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