dif- ference is not as significant as the discrepancy between Krdyer's T. impressus and Nord- mann's T. impressus. Consequently, Wilson identified his specimens collected in Beaufort, N.C., as T. impressus, and created T. verru- C0S7IS for Nordmann's T. impressus. The total length (1.67 mm.) given by Wilson (1911: 356) for the species of T. impressus is too small for the Atlantic type. I have measured all 30 specimens that were identified by Wilson as T. impressus in USNM collections. The col- lections, number of specimens, and maximum sizes are : Catalogue Number of Smallest Largest number specimens (mm.) (mm.) USNM 38625.. 11 (3 immature) 1.13 1.69 USNM 38627.. 8 2.04 2.23 USNM 38628.. 11 1.86 2.49 It is obvious, therefore, that Wilson took into consideration only the specimens in USNM 38625. This collection unfortunately contains no fully grown ovigerous females (judged by the length of the egg sac). One of the three immature adult females in this collection was described by Wilson as a male, and the meas- urements given for it are (Wilson, 1911 : 357) : Total length, 1.27 mm.; cephalothorax, 0.3 by 0.5 mm. ; trunk 0.75 by 0.51 mm. ; and width of genital segment, 0.25 mm. Thest; figures lie within the range of the im- mature adult female with a longer (guitar- shaped) trunk given in table 2. The 10 specimens of T. impressus reported by Nordmann (1864) were taken from a "fleckigen Diodon-Art." According to Nord- mann's description on p. 491, these parasites measure about 5 mm. long including the egg sac. Judging from his illustration of a complete parasite in pi. VI, fig. 7, the body is about 3.15 mm. long and the egg sac, 1.85 mm. ; therefore, the size is about that of the Atlantic type of the T. impressus. According to what Nordmann described (pp. 491-494) and illustrated (pi. VI, figs. 7-10), however, this West African species of Tucca is definitely different from all three types of T. impressus in the North and South American waters. ACKNOWLEDGMENTS Two field collections and the subsequent laboratory study have been aided by a grant (GB-1809) from the National Science Founda- tion of the United States to Arthur G. Humes, who also critically reviewed the first draft of this report. Roger F. Cressey, Division of Crustacea, U.S. National Museum, Washington, D.C., loaned the USNM collections of the speci- mens of Tucca i7npressus Kroyer; W. Ver- voort, Rijksmuseum Van Naturlijke Historie, Leiden, The Netherlands, and P. Illg, Uni- versity of Washington, Friday Harbor, Wash., reviewed the manuscript, as did Kenneth Sher- CYCLOPOID COPEPODS OF GENUS TUCCA 297 man. Bureau of Commercial Fisheries Biol- ogical Laboratory, West Boothbay Harbor, Maine. Jack Rudloe, the owner of the Gulf Specimen Company, Panacea, Fla., pro- vided certain specimens and helped me dur- ing the summer of 1965 while I was collecting parasitic copepods in Apalachee Bay, Fla. The Bureau of Commercial Fisheries (Region 2), Fish and Wildlife Service, gave me the oppor- tunity to collect parasitic copepods from fishes taken by the R/V Oregon during Cruise 105, November 16 to December 2, 1965. LITERATURE CITED Bassett-Smith, p. W. 1899. A systematic description of parasitic cope- pods found on fishes, with an enumeration of the known species. Proc. Zool. Soc. London 2: 438- 507. Bere, Ruby. 1936. Parasitic copepods from Gulf of Mexico fish. Amer. Midland Natur. 17 (3): 577-625. CARVALHO, J. DE PAIVA. 1951. Notas sobre alguns copepodos parasites de peixes maritimos da costa do Estado de Sao Paulo. Bol. Inst. Paulista Ocean. 2(2): 135- 144. Causey, David. 1955a. Parasitic Copepoda from Gulf of Mexico fish. Occas. Pap. Mar. Lab., Louisiana State Univ. 9: 1-19. 1955b. The external morphology of BUas prionoti Kr0yer, a copepod parasite of the .sea robins (Prionotiis). Publ. Inst. Mar. Sci. 4(1): 5-12. Heegaard, p. 1947. Contribution to the phylogeny of the Arth- ropoda. Spolia Zool. Mus. Hauniensis (Skr. Univ. Zool. Mus. K^benhavn) 8: 1-236. Kroyer, Henrik. 1837. Om Snyltekrebsene, isear med hensyn til dem danske Fauna. Naturh. Tidsskr., ser. 2, 1: 172-208, 252-304, 476-504, 605-628. Milne-Edwards, H. 1840. Ordre des Copepodes. hi: Histoire Nat- urelle des Crustaces, comprenant I'Anatomie, la Physiologie et la Classification de ces Animaux 3: 411-529. NORDMANN, A. VON. 1864. Neuw Beitrage zur Kenntnis parasitischer Copepoden. Erste Beitrage. Bull. Soc. Nat. Moscou 37: 461-520. Pearse, a. S. 1952. Parasitic Crustacea from Alligator Harbor, Florida. Quart. J. Florida Acad. Sci. 15(4): 187-243. Sewell, R. B. Seymour. 1949. The littoral and semiparasitic Cyclopoida, the Monstrilloida and Notodelphyoida. The John Murray Exped., Sci. Rep. 9(2) : 17-199. Vervoort, W. 1962. A review of the genera and species of the Bomolochidae (Crustacea, Copepoda), including the description of some old and new species. Zool. Verhandel. 56: 1-111. Wilson, Charles Branch. 1908. North American parasitic copepods: A list of those found upon the fishes of the Pacific coast, with descriptions of new genera and spe- cies. Proc. U.S. Nat. Mus. 35: 431-481. 1911. North American parasitic copepods belong- ing to the family Ergasilidae. Proc. U.S. Nat. Mus. 39: 263-400. 1913. Crustacean parasites of West Indian fishes and land crabs, with descriptions of new genera and species. Proc. U.S. Nat. Mus. 44: 189-277. 1932. The Copepoda of the Woods Hole region, Massachusetts. Bull. U.S. Nat. Mus. 158: 1-635. Yamaguti, Satyu. 1963. Parasitic Copepoda and Branchiura of fishes. Interscience Publishers, New York, pp. 1-1104. 298 U.S. FISH AND WILDLIFE SERVICE MORPHOLOGY AND DISTRIBUTION OF LARVAL WAHOO ACANTHOCYBIUM SOLANDRI (CUVIER) IN THE CENTRAL PACIFIC OCEAN B> Walter M. Matsumoto, Fishery Biologist Bureau of Commercial Fisheries Biological Laboratory, Honolulu, Hawaii 96812 ABSTRACT Descriptions are presented of the early developmental stages of the wahoo, Acanthocybium solandri (Cuvier), ranging from 2.8 to 23.7 mm. in standard length. Develop- mental changes in body pigmentation, body form, fin formation, and ossification of bones and other hard parts were studied for 38 larvae collected in the central Pacific Ocean. Drawings of larvae at various sizes are included. Certain adult characters are discussed, such as: the number of vertebrae and the vertebral formula; the number of spines and rays in the first dorsal, second dorsal, and anal fins; and the number of dorsal and anal finlets. Larval and adult wahoo live in the open ocean as well as near land. The adults spawn throughout the tropical and subtropical waters between lat. 30° N. and 25° S. The species spawns throughout the year in the equatorial waters between lat. 14 N. and 15° S., and during the northern and southern summer in areas farther from the Equator. Although considerable knowledge has been gained in recent years about the early life his- tory of the commercially important mackerels and tunas, the larval and juvenile stages of many scombroid fishes are poorly know^n. This is particularly true of the larvae of the wahoo, AcantJiocybium solandri (Cuvier). The smal- lest wahoo previously recorded was a 23.7-mm. juvenile from the central Pacific Ocean (Stras- burg, 1964). Prior to this record, the only mention of a young wahoo in the literature was a 28-cm. juvenile caught off Japan in 1917 (Kishinouye, 1923). The wahoo, a member of the Scombridae, is usually taken in small quantities as incidental catches on the longline, and in larger quantities by surface trolling (Iversen and Yoshida, 1957). It is found in tropical and subtropical areas of the oceans. While sampling for tuna larvae in the central Pacific Ocean from 1950 to 1962, the Bureau of Commercial Fisheries Biological Laboratory, Honolulu, Hawaii, collected 38 young wahoo from 2.8 to 17.8 mm. SL (standard length) in plankton net hauls. The morphology and dis- Published October 1967 tribution of the larvae were studied to increase our knowledge of the early life history of the scombroid fishes. This paper describes the developmental changes in body pigmentation, fin formation, and ossification of various bones. It also dis- cusses certain adult characters that require definition, such as the number of vertebrae and the vertebral formula, the number of spines and rays in the dorsal and anal fins, and the number of finlets. The growth rates of various body parts are included, as well as new in- formation on distribution of the species in the central Pacific Ocean, as determined from captures of the larvae and adults. COLLECTION AND TREATMENT OF MATERIAL Plankton hauls were made with a 1-m. plank- ton net on 32 cruises of the Bureau of Com- mercial Fisheries research vessels Hiigh M. Smith and Charles H. Gilbert from May 1950 to July 1962. The types of hauls varied slightly over the years; generally, the net was hauled obliquely from a depth of 200 m. to the surface before 1956, but from 1956 to 1962, it was FISHERY BULLETIN: VOLUME 66, NO. 2 299 hauled obliquely from 140 m. to the surface. On a few cruises the net was hauled obliquely from a depth of 60 m. and horizontally near the surface. Each haul lasted 30 minutes. Of the 1,643 plankton .samples obtained, 600 were taken near the Hawaiian Islands (lat. 15 to 30° N., long. 150° to 165° W.) and most of the others in the equatorial region (lat. 10' N. to 10^ S.. long. 110 to 170 W.). The 38 wahoo larvae were found in 34 plankton samples; 2 larvae were found in two samples and 3 larvae in one (fig. 1 and table 1). All 'r.\BLK 1. — Record of Itnval and juvenile wahoo captured in plankton net hauls in the central Pacific Ocean, 1950-62 ■ • NONTH «*'»COUATORI« CURRENT CA ,;" ';♦ EOlATORIAl , COUNTER ' SOWIH EOVATORIAL -■ 1 CURKCNI ' 1 , -' -.'■:-■ ■iyE^^ „ rw «»- n-- If,- ISI- lirr- L*- 1 .. Ill- 1 Figure 1. — Locations at which wahoo larvae were taken in plankton net hauls. Each star represents a single larva. A star and number show catches of two or more larvae. Broken lines indicate a distance of 110 kilometers from land. Major currents of the central Pacific Ocean arc shown. the larvae were preserved in a 10 percent solu- tion of Formalin.' The wahoo larvae ranged from 2.8 to 17.8 mm. SL. The size i-ange was extended to 23.7 mm. by the inclusion of a juve- nile (collected in a midwater trawl) described by Stra.sburg (1964). Of the 38 larvae, 21 were cleared in a weak (1-2 percent) solution of potassium hydroxide and stained with alizarin in the technique de- scribed by Lipman (1935). Standard length and various body parts were measured, and counts of fin rays and spines, myomeres, teeth. and branchiostegal rays were made before clearing and rechecked after staining. Verte- ' Trade rianifs reftrrcd to in thi^ iiublic-utiuti do not iinijly endorse- iiicni of coinmereiai products. Sur- Dis- Limits o( Local face tance Stand- latitude Cruise I'ositlon time tem- from ard imd drtto ~ pera- ture land length Lnt. li" SO" N. Lai. LoiiQ. ° C. Km. Mm. May 17, 1950. HMS-4 ' 2I°06'N. lei-oc w. 0925 24.6 139 3.4 -May 18. ig.W- nMS-4 19°25' N. 159°50' W. 1048 24.8 185 4.1 May 24, ly.W aM.S-4 21°.53' N. 159°09' W. 0120 22.8 93 2.8; 4.2 June 17. lO.'H. IIMS-2fi 20°5r N. 16r,59' \V. 0512 25.9 222 3.1 July 1. 19.10.. H.MS 5 22°.58'.N. 173°00' W. 2312 26.1 333 7.4 July 12, 1902. CIl()-.'.8' 22°44' N. 160°47' W. 0100 25.5 139 5.8 July 22, 1951. HMS-10 22°27' N. 159° 1.5' W. 0715 25.6 37 4.8 July 30, 1951.. nMS-10 21''25' N. 156°30' W. 1343 25.5 110 4.3 .VUB. .1, 19.53.. ll.MS 21 21'>lfl' N. 157°2S' W, 0240 25,4 28 5.5 .Vug. 7, 1953.- UMS-21 2r53' N. 155021' W. 0315 24.8 157 3.8 .VUR. 10. 1900. C1IO-4S 24°1.5'N. 178°.54' E. 1800 26.7 592 8.7 \ug. 19. 1950 . HMS-0 20°15'.N. 158'>24' W. 2340 20.7 130 8.4 .Vug. 21. 19.53. IIMS-21 17°40' N. LIS'SO' \V. 0120 25.0 130 4.6 .Vug. 2!, 19,53. HMS.21 18=14' .N. 157°09' \V. 2038 25.8 148 4.3 Sept. 10, 1952. nMS-17 21°48' N. 157-18' W. 1149 25.2 37 4.4 Oct. 4, 1951... HMS-11 19°00' .\". 151019' W. 0750 25.4 389 5.7 Lai. W N.- 14° .S. Jan. 12, 1959.. HMS-50 12°42' N. 150024' \V. 2000 26.1 890 7.6: >7.9 Keb. 9, 19.58. - . llMS-43 7°35' S. 139040' W. 2110 28.5 46 7.8 Mar. 13, 19,54. CIKl-lS S°42' .S. 115039' W. 1954 ■26.4 2,057 5,8 .Mar. 10, 19.5t;. HMS-33 6°03' S. 140O16' W. 1949 26.3 315 5.2: 0.2 .Vpr. 10, 19,54. CHG-15 2009' N. 157005' W. 1932 27.7 28 '6.0 4.3 .Vpr. 23, 1958.. CHa-38 2°34' S. 144012' W. 2003 28.4 732 17.8 .\pr. 2.5, 19,58.. May 29, 1952. CHG-38 3°15' N. 147=40' W. 2003 28.6 1,047 9.0 HMS-15 0°30' N. 139044' W. 1035 27.7 1,550 0,6 June 10, 19,54. nMS-26 14°33' N. 168025' W. 1739 26.6 232 10.7 June 19. 1958. HMS-45 12O02' N. 149003' W. 2010 27.0 1,055 5.9 July 6. 1950- -. nMS-5 8'=54' N. 172O00' W. 2012 26.9 912 4.5 Vug. 30, 1950 CHO-30 9°22' .S. 137001' W. 2000 25.2 167 9.2 Oct. 0, 1957... CHG-35 ir42' N. 1 51=30' W. 0000 27.6 908 5.2 Dec. I. 1957.-- CHG-35 9°33' S. 139=51' W. 0839 27.5 56 10.2 Dec. 1, 1957-.. CHG-35 9°34' S. 13906O' W. 1406 28.6 56 3.9 Lat. 15°- SS" S. reb. 14, 1902. CHG-55 20°34' S. 175=29' E. 2004 27.0 145 13.2 .Mar. 5. 1957.. nMS-38 17°56'S. 140=28' W. 2030 28.1 28 2.8 Mar. 13, 1962. CHG-55 IS'OS'S. 170=48' VV. 2000 27.9 56 C.8 1 Bureau of Commercial Fislicrics research vessels, Ilugti M. Smith (IIMS) and Charlta H. Oilhrrt (CHG). -Length nr7.9.nnn, and fi.O.mm. specimens estimated, brae were counted after the specimens had been stained. In a few instances where the body was slightly bent, a small piece of glass slide was placed over the specimen to straighten it before measuring. The following measure- m.ents were made on each specimen : Standard length : The distance from anter- iormost tip of snout to posterior end of noto- chord ; after the notochord had flexed dorsad, the distance from tip of snout to posterior edge of hypural complex was measured. Head length : The distance from anterior- most tip of snout to dorsal end of gill cover. Snout length : The distance from anter- iormost tip of premaxillary to anterior edge of orbit. Orbit diameter : The greatest distance meas- ured along the longitudinal axis of the body. 300 U.S. FISH AND WILDLIFE SERVICE This measurement was chosen over eye dia- meter because the eyes tend to shrink in preservative, and some had been lost. Premaxillary length: The distance from tip to anterior edge of the mesethmoid, measured along the dorsal profile of the snout. Upper and lower jaw lengths: The distance from anteriormost tip to posterior edge of maxillary and mandible. Body depth : Vertical distance immediately behind the anus. Snout to anus distance: The straight-line distance from anteriormost tip of snout to posterior edge of anal opening. Snout to first and second dorsal fins: the straight-line distance from anteriormost tip of snout to origins of the fins. DEFINITION OF CERTAIN ADULT CHARACTERS In the identification of larval and juvenile fishes, it is important that the adult characters be defined accurately because they are often ap- plicable to the young as well. One useful skele- tal character for the diagnosis of adult scrom- brids is the number of vertebrae. Previous literature on the wahoo contains a wide varia- tion in vertebral number, which is unusual in the family Scombridae. The number of verte- brae among most scombrids is known to be nearly constant (Ford, 1937; Godsil and Byers, 1944; Clothier, 1950), but a number of